PROCEEDINGS
The Tenth Annual
Waste
Testing
& Quality
Assurance
Symposium
July 11-15, 1994
Hyatt Regency
Crystal City
Arlington, VA
-------
TABLE OF CONTENTS
QUALITY ASSURANCE
Paper Page
Number Number
1. Florida's Laboratory Certification Program: 1994 and Beyond 1
C. C. Kircher, S.E. Garlington, P.J. Legge, C.M. Oakley
2. The Significance of Sample Preparation in the Analytical Process 14
G.L. Robertson, S. Kumar
3. Idaho National Engineering Laboratory Analytical Services Performance Evaluation Plan 15
J.M. Connolly, S.J. Sailer, R.P. Wells, D.A. Anderson
4. Quality by Design—It Can Be Achieved! A Review of Two Projects with Different Levels of Project Planning 30
R. Thielke, K. Luka
5. Ensuring Comparability of Data Generated by Multiple Analytical Laboratories for
Environmental Decision Making at the Fernald Environmental Management Project 45
C. Sutton, B. Campbell, R. Danahy, T. Dugan, F.K. Tomlinson
6. Utilizing Statistics in Sampling to Meet Data Quality Objectives at Sand Creek Superfund Site 46
E. Acheson, R. Edmonds
7. Development of a Laboratory Methods Comparison Program: A Total Quality
Management Approach to Resolving a Laboratory Methods Comparison Question 67
T.Y. Hosaka, O.P. Bredt, K.J. Kuhl-Klinger, K.N. Pool, A.D. Rice, C. Stacey, L.H. Taylor, W.I. Winters
8. Assessing Data Quality for a Federal Environmental Restoration Project:
Rationalizing the Requirements of Multiple Clients 68
V.R. Kiszka, T.M. Carlsen
9. Exploratory Data Analysis Using EPA's Contract Laboratory Program Data 67
P. Novick, M. Kanaan, M.V. Chacko, R. Litow, D. Eng
10. Cost & Quality Effectiveness of Objective-Based and Statistically Based Quality Control
for Volatile Compounds Analyses of Gases 80
J.T. Bennett, C.A. Crowder, S.J. Sailer, M.J. Connolly
11. A Case Study Comparison of a HTW Data Evaluation Procedure Using EPA SW-846 and Army Corps
of Engineers Engineering Protocol and the EPA CLP Functional Guidelines Method of Data Evaluation 95
A.M. Ilias, G.J. Medina
12. Contract Management Strategies for Overseeing Laboratory Analyses—Special Analytical Services 108
S.J. Kolb
13. Creation of a Site-Specific Soil Laboratory Control Sample for the Idaho
National Engineering Laboratory 109
S.J. Sailer, J.M. Connolly, T.R. Meachum, R.B. Chessmore
14. Audit Standards for Field Sampling and Field Measurement 110
M. Johnson, B. Newberry, D. Bottrell
15. Quality Assurance for the EPA Region V ESAT Field Analytical Support Program Ill
L. Kranz, D. Miller, J. Thakkar
16. The Quantifications of Data Quality with Explicit Examples from Organic,
Inorganic, and Radiochemical Methods 125
A.D. Sauter
17. Rapid Site Assessment Using the QTM Service:A Quality Assurance Perspective 126
M.C. Eldridge, B.E. Lane, L.E. Minnich
18. Studies of Method Detection Limits in Solid Waste Analysis 138
D.E. Kimbrough, J. Wakakuwa
19. How Region II RCRA Quality Assurance Outreach Program Has Assisted
Industry to Minimize RCRA Compliance Costs 139
L. Lazarus, P. Flax, T. Ippolitto
-------
20. An Evaluation of QC Requirements in the CLP for Furnace Atomic Absorption Analyses 153
D.C. Hillman, D.M. Boyer, L.C. Butler, J.T. Rowan
21. The Quality Control Level: An Alternative to the Detection Levels 165
D.E. Kimbrough, J. Wakakuwa
22. Detection Limits 166
F. Moghadami, K.L. Wright, R.J. Bath
SAMPLING AND FIELD
24. Effect of Transformer Oil, Petroleum Hydrocarbons, and Inorganic Salts as Interferences
in Field Screening for PCB Contamination in Soil 167
A. Gaskill
25. Field Testing of a Portable Trichloroethylene and Chloroform Fiber-Optic Chemical Sensor. 179
J. Wells, D. Gobeli
26. Improvements in Sample Recovery of XAD-2 Resin from Method 0010 Trains 180
M. Jackson, L. Johnson, J. McGaughey, D. Wagoner, J. Bursey, R. Merrill
27. Evaluation of a New Non-Immunoassay Field Test Kit for Total Petroleum Hydrocarbons in Soil 192
S. Finch, T.B. Lynn, L. Sacramone, K. Wright
28. Multi-Element, Multi-Media Analysis of Airborne Emissions at Secondary Lead Foundries 198
D.E. Kimbrough, I.M. Suffet
29. NEIC Forms II: Automating Field Operations Records , 199
K. Mathews, J.C. Worthington, D.T. Lark, C.Y. Teng, T.A. Phillips, B.K. Patlen
30. Field Verification of Metals Contamination in Sediment by Stripping Analysis 210
K.B. Olsen, J. Wang, R. Setiadji, J. Lu
31. Comparison of the Response of PCB Field Test Methods to Different PCB Aroclors 211
S. Finch
32. Guiding Field Activities by Using Rapid Cost-Effective Analyses Performed in a Fixed Laboratory 216
D.A. Loring, R.T. Gomez
33. Reducing Laboratory Costs with a Field-Portable Ion Chromatography System 226
W. Mozer
ENFORCEMENT
34. EPA Compliance Program in Waste Testing 227
F. Liem, F.L. Siegelman
35. How to Prepare for and Manage a Laboratory Inspection 235
R.L. Cypher, M.M. Uhlfelder, M.M. Robinson
36. Improper Hazardous Waste Characterizations—Financial and Compliance Implications 244
R.M. Walka, F.A. Langone, R.P. Russell, T.J. Watson, W.F. Cosulich
INORGANICS
37. A Comparison of Methods of Measurement of Cr+6 in Wastewaters, Soils, and Sediments:
UV-Visible Spectrometry and Ion Chromatography 261
S.J. Nagourney, J. Birri, L. Vu
38. Hexavalent Chromium Methodology for Soils: Results of Extraction Comparison
Research and Multi-Laboratory Holding Time Study 263
R.J. Vitale, B. James, G. Mussoline, J. Petura
39. Enzyme-Linked Immunoassy (ELISA) for the Detection of Mercury in Environmental Matrices 275
C. Schweitzer, L. Carlson, M. Riddell, D. Wylie, B. Holmquist
40. Application of New Mercury Speciation Technique to Samples from Sites Heavily
Contaminated with Mercury 293
E.L. Miller, D.E. Dobb, D. Cardenas, E. M. Heithmar, K.W. Brown
41. "Almost Digestions" or Hot-Acid Leaches with Continuous Flow Microwave Sample Preparation 294
E.E. King, D. Barclay, J.D. Ferguson, L.B. Jassie
-------
42. An Inter-Laboratory Comparison of Instruments Used for Analysis of Elements in Acid
Digestates of Solids 299
D.E. Kimbrough, J. Wakakuwa
43. An Evaluation of Interelement Correction Factors: Uses and Limitations 300
G.A. Laing, D.E. Dobb, J.T. Rowan, F.C. Garner, L.C. Butler, L.J. Ottmar
44. Improved Detection Limits with Axial Plasma 308
V.J. Luciano, R. Foster
45. Analysis of Environmental Samples Using the TJA 61E Trace ICP. 309
R.E. Wolf, O.K. White
46. Ultrasonic Nebulization and Arsenic Valence State Considerations Prior to Determination
via Inductively Coupled Plasma Mass Spectrometry 323
C.A. Brockhoff, J.T. Creed, T.D. Martin
47. Alternatives to the Use of ASTM Type II Reagent Grade Water 326
G.H. Kassakhian, S.J. Pacheco
48. Improvements to EPA Method 335.2 for Determination of Total Cyanide Achieved by
Oxidation of Interferences 334
B. Potter, M. Goldberg, A. Clayton
Analysis of As, Pb, Se and Ti in Solid Waste Using the TJA 61E Trace Analyzer ICP. 335
M.E. Tatro, L. Camp, R. Christenberry
Trace Metal Valence State Consideration When Using Ultrasonic Nebulizer and Inductively
Coupled Plasma Atomic Emission Spectrometry (ICPAES) 342
C. Brockhoff, J.T. Creed, T.D. Martin
51. Analysis of Silver & Other Elements by Toxicity Characterization Leaching Procedure 343
D.E. Kimbrough, J. Patel, J. Wakakuwa
52. Introduction to the USEPA Method 245.7, Determination of Mercury by Automated
Cold Vapor Atomic Fluorescence Spectrometry 344
B. Potter, W.H. McDaniel, J. Scifres, M.A. Wasko, W.J. Bashe, M.D. Castellanos
53. Determination of Selected Metals by Portable X-Ray Fluorescence (XRF) Spectroscopy 351
V. Taylor
54. Improving Mercury Detection Limits Using a Dedicated Flow Injection System 363
Z. Grosser, S. Mclntosh, S. Sauerhoff
55. Rapid Micro Distillation of Total Cyanide Using Ligand Displacement and Determination by
Flow Injection Analysis 367
W. Prokopy, M. Stone
56. Advances in Quality Assurance for XRF Determination of Lead 368
H.A. Vincent, J.E. Kilduff
57. The Stability of Calibration Standards for ICP/AES Analysis: A Two-Year Study 369
D.R. Huff, E.A. Huff
58. CLP-Type Analysis Using an Axial Plasma ICP-OES 384
M. Paustian, K. Barnes, Z. Grosser
59. Determination of Mercury by ICP-MS 387
D.E. Dobb, J.T. Rowan, D. Cardenas
60. Errors Eliminations and Quality Assurance Procedures for the Determination of Germanium,
Arsenic, and Selenium in Biological Samples by Inductively Coupled Plasma Mass Spectrometry 395
T. Yeh, F.H. Ko
ORGANICS
61. Immunoassay Methods: Development and Implementation Program at the US EPA 409
B. Lesnik
62. Quantitative Environmental Immunoassay: The Next Step? 419
P. Marsden, F. Tsang
63. Toxaphene in Soil/Trichloroethylene/Rapid Dioxin Screening by Immunoassay 433
T.S. Fan, B. Young, D. Grouse, H. Allen, R.O. Harrison, H. Shirkhan, R.E. Carlson
64. TNT and RDX by Immunoassay 434
G.B. Teaney, J.M. Melby, J.W. Stave, R.T. Hudak
-------
65. Commercial Laboratory Position on the Use of SW-486 435
J.L. Parr
66. Surrogate-Based Corrections of Data for Method 5032 436
M. Hiatt, C. Fair
67. Static Headspace Analysis of Soil Samples 437
J. Purcell, A. Tippler, G. McClure, J. Ryan
68. Solid-Phase Extraction of TCLP Leachates 438
C. Markell
69. Supercritical Fluid Extraction of Organochlorine Pesticide Residues from Soils 439
D. Bennett, B. Lesnik, S.M. Lee
70. Supercritical Fluid Extraction (SFE) of Organochlorine and Organophosphate Pesticides from Soils 440
JX.Snyder
71. SFE in a Production Environmental Lab 455
M.L. Bruce, M. Miille
72. Pre-Extraction Holding Times for Nitroaromatics and Nitramines in Soils 462
T.F. Jenkins, C.L. Grant, K.F. Myers, E.F. McCormick
73. Widespread and Systematic Errors in the Analysis for PCBs in Soils 474
D.E. Kimbrough, R. Chin, J. Wakakuwa
74. An Evaluation of Gas Chromatography/Ion Trap Mass Spectrometry for Analysis of
Environmental Organochlorine Pesticides 475
G. Robertson, J. Barren
75. Rapid & Cost-Effective Analysis of 2,3,7,8-TCDD Using the "Dioxin RISc" Immunoassy Test Kit 476
R. Allen, T. Stewart, D. Reynolds, S. Friedman
76. Automated Soxhlet Extraction and Concentration of Semi-Volatile Analytes in Soil Samples 477
E. E. Conrad, K.P. Kelly
77. Evaluation of a Quantitative Immunoassay Field Screening Method for the Determination of
Pentachlorophenal in Soil & Water 480
M. Hayes, S. Jourdan, D. Herzog
78. Screening for Low-Level Chlorinated Pesticides Using Solid-Phase Extraction 494
L. Nolan
79. Contaminant Degradation Study at the Hanford Site 1100-EM-I Operable Unit 505
K.M. Angelos, R.A. Bechtold
80. The Use of Automated Supercritical Fluid Extraction/GC-MS for the Quantitative
Determination of PAHs in Soil 518
J.M. Levy, L.A. Dolata, R.M. Ravey, A. Cardamone
81. Round-Robin Study of Performance Evaluation Materials for the Analysis of Volatile
Organic Compounds in Soil: Preliminary Assessment 522
A.D. Hewitt, C.L. Grant, T.F. Jenkins, M. Stutz
82. Environmental Analysis Using HPLC with On-Line Photodiode Array Mass Spectrometer Detection 531
J.P. Romano, M.P. Balogh, R.C. Cotter, E. Bouvier, S. Oehrle
83. Microwave-Assisted Extraction of Organic Compounds from Soils & Sediments 532
W.F. Beckert, V. Lopez-Avila, R. Young, R. Kim, J. Benedicto, P. Ho
84. Detection of Toxaphene in Soil by Immunoassay 534
T.S. Fan, B. Young, H. Allen, D. Grouse
85. Detection of DDT and its Metabolites in Soil by Enzyme Immunoassay , 540
K.A. Larkin, J. Matt, B. Ferguson, H.LBeasley, J.H. Skerritt
86. Determination of Polynuclear Aromatic Hydrocarbons (PAHs) in Soil by a Magnetic
Particle-Based Enzyme Immunoassay 55 \
P.M. Rubio, T.S. Lawurk, C.E. Lachman, D.P. Herzog, J.R. Fleeker
87. Laboratory Evaluation of Immunoassay PCB Tests 555
M.L. Bruce, R.P. Swenson, Jr., K.R. Carter, R.M. Risden
88. Determination of TCE by Immunoassay 573
T.S. Fan, B. Young, D. Grouse, R.E. Carlson
89. Important Factors in Enhancing Supercritical Fluid Extraction Efficiencies for Environmental Applications 574
J.M. Levy, L.A. Dolata, R.M Ravey, V. Danielson, A. Caradamone
-------
90. Applications & Performance of the D-Tech TNT Environmental Testing System 582
G.B. Teaney, J.M. Melby, J.W. Stave
91. Screening Waste by FT-IR 592
D.L. Bumgarner, L.S. Bucina
92. An Immunoassay for the Detection of Benzene in Water 609
S. Friedman, T. Stewart, R. Allen
93. Characterization of Multiphase Partitioning 610
H.M. Liljestrand, Y. Shimizu, I. Lo, YD. Lee, J.K. Koo
94. Validation of an Enzyme Immunoassay-Based Field-Screening System for the Detection
of RDX in Soil and Water Samples 611
G.B. Teaney, R.T. Hudak
95. Direct Analysis by Raman Scattering: An Emerging Technology for Waste Remediation 624
C.K. Mann, T.J. Vikers
96. Integrated Sample Automation for Concentration, Extraction & Cleanup 634
M. Cava
97. Analysis of Ash Residues for Volatile Organic Constituents by Using a Modified SW-846 Method 8260 635
R.P. Grese, M.D. Edwards
98. PCB Field Experience Using the D-Tech PCB Immunoassay Field-Screening Test 647
J.M. Melby, B. Finlin, M. Knowles, T. Young
99. Identification of High Molecular Weight Biogenic n-Alkanes, n-Alkanols, H-Alkanals, and
Plant Sterols in Environmental Samples and Determination of Lee Retention Indices by GC/MS
Equipped with Electronic Pressure Control 648
P.H. Chen, W.A. Van Ausdale, A. Tomaras, D.F. Roberts
100. Rapid Dioxin Screening by Enzyme Immunoassay 659
R.O. Harrison, H. Shirkhan, R.E. Carlson, T. Keimig, W.E. Turner
101. Extraction of Environmental Analytes Using a Carbon Membrane 666
S. Miller, G.L. Nixon
102. A Preliminary Investigation of Retention Times and Analyte Recoveries when Using
High-Efficiency Gel Permeation Chromatography Cleanup 667
E.E. Conrad, N.L. Schwartz, K.P. Kelly
103. Determination of PAHs by Enzyme Immunoassay 669
R.O. Harrison, R.E. Carlson
104. Immunoassay Detection of Polycyclic Aromatic Hydrocarbons Simplifies Field Analysis
of Soil and Water 677
M. Mullenix, J. Stave, R.T. Hudak
105. Extraction Conditions in Supercritical Carbon Dioxide Resulting in Partial Breakdown of
DDT from Contaminated Soils 687
A.L. Wilson, B.M. Austern
AIR AND GROUND WATER
106. ABS, FEP, FRE, and FRP Materials: Ability to Withstand Attack by Organic Solvents
and Sorption of Trace-Level Organics 694
L. Parker, T.A. Ranney
107. Sampling and Analytical Methods for House Dust and Dermal Exposure 709
J.P. Hsu, D. Camann, P. Geno
108. CO2 Management for TO-14 Analyses Using a Controlled Desorption Trap (CDT) 710
R.A. Jesser, S. Reiss
109. Status and Need for Fuel Toxicity Characteristics Leaching Procedure (TCLP) 711
A.M. Ilias, G.J. Medina
110. Analysis of Air Canister Samples for Polar and Volatile Compounds Using Modified TO-14 719
P.E. Kester, E.T. Lewis, A.T. Madden, V. Naughton
-------
RADIATION
111. Radiochemical Methods and Detection Limits 720
E. Baratta
112. Progress Toward Maturity of DOE Methods for Evaluating Environmental and Waste
Management Samples 732
S.C. Goheen, M. McCulloch, R.G. Riley, S.K. Fadeff, B.L. Thomas,
G.M. Mong, G.K. Ruebsamen, W.C. Cosby, D.S. Sklarew
113. QC Data Review for Gamma-Ray Liquid Scintillation Analyses 740
R. Litman
114. National Institute of Standards and Technology's Measurement Quality Assurance Programs
for Ionizing Radiation 755
K.G.W. Inn, J. Humphreys, J. Shobe
115. Development of a Mixed Analyte Performance Evaluation Program for the Environmental
Restoration and Waste Management Office of the U.S. Department of Energy 756
S. Morton
116. Implementation of an Fourier Transform Infrared Spectrophotometer for the Determination of
VOCs in Waste Drum Headspace 758
W. Bauer, M.J. Connolly, D. Gravel, A. Rilling, F. Baudai
117. Hanford Environmental Restoration Data Validation Process 773
R.A. Bechtold, M.R. Adams, K.M. Angelos
GENERAL
118. Corrositex—New Solution for Determining Corrosivity of Products and Waste 787
V.C. Gordon, R. Wei
119. Reactivity Characterization by Differential Scanning Calorimetry 802
M.L. Siao, J.H. Lowry
120. Overview of NIST SRM Activities for Inorganic Environmental Studies 808
J.S. Kane
-------
QUALITY ASSURANCE
-------
FLORIDA'S LABORATORY CERTIFICATION PROGRAM: 1994 AND BEYOND
S. E. Garlington, Laboratory Consultant, P. J. Legge,
Laboratory Consultant, C. M. Oakley, Laboratory Consultant,
and C. C. Kircherr Program Administrator, FL Department of
Health and Rehabilitative Services, Office of Laboratory
Services, Jacksonville, Florida 32231
ABSTRACT
The Florida Department of Health and Rehabilitative Services
(HRS) operates two laboratory certification programs.
Currently, 300 Safe Drinking Water testing laboratories and
500 Environmental testing laboratories are certified in each
program. Because these programs are fairly large and
involve laboratories nation-wide that support Florida
projects, the following issues regarding certification are
routinely addressed:
- Reciprocal certification with other states and the
National Environmental Laboratory Accreditation Program.
- Use of approved, standardized, performance-based, or
alternate analytical testing methods.
Expanding certification for newly regulated
contaminants, such as asbestos, dioxin, and disinfectant by-
products.
- Formats and documentation of quality systems, both
for laboratories and for accreditation agencies.
- Representative on-site laboratory surveys and scope
of accreditation among various states and 3rd party
agencies.
Many of these issues are common to other state
laboratory certification programs. The purpose of this
paper is to present Florida's approach to these issues and
to stimulate discussions among the regulatory agencies,
testing laboratories, water supply systems, and the public.
INTRODUCTION
The State of Florida offers two certification programs to
accredit laboratories. One program is for Safe Drinking
Water (SOW) testing, and the other is for Environmental
Water (ENV) testing. The Department of Health and
Rehabilitative Services (HRS) Office of Laboratory Services
-------
has assumed primacy for the enforcement of the Safe Drinking
Water Act. The SDW program thus serves to fulfill the
requirement that drinking water testing be performed by
certified laboratories. Approximately 300 laboratories are
certified or pending certification in this program.
Statutory authority for the ENV program is derived from the
Florida Air and Water Pollution Control Act. Laboratory
certification for this program is required only for
analyzing domestic effluents and wastewater during treatment
and discharge. Due to laboratory demand, certification is
also offered for hazardous waste, soils and sediments,
groundwater, and other environmental systems. Defining
"environmental water samples" as essentially any source that
can impact Florida water quality allows our program to offer
this expanded certification. Thus, although the ENV
certification is largely voluntary, many laboratories choose
to become certified because this accreditation enhances
their marketability, public trust, and client relations.
Approximately 500 laboratories are certified or pending
certification in this program.
The legal framework for Florida's certification programs is
encompassed in chapter 10D-41 Florida Administrative Code
(FAC). Sections 10D-41.050 through 10D-41.062 FAC pertain
to the SDW program, and sections 10D-41.100 through 10D-
41.113 FAC pertain to the ENV program. HRS shares
responsibilities with the Florida Department of
Environmental Protection (DEP) to establish the criteria for
laboratory accreditation.
Florida's laboratory certification program is one of the
larger programs in the nation. Support for the program is
based on the extreme importance Florida places on
maintaining the quality of its surface waters and
groundwater. Based on regular collaborations with other
Florida officials and other state certification agencies,
our program routine addresses issues that relate to
Reciprocal Certification with other states, National
Environmental Laboratory Accreditation, and Performance-
based, newly approved, or Alternate Analytical Methods. The
purpose of this paper is to present our approach to date on
these issues, to describe the implementation and operation
of our program, and to stimulate discussions so that the
direction of this program will meet the needs of Florida's
citizens, other accreditation agencies, U.S. EPA, and the
public.
-------
CERTIFICATION CATEGORIES AND REQUIREMENTS
Certification for Safe Drinking Water Testing Laboratories
is offered in the following categories:
PRIMARY INORGANIC CONTAMINANTS
Metals Lead and Copper
Nitrate and Nitrite Fluoride
Cyanide Asbestos
SECONDARY INORGANIC CONTAMINANTS
PESTICIDES AND PCB's
Insecticides Herbicides
Carbamates Disinfectant BP's/VOC's
Miscellaneous SOC's Adipates and Phthalates
PCB's PAH's
OTHER REGULATED CONTAMINANTS
VOC'S THM'S
GROUP I UNREGULATED CONTAMINANTS
Herbicides Carbamates
GROUP II UNREGULATED CONTAMINANTS
Purgeables Acid Extractables
Base/Neutral Extractables
DIOXIN
MICROBIOLOGY
RADIOCHEMISTRY
These categories and subcategories are generally organized
according to SOW regulatory program, monitoring
requirements, and analytical technique.
Certification for Environmental Water Testing Laboratories
is offered in the following categories:
METALS
NUTRIENTS
DEMANDS
EXTRACTABLE ORGANICS (GC, GC/MS, HPLC)
GENERAL CATEGORY I
GENERAL CATEGORY II
MICROBIOLOGY
PESTICIDES-HERBICIDES-PCB'S (GC, GC/MS, HPLC)
PURGEABLE ORGANICS (GC, GC/MS)
BIOASSAY
HAZARDOUS WASTE CHARACTERIZATION
RADIOCHEMISTRY
BASIC ENVIRONMENTAL LABORATORY
-------
Some of the analytes offered in the Metals and Nutrients
categories are also offered in General Category I and
General Category II. The intent for inclusion in the
general categories is for smaller laboratories to analyze
these contaminants with colorimetric/titrimetric methods or
with test kits where ultratrace sensitivities are not
required. Also, to save smaller, in-house water utility and
wastewater treatment laboratories the expense of being
certified in 4-5 categories, selected contaminants from
Nutrients, Demands, Microbiology, General Category I, and
General Category II categories are combined in the Basic
Environmental category. However, laboratories certified in
this category may also be certified in only one other
additional category of choice. Otherwise, the laboratory is
not considered "basic" and must select analytes from the
other categories in order to become certified.
As with other state certification agencies, Florida's
laboratory certification requirements include completion of
an application form, satisfactory analysis of proficiency
test samples, formulation of a quality assurance plan,
demonstrating compliance with certification requirements
during an on-site laboratory inspection, and payment of
certification fees. The applications are reviewed for
completeness (including the Personnel and Quality Assurance
(QA) sections), for original signatures from the Laboratory
Director and QA Officer attesting to compliance with
certification rules and regulations, and for appropriate
analytical methods with each pending analyte. All regulated
SOW contaminants must be analyzed with EPA-approved methods.
Analytical methods selected for other pending analytes must
include the analytes in their titles or their compound
lists, or else be referenced by EPA in the Code of Federal
Regulations (CFR).
Pending SOW laboratories are enrolled in EPA's Water Supply
(WS) proficiency testing (PT) program, and pending ENV
laboratories are enrolled in the EPA Water Pollution (WP)
program. To be eligible for certification for a particular
analyte, the laboratory must produce acceptable results at
all available concentration levels for that analyte during
the latest WS or WP testing round. If a pending analyte is
not available during a round, the laboratory is still
eligible for certification for that analyte if the
laboratory passes at least 75% of all available analytes
pending certification that are in the same category as the
unavailable analyte. In general, as a Standard Operating
Procedure (SOP), our program will classify a pending
laboratory as ready for the on-site inspection if the
laboratory passes at least 75% of all available pending
analytes at all available concentration levels in any one
-------
category during the latest PT testing round. To maintain
certification, a laboratory must produce acceptable results
at least once per year for each available certified analyte
at all available concentration levels. With prior
permission from HRS, a laboratory may analyze commercially
available PT samples that are determined to be equivalent to
the corresponding EPA program.
For SOW certification, the laboratory must have a QA plan
available for review during the on-site inspection, if
requested. For ENV certification, the laboratory QA plan
must be reviewed and approved prior to the on-site
inspection. If the pending ENV laboratory is analyzing only
domestic effluents and wastewater, HRS is authorized to
review and approve the QA plan. If the laboratory is
analyzing other types of samples, such as hazardous waste,
then the QA plan must be reviewed and approved by the
Florida DEP Quality Assurance Section.
Laboratories located in Florida are generally inspected on-
site once per year; out-of-state laboratories are surveyed
once every two years. The duration of the on-site survey
ranges from a half-day for SDW Microbiology and Basic
Environmental laboratories to two days for full-service SDW
and ENV laboratories. In general, the laboratory surveys
are conducted to verify that the laboratory has performed
the Initial Demonstration of Capability, instrumental
calibration, and on-going QA requirements in the analytical
methods and to verify fulfillment of sample preservation,
holding time, and data reduction/reporting requirements for
the regulatory compliance programs. Our certification
program uses in-house checklists as guidelines to assist in
the inspection process. Laboratory consultants are trained
by EPA as certification officers in both Chemistry and
Microbiology; consequently, one inspector generally conducts
the entire survey of the laboratory. Deficiencies observed
during the survey are noted on a standard report form, and
the laboratory must respond in writing to each deficiency
with a Plan of Correction and completion date.
When the above requirements are met, the laboratory is
invoiced for certification fees. For each program, SDW and
ENV, these fees are $400 per category up to a maximum of
$1500 for 4 categories or more. If the laboratory meets the
requirements in the middle of the Florida fiscal year, the
certification fees are prorated on a quarterly basis. Out-
of-State laboratories are also invoiced for the inspector's
travel expenses to conduct the on-site survey. Because
Radiochemistry certification for both programs is handled
through the HRS Office of Radiation Control rather than the
Office of Laboratory Services, certification fees for
-------
Radiochemistry are $400 per year in addition to any other
certification fees assessed for Chemistry and Microbiology-
Laboratory certification is valid for the current fiscal
year between July 1 and June 30, and the laboratory is
assessed the same certification fees for renewal for the
next fiscal year.
When the certification fees are paid and the Plan of
Correction is accepted, HRS assigns the laboratory a 5-digit
certification number based upon the laboratory's geographic
location, type of laboratory (commercial, public utility,
research institution, etc.)/ and numerical sequence relative
to previously certified laboratories. A certificate showing
all categories certified and an Analyte Sheet showing all
analytes and test methods certified are mailed to the
laboratory.
PROGRAM ORGANIZATION
The laboratory certification program consists of fifteen
full-time employees: four clerical staff, four in-house
professionals, six laboratory consultants, and one program
administrator. The clerical staff of secretaries and data
entry personnel are responsible for typing correspondence,
keying in laboratory and PT data into databases and
spreadsheets, maintaining office equipment and supplies, and
filing data into various hardcopy files. The in-house
professionals are chemists and microbiologists who review
certification applications, coordinate PT samples and
results, update computer databases and hardcopy files, and
provide administrative consultation to laboratories and the
public. The laboratory consultants are chemists and
microbiologists who are responsible for conducting the on-
site laboratory surveys, reviewing QA plans, and providing
technical consultation to laboratories and the public. The
program administrator is a senior executive who is
responsible to the Chief of Laboratory Services and who
directs the policy decisions and procedural implementations
of the certification program, resolves disputes and
complaints from laboratories and the public, delegates
defined activities for the program to individuals or
committees, and promotes a working environment where
accreditation decisions are free from conflict-of-interest
or other influences.
Information on pending and certified laboratories is stored
in hardcopy files and electronic computer files. For the
hardcopy files, data for each certified laboratory is
organized in color-coded sections that correspond to
certificate, invoices, and certification-decertification
correspondence; certification application; analyte sheet; PT
-------
results; on-site survey reports and Plans of Correction; QA
plan; and miscellaneous correspondence. Information on
certified laboratories that is over two years old is stored
in purge files. Information on applicant laboratories is
stored in pending file folders. Generally, separate pending
files are kept for laboratories pending certification and
for certified laboratories that have applied for additional
certification. Information on laboratories that are no
longer involved in the certification program are stored in
inactive files.
Laboratory information is also stored electronically in
various computer databases and spreadsheets. The
certification program operates a dedicated Novell network
(Version 1.22) in which a fileserver is linked to eight
personal computers, four printers, and one modem. Access to
the network and to databases is controlled through logon
ID'S and passwords. Most laboratory information is
organized in the databases through dBase-IV, Version 1.1.
This particular software allows searching for specific
records, sorting and organizing laboratories according to
specific criteria, and varying the output formats for
reports and mailing labels as needed. Analyte Sheets, PT
results, and certification status based on PT results are
stored in spreadsheets created through Lotus 123, Version
2.2. Standard forms and templates have been prepared with
Microsoft Word, Version 5.0.
Florida's laboratory certification program is committed to
operating this program in accordance with generally accepted
international standards of quality. Because many
laboratories use this program as an integral part of their
individual quality systems, the certification program
documents its quality system in various Standard Operating
Procedures (SOP's) and in Appendices that contain the
various forms, templates, checklists, and other materials
used by the program. This quality system is intended to
comply with the guidelines established in ISO Guides 25, 54,
and 55d-3) and in NISTIR 4576W.
RECIPROCAL CERTIFICATION FOR OUT-OF-STATE LABORATORIES
Florida's certification program is one of the larger
programs in the nation and currently certifies laboratories
in 34 states and Puerto Rico. Additional laboratories in
Canada and Mexico are pending Florida certification.
Consequently, this program deals with reciprocal
certification with out-of-state laboratories on a regular
basis. Some states are known to certify their out-of-state
laboratories based upon the laboratories' Florida
certifications. However, Florida's current legal system
-------
does not allow for third-party reciprocal certification, and
a laboratory must be certified by its home state for the
same analytes and test methods to be eligible for reciprocal
certification in Florida.
In general, Florida's program handles a laboratory's request
for reciprocal certification on a case-by-case basis. The
applicant laboratory must submit copies of its state's
certificate, analyte sheet, PT results (if applicable), QA
plan (if applicable), survey report from the most recent on-
site laboratory inspection, and certification rules and
regulations. This information is reviewed to verify
compliance with all sections of 10D-41 FAC and to ascertain
whether Florida's laboratory consultants would observe the
same condition of the laboratory as noted in the home
state's survey report if Florida's consultants had conducted
the inspection. If these requirements are met, the
requirements of the on-site laboratory inspection and the
laboratory paying for the inspector's travel expenses are
waived, and the laboratory is recommended for Florida
certification by reciprocity.
Reciprocal certification occurs most frequently with out-of-
state laboratories that are seeking Florida certification
for a specific analytical function, such as Bioassay,
Asbestos analysis, or Dioxin analysis. In these cases, it
is comparatively easy to confirm the home state's
certification and to verify the representativeness of the
inspection criteria used. In other cases, Florida HRS can
grant certification by reciprocity for some of the analytes
and test methods, but the on-site survey by Florida's
consultants would be required for the other pending analytes
in the laboratory's application.
Florida's program does accommodate another state's
certification by compound class or analytical technique. As
an example, an out-of-state laboratory certified in its home
state to analyze drinking water by Purge-and-Trap GC/MS
would be reciprocally certified by Florida for all Volatile
Organic Contaminants, Trihalomethanes, and Purgeable
Organics that Florida regulates, by EPA Method 524.2.
Florida's certification program is currently prepared to
enter into a Memorandum of Understanding (MOU) with other
state programs that require a formal reciprocity agreement.
This MOU confirms that the laboratory must meet each state's
certification requirements and specifies the differences
between the two states' requirements. For example, Florida
would require formal approval of the out-of-state
laboratory's QA plan, but the Florida laboratory would have
to fulfill the other state's existing PT or application
-------
requirements (which Florida agrees is equivalent to or more
stringent than its own requirements).
NATIONAL ENVIRONMENTAL LABORATORY ACCREDITATION
From the State of Florida, Dr. E. C. Hartwig, Chief of HRS
Laboratory Services, and S. Labie, head of DEP's Quality
Assurance Section, currently serve on various task forces
and working groups in EPA's Committee for the National
Accreditation of Environmental Laboratories (CNAEL). Both of
these individuals have close associations with the
certification program; consequently, there has been
considerable information exchange on CNAEL's activities and
providing input for task forces and working groups.
Despite the current situation with different states
implementing their own certification programs, there is
considerable precedent and framework already in place for a
national accreditation program. 49 out of 50 states have
assumed primacy for the enforcement of the Safe Drinking
Water Act and thus have laboratory certification programs of
various sizes. In addition, 43 of these programs utilize
EPA's WS and/or WP PT samples as integral parts of their
certification requirements. Furthermore, all other states'
rules and regulations reviewed thus far contain various
forms of application, proficiency testing, quality
assurance, on-site laboratory inspection, and fee payment
requirements for certification.
From a laboratory certification perspective, two of the
biggest barriers to national laboratory accreditation are
the scope of the certification that will be offered and the
representativeness of the on-site laboratory survey among
the various organizations that will operate the
accreditation program. The scopes of accreditation among
the states vary widely. States certify laboratories by
categories, chemical compound class, analytical technique,
individual analytes, test methods, and/or sample matrix.
One way to resolve these differences is to adopt a
hierarchical structure for national laboratory accreditation
such as the one proposed by the CNAEL.(5)
Because the on-site laboratory survey is important in
determining a laboratory's certification status, ensuring
that the inspection criteria are completely defined and
uniformly applied is critical to the efficacy of a national
program. Florida HRS has requested and obtained inspection
checklists from other states' programs and has compared
these checklists with our own. These checklists have been
helpful for improving our own checklists and for deciding
-------
whether an out-of-state laboratory should obtain reciprocal
certification.
In addition, Florida's certification program has reviewed an
EPA draft document "Data Audits for Drinking Water
Laboratories. "(6) This document was authored by EPA
personnel and is being proposed for inclusion as a chapter
or appendix in the EPA "Manual for the Certification of
Laboratories Analyzing Drinking Water. "O7) Although many of
the guidelines need to be tailored to specific analytical
methods, the "Data Audit" document has excellent quality in
content, organization, and presentation. Its details are
pertinent to a laboratory performance audit as well as a
conventional data audit. Incorporation of a document such
as this one in the National Environmental Laboratory
Accreditation Program and the implementation of its
guidelines by all accrediting organizations will go a long
way in reinforcing the viability of the National Program,
addressing the concerns of laboratories regarding uniform
treatment, and serving the needs of the public in health and
safety.
Thus, from a laboratory certification perspective, the
authors see the national accreditation program as having
three different components, oversight, implementation, and
operation. In general, CNAEL's conclusion of the national
program having federal oversight and implementation by the
various state agencies is consistent with this view.
Specific roles and responsibilities will be assigned to each
component, and particular interactions among the components
will be defined. The infrastructure of the state
accrediting agencies that is already in place would be
preserved, and state agencies would have the option of
operating the national program on behalf of their respective
states, or of delegating these operations to a third party
accrediting agency.
EXPANDING CERTIFICATION PROGRAMS
Florida's certification program attempts to be as proactive
as possible to changing monitoring requirements in the SOW
Act and to updates in CFR. HRS has the fudiciary
responsibility and statutory authority to amend its
certification rules whenever EPA or Florida DEP promulgate
their rule amendments. Certification personnel gain
technical competence in new and revised test methods through
in-house training in the HRS Central Laboratory (the SOW
Primacy laboratory for Florida) and through participation in
workshops sponsored by EPA's Environmental Monitoring and
Systems Laboratory in Cincinnati and by EPA Region IV. When
these options are not available, we can employ outside
10
-------
consultants to assist in the necessary training for the on-
site inspection process. For example, J. Webber from the
New York Department of Health instructed our laboratory
consultants in the on-site surveys of the first Florida-
based laboratories to be certified for Asbestos in drinking
water with the Transmission Electron Microscope method. We
are attempting to develop expertise in the analyses of
Giardia, Cryptosporidia, viruses, disinfectant by-products,
and other contaminants should our certification program
become involved in any laboratory approval process
associated with the Information Collection Rule, Phase VIb,
or other SOW monitoring changes.
ALTERATIVE AND PERFORMANCE-BASED ANALYTICAL METHODS
Because Florida laboratory certification is based on test
method as well as analyte, our program relies greatly on the
references of approved methods in CFR and the references of
analytes in each method. In current practice, the use of
alternative test methods is allowed in two cases. The most
common situation occurs when laboratories exercise their
options of modifying approved analytical methods to
consolidate analysis procedures and improve performance.
For example, a laboratory can use a mass spectrometer as
allowed in Sections 6.8.3 and 10.4 of EPA Method 507 as long
as the laboratory achieves the Initial Demonstration of
Capability precision and accuracy objectives in Section 10.3
and obtains method detection limits that are reliably and
consistently below the regulated Maximum Contaminant Level
for each analyte. But the laboratory's Analyte Sheet will
show certification for these analytes by EPA Method 507.
Alternatively, the laboratory may elect to submit its
analytical method to the Alternative Test Method approval
process, or to obtain a variance from EPA Region IV allowing
the use of its method. When the Alternative Test Method is
published in Federal Register or a letter of variance is
received from EPA officials, we can offer certification for
that method. Examples of variances received to date include
analyzing wastewater samples for anions by Ion
Chromatography and for metals by ICP/MS.
Our program recently reviewed an EPA draft document
"Guidance on the Evaluation of Safe Drinking Water Act
Compliance Monitoring Results from Performance-Based
Methods."® The possible use of Performance-Based Methods
(PBM's) represents EPA's approach to rapid changes and
improvements in scientific analytical technology, and EPA's
departure from the "command and control" style of
environmental rule enforcement. Technically, PBM's are most
useful in analyses of environmental samples that are
11
-------
impossible to complete with approved methods. Drinking
water samples are less susceptible to matrix interferences,
and the current approved SOW methods do allow for
considerable flexibility already- Nevertheless, the
promulgation of PBM's may result in lower laboratory
operating costs and reward the laboratories economically for
developing new methods to meet data quality objectives. The
chief disadvantage to PBM's is that reliable conclusions may
not be reached from data among different water supplies and
geographical regions, especially since different PBM's may
not be universally applicable to all sampling conditions,
site locations, sets of analytes, concentration ranges, or
time of year. PBM's will be a challenge to implement in a
certification program, and the proposed national
accreditation program may be affected if primacy states
choose to set their own data quality objectives.
CONCLUSIONS
Florida HRS is committed to operating its laboratory
certification program to meet the needs of Florida citizens,
the laboratories, US EPA, and the public. The primary
purpose of the program is to ensure that laboratories
perform analyses according to prescribed methods, produce
data that meet defined standards of quality and
defensibility, and comply with federal and Florida
regulatory programs. Nevertheless, the philosophy of all
personnel in the program is to provide cooperation and
assistance to the laboratories in meeting these requirements
and, when requested, provide technical assistance to the
laboratory to improve analytical performance. Because the
laboratory pays fees for its accreditation, Florida
certification personnel want the product, the Certificate
that the laboratory displays, to communicate the
laboratory's trust and integrity to the public and quality
and commitment to its clients. By pursuing these
objectives, Florida's certification program is poised to
offer laboratory accreditation that will meet present and
future needs, benefit local communities, and provide for
testing laboratory acceptance in international markets.
REFERENCES
(1) Guide 25, "General Requirements for the Competence of
Calibration and Testing Laboratories," 3rd Edition,
International Organization for Standardization, Switzerland,
1990.
(2) Guide 54, "Testing Laboratory Accreditation Systems -
General Recommendations for the Acceptance of Accreditation
12
-------
Bodies," 1st Edition, International Organization for
Standardization, Switzerland, 1988.
(3) Guide 55, "Testing Laboratory Accreditation Systems -
General Recommendations for Operations," 1st Edition,
International Organization for Standardization, Switzerland,
1988.
(4) NISTIR 4576, "Laboratory Accreditation in the United
States," National Institute of Standards and Technology,
1991.
(5) R. Stephens, "Scope of a Proposed National Environmental
Laboratory Accreditation Program," EPA Committee for the
National Accreditation of Environmental Laboratories, DRAFT
10-21-91.
(6) "Data Audits for Drinking Water Laboratories," EPA
Draft, Oct. 1, 1993.
(7) "Manual for the Certification of Laboratories Analyzing
Drinking Water: Criteria and Procedures, Quality
Assurance," 3rd Edition; EPA/570/9-90/008, April 1990;
Change 1, EPA/570/9-90/008A, October 1991; and Change 2,
EPA-814B-92-002, September 1992.
(8) "Guidance on the Evaluation of Safe Drinking Water Act
Compliance Monitoring Results from Performance-Based
Methods," EPA Draft, January 14, 1994.
13
-------
The Significance of Sample Preparation in the
Analytical Process
Gary Robertson. U.S. EPA Environmental Monitoring
Systems Laboratory, Las Vegas, NV 89193
Sujith Kumar, Lockheed Environmental Systems &
Technologies Company, Las Vegas, NV 89119
The U.S. EPA is concerned about the quality of
analytical data used for making decisions about
environmental cleanups. A research project at the
Environmental Monitoring Systems Laboratory-Las Vegas
relates the results of quality control (QC) activities
to specific data quality parameters. Historically,
laboratories have placed emphasis on the instrumental
measurement system in environmental analysis with less
attention paid to sample extraction and cleanup.
Generally, in commercial laboratories, extraction
personnel are less experienced, less educated and lower
paid than instrument operators. This can have a severe
impact on data quality unless appropriate controls are
in place.
The object of this study was to differentiate
between sample preparation errors and instrumental
analysis errors and examine the effect of the sample
preparation errors. QC data from the Contract
Laboratory Program quarterly performance evaluation
samples were examined and related to the analytical
results obtained by program laboratories. Data from
surrogates, internal standards, calibrations, and tunes
were evaluated for laboratories with poor or failing
scores on the semivolatile and pesticide fractions.
The errors relating to the extraction portion of the
analysis will be discussed relative to their effect on
the analytical results. Case studies will be presented
which illustrate the relationship between the QC and
the analytical results, with a discussion of the errors
likely to be responsible in each case.
The types of data quality defects resulting from
sample preparation errors on performance evaluation
samples are assumed to also occur from similar sample
preparation errors on other samples. Therefore, an
increased emphasis on procedures and training in sample
preparation can make a significant improvement in the
quality of data produced by a laboratory.
14
-------
IDAHO NATIONAL ENGINEERING LABORATORY
ANALYTICAL SERVICES PERFORMANCE EVALUATION PLAN*
Joan M. Connolly. Senior Scientist, S. J. Sailer, R. P. Wells, EG&G Idaho, Inc., Idaho National
Engineering Laboratory, P.O. Box 1625, Idaho Falls, Idaho 83415; D. A. Anderson, Battelle Pacific
Northwest Laboratories, P.O. Box 999, MSIN P7-27, Richland, WA 99352
ABSTRACT
Consistent assessment of laboratory (mobile and permanent) and validator analytical support supplier
performance across regulatory programs is a concern for sample management organizations across
the Department of Energy (DOE) complex. The use of suspect analytical data for decision making
has serious legal and financial implications to the DOE and its contractors. A joint statement issued
by DOE, U.S. Environmental Protection Agency (USEPA) and the Department of Defense, and a
subsequent DOE memorandum, emphasi/e the need to develop and implement a consistent and
innovative means of preventing suspect data. Past occurrences of fraud in the generation of
environmental and waste characteri/ation data have been detrimental to the credibility of the DOE.
A recent Federal Register proposed rule indicated that DOE will not be able to rely on other agencies
for notification of suspect laboratories or validators.
The Idaho National Engineering Laboratory (INEL) Analytical Services Performance Evaluation Plan
(ASPEP) proposes a strategy to ensure that analytical service suppliers provide products of known
and appropriate quality for all DOE Environmental Management (EM) activities. The ASPEP
emphasizes supplier product quality improvement through integration of real-time with periodic
performance evaluation. Periodic evaluation of performance is accomplished using data generated
from existing performance evaluation programs (i.e., those administered by DOE, the USEPA, and
the Army Corp of Engineers) and audits. Real-time performance assessment is accomplished through
evaluation of routine laboratory quality control data, blind performance evaluation samples, INEL-
specific performance evaluation materials, deliverable^ and supplier management practices.
This paper describes the general approach and specific components outlined in the ASPEP for the
assessment of analytical laboratory and data validator performance.
INTRODUCTION
The primary responsibility of the Idaho National Engineering Laboratory (INEL) Sample
Management Office (SMO) is to ensure that data of known quality are supplied to the INEL by
analytical chemistry service organizations. Because high-quality analytical support is vital to the
success of the Department of Energy (DOE) Environmental Management (EM) programs at the
Work supported by the U.S. Department of Energy, Assistant Secretary for Environmental Restoration, Under
Idaho Operations Office Contract I>E-AC07-76II)01570.
15
-------
INEL, the performance of organizations providing these services must be routinely monitored and
assessed. This will be accomplished through the implementation of a Performance Evaluation (PE)
Program.
The INEL Analytical Services Performance Evaluation Plan (ASPEP) documents the approach of the
INEL PE program. The framework described in this document will ensure that laboratory and
validator analytical service suppliers used by the SMO meet requirements and maintain an
appropriate quality level. The ASPEP was developed to provide a proactive tool for prevention and
detection of suspect analytical support supplier products. This system provides an objective means
to measure and assess performance.
Program Scope
The INEL ASPEP is an integrated approach that assesses analytical supplier performance supporting
all EM programs (i.e., environmental monitoring, environmental restoration, and waste management)
at the INEL. The approach includes specific supplier evaluations of Comprehensive Environmental
Response, Compensation, and Liability Act (CERCLA), Safe Drinking Water Act (SDWA), Clean
Water Act (CWA), Resource Conservation and Recovery Act (RCRA) and Clean Air Act (CAA)
analytical services.
Analytical service suppliers evaluated through the INEL PE program include analytical laboratories
and data validators from both private and government sectors that have achieved SMO approval.
The analytical disciplines for which ongoing performance is monitored are 1) metals, 2) organics, 3)
radiochemistry, and 4) classical analyses.
Program Philosophy
The intent of the INEL PE program is to augment existing PE programs administered by the
Environmental Protection Agency (EPA), the Department of Defense (DOD) and DOE by expanding
into areas not addressed by these programs. Supplier performance monitoring based solely on
existing PE program results occurs too infrequently to allow prompt correction of quality problems.
There are no formal programs addressing the evaluation of data validation supplier performance.
Therefore, the ASPEP integrates periodic and real-time supplier performance evaluation with quality
improvement tools for a comprehensive approach to quality management of laboratory and data
validation analytical services.
Multiple tools are available for assessing performance. The performance assessment tools chosen for
this program fit into one of the following approaches to quality assurance:
1. Periodic Supplier Performance Evaluation: consists of evaluation of laboratory
performance data generated from participation in existing PE programs administered by
EPA or DOE, supplier audits, and evaluation of data validator performance on blind
prevalidated data packages.
2. Real-time Supplier Performance Evaluation: consists of evaluation of laboratory quality
control (QC) and blind PE sample data generated concurrently with INEL field samples,
supplier management practices, and evaluation of supplier deliverables.
16
-------
3. Quality Improvement: consist of matrix-specific QC materials provided to suppliers for
use as known laboratory controls and feedback to suppliers on performance to encourage
process improvements.
Data generated by participation in existing PE programs will be used as a tool to determine the
general ability of a supplier to perform the required work. However, existing PE programs
administered by EPA or DOE primarily manage quality by inspection because they are too infrequent
to ensure that real-time control of data quality is maintained. The INEL ASPEP addresses quality
through prevention, and augments data available from existing PE programs by:
1. Use of real-time quality assurance tools to monitor laboratory and validator performance
when INEL samples or data packages are being processed
2. Performance assessment of validation support suppliers
3. Creation and use of INEL-specific PE soil samples to identify and correct analytical
problems affecting actual INEL samples
4. Use of commercially available PE samples
5. Tracking performance indicators that reflect supplier management practices.
PROGRAM COMPONENTS
The program approach to performance evaluation is schematically represented in Figure 1. The
components of the approach are performance tools, performance indicators associated with these
tools, performance criteria for each indicator, an indicator assessment process and determination of
supplier performance status. The status of analytical service supplier performance is determined by
their records of conformance to criteria. Individual sections of supplier organizations will have a
unique performance status based on the nature of their nonconformances. This performance status
will determine the eligibility of a supplier to receive work from the INEL SMO.
Performance Evaluation Tools
Several performance evaluation tools and the required frequency for their use have been identified
for use in evaluating analytical support suppliers. The analytical laboratory performance tools
include: general laboratory operation assessment tools (audits or desk evaluations, assessment of
deliverables, time management information) and method or analytical discipline-specific performance
tools (existing PE program sample data, blind real-time PE sample data, and routine laboratory QC
sample data, field split data, and field spike sample data). Validation supplier performance tools
include: general assessment tools (audits or desk evaluations) and analytical discipline-specific
assessment tools (assessment of deliverables, dual validation of data packages, and blind test data
packages). These tools were selected to provide information regarding the supplier's management
practices as well as their performance in specific analytical disciplines.
17
-------
For purposes of this discussion, PE tools are referred to as "periodic" or "real-time", depending on
the nature of their use. Periodic tools are used at a set frequency to provide information on supplier
performance, either general or discipline-specific, which is not specifically associated with a particular
batch of INEL field samples. Information supplied by periodic PE tools is indicative of overall
performance or capabilities in a particular analytical discipline or method. Information provided by
real-time PE tools is directly associated with supplier performance on specific batches of INEL field
samples. Real-time tools are those which are either processed by the supplier at the same time as
INEL field samples (e.g., a blind PE sample or laboratory QC) or data (e.g., a blind previously-
validated data package) or involve assessment of supplier deliverables for actual INEL field samples.
Participation in existing periodic performance evaluation programs administered by EPA, DOE, or
DOD is one tool used to assess laboratory performance. Supplier participation in these PE programs
is dependant on the scope of the supplier's support to INEL (e.g., a laboratory supporting the
drinking water program exclusively would not be required to participate in the Water Pollution
program). Table I reflects the participation requirements associated with existing PE programs.
Performance Indicators and Performance Criteria
Performance indicators are associated with each of the tools. These performance indicators are the
basis for qualitative and quantitative assessment of analytical laboratory and data validator
performance. Acceptable performance criteria are defined for each performance indicator. The
indicators and acceptable performance criteria are organized into general laboratory performance,
specific laboratory analyses, and data validation performance categories. Analytical service supplier
performance for each indicator is tracked, trended, and compared to the acceptable performance
criteria. The frequency of nonconformance to these performance criteria are the basis for evaluating
supplier performance.
General laboratory performance indicators are used to assess laboratory management and operational
processes. Areas of assessment include holding times, turnaround times, completeness and accuracy
of deliverables, audit results, and responsiveness. These parameters are assessed for either the
laboratory as a whole, or for specific analytical disciplines and analysis types (e.g., VOCs, SVOCs,
ICP metals, GFAA metals) as appropriate. The general laboratory performance indicators and
performance criteria are listed in Table 2.
Performance indicators for specific laboratory analyses are listed by analytical discipline in Table 3.
The listed indicators are applicable for all sample matrices listed in the Applicable Matrices column
of the table unless otherwise indicated by a parenthetical clarification.
Specific laboratory perfonnance indicators are identified for existing PE programs, INEL-sponsored
blind PE samples, and routine laboratory QC. These indicators provide data for overall discipline
performance and method and analyte-specific performance. The list may be subject to modification
as the Performance Evaluation Program evolves and the usefulness of each indicator is assessed.
Perfonnance indicators for analysis of other matrices not specified in this section will be developed
as needed.
Areas of data validation supplier performance assessment are turnaround times, deliverable accuracy
and completeness, audit results, and responsiveness. The specific indicators with their associated
18
-------
acceptable performance criteria are presented in Table 4. These parameters are indicators of general
organization management and technical performance, and are assessed for the organization as a
whole or on an analytical discipline or analysis type basis, as appropriate.
Performance Indicator Assessment Process
Performance indicator results and compliance with the associated acceptable performance criteria will
be tracked for each supplier over time, using frequency histograms, control charts, and statistical
analysis, as appropriate. Failure to meet the acceptable performance criteria for any performance
indicator constitutes a supplier nonconformance (SNC). These SNCs will be communicated to the
supplier, and corrective action approaches and closure times negotiated, documented, and tracked.
Additionally, trends that indicate possible future problems (e.g., analyte recoveries dropping over
time) will be communicated to the supplier. A form has been created to simplify and expedite these
communication processes.
In certain limited instances, suppliers will he granted a variance from the acceptable performance
criteria for circumstances which would incur a SNC through no fault of their own. For example, if
samples arrive at a laboratory after holding times have expired, the laboratory would be granted a
variance from the holding time performance criteria for those samples.
Supplier Performance Status
Supplier performance status categories have been established and are defined as follows: satisfactory,
probation (warning), suspension (work stopped), and termination of SMO approval. The
performance status of a supplier may be changed for part or all of their organization. Individual
functions within the supplier organization may be affected independent of one another. The current
performance status for each SMO-approved analytical service supplier is determined from results of
performance indicator assessments. The nature and number of the SNCs incurred in the indicator
assessment process determines the status categories assigned to individual supplier areas.
Tables 5 and 6 are performance status matrices constructed to outline the grounds for changing
supplier performance status. The matrices reflect the scope and severity of performance problems
causing status change. Progression from left to right across the matrices reflects increasingly
pervasive impacts due to a performance problem. Progression from top to bottom indicates
increasing severity of a performance problem. An example of increasing pervasiveness is
nonconformances affecting several analytes, which in turn impact an entire multi-analyte method,
and which ultimately, if unresolved, could affect the entire analytical discipline within a laboratory.
The severity of a performance problem increases with multiple occurrences or unresolved corrective
actions in the preceding categories. Repeated or unresolved performance problems will roll a
supplier's status further toward the bottom right of the matrices.
The manner in which the performance indicators are tracked dictates the entry point into the status
change matrices. For analytical laboratories, performance status changes may be invoked on an
analyte, method or analytical discipline basis, or for the entire laboratory. In some cases, laboratory
performance status changes may also be invoked for certain analysis types within the discipline (e.g.,
VOC, SVOC, GFAAS methods), where the performance status change effects more than one method
but not the entire discipline. The performance status of data validators may be changed on an
19
-------
analytical type or analytical discipline basis as well as for the entire organization. AH changes in
supplier performance status involve procurement concurrence; changes involving Stop Work orders
(i.e., suspension and termination of approval) are handled according to DOE procurement practices.
Table 7 summarizes actions and responsibilities associated with performance status changes. If a
supplier (or area within the supplier organization) has had a probation or suspension status change
invoked, corrective actions (CAs) are mandatory to reinstate satisfactory status. CAs (CA) are
negotiated between the supplier and the INEL SMO and closure times established. Contingencies
exist for suppliers to appeal status change actions.
Status records are maintained for analytical laboratory general operations and for analytes, methods,
analysis types, and analytical disciplines, as appropriate for the laboratory's support to the SMO.
Status records for data validation suppliers are maintained for analytes, analysis types, analytical
disciplines and overall performance. Since recurrences of performance problems can immediately
force a supplier into probation or suspension, provisions exist to recognize successful corrective
action. When sufficient time lapses after CA closure without a repeat occurrence of the same
performance problem, the past occurrences will not be held against the supplier when determining
future status changes.
SUMMARY
The ASPEP proposes a strategy for ensuring that analytical service suppliers provide products of
known and appropriate quality to support INEL EM programs. This strategy is a proactive objective
approach to identify and correct situations that may lead to the generation of suspect analytical
products. Objective data for all analytical service suppliers will be compiled and assessed to ensure
that suspect suppliers are not used by INEL projects. The iterative nature of the approach with an
emphasis on communication between the supplier and the customer ensures that total quality
management is acheived by all parties involved.
REFERENCES
1. J. Fisk, EPA Analytical Operations Branch, to CLP and SAS Laboratories, Subject: "Joint
Statement on Fraud in the Community of Contractors for Analysis of Environmental Samples,"
August 9, 1991.
2. Code of Federal Regulations, 40 CFR Part 300, "OSWER Procedures for Contract Laboratory
Program Investigations," Office of the Federal Register, May 1992.
3. EPA, Contract Laboratory Program Statement of Work for Inorganic Analysis Multi-Media Multi-
Concentration, ILMOl.x, most recent revision.
4. EPA, Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, S W-846, most recent
updates.
20
-------
5. EPA, Methods for Chemical Analysis of Water and Wastes, PB84-128677, March 1983.
6. EPA, Contract Laboratory Program Statement of Work for Organic Analysis Multi-Media Multi-
Concentration, OLMOl.x, most recent revision.
7. DOE, Quality Assurance Program Plan for the Waste Isolation Pilot Plant Experimental-Waste
Characterization Program, DOE/EM/48063-1, July 1991.
8. DOE, Performance Demonstration Program Plan for the W1PP Experimental-Waste
Characterization Program, DOE/W1PP 91-016, Waste Isolation Pilot Plant, Carlsbad, New
Mexico.
9. EG&G Idaho, Inc., 1993, Quality Assurance Project Plan for the Development of Idaho National
Engineering Laboratory Site-Specific Performance Evaluation Soil Materials, EGG-ER-10720,
May 1993.
10. EG&G Idaho Inc., 1993, Statement of Work: The Preparation of Idaho National Engineering
Laboratory-Specific Soil Performance Evaluation Samples, ER-SOW-126, February 1993.
11. EG&G Idaho, Inc., 1993, Statement of Work: Inorganic and Radiological Characterization of the
1NEL Laboratory Control Sample Performance Evaluation Soil Material, ER-SOW-141, August
1993.
12. EG&G Idaho, Inc., Quality Program Plan for the Environmental Restoration Program, QPP-149,
Revision 3, 1993.
13. EG&G Idaho Inc., Quality Manual, current issue.
21
-------
Evaluation Tools
• Periodic
• Real-time
Provide
Performance Indicators
• General
• Specific
Associated
Performance Criteria
• Indicator-specific
• Acceptable performance
Basis for
to
ro
Performance Assessment Process
• Track & trend data
• Compare data to criteria
• Supplier non-conformances
• Variances
Establishes
Performance Status
• Satisfactory
• Probation
• Suspension ^.
• Approval Termination"]
Figure 1. Performance Evaluation Approach
-------
Table 1. Required participation in existing PE programs.
Existing Performance Evalu
and
Required Participation
EMSL RIS
Semiannual
EML QAP
Semiannual
USEPA CLP QB
Semiannual
EMSL-CI WP
Semiannual
EMSL-CI WS
Semiannual
WIPP POP
Semiannual
Corp. of Eng.
PE Program for High
Explosive Residues
As Needed
RESL MAPEP
Under Development
Semiannual
Commercial PE Program
As Needed
ation Programs
Frequency
Mandate
Source
DOE
DOE
CERCLA &
INEL ASPEP
CWA & INEL
ASPEP
SDWA&
INEL ASPEP
WIPP&
INEL ASPEP
INEL ASPEP
DOE
INEL ASPEP
Idaho National Engineering Laboratory Environmental Program
Environmental Restoration
Rad
F,W
F,W
S,B
W,S
W,S
Org
W,S
W,S
W
G
(HCV)
Metals
W,S
W,S
W
w,s
w,s
Other
W,S
(CN)
w,s
W
s,w
(HER)
Waste Management
Rad
F,W
F,W
S,B
W,S
W,S
Org
W,S
W,S
G
(NVG)
Metals
W,S
W,S
w,s
w,s
Other
W,S
G
S,W
(HER)
Environmental Monitoring
Rad
F,W
F,W
S,B
W,S
W,S
Org
W,S
W
Metals
W,S
W
W,S
W,S
Other
W,S
W
ro
CO
Sample Matrix Key: F = Air Filter, W = Water, S = Soil, B = Biota, G = Gas
Special Analyses Key: HER = High Explosive Residue; HCV = High Concentration Volatiles; NVG = Non-VOC Gases; CN = Cyanide
-------
Table 2. General laboratory operations performance indicators and acceptable performance criteria.
Area of Assessment
Performance Indicator
Performance Criteria
Tracked by:
Holding Time &
Turnaround Time
• Total number of sample delivery groups (SDGs)
having holding time violations per analysis type
• Total number of SDGs with turnaround time
violations
• Zero (0) SDGs with holding
time violations
• Zero (0) SDGs with
turnaround time violations
Analysis type
Analysis type
Completeness &
Accuracy of Deliverable
• Number of SDGs for which resubmissions are
requested due to missing, incomplete, or inconsistent
forms and COCs, during Level B or Level C validation
per analysis type
• Number of SDGs for which resubmissions are
requested by the data validator due to missing,
incomplete, or inconsistent forms, COCs, or raw
data) during Level A validation per analysis type
• Number of SDGs requiring full resubmission per
analysis type
• Number of SDGs in which specified methods were
not used (i.e., method change not approved by SMO)
per analytical discipline
> < 1 SDG for which
resubmissions are requested
• < 1 SDG for which
resubmissions are requested
' Zero (0) SDGs requiring full
resubmission
' Zero (0) SDGs in which
incorrect methods were used
•Analysis type
Analysis type
Analysis type
• Analytical
discipline
Audits
Number of findings per audit or desk evaluation
repeated from a previous audit or desk evaluation
Zero (0) repeated findings
Organization
Responsiveness
• Number of corrective action responses not received
within required response time frame
Number of corrective actions not closed within the
required closure time frame
• Zero (0) corrective action
responses not received within
required response time frame
• < 3 corrective actions not
closed within the required
closure time frame
Organization
Organization
-------
Table 3. Analysis-specific laboratory performance indicators and acceptable performance criteria.
Analytical
Discipline
Radiochemislry
Mcials
Organics
Applicable
Matrices
W, S, F, B
W, S, F
W, S, G
PE Program and INEL-Sponsored Blind PE Sample Results
Overall Performance (Analytical Discipline or Analysis Type)
Performance Indicators
• Number of analytes misqualified
(i.e., not correctly identified) per
PE sample
• Number of analytes misquant'ified
per PE sample
• CLP QB score, when applicable
* Number of analytes misqualified
per PE sample
• Number of analytes misquantified
per PE sample
• CLP QB score, when applicable
' Number of analytes misqualified
per PE sample
' Number of analytes misquantified
per PE sample
Performance Criteria
• £ 1 per PE sample
• £ 10% of total number
of requested analytes per
PE sample
• CLP QB Score £ 75
• £ I per PE sample
• MRDL or CRDL
• Number of SDGs having
analytical yields outside of
required limits
• % Recovery of laboratory LCS
• % Recovery of LCS PESM (soil)
• % Recovery of detection-level
QC standard (e.g., CRI.CRA)
• Number of reported blanks
> MRDL or CRDL
• % Recovery of surrogate spikes
(fillers)
• Number of SDGs having ICV,
CCV, or ICSAB QC results
outside required limits
* % Recovery of laboratory LCS
(gas)
• Number of repotted blanks
> CRQL or PRQL
• % Recovery of surrogate
standards (water and soil)
• RPD for MS/MSD
(water and soil)
• Number of SDGs wflCAL, CCAL or
internal standard areas outside of
required limits
Performance Criteria
• Within method/SOW requirements
• Within established control limits
• Within method/SOW requirements
• £ 1
• £ 1 SDG affected by any
out-of-conlrol analytical yield
• Within melhod/SOW requirements
• Within established control limits
• ± 50% of true value
• £ 1
• Within method/SOW requirements
* £ 1 SDG affected by any out-of-
conlrol ICV, CCV, or ICSAB
• Within method/SOW requirements
• £ 1
• Within method/SOW requirements
• Within method/SOW requirement
• £ 1 SDG affected by any out-of-
control ICAL, CCAL, or internal
standard areas
10
01
Matrices: W = Water, S = Soil; G = Gas; F = Air Fillets; B = Biota
-------
Table 4. Data validation performance indicators and acceptable performance criteria.
AREA OF
ASSESSMENT
Turnaround Times
Deliverable
Completeness &
Accuracy
Deliverable
Assessment of Periodic
Blinds and Real-Time
Dual Validation
Audits
Responsiveness
PERFORMANCE INDICATORS
• Total number of turnaround time violations per
analytical discipline
• Total number of L&V Reports for which
resubmissions are requested by the SMO per analysis
type
• Number of problems associated with the data
package identified by the data validator as a
percentage of the total number of problems identified
by referee validation per analysis type
• Number of L&V Reports affected by errors evaluating
the following critical parameters: COC, rejected data
points, and incorrectly assigned data qualifier flags
(i.e., those assigned which should not have been
assigned, and those which were not assigned when
they should have been) per analysis type
• Number of findings per audit or desk evaluation
repeated from a previous audit or desk evaluation
• Number of corrective action responses not received
within required response time frame
• Number of corrective actions not closed within the
required closure time frame
PERFORMANCE CRITERIA
• Zero (0) L&V reports with
turnaround time violations
• < 1 L&V for which
resubmissions are requested
• 95% - 100% of problems
identified by referee validation
• < 1 L&V Report affected by
critical parameter errors
• Zero (0) repeated findings
• Zero (0) corrective action
responses not received within
required response time frame
• < 3 corrective actions not
closed within the requried
closure time frame
TRACKED BY:
• Analytical
discipline
• Analysis type
• Analysis type
• Analysis type
• Organization
• Organization
• Organization
ro
O)
-------
Table 5. Grounds for change in laboratory performance status.
Performance
Status
Satisfactory
3robation
Suspension
Termination
of Approval
Scope of Impact
Single Analyte Method or
Single Analyte from Multiple-
Analyte Method
• Failure to meet performance
criteria for any specific
laboratory analysis indicator
tracked by analyte or method
• SNC on 2 or more indicators
• 2 SNCs on a single indicator
• SNC on an indicator which was
the cause of previous probation
or suspension
• Failure to close probation CA
within allowed response time
frame
• 2 open probation corrective
actions on a single indicator
• Open probation corrective
actions on 2 or more indicators
• Failure to resolve suspension
CA within negotiated time frame
• 3 suspensions on a single
indicator
Multiple-Analyte Method
• Failure to meet performance
criteria for any specific
laboratory analysis indicator
tracked by method
• 2 or more probations
affecting a single analyte
• SNCs on multiple analytes
• SNC within a method which
has had a previous probation
or suspension
• Failure to close CA for
mutilple-analyte method
probation within allowed time
frame
• 2 open probations affecting
the method
• Suspension of 1 or more
analytes
• Failure to close method
suspension CA within
allowed time frame
• 3 method suspensions
• Termination of any analyte
Analysis Type
• Failure to meet assessment
criteria for any general
performance indicator tracked
by analysis type
• Probation on multiple
methods within analysis type
• SNCs in 2 or more single-
analyte methods within
analysis type
• 2 SNCs on a single general
indicator tracked by analysis
type
• SNCs on multiple general
indicators tracked by analysis
type
• SNC on a general indicator
which was cause of previous
probation or suspension
• Failure to close analysis type
probation CA within allowed
lime frame
• 2 open analysis type
probation corrective actions
• Multiple method suspensions
within analysis type
• Failure to close analysis type
suspension CA within
allowed time frame
• 3 analysis type suspensions
• Termination of multiple
methods within the analysis
type
Analytical Discipline
• Failure to meet performance
criteria for any general
performance indicator
tracked by discipline
• Probations in multiple
analysis types in discipline
• 2 SNCs on a single general
indicator tracked by
analytical discipline
• SNCs on muliple general
indicators tracked by
analytical discipline
• SNC on a general indicator
which was cause of
previous probation or
suspension
• SNC on CLP QB Score
• Failure to resolve analytical
discipline probation CA
within allowed time frame
• 2 open analytical discipline
probations
• Suspension of multiple
analysis types in discipline
• Sequential SNCs for any WS
PE sample analyte recovery
indicator
• Failure to close analytical
discipline suspension CA
within allowed time frame
• 3 analytical discipline
suspensions
• Termination for multiple
analysis types per discipline
Analytical Support Supplier
Organization
• Failure to meet performance
criteria for any general
performance indicator
tracked by organization
• Probation in 2 or more
disciplines
• Multiple SNCs on a general
performance indicator
tracked by organization
• SNC on multiple general
performance indicators
tracked by organization
• SNC on a general indicator
which was the cause of
previous probation or
suspension
• Failure to close CA for
supplier organization
probation within allowed time
frame
• 2 open supplier probation
corrective actions
• Suspension in 2 or more
disciplines
• Falsification of records/data
• Failure to close CA from
supplier organization
suspension within allowed
time frame
• Termination of 2 or more
disciplines
CA = Corrective Action; SNC = Supplier Nonconformance
-------
Table 6. Grounds for change in data validator performance status.
Performance
Status
Scope of Impact
Analysis Type
Analytical Discipline
Analytical Support Supplier Organization
Satisfactory
Failure to meet performance criteria
for any indicator tracked by analysis
type
• Failure to meet performance criteria
for any indicator tracked by analytical
discipline
Failure to meet assessment criteria for
any indicator tracked by organization
Probation
• 2 SNCs on a single indicator
• SNCs on multiple indicators within
the analysis type
• SNC on an indicator which was the
cause of previous probation or
suspension
• 2 SNCs on a single indicator (tracked
by discipline)
• SNCs on mulitiple indicators tracked
by discipline
• 2 or analysis type probations within
the discipline
• SNC on an indicator (tracked by
discipline) which was the cause of
previous probation or suspension
• 2 SNCs on a single indicator (tracked
by organization)
• SNCs on mulitiple indicators tracked
by organization
• Probation in multiple analytical
disciplines
• SNC on an indicator (tracked by
organization) which was the cause of
previous probation or suspension
oo
Suspension
Failure to close probation CA within
allowed time frame
2 open probation CAs within the
analysis type
Failure to close CAs for discipline
probation within the allowed time
frame
2 open probation CAs within for
indicators tracked by discipline
Suspension of multiple analysis types
within the discipline
• Failure to close CA for supplier
organization probation within the
allowe time frame
• 2 open probation CAs within for
indicators tracked by organization
• Suspension of multiple disciplines
Termination of
Approval
• Failure to close suspension CA
within allowed time frame
• 3 suspensions within the analysis
type
• Failure to close CA due to discipline
suspension within the allowed time
frame
• 3 suspensions for an indicator tracked
by analytical discipline
• Termination of multiple analysis types
within the discipline
• Falsification of L&V reports
« Failure to close CA for supplier
organization suspension within allowed
time frame
• 3 suspensions for an indicator tracked
by organization
• Termination multiple disciplines
-------
Table 7. Supplier performance status resolution.
Condition
Trend Condition
Supplier
Nonconformance
Probation
Suspension
Termination of
Approval
Impact on
Data Quality
None
Minor
Major
Critical
Fatal
SMO Approval
Status
Satisfactory
(Not Affected)
Satisfactory
(Not Affected)
Conditional
(Qualified)
Suspended
(On Hold)
Terminated
(Revoked)
Actions
PE Program Office
• Document
• Counsel supplier
• Document
• Counsel supplier
• Review disposition
• Review CA (if needed)
• Document
• Obtain procurement
concurrence
• Notify supplier
• Evaluate appeal
• Approve CA
• Resolve probation
• Document
• Notify Procurement
• Evaluate appeal
• Approve CA
• Resolve suspension
• Document
• Notify procurement
• Recind supplier
approval & PE program
participation
EG&G Procurement
None
None
• Sign probation notice
• Notify Supplier
• Issue Stop Work Order
• Forward CA to PE
Office
• Lift Stop Work Order
• Notify Supplier
• Issue Stop Work Order
• Terminate contract
(discretionary)
Supplier
Discretionary
• Respond by dispositioning
SCN
• Submit & implement
CA (discretionary)
• Respond by:
Appeal
or
Submit & implement
CA (mandatory)
• Respond by:
Appeal
or
Submit & implement
CA (mandatory)
• Reapply through
supplier approval
process
CO
-------
QUALITY BY DESIGN - IT CAN BE ACHIEVED!
A Review of Two Projects with Different Levels of Project Planning
Robert J. Thielke. Technical Operations Manager, Kerri G. Luka, Quality Assurance
Manager, QuantaLex Inc., 300 Union Boulevard, Suite 600, Lakewood, Colorado 80228
ABSTRACT
Because data validation services are often obtained late in the remediation process after
sample analysis has occurred, these quality control activities are too frequently viewed as
added costs to project budgets as well as frustrating delays to project schedules. These
impressions are often valid. Data validation performed under this scenario serves as
either an expensive inspection function that limits the usability of data or as an even
costlier data rehabilitation activity. However, if quality control activities are designed
into a project from the beginning, data validation activities are transformed from an
expensive inspection process to a more efficient, less costly evaluation and assessment
process. The process then becomes one in which project data quality objectives (DQO's)
are utilized to determine that they were met as planned in the design process. Therefore,
a most effective and efficient use of data validation experts in environmental activities is
to utilize their experience and knowledge in the up-front project planning process.
The same expertise used in the design of a validation system that satisfies project-specific
needs should also be used to develop the following:
• Comprehensive analytical statements of work
• Usable data management systems
• On-site laboratory evaluation plans
• Performance evaluation sample programs/test data packages for compliance to
statements of work
• Laboratory surveillance programs based upon validation results
• Project planning start-up sessions to communicate project requirements in advance
to all project participants (DQO development process)
The value added to the environmental project planning process by involving highly-
skilled validation staff up-front in the project design process is demonstrated through two
scenarios of environmental remediation projects. The first scenario is a medium-sized
project in which the validation staff were introduced into the process after the
procurement of analytical services. A significant amount of the data in this project
proved to be unusable because of problems that could have been prevented if adequate
pre-planning of the project had occurred. The unusable data resulted in the need for
expensive re-sampling efforts and jeopardized important project schedules. The second
scenario is a larger, more complex project in which the validation staff were involved in
all of the pre-planning and surveillance activities. The data for this project satisfied all
project DQOs as a result of the pre-planning activities and project schedules were not
compromised due to re-work.
This paper also describes the necessary elements of each of the pre-planning activities
and describes the skills that knowledgeable Project Managers should look for when
selecting validation personnel to participate project planning.
30
-------
INTRODUCTION
Analytical quality control activities are too frequently viewed as added costs to project
budgets and the cause of frustrating delays to project schedules. These impressions occur
because data validation and analytical support services are frequently obtained late in the
environmental project process, often after laboratory analyses have been completed.
Unfortunately, these impressions are usually valid because data validation and other
quality control actions performed under this scenario serve as either expensive inspection
functions that limit the usability of data or as even costlier data rehabilitation activities.
However, when quality planning activities are designed into a project from the beginning
of the project, data validation and other quality control activities are transformed into cost
effective verification activities. Using qualified analytical support services to implement
project quality planning activities can reduce project cost overruns and scheduling delays.
The cost effectiveness of analytical quality planning can be demonstrated by performing
cost-benefit analyses of two environmental project scenarios. The first scenario is on an
environmental project that did not have sufficient up-front quality planning of
environmental chemical analyses and the second scenario is on a project that had
sufficient analytical quality planning built into the project from the beginning of the
project. The scenarios presented are derived from actual environmental projects and
demonstrate the financial risks, scheduling risks, and other risks that can occur when
planning is not designed into projects. However, the costs of project activities are
estimated based upon similar available project information.
Determining the optimal cost of analytical quality planning can be accomplished by
performing a cost-benefit analysis on the cost of analytical quality planning. The analysis
will consist of identifying the costs of analytical quality planning, determining the cost-
benefits of analytical quality planning, and specifying the desired probability of success.
The probability of success can be factored into the quality planning evaluations by
performing a sensitivity study of the costs and benefits at a low, medium, and high
probability of success.
Performing quality planning on a project does not guarantee success on a project.
Conversely, a project is not doomed to fail if planning does not occur. However, quality
planning activities generally increase the probability of project success. This probability
can be increased if the activities are designed and performed with the help of
knowledgeable analytical services personnel. Identifying the most effective analytical
quality planning tools and understanding their intended purposes are crucial to
maximizing the cost effectiveness of analytical quality planning.
PROJECT SCENARIO NUMBER 1
The analytical requirements for this environmental project consisted of 50 groundwater
samples and 200 surface soil samples. The site consisted of a singular operable unit and
was not associated with any other ongoing environmental project. The samples were
taken to determine the level of metals contamination in an effort to determine whether
further cleanup was required. The samples were collected over a ten week period and
sent to the laboratory for analysis. The metals analyses were diluted to the point that the
detection limits were above the levels required to determine whether further action was
necessary. In addition, the appropriate frequency of duplicate precision and other quality
control analyses specified in the quality assurance plan were not met.
31
-------
As a result, all the metals data points were rendered useless and new samples and
analyses were required. Re-analysis of the previous samples was not possible because
the laboratory disposed of the samples shortly after analysis. The quality criteria
identified in the project Quality Assurance Plan were not addressed in the laboratory
statement of work (SOW) and these requirements were not conveyed to the laboratory.
In addition, neither the laboratory operations or analytical techniques were evaluated
prior to the project. The only analytical quality planning that was performed was to give
laboratory personnel the opportunity to review and comment on the Quality Assurance
Plan.
The costs of quality planning for this project scenario were as follows:
Activity
QA Plan Review
Hours
8
Labor Rate
$50/hour
Extended Cost
(Labor X Hours')
$400
Labor Cost Total
$400
There were no travel or non-labor costs associated with quality planning for this project.
The costs associated with analytical failure on this project are the actual costs of
repeating the sampling and analysis of these samples and other external costs are not
easily quantifiable. The Project Manager decided to re-sample 190 of the soil samples
and 45 of the groundwater samples because the Quality Assurance Plan specified that 90
percent of the data must be usable. Repeating 95 percent of the samples allowed for a
small amount of unusable data in the second sampling episode. The costs of re-taking all
of the samples that were rendered useless for this scenario were as follows:
Labor Costs
Activity
Water sampling1
Soil sampling2
Hours
275
380
Labor Rate
$120/hour
$120/hour
Labor Cost Total
Extended Cost
(Labor X Hours)
$33,000
$45,600
$78,600
Additional costs associated with re-sampling for this scenario were:
Laboratory analysis3
Travel4
Re-validation5
Total Additional Costs
Total Costs of Re-sampling
$110,500
$7,600
$8,225
$126,325
$204,925
The cost of re-sampling must also be adjusted for the time value of money because the
costs of planning occur in the present while the costs of re-sampling occur in the future.
The adjusted cost of re-sampling becomes $198,961 if the current annual interest rate is 3
percent and the re-sampling will take place approximately one year after quality planning
activities. However, the cost of failing to obtain acceptable data on this project due to
insufficient quality planning was manifested in more than just financial terms.
32
-------
The time needed to take, analyze, and evaluate the additional samples caused project
deadlines to be missed. Determining the impact of a missed deadline will vary between
projects. However, it is safe to assume that impacts of missed deadlines are all negative
and could include additional indirect project costs, a loss of future work, and reduced
public confidence in the cleanup effort. Although these costs are not easily quantifiable
they must be considered when determining the level of quality planning to perform on a
project.
This scenario is an extreme example of what can go wrong on an environmental project.
However, it is an indicator of why quality planning is necessary.
PROJECT SCENARIO NUMBER 2
This environmental project consisted of an operable unit that was part of a larger site
with other operable units. The analytical requirements for this environmental project for
purposes of comparison also consisted of 50 groundwater samples and 200 surface soil
samples taken over a ten week period. The samples were taken to determine the level of
metals contamination in an effort to determine whether further cleanup was required.
The analytical quality planning that was performed for this site included the following
activities:
Development of a laboratory SOW that specified detection limits, quality
control requirements, reporting requirements, and sample handling
requirements.
Pre-award laboratory evaluation to determine if laboratory procedures,
personnel, and equipment were adequate to perform the work.
Project set-up meeting with all project participants to discuss the
requirements of the statement of work and clarify any ambiguous issues.
Analysis of test samples by the laboratory. The analytical results were
then sent to qualified data validation staff to determine if the data satisfied
all project requirements. The findings of this evaluation were discussed
with the project participants prior to beginning work on samples collected
from the site.
Less than 2 percent of the 6800 metals data points were rejected and only 3 of the
groundwater samples required re-analysis.
The costs of pre-planning for this project were as follows:
Labor Costs
Activity
Prepare SOW
Project Set-up
Perform Audit
Evaluate Test Data
Hours
60
16
24
8
80
10
Labor Rate
$60/hour
$50/hour
$60/hour
$100/hour
$60/hour
$60/hour
Labor Cost Total
Extended Cost
(Labor X Hours)
$3,600
$800
$1,440
$800
$4,800
$600
$12,040
33
-------
Additional costs associated with planning were:
Laboratory analysis6 $1,000
Two Test Samples7 $500
Total Additional Costs $1,500
Total Costs of Analytical Quality Planning $13,540
Costs of Re-sampling
The costs associated with analytical failure on project scenario 2 were as follows:
Labor Costs
Extended Cost
Activity Hours Labor Rate (Labor X Hours)
Water sampling8 20 $120/hour $2,400
Labor Cost Total $2,400
Additional costs associated with re-sampling were:
Laboratory analysis9 $1,350
Travel10 Not Applicable
Re-validation11 $105
Total Additional Costs $1,455
Total Costs of Re-sampling $3,855
The cost of re-sampling adjusted for the time value of money $3,743 if the current
annual interest rate is 3 percent and the re-sampling will take place approximately one
year after quality planning activities.
This scenario demonstrates the additional sampling costs that can be avoided if effective
quality planning is performed on a project.
Both of these project scenarios illustrate the ultimate cost of failing to build in quality
planning into the project design and the possible cost savings if adequate analytical
quality planning is performed. The level of analytical failure for an environmental
project will fall somewhere in between these two scenarios. Judicious project
management would dictate that some level of analytical quality planning should occur on
a project.
DETERMINING THE OPTIMAL LEVEL OF QUALITY PLANNING
Although a certain level of quality planning does not guarantee project success, the
probability of success will increase as the level of quality planning increases. The
optimal amount of quality planning for a project is determined when the marginal cost of
quality planning equals the marginal benefits of the quality planning at a desired
probability of success. Each Project Manager should evaluate the need for quality
planning and the costs of project failures. The amount of resources committed to quality
planning should vary from project to project. The benefits of analytical quality planning
34
-------
are the cost savings of not having to re-sample and the avoidance of other social costs.
Some relatively small quality planning expenditures will have a large benefit by
improving data quality. However, additional expenditures on quality planning will
eventually exceed the benefit obtained after a certain level of quality is reached. An
example curve of quality costs versus quality benefits was developed by Harris and
Chaney12 and has been adapted for this paper and included as Figure 1. The optimal
level of quality planning is determined by performing a cost-benefit analysis of the costs
of quality planning, the costs of analytical failure, and the required probability of success.
High costs of quality planning in relation to the costs associated with re-sampling would
encourage a lower level of quality planning. Conversely, high costs associated with re-
sampling in relation to the costs of quality planning would encourage a higher level of
quality planning.
The costs of quality planning are usually easily determined since the same basic quality
planning activities are generally performed for each project. The various types of
analytical quality planning activities that can be performed are discussed later in this
paper.
The costs of analytical project failure are more difficult to evaluate because of the
variability of projects, the uses of the data, and the impact of external costs. An
evaluation of the costs of re-sampling should incorporate a number of project variables.
These variables include, but are not limited to, the following:
The number of samples taken and the requested analyses for each sample.
Some analyses can be very expensive or have very short holding times
that would prevent the original sample from being re-analyzed if the first
analysis failed.
The cost of re-sampling different types of samples. For instance, the cost
of drilling additional boreholes are much greater than obtaining additional
surface water samples.
The location of the site can have a significant impact on the costs of re-
sampling. Remote or inaccessible locations would be more expensive to
re-sample than nearby or accessible locations.
The geological and hydrological conditions of the sampling location can
increase the costs of re-sampling. Deep bedrock wells would be more
difficult and more expensive to re-sample than shallow alluvial wells.
Some types of samples can realistically only be taken one time because of
the nature of the sample or the sampling process. These types of samples
would include samples taken to test singular events, or samples taken from
studies or situations that could only be duplicated at extreme costs.
External costs of failure would include the public sensitivity towards the
project, fines, and the impact of schedule delays on future project
activities.
An initial judgment of the benefit of planning activities would indicate that the
investment would be extremely prudent. However, spending money on planning and
quality assurance does not guarantee success on a project. In addition, the probability of
predicting success on a project is difficult and will vary from project to project.
35
-------
Costs of Quality Planning
40000
0
0.2 0.4 0.6 0.8
Probability of Project Success
Figure 1
36
-------
Mansfield indicated that economic decisions based upon risk are a function of the
probability of success and failure and the economic return associated with each outcome.
The benefits of quality planning are best determined by performing a prior risk analysis
based upon the costs of re-sampling and the probability that the project will fail. The
probable adjusted cost savings are estimated based upon the assumption that the Project
Manager is indifferent to risk and will choose the option that maximizes monetary
wealth. The calculation will change based upon the Project Manager's aversion to risk or
acceptance of risk.
A sensitivity study evaluating the costs and benefits of planning at different probabilities
of success would demonstrate the effectiveness thresholds of quality planning on a
project. A benefit analysis of quality planning at a 90 percent (high) probability of
success, a 50 percent (medium) probability of success, and a 10 percent (low) probability
of success would demonstrate the probability at which analytical quality planning
activities would be cost effective. The probable cost savings are determined by:
C(a) = C(r) x P
Where:
C(a) = The probable adjusted cost savings of quality planning
C(r) = The adjusted cost of re-sampling
P(s) = The probability of avoiding re-sampling because of pre-planning13
The table below illustrates the cost savings at low, medium, and high probabilities of
avoiding re-sampling costs for project scenario 1 if the analytical quality planning
activities in project scenario 2 had occurred.
Quality
Planning Cost
$13,540
$13,540
$13,540
P(s)
0.9
0.5
0.1
Probable Adjusted Cost Saving
of Quality Planning
$179,065
$99,481
$19,896
For this scenario the benefits of quality planning exceeds the costs at each level of
sensitivity. Analytical quality planning would be beneficial if the cost of the quality
planning did not exceed the benefit gained as a result of project success at a given
probability of success. This decision would be influenced by the willingness of the
Project Manager to accept the risk of failure. The benefits of quality planning would
exceed the costs for this example if the quality planning had greater than a 6.8%
probability of success. The Project Manager would have to decide if the costs of quality
planning would be beneficial in comparison to the costs of project failure and the
probability of success. The levels of quality planning for a project will vary based upon
the scope of the project and the potential costs of failure. Small projects or projects with
minimal re-sampling costs (such as a preliminary site assessment for a small site) may
not require the same level of quality planning costs as a politically sensitive project or a
project with extremely high costs associated with re-sampling (such as pilot studies or
borehole soil sampling in remote locations.) The desired probability of project success
must also be factored into the cost of project failure. Although a Project Manager would
not usually be willing to pay for complete assurance of project success, tight budgets and
tight schedules would require some level of quality planning with a reasonably high
37
-------
probability of success. The required probability of success becomes much greater at
sensitive or high profile sites.
Determining the marginal benefit associated with a marginal unit of quality planning cost
is difficult to predict. This objective probability of marginal benefit is most easily
determined by the conjecture of probability 14 The example Sinn uses to explain this
statement is that of oil exploration. The objective probability of finding oil in a location
is determined by the relative rate of success in locations with similar geological and
topographical characteristics. A conjecture is made as to the probability of success based
upon the probability of success at similar locations. Obtaining more information about
the location or adding more controls to variables will increase the objective probability of
finding oil. Eventually the amount of information that can be obtained about a location
will be limited by technology and cost. This analogy can be applied to environmental
projects. Based upon the characteristics of a project, the Project Manager can conjecture
about the probability of success by assessing the level of success of similar projects using
a defined level of quality planning. The project will have a higher expected probability
of success as more quality planning is performed. Some measures of control such as
detailed SOWs with specific liquidated damages upon failure to perform have proven to
be a low cost quality planning measure with a high probability of success. Other quality
activities such as inter-laboratory comparison studies may be considerably more
expensive and provide limited quality benefits. A Project Manager should evaluate the
effectiveness of quality planning activities on similar projects to determine which
activities provide the greatest benefit to cost ratios.
QUALITY PLANNING ACTIVITIES
Analytical quality planning activities can help alleviate the probability of project failure
if they are implemented by experienced analytical support staff and in the proper
sequence of events. Experienced analytical support staff could help effectively develop
and implement the following analytical quality planning activities:
• Comprehensive analytical statements of work
• Usable data management systems
• On-site laboratory evaluation plans
• Performance evaluation sample programs/test data packages for compliance to
statements of work
• Laboratory surveillance programs based upon validation results
• Project planning start-up sessions to communicate project requirements in advance
to all project participants (DQO development process)
ANALYTICAL STATEMENTS OF WORK
Comprehensive laboratory analytical SOWs are the primary method of specifying:
• Project objectives
• Analytical methods
• Required detection limits
• Analytical quality control requirements
• Documentation requirements
• Data package compilation requirements
38
-------
Analytical SOWs should reflect the requirements of all primary quality assurance
documents such as the project Quality Assurance Project Plans. The analytical SOW and
the analytical services contract are also the appropriate documents to reference
contractual provisions that can either enhance the quality of the analysis or help recover
costs if the analysis fails due to laboratory error. The types of provisions that should be
included in these procurement documents are analytical holding times, reporting
turnaround times, sample disposition and disposal requirements, and quality control
limits that would prompt re-analysis or re-preparation of samples.
Qualified analytical personnel should help develop the analytical SOW because they
have an understanding of the analytical methods and documentation requirements of the
project. Less experienced personnel may not be aware of the relative weaknesses and
strengths of various methods.
All too often less experienced personnel develop analytical SOWs using modifications of
SOWs previously developed for projects. Revised SOWs are functional as long as all
non-relevant information is removed from the document. Information mistakenly left in
SOWs is usually confusing and costly. Independent experts can assist in developing draft
SOWs that are both effective and concise.
DATA MANAGEMENT SYSTEMS
Most sampling projects should utilize data management systems to track project progress
and to trend laboratory performance. All data management systems benefit from design
and implementation input provided by independent analytical support experts. Analytical
data management systems are often designed without the input of the laboratory or data
validation experts. As a result, the systems are often inefficient or unusable for their
intended purpose. Data management systems are frequently designed by computer
programmers and Project Managers that may not have specific knowledge of
environmental analysis and, as a result, add fields to document unnecessary analytical
information or omit fields that are critical to documenting analytical results, validation
results, and data usability information. Unnecessary information in a database that
contains millions of data points has a multiplying cost factor because each data element
must be generated and reviewed at several points by the analytical, validation, and
assessment personnel to ensure the accuracy of information in the database. This data
entry and review becomes very expensive as the size of the data base grows. Large data
files which house unnecessary information may slow software systems thereby
decreasing the efficiency of the data entry and data management activities. In addition,
expert analytical personnel could work with computer programmers to develop data
management systems that are both streamlined and compatible with existing software
used by laboratories to manage and report analytical data to more efficiently.
LABORATORY EVALUATIONS
A critical element of laboratory procurement is the on-site laboratory evaluation. The
audits should be of sufficient detail and scope to determine if the laboratory has adequate
facilities and equipment to perform the analyses required for the project. The auditors
should also have sufficient experience in the laboratory to assess the experience and skill
of the laboratory personnel to perform the required analyses. Experienced analytical
support personnel could be used to develop comprehensive evaluation checklists and
perform the on-site evaluations.
39
-------
Large sampling programs typically have routine performance evaluation (PE) sample
programs that are used to monitor the performance capabilities at analytical laboratories.
The composition and content of the PE samples should be consistent with the
environmental samples that will be taken. In addition, several different PE samples
should be used to provide a representative profile of the conditions at the site. The PE
samples should be sent to the laboratory prior to contract award and may also be sent
prior to project initiation, after significant changes in laboratory management and
staffing, and as a result of laboratory performance as determined through data validation
and surveillance results. PE samples may be sent more frequently to laboratories that
have had previous performance problems or to laboratories that are receiving critical
samples. Experienced analytical services personnel would be helpful in determining the
appropriate number and composition of PE samples in conjunction with the Project
Manager. PE samples are valuable for determining a laboratory's analytical capability
under optimal conditions. They are also useful prior to procurement because they can
help both the evaluator and the laboratory management to determine whether the
laboratory personnel have a complete understanding of the necessary documentation and
package compilation procedures as specified in the laboratory.
On-site laboratory surveillances are also useful in evaluating laboratory performance and
capabilities. With shrinking project budgets, regularly scheduled quality assurance
activities should be replaced with performance-based quality assurance activities. Data
validation can be used to trend laboratory performance by which to base such
surveillances. Project Managers can remain apprised on the status of the project,
compliance to quality objectives, and the need for additional quality assurance activities
(e.g., surveillances and PE samples) through the use of control charts and trending
reports. As with audits, surveillances should be of a scope commensurate with the data
validation results. Individuals conducting surveillances should have adequate experience
to effectively evaluate the laboratories.
PROJECT MEETINGS
Dialogue in the form of meetings is the another effective tool to ensure that analytical
and data validation services are designed and delivered in a manner that best serves the
needs of the project. Informal project start-up sessions can be set-up by independent
experts to ensure that:
• Project participants feel comfortable to freely discuss the requirements and status of the
project
• Project participants function as a team
• Project participants have a forum in which questions can be asked and answered
without delay
• Problems are discussed and resolved in a timely manner
The frequency and scope of project meetings should be based upon the project funding,
project sensitivity, and project size.
40
-------
INDEPENDENT ANALYTICAL SUPPORT FIRM QUALIFICATIONS
When contracting for an independent analytical support firm, organizations and agencies
should consider the following qualifications:
Is the firm truly independent?
The firm should have no vested interest in the procurement of the laboratories if the
contract includes the development of a draft SOW. The firm should also have no interest
in the outcome of the data validation and analytical results.
Are the Data Validators qualified to perform quality planning activities?
Data Validators should be knowledgeable of the available methods so that they can
provide guidance when selecting methods that best suit the needs of the project.
Laboratory experience is a qualification requirement that is often solicited when
evaluating data validation firms. While laboratory experience can be valuable in the
conduct of on-site evaluations, it can be mutually exclusive of method knowledge and
general data validation experience. Laboratory experience therefore, may not be an
indicator of validator qualifications.
Does the firm have previous experience developing sample data development and
tracking systems ?
Firms that have previous experience developing such systems have knowledge of what
works well and what does not. It is far more beneficial to select a firm that has
experience and knowledge of "lessons learned" from other similar projects than to re-
invent the wheel with each new project.
Does the firm have personnel with adequate verbal and written communication skills?
Developing a sampling, analysis, and validation program with the quality planning
activities previously described demands that the individuals assisting in the development
of the program have exemplary communication skills.
Does the firm have a computer software development department that is knowledgeable
of programming methodologies that will serve the needs of the project?
Organizations and agencies should seek-out firms that have the expertise in the types of
hardware and software required by the project. Software systems can be developed to be
compatible with existing software systems. Describing these pre-existing conditions up-
front with the firm is essential to ensure that systems will be as functional as possible.
Integrating existing systems and newly developed software can be very cost effective and
efficient. Software developers should write code in a generic, modular fashion. If
function-oriented programming and object-oriented programming methodologies are
used, libraries of code should be maintained. A complete library of functions or objects
can reduce the cost of software development by reducing the time to write and test code.
Object-oriented programming is more easily modified for specific projects than other
methodologies because the modules can be re-used without modification. Over time a
complete object-oriented library can reduce costs significantly.
41
-------
SUMMARY
Performing analytical quality planning cost-benefit analysis on environmental projects
indicates that quality planning is a worthwhile expenditure even at low probabilities of
success, However, as evidenced by the scenario presented in this paper adequate levels
of quality planning are not always performed. These failures occur because Project
Managers lack the information to choose and execute adequate quality planning
activities, because Project Managers fail to account for the possibility of failure when
developing project schedules and budgets, and because the irreversible nature of
sampling and analysis is not understood. Undertaking adequate cost-benefit analyses and
allowing expert analytical support personnel to participate in analytical quality planning
activities would enhance the probability of project success. The cost analysis would not
be definite because of the uncertainty associated with the project and because of the
difficulty in quantifying non-fiscal costs. However, based on knowledge from other
projects the Project Manager could determine the objective probability of success for
quality planning activities. Using qualified analytical experts will help to ensure that the
quality planning activities are effective in meeting project objectives. The current
climate in the environmental arena demands that projects are not over budget and that the
work is done in a reasonable and timely manner. Analytical quality planning is an
essential element of helping to ensure project success.
42
-------
ENDNOTES
3.3 wells could be sampled by a two person sampling crew each 10 hour day and the depth of the wells
averaged 50 feet. The time requirements include all set-up, decontamination, documentation, and shipping
activities. These cost estimates are based on a theoretical sampling costs profile developed by CDM-FPC
(Denver). CDM-FPC was not involved in any aspect of this project.
-^
The soil samples were taken within the first six inches of soil. Ten soil samples could be taken by a two
person sampling crew each 10 hour day. The time requirements include all set-up, compositing,
decontamination, documentation, and shipping activities. These cost estimates are based on a theoretical
sampling costs profile developed by CDM-FPC (Denver). CDM-FPC was not involved in any aspect of
this project.
o
The laboratory analysis consisted of 45 groundwater samples and 190 soil samples analyzed for CLP
metals with full documentation at a cost of $450/sample for aqueous samples and $475/sample for soil
samples. The prices for analysis are used from an actual price guide for a laboratory. However, the
identity of the laboratory will not be revealed due to business confidentiality reasons.
The cost of travel was estimated at two airfares at $1,000 per roundtrip airfare and per diem for two
samplers at $100/day for 45 days.
The cost of data validation on the additional 360 samples for full documentation CLP metals is
$35/sample.
The analysis cost of two CLP metals samples with a full CLP data package for the performance
evaluation samples.
The cost of two standard metals performance evaluation samples.
o
0 3.3 wells could be sampled by a two person sampling crew each 10 hour day and the depth of the wells
averaged 50 feet. The time requirements include all set-up, decontamination, documentation, and shipping
activities. These cost estimates are based on a theoretical sampling costs profile developed by CDM-FPC
(Denver). CDM-FPC was not involved in any aspect of this project.
The laboratory analysis consisted of 3 groundwater samples analyzed for CLP metals with full
documentation at a cost of $450/sample for aqueous samples. The prices for analysis are used from an
actual price guide for a laboratory. However, the identity of the laboratory will not be revealed due to
business confidentiality reasons.
There were no travel costs because of the site location.
The cost of data validation on the additional 3 samples for full documentation CLP metals is
$35/sample.
Harris, Douglas and Chaney, Fredrick. Human Factors in Quality Assurance, pg 19.
*3 Mansfield, Edwin. Statistics for Business Economics, pg. 520.
Sinn, Hans Werner. 1983. Economic Decisions Under Uncertainty, pg 12.
43
-------
REFERENCES
Harris, Douglas and Chancy, Fredrick. 1969. Human Factors in Quality Assurance. New York: John
Wiley & Sons, 228p.
»
Sinn, Hans Werner. 1983. Economic Decisions Under Uncertainty. Amsterdam: North Holland
Publishing Company, 350p.
Mansfield, Edwin. 1980. Statistics for Business Economics. W.W. Norton & Company, 580p.
44
-------
ENSURING COMPARABILITY OF DATA GENERATED
BY MULTIPLE ANALYTICAL LABORATORIES
FOR ENVIRONMENTAL DECISION MAKING AT THE
FERNALD ENVIRONMENTAL MANAGEMENT PROJECT
Chris Sutton. Manager, Environmental Division, Barbara Campbell,
Supervisor, Environmental Division, Raymond J. Danahy, Manager,
Environmental Division, Thomas Dugan, Senior Environmental Scientist,
Environmental Division, F. Keith Tomlinson, Senior Environmental
Scientist, Environmental Division, Fernald Environmental Restoration
Management Corporation (FERMCO), P.O. Box 298704, Cincinnati, Ohio 45239-
8704
ABSTRACT
The Fernald Environmental Management Project is a U. S. Department of
Energy (DOE)-ovmed facility located 17 miles northwest of Cincinnati,
Ohio. From 1952 until 1989, the Fernald site provided high-purity uranium
metal products to support United States defense programs. In 1989 the
mission of Fernald changed from one of uranium production to one of
environmental restoration.
At Fernald, multiple functional programs require analytical data.
Inorganic and organic data for these programs are currently generated by
seven laboratories, while radiochemical data are being obtained from six
laboratories. Before final cleanup of the Fernald site occurs, numerous
additional laboratories may well have provided analytical data for
environmental programs.
Quality Assurance (QA) and Quality Control (QC) programs have been
established to help ensure comparability of data generated by multiple
laboratories at different times. The quality assurance program for
organic and inorganic measurements specifies which analytical
methodologies and sample preparation procedures are to be used based on
analyte class, sample matrix, and data quality requirements. In contrast,
performance specifications have been established for radiochemical
analyses.
A blind performance evaluation program for all laboratories, both on-site
and subcontracted commercial laboratories, provides continuous feedback on
data quality. The necessity for subcontractor laboratories to participate
in the performance evaluation program is a contractual requirement.
Similarly, subcontract laboratories are contractually required to generate
data which meet radiochemical performance specifications. The Fernald on-
site laboratory must also fulfill these requirements. Laboratories with
NOTICE: THIS PAPER WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED BY AN AGENCY OF THE UNITED STATES
GOVERNMENT. REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY TRADE NAME,
TRADEMARK, MANUFACTURER, OR OTHERWISE DOES NOT CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR
FAVORING BY THE UNITED STATES GOVERNMENT OR ANY AGENCY THEREOF. THE VIEWS AND OPINIONS OF AUTHORS
EXPRESSED HEREIN DO NOT NECESSARILY STATE OR REFLECT THOSE OF THE UNITED STATES GOVERNMENT, OR ANY
AGENCY THEREOF OR FERNALD ENVIRONMENTAL RESTORATION MANAGEMENT CORPORATION, ITS AFFILIATES OR ITS
PARENT COMPANIES.
45
-------
persistent problems or which generate data falling outside specified
control limits are judged to be failing to generate comparable data.
In summary, although data used for environmental decision making purposes
at Fernald are generated by multiple laboratories, comprehensive quality
assurance and quality control programs have been formulated and
implemented to ensure data comparability.
INTRODUCTION
Management of analytical laboratory services at a Department of Energy
(DOE) site undergoing environmental restoration presents a variety of
challenges. One of the major challenges is ensuring that sufficient
analytical capacity is available to meet site customer schedule
requirements as well as to meet performance criteria with regard to data
quality objectives. Because of the complexity and lengthy duration of
restoration activities at a DOE site, numerous analytical laboratories may
be called upon to provide data during the entire cleanup phase. An
implicit requirement for the use of multiple analytical laboratories is
the criterion that data generated by those laboratories must be
comparable. This paper outlines certain aspects of the program
implemented at Fernald to assess data comparability.
Background
The Fernald site, formerly known as the Feed Materials Production Center,
is a U.S. Department of Energy (DOE)-owned facility located about 17 miles
northwest of Cincinnati, Ohio. From 1952 until 1989, the Fernald site
provided high-purity uranium metal products to support United States
defense programs. DOE is now conducting cleanup activities at the site
under its Environmental Restoration and Waste Management Program.
Environmental remedial actions at the Fernald site are being carried out
in accordance with the Comprehensive Environmental Response, Compensation,
and Liability Act of 1980 (CERCLA) , as amended by the Superfund Amendments
and Reauthorization Act of 1986 (SARA); and also in accordance with the
Resource Conservation and Recovery Act of 1976 (RCRA), as amended by the
Hazardous and Solid Waste Amendments Act of 1984 (HSWA). CERCLA response
activities are being implemented at the Fernald site pursuant to the terms
of an Amended Consent Agreement between DOE and the U. S. Environmental
Protection Agency (USEPA), and RCRA activities are being conducted under
the provisions of a Consent Decree with the State of Ohio. Thus, CERCLA
and RCRA are two principal drivers associated with analytical support
services. In 1992 DOE implemented an Environmental Restoration Management
Contract or ERMC approach to the Fernald site, with the Fernald
Environmental Restoration Management Corporation (FERMCO), a wholly owned
subsidiary of Fluor Daniel, Inc., being awarded the ERMC contractor.
46
-------
FEMP PROGRAMS REQUIRING ANALYTICAL DATA
Five broad functional programs require analytical data at the Fernald
site: utilities operations and National Pollutant Discharge Elimination
System (NPDES) operations, CERCLA characterization activities, RCRA
characterization and monitoring activities, sitewide environmental
monitoring, and CERCIA Treatability Studies (coordinated with DOE
technology demonstration programs).
Utilities and NPDES analyses are performed on waste water and discharge
water for a series of inorganic compounds. These include nitrates,
nitrites, total uranium, total thorium, ammonia, fluoride, metals, and
total dissolved solids.
Analyses associated with CERCLA activities often require a full suite of
organic and inorganic analyses performed in accordance with EPA SW 846 or
Contract Laboratory Program (CLP) protocols at high data quality levels.
Radiochemical analyses associated with CERCLA activities may include
strontium-90; technetium-99; americium-241; cesium-137; lead-210;
polonium-210; actinium-227; neptunium-237; radium-226, 228; thorium-227,
228, 230, 232; uranium-234, 235/236, 238; and plutonium-238, 239/240, 241.
Analyses conducted for RCRA operations are performed to characterize
drummed waste and construction waste and to determine if various
substances and materials should be classified as RCRA hazardous. Organic
and inorganic analyses are performed in accordance with SW 846 analytical
protocols. Additionally, Toxic Characteristic Leaching Procedure (TCLP)
analyses are routinely performed for specific organic compounds and
metals.
Environmental monitoring programs at the Fernald site consist of periodic
sampling and analysis of surface waters, ground waters, air, home owner
wells, sediments, and selected biota to determine the impact of the site
on the surrounding environment and to establish a baseline against which
the progress of remediation efforts can be measured. The same types of
organic, inorganic, and radiochemical analyses required for the CERCLA
programs are required for these programs as well, although detection
limits and data quality levels may be different between the CERCLA and
environmental monitoring programs.
Bench-scale treatability studies are initiated to determine the optimum
chemical, physical, and engineering parameters for conducting a specific
type of remediation; the chemical and physical characteristics of products
resulting from remediation; and the handling characteristics of
remediation products. Pilot-scale technology demonstration programs
delineate the feasibility of processes prior to field implementation.
Inorganic and organic data are currently being generated by seven
laboratories six commercial laboratories and the Fernald on-site
laboratory. Radiochemical data are provided by six laboratories five
commercial laboratories and the Fernald on-site laboratory. Before final
clean-up of the Fernald site occurs, numerous additional laboratories may
well have provided analytical data for environmental programs.
47
-------
COMPARABILITY
Comparability is one of the five PARCC (precision, accuracy^
representativeness, completeness, and comparability) parameters used as
indicators of data quality.(1) A typical definition of comparability is
given below:
"Comparability is a qualitative parameter expressing the confidence
with which sample measurement data can be compared with measurement
data for similar samples and sample conditions." (2)
However, such a definition is too generic to implement a practical program
to assess data comparability for multiple laboratories. Accordingly,
FERMCO has defined comparability (in the Fernald sitewide sampling and
analysis QA plan discussed later in this paper) in terms which provide a
quantitative statistical basis for assessing data comparability.
"Comparability refers to one of five elements identified by the
USEPA to describe data quality. It is an expression of the
confidence with which one data set can be compared to another.
Analytical data for the same analyte generated by the same
analytical procedure (whether by the same laboratory at different
times or by different laboratories) are comparable provided that
specified acceptance criteria for quality control parameters such as
detection limits, accuracy, precision, matrix spikes, etc. are met
or exceeded. Data for the same analytes generated by different
analytical procedures are also comparable provided that specified
acceptance criteria for quality control elements such as those
listed above are met or exceeded."
This definition of comparability has three advantages. First, it relates
comparability to common analytical quality control parameters. This in
turn has the benefit that common tools to assess data quality can be used
to assess data comparability. Second, the definition implies that
accurate and precise data are comparable; inaccurate and imprecise data
are not. Thus, to the extent that data can be judged to be accurate and
precise, data can also be judged to be comparable or incomparable. Each
of the two advantages listed above provides a means of evaluating
comparability for organic and inorganic data. Third, the definition opens
the door for use of performance based methods by emphasizing the meeting
of acceptance criteria for quality control parameters. The use of
performance based methods for radiochemistry is an integral part of the
analytical services program at Fernald as discussed below.
DATA COMPARABILITY ASSESSMENT PROGRAM
The program which FERMCO has implemented to address data comparability
contains six essential elements: a quality assurance plan for
environmental sampling and analysis, performance based methods for
radiochemical analyses, contractual requirements for analytical
subcontractor laboratories, performance evaluation, laboratory QA and QC
audits, and data verification and validation. The first four of these
elements are discussed in detail below.
48
-------
Quality Assurance Plan
A quality assurance plan governing all environmental sampling and analysis
activities was implemented at Fernald. The plan consolidates QA and QC
requirements for a variety of environmental programs.
The functions of this QA plan are primarily three-fold: 1) to set minimum
performance standards for sampling and analysis activities, 2) to direct
that all Fernald environmental programs and activities follow these
standards to ensure programmatic and temporal consistency, and 3) to allow
data gathered under one cleanup program to be used by other cleanup
programs.
One of the key elements in the Fernald QA (referred to as the Sitewide
CERCLA Quality Assurance Project Plan [SCQ]) plan that significantly
promotes comparability of current and future data are method selection
tables for organic, inorganic, wet chemical, and historic Fernald site
methods. The organic, inorganic, and wet chemical methods listed in the
method selection tables are EPA or other standard methods commonly used
for CERCLA and RCRA activities and commonly performed by the commercial
analytical community. EPA methods may include 200-500 series methods
(40CFR141), 600 series methods (40CFR136), SW 846 methods (40CFR261) and
CLP-SOW methods. Other standard methods include those listed in "Standard
Methods for the Analysis of Wastewater" (3), and those listed in American
Society for Testing and Materials (ASTM) publications.
Because of the presence of radionuclides at the Fernald site, specific
methods have been developed for radiochemical and chemical analysis of
certain elements (e.g. uranium and thorium). Although these methods have
a long history of use, they have not been promulgated nor have they been
compiled as "standard methods" due to potential limited applicability.
These methods are called historic Fernald site methods.
TABLE 1 gives an example of the Fernald site QA plan method selection
tables. These tables specify which analytical methodologies and sample
preparation procedures may be used at the Fernald site based upon analyte
class, sample matrix, and data quality level. (Data quality levels at the
Fernald site are called Analytical Support Levels, or ASL Levels. ASL
Levels A, B, C, D are generally equivalent to EPA data quality levels I,
II, III, and IV.) The primary intent of the method selection tables are
to ensure uniformity and consistency of method selection and application.
Such uniformity and consistency is an essential element toward the
promotion of comparability.
Performance Based Radiochemistrv Methods
Unlike inorganic and organic analyses, no single compilation of
promulgated standard methods exists for radiochemistry determinations.
Multiple sample preparation methods and multiple instrumental detection
techniques are available commercially for many of the radionuclide
analytes. Additionally, standard established quality assurance/quality
control requirements and acceptance criteria have not been established for
environmental radiochemistry analyses. Nonetheless, inter-laboratory
49
-------
comparison studies in performance evaluation programs have demonstrated
that accurate and comparable radiochemical data are attainable even though
different analytical procedures are used.
To alleviate concerns over a potential lack of data comparability from
multiple laboratories performing radiochemical analyses, FERMCO has
adopted the approach of utilizing performance based methods for
radiochemical analyses. FERMCO radiochemists (in conjunction with USEPA
and commercial laboratory radiochemists) developed a set of performance-
based criteria which cover a range of radiochemical analytes, matrices,
and data quality levels (ASLs).
To establish such criteria, FERMCO formulated performance requirements for
minimum detectable concentrations, tracer/chemical recovery, matrix spike
recovery, method blank concentration, precision of duplicates, and
accuracy of laboratory control samples. An example of such performance
criteria is shown in Table 2. USEPA Region V has approved the use of
these performance criteria, and they are routinely used in statements of
work for subcontracted radiochemistry analyses. Contractual requirements
for analytical services are discussed in more detail in the ensuing
sections.
Contractual Requirements
As mentioned earlier, currently seven laboratories provide organic,
inorganic, and wet chemical analytical data to the Fernald site; while six
laboratories provide radiochemical data. Because the primary contaminants
of concern at the Fernald site are radionuclides, this section will
discuss radiochemical contractual requirements that directly affect data
comparability.
Radiochemical analyses are performed by commercial laboratories through
the use of Task Order Subcontracts. Qualified laboratories (the
qualification process falls outside the scope of this discussion) bid to
supply specific radioanalytical services which are defined in Task Orders.
General-type requirements in the original statement of work as well as
detailed specifications in each task order bearing upon data comparability
must be met. The general requirements are delineated below:
1. Subcontractor Quality Assurance Plan
The vendor shall have a Quality Assurance (QA) program which
addresses the applicable requirements of the most recent
version of the Fernald Sitewide CERCLA Quality Assurance
Project Plan (SCQ) and ANSI/NQA-1. The SCQ must be a
contract-specified attachment to the laboratory-specific QA
plan and is the governing document in all matters relating to
subcontractor QA programs. If there is any disagreement
between a subcontractor QA program and the SCQ, the latter
document shall govern.
50
-------
2. Preaward Verification of Analytical Capabilities
In evaluating each subcontractor laboratory to be included on
an approved list under the Radioanalytical Laboratory Services
Task Order Subcontract (RLS-TOS), each one must submit a
series of tables summarizing recent data from their laboratory
for two matrices and each analyte of interest that
demonstrates they can meet the FERMCO radiochemical analysis
performance specifications. An example of the required
summary tables (16 required) is shown in Tables 3A and 3B.
3. Postaward Confirmation of Analytical Performance
After all analyses have been completed for a specific task
order, summary tables similar to those described above in Item
#2 must be provided which demonstrates the extent to which the
laboratory complied with the radiochemical analysis
performance specifications designated in the SCQ. These data
must be provided for all six Performance Parameters (ASL C or
D) for each analyte/sample matrix combination included in the
task order. The data must be provided in summary form. The
Performance Parameters (if applicable) which must be addressed
for each analyte/sample matrix combination are:
• Highest Allowable Minimum Detectable Concentration
(HAMDC);
• Percent Overall Tracer/Chemical Recovery;
• Percent Matrix Spike Recovery;
• Method Blank Concentration;
• Laboratory Control Sample: Percent of Known Value; and
• Precision Requirements for Duplicate Samples (RER).
This information must be provided before work under each task
order is considered complete.
4. Participation in External QA Programs
Subcontractor laboratories must participate in QA programs
conducted by the DOE and the USEPA. The DOE's Environmental
Measurements Laboratory (EML) conducts a Quality Assessment
Program (DOE/EML-QAP) designed to evaluate the capabilities of
laboratories to perform accurate environmental radiochemical
analyses. FERMCO has established a quantitative scoring
system, which includes specific pass/fail criteria, for
evaluating each laboratory's performance in the DOE/EML-QAP.
This system was used to initially qualify laboratories for
providing radioanalytical services. In addition, this
evaluation system also sets minimum performance criteria for
51
-------
subsequent rounds of the DOE/EML-QAP which a laboratory must
meet in order to remain qualified for bidding on task orders.
Laboratories must also participate in the USEPA's
Environmental Measurements and Support Las Vegas Laboratory
(EMSL-LV) Intercomparison Study evaluation program. A
laboratory's performance in the EMSL-LV program is evaluated
to determine if significant analytical problems exist.
However, no pass/fail criteria have been established for this
QA program.
5. FERMCO Administered Performance Evaluations
FERMCO conducts continuing performance evaluations by
submitting quality control samples. Samples for ongoing
performance evaluation are provided by FERMCO or a third party
as part of a performance evaluation program.
Item 1 above ensures that all commercial laboratories utilize the same
radiochemical QC elements, radiochemical performance specifications, and
acceptance criteria for those specifications. Item 2 ensures that
contractors can actually meet the performance specifications prior to
performing any analyses for the Fernald site. Tables 3A and 3B are
examples of proof of capability to meet radiochemical performance
specifications (FERMCO audits verify data authenticity). Item 3 ensures
that radiochemical data generated for the Fernald site in each task order
meet the QC performance acceptance criteria. This item in particular
contributes significantly toward ensuring data comparability. Item 4
helps ensure that well qualified contractor laboratories are selected
initially and that these laboratories subsequently maintain an acceptable
level of performance. Finally, Item 5 specifies that contractor
laboratories (and the Fernald on-site laboratory) must participate in a
FERMCO administered analytical performance evaluation program. As
discussed in the following section, this program is another key element in
assessing data comparability.
Performance Evaluation Program
The purpose of the FERMCO performance evaluation program is to assess
comparability of data as well as the performance of multiple laboratories
over a long period of time. As stated above, all commercial laboratories
performing analyses for the Fernald site must participate in this program.
The Fernald on-site laboratory is also included in this program.
Spiked samples (typically soils and water) are sent on a monthly basis to
each contracted laboratory as well as to the Fernald on-site laboratory.
The quantity of performance evaluation samples sent to individual
laboratories is based upon the number of field samples sent to each
laboratory for analysis. Identical sets of PE samples are sent to each
participant in the program.
52
-------
Basically, two sets of statistical parameters are calculated. The first
is percent recovery defined as:
Percent Recovery = Measure^ Concentration of Added Analyte Found x ^QO
Known Concentration of Analyte Added
The second is the deviation from the group mean which is defined as:
Deviation from Mean = |xi - x[
S
| I = absolute value
Xi = value of individual result
X = group mean
S = standard deviation of the mean
Control charts are maintained for percent recovery where warning limits
are set at the 95% confidence limit and control limits are set at the 99%
confidence limit. Figure 1 shows a typical control chart for multiple
laboratories based upon available data.
Finally, laboratories are ranked each month according to their
performance. Rankings are based both upon percent recovery and average
deviation from the mean. The closer a laboratory is to 100 percent
recovery, the higher it is ranked. Thus, for a given analyte the
laboratory with the absolute value of recovery closest to 100 is ranked
number one, while the laboratory with the absolute value of recovery
furthest from 100 is ranked lowest. Similar rankings are established for
each analyte using the deviation from the mean. Table 4 shows an example
of rankings of laboratories for the month of March, 1994. The overall
rankings in Table 4 are based upon the combined % recovery and deviation
from the mean rankings.
Through the use of statistical processes, control charts, and rankings,
the performance evaluation data are used to identify trends, outliers, and
problem areas. Laboratories with persistent problems or which
consistently generate data falling outside control limits are judged to be
failing to generate comparable data. Until corrective actions are taken
to remedy problems to the satisfaction of FERMCO, no further samples will
be sent to these laboratories.
CONCLUSIONS
Comparability of data is essential for environmental decision making at
Fernald because multiple laboratories currently supply analytical data,
and use of multiple laboratories will continue throughout the lengthy
duration of the environmental restoration process. A definition of
comparability that involves commonly used quality control parameters
provides a basis for implementing a data comparability assessment program.
The key elements of the program discussed above include a comprehensive QA
plan for sampling and analysis activities, performance based methods for
53
-------
radiochemistry analyses, contractual requirements for analytical
subcontractor laboratories, and performance evaluation programs . Although
the data comparability program is a relatively recent development at
Fernald, all indications are that it is working well. Consequently, data
being generated now and in the future will facilitate environmental
decision making.
REFERENCES
1. "Data Quality Objectives for Remedial Response Activities",
EPA/540/G-87/003, March 1987, EPA Contract Number 68-01-6939.
2. Goheen, S. C.; McCulloch, M.; Thomas, B. L.; Riley, R. G.; Sklarew,
D. S.; Mong, G. M. ; Fadeff, S. K. , (eds.); "DOE Methods for
Evaluating Environmental and Waste Management Samples", DOE/EM-
0089T, March 1993, Contract Number DE-AC06-76RLO 1830.
3. Greenberg, A. E. ; Clesceri, L. S.; and Eaton, A. D., (eds.);
"Standard Methods for the Examination of Water and Wastewater" , 18th
edition, American Public Health Association, Washington, D. C.,
1992.
54
-------
TABLE 1
EXAMPLE OF FEMP SCQ METHOD SELECTION TABLES
1
AHALYTE OR
CLASS OF AHALYTE
la. VOCs
Ib. VOCs
(Drinking Water)
2. Metals by GFAA
3. Metals by ICP
4. Cyanide (Tot)
5. Cyanide (Low)
6. Alkalinity
7. Oil & Grease
8. Thorium, Low Level
9. Uranium, Low (ppm) Level
10. Uranium, High Level
ASL
B
C,D
B
B
C,D
B
C,D
B
B
B
B
B
B
B
MATRICES AND METHODS
HATER & HASTEHATER
PREP
METHOD(S)'-2
W
W
W
SW 846-3020
or 7060<®,
7740(S> or
7761<">
W
SW 846-3010
or 7760m
W
W
W
W
W
W
W
W
AKALYTICAL
METHOD(S)
SW 846-8260
CLP*
EPA 524.2
SW 846-7000
series or
3500<4) series
CLP*
SW 846-6010 or 3500(4>
series
CLP*
335.2<5)
335. 3<»
3 10. I**
or 2320B<">
SW 846-9070
EPM 10800,
3059®, 3063(S)
EPM 3002<*>
EPM 1039<»
son. & SOLIDS
PREP
METBOD(S)"
W
W
N/A
SW 846-3050
or 7761<«
W
SW 846-3050
or 7760m
W
W
W
N/A
W
W
W
W
ANALYTICAL
METHOD(S)
SW 846-8260
CLP*
N/A
SW 846-7000
series or
3500(4) series
CLP*
SW 846-6010 or 3500'41
series
CLP*
335. 2<"
335. 3*
N/A
SW 846-9070 or 9071
EPM 1080",
3059(5), 306318
EPM 3002(5>
EPM 1039(5)
cn
Ol
SW 846-1311 (TCLP) could be a prep; however, it is not necessary in all cases.
"W" signifies that preparation is contained in the analytical method.
Methods for Chemical Analysis of Water and Wastes. EPA 600/4-79-020.
Standard Methods for the Analysis of Water and Wastewater. 17th ed.
Historic Fernald site method.
7060 contains the preparation for As, 7740 for Se, and 7761 for Ag.
7760 contains the preparation for Ag.
USEPA Contract Laboratory Program Statement of Work, most recent.
-------
TABLE 2
EXAMPLE OF RADIOCHEMICAL ANALYSIS PERFORMANCE SPECIFICATIONS FOR ANALYTICAL SUPPORT LEVELS C AND D
ANALYTE:U-234, U-235/236,
U-238
PERFORMANCE PARAMETERS
Highest Allowable
Minimum Detectable
Concentration (HAMDC)(1)
Percent Overall
Tracer/Chemical
Recovery(6)
Percent Matrix Spike
Recovery(6)
Method Blank
Concentration
Laboratory Control
Samples: Percent of
Known Value (6)
Precision Requirements
for Duplicate Samples
SAMPLE MATRIX
WATER
0.2 pCi/L
50-100%
50-100%
3, take corrective actions and
reanalyze the batch of samples.
(6) Recoveries or percentages of known values which are 15% above or below the ranges listed are acceptable
on an infrequent basis, i.e., less than 15% of the time. These occurrences must be investigated and
explained. If more than 15% of the recoveries are outside the ranges listed, take corrective actions and
reanalyze samples.
-------
TABLE 3A
LABORATORY "A" DATA DEMONSTRATING CAPABILITY TO MEET FERMCO
SCQ RADIOCHEMICAL ANALYSIS PERFORMANCE SPECIFICATIONS
AT ASLs C/D
HEPTDNIUM - 237
MATRIX
WATER
SOIL
NUMBER OF
SAMPLES
12
27
15
12
15
12
12
24
15
12
15
12
PERFORMANCE
PARAMETER
HAMDC
Tracer Recovery
MS Recovery
Blank Cone .
LCS Recovery
MD RER
HAMDC
Tracer Recovery
MS Recovery
Blank Cone .
LCS % Recovery
MD RER
SCQ
REQUIREMENT
0.5 pCi/L
50-100%
50-100%
-------
TABLE 4
OVERALL RANKING OF LABORATORIES PARTICIPATING IN THE FERMCO PERFORMANCE EVALUATION PROGRAM
FOR THE MONTH OF MARCH, 1994
LAB
D
F
B
C
A
E
G
METAL/
SOIL
2**
1
4
6
7
5
3
VOA/
WATER
5
1
1
6
4
3
7
VOA/
SOIL
3
6
2
3
1
5
7
WET/
CHEM
2
5
5
***
1
3
4
SEMI-
VOA/S
1
***
4
1
4
***
3
PEST/
SOIL
1
***
3
1
4
5
6
RAD/
SOIL
4
1
2
3
6
***
5
SUM OF
RANKING
18
14
21
20
27
21
35
*
AVERAGE
RANKING
2.6
2.8
3.0
3.3
3.8
4.2
5.0
ABSOLUTE
RANKING
1
2
3
4
5
6
7
en
oo
* = Sum of rankings divided by number of analyte groups analyzed.
** = Ranking for a specific Analyte Class
*** = Laboratory did not participate in or submit results on this sample set.
-------
FIGURE 1
01
CD
LLJ
>
o
o
130
120
110
100
90
80
ARSENIC
BY PERCENT RECOVERY
A
,*.
O
70
DECEMBER 1994
A B
* O
95% CONFIDENCE LIMITS 87 - 121%
99% CONFIDENCE LIMITS 82 - 126%
MEAN 103.9% POINTS 13
JANUARY 1994
SAMPLE MONTH
c
n
D
A
E
A
UPPER CONTROL LIMIT
MEAN
LOWER CONTROL LIMIT
FEBRUARY 1994
-------
UTILIZING STATISTICS IN SAMPLING
TO MEET DATA QUALITY OBJECTIVES AT THE
SAND CREEK SUPERFUND SITE
by: Erna Acheson. Remedial Project Manager (RPM)
Rick Edmonds, Quality Assurance Manager (QAM)
U.S. Environmental Protection Agency Region VIII
999 18th Street, Suite 500
Denver, Colorado 80202-2466
ABSTRACT
The 300 acre Sand Creek Superfund Site, located north of
downtown Denver, Colorado is in a heavy industrial area.
There are few residences within the vicinity of the Site but
the population swells during the day to a few hundred people
working in the area. The focus of this paper is to discuss
the pesticide contaminated soils located at this Site.
Poor pesticide manufacturing practices during the 1960's and
1970's resulted in approximately 10,000 cubic yards of
contaminated soil. Pesticides manufactured at this Site have
been banned from production for over 10 years in part due to
their persistence in the environment.
Years of investigations at this Site yielded a Record of
Decision (ROD) for cleanup of 14,000 cubic yards of
contaminated soils by soil washing. The cleanup cost estimate
in the ROD was for 5 Million dollars. The volume of soil to
be cleaned up detailed in the ROD was primarily based on only
17 soil samples. Subsequent to the ROD, a soil washing
treatability study revealed that the cost of cleanup by this
method would be 13 million dollars.
The goal in presenting this paper is to provide a "horror
story beginning" with a "success story ending". The horror
story begins with over double the cost estimate for cleanup
and only 17 soil samples for data. The story end successfully
through use of statistics, planning for meeting data quality
objectives, and following "current" Superfund Accelerated
Cleanup Model (SACM) guidance.
INTRODUCTION
The Sand Creek Superfund Site (Site) is located in Commerce
City, Colorado, a suburb northeast of Denver, Colorado. The
300 acre Site has been the site of various industrial
activities for more than 40 years. The Colorado Organic
Chemical Company (COCC) manufactured pesticides and
insecticides in the 1960s and 1970s which tainted the soil.
Site inspections in the 1960's and 1970's by the Colorado
Department of Health (CDH) revealed violations in the storage
60
-------
and handling of pesticide products and waste. Subsequently,
EPA conducted several investigations at the Site in the early
1980s. Sampling and analysis indicated the presence of
chlorinated organic compounds, pesticides and metals. This
information was used to place the Site on the National
Priorities List (NPL) in 1982.
Several investigations were conducted to facilitate
remediation of contaminated soil. These included a Remedial
Investigation (1988), Feasibility Study (1989), Remedial
Design (RD) (1993), and a pilot scale soil washing
treatability study (1992) . In addition, demolition and removal
of selected COCC area buildings, tanks, rail cars and chemical
waste plant residues occurred in 1991-1992. The Record of
Decision (ROD) specified soil washing to remediate 14,000
cubic yards of contaminated soils. The cost estimate in the
ROD to perform the washing of contamination from the soils was
$5 million. Based on the results from the treatability study
(1992) , EPA determined that a more accurate cost estimate for
soil washing was $13 million.
The expected site remediation costs were thus some $8 million
greater than originally anticipated. The soils washing pilot
tests also showed that soil washing would not reduce the
concentrations of dieldrin and heptachlor to the action levels
specified in the ROD. As a result of this treatability study,
another sampling effort was conducted to better characterize
the current extent of the contamination. This was done in
order to reduce the area needing remediating and possibly
modifying the action levels upward. Unbelievable as it may
seem the original cleanup estimate of $5 million was based on
the analysis of only 17 soil samples.
As a solution to this problem the EPA contractors were
directed to prepare a proposal for a sampling program to
better define the contamination problem; the draft proposal
recommended the collection of 2820 soil samples. A desperate
situation existed. The cleanup was to be completed within 1.5
years at a cost near the original estimate. This was a large
number of samples for Remedial Design, that would take a great
deal of time to complete and could disrupt the cleanup
schedule (this is may be an appropriate number of samples for
a Remedial Investigation, but not for a Remedial Design).
The focus of this paper is to provide a "horror story
beginning" with a "success story ending." The horror story
begins with over double the estimated cost for cleanup based
on only 17 soil samples. The story ends with utilizing a
sampling strategy / methodology that follows "current"
Superfund Accelerated Cleanup Model (SACM) guidance.
61
-------
CLEANUP GOALS
Chemicals of Concern were identified in the ROD: dieldrin;
heptachlor; chlordane; DDT; 2,4-D; arsenic; and chromium.
Based on data from the Sand Creek Remedial Investigation (RI;
1988) and OU5 Risk Assessment (RA; 1990) , dieldrin and
heptachlor were chosen as driver compounds for remediation of
OU5 due to their carcinogenicity and concentrations. By
treating dieldrin and heptachlor to action levels of 0.155 and
0.553 mg/kg, respectively an acceptable overall carcinogenic
risk (for the occupational soil-ingestion pathway) of 2.7E-05
would be achieved for the site. The extent of contamination
from other COCs was believed to generally coincide with
dieldrin and heptachlor contamination. Therefore, the overall
site risk for industrial workers would be lowered to at least
the acceptable lifetime excess cancer riskrange of l.OE-04 to
l.OE-06 by focusing on reducing dieldrin and heptachlor
concentrations to the specified action levels.
SAMPLES
The jump from 17 samples to 2820 samples was clearly not
reasonable. As stated earlier, 2820 samples for a Remedial
Investigation may be warranted; but, this work was to be
performed as part of the Remedial Design. Therefore, the QAM
was contacted by the RPM to aid in reaching a reasonable
sampling strategy.
The QAM and RPM, together with two other EPA scientists,
planned a sampling strategy to utilizing statistics to reduced
the number of required samples. Statistics as the basis for
radom grid sampling performed during the initial sampling
event. This initial phase of sampling consisted of collecting
10 grab samples and one composite sample from within a 50' by
50' grid of 10' by 10' cells. This random sampling was
performed to determine the range of pesticide contamination
within the cells and across the site. This sampling strategy
provided the variability of heptachlor and dieldrin
concentration levels in soils within the 50' by 50' grid, as
well as between the cells.
The initial sampling was done to determine the variability of
the soil across one "exposure unit" (EU). An EU was defined
as one 50 X 50 feet square which is 5 feet deep. The EU's
sampling depth horizons were 0 to 1 foot, 1 to 3 feet, and 3
to 5 feet. Seven one foot samples were taken by random
(calculated location of random) for three separate EUs and
pesticide levels measured for each sample. Additionally, a
split of each seven foot sample EU was mixed together for a
composite measurement of pesticides in each EU. Variability
in levels of pesticides within each EU and between EUs was
then calculated.
62
-------
The initial goal was to arrive at an optimum amount of
sampling in order to proceed with the final Remedial Design.
It was decided that an additional goal of this initial
sampling was to provide information on field immunoassay which
would analyze for pesticides in the field. Unfortunately, the
field immunoassay testing was not performed due to the lack of
test kits available on the market today. Therefore the second
phase of sampling, which would have evaluated the quality and
usefulness of immunoassay field kits, was not carried out.
The number of samples required depended in part on the results
of the initial sampling program. The final phase of sampling
used recognized statistical methods with input from those
persons familiar with pesticide contaminations. In order to
determine which EUs to remediate, the team worked in
conjunction with the toxicologist; the toxicologist's input in
part was based on an evaluation of "hot spots".
Initially the COCC processing area was investigated by the
same methodology as the adjacent railroad corridor. Because
of several factors it was determined that the railroad
corridor area needed to sampled under a different sampling
stategy than the COCC property. These factors included risk
determinations, cost of remediation, and difference in origin
of contamination. The risk evaluation on the railroad
corridor was determined by a child passing through the
corridor scenario which is different than the on-site worker
scenario utilized for the COCC facility risk evaluations. The
cost of remediating the railroad corridor past the top foot of
soil woud be excessive for there are three high pressure fuel
product lines paralleling the tracks as well as a 72 inch
sewer main serving metro Denver. In addition, the source of
much of the railroad corridor contamination was from surface
water drainage from the COCC processing area, which is
considered a secondary residual waste. The type of pesicides
found at the Site tend not to be very mobil. Therefore, the
assumption was made that the higher levels of contamination
would be found in the top foot of soil on the COCC property
and that the COCC area would require more extensive sampling
than the railroad corridor.
In addition to separating the Site into two distinct sampling
areas, a more comprehensive list of chemicals were analyzed
for in order to better define the apportionment of risk. The
goal was to identify chemicals that contribute significantly
to the overall human health risk at the Site so that a
comprehensive list of cleanup levels would achieve site
restoration goals.
SAMPLING METHOD
Soil samples were obtained for pesticide and metals analysis
in soils for each cell. Samples obtained from the railroad
corridor were collected by hand using a stainless steel bucket
auger. The soil was then mixed or homogenized in a stainless
63
-------
steel bowl and then transferred to sample jars, except for
that portion of the sample to be analyzed for volatile organic
compounds (VOC), which was transferred to a jar before
homogenization. The samples were then placed in ice, labeled,
and sent to a Contract Laboratory Program (CLP) laboratory for
analysis. Samples obtained from the COCC area were collected
by using a stainless steel split spoon sampler advanced with
the aid of a truck mounted drill rig. A hand auger was used
in cells inaccessible by the truck mounted rig. Samples
collected for VOC analysis were collected from each depth
interval at every third cell, from cells with cell numbers
divisible by three. One blind field duplicate for each set of
20 samples was collected for VOC, metals, and pesticide
analysis.
The sampling stategy utilized for the railroad corridor was to
sample the top foot of soil in each grid. No deeper samples
were obtained. There were some earlier samples that were
taken from depths greater than the first foot of soil. These
samples indicated lower levels of contamination than samples
taken at the surface.
The sampling stategy utilized for the COCC area is called "a
decision tree methodology" in which samples are obtained from
three horizons: A-top, B-middle, and C-lower. Then, analysis
was performed on all soil in the B-middle horizon. If the
soil was found to be contaminated above the "action levels"
then it was assumed that the A-top horizon was contaminated,
and the C-lower horizon was analyzed to determine if the
contamination continued past the B-middle horizon. However,
if the B-middle soil was clean, then the A-top horizon was
analyzed to determine if contamination was present. This
methodology greatly reduced the number of samples requiring
analyses.
Three horizons were sampled on the COC property. The A-top
horizon is defined as the surface to one foot depth, the B-
middle horizon is from one to three foot depth, and the C-
bottom horizon from three to five foot depth. Samples taken
at one time from one to three and three to five foot depths do
not impact the cost as much as re-sampling at different times.
The greatest cost savings are that the "decision tree
methodology" allows for the reaonable assumption of where
contamination would likely exist depending on the results of
the B-middle horizon analysis. In this case there were 157 B-
middle horizon samples analyzed, 97 A-top horizon samples
analyzed and 58 C-bottom horizon samples analyzed. This
methodology saved analyzing 159 samples. The cost savings
from this methodology was approximately $330,000. In this day
of limited resources and tight budgets, this savings should
not be considered lightly.
As with any assumption, the assumptions made for this
methodology were checked against the analytical findings. The
64
-------
following tables indicate that in this case the assumptions
made were reasonable. The A-top horizon has higher levels of
contamination than B-middle and C-bottom horizons.
Additional field studies were undertaken in 1992 and 1993 as
part of the RD and the results of the treatability study were
impetus to preparing the ROD Amendmant (1993). The field
sampling provided updated contaminant data and to fill in data
gaps resulting from the focussed nature of earlier studies.
In addition to reducing the area requiring remediation by over
25% EPA determined in 1993 that low temperature thermal
treatment (LTTT) would be more cost-effective than soil
washing in remediating tese soils. Unbelievable as it may
seem the current estimate for cleanup is $3 million which is
$2 Million less than original cost estimate. The agency goal
for completion of the cleanup by LTTT is September 1994. The
LTTT began in mid-May 1994.
Utilizing statistics to meet data quality objectives for the
sampling effort at the Sand Creek Site was clearly successful.
In addition the knowledge gained allowed the team to make a
management decision for a more cost-effective approach to
remediating the contaminated soils at the Site.
The additional field sampling satisfied the following
objectives:
* Reduced the volume of soil needing remediation by
approximately 25%.
* Provided a comprehensive and thorough knowledge of what
areas require remediation, i.e. supplement a 1988 site-
wide RI with more current and complete data.
* Reduced the number of samples to be analyzed through a
phased sampling strategy and decision tree methodology.
* Reduced the time and costs from the initial plan (2820
samples) substantially.
* Provided data to re-evaluate contaminants so that more
comprehensive health based cleanup levels could be
determined.
The methodology used in this "success story" can readily be
applied at other sites which at first glance may appear to
require substantial sampling.
Savings realized from the initially proposed 2820 samples
versus the 332 samples were substantial. Considering
personnel costs, travel, oversight, equipment, supplies and
analytical costs, the total savings was in the area of
$500,000. Several months time was saved in the field sampling
and several additional months in analytical work. It should
65
-------
also be noted that several laboratories would have had to be
involved if the initially proposed program had been carried
out. The agency's goal regarding the time for cleanup also
would not have been attained if the inital proposal was
accepted.
Further cost-savings may have also been realized if the
technology of immunoassay kits had been available for this
study. This was initially investigated; however, the
immunoassay is contaminant-specific and the pesticides that
were of concern, e.g. heptachlor, dieldrin, etc. were not
available at that time for the method. We encourage industry
to develop the specific analyte capability for most
pesticides; because, this could be a cost-savings tool for all
of EPA in pesticide investigations. This would quickly
identify "hot-spots" and provide a more timely
characterization of where cleanup efforts should be
concentrated.
66
-------
Abstract
DEVELOPMENT OF A LABORATORY METHODS COMPARISON PROGRAM: A
TOTAL QUALITY APPROACH TO RESOLVING A LABORATORY METHODS
COMPARISON QUESTION
0. P. Bredt, T. Y. Hosaka, K. J. Kuhl-Klinger;Pacific Northwest Laboratory1 (PNL),
and K. N. Pool, A. D. Rice, C. Stacey, L. H. Taylor, W. I. Winters; Westinghouse
Hanford Company.
Two primary laboratories, the Pacific Northwest Laboratory's, Analytical Chemistry
Laboratory (ACL) and Westinghouse Hanford Company's Hanford Analytical
Services (HAS) laboratory, analyze Hanford High-Level (radioactive) Tank Waste
samples. Data from both laboratories feed into various Tank Waste Remediation
programs where decisions on treatment, storage, and disposal will ultimately be
made. The fact that both laboratories do not use the same procedures for
preparation and analysis of tank material has led to questions regarding the
comparability of analytical results between the two laboratories. To answer these
questions of comparability, a collaborative was formed. This collaborative, called
the Quality Assurance/Quality Control (QA/QC) Triad, includes participants from
each laboratory and the Hanford sample-management office. The QA/QC Triad
initiated a program that helped to answer the question of comparability and also
provided insight on how analytical processes could be improved. This program,
called the Sample Exchange/Evaluation (SEE) program, combines elements of both
a performance evaluation and methods-comparison study (on real world samples).
It provides the requisite information to make meaningful changes at the laboratory
level and provides information to data users to ensure common understanding of
results generated.
Keys to the SEE program success are attributed to the involvement of all the key
players (QA/QC Triad) in the decision-making process, a "common sense"
approach to resolving our common problem, and an emphasis on continuous
improvement. Our poster will describe the process employed in developing this
highly successful program, and will point to ways in which this program can be
"cloned" at both commercial and other DOE analytical laboratories.
1 Pacific Northwest Laboratory is operated for the United States Department of Energy by
Battelle Memorial Institute under Contract DE-AC06-76RLO 1830.
67
-------
8
ASSESSING DATA QUALITY FOR A FEDERAL
ENVIRONMENTAL RESTORATION PROJECT:
RATIONALIZING THE REQUIREMENTS OF MULTIPLE
CLIENTS
Valerie R. Kiszka. Quality Assurance Coordinator, Environmental Restoration Division,
and Tina M. Carlsen, Environmental Chemistry and Biology Group Leader,
Environmental Restoration Division, Lawrence Livermore National Laboratory,
7000 East Avenue, L-619, Livermore, California 94550.
ABSTRACT
Most environmental restoration projects at federal facilities face the difficult task
of melding the quality assurance (QA) requirements of multiple clients, as well as dealing
with historical data that are often of unknown quality. At Lawrence Livermore National
Laboratory (LLNL), we have successfully integrated the requirements of our multiple
clients by carefully developing a QA program that efficiently meets our clients' needs.
LLNL is operated by the University of California for the Department of Energy (DOE).
The Site 300 Experimental Test Site is operated by LLNL in support of its national
defense program. The responsibility for conducting environmental contaminant
investigations and restoration at Site 300 is vested in the Site 300 Environmental
Restoration Project (Site 300 ERP), which is part of LLNL's Environmental Restoration
Division (ERD). LLNL Site 300 ERP must comply with the QA requirements of several
clients, which include: the LLNL Environmental Protection Department, the DOE, the
U.S. Environmental Protection Agency-Region IX (EPA), the California Regional Water
Quality Control Board-Central Valley Region, and the California Department of Toxic
Substances Control. We will present a hierarchy of QA documents prepared for the
various clients, how they interact, and how they are implemented within the LLNL Site
300 ERP. This comprehensive QA program was used to determine the acceptability of
historical data. The Site 300 ERP began soil and ground water investigations in 1982.
However, we did not begin receiving analytical quality assurance/quality control
(QA/QC) data until 1989; therefore, the pre-1989 data that were collected are of unknown
quality. The U.S. EPA QAMS-005/80 defines data quality as the totality of features and
characteristics of data that bears on its ability to satisfy a given purpose. In the current
context, the characteristics of major importance are accuracy, precision, completeness,
representativeness, and comparability. Using our established QA program, we
determined the quality (as defined by EPA QAMS) of this historical data based on its
comparability to the post-1989 data. By accepting this historical data, we were able to
save a considerable amount of money in recharacterization costs.
INTRODUCTION
Environmental contaminant investigations of the Site 300 Experimental Test Site began
in 1982 to investigate the impact to the site's soil, rock, and ground water by the
operation of nine solid waste landfills, and to determine the extent of contamination of
trichloroethene (TCE) from past waste handling practices. Ground water, soil, and rock
chemistry and field measurement data were required to assess the extent of
68
-------
contamination, understand the hydrogeologic characteristics of the site, determine the
nature and location of possible sources of contamination, develop public health and
ecological assessments, evaluate potential remedial action alternatives and engineering
designs, and characterize baseline conditions.
In 1986, environmental investigations, routine environmental surveillance, and routine
environmental monitoring were consolidated in the Environmental Protection Department
(EPD) of Lawrence Livermore National Laboratory (LLNL). By 1987, all Site 300 soil,
rock, and ground water investigations were performed by what is now the Site 300 ERP,
which is a section of ERD within EPD. Also, by this time, a major effort was underway
to consolidate and centralize all environmental chemical analytical data collected
previously during various Site 300 activities. A centralized data management team was
formed within ERD, and a centralized database on a DEC VAX mainframe computer was
created.
The data management team collected all hard copy reports of analytical chemical data.
Assisted by ERD chemists and geologists, this team verified all historical analytical
reports, ensuring that proper sampling, handling, and analytical protocols were followed,
and that proper documentation concerning the sample was available. After verification
was complete, the data were entered into the centralized database. Analytical results
failing such verification were excluded from the database or properly annotated. Samples
were analyzed by both on-site LLNL laboratories and off-site commercial laboratories
that were California state certified. Site 300 ERP started receiving QA/QC
documentation from the off-site analytical laboratories in 1989. The QA data generated
by on-site LLNL laboratories continue to be archived by them and are available for
review by the Site 300 ERP QA chemists.
Prior to 1989, reports from off-site analytical laboratories contained minimal QA
information. However, these reports did contain sufficient information required for data
acceptance into the database. Reports always included LLNL sample identification,
analytical laboratory identification, sample matrix, date sampled, date analyzed, and
analytical results. Generally, the reports also included the analytical method, reporting
detection limit, and certification by the laboratory manager. If the validity of a particular
result was questioned, the laboratory was requested to provide all associated QC data for
review by Site 300 ERP QA chemist.
From 1982 through 1989, field investigations were conducted under the guidance of the
California Regional Water Quality Control Board (RWQCB)-Central Valley Region.
Site 300 was placed on the National Priorities List on August 30, 1990, and guidance and
oversight of the cleanup was transferred to the Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA) regulations under EPA Region IX, the
RWQCB, and the California Department of Toxic Substances Control (DTSC). In 1990,
a formal manual of standard operating procedures (SOPs) and quality assurance project
plan (QAPjP) were prepared for the Site 300 ERP. Both documents were revised in
1992. The Site 300 ERP's data must meet the specific quality criteria documented in the
QAPjP to be considered meaningful to the data users.
69
-------
Presently, the investigations at Site 300 require the sampling and analyses of more then
430 monitor wells for various parameters with the emphasis on volatile organic
compounds (VOCs), inorganics, high explosive (HE) compounds, and radionuclides,
using methods and procedures functionally equivalent to the methods and procedures
used in the EPA Contract Laboratory Program (CLP) and the California DTSC Certified
Laboratory Program. California state certified commercial laboratories are being utilized
for the majority of analyses. However, on-site laboratories are used for QA/QC purposes
to analyze collocated samples for HE compounds and for special radiochemistry analyses.
Figure 1 details the environmental regulatory and QA/QC history of the Site 300 ERP.
DISCUSSION
Figure 2 depicts the document hierarchy that establishes the QA requirements governing
the Site 300 ERP The Department of Energy (DOE) initiated DOE Order 5700.6C in
1991 to improve the safety and reliability of the Department's programs, projects, and
facilities. The 10 criteria of DOE Order 5700.6C direct organizations to develop,
implement, and maintain a written quality assurance program. To comply with the DOE
order, EPD within LLNL, developed a Quality Assurance Management Plan (QAMP)
based on the 18 criteria of the American Society of Mechanical Engineers Nuclear
Quality Assurance (ASME NQA-1) and satisfying the 16 elements of the Environmental
Protection Agency Quality Assurance Management Staff (EPA QAMS). The EPD
QAMP ensures that EPD management provides planning, organization, direction, control,
and support to achieve EPD's objectives; that the line organizations achieve quality; and
that overall performance is reviewed and evaluated using a thorough assessment program.
The ERD developed the ERD Quality Assurance Plan (QAP), also based on NQA-1 and
EPA QAMS, as required by the EPD QAMP. The ERD QAP establishes and presents the
framework of requirements that shall be met in planning, performing, documenting, and
verifying ERD quality-affecting activities. Since the Site 300 ERP is a part of ERD, it
must comply with all the QA requirements of the QAP through the development of
Quality Implementing Procedures (QIPs) and SOPs. The Site 300 ERP is also required
by its Federal Facility Agreement (FFA) and EPA QAMS, to write a QAPjP that
documents the policies, organization, objectives, functional activities, and specific
QA/QC activities designed to achieve the data quality goals of the restoration project.
Along with the revision of the QAPjP in 1992 and the development of the ERD QAP in
1993, came many quality improvements and implementation of new quality affecting
procedures. For example, 100% of the data generated by analytical laboratories is
reviewed by the Site 300 QA chemist before being entered into the database. Figure 3
shows the flow of Site 300 project data. The QA chemist reviews the data for internal
consistency, technical adequacy, and quality. The quality of the data is judged by the
chemist's review of the QC data generated by the laboratory. The off-site commercial
laboratories are contractually required to provide method blank, laboratory control
sample, matrix spike, and matrix spike duplicate results with every analysis. Calibration
information is made available upon the request of the QA chemist. If problems are
found, the QA chemist can qualify the data by assigning CLP-like data qualifier flags that
are entered into the database. All data are flagged in the database with the appropriate
analytical level (I-V). Periodic statistical analyses are performed on all the data in the
database to identify outliers. When outliers are flagged, the QA chemist investigates the
70
-------
possibility of data entry or analytical laboratory errors. Electronic validation/verification
procedures are being developed to expedite the review of data by the QA chemist.
The Site 300 ERP QAPjP requires that 10% of all samples collected be collocated. Five
percent are sent blind to the primary laboratories and 5% are sent to the QA/QC
laboratories to assess interlaboratory and intralaboratory precision. The QAPjP also
specifies that Site 300 ERP will conduct performance checks twice a year and a systems
audit annually. Currently, the off-site analytical laboratories are receiving blind VOC and
metals performance check samples on a quarterly basis.
Before an off-site laboratory is awarded an analytical contract, they must be California
state certified, and their QA program and operating procedures are evaluated for technical
adequacy. On-site LLNL laboratories must be evaluated by the EPD QA Office and
placed on a Qualified Suppliers List before work can begin. The EPD QAMP requires
that off-site analytical laboratories also be subjected to audits before a contract can be
awarded. The auditors verify that the laboratory is in compliance with its internal
procedures and QA program, and that all DOE and EPD requirements are met.
The Site 300 ERP QA program is frequently assessed and/or audited by DOE, EPD, and
through ERD's self assessment program. The Site 300 ERP monitors the QA program by
submitting QA reports to Site 300 ERP management at least annually. These QA reports
may summarize inter- and intralaboratory precision, performance check sample results,
qualified data, outiiers, any audit findings, completeness, accuracy of laboratory control
samples and matrix spikes, and progress toward future quality goals.
While the majority of the data collected prior to 1989 was not reviewed by the QA
chemist for compliance with matrix spike, matrix spike duplicate, and laboratory control
sample precision and accuracy acceptance limits, the data in most cases were analyzed by
California state certified laboratories using standard analytical methods. Since the
laboratories were certified by the state and had met the strict state criteria, we assumed
that the analytical laboratory chemists had already reviewed the data for quality and
technical adequacy.
By melding all the Site 300 ERP clients' QA requirements and our own QA/QC
procedures, the Site 300 restoration project data produced since 1989 are legally
reproducible, defensible, and of known quality. The pre-1989 data were then visually
compared to the usable current data, looking for variances and anomalous trends. Figure
4 contains a graph of four monitoring wells that have been sampled and analyzed for TCE
since 1982. The graph demonstrates the natural variability of TCE within the ground
water. On the basis of this examination and comparison, the Site 300 ERP has
determined that the majority of the pre-1989 data is internally consistent with the
post-1991 data. Historical data were accepted and used together with more recent data of
known quality for delineation of the nature and extent of contamination at Site 300 and
for use in the baseline quantitative risk assessment.
71
-------
CONCLUSION
The Site 300 ERP has successfully implemented all of its clients' QA requirements and is
continually striving to improve the quality of the project's data and the availability of the
data to the user. By showing the historical data to be equivalent to the more recent data
and therefore acceptable, the Site 300 ERP was not required to duplicate their previous
site characterization effort and consequently saved the DOE a considerable amount of
money.
72
-------
Figure 1. Environmental regulatory and QA/QC history of LLNL Site 300
Environmental Restoration Project.
Date
1982
1982-1989
1988
March 1989
April 1989
July 1989
July 1989
1990
August
1990
October
1990
October
1990
October
1990
1991
1992
June 1992
January
1994
Present
Event
Site 300 seeks informal guidance from the California RWQCB
after discovering soil and ground water contamination.
Field investigations under the guidance of California RWQCB.
Started to receive QA/QC monthly from the California state
certified analytical laboratories for all VOC analyses.
RCRAa 3008(h) Order: required LLNL to investigate all
releases of hazardous waste or hazardous constituents at the
site and to determine if corrective actions under RCRA were
required.
Draft Cleanup and Abatement Order: required LLNL to
continue environmental restoration efforts already underway.
RCRA 3004(u) Corrective Action Order: required LLNL to
comply with provisions of RCRA 3008(h) Order.
Site 300 proposed for inclusion on National Priorities List.
Started to receive QA/QC with individual analytical reports.
Standard Operating Procedures (SOPs) and Quality Assurance
Project Plan (QAPjP) developed.
Site 300 placed on NPL; guidance and oversight of Site 300
cleanup transferred to CERCLAb regulations.
RCRA 3008(h) negotiations suspended in favor of a CERCLA
Federal Facility Agreement (FFA).
Draft Letter of Agreement issued to cover Site 300
environmental restoration until FFA is signed.
Letter of Agreement signed by U.S. DOE and U.S. EPA.
Began receiving QA/QC data electronically by special request.
EPD QAMP implemented.
Began assigning data qualifier flags and analytical levels to
analytical data. SOPs and QAPjP revised and approved by
regulatory agencies.
Federal Facility Agreement signed by U.S. DOE, EPA, DISC,
and RWQCB-Central Valley.
ERD QAP approved.
ERD working with analytical laboratories to develop electronic
QA/QC data transfer protocol.
Lead
agency
California
RWQCB
California
RWQCB
EPA IX
California
RWQCB
EPA IX
EPA IX
EPA IX,
RWQCB,
DISC
EPA IX
EPA IX
EPA IX
EPA IX
a RCRA = Resource Conservation and Recovery Act.
b CERCLA = Comprehensive Environmental Response, Compensation, and Liability Act.
73
-------
Figure 2. The Lawrence Livermore National Laboratory's Environmental
Protection Department's Quality Assurance Document Hierarchy.
DOE Order S700.6C
i
ASME NQA-1-1989
EPD Quality Assurance Management
Plan (EPD QAMP)
Developed by EPD QA Office
EPA QAMS-004/80
AND - 005/80
ERD Quality Assurance
Plan (ERD QAP)
Developed by ERD QA Office
I
Quality Implementing Procedures
(QIPs)
Developed by ERD QA Office
Applicable or Relevant and
Appropriate Requirements
of Federal, State and
local Environmental
Regulations
I
I
Standard Operating Procedures
(SOPs)
Developed by ERD Sections/Groups
Task/Activity Plans; including QA
Project Plans (QAPjPs)
Developed by ERD Sections/Groups
±
Task-Specific Procedures/Instructions
(when necessary)
Developed by ERD Sections/Groups
74
-------
Figure 3. Flow of LLNL Site 300 Environmental Restoration Project data.
Chemistry Measurements
(Soil and Ground Water
Analyses)
Field Meassurements
(Geophysical Logging, Water Level
Measures and Physical
Parameters, and
Pump Test Results)
t
Sample collection and completion
of required field notes
I
Chain of custody documentation
Perform field measurement and
complete required field notes
I
Archive field measurement forms
Electronically record reported measurements
I
Sample analysis
Chemistry Group
Data Manager
Field Geologist
Hydrogeology Group
Data Manager
Data validation/verification
Data verification
Entry into LLNL Ground Water
Project relational
database
75
-------
Figure 4. Time trend graph of trichloroethene concentrations in four
monitoring wells located at Site 300.
105
3
»nt
s
1
•£
8
,0 1000
•g
£
o
c
o
1
8 10
§
u
01
1 1 1 1 1 1 1
3 0
Q O o 08 on
Q _ OfiOQ _o ~~ r\ &j> &O& O O
o o 8 o ou $90 ^->
0 r>n ° xv O «° ^ *
-------
REFERENCES
Carlsen, T.; Ridley M.; Kiszka, V., "Quality Assurance Project Plan, Lawrence
Livermore National Laboratory Site 300 Environmental Restoration Project", Lawrence
Livermore National Laboratory, Livermore, CA. 1992; UCRL-AR-103160 Rev. 1.
Environmental Protection Department, "Environmental Protection Department Quality
Assurance Management Plan, Rev. 3", Lawrence Livermore National Laboratory,
Livermore, CA. 1993.
Environmental Restoration Division, "Environmental Restoration Division Quality
Assurance Plan, Rev. 0", Lawrence Livermore National Laboratory, Livermore, CA.
1993.
U.S. Department of Energy, "Quality Assurance", Office of Nuclear Energy & Office of
Environment, Safety, and Health, U.S. Department of Energy, Washington D.C. 1991;
DOE Order 5700.6C.
U.S. Environmental Protection Agency, "Interim Guidelines and Specifications for
Preparing Quality Assurance Project Plans," Office of Monitoring Systems and Quality
Assurance, U.S. Environmental Protection Agency, Washington, D.C. 1980; 20460,
QAMS-005/80.
U.S. Environmental Protection Agency Region 9, California Department of Toxic
Substances Control, Central Valley Regional Water Quality Control Board, and U.S.
Department of Energy, LLNL Site 300 Federal Facility Agreement, Lawrence Livermore
National Laboratory, Livermore, CA.
Webster-Scholten, C.P., et al., "Draft Final Site-Wide Remedial Investigation Report
Lawrence Livermore National Laboratory Site 300", Lawrence Livermore National
Laboratory, Livermore, CA. 1993; UCRL-AR-108131dr.
Work performed under the auspices of the U.S. Department of Energy by the Lawrence
Livermore National Laboratory under Contract W-7405-Eng-48.
77
-------
EXPLORATORY DATA ANALYSIS USING EPA'S CONTRACT LABORATORY
PROGRAM DATA
Peter Novick, Muhannad Kanaan, Roger Litow, M. V. Chacko, DynCorp Viar, Inc., 300
North Lee Street, Alexandria, Virginia 22314-2695
David Eng, U S EPA Analytical Operations Branch, Hazardous Site Evaluation Division,
1235 Jefferson Davis Highway (52046), Arlington, VA 22202
The EPA regulates and manages one of the largest environmental analytical data
repositories in the nation. All data in electronic media submitted to the EPA under the
national Contract Laboratory Program (CLP) is stored in this repository - the CLP
Analytical Results Database (CARD). The database contains analytical results for over
three hundred thousand samples collected over a seven year period. This vast database
is an ideal source of data for exploratory searches catering to numerous interests
whether they be studies on instrumental performance, quality control parameters,
analytical methods, spiking compounds, or even searches for possible trends or patterns
across multiple analytical protocols and databases. Exploratory data analysis (EDA),
when appropriately controlled and carefully balanced by classical statistical methods, has
contributed to scientific progress.
This paper discusses the prospects of conducting such an exploratory data analysis
experiment using CARD. In our search for a non-trivial subject of interest, we decided
to search for possible trends or patterns that would suggest an effect on the quantitation
and recovery of certain matrix spiking compounds due to the presence of inorganic
elements. Although prior studies have established that a correlation can exist, our initial
searches for the same in CARD did not suggest this. As indicated by other studies, the
analytical and environmental effects of these elements is related to several other factors,
including the pH of the environment, the soil matrix, as well as the metal species and
anions present and their concentrations. This was complicated further by the fact that the
database did not contain some of the information necessary to ascertain the state of these
factors.
We proceeded to take control of some of these factors in our further probing of the data.
Resistant methods were applied to pay less attention to outliers. The inorganic analytes
were segregated based on their chemical properties; and the organic sample spike
recoveries were examined at three different pH ranges. We began to see interesting
patterns. The next step is to establish a resistant fit that may include the bulk of the data,
and if one is found, to separate this fit from the data, leaving behind a resistant residual.
These residuals can warn of important systematic aspects of data behavior crucial to
further explorations, aiding us in confirmatory data analysis that would assess the
reproducibility of the observed trends.
Finally, the data behavior can be displayed via a variety of visual/graphical representation
techniques that would enable the viewer to grasp the unexpected results as well as the
more familiar features of the data as determined by the EDA experiment.
78
-------
EXPLORATORY DATA ANALYSIS USING EPA'S CONTRACT LABORATORY
PROGRAM DATA
Peter Novick, Muhannad Kanaan, Roger Litow, M. V. Chacko, DynCorp Viar, Inc., 300
North Lee Street, Alexandria, Virginia 22314-2695
David Eng, U S EPA Analytical Operations Branch, Hazardous Site Evaluation Division,
1235 Jefferson Davis Highway (52046), Arlington, VA 22202
The EPA regulates and manages one of the largest environmental analytical data
repositories in the nation. All data in electronic media submitted to the EPA under the
national Contract Laboratory Program (CLP) is stored in this repository - the CLP
Analytical Results Database (CARD). The database contains analytical results for over
three hundred thousand samples collected over a seven year period. This vast database
is an ideal source of data for exploratory searches catering to numerous interests
whether they be studies on instrumental performance, quality control parameters,
analytical methods, spiking compounds, or even searches for possible trends or patterns
across multiple analytical protocols and databases. Exploratory data analysis (EDA),
when appropriately controlled and carefully balanced by classical statistical methods, has
contributed to scientific progress.
This paper discusses the prospects of conducting such an exploratory data analysis
experiment using CARD. In our search for a non-trivial subject of interest, we decided
to search for possible trends or patterns that would suggest an effect on the quantitation
and recovery of certain matrix spiking compounds due to the presence of inorganic
elements. Although prior studies have established that a correlation can exist, our initial
searches for the same in CARD did not suggest this. As indicated by other studies, the
analytical and environmental effects of these elements is related to several other factors,
including the pH of the environment, the soil matrix, as well as the metal species and
anions present and their concentrations. This was complicated further by the fact that the
database did not contain some of the information necessary to ascertain the state of these
factors.
We proceeded to take control of some of these factors in our further probing of the data.
Resistant methods were applied to pay less attention to outliers. The inorganic analytes
were segregated based on their chemical properties; and the organic sample spike
recoveries were examined at three different pH ranges. We began to see interesting
patterns. The next step is to establish a resistant fit that may include the bulk of the data,
and if one is found, to separate this fit from the data, leaving behind a resistant residual.
These residuals can warn of important systematic aspects of data behavior crucial to
further explorations, aiding us in confirmatory data analysis that would assess the
reproducibility of the observed trends.
Finally, the data behavior can be displayed via a variety of visual/graphical representation
techniques that would enable the viewer to grasp the unexpected results as well as the
more familiar features of the data as determined by the EDA experiment.
79
-------
10
COST AND QUALITY EFFECTIVENESS OF OBJECTIVE-BASED AND STATISTICALLY-BASED
QUALITY CONTROL FOR VOLATILE ORGANIC COMPOUNDS ANALYSES OF GASES
J. T. Bennett, Scientific Specialist, Engineering Research and Applica-
tions Department (ERAD), C. A. Crowder, Senior Scientist, ERAD, S. J.
Sailer, Senior Scientist, ERAD, and M. J. Connolly, Scientific Specialist,
Environmental Restoration and Waste Management Department, Idaho National
Engineering Laboratory, EG&G Idaho, Inc., P.O. Box 1625, Idaho Falls,
Idaho 83415-4107
ABSTRACT
Gas samples from selected drums of radioactive waste at the U.S.
Department of Energy (DOE) Idaho National Engineering Laboratory are being
characterized for 29 volatile organic compounds (e.g., acetone, bromoform,
tetrachloroethylene) to determine the feasibility of storing the waste in
DOE's Waste Isolation Pilot Plant (WIPP) in Carlsbad, New Mexico. Quality
requirements for the gas chromatography and gas chromatography/mass
spectrometry chemical methods used to analyze the waste are specified in
the Quality Assurance Program Plan for the WIPP Experimental Waste
Characterization Program. This document was prepared by DOE with input
and review by U.S. Environmental Protection Agency. Quality requirements
consist of both objective criteria (i.e., data quality objectives, DQOs)
and statistical criteria (i.e., process control). The DQOs apply to
routine sample analyses, while the statistical criteria serve to determine
and monitor precision and accuracy (P&A) of the analysis methods and are
also used to assign upper confidence limits to measurement results close
to action levels.
After over two years and more than 1000 sample analyses there are two
general conclusions concerning the two approaches to quality control:
(1) Objective criteria (e.g., + 25% precision, + 30% accuracy) based on
customer needs and the usually prescribed criteria for similar EPA-
approved methods are consistently attained during routine analyses.
(2) Statistical criteria based on short term method performance are
almost an order of magnitude more stringent than objective criteria and
are difficult to satisfy following the same routine laboratory procedures
which satisfy the objective criteria. Statistical P&A criteria are
initially established from 30 replicate analyses of a standard sample over
a period of several days. System performance is then tested semi annually
by analysis of seven replicates whose results are compared to the results
of the initial 30 replicates. The inability to obtain statistically
equivalent data sets at the 95% confidence level for the majority of
analytes arises primarily from short term (i.e., few days to few weeks)
precision being more definitive than long term (i.e., few weeks to few
months) excersions in accuracy even though the accuracy is always well
within the DQOs.
A more cost effective and representative approach to establishing
statistical method performances criteria would be either to utilize a
moving average of P&A from control samples over a several month time
80
-------
period or to determine within sample variation by one-way analysis of
variance of several months replicate sample analysis results or both.
Confidence intervals for results near action levels could also be
determined by replicate analysis of the sample in question.
INTRODUCTION
Analytical chemistry laboratories conducting repeated chemical
measurements of environmental, industrial, and waste samples must utilize
standardized and formalized quality control (QC) and quality assurance
(QA) procedures to continuously control and evaluate the results from
their analysis activities. In most laboratories these QA/QC procedures
are derived from a combination of objective-based and statistically-based
criteria. Objective-based criteria are usually measurement control limits
that are designed to provide data of acceptable quality and agreed to by
both the laboratory and the customer prior to analyses being conducted.
These objective criteria are often referred to as data quality objectives
(DQOs) and are established considering two primary factors:
(1) Customer's intended use of the data
(2) Capabilities and limitations of the chemical analysis method
Statistically-based criteria are also employed as measurement control
limits, but compared to DQOs these criteria are less arbitrary and based
more on actual analysis method performance. The control limits in this
case are established and periodically revised from the measured
performance of the analysis system (e.g., repeated measurements of a
sample to establish precision control limits). This statistically-derived
form of QA/QC is also referred to as statistical process control (SPC).
For an analytical chemistry laboratory to consistently produce analysis
results which are of acceptable quality, strict adherence to standardized
QA/QC procedures is essential; however, cost effectiveness and quality
effectiveness of a laboratory's QA/QC procedures can vary substantially
depending on the specific manner in which the procedures are implemented
and maintained. At the U.S. Department of Energy (DOE) Idaho National
Engineering Laboratory (INEL) a program is underway to determine the
feasibility of storing radioactive waste in DOE's Waste Isolation Pilot
Plant (WIPP) in Carlsbad, New Mexico. One aspect of this program is to
chemically analyze gas samples taken from selected drums of radioactive
waste currently stored at INEL. The QA/QC requirements for these chemical
analyses are a combination of objective and statistical criteria and are
described in detail in the Quality Assurance Program Plan for the WIPP
Experimental Waste Characterization Program (USDOE, 1991). Volatile
organic compounds are one of the types of compounds being analyzed for,
and this paper summarizes and evaluates the quality effectiveness and cost
effectiveness of the QA/QC procedures for such analyses conducted between
February 1992 and March 1994.
81
-------
QUALITY REQUIREMENTS
The QA/QC requirements for volatile organic compound (VOC) analyses for
the WIPP Experimental Waste Characterization Program vary depending on the
chemical analysis methods, which for this investigation were as follows:
• WIPP Method 430.1: Modified Method TO-14 for the Determination
of Volatile Organic Compounds in Waste Container Headspace Using
SUMMA® Passivated Canister Sampling and Gas Chromatographic/Mass
Spectrometric Analysis
• WIPP Method 440.1: Gas Chromatography-Flame lonization Detector
Determination of Alcohols and Ketones in Waste Container
Headspace Collected Using SUMMA® Passivated Canisters
WIPP Method 430.1 is a gas chromatographic technique with mass
spectrometric detection (GC/MS) using mass flow controller for gas
injection. This method is derived from U.S. Environmental Protection
Agency (EPA) Ambient Air Method T014 (USEPA, 1988a and 1988b). WIPP
Method 440.1 is a gas chromatographic technique with flame ionization
detection (GC-FID) using a sample loop for direct gas injection. The
principal VOCs, which these methods are designed to identify and measure,
are listed in Table 1.
As previously stated the QC requirements are a combination of objective
criteria and statistical criteria. The objective-based QC apply to
routine sample analyses, while the statistically-based QC serves to
determine and monitor precision and accuracy (P&A) of the analysis methods
and is also used to assign upper confidence limits to measurement results
close to action levels.
Objective-Based Quality Control
The objective-based quality control procedures, which will also be
referred to as data quality objectives (DQOs), are primarily numeric
control limits for QC measurements. These QC measurements and associated
evaluations are required during all phases of analysis system calibration
and sample analyses and are summarized in Tables 2 and 3. The specific
numeric values of the control limits were chosen to satisfy DOE's data
usability needs and also considering the inherent capabilities and
limitations of the analysis techniques. The DQOs listed in Tables 2 and
3 are very similar in type and magnitude to those generally used for
similar published and commonly used analysis techniques, such as EPA
SW-846 Method 8240 (USEPA, 1990).
Brief descriptions of DQOs and explanations of how their indicators listed
in Tables 2 and 3 are calculated are provided in the following
subsections:
• Precision
• Accuracy
• Completeness
82
-------
Table 1. VOC analysis Target Compound List (TCL) and Program Required
Quantitation Limit (PRQL).
Volatile Organic Compounds
1. Acetone
2. Benzene
3. Bromoform
4. 1-Butanol
5. 2-Butanone
6. Carbon tetrachloride
7. Chlorobenzene
8. Chloroform
9. Cyclohexane
10. 1.1-Dichloroethane
11. 1,2-Dichloroethane
12. 1.1-Dichloroethene
13. cis-1.2-Dichloroethene
14. Ethyl benzene
15. Ethyl ether
16. Met Hanoi
17. Hethylene chloride
18. 4-Methyl-2-pentanone
19. 1.1.2.2-Tetrachloroethane
20. Tetrachloroethene
21. Toluene
22. 1.1.1-Trichloroethane
23. Trichloroethene
24. 1.1.2-Trichloro-lf2.2-trifluoroethane
25. 1.3.5-Trimethylbenzene
26. 1.2.4-Trjmethylbenzene
27. m-Xyleneb
28. o-Xylene
29. p-XyleneD
CAS
Number
67-64-1
71-43-2
75-25-2
71-36-3
78-93-3
56-23-5
108-90-7
67-66-3
110-82-7
75-34-3
107-06-2
75-35-4
156-59-2
100-41-4
60-29-7
67-56-1
75-09-2
108-10-1
79-34-5
127-18-4
108-88-3
71-55-6
79-01-6
76-13-1
108-67-8
95-63-6
108-38-3
95-47-6
106-42-3
PRQL8
(ppmv)
100
1
1
100
100
1
1
1
1
1
1
1
1
1
1
100
1
100
1
1
1
1
1
1
1
1
1
1
1
a Values based on delivering 10 mL to the analytical system.
b These xylene isomers cannot be resolved by the analytical methods
employed in this program.
83
-------
Table 2. Calibration requirements for VOC analyses.
Method
430.1
(GC/MS)
430.1
(GC/MS)
430.1
(GC/MS)
440.1
(GC-FID)
440.1
(GC-FID)
Procedure
Bromof 1 uoro-
benzene (BFB)
tune
5-pt initial
calibration
Continuing
calibration
3-pt initial
calibration
Continuing
calibration
Frequency of
Procedure
Every 12 hours
Initially and
as needed
Every 12 hours
Initially and
as needed
Every 12 hours
Acceptance Criteria
Within specified key
ion abundance ranges
Percent relative
standard deviation
(RSD) for all compounds
<35%
Relative percent
differences (RPD) for
all compounds within
30% of initial
calibration
RSD all compounds <30%
RPD for all compounds
within 30% of initial
calibration, retention
time within 3 standard
deviations of initial
calibration
Precision: Precision is a measurement of the random error in an
analytical measurement process (i.e., the degree of agreement between
independent measurements determined by the analysis of replicate samples).
When calculated for duplicate sample analyses, precision is expressed as
the relative percent difference (RPD), which is calculated as
RPD(%) =
S - D
where
S
D
(S + D)/2
x 100
= first sample value (original)
= second sample value (duplicate)
When precision is calculated for three or more replicate determinations,
the relative standard deviation (RSD), also known as the coefficient of
variation (CV) expressed in units of percentage, is used. This is an
expression of the spread of the data relative to the mean value, X, of the
determinations. The specific formulas used for calculation of the RSD are
84
-------
Table 3. VOC analysis data quality objectives.
Compound
Method 430.1 (GC/NS)
Benzene
Bromoform
Carbon tetrachloride
Chloro benzene
Chloroform
Cyclohexane
1.1-Dichloroethane
1 . 2-Dichloroethane
1 . 1-Dichloroethene
cis-l,2-Dichloroethene
Ethyl benzene
Ethyl ether
Methylene chloride
1,1,2. 2-Tetrachloroethane
Tetrach loroethene
Toluene
1,1, 1-Trichloroethane
Trichloroethene
1.1.2,-Trichloro-l,2,2-
trif luoroethane
1,3, 5-Trimethylbenzene
1.2, 4-Tr Imethy 1 benzene
ra-Xylene
o-Xylene
p-Xylene
Method 440.1 (GC-FID)
Acetone
1-Butanol
2-Butanone
Met Hanoi
4-Methyl-2-pentanone
Precision
(RSD or RPD)
<25%
<25%
Accuracy
Recovery
(R)
70-130%
70-130X
Detection
(HDL)
8 ng
or less
50 ng
or less
Completeness
(C)
90%
90%
RSD - Relative standard deviation
RPD = Relative percent difference
MDL = Method detection limit (total number of nanograms delivered to the
analytical system per sample)
85
-------
X =
n
E
i = 1
and
n
i = 1
(x,- - X)
n - 1
1/2
RSD(%) = CV =
x 100
where
xi
n
s
result value for the ith measurement
total number of measurements
standard deviation, expresses the variability of data
about the mean, X
Accuracy: Accuracy (bias) is a measurement of the extent to which a
measured value of a quantity (parameter or analyte) agrees with the
accepted value of that quantity. Accuracy is assessed by the analysis of
samples of known concentration (e.g., laboratory control samples,
calibration samples, field reference standards, or additional QC samples)
for the analyte of concern, or by spiking samples with a known quantity of
the analyte of concern before analysis. In both instances, accuracy is
quantified by calculating the percent recovery (R) of the known quantity
(true value, TV) of analyte. The general equation used for this
calculation is
Measured Value Background Value
True Value of Sample or Spike
x 100
Method detection limits (MDLs) are determined for each analyte for each
method. These MDLs are determined by (a) conducting replicate analyses of
standards at quantities approximately one to five times the estimated MDL,
(b) determining the standard deviation, s, of the replicate measurements,
and (c) calculating the MDL from
MDL = t
(n 1, 1 -
0.99)
where
n = the number of replicate analyses
-(n - 1, 1 - « = 0.99)
The t distribution value appropriate to a 99%
confidence level (one-tailed) and standard
deviation estimate with n - 1 degrees of freedom
86
-------
The MDL calculated in this manner represents the minimum amount of a
substance that can be measured and reported with 99% confidence that the
analyte quantity is greater than zero.
Completeness: Completeness (C) of the reported data (expressed as a
percentage) is calculated as
V
C(%) = — x 100
T
where
V = number of measurements judged to be valid (meets all QA/QC
requirements)
T = total number of measurements expected (based upon number of
samples submitted for analysis)
Statistically-Based Quality Control
A summary of the program's required approach to statistically evaluating
analysis method performance is shown in Figure 1. Initially, precision
and accuracy are assessed using analysis method performance data from
analysis of 30 laboratory reference standards (i.e., initial P&A data
set). Thereafter, precision and accuracy data for each analysis method is
continuously monitored and evaluated by analysis of at least seven
replicates every six months (i.e., continuing P&A data set). Corrective
actions must be taken if consecutive data sets are not statistically
equivalent.
Analysis of Precision: The F test is used to determine if the precisions
obtained from different data sets are statistically the same. Comparison
of data sets is done at the 95% confidence level (two-tailed). A
calculated F value is determined by taking the ratio of the two data set
variances:
where
v1,v2
- S'/S
v1 = number of degrees of freedom for data set 1
v2 = number of degrees of freedom for data set 2
s, = standard deviation of data set 1
s2 = standard deviation of data set 2
When calculating F values, s, and s2 are chosen so that s, is greater than
therefore, - 2/- 2
''/s2£ is always greater than one. The calculated F value
is then compared to the critical Fc value found in tables of critical F
values. The Fc value is at the 95% confidence level with the degrees of
freedom (v,v2) such that v1 is degrees of freedom for the numerator and v2
is the degrees of freedom of the denominator. The precisions obtained
from two data sets are statistically equivalent if the calculated F value
is less than the critical Fr value.
87
-------
Determine initial performance
by analyzing 30 replicates
Calculate precision
and accuracy
Initial performance
for assigning upper
confidence limits
Participate in program
for 6 months
Determine continuing performance
by analyzing 7 replicates
Calculate precision
and accuracy
No
Compare continuing
7 replicates performance data with
initial/pooled/last 7 replicates
performance data
This
is second
set of 7
replicates for
semiannual
evaluation?.
Does
precision o
accuracy of
data sets
differ?
Pool precision
and accuracy
data
corrective
Continuing performance
for assigning upper
confidence limits
Figure 1. Statistical analysis of method performance data.
88
-------
from two data sets are statistically equivalent if the calculated F value
is less than the critical Fc value.
Analysis of Accuracy: The accuracy obtained from two data sets is
evaluated at the 95% confidence level (two-tailed) by comparing the data
set averages using the t-statistic. If the standard deviations of two
data sets are statistically equivalent by the F-test, a pooled estimate of
the standard deviation is calculated as follows:
s2 = [(n, - l)Sl2 + (n2 - 1) s22]/(n1 + n2 - 2)
where
n, = number of samples in data set 1
n2 = number of samples in data set 2
The t value is given by
tv = (XT - X2)/[s(l/n, + l/n2)1/2]
where
v = number of degrees of freedom (n,, + n2 - 2)
X., = mean concentration of data set 1
X2 = mean concentration of data set 2
If the standard deviations of the two data sets are significantly
different, the t value is given by
/S 2
tv = (X1 - X2)/[— + —
V"1
with the degrees of freedom, v, given as
[ ("' *
s2 V 1
2
V ", n2 /
(s,2/n,)2
0,, + 1
(s22/n2)2
n2 + 1
- 2
For either case, the mean concentrations of the two data sets are not
significantly different if the t value obtained is less than the critical
value of tc. If the mean concentrations do not differ significantly, then
the accuracies for the two data sets do not differ.
89
-------
RESULTS AND DISCUSSION
During routine analysis of samples between February 1992 and March 1994,
associated quality control (QC) measurements were over 99% compliant with
program data quality objectives (DQOs). This near zero defects
performance was due primarily to two factors: (1) representativeness and
reasonableness of QC measurement acceptance criteria and (2) good
laboratory practice. The general manner in which analysis method QC
procedures are applied also contributes to avoiding DQO deficiencies. For
example, QC requirements for analysis method calibration must be satisfied
before conducting sample analyses, and this phased approach greatly
reduces the probability the noncompliant QC measurements as the analysis
process proceeds.
Based on laboratory control sample (LCS) analysis results during routine
sample analyses, average precision and accuracy (P&A) for the 29 compounds
of interest was relative standard deviation (RSD) precision of 7.5% (7.0%
for Method 430.1, 9.7% for Method 440.1) and recovery (R) accuracy of
97.8% (98.2% for Method 430.1, 97.8% for Method 440.1). These average
values are well within the corresponding program DQOs of 25% RSD and 70-
130% R. Even if individual compounds are considered, the compounds with
largest deviations from the ideal results (i.e., maximum RSD: 15.2% for 1-
butanol, minimum R: 85.3% for 1,3,5-trimethylbenzene, maximum R: 106.1%
for 1,2-dichloroethane) are still well within the program's DQOs.
The described sample analysis performance was achieved during normal
laboratory operations, and instances where work had to be repeated to
satisfy QC requirements were very infrequent (i.e., less than 3% of total
analysis time). Since analysis costs averaged $800 per sample, the cost
for rework (i.e., price of nonconformance) was about $24 per sample. For
the time period considered, close to 1,000 samples were analyzed, so total
analysis costs were about $800,000 of which the cost for rework to satisfy
QC requirements was at most $24,000.
In contrast to program DQOs being rather simply and inexpensively
satisfied through good laboratory practices, satisfaction of program
statistically-based QC requirements proved to be both difficult, if not
impossible, and expensive. Summaries of statistical test results are
provided in Tables 4 and 5.
As is the case for the LCS analysis results, average P&A results for the
statistical test results are all well within program DQOs. Additionally,
the statistical test results show:
(1) average precision for each data set is generally much better
(i.e., smaller RSD) than that compiled from all LCS analyses
(2) magnitude of dat set precisionis directly related to the length
of time for conducting data set analyses
(3) data set precision approaches overall LCS analysis precision for
data set analysis time intervals of about six months or more
90
-------
Table 4. Method 430.1 (GC/MS) statistical test results for VOC analyses.
CO
Statistical Test Results for Compounds
Data Set Replicate Average Average Data Set Statistically Equivalent (X)
Identification Analysis Time Period Analyses Precision Accuracy Compared To Precision &
Number
1
2
3
4
5
6
7
8
9
(dates) (days)
02/18/92
10/05/92
10/05/92
10/05/92
05/15/93
06/22/93
12/22/92
11/17/93
02/23/93
- 02/24/92
- 10/05/92
- 10/23/92
- 04/15/93
- 06/18/93
- 08/10/93
- 08/10/93
- 03/18/94
- 03/18/94
7
1
19
30
34
49
231
121
388
(number)
30
7
7
30
7
7
30
7
30
(X
3
2
RSD)
.62
.69
4.30
5
3
2
5
4
.54
.86
.03
.45
.92
5.55
(X R)
99.
105.
97.
99.
98.
99.
99.
92.
97.
2
6
2
8
7
5
6
4
1
(ID)
NA
1
2
NA
4
5
NA
7
NA
Precision Accuracy Accuracy
78 30 17
96 48 43
61 61 39
61 74 39
69 17 13
NA = Not applicable. New initial (i.e., 30 replicate analyses) data set.
-------
Table 5. Method 440.1 (6C/FID) statistical test results for VOC analyses.
CD
ro
Statistical Test Results for Coovounds
Data Set
Identification Analysis Tine Period
Nuifcer
1
2
3
4
5
6
7
8
9
(dates)
03/11/9Z
09/14/92
09/18/92
03/04/93
06/30/93
08/11/93
04/09/93
01/20/94
04/22/93
- 03/17/92
- 09/14/92
- 09/18/92
- 04/26/93
- 08/10/93
- 10/06/93
- 10/06/93
- 03/10/94
- 03/10/94
(days)
7
1
1
53
41
56
180
49
322
Rep 1 i cate Average
Analyses Precision
(number)
30
7
7
30
7
7
30
7
30
(X RSD)
1.40
0.91
0.92
7.25
6.13
4.45
7.80
3.21
11.86
Average Data Set Statistically Equivalent (XI
Accuracy Compared To Precision &
(XR)
99.5
101.7
101.4
94.1
84.1
90.1
86.7
96.4
89.8
(ID)
NA
1
2
NA
4
5
NA
7
NA
Precision Accuracy Accuracy
100 20 20
100 100 100
80 40 40
100 20 20
60 20 20
NA = Not applicable. New initial (i.e., 30 replicate analyses) data set.
-------
In five comparison between data sets, however, statistical equivalency
between two data sets at the 95% confidence level was never satisfied for
all 29 compounds. Examination of the statistical test results indicate
the inability to obtain statistical equivalent data sets was due mainly to
data set precision being more stringent than differences in accuracy
between data sets. Apparently, P&A data generated from replicate analyses
over short time intervals as specified by the program are not
representative of actual method performance during routine sample
analyses. This also means the accuracy of sample confidence limits
estimated from such P&A data is questionable.
In conducting the statistical tests, data sets 1-4 were compiled by
replicate reference standards analyses conducted independently of routine
sample analyses. This required at least 60 analyses with costs of about
$48,000 for analyses and $6,000 for data compilation and evaluation. For
the time period considered the statistical tests cost was about 15% of
total laboratory costs. In early 1993 it was concluded that data sets
analyzed over short period of time (i.e., few days to few weeks) were not
providing representative P&A data, so subsequent data sets were compiled
from LCSs analyzed during routine sample analyses. Using LCS analysis
results for the statistical tests greatly reduced costs for the tests to
less than 2% of total laboratory costs; however, the data sets were still
not statistically equivalent. Without incurring substantial additional
expense to implement extremely stringent QC procedures (i.e., about 10
times more restrictive than current DQOs), it is unlikely the statistical
tests as specified by the program could ever be satisfied and the data
representative of routine analysis method performance.
A more practical approach to obtaining representative P&A data would be a
moving average of LCS analysis results; this would also be the most cost
effective approach since LCS analyses are required as part of routine
sample analyses. Since the primary purpose of the P&A data is for
assigning upper confidence limits to selected sample analysis results,
replicate analyses of the sample in question would be a much more direct
and representative way to obtain that information. Another potential
technique for obtaining representative precision information is one-way
analysis of variance (one-way ANOVA). By subjecting all replicate sample
and replicate LCS analysis results to one-way ANOVA, representative
analysis precision would be provided by the one-way ANOVA error sum of
squares term (i.e., variation within individual samples).
CONCLUSIONS
Comparisons of objective-based and statistically-based quality control
(QC) measurement results from volatile organic compounds analyses of gases
revealed some distinct differences. In general for the two year time
period and the gas chromatography and gas chromatography/mass spectrometry
analysis techniques considered, objective-based QC was effective and
inexpensive while statistically-based QC was ineffective and expensive.
93
-------
More specifically, key observations and differences between the two
approaches are as follows:
Objective-Based Quality Control
• Effective - over 99% compliant with QC specifications
• Inexpensive - less than 3% of total analysis costs
• Representative - measurement QC limits consistent with analysis
method capabilities
• Workable - measurement QC limits easily and consistently satisfied
during routine laboratory operations
Statistically-Based Quality Control
• Ineffective never fully compliant with program QC requirements
• Expensive - up to 15% of total analysis costs
• Unrepresentative - short period (days) precision and accuracy (P&A)
data not representative of long period (months) analysis method
performance
• Unworkable statistical equivalency of P&A data between data sets
from replicate analyses not demonstrable at 95% confidence level
ACKNOWLEDGEMENTS
Work supported by the U.S. Department of Energy, Assistant Secretary for
Environmental Restoration and Waste Management, under DOE Idaho Operations
Office Contract DE-AC07-761D01570. The authors wish to thank A. D.
Chapman, B. C. Jensen, and B. E. Dates for assistance in analyzing samples
and F. A. Sabel for typing the manuscript.
REFERENCES
USDOE, 1991, Quality Assurance Program Plan for the Waste Isolation Pilot
Plant Experimental Waste Characterization Program, DOE/EM/48063-1, U.S.
Department of Energy, Washington, D.C., July 1990.
USEPA, 1988a, Compendium of Methods for the Determination of Toxic Organic
Compounds in Ambient Air, U.S. Environmental Protection Agency, Quality
Assurance Division, Environmental Monitoring Systems Laboratory, Research
Triangle Park, North Carolina, May 1988.
USEPA, 1988b, The Determination of Volatile Organic Compounds (VOCs) in
Ambient Air Using SUMMA Passivated Canister Sampling and Gas
Chromatographic Analysis, Compendium Method TO-14, U.S. Environmental
Protection Agency, Quality Assurance Division, Environmental Monitoring
Systems Laboratory, Research Triangle Park, North Carolina, May 1988.
USEPA, 1990, Test Methods for Evaluating Solid Waste, Physical/Chemical
Methods, SW-846, 3rd Edition, Final Update 1, U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response,
Washington, D.C., November 1990.
94
-------
11
A CASE STUDY COMPARISON OF A HTW DATA EVALUATION PROCEDURE
USING EPA SW-846 AND ARMY CORPS OF ENGINEERS ENGINEERING
PROTOCOL AND THE EPA CLP FUNCTIONAL GUIDELINES METHOD OF DATA
EVALUATION
Ajmal M. Ilias, Chemist and George J. Medina, Chemist, U.S. Army Corps of Engineers,
1491 N.W. Graham Ave. Troutdale, Oregon, 97060-9503
ABSTRACT
Historically, EPA functional guidelines have been used for data evaluation of
environmental studies with mixed reviews from the analyst and regulators. The CLP
methods and functional guidelines have been effective for certain matrices and projects;
however, it has come under scrutiny as being, excessively voluminous with
documentation and not cost effective. There is a need for a comprehensive, yet practical
data evaluation procedures, particularly for complex matrices, that is also cost effective.
Using EPA SW-846 methods and quality control protocols, the Corps of Engineers has
established a quality assurance program that is comprised of the evaluation often -
percent quality assurance splits by an impartial referee laboratory (inter-laboratory) for
the evaluation of precision, accuracy, reproducibility, comparability and completeness
(PARCC). A case study of a remedial action with specific EPA clean-up goals that uses
the CLP and SW-846 data evaluation procedures in conjunction with Corps of Engineers
guidelines, is presented and discussed.
INTRODUCTION
The contract laboratory program (CLP) and SW-846, for the majority of analytical
methods, have common functional data evaluation elements, such as dealing with
laboratory blanks, detection limits, surrogate recoveries, matrix spike (MS), matrix spike
duplicate (MSD) recoveries and relative percent differences (RPDs) obtained either from
duplicate analysis or from the MS and MSD recoveries. In spite of the above mentioned
controls, the following elements are not completely covered by either system; (a)
confidence, (b) reliability, (c) comparability and (d) completeness. Both systems have
stringent guidelines about data precision and accuracy but are monitored or approached
differently. CLP has established guidelines for accuracy, which is obtained from
surrogate and MS/MSD recoveries. Precision is obtained from the analysis of duplicate
samples or calculated from the MS and MSD recoveries. In the case of SW-846,
laboratories are suppose to establish their own acceptance criteria through a series of
experimental analysis for each method of interest and matrices. In the case of CLP,
accuracy and precision of the laboratories vary from laboratory to laboratory and the
method does not articulate specific evaluation criteria. In addition to differences in
precision and accuracy by both systems, there are quite a few variability in protocols for
the determination of detection limits, the number of targeted analytes covered, the
standards to be used and the execution of the initial and continuing calibration, holding
95
-------
times and sample preservation. Because of this ambiguity, there is a tendency, on the part
of the scientific community and environmental laboratory staff, to use the interchangeable
guidelines of CLP and EPA SW-846.
In the interest of expediency, only notable difference that impact data evaluation will be
addressed. The Army Corps of Engineers approach is to require the use EPA SW-846
Methods and criteria by laboratories performing analytical work. Ten percent of the
samples are pulled in triplicate, with blind duplicate splits (intra-laboratory), called
quality control (QC) samples being submitted to the primary laboratory and the
remaining split, called the quality assurance (QA) sample being sent to an impartial
referee laboratory (inter-laboratory). Data obtained in this manner is legally defensible
with a high degree of confidence derived from PARCC evaluation and assessment. Data
from eleven ground water studies from Ft. Lewis Logistics Center, Washington, are
chosen to demonstrate the effective use of data evaluation procedures using CLP protocol
and EPA SW-846 in conjunction with the Army Corps of Engineers QC and QA
approach.
EXPERIMENT AND DATA EVALUATION
An ongoing ground water monitoring study, Fort Lewis Logistics Center, Deep Aquifer
Study, is chosen to demonstrate the effectiveness of the two systems of data evaluation.
Eleven groundwater samples including one blind duplicate, three project trip blanks and
one project rinsate were collected by an architectural engineering firm (AE) for the
project laboratory. One QA sample with a trip and rinsate blank were collected by the
same AE firm at the same time for the QA laboratory to evaluate and compare against the
project blind duplicate data generated by the project laboratory, a sub-contractor of the
AE.
The analytical methods used are Methods For Chemical Analysis Of Water and Waste
(1), Test Method For Evaluating Solid Waste (2), and Method For The Determination Of
Organic Compounds In Drinking Water (3).
All samples were analyzed for volatile organics (VOC) by EPA Method 524.2 or 8260,
semi-volatile organics (BNA) by EPA Method 8270, chlorinated pesticides/PCB by EPA
Method 8080, twenty-three total and dissolved metals using EPA Method 6000/7000
series and cyanide by EPA Method 9010. Trip blanks were analyzed for VOC only. The
data evaluation procedures used CLP functional guidelines and SW-846 method required
guidelines along with Corps of Engineers inter and intra data comparison procedures,
which are detailed in Table 1. Inter and intra data comparisons of VOC in Table 2 and
total metals in Table 3 are presented. The data comparisons of other methods are not
presented since no disagreements between the data of the two laboratories were found.
Because of limited space, data of trip and rinsate blanks are not shown but are discussed
later in the data evaluation comparison in Table 1. The Corps of Engineers collects and
analyzes ten percent of all samples or at least one duplicate or sequential samples,
whichever is greater for each matrix, for the comparison and evaluation of the project
96
-------
laboratory's data. In this case study, the QA data were evaluated by the Corps of Engineers,
similar to project data evaluation presented in the left column of Table I. After evaluation,
the QA data were compared with project blind duplicate data. Tables 2 and 3 evaluate and
compare the QA and the project data to assess comparability, completeness and confidence.
97
-------
Table 1. Comparison of Data Evaluation Procedures
Evaluated by Corps of
Engineers
QC Items Covered
a. Surrogates:
Use of SW-846 Method
required QC and Corps of
Engineers inter and intra
laboratory data comparisons
approach.
Surrogate recoveries in SNA
and pesticides/PCB were within
laboratory established (LE) or
method required QC limits and
were accepted. The surrogate
recoveries in VOC samples LC-
35D,-41D,-41F,-21C,-41E,-
67D, -26D, -35DR, -21ER, LC-
26DMSD, -4ID and -4IF (after
dilution) were above LE or
method required QC limits. Of
these out of control recoveries,
the data of re-run samples LC-
21CR, LC-41ER and LC-67DR
were acceptable as one of the
two surrogate were marginally
out side the QC limits and are
acceptable. The surrogate
recoveries in all trip, rinsate and
laboratory blanks of EPA
Method 524.1 were within
method required QC limits and
are acceptable.
Evaluated by Architectural
Engineering Firm
EPA CLP Statement of Work for
organic (No. OLM01) and inorganics
multimedia document and (OLM02.1)
1991.
All field BNA and associated control
samples of surrogate recoveries were
within special QC criteria. No
surrogate recoveries of chlorinated
pesticides/PCBs were mentioned.
Surrogate recoveries for volatile
organic analysis exceeded criteria in
samples LC-35-D, LC-41D, LC-41F,
LC-21C, LC-41E, LC-67D and LC-
26D. Re-analysis of the samples
revealed the same high surrogate
recoveries exceeding criteria.
Samples LC-21C and LC-41E were
re-analyzed at a 1:10 and 1:20
dilution, respectively, due to high
analyte concentrations. Surrogate
recoveries of these dilutions were also
high. Sample LC-26D had high
surrogate recovery but was not re-
analyzed in accordance with the CLP
protocol. All surrogate results with
recoveries out of specification were
qualified as estimated (J). LC-21C re-
analyses results were not qualified as
only one surrogate slightly exceeded
the QC limit.
98
-------
b. Matrix Spike
and Matrix
Spike Duplicates
(MSP) Recoveries:
Evaluated by Corps of
Engineers
Six out often MS and MSD of
Method 524.1 were above upper
QC limits. VOC data are
questionable due to surrogate,
MS and MSD recovery failures.
Three out of twenty-two MS
and MSD of Method 8270 were
above QC limits; data were not
affected as no targeted analytes,
with the exception of (bis 2-
Ethylhexyl) phthalate, detected
in one sample. MS and MSD of
pesticides/PCB's were within
QC limits and are acceptable.
MS recoveries of about one-half
of the metals were outside QC
limits. Data are accepted based
on acceptable post-digestion
and laboratory control
recoveries.
99
Evaluated by Architectural
Engineering Firm
Sample LC-26D was analyzed as a
volatile organic compound MS/MSD.
Six out often spike recoveries
exceeded the method QC limits. All
percent RPDs were within specified
criteria except for chlorobenzene.
Based on the result of MS/MSD and
surrogate analyses, field sample with
surrogate recoveries out of
specification were qualified as
estimated (J). All volatile organics
analyses are acceptable. Three out of
twenty-two BNA spike recoveries
exceeded method QC limits. The N-
nitro di-n-propylamine recovery was
fourteen percent, but was outside the
advisory QC limits. Field sample data
was not qualified on the basis of
MS/MSD criteria in accordance with
the functional guidelines. All BNA
data are acceptable. Pesticides and
PCBs MS/MSD results met all QC
criteria and indicated acceptable
precision and accuracy. Samples LC-
26D and T-9E were analyzed as total
metals matrix spike samples. Samples
LC-41E and T-9E were analyzed as
dissolved metals matrix spikes. LC-
35D and LC-47D were analyzed as
total and dissolved mercury matrix
spikes, respectively. QC criteria for
metals is 75-125 percent. In
accordance with data validation
guidelines, arsenic, beryllium,
cadmium, antimony, manganese and
vanadium results greater than the IDL
for total metals were assigned
estimated (J) qualifiers. Thallium
total metals results were qualified as
usable (R) for all non-detected
values. For the dissolved metals,
cadmium, thallium and antimony non-
-------
detected values were qualified as
unusable (R) due to low recoveries.
Manganese sample results below the
IDL and positive detects were
qualified as estimated (UJ) for both
total and dissolved samples. Mercury
dissolved and total analytical results
below the IDL were also qualified as
estimated (UJ). Cyanide percent
recoveries were not reported.
100
-------
c. Laboratory
Duplicate Analysis:
d. Blind
Duplicates:
e. Laboratory
Blanks:
Evaluated by Corps of
Engineers
The relative percent differences
(RPDs) of VOCs, BNAs and
pesticides/PCB's were within
QC limits except one out of five
VOCs and six out often BNAs
which were above QC limits.
BNA data were not affected as
no targeted analytes except
phthalates were detected. The
RPD of beryllium, barium, iron,
zinc, aluminum, antimony,
cadmium and copper ranged
from 31 through 200-percent,
indicating large analytical
variation. Data of these metals
should be considered estimates.
Blind duplicate data are shown
in Tables 2 and 3. All data
agree except for methylene
chloride and toluene in Table 2,
due to varying degree of
laboratory contamination.
Blind duplicate data of other
methods are shown in Corps of
Engineers unpublished chemical
quality assurance report (4). All
blind duplicate results were
found within a factor of three to
each other and were considered
comparable.
VOC laboratory blanks were
contaminated with methylene
chloride, toluene and BNA.
Blanks were contaminated with
1-methylbenzene. Data of these
analytes are not applicable for
site evaluation. Laboratory
blanks of all other methods
were free of targeted analytes.
101
Evaluated by Architectural
Engineering Firm
RPDs of organics were discussed
along with MS/MSD recoveries.
Inorganic duplicates analyses
discussed as follows. Duplicate
sample analyses was performed on
samples LC-26D (total) and T-9E
(total and dissolved). Applying the
control limits of+/- 20 percent relative
percent difference (RPD) for sample
5XCRDL, zinc, cadmium and iron
results in sample T-9E (total) were
outside criteria. An estimated
qualifier J was assigned to zinc,
cadmium and iron results for the total
metals only.
Blind duplicate results were not
presented.
Methylene chloride and toluene were
detected in VOC blank. Other
laboratory blanks were free from
targeted analytes.
-------
Evaluated by Corps of
Engineers
Evaluated by Architectural
Engineering Firm
f. Trip Blanks:
g-
Rinsate Blanks:
Up to ten VOC analytes,
ranging in concentration from
detection through 3.24 ppb were
detected.
Seven VOCs, ranging in
concentration from 0.9 through
2.04 ppb were found. 30 ppb of
methylbenzene was found in the
BNA rinsate and up to 858 ppm
of four alkali and alkaline earth
metals were found in the filtered
and unfiltered rinsates. Data of
these analytes should be
considered with caution.
h. Holding Times: Discussed
I. Detection limits,
tuning and mass
calibration
j. Initial and
Continuing
Calibration:
k. ICP Interference
Check:
All met method requirements
and are acceptable
Not discussed in this project.
Inadvertently left out.
Not discussed
1. Overall
Evaluation of the
Laboratory's Data:
Project data were accepted
except for the following.
I. Ten VOC analytes found in
the trip blanks are identical in
102
Not discussed/Not included with data
evaluation.
Not discussed/Not included in the data
evaluation.
Discussed
Not included in data evaluation
package or not discussed
Discussed in detail for VOC and BNA
methods. No discussion of initial or
continuing calibrations were found for
chlorinated pesticides/PCB, metals or
cyanide.
Interference check sampler (ICS) were
run at the beginning and end of each
sample analysis. Sodium and
potassium analysis did not include an
ICS at the end of the sample run. All
reported recoveries were within 20 %
of the true value. All laboratory
control sample results were within 80-
120 % recovery.
No data were rejected. All data were
used with some sort of qualifier.
-------
concentration to that found in
most samples except for
trichloroethene (TCE) and cis-
dichloroethene (cis-DCE) in
samples LC-41D and LC-41F.
The analytes found in the trip
blanks should be treated with
caution.
II. Seven VOC analytes were
found in the rinsate blanks; data
of these analytes should be
viewed with caution.
III. The data of methylene
chloride and methylbenzene
should be discarded due to their
presence in all laboratory
blanks.
IV. Certain metals were found
at higher levels in the filtered
samples than the unfiltered
samples. Recommend
rechecking for possible switch
between filtered and unfiltered
samples.
V. Data of beryllium, barium,
iron, zinc, aluminum, antimony,
cadmium and copper should be
considered estimates due to the
high degree of analytical
variability.
103
-------
Table 2: COMPARISON OF PROJECT BLIND DUPLICATE AND QA RESULTS
Project: Ft. Lewis Logistics Center. Deep Aquifer Study Matrix: Water Prefix: LC-41
Project Laboratory: Hittman Ebasco QA Laboratory: CAS. Inc.
Method: Volatile Organics fEPA 524.2^ Units: ug/L (ppb)
Project Lab
Analytes -D -F Detection
Detected Initial/Re-analvsis Initial/Re-analvsis Limits
Methylene Chloride 2.28
cis-l,2-Dichloroethene 22.4
Benzene 0.4 J
Trichloroethene 137
Toluene ND
Ethylbenzene ND
Total Xylenes ND
1,1,1-Trichloroethane ND
Trans- 1,2-
Dichloroethene ND
Chloroform ND
13.8
21.4
ND
155
1.89
0.51 J
0.58 J
ND
ND
ND
2.13
19.8
ND
119
0.24
ND
ND
0.68
ND
ND
27.2
20.0
ND
145
ND
ND
ND
ND
ND
ND
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
0.10-2.00
QALab
Detection
-F Limits
ND
16
ND
130
ND
0.2
ND
0.6
0.2
0.1
2.0
0.1
0.1
0.1
0.1
0.1
0.3
0.1
0.1
0.1
Tentatively Identified Compounds
Hexane 0.4 ND 50 ND ND
Methylcyclopentane ND ND 8 ND ND
J = Estimated concentration
ND = None detected
SUMMARY: The project blind duplicate and QA data agree for all targeted analytes
except for methylene chloride, toluene and 1,1,1-trichloroethane. Data discrepancies for
methylene chloride and toluene are probably due to laboratory contamination of the project
laboratory. The 1,1,1-trichloroethane discrepancy is due to analytical variations of the
project laboratory, as it was found in one out of four trials at close to that reported by the
QA laboratory. Data comparisons at close to or below detection limits are not significant.
104
-------
Table 3: COMPARISON OF PROJECT BLIND DUPLICATE AND QA RESULTS
Project: Ft. Lewis Logistics Center. Deep Aquifer Study Matrix: Water Prefix: LC-41
Project Laboratory: Hittman Rhascn QA Laboratory: CAS. Inc.
Method: Dissolved Metals (EPA 6000/7000 Series^
Analytes
Detected
Arsenic
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Potassium
Selenium
Silver
Sodium
Thallium
Vanadium
Zinc
Project
-D
ND
ND
1.3
19.4
ND
ND
9500
ND
ND
ND
ND
ND
5010
ND
ND
ND
978
ND
ND
9050
ND
ND
27.9
Lab
-F
ND
ND
1.3
17.8
ND
ND
9480
ND
ND
ND
ND
ND
4970
ND
ND
ND
1170
ND
ND
ND
ND
ND
15.6
Detection
Limits
97.0
3.0
~
—
0.2
0.1
—
2.2
8.7
11.0
13.0
3.0
~
3.4
0.2
22.0
~
3.3
2.8
~
1.0
4.0
~
QALab
-F
80
ND
ND
ND
ND
0.7
11,900
ND
ND
ND
140
4
6240
6
ND
ND
ND
ND
ND
5900
ND
ND
140
Detection
Limits
50
3
1
5
0.2
0.1
50
5
10
10
20
2
10
5
0.5
20
2000
5
10
100
5
4
10
— = Not reported
ND = None detected
SUMMARY: The project blind duplicate and QA data agree within a factor of three to
each other or their detection limits except for zinc. A higher zinc level was found in the QA
filtered samples than the unfiltered samples, where no zinc was detected at 10 ppb. This
anomaly could be due, in part, to field filtration procedures, sample bottles used or a sample
switch that may have occurred during sample labeling or loading into auto-sampler in the
laboratory. Iron data discrepancies could not be resolved analytically. Both laboratories
had acceptable internal QC data
105
-------
RESULTS AND DISCUSSIONS:
An advantage to the CLP data evaluation procedure is the thoroughness of the review and
validation of data by going back to the raw data. The CLP evaluates, at a minimum, ten
percent of all raw data. Recalculation of data serves as a check to assure data validity.
Unlike the Corps of Engineers, the CLP procedure evaluates the initial and continuing
calibration of GC/MS analyses and reviews and confirms that GC/MS tuning parameters
are within the specified acceptance criteria. Similarly, the CLP validates that matrix
interference checks have been used when EPA Method 6010 metals are reported, where the
Corps of Engineers does not. This effort on the part of CLP assures that false positives, due
to positive matrix interference, are not reported.
The Corps of Engineers QA report preparation comprises review and evaluation of all
internal QC data. There is no re-calculation of raw data since contractually three levels of
review are required. Both the CLP and SW-846 Corps of Engineers analytical programs
emphasize demonstration of precision and accuracy. However, it is the opinion of the
authors that reproducibility, comparability and completeness is not emphasized sufficiently
by the CLP program. In addition to the mandated demonstration of accuracy and precision,
the Corps of Engineers requires ten percent inter and intra laboratory data comparison. This
requirement is key to meeting the extended requirements of aforementioned PARCC.
The trip and rinsate blanks data checks for potential cross contamination during sample
shipment/storage and collection, respectively. Up to ten VOC and about one and a half
dozen of the targeted analytes were detected in the trip and rinsate blanks (this data is not
provided in this paper). Because the CLP data evaluation process does not address
trip/rinsate blanks sufficiently, erroneous false positives can skew the data assessment.
Unaccountable cross contamination can impact engineering decisions in a negative way.
Data reproducibiliry and comparability are shown in Tables 2 and 3. In about ten detected
analytes of VOCs (Table 2), data of methylene chloride, toluene and 1,1,1-Trichloroethane
did not agree due, in part, to the project laboratory's varying degree of laboratory cross
contamination and variation of data close to the detection limits. All twenty-three metals in
Table 3 agree within a factor of three to each other or their detection limits except for the
data of inorganic zinc. These data disagreements or anomalies could be due, in part, to
cross contamination encountered in the field, sample switching (total/dissolved), sample
mis-labeling, or incomplete filtration of dissolved samples. Two out of three project and
QA samples were compared to the data of the dissolved samples portion present in this
table. CLP data evaluation procedure often do not account for these types of anomalies.
Overall, Corps of Engineers data evaluation coupled with internal QC elements of SW-846
provide data that would meet PARCC requirements.
ACKNOWLEDGMENTS:
The authors are grateful to Mr. Timothy Seeman, Director, U.S. Army Corps of Engineers,
North Pacific Division Laboratory.
106
-------
REFERENCES:
1. EPA-600/4-79-020, Methods for Chemical Analysis of Water and Wastes, March 1983.
2. EPA SW-846, Third Edition, Test Methods for Evaluating Solid Waste, November,
1986.
3. EPA-600/4-88-039, Methods of the Determination of Organic Compounds in Drinking
Water, December, 1988.
4. Ft. Lewis Logistic Center, Deep Aquifer Study; Chemical Quality Assurance Report, No.
90-HM-232, 20 November 91; North Pacific Division Laboratory, Troutdale, Oregon,
107
-------
12
Contract Management Strategies for Overseeing Laboratory Analysis - Special Analytical
Services
Sean J. Kolb, DynCorp Viar, Alexandria, Virginia 22314
ABSTRACT
The Special Analytical Services (SAS) program has provided CLP data users access to non-
routine analytical services since 1982. These services are based on technical requirements
developed by the data user and have been obtained by subcontracting with commercial
environmental testing laboratories through the US EPA's Sample Management Office contractor.
In several instances, .the subcontractor community may be unable to provide the services as
requested by the Regional client. These situations can lead to unusable data that can cause
delays in Superfund site remediation activities. Since 1990, numerous modifications have been
made to the contract management system used by the SMO contractor in managing these
activities. This paper will outline the overall SAS program and will examine six separate case
studies of highly visible SAS projects. These case studies will include a summary of the scope
of each project and the issues involved. The issues in these case studies include the data users
need for fast turnarounds and/or low detection limits, method/laboratory performance problems,
complex sample matrices, and data reporting requirements for the projects. These case studies
will also review the problem-solving approaches used by the Region, SMO contractors and
laboratory subcontractor for addressing these issues, and how the results of these projects have
been applied to the continuous improvement of the SAS program. These results of these case
studies show that diere are common characteristics and strategies for successfully managing and
obtaining these types of services.
108
-------
13
CREATION OF A SITE-SPECIFIC SOIL LABORATORY CONTROL SAMPLE
FOR THE IDAHO NATIONAL ENGINEERING LABORATORY
S. J. Sailer. J. M. Connolly, T. R. Meachum, EG&G Idaho, Inc., Idaho National Engineering
Laboratory, P.O. Box 1625, Idaho Falls, Idaho 83415; R. B. Chessmore, RUST Geotech, Inc.,
P.O. Box 14000, Grand Junction, Colorado 81502-5504
ABSTRACT
Natural soil matrix quality control (QC) materials are needed to evaluate analytical method
performance in support of Environmental Management (EM) projects. Commercially available
materials are usually synthetically prepared and may not adequately mimic the specific
chemical and physical characteristics of actual field samples. The use of site-specific natural
soils as QC materials ensures that analytical method performance will be comparable between
the QC materials and the site's routine field samples. The Idaho National Engineering
Laboratory (INEL) has developed a site-specific soil material for use as a laboratory control
sample (LCS) which will be provided to laboratories supporting EM projects as a quality
improvement tool.
The LCS was created from residual samples of radioactively contaminated soil collected as part
of characterization activities at the INEL during 1990. Data from the original analyses of the
residual samples provided baseline information to develop a formulation scheme for the LCS.
Certain metal analyte concentrations were enhanced to desired levels through the addition of
natural mineral source materials.
This paper discusses the protocols used for the preparation, verification of homogeneity , and
characterization of this material. Characterization of the material was limited to the 23 metal
target analytes of the Contract Laboratory Program (CLP) and 10 radionuclides historically
detected at the INEL. Metal analyte concentrations were characterized using the CLP methods
that are routinely used on INEL field samples. Radionuclide characterization utilized total
dissolution procedures with the determinative method selected by the participating laboratories.
Analyte concentration control limits for the LCS material were statistically derived from the
characterization data, and will be updated as the material is used. Results of the
characterization and control limit determination are presented, along with lessons learned
during the preparation and characterization processes.
109
-------
14
AUDIT STANDARDS FOR FIELD SAMPLING AND FIELD MEASUREMENT
Bob Newberry and Dave Bottrell, Office of Environmental Restoration and Waste Management,
EM-563
Michael Johnson, United States Department of Energy, Environmental Measurements
Laboratory, New York, NY 10014.
ABSTRACT
A well established and supported audit program is critical to the success of the
characterization, remediation and post-closure monitoring activities at DOE facilities. The
Office of Environmental Restoration and Waste Management (EM), Laboratory Management
Division (EM-563), is responsible for assuring that all EM operations are effectively performing
the environmental sampling and field measurement services required. In support of this goal
EM-563, along with the Department of Energy's, Environmental Measurements Laboratory,
(EML), has developed a series of audit standards. These standards provide detailed questions
related to various Quality Assurance (QA), operational and technical aspects of the operation
of many of the most commonly used field sampling and field measurement equipment and
procedures. In support of these audit standards these is a companion guide is being
developed that explains the technical reason for the question, the intended response and
typical problems and limitations associated with using the specific field sampling or field
measurement devise.
The standards cover field sampling for liquid, soil, sediment, sludge, air, surface and
vegetation matrices. The standards address devises such as COLIWASAs, direct emersion
samplers, lysimeters, scoops, augers, corers, triers, split spoon samplers, various types of
dredges, filters, impingers, canisters, traps and various types of equipment for soil gas
sampling. The standards for field measurement cover the same matrices as the sampling
equipment and address devises such as oxygen meters, portable combustible gas indicators,
portable Flame lonization Detectors (FIDs), portable Photoionization Detectors (PIDs),
calorimetric indicator tubes, electrodes, enzyme test kits, portable Dissolved Oxygen (D.O.)
meters, spectrophotometers and X-ray fluorescence equipment. Also covered are field
radiation measurements using devises such as proportional counters, G-M counters,
semiconductor detectors, solid scintillation and liquid scintillation detectors.
These audit standards are being designed for assessments of DOE/EM contractors. They can
also serve as guidance for an internal assessment program at any facility performing field
sampling and field measurement activities.
110
-------
15
QUALITY ASSURANCE FOR THE EPA REGION V ESAT
FIELD ANALYTICAL SUPPORT PROGRAM
Lewis Kranz and Dennis Miller, Lockheed Environmental Services
Assistance Team, Region V, Chicago, IL 60605 and Jay Thakkar,
USEPA, Regional Project Officer, CRL Region V, Chicago, IL
60605
Abstract
To achieve the goal of properly documenting the quality of the
analysis data generated, an effective mobile laboratory
program for on-site field analyses requires a quality
assurance/quality control program containing not only standard
operating procedures (SOPs) for sample preparation and
analysis, but also for sample tracking, data management, data
validation, data communication and other logistical support.
The QA/QC program must be both flexible and stringent to
accommodate variable reporting timeframes for appropriately
reviewed data for both short term (one to two weeks) projects
and long term (three to ten weeks) projects involving analyses
of generally site specific target compounds (i.e. VOCs, PAHs,
PCBs, pesticides, etc.) in multiple matrices. The QA/QC
program must include uniform procedures and checklists easily
utilized by different analysts performing different analyses.
Environmental Services Assistance Team (ESAT) chemists utilize
EPA Region V Central Regional Laboratory mobile laboratory
equipment and instrumentation to support field screening
analysis needs of Superfund sites. A quality assurance program
has been developed and implemented which effectively documents
the complete field laboratory process tracking from sample
receipt to the approval of the final deliverable. The overall
sophisticated QA process allows for stepwise checks and
balances through interlocking SOPs encompassing sample
tracking, extraction, analysis, data handling, rapid field
data review, timely field data reporting and final QC data
approval. An examination of the ESAT field QA preparation,
review and approval processes demonstrates that field
screening analyses in the laboratory can provide data
equivalent to that of a static laboratory. Effective
utilization of definitive operational procedures and
appropriate quality control measures for an on-site mobile
laboratory program can provide the associated documentation of
laboratory processes necessary to ensure the defensibility of
field screening data for litigation purposes.
111
-------
Introduction
Lockheed utilizes the EPA Region V ESAT Field Analytical
Support Program (FASP) mobile laboratories to provide on-site,
real-time analysis of soils, sediments, soil gas, waters, oils
and wipes for selected organic and inorganic analytes. Field
operations primarily support Superfund activities and may
include field screening analyses for site characterization,
cleanup, monitoring and spill response. Sample analyses are
conducted using modified static laboratory protocols which can
accommodate turnaround of analytical results within 24 hours
of sample receipt. FASP analyses are conducted within a
quality assurance program which maintains quality control
through a progression of checks and balances administered
throughout the field operations. The chain-of-custody
established during field sampling is maintained by in-
laboratory chain-of-custody processes throughout the entire
analytical sequence from sample receipt to the submission of
hardcopy data package to the data user.
The mobile laboratories are maintained in a "road-ready"
status and have the capability to rapidly move on-site and be
utilized for multiple analyte analysis of a large number of
samples in either a short or extended timeframe. The field
screening routinely analyses routinely performed in the mobile
laboratories include volatile organic compounds (VOCs),
polycyclic aromatic hydrocarbons (PAHs), polychlorinated
biphenyls (PCBs) and selected metals by x-ray fluorescence
(XRF). The site specific target compounds or elements of
interest and the required detection limits are determined
prior to mobilization to the site. Table 1 gives a
representative list of sites where FASP mobile laboratories
have been utilized for field screening analyses.
Program Development
The FASP laboratories have primarily been utilized at
Superfund sites but the analysis services are also available
to NPDES and RCRA projects.
1. Laboratory Personnel
FASP mobile laboratory personnel are degreed chemists
experienced in sample preparation and GC or GC/MS analysis.
The FASP chemists perform routine "CLP type" sample analyses
at the Region V Central Regional Laboratory and
review/validate CLP and PRP data packages when they are not in
the field. When new technology, new instrumentation or new
analysis procedures are incorporated into existing FASP SOPs,
the field chemists are thoroughly trained and must demonstrate
proficiency before being allowed to utilize the "new methods"
112
-------
for sample analysis. Training and proficiency evaluation are
appropriately documented. Immunoassay test kit use is being
incorporated into the PAH and PCB analysis SOPs. Two FASP
chemists were trained and certified by Ensys to utilize their
immunoassay test kits for SW-846 field screening methods 4020
and 4035 for PCBs and PAHs, respectively
A minimum of 2 chemists (for safety considerations) are used
to staff the laboratory during any effort, but 3 chemists may
be required when multiple analyses are being conducted.
In addition to the laboratory personnel, an available in-
office member of FASP is made the Project Lead for any site
the FASP laboratory is servicing. This Project Lead provides
an in-office contact, coordinates and organizes the analytical
data as received from the FASP laboratory into a final report
and ensures proper document archiving.
Additionally, the ESAT QA/QC Coordinator is involved in the
preplanning of the site project and serves as a check and
balance in the review/approval of individual data sets and
approval/release of the hardcopy site data set package. To aid
in the planning, a questionnaire (see Figure 1) requesting
basic information of the site, the analytes of interest,
report deliverables and the intended usage of FASP is supplied
to ESAT in preparation of on-site operations.
2. Standard Operating Procedures (SOPs)
FASP SOPs have been written for both sample analysis and data
quality control (data reporting, logbooks) procedures. A
typical analysis SOP includes sample preparation procedures,
calibration and analysis procedures, and quality assurance/
quality control (QA/QC) requirements for sample preparation
and analysis. Operational procedures identify chain-of-custody
policies, sample tracking requirements, report deliverables
and data handling protocols. The FASP SOPs currently in use
are listed in Table 2.
A FASP mobile laboratory is normally utilized to
simultaneously perform field screening analysis for more than
one analyte type and thus organizational systems are required
to direct specific samples for the proper analyte analysis and
report the corresponding sample data in both a daily and
comprehensive final manner. Therefore, suitable QA/QC steps
with realistic turnaround times have been incorporated into a
complete monitoring process which progresses from sample
receipt through calibration and analysis to data reporting and
data package review.
To achieve this goal, field samples received under chain-of-
113
-------
custody by the mobile laboratory are logged into a
computerized program designed to track individual samples as
they proceed progressively through the preparation, analysis
and reporting. Table 3 is an example of sample tracking at a
recent site where VOA analyses were performed. The samples are
entered by their assigned sample ID number along with the date
of receipt, analysis parameter and matrix type. Additionally,
samples requiring extraction are entered into a site specific
extraction logbook and are assigned an unique GC autosampler
vial number.
All samples, whether requiring extraction or direct analysis,
are analyzed using the sample ID being entered into the
instrument run log. The instrument run log is retained in the
site logbook and is utilized as a monitoring tool for
documentation of the sample trail.
An overall data reporting system was developed using daily
data file names for sample tracking and incorporates
information derived from the sample receipt date, sample site
and analytical parameter. This system is utilized to segregate
and compile the individual daily data set packages. The daily
list of samples has the corresponding file name assigned
permits sorting and listing by ascending sample ID order. This
list provides a convenient method for compiling, assembling
and reviewing the sample data for subsequent levels of QC
review. These data set packages are normally released to the
data user in the field within twenty-four hours after sample
receipt. This tracking system is updated daily and allows for
a check and balance monitoring of incoming samples and
outgoing analysis results to ensure that received samples are
analyzed and reported within the targeted turnaround time.
Using the established tracking systems, a reviewer can quickly
inspect the documentation to ascertain the exact status of any
samples received, where a sample is in the analytical process,
what level of QC review has been conducted and when/if the
data set package has been reported.
3. Deliverables
Due to the timely reporting of the FASP sample analysis data,
the FASP reporting deliverables are tailored to efficiently
present the analytical results without raw data. The
analytical data is summarized in a data set report which
consists of a narrative identifying samples, describing
analytical problems or anomalies, listing matrix spike
recoveries and includes a discussion of the calibrations, the
associated blank results and the analytical results listed in
tabular form.
114
-------
The daily analytical data packages are compiled and assembled
in the following order: the narrative describing analytical
problems or anomalies associated with the samples and/or QC,
the blank results, the sample results, QC data package
checklist and matrix spike recoveries accompanied by a
completed chain-of-custody transfer form.
The final FASP deliverable, prior to dissemination of any
information, is reviewed and scrutinized by the two field
chemists and the QA/QC Coordinator. This review is aided by
the use of a checklist (Figure 2) which identifies the major
elements of the data set report and the specific areas of QC
focus. Completion of the QC checklist provides documentation
that a three tier check encompassing all the pertinent data
information has been completed and assembled. The QC checklist
is utilized in the data report assembly and data review
process, but it is not part of the final data set report.
4. QA/QC Review Process
The samples, calibrations, blanks and spikes are analyzed
during the course of a daily analytical run and the
corresponding data set is compiled and the appropriate data
reporting forms are completed. A narrative detailing the
specifics of the analytical results and associated QC is
produced on-site by the chemist (designated as QC1) conducting
the analysis. The data set package is assembled in the order
specified in the appropriate SOP by QC1 chemist using the QC
checklist. The entire data set package then undergoes an in-
field review by another on-site chemist (QC2). The data set
package is either (1) approved by QC2 and the narrative is
signed by QC2 and it is then released to the data user as a
daily field report or (2) the data set package is returned to
QC1 for corrections and the process is repeated until an
acceptable data report is signed by QC2. The original field
daily report is placed in an individual folder with all the
associated supporting raw data as a unique data set package.
The daily data set packages generated during field operations
are returned to the ESAT home office for a final review by the
QA/QC coordinator (QC3) . After all of the daily data set
packages are approved by the QA/QC Coordinator, each data set
package is collated in chronological order to prepare the
final data set site report. A case narrative summarizing all
of the analytical problems or anomalies associated with the
site samples and/or QC is written and the final site report
and a chain-of-custody transfer form bearing the appropriate
release approvals are delivered to the Lockheed ESAT Team
Manager for his review and approval. Flowchart 1 illustrates
the routing of samples and analysis under chain-of-custody
procedure. The final site report is produced within five days
of the conclusion of FASP on-site support activities.
115
-------
5. Archiving
The raw data and supporting documentation for a FASP operation
is archived by site name and date of activity- The hardcopy
data set packages are packed into file boxes in chronological
order with a copy of the final report and all computer files
generated during the effort. In the instances where multiple
analyses were conducted during the effort, the data set
packages are grouped by analysis type and arranged in
chronological order.
6. Audits
The ESAT QA/QC Coordinator has conducted several on-site
audits of the mobile laboratory during field operations. The
audits have been identified QA/QC areas requiring improvements
or modification. The audits also helped highlight deficiencies
in extraction and analysis procedures which are being
addressed through method development.
The use of PE samples as QC blinds is also utilized to assess
the accuracy of the FASP field screening methods. In other
cases, representative field samples or extracts were brought
back to the CRL for GC/MS analysis to more appropriately
qualify field screening results. Modifications are made to
analysis procedures when severe matrix problems arise. For
example, most Continental Steel soil samples contained
significant amounts of petroleum type hydrocarbons which
complicated PAH analysis using a GC/FID system. To minimize
the number of false positive analytical results, a dual column
confirmation procedure was implemented for the dirty samples
and a larger extract volume was employed. The SOP changes were
documented and appropriately higher detection limits were
utilized for reporting non-detects.
7. Method development
FASP flexibility and adaptability has been demonstrated at
sites requiring the measurement of analytes not found on the
SOP target compound list. For our last field project,
precision, accuracy and method detection limit studies were
performed for acetone, tetrahydrofuran and methyl ethyl ketone
so that these compounds could be incorporated into the site-
specific target compound list. Chromatographic parameters were
optimized to achieve adequate compound resolution and
calibration levels were identified to meet site measurement
needs. The generic FASP VOC analysis SOP modifications were
documented and the on-site field screening analyses were made
using the modified SOPs. Empore disc solid phase extraction
(SPE) procedures are being developed and validated for water
samples containing PAHs and PCBs because the FASP
116
-------
liquid/liquid extraction procedures are cumbersome, time
consuming and generally have lower analyte recoveries than the
SPE procedures.
Conclusions
The sample results reported in FASP data packages generated
from on-site field analysis are reviewed 100%. The data review
process entails three levels of data scrutiny, approval and a
positive release system for the final reported results. Chain-
of- custody is maintained throughout the operation from sample
receipt to final data set package delivery. The level of
documentation of the entire analytical and reporting process
coupled with the levels of review assures the capacity of
legally defensible data.
117
-------
TABLE 1
SITES WHERE FASP HAS BEEN UTILIZED
SITE
LOCATION
ANALYTES
Scrap Processing
Evergreen Manor
Medford, Wisconsin VOAs, PAHs, PCBs, Pb
Roscoe, Illinois
Rockford Groundwater Rockford, Illinois
Chlorinated VOAs in soil gas and
groundwater
Chlorinated VOAs in groundwater
Allied Paper
Conrail Rail Yard
Prestolite Battery
Continental Steel
Circle Smelting
Kalamazoo, Michigan PCBs in soil & paper waste
Elkhart, Indiana
Vincennes, Indiana
Kokomo, Indiana
Beckmeyer, Illinois
Galen Myers Dump Osceola, Indiana
Maumee River Basin Toledo, Ohio
Brooklyn Park Dump Brooklyn Park, MN
Tomah Armory
Tomah Fairgrounds
Belvidere Landfill
Tomah, Wisconsin
Belvidere, Illinois
Chlorinated VOAs
Chlorinated VOAs
VOAs, PAHs, PCBs
Pb, Cu, Zn and Fe in sediments
by field portable XRF
VOAs, PAHs
VOAs, PAHs, PCBs
Geoprobe collection of soil
screening samples
Geoprobe collection of
groundwater screening samples
Geoprobe collection of
groundwater screening samples
118
-------
TABLE 2
ESAT-5-095.0
ESAT REGION V STANDARD OPERATING PROCEDURES LIST
Updated May 11, 1994
SOP NUMBER
022
023
024
025
026
027
030
SOP NUMBER
021
008
012
016
017
028
029
031
038
SOP NAME
FSOILGAS-02
FSP-CAL-00
FSP-PAH-03
FASP-PCB-03
FSP-VOA-03
FSP-XRF-00
FSPPESTC-00
SOP NAME
FSLOGBOK-00
CASENARR-00
CUSTRANS-00
DATASET-00
DISTSLOG-00
FSPCHLST-00
FASP-HND-00
GEN-APP-00
LABAUDIT-00
METHOD SOP TITLES
FASP SOIL GAS METHOD
FASP CALIBRATION STANDARDS PREPARATION
FASP POLYAROMATIC HYDROCARBON METHOD
FASP POLYCHLORINATED BIPHENYLS METHOD
FASP VOLATILE ORGANIC ANALYSIS METHOD
FASP X-RAY FLUORESCENCE ANALYSIS METHOD
FASP CHLORINATED PESTICIDE ANALYSIS METHOD
QA ADMINSTRATION SOP TITLES
FASP LOGBOOK MAINTENANCE
CASE NARRATIVE GENERATION PROCESS
DATA SET CUSTODY (COC) TRANSFER FORM
DATA SET PACKAGE ASSEMBLY
SOP DISTRIBUTION RECORDKEEPING LOG
FASP DATA SET CHECKLIST
FASP SAMPLE AND DATA SET PACKAGE HANDLING
SOP GENERATION AND APPROVAL PROCESS
LABORATORY ANAYLSIS AUDITS
119
-------
TABLE 3
LOCKHEED FASP MOBILE LABORATORY
DOUGLAS ROAD LANDFILL, IN
APRIL 21, 1994
RECEIPT
DATE
04/11/94
04/11/94
04/11/94
04/11/94
04/11/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/12/94
04/13/94
04/13/94
04/13/94
04/13/94
04/13/94
04/13/94
04/13/94
04/13/94
04/13/94
04/13/94
04/14/94
04/18/94
04/18/94
04/18/94
04/18/94
04/18/94
04/18/94
04/18/94
04/18/94
04/18/94
SAMPLE ID
DRLGG0820
DRLGG0835
DRLGG0850
DRLRW0100
DRLRW0200
DRLGG0624
DRLGG0624D
DRLRW0400
DRLGG0724
DRLRW0300
DRLRW0500
DRLRW0700
DRLGG0221
DRLRW0321
DRLRW0421
DRLRW0521
DRLRW0600
DRLGG0118
DRLGG0121
DRLGG0921
DRLGG1021
DRLGG0118D
DRLMWE011
DRLGG1224
DRLGG1321
DRLGG1421
DRLRW0502
DRLGG1521
DRLGG1624
DRLGG1624D
DRLGG1724
DRLGG0236
DRLGG0251
DRLGG2518
DRLGG2624
DRLGG1918
TARGET ANALYTE
VOA MATRIX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
WATER
VOA DATA QC2
REPORT FILE CHECK
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V11
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V11
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V12
DRLVOA.V13
DRLVOA.V13
DRLVOA.V13
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
DRLVOA.V21
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
120
-------
Requestor
FIGURE 1
ESAT/FASP USAGE QUESTIONNAIRE
Mailing Address or Mail Code
Phone Number
Projected date(s) when services are needed
(3-weeks is the maximum amount of time for a single project)
Name of site
Location of site (State and nearest city)
Brief site history (contaminants & concentrations)
What are your data quality objectives?
Soil
Matrix(ces) and Estimated Water
Sample Number Sediment
Target analytes
Requested turn-around time for: field analysis results
final deliverables at project completion
Form of final deliverables
Will CLP/CRL confirmation be required?
If yes, at what frequency?
Do you require technical assistance in other areas (SAS analyses, data validation)?
Briefly describe
Please return completed form to: Jay Thakkar
ESAT Regional Project Officer
Central Regional Laboratory
536 South Clark Street
Chicago, Illinois 60605
Mail Code: SL-10C
121
-------
FIGURE 2
Page 1 of 2 ESAT-5-055.0
FA8P ANALYTICAL CHECKLIST FOR DATA SET PACKAGES
FASP Method: TID #
Site Name: Charge Number
Analyst/Date (QC1):
Task Lead/Date (QC2):
QA Personnel/Date (QC3):
Approvals
QC1 QC2 QC3
MATRIX (CONTROL)/MATRIX SPIKE DUPLICATE
1. The associated samples are properly listed
2. The header form is correct
3. Sample QC data matches the quantitation reports
4. Target analyte and concentrations are listed
on MS/MSD form
5. MS/MSD recoveries and RPDs are correctly
reported
II METHOD BLANK
1. The Method Blank Summary is correct _
2. The associated samples are properly listed _
3. The header form is correct ~
III INITIAL CALIBRATION
1. The initial calibration is present _
2. The initial calibration is within method ~~
criteria
3. External calibration checks are reported ~~
IV CONTINUING CALIBRATION
1. Continuing calibration ran at proper intervals _
2. Each continuing calibration is listed on a form ~
3. The concentration and analytes are listed on the~
continuing calibration form
4. The %D's are reported ~~
5. Continuing calibrations that do not meet criteria"
are discussed in the case narrative
V FINAL CALIBRATION
1. Results of the final calibration are listed on
a form
2. The %D's are reported
3. If final calibration does not meet criteria
samples are "J" flagged and outlier is
discussed in the case narrative
122
-------
FIGURE 2
Page 2 of 2 ESAT-5-055.0
FASP ANALYTICAL CHECKLIST FOR DATA SET PACKAGES
Approvals
QC1 QC2 QC3
VI SAMPLE RESULTS
1. A report form is present for all samples/blanks
2. "D" flags and appropriate dilution factors
are present for all diluted samples
3. All reported results are properly rounded
4. All MDLs are properly listed
5. The "F" flag is reported for all samples/blanks
VII QUANTITATION REPORTS
1. Quant reports/chromatograms for all standards
2. Quant reports/chromatograms are present for all
samples, blanks and spikes
3. Samples are quanted after calibrations
VIII MISCELLANEOUS
1. Copies of all extraction records are present
2. Copies of all run logs are present
IX CASE NARRATIVE
(Form ESAT-5-007.0 - Case Narrative)
1. Method, Date, Author are on narrative
2. The narrative is consistent with the facts
3. The narrative has correct spelling
4. The rhetoric of the narrative is correct
X DATA CUSTODY
1. The data set has been placed in storage
XI COMMENTS
Data Package approved for release: Yes No (Circle) Date
Returned for correction of deficiencies: Yes No (circle) Date
123
-------
FLOWCHART 1
SAMPLE/DATA ROUTING CHAIN-OF-COSTODY FLOWCHART
Step 1. Samples received
with traffic report.
under cbain-of-cu«tody
Originating analyst
makes corrections
T
No
Step 6. Does data set package
require corrections.
-> Step 2. Samples are entered
into the SI/DO systc
Step 5. Review of data set <—
package by 2nd analyst.
-> Step 3. Lab analyses
of samples
Step 4. Compile daily data set
package with QC.
Yes
I
Step 7. Draft data set report
delivered to the customer
in the field
Step 8. FASP coordinator receives
reviewed data set package.
-> Step 9. QA/QC coordinator
reviews data set package.
Originating analyst
makes corrections
Yes
Does data set package
require corrections
Step 12. Group Lead archives all ^
data and information
generated during the project.
Step 11. Group Lead generates the ^—
site Analytical Data Report
as the final deliverable.
No
4
Step 10. Group Lead receives
data set package.
-------
16
The Quantification of Data Quality with Explicit Examples
from Organic, Inorganic and Radiochemical Methods
A.D. Sauter, PAR Enterprises, Inc., 217 Garfield Drive,
Henderson, Nevada 89014
ABSTRACT
Knowledge of the quality of sets of environmental data is essential
for interpreting test results. While there are those that claim to hold
the holy grail of data quality calculations, we show in this paper
that data quality calculations can not be described by any one
approach. Rather, we demonstrate that environmental data quality
must be calculated using different metrics to provide the necessary
alternate perspectives on various data features. Using GC, GC/MS,
ICP and Gamma Ray Spectrometry environmental data, we show
how alternate data quality calculations compliment each other and
how they provide insight on given test sets. We also demonstrate
how our approach whose essence arises, from exploratory data
analysis techniques can be utilized to train junior personnel and
others to easily provide environmental data quality insight with off-
the-shelf software products.
125
-------
17
RAPID SITE ASSESSMENT USING THE QTM SERVICE
A QUALITY ASSURANCE PERSPECTIVE
M. Catherine Eldridae, Environmental Chemist/Analyst, Bruce E. Lane,
Technical Analyst, and Loren E. Minnich, Technical Manager/Analyst,
DynCorp Viar, Inc., 300 North Lee Street, Alexandria, Virginia 22314-
2695
ABSTRACT
In the initial phase of site assessment, it is crucial to identify any
further action that may be required to mitigate threats to the
environment and to human health. Loss of time and effort due to
questionable analytical results may have potentially serious legal and
financial consequences for the regulatory and/or enforcement
communities. A service which promotes efficiency in collecting samples,
laboratory analysis, obtaining data, and making decisions while assuring
high levels of quality is essential. The Quick Turnaround Method (QTM)
Service provides this necessary degree of quality assurance (QA) while
fulfilling the requirements for rapid analysis, reporting, and decision
making.
The QTM Service is a powerful tool in screening, monitoring, and
performing hazardous waste site assessment activities, providing full
organic analyses of twenty volatile organic compounds, fifteen
polynuclear aromatic hydrocarbons, fourteen phenolic compounds, nineteen
organochlorine pesticides, seven aroclors and toxaphene. Under routine
conditions, laboratories providing analytical services for the QTM
Service are required to analyze and report analytical results, in an
electronic format, within 48 hours (depending upon batch size) from the
time of sample receipt. Delivery of the complete, Contract Laboratory
Program (CLP) format hardcopy data packages is accomplished within seven
(7) days.
To ensure that the laboratories participating in the program are able to
meet the requirements of the QTM Service, each of the laboratories
successfully performed a stringent prequalification phase. This phase
of QA consisted of the analysis of multiple performance evaluation
samples by the laboratories and the demonstration of their capabilities
through the completion of an on-site laboratory evaluation. When
quality laboratory services are linked with the QTM analytical methods,
which are designed to permit effective and rapid extractions and
instrumental analyses with rigorous quality control (QC) requirements,
the user is provided with analytical results of high quality.
Additionally, when integrated with the QTM Software systems,
incorporating several automated data review functions, the user is able
to receive that data, with a variety of validation reports, in a very
short period of time.
126
-------
Through providing analytical results of known and documented quality
within 48 to 72 hours (depending upon batch size) with rigorous levels
of QA/QC, the QTM Service is a useful tool within the Superfund Program.
As the analytical results are available to users rapidly, and are of
high quality, the results can be applied to direct sampling efforts
and/or be used to monitor the effectiveness of treatability operations,
cleanups, and removals with a high level of confidence and efficiency.
INTRODUCTION
In an effort to increase the productivity of the Superfund Program, the
United States Environmental Protection Agency (EPA) initiated the
Superfund Accelerated Cleanup Model (SACM). SACM was conceived to make
hazardous waste cleanups more timely and efficient by focusing on the
initial phases of the process, including rapid site assessment,
remediation, and removal. As a result, services that support SACM must
provide a faster and more efficient process for obtaining defensible,
high quality analytical data to support decisions during all phases of
the SACM process.
In response to the SACM initiative, the Quick Turnaround Method (QTM)
Service was designed to meet a specialized need for keeping site work
underway during site characterization studies, treatability studies, and
site cleanup. The QTM Service was developed to provide analyses of
large numbers of samples in short periods of time. This was made
possible by streamlining analytical methods to enable a laboratory to
perform many analyses in a rapid turnaround period. Additionally, the
QTM Service was created to report analytical results electronically and
incorporate automated data review and report generation features to
increase the speed of data reporting and validation.
To compliment the efficiency of the QTM analytical procedures and
reporting functions, stringent quality control (QC) and quality
assurance (QA) activities were implemented. Rigorous QC measures were
incorporated into the analytical methods to ensure that the results
being reported by the laboratories were of the highest technical
quality. The data reporting software applied an automated QA module
which aided the laboratory and the user in identifying possible
contractual and/or performance issues. Additionally, to assist the user
with the data validation and decision-making processes, the software
included an automated data validation module.
Once the initial versions of the QTM analytical procedures and software
were completed, a pilot program was implemented to serve as a working
prototype to evaluate the service. QA of a similar scope to the
analytical methods was included in the pilot program to ensure the use
of highly qualified laboratories. This function consisted of each QTM
candidate laboratory completing multiple performance evaluation sample
analyses and an on-site laboratory evaluation. After the placement of
contract awards, data deliverables submitted by the laboratories were
127
-------
routinely assessed for contract compliance and overall laboratory
performance throughout the lifetime of the pilot program.
The implementation of the pilot program provided the EPA with a
mechanism which would satisfy the Superfund Program's immediate needs
for a fast turnaround analytical screening service that included
considerable QC and QA measures in the same timeframe. Additionally,
the pilot program gave the EPA the opportunity to evaluate and modify
the analytical requirements and software systems and ensure that the
service continued to provide data of known and documented quality to the
various EPA clients.
DynCorp Viar, Inc. worked closely with the EPA on the QTM pilot program
by dedicating a team experienced in laboratory procurement, software
development, and analytical chemistry. Our role in the pilot program
included procuring laboratory services, assisting in the development of
automated systems, and monitoring laboratory performance throughout its
duration.
THE PILOT PROGRAM
LABORATORY SERVICES
With the initiation of SACM, the U.S. EPA Analytical Operations Branch
(EPA AOB) conducted a survey of the EPA Regions to determine the need
for a CLP analytical screening service to support the Superfund
Program's remedial, removal, and site assessment activities. Based on
the results, EPA AOB and the EPA Regions began the development of quick
turnaround methods and procedures that would satisfy their needs.
During the summer of 1991, the first procurement of analytical
laboratory services for the QTM Service pilot program began. The scope
of the project offered the EPA Regions full organic analysis for a
variety of matrices, including water, soil/solid, oil/oily, and wipe
samples, within 48 hours (depending upon batch size) from the time of
sample receipt at the laboratory. Under routine conditions, the
analytical results would be transmitted directly from the laboratory to
the EPA user in an electronic format with complete CLP hardcopy data
deliverables being submitted within seven (7) days.
In September 1991, the pilot program was implemented with two
laboratories providing a total capacity of up to 9600 fractional
analyses over an eighteen month period. The results of the pilot
program enabled EPA AOB to test the analytical methods with true
environmental samples and identify whether these methods would be
adequate to meet EPA Regional needs. Technical comparisons between the
QTM results and those of similar EPA methods used for other projects
within the CLP provided EPA AOB with additional method performance and
comparison information useful in evaluating whether the technical
requirements would be achieved. The QTM Regional and Laboratory
Software were also enhanced to include valuable QA procedures. The
software modules were designed to assemble the analytical results in the
128
-------
appropriate CLP format, enabling the laboratory to check the data for
noncompliance, and reprocess the data, if necessary, prior to submission
to the EPA mainframe system. Software was also developed to process the
results through an automated data review program which enabled EPA
clients to obtain customized, validated reports directly from the
laboratory results.
In September 1992, the scope of the QTM pilot program was expanded to
enable the EPA AOB to refine the analytical methods, complete testing of
software systems, and stress-test both aspects to determine any weak
points in the QTM Service. To facilitate this expansion and ensure
continued QA, the following processes were developed:
Specific analytical and data deliverable criteria were developed to
assess the abilities of potential QTM laboratories prior to field sample
analysis. This included the incorporation of a two-step laboratory
prequalification period designed to assess a laboratory's ability to
meet not only the analytical requirements of the project, but also their
ability to deliver compliant electronic and hardcopy data deliverables
within the timeframe required. Three laboratories participated in this
validation process. In Part 1 of this process, the laboratories
completed the analysis of laboratory-spiked samples within an extended
(14 day) period of time. This set of analyses was intended to gauge
each laboratory's ability to perform and meet all of the technical
requirements of the analytical methods. Upon completion of those
analyses and obtaining feedback regarding their performance, the
laboratories began the analysis of similar samples under real-time
conditions (Part 2). This phase was intended to measure each
laboratory's ability to meet both the analytical and electronic data
transfer requirements of the QTM Service.
The analytical data submitted by the laboratories for Part 1 underwent a
detailed technical review process. The review, performed by a QA
chemist dedicated to the QTM pilot program, consisted of a manual review
of all data reporting forms and associated raw data for accuracy,
consistency, and verification of calculations. Chromatogram data were
checked for compliance with analytical and contractual requirements.
The results were reported to each laboratory prior to preceding with
Step 2 of the validation phase so that procedural adjustments could be
made accordingly.
Upon each laboratory's completion of Part 2, the results were reviewed
to determine whether each laboratory successfully transmitted the
electronic data deliverables within the required timeframe and if the
analytical results were compliant with all contractual and technical
requirements of the project. As the analyses for Part 2 were performed
on samples similar to those employed for Part 1, and to accelerate the
prequalification phase, a streamlined data review process was used,
focusing primarily on the issues raised previously and the results of
the Contract Compliance Screening (CCS) module of the QTM Software.
Again, the results were provided to each laboratory prior to the
129
-------
performance of on-site laboratory evaluations so that all issues could
be resolved at that time.
The final step in the QA process was to verify the analytical
capabilities and instrumentation of each of the candidate QTM
laboratories. To achieve this, a detailed on-site evaluation was
performed. Through completion of the on-site evaluations, it was
confirmed that each of the laboratories had the capabilities, analytical
instrumentation, computer hardware, and expertise required to meet all
the analytical and data deliverable requirements and successfully
participate in the QTM pilot program.
Through the two-step analytical prequalification phase and the on-site
evaluations, each of the three candidate laboratories successfully
demonstrated their ability to adequately perform the analyses as
required for the QTM Service and were subcontracted to supply the EPA
with the necessary laboratory capacity.
In June 1993, after the completion of the procurement process, the QTM
pilot program was expanded with three laboratories providing analytical
services for a total capacity of up to 7500 fractional analyses over a
one (1) year period. Field samples from a number of EPA Regions were
scheduled with each of the laboratories and the Service continued
routinely. To ensure that overall laboratory performance remained at
the high levels expected within the CLP, on-going technical data review
activities were completed. This form of QA monitoring assisted the EPA
in gauging not only laboratory performance but method and software
performance as well. The data obtained from the reviews was also used
to further advance the method and electronic facets of the Service.
ANALYTICAL METHODS
The QTM Service offers full organic analyses using gas chromatography
(GC) for the following fractions: volatile organic compounds (VOAs)
using heated headspace sampler instrumentation with GC/photoionization
detector (PID) and GC/electrolytic conductivity detector (ELCD) ,-
polynuclear aromatic hydrocarbons (PAHs) using solvent or solid phase
extraction (SPE), with GC/flame ionization detector (FID); phenols
(PHENs) using solvent or SPE, with GC/FID or GC/PID; and organochlorine
pesticides (PESTs) and polychlorinated biphenyls (PCBs) using solvent or
SPE, with GC/electron capture detector (BCD)1. Table 1 provides an
overview of the QTM analytical methods.
In order for the instrumental analyses to be rapid and effective, a 24-
hour analytical sequence is used. It consists of two (2) types: an
initial calibration analytical sequence and a daily calibration check
analytical sequence. The analytical sequences consist of the following
requirements:
130
-------
Table 1.
QTM Analytical Methods Overview
Fractions :
No . Compounds : '
CRQLs (unadjusted) :'
- water (pig/L)
- soil/solid (/zg/kg)
oil2 (/ig/L)
(miscible)
oil2 (pig/kg)
(non-miscible)
Sample Size:
water (mL)
- soil/solid (g)
- oil2 (/zL)
(miscible)
- oil2 (g)
(non-miscible)
Sampling Volume:
- water samples
soil/solid samples
high concentration
(water or solid)
VOA
21
20
40
400
400
2
1
100
6
40 mL
120 mL
6 oz
PAH
16
20
330
--
20,000
100
6
--
1
1/2 gal
6 oz
6 oz
PHEN
16
50
830
--
830
100
6
--
6
1/2 gal
6 oz
6 oz
PEST
20
0.1
1.7
--
100
100
6
--
1
1/2 gal
6 oz
6 oz
PCB
8
1-5
17-83
--
1000-
5000
100
6
--
1
1/2 gal
6 oz
6 oz
Target Compound Lists and Contract Required Quantitation Limits
(CRQLs) listed in the draft QTM SOW, 2/93.
Miscible: methanol miscible/aqueous miscible oil samples.
Nonmiscible: methanol non-miscible and methanol miscible/aqueous
non-miscible oil samples.
131
-------
Initial Calibration Analytical Sequence1
Three-point Initial Calibration (high standard proceeded by
mid and low standards);
Instrument Blank;
Laboratory Control Sample(s) ;
Method Blank (s) ,-
Field Sample(s);
Instrument Blank(s); and
Performance Verification Standard(s).
Daily Calibration Analytical Sequence
Instrument Blank;
Calibration Check Standard (mid standard),-
Method Blank(s);
Laboratory Control Sample(s);
Field Sample(s);
Instrument Blank(s); and
Performance Verification Standard(s).
The QC requirements of the QTM analytical methods are structured to
provide consistent results of known and documented quality. Data
reviewers are able to determine the quality of data, as well as the
applicability to each sampling activity. The minimum QC requirements of
the QTM Service are outlined in Table 2.
ELECTRONIC DATA TRANSFER CAPABILITY
In order to provide real-time computer-reviewed and computer-validated
electronic summary data directly to the EPA Regions, a QTM electronic
data delivery and reporting system is used, consisting of two
components; the QTM Laboratory Software and the QTM Regional Software.
Data are entered into the QTM Laboratory Software, checked for format,
contractual requirements, data reviewed and evaluated, and transmitted.
Data receipt is accomplished by downloading the validated electronic
summary data from a user-secured mailbox on the EPA mainframe using the
QTM Regional Software . This secured mailbox is necessary to ensure
that the integrity of the data retrieved by the user is not compromised.
In addition to downloading data, the QTM Regional Software can produce
reports that generate the analytical data package, automated data review
results, CCS results, and other information that are helpful in the data
review process at the EPA Regional level. Figure 1 illustrates the flow
of electronic data deliverables from the QTM laboratory to the EPA user.
Under the program, the laboratories are required to transmit data to the
EPA Regions in electronic format 48-hours after the validated time of
sample receipt (VTSR) of the last sample in the batch. (If more than
three fractional analyses were requested, data must be transmitted
within 72-hours after VTSR of the last sample in the batch.) By using
132
-------
Table 2. Minimum Quality Control Requirements of the QTM Analytical Methods
QC
Initial Calibration
Calibration Check
Standard
Method Blank
Laboratory Control
Sample
Instrument Blank
Performance
Verification
Standard
System Monitor
Compound
Frequency
Initial 3 -point and
when CCS or PVS
criteria not met
Start of 24 -hour
sequence
Per matrix
Per matrix
At least two per
sequence
At least one per
sequence
Spiked into each
field sample
Purpose
Analysis of three different concentration standards at
the beginning of the sequence and whenever QC criteria
not met. The initial calibration assesses the
instrument's linearity, generates calibration factors
for sample quantitation, and establishes target
compound identification windows.
Analysis of the calibration check standard (CCS) is to
determine whether the initial calibration is still
valid, to ensure linearity of the curve, and to
evaluate the system for retention time shifting.
Analysis of the method blank is to evaluate the level
of laboratory background contamination. The method
blank is subjected to the entire analytical procedure.
Analysis of the Laboratory Control Sample (LCS) is to
demonstrate that the analytical system is in control.
The LCS is subjected to the entire analytical procedure
to assess the instrument's capability of producing
consistent data.
Analysis of the instrument blank is to evaluate
instrument cross -contamination.
Analysis of the Performance Verification Standard (PVS)
is to ensure that instrument stability and sensitivity
was maintained during sample analyses .
System Monitor Compound (SMC) is added to every field
sample and QC sample to assess analytical efficiency by
evaluating target compound retention time shift and
recoveries .
CO
CO
-------
Analysis of QTM Sample(s)
I
cc
o
DO
Lab QTM Software
Data Entry of Analytical Results
Structural Checker
Contract Compliance Screening
Computer-Assisted Data Review and Evaluation
Upload
EPA Mainframe
Region QTM Software
Download
Generation of Reports:
cc
LU
Q
111
Spreadsheet
Criteria/Outlier
Narrative
QC Forms (II-VI)
Forms I
Surrogate Recovery Summary
Contract Compliance Screening
Determination of Data Useability
Figure 1. QTM Electronic Data Deliverable Process
134
-------
the QTM Laboratory Software, the laboratory can transmit the data
quickly and efficiently. The QTM Laboratory Software contains the
following five (5) components2:
Component I. Data Entry Software allows entry of required analytical
data and performance information. In order to further advance the
efficiency of the electronic reporting, the laboratories are encouraged
to develop in-house software that can produce data in the required
format directly from their instrumentation.
Component II. Structural Checker ensures that the final data product
submitted by the laboratory can be loaded into the CLP Analytical
Results Database (CARD) and be adequately evaluated by other automated
systems such as CCS and Computer-Assisted Data Review and Evaluation
(CADRE). This ensures that certain variables are in the correct
positions and in the required format. If errors are encountered, an
error report is generated, the transmission process is stopped, and the
laboratory must make the appropriate corrections and reprocess the data.
This cycle continues until the data passes this phase of the
transmission process, thereby ensuring that the data being delivered is
structurally and contractually compliant.
Component III. CADRE occurs at the laboratory level prior to
submission. CADRE is based on the Draft of the National Functional
Guidelines for QTM Data Review3, dated May 1993, which is concerned with
QC performance regarding aspects that are within the laboratory's
control. The QC criteria used to evaluate the data includes: initial
calibration; calibration check standard; performance verification
standard; laboratory control sample,- system monitor compound; laboratory
blanks; analytical sequence; and quantitative results verification.
CADRE is blind to the laboratory, but the results are saved and
transmitted to the EPA Region along with the required electronic data
package information. This feature saves data processing time at the EPA
Region.
Component IV. CCS ensures that data analysis procedures occurred
according to the specifications of the contract. This feature produces
a report for the laboratory so that they can correct any identified
defects prior to data submission to the EPA mainframe.
Component V. Data Transmission is the final data processing phase.
This part of the software sends a copy of the laboratory's analytical
results data package plus CADRE results to the EPA mainframe via
standard telephone lines. As soon as the data transmission process is
complete, the data are available for EPA user access.
Once the data are uploaded to the EPA mainframe by the QTM laboratory,
the EPA user, with a valid system password, can retrieve the analytical
results to an IBM AT, 386 or compatible personal computer through the
use of the QTM Regional Software. The software also features view or
hardcopy options for generating QTM electronically-produced reports4.
135
-------
These reports include:
Data Spreadsheet. A CADRE-generated report that includes a compilation
of sample result information based on Form QIs with CADRE-assigned data
review final qualifiers. '. '
Criteria/Outlier Report. A CADRE-generated report that includes analyte
concentration, CADRE-assigned data review qualifiers, qualification
parameters, and associated criteria. This report can be used in
assisting with site decisions or determining samples for reanalysis.
Batch Narrative. A report detailing documentation provided by the
laboratory of any QC, sample shipment, and/or analytical problems
encountered in processing and analyzing the samples reported in the data
package. The laboratory also lists the samples analyzed and describes
the associated sample matrices.
Form QIs. The data reporting forms that include tabulated analytical
results and retention time information.
Quality Control Forms. The data reporting forms (II through VI) that
include tabulated analytical results of the initial calibration,
calibration check standard, LCS, and PVS, and a summary of the date and
time of analyses.
Contract Compliance Screening Results. A report that identifies
contract compliance discrepancies by fraction, field sample, code, and
criterion.
Surrogate Recovery Summary Report. A report that includes a summary of
the System Monitor Compound (SMC) recoveries for all field and QC
samples.
When these electronic data review processes and automated reports are
used in conjunction with the hardcopy data deliverables (including mass
spectra and other raw data) and manual data review and validation
processes, the user is provided with an efficient and comprehensive
approach to ensuring a high level of confidence and usability of the
data generated under the QTM Service.
STATUS/SUMMARY
Over the past several years the QTM Service has continued to develop and
expand, adding more laboratory capacity and EPA Regional users. To
date, four (4) subcontracted environmental laboratories and five (5) EPA
Regions have participated in recruitment and analysis of approximately
5,500 samples within the QTM Service. Multiple special method studies
and software development initiatives have been completed, each with the
intent of making a stronger, more efficient program. At the foundation
of all the development and enhancements of the QTM Service has been the
focus on the production of high quality, defensible data. This function
136
-------
has been aided, in large part, by the depth of QC in the analytical
methods, the QA mechanisms within the software systems, and the overall
mission of quality throughout the QTM Service.
REFERENCES
1 U.S. EPA/Analytical Operations Branch. February 1993 (Draft).
USEPA Contract Laboratory Program Draft Statement of Work for
Quick Turnaround Analysis.
2 U.S. EPA/Analytical Operations Branch. Fall 1991. Quick
Turnaround Method Regional Training Manual.
3 U.S. EPA/Analytical Operations Branch. May 1993 (Draft).
National Functional Guidelines for Quick Turnaround Method Data
Review.
4 U.S. EPA/Analytical Operations Branch. June 1993 (Draft). User's
and Sampler's Guide to the Quick Turnaround Method (QTM)
Analytical Service.
137
-------
18
Studies of Method Detection Limits in Solid Waste Analysis
David E. Kimbrough*, Public Health Chemist, and Janice
Wakakuwa, Supervising Chemist, California Department of Toxic
Substances Control, Southern California Laboratory, 1449 W.
Temple Street, Los Angeles California 90026-5698.
Abstract
A two part study is presented to assess the
applicability of the USEPA's Method Detection Limit (MDL) to
analysis of solid materials. The first part compares MDLs
calculated for arsenic, cadmium, molybdenum, selenium, and
thallium in soil with the actual method performance on real
spiked soils with these analytes at concentrations above and
below the calculated MDL. The MDL method is examined both
for its empirical suitability for solid waste analysis and
whether it is the proper theoretical tool for solid matrices.
The criteria are the precision and accuracy of results. The
results show that the MDL method produces accurate and
precise results only in interference free conditions.
This investigation was extended to an inter-laboratory
study which included the same five soils used above and five
other soils spiked with PCBs. 160 accredited environmental
laboratories participated in this study. The applicability
of the MDL was assessed by measuring the number of
qualitative and quantitative errors produced by the these
laboratories. The results indicate that approximately two
thirds of the reported MDLs produced significant errors.
138
-------
19
HOW THE REGION II RCRA QUALITY ASSURANCE OUTREACH PROGRAM
HAS ASSISTED INDUSTRY TO MINIMIZE RCRA COMPLIANCE COSTS
L. Lazarus. U.S. Environmental Protection Agency, Region II,
Edison, New Jersey 08837; P. Flax, U.S. Environmental
Protection Agency, Region II, New York, New York 10278; and T.
Ippolitto, U.S. Environmental Protection Agency, Region II, New
York, New York 10278.
ABSTRACT
The goal of the EPA Region II RCRA quality assurance outreach
program is to help the regulated community understand how to
comply with RCRA requirements and minimize RCRA compliance
costs.
The FY'93 outreach program consisted of the following seminars:
First Quarter FY'93 Outreach Efforts
1. Organic Data Validation. This five day course was offered
October 19-23, 1992 at Westchester County College in Valhalla,
New York. This course, which is scheduled several times each
year, illustrates how federal, state and local government
agencies can cooperate to assist the regulated community comply
with environmental regulations in a cost effective manner.
Members of the Region II Environmental Services Division's QA
staff provided guidance in developing the curriculum used for
this course, which is based on the Region II Organic Data
Validation Training Manual. Westchester County College is a
local government agency. The New York State Department of
Environmental Conservation (NYSDEC) allows individuals to
demonstrate their proficiency at validating data by
successfully completing the data validation training course.
2. Conference on Quality Assurance in Environmental
Monitoring. This conference was sponsored by Westchester
County College and the New York Water Environment Association -
Lower Hudson Chapter on November 18, 1992 in Yorktown Heights,
New York. The conference was hosted by the IBM T. J. Watson
Research Center, and was open to the public. Topics presented
included quality assurance concepts, data quality objectives,
ground water monitoring, data validation, quality assurance
requirements in New York State, data management, sampling
heterogeneous media, and data assessment. The Region II QA
staff were involved in developing the agenda for this
conference as well as presenting some of the sessions. On
November 19, 1992 a modified version of this presentation was
offered for the New York State Department of Environmental
Conservation in Latham, New York.
139
-------
Third Quarter FY'93 Outreach Efforts
1. April 19-23, 1993 Twenty five scientists and engineers
attended Westchester County College's inorganic data validation
training course in White Plains, NY. The NYSDEC recently
recognized individuals who pass the inorganic data validation
final examination as acceptable to perform data validation for
the Division of Hazardous Waste.
2. April 27-29, 1993 Twenty nine individuals attended the EPA
Region II Monitoring Management Branch's " How to write/review
a RCRA quality assurance project plan training course" in
Edison, New Jersey-
3. May 11, 1993 The TCLP training manual was finalized. The
manual was edited by an engineer from IBM's T. J. Watson
Research Center. IBM donated his time to the outreach project.
State agencies with RCRA primary enforcement responsibility
provided written copies of their TCLP policies. Partnerships
of federal and state agencies, with private industry, produced
more effective, and less expensive, outreach seminars.
4. June 7-11, 1993 Twenty eight scientists and engineers
attended Westchester County College's organic data validation
training course in White Plains, New York.
5. June 7-8, 1993 130 individuals representing, state
agencies, environmental laboratories, generators, and TSDFs
attended the RCRA Toxicity Characteristic Leaching Procedure:
Implementation and Compliance (TCLP) seminar in Albany. On
June 8, 1993, 119 individuals attended the TCLP seminar in
Rochester.
6. June 14, 1993 100 individuals attended the TCLP seminar in
San Juan, PR. On June 15-16, 1993, fifty seven individuals
attended the how to write/review a RCRA quality assurance
project plan training course in San Juan, PR.
7. June 21-23, 1993 110 individuals attended the Princeton,
NJ TCLP seminar on June 21; 88 people attended the seminar in
Edison, New Jersey on June 22; and 101 people attended the
seminar in New York City on June 23.
Fourth Quarter FY'93 Outreach Efforts
1. The TCLP seminar was offered twice at the Waste Testing and
Quality Assurance Symposium in Alexandria, VA. The Region II
organic, inorganic, dioxin, and TCLP data validation protocols,
and the Region II TCLP and CERCLA quality assurance manuals,
were uploaded onto OSWER's CLU IN computer bulletin board
system. This saves copying and mailing costs, and allows the
140
-------
regulated community to obtain copies of requested documents
immediately.
2. _ The October 21, 1993 Conference on Quality Assurance in
Environmental Monitoring was organized. The conference was
sponsored by Westchester County College and the New York Water
Environment Association, and was held in Yorktown Heights, New
York. The conference was hosted by the IBM T. J. Watson
Research Center. The goal of the conference was to teach
regulators and the regulated community how to reduce monitoring
costs. Conference topics included demonstrating EPA quality
assurance software, auditing environmental laboratories,
selecting monitoring well construction materials, and utilizing
immunoassay analyses.
FY'94 Outreach Efforts
We are arranging a series of QA monitoring conferences, TCLP
seminars, and quality assurance project plan training courses
to reduce the cost of environmental monitoring in the RCRA
program.
Discussion
The USEPA Region II RCRA quality assurance outreach program
assists the regulated community to minimize costs when
complying with RCRA regulations. There are four types of
training courses offered by the outreach program:
I. One day symposia on quality assurance in environmental
monitoring. These symposia offer two concurrent sessions on
monitoring and compliance issues. Conference attendees receive
EPA quality assurance training software on various topics, such
as data quality objectives (DQO), QA/QC principles, and
preparation of quality assurance project plans. A typical
conference agenda is attached.
II. One day toxicity characteristic leaching procedure (TCLP)
training courses. These courses illustrate how to use the DQO
process to generate valid data. The courses also demonstrate
how to reduce hazardous waste generator costs by explaining
when TCLP data are inapplicable, and how to obtain corrective
action management unit (CAMU) variances from the Land Ban
regulations.
The following synopsis describes four inappropriate uses of
TCLP, and how to obtain a CAMU variance:
141
-------
A. Invalid Risk Assessments
The TCLP model assesses risk to ground water when potentially
hazardous toxicity characteristic (TC) waste is co-disposed
with garbage into sanitary landfills. The TCLP model does not
assess risk when potentially TC waste is disposed in any other
matrix. Waste which is hazardous because it exhibits the
toxicity characteristic for mercury may be solidified by adding
cement and water to the waste in a tank. In this example, the
resultant concrete in not hazardous, and will be disposed at
the facility where it was generated. The waste concrete is not
hazardous because the concentration of mercury in the
concrete's TCLP extract is not a risk to ground water beneath
a sanitary landfill. However, because this concrete is non-
hazardous, it may emplaced where it was generated (at a non-
hazardous disposal site) . We do not know if the concrete poses
a risk to the ground water at the non-hazardous disposal site.
The resultant concrete may only be emplaced where it was
generated if it does not exhibit the toxicity characteristic
for mercury, unless a CAMU is obtained. The TCLP model
discloses no information about potential risks to groundwater
at the disposal site where the mercury immobilized in concrete
is emplaced. EPA is designing site specific risk assessment
models, but these will not be promulgated for several years.
When determining whether to use the TCLP for risk assessment,
it is important to remember that TCLP simulates worst case
management of hazardous waste in a landfill. Much caution must
be used before TCLP data are used in risk assessment because
the TCLP conditions rarely reflect actual site conditions.
EPA's Science Advisory Board Report outlines many limitations
of using TCLP for risk assessment at industrial sites. The
Science Advisory Board recommends developing leach tests which
emulate site conditions.
EPA's 1991 Science Advisory Board report on Leachability
Phenomena concluded that:
1. Many of the proposed uses of the EP and TCLP test have been
inappropriate because the waste management scenarios of concern
were not within the range of conditions used in the development
of the tests themselves. In most cases of inappropriate use of
the EP or TCLP tests, the justification given was that it was
necessary to cite "standard" or "approved" methods. Even if it
is acknowledged that the tests cannot be applied without
significant change in test protocol itself, the need to use a
previously "approved" test has been cited, (page 3)
142
-------
2. A variety of contaminant release tests and test conditions
which incorporate adequate understanding of the important
parameters that affect leaching should be developed and used to
assess the potential release of contaminants from sources of
concern. In scientific terms, no "universal" test procedure is
likely to be developed that will always produce credible and
relevant data for input to all decision making exercises.
(pages 7-8)
3. Leach test conditions appropriate to the situations being
evaluated should be used for assessing long-term contaminant
release potential. The best way to estimate the extent of
contaminant release from a waste matrix of interest is to have
a test that reflects realistic field conditions, (page 13)
4. To facilitate the evaluation of risk implications of
environmental releases, the EPA should coordinate the
development of leach tests and the development of models in
which the release terms are used. (page 17)
In addition, the TCLP test cannot predict the potential for
toxic chemicals to leach from oily waste, through soil, to
contaminate ground water. This applies to both sanitary
landfills and industrial sites. EPA and the American Society
of Testing and Materials (ASTM) have formed a workgroup to
develop a site specific risk assessment model for oily waste.
At a minimum, the model will incorporate physical and chemical
characteristics of the oily waste and the soil. However, this
model is not expected to be approved by EPA for several years.
Until EPA approves this site specific model for oily waste risk
assessments, oily waste site assessments should be based on
total constituent analysis, not TCLP extract analysis.
B. Unnecessary Hazardous Waste Determinations;
1. Generator's knowledge of waste (e.g. chocolate ice cream).
2. Exempt waste (e.g. household garbage).
3. Material is not a solid waste (e.g. clean sand, laundry
detergent).
4. Generator's testing of waste (total constituent analysis is
available).
5. The solid waste is a listed hazardous waste.
C. Unnecessary Land Ban Determinations:
1. Some Land Disposal Restrictions (LDRs) are for total
constituent instead of TCLP extract concentrations.
2. Generator's testing of waste (total constituent analysis is
available).
3. Pure liquid waste samples (waste is TCLP extract; waste
would fail paint filter test).
143
-------
D. Determination of Corrective Action Clean-up Levels and
Clean Closures
Corrective action clean up levels and clean closures are site
specific. They are not based on toxicity characteristic
regulatory action levels.
E. LDR Variances; CAMUs are Designed to Reduce the Cost of
On-Site Remediation
Corrective action management unit (CAMU) regulations are
enumerated in 40CFR260.10 and 40CFR270.2. CAMU is an area
within a facility designated by the EPA Regional Administrator
for implementing CERCLA or RCRA corrective action requirements.
A CAMU may only be used for the management of remediation
wastes pursuant to corrective action requirements at a
facility.
Placement of hazardous remediation waste into a CAMU will not
automatically trigger LDRs. This variance from the LDRs can
result in substantial cost reductions. CAMU boundaries are not
confined to where contamination exists at the site; CAMU
boundaries are based on where remediation waste will be
managed.
i
Limitations and Conditions Applicable to CAMU Designations
1. The CAMU shall facilitate the implementation of reliable,
effective, protective, and cost-effective remedies;
2. Waste management activities associated with the CAMU shall
not create unacceptable risks to humans or to the environment
resulting from exposure to hazardous wastes or hazardous
constituents;
3. The CAMU may only include uncontaminated areas of the
facility if the incorporated area is more protective than
management of such wastes at contaminated areas of the
facility;
4. Areas within the CAMU, where wastes remain in place after
closure of the CAMU, shall be managed and contained so as to
minimize future releases to the extent practicable;
5. The CAMU shall expedite the timing of remedial activity
implementation, when appropriate and practicable;
6. The CAMU shall enable the use, when appropriate, of
treatment technologies (including innovative technologies) to
enhance the long-term effectiveness of remedial actions by
reducing the toxicity, mobility, or volume of wastes that will
remain in place after closure of the CAMU; and
7. The CAMU shall, to the extent practicable, minimize the
land area of the facility upon which wastes will remain in
place after closure of the CAMU.
144
-------
III. Five day organic and five day inorganic data validation
training courses at Westchester County College.
The EPA Region II CERCLA QA manual defines data validation as
"a systematic process for reviewing a body of data against a
set of criteria to provide assurance that the data are adequate
for their intended use. Data validation consists of data
editing, screening, checking, auditing, verifying, certifying,
and reviewing." Data validation reduces false negatives, false
positives, and misquantitation in reported data.
Misquantitation includes both laboratory arithmetic errors, and
data qualified as estimated or presumptively present because of
analytical problems. Agendas for both data validation courses
are attached.
IV. The three day course in preparing and evaluating RCRA
quality assurance project plans has been condensed into a one
day workshop on generating scientifally valid and legally
defensible data. A typical agenda is attached.
Outreach Activities on OSWER's CLU-IN Computer Bulletin Board
System
Administrator's Update #5, issued on September 24, 1993,
discussed cost-saving measures the government can take to
reduce useless expenditures, such as increased utilization of
electronic reporting and communication. In order to comply
with Update 5, we have uploaded our TCLP manual, CERCLA quality
assurance manual, and data validation protocols onto the OSWER
CLU-IN computer bulletin board system. When people call Region
II for copies of these protocols or documents, we tell them how
to download the files from the CLU-IN bulletin board system.
This saves copying and mailing costs, and people obtain copies
of these documents immediately.
Proposed 1995 RCRA Outreach Activities
In 1995, we plan the following additional outreach activities:
1. Modify EPA's QASPER 4.0 quality assurance project plan
preparation software to properly identify and characterize some
types of hazardous waste. Assisting hazardous waste generators
identify and characterize their waste in a cost effective
manner is a fundamental objective of the Region II RCRA QA
outreach program.
2. OSWER's CLU-IN computer bulletin board system will be
reconfigured in the near future. In 1995, the Region II RCRA
Outreach program will have an easily accessible category of the
bulletin board system devoted to increasing compliance and
reducing costs.
145
-------
Course Itinerary for Generating Scientifically Valid and
Legally Defensible Data Workshop at 1994 Waste Testing and
Quality Assurance Symposium
How to Use Proper QA/QC Procedures, and the DQO process, to
Generate Scientifically Valid and Legally Defensible Data
Leon Lazarus and Dr. Margo Hunt, USEPA Region II,
Environmental Services Division.
This course is designed for project managers who are
responsible for sampling potentially hazardous waste. The
course consists of lectures, hands on computer training, and
brief discussions on quality assurance project plan training
materials which will be disseminated.
As buzz words, Quality Assurance (QA), and Quality Control
(QC), seem to leave most individuals in the dark as to their
meaning, their difference, and their significance. A brief
introduction to both QA and QC will clarify the subject and
give you insights into their importance at EPA. Also to be
discussed is the Data Quality Objectives (DQO) process
recommended by EPA. Most individuals recognize that the
evolution of a project is a multistep process which entails
several individuals that must function together. The DQO
process is a systematic planning tool for ensuring that the
type, quantity, and quality of data collected will correspond
to the decision to be made, and the importance of making the
right decision. For example, a very low detection limit is
needed if drinking water is sampled for volatile organic
compounds whereas soils on a Superfund site would not need
such low detection limits since contaminant regulatory action
levels are much higher.
Appropriate QA/QC procedures are also delineated in the DQO
process. Proper containers, preservatives, and holding times
are essential if the data generated are assumed to be valid.
For example, if sampling equipment or containers are not
cleaned properly, samples may show the presence of
contaminants that originated from the equipment or containers
themselves. If monitoring wells are not properly constructed
and evacuated, the resultant sample data may be too high or
too low, even if the lab analyzed the sample perfectly- If
proper chain of custody procedures are not followed, data
will not be considered valid in court.
146
-------
Course Outline
8:00 am Quality Assurance Concepts
9:00 am Hazardous Waste Field Sampling Computer Based
Training hands-on demonstration
10:30 am Coffee Break
10:45 am Data Quality Objectives
11:30 am Data Quality Objectives Computer Based Training
hands-on demonstration
12:30 pm Additional Quality Assurance Materials
(provided on 3 1/2 inch computer disks):
1. New York State Department of Environmental
Conservation RCRA QA Manual
2. USEPA Region V RCRA QA Project Plan
Preparation Guidance
3. Data Quality Objectives and Sampling Design,
from USEPA Region II TCLP Manual
4. USEPA QASPER 4.0 DQO based Project Plan
Preparation Software
5. USEPA Region II SOP for Writing SOPs.
147
-------
The Professional
Development
Center
Programs for Career and Professional Growth
SUNY/WESTCHESTER
COMMUNITY COLLEGE
Courses & Programs on
Environmental Monitoring:
Practices & Procedures
Fall 1993
These seminars and courses are
offered to environmental engineers,
project managers, technicians and
related personnel to enable them to
comply with governmental
regulations relating to environmental
data.
148
-------
Organic Data
Validation
Course # XAP-555
November 15-19,1993
8:30 a.m. to 4:30 p.m.
Tuition: $1,100, includes
all course materials,
continental breakfast, and
lunch daily.
This 35 hour course in the Performance of Data Validation Techniques For
Organic Analyses As Required By The Superfund Program includes a four
hour final examination. Techniques will include reviews of both CLP and non-
CLP generated data and the data will be reviewed for both validity and
usability. A certificate will be issued by the College to those passing the
examination. Holders of this certificate are recognized by the NYS
Department of Environmental Conservation as acceptable to perform Organic
data validation for ENCON's Division of Hazardous Waste.
Recommended Prerequisite: B.A. or B.S. in Science or Engineering and at
least 6 mos. experience in the chemical analysis of environmental samples; or
college level instrumentation course in analytical chemistry (GC or GC/MS).
Participants should bring a scientific calculator.
The instructor for the course will be John H. Samuelian, Ph.D. Dr. Samuelian
is a certified organic data validator and has over five years experience in data
validation in federal and state Superfund programs. He participated in the
development of the first training manual and has peer reviewed the current
training manual. He is currently serving as program manager for EA
Engineering, Science and Technology's data validation program.
Topics will include:
Intro to Sampling Analysis Plans and
Data Quality Objectives (SAPs & DQOs)
Intro to Contract Laboratory
Procedure Analysis (CLP)
Intro to Data Package
Intro to Validation
Holding Times
Blanks
Calibration
Internal Standards
Contact Required
Quantitation Limits (CRQLs)
Data Calculations
Surrogates
Special ID
Field Duplicates
Matric Spike/Matrix
Spike Duplicates
Pest Holding Times
Pest Surrogate Times
Pest MS/MSD
Pest Blanks
Pest Calibration
Pest Analytical Sequence
Pest Cleanup Verification
Pest ID
Pest Calculations
CLP Contract
Inorganic Data
Validation
Course* XAP506
April 19-23,1993
8:30 a.m. to 4:30 p.m.
Tuition: $1,100, includes
all course materials,
continental breakfast, and
lunch daily.
This 35 hour course in the Performance of Data Validation Techniques For
Inorganic Analyses As Required By The Superfund Program includes a four
hour final examination. Techniques will include reviews of both CLP and non-
CLP generated data and the data will be reviewed for both validity and
usability. A certificate will be issued by the College to those passing the
examination. Holders of this certificate are recognized by the NYS
Department of Environmental Conservation as acceptable to perform
inorganic data validation for ENCON's Division of Hazardous Waste.
A B.A. or B.S. in Science or Engineering, plus one year of College Chemistry
are minimum requirements for registration.
Participants should bring a scientific calculator.
The instructor for the course will be Dale S. Boshart. Mr. Boshart is currently a
Team Manager with Lockheed Engineering, Managing Region III ESAT Project
in support of USEPA's commitment to the Superfund program. He was
formerly associated with Roy F. Winston Corporation, ESAT, USEPA Region II
as senior Inorganic Data Reviewer.
149
-------
3D311ODAllNnWWOD
Professional Development Center Registration Form
Name Address
Company Name & Address
Day Telephone Social Security #
Please register me in the following course/courses. My check or company voucher is enclosed. Make checks payable
to Professional Development Center.
Please register me in the following course via credit card • Signature
Master Card Exp. date Visa Exp. date
Course* Quality Assurance Conference • October 21,1993 • Course fee: $95
Course* Organic Data Validation • November 15-19,1993 • Course fee: $1,100 • Limit 25
Registrations must be received at least one week prior to program date.
Mail to: Professional Development Center, Administration Building, Westchester Community College,
75 Grasslands Rd, Valhalla NY 10595
For further registration information call Elaine Sail 914-285-6659
150
-------
SESS/ON C
Ul
Workshop on How to Use Proper
QA/QC Procedures, and the DQO
Process, to Generate Scientifically Valid
and Legally Defensible Data
—Leon Lazarus, USEPA Region II
9:00 am - 10:00 am Quality Assurance
Concepts
10:00 am - 11:00 am Hazardous Waste Field
Sampling Computer-
Based Training
Hands-On Demonstration
11:00 am - 11:30 am Coffee Break
11:30 am - 12:30 pm Data Quality Objectives
12:30 pm - 1:30 pm Lunch
1:30 pm - 2:30 pm Data Quality Objectives
Computer-Based Training
Hands-on Demonstration
2:30 pm - 3:00 pm Questions'and Answers
This workshop is designed for project
managers who are responsible for sampling
potentially hazardous waste and for envi-
ronmental attorneys who need to under-
stand the uncertainties in environmental
data. The course consists of lectures, hands
on computer training, and brief discussions
on quality assurance project plan training
materials which will be disseminated.
As buzz words, Quality Assurance, QA,
and Quality Control, QC, seem to leave most
individuals in the dark as to their meaning,
their difference, and their significance. A
brief introduction to both QA and QC will
clarify the subject and give you insights into
their importance at EPA. Also to be dis-
cussed is the Data Quality Objectives (DQO)
process recommended by EPA. Most indi-
viduals recognize that the evolution of a
project is a multistep process which entails
several individuals that must function togeth-
er. The DQO process is a systematic plan-
ning tool for ensuring that the type, quantity,
and quality of data collected will correspond
to the decision to be made and the impor-
tance of making the right decision. For
example, a very low detection limit is needed
if drinking water is sampled for volatile
organic compounds, whereas, soils on a
Superfund site would not need such low
detection limits since contaminant regulatory
action levels are much higher.
Appropriate QA/QC procedures are also
delineated in the DQO process. Proper con-
tainers, preservatives, and holding times are
essential if the data generated are assumed to
be valid. For example, if sampling equip-
ment or containers are not cleaned properly,
samples may show the presence of contami-
nants that originated from the equipment or
containers themselves. If monitoring wells
are not properly constructed and evacuated,
the resultant sample data may be too high or
too low, even if the lab analyzed the sample
perfectly. If proper chain of custody proce-
dures are not followed, data will not be
considered valid in court.
$&»W£DNESDA YtfMA Y, 118*1994 .•»•
(Jonfe
Quality
Environmental J^onitoring
erence on
ssurance in
on (generating
Scientifically \falid and
J^gally Defensible
Sponsored by
SETI, LTD. and ALFRED UNIVERSITY
Alfred, NY 14802
Conference location:
Environmental Technologies Information Center
F. W. Olin Building, College of Business
Alfred University
Alfred, NY 14802
WEDNESDA Y#MA
-------
cn
ro
The conference topics will be
relevant to managers and staff from
private industry, consulting firms, and
regulatory agencies. For registration
information, contact Roland Hale at
(607) 587-8377. For program
information, contact Leon Lazarus,
USEPA Region IIRCRA Quality
Assurance Officer, at (908) 321-6778.
Conference Chairmen: Leon Lazarus
and Phil Flax, USEPA Region II
Return reservation form by May 10,
1994
Conference Fee: $95/person
Total amount enclosed: $
Name:
Title:
Organization:_
Address:
City
Slate
ZIP
Telephone: ( )
Please attach other names on a separate sheet.
Make checks payable to: SET! LTD.
Send reservations to:
SETI, LTD.
Ceramic Corridor Innovation Center
200 N. Main Street
Alfred, NY 148O2
SESSION A
Two concurrent sessions (A and B) will
be held to present principles applicable to waste,
soil, and water environmental monitoring.
Media-specific monitoring topics will also be
addressed.
8:50 am - 9:00 am Opening Remarks
9:00 am - 10:00 am
10:00 am - 10:30 am
10:30 am - 11:30 am
11:30 am - 12:30 pm
Do Low Matrix Spike
Recoveries Equal Bad Data:
A Case Study of Hexavalent
Chromium in Soil
—Rock Vitale, Environmental
Standards. Inc
Coffee
Regaining Control of Your
Environmental Investigation
Throufh Auditing Your
Environmental Laboratory
—Rock Vitale, Environmental
Standards, Inc.
How to Utilize Qualified
Data in the SPDES Program
-Larry Bailey, NYSDEC
12:30 pm - 1:30 pm Lunch
l:30pm- 2:30 pm
The Laboratory/Client
Relationship: Two-way
Communication for
Selecting Appropriate
Methodology & Deliverables
—Gale Sutton, Gabon
Corporation
2:30 pm - 3:00 pm Coffee
3:00 pm - 4:00 pm Immunoassay Demonstration
—Bob Bednar, Ensys
Corporation
SESSION B
8:50 am - 9:00 am Opening Remarks
9:00 am - 10:00 am ERA'S Guidance for
Sampling Demolition
Debris Containing Lead
Based Paint
—John Hansen, USEPA
Region II
10:00 am - 10:30 am Coffee
10:30 am - 11:30 am Improper Hazardous
Waste Characterizations:
Financial and Compliance
Implications
—Richard Walka, William
F. Cosulich Associates
11:30 am- 12:30 pm When is TCLP
appropriate? When is
TCLP not applicable?
—John Hansen, USEPA
Region II
12:30 pm- 1:30 pm Lunch
l:30pm- 2:30pm Immunoassay
Demonstration
—Bob Bednar, Ensys
Corporation
2:30 pm- 3:00 pm Coffee
3:00 pm - 4:00 pm Hazardous Waste
Minimization: Regulatory
Requirement or
Prescription for Survival?
—Richard Walka, William
F. Cosulich Associates
-------
20
AN EVALUATION OF QC REQUIREMENTS IN THE CLP FOR FURNACE ATOMIC
ABSORPTION ANALYSES
D.C. Hillman. J.T. Rowan, and D.M. Boyer, Lockheed Environmental Systems and Technologies
Company, Las Vegas, NV 89119, and LC. Butler, Environmental Protection Agency, Las Vegas,
NV 89119
ABSTRACT
Under its mission of performing Superfund QA research for the Office of Research and
Development (ORD), the Quality Assurance Branch of the EPA EMSL-LV has reviewed the method
quality control (QC) for furnace atomic absorption analysis (AAS). The QC requirements for
furnace AAS in the CLP program are quite stringent and contribute significantly to making furnace
AAS the slow point in metals analysis. The QC could be modified to increase throughput and
decrease reanalyses without significantly affecting the resulting data quality. Areas for modification
include the duplicate-injection requirement, the criteria for the concentration of the analytical spike,
and the corrective action dictated by a poor spike recovery.
Currently, the Superfund CLP Inorganic Statement of Work (SOW) requires that all furnace AAS
analyses be performed using duplicate injections, which doubles the analysis time from 2-3
minutes/sample to 4-6 minutes/sample. Historically, this requirement was included in the analytical
procedure because it was common for the single-measurement precision to be poor due to an
atypical measurement or poor injection. Modern instrumentation, though, does not suffer from poor
precision. Data from routine sample analyses over 6 months on multiple instruments was reviewed
and the frequency of unacceptable results was less than 1% for As, Pb, Se, and Tl (>5000
injections). For the same data, precision and accuracy are equivalent for single and double
measurements. The data also demonstrate that bad injections and atypical measurements are
adequately detected using the post-digest spike recovery. Therefore, considerable time savings
would be realized if only single measurements were required and data quality would not be
affected.
The furnace AAS QC requirements specify that a post-digest analytical spike (PDS) must be
analyzed for every sample. The PDS recoveries are used to determine if matrix interferences are
affecting the quantification. If a significant effect is detected, the sample must be reanalyzed by
the method of standard additions (MSA) or flagged as "W" or "E" (out-of-control). However, under
the current procedures a significant fraction of PDS recoveries can be outside the acceptance
criteria due to measurement variability rather than matrix effects. Additionally, the relative
acceptance criteria can be more stringent than the analytical precision. Both issues lead to data
flags or MSA analyses for reasons other than matrix interferences. By modifying the PDS
concentration and the procedure for interpreting the recoveries, unnecessary MSA analyses and
data flags can be minimized, with the benefit of reduced analytical costs and equivalent, or better,
data quality. Proposed modifications are provided in this paper.
NOTICE: Although the research described in this article has been funded wholly by the U.S.
Environmental Protection Agency through Contract 68-CO-0049 to Lockheed Environmental
Systems & Technologies Company, it has not been subjected to Agency review. Therefore, it
does not necessarily reflect the views of the Agency.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
153
-------
INTRODUCTION
Quality control (QC) plays an important role in analytical procedures performed under the Superfund
CLP program. QC data is used to evaluate the quality of associated sample data and to judge its
usability. The amount of QC must be balanced with cost. Too much QC increases costs without
providing any more information regarding data quality or usability, while too little decreases
analytical costs but results in sample data of unknown quality. In support of the Office of Research
and Development (ORD), EMSL-Las Vegas reviews the CLP analytical methodology and associated
QC on a continuing basis to ensure that both are current with today's rapidly improving technology.
This includes evaluating and balancing QC requirements. In this paper, the QC required for CLP
furnace atomic absorption analysis (AAS) is evaluated. The QC requirements for furnace AAS in
the CLP program are quite stringent and contribute significantly to making furnace AAS the slow
point in metals analysis. However, technology advances over the years may permit modifying the
current QC without impacting data quality or useability. Two specific aspects of furnace AAS QC
are considered, the duplicate injection requirement and the post-digest analytical spike (PDS)
procedure.
Under the Superfund CLP Inorganic Statement of Work (SOW) all furnace AAS analyses must be
performed using duplicate injections, which doubles the analysis time from 2-3 minutes/sample to
4-6 minutes/sample. Historically, this requirement was included in the analytical procedure because
it was common for the single-measurement precision to be poor due to an atypical measurement
or poor injection. Modern instrumentation, though, does not suffer from poor precision or atypical
measurements. Considerable time savings would be realized if only single measurements were
required. In this paper, the precision and accuracy of single and double injection measurements
are compared using data obtained from routine analyses over a period of several months. This
paper discusses the data quality implications of removing the double-injection requirement and how
the furnace AA QC decision tree could be changed to accommodate single measurements.
The furnace AAS QC requirements also specify that a PDS must be analyzed for every sample.
The PDS recoveries are used to determine if matrix interferences are affecting the quantification.
If a significant effect is detected, the sample must be reanalyzed by the method of standard
additions (MSA) or flagged ("E" or "W"). However, with the current spiking levels, measurement
variability can result in significant PDS recovery variability. As the variability of the recovery
increases, an increasing number of recovery values will be outside the acceptance criteria due, not
to a matrix effect, but to measurement variability. Consequently, samples are unnecessarily
reanalyzed by MSA or the results flagged, which increases analysis costs and time, and can also
taint the data useability. This paper discusses how the measurement variability affects the recovery
values and how to minimize the effect. By modifying the concentration of the analytical spike and
the procedure for interpreting analytical spike recoveries, unnecessary MSA analyses and data
flags can be minimized, with the benefit of reduced analytical costs and equivalent, or better, data
quality. Proposed modifications (with examples) will be provided in this paper.
EXPERIMENTAL
Lockheed has a commercial analytical laboratory equipped with 5 Perkin-Elmer and 1 Hitachi
furnace AAS instruments. Each of these is busy 24 hours a day performing routine analyses for
As, Pb, Se, and Tl in all types of environmental samples. Both CLP-type and non-CLP protocols
are followed, depending upon the client. The non-CLP protocols utilize single-injection
measurements and higher PDS concentrations.
154
-------
To compare the precision and accuracy of single and double injection analyses, the QC data (ICVs,
ICBs, CCVs, CRAs, CCBs) from single and double injection analytical runs were examined over
the time period November 1993 - April 1994.
The PDS procedure was examined by first determining the theoretical effect of measurement
variability upon the PDS recovery using progression of error calculations. The analytical spike data
from the same analytical runs used above were then examined for PDS recovery distributions and
the effect of measurement variability.
RESULTS & DISCUSSION
Double Injection vs. Single Injection Measurements
Table 1. Bad Injection Summary
Analyte
As
Pb
Se
Tl
No. Injections
3984
3438
1830
1628
% bad injections'
0.8
0.6
0.3
0.2
RSD > 20% and concentration > CRDL
Modern instrumentation is very reliable. If the
frequency of bad injections is very low, then
having a QC step to detect the occurrence of a
bad injection is not very useful or cost-effective.
Routine data from double-injection analytical
runs were examined to determine the frequency
of bad injections. The data, summarized in
Table 1, indicate that the overall frequency of
bad injections is less than 1%. This low
frequency for bad injection occurrences
suggests that the double-injection requirement should be dropped. If the requirement is dropped,
the question "What would be the effect on the data quality?" is raised. To answer this question,
the precision and accuracy for both double- and single-injection analyses were compared using the
results for ICB/CCB, ICV/CCV, and CRA QC samples. The results, summarized in Table 2,
indicate equivalent precision and accuracy for double and single injection analytical runs.
Table 2. QC results for Single and Double Injection Analyses
Analyte
As
Pb
Se
No. injections
Single
Double
Single
Double
Single
Double
ICB/CCB*
(ppb)
-0.03 ±1.3 (140)
0.3 ± 1.1 (234)
0.07 ± 0.86 (234)
0.01 ± 0.79 (225)
0.55 ± 0.89 (154)
0.36 ±0.70 (115)
ICV/CCV*
(% recovery)
100.5 ± 5.5 (111)
103.3 ±5.6 (186)
98.8 ± 4.7 (242)
99.2 ± 3.7 (222)
101.8 ± 5.0 (153)
102.4 ± 4.8 (113)
CRA*
(% recovery)
103 ± 19 (23)
106 ± 14 (39)
106 ±27 (36)
114 ± 31 (46)
95 ± 26 (22)
101 ± 13 (18)
* The number of observations is indicated in ()
The analytical spike data for samples analyzed in the same analytical runs can also be examined
to determine if a difference exists between single and double injection measurements. Table 3
summarizes the distribution (by %) of analytical spike recoveries from double injection analyses and
the results that would be obtained from the same analyses if only the first injection was utilized.
The table shows that there is effectively no difference between double and single injection
analyses.
155
-------
Table 3. Distribution of analytical spike recoveries (by %)
Analyte
(# samples)
Arsenic
n = 455
Lead
n = 420
Selenium
n = 246
Thallium
n = 312
# inj.
Double
Single
Double
Single
Double
Single
Double
Single
85<%R<115
54
55
61
64
60
59
85
81
%R < 40
2
3
4
4
6
9
0
0
40 < %R < 85
or
%R> 115
44
42
35
27
34
32
16
19
Another concern for single-injection analyses is "How is a bad injection detected during single-
injection analyses?". During an analytical run there are essentially two types of samples, calibration
QC samples (ICVs, CCVs, CCBs, CRAs) and "other" samples (field and related QC samples). A
bad injection for a calibration QC sample will be apparent from the recovery of the true value. For
the "other" samples, the PDS recovery, rather than injection precision, can serve as an indicator
of a bad measurement. Bad measurements will affect PDS recovery values, which will trigger
corrective action as specified by the furnace AAS QC flowscheme. The bad injection data from
samples identified as having unacceptable duplicate precision in Table 1 were evaluated against
the QC flowscheme. Examples are given in Table 4. In each case, the bad injection is detected
and resolved by the required QC action.
Table 4. Examples of Bad Injection Data
Item
Sample 1 (ppb)
Sample 1 + Spike (ppb)
% Recovery
Analyte
As
Injection 1
42.0
39.4
-12.9
Injection 2
16.5
37.4
104.6
RSD
61.0
3.7
Discussion: Bad injection (high bias) resulted in negative recovery, requiring reanalysis. Subsequent reanalysis of
this sample resulted in acceptable recovery (104%) and verified the sample result for injection 2 (18.1 ppb).
Sample 2 (ppb)
Sample 2 + Spike (ppb)
% Recovery
Pb
2U
22.2
110.8
-2.9
-3.5*
-17.5
NA
194
Discussion: Bad injection (low bias) resulted in negative recovery, requiring reanalysis. Subsequent reanalysis of
this sample resulted in acceptable recovery (91%) and verified the sample result for injection 1 (2U ppb).
Sample 3 (ppb)
Sample 3 + Spike (ppb)
% Recovery
Se
11.8
20.2
83.6
20.6
15.5
-50.8
41.4
18.6
Discussion: Bad injection (high bias) resulted in low recovery, requiring reanalysis. Subsequent reanalysis of this
sample resulted in acceptable recovery (98%) and verified the sample result for injection 1 (9.8 ppb).
Sample 4 (ppb)
Sample 4 + Spike (ppb)
% Recovery
Tl
2.7
22.8
100.4
2.0
14.4
61.8
NA
26.1
Discussion: Bad injection (low bias) resulted in low recovery, requiring reanalysis. Subsequent reanalysis of this
sample resulted in acceptable recovery (92%) and verified the sample result for injection 1 (2 ppb).
Suspected bad injection
156
-------
In summary, the data support dropping the double-injection requirement. Considerable time savings
would be realized if only single measurements were required and data quality would not be
affected. The precision and accuracy are equivalent to that for double injection measurements and
bad injections (measurements) are still detected by other QC samples. An additional, minor
requirement should be added for single-injection analyses in that the PDS sample must also be
prepared and analyzed for the pre-digest matrix spike sample (in order to detect a bad injection).
Analytical Spike Procedure and Interpretation
A PDS must be prepared and analyzed for each element analyzed by furnace AAS in all samples
(except the pre-digest matrix spike sample). For each PDS, a recovery is calculated using the
following equation;
% R = x 10Q
SA
where
SSR = analytical spike sample result
SR = sample result
SA = spike added
The PDS recovery calculation endeavors to reflect the impact of the sample matrix on the reliability
of an analytical determination. The recovery value is used to answer the question "Is the matrix
significantly affecting the quantification of the analyte using the existing external calibration?".
When a significant effect is detected, samples are either flagged or must be reanalyzed by MSA.
A significant effect is defined using a fixed acceptance window. When PDS recoveries are within
the window, the measurements are accepted as reliable, otherwise they are not. The acceptance
windows should be wide enough to permit for reasonable precision and normal variability in the
analytical measurements, but not so wide that significant matrix effects can go undetected. The
spike should also be large enough that the method can reliably measure the difference between
the spiked and unspiked samples. However, the current requirements for PDS samples and
recovery interpretation do not address the issue of simple measurement variability.
To understand the impact of measurement variability, the error in %R must be considered. Through
progression of error calculations, an estimate of the relative error in %R is calculated as follows;
=
%R SSR-SR
where
s%R = estimated standard deviation for %R
SSSR = estimated standard deviation for SSR, and
s,B = estimated standard deviation for SR.
On
The terms s2^ and s2SR tend to increase with increasing concentration. Spike levels (SA) are fixed
for each analyte, and the term SSR-SR is equal to SA at 100% recovery. From this it is apparent
that the variability of %R is very dependent upon both the sample concentration and the PDS
concentration. As the variability increases, the chances of a recovery value being outside the
acceptance criteria due to random error (i.e., measurement variability) also increase. Figure 1 is
a plot of the frequency at which spike recoveries (assuming a mean of 100%) fall outside the
157
-------
acceptance windows due to random errors as a function of the sample/PDS concentration ratio.
The frequency was determined by calculating the standard deviation (s%R) of the PDS recovery
(using the relative error equation above) and assuming a normal distribution around the mean.
From this it is possible to determine the expected percentage of the values that lie outside of any
given window around the mean. s%R was calculated for several values of SA and SR. Since %R
= 100, SSR-SR is defined by SA. The standard deviations of SR and SSf? were assumed to follow
a normal distribution around the mean and are based on a 1 ppb error for concentrations below 20
ppb and 4% error at concentrations above that. From Figure 1 it is apparent that the relative error
becomes very significant as SR/SA increases. Consequently the frequency of recovery values
outside of the acceptance windows due solely to random error increases as spike concentration
decreases and as sample concentration increases.
0.6
8°'4
CD
;g
' 20 ppb
and 1 ppb error when concentration < 20 ppb)
Figure 1. Expected Fraction of PDS Recoveries Outside of QC Limits vs. the SR/SA Ratio
When the ratio of SR/SA is large, a considerable portion of the unacceptable spike recoveries could
be due to measurement variability rather than to matrix effects. Using lead as an example, Table
5 lists the theoretical frequency of false QC hits due to measurement variability for the current
spiking level and for a higher spiking level.
158
-------
Table 5. Frequency of false QC hits due to measurement variability
SR (ppb)
10
20
40
100
5/4 = 20 II 5/4 = 100
SR/SA
0.5
1.0
2.0
4.0
%>QC
20.0
9.0
29.0
62.0
SR/SA
0.1
0.2
0.4
1.0
%>QC
0.1
0.1
1.0
6.0
Direct evidence that current spiking levels and acceptance windows result in inappropriate QC hits
can be obtained by examining the absolute recovery (rather than the relative recovery) for analytical
spike samples. The absolute recovery is defined by the following equation:
D = | (SSR-SR) - SA\
If the value of D is not significant, then the recovery should be deemed acceptable (i.e., no
significant matrix effects). "Not significant" is defined here as 2*IDL, which is based upon the 99%
upper confidence interval for the IDL (2 ppb for each analyte in this work). The definition of an
acceptable D at a given value of SA effectively defines the acceptance window for %R. For As,
Pb, and Tl (IDL=2 ppb and SA=2Q ppb) a 2*IDL (or 4 ppb) absolute window corresponds to an 80-
120% acceptance window for %R. Similarly for Se (IDL=2ppb and SA=10 ppb), a 2*IDL absolute
window translates to a 60-140% acceptance window for %R. Considering that the current windows
are fixed at 85-115% regardless of analyte or spiking level, it is evident that the recovery values
outside of these fixed windows will not always be due to a significant matrix effect.
Table 6 lists the distribution of PDS recovery data for two different PDS concentrations from a
number of routine analyses. Included for each fraction outside the acceptance criteria is the
percentage of the fraction for which the absolute recovery is acceptable. The data clearly show
that, due to measurement variability, the value of %R is not always a reliable indicator of matrix
interferences with the current spiking levels and %R acceptance criteria. Combined, the data in
Figure 1 and Tables 5 and 6 indicate that there is a marked reduction in the number of
unacceptable spike recoveries due to random errors as the concentration of SA is increased and
if the absolute recovery is considered. Therefore, it is apparent that the way to minimize the effect
of measurement variability and simplify interpretation of the %R values is to raise the value of the
spike concentration. The concentration of the spike should be high enough such that the absolute
recovery is not a factor and low enough so that spiked samples do not often require dilution for
analysis. Increasing the value of SA will significantly reduce the number of data flags and MSA
analyses that result from reasons other than matrix interferences, namely measurement variability.
In order to estimate suitable concentrations for SA, frequency distributions for As, Pb, Se, and Tl
in both soil and water digests were obtained from the CLP Analysis Results Database (CARD), and
are listed in Table 7. Suggested values for SA are included in the table. For Se, although the
SR/SA ratio is less than 1 for most samples (-75%), the spiking level is close enough to the
detection limit that the variability in the spike recovery is impacted significantly (from figure 1 it is
estimated that greater than 20% of the recoveries could be outside the window due to chance
alone) If the spiking level were raised to 40 ppb, more of the samples would have low SR/SA
159
-------
Table 6. Distribution of %R Values
Anal.
As
Pb
Se
Tl
spike (ppb)
20
40
20
40
10
40
20
40
n
457
290
420
474
246
365
312
154
Distribution of %R values*
<40
3(0)
2(0)
4(0)
6(0)
7(0)
4(0)
0
2(0)
40-85
14(7)
17(0)
17(11)
9(0)
25 (25)
10(0)
13(11)
10(0)
85-115
53
71
60
80
59
71
85
84
>115
30 (20)
10(0)
19 (18)
5(0)
9(9)
15(0)
2(2)
4(0)
* Value in () is the % for which the absolute recovery is OK
values (~95%) and the effect of variability near the detection limit would not adversely affect the
recovery values. For As and Pb in water samples, spiking at 20 ppb results in acceptable SR/SA
values for most samples (-75%). For soil samples, a 20 ppb spike will result in a large number of
false QC hits. Increasing the spike level to 40 ppb for As and Pb would significantly reduce the
chances of a false QC hit. For simplicity, a value of 40 ppb is also suggested for water samples.
For Tl, spiking at 20 ppb results in acceptable SR/SA values for most samples (~95%). The linear
ranges of most modern instrumentation should be able to accommodate spikes at these levels
without requiring excessive dilutions.
Table 7. As, Pb, Se, and Tl in real samples
Analyte
Matrix
Distribution (ppb by percentile)
25th
50*
75*
95th
Suggested SA
ppb in digestate
As
Pb
Se
Tl
Water
Soil
Water
Soil
Water
Soil
Water
Soil
3.1
10
2.4
34
2
3
1.2
1.6
6.4
24
4.7
92
3
5
2.1
2.4
16
53
15
473
10
11
5.2
4
86
400
123
9700
31
42
20
12
40
40
40
40
40
40
20
20
Although the linear range of furnace AA instruments is a limiting factor on the maximum PDS
concentration, another factor is the matrix interference mechanism. If the interference causes equal
suppression at all concentrations, then the PDS concentration is not a concern. However, if the
mechanism is not linear with concentration, then it may be possible to overwhelm the effect by
spiking so high that the effect, though significant at the level of the sample, is undetectable in the
spiked sample. For example, a 10 ppb suppression on a 10 ppb spike will cause a 0% recovery
160
-------
and corrective action will be necessary. That same suppression on a 100 ppb spike will yield a
90% recovery and the sample data will be reported down to the IDL. The data in this database
does not suggest that such a mechanism commonly presents any difficulties (spike concentrations
up to 60 ppb were evaluated), so the suggested spike concentrations should be acceptable.
In addition to increasing the PDS concentration, another simple change can be made to improve
th furnace AAS procedure. Currently, when the PDS recovery is out of the acceptance criteria, the
sample is diluted, reanalyzed, flagged ("E" or "W"), or reanalyzed by the method of standard
additions (MSA). The institution of reanalyses and dilutions for samples that would otherwise have
been flagged as "W" or analyzed by MSA will further reduce the time and number of injections
involved in GFAA analyses by resolving analytical difficulties in ways that are less labor intensive
and more appropriate. The GFAA database shows that a significant proportion of poor recoveries
can be remedied simply by repeating the analysis (an indication that the original recovery was out
of criteria due to measurement variability). Although poor recoveries that result from matrix effects
can be handled by MSA, this is a time consuming procedure that is often unnecessary. An
alternative to MSA for samples with analyte concentrations well above the IDL is simple dilution and
reanalysis, which is much less labor intensive. Dilution is often effective because the matrix
components causing the interference are diluted to insignificant levels while the analyte is
maintained above the IDL.
Revised QC Procedure
Based upon the changes recommended above (single injections, increase spike concentrations,
and alternate corrective action), the overall furnace AAS QC procedure should be revised. A
suggested "Furnace AA Analysis Scheme" is pictured in Figure 2. Each step is discussed below.
The functional changes to the tree consist mainly of eliminating the double injection requirement,
allowing repeat analyses to compensate for suspected "bad" measurements and random error, and
permitting the dilution of sufficiently concentrated samples as an alternative to MSA. These
modifications reduce time, injections, and expense while maintaining the same level of quality
assurance as is currently in place.
[1] Prepare a post-digest analytical spike for every sample except the calibration QC samples
(ICV, ICB, CCV, CCB, CRA) and analyze along with the unspiked sample. The required
spiking concentrations are listed below. The spiked sample can be prepared by the
furnace AAS instrument directly in the furnace tube or manually by the operator. To
prepare in an automated fashion using the instrument, consult the operator's manual. To
prepare manually, add a known quantity of the analyte to an aliquot of the digested sample
and the same quantity of acidified ASTM Type II water to another aliquot of the digested
sample. The volume of spiking solution must not exceed 10% of the sample aliquot
volume.
Post-digest analytical spike concentration (ppb)
As
40
Pb
40
Se
40
Tl
20
[2] The concentration of analyte in the spiked and unspiked samples must fall within the
calibrated range of the furnace AAS instrument. If not, the sample must be diluted,
respiked, and reanalyzed (refer to Step 8).
161
-------
Reanalyze sample
and spiked sample
[4c]
Dilute sample 5-1 OX,
respike, and reanalyze
[8]
Prepare and Analyze
Sample and One Spike
[1]
Dilute sample 5-10X,
respite, and reanalyze
I8]
NO
Report result* down to
Id-
Rag data with an "E".
Quantify from cal. curve.
Report down to IOL
Report result* down to
IOL.
Rag with a W.
Report MSA result*.
Rag data with"*"
Report MSA result*.
Rag data with "S"
Figure 2. Furnace AAS QC Flow Diagram
162
-------
[3a,b] If the analytical spike recovery is less than 40%, the sample must be diluted, respiked, and
reanalyzed once. If after dilution and reanalysis, the analytical spike recovery is still less
than 40%, the result is reported down to the IDL and flagged with an "E" to indicate matrix
interference problems (refer to Step 8).
[4a-c] If the analytical spike recovery is within the windows of 85-115%, the results by direct
quantification are "acceptable" and are reported down to the IDL. If the recovery is not
within the acceptance limits, the analyst has the option of reanalyzing the sample and spike
if this is the first analysis.
[5] If the analytical spike recovery is outside of the windows 85-115% and the sample
concentration is less than half of the spike concentration, the results are reported down to
the IDL and flagged with a "W". The "W" flag indicates that the analytical spike recovery
is out-of-control for unspecified reasons (eg, slight matrix problems or poor spiking
procedure). Because of the sample concentration, additional effort to resolve the problem
is not expected to result in a better number.
[6] If the analytical spike recovery is outside of the windows 85-115% and the sample
concentration is greater than half of the spike concentration and greater than 10 times the
IDL, the sample is diluted, respiked, and reanalyzed (refer to Step 8).
[7a-c] If the analytical spike recovery is outside of the windows 85-115% and the sample
concentration is greater than half of the spike concentration but less than 10 times the IDL,
the sample is quantified by the method of standard additions (MSA). Samples for MSA
analysis are prepared manually by the operator. Alternatively, the MSA aliquots can be
prepared in an automated fashion by the furnace AAS instrument if it has the capability.
In either case, the following steps must be incorporated into the MSA analysis.
a. Data from MSA calculations must be within the linear range as determined by the
calibration curve generated at the beginning of the analytical run.
b. The MSA analysis is performed by consecutively analyzing the sample and three
spikes.
c. The spikes must be prepared such that spike 1 is approximately 50% of the sample
concentration, spike 2 is approximately 100% of the sample concentration, and spike
3 is approximately 150% of the sample concentration.
d. The data for each MSA analysis must be clearly identified in the raw data
documentation using added concentration as the x-variable and absorbance (or found
concentration) as the y-variable along with the slope, x-intercept, y-intercept, and
correlation coefficient (r) for the least squares fit of the data. The results must be
reported on Form 8.
e. If the correlation coefficient (r) for an MSA analysis is less than 0.995, the MSA
analysis must be repeated once. If r is still less than 0.995, report the results of the
run with the greater correlation coefficient "r" on Form 1 and flag the result with a "+",
indicating that the results were obtained by MSA and that r was not acceptable. If r
is greater than 0.995, report the results and flag the result with a "S", indicating that
the results were obtained by MSA and that r was acceptable.
163
-------
[8] If called for by steps 2, 3, or 6, samples are diluted, respiked, and reanalyzed. Generally,
dilutions of 5-10 are acceptable. However, analyst judgement may be used to perform
other dilutions. However, the sample must not be diluted so that the analyte is less than
the IDL. If the sample is diluted below the IDL, it must be reanalyzed using a lower dilution
factor, if possible.
CONCLUSIONS
A review of the GFAA data from this laboratory confirms that PDS recoveries are adequate
indicators of data quality and that duplicate measurements are not necessary for the detection of
bad injections. Modifications are warranted, however, to ensure that the PDS provides meaningful
information. Current requirements are such that random measurement variability can cause a
significant proportion of all GFAA analyses to be needlessly flagged or reanalyzed by MSA.
Increasing the spike concentrations would alleviate the impact of random variability. Reanalysis
and dilution are additional features of this proposal that would provide laboratories with sufficient
latitude to deal with simple analytical problems so that data useability is not compromised and
without having to resort to MSA when another approach would be more appropriate. Since the data
quality would not be affected by these changes, the potential savings in time and cost as well as
increases in productivity should provide ample driving force to get these changes incorporated into
modern methodology.
164
-------
21
Th« Quality Control Level: An Alternative to Detection Levels
David E. Kimbrough*, Public Health Chemist, and Janice
Wakakuwa, Supervising Chemist, California Environmental
Protection Agency, Department of Toxic Substances Control,
Southern California Laboratory, 1449 W. Temple Street, Los
Angeles California 90026-5698.
Abstract
Existing methods for the determination of reporting
limits (e.g., MDL, LOD, etc) are based on a binary procedure
that determines that lowest concentration of an analyte that
is either detected or not detected within specified
confidence limits. There is no assessment of the accuracy or
precision of the results detected. An alternative procedure
is presented, the Quality Control Level which determines the
lowest concentration that meets the data quality objects of
the data user in term of the minimum acceptable precision and
accuracy.
To examine this hypothesis, a series of 15 soils and
aqueous liquids were prepared with successively smaller
concentrations of 16 toxic regulated elements. The range of
the concentrations were over four orders of magnitude for
both. Each of these liquids and soils were analyzed eight
times and the accuracy and precsion of each analyte was
measured against concentration. This paper will show that it
is possible to use a quality control approach to reporting
levels.
165
-------
22
DETECTION LIMIT
Kelvin L. Wright, Waste Policy Institute, Gaithersburg, MD
Farideh Moghadami, Science Applications International Corp. (SAIC), Gaithersburg, MD
Raymond J. Bath, Ph.D., Waste Policy Institute, Gaithersburg, MD
ABSTRACT
Over the past thirty years, the definition of detection limits has been molded and designed by
various groups to fit a specific need and purpose. The EPA, NRC, ANSI and instrument
manufacturers have all developed or adopted generic definitions for detection limits, and each
identifies a slightly different safety factor or component that should be included in the equations.
Since the end users of environmental data are often not chemists, this has lead to increased
confusion about detection limits. Generally, lawyers, engineers, hydrologists, and geologists are
the frequent end users of environmental data and lack the background to interpret the intended
meaning of "detection limit" This confusion and lack of understanding of detection limits can be
costly. This paper will examine the various definitions of detection limits and attempt to qualify
the differences between statistical approaches. It will also examine how selecting the proper
method for estimating detection limit on environmental restoration projects will minimize the
possibility of making incorrect decisions and possibly reduce sampling requirements.
166
-------
SAMPLING/FIELD
-------
24
EFFECT OF TRANSFORMER OIL, PETROLEUM HYDROCARBONS
AND INORGANIC SALTS AS INTERFERENCES IN FIELD SCREENING
FOR PCB CONTAMINATION IN SOIL
Alvia Gaskill. Jr.r President, Environmental Reference
Materials, Inc., PO Box 12527, Research Triangle Park,
North Carolina 27709
ABSTRACT
A study was conducted to determine the effect of
transformer oil and petroleum hydrocarbons on the accuracy
of field test methods commonly used to determine PCBs in
soil. Laboratory prepared soil samples spiked with Aroclor
1242 over the range 0-100 ppm were tested at both varying
and constant levels of transformer oil (0-10%), diesel fuel
oil (0-4%) and gasoline (1%). The effect of inorganic
chloride on the accuracy of the field test methods was also
determined. Samples were analyzed using draft SW-846
Method 4020, Soil Screening for Polychlorinated Biphenyls
by Immunoassay, by a method based on the L2000 PCB
Chloride Analyzer, and by gas chromatography. Testing at
action levels of 2 and 10 ppm Aroclor 1242, the L2000 and
GC methods correctly classified all soil samples as to
containing greater than or less than the action levels,
even in the presence of 2-10% transformer oil, 0.25-4%
diesel fuel and 1% gasoline. Method 4020 failed to
correctly classify such soils due to a negative
interference caused by the hydrocarbons. Because it isn't
possible to visually determine if a soil contains more than
2% transformer oil or 0.5% diesel fuel (the levels at which
Method 4020 began to fail to detect Aroclor 1242), Method
4020 cannot be used to rule out PCB contamination when
other hydrocarbons are present as is frequently the case at
waste dumps and spill sites. The full extent of the class
of compounds capable of causing this interference is not
known and should be the subject of future studies. Both
the L2000 method and Method 4020 were able to correctly
classify soil/salt mixtures containing 4 ppm Aroclor 1242
at the 2 ppm action level in the presence of up to 100%
sodium and calcium chloride.
INTRODUCTION
Soil can become contaminated with PCBs through accidents
involving the removal and maintenance of transformers and
capacitors or through improper disposal of PCB containing
substances. Accurate determination of the PCB content of
soils suspected to be contaminated is necessary in order
for the responsible parties to make the appropriate
decisions regarding site clean-up and remediation.
167
-------
Several proven laboratory analytical techniques have been
used for nearly a decade to meet this need. Most of these
have involved a gas chromatographic analysis of a
cleaned-up extract of the soil. More recently, field test
kits and portable laboratory systems using test kits have
become a popular way to identify PCB "hot spots" on-site in
suspected contaminated areas, thereby reducing the number
of samples requiring the more expensive and time consuming
laboratory tests and limiting the extent of soil excavation
required in cleaning up the site. These have included test
kits based on solvent extraction of the PCBs from the soil,
followed by chemical dehalogenation of the PCBs and
analysis by either colorimetric reaction or specific ion
electrode determination of the resulting chloride (the
L2000 PCB Chloride Analyzer of the Dexsil Corporation). A
method based on the L2000 kit, "Screening Test Method for
Polychlorinated Biphenyls in Soil" has been submitted to
EPA for approval.
Recently kits based on enzyme-linked immunosorbent assays
(ELISA) in which a competitive reaction between PCBs and a
PCB conjugate is used to determine the PCBs in a sample,
have been used. A kit^method based on one manufacturer's
ELISA kit (PCB RISc Soil Test System from Ensys, Inc.)
has recently received a de facto endorsement by the EPA as
draft Method 4020, Soil Screening for Polychlorinated
Biphenyls by Immunoassay (1). Similar immunoassay-based
(IA) kits have been developed and marketed by other
companies who are also seeking EPA approval of their kits.
However, there are at present no legal or regulatory
requirements to use any particular test kit or kit-based
method for the determination of PCBs in soil.
EPA's Office of Solid Waste Methods Hotline gives a
recorded message that Method 4020 should not be used for
regulatory purposes without the approval of a regulator.
It is important that these test kit methods be properly
validated so that regulatory agencies and industry will
know which ones to recommend or require to meet their
testing needs.
EXPERIMENTAL
Study Design
An examination of the published information on the
ELISA-based kits did not indicate whether transformer oil
above a given level would adversely affect the
determination of PCBs in a soil sample. Ensys has reported
(1, 2) that transformer oil, diesel fuel oil and gasoline
do not result in false positive interferences with their
168
-------
kits at levels greater than 1%, but the possible effect of
such hydrocarbons as negative interferents has not been
addressed. Baek (3) reported that a high concentration of
substances such as oil, may hinder PCB extraction from soil
by either saturating an extractant or blocking contact with
an extractant. To test for possible negative
interferences, laboratory generated soil samples
contaminated with transformer oil, diesel fuel oil,
gasoline and Aroclor were analyzed using this kit-based
method. Comparison tests were performed using ASTM Method
D3304 for PCBs in soils and by the method based on the
Dexsil L2000 field test system for PCBs in soil and oil.
Because the possible effect of inorganic chloride as an
interferent was not investigated by the EPA in their
validation of Method 4020 and is a possible positive
interferent in the L2000 method, laboratory generated soil
samples contaminated with inorganic chloride and Aroclor
1242 were also tested using the L2000 method and Method
4020.
The experiments were designed to simulate PCB contaminated
oil spills covering a range of PCB and transformer oil
concentrations as well as soils contaminated with diesel
fuel oil and gasoline to simulate soils found in
uncontrolled waste disposal sites or those associated with
leaking underground storage tanks. Inorganic chloride
interference experiments were designed to model several
scenarios: that of soil contaminated with salt water, that
of soil contaminated with road de-icing salt, and that of
pure salt taken as a sample to model the worst possible
case of inorganic chloride contamination in a sample.
All experiments involving IA testing were carried out using
IA kits with detection levels of either 2 or 10 ppm Aroclor
1242. Aroclor 1242 was the only PCB used in this study as
the IA kits used for this purpose must be ordered
calibrated by the manufacturer specifically for each
Aroclor. If a different Aroclor is tested than the one the
kit is calibrated for, the results can vary by more than an
order of magnitude.
Two types of hydrocarbon contaminated soils were prepared:
those contaminated with Aroclor 1242 only and those
contaminated with 1242 and transformer oil, diesel fuel oil
or gasoline. Levels of 1242 in the soil containing only
1242 varied from 0 to 100 ppm. The co-contamination
experiments involving transformer oil were conducted on
soils contaminated with 1242 at constant levels while
varying the concentration of the oil and at varying
concentrations of 1242 while keeping the oil concentration
constant.
169
-------
To simulate a field testing scenario, two action levels
were chosen: 2 and 10 ppm Aroclor 1242. All of the results
for the IA method were interpreted based on these pre-set
action levels. A total of 27 soil samples were analyzed
by each of the methods and the results compared to the
expected concentration of the Aroclor in the soil based on
gravimetric preparation. The total numbers of both correct
and incorrect classifications that resulted by using each
of the methods were tabulated and used to evaluate each
method. To avoid ambiguities due to possible sampling
variations, the results from the analyses of soils
contaminated at the action levels (i.e., 2 and 10 ppm) were
used only as an indicator of the precision of the
analytical methods. For concentrations other than the
action level, the data were used to determine the
percentage of correct determinations (i.e., whether or not
the PCB concentration was identified corrrectly as greater
or less than the action level).
The contamination levels used for the soil containing
Aroclor 1242 and no transformer oil were 0, 10, 20, 50 and
100 ppm. The first series of co-contaminant experiments
was conducted using soil to which 1242 had been added at
two times the 10 ppm action level (20 ppm) and to which
Shell Diala A transformer oil was added at 0.5, 1, 2, 4,
and 10 percent. Based on the results of the first series,
a second series of co-contaminant experiments was conducted
using soil containing transformer oil at 4% and 1242 levels
of 0, 10, 50, and 100 ppm. All of the above soils were
analyzed using the L2000 PCB Analyzer Method, the ASTM GC
Method and the PCB RISc kit method (4020) at an action
level of 10 ppm.
A third set of experiments was conducted at an action level
of 2 ppm 1242 using soils contaminated with 4% Diala A oil
as the co-contaminant. The 1242 levels were 2, 5, 10 and
2 0 ppm.
The contamination levels for the soil containing Aroclor
and diesel fuel or gasoline included levels of Aroclor 1242
of 4 ppm to which the diesel level was either 0.25 or
0.50%, levels of Aroclor 1242 of 10 ppm to which the diesel
level was 1 or 4% and levels of Aroclor 1242 of 20, 40, and
50 ppm to which the diesel level was 4%. Gasoline was
tested at a level of 1% with an Aroclor level of 10 ppm.
The contamination levels used to determine if inorganic
chloride is an interference in these methods involved
either samples of soil containing 10% or 100% sodium or
calcium chloride and no Aroclor 1242 or 4 ppm Aroclor 1242.
170
-------
Preparation of Spiked Soil Samples
To simulate typical soils, an 8 kg mixture of clay soils
and sand, approximately 75:25 w/w was prepared. The soils
and sand were obtained from residential areas and were
determined to contain less than 0.1 ppm Aroclor 1242 and
less than 1 ppm total organic chlorine. In addition, the
base soils were analyzed for total petroleum hydrocarbons
(TPH) using EPA Method 418.1 and found to contain less than
10 ppm TPH. The clay soil was broken up by hand and
allowed to air dry for 24 hours. The clay and sand were
sieved to pass a 0.850 urn sieve, mixed together, and
tumbled for 24 hours in a rotating pail. Most of the clay
particles were observed to be considerably smaller than
0.850 urn. The water content of the freely flowing mixture
was 1%. The soil was transferred to aluminum cake pans
prior to addition of contaminants. Subsampling to generate
homogeneous splits for spiking was performed according to
the procedure described by Schumacher (4) in which soil was
transferred from one pan to another in random order five
times. The soils were then spiked .with Aroclor 1242,
either in hexane or in Diala A (Shell) transformer oil to
generate known levels of the Aroclor and the oil in the
soil. The spikes were slurried with the soils and hexane
was added to facilitate mixing. The soil mixtures were air
dried to a constant weight, bottled in previouslMMunused
clean Qorpak 8 oz. glass bottles with Teflon lined
caps and tumbled for 4 hours on a rotating tumbler.
Samples containing diesel fuel were also prepared according
to this procedure. Samples containing gasoline or salt
were prepared individually by spiking samples of Aroclor
containing soil. Diesel levels in the spiked samples were
determined by Method 8015 to be >85% of the gravimetric
value.
RESULTS AND DISCUSSION
The antibody test upon which the ELISA-based kits rely
requires that the PCBs first be extracted from a 10 g soil
sample into methanol. The methanol extract is diluted with
more methanol, then an aqueous buffer, and is then added to
the antibody coated reaction tube. A solution of enzyme
conjugates is also added to the solution and the resulting
mixture is allowed to equilibrate. After the solution
phase is removed and the vessel washed, a color developing
agent is added. The greater the color obtained, the lower
the PCB content of the original sample. The test results
are interpreted by comparing optical densities (OD)
obtained for the sample with that of a standard. If the OD
of the sample is less than that of the standard, the sample
contains more than the level of Aroclor set for the kit by
171
-------
the manufacturer. If the OD is greater than that of the
standard, the sample contains less than the pre-set level
of Aroclor. Other concentration ranges can be measured by
serial dilution of the diluted sample or by use of kits set
to respond at other thresholds.
The specific IA kits evaluated in this study were set to
respond to Aroclor 1242 at 2 ppm and at 10 ppm. While the
distributors of the kit recommend that the user perform a
preliminary GC screen to identify the specific Aroclor
present, Method 4020 provides no such guidance. There have
also been at least five versions of the kit issued and two
drafts of Method 4020 since EPA's SW-846 Organic Methods
Working Group endorsed 4020 for PCBs in July 1992. Method
4020 is based on a 5 ppm threshold for PCBs and refers the
user to the manufacturer (in this case Ensys) for specific
instructions. The various generations of the kit have
added, in succession, a QC step which may disqualify an
entire set of analyses and a series of dilutions which
complicate the implementation of the test. The latest
version of the kit (Revision 5, 9/1/93) includes two QC
criteria which are contradictory. One calls for rejection
of test results if the optical densities obtained on
duplicate standards run with the samples differ by greater
than 0.20 absorbance units as read from the portable
spectrophotometer supplied with the kit. The other calls
for rejection if the results differ by more than 0.30
units. These changes make the evaluation of Method 4020
and the Ensys kit difficult and the comparison with
previously published validation data of concern. The
testing described in this study at the 10 ppm level for
transformer oil was based on the Ensys kit involving the QC
requirements and the double dilution of the sample extract
as described in the December .1992 information supplied by
the company (5) and the kits used were those set to expire
in August 1993. Testing at the 2 ppm level for transformer
oil, gasoline and diesel (1%) was based on the same kit
system, but without one of the dilution steps, as per the
instructions accompanying the kit. Thus, the evaluation at
the 2 and 10 ppm levels involves the same kit, but with a
different dilution of the sample extract prior to the IA
reaction.
The testing described in this study for evaluating the
effect of diesel fuel (all levels except 1%) and inorganic
chloride at the 2 ppm level was based on the Ensys kit
involving the contradictory QC requirements and the single
dilution of the sample as described in Revision 5 of the
kit. The kits used were those set to expire in April, 1995.
The Dexsil Corp. L2000 Chloride Analyzer System was used
according to the draft method and the instructions
172
-------
supplied with the unit (6). As was the case with the IA
method, a 10 g sample of soil is extracted with a solvent.
As such, both kit methods experienced equivalent sample
sizes and sample homogeneities. The extract is reacted in
plastic tubes with a sodium metal dispersion which
dehalogenates the chlorine from the PCBs and converts it to
water soluble chloride. The chloride is then measured by a
chloride specific ion electrode system included with the
kit. The electrode response is reported by the unit in
terms of either specific Aroclors or chloride. When the
specific Aroclor(s) of interest are unknown, the
instructions recommend that results be reported in terms of
Aroclor 1242 to cover the worst case likely when
transformer oil is present. The method is simple to use
and the reaction and measurement steps are not as time
dependent as for the IA method. While the IA method
requires strict adherence to specific reaction time
schedules set forth in the instructions, the L2000 method
seems less sensitive to reaction times so long as the
minimum times are met. The electrode system in the L2000
method requires frequent recalibration, which is prompted
automatically by the unit. A system blank, which is to be
run daily and subtracted from all results, is typically
around 1-4 ppm expressed as Aroclor 1242.
ASTM Method D3304, which involves a Soxhlet extraction of
the PCBs using iso-octane, followed by gas chromatographic
analysis of the extract for specific Aroclors, was used as
written. The method is essentially the same as EPA Method
3540A with iso-octane substituted as the extraction
solvent, followed by analysis by Method 8080 and was
developed specifically for determining PCBs in soil
contaminated with insulating fluids. Approximately 30-40 g
of each sample were extracted and analyzed in triplicate.
The effect of transformer oil as an interferent in the
determination of Aroclor 1242 concentrations is presented
in Table 1. Results of analyses of three replicate
aliquots from each sample by each kit method and by Method
D3304 (identified as "GC" in the table) are shown.
Recoveries by the GC and L2000 methods were higher in the
samples containing transformer oil than in those without
transformer oil, probably because the PCBs tend to remain
associated with the oil phase. When the oil is not
present, the PCBs are more likely to become adsorbed to
soil surfaces and thus more difficult to extract. While
neither the GC nor L2000 methods result in 100% recoveries
of the gravimetric levels of Aroclor 1242 added to the
soil, in most cases the 95% confidence intervals for
results by each include the gravimetric value, indicating
that both methods produce acceptable quantitative
173
-------
measurements of Aroclor 1242 in these samples. The mean
values for the GC and L2000 results differ significantly at
the 95% level of significance for about half the samples,
indicating that the GC and L2000 results cannot always be
considered identical.
When the results from the L2000 and IA methods are
compared with the expected values based on gravimetric
preparation, both methods correctly identified soil
containing 20 ppm 1242 as being above the action level of
10 ppm when the transformer oil concentration was less than
or equal to 1%. At transformer oil levels of 2% or greater
and 20 ppm Aroclor 1242, the IA method failed to detect the
Aroclor as >10 ppm. At 4% transformer oil the interference
due to the oil (when testing using a kit set to respond at
10 ppm) disappears when the Aroclor content increases to
>50 ppm. This interference is quite severe when it is
considered that the real threshold for the IA kits is
around 1 ppm (2). Thus, 95% of the response to PCBs must
be blocked for a false negative to occur. There were no
false negatives reported by the L2000 or GC methods.
Because it isn't always possible to simply look at a soil
and determine if it has >2% transformer oil, another means
must be found to identify such samples prior to testing
using this IA method in order to avoid excessive false
negative classifications.
Subsequent testing of soil samples containing 2, 5, 10 and
20 ppm Aroclor 1242 and 4% transformer oil using the IA kit
set to respond at 2 ppm, failed to show the presence of the
Aroclor at actual Aroclor levels as high as ten times the
action level of the kit. As the only difference between
the 2 ppm and the 10 ppm kit is the dilution made, it can
be concluded that the oil is once again responsible for the
interference. The L2000 method properly classified all of
these contaminated soils as did the GC method.
The effect of diesel fuel oil and gasoline as interferents
in the determination of Aroclor 1242 is presented in Table
2. While the L2000 and GC methods correctly classified
soils contaminated with 0.5-4% diesel fuel and gasoline at
1242 levels from 0-50 ppm, the IA method failed to
correctly classify these soils at 1242 levels up to 20
times the action level of the kit. Levels of diesel oil as
low as 0.25% interfered with the determination of 4 ppm
Aroclor 1242. At higher levels of oil (4%) the
interference persisted until the Aroclor concentration
reached 50 ppm. Once again, as the real threshold for the
kits is around 1 ppm PCBs, an interference at 40 ppm
suggests that 98% of the response to PCBs must be blocked
for a false negative to occur. Thus, the interference due
174
-------
to diesel oil is even more severe than that observed for
transformer oil. Similarly, because a visual observation
of a soil sample will not identify diesel contamination at
0.5%, the IA kits cannot be used to correctly classify the
Aroclor content of such samples.
Salt, either as sodium chloride (a model compound for salt
water contamination of a soil) or calcium chloride (a model
compound for road de-icing salt contamination of a soil)
had no adverse effect on the correct classification of the
Aroclor 1242 content of soils by either the L2000 or IA
method. The results presented in Table 3 demonstrate that
even with 100% salt, the kit methods were able to
effectively detect Aroclor 1242 at levels as low as 4 ppm.
This is especially important for the L2000 method as it is
based on a total organic chlorine measurement and includes
several steps to remove inorganic chloride from the soil
sample prior to the determinative step. Some difficulty
was experienced in filtering the 100% calcium chloride
methanol extract in the IA test due to its viscosity.
While it is unlikely that field samples containing 100%
salt will be tested using these kits, the results from this
study demonstrate that the incidental presence of inorganic
chloride from seawater or estuarine water contamination of
soils or from road de-icing salts, does not constitute a
serious interference.
CONCLUSIONS
The results from the L2000 PCB Analyzer Method were not
adversely affected by the presence of either transformer
oil or fuels in the samples. This suggests that other
hydrocarbons would also not present a problem in the
analysis of contaminated soils. The co-contamination of
the soil with transformer oil does appear to enhance the
efficiency of the L2000 method's extraction of PCBs from
soil. The GC results also show an extraction efficiency
enhancement due to the presence of transformer oil as a
co-contaminant. In neither of these two methods did the
extraction efficiency enhancements affect the accuracy of
the soil classifications. Inorganic chloride was also not
found to be an interference, even when 100% salt was
tested.
The IA method results, however, suggest that the
ELISA-based kits of Method 4020 suffer from a severe
negative interference due to transformer oil, diesel fuel
and gasoline, although not inorganic chloride, at
levels typically found in spill site and landfill samples.
The full extent of the class of compounds capable of such a
strong negative interference is not known and should be the
175
-------
subject of future studies. It is widely known that
immunoassays can be very specific in their positive
response to the analyte of interest. These experiments
have highlighted a less well advertised aspect of IA
systems in that they involve complex, large molecules and
may be susceptible to non-specific interferences,
especially when used with non-aqueous solutions. Reduction
of the interference by dilution alone may not be possible,
as the kits as presently configured are responsive over a
rather narrow range of parameters. Further dilutions on
the order of those predicted necessary to eliminate the oil
interference would likely render the kits insufficiently
sensitive to determine PCBs at the levels of interest to
government and industry. While the exact mechanism of the
interference observed here and its correction are beyond
the scope of this study, the EPA needs to fully investigate
the potential of all such possible co-contaminants to
produce false negative results. The EPA or the
manufacturer of the IA kits has, as yet, made no published
determination of the negative interference of any compounds
and should do so immediately. Oils and fuels represent
only a small fraction of the possible interferents likely
to be present in environmental soil samples. This will
undoubtedly require reanalysis of some field samples
already determined non hazardous by using draft Method
4020.
REFERENCES
1. Environmental Protection Agency, Office of Solid Waste.
EPA Method 4020, Soil Screening for Polychlorinated
Biphenyls by Immunoassay, Revision Draft 1, October
1992. Washington, D.C.
2. N. Stewart, R.M. Mudd, J.P- Mapes, P.A. Reddy, W.B.
Manning, W.B. Studabaker, K.D. McKenzie, L.R.
McClelland, and S.B. Friedman, "PCB-RISc-An On Site
Immunoassay for Detecting PCB in Soil." In: EPRI PCB
Seminar, October 8-11, 1991, Baltimore, Maryland, EPROI
TR-100503 May 1992 pp. 49-1 to 49-4.
3. N.H. Baek, Evaluation of Immunoassay Tests in Screening
Soil Contaminated with Polychlorinated Biphenyls,
Bulletin of Environmental Contamination and Toxicology,
1993. 51, pp. 844-851.
4. B.A. Schumacher, K.C. Shines, J.V. Burton, and M.L.
Papp. "A Comparison of Soil Sample Homogenization
Techniques." in "Hazardous Waste Measurements." M.S.
Simmons, ed. Chelsea, Michigan: Lewis Publishers, 1991,
pp. 53-68.
176
-------
5. PCB RISc Soil Test System User's Guide, Ensys, Inc.
1993.
6. Instructions for the L2000 PCB/Chloride Analyzer, Dexsil
Corp. January 1991.
Table 1. Comparison of GC, L2000 and IA Kit Method Accuracy
in Classifying Aroclgrb1242 Levels in Soils Contaminated
with Transformer Oil '
Aroclor 1242,
Gravimetric
Value, ppm % oil
correct
Method Results,
GC L2000
ppm
GC &
L2000
10 ppm action level
IA
0 0
10 0
20 0
50 0
100 0
0.110.04
8.712.4
15.111.7
40.511.8
1.410.6
9.5+1.2
13.412.3
40.314.1
77.312.0
78.912.3
100
NE
100
100
100
100
NE
100
100
100
20
20
20
20
20
0
10
50
100
2
5
10
20
0.5 16.112.8
1 17.310.9
2 17.311.5
4 18.011.0
10 17.312.7
4
4
4
4
4
4
4
4
18.911.5
19.511.7
23.210.9
21.510.7
25.511.9
0.110.04 0.010.0
9.610.2 8.310.4
46.114.6 49.810.2
92.714.7 105.813.0
100
100
100
100
100
100
NE
100
100
2 ppm action level
1.8±0.1 2.810.6 < 2 NE
4.310.2 5.210.7 < 2 100
8.310.1 9.410.2 < 2 100
18.011.0 21.510.7 < 2 100
100
100
0
0
0
100
NE
100
100
NE
0
0
0
f;all results uncorrected for % water
mean 1 1 standard deviation for triplicate GC and L2000
determinations; all three results from each IA method
determination agreed in all cases
NE: not evaluated because expected value is the action
level
177
-------
Table 2. Comparison of GC, L2000 and IA Kit Method Accuracy
at the 2 ppm Action Level in Classifying Aroclor 1242 Levels
in Soil Contaminated with Diesel Fuel and Gasoline ' '
Aroclor 1242,
Gravimetric Method Results, ppm % Correct
Value, ppm % Diesel GC L2000 IA GC L2000 IA
0 00 0.0±0.0 <2 100 100 100
4 0.25 2.9 3.7±0.8 >2,<2 100 100 50
4 0.5 3.0 3.9±0.9 <2 100 100 0
10 1 NA 9.0±0.1 <2 NA 100 0
10 4 9.1 11.0±1.6 <2 100 100 0
20 4 16.4 20.3±1 .4 <2 100 100 0
40 4 NA 41.2±0.4 <2 NA 100 0
50 4 37.7 53.1±5.9 >2 100 100 100
10 (gasoline) 1 NA 9.2±1.4 <2 NA 100 0
.duplicate determinations
all results uncorrected for % water
Cmean ± 1 standard deviation for L2000 results; both results
from each IA method determination agreed unless otherwise
noted
NA: not analyzed
Table 3. Comparison of L2000 and IA Kit Method Accuracy at
the 2 ppm Action Level in Classifying Aroclor 1242 Levels
in Soils Contaminated with Inorganic Salts
Aroclor 1242,
Gravimetric Method Results, ppm % Correct
Value, ppm %Salt L2000 IA L2000 IA
0 10% NaCl o£ <2 100 100
0 10% CaCl? Of* <2 100 100
0 100% NaCl 0 <2, 100 100
0 100% CaCl? 1.0 <2 100 100
4.1 10% NaCl 4.4 >2 100 100
4.2-4.6 100% NaCl 10.0 >2 100 100
4.1 10% CaCl- 3.8 >2, 100 100
4.3-7.0 100% CaCl^ 9.8C >2a 100 100
, single determinations unless otherwise noted
duplicate determinations
,expected value 7.0
expected value 4.3
178
-------
25
FIELD TESTING OF A PORTABLE TRICHLOROETHYLENE AND
CHLOROFORM FIBER OPTIC CHEMICAL SENSOR
James C. Wells and David A. Gobeli
Purus, Inc.
2713 North First Street
San Jose, CA 95134-2010
Purus, Inc. has developed and field tested a portable fiber optic chemical sensor for
the semi-specific determination of ppb levels of trichloroethylene and chloroform in water,
soil and gaseous samples. This sensor configuration is an extension of our original
laboratory fiber optic chemical sensor and allows the sensor to be used as a field screening
device. The sensor is contained in a carrying case measuring 46 x 30 x 20 cm. and
weighing less than 10 kg.
The sensor consists of an flow optrode, reagent delivery and recovery system, fiber
optic transmitter-receiver, embedded micro controller, display, and communication port.
The optrode is a miniature reaction chamber through which the chemical reagents are
pumped. The reagents react with gaseous halogenated compounds that diffuse in through a
gas permeable membrane to form a colored product, and the product is detected by its
absorbance of light from a 560 nm diode. The reagents are based on the Fujiwara (K.
Fujiwara, Chem. Abstr. 11:3201 (1917)) alkaline pyridine chemistry and optimized to
measure trichloroethylene or chloroform. The minimum analysis time is 2.5 minutes at the
detection limits of a few ug/L in water, and may be shortened at higher concentrations or
by further refinements to the hardware.
Recent field trial results for chloroform in drinking water have demonstrated
detection limits of 3 ug/L and repeatibility of ± 5% at the 40 ug/L level. The results of field
monitoring trials for TCE contaminated groundwater will be presented that demonstrated a
detection limit of 2 ug/L Field trial results for TCE contaminated soil samples demonstrate
a detection limit of 5 ug/kg. Quantitation results will be presented that demonstrate the
viability of this fiber optic chemical sensor as low cost screening tool for site assessment,
monitoring, and process control applications.
179
-------
26
IMPROVEMENTS IN SAMPLE RECOVERY OF XAD-2 RESIN FROM
METHOD 0010 TRAINS
Merrill D. Jackson and Larry D. Johnson, Atmospheric Research and
Exposure Assessment Laboratory, U. S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711, James F. McGaughey,
Raymond G. Merrill, Jr., Joan T. Bursey, and Denny E. Wagoner, Radian
Corporation, P.O. Box 13000, Research Triangle Park, North Carolina 27709
ABSTRACT
In a field evaluation study for semivolatile halogenated organic compounds
listed in Title III of the Clean Air Act Amendments of 1990, dynamic spiking
experiments using a liquid solution were performed in the field. Two of four
quadruple sampling trains were spiked for eight sampling runs. Method 0010
train components were prepared and analyzed in three parts: filter/front half
rinse, XAD-2® resin, and condensate/condensate rinse. In sixteen spiked trains,
spiked analytes were detected with reasonable recoveries (> 50%) in only four
runs. In general, surrogate compounds spiked during preparation of the
samples showed low recoveries from XAD-2®, and recoveries of spiked analytes
which were observed ranged from 4 to 63 percent. Because these results were
at variance with results obtained for analytes spiked in laboratory studies and
a previous field study, the sample preparation process was investigated in
detail. Sample preparation procedures had followed Method 0010, but use of
some procedures which were not specifically prohibited by Method 0010 had
depressed compound recoveries. Laboratory studies were performed to
evaluate the effects of various sample preparation parameters on compound
recoveries. To ensure thatthe sample preparation procedures for Method 0010
train components were clear and unambiguous, a new protocol to address
preparation of Method 0010 train components for Method 8270 analysis was
written. The new protocol has been used in a subsequent field study with
excellent results.
INTRODUCTION
In order to evaluate the performance of SW-846 Method 0010 for sampling
and Method 8270 for the analysis of semivolatile halogenated organic
compounds listed in Title III of the Clean Air Act Amendments of 1 990, a field
study was performed using dynamic spiking techniques to establish the
precision and bias of the overall methodology. Using the guidelines of EPA
Method 301 (Protocol for the Field Validation of Emission Concentrations from
Stationary Sources) for statistical design of the field testing experiments,
180
-------
quadruple Method 0010 sampling trains with four collocated probes were used.
Dynamic spiking equipment and procedures had been developed and evaluated
to allow dynamic spiking of a methylene chloride solution of the compounds of
interest for the duration of each Method 0010 sampling run. According to the
guidelines of Method 301, two trains were spiked and two trains were
unspiked.
EXPERIMENTAL
The field evaluation study was conducted at a chemical manufacturing
facility where waste chemicals were incinerated in a coal-fired boiler. A
"biosludge" consisting of 10 percent organic matter and 90 percent water was
fed continually to the incinerator. A site presurvey, when preliminary samples
were taken, showed that none of the proposed analytes was present in the
background emissions from the boiler, and that the emissions were wet
(approximately 10 percent moisture). Method 0010 sampling trains were
recovered in the field, and components were shipped to the laboratory for
preparation and analysis. Extracts (three per sampling train) were generated
from methylene chloride extractions of the following train components:
• Filter/front half rinse;
• XAD-2® sampling module; and
• Condensate/condensate rinse.
The final extract volume for these sampling train components was 5 mL,
rather than the 1 mL final volume specified by Method 8270.
Results for the GC/MS analysis are summarized in Table I. To perform a
thorough statistical analysis according to Method 301 procedures, results from
six paired spiked runs are required. Eight sampling runs using quadruple trains
had been performed in the field; acceptable results were obtained for only four
runs (1,2,3,6). For those four runs, for most compounds results appear
generally comparable to laboratory and field results obtained previously
(Table II). However, results from other sampling runs showed very low
recoveries for the surrogate compounds and many of the spiked compounds
were not detected.
RESULTS AND DISCUSSION
Careful examination of the data for all of the sampling runs showed that, in
general:
181
-------
• Recoveries of the surrogate compounds spiked in the laboratory were
low for the XAD-2®, where most of the organic compounds were
expected to be retained;
• Isotopically-labeled compounds spiked in the laboratory to track
recovery were frequently not observed at all; and
• The majority of the analytes spiked in the field were not observed.
Recoveries for field-spiked analytes that were observed ranged from
4 percent to 63 percent.
Since the surrogate compounds and isotopically-labeled compounds are
spiked in the laboratory after return of the sampling train components, problems
were obviously encountered in the laboratory preparation rather than in the field
spiking.
The critical parameter is recovery of spiked compounds from XAD-2®.
Recovery results for these field samples were sufficiently at variance with
previous recovery results from a laboratory study1 and a field study2 that an
explanation for the low recoveries was pursued. Quality Control results from
Method Blanks were examined. Method Blanks consist of sampling train media
(filters, water, solvents, XAD-2®) that are spiked with surrogate compounds in
the laboratory, extracted, and analyzed. Recoveries from Method Blanks were
acceptable to high, indicating that general laboratory sample preparation and
analysis procedures were done properly.
Method Spike recovery data were also examined. Method Spikes consist of
train components spiked with analytes and surrogate compounds in the
laboratory. The Method Spikes are extracted and analyzed with the field
samples. The results obtained for the XAD-2® Method Spikes are typical
(Table III): acceptable to high recoveries indicated that surrogate and sample
spiking, preparation, and analysis procedures were in control.
From an examination of the Quality Control samples, we concluded that a
systematic error in sample spiking, sample preparation, or analytical procedures
did not appear to be the cause of the low recoveries: Method Blanks and
Method Spikes were prepared and analyzed with the field samples, using the
same spiking solutions and the same procedures. The original extracts, which
had been archived after mass spectral analysis, were next examined visually to
determine if the appearance of these extracts was qualitatively or quantitatively
different from the appearance of the Quality Control samples. Several key
differences were observed:
182
-------
• Method Blanks and Method Spikes were light yellow in color and had
the appearance of several ml of clear organic solvent. The color of
field sample extracts ranged from clear to nearly brown.
• Some of the field extracts were clearly completely aqueous, with only
small pools of organic liquid floating on top;
• Two phases were clearly visible in some of the field extracts; and
• Many of the field samples were not methylene chloride extracts, since
only a slight odor of methylene chloride was detected when vials
were opened.
Laboratory sample preparation procedures and observations were carefully
reviewed with laboratory staff. The observation was reported that many of the
field samples required far longer (3-4 hours) than the usual amount of time
(20-30 minutes) to achieve concentration to 5 ml using Kuderna-Danish
concentration procedures.
The obvious difference between the Quality Control samples and the field
samples was that the laboratory-generated sampling train media were dry, while
the field XAD-2® samples were wet because of the moisture content of the
source. Dry XAD-2® can simply be poured from the sampling module to the
Soxhlet extraction apparatus. Wet XAD-2® does not pour: the wet resin sticks
to the glass walls of the sampling module and is not readily moved from the
sampling module with methylene chloride rinses. Typical procedures used for
the removal of wet XAD-2® from the sampling module include repeated rinses
with methylene chloride, which frequently leaves significant amounts of the wet
XAD-2® in the sampling module, or tapping the sampling module against the
laboratory bench top, which often results in breakage of the sampling module.
Laboratory staff had tapped the XAD-2® from the modules to remove as much
as possible, rinsed the walls of the module with methylene chloride to remove
as much of the remaining wet XAD-2® as possible, and performed a final rinse
of the sampling module with methanol to remove all of the remaining XAD-2®.
If a sufficiently large amount of methanol is present when sample concentration
is performed, methylene chloride will be driven off rather than methanol, and
the final extract will consist of a methanol solution with significant losses of
surrogate compounds and analytes.
The rinses used in the field recovery of Method 0010 train components
consist of 50:50 methylene chloride: methanol, which form a homogeneous
solution. The methanol can be separated from the methylene chloride only if
sufficient water is added to create two distinct phases. However, 100 mL of
183
-------
methylene chloride can hold up to 15 ml of water without separating into two
distinct phases. According to the method, sample extracts are dried by filtering
through a bed of dry sodium sulfate. If sufficient water is present, the sodium
sulfate will cake and will not dry the extract efficiently. Thus, after drying, if
the sodium sulfate cakes, an extract may consist of methylene chloride, water,
and methanol, all in one phase. If a solution of this composition is
concentrated, methylene chloride will be lost before the water and methanol are
lost, resulting in a water/methanol solution if sufficient quantities of water and
methanol are present in the original extract. However, if sufficient water
(50-100 mL) to effect separation of phases is added prior to extraction, the
methanol will be driven into the aqueous phase and excellent recoveries of
spiked surrogate compounds and analytes can be obtained.
Laboratory experiments were conducted to reproduce the conditions under
which the field samples had been extracted. Replicate samples of dry XAD-2®
were spiked with surrogate compounds and analytes to provide a baseline for
recovery. Excellent recoveries and good reproducibility were obtained. Next,
wet XAD-2® was prepared and spiked with surrogate compounds and analytes.
The 40 g quantity of XAD-2® which is contained in the sampling module of the
Method 0010 train retains approximately 50 mL of water when water is poured
through the resin bed. This 50 mL of retained water does not produce a
distinct water layer when the spiked wet XAD-2® is extracted and analyzed.
When the extracts from the wet XAD-2® were concentrated and analyzed,
recoveries were slightly lower than the recoveries obtained with dry XAD-2®
and reproducibility was slightly poorer, but both recovery and reproducibility
were acceptable. The wet XAD-2® was prepared and spiked in the Soxhlet
extractor, so no transfer of wet XAD-2® was required. Wet XAD-2® alone does
not depress recoveries significantly.
The major problem appeared to occur in the transfer of the wet XAD-2®. A
procedure was therefore developed to transfer the wet XAD-2® without the use
of methanol. The apparatus shown in Figure 1 is used to transfer the XAD-2®
if the resin is too wet to pour. The glass wool is removed from the end of the
sampling module and placed in the Soxhlet extractor to ensure extraction. A
small piece of pre-cleaned glass wool is placed in the arm of the Soxhlet
extractor to ensure that no XAD-2® enters the side-arm. The XAD-2® sampling
module is inverted (glass frit up) over the Soxhlet extractor, approximately 5-10
mL of methylene chloride is added above the glass frit, and air pressure created
by squeezing the rubber bulb shown in Figure 1 is used to gently but firmly
push the methylene chloride through the frit, forcing the XAD-2® out of the
sampling module. This process is repeated 3 to 5 times, and a Teflon® wash
bottle containing methylene chloride is used to rinse the walls of the sampling
module to transfer XAD-2® which adheres to the walls of the sampling module.
184
-------
After 3-5 methylene chloride rinses, no more than a monolayer of XAD-2®
usually remains in the sampling module. This XAD-2® transfer procedure has
been used successfully to transfer XAD-2® from sampling modules used in
sampling a source with 55 percent moisture: excellent recoveries of both
surrogate compounds and spiked analytes were obtained. In addition, this
procedure is far more efficient than the procedure of tapping the resin out of
the sampling module: three transfers using the rubber bulb can be performed
in one or two minutes.
The investigation with subsequent laboratory study illustrates the value of
sufficient Quality Control data in determining the cause of a problem with data
quality. A new procedure for the preparation of Method 0010 train components
for analysis by SW-846 Method 8270 has been written. A flowchart for the
overall method is shown in Figure 2. In this procedure, the use of methanol in
the laboratory is directly and specifically prohibited to ensure that the final
extracts consist of methylene chloride, not a mixture of methylene chloride and
methanol. Also, addition of sufficient water to ensure that two distinct phases
are produced when both water and methanol are components of the solution
(for example, in the sampling train rinses of the front half and the condensate)
is a required part of the procedure. This procedure is being subjected to EPA
review.
REFERENCES
1. Laboratory Validation of VOST and SemiVOST for Halogenated
Hydrocarbons from the Clean Air Act Amendments. Volume 1 and 2. EPA
600/R-93/123 and b. NTIS PB93-227163 and PB93-227171. U. S.
Environmental Protection Agency. July, 1993.
2. Field Test of a Generic Method for Halogenated Hydrocarbons.
EPA 600/R-93/101. NTIS PB93-212181. U. S. Environmental Protection
Agency. June, 1993.
DISCLAIMER
The information in this document has been funded wholly by the United
States Environmental Protection Agency under contract 68-D1-0010to Radian
Corporation. It has been subjected to the Agency's peer review and
administrative review, and it has been approved for publication as an EPA
document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
185
-------
Table I
Summary of Results for All Eight Runs and All Sampling Trains,
Using Surrogate-Corrected Data
Run
1
2
3
4
5
6
7
8
Train A
Sp&«tf
X
y
y
y
n
Z
y
n
n
€
y
y
y
y
y
y
n
y
F
y
y
y
n
y
n
n
Z
Train B
SpiKed
H!
y
y
y
n
Z
y
y
y
C
y
y
y
y
y
n
y
y
V
y
y
y
n
n
n
y
y
Train C
tln$pi1ted
X
y
y
y
y
y
z
z
z
C
y
y
y
y
y
y
y
y
r .
y
y
y
n
y
y
z
y
Train D
Dhspiked
X
y
n
y
y
y
z
y
y
C
y
y
y
y
y
y
y
y
F
y
y
y
y
y
y
z
z
Note: Recoveries for C and D Trains refer to recoveries of surrogate compounds and isotopically-labeled analogs.
X = XAD-2® module.
C = Condensate fraction.
F = Filter fraction.
Z = Partial success; some but not all analytes detected.
y = All analytes detected.
n = No analytes detected.
186
-------
Table II
Comparison of Percent Recoveries of Semivolatile Halogenated
Organic Target Compounds in Laboratory and Field Studies
(Uncorrected for Surrogate Recoveries)
Compound
Bis(chloromethyl)ether
Epichlorohydrin
cis- 1 ,3-Dichloropropene
trans- 1 ,3-Dichloropropene
1 , 1 ,2-Trichloroethane
1 ,2-Dibromoethane
Tetrachloroethene
Chlorobenzene
Bromoform
1 , 1 ,2,2-Tetrachloroethane
Dichloroethyl ether
1 ,4-Dichlorobenzene
Benzyl chloride
Hexachloroethane
1 ,2-Dibromo-3-chloropropane
1 ,2,4-Trichlorobenzene
Hexachlorobutadiene
Benzotrichloride
2-Chloroacetophenone
Hexachlorocyclopentadiene
2 ,4 ,6-Trichlorophenol
2,4,5 -Trichlorophenol
Hexachlorobenzene
Pentachlorophenol
Pentachloronitrobenzene
Chlorobenzilate
3,3' -Dichlorobenzidine
Mean R«s»lts
Laboratory1
18.3
75.2
21.9
20.4
53.1
66.3
49.7
76.0
99.3
81.1
75.8
68.2
78.7
85.4
66.2
58.2
58.3
67.0
79.7
513.0
45.6
52.7
32.9
8.9
38.2
43.6
86.4
Field 1*
0.0
6.0
49.1
52.0
56.4
58.9
53.2
62.3
59.8
64.0
60.9
56.2
67.4
74.0
44.8
59.5
65.4
60.1
56.0
42.3
49.8
62.7
44.6
42.4
43.4
40.7
4.4
Field 2*
0.0
13.4
50.3
79.8
60.3
62.5
49.4
65.1
69.3
73.9
77.0
73.5
73.9
70.9
73.8
76.1
77.1
72.4
79.5
59.6
75.4
76.6
56.5
60.3
58.5
61.8
0.6
'Mean of 16 replicates.
2Mean of 12 replicates.
3Mean of 4 replicates.
187
-------
Table III
Spiked Compounds and Surrogates Recovered
from Dry Method 0010 XAD-2® Traps
CaropoujwJ
Surrogate
2-Fluorophenol
Phenol-d;
Nitrobenzene-d5
2-Fluorobiphenyl
2,4,6-Tribromophenol
Terphenyl-d,4
Epichlorohydrin-d5
Chlorobenzene-d5
1 , 1 ,2,2-Tetrachlorotthane-d2
Bis(chloroethyl)ether-d8
Benzyl chloride-d7
2,4,5-Trichlorophenol-d2
Targets
Epichlorohydrin
cis- 1 ,3-Dichloropropene
trans- 1 ,3-Dichloropropene
1 , 1 ,2-Trichloroethane
1 ,2-Dibromoethane
Tetrachloroethene
Chlorobenzene
Bromoform
1 , 1 ,2,2-Tetrachloroethane
Bis(chloromethyl)ether
1 ,4-Dichlorobenzene
Benzyl chloride
Hexachloroethane
1 ,2-Dibromo-3-chloropropane
1 ,2,4-Trichlorobenzene
Hexachlorobutadiene
Benzotrichloride
2-Chloroacetophenone
Hexachlorocyclopentadiene
2,4,6-Trichlorophenol
2,4,5-Trichlorophenol
Hexachlorobenzene
Pentachlorophenol
Pentachloronitrobenzene
Chlorobenzilate
3,3'-Dichlorobenzidine
Theoretical
Amount
(wr>
991
1010
509
490
997
501
250
350
254
333
244
129
0*8)
199
159
34
195
196
195
200
202
200
252
226
202
185
272
198
200
199
229
204
237
194
222
202
216
200
190
% $te<*>wry
MS-A
107
112
112
119
67
135
99
94
114
104
103
ND
MS*B
99
106
95
115
74
112
68
91
93
91
122
ND
MS-C
108
113
104
122
73
115
76
106
99
95
130
106
MS4*
102
108
98
111
66
108
71
93
91
87
117
ND
% Recovery
991
87
365
98
95
86
99
101
101
80
96
102
107
103
104
107
106
112
135
109
101
102
83
101
116
142
68
67
77
77
84
82
92
104
84
70
119
95
103
109
120
126
126
108
133
121
127
110
100
106
110
140
72
71
80
84
94
92
96
120
91
72
125
105
112
118
132
139
141
116
133
129
130
124
87
113
123
171
74
76
86
86
95
92
100
127
92
74
131
104
114
121
135
148
142
120
133
129
139
121
54
114
130
158
188
-------
Rubber
Bulb
MeCJ2is added to XAD-2® Trap
Glass
Frit
XAD-2®
Glass Wool
Soxhlet
Free leaned
Glass
Wool
Round
Bottom
Flask
Teflon®
Tube
Sovirel®
Fitting
Ground Glass
Ball Joint
Figure 1. Transfer of Wet XAD-2®
189
-------
XAD-2®
(Container 3)
Spike with Surrogates (and
Isotopically-Labeled Analogs)
Soxhlet Extraction
XAD-2®
Extract
Add Sufficient Water
to Separate into Two
Phases; Separate
Extract Water Layer with
CH2CI2; Adjust pH and do
Acid/Base or Base/Acid Extraction
Combine CH2CI2 Extracts
I Remove Moisture
| with Na2S04
Rinse all of Glassware Between
Back Half of Filter Holder and
XAD-2® (Filter Holder Back Half
Connector, and Condenser)
with CHjCyCHpH
(Container 5)
Impinger Contents
(Impingers 2 and 3)
Archive
Silica Gel
(Impinger 4)
(Container 6)
Weigh in the Field
Regenerate
Re-use
Concentrate to 5mL
Analyze by GC/MS
Figure 2. Sample Preparation Scheme for Method 0010 Train Components
190
-------
Condeneate H
(Impingerl)
(Container 4)
Rinse of Impinger 1
CH2CI2/CH3OH
Combine
Participate Matter
Filter
(Container 1)
Spike with Surrogates and
iBOtopically-Labeled Analogs
Spike with Surrogates,
Itotopically-Labeled
Analogs
Separately Funnel Extraction
(Add H2O if necessary to
separate phases)
Front Half Rinse,
Front Half of Filter
Holder, Probe and Nozzle
CH2CI2/CH3OH
(Container 2)
Filter; Add Filter to
Paniculate Matter Filter
Soxhtet Extraction
CH2CI2
Extract Water Layer with
CH2CL,; Adjust pH and do
Acid/Base or Base/Acid Extraction
Separate
CH2CI2
Extract
Separatory Funnel
Extraction of Filtrate
(Add HjO if necessary to
separate phases)
Extract Water Layer with
CH2Cl2; Adjust pH and do
Acid/Base or Base/Acid Extraction
Combine CH2CI2 Extracts
Combine CH2CI2
Extracts
Remove Moisture with
Na2S04
Save CH £l 2 Layer
(Bottom)
Remove Moisture
with Na2SO4
Concentrate to 5mL
Concentrate to 5mL
Analyze by GC/MS
Analyze by GC/MS
Figure 2. (Continued)
191
-------
27
EVALUATION OF A NEW NON-IMMUNOASSAY FIELD TEST KIT FOR
TOTAL PETROLEUM HYDROCARBONS IN SOIL
T.B. Lynn, S. Finch, L. Sacramone, Dexsil Corporation, One Hamden Park Drive,
Hamden, CT 06517 and K.A. Wright, 2094 Hideaway Ranch Road, Placerville CA
95667
Abstract
A new TPH field test has been evaluated. Soils spiked with known amounts of both
diesel and number 6 fuels were used to determine the relative response factors for the
two different analytes as well as the method precision and bias. The results indicate that
the response factor for number 6 fuel is 90% of the diesel fuel calibrator. Very little
bias was found and the precision is better than +/- 10% for both analytes.
Introduction
A novel new analytical procedure has been developed to determine the hydrocarbon
content of soil samples using environmentally safe reagents which will be simpler and
less expensive than alternative methods. The data presented here illustrates the
effectiveness of the new Petro-Flag™ technology on two analytes; diesel fuel and number
6 fuel oil.
The Petro-Flag technology has been in use in the field in a non-commercial form for over
two years. A commercial version will be available soon for use in the field by
environmental professionals. The colorimetric test is easy to use and contains no
hazardous chemicals. A specially designed hand-held colorimeter will be available to
provide a digital readout in ppm of the analyte. Using the prepackaged reagents 10 to
20 samples can be run in one batch in under 30 minutes. The anticipated cost per test
is $10 to $15 and the colorimeter will cost under $300.
The patent pending kit chemistry relies on a unique system of extraction solvents and
color-forming reagents. Because of its broad linear response range the Petro-Flag test
kit can be used on a wide variety of hydrocarbon analytes including fuels, lubricants,
hydraulic fluids and greases. It does not only test for specific compounds or aromatics
but all petroleum hydrocarbons which makes the kit useful as a low cost general
screening tool as well as for quantitative determination of TPH concentrations is easily
192
-------
performed at contaminated sites requiring lateral and vertical definition, as well as for
soil remediation projects to monitor the status of the remediation and for determining
when samples should be taken for expensive laboratory analysis for final closure. The
field test is especially useful during underground storage tank removals to screen the
excavated area during the excavation and prior to collecting confirmatory samples. The
test uses field calibration standards to achieve a high degree of accuracy in a large
variety of soil types.
The PETRO-FLAG technology is currently in the beta-testing stage. Evaluations on
different soil types with different analytes indicates that the extraction efficiencies are
very high for most petroleum hydrocarbon contaminants.
Preparation of Spiked Soil Samples
To simulate typical soils, an 8 kg mixture of clay soils and sand, approximately 75:25
w/w was prepared. The soils and sand were obtained from residential areas and were
analyzed for total petroleum hydrocarbons (TPH) using EPA Method 418.1 and found
to contain less than 10 ppm TPH. The clay soil was broken up by hand and allowed to
air dry for 24 hours. The clay and sand were sieved to pass a 0.850 um sieve, mixed
together, and tumbled for 24 hours in a rotating pail. Most of the clay particles were
observed to be considerably smaller than 0.850 um. The water content of the freely
flowing mixture was 1%. The soil was transferred to aluminum cake pans prior to
addition of contaminants. Sub-sampling to generate homogeneous splits for spiking was
performed according to the procedure described by Schumacher (4) in which soil was
transferred from one pan to another in random order five times. The soils were then
spiked with either diesel fuel or number 6 fuel oil. The fuels were dissolved in an excess
of hexane to facilitate uniform mixing through out the soil. The soil mixtures were air
dried to a constant weight, bottled in previously unused clean 8 oz. glass bottle with
PTFE lined caps and tumbled for 4 hours on a rotating tumbler.
Analysis
Each of the spiked soils were analyzed in triplicate using the Petro-Flag test kit. Using
the Petro-Flag procedure, 5 grams of soil were weighed into the extraction tube, 10
grams of extraction solvent were added and the sample was extracted for 5 minutes with
intermittent shaking. The extract was then filtered using a 0.2 um filter fitted with a
glass wool pre-filter. The filtered extract was added directly to the analysis vial
193
-------
containing the premeasured color reagent. The vial was then capped and shaken
vigorously to ensure mixing. The color was allowed to develop for a minimum of 10
minutes, shaking intermittently. After the color development step, the vial was placed
in the colorimeter. The absorbance reading was then used to quantify the TPH content
of the sample using the standard calibration curve.
The calibration solutions were made up using diesel fuel in the extraction solvent at 50
ppm and 250 ppm. For the 5 gram sample size used for this study this is equivalent to
100 ppm and 500 ppm in the soil. In the field the Petro-Flag kit will use a soil spike at
two levels to determine the response factors and background corrections for site specific
soil samples. For this study, using solvent standards allowed for an estimation of the
extraction efficiencies for the new method.
Results and Discussion
The diesel fuel results are presented in table 1. As indicated by the relative standard
deviation between replicates the method is very reproducible. The comparison with the
gravimetric value indicates that the extraction efficiency is greater than 95 %. There is
also very little bias.
Table 1 Petro-Flag Diesel Results
Cone.
(ppm)
54
106
255
516
Trial A
(ppm)
69
107
239
504
Trial B
(ppm)
69
114
254
507
Trial C
(ppm)
71
114
245
516
Mean
(ppm)
70
112
246
509
Std. Dev.
(ppm)
1.09
3.82
7.58
6.43
The number 6 fuel oil results shown in table 2 indicate that the response factor is
approximately 90% at 500 ppm. The repeatability as indicated by the standard deviation
is again at least +/- 10%.
194
-------
Table 2 Petro- Flag Number 6 Fuel Results
Cone.
(ppm)
50
100
251
500
Trial A
(ppm)
75
105
222
438
Trial B
(ppm)
71
117
234
449
Trial C
(ppm)
75
117
234
449
Mean
(ppm)
73
112
230
446
Std. Dev.
(ppm)
2.18
6.55
6.55
6.55
Summary
The data for the two analytes investigated show that the method is very reproducible.
The 95 % confidence intervals for the replicates indicate the method should be expected
to have repeatability of better than 10%. Although the two analytes differ in
composition they give a response that is within 10% at the 500 ppm level.
195
-------
600
CD
O>
Trial A x- Trial B -+- Trial C
100
200 300 400
Actual Cone, (ppm)
500
600
Figure 1: Petro-Flag Diesel Results
-------
600
CD
Trial A x- Trial B ^ Trial C
100
200 300 400
Actual Cone, (ppm)
500
600
Figure 2: Petro-Flag Diesel Results
-------
28
Multi-Element Multi-Media Analysis of Air-Borne Emissions at Secondary Lead Foundries
David Eugene Kimbrough, Public Health Chemist II, California Environmental Protection
Agency, Department of Toxic Substances Control, Hazardous Materials Laboratory-Southern
California 1449 W. Temple Street, Los Angeles, California 90026-5698
Dr. Irwin "Mel" Suffet, Professor of Environmental Health, UCLA School of Public Health,
Department of Environmental Health Sciences, 10833 Le Cante Ave. Los Angeles, CA 90024-
1772
Abstract
The principal focus of public health concern regarding air borne emissions from secondary
lead smelters has been lead. Antimony, arsenic, cadmium, and copper also are present in the
metallic part of these batteries, as are selenium and silver in some cases. Additionally, air
samplers can also pick up wind blown dusts and soils which can contain the above mentioned
elements as well as barium, cobalt, chromium, nickel, vanadium, and zinc.
In order to fully distinguish the sources of air borne lead (smelters, automobiles, the soil)
and to fully characterize the potential health effects from all toxic elements that may be emitted by
a source it is essential to analyze for all toxic elements. In this paper, a method for the
solublization of sixteen regulated toxic elements from glass fiber filters and analysis by ICP-AES
is presented. The method is an application of USEPA SW-846 method 3055, an agua regia
method. The results from high volume field samplers from two secondary lead smelters located
on and off site are shown similar patterns of concentration and type of elements as found in the
parent material and in the surrounding soil.
198
-------
29
NEIC FORMS II:
AUTOMATING FIELD RECORDS MANAGEMENT
Kaye Ir Mathews, Quality Assurance Manager, EPA National Enforcement
Investigations Center, Denver Federal Center, Denver, Colorado 80225; Jeffrey C.
Worthington, CQA, Director of Quality Assurance; David T. Lark, Senior
Programmer/Analyst; Chia-Ying Teng, Senior Programmer/Analyst; Tracy A. Phillips,
Audit Program Manager; Brian K. Fallen, Project Manager; TechLaw, Inc., 12600 W.
Colfax Avenue, Suite C-310, Lakewood, Colorado, 80215.
ABSTRACT
Application of technology to field sample collection has come of age with development
of a specialized software system that automates routine field sampling tasks. Originally
conceived as part of EPA Analytical Operation Branch's (AOB) Quick Turnaround
Method (QTM), EPA's National Enforcement Investigations Center (NEIC) has
broadened the scope of the project to include the Contract Laboratory Program (CLP)
and make it flexible enough to handle most sampling scenarios.
The Field Operations Records Management System II (FORMS II) was developed by
NEIC with support provided by the Contract Evidence Audit Team (CEAT). The
initial prototype was developed by EPA Region 4.
FORMS II automates evidence-related field sampling documentation including bottle
labels, sample tag labels, traffic reports, chain-of-custody records, and custody seals.
FORMS II enables field personnel to download information to the laboratory, RSCC,
and regional users. FORMS II improves field time management, standardizes
information management, and captures collection information in an electronic format
early in the field sampling process.
The authors provide a summary of FORMS II development activities and review
features of the system. Directions for obtaining FORMS II are provided in the
summary of this paper.
199
-------
INTRODUCTION
FORMS II was developed to:
Q Facilitate capture of field information during sampling events, and
Q Automate production of bottle labels, sample tags, bottle-specific custody seals,
chain-of-custody seals, chain-of-custody records, cooler seals, PRP sample receipt
records, and field reports.
The background and history of FORMS II development is presented here, followed by
a discussion of EPA Life Cycle Documentation Guidance. Design criteria are
provided. A brief description of the development process is presented. Three field
tests of FORMS II are discussed, followed by a summary description of system
operations and features.
BACKGROUND AND HISTORY
The EPA NEIC conducts environmental enforcement investigations and develops and
implements EPA enforcement strategies for EPA and other federal and state agencies.
NEIC functions include forensic laboratory services, field investigations, litigation
support, information services, and training.
In the early 1980s, NEIC established the Evidence Audit Program for the Superfund
Contract Laboratory Program (CLP). This program established documentation and
sample-handling requirements for laboratories generating environmental data for the
Agency based on NEIC evidentiary policies and procedures.
NEIC experience with evidence issues provided a basis for a request to assist in an
effort to streamline the resource intensive field documentation functions that occur
during sampling activities. The EPA AOB requested that NEIC develop a state-of-the-
art software system especially for the model CLP QTM analytical service. NEIC
directed their Contract Evidence Audit Team to provide system development in support
of this project.
200
-------
NEIC participated in QTM workgroup meetings which focused on considerations to
streamline each analytical service phase from sampling effort through data delivery
using uniquely tailored software. Although automation had not been widely used to
support EPA sampling activities, a prototype system for automating the documentation
task had been developed in EPA Region 4. The Region 4 system was ultimately used
as the basis for development of the NEIC FORMS II system.
EPA SYSTEM LIFE CYCLE MANAGEMENT
FORMS II was developed using life cycle documentation guidance prescribed by EPA.
OSWER's Life Cycle Management Guidance provides a structured approach for the
solution of information management problems, particularly those that require
consideration of automated solutions.
The benefits of system life cycle management include:
Q Full consideration of a system's operating environment, and associated systems and
data requirements.
Q Early identification of technical and management issues.
Q Early assessment of resource needs.
Q Realistic expectations of user community.
Q Balanced consideration of all aspects of proposed system modifications.
Q Periodic evaluations of system effectiveness.
The stages of a system life cycle include initiation, concept, definition, design,
development, implementation, production, evaluation, and archival.
The following life cycle documents were developed for this pilot project:
G System Concept Paper
Q Detailed Data Requirements
Q Detailed Functional Requirements
201
-------
Q Data Management Plan
Q Project Management Plan
Q System Test Plan
Q Data Dictionary
Q Users Manual/Operation and Maintenance Manual
DESIGN CRITERIA
During requirements analysis in initiation and concept stages of FORMS II, a variety of
design criteria were identified and incorporated into the FORMS II design. Following
is a discussion of design criteria.
C/C++ language
FORMS II was programmed in C/C++ language. C/C++ language is a third
generation universal programming language which is widely used in the programming
industry.
Object-oriented techniques
Object-oriented programming techniques were employed in FORMS II design. This
approach allows programmers to design and program "objects." Objects perform
individual functions and they may be re-used at any time or in other programs and
systems. Objects are saved in a library of objects and often include data elements that
objects act on.
When a new object is needed, coders may use an old object to make a new object
which "inherits" designated characteristics from previous objects. The technique
results in programs which may be easily modified. New programs may be designed
and programmed faster using previously created objects.
Hardware portability/compatibility/versatility
Because field samplers cannot guarantee access to an AC power source or a stable
computer working environment, FORMS II is compatible with existing portable
hardware including portable computers, portable printers, and portable bar code
202
-------
scanning devices. While FORMS II is mouse-compatible, unknown field conditions
prohibit FORMS II reliance on a mouse.
Field conditions can pose considerable obstacles for many hardware components.
Specific hardware units were selected for field testing to minimize likelihood of failure
or downtime.
Bar code application
Use of bar code technology was established as a priority because bar coding can
accelerate the sample packing process for sample shipment to laboratories. Also,
laboratory personnel may use sample bottle bar codes to facilitate receipt and associated
records management activities.
Flexibility for multiple samplers/samples
Field samplers and field sampling organizations often use unique numbering and
identification schemes when collecting samples. They also vary in their approach to
many other activities. For that reason, FORMS II design includes choices for:
Q Identification scheme
Q Activity names
Q Choices for label information
Q Choices for number and types of labels/tags/seals
Sample Definitions
FORMS II design criteria is based on the following definitions of a sample and related
sample types:
Q Sample "environmental sample," a single, discrete portion or piece of the
environment collected from a specified physical location at a specific time. The
single sample may be placed in multiple vessels. Parts of the same sample in
different vessels are not referred to as separate samples.
Q Fraction a part of an environmental sample which may be designated for a specific
type of analysis or analyses. Multiple fractions may be placed in a single vessel.
One fraction may be in several vessels.
203
-------
Duplicate - represents additional amounts of an environmental sample which are
placed into additional vessels identical to the vessels used for the sample. This
additional sample is not a different sample. The same identification number is used
and the duplicate may be designated by a "D" suffix. FORMS II creates a separate
sample record for a duplicate.
Blind duplicate a duplicate where the identification number used is not identical to
the identification number of the original environmental sample. FORMS II creates
a separate sample record for a blind duplicate.
Split "PRP" split, a duplicate which is numbered with the identification number
of the original environmental sample and also marked with a PRP designation.
FORMS II creates a separate sample record for a split.
DEVELOPMENT PROCESS
During FORMS II development, existing software was reviewed to determine
compatibility with proposed EPA uses. No commercially available software met EPA
needs at that time. Research was conducted to identify and itemize data and functional
requirements for FORMS II.
Data entities and their relationships were drafted and flow charts were presented to
field samplers for review. Borland C++ Turbovision™ was selected for the
programming language based on the object-oriented approach inherent in the language
and ease of use for developing applications. Code Base was selected as the data base
manager. A bar code library was purchased to print bar codes on labels. Universal
Code 39 was selected as the bar code format.
Internal, unit, and integration testing were performed as described in the FORMS II
Test Plan. System screens were presented to field samplers for review. Example
system screens are provided in the attachment to this paper. Draft copies of FORMS II
on disc were also presented for review.
204
-------
FIELD TESTS
The final phase of software testing is system testing. An important part of system
testing for FORMS II included field tests in three EPA Regions. A discussion of the
field tests in each region is provided below.
Region 2 Field Test
FORMS II was used at a former landfill site in Region 2. Sampling efforts consisted of
collection of stream and other surface water sources. Sampling was performed by an
EPA contractor. EPA personnel from Region 2 and Headquarters observed the field
test. FORMS II was tested concurrently with the field sampler's manual documentation
efforts to compare system efficiency and error rates.
Region 3 Field Test
Region 3 contractors and EPA personnel reviewed FORMS II operations and outputs
during a sample collection event at a former transportation, storage, and disposal
(TSD) facility. Samples were collected from large storage tanks. FORMS II was used
in parallel to actual sample collection efforts in a staging trailer adjacent to field
collection operations. Use of portable equipment was not necessary because there was
a power source in the trailer and a stable computer environment.
Region 8 Field Test
FORMS II was tested at a former mine site in the Rocky Mountains. Region 8
contractors were collecting soil samples in an ongoing sampling effort. FORMS II
hardware was tested during cold weather conditions which resulted in alternate label
selection.
SYSTEM OPERATION
FORMS II operates in a modular fashion. Prior to entering the first module, field
samplers may customize the system for their own use by doing the following:
Q Choosing the number of labels to be printed for bottle labels, sample tag labels, and
bottle-specific custody seals.
205
-------
Q Designing analysis fractions and customizing sample parameters such as bottle size,
preservative, and number of bottles.
Q Entering names of all possible field samplers.
Q Entering names of laboratory destinations.
Q Designing analysis requests (a series of fractions tied together).
After customizing the system to individual preferences, samplers can enter data
working down through the system from the largest entity to the smallest. Site
information can be entered first. If site information is not available to enter into every
field provided, a sampler can simply name the site (with any name) which is the
minimum data requirement for this module.
The activity level is the module below the site module and is simply a management tool
for the sampler to organize sample collection events. Some sample collection efforts
are time-dependent (activities = spring, summer, etc.), some are matrix-dependent
(activities = soils, waters, etc.), some are physical location-dependent (activities =
west side, east side, hill area, tank area, etc.), and some are regulatory-dependent
(activities = remedial investigation, field study, etc.). Individuals reading this will no
doubt have their own additions to this list.
Samplers are not required to name an activity. If this section is by-passed, FORMS II
will assign a number as the name of the activity.
The plan level is the module below the activity module and is also a management tool
to further organize a specific collection event, whether the event is one day or more.
Samplers can name the plan. Plans are much like activities. If samplers want to use
this as an organization tool to distinguish between different sampling activities, they
can do so.
The default module is actually a part of the plan module. If specific information about
samples will be very similar (e.g., matrix, fractions designated for analysis, sample
date, level), defaulting these items in FORMS II will help accelerate entry of
information. In a straightforward sampling event, it is possible that the sampler may
need to enter only the sample ID and time if all other information was entered in the
default mode.
206
-------
The collection mode is within the plan mode and is the place where a sampler can enter
all information about an individual sample if the information has not already been
defaulted. Defaulted information can be changed on a sample-specific basis.
After entering sample collection information, the sampler can elect to print labels.
Alternatively, if the field sampler wants to print labels in advance in a staging area or
the hotel room and they have advance knowledge of all samples to be collected, they
can print labels which contain all required information except date and time. The date
and time can be handwritten on the labels and later entered into FORMS II. Label
formats are provided in the attachment to this paper.
After sample bottles are labeled, the field sampler is essentially in the same situation as
any person who has an inventory to prepare for shipment. At this point, a peripheral
bar code scanning device can increase the efficiency of packing sample shipments.
Bottle bar codes can be scanned as the bottles are loaded into coolers. The information
is attached to the Traffic Report/Chain-of-custody file and will automatically print on
the shipment forms. Traffic Report/Chain-of-custody formats are provided in the
attachment to this paper.
Estimated training for new FORMS II users is less than two hours. If new users do not
need to enter customized information, training time may be reduced.
SUMMARY
NEIC, with system development support by the CEAT, developed FORMS II to:
G Support quick turnaround sample collection,
Q Automate sample collection documentation,
Q Improve time management in the field,
Q Standardize management of field information, and
Q Capture collection information in an electronic environment early in the process.
FORMS II is currently managed by David Eng, AOB National Automated Data
Processing Manager of Superfund Analytical Information. If you are interested in
using FORMS II, please contact David at (703) 603-8827.
207
-------
Site Activity Plan Default Collection shipment Reports Utilities
= View/Edit Container Information ======
Container No COOLER 1 Project Code 84644
Traffic No 834576 Account Code 84566
COC Seal No 8735 Regional Info
Non-Superfund Program RCRA
Shipper Lark
Destination IT Analytical Services-Export
Carrier Federal Express
Airbill NO 98456-5468576
Date Shipped 04/11/1994
Tine Shipped 16:00
Available Bottles
AW345 04/BNA
AW345 05/BNA
AW345 06/VOA
AW345 07/VGA
AW345 08/VOA
Cancel
Bottles Packed
MAW345 01/TOT METALS
MAW345 02/CN
MAW345 03/DIS METALS
MAW346 01/TOT METALS
MAW346 02/CN
Alt-X Exit Fl Help | Enter Container Number (required) 16:36:24 105664
Site Activity Plan Default Collection shipment Reports Utilities
View/Edit Sample Information ====^===
Site: M.G.D Landfill Site
Activity: Spring Sampling
Plan: Monday Sampling
System ID: 259487743-0001 Related QC System ID: N/A
Regional ID: DTL-01-001 Field ID: SW-041194-01
QC Designator: —
Default: Surface Water
Anal. Request: Std. Surface Water
Station ID SWOl
Sample Type Grab
Sampler Lark
Con. Level Low
Status: Uncollected
Protocol IDs • Mark Collected
Matrix: Sur. Water (Aq)
Date Began 04/11/1994
Ended / /
Time Began :
Ended :
Number of Bottles: !
- Bottles List m
Notes _ Print Labels _
Saiple 10:258487743-0002 Date: Tue:
Frac.sVOA Hatruibur. Water (Aq)
Regional ID:DTL-01-002
RAS Bottle ID: AU346 08
Saeple 10:259487743-0002 Date: Tue;
Frac.:VOA HatrixiSur. Water (fiqi
Regional 1D:DTL-01-002
hi DOtue lu: AM34b 07
Sasple ID:259487743-0002 Date: __ Tiie;_
Frac.iVGA Hal7Tx:iJur. Water tuq)
Regional 1D:DTL-0!-002
RfiS settle lb: AU346 06
Sample 10:259487743-0002 Late: Ti*e:
Frac.:6Nfl RatrixfSur. kiter
Reqional ID:DTL-01-002
RAS Bottle ID: AU34b Ob
Sa«ple 10:259487743-0002 Date; Tue;
Frac.:BNA » HaTrTxTSurT WaterIBqT
Regional !D:DTL-01-002
RAS Bottle ID: AU346 04
Saiple 10:259487743-0002 Date: Tue:
Frac.iDIS HETALS «atfix:Sur. HaterTBqT
Reqional ID:DTL-01-002
RAS Bottle ID: HAW346 03
CO
o
Alt-X Exit Fl Help | Press F3 for a list
16:33:25
111296
-------
Traffic Report and Chain-of-Custody Record
RAS Inorganic
Project Information Shipping Information
Region: 1 Case No.: 84566 [ ] Case Complete
Lead: Superfund Early Action: RI Long Term Action: FS
Project Code: 84644 Account Code: 84566
Regional Information:
Non-Superfund Program: RCRA
Date Shipped: 04/11/1994
Carrier: Federal Express
Airbill No.: 98456-5468576
To: David Dunlap
IT Analytical Services-Export
5103 Old William Penn Highway
Export, PA 15632-
Site: H.G.D Landfill Site
Boston, PA
Site SpiII ID: MA86
Sample ID
HAU345
MAW345
MAW345
MAU346
MAW346
MAW346
MAW347
MAW347
HAW347
Matrix/ Preserved/
Sample Type Cone Level Analysis
Sur. Water (Aq) HN03 DIS METALS
Grab Low
Sur. Water (Aq) HN03 TOT METALS
Grab Low
Sur. Water (Aq) NaOH CN
Grab Low
Sur. Water (Aq) HN03 DIS METALS
Grab Low
Sur. Water (Aq) HN03 TOT METALS
Grab Low
Sur. Water (Aq) NaOH CN
Grab Low
Sur. Water (Aq) HN03 DIS METALS
Grab Low
Sur. Water (Aq) HN03 TOT METALS
Grab Low
Sur. Water (Aq) NaOH CN
Grab Low
Tag Numbers
3-84568778
3-84568776
3-84568777
3-84568786
3-84568784
3-84568785
3-84568794
3-84568792
3-84568793
Bottle Collection
Number Station ID
3 SW01
1 SW01
2 SW01
3 SW02
1 SW02
2 SU02
3 SW03
1 SU03
2 SW03
Date/ Sampler/
Time QC Designator
04/11/94 Lark
08:45
04/11/94 Lark
08:45
04/11/94 Lark
08:45
04/11/94 Lark
09:30
04/11/94 Lark
09:30
04/11/94 Lark
09:30
04/11/94 Lark
10:45
04/11/94 Lark
10:45
04/11/94 Lark
10:45
Sampling Company: Sample-It,Inc.
Sampler Name
Sampler Signature
Additional Signature
Additional Signature
CHAIN-OF-CUSTODY RECORD
Relinquished by: (Sign)
Date
Time
Received by: (Sign)
Relinquished by: (Sign)
Date
Time
Received by: (Sigi
Relinquished by: (Sign) Date Time Received by: (Sign) Relinquished by: (Sign) Date Time Received by: (Sig
Relinquished by: (Sign) Date Time Lab Received by: (Sign) Date Time Remarks: Is custody seal Intact? ( Y / N / none
Custody Seal Number: 8735
Traffic/COC Report Number: 834576
REGION/SMO
PAGE: 1
209
-------
30
FIELD VERIFICATION OF METALS CONTAMINATION IN SEDIMENTS BY STRIPPING ANALYSIS
Khris 01 sen, Pacific Northwest Laboratory, P.O. Box 999, Richland, Washington
99352 and Professor Joseph Wang, Rossi Setiadji, and Jianmin Lu, Department of
Chemistry, New Mexico State University, Las Cruces, New Mexico 88003.
ABSTRACT
Stripping analysis is targeted for the determination of chromium, lead,
cadmium, copper, and zinc concentrations in sediment samples in the field
during characterization and remedial activities at hazardous waste sites.
Sediment samples are dried and digested using microwave digestion procedures
tailored to meet the needs of field activities and electrochemical
measurements. Adsorptive stripping voltammetry was used
to determine chromium concentrations. Conventional anodic stripping
voltammetry and potentiometric stripping analysis were used for cadmium, zinc,
copper, and lead determinations in sediment leachate samples. Stripping
analysis was demonstrated in the field during site characterization activities
at the Unlined Chromic Acid Pit (UCAP) and 1960's Pits located at Sandia
Laboratory's Chemical Waste Landfill. Stripping analysis results of sediment
leachate solutions were compared to conventional atomic or mass spectroscopy
methods approved by the U. S. Environmental Protection Agency (EPA).
Stripping analysis for chromium contamination in sediments was conducted at
UCAP. Maximum chromium concentrations observed at UCAP were approximately
10,000 ppm. Stripping analysis for chromium, cadmium, zinc, copper, and lead
contamination was conducted at the 1960s Pits. Only chromium and copper
contamination were identified at the 1960s Pits. Maximum concentrations
observed were 5313 ppm and 1749 ppm, respectively.
Stripping analysis has been successfully employed for field verification of
metals contamination in soils and sediments at a hazardous waste site. The
results demonstrate that stripping analysis is capable of onsite
identification of contaminate layers in soils and sediments. Concentration
values measured by stripping analysis correlated well with those obtained by
EPA approved methods. The remarkable sensitivity, portability, low power
need, and low cost makes stripping analysis an attractive choice for onsite
analysis of selected metals during site characterization and remediation
activities.
This work was funded by Department of Energy's Office of Technology
Development through the Mixed-Waste Landfill Integrated Demonstration Project
at Sandia National Laboratory through a contract with Pacific Northwest
Laboratory. Pacific Northwest Laboratory is operated for the Department of
Energy by Battelle Memorial Institute under Contract DE-AC06-76RLO 1830.
210
-------
31
COMPARISON OF THE RESPONSE OF PCB FIELD TEST METHODS TO
DIFFERENT PCB AROCLORS
Stephen Finch, Dexsil Corporation, One Hamden Park Drive, Hamden, CT 06517
ABSTRACT
Polychlorinated biphenyls (PCB) are one of several environmental analytes that are not
composed of single compounds but rather groups of related compounds. Because the
analyst is looking for a number of different compounds, he or she must be aware of
exactly what a particular analytical technique is detecting. To evaluate how well several
popular field methods (2 immunoassays and one chemical method) can test over the range
of possible Aroclors, a study was performed where each of the three methods was used
to test a broad range of available Aroclors. Results show that on the lower chlorinated
Aroclors (e.g. 1221) and the more highly chlorinated Aroclors (e.g. 1268) the chemical
method may be off by a factor of three and the immunoassay methods by a factor of 100.
Analysts using these techniques, therefore, should know ahead of time exactly what
Aroclors they are dealing with or should implement proper correction factors to eliminate
the chance of false negative results.
INTRODUCTION
Several methods currently exist to test for PCBs in soil samples. The most established
and most quantitative is gas chromatography (GC), usually capitalizing on the high
sensitivity of the electron capture detector (SW-846 method 8080). GC is an excellent
technique for quantifying PCBs because it separates out different congeners and quantifies
them individually, alerting the analyst to any Aroclor mixtures or weathering that may
have occurred while the PCBs have been exposed to the environment.
Field screening methods usually do not quantify individual compounds when testing for
PCBs but make an estimate based on one or more characteristics of the target analyte.
Therefore, field testing methods may give results that differ from other test methods even
though they are operating exactly as designed. Three such field methods were compared
on soils contaminated with a variety of Aroclors to see how they would respond in
relation to each other. Two of the methods tested are immunoassay (IA) based tests
(Millipore EnviroGard™, Ensys PCB RISc™) and one is a chemical based testing device
(Dexsil L2000 PCB Analyzer™).
BACKGROUND
Immunoassay based test kits (ELISA) that are currently available for PCB analysis are
specific devices that are designed to test exclusively for PCBs. When an animal is
211
-------
immunized to produce antibodies for PCBs, it is injected with a derivitive of a single or
several PCB congeners, but not all 209. Therefore, the antibodies that it produces will
be sensitive to specific congeners, but not to all PCBs. For instance, if antibodies are
produced to respond to 3,4, 3', 4' tetrachlorobiphenyl, the test kit that utilizes this
antibody will be highly sensitive to 3, 4, 3 ',4' tetrachlorobiphenyl but less sensitive to
PCBs that contain different numbers of chlorine atoms or have chlorine atoms at different
locations on the biphenyl molecule. As a result of this variation in sensitivity to different
PCB congeners, the analyst using an I A test kit may obtain vastly different responses to
different Aroclors.
The L2000 PCB Analyzer is not based on an immunoassay, but instead, chemically
detects the presence of PCBs by analyzing the sample for total organic chlorine and
translates the amount of chlorine detected into ppm PCBs. All PCBs contain some
chlorine and therefore, if the percent chlorine in the PCB being analyzed is known, the
amount of PCB present can be easily quantified. The percent chlorine contained in a
specific Aroclor is usually given by the last two numbers in the four digit Aroclor
designation, e.g. , Aroclor 1260 is composed of 60% chlorine. Aroclors vary in chlorine
content from 21 to 68 percent. This means that for a given concentration of PCB, the
amount of chlorine will vary by about a factor of three.
Because both the I A methods and L2000 method may vary in response among Aroclors,
a study was designed to determine what that variation might be. If the analyst is testing
at a specific level for a certain Aroclor, then what levels of the other Aroclors would
need to be present to avoid a false negative? or to avoid a false positive?
PREPARATION
Each field method was purchased or calibrated to test for Aroclor 1242 at a level of 2
ppm. The following Aroclors were included in the study:
1221 1232 1016 1242 1248 1254 1260 1268
Neat standards from General Electric (1254, 1260), Ultra Scientific (1268), Analabs
(1248, 1242, 1016), Monsanto (1232) , and Chem Services (1221) were used to make
standards in hexane at a level of 1000
A standard soil was made by mixing 6 kg dried clay with 2 kg dried sand after passing
each through a 850 /urn sieve. The mixture was then tumbled overnight to assure
uniformity. The mixture was analyzed by method 8080 to assure that it was PCB free.
Soil standards were prepared by placing 200 g of soil on an aluminum pan and spiking
with the appropriate amount of PCB in hexane standard. Enough additional hexane was
added to form a slurry. Samples were mixed and allowed to dry over night in a fume
hood. Samples were then placed in glass jars and tumbled for four hours to assure
uniformity.
212
-------
Soil samples were prepared at the following concentrations:
Aroclor Concentrations (qg/g)
1221 40, 200
1232 20
1016 5
1242 2, 5
1248 1, 2
1254 5
1260 10
1268 10, 100
PROCEDURE
Each field test was run according to the instructions supplied by each manufacturer. All
the Aroclors were run on each test and the PCB concentration in each soil was adjusted
and reanalyzed until a result was obtained that gave a response equal to or just greater
than the response obtained from 2 ppm of Aroclor 1242. Soil samples were initially
tested at concentrations determined from the "detection limit" information provided by
each manufacturer. The levels at which the L2000 was tested were simple to calculate
because the percent chlorine of each Aroclor is well known. The levels for the IA kits
were more difficult to choose because predicting the response of the kits to various
Aroclors is not straightforward. This involved an iterative process of lowering or raising
the PCB concentrations until a response equal to or greater than that of 2 ppm 1242 was
obtained. PCB concentrations below those in the originally prepared soil samples were
made by cutting the soil samples with the appropriate amount of blank soil to arrive at
the final concentration. For example, a 6 ppm 1232 sample was prepared by mixing 3
g of 20 ppm 1232 standard with 7 g of blank soil.
RESULTS
For each of the eight Aroclors tested, Table 1 lists the PCB concentration that was
required to yield a response equal to that of 2 ppm Aroclor 1242. For both of the IA
kits, a level of 40 ppm 1221 was required to yield a positive test result. The L2000
provided a positive result at 4 ppm of the same Aroclor. The most sensitve Aroclors for
the Millipore test were 1248 and 1254 which yielded positive results at 0.9 ppm. The
most sensitive Aroclor for the Ensys test was 1260 which resulted in a positive test at a
level of only 0.4 ppm. The L2000 exhibited the greatest sensitivity to the most highly
chlorinated Aroclor, 1268, and gave a positive response at a level of 1.2 ppm.
213
-------
Table 1
Aroclor Millipore Ensys L2000
1221 40 40 4
1232 7 3 2.6
1016 332
1242 222
1248 0.9 1.1 1.8
1254 0.9 0.7 1.6
1260 1.5 0.4 1.4
1268 25 3 1.2
The sensitivity ratios for each method, defined as the ratio of the concentrations required
to yield a positive test between the most sensitive and least sensitive of the Aroclors, was
determined to be the following:
For the Millipore test, 1221:1248 = 40:0.9 = 45
For the Ensys test, 1221:1260 = 40:0.4 = 100
For the L2000 test, 1221:1268 = 4:1.2 = 3.3
This means, that depending on the method, a specific test may require that one type of
PCB be at a concentration 100 times greater (Ensys) than another type in order to yield
the same response. This ratio should remain a constant for each method and will not
vary with a change in calibrating Aroclor or concentration.
CONCLUSION
What are the consequences of these results? Suppose an analyst is field testing for PCBs
at a site known to contain a variety of Aroclors, some of them partially weathered. The
regulator has stated that the site must be cleaned up to a level of PCBs no greater than
2 ppm. Now the analyst has a decision to make. If he or she uses an immunoassay test
should the test be calibrated using the most sensitive Aroclor, least sensitive Aroclor, or
something in between? By calibrating on the Aroclor with the highest sensitivity, (1248
or 1254 for Millipore and 1260 for Ensys) if Aroclor 1221 is present, the test will not
yield a positive result until the level reaches 90 ppm for the Millipore test or 200 ppm
for the Ensys test meaning that there is a very high probability of obtaining a false
negative result. If the analyst chooses to calibrate on the least sensitive Aroclor (1221)
in an effort to avoid false negatives, then false positives would result for anything above
a level of 0.045 ppm 1254 for the Millipore test and for anything above 0.02 ppm 1260
for the Ensys test. The odds of obtaining a false positive result are huge! If an Aroclor
with average sensitivity is choosen, then the false positive/false negative debate is split
down the middle and the potential for either one is still quite high.
214
-------
As an alternative to either one of these methods, the L2000 could be used and calibrated
to yield a positive test at 2 ppm 1221. Because the sensitivity ratio for the L2000 is so
low (3.3), even the most sensitive Aroclor for this method, 1268, would have to be at
a level at least as high as 0.6 ppm before a false positive result is obtained and there
would be no chance of a false negative for any of the Aroclors.
The most important point to remember is that just because a particular method delivers
acceptable results for the Aroclor on which it is calibrated, it does not mean that these
results will be the same across the entire range of Aroclors. Unless it is known that a
particular site contains one and only one Aroclor the analyst must take care to be sure
that 1) less sensitive Aroclors are not missed and 2) more sensitive Aroclors do not cause
excessive amounts of soil to be removed or remediated.
Chart 1
The concentration of each of eight Aroclors required to
yield a positive result when test methods are calibrated
to give a positive result at 2 ppm Aroclor 1242
Millipore
Ensys
D L2000
1221 1232 1016 1242 1248
Aroclor
1254
1260 1268
REFERENCES
Instruction manual for EnviroGard PCB Test Kit, Millipore Corp, Bedford, MA.
Instruction manual for PCB RISc Soil Test System, Ensys, Inc., Morrisville, NC
Instruction manual for L2000 PCB Analyzer, Dexsil Corp., Hamden, CT
215
-------
32
GUIDING FIELD ACTIVITIES BY USING RAPID, COST-EFFECTIVE ANALYSES
PERFORMED IN A FIXED LABORATORY
Gomez, Richard T., Project Manager, Inchcape Testing Services, Aquatec Laboratories, 55
South Park Drive, Colchester, Vermont, 05446.
Loring, Deborah A., President, Loring Environmental Associates, 10 Graham Ter, West Roxbury,
Massachusetts 02132-1822.
I. INTRODUCTION
Much has been documented in previous years regarding the availability, reliability, and cost-
effectiveness of analytical results derived from field screening techniques. The application of
screening methodology using both portable equipment and test kits can minimize both cost and
turnaround time involved in preliminary site assessments and field and sampling efforts
performed in support of RCRA and CERCLA investigations.
However, very little focus has been placed on the ability and willingness of a fixed laboratory to
provide high quality, low cost, quick turnaround analyses to support field efforts. In fact,
laboratories have been somewhat restricted in their ability to use their technical expertise to
develop new or modified methodology and reporting formats based on project specific data
quality objectives. When performing analyses in support of both RCRA and CERCLA programs,
laboratories have historically been required to use traditional SW-846 and/or CLP methodology,
and in many cases have been obliged to supply extensive data packages. Given the complexity
of the methodologies and data reporting requirements, the cost and turnaround times have been
commensurate with the level of effort required in performing the method and assembling the
data packages.
This paper will document, through a real example, how a client was able to use the expertise of a
fixed laboratory to provide cost-effective, quick turnaround results by allowing the lab to use
modified EPA methodologies with specified objectives in terms of sensitivity, precision, and
accuracy; and a shortened reporting format.
Based on the client's objectives in terms of the required target analyte list, accuracy and
precision, detection limit, cost, level of documentation and turnaround time, the laboratory was
able to critically review the options available in terms of sample preparation, traditional and
modified analytical methodology, and reporting formats, and then to design an analytical method
and reporting format that met the client's objectives.
The target analytes included volatile organic compounds (VOCs), semivolatile organic
compounds (SVOCs), and 22 metals. The required detection limits were those found in the
current Contract Laboratory Protocol (CLP) methodology, with the exception of those for lead,
arsenic, selenium, and thallium. Reports were generated automatically, using a combination of
internal software developed at the laboratory and reports generated directly from the Hewlett
Packard (HP) ChemServe software. The reports included target compound results, internal
standard area and retention time summaries, Laboratory Control Sample (LCS) results,
surrogate recovery summaries, and chromatograms. The laboratory transmitted VOC and
metals reports to the field office in 24 hours and SVOC reports in 48 hours. The cost for each
analysis was 30-50% of the cost of providing the traditional methodology and data package. The
screen results were subsequently used to direct field activities and focus the sampling efforts at
the site, while providing technically sound, defensible data upon which decisions were made.
216
-------
This paper will also examine further options available to the client that can minimize overall
project costs and will point out the advantages of using a fixed laboratory for certain types of
analytical activities.
II. CLIENT OBJECTIVES
The purpose of the field program was to determine the extent of contamination and delineate the
migration of volatile organics, semivolatile organics and a number of metals known to be present
at the site. Due to the size of the site and the tight field sampling schedule imposed, the client
initially considered directing the daily activities of several drill rigs on-site using data supplied
from a mobile laboratory.
The duration of the field program was estimated to be 6 weeks with sampling taking place daily.
Approximately 6-20 samples were to be collected 5 days a week and analyzed for VOCs,
SVOCs, and metals. Given the expense of maintaining a mobile laboratory equipped to perform
the sophisticated analyses requested, the client opted to have the analyses performed in the
fixed laboratory, assuming the laboratory could meet the following goals:
• The rapid data delivery schedule of the project could be met.
• The fixed laboratory could provide documented data of a higher quality than achievable in a
mobile laboratory.
• The first two goals could be achieved at a cost equal to or less than the mobile laboratory's
rate over the same time period.
ITS-Aquatec achieved all of the goals established for this study by:
• Designing the appropriate analytical methods in order to achieve the tight data delivery
schedule established without compromising the integrity and quality of the data.
• Assigning a fixed amount of time to each task beginning with sample login to ensure that
turnaround times were met, and strictly adhering to this timetable.
• Establishing a rapid response team within the laboratory to execute the program.
• Committing to the success of the program.
Different stages of planning were required in order to fulfill the project goals. Certainly,
designing the "Rapid Analysis" (RA) program was a prerequisite to the overall success of the
study.
III. LABORATORY APPROACH BASED ON THOSE METHODS
Based on the compound list, the detection limits, and the required level of precision, accuracy,
and documentation, the laboratory was able to modify the EPA CLP methods and integrate these
with their own computer data systems, so that in most cases, reports were automatically
generated overnight. Minor modifications were made to the methodologies to allow the
laboratory to turn the work around faster, while sacrificing little in the way of precision, accuracy,
or sensitivity. The approaches that the laboratory took are outlined below:
217
-------
Metals
The Metals Target Analyte List (TAL) and required detection limits are listed in Table 1. One of
the critical aspects that allowed the laboratory to provide such a quick turnaround was the
flexibility in the metals detection limits. Often, EPA CLP Contract Required Detection Limits
(CRDLs) are requested for site activities by default, whether or not these detection limits are
necessary. In general, meeting the CRDLs requires that two separate digestions be performed,
one for Inductively Couple Plasma (ICP) analysis and one for Graphite Furnace Atomic
Absorption (GFAA) analysis, and five separate analytical runs (one ICP run and four GFAA runs)
be performed. The extra digestate and 4 extra analyses are necessary because analysis for
arsenic, selenium, lead, and thallium by ICP does not meet the CLP CRDLs1.
Table 1 Metals Target Analvte List and Detection Limits (in ug/L)
Laboratory ICP CLP Required
Element Detection Limits Detection Limits
Aluminum 29 200
Antimony 26 60
Arsenic 41 10
Barium 22 200
Beryllium 0.6 5
Cadmium 3 5
Calcium 440 5000
Chromium 1 10
Cobalt 4 50
Copper 3 25
Iron 14 100
Lead 23 3
Magnesium 548 5000
Manganese 1 15
Nickel 6 40
Potassium 1068 5000
Selenium 39 5
Silver 3 10
Sodium 532 5000
Thallium 36 10
Vanadium 4 50
Zinc 2 20
Because this site assessment did not require the low arsenic, selenium, thallium, and lead
detection limits, the laboratory was able to cut the turnaround time considerably by using a single
digestion and single analysis to generate all of the metals data.
The methodology used was based on the EPA CLP Methodology. Some modifications were
made to the methodology that did not significantly affect the accuracy of the data, but did
enhance the laboratory's ability to meet the 24 hour turnaround time by limiting the number of
reanalyses required.
218
-------
The modifications made by the laboratory are outlined in Table 2:
Table 2
EPA CLP Method ILM01.0 Modifications
Method Element
Modified
EPACLPILM01.0
Digestion
Instrument Detection
Limits (IDLs)
Lead
Selenium
Thallium
Arsenic
Initial Calibration
Interference Check (ICSA/AB)
Detection Limit Standard (CRI)
Initial Calibration Verification (ICV)
Initial Calibration Blank (ICB)
Continuing Calibration Blank (CCB)
Continuing Calibration Verification (CCV)
Final Values
QC Samples (LCS/MS/MD)
1 for ICP
23 ug/L
39 ug/L
36 ug/L
41 ug/L
ILM01.0
ILM01.0
ILM01.0
ILM01.0
ILM01.0
ILM01.0
90% to 110% .allowed
upto85%-115%ifthat
element is not detected
in bracketed samples.
Wet Weight
ILM01.0
1 for ICP
1 for GFAA
3 ug/L
5 ug/L
10 ug/L
10 ug/L
ILM01.0
ILM01.0
ILM01.0
ILM01.0
ILM01.0
ILM01.0
90%-110%
Dry Weight
ILM01.0
Orqanics
The Contract Required Quantitation Limits (CRQLs) were not the determining factor in the ability
of the laboratory to perform volatile and semivolatile organics at a faster rate. The main factor
influencing the laboratory turnaround time was the modification of the quality control criteria in
the methodology.
In the CLP methodology, matrix effects are required to be "proven" (i.e., the chemist must rerun
any sample that does not meet internal and surrogate standard QC limits to prove that it is the
sample matrix, and not laboratory error, that is responsible for the QC outages observed).
However, a trained chemist can often conclude that an instance of exceeding the QC criteria is
caused by a matrix effect without rerunning the sample to prove the theory. This can often be
done by examination of total ion chromatograms, mass chromatograms, mass spectra of
interfering peaks, extraction notes, etc.
Examination of the total ion chromatogram and the specific mass chromatograms and mass
spectra can often times provide sufficient evidence that a particular internal or surrogate
standard's response is suppressed or increased due to an unresolved mixture. A GC/MS
chemist trained in mass spectral interpretation can often times reach this conclusion without
reanalyzing the sample. Requiring that this be proven by reanalysis can often be unnecessary,
as long as the decision is documented and is technically supportable.
Additionally, surrogate recovery ranges are very narrow in the VOC methods, particularly in
comparison with the related requirements for response factors in the standards. For example,
the response factor for Bromofluorobenzene (BFB), one of the VOC surrogates, is allowed a
variance of up to 25.0% in the continuing calibration standard. If the intial calibration is assumed
219
-------
valid, then the "acceptable" variance in the actual quantitation of BFB is at least 25.0%, without
taking into account cumulative error from other steps in the analytical process. The QC criteria
for the percent recovery of BFB in the sample is 86% -115%, which is restrictive in comparison
with the level of error allowed in the response factor. Given the fact that the sample analysis
adds variance due to the sample matrix, the approximate variance of + 15% in the surrogate
may not always be necessary to meet project objectives.
Other modifications were made to the method that did not significantly affect the accuracy of the
results reported, but which allowed the laboratory to achieve quicker turnaround times. Dilutions
were only made if the mass spectrometer was saturated. Tentatively Identified Compounds
(TICs) were not reported, and all final results were reported directly from the HP ChemServe
Data Station in wet weight. Percent solids resulted were reported separately. Gel Permeation
Chromatography (GPC) cleanup was not performed on the semivolatiles fraction prior to
analysis. Modifications that allowed the laboratory to achieve the turnaround time of 24 hours for
volatiles and 48 hours for semivolatiles are outlined in Table 3. The Organics Target Compound
List (TCL) and required detection limits are listed in Table 4.
Table 3 Organics EPA CLP OLM01.0 Method Modifications
Method Element
Modified
EPACLPOLM01.0
Volatiles
Tune
Initial Calibration
Continuing Calibration
Method Blank
Internal Standard (ISTD) Area
Internal Standard Retention Time
Surrogate Standard Area
Dilutions
TICs
Final Value
QC Samples (MS/MSD)
Semivolatiles
Same as above except:
GPC
OLM01.0
OLM01.0
OLM01.0
OLM01.0
Rerun if necessary,
based on professional
judgment. Flag data if
outside limits.
OLM01.0
Rerun if necessary,
based on professional
judgment. Flag data if
outside limits.
Dilute if detector is
saturated.
Not reported
Not reported
Wet Weight
OLM01.0
Not performed
OLM01.0
OLM01.0
OLM01.0
OLM01.0
Rerun if any ISTD
varies by more than a
factor of 2.
OLM01.0
Rerun if outside QC
limits.
Dilute if above
calibration range.
Report 10-VOC
Report 20 - SVOC
Dry Weight
OLM01.0
Required
220
-------
Table 4 Organics Target Compound List and Quantitation Limits
Compound
Chloromethane
Bromomethane
Vinyl Chloride
Chloroethane
Methylene chloride
Acetone
Carbon Disulfide
Quantitation Limit-ug/L
10
10
10
10
10
10
10
trans-1,2-dichloroethene 10
1,1-dichloroethene 10
1,1-dichloroethane 10
1,2-dichloroethene (t) 10
Chloroform 10
1,2-dichloroethane 10
2-butanone 10
cis-1,2-dichloroethene 10
1,1,1-trichloroethane 10
Carbon tetrachloride 10
Bromodichloromethane 10
1,2-dichloropropane 10
cis-1,3-dichloropropene 10
Trichloroethene 10
Dibromochloromethane 10
1,1,2-trichloroethane 10
Benzene 10
t-1,3-dichloropropene 10
Bromoform 10
4-methyl-2-pentanone 10
2-hexanone 10
Tetrachloroethene 10
1,1,2,2-tetrachloroethane 10
Toluene 10
Chlorobenzene 10
Ethylbenzene 10
Styrene 10
Xylene (m,p) 10
Xylene (o) 10
Xylene (total) 10
Phenol 10
bis(2-chloroethyl)ether 10
2-chlorophenol 10
1,3-dichlorobenzene 10
1,4-dichlorobenzene 10
1,2-dichlorobenzene 10
2-methylphenol 10
2,2'-oxybis (1-chloropropane) 10
4-methylphenol 10
N-nitroso-di-n-propylamine 10
Hexachloroethane 10
Nitrobenzene 10
Isophorone 10
2-nitrophenol 10
Compound Quantitation Limit-ug/L
2,4-dimethylphenol 10
bis(2-chloroethoxy)methane 10
2,4-dichlorophenol 10
1,2,4-trichlorobenzene 10
Naphthalene 10
4-chloroaniline 10
Hexachlorobutadiene 10
4-chloro-3-methylphenol 10
2-methylnaphthalene 10
Hexachlorocyclopentadiene 10
2,4,6-trichlorophenol 10
2,4,5-trichlorophenol 25
2-chloronaphthalene 10
2-nitroaniline 25
Dimethylphthalate 10
Acenaphthylene 10
2,6-dinitrotoluene 10
3-nitroaniline 25
Acenaphthene 10
2,4-dinitrophenol 25
4-nitrophenol 25
Dibenzofuran 10
2,4-dinitrotoluene 10
Diethylphthalate 10
4-chlorophenyl-phenylether 10
Fluorene 10
4-nitroaniline 25
4,6-dinitro-2-methylphenol 25
N-nitrosodiphenylamine 10
4-bromophenyl-phenylether 10
Hexachlorobenzene 10
Pentachlorophenol 25
Phenanthrene 10
Anthracene 10
Carbazole 10
Di-n-butylphthalate 10
Fluoranthene 10
Pyrene 10
Butylbenzylphthalate 10
3,3'-dichlorobenzidine 10
Benzo(a)anthracene 10
Chrysene 10
bis(2-ethylhexyl)phthalate 10
Di-n-octylphthalate 10
Benzo(b)fluoranthene 10
Benzo(k)fluoranthene 10
Benzo(a)pyrene 10
lndeno(1,2,3-cd)pyrene 10
Dibenz(a,h)anthracene 10
Benzo(g,h,i)perylene 10
221
-------
IV. REPORT FORMAT
A key element that allowed ITS-Aquatec to meet the turnaround time for this project was the
ability to rely heavily on the computer to produce reports overnight upon completion of the
autosampler runs. For organics, which were run entirely on HP GC/MS systems, the HP
ChemServe software reports were used without change. By allowing minor variations from the
CLP Forms, the reports can be automatically produced from the instrument, thus saving a
considerable amount of time and effort in uploading to another computer system.
The metals data were uploaded directly from the ICP to the laboratory's VAX system, and
inorganics reports were produced from it. To accomplish this, the laboratory's Information
Services group was able to write software that automatically uploaded the data, and after an
overnight autosampler run, the reports were waiting for review in the morning when the analysts
arrived.
By allowing the lab a small amount of flexibility in the reporting format, the reports could be
produced automatically without manipulation through another computer system, and then
reviewed, approved, and faxed directly to the field office. The analytical report generated for this
study provided far more quality control information than that attained with field screening
instrumentation. In addition, the data were initially reviewed by in-house data review groups for
validity prior to release to the field office on site. Most laboratories archive all sample data,
standards data, and related QC logs (i.e., instrument logs, standards records, etc.) for a specified
number of years, and will guarantee their ability to provide a data package at a later date if
required.
The laboratory's reports contained the following information:
1. Cover letter
2. Chain of Custody
Volatiles and Semivolatiles
3. Form I's - HP ChemServe Software
4. Surrogate Recovery Report - HP ChemServe Software
5. Matrix Spike Recovery Report - HP ChemServe Software
6. Internal Standard Area and Retention Time Summary - HP ChemServe
Software
7. Chromatograms
Metals
8. Analytical Results Sheet - VAX generated, direct upload from ICP
V. REASONS FOR AN ALTERNATIVE APPROACH
By allowing the laboratory to make minor deviations in "standard" methodology and reporting
formats, laboratories are able to overcome most of the obstacles which cause long turnaround
times and avoid the escalating costs of producing data in support of various types of site
activities. However, this flexible approach is rarely, if ever taken.
One reason may be the lack of awareness that these options exist. In order to stay competitive
in an increasingly cost-conscious market, laboratories must adapt their approaches to site
specific investigations, and do neither more nor less work than is required to achieve site specific
goals.
222
-------
Some of the burden for this approach must lie with the various regulatory agencies and
commercial Quality Assurance firms, who through audits, data validation criteria, and specific
project management approaches, have made the laboratories painfully aware of what can
happen if they deviate from a standard approach. Often, laboratories have had data rejected, or
have been forced to use inferior methodology and technology as a result of data validation or
walk through audits by both government and commercial personnel. Audit criteria are generally
derived by using a specific set of regulatory guidelines, often developed under the CLP program,
whether or not these criteria are relevant to a specific project with specific goals in terms of
compound lists, detection limits, and precision and accuracy requirements.
Project specific goals must drive analytical data gathering activities, instead of the present
situation, in which the application of CLP or SW-846 methods drive data gathering, and use one
level of precision, accuracy, and sensitivity for all data generation regardless of the data quality
objectives for that particular site activity.
This approach is particularly off-course in the application and validation of data generated under
RCRA, in which SW-846 should be used as a guidance document only (with the four exceptions
as outlined in the regulations, in which the methods must be followed). Unfortunately, there are
many state agencies as well as private companies who require that the methods as contained in
SW-846 be followed line by line, without deviation. And to further the misuse of the available
documents, data generated using the SW-846 methodology as a "guide" is often validated using
CLP validation criteria.
VI. PROJECT SUMMARY
The case study presented here took place over a 5 week period. During this time, approximately
175 soils and 75 groundwaters were analyzed for organic and inorganic constituents. Metal and
volatile organic compounds were determined and reported within 24 hours of sample receipt and
semivolatile organics were reported in 24-48 hours. The success of the analytical approach was
measurable in 3 principal ways:
1. Data of known quality were generated and reported in near real time allowing field activities
to advance based on sound analytical information.
2. The turnaround times established in the beginning for this program were met in nearly all
cases and did not delay the project's rapid field sampling schedule.
3. The results of the RA program complimented the results of the QC sample analyses
performed by the CLP protocols.
The success of this laboratory program could be attributed to several other factors as well.
Participation in the planning stage of the field program provided valuable insight into the goals of
the client. Laboratory methodologies could be modified to enhance the analysis and reporting
speed and yet still maintain the key elements of the methods that would allow for cross
correlation with QC sample results. A project specific, rapid response team was formed to focus
on the logistics of the project to ensure that the turnaround times were met. Above all, the
willingness of the laboratory to operate outside of the normal day to day production mode in
order to serve the client's needs was an important factor that cannot be overlooked.
223
-------
VII. ADVANTAGES OF A FIXED LABORATORY
There are a number of advantages in using a fixed laboratory, and allowing flexibility in
application of standard methods to guide field activities. Certainly, the quality of data
achievable in a fixed laboratory is most often not routinely achievable in the field. Some issues
affecting the quality of the data include:
• First, in a fixed laboratory, analyses are performed in a controlled environment, free from
potential air contamination. Many analyses require the use of sophisticated instrumentation,
in which stable power sources and controlled environments (in terms of air contamination,
temperature, and humidity) are required. ITS-Aquatec's laboratory processes incoming air to
the building through charcoal filtration and all of the laboratories are positive or negative
pressure rooms as appropriate. This is crucial, particularly when performing volatile organics
analysis.
• A fixed laboratory generally possesses redundant instrumentation. This may not be an
option when generating data on-site. As such, rapid laboratory analyses can be performed
on schedule despite any unforeseen problems that may cause an instrument such as a
GC/MS system to go down. ITS-Aquatec committed 5 GC/MS systems to this study to
ensure the client's goals and data delivery schedule were met. In addition, an in-house
service technician was on call in the unlikely event that an instrument required service. It is
important to assess the ability of the laboratory or field crew to respond to instrument failure
during the planning stage of the field program.
• To ensure the quality of the data generated, a team approach was developed for this study.
A technical project director was assigned to oversee the project and the delivery schedule.
Since the methods employed were modified CLP protocols, final data review was
performed by PhD's specializing in inorganic chemistry and mass spectrometry.
Rapid analyses can often times be performed more cost effectively in a fixed laboratory
because the unit costs are determined in advance and are fixed. The client does not have to pay
an on-site analytical testing crew if there happens to be down time in the field. For the purposes
of this study, a unit cost was determined for each analysis to be performed.
In addition and most importantly, the overall cost of the analytical program can be reduced by
allowing the laboratory the flexibility to modify existing methods where appropriate. By slightly
modifying the "official" methods slated for an investigation and allowing the laboratory to judge
the validity of the experiment performed, one can more closely correlate the "screen" data to that
generated by "official" methods. In other words, one may only have to analyze 10 or 20% of the
samples collected by these official methods and still be able to correlate the results to those of
the screen, thereby reducing the overall cost of the project.
VIM. CONCLUSION
As was stressed in the introduction, laboratories have been burdened with unnecessarily strict
QC protocols, excessive reporting requirements, and very little flexibility to adapt methodologies
to meet project specific goals. In some cases, the laboratories themselves have perpetuated this
problem by standardizing the lab to such a production line approach that flexibility no longer
exists. In the last ten years, environmental chemists working in a commercial laboratory have
had fewer and fewer options in their technical approach to solving specific problems. A point
224
-------
that should be emphasized is that environmental chemistry is not so straightforward and routine
that it can be run through a production lab. Chemists must be allowed to use their scientific
training, to develop methodology, to use new products and innovations, and to customize their
approach to help laboratory clients meet their goals.
Strong QA functions exist at the government, private, and laboratory levels. Data validation and
systems and data auditing should be used to ensure that data generated in support of specific
field activities is scientifically sound and legally defensible, while allowing the chemist the
flexibility and creativity that they need to approach and solve environmental problems.
There are a myriad of screening techniques that a laboratory can utilize to expedite data
generation and reporting. As an example, GC/MS Single/Multiple Ion Monitoring can be used to
identify specific organic compounds in very complex mixtures at ppb and ppt levels. This is an
especially quick, effective tool, in both the field and the laboratory. In the example as outlined,
ICP was used quite effectively with minor increases in some detection limits, and resulted in the
laboratory generating data with approximately 25% of the effort that would be required in the
traditional approach.
Finally, the case presented here is only one example of an approach to direct field activities.
Many fine field screening instruments and immunoassay techniques can be used to provide
valuable, real time answers in the field. The most appropriate approach is generally chosen on a
project specific basis. However, by allowing a fixed laboratory more flexibility, considerable time
and money can be saved in generating analytical data of known quality to support field activities.
IX. FOOTNOTES
1 Detection limits achievable with specific ICPs vary depending on the instrument and laboratory.
The new trace ICPs reach the CLP CRDLs for these elements, but these have just recently been
put into operation in a limited number of laboratories, and thus are not yet a common option.
X. ACKNOWLEDGMENTS
The authors would like to thank Dr. R. Mason McNeer, of ITS-Aquatec Laboratories, Colchester,
Vermont, for reviewing this paper and providing helpful suggestions.
225
-------
33
Reducing Laboratory Costs with a Field Portable Ion Chromatography System
W. Mozer, Analytical Magnetics, Inc. The Mill,. 73 Princeton Street, North Chelmsford, MA
01863
ABSTRACT
Field portable Ion Chromatography (1C) has become a reality with the advent of improved
instrument design and microprocessor technology. It is now possible to determine anion and
cation levels in surface and waste waters at field sites without compromising analytical
performance. This paper will demonstrate that in-field and laboratory results for various water
matrices are analytically equivalent. The benefits of on-site work are obvious: the need to collect
large numbers of samples to be brought to a central laboratory for analysis, and the costs
associated with this procedure, is no longer required. Samples can now be analyzed as collected
and only a few "check" samples need be returned to the laboratory for verification. Should the
need to review the in-field work arise, the instrument has all the chromatographs stored in the
microprocessor and the results may be reworked at any point using a personal computer. Thus,
with the need to obtain more rapid, site specific information, reduce manpower and the costs
involved with sample collection a field portable 1C proves to be a rapid, low cost means to
accomplish these goals.
226
-------
ENFORCEMENT
-------
34
EPA COMPLIANCE PROGRAM IN WASTE TESTING
Francisca E. Liem, Chief, Laboratory Data Integrity Branch, Office
of Compliance, U.S. Environmental Protection Agency, Washington, DC
20460, Frederic L. Siegelman, Chemist, Office of Compliance, U.S.
Environmental Protection Agency, Washington, DC 20460.
ABSTRACT
The U.S. Environmental Protection Agency (EPA) uses data that are
submitted or self-reported. In the Resource Conservation and
Recovery Act (RCRA), the regulated communities are required to
identify themselves, obtain permits and periodically sample,
analyze and self-report. Concerns exist about the quality and
integrity of self-reported data. Congress asked the General
Accounting Office (GAO) to assess the RCRA program at EPA.
EPA has undertaken a reorganization of the various enforcement and
compliance monitoring functions into a single office, the Office of
Enforcement and Compliance Assurance (OECA). OECA will be the
national program manager for all inspection activities. OECA will
use a broad range of tools to achieve compliance with environmental
statutes. The Office of Compliance (OC) within OECA will have the
responsibility for developing targeting, training and national
guidance for inspections and will have the lead within OECA for
compliance assurance and assistance. The Laboratory Data Integrity
Branch (LDIB) of OC will be responsible for compliance monitoring
and quality assurance of laboratory data submitted or self-reported
by the regulated community. Compliance assurance of laboratory
data for the RCRA program will be an LDIB responsibility. LDIB can
be anticipated to increase the scrutiny of the laboratories
providing health and environmental data for the various programs.
The RCRA program does not require inspections of laboratories or
the determination of the laboratory's capability to perform
analysis. The RCRA program needs to institute controls over the
laboratories, including inspections and performance evaluations.
A laboratory accreditation program may be the answer. Mechanisms
are also needed for overseeing quality assurance within the
regulatory agencies themselves.
Because of the GAO audit and organizational changes at EPA the RCRA
program is at a crossroad. This paper only discusses data
integrity issues of laboratories that are analyzing groundwater
samples.
227
-------
- 2 -
INTRODUCTION
The U.S. Environmental Protection Agency (EPA) uses data for
several programs that are submitted or self-reported by regulated
communities. In some programs, such as the Resource Conservation
and Recovery Act (RCRA) and the National Pollutant Discharge
Elimination System (NPDES) programs, members of the regulated
communities are required to identify themselves, obtain permits and
periodically sample, analyze and self-report according to
regulations and specific permits. Similarly under the Federal
Insecticide, Fungicide and Rodenticide Act (FIFRA), members of the
regulated community must carry out required studies and submit the
results. Concerns exist about the quality and integrity of
submitted or self-reported data. EPA has undertaken a
reorganization of the various enforcement and compliance monitoring
functions into a single office, the Office of Enforcement and
Compliance Assurance (OECA). As part of this reorganization, the
Laboratory Data Integrity Branch (LDIB) of the Office of Compliance
(OC) within OECA will be responsible for compliance monitoring and
quality assurance of laboratory data submitted or self-reported by
the regulated community.
EPA and forty-seven authorized states and U.S. territories that
carry out the environmental program under RCRA rely on information
that approximately 300 permitted land disposal facilities
themselves submit to determine their compliance with environmental
laws. However, some facilities may not voluntarily invest the time
and resources to obtain accurate data and report environmental
violations. It is EPA's responsibility to determine that self-
reported data are accurate and valid, and to determine that
facilities are complying with environmental regulations.
RCRA facilities are required to monitor the groundwater underlying
their facilities in order to detect any contamination. These
results must be reported to EPA or an authorized state once a year.
However, if the analyses show contamination, facilities must
immediately notify authorities and begin more extensive monitoring.
Congress was concerned about the potential avoidance of regulation
or submission of inaccurate or fraudulent data to EPA, and so, in
1992 has asked the General Accounting Office (GAO) to assess the
RCRA program at EPA. GAO reviewed EPA and authorized states
procedures to ensure that (1) subjected facilities identify
themselves, (2) sampling results are representatives of its
compliance, and (3) oversight of facilities collecting and
laboratories analyzing samples is adequate to prevent error and
fraud.
Because of the GAO audit and organizational changes at EPA the RCRA
program is at a crossroad: What is EPA Headquarters role in the
program? This paper only discusses data integrity issues of
laboratories that are analyzing groundwater samples.
228
-------
- 3 -
DISCUSSION
The OECA reorganization
OECA includes the various compliance assurance and enforcement
capabilities of EPA. The proposed organization is made up of the
Office of Criminal Enforcement, the National Enforcement
Investigations Center, Office of Regulatory Enforcement, Office of
Compliance (OC), Office of Site Remediation Enforcement, Office of
Federal Activities as well as Administration and Resource
Management Staff, Enforcement Capacity and Outreach Office and
Federal Facilities Enforcement Program. OECA's incorporation of
the capabilities of compliance and enforcement into a single
organization recognizes that the primary goal is to protect human
health and environment by obtaining compliance and that enforcement
is a part of compliance. OECA will use a broad range of tools to
achieve compliance with environmental statutes. OECA will measure
success based on compliance and environmental results. Traditional
enforcement should be seen as a tool for achieving the broader goal
of compliance and not as an end unto itself. Risk-based strategies
are necessary.
OECA will be the national program manager for all inspection
activities including compliance assistance. OC will have the
responsibility for developing targeting, training and national
guidance for inspections and will have the lead within OECA for
compliance assurance and assistance. The laboratory data integrity
function will be established as a branch, Laboratory Data Integrity
Branch (LDIB), in the Agriculture and Ecosystems Division in OC and
will be responsible for compliance assurance of laboratory data
including inspections, audits, targeting and evaluation.
Education and assistance should be recognized as an important tool
for achieving compliance. Resources should be directed toward the
greatest risks to human health and the environment. EPA needs to
encourage and promote voluntary compliance. The need for effective
enforcement, and the ability to deal with serious offenders will
persist. If serious offenders are not punished for their
violations, the credibility of a compliance program is seriously
endangered. Successful enforcement "catches" the worst offenders
and helps achieve compliance by the population as a whole.
OC will be organized on a sector approach. Compliance assistance
activities will complement traditional enforcement and program
efforts. National enforcement strategies will increasingly be
oriented around sectors of the economy. The sector perspective
should allow change and the development of innovative compliance
and enforcement strategies. OC will be responsible for setting
national priorities and collecting and integrating quality
compliance data. In addition, OC will develop effective compliance
assurance programs to support inspections and self reporting and
229
-------
- 4 -
will build capacity for more effective compliance assistance to
the regulated community. Finally, OC will work in partnership
with Regions, States, municipalities, citizens and industry.
Compliance assistance is information or technical advice provided
to help the regulated community and interested parties understand
and meet statutory and regulatory requirements. Compliance
assistance includes outreach (publications, training seminars etc.)
and technical assistance. Inspectors can play an important role in
providing on-site assistance. Compliance assistance should help
prevent as well as correct violations.
The historical major activity of the LDIB staff was the inspection
and monitoring of the Good Laboratory Practice (GLP) standards
regulations under FIFRA and the Toxic Substances Control Act
(TSCA). The GLPs are management standards for operating
laboratories active in environmental matters and are a major factor
in efforts to assure high quality data primarily in support of
pesticide and toxic chemical registrations. Compliance assurance
issues related to laboratory data integrity in other areas also
will be LDIB's responsibility. All of the compliance assurance
activities that used to be in the program offices will be combined
within OECA. At EPA Headquarters, compliance assurance of
laboratory data for the RCRA program will be an LDIB
responsibility.
LDIB responsibility will be to ensure data quality and integrity by
assuring compliance through a variety of activities including
inspections, compliance oversight of delegated (e.g., state and
regional) efforts, outreach, and compliance assistance. Inspectors
will take part in compliance assistance efforts when they are part
of a predetermined strategy.
EPA's Current Hazardous Waste Enforcement Program
EPA at Headquarters in Washington, DC provides guidance for
management of the compliance assurance and enforcement of the RCRA
program. The RCRA compliance and enforcement program at
Headquarters was situated in the Office of Solid Waste and
Emergency Response, Office of Waste Programs Enforcement, RCRA
Enforcement Division. Inspection and audit manuals to be used by
EPA regional and authorized state inspectors have been established.
EPA regions write the permits and carry out the inspections in
unauthorized states and are also responsible for overseeing that
authorized states meet federal requirements. The RCRA program is
responsible for developing quality assurance techniques, guidance,
and technical support to EPA regions and authorized states to use
and specific sampling and analysis procedures for incorporation
into a facility's permit. The program officials are also
responsible for developing oversight procedures that EPA regions
230
-------
- 5 -
and authorized states can implement to ensure that facilities carry
out sampling and analysis properly.
EPA's quality assurance guidance requires that EPA and state
regulatory agencies inspect facilities periodically to ensure that
facilities are collecting samples properly. The EPA quality
assurance guidance also calls for inspections of laboratories that
analyze samples and testing their performance by requiring them to
analyze blind samples. EPA's RCRA program, however, does not
require inspections of laboratories or the determination of the
laboratory's capability to perform analysis.
The RCRA program has two types of inspections that are designed to
verify sampling procedures and techniques at the facility.
According to EPA policy, all RCRA land disposal facilities are
required to receive one of these inspections at least once every
three years. The operation and maintenance inspection is to
determine whether the groundwater monitoring system is adequate to
produce accurate data. The comprehensive groundwater monitoring
inspection is to evaluate and review the groundwater monitoring
system in more detail and often includes sampling. Most authorized
states routinely review sampling procedures during their
inspections.
Considerations for EPA's RCRA program in the Office of Compliance
EPA recognizes that errors can occur at any point during sampling
and analysis. Potential errors in sampling and analysis are in the
design of a sampling strategy, collecting, handling, preparing and
analyzing samples. Errors could even be made during the
interpretation of data.
EPA has sent guidance to the regions, that outlines the most recent
Data Quality Objective/Quality Assurance Project Plan for
groundwater monitoring and corrective action. In addition, EPA is
developing a Compliance Monitoring Evaluation interactive training
video that will emphasize the importance of complete and effective
reviews of groundwater sampling procedures.
To ensure that facilities are collecting and analyzing samples
according to the sampling design and procedures, EPA or the
authorized state periodically inspects facilities where samples are
taken and should inspect laboratories where samples are analyzed.
In addition to facility level controls and oversight, certain
mechanisms are needed for overseeing quality assurance within the
regulatory agencies themselves, i.e. EPA Headquarters oversight of
the regions and authorized states.
231
-------
- 6 -
Although EPA regions and most states were reviewing sampling
procedures during their inspections, EPA regional officials who
observed state inspections found problems with their quality, e.g.,
the state inspector used improper sampling equipment, did not have
sufficient knowledge of the facility's sampling and analysis plan,
and did not adequately review the facility's hydrogeology or site
characterization information. Because of improper collection and
preservation of samples for the duration from the field to the
laboratory, there were difficulties in analyzing samples properly.
These deficiencies could effect the ability of state inspectors to
detect errors in the procedures used and the data reported by the
facilities. Because of these problems, the RCRA program decided to
emphasize oversight inspections and to increase regional oversight
of state inspections to ten percent.
At laboratories, regulatory agencies can perform several types of
oversight. The first is an inspection to determine the
availability of analytical instruments and controls at the
laboratory- The objective of the inspection is to determine the
laboratory's capability to perform analysis and generate reliable
and valid data. This type of inspection would review calibration
and maintenance records of analytical instruments, general
cleanliness of the laboratory, condition of equipment and facility,
documentation of procedures and other related components.
A second oversight mechanism is a performance evaluation to
determine whether the laboratory can perform analysis and produce
reliable and valid data. To conduct a performance evaluation,
samples of known content and quantity are prepared and sent to the
facility for analysis. The facility sends the samples to the
laboratory that it usually uses for analysis. After the laboratory
analyzes the samples, the state or EPA compares the results with
the known values to determine the laboratory's performance in
analyzing the samples.
The RCRA program has never required inspections or performance
evaluations of laboratories used by RCRA land disposal facilities.
The rationale for not implementing laboratory inspections or not
developing performance evaluations is the same, the potential for
error is greater during sampling at the facility than during
analysis of the samples at the laboratory and the EPA has therefore
focused its attention on the sampling program. In addition,
conducting performance evaluations in the RCRA program would
require the program to develop a complex set of samples for use in
performance evaluations, which is beyond its current resources.
However, some EPA officials believe that the RCRA program needs to
institute controls over the laboratories, including inspections and
performance evaluations. A laboratory accreditation program may
be the answer.
232
-------
- 7 -
Participation in the proposed National Environmental Laboratory
Accreditation Program is being considered. The Agency is pursuing
a national environmental laboratory accreditation program. A
committee was established to solicit information from outside the
Agency on the need for and operation of a national laboratory
accreditation program. The committee with representatives from
states, federal agencies, the laboratory and the regulated
community, recommended establishment of a voluntary national
accreditation program for environmental laboratories.
In late Fiscal Year 1994 or early Fiscal Year 1995, EPA plans to
sponsor a national meeting with federal and state officials and
private sector representatives to develop consensus standards for
a national environmental laboratory accreditation program. The
result will be published in the Federal Register. It is
anticipated that EPA will adopt these standards without going
through a separate regulatory process, and that states will
voluntarily modify their own accreditation program to be consistent
with these consensus standards.
In 1988, EPA's RCRA Groundwater Task Force found that laboratories
analyzing groundwater data did not have quality assurance/quality
control procedures and did not always use correct analytical
methods. Consequently, the task force recommended that the agency
develop the resources and expertise to conduct inspections at these
laboratories for use by both EPA and authorized states.
The RCRA program agreed to implement this recommendation and
developed a guidance for a new inspection called the laboratory
audit inspection. This type of inspection is to detect the use of
improper procedures, identify violations and provide a mechanism to
investigate anomalies in the facility's groundwater data or other
concerns over the quality of data by individual laboratories, and,
determine whether laboratories were capable of generating high-
quality analytical data.
A third oversight mechanism could be an audit of laboratory raw
data and records to determine if they validate the reported results
and if the results were obtained as planned and reported. The
Agency is developing a module on laboratory fraud in the RCRA
Inspector Institute and Advanced RCRA Institute. The module will
train inspectors on laboratory fraud and data quality problems to
enable inspectors inspect laboratories and determine any data
quality problems.
LDIB can be anticipated to increase the scrutiny of the
laboratories providing health and environmental data for the
various programs. This will be accomplished by direct inspections
by LDIB staff, oversight of state and regional inspections, review
of EPA regions and states management of their quality assurance
programs, and inspector training and guidance efforts.
233
-------
- 8 -
CONCLUSION
Based on GAO recommendations and EPA's reorganization, changes to
the RCRA compliance assurance program on laboratory data integrity
issues have been considered: (1) Inspections of facilities include
complete and effective review of groundwater sampling procedures.
(2) Regional oversight of state inspectors is emphasized. (3) EPA
Headquarters oversight of regions and authorized states. (4)
Inspectors will be trained in quality assurance, sample collection
techniques, data quality problems and laboratory fraud. (5)
Laboratory inspections will be carried out. (6) The Agency is
pursuing a national laboratory accreditation program for
environmental laboratories.
REFERENCES
General Accounting Office Report, GAO/RCED-93-21, Environmental
Enforcement - EPA Cannot Ensure the Accuracy of Self-Reported
Compliance Monitoring Data, March 1993
Response to GAO's report, April 6, 1994
NOTICE
This paper contains the views and opinions of the authors. It does not
necessarily reflect the current position or policy of the USEPA.
234
-------
35
HOW TO PREPARE FOR
AND
MANAGE A LABORATORY INSPECTION
Robert L. Cypher. Director of Quality Assurance, EA Engineering, Science, and
Technology, Inc., 11019 McCormick Road, Hunt Valley, Maryland 21031; Mary M.
Uhlfelder, Manager of Quality Services, EA Laboratories, 19 Loveton Circle, Sparks,
Maryland 21152; Michael M. Robison, Quality Assurance Officer, Maryland Spectral
Services, Inc., 1500 Caton Center Drive, Baltimore, Maryland 21227.
ABSTRACT
The majority of environmental analytical laboratories will experience several laboratory
inspections over the course of a year. The nature of these inspections are to initially qualify,
certify, or validate the laboratory or provide annual renewal by outside groups. The majority
of the inspecting groups will be from a federal or state agency and the minority will be
commercial or private clients. Most inspections will last a day, or two, with usually one,
sometimes two or more inspectors.
The objective of this presentation is to give you some insights into how to prepare for and
manage a laboratory inspection. In this presentation, we will ask a number of questions such
as: What should you be doing before, during and after each inspection? Does your
laboratory have a company policy and procedure for handling inspections? Do you have a
statement of intent? Have you designated an individual to function as an official escort or
representative? Do you conduct training sessions and pass out handouts regarding
inspections? Has the receptionist been briefed on what to do when an inspector arrives?
Who in the laboratory is to be notified? Do you have an inspector's survival kit? What are
the "Dos" and "Don'ts" of an inspection? Do you have a debriefing at the completion of
every inspection? We will provide answers to these questions and provoke other questions
that should stimulate your thinking as to how you handle your future inspections.
INTRODUCTION
To perform analytical work in the environmental arena today, your laboratory will be
required to be certified or validated by a federal or state agency. All of your federal and
most of your state certification/validation procedures will include an on-site inspection of
your laboratory. The first impression that you as a laboratory make on the inspecting team
must be a good one, because that impression sets the tone for the rest of the inspection.
From our viewpoint, any inspection is an opportunity for your laboratory to show off your
facility and qualifications. Inspections are to be taken seriously, but at the same time use
them as a learning experience, a chance to correct any deficiencies and to have another pair
of eyes examine your systems and procedures. For most state certifications you have to pay
for your inspection, so you might as well make the most of it. The best way to be prepared
235
-------
for any inspection is to practice. Over the next couple of sections I'll be discussing some of
our experiences and how we handle some of the situations. We can't say it is the best way,
but it is the way that works for us.
"WHEELS OF MOTION"
The phrase "Wheels of Motion" is used to get you to start thinking about inspections before
they happen, to develop an action plan, and to test the system. This is where I discuss the
"before, during, and after" the inspection; you're constantly in an evolving circle.
Inspections are a way of life in most laboratories because of regulations, regulators, and
demands of your clients. Also, because of the recent incidence of laboratory fraud you may
find an increased frequency of inspections.
BEFORE THE INSPECTION
A laboratory can do a lot in preparing for any inspection by building its inspection
procedures into its daily laboratory activities. First, develop a policy and procedure on
inspections that will serve as a road map for everyone to follow in your laboratory.
A brief example of a company policy and procedure for inspections is presented in Table 1.
Second, maintain a "high state of readiness"; this means that if an inspector walked in
unannounced you would be prepared. Many times, a laboratory will receive prior
notification of an inspection, so they usually devote the day before cleaning up, checking the
books, conducting a quick lab inspection, and looking for clean lab coats. Although rare,
there are unannounced inspections that count just the same. Third, conduct "mock"
inspections to test the system and get the general laboratory population exposed to what
possibly could be the "real thing." Fourth, select a room where you will house the
inspectors during the course of the inspection? Things to consider when selecting a room:
• Size, table, and chairs
• Blackboard (requests, schedules)
• Projection screen
• Telephone
• Proximity to rest rooms
• Availability of refreshments
• Isolated from the work flow
DURING THE INSPECTION
The arrival of the inspectors at the laboratory is the first critical step in the inspection
process. From an inspector's viewpoint there are three critical periods:
• During the reception
• In the laboratory
• At the completion of the inspection
236
-------
During each of these critical periods an "impression" is formed about the laboratory, the
personnel, systems, and procedures. As an example, if your laboratory is dirty and
cluttered, the natural tendency is to think bad data. The inspector may decide to look for
problems or make issues of minor items. This is not to say that a clean looking laboratory
wouldn't have problems either.
Once the inspectors have been received and taken to the designated room, their credentials
should be checked. If they are not presented you should ask to see them. You should satisfy
yourself that the inspector is an authorized employee of the inspecting agency.
Next, the company representative (escort) should request an explanation from the inspector as
to the purpose of the inspection. With a Food Drug Administration (FDA) or an
Environmental Protection Agency (EPA) inspector they will usually present a "Notice of
Inspection." Discussing the reasons for the visit is beneficial to each of the parties, because
this can increase the efficiency of the inspection. Note that, if the inspection is directed
towards a particular client, the laboratory needs to get authorization from the client before
the inspector can review their data.
After the purpose of the inspection has been discussed, the company representative should
describe the company's Statement of Intent (Table 2), the so-called ground rules for the
inspection. At this point, the inspectors usually will take an orientation tour of the facility or
just start the inspection.
Prior to the reception of the inspectors, the company escort (or alternate) should have made a
quick tour of the laboratory noting any problem areas, talking to each supervisor about work
schedules and status of personnel.
The key to managing any inspection is having a qualified escort (Table 3). This individual
can make or break any inspection. Another key component of an inspection is what I call an
"Inspector's Survival Kit" (Table 4). It is a cardboard box with all of the information that I
carry from inspection to inspection.
During any inspection, there are certain guidelines that you should follow. I call them the
"dos and don'ts" (Table 5) of an inspection.
During the inspection, the escort should be taking notes of the areas inspected, personnel
interviewed, records reviewed, any problems noted, etc. As you're progressing through the
inspection you should be developing a "gut feeling" for how the inspection is going and the
direction the inspector is taking. Anytime that the escort perceives that there may be a
problem, the escort should ask the inspector for more details to determine whether it is or
isn't a problem. At this point, you should be passing the word to the other areas that will be
inspected as to what to expect and the approach being taken. If the inspection lasts for more
than a day, at the end of the day the escort should get a quick debriefing from the inspector
and confirm the schedule for the next day. If it's only a one-day inspection, during the lunch
break, you may be able to get a quick debriefing. Why should you be concerned about
getting a debriefing while the inspection is on-going? Because, if there is any way of
237
-------
correcting a situation while the inspector is still in the facility and has not given a final
debriefing, you may be able to have the problem removed from the list.
During the course of the inspection, if copies of data are requested by the inspector, a
duplicate copy (for your record) should be made, the data stamped as "Company
Confidential" and a receipt prepared. Many laboratories may not know that any information
a federal agency acquires during an inspection becomes part of the agency record, which
may be disclosable to the public (Freedom of Information Act) or used as evidence by the
agency to support application of civil penalties, seizure, injunction, or criminal penalties.
At the completion of the inspection, a debriefing should be held. The laboratory should
determine the list of attendees, designate a spokesperson and a recorder. The laboratory
should discuss each finding point-by-point. The inspector may document your response and
could delete a finding if evidence can be produced. The inspector will usually indicate that a
report will follow in so many weeks and that you will have a specified number of days to
respond.
AFTER THE INSPECTION
After the inspector has left the facility, you should have the debriefing notes typed and hold a
meeting with key personnel to go over the findings and outline corrective actions, if any, that
need to be taken. Set up an inspection file. Upon receipt of the inspector's letter, you
should make copies and distribute them to key personnel, set up a meeting, and develop an
action plan. Usually the quality assurance group will oversee this function and perform a
follow-up inspection to verify that the corrective actions have been implemented.
The best way for a laboratory to learn from each inspection is to conduct training sessions
with your laboratory staff. Go over the findings from past inspections and have the escort
discuss what its like to be with an inspector and what they look for. The best way to handle
any inspection is to have your laboratory personnel educated.
SUMMARY
We realize that this may seem like a lot to ask of any laboratory, but if you want to prepare
for and manage your laboratory inspections some type of system needs to be in place. Most
of your laboratory inspections are the responsibility of the quality assurance group, but we
want to remind you that the real responsibility falls upon everyone that works in the
laboratory. If you don't pass an inspection it may have repercussions that could affect the
future of the laboratory.
238
-------
TABLE 1
COMPANY POLICY AND PROCEDURE FOR INSPECTIONS
Every company should have a policy and procedure on inspections by outside organizations.
POLICY
• Specify the nature of Company X's activities that are subject to federal, state, and
local regulations.
• Be cooperative, courteous, and candid.
• Obtain authorization from client.
• Identify circumstances under which legal counsel is to be notified.
• Check all questions of legal rights with legal counsel.
• Identify disclosable and nondisclosable documents.
• Make company personnel familiar with Policy and Procedure.
• Maintain copy of each document supplied to inspector.
• Assign responsibility for coordination of inspections:
- Reception
- Designation of company escorts and alternates
- Notification of key personnel
- Conduct of inspection
- Approval for release of requested data
- Debriefing
- Post-inspection report
- Corrective action plan
- Follow-up of plan
• Prepare a set of "ground rules" for inspector.
- Statement of Intent
239
-------
TABLE 2
STATEMENT OF INTENT
1. Company X recognizes the proprietary nature of its programs for clients and will not release
information without authorization of the client.
2. Company X recognizes the authority of regulatory agencies to inspect—at reasonable times
and in a reasonable manner (state normal operating hours).
3. A representative of a regulatory agency wanting to inspect will provide the following
information:
• Personal Identification
• Identification of the Agency
• Purpose of the Inspection
• Approximate Duration of Inspection
4. A representative (escort) of the company will accompany the inspector(s) at all times. The
inspector(s) requests for information, documentation, or interviews will be handled through
the escort.
5. Company X forbids photographs, videotapes, or tape recordings to be taken.
6. The inspector will sign a receipt for data or samples taken. Data may be stamped as
"Company Confidential Information." At the discretion of management a reasonable fee
may be charged for collection of data or samples.
7. The inspector(s) will not be given access to personnel files, financial information, or internal
audit reports.
8. The inspector(s) will follow all company safety rules and regulations at all times when on
company property.
9. Affidavits may not be signed upon ADVICE of legal counsel.
10. At the conclusion of the inspection a debriefing will be held.
240
-------
TABLES
QUALIFICATIONS OF A DESIGNATED ESCORT
• Excellent Reputation - Both Inside and Outside Company
• Knowledgeable about:
- Federal, State and Local Regulations
- Company Policies and Procedures
- Standard Operating Procedures, Methods
- Personnel
- Laboratory Functions/Areas
• 6th Sense—ability to anticipate an inspector's interest, reactions, and conclusions
• Pleasant, but Business-like
• Persuasive, but not Argumentative
• Alert and Observant
• Clear in Speech and Writing
• Experienced in Laboratory and Technical Area
• Diplomatic and Flexible
241
-------
TABLE 4
CONTENTS OF AN INSPECTOR'S SURVIVAL KIT
• History of the company
• Organization charts
• Floor plans
• Last inspection findings with corrective action
• Index of standard operating procedures/methods/procedures, etc.
• CVs of key personnel
• List of key personnel with phone numbers
• Company policy and procedure on inspections—ground rules
• Authorization as designated escort
• Inspection checklist
• Work flow patterns/charts
• Security/safety procedures
• Validation records (data, LIMS, environmental, etc.)
• Employee training records
• Pencils, pens, paper, and clipboard
242
-------
TABLES
"DOS" AND "DON'TS"
• Take a lawyer's approach to questions: "yes", "no" or "I don't know"
• Give positive constructive answers—but don't stretch the truth
• Do not lie or ask others to do so
• Do not hide or make changes in requested documents
• Be helpful—but don't volunteer unnecessary information
• Respond promptly—but don't guess or pretend to know answers
• Ask for feedback—but don't concede violations
• Explain company ground rules in advance—but don't be quarrelsome or obstructive
• Make suggestions—but don't dominate or push
• Prepare other personnel for questions—but don't "huddle secretively"
• Show that you are knowledgeable about the law and regulations—but don't be rude
• Recognize the inspector's position—but don't be servile
• Do not offer gifts
243
-------
36
IMPROPER HAZARDOUS WASTE CHARACTERIZATIONS
FINANCIAL AND COMPLIANCE IMPLICATIONS
Richard M. Walka. Senior Associate, William F. Cosulich Associates, 330 Crossways Park
Drive,Woodbury, New York, 11797-2015;
Frank A. Langone, International Business Machines Corporation, Thomas J. Watson
Research Center, Mail Code 31061, PO Box 218, Yorktown Heights, New York, 10598;
Richard P. Russell, Project Engineer, William F. Cosulich Associates, 330 Crossways
Park Drive, Woodbury, New York, 11797-2015.
ABSTRACT
Generators of waste often assume that "when in doubt" concerning whether a waste
material is hazardous or nonhazardous, transporting and disposing of the material as a
hazardous waste, with an accompanying manifest, affords them some "protection" that
ordinarily would not otherwise exist. The purpose of this paper is to present an array of
potential environmental regulatory and financial liability implications regarding generators
who subscribe to this "protective filer" approach in lieu of proper waste characterizations.
As the paper presents, those generators who do not render accurate and well-founded
waste characterizations pursuant to appropriate federal and/or state requirements, can
actually increase their liability. This is particularly important when the waste is not a
hazardous waste but is managed as if it were by electing to use a hazardous waste
manifest accompanying its management.
Based on a review of various federal regulatory and certain state regulatory and financial
requirements, the paper reviews the responsibilities attendant to the use of a manifest and
the costs associated with these responsibilities. Basically, generators of hazardous waste
must comply, at some cost, with the hazardous waste management regulations of the
federal government, or an authorized state as prescribed by RCRA. These requirements
include, but are not necessarily limited to, the following:
• Increased accumulation and storage costs.
• Compliance with hazardous waste generator requirements, including:
- EPA ID Number
Accumulation requirements
Storage requirements
- Inspections
- Manifest
- Record Keeping
244
-------
• Increased transportation costs.
• Increased disposal costs.
• Waste minimization certification and associated waste minimization implica-
tions.
• Increased financial liability associated with taxes and regulatory fees.
• Cost associated with non-compliance.
Overall, the paper supports the approach that proper and judicious hazardous waste
characterizations can go a long way toward reducing regulatory compliance and financial
liabilities for generators of waste. Toward that end, the design and implementation of a
properly crafted and implemented Quality Assurance/Quality Control Plan with clear Data
Quality Objectives is likely to be the cost-effective approach in reducing overall long-term
waste management costs.
INTRODUCTION
This paper is not a "how to" on when or under what circumstances one is required to
make a hazardous waste determination, or a primer reviewing a step by step approach to
rendering a proper characterization. Rather, the purpose of this paper is to illustrate why
generators of waste should render accurate and well-founded determinations as to whether
the material is a hazardous waste pursuant to appropriate federal and/or state requirements.
The paper will also review the potential financial and compliance implications of
managing nonhazardous waste as hazardous waste in New York State. When waste is
classified as hazardous, its management requires the use of a manifest for off-site
transportation. Utilizing a manifest results in a number of explicit and implicit protections
and liabilities. The implications of using the hazardous waste manifest will be a major
portion of the discussion.
The recognition of the uniform hazardous waste manifest among waste generators across
the nation is probably second only to the IRS 1040 Form. Since its creation by the
Environmental Protection Agency in the late 1970's, the hazardous waste manifest and the
vast information network it supports, remains the cornerstone of RCRA's national "cradle
to grave" hazardous waste tracking system. We will explain why the manifest possibly
has more responsibilities than protections associated with its use.
It has been our experience that some clients firmly believe that "when in doubt" waste
should be classified as hazardous. These clients assume that transporting the material
as a hazardous waste, with an accompanying manifest, is always "safer," affording them
some protection that ordinarily would not otherwise exist. This "protective filer"
mentality has a number of enforceable regulatory liabilities, as well as financial
245
-------
implications, associated with it which can be burdensome when the facility actually does
not generate any hazardous waste. These considerations are in addition to the high cost
of hazardous waste disposal.
Obviously, this paper supports utilizing the hazardous waste manifest when appropriate,
and fosters the concept of proper and judicious hazardous waste characterizations.
However, we are opposed to using hazardous waste manifests when solely
generating/transporting, nonhazardous industrial waste.
Before we discuss the financial and compliance liabilities associated with transporting
manifested nonhazardous waste, we will briefly describe the hazardous waste program.
WHEN IS A WASTE A HAZARDOUS WASTE?
Section 1004(5) of RCRA defines hazardous waste as a "...solid waste, or combination
of solid wastes, which because of its quantity, concentration or physical, chemical or
infectious characteristics may:
(A) cause or significantly contribute to an increase in mortality or an increase
in serious irreversible, or incapacitating reversible illness;
or
(B) pose a substantial present or potential hazard to human health or the
environment when improperly treated, stored, transported or disposed of,
or otherwise managed."
While this statutory definition is subjective, it clearly states that in order for a material
to be a hazardous waste, it must first be a "solid waste," as defined by RCRA. In order
for a solid waste to be defined as a hazardous waste, it must meet the following
conditions:
• Is not excluded from regulation as a hazardous waste, and;
• Exhibit any of the characteristics of a hazardous waste, and/or;
• Be named a hazardous waste and listed by regulation as such, or;
• Is a mixture containing a characteristic waste/listed hazardous waste and a
nonhazardous solid waste, unless the mixture is specifically excluded or no
longer exhibits any of the characteristics of hazardous waste. A mixture
containing a nonhazardous waste and a listed hazardous waste will remain a
listed hazardous waste.
246
-------
In the preceding section, we provided a general definition of hazardous waste. However,
there are two principle mechanisms to determine whether a waste is a hazardous waste.
First, one must determine if it is a "listed waste," so named because it is specifically listed
as such by the EPA or a State as part of its hazardous waste regulations. Secondly, based
on knowledge or laboratory analysis, one must determine if it exhibits any characteristics
of a hazardous waste: ignitability, corrosivity, reactivity and toxicity.
CHARACTERISTIC WASTE
Ignitabilitv
A solid waste that exhibits any of the following properties is considered a hazardous
waste due to its ignitability:
• A liquid, except aqueous solutions containing less then 24 percent alcohol, that
has a flash point less than 60°C(140°F);
• A nonliquid capable under normal conditions of spontaneous and sustained
combustion;
• An ignitable compressed gas in accordance with Department of Transportation
(DOT) regulation;
• An oxidizer per DOT regulation.
EPA's reason for initially including ignitability as a characteristic was to identify waste
that could cause fires during transport, storage or disposal.
Corrosivity
A solid waste that exhibits any of the following properties is considered a hazardous
waste due to its corrosivity:
• An aqueous material with pH less than or equal to 2.0 or greater than or equal
to 12.5;
• A liquid that corrodes steel at a rate greater than 0.25 inch per year at a
temperature of 55°C (130°F).
EPA chose pH as an indicator of corrosivity because waste with high or low pH can react
dangerously with other waste or cause toxic contaminants to migrate from certain waste.
Steel corrosion was chosen because waste capable of corroding steel can escape from its
container.
247
-------
Reactivity
A solid waste that exhibits any of the following properties is considered a hazardous
waste due to its reactivity:
• Normally unstable and reacts violently without detonating;
• Reacts violently with water;
• Forms an explosive mixture with water;
Generates toxic gases, vapors or fumes when mixed with water;
• Contains cyanide or sulfide and generates toxic gases, vapors or fumes at a pH
of between 2 and 12.5;
Capable of detonation if heated under confinement or subjected to strong
initiating source;
• Capable of detonation under standard, temperature and pressure;
• Listed by DOT as Class A or B explosive.
Reactivity was chosen as a characteristic to identify unstable waste that can pose a
problem at any stage of that waste management cycle.
Toxicity
The toxicity characteristic test is designed to identify waste likely to leach particular toxic
constituents into the groundwater as a result of improper management.
To ascertain if a solid waste is hazardous because of the toxicity characteristic,
constituents are extracted from the waste in a manner designed to simulate the leaching
action which occurs in landfills. The extract is then analyzed to determine if it possesses
any hazardous constituents listed in Table 1. If the concentrations of the toxic
constituents are equal to or exceed the regulatory levels listed, the waste is classified as
hazardous.
Characteristic hazardous wastes are defined by certain physical/chemical criteria which
may require a representative waste sample analysis by the generator. The Toxicity
subcategory is more likely than the other characteristics to require chemical analysis.
Consequently, a waste generator must be very careful in selecting/paying for the correct
protocols to characterize his waste for Toxicity.
248
-------
Table 1
TOXICITY CHARACTERISTIC CONTAMINANTS
AND REGULATORY LEVELS
EPA Hazardous
Waste Number
Contaminants
Chronic Toxlcity
Reference Level (mg/11
Basis*
Regulatory
Level (meflf
D004
D005
D018
D006
DOI9
D020
D021
D022
D007
D023
D024
D025
D026
D016
D027
D028
D029
D030
D012
D031
D032
D033
D034
D008
D013
D009
D014
D035
D036
D037
D038
D010
D011
D039
DO! 5
D040
D041
D042
D017
D043
Arsenic
Barium
Benzene
Cadmium
Carbon tetrachloride
Chlordane
Chlorobenzene
Chloroform
Chromium
o-Cresol
m-Cresol
p-Cresol
Creiol
2,4-D
1,4-Dichlorobenzene
1,2-Dichloroethane
1,1 -Dichloroethylene
2,4-Dinitrotoluene
Endrin
Heptachlor (and its hydroxide)
Hexachlorobenzene
Hoxachloro-1,3-butadiene
Hexachloroethane
Lead
Lin dan c
Mercury
Methoxyclor
Methyl ethyl ketone
Nitrobenzene
Pentachlorophenol
Pyridine
Selenium
Silver
Tetrachloroethylene
Toxaphene
Trichloroethylene
2,4,5-Trichlorophenol
2,4,6-Thchlorophenol
2,4,5-TP (Silvex)
Vinyl chloride
0.05
1.0
0.005
0.01
0.005
0.0003
1
0.06
0.05
2
2
2
2
0.1
0.075
0.005
0.007
0.0005
0.0002
0.00008
0.0002
0.005
0.03
0.05
0.004
0.002
0.1
2
0.02
1
0.04
0.01
0.05
0.007
0.005
0.005
4
0.02
0.01
0.002
MCL
MCL
MCL
MCL
MCL
RSD
RED
RSD
MCL
RtD
RfD
RfD
RfD
MCL
MCL
MCL
MCL
RSD
MCL
RSD
RSD
RSD
RSD
MCL
MCL
MCL
MCL
RfD
RfD
RfD
RfD
MCL
MCL
RSD
MCL
MCL
RfD
RSD
MCL
MCL
5.0
100.0
0.5
1.0
0.5
0.03
100.0
6.0
5.0
200.01
200.0'
200.0'
200.0'
10.0
7.5
0.5
0.7
0.13'
0.02
0.008
0.13*
0.5
3.0
5.0
0.4
0.2
10.0
200.0
2.0
100.0
5.0"
1.0
5.0
0.7
0.5
0.5
400.0
2.0
1.0
0.2
* MCL = Maximum Contaminant Level or National Interim Primary Drinking Water Standard
RSD " Risk-Specific Dose
RfD = Reference Dose
* The regulatory level equals the chronic toxicity reference level times a dilution/attenuation factor (DAF) of 100, unleis otherwise noted.
' If o-, m-, and p-cresol concentrations cannot be differentiated, the total cresol (D026) concentration is used. Note that D026 was added to the final
rule for this purpose, but is not a new constituent
b The quantitation limit (i.e., five times the detection limit) is greater than the calculated regulatory level; thus, the quantitation limit becomes the
regulatory level.
Source: 55 FR 11804 and 11815-11816.
1-24-94
249
-------
There are financial implications tied to Toxicity wastes in two areas. First, the prescribed
leaching protocol (Toxicity Characteristic Leaching Procedure - TCLP) is an expensive
protocol to run. However, it may not be required if the waste is a 100% solid matrix or
a solid/water matrix, for which a total constituent analysis has been performed. The
extraction is also not required if the waste is a liquid with no solid phase. Thus, an
understanding of when the TCLP is needed has a significant influence on the cost of
waste characterization.
Secondly, it is important to compare the appropriate Toxicity analytical result with the
regulatory level to avoid incorrect hazardous waste determinations (false positives) which
will result in the compliance/financial implications that will be discussed below. For
example, the total constituent analysis of a 100% solid waste sample would be reduced
by a factor of 20, before comparing it to the regulatory level (the 1 to 20 factor is derived
from the 1 to 20 dilution factor that is part of the TCLP protocol).
A good source of information on the entire Toxicity category and, in particular, on
correctly using the TCLP procedure/interpreting analytical results is: "Technical
Assistance Document for Complying with the TC Rule and Implementing the Toxicity
Characteristic Leaching Procedure (TCLP)," May 1993, U.S. EPA-Region II.
LISTED WASTE (SPECIFIC/NONSPECIFIC)
As we mentioned above, solid waste is considered hazardous waste if it is "listed" as one
of the following:
• Nonspecific source waste - These are generic wastes, commonly produced by
manufacturing and industrial processes. Examples from this list include spent
halogenated solvents used in degreasing and wastewater treatment sludge from
electroplating processes.
• Specific source waste - This list consists of wastes from specifically identified
industries such as wood preserving, petroleum refining and organic chemical
manufacturing. These wastes typically include sludges, still bottoms,
wastewaters, spent catalysts and residues; e.g., wastewater treatment sludge
from the production of pigments.
• Commercial chemical products - The third list consists of specific unused
commercial chemical products or manufacturing chemical intermediates. The
key criterion for this category is that the waste is unused, e.g., off-spec or
spilled materials. This list includes chemicals such as chloroform and
creosote, acids such as sulfuric acid and hydrochloric acid, and pesticides, such
as DDT and kepone.
250
-------
These lists were developed by examining different types of waste and chemical products
to ascertain if they:
• Exhibit one of the four characteristics of a hazardous waste(listed above);
• Meet the statutory definition of hazardous waste;
• Are acutely toxic or acutely hazardous;
• Are otherwise toxic.
It should be noted that individual States may designate additional materials as listed
hazardous wastes. For example, New York State regulates PCB wastes as listed
hazardous wastes (B-codes).
REGULATORY REQUIREMENTS
If a facility produces hazardous waste based on the regulatory criteria discussed above,
it may be classified as a hazardous waste generator. Hazardous waste generators are the
first link in the "cradle to grave" hazardous waste management system established
pursuant to the Resource Conservation and Recovery Act (RCRA). Generators of 100
kilograms of hazardous waste or 1 kilogram of acute hazardous waste per month must
comply with certain enforceable generator standards.
The pretransport regulatory requirements for hazardous waste generators include:
• Obtaining an EPA ID number. One way that EPA monitors and tracks
generators is assigning each generator a unique identification number.
Without this number, the generator is barred from treating, storing, disposing,
transporting or offering for transport any hazardous waste to any transporter
or treatment, storage or disposal facility;
• Adhering to procedures for handling hazardous waste before transport;
• Manifesting hazardous waste for off-site transportation;
• Maintaining a 24-hour Emergency Contact for each shipment;
• Record keeping and reporting;
• Proper packaging to prevent leakage of hazardous waste during normal
transport conditions and in potentially dangerous situations (e.g., when a drum
falls out of a truck);
251
-------
• Identifying the characteristics and dangers associated with the waste being
transported through labeling, marking and placarding of the packaged waste;
• Preparing applicable Land Disposal Restriction (Land Ban) shipping notices.
It is important to note that these pretransport regulations only apply to generators shipping
waste off-site.
In addition to the requirements outlined above, EPA and authorized states also developed
pretransport regulations for accumulation of waste prior to transport. A generator can
accumulate hazardous waste on-site for 90 days or less without a permit, as long as the
following requirements are met:
• Proper Storage - The waste is properly stored in containers or tanks marked
with the words "Hazardous Wastes" and the date on which accumulation
began. The waste must also be inspected at least weekly and inspections
records maintained.
• Emergency Plan - A contingency plan and emergency procedures to use in an
emergency must be developed.
• Personnel Training - Facility personnel must be trained in the proper handling
of hazardous waste.
• Preparedness and Prevention Measures - Providing adequate security measures,
signage, and communication systems.
The 90-day period allows a generator to collect enough waste to make transportation more
cost-effective; that is, instead of paying to haul several small shipments of waste, the
generator can accumulate waste until there is enough for one large shipment.
THE HAZARDOUS WASTE MANIFEST
The manifest is the fundamental element of the hazardous waste tracking system. The
uniform hazardous waste manifest is the document which accompanies shipments of waste
and tracks the material from the generator (the cradle) to the ultimate disposal facility (the
grave). The RCRA manifest requires the following information:
• Name and EPA identification number of the generator, transporter(s) and the
facility where the waste is to be treated, stored or disposed;
• U.S. DOT description of the waste being transported;
• Quantities of waste being transported; and
252
-------
• Address of the treatment, storage or disposal facility to which the generator is
sending the waste.
• 24-hour emergency contact telephone number.
It is especially important for the generator to prepare the manifest properly, since the
generator is responsible for the hazardous waste produced and its ultimate disposition.
Waste Minimization
When Congress passed the Hazardous and Solid Waste Amendments (HSWA) in 1984,
it established a framework aimed at eliminating specific forms of waste management such
as land disposal, in favor of more technologically advanced, permanent destruction
methods, such as incineration.
Among its complex and far reaching provisions, the HSWA contained a statutory
provision which initiated waste minimization criteria. Sections 3002(a)(6), regarding the
preparation of biennial waste reduction reports, and 3002(b) of HSWA, entitled, "Waste
Minimization," require generators of hazardous waste to practice and report on waste
minimization activities. It requires generators of hazardous waste to sign a specific
certification on the manifest indicating that they are doing all that is "economically
practicable" towards reducing the volume or quantity and toxicity of the hazardous waste
generated at a facility. It is a signed certification that generators are in full compliance
with HSWA waste minimization criteria.
Section 3002(b) of HSWA, entitled "Waste Minimization," reads as follows:
"(b) Waste Minimization - Effective September 1, 1985, the manifest required by
subsection (a)(5) shall contain a certification by the generator that -
(1) the generator of the hazardous waste has a program in place to reduce the
volume or quantity and toxicity of such waste to the degree determined by the
generator to be economically practicable; and
(2) the proposed method of treatment, storage, or disposal is that practicable
method currently available to the generator which minimizes the present and
future threat to human health and the environment."
With regard to the requirement for biennial reporting of waste reduction efforts to
regulatory agencies, Section 30021(a)(6) reads as follows:
"(6) submission of reports to the Administrator (or the State agency in any case
in which such agency carries out a permit program pursuant to this subtitle) at
least once every two years, setting out -
253
-------
(A) the quantities and nature of hazardous waste identified or listed under
this subtitle that he has generated during the year;
(B) the disposition of all hazardous waste reported under subparagraph
(C) the efforts undertaken during the year to reduce the volume and toxicity
of waste generated; and
(D) the changes in volume and toxicity of waste actually achieved during the
year in question in comparison with previous years, to the extent such
information is available for years prior to enactment of the Hazardous
and Solid Waste Amendments of 1984."
In addition to these enforceable waste reduction mandates brought about by the HSWA
manifest certification, many states have already passed additional statutes and regulations
to address pollution prevention and waste reduction.
For example, in August 1990, New York State passed a law requiring facilities that
generate and have the potential to release hazardous wastes and toxic substances into the
environment reduce, to the maximum extent possible the volume or quantity and toxicity
of wastes, whether emitted into the air, discharged into the waters, or treated and disposed
of in a permitted facility. The waste reduction may be achieved by implementing
technically feasible and economically practicable waste reduction technology, process or
operation changes. The legislature declared that implementing such measure will help the
State achieve an overall reduction in the generation and release of hazardous waste of
fifty percent over the next ten (10) years.
This law requires generators of hazardous wastes to prepare, implement and submit a
Hazardous Waste Reduction Plan (HWRP) to the New York State Department of
Environmental Conservation (NYSDEC). The HWRP, which is reviewed for acceptance
by NYSDEC, must be updated biennially and annual status reports must be submitted.
Failure to submit an acceptable plan precludes the generator from signing the hazardous
waste manifest certification.
The requirements of the Hazardous Waste Reduction Plan include, but are not limited to,
the following:
Quantification of hazardous waste(s)
• Description of hazardous waste source(s) of generation and disposal method(s)
• Indices of hazardous waste generation to production (i.e., output from, or input
to, the process generating the waste stream)
254
-------
• Submission of a hazardous waste generator summary
• Cost estimate(s) for managing each waste
• Evaluation of technical feasibility and economical practicability of
implementing waste reduction options
• Listing of technically feasible and economically practicable waste reduction
measures and schedule for implementing identified waste reduction measures
• Description of corporation's and facility's waste reduction policy
• Identification of party responsible for implementation of waste reduction plan
• Identification of waste reduction measurement(s)
• Identification of employee training programs
• Estimate of anticipated hazardous waste reduction
• Estimate of anticipated transference of hazardous waste into other
environmental media
• Submission of Hazardous Waste Reduction Program Summary (HWRP)
• Biennial updates of HWRP
• Annual status reports
In addition, to the state statute, EPA has published its interim final rule regarding waste
minimization program requirements. Although published in the Federal Register as an
interim final rule, the guidance puts "additional enforcement teeth" behind the hazardous
waste manifest certification requirements mandated by the Hazardous and Solid Waste
Amendments.
TAX ASSESSMENT/REGULATORY FEES
In addition to the regulatory/compliance implications discussed above, there is an
additional liability when hazardous waste is generated at your facility and properly
managed via a manifest, or a nonhazardous waste is accompanied by a manifest. That
liability is a special assessment and regulatory fee. In New York State, these assessments
and fees are administered by the New York State Departments of Taxation and Finance
and Environmental Conservation, respectively. The "bottom line" is that the generation
255
-------
of hazardous waste in New York State can affect your "bottom line." Another good
reason to make sure your hazardous waste is in fact, hazardous. Let's briefly review the
two revenue programs.
As mentioned above, the first tax, entitled, "Special Assessments on Generation,
Treatment or Disposal of Hazardous Waste in New York State" is administered by the
Department of Taxation and Finance and is self reported by generators and TSDFs within
the state on Form TP-550. This self reporting program is managed comparable to other
state taxes, that is, it is prepared and reported by the generator and subject to review and
audit by the Department of Taxation and Finance. In its simplest form, the Special
Assessment, or "waste end assessment" as it is commonly referred to, is calculated by
generators based on the tons of hazardous waste generated in New York State that
received on site treatment or disposal or that were designated for removal or removed
from the site of generation for treatment or disposal or for storage prior to such treatment
or disposal during the reporting period. With regard to treatment and/or disposal
facilities, these entities are only required to report the tons of hazardous waste received
from generators outside New York State for treatment disposal or for storage prior to such
treatment or disposal (this avoids double counting.)
In accordance with the Environmental Conservation Law (§27-0923 Special Assessments
on Hazardous Wastes Generated) the following assessment rate schedule currently
applies:
Assessment Rate
Category (in dollars per Ton)
Tons disposed of in landfill on-site of generation $27
Tons designated for removal or removed from the site of
generation for disposal in a landfill or designated for
removal or removed from the site prior to disposal in a landfill $27
Tons designated for removal or removed from the site of generation
for treatment or disposal (except by landfill or incineration), or for
storage prior to such treatment or disposal $16
Tons designated for removal or removed from the site of
generation for incineration or for storage prior to incineration . $9
Tons incinerated on the site of generation $2
Tons received from out-of-state for landfill
disposal or for storage prior to such disposal $27
256
-------
Tons received from out-of-state for treatment or disposal other than
landfill or incineration, or for storage prior to such treatment
or disposal $16
Tons received from out-of-state for incineration or
for storage prior to incineration $9
As can be seen from the table above, the rates are structured to provide an incentive for
disposal/treatment of waste by incineration and discourages disposal via landfilling. A
deduction may be taken for waste that is reclaimed.
These special assessments are paid quarterly and are due to the Department by the 20th
day of the month after the end of each calendar quarter. While the fees may not seem
onerous at first glance, it does not take much material to achieve a ton of waste. For
instance, a 55 gallon drum of water (the density used to calculate the fee) weighs
approximately 459 Ibs or approximately .23 of a ton. So one can see how quickly the
special assessments can take a bite of your bottom line.
The second financial implication of generating hazardous waste is the Regulatory Fee,
which is administered by the New York State Department of Environmental Conservation
pursuant to 6 New York Codes, Rules and Regulations Parts 480 through 486 (Revised
1991).
Unlike the "waste end assessment" discussed above, the Hazardous Waste Program Fee
prescribed in Part 483 is an invoice prepared and sent by the Department based on data
from annual generator reports and manifest documents submitted.
Basically, for generators of hazardous waste, the hazardous waste program fee is currently
determined as follows:
• $1000.00 for generators of equal to or greater than 15 tons per year and less
than or equal to 100 tons per year of hazardous waste,
• $6,000.00 for generators of greater than 100 tons per year and less than or
equal to 500 tons per year of hazardous waste,
• $20,000.00 for generators of greater than 500 tons per year and less than 1,000
tons per year of hazardous waste, and
• $40,000.00 for generators of greater than 1,000 tons per year.
In addition to the above, generators of equal to or greater than 15 tons per year of
hazardous wastewater are assessed $3,000.00.
257
-------
The use of a manifest New York State should only accompany shipments of hazardous
waste as defined under Part 371. In fact, Part 372.2(b)(6) states ..."Use of a Uniform
Hazardous Waste Manifest constitutes a determination by the generator that the solid
waste is a hazardous waste in New York and/or the state of generation."
Therefore, the utilization of a manifest accompanying a waste material that is clearly not
a hazardous waste material may be interpreted as a violation of New York State
hazardous waste regulations and therefore could be enforceable.
SUMMARY
We have just reviewed a number of regulatory and financial requirements that generators
of hazardous waste in New York State and elsewhere across the country are obligated to
comply with in order to protect human health and environment. Presentations such as this
are typically offered to assist hazardous waste generators in achieving and maintaining
regulatory compliance... and thereby... minimize liability from regulatory violations and
associated fines.
However, as we stated at the outset, the objective of this presentation is quite the
opposite. The focus here is to identify how one's liability is actually increased by
utilizing the same regulatory system discussed above (i.e., manifest document, etc.) when
it simply is not required because the waste is not truly a hazardous waste.
Example. Facility A uses a water soluble alkaline powdered product to aid in degreasing
engine electric motors and transmission parts prior to rebuilding as part of its scheduled
maintenance program. Rather than establish a proper waste characterization program with
appropriate data quality objectives and quality assurance/quality control program, the
facility manager characterizes the spent solution as a characteristic hazardous waste and
ships it off-site via a licensed transporter with a signed manifest. The waste is
characterized as corrosive (D002). This practice continues for a number of years, with
manifests documenting ten's of thousands of gallons of "hazardous waste" being generated
at the facility. (In fact, subsequent analytical data and proper waste characterization
determined the material not to be hazardous.)
What are the liabilities associated with this scenario?
First, the use of a manifest signifies that the entity is a hazardous waste generator and,
as such, is required to comply with appropriate federal and state generator requirements.
Most important among these requirements is obtaining an EPA ID number. The ID
number must be obtained by the generator in order to use a manifest. Now that you have
declared generator status, the following enforceable requirements are applicable:
• compliance with specific procedures for handling waste
258
-------
record keeping and reporting
• proper labeling, marking and placarding
develop and implement an emergency plan
• personnel training
• Preparedness and prevention measures
• annual generator report.
The above requirements are enforceable and may be subject to fine if determined not to
be satisfactory to government inspectors. However, if the generator manifests wastes as
"non-hazardous," the above requirements are not applicable.
Remember, the signature box on the manifest provides for a certification indicating that
the generator has a program in place to reduce the volume and toxicity of waste, and that
the method of treatment, storage or disposal minimizes present and future threats to
human health and the environment. While these are commendable objectives, they are
not applicable for a small manufacturing facility that does not generate hazardous waste
in the first place.
There are also hazardous waste reduction plan requirements. Hazardous waste manifest
documents are utilized to quantity the amounts of hazardous waste being generated at a
facility. This process can include an otherwise non hazardous waste generator on the list
of hazardous waste generators required by state law to prepare a Hazardous Waste
Reduction Plan. The use of manifests and the submission of annual generator reports at
a facility that does not generate hazardous waste in the first place can be an unnecessary
regulatory burden. It costs resources to develop, submit and implement a hazardous waste
reduction plan.
Last but not least, there are the "waste end assessments" reportable and payable to the
Department of Taxation and Finance in New York State and regulatory fees assessed
directly by the Department of Environmental Conservation. While perhaps not perceived
as an overwhelming burden, when taken as a whole in consideration with the other
prescribed regulatory requirements, improper hazardous waste characterizations can take
a bite from your bottom line.
In short, the time, money and resources spent up front in proper waste characterizations
including the development and implementation of clear data quality objectives and a
quality assurance/quality control program, can go a long way toward reducing
environmental compliance and financial liability.
259
-------
References
Technical Assistance Document for Complying with the TC Rule and Implementing the
Toxicity Characteristic Leaching Procedure (TCLP), May 1993, US EPA Reg. II, DCN
EPA 902-b-93-001.
260
-------
INORGANICS
-------
37
A Comparison of Methods of Measurement of Cr+6 in
Wastewaters, Soils and Sediments: UV-Visible Spectrometry and
Ion Chromatography
10th Annual Waste Testing and Quality Assurance Symposium
Arlington, Va. July 1994
Stuart J. Nagourney, NJDEPE, and John Birri and Leonard Vu,
USEPA, Region II
Cr is one the most interesting and chemically diverse
elements in the periodic table. It routinely exists in the
environment in inorganic compounds or organometallic
complexes in several oxidation states. The most common
species are the +3 and +6 forms; in its trivalent form, Cr is
often a nutrient while Cr 6 is well recognized as a probable
agent of lung cancer in man1 as well as having been reported
to produce gastrointestinal disorders, dermatitis and
ulceration of the skin.
From the 1900's to the early 1970's, it is estimated that
more than 2 million tons of chromite ore processing residue
were created by a variety of manufacturing processes in
Hudson County, New Jersey. Some of this material was used as
fill material. Cr+6 is also a by-product of several
industrial processes and is present at low concentrations in
some commercial products. Public concern about the possible
health affects associated with skin contact and inhalation
from soluble and inhalable Cr "-containing media has driven
efforts to enhance the understand of how Cr behaves in the
environment and ways to accurately measure its speciated
forms. The measurement of Cr 6 is now reguired as part of
many NJ Pollution Elimination Discharge System (NJPDES)
permits.
Any study to accurately assess the levels of speciated Cr in
ambient air, waters, soils, sediments or other media requires
an understanding of several contributing factors: the ability
to obtain a representative sample, the means to
quantitatively remove the Cr from the media of interest
without altering its indigenous oxidation state and the
capability to accurately measure specific ionic forms. The
sampling and sample preparation aspects have been studied by
many investigators. USEPA Method 7196a3 is a colorimetric
method originally designed for water and wastewaters that is
routinely used to measure Cr+6 in non-aqueous extracts.
USEPA Method 218.6 was originally designed to utilize Ion
Chromatography for the measurement of Cr+b in emissions from
incinerators burning municipal sludge; it has been applied to
drinking, ground, waste and seawaters with considerable
success.
261
-------
Most users of the colorimetric method report few problems
with its applicability when Cr+6 sample concentrations are
high or when there are limited concentrations of organic
material present in the matrix. At lower Cr^
concentrations, the ability to quantitate Cr 6 may be
difficult; this is exacerbated when matrix components of non-
aqueous samples produce darkly colored extracts, making
comparison of sample and blank often next to impossible.
This paper compares the results obtained by Methods 7196a and
218.6 on two types of samples that offer the greatest
challenges to detection and quantitation of Cr . These
sample types are:
- effluents from industrial and domestic treatment plants
where permit limits require accurate measurements to insure
regulatory compliance
- sediments and soils from New Jersey which contain
variable levels of Cr 6 and organic material and where the
oxidation state of Cr may have been altered over time due
to natural processes.
The samples were prepared for measurement and aliquots
analyzed by both methods. The comparison of the results by
both methods will offer new insight on options and
limitations for the increasingly-important determination.
References
1. Browning, E., In Toxicity of Industrial Metals, 2nd
Edition, Appleton-Century-Crofts, New York, 1979.
2. The Merck Index, Tenth Edition, Merck & Co. Inc,
Rahway, New Jersey, 1983.
3. Test Methods for Evaluating Solid Waste,
Physical/Chemical Methods, SW-846, 3rd Edition, USEPA,
Office of Solid Waste, Method 7196a
4. USEPA Report, "Determination of Dissolved Hexavalent
Chromium in Drinking Water, Ground Water and Industrial
Wastewater Effluents by Ion Chromatography: Collaborative
Study", Environmental Monitoring Systems Laboratory -
Cincinnati, Report #1452, 1991.
262
-------
38
Hexavalent Chromium Methodology for Soils: Results of Extraction
Comparison Research and Multi-Laboratory Holding Time Study
R. Vitale. Quality Assurance Specialist/Principal, G. Mussoline, Senior Quality
Assurance Chemist, Environmental Standards, Inc., Valley Forge, Pennsylvania
19482; J. Petura, Principal, Applied Environmental Management, Inc., Malvern,
Pennsylvania 19355; and B. James, Ph.D., Associate Professor, University of
Maryland, College Park, Maryland 20742
ABSTRACT
Hexavalent Chromium [Cr(VI)] analysis has been performed on a significant number
of soil samples suspected of being contaminated with chromite ore processing residue
(COPR) using a modified version of the SW846 (2nd Edition) Method 3060 followed
by Method 7196A as the extraction and analysis procedures, respectively. The authors
previously evaluated Modified Method 3060 using eight different soil materials,
including field-moist and dried COPR. More recently, Modified Method 3060 has
been compared to four alternate extraction techniques — (1) distilled deionized water,
(2) phosphate buffer extraction (pH 7), (3) Modified Method 3060 alkaline extraction
(without heat), and (4) alkaline extraction with sonication (without heat) — to compare
the recoveries of soluble and insoluble Cr(VI) matrix spikes in four different soil
types. For each sample type, Modified Method 3060 extracted the largest quantity of
Cr(VI).
Additionally, a study was performed to establish a suitable holding time for soil
samples prior to the analysis of Cr(VI) using Modified Methods 3060/7196A.
Thoroughly homogenized, well-characterized samples of COPR were analyzed over a
one-month period after collection by one commercial laboratory and subsequently over
an eight-month period by four different commercial laboratories. The results of the
round-robin testing over time revealed that Cr(VI) is stable in both field moist and
dried COPR samples for at least one month after collection. Complementary testing
showed that Cr(VI) is stable in the Method 3060 alkaline solution for at least 96 hours
after digestion prior to analysis.
This paper summarizes the study design and performance results obtained from the
five different extraction solutions that were used as the preparatory step for Cr(VI)
analysis. It will also present and discuss the data obtained from the holding time study
with regard to (1) comparison of results using Modified Methods 3060/7196A among
the four commercial laboratories, (2) the stability of Cr(VT) in COPR as a function of
holding time after soil sample collection, and (3) the stability of Cr(VI) in the Method
3060 alkaline digestate (extract) prior to analysis with respect to holding time.
263
-------
page 2
INTRODUCTION
Quantification of hexavalent chromium [Cr(VI)] in soils without inclusion of Cr(m)
has significant relevance due to the difference in the toxicity of the +6 and the +3
valence states. An effective and reliable means of extracting Cr(VI) from soil samples
is required. A novel aspect of this requirement is that the media used to extract
Cr(VI) from soil samples not only has to be an efficient extraction media, it must also
be conducive to maintaining Cr in the +6 valence state. The maintenance of Cr in the
+6 valence state necessitates the use of an extraction medium that does not influence
both the reduction of Cr(VI) and the oxidation of Cr(m). This study evaluated the use
of five different extraction methods in the preparation of soil samples for Cr(VI)
analysis by SW846 Modified Method 7196A. Both soluble and insoluble Cr(VI)
matrix spikes were used to evaluate the extraction efficiency of each method.
The extraction methods included: (1) an extraction using water, (2) a phosphate buffer
extraction solution, (3) an alkaline extraction solution employing the use of sonication,
(4) a modified version of SW846 Method 3060 employing heat, and (5) a modified
version of SW846 Method 3060 without heat. These five different extractions were
tested with the same four soil matrices; (1) quartz sand, (2) Cr(VI) chromite ore
processing residue (COPR), (3) anoxic sediment and (4) loam soil. Soluble and
insoluble Cr(VI) matrix spike compounds, as well as Cr(IQ) matrix spike compounds,
were used to evaluate the extraction media on the different soil types. The Cr(VI)
recoveries of the matrix spike compounds after sample digestion and analysis were
assessed to evaluate the efficacy of each solution for extracting Cr(VI), minimizing the
reduction of Cr(VI) to Cr(m), and inhibiting the oxidation of Cr(m) to Cr(VI). The
results of this portion of the study were evaluated to determine (1) the best of the
methods evaluated for extracting all Cr(VI) (both soluble and insoluble) from soils, (2)
if oxidation and/or reduction of Cr occurs during the extraction process and (3) if a
sparingly-soluble form of Cr(VI) is dissolved and recovered from soils.
Additionally, a study was performed in an attempt to evaluate and establish a suitable
holding time for soil samples prior to the analysis of Cr(VI) using SW846 Modified
Methods 3060/7196A. Thoroughly homogenized, well-characterized samples of
COPR and laboratory-prepared synthetic chromium-bearing soil were analyzed over a
one-month period after collection by one commercial laboratory and subsequently over
an eight-month period of time by four different commercial laboratories.
264
-------
page 3
EXPERIMENTAL
Extraction Comparison
For each of the soil materials, 2.5 g of oven-dry soil material (sand, COPR, anoxic
sediment and air-dried loam) was weighed into 60, 250-mL beakers. The beakers
were then divided into five groups of twelve beakers for the five extraction media, and
each group of 12 beakers was divided into four groups of three each for the four soil
matrices. For each group of 12 beakers, the following four spikes of Cr were added
in triplicate:
(1) None
(2) Cr2O3: (36.8 mL of 1.00 g Cr2O3/L suspension per beaker; equivalent
to a spike of 10,000 mg Cr(m)/kg soil material)
(3) BaCrO4: 0.0100 to 0.0200 g of solid compound was added to each
beaker and the exact weight recorded; equivalent to a spike of
approximately 1000 mg Cr(VT)/kg soil.
(4) K2CrO4: 1.25 mL of 2000 mg Cr(VI)/L solution was added to each
beaker; equivalent to a spike of 1000 mg Cr(VI)/kg soil.
To each group of 12 beakers prepared with the four spikes, 50 mL of the following
extracting solutions were added:
(1) Distilled water (pH 5.7)
(2) Phosphate buffer: 0.005 M K2HPO4/0.005 M KH2PO4 (pH 7)
(3) 3060 Extraction Solution: 0.28 M Na2CO3/0.5 M NaOH (pH 11.8)
Added to two sets of 12 beakers; one to be heated and the other to
remain at room temperature.
(4) Sonication Solution: 0.1 M NaOH
All suspensions were swirled and allowed to stand for 60 + 5 minutes before further
treatments were applied as follows:
265
-------
page 4
(1) One-half of the soil suspensions in the Method 3060 solution was heated at 90-
95° C on a hot plate, with stirring, for 60 minutes, and the other half was
allowed to stand at 25 ±2°C.
(2) The water and phosphate buffer suspensions were stirred for 60 minutes at
room temperature,
(3) The suspensions in 0.1 M NaOH were placed in a sonicating bath for 30
minutes, and allowed to stand at room temperature for an additional 30
minutes.
At the end of these treatments, all beakers were brought to a total solution volume of
120 mL of water by weight (assuming the density of water to be 1.0 g/mL) using
distilled deionized water. After centrifuging a portion of the suspensions (25°C, 10
min., 10,000 x g), which is equivalent to the filtration step in Method 3060, the
diphenylcarbazide (DPC) colorimetric method (SW846 Method 7196A) was used to
measure Cr(VI) in the centrifugates after making appropriate dilutions using distilled
water.
Holding Time Study
Three different soil types [a high-level Cr(VI) COPR sample (field moist and dry), a
moderate-level Cr(VI) COPR sample and a synthetic moderate-level Cr(VI) sample]
were analyzed over a period time to determine a time frame over which Cr(VI)
remains stable from sample collection to sample preparation and analysis. In addition
to this, the soil types were evaluated to determine a time frame over which Cr(VI)
remains stable in the alkaline digestate after sample preparation, prior to analysis.
All of the sample types were prepared and analyzed following Modified Methods
3060/7196A.
RESULTS AND DISCUSSION
Extraction Comparison
Absorbance readings were measured at 540 nm (as specified in Method 7196A) to
determine the concentration of Cr(VI) in each extract, including samples spiked with
soluble and insoluble forms of chromium. Table I shows the results of these analyses
readings.
266
-------
Table 1
Sample Type
Extraction Media
Water Solution a
b
c
(1
Phosphate Buffer Solution a
h
c
(1
3060 Solution with Heat a
h
c
d
3060 Solution without Heat a
b
c
d
Sonication a
h
c
d
Sand
Mean Cr(VI)
Cone.
(m«/ks)
<1¥
< 1
31.3
1100
< 1
<1
10.3
1100
<1
< 1
1260
1100
< 1
<1
180
1090
< 1
<1
126
1070
Mean Cr(VI)
Percent
Recovery
0
0
2
110
0
0
1
110
0
0
82
110
0
0
12
109
0
0
9
107
COPR
Mean Cr(VI)
Cone.
(m«/k«)
992
925
916
1760
1120
1090
1130
2110
1400
1400
2220
2500
1250
1260
1500
2290
1190
1160
1430
2250
Mean Cr(VI)
Percent
Recovery
0
0
-6
77
0
0
1
100
0
0
64
110
0
0
22
104
0
0
18
106
Sediment
Mean Cr(VI)
Cone.
(m«/ku)
<1
< 1
< 1
<1
< 1
<1
<1
<1
<1
<1
< 1
< 1
< 1
<1
< 1
< 1
< 1
<1
<1
<1
Mean Cr(VI)
Percent
Recovery
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Loam
Mean Cr(VI)
Cone.
(m«/k«)
< 1
<1
105
1070
<1
<1
37
1100
<1
< 1
512
1070
<1
<1
206
1100
<1
<1
419
1080
Mean Cr(VI)
Percent
Recovery
0
0
5
107
0
0
3
110
0
0
39
107
0
0
18
110
0
0
35
108
h-
CD
CM
Notes:
¥ - All values are means of three replications
a - Unspiked
b - Spiked with Cr(III)
c - Spiked with insoluble Cr(VI)
d - Spiked with soluble Cr(VI)
-------
page5
None of the extraction procedures was able to extract any Cr(VI) from the unspiked
sand, loam and sediment. The soluble Cr(VI) spikes that were added to the sand and
loam were recovered completely by all four extraction procedures and zero percent
recovery was observed in the anoxic sediment, as expected, due to its strongly reduced
condition. The water and phosphate buffer solutions extracted Cr(VI) from the COPR
samples; however, the water extraction procedure recovered only 71 % and the
phosphate buffer solution recovered only 79 % of the Cr(VI) that was extracted by
Modified Method 3060 (with heat). The alkaline digestion solution using sonication
(without heat) and Modified Method 3060 (without heat) also extracted Cr(VI) from
the COPR sample; however, the alkaline digestion solution with sonication recovered
only 85% and Modified Method 3060 (without heat) recovered only 89% of the
Cr(VI) that was extracted by Modified Method 3060 (with heat). Figure 1 shows a
comparison of the amount of Cr(VI) extracted from the unspiked COPR by the five
different extraction techniques.
Figure 2 presents the Cr(VI) percent recoveries observed from the five extraction
techniques following the addition of soluble and insoluble forms of Cr(VI) to the four
different soil types. Modified Method 3060 (with heat) was capable of removing the
greatest amount of Cr(VI) [both soluble and insoluble forms of Cr(VI)]. Another
noteworthy item from Figure 2 is that in all instances, the anoxic sediment displayed
0% recovery. This 0% recovery was observed despite the extraction technique that
was used. This indicates that the soil type (i.e. sediment), and not the extraction
technique, was responsible for the reduction of the Cr(VI) spikes.
Holding Time Study
Thoroughly homogenized, well-characterized samples of COPR (both field moist and
dried) were analyzed over a one-month period after collection by one commercial
laboratory and subsequently over an eight-month period by four different commercial
laboratories. An additional sample of COPR (passed through a 0.85/xm sieve) and a
soil sample (passed through a 0.85/xm sieve) spiked with Cr(VI) by a commercial
laboratory have been repeatedly analyzed by three different commercial laboratories
over time. The results of the round-robin testing over time revealed that Cr(VI) is
stable in both field moist and dried COPR samples for at least one month after
collection, indicating that a holding time of at least one month for Cr(VI) is
appropriate. Figure 3 shows the results of the Cr(IQ) analysis over the eight-month
period.
Complementary testing showed that Cr(VI) is stable in the Method 3060 alkaline
solution for at least 96 hours after digestion, prior to analysis. The commercial
laboratories involved in the study were requested to analyze each sample for Cr(VI)
268
-------
Figure 1
Cr(VI) Concentrations for COPR Using Different Digestion Solutions
CJ
s
en
CD
CM
Water Solution
Phosphate Buffer Solution 3060 Solution (Heat) 3060 Solution (No Heat)
Digestion Solution
Sonication
-------
Figure 2
Cr(VI) Recovery for Sample Types Spiked with Potassium Chromate as a
Function of the Extraction Procedure
Quartz Sand
COPR Anoxic Sediment
Sample Type
Loam
! D Water Solution
Q Phosphate Buffer Solution • 3060 Solution with Heat D 3060 Solution without Heat D Sonication
O
r^
CM
Cr(VI) Recovery for Sample Types Spiked with Barium Chromate as a Function
of Extraction Procedure
Quartz Sand
COPR Anoxic Sediment
Sample Type
Loam
Water Solution
I Phosphate Buffer Solution • 3060 Solution with Heat D 3060 Solution without Heat D Sonication
-------
Figure 3
High Field Moist COPR Cr(VI) Cone, over Time
700 j
600
If 50°
"to
" 400
c
o
•i 300
U 200 H
f> r*l
-t Tf •» -J- f TT -r
C\ C\ O-v C\
5; Si
fl r"?
«S fS
o o
I -t---l- ---I
f) <^1
>-i 06 OD o5 55
C-< ^" t"^ "^ V—\
O *•"< »~* »™1
-------
page 6
upon receipt and again at varying time periods from the initial analysis. This testing,
as summarized on Figure 4, revealed that the Cr(VI) concentration did not change
over time once the sample had been digested by the alkaline digestion procedure
(Modified Method 3060). The longest holding period for the digestates was 96 hours,
and consistent results were obtained among the participating laboratories.
CONCLUSIONS
The results of this study demonstrated that a modified version of SW846 Method 3060
(with heat) was the most effective procedure for removing Cr(VI) from the four soil
types investigated. The results also confirmed that Modified Method 3060 (with heat)
was the most capable method of extracting all forms of Cr(VI) (soluble and insoluble)
of the extraction methods investigated. Matrix spike recoveries also revealed that this
method is capable of maintaining Cr(VI) in the +6 oxidation state.
Further study has revealed that Cr(VI) in COPR samples is stable for up to a one-
month period from collection. Subsequent studies have shown that Cr(VI) is stable in
the alkaline 3060 Solution (after digestion) for at least 96 hours.
272
-------
Figure 4
Mean Cr(VI) Concentrations Observed by Four Commercial
Laboratories
600 -r
500 -
S 400
u
c
o
U
a-
§
u
300 -
200 --
100
CO
1^
C\J
D
Laboratory
Mean Cr(VI) Cone, (mg/kg) (First Analysis) D Mean Cr(VI) Cone, (mg/kg) (Second Analysis)
Notes:
Laboratory A reanalyzed after 96 hrs.
Laboratory B reanalyzed after 72 hrs.
Laboratory C reanalyzed after 48 hrs.
Laboratory D reanalyzed after 24 hrs.
-------
USEPA (United States Environmental Protection Agency), 1992 Test Methods for
Evaluating Solid Wastes. Physical/Chemical Methods. SW-846, Third Edition, Final
Update, Office of Solid Waste and Emergency Response, Washington D.C.
USEPA (United States Environmental Protection Agency), 1982. Test Methods for
Evaluating Solid Wastes. Physical/Chemical Methods. SW-846, Second Edition.
Office of Solid Waste and Emergency Response, Washington D.C.
James, B.R., and R.J. Bartlett. 1988. Chromium in the Natural and Human Environments.
p. 267. John Wiley & Sons, Inc.
274
-------
39
ENZYME-LINKED IMMUNOASSAY (ELISA) FOR THE DETECTION OF
MERCURY IN ENVIRONMENTAL MATRICES
Craig Schweitzer. Product Development Manager, BioNebraska, Inc., Lincoln, Nebraska,
68524, L. Carlson, Scientist, BioNebraska, Inc., Lincoln, Nebraska, 68524, B. Holmquist,
Vice President of Research and Development, BioNebraskaJnc., Mai Riddell, Managing
Director, BioNebraska, Inc., D. Wylie, Consultant, BioNebraska, Inc.
ABSTRACT
Immunochemical-based analytical methods are widely used in the medical diagnostic field,
but they have only recently been adapted for field-portable environmental applications.
BioNebraska has developed such an immunoassay for the detection of mercury in
environmental samples. The assay allows for real-time, user-friendly generation of data at
a fraction of the cost of traditional methods. The assay is available in two formats, a
microplate format for large volume, quantitative analysis of samples in the laboratory and
a tube format for rapid semi-quantitative analysis in the field.
The environmental sample, typically 5 grams of soil or sediment, is extracted for ten
minutes with a mixture of hydrochloric and nitric acids. A buffer is added and the sample
is filtered and diluted by means of a dropper bottle with an enclosed filter. The samples
are then ready for analysis by the immunoassay. NIST traceable reference samples are
included for extraction and comparison with the test samples. After extraction, the assay
itself can be done in less than twenty-five minutes and consists of four reagents, each of
which is added for a 5-minute incubation. The reagents include the extracted sample, a
monoclonal antibody specific for mercuric ions, a secondary enzyme-conjugated antibody
specific for the monoclonal, and a substrate solution which is oxidized by the enzyme on
the secondary antibody. The enzymatic reaction which produces the color development is
terminated, and the mercury concentration is determined relative to the reference
standards by means of battery-operated, field-portable, differential photometer available
from BioNebraska. The dilutions and additions of reagents are facilitated by dropper
bottles provided in the kit. The short times required for extraction and immunoassay
allows the user to produce real-time data in less than 40 minutes. The convenience and
rapidity of the assay allow field testing of multiple samples to be analyzed in a short time
frame, resulting in lower site evaluation and clean up costs, and fewer samples that require
analysis by slower, more expensive methods. Based on in-house and independent field
results, the assay appears to be well-suited for low-cost, real-time, user friendly field
screening of mercury in the environment. Other substances present in environmental
matrices do not appear to interfere with the assay, and the results correlate well with
traditional analytical methods. In addition, preliminary data suggests that the
immunoassay can be applied to the measurement of mercury in seafood and animal tissues,
so that potential problems resulting from biomagnification of mercury can be identified
before the contaminated food sources are used for human consumption. In summary, the
275
-------
results show that the BiMelyze mercury immunoassay is a reliable alternative to more
expensive, time-consuming analytical methods.
INTRODUCTION
Heavy metal contamination in the environment is recognized as a serious danger to
humans and wildlife. Accordingly, the use of toxic heavy metals has become more strictly
regulated; but careless practices in the past have led to massive deposits of these toxins in
the environment. Metals are of special interest because they bioaccumulate up the food
chain and have long half-lives in biologic tissues. Mercury, one of the most toxic heavy
metals, causes severe behavioral, reproductive and developmental problems (1). Although
a significant portion of the mercury present in the environment is caused by the natural
degassing of the earth's crust (2), most mercury derives from anthropogenic sources,
including mining, smelting, chloralkali industries, electrical equipment, paint industries,
military applications, agriculture, and medicine (3). Mercury is also directly released to
the environment by the burning of fossil fuels and from municipal waste incinerators.
Analytical tools that can measure environmental contaminants in the field are central to the
ability to regulate, manage, and decontaminate sites. Conventional analytical methods,
such as atomic absorption and X-ray fluorescence, although very precise, can be used only
in a laboratory setting. Immunochemical techniques provide sensitive and specific
methods capable of measuring analytes of interest in complex biological matrices. The
medical laboratory community has long recognized these qualities in immunoassays, but
only in the last few years have they been adapted for use in detecting environmental
contaminants.
BioNebraska has developed an immunoassay for the detection of mercury in
environmental samples. The assay exists in two formats: a plate format which is
quantitative and best suited for laboratory analyses, and a tube format which can be used
for semi-quantitative measurements in the field. The immunoassay can specifically and
quantitatively detect mercuric ions in several different environmental matrices. Analyses
of laboratory and field samples using either format gives results that are in good
agreement with those obtained by more conventional analytical methods, such as cold-
vapor atomic absorption, neutron activation analysis, and X-ray fluorescence. The assay is
not affected by other metals at concentrations likely to be encountered in environmental
samples.
EXPERIMENTAL
Monoclonal antibodies.
The production and characterization of the mercury-specific monoclonal antibodies used
in these analyses were described previously (4,5).
276
-------
Extraction of samples for immunoassay
Before the BiMelyze® immunoassay can be used for environmental analysis. The mercury
is extracted from the sample using a kit available from BioNebraska. The procedure
requires digesting a 5-gram sample, representative of the area being tested, in a solution of
hydrochloric acid, nitric acid, and water (2:1:1) for ten minutes with intermittent, gentle
agitation. The acid for the extraction is provided by the end user or, alternatively, is
available from an independent supplier. After extraction, the sample is buffered, filtered,
and diluted by means of filter-tipped dropper bottles provided in the soil extraction kit,
then analyzed by the immunoassay. National Institute of Standards and Technology
(NIST)-traceable standards are included in the kit and are extracted at the same time as
the unknown samples to provide comparisons for semi-quantitative determination of
mercury in the field.
BiMelyze Mercury Assay
The assay has been developed in two formats: a quantitative 96-well, plate method for
analyzing large numbers of samples in the laboratory, and a semi-quantitative, field-
portable tube method. Both formats consist of sequential addition of four reagents, with a
five-minute incubation period for each. The BiMelyze immunoassay is based on the initial
binding of the mercuric ion from a sample to the mercaptan of glutathione that has been
covalently linked to bovine serum albumin and bound to a solid support (Figure 1). After
a water rinse, a mouse anti-mercury antibody is added which binds to the mercury. The
tubes/wells are washed with a detergent then rinsed with water to remove unbound
antibody. The amount of anti-mercury antibody bound is detected by binding horseradish
peroxidase-labeled rabbit secondary antibodies to the mouse antibodies. After washing as
above, substrate, which is oxidized by the peroxidase-labeled antibodies, to produce a
green chromogen is added. The amount of color that develops is a function of the amount
of mercury in the initial sample. Color development is terminated by addition of stop
solution, and the tubes are read anytime within an hour. The absorbance of the color in
the tubes/plate is measured at 405 nanometers. Microplate readers, present in most larger
laboratories, are used to read the color development of the plate assay. A field-portable,
battery-powered, differential photometer, available from BioNebraska, allows the
absorbance in the tube assay to be quantified.
Metal specificity
Standard assays were performed as above except that various concentrations of additional
metal salts were added to the samples prior to assay, and their effect on color development
in the assay was monitored.
RESULTS
The lower limit of quantitation of mercuric ions with the BiMelyze immunoassay in an
aqueous sample is 0.25 parts per billion (ppb) for both the tube and plate assays. Figure 2
277
-------
shows a dose-response curve in which Mercury Standard Reference Material 3133 from
the National Institutes of Standards and Technology was diluted to various concentrations
in 0.1 M HEPES buffer, pH 7.0, and analyzed with the tube assay. The results
demonstrate that the absorbance in the ELISA is linear and reproducible over the range
from 0.25 to 25 ppb with all coefficients of variation below 8%. Similar results have been
reported previously for the mercury-specific plate assay (5).
Since the soil samples must be acid extracted and neutralized before assay, it was
necessary to know the pH dependence of mercury binding to the ligand-coated tubes.
This was tested by using a three-point standard curve with mercury concentrations of 0,
and 1 ppb in buffers at pH 2, 4.75, 6, 7, 8, 8.75, 10, and 11.8 (Figure 3). Samples of 0.5
ml were assayed according to the tube protocol. The assay is essentially unaffected over
the pH range 4.75 to 8.75. Consequently, the pH of the samples is adjusted to between 6-
8 for routine analysis.
Because many metals are ubiquitous in the environment, their effect on the reliability of
the mercury-specific assay was characterized in detail using several approaches. First, a
standard curve was constructed in which known concentrations of mercury were diluted in
a multi-metal mixture and measured in the immunoassay. The composition of this mixture
corresponded to that of the EPA extract metals quality control sample formerly available
from the Environmental Protection Agency. It contained 100 mg/L barium nitrate, 1 mg/L
cadmium sulfate, 5 mg/L lead nitrate, 5 mg/L silver nitrate, and 5 mg/L chromium
trioxide. Solutions of mercury at concentrations of 20, 2, 0.5 and 0.2 ppb were prepared
in the metal mixture and used in the immunoassay. The results obtained with these
samples were compared to a standard with mercury at the same concentrations in a buffer
at pH 7.0 which did not contain the other metals. The standard curves obtained with
mercury in these two diluents (Figure 4) are essentially identical, indicating that none of
these metals has an effect on the mercury immunoassay.
The potential interference by individual metals was examined over a wide concentration
range, to a level higher than would normally be present in field samples (Figure 5).
Standard curves were constructed in which mercury at 100, 50, 10, 5 and 0.5 ppb was
diluted into solutions containing the indicated concentrations of these metals. According
to the experimental design, for each metal there were five separate mercury-specific
standard curves, each containing 1 mM, 10 uM, 100 nM, 10 nM and 1 nM of a potentially
interfering metal. A control standard curve was also included in which the mercury was
diluted to the same concentrations in an equal volume of metal-free buffer. The metals
examined were: arsenic trioxide, barium nitrate, cadmium chloride, chromium nitrate,
cupric chloride, gold trichloride, iron sulfate, lead chloride, nickel chloride, silver nitrate,
sodium bicarbonate, sodium chloride, strontium nitrate, thallium nitrate and zinc chloride.
Barium nitrate (Figure 5a), which gave results typical of most metals, shows no
interference even at the highest concentrations employed. Only three of the metals tested
affected mercury detection, but they did so only at high concentrations. Gold trichloride
inhibited the response at the two highest concentrations (Figure 5b). The highest
278
-------
concentration of gold trichloride caused a purple precipitate when added to the tube in the
first step of the assay (data not shown). Silver nitrate produced an increase in absorbance
at the two highest concentrations (Figure 5c), which might be related to silver salt
precipitation. An increase in signal was also seen with 1 mM chromium nitrate (data not
shown).
Finally, the mercury content of several soil samples containing certified amounts of various
metals was measured. The descriptions of these samples and their metal compositions are
shown in Table 1. Triplicate samples were extracted according the BiMelyze protocol,
and analyzed with the BiMelyze Mercury Tube Assay. By comparison with soil standards
containing either 4 ppm or 15 ppm, the results were interpreted as less than 4 ppm,
between 4 and 15 ppm, and greater than 15 ppm. As shown in Table 2, the immunoassay
correctly predicted the mercury concentrations of these samples in almost all cases. The
only incorrect determination was in experiment #3 with soil sample #4.
The reproducibility of the soil assay was examined with five-gram aliquots of reference
soils containing 0, 1, 2, 3.2, 4, 4.8, 6, and 8 ppm mercury, as determined by cold-vapor
AA. Seven replicate analyses were done of each. The results (Table 3) are presented as
the differential absorbance of each mercury concentration relative to a 4-ppm standard.
The data demonstrate the ability of the test to distinguish between small differences in
mercury concentration in soils. They also show the reliability of the assay, since only one
false-negative and one false-positive were obtained in these analyses.
To demonstrate the usefulness of the assay under field conditions, the BiMelyze Mercury
Tube Assay was used to analyze ten environmental samples, whose mercury
concentrations were also measured by both neutron activation analysis (NAA) and X-ray
fluorescence (XRF). The general format of the study was to compare the unknown soil
samples to the 5 and 15 ppm mercury-in-sediment standards. The data for the tube assay
were interpreted as <5, 5-15, or >15 ppm by comparison of the absorbance of each sample
to that of the standards. The tube assay gave excellent agreement with the reference
methods, differing from neutron activation analysis in only two samples (#3 and #5).
However, X-ray fluorescence analysis of sample #5 agreed with the immunoassay rather
than with NAA. Sample #3 was not analyzed by XRF.
Another field study was conducted by an environmental testing company who collected
samples and analyzed them with the BiMelyze Mercury Tube Assay, and by cold-vapor
AA. The description of the samples, along with the results of the analysis (Table 5)
indicate a good agreement by both methods, showing a disparity with only one sample.
However, even with that sample, the difference was not large, since the AA value was 14
ppm and the BiMelyze results were >15 ppm. This independent analysis tested matrices
which have not been tested by BioNebraska (e.g., paint, cinderblock, and sludge), but
suggests the versatility of the method. The matrices for which the BiMelyze assay is
applicable appears to be limited only by the ability of the acid mixture to disassociate and
oxidize mercury to the mercuric form.
279
-------
DISCUSSION
The BiMelyze mercury tube immunoassay provides an accurate, reliable method for
detecting mercury in a variety of matrices. Under laboratory conditions the assay is
quantitative for mercury in aqueous solution (Figure 2). The tube assay was designed as a
field test that could be used for on-site evaluation of environmental samples. Under these
conditions, the assay is semi-quantitative. The mercury concentration is determined by
direct comparison to a standard with a known amount of mercury. When used with both
field samples and NIST-traceable soil standards, the tube assay results agreed with those
obtained by a reference method (either NAA or AA) in at least 20 of 22 various samples
(Tables 4 and 5). The importance of determining the mercury concentration by
comparison to a standard analyzed at the same time as the unknown samples must be
emphasized. The mercury assay is an enzyme-linked immunosorbent assay (ELISA)
whose interpretation depends on the absorbance obtained by enzymatic conversion of a
colorless substrate to a colored product (Figure 1). The absorbance obtained with
samples having the same mercury concentration can vary from day to day (Table 2), since
the enzyme activity of the assay is affected by ambient conditions. Analysis of a reference
standard at the same time as the unknown samples controls for this variability.
With environmental samples, acid digestion is needed to extract total mercury from
constituents of the matrix. This treatment oxidizes mercury to mercuric ions, for which
the antibody is specific. The extraction method used here consists of treatment of the
sample with a mixture of hydrochloric acid, nitric acid and water (2:1:1) for ten minutes.
This extraction method is as efficient as E.P.A. Method 7471 for extracting most forms of
mercury, except for methylmercury (data not shown). Excellent correlation has been
reported previously when the plate assay was used for analysis of environmental samples
and compared to other analytical methods (6). The reliability of the results obtained in the
previously reported study led to the inclusion of the BiMelyze Mercury Plate Assay into
the Department of Energy Methods Compendium as Method MB 100. The results from
analyses of environmental samples with the BiMelyze Mercury Tube Assay, as reported
here, are equally reliable. The method can accurately measure mercury in a variety of
environmental samples. Independent organizations have tested cinder block and paint
chips in addition to various soils, and it is likely to be applicable to other matrices which
have not previously been assayed successfully (Table 5). We are currently working on an
efficient, rapid, and user friendly extraction technique for methylmercury, which would
have implications for measurement of mercury in biological samples, such as fish tissue.
In the past, accurate testing of environmental samples for mercury has been limited by the
availability of analytical methods, such as AA, that utilize expensive equipment requiring
highly trained personnel for proper operation. Another disadvantage of these procedures
is the lag time between sample collection and acquisition of the results, and problems
arising from the instability of the sample, since, in most cases, the samples must be sent off
to reference laboratories for analysis. In contrast, the BiMelyze mercury assay provides a
convenient, cost-effective, real-time method for monitoring and surveying environmental
sites for mercury that can be performed in the field by personnel with minimal training. Its
280
-------
use can thus reduce the number of samples that must be analyzed by more expensive,
traditional methods. The real-time data acquisition reduces potential re-mobilization costs
which can occur if initial remediation is insufficient. The method is ideal for measurement
of mercury in remote areas where sample storage, inventory and transportation present
logistical problems. The kit is stable for at least six months at 4°C and for shorter periods
of time at elevated temperatures. The only instrumentation required is a field-portable
spectrophotometer that is inexpensive (<$ 1,000) compared to the instrumentation needed
for traditional analytical methods. The method has a high selectivity for mercury and is
not affected by metals likely present in environmental samples (Figure 5).
With increased awareness on the part of both the general public and various governmental
regulatory agencies concerning toxic chemicals in the environment, the demand for
convenient, reliable methods for their detection will certainly increase. Although
environmental immunoassay technology is relatively new, it provides an excellent way for
local, state, and federal agencies to implement effective monitoring programs on
increasingly tight budget constraints. Immunoassays can be used in conjunction with
traditional analytical methods to allow a larger number of samples to be analyzed at a
lower total cost.
281
-------
LITERATURE CITED
1. Friberg, L., Nordberg G.F., & Voulk V.B. (1986) Handbook on the toxicology of
metals 2, 389-396.
2. World Health Organization Environmental Health Criteria 86, Mercury-
Environmental aspects, (1989) WHO, Geneva.
3. Fitzgerald, W.F. (1986) The world & I, "Mercury as a global pollutant". 192-199.
4. Wylie, D.E., Lu, D., Carlson, L.D., Carlson, R., Babacan, K.F., Shuster, S.M. &
Wagner, F.W. (1992) Proc. Natl. Acad. Sci. USA. 89, 4104-4108.
5. Wylie, D.E., Lu, D., Carlson, L.D., Carlson, R., Wagner, F.W. & Shuster, S.M.
(1991) Anal. Biochem.194, 381-387
6. Department of Energy Methods for Evaluating Environmental and Waste Management
Samples. (1993) Method MB100, "Immunoassay for mercury in soils", pp. 1-9.
282
-------
Table 1.- Metal compositions of certified reference soils.
Sample metal composition, ppm
METAL
Aluminum
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Molybdneum
Nickel
Potassium
Selenium
Silver
Sodium
Thallium
Vanadium
Zinc
1"
4090
<2
50.3
-------
Table 2. Analysis of certified reference soils using BiMelyze Mercury Tube Assay.
Soil Sample
1
2
3
4
5
6
7
8
[Mercury]
(ppm)
<0.10
1.40
1.47
2.36
4d
15d
50d
122
Exp. 1
0.12
1.01
0.78
1.54
1.76
1.99
2.04
2.55
Aios"
Exp. 2
0.05
0.64
0.41
0.84
1.01
1.45
1.73
2.55
Exp. 3
0.08
0.47
0.19
0.925C
0.83
1.59
2.02
2.55
Interpretation
b
<4
<4
<4
b
b
>15
>15
a Absorbance at 405 nanometers with the BiMelyze Differrential Photometer
b Standard reference point, no interpretation
0 Only value which gives incorrect conclusion
d NIST-SRM solid phase diluted from an initial concentration of 107 ppm.
Table 3. Reproducibility of BiMelyze Mercury Assay Tube Kit with extracted soil samples
Seven replicate extractions of 5 gram soil samples with a 4 ml mixture of 2:1:1
hydrochloric, nitric acid and water. The samples were then analyzed by both the tube
assay and cold vapor atomic absorption. CVAA data represents an average of the seven
analysis and the immunoassay data are presented as the difference relative to a 4 ppm
standard.
[Hg]
ppm
0.0
1.0
2.0
3.2
4 0
i . \J
4.8
6.0
8.0
[AA]
ppm
0.0
1.05+07
2.08±.12
3.27±14
A 1 fi+ 94
H. 1 UZL.^H
4.93+.31
6.11+32
7.97+36
TUBE ASSAY EXPERIMENTS
1
-1.38
-0.63
-0.25
-0.06
a
+0.26
+0.23
+0.23
2
-1.32
-0.62
-0.38
-0.26
+0.05
-0.10
+0.15
3
-1.02
-0.57
-0.40
-0.17
+0.03
+0.07
+0.30
4
-0.57
-0.34
-0.16
+0.28
+0.46
+0.66
+0.92
5
-1.05
-0.69
-0.47
-0.22
+0.16
+0.21
+0.25
6
-0.98
-0.51
-0.37
-0.05
+0.15
+0.22
+0.37
7
-1.03
-0.41
-0.08
-0.04
+0.02
+0.21
+0.19
4.0 ppm used as standard
284
-------
Table 4. Analysis of mercury in soils using the BiMelyze Mercury Assay Tube Kit.
Concentration bv
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
NAA
ppm
na
na
na
116
<3.3
11
<2.1
<5.3
<1.5
<7.7
87
19
121
XRF
ppm
na
na
na
90-116
na
na
na
22-52
na
na
150-159
na
118-122
Absorbance
at410nm
0.077
0.131
0.216
0.357
0.104
0.293
0.096
0.292
0.127
0.088
0.337
0.259
0.264
Interpretation
a
>15
0-5
>15
0-5
>15
0-5
0-5
>15
>15
>15
a Standard controls, no interpretation
285
-------
Table 5. Independent analysis of mercury in samples using the BiMelyze Mercury Assay
Tube Kit at an abandoned battery reclamation site.
SAMPLE DESCRIPTION
4 Cold Vapor Atomic Absorption
TEST KIT
CVAA"
Process Room
Dust from process room
Groundwater -unfiltered
Soil, alkaline
Sludge from tank
Sump sludge
Cinderblock
Cinderblock- duplicate
Soil
Paint
Backround Cinderblock
Backround paint
Debris from CC>2 blast
<5 ppm
< 5 ppm
< 0.5 ppb
< 5 ppm
> 15 ppm
5 >15 ppm
< 5 ppm
< 5 ppm
5 >15 ppm
> 15 ppm
< 5 ppm
> 15 ppm
> 15 ppm
0.83 ppm
> 4.5 ppm
< 0.4 ppb
0.93 ppm
4,400 ppm
14 ppm
3 ppm
14 ppm
34 ppm
1.4 ppm
14 ppm
19 ppm
286
-------
FIGURE LEGENDS
Figure 1. Basis of the BiMelyze Mercury Immunoassay "ELISA". The schematic shows
the major steps and the reagents used to perform the assay.
Figure 2. Dose response curve for mercury detection by the BiMelyze Mercury Tube
Assay. Mercuric nitrate was diluted to final concentrations of 0, 0.25, 1, 5, 10, and 25
ppb in pH 7.0 buffer. The mercury solutions were then analyzed by ELISA as described in
the Experimental section, with six replicates for each concentration. The results
demonstrate both the linearity and reproducibility of the assay for mercury concentrations
between 0.25 and 25 ppb when graphed as the ELISA absorbance versus the log of the
mercury concentration.
Figure 3. Effect of pH on mercury detection by ELISA. Solutions containing either 0 or
1 ppb mercuric nitrate in buffer adjusted to the indicated pH with either 1 N HC1 or 1 N
NaOH. The solutions were then used in the ELISA as described in the Experimental
section.
Figure 4. Detection of mercury in EPA Extract Metals Quality Control Sample. Mercuric
nitrate was diluted to final concentrations of 0, 0.2, 2, or 20 ppb in a solution containing
metals at the concentrations present in the EPA Extract Metals Quality Control
Sample(H), as described in the Experimental section. A standard curve was then obtained
by analysis of these solutions in the ELISA and compared to that obtained with mercury
diluted to the same concentrations in 0.1 M HEPES buffer, pH 7.0(0). A sample
containing metals at the same concentrations as in the EPA quality control sample but
without mercury was also included in the assay(A).
Figure 5. Effect of non-mercury metal salts on mercuric ion detection by ELISA.
Standard curves of mercury concentrations ranging from 0.5 to 100 ppb were assayed in
solutions containing potentially interfering metals at final concentrations of 1 nM(0), 10
nM(X), 100 nM(A), 10 |iM(B), and 1 mM(*). A control curve in which the mercury
was diluted into pH 7.0 buffer was also included(D) These solutions were then used in
the ELISA as described in the Experimental section. For each concentration of metal salt,
a control point containing the same concentration of metal salt but with no added mercury
was included and is represented by the bottom-left points. Figure 5a represents typical
results obtained with the listed metals. Figure 5b shows the inhibition by gold trichloride,
and Figure 5c shows the increase in signal with silver nitrate.
287
-------
to
CD
CO
ffi ffi ffi ffi
CO GO CO 00
9999
WELL
X
m
//\ A\
60 £?
E ^? PC ^?
CO CO CO CO
9999
+ SAMPLE
+HRP-SECOND
ANTIBODY
+HRP SUBSTRATE
CO CO CO CO
9999
t
+ ANTI-Hg
ANTIBODY
E * E «
CO CO CO CO
9999
Color
-------
ro
c»
CD
a
a
10
o
"
o
a
O
C/5
0.5
0
, L
0.1
0.3
I . .
1 3
[Mercury], ppm
10
30
-------
ro
CO
o
Control (no Hg)
-------
l\3
CD
0.8
I 0.6
0.4
•e
0.2 -
o
o.i
0.3
1 3
[Mercury], ppm
10
30
-------
1 3 10
[Mercury], ppm
30 100
292
-------
40
APPLICATION OF A NEW MERCURY SPECIATION TECHNIQUE TO SAMPLES
FROM SITES HEAVILY CONTAMINATED WITH MERCURY
Eric L. Miller, David E. Dobb, and Dolf Cardenas, Lockheed Environmental Systems & Technologies
Company, Las Vegas, Nevada 89119; and EdwardM. Heithmar and Ken W. Brown, U.S. Environmental
Protection Agency, EMSL-LV, Las Vegas, Nevada 89119.
ABSTRACT
Mercury contamination in the environment poses a serious health risk, especially when the contaminants
are in a toxic form (such as organomercury compounds) that can accumulate in organisms. Other species
(such as mercury sulfide) are geologically stable and are of less concern. Mixtures of these compounds
can occur naturally or anthropogenically at a given site. Therefore, an accurate method for speciating
between types of mercury compounds is an absolute necessity for obtaining proper risk assessment data.
EPA's Environmental Monitoring Systems Laboratory at Las Vegas is developing methods for speciating
mercury and other metals to enable the Agency to make better risk-based decisions. In the work
described here, a method has been developed and applied to soil and sediment samples from three sites
heavily contaminated with mercury. The samples were analyzed for total mercury and for five types of
mercury compounds.
Separation of the mercury types was based on sequentially leaching the soil samples with a series of
increasingly acidic and oxidizing extraction solutions. Specifically, the types of mercury compounds and
their associated extractants (as applied in sequential order) included: organic (soluble in toluene in the
presence of chloride, e.g., CH3Hg+); "water soluble" (soluble in 0.01M K2SO4 and 0.01M KC1, e.g.,
HgCl2 and soluble Hg2+ salts); dilute-acid soluble (soluble in 0.2M HNO3, e.g., HgO); concentrated-acid
soluble (soluble in 3.9M HNO3, e.g., free and amalgamated elemental mercury); oxidizing concentrated-
acid soluble (soluble in 3.9M HNO3, 0.7M HC1, e.g., HgS and Eg2Cl2). Concentrations of mercury in
total mercury extracts and specific mercury compound extracts were measured by inductively coupled
plasma-mass spectrometry (ICP-MS), with confirmatory determinations by anodic stripping voltammetry
(ASV), cold vapor atomic absorption spectrophotometry (CVAAS), and X-ray fluorescence (XRF).
Samples from the sites were found to contain total mercury concentrations ranging from < 1 mg/kg to
3000 mg/kg. At one site where samples were taken from two depths, deeper samples contained higher
total mercury concentrations than shallow samples. In shallow samples, about 80 to 90 percent of the
mercury was extracted as elemental mercury. In deeper samples, mercury was present predominantly
as a mixture of elemental mercury and mercury sulfide in approximately equal concentrations. Minor
amounts of dilute-acid soluble mercury (HgO) were present in most of the samples. Water soluble
mercury compounds were typically < 1 mg/kg, but detectable. Organic mercury compounds were
typically undetectable at < 0.005 mg/kg. The new speciation procedure offers significant improvements
in accuracy and throughput over previous methods, enabling analysis of many samples at low cost.
NOTICE: Although the research described in this presentation has been funded wholly or in part by the
U.S. Environmental Protection Agency through Contract 68-CO-0049 to Lockheed ESAT, it has not been
subjected to the Agency's review. Therefore, it does not necessarily reflect the views of the Agency.
Mention of any trade names or commercial products does not constitute endorsement or recommendation
for use.
293
-------
41
"Almost Digestions" or
Hot-Acid Leaches with Continuous Flow Microwave Sample Preparation
by
Edward E. King, David A. Barclay, J. Douglas Ferguson, Lois B. Jassie,
CEM Corporation, 3100 Smith Farm Road, Matthews, NC 28105.
Abstract
An important advance in sample preparation for atomic spectroscopy is
the use of flow injection coupled with microwave digestion to produce a
continuous flow microwave sample preparation system. This approach
automates the most time consuming step in the analytical process: sample
preparation. The SpectroPrep™ system embodies this hybrid technique and has
been applied to the preparation of environmental samples prior to metals
analysis.
Introduction
In environmental analyses for determination of metals, total digestion of
the matrix is not practical nor always needed. Partial digests, or leaches, have the
ability to extract analytes of interest, however, reproducibility of these extractions
is not well documented. Because reaction conditions in the SpectroPrep™ system
are reproducible through consistent temperatures, the analyst has the ability to
tailor the leaching procedure to the matrix and to the elements of interest. The
broad applicability of the SpectroPrep™ system for achieving needed consistency
is demonstrated in the preparation and analysis of soils, sludges, and other solids
at slurry concentrations to 1% in a variety of acid media.
The objectives of this study were to
• Demonstrate the comparability of hot-acid leaching continuous flow
microwave sample preparation to EPA SW-846 Method 3051 (batch) for
the analysis of metals in environmental samples
• Assess the reproducibility of hot nitric acid leaching for the
preparation of soils, sediments, sludges, and oily wastes, and
• Evaluate the distribution of analytical results for extractable metals in
a set of solid environmental samples over their dynamic range
These objectives have been accomplished by implementing the following
strategy:
• Use of Standard Reference Materials (SRMs) to assess the accuracy
and precision of both methods
• Use of mixtures of known materials (SRMs) to establish the dynamic
range of the techniques
294
-------
• Real life testing with subset of real world and certified reference
material samples
Experimental Protocol
Materials
Materials selected for this study were drawn from a variety of sources to
simulate a realistic environmental laboratory's sample distribution. Such solids
might constitute the bulk material in a hazardous waste sample. In addition to
wet soil, fuel-soaked soil, clay, and dewatered sludge, five reference materials
representing three types of matrices were included. Two sediments- SRM 2704,
BCSS-1; two soils-SRM 2710 and SRM 2711; and one rock, CANMET MRG-1 were
analyzed. Average particle size was reduced to < 250 microns to allow passage
through the tubing and filters in the continuous flow system.
All samples were analyzed routinely for twenty elements; Al, Ag, As, Ba,
Be, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, P, Pb, V, and Zn.
Leach reagent (acid) to sample ratio for batch Method 3051 is 10 mL of
concentrated nitric acid (HNO3) for 0.5 g of solid. Continuous flow sample to
reagent ratio is 0.5 g of solid in 50 mL of 20 % nitric acid. This represents a slurry
concentration of 1%. A 10 ppm yttrium standard is incorporated into the acid
reagent to track the continuous flow dilution factor.
Methodology
Method detection limits were established for the SpectroPrep™device in
20% acid solution to match the final concentration of the analyzed solutions
from both the continuous flow and from the diluted batch digested samples. Six
samples was the normal complement for Method 3051. Four replicates of every
material were prepared for analysis and method blanks were included in the
batches to round out a complement. On average, one blank was analyzed for
every 4 samples. For the continuous flow, 2 replicates of each material were
weighed out and each replicate was sampled twice. One blank was analyzed for
every 10 samples run through the system.
Temperature and pressure conditions for the two methods were matched
as closely as posssible. Method 3051 stipulates the following performance criteria:
1) reach 175 °C in 5.5 minutes, and 2) remain between 170-180 °C for the balance
of ten minutes. Under these conditions, the pressure varies from 140 psi in a
completely inorganic material like the CANMET rock, to 175 psi for a fuel-soaked
soil which contains ~ 8% organic material. In the continuous flow system the
estimated temperature of ~ 200 °C is achieved by maintaining a backpressure of
175 psi at the last pump. For the normal 10 mL sample loop the average
residence time of this sample stream is 2 minutes and 15 seconds.
Metal concentrations were determined by inductively coupled atomic
emission spectroscopy (ICP-AES) on an ARL model 3560 sequential spectrometer
equipped with a mini-torch and Meinhard nebulizer. In continuous flow, 8-9 mL
are normally collected from the 10 mL sample loop which represents a 1.3
dilution. All of the batch samples are diluted to 50 mL so that the acid
concentration presented to the ICP instrument is 20%.
295
-------
Assessment of the dynamic range was achieved by carefully mixing two of
the SRMs containing low and high concentrations of the metals of interest
according to the proportions in Table I. This resulted in samples whose Mn
concentration, ranged from ~ 20-8000 ppm with similar ranges for Zn, Cu and Pb.
Only the managanese data will be presented here.
Table I. Mixture Composition for Dynamic Range Study
Composition, %
Material
Low
Mix I
Mix II
Mix III
High
SRM 2710
0
25
50
75
100
SRM 2711
100
75
50
25
0
Results and Discussion
Comparability of leaching by the two microwave methods is shown by the
generally good agreement between the elemental values obtained by continuous
flow and the batch reactor as seen in Table H
Table II. Elemental Concentrations in Mixture II of Montana Soil SRM 2710 and
SRM 2711 (1:1) Prepared by Microwave Leaching with Nitric Acid
concentration, ppm
continuous flow Method 3051
Element
Al
As
Ba
Ca
Cd
Co
Cr
Cu
Fe
K
Mg
Mn
Na
Ni
P
Pb
V
Zn
mean
23500
373
292
12980
40.1
10.4
24.2
1478
23272
6562
6913
4426
861
20.3
866
3207
54.6
3276
%RSD
0.51
0.72
0.65
0.26
0.37
1.21
1.29
0.47
1.00
0.60
0.47
0.23
0.64
1.15
0.45
0.38
0.37
0.85
mean
17772
337
258
12615
36.3
8.57
21.9
1420
23581
4876
5997
4153
398
19.8
787
3081
40.4
3303
%RSD
5.09
0.73
1.11
0.54
0.25
8.02
3.66
0.49
1.66
3.02
1.76
1.04
4.98
2.01
1.89
0.42
4.38
1.88
Exp Range
12000-24500
289-355
235-330
11900-14900
22.6-36
6.65-12
15-24
1245-1755
19500-29000
3200-5200
5800-7500
3300-4810
345-455
11.4-17.5
830-1000
2615-4250
35.5-50
2745-3620
296
-------
For many of the elements, the continuous flow data are ~ 3% high on average.
For vanadium at 55 and 44 ppm and cobalt at 10.4 vs 8.6 ppm respectively, the
bias is ~ 20% for SpectroPrep™ values. Individual manganese values obtained
for the reference materials that make up the mixture are in good agreement
however, with leach values published in a certificate addendum issued by NIST,
as Table EH demonstrates.
Table III. Manganese Leached by Nitric Acid from SoilsReference Materials
Prepared by Microwave Sample Preparation
material
concentration, M-g/g
Method3051 continuous flow NIST range
2710
2711
7570
496
7895
501
6600-8900
470-570
Despite this bias in favor of the SpectroPrep™ results, inspection of Figure 1
demonstrates that the dynamic range for managanese is linear over nearly four
orders of magnitude.
Q.
Q.
0>
u
C
O
u
**
O
_J
u.
CO
3
O
O
O
10000
8000 -
6000 -
4000 -
2000 •
Manganese Leached from Environmental Solids
DYNAMIC RANGE STUDY
y = 1.0317 x - 1.4525 R2=0.998
+ dynamic range data
o all other manganese values
• unity line
2000 4000 6000 8000
EPA METHOD 3051, concentration (ppm)
297
-------
Q.
C
0
32
2
+*
c
o
o
c
o
o
•*
o
O
D
Z
P
Z
O
o
EXPLODED VIEW OF LOWER DYNAMIC RANGE
MANGANESE LEACHED FROM ENVIRONMENTAL SOLIDS
A Comparison of Continuous Flow vs EPA SW-846 Method 3051
600
500
400
300
200
100
y = -7.6454 + 0.94306x RA2 = 0.940
wet soil
EPA Sludge
BCSS-1
X
oily soil
SRM2704
o CanMet
o all other manganese values
• Mn conc/SP
100 200 300 400 500
METHOD 3051, concentration (ppm)
600
Figure 2 demonstrates that the linearity is as good between 25 and 500 ppm as it is
between 500 to 8000 ppm.
For all samples prepared by continuous flow the average relative standard
deviation (RSD) is 4.43 % compared to 5.15 % RSD for EPA Method 3051. The
RSD ranges from 0.82% to 10.37 % for SpectroPrep™ where the smallest value is
for MIX n and the highest was for the CANMET material. For the batch method,
the lowest value is for Montana Soil SRM2710 and the highest imprecision is for
the oily soil matrix. For all elements the %RSD for continuous flow is 4.37 %
compared to 5.26 % RSD for EPA Method 3051. %RSD ranges were 2.19-8.61% for
the continuous flow and 2.50-8.71% for the batch method.
Conclusions
Both continuous flow and batch microwave hot acid leaching are comparable
sample preparation methods for the determination of metals in soils, sediments,
sludges, and oily wastes. Data using Standard Reference Material mixtures whose
concentrations span 4 orders of magnitude show the validity of the continuous
flow procedure. Average < 5% RSDs for acid leaching procedures indicate that
the goal of reproducibility has indeed been achieved.
298
-------
42
An Inter-Laboratory Comparison of Instruments Used for the
Analysis of Elements in Acid Digestates of Solids
David Eugene Kimbrough, Public Health Chemist II, Janice
Wakakuwa, Supervising Chemist, California Environmental
Protection Agency, Department of Toxic Substances Control,
Southern California Laboratory, 1449 W. Temple Street, Los
Angeles, California, 90026-5698
Abstract
This paper presents data from an inter-laboratory study
of 160 accredited hazardous materials laboratories comparing
the accuracy and precision of four different analytical
instruments, inductively coupled plasma - atomic emission
spectroscopy, inductively coupled plasma - mass spectroscopy,
flame atomic absorbance spectroscopy, graphite furnace atomic
absorbance spectroscopy, and hydride generation atomic
absorbance spectroscopy. Each laboratory performed a mineral
acid digestion on five soils spiked with different
concentrations of arsenic, cadmium, molybdenum, selenium, and
thallium. The resulting digestates were analyzed on one of
the above instruments. Results show that at most
concentrations that ICP-AES has significantly better
precision and accuracy than the other techniques but had the
highest rates of false positives and negatives. HGAAS and
GFAAS consistently showed the poorest precision and accuracy.
FAAS better precision and accuracy than either HGAAS or GFAAS
but performed poorer than ICP-AES. FAAS however had fewer
false positives and negatives than ICP-AES.
299
-------
43
AN EVALUATION OF INTERELEMENT CORRECTION FACTORS: USES AND LIMITATIONS
Guy A. Laing, L.J. Ottmar, D.E. Dobb, F.C. Garner, J.T. Rowan, Lockheed
Environmental Systems & Technologies Company, Las Vegas, NV 89119, and
L.C. Butler, U.S. Environmental Protection Agency, Las Vegas, NV 89119.
Abstract
The Environmental Monitoring Systems Laboratory in Las Vegas, under the
Office of Research and Development, is continually evaluating and
improving quality assurance parameters so as to obtain more reliable data
from EPA methods. One quality assurance parameter associated with the
data received from an inorganic Statement of Work (SOW)is the interelement
correction factors (lECs) used by Inductively Coupled Plasma - Atomic
Emission Spectrometry (ICP-AES). lECs are a procedural and contractual
requirement for the use of ICP-AES in the Contract Laboratory Program
(CLP), and are used to remove the spectral interferences that are present
in the ICP-AES method. A similar type of correction is also performed for
Inductively Coupled Plasma - Mass Spectroscopy (ICP-MS), which is being
considered for use in future SOWs. In both methods the analytical data
are subject to corrections to the analytical signal based upon responses
from interfering constituents. These correction factors are based upon
measured or known relationships between non-analyte peak intensities and
the corresponding magnitude of interference on an analyte peak. This
study evaluates correction factors in both ICP-AES and ICP-MS to determine
their uses and limitations. Real data are examined in order to illustrate
the impact of error in interelemental corrections as well as to suggest
methods for detecting inappropriate corrections. Limitations on the use
of corrections are considered and alternative methods for applying
corrections are proposed. A detailed statistical evaluation of the errors
is presented, along with an evaluation of the limitations of the
interferences based upon the amount of interferant present. Each of the
sources of error are examined and statistical models were evaluated which
manipulate the relevant factors and evaluate their relative importance as
sources of error.
INTRODUCTION
In the CLP's Inorganic SOW(l) one of the areas that would benefit from
increased quality control is the determination and application of lECs.
The current SOW stipulates only that lECs must be determined for spectral
interferences due to aluminum, calcium, iron, and magnesium at all
wavelengths used for each analyte reported by the method. Other lECs must
be reported only if they are applied. The SOW makes no requirements as to
how the lECs should be determined, what the basis for their determination
is, and does not stipulate a standardization procedure for the factors
themselves. The only other reference to interferences states that the
laboratory must assume responsibility for verifying the absence of
NOTICE
Although the research described in this article has been funded wholly
by the U.S. Environmental Protection Agency through Contract 68-CO-0049
to Lockheed Environmental Systems & Technologies Company, it has not
been subjected to Agency review. Therefore, it does not necessarily
reflect the views of the Agency.
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
300
-------
spectral interferences from interfering elements that may be present in a
sample but for which there is no channel in the instrument array. Because
no requirement for evaluation of those spectral interferences exists in
the SOW, they are generally ignored, as evidenced historically by a lack
of statements about spectral interferences in case narratives supplied to
the EPA.
A related quality assurance parameter that is designed to test the lECs is
the Interference Check Solution (ICS). Although there is not a direct
contractual link between the ICS and the lECs, the QC requirements
associated with the ICS require that the lECs have values that will allow
the ICS QC criteria to be met. The results of this study demonstrate that
the ability to pass the ICS criteria does not evaluate the lECs
sufficiently.
lECs can be error prone requiring a greater measure of quality assurance
criteria than is currently felt necessary. In fact, the lECs incorporate
error from each measurement included as part of the final analytical
determination, and the total error depends on the number of measurements,
the amount of interferant, the accuracy and magnitude of the correction
factor(s), and the precision of the individual measurements.
Although ICP-AES and ICP-MS are similar in the use of lECs, ICP-MS does
offer some advantages over the ICP-AES primarily because the number of
interferences is greatly reduced in ICP-MS(2). Additionally, the lECs
themselves are generally based upon different properties than in ICP-AES.
In ICP-AES the lECs are a result of spectral interferences which are
affected by spectrometer settings and plasma conditions, as well as by
relative concentrations between the analyte and the interferant(3).
Therefore, the exact relationship between the analyte and interferant can
vary from day to day. In ICP-MS on the other hand, the lECs (called
elemental expressions) are primarily determined by the naturally occurring
abundances of the isotopes of the elements(4). Because the abundances of
the isotopes are independent of the plasma conditions, are constant, and
are well defined, they can be applied independently of the exact
instrument make or model used. These isotope ratios are also applicable
even if the interference is due to a molecular interference. In order to
determine if a molecular correction is required, it is only necessary to
examine for the presence of the elemental constituents.
EXPERIMENTAL
A Fisons ARL Model 3560 ICP-AES was used to evaluate the lECs for ICP-AES.
The linearity of the lECs was tested by running a single element solution
of iron at six different concentrations a total of three different times.
The curves of apparent analyte concentration were plotted against the iron
concentration for a number of different analytes. The plots were examined
for linearity and noise associated with each individual measurement.
The same ICP was used to evaluate the presence of interferants which are
not typically examined by laboratories performing work for the CLP. These
interferences were assessed by evaluating single element solutions of
cerium, copper, manganese, titanium, and zirconium. A spectral scan was
obtained for each solution around the silver 328.070 nm wavelength and the
apparent silver concentration was evaluated. Additionally, each solution
was screened for the presence of silver using a Perkin-Elmer 5000 ICP-MS
to validate the absence of silver in the single-element solutions.
301
-------
Iron Interference on Cobalt
80
60
40
0
§ 20
u
o
"S
§,-40
-60
-80
X Measured Cobalt
Regression Line
0
500 1000
Iron Cone, (ppm)
1500
2000
Figure 1
RESULTS AND DISCUSSION
To evaluate the use of lECs, the linear relationship between the analyte
peak intensity and the magnitude of the interference was tested. The
assumption that a linear relationship exists may often be false, as
demonstrated by Figure 1 which represents the cobalt interference from
iron. It is clearly seen that the cobalt interference does not occur at
all until sometime after the iron concentration exceeds 1,500 ppm in
solution and peak broadening causes it to be detected. Therefore it would
clearly be inappropriate to use the IEC which would be determined when the
iron concentration is less than 1,500 ppm for iron concentrations greater
than 1,500 ppm. Likewise it would also be incorrect to use an IEC derived
from the iron solution at 2,000 ppm for iron concentrations less than
1,500 ppm. To evaluate the relevance of interferences of this type,
additional information needs to be available to assess the frequency with
which samples are subject to interferences of this type. The levels at
which the interferences are tested should then be evaluated. Table 1
contains results of the CLP Analysis Results Database (CARD) which show
the concentrations at the listed rankings typically found in various
matrices.
302
-------
Table 1. Statistical ranking of iron concentration in the CLP by matrix
(Concentration is reported in mg/L of the solution after digestion).
matrix
ground water
other water
soil /sediment
other soil
25th Pet.
0.1
.034
33.9
16.5
50th
Pet.
1.21
.086
72
39.3
75th
Pet.
13.2
.675
126.5
61.5
95th Pet.
105
53.6
309.5
98.5
Maximum
35,500
2,050
17,200
1,160
Tables like these may be useful, once the interferant curve has been
determined, in evaluating the frequency that errors in quantitation occur
due to the presence of interferants. This information will be essential
for those who make policy to determine the extent of potential problems,
which will be relevant as quality assurance procedures are evaluated.
Another situation in which the interferant level can be demonstrated to
have a deleterious effect on data quality, is when the interferant exceeds
the capacity of the detector. The resulting correction would be
inappropriate in two scenarios. First, if the instrumentation does not
notify the operator of detector saturation, then the saturated signal
would be applied to the IEC and the result would be biased high because
the interferant was not properly quantified. Modern instrument software
systems generally notify the operator of this situation, but as recently
as 1989 this interference was observed in the MuItilaboratory Evaluation
of Method 6020 CLP-M(2). The second scenario occurs when the operator
identifies the presence of a large interferant and performs a dilution to
correctly quantitate the interferant but does not apply the correction to
the original undiluted sample. This procedure would have to be performed
manually, and should be identified in the case narrative. This has rarely
been identified in case narratives and therefore the extent of this
problem cannot be determined.
Another error that may occur in the use of lECs is the failure to make a
correction when one is needed. This may take two forms. The first is
demonstrated in ICP-AES by the interference of manganese on silver. This
interference was observed in data presented in the Multilaboratory
Evaluation of Method 6020 CLP-M Inductively Coupled Plasma - Mass
Spectrometry(2) when the ICP-AES instruments found silver levels present
at levels in one sample which were 4 times greater than the CRDL and 8
times greater than the typical IDL. The failure to perform an
interelement correction was due to an unidentified spectral interference.
Not only can this happen when a channel for the interferant is present in
the spectrometer, but could also happen when a channel for the interferant
is not present in the spectrometer. Table 2 demonstrates this by
evaluating the interference on silver by a variety of interferants. The
absence of silver contamination in the single-element solutions were
verified by the use of ICP-MS.
Table 2. Interference effects on silver.
Interferant
Cerium
Copper
Manganese
Titanium
Zirconium
Interferant
Concentration (ppm)
100
100
10,000
100
75
Apparent Silver
Concentration (ppb)
-59.68
-2.83
1,316
800
196
303
-------
Several observations from the data in Table 2 can be made. First is that
the correction for silver from titanium is quite high. This demonstrates
when an interference is typically not determined even though the
capability to perform the correction is usually available. Secondly, the
interference from zirconium is also high but because an optical channel is
generally not available, the interference cannot be assessed. Third is
the negative interference from cerium, which means that the spectral
interference is actually at the background correction point, not at the
analytical wavelength. Fourth, the interference from copper appears to be
negligible even though there are two copper emission lines in the vicinity
of the analytical line. One line at 327.40 nm is below the silver analyte
line at 328.07 nm, and the other is at 328.27 nm, just above the silver
line. This raises the possibility that the background correction point
circumstantially corrected for the copper interference and if the copper
concentration had varied, the IEC of zero would not have worked. Finally,
the data in Table 2 identifies a manganese correction that would need to
be applied to the silver data once the manganese level was correctly
measured. Most laboratories do not correct for this interference when the
amount of the analyte in solution greatly exceeds the linear range for the
analyte.
Another example of when an interelement correction is not performed is
when the interference is difficult to measure. The chromium interference
on antimony using ICP-AES has long been known, but many laboratories still
fail to correct for it. This has consistently resulted in false-positive
reports for antimony in samples with high levels of chromium. Table 3
presents the data from one of the samples evaluated in the Multilaboratory
Evaluation of Method 6020 CLP-M(2). The data demonstrate that even though
each of these laboratories used identical wavelengths to perform the
analysis, not only is the interelement correction performed inconsistently
between laboratories, but the magnitude of the correction varies widely.
It can also be seen from the data that the results varied widely and were
not dependant upon the application or non-application of an interelement
correction.
Table 3 - Antimony Results Comparing ICP-MS and ICP-AES
ICP-MS
Laboratory
A
B
C
E
F
G
H
J
Mean
Std. Dev.
Percent
RSD
Sb Result
50.6
45
51.2
46.11
41.8
53.6
54.2
53.5
51.48
4.84
9.4
Duplicate
Sb Result
54.4
52.6
48.7
47.76
50.1
58.6
55.7
59.8
ICP-AES
Laboratory
M
0
P
Q
R
S
U
V
X
Sb result
245
81.3
3150
1100
1190
4580
864
926
1920
1553
1473
94.8
Duplicate
Sb Result
252
93.8
3140
1020
1000
4980
623
979
1810
Cr IEC
value
.00515
none
none
.006717
none
.255
none
none
none
Note - Results were excluded if they were either less than the IDL, or the
laboratory significantly missed the calibration QC requirements.
Another issue not addressed by the quality assurance for lECs, is the
304
-------
assumption that the solutions used to determine the lECs are free from
contamination. In the course of this investigation it was noted that
several laboratories were using interelement correction factors which
could not be substantiated by any spectral interferences listed in the
reference wavelength tables. One of the most striking examples was a
manganese interference on nickel. In this particular case it is difficult
to chemically separate the nickel from the manganese and most manganese
will contain some nickel as an impurity. This would lead to the use of a
correction factor because of a nickel impurity in the single element
manganese solution, not because of a spectral interference. If the
solution being evaluated to determine the IEC for an analyte is
contaminated, then the use of the solution would lead to the use of an IEC
when one is not warranted. The ICP-AES method is particularly prone to
this error because generally no additional checks on the validity (i.e.
spectral verification of contamination) of the correction factor are
performed. Currently the SOW does not specify that any additional checks
must be performed, therefore the extent of the problem can not be
determined at this time. In ICP-MS this problem is not as widespread due
to the less empirical nature of elemental expressions.
30
20
Iron Interference on Copper
ex
ex
8
ID
ex
ex
o
a
-------
Figure 2 demonstrates the fallacy of making that assumption. It can be
clearly seen that when the iron concentration is 100 and 1,500 ppm, the
apparent copper concentration is less than the apparent copper
concentration when the iron concentration is 1,000 ppm. This may be due
to the iron interfering primarily with the background correction point
(the background correction would be abnormally high which, when
subtracted, would result in a negative contribution on the part of iron to
copper). At a iron concentration higher than 1,500 ppm the wing of an
iron line interferes with the analytical wavelength giving an apparently
high copper value. Figure 2 also demonstrates the difference that two
different determinations would make on the IEC. In fact simply using the
three data points from the 100 ppm iron concentration, the slope varies
from 0.00559 to 0.01956, representing differences of a factor of three
between the determinations.
Figure 2 also shows that a single point evaluation of the interference
would have a negative intercept. Without a thorough evaluation of the
intercept of the regression line, the copper interference would have a
negative slope, which would clearly be incorrect.
Finally, even when the relationship between an interferant and the
interference has been accurately defined, then errors are still
contributed by the imprecision associated with the measurements of each
associated peak. This error is dependant upon the size of the correction
terms and the magnitude of the interferant.
Because the lECs in ICP-AES and ICP-MS are mathematically similar, an
evaluation of how the factors used for lECs could be applied independently
of the method used was studied. A statistical model was developed to
evaluate the effects of the magnitude of the correction terms in lECs on
the analyte signal. The model consists of the following.
If
X = the concentration of the target analyte
Yj = the concentration of the ith interferant
and
Z = is the observed response at the primary mass or wavelength of the
target analyte
where
a = the calibration slope
ai = the correction factor, and
b = the coefficient of variation (%RSD) in intensity measurement
k = the number of potential interferants
Y is the estimate of Yj
then
Z = aX + a-L^ + a2Y2 + . . . akYk + e .
We have estimates of Yir a, and ai
The estimate of X is
Z - Va,Y,
X = ^ * J
a
and the variance of the estimate is
306
-------
V (X) =
If the relative standard deviation of the response measurement is
approximately constant for all responses,
i.e. RSD = 100 * b
then
V(X) _
This assumes that Z and Yi are all independent measurements.
Once these factors are determined then the errors can be estimated by the
above equation. As the equation demonstrates, the magnitude of the
interferant peak contributes error to the analytical peak.
CONCLUSIONS
Based on the observations cited above, a need exists for the definition of
additional quality assurance requirements on the interelement corrections
factors used for ICP-AES. It is apparent that the traditional reliance
upon the ICS criteria to fully evaluate the lECs does not evaluate the
interferences that have been mentioned. Requirements for lECs should also
be used for ICP-MS when that technique becomes commonly used in the CLP.
The factors which should be defined consists of criteria which evaluate
the ruggedness of the IEC terms being used. These include, an evaluation
of the solutions used to determine the lECs to prevent assigning lECs that
are simply due to contamination, a precedure to obtain better estimates of
the IEC factors by running replicates of the IEC determinations, a
procedure to standardize the factors so that comparisons can be made
between instruments, and means by which a more thorough evaluation of the
spectral characteristics and limitations of the lECs is performed.
The use of an error analysis statistic may also be useful either in data
review or data validation. The statistic presented here appears to
describe the analytical system quite well for ICP-AES and ICP-MS, but
before it receives widespread use, its characteristics should be more
thoroughly examined.
REFERENCES
(1) U.S. EPA, USEPA Contract Laboratory Program Statement of Work for
Inorganics Analysis Multi-Media Multi-Concentration, Document Number
ILM03.0, 1993.
(2) Laing, G.A., Stapanian, M.A., Aleckson, K.A., Dobb, D.E., Rowan,
J.T., and Garner, F.C., Final Report of the Multi-Laboratory
Evaluation of Method 6020 CLP-M Inductively Coupled Plasma - Mass
Spectrometry, An internal report submitted to USEPA EMSL-LV and AOB,
1989.
(3) Winge, R.K., Fassel, V.A., Peterson, V.J., and Floyd, M.A.,
Inductively Coupled Plasma - Atomic Emission Spectroscopy - An Atlas
of Spectral Information, Elsevier Science Publishing, New York, New
York, 1985.
(4) Houk, R.S., Analytical Chemistry, 58, 97A, 1986.
307
-------
44
IMPROVED DETECTION LIMITS WITH AXIAL PLASMA
Vincent J. Luciano, Thermo Jarrell Ash Corporation, 8E Forge Pkwy.,
Franklin, MA 0203 8
ABSTRACT:
The last year has seen a switch away from the conventional side on viewing of the ICP to
axial viewing as a direct result of improvements in detection limits for all elements capable
of ICP determination. With the proper optical configuration, detection limits
improvements of up to 20x can be achieved with conventional pneumatic nebulization.
Alternative nebulization devices (ultrasonic, thermospray, electrospray) reduce these limits
even further. The axial design has now been incorporated into both simultaneous and
sequential photomultiplier instruments and into solid state detector systems. Data will be
presented in several matrices for each instrument type and a summary comparison given
between conventional and axial viewing.
308
-------
ANALYSIS OF ENVIRONMENTAL SAMPLES USING THE TJA 61E TRACE ICP
Dr. Ruth E. Wolf. Inorganic Technology Manager, Enseco-Rocky Mountain
Analytical Laboratory, Debra K. White, Chief Inorganic Scientist, Enseco
Inc., 4955 Yarrow Street, Arvada, Colorado 80002
ABSTRACT
For many years, the inability of most commercial ICP-OES (Inductively
Coupled Plasma-Optical Emission Spectroscopy) instrumentation to achieve
the detection levels for arsenic, lead, selenium, and thallium needed for
environmental decision making has resulted in the routine use of graphite
furnace atomic absorption spectroscopy (GFAAS) for the analysis of these
four elements. The use of both ICP-OES and GFAAS on environmental samples
requires the preparation of each sample by two separate digestion methods
for CLP work and up to three separate digestion methods for SW-846
Methods, increasing both analysis costs and waste generation. Since GFAAS
is a single element technique, each sample must be analyzed independently
for each desired element. This makes GFAAS analysis both time and labor
intensive. In addition, GFAAS is prone to matrix interferences which in
many cases affect the quality of the data and the quality control sample
burden for CLP GFAA analyses is also high due to the post-digest
analytical spike required on every sample.
The introduction of the 61E Trace ICP by the Thermo Jarrell Ash
Corporation (Franklin, MA) represents an opportunity to decrease the use
of GFAAS for the analysis of arsenic, lead, selenium, and thallium. The
61E Trace ICP is capable of achieving detection levels comparable to GFAAS
for As, Pb, Se, and Tl allowing the use of ICP-OES for quantitation of
these elements at levels of environmental concern. The elimination of the
separate GFAAS sample digestion procedures and the individual analyses of
these elements by GFAAS should result in an increase in sample throughput
and a decrease in analysis costs, turnaround time, and waste generation.
There were, however, some concerns about the comparability of the data
from the ICP-OES analysis to results generated by traditional GFAAS
analyses. Clients with extensive historical data based on GFAAS
methodology have been especially concerned. In order to evaluate the
impact of the Trace ICP on data quality, data comparability, and analysis
costs an extensive study was undertaken at Enseco-RMAL. Client and
reference samples were analyzed using conventional ICP-OES and GFAAS
methods as well as with the new TJA 61E Trace ICP. Data from both ICP and
GFAAS preparation methods were also generated and compared. Finally,
analysis time and costs were tracked and compared to evaluate the
financial impact of the Trace ICP on a large production environmental
laboratory such as Enseco-RMAL. The results generated from these studies
will be presented and discussed as well as several technical
considerations that must be addressed when using the 61E Trace ICP.
309
-------
INTRODUCTION
Although the use of Inductively Coupled Plasma Optical Emission
Spectroscopy (ICP-OES) has become widely used in the environmental
industry since the late 1970's, only recently has it become possible to
analyze for the low levels of arsenic, lead, selenium, and thallium that
are environmentally important using ICP-OES. These four elements have
routinely been analyzed by environmental laboratories using Graphite
Furnace Atomic Absorption Spectroscopy (GFAAS), since the levels of
regulatory concern are typically in the 1-2 parts per billion range, a
detection level that is routinely achieved using GFAAS. The use of GFAAS
for arsenic, lead, selenium, and thallium and ICP-OES for elements such as
aluminum, antimony, barium, beryllium, cadmium, calcium, chromium, cobalt,
copper, iron, magnesium, manganese, nickel, potassium, sodium, silver,
vanadium, and zinc has lead to a highly labor intensive analysis scheme.
In a typical laboratory following SW-846 methodology this would require
two separate sample digestions for GFAAS analysis: 1) for As and Se and
2) for Pb and Tl. In addition a third sample preparation must be
performed for the ICP-OES analysis of the remaining elements. Once the
samples are prepared, arsenic, lead, selenium, and thallium must be
analyzed individually using GFAAS. This results in a total of three
sample preparation procedures and a minimum of 5 analyses to provide
analytical data on the metals content of a single sample.
Other disadvantages of the GFAAS technique include its susceptibility to
matrix interferences which require the use of matrix modifiers and its
comparatively narrow linear range (typically 1-100 ppb) which can lead to
many sample dilutions. Furthermore, most laboratories perform analytical
spikes on each sample when GFAAS is used in order to assess the effect of
the matrix on the sample result. This leads to a relatively high Quality
Control sample burden as well.
In March of 1993, the Thermo Jarrell Ash Corporation (Franklin, MA)
introduced an ICP-OES instrument that could routinely meet detection
levels comparable to those of GFAAS for arsenic, lead, selenium, and
thallium. This would allow the use of ICP-OES for the quantitation of
these particular elements at levels of regulatory concern. The TJA ICAP
61E Trace is an improved version of the TJA 61E ICAP that has long been
used by the environmental industry. The specifications of the TJA ICAP
61E Trace are given in Table 1.
Use of ICP-OES for the analysis of all the environmentally important
metals including arsenic, lead, selenium, and thallium would eliminate the
need for the two separate GFAAS sample digestions and the four separate
GFAAS analyses for these elements. Overall, the ability to use ICP-OES
for all metals analyses would lead to advantages for both the
environmental laboratory and the client (e.g. government agencies,
engineering firms, regional EPA offices, and remediation programs). These
advantages include the following: reduction in analysis costs, reduction
in waste generated by the sample preparation procedures, decreased
analytical turnaround times, and decreased data validation costs.
310
-------
There were, however, some concerns about the comparability of the data
from ICP-OES analysis to results generated by traditional GFAAS analyses.
Clients with extensive historical data were especially concerned. There
were also questions on the performance of the instrumentation including
its ability to meet regulatory detection levels, linear ranges, ease of
use, and if there were any operational difficulties not observed in the
traditional horizontally viewed ICP-OES instrumentation. To answer these
questions, Enseco performed an extensive evaluation of the 61E Trace ICAP
with the cooperation of Thermo Jarrell Ash over a period of about 3
months. The results of these studies will be discussed.
EXPERIMENTAL
Materials and Instrumentation: A TJA 61E Trace ICP with the analytical
lines listed in Table 2 was used for the study. Additional lines were
installed to compensate for interferences and a yttrium line was also
installed to be used as an internal standard. Standard ICP stock
solutions were used from several different commercial standards vendors as
well as actual client samples and NIST Standard Reference Materials (SRMs)
1643c and 2711. A commercially available reference soil from
Environmental Resource Associates was also used in the study.
RESULTS AND DISCUSSION
Detection Limits and Linear Ranges: Detection limits studies were
performed by analyzing a low level standard (10 ppb) with rinsing in
between for a total of 7 measurements per day for three non-consecutive
days. The average of the daily standard deviations was then multiplied by
3 to obtain the Instrument Detection Limit (IDL). Linear ranges were
obtained by analyzing standards of successively higher concentrations
until the highest standard was found where the result read within ±5% of
the true value. The results of the detection limit and linear range
studies are listed in Table 3. It should be noted that these were
obtained with the use of an internal standard to compensate for any
viscosity differences.
The IDLs obtained following the protocol from the Environmental Protection
Agency's (EPA) Contract Laboratory Program (CLP) Statement of Work (SOW)
as described above were generally within a factor of two of those obtained
using the same CLP IDL methodology for GFAAS. The linear ranges obtained
using the Trace ICP are however, much higher - generally by a factor of
one hundred.
Physical interferences/viscosity effects: It has long been known that
greatly differing acid concentrations between samples and standards in
ICP-OES analysis can bias sample results. Most regulatory methods suggest
that users matrix match all samples and standards to avoid such biases.
In order to determine whether such viscosity effects were a significant
311
-------
problem on the Trace ICP a simple study was conducted. The instrument was
calibrated with a standard prepared in a 10% nitric/10% hydrochloric acid
matrix. Standards containing 100 ppb of As, Pb, Se, and Tl were then
prepared in differing acid combinations and concentrations and the
measured concentration recorded. Under SW846 Method 6010A a ±10% control
is established on the calibration. As a result, any biases seen in this
study that were greater than ±10% of the true value were deemed
significant. The results shown in Tables 4 and 5 indicated that
significant biases can occur if the sample acid content differs from that
of the standards. The largest differences occurred when the acid content
was 2% or less and when only hydrochloric acid was used. A further study,
in which yttrium was used as an internal standard to compensate for
nebulization changes caused by differing acid content showed that the use
of such an internal standard element can significantly reduce or eliminate
this type of bias completely in most cases.
Dependance of IDLs on Nebulizer and Conditions: It was also shown that
the Instrument Detection Limits obtained are highly dependant on the
nebulizer used and the optimization conditions of the nebulizer. The
original Meinhard K Type nebulizer installed with the instrument was
replaced with a new Meinhard K Type nebulizer obtained from Thermo Jarrell
Ash and the IDLs determined using the argon and solution flow rates
established with the previous nebulizer. It was found that for some
elements the IDL obtained nearly doubled, as is illustrated in Table 6.
The new nebulizer was able to achieve similar detection limits to the
original nebulizer only after an optimization of argon flow rates was
undertaken (about 4 hours are required for this procedure). As a result
of this study, IDL checking procedures were incorporated into Enseco's
Standard Operating Procedures as part of the optimization routine that
must be followed whenever a new nebulizer is installed.
Data Comparability Studies; Tables 7-9 illustrate results obtained by
Trace ICP, regular ICP-OES, ICP-MS, and GFAA results on a variety of water
and soil samples. For the most part the GFAA and Trace ICP results for
the aqueous samples compare well. The less than symbol indicates that the
measured value was less than the reporting limit. In the case of the GFAA
analyses the Reporting Limits were raised by a factor of 50 due to
dilutions because of poor analytical spike recovery. The ICP-MS results
for Se on the aqueous samples are higher than the GFAA results due to the
presence of Br in the samples that was not corrected for at the time of
analysis.
The results by Trace ICP for NIST SRM 1643C in Table 9 are all generally
within ±10% of the certified value, which is within the limitation imposed
by the calibration verification criterion of Method 6010A of ±10%.
The soil results for client samples show some differences in the lead and
arsenic results, depending on whether the ICP or the GFAA digestate from
Method 3050 was analyzed for some samples. This difference is a result of
the different sample preparation method and not a result of the
312
-------
instrumentation differences.
The results for the PriorityPollutnT - CLP Soil, Lot 213 compare well and
are within the advisory ranges established by the manufacturer for all
elements under any sample preparation conditions. The certified values
were established by a round robin study of 8-15 laboratories performing
SW-846 Method 3050 and Method 6010A for analysis.
The results obtained for NIST SRM 2711 compare very well, generally within
±10%, with the average results obtained from a NIST Interlaboratory Study
of 17 U.S. EPA CLP laboratories.
SUMMARY
The TJA 61E Trace ICP did perform up to the expectations established by
the manufacturer of: 1) IDLs comparable to GFAA, 2) wide linear range, 3)
data comparable to GFAA analysis for most samples, and 4) realization of
decreased analysis costs and turnaround times.
Several recommendations can be made after the 3 month evaluation of the
TJA 61E Trace ICP. First of all, since the IDLs were found to vary with
the individual nebulizer used and that each nebulizer must be
independently optimized to achieve the best IDLs, IDL optimization and
checking procedures should be included in any laboratory Standard
Operating Procedures describing the operation of this instrument.
In addition, the use of an internal standard element, such as Yttrium, to
compensate for acid viscosity effects that can contribute to result bias
due to differing acid concentrations is highly recommended. Although the
study showed that the internal standard could not entirely negate the
problem at vastly different acid concentrations, in the vast majority of
acid concentrations seen in analyses done under SW-846 protocols, the
problem could be alleviated.
Data comparability with traditional GFAA and ICP-OES results was generally
good with real client samples and certified reference materials. Some
differences were observed that appear to be caused by the change in sample
preparation method, rather than due to the change in the analysis
technique
313
-------
Table 1. TJA ICAP 61E Specifications
3/4 meter Rowland Circle, Paschen-Runge Mount
Vacuum Optics
2400 groove/mm Holographic Grating
10 micron entrance and exit slits
Resolution: 0.008 nm 2nd order
0.016 nm 1st order
Axial plasma viewing
Cyclone spray chamber
Meinhard concentric nebulizer
27.12 MHz crystal controlled RF generator
Crawford/Kunselman noise reduction technique
314
-------
Table 2. Configuration of 61E Trace ICAP used in the study.
Element
As
Se
Pb
Tl
Sb
Cd
Al
Ca
Mg
Fe
Mn
V
Cr
Mo
Y
Order
2
1,2
1,2
2
1,2
2
1
1
1
1
1
1
1
1
1
Line (nm)
189.042
196.026
220.353
190.864
206.838
226.502
308.215
317.953
279.078
271.441
257.610
292.402
267.716
202.030
371.030
Use
Analytical
Analytical
Analytical
Analytical
Analytical
Analytical
Interference
Interference
Interference
Interference
Interference
Interference
Interference
Interference
Internal Standard
Table 3. Detection Limits and Linear Ranges
Element
As
Pb
Se
Tl
Sb
Cd
CLP Detection Limit Linear Range
3.0 ppb
2.0 ppb
ppb
ppb
ppb
3.7
2.5
2.3
0.5 ppb
10 ppm
30 ppm
20 ppm
30 ppm
10 ppm
10 ppm
315
-------
Table 4. Physical interference effects caused by acid concentration
Nitric
Cone.
1%
5%
10%
20%
1%
5%
10%
Hydrochloric
Cone.
1%
5%
10%
5%
10%
20%
Arsenic
Recovery
(%)
110.9
105.9
101.2
94.5
109.3
105.4
100.7
111.1
110.3
106.4
Lead
Recovery
(%)
111.6
109.1
109.4
105.7
111.7
108.5
106.5
110.7
107.9
105.7
Selenium
Recovery
(%)
114.2
102.3
94.7
85.9
117.2
106.9
98.4
121.9
121.6
120.9
Thallium
Recovery
(%)
116.3
106.7
101.9
90.7
116.4
116.4
102.0
120.7
117.8
115.2
Notes:
1. Measurements are percent recoveries of a 100 ppb spike in the
listed acid concentration.
2. No Internal Standard was used.
3. Instrument was calibrated at 1 ppm in 10% nitric/10%
hydrochloric acid.
316
-------
Table 5. Use of Internal Standard to correct for physical interference
effects caused by acid concentration
Nitric
Cone.
1%
5%
10%
20%
1%
5%
10%
Hydrochloric
Cone.
1%
5%
10%
5%
10%
20%
Arsenic
Recovery
C/0
101.9
96.9
94.8
91.0
101.2
98.1
95.0
101.3
102.4
99.1
Lead
Recovery
(%)
102.3
99.5
102.3
102.0
103.2
100.7
101.0
100.5
99.9
98.3
Selenium
Recovery
('/•)
104.4
93.1
88.5
82.9
108.1
99.2
93.3
110.5
112.4
112.3
Thallium
Recovery
(%)
107.3
98.1
95.6
87.2
108.1
99.8
96.8
110.5
109.5
107.6
Notes:
1. Measurements are percent recoveries of a 100 ppb spike in the
listed acid concentration.
2. 2 ppm Yttrium was used as Internal Standard.
3. Instrument was calibrated at 1 ppm in 10% nitric/10%
hydrochloric acid.
317
-------
Table 6. Nebulizer Dependence on IDLs
IDLs
Meinhard #1 (Ar @ 0.56 L/min)
Meinhard #2 (Ar @ 0.56 L/min)
Meinhard #2 (Ar 0 0.60 L/min)
As Pb Se T1
2.1 1.1 2.0 1.9 (optimized)
3.9 0.8 4.8 4.0 (not optimized)
2.3 1.0 2.4 2.9 (optimized)
318
-------
Table 7. Data Comparability - Aqueous Samples
Sample
846-15
846-16
846-17
As Results (ppb)
GFAAS
260
<500
240
Trace ICP
297
243
251
Pb Results (ppb)
GFAA
<50
<50
<50
Trace ICP
<3
<3
<3
Sample
3604-14
3604-15
3604-16
3604-16D
3604-17
3604-18
As
Results (ppb)
ICP-MS
68
42
32
31
47
39
TRACE
75
42
35
35
49
44
Pb
Results (ppb)
ICP-MS
640
150
300
290
400
140
TRACE
670
150
320
320
520
140
Se
Results (ppb)
ICP-MS
8.8
4.0
5.7
5.1
7.4
<3
TRACE
4.4
<2
<2
2.2
4.2
<2
Note: Se values by ICP-MS are high due to a Br interference on
82Se that was not corrected for during analysis.
319
-------
Table 8. Data Comparability - Soil Results
Sample
155-01
155-02
155-03
155-04
155-05
155-06
155-07
155-08
DCS*
DCS*
Arsenic Results (ppb)
GFAA
8.6
12
5.6
<10
18
<10
7.4
33
1450
1360
TRACE ICP
GFAA Prep
15
25
11
21
23
29
12
42
1343
1274
ICP Prep
28
34
15
31
24
32
46
54
1308
1300
Lead Results (ppb)
GFAA
78
138
90
145
106
73
82
140
1360
1360
TRACE ICP
GFAA Prep
78
116
94
142
107
156
83
130
1464
1383
ICP Prep
98
271
94
274
135
170
113
145
1378
1361
Environmental Resource Associates PriorityPollutnT - CLP Soil Lot 213
320
-------
Table 9. Reference Material Results
Results for NIST SRM 1643C - Trace Elements in Water
El ement
As
Pb
Se
Sb
Tl
Cd
Trace ICP
Average Results
in ppb, (%RSD), n=2
88.6 (1.6)
31.2 (2.1)
10.3 (1.1)
ND
11.3 (0.9)
13.1 (0.1)
NIST Certified Value
Total Concentration
in ppb, (± error)
82.1 (1.2)
35.3 (0.9)
12.7 (0.7)
ND
7.9*
12.2 (1.0)
* Result for information only, not certified.
ND = Not Detected
Results for NIST SRM 2711 - Montana Soil (EPA CLP Leach Study Results)
Element
As
Pb
Se
Sb
Tl
Cd
Enseco-RMAL
Trace ICP
Method 3050
n=3
rag/kg, (%RSD)
89.0 (3.6)
1050.0 (0.4)
nr
14.0 (0.8)
nr
39.2 (0.2)
NIST Interlab
Leach Results
Range
mg/kg
88-110
930-1500
nr
nr
nr
32046
NIST Average
Method 3050
Leach
Results
mg/kg
90
1100
nr
<10
nr
40
Enseco-RMAL
61E Results
mg/kg
102.6
1100
nr
12.1
nr
37.8
nr = no recovery reported.
321
-------
Results for Environmental Resource Associates PriorityPollutnT - CLP Soil
Lot 213
El ement
As
Pb
Se
Tl
Sb
n=
2
2
2
2
2
GFAA
Results
mg/kg
Avg (%RSD)
140 (4.5)
136 (0)
158 (6.7)
86.7 (3.0)
89.2 (17)
Trace ICP
mg/kg
GFAA Prep
Avg (%RSD)
130 (3.8)
142 (4.0)
123 (3.4)
81.9 (3.1)
85.9 (2.1)
Trace ICP
mg/kg
ICP Prep
Avg (%RSD)
130 (0.5)
137 (1.0)
120 (0.5)
80.0 (0.8)
74.2 (0.9)
Certified
Value
mg/kg
145
148
143
85.1
55.2
Advisory
Range
mg/kg
86-204
97-200
97-189
44-126
10-200
322
-------
ULTRASONIC NEBULIZATION AND ARSENIC VALENCE STATE CONSIDERATIONS PRIOR
TO DETERMINATION VIA INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY.
Carol A. Brockhoff, US Environmental Protection Agency, Chemical
Research Division, Environmental Systems Laboratory, Cincinnati, OH
45268.
John T. Creed, USEPA, CRD, EMSL, Cincinnati, OH 45268 and Theodore D.
Martin, USEPA, CRD, EMSL, Cincinnati, OH 45268
Introduction
Arsenic is currently regulated under the National Primary
Drinking Water Regulations (NPDWR) with a maximum contaminant level
(MCL) of 50 ppb. The drinking water regulations associated with
arsenic are in the process of being updated. This update will include
a revised MCL for arsenic. If the revised MCL for arsenic is set
below 2 ppb, the approved analytical methodologies for routine
monitoring will need sub-ppb detection capabilities. This requirement
alone will limit the number of acceptable analytical
methodologies/techniques which could be utilized in routine
monitoring. One technique which shows considerable promise for the
detection of As in the low to sub-ppb range is Inductively Coupled
Plasma Mass Spectrometry (ICP-MS). One disadvantage of ICP-MS for the
detection of Arsenic is an isobaric interference (40Ar35Cl) caused by
chloride containing matrices. This sensitivity requirement in
combination with the inherent isobaric interference may exceed the
detection capability of ICP-MS if conventional nebulizers are
utilized. The low dissolved solids associated with drinking water
matrices could allow an ultrasonic nebulizer to be utilized to obtain
the added sensitivity for monitoring arsenic in the sub-ppb
concentration range.
Results and Discussion
An ultrasonic nebulizer is unlike a pneumatic nebulizer in that
the aerosol is allowed to traverse a heated chamber followed by a
condenser which desolvates the aerosol prior to its introduction into
the plasma. This desolvation of the aerosol produces an arsenic
response difference between the two predominate valence states of
arsenic, As(III) and As(V). A SOppb As(III) solution in 0.4% HN03 will
produce a response which is approximately 40ppb if a SOppb As(V)
solution is used for calibration. This response difference observed
in a nitric acid only matrix can be eliminated by adding 0.2% HC1 to
the 50ppb As(III) solution. This response difference between As(III)
and As(V) as a function of HC1 was further investigated by collecting
samples at various points along the desolvation pathway of the
ultrasonic nebulizer.
Figure 1 is a schematic of an ultrasonic nebulizer with four
sampling points labelled. These sampling points are: 1.) prior to
nebulization 2.) the spray drain 3.) the condenser drain 4.) the
323
-------
aerosol exit from the condenser. Two sets of samples were collected at
each of the four sampling points. One set of four samples was
collected while nebulizing As in 0.4% HN03 and the other set of four
samples was collected while nebulizing a mixed acid matrix of 0.4%
HN03 and 0.2% HC1. The eight fractions collected at these four
sampling points were then chromatographed [1] to separate the As(III)
from the As(V). These chromatographic results are summarized in
figure 1. The chromatographic results associated with the solution
prior to nebulization indicated that equal amounts of As(III) and
As(V) were found in both matrices. This indicates that the response
difference initially observed was not induced by a preoxidation caused
by the HC1 prior to nebulization. The chromatographic results from
the samples collected from the spray chamber drain indicate that the
mixed acid sample contained a disproportionate amount of As(V). This
may indicate that the heat from the transducer in combination with the
HC1 is causing some accelerated oxidation relative to the sample
containing HN03 alone. The chromatographic results from the
collection of the condenser drain indicate that after traversing the
heating chamber and the condenser of the ultrasonic nebulizer all of
the As(III) has converted to As(V) in the mixed acid matrix. While,
the HN03 acid matrix showed very little additional oxidation. The As
in the mixed acid sample is preferentially being oxidized as it
traverses the heater tube and condenser. The final sample collected
at the exit to the condenser was collected via a bubbler. The two
samples collected at the exit to the condenser are essentially the
same as those collected from the drain of the condenser. The As(III)
from the exit of the condenser is completely oxidized in the presence
of 0.2% HC1 but is only partially oxidized in the presence of 0.4%
HN03.
Conclusion
The ultrasonic nebulizer unlike a conventional pneumatic
nebulizer, produces a response difference between As(III) and As(V) in
a pure nitric acid matrix. This response difference can be eliminated
by adding 0.2% HC1 to both the As(III) and the As(V) solutions. The
HC1 does not oxidize the As(III) to As(V) prior to nebulization, but
rather the As(III) is converted to As(V) as it traverses the
ultrasonic nebulizer. In a pure nitric acid matrix this oxidation by
the ultrasonic is not observed and therefore a response difference is
observed.
References
1.) Raimund Roehl and Maricia M. Alforque, 1992 Winter Conference on
Plasma Spectrochemistry, San Diego, Ca.
324
-------
FIGURE 1
CHROMATOGRAPH1C SEPARATION AS A FUNCTION OF TRANSPORT
4 A*(IU) A«(V)
0.4X HMO, 39X 61X
O.tt HMO,
0.2X HCL
OX
Co
ro
01
HEATED TUBE-''
CONDENSER
As(III) As(V)
59X
CONDENSER DRAIN °'**
0.4X HMO.
0.2X HCL
DESOLVATED
AEROSOL TO
PLASMA
SAMPLE INLET
SPRAY CHAMBER DRAIN
2 Ac(III) As(V)
0.4X HMO, 42X 58X
1 As(III) As(V)
0.4X HMO, 73X 27X
0.4X HMO, 73X 27X
0.2X HCL
0.4X HMO, 23X 77X
0.2X HCL
-------
47
ALTERNATIVES TO THE USE OF ASTM TYPE II REAGENT GRADE WATER
Garabet H. Kassakhian. Manager, Quality Assurance, and Stephanie J. Pacheco, Assistant
Manager, Quality Assurance, Tetra Tech, Inc., 348 West Hospitality Lane, Suite 300, San
Bernardino, California 92408-3216
ABSTRACT
The American Society for Testing and Materials (ASTM) has established standard specifications
for reagent grade water. ASTM Type II Reagent Grade water is often designated by quality
assurance project plans (QAPjP) for use during field decontamination procedures, and for the
preparation of trip and ambient blanks. The United States Environmental Protection Agency
(USEPA) recommends its use in the preparation of reagents, and for general laboratory
procedures, and it is also claimed by most laboratories as part of their quality assurance
programs. The specifications of the most recent ASTM Type II (ASTM D-l 193-91) for
chlorides, sodium, total silica, and total organic carbon are inconsistent, incomplete and
unattainable by routine analyses. A survey of laboratories engaged in federal Remedial
Investigation/Feasibility Study (RI/FS) programs indicates that none can attain the quantitation
limits stated in the ASTM specifications. It is recommended that an alternative standard
specification for Reagent Grade Type II water by the American Public Health Association
(APHA) be used. Although no commercially available purified water is certified to meet ASTM
Type II specifications, the analyses of seventeen lots of purified water for project specific analytes
of concern, namely metals, and volatile and semi-volatile organics, yielded results free of
significant contamination. It is recommended that commercially available purified water be used
in lieu of Reagent Type II water, on the condition that prior to use each lot of 100 to 200 gallons
be tested to comply with the APHA specifications and be free of any analytes of concern.
INTRODUCTION
American Society for Testing and Materials (ASTM) Type II Reagent Grade water is widely
recommended for use by the United States Environmental Protection Agency(USEPA) for use in
laboratory analyses, and by other U.S. Government Departments, e.g. the U.S. Air Force, for
use in field decontamination procedures1'2. Commercial procurement of ASTM Type II water
is a challenge yet to be surmounted, while analyses for verification that the procured water indeed
conforms to the ASTM Standard Specification for Reagent Water3 has proved to be an impossible
task.
The long standing ASTM Standard Specification for Reagent Water D-1193-774, that had been
reapproved in 1983, was finally supplanted in 1991 by D-l 193-91 (Table I). The current D-
1193-91 standard specification requires that chlorides, sodium, total silica and total organic
carbon be determined at very low detection levels. A close examination of the recommended test
methods indicates that there is inherent inconsistencies in the recommended test methods, as
indicated in Table II.
326
-------
TABLE I
ASTM Specification for Type II Reagent Grade Water
CO
Parameter
Total matter, max. mg/L
Electrical conductivity, max
/tmho/cm (/iS/cm) at 298 °K
(25°C)
Electrical resistivity, min.
M«-cm at 298°K (25°C)
Minimum color retention time
of potassium permanganate,
minutes
Total Organic carbon (TOC),
max. jig/L
Maximum soluble silica
Total Silica, max. /j.g/L
Sodium, max. /ig/L
Chlorides, max. /ig/L
ASTM Recommended
Test Method
Method B of Test
Methods D-1888
D-1125
D-1125
D-l 193-77, Section 7.4
D-4779
D-859
D-4517
D-1428
Under development per
D-l 193-91, footnote 63
Specification
ASTM D-1193-77
0.1
1.0
1.0
60
Not applicable
Not detectable
Not applicable
Not applicable
Not applicable
Specification
ASTM D-1193-91
Not applicable
1.0
1.0
Not applicable
50
Not applicable
3
5
5
-------
TABLE II
Status of ASTM Test Methods for ASTM Type II Water Analysis
co
N>
CD
Parameter/
ASTM Recommended
Test Method (current)
Chlorides
Sodium/D-1428-82
Total Silica/D-4517-85
Total Organic Carbon/
D-4779-93
Specification per
ASTM D-1193-91
5/ig/L
5^g/L
3/£g/L
50 ngfL
Method Status
Method under development
Method discontinued in 1990
Method is not applicable; its
range is from 25 /ig/L to 250
Mg/L
New, unproven method, poor
reproducibility. Type II water
typically contains organic
carbon in the range of 0.2
mg/L or less.
Reference
Footnote no. 64
p. 591, 1992 Annual
Book of ASTM
Standards5
p.l, Section 1.26
p.2, Section 8.27
-------
A number of nationally known laboratories currently engaged in federal RI/FS programs, were asked to
supply their currently applicable methods and the relevant practical quantitation limits for the Table II
analytes of interest, as presented in Table III.
Of the six laboratories surveyed none uses ASTM D-4779-93 to determine total organic carbon.
It is evident from Table III that detecting the ASTM required maximum permissible concentrations for
chlorides, sodium, total silica, and total organic carbon is impossible under current analytical day-to-day
operations of the laboratories. It comes as no surprise then that no commercial vendor of Type II water
is able to provide lot or batch test certificates for their products.
During its RI/FS operations, Tetra Tech, Inc. chose to purchase commercially available purified water,
and analyze each lot of 100 to 200 gallons for analytes of interest. The latter were selected on the basis
of detected or suspected contaminants for the sites under investigation. Some or all of the following tests
were conducted on the water samples, which were delivered to the laboratory unopened, and in their
original containers.
Parameters Test Methods
General Minerals California Title 22
ICP Metals SW6010
Arsenic SW7060
Lead SW7421
Mercury SW7470
Selenium SW7740
Total Organic Carbon 415.1
Silica as SiO2 SW6010
Volatile Organics SW8260
Semivolatile Organics SW8270
Specific Conductance E120.1
Four independent laboratories analyzed 17 samples of purified water. No analytes of interest as identified
in past work were detected in the water analyzed. Although the purified water was supplied in heavy
gauge 5 gallon polypropylene containers, the volatile and semivolatile organic tests did not indicate the
presence of contaminants in the purified water traceable to the container material (Tables IV and V).
CONCLUSIONS
Since the current ASTM Type II reagent grade water specifications are difficult to adhere to and verify
by readily available analytical means, it is recommended that commercially available purified water be
used instead of ASTM Type II water, after testing a lot sample for analytes of interest.
It is also recommended that until ASTM clarifies the status of its Reagent Water Specification and the
status of the related tests, a substitute specification be used, namely the one in use by the American Public
Health Association (APHA).
Currently APHA, the American Water Works Association (AWWA), and the Water Environment
Federation (WEF) use a less stringent Type II Reagent Water Specification8, namely:
329
-------
TABLE III
ASTM TYPE II WATER ANALYTE
PRACTICAL QUANTITATION LIMITS FOR SELECTED LABORATORIES
Parameter
Silica
Chlorides
Sodium
Total
Organic
Carbon
ASTM Specified
Maximum
Permissible
Concentration in
/ig/L
3 /ig/L
5 /ig/L
5 /ig/L
50 /tg/L
Laboratory
Suggested
Method
SW6010
160.3
SW6010
415.2
Lab 1 located in
EPA Region IX
200 /tg/L
20 /ig/L
2,000 /ig/L
1,000 /ig/L
Lab 2 located in
EPA Region X
20 /ig/L
1,000/ig/L
(Method 325.2)
200 /tg/L
500 /ig/L (Method
SW9060)
Lab 3 located in
EPA Region IX
200 /ig/L
Not applicable
not applicable
1,000 /ig/L
(Method 415.1)
Lab located in
EPA Region V
Not Applicable
100 /ig/L
(Method 300.0)
240 /ig/L
5,000 /ig/L
(Method 415.1)
CO
co
o
-------
TABLE IV
Extraction Method: EPA Method 3520
Analytical Method: EPA Method 8270
Malriii Water
Unit*: ug/L
Test ReiulU of SemlvobtUt Organic* in Purified Water
method 8270
Parameters
PQL MDL AUP-6 AUP-7 AUP-8 AUP-9 AUP-10 AUP-11 AUP-12 AUP-13 AUP-14
4/6/93 6/17/93 8/11/93 9/9/93 12/16/93 12/16/93 2/17/94 4/7/94 4/7/94
Acenaphthene
Acenaphthylene
Anthracene
Benzo (a) anthracene
Benzo (b) riuoranthene
Benio (V) Diioranlhene
Benio (g,h,l) perylene
Benzo (a) pyrene
Benzyl alcohol
Bis (2-Chloroethoxy) met
Bis (2-Chloroisopropyl) e
Bis (2-Chloroethyl) ether
Bis (2-Elhylhexyl) phthal
4-Bromophenyl phenyl et
Butyl benzyl phthabte
4-Chloraaniline
2-Chloronaphthalene
4-Chlorophenyl phenyl et
Chrysene
Dibenzo (a,h) anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dkhlorobenzene
1 ,3-Dichlorobenzene
1,4-Dkhlorobenzene
3^'-Dichlorobenzidine
Dimethyl phthalaU
Diethyl phthalate
2,4-Dinltrotoluene
2,6-Dinitrotoluene
Di-n-octyl phlhalale
Fluoranthene
Fluorene
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadie
Hexachloroethane
Indeno (1,2,3-cd) pyrene
bophorone
2-Methylnaphthalene
Naphthalene
2-Nitroanlllne
3-Mtroanlllne
4-Nitroanillne
Nitrobenzene
N-NUrosodiphenylamine
N-Nitroso-di-n-propylami
Phenanthraie
Pyrene
1 ,2,4-Trkhlorobenzene
Benzok Acid
4-Chloro-3-methylphenol
2-Chlorophenol
2,4-Dfchlorophenol
2,4-Dimethylphenol
4,6-Dlnitro-2-mcthylphen
2,4-Dinllrophenol
2-Methylphenol (o-Crool
4-Methylphenol (p-Cresol
2-Nltrophenol
4-Nitrophenol
Pentachlorophenol
Phenol
2,4,5-Trtchlorophenol
2,4,6-Trichlorophenol
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
11
10
10
10
10
10
10
10
10
20
10
10
10
10
10
10
10
10
10
10
10
10
10
10
11
50
50
50
10
10
10
10
10
10
50
10
10
10
10
50
50
10
10
10
50
30
10
50
10
8.8
6
3
3
4
3
4
3
2
2
4
1
7
3
4
5
11
6
4
4
6
3
8
6
10
3
3
4
3
3
4
3
4
3
8
10
6
4
2
8
11
2
3
10
2
2
2
3
3
9
4
2
2
I
2
4
6
4
2
3
5
3
1
3
2
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.U
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
J- DETECTED BETWEEN THE MDL AND PQL
331
-------
Extraction Method: EPA Method 5030
Analytical Method: EPA Method 8260
Matrix: Water
Units: ug/L
TABLE V
Test Results ofVolatileOrganics in Purified Water
Method 8260
CO
co
ro
Parameters
Chloromethane
Bromomethane
Vinyl chloride
Chloroethane
Methylene Chloride
Acetone
Carbon disulfide
1,1 -Dichloroethene
1,1-Dichloroethane
cis-l,2-Dichloroethene
trans-l,2-Dichloroethene
Chloroform
1,2-Dichloroethane
Methyl Ethyl Ketone
1,1,1-Trichloroethane
Carbon Tetrachloride
Vinyl Acetate
Bromodichloromethane
1,1,2,2-Tetrachloroethane
1,2-Dichloropropane
trans-1,3-Dichloropropene
Trichloroethene
Dibromochloromethane
1,1,2-Trichloroethane
Benzene
cis-1,3-Dichloropropene
2-ChloroethyI vinyl ether
Bromoform
2-Hexanone
Methyl Isobutyl Ketone
Tetrachloroethene
Toluene
Chlorobenzene
Ethylbenzene
Styrene
o-Xylene (1,2-Dimethylbenzene)
m,p-Xylene (Sum of Isomers)
PQL MDL
10
10
10
10
s
10
5
5
5
5
5
5
5
10
5
5
10
5
5
5
5
5
5
5
5
5
10
5
10
10
5
5
5
5
5
5
5
3
2
3
2
2
10
5
3
2
5
3
2
1
3
3
3
10
1
3
1
1
3
2
2
2
1
10
1
4
4
3
2
2
2
2
5
5
AUP-6
4/6/93
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
5J
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP-7
6/17/93
ND
ND
ND
ND
19B
12
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP-8
8/11/93
ND
ND
ND
ND
10
ND
ND
ND
ND
ND
ND
ND
ND
4J
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP-9
9/9/93
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
4JB
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP-10
12/16/93
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP-11
12./16/93
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP-13
4/7/94
ND
ND
ND
ND
ND
10
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
AUP14
4/7/94
ND
ND
ND
ND
ND
12
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
-------
APHA/AWWA/WEF Reagent Water Specification
Quality Parameter Type II
Bacteria,CFU/mL less than 1000
pH not specified
Resistivity, megohm-cm at 25 °C greater than 1
Conductivity, /imho/cm at 25 °C 1
SiO2,mg/L less than 0.1
Particulate matter not specified
Organic contaminants not specified
ACKNOWLEDGMENTS
The authors would like to thank Mr. Arlen Saxton, Ms. Donna M. Charles, and Ms. Lisa Arrasmith for
their assistance during the preparation of this paper.
REFERENCES
1. U.S.EPA, "Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846", pp.
One-28 and 29, 3rd Edition, Final Update I, July 1992, U.S.EPA, Washington, DC 20460
2. U.S. Air Force Center For Environmental Excellence," Handbook For the Installation
Restoration Program (IRP) Remedial Investigations and Feasibility Studies(RI/FS)", p.2-29,
September 1993, Environmental Services Directorate, Brooks Air Force Base, Texas 78235
3. American Society for Testing and Materials, "Standard Specification for Reagent Water D-1193-
91", November 1991, ASTM, Philadelphia, Pennsylvania 19103
4. American Society for Testing and Materials, "Standard Specification for Reagent Water D-1193-
77 (Reapproved 1983)", March 1977, ASTM, Philadelphia, Pennsylvania 19103
5. American Society for Testing and Materials, "Standard Test Methods for Sodium and Potassium
in Water and Water-Formed Deposits by Flame Photometry, D-1428-82" in the " 1992 Annual
Book of ASTM Standards, Section 11, Water and Environmental Technology", 11.01, p. 591,
ASTM, Philadelphia, Pennsylvania 19103
6. American Society for Testing and Materials, "Standard Test Method for Low Level Total Silica
in High-Purity Water by Flameless Atomic Absorption Spectroscopy D-4517-85", October 1985,
ASTM, Philadelphia, Pennsylvania 19103
7. American Society for Testing and Materials, "Standard Test Method for Total, Organic, and
Inorganic Carbon in High Purity Water by Ultraviolet (UV) or Persulfate Oxidation, or Both, and
Infrared Detection", November 1993, ASTM, Philadelphia, Pennsylvania 19103
8. American Public Health Association, American Water Works Association, Water Environment
Federation, "Standard Methods For The Examination of Water and Wastewater", p. 1-32, 18th
Edition, 1992, APHA, Washington, DC 20005
333
-------
48
IMPROVEMENTS TO EPA METHOD 335.2 FOR DETERMINATION OF TOTAL
CYANIDE ACHIEVED BY OXIDATION OF INTERFERENCES
B.B. Potter, U.S. Environmental Protection Agency, Office of Research and Development,
Cincinnati, OH 45268, M Goldberg and A. Clayton, Research Triangle Institute, Research
Triangle Park, NC 27709
ABSTRACT
EPA Method 335.2 For the determination of total cyanide works well for simple cyanide
salts in the absence of method interferences. However, industrial effluents may contain several
interfering chemicals, including sulfide, thiocycanate, phenols, amines, oxidants, carbonates,
aldehydes, sugars, fatty acids, and nitrite. Although other researchers have studied the effects of
these interferences individually, no one has previously examined the effect of multiple
interferences present simultaneously. We prepared test samples that simulated sewage treatment
plant effluent and contained sulfide, thiocyanate, caffeine, nitrite, and hypochlorite. Samples were
analyzed for cyanide content using Method 335.2 with a colorimetric endpoint. Results indicated
that the relationship between measured cyanide concentration and the known amounts of cyanide
and interferences is complex. Not only do the interferences exhibit a significant effect
individually, they also exhibit significant interaction effects. Experiments performed using
potassium permanganate and sodium vanadate indicated that either oxidant effectively removes
most of the interference effects. However, an exact stoichiometric amount of permanganate was
required for accurate cyanide; the presence of excess permanganate caused oxidation of cyanide.
and thus low recovery. Excess vanadate was well tolerated by the method, and accurate analysis
of cyanide was achieved even when sulfide, caffeine, nitrite, and p-cresol were present in the
cyanide sample. However, neither oxidant was capable of removing the interfering effects of
thiocyanate, and the use of vanadate can only be recommended in the absence of thiocyanate.
334
-------
ANALYSIS OF As, Pb, Se and Tl IN SOLD) WASTE
USING THE TJA 61E TRACE ANALYZER ICP
Mark E. Tatro. Spectra, Inc., P.O. Box 1176, McAfee, NJ 07428; Larry Camp and Rick
Christenberry, Progress Environmental Laboratory, 4420 Pendola Point Road, Tampa, FL
33619
ABSTRACT
The Thermo Jarrell Ash [TJA] 61E Trace Analyzer ICP has been claimed by the manufacturer
to give detection limits comparable to Furnace AA for most elements. Typically most
environmental laboratories use Furnace AA for the analysis of As, Pb, Se and Tl to meet the
EPA Contract Laboratory Program (CLP) required detection limits of 10,3,5 and 10 ng/ml,
respectively (1). We investigated the manufacturer's claims by developing methods for the
analysis of As, Pb, Se and Tl in wastewater and solid waste to verify that the Trace Analyzer
could replace Furnace AA analysis. Our results showed that with careful programming of
background correction and inter-element correction factors, that method detection limits of
2.4,1.4, 1.7 and 1.4 ppb for As, Pb, Se and Tl respectively can be obtained, that recoveries
for these four analytes in the CLP ICP Interference Check Sample are in the 90% range, and
that recoveries of these four analytes from real world spiked solid waste samples are very
acceptable.
INTRODUCTION
Progress Environmental Laboratory [PEL] purchased a TJA 6IE Trace Analyzer to fulfill
the requirements of a contract to analyze solid waste samples for a number of analytes
including As, Pb, Se and Tl. The manufacturer claims that the 61E Trace Analyzer can meet
Furnace AA detection limits and PEL saw the obvious productivity advantages of analyzing
for all elements by simultaneous ICP. PEL called in SPECTRA Spectroscopy &
Chromatography Specialists, Inc. [SPECTRA] to develop the ICP methods for the 12
required contract elements including As, Pb, Se and Tl.
The problem with analyzing As, Pb, Se and Tl by ICP for environmental applications are
typically twofold. One is that the signal to noise ratio is too low to meet the required detection
limits; and, the other is that the emission from aluminum and iron, usually present at high
concentrations, interferes with the analyte wavelength emission particularly for Pb and Se.
The 6IE Trace Analyzer solves the signal to noise drawback by mounting the plasma torch
horizontally so that the optical path is coincident with the plasma. This allows a larger
segment of the plasma to be viewed than in the typical vertical mounted torch configuration.
A redesign of the transfer optics improves the analyte signal intensitywithout increasing the
background intensity (2).
335
-------
The blank corrected signal for 1 ug/ml Cd using the conventional vertical torch configuration
of the T JA 61ICP is 703 intensity units as measured by SPECTRA during a previous methods
development course. The blank corrected signal for 1 ug/ml Cd using the horizontal torch
configuration of the 61E Trace Analyzer is 39,568 as measured here. Therefore, the 61E
Trace Analyzer ICP provides a blank corrected signal enhancement of 56X for cadmium.
The 61E Trace Analyzer solves the background correction from aluminum and iron on Pb
and Se by a simultaneous background correction technique where the analyte peak is
measured at one wavelength order and the background is measured at another wavelength
order simultaneously and by an improved wavelength resolution from the incorporation of
a holographic grating. A dummy "S" channel is used to monitor the first and second order
wavelengths of Pb and Se.
EXPERIMENTAL
The 6 IE Trace Analyzer ICP is equipped with a 2kW RF generator controlled at 27.12 MHz,
built in peristaltic pump, concentric glass nebulizer, glass spray chamber and demountable
quartz torch. The optics consist of a 0.75 m Rowland Circle with a Paschen-Runge mount
2400 grooves/mm holographic grating blazed at 500 nm. The system is interfaced to an IBM/
AT 386-33 computer using the ThermoSPEC software. The ICP conditions used for analysis
are detailed in Table 1.
The logic followed by SPECTRA for setting ICP background correction intervals is to use
the graphics capabilities to' 'let the instrument speak for itself' (3). Standards were prepared
for setting background correction positions using 1,000 and 10,000 ug/ml plasma grade
standards from J.T. Baker. The standards were (a) matrix blank; (b) 1 ug/ml As, Pb, Se and
Tl; (c) 10 ug/ml Cr, Mn and V; and (d) 500 ug/ml Al, Ca, Mg and 200 ug/ml Fe. Since the
samples would be in a matrix of 1% HNO3 + 5% HC1 following the SW-846 Method 3050
digestion procedure, all standards were prepared in this same matrix using J.T. Baker Instra-
Analyzed grade acids. Scans were made using the four multi-element standards and
overlayed to set the background correction positions for As, Pb, Se and Tl. The background
correction positions chosen are detailed in Table 2. The two separate positions for Pb and
Se represent the first and second order wavelength lines for these two elements.
To determine lEC's, the ICP was programmed with a series of four standards in a water
matrix. We did not use the acid matrix to avoid contaminating the standards with Instra-
Analyzed grade acids. The calibration standards used were (a) blank; (b) 1000 ng/ml As, Pb,
Se, Tl; (c) 500 ug/ml Al, Ca, Mg and 200 ug/ml Fe; and (d) 10 ug/ml Cr, Mn and V. After
calibration, the following standards were analyzed to calculate the lEC's: (a) 500 ug/ml Al;
(b) 200 ug/ml Fe; (c) 500 ug/ml Ca; (d) 500 ug/ml Mg; (e) 10 ug/ml Cr; (f) 10 ug/ml Mn;
(g) 10 ug/ml V.
336
-------
We found it very useful to use the command function of the ThermoSPEC software to
evaluate the effectiveness of the lECs. After the analysis of each standard, e.g. the 500 ug/
ml Al, we saved the results that had not been corrected by lECs using the command function
srd"500Al". After entering the lECs for aluminum on the affected analytes into the
program, we used the command function Ird'' 500 Al' 'rdpv. This command applies the lECs
to all the affected analytes and prints out a data sheet that shows the analytical results with
the lECs applied. In this way we could estimate immediately if the lECs corrected
interferences.
Table 3 details the lECs that were calculated from our study. We used the analyzed value
for the interferent rather than the theoretical value. We were surprised to see lECs with values
such as -2.4202 for the affect of V on Tl since we have been used to seeing lECs with values
of 10"3 and 10"4. The large lECs are due to the fact that for the traces such as As, Pb, Se and
Tl the standards are programmed as ppb and the interferents are programmed as ppm.
Therefore, the IEC for V on the Tl line was calculated as -24.71 ppb Tl/10.21 ppm V = -
2.4202.
The proof of correct entries for background correction and lECs is the analysis of the EPA
ICS [AB] standard. We prepared an ICS [AB] standard and spiked it with 500 ppb As, Se
and Tl. The concentration of Pb in this standard is set at 1000 ppb. Table 4 details the
recoveries obtained on the ICS [AB] standard. The EPA Contract Laboratory Program
recommends recoveries between 80 to 120% for this standard. Our results were all well
within these limits.
To determine method detection limits [MDLs] a multi-element standard was prepared
containing all the elements of interest at a concentration of 3-5 times the reported instrument
detection limit as required by EPA (4). Seven aliquots of the standard were digested using
SW846 Method 3050 and analyzed as separate samples. Table 5 details the MDL
concentrations determined from this study. The MDLs demonstrate that the 6IE Trace
Analyzer can meet EPA CLP detection limits for As, Pb, Se and Tl.
Over the course of one month, 20 individual 1.0 gram aliquots of different soil samples from
the contract study at PEL were pre-digestion spiked with 250uLofalOO ug/ml multi-element
standard and brought up to 100 mis following digestion by the staff at PEL. The spike levels
were, therefore, 250 ng/ml in solution or 25 ug/gm on a w/w basis. Each soil sample was
analyzed on a separate day and the recoveries calculated to formulate a Shewart Accuracy
Control Chart. The average recoveries with 1 sigma standard deviations are detailed in Table
6.
337
-------
SUMMARY
The TJA 6 IE Trace Analyzer ICP system is capable of replacing Furnace AA for the analysis
of As, Pb, Se and Tl in solid waste samples. We found that the greatly improved sensitivity
of this instrument providing MDLs at the < 1 ppb levels, requires the laboratory personnel
to pay special attention to purity of standards and acids as well as contamination from the
environment.
ACKNOWLEDGEMENTS
The authors wish to thank Vince Giampa of Progress Environmental Laboratory for his
valued support throughout this project.
REFERENCES
1. USEPA Contract Laboratory Program Statement of Work for Inorganic Analysis,
Document Number ILM01.0
2. Bob Foster, TJA, personal communication.
3. Tatro, M.E., Methods Development Logic for ICP-AES, Spectroscopy, 1(4), 21
(1986).
4. Appendix B, Part 136, Definition and Procedure for the Determination of the Method
Detection Limit, Revision 1.11, 40 CFR, Chapt. 1 (7-1-90).
338
-------
TABLE 1
TABLE 2
PARAMETER
Number of Repeats
Integration
Flush Time
Torch Gas
Auxilary Gas Flow
Nebulizer Pressure
RF Power
Analysis Pump Rate
Flush Pump Rate
Relaxation Rate
Pump Tubing
SETTING
3
15 sec
45 sec
High Flow
Low Flow
30psi
950 W
130
130
0
Tygon-Orange
ELEMENT
As
Pb(l:+10SS)
Pb (2: - 10 SS)
Se(l:+10SS)
Se(2:-10SS)
Tl
BGC OFFSET
-15
-10
+10
-10
+10
-15
339
-------
TABLES
ANALYTE
As 1890
Pb2203-l
Pb2203-2
Se1960-1
Sel960-2
Til908
"S" CHANNEL
Pb 2203
Se 1960
INTERFERENT
Fe2714
A13082
Ca3179
Mg2790
G-2677
Fe2714
A13082
Mg2790
Fe2714
A13082
Mg2790
Cr2677
V2924
Fe2714
Fe2714
Mn2576
Fe2714
Cr2677
Mn2576
V2924
WAVELENGTH
2203-1
2203-2
1960-1
1960-2
- 0.071890
0.008193
0.026903
0.004294
0.165659
0.089876
0.394050
0.003921
0.050755
•0.185810
0.004528
0.244488
•0.314890
• 0.026270
• 0.240540
0.283170
•0.212650
0.220098
0.215604
• 2.420200
kl
• 0.333
• 0.667
• 0.333
• 0.667
340
-------
TABLE 4
TABLES
TABLE 6
ELEMENT
As
Pb
Se
Tl
% R FROM ICS [AB1
94
92
90
90
ELEMENT
As
Pb
Se
Tl
MDL (ppb)
2.4
1.4
1.7
1.4
ELEMENT
As
Pb
Se
Tl
SOIL SPIKE RECOVERY (%
92.9 ± 10.9
102.8 ± 7.4
86.8 ± 16.7
102.8 + 8.5
341
-------
50
TRACE METAL VALENCE STATE CONSIDERATION WHEN USING
ULTRASONIC NEBULIZER AND INDUCTIVELY COUPLED PLASMA
ATOMIC EMISSION SPECTROMETRY (ICP-AES).
Carol A. Brockhoff, U.S. Environmental Protection Agency,
Chemistry Research Division, Environmental Monitoring Systems
Laboratory, Cincinnati, Ohio 45268, Theodore D. Martin, U.S.
EPA, CRD, EMSL-Cinti, Cinti, OH 45268, and John T. Creed, U.S.
EPA, CRD, EMSL-Cinti, Cinti, OH 45268
ABSTRACT
In recent years there has been a need to measure trace
metal concentrations in ambient and drinking waters at lower and
lower concentrations. Under the Safe Drinking Water Act the
maximum contaminant level (MCL) of some primary metal
contaminants have been lowered to the exclusion of ICP-AES from
the list of approved methodology for compliance monitoring of
these analytes. New regulations based on water quality criteria
have placed new sensitivity requirements on methodologies that
can be used for the monitoring of ambient waters. Although
widely used for many types analyses, ICP-AES has limited
acceptance under these more demanding conditions. The utility
of ICP-AES can be extended by utilizing an ultrasonic nebulizer
to achieve improved sensitivity.
The 3-10 fold decrease in detection limits achieved with
the ultrasonic nebulizer are not accomplished, however, without
some analytical limitations. In this paper data will be
presented demonstrating the limitation induced by the
desolvation system on the determination of arsenic, chromium,
and selenium in aqueous matrices. The desolvation step used in
ultrasonic nebulization induces a different signal response for
the two valence states of arsenic [As(III) vs As(V)] and
chromium [Cr(III) vs Cr(VI)], a phenomenon which does not occur
with pneumatic nebulization. For 1 mg/L solutions As(V) showed
a 60% enhancement over As(III), while Cr(III) showed a 30%
enhancement over Cr(VI). If hydrogen peroxide is added to a
mixed acid (HNO, + HC1) solution of these analytes, As(III) is
oxidized to As(V) and Cr(VI) is reduced to Cr(III). This simple
addition of H202 to both sample and standard solutions alike
brings these analytes to a common valence state and eliminates
the difference in signal response.
In the case of selenium the enhanced analytical response is
dependent on the presence of concomitant elements. The degree
of enhancement is not only dependent on the particular
element(s) in solution, but also their concentration. Different
concomitant mixtures will produce different degrees of
enhancement. A common drinking water matrix or a 1 mg/L
solution of ICP-19 will produce w 100% increase in signal over
a 1 mg/L single element solution of selenium. One way to
circumvent this matrix interference is to prepare a matrix
matched calibration standard that approximates routine samples.
These analytical limitations are discussed in greater
detail in the poster, a must stop on your poster paper tour.
342
'0
-------
51
Analysis of Silver and Other Elements by
Toxicity Characterization Leaching Procedure
(TCLP, SW 846 1311)
David Eugene Kimbrough*, Public Health Chemist II, Jayavant
Patel, Public Health Chemist I, Janice Wakakuwa, Supervising
Chemist, California Environmental Protection Agency,
Department of Toxic Substances Control, Southern California
Laboratory, 1449 W. Temple Street, Los Angeles California
90026-5698.
Abstract
Under Federal and State law a material is considered
hazardous if exhibits the "characteristic of toxicity" as
determined by the Toxicity Characterization Leaching
Procedure, or TCLP, (SW 846 as method 1311). There are eight
metals that can be tested by the TCLP which are silver,
arsenic, barium, cadmium, chromium, lead, mercury, and
selenium. This test using one of two acetic acid buffers to
the amount of these elements which will leach in a landfill.
A study is presented on the solubility of silver and
fifteen other elements using both of the TCLP buffers and a
citric acid buffer used in similar leaching test mandated by
California law (the Waste Extraction Test). The solubility
of silver nitrate, silver chloride, and silver sulfate are
all tested alone and with soils. Real world and laboratory
control samples are also tested.
While reagent grade silver nitrate and sulfate are very
soluble, silver chloride is high insoluble. During the
extraction of real world materials, buffer silver nitrate and
sulfate form highly insoluble silver salts. Thus, it is
enormously unlikely for any real world material every to be
determined to be hazardous using this test.
343
-------
52
INTRODUCTION TO THE USEPA, METHOD 245.7, DETERMINATION OF MERCURY
BY AUTOMATED COLD VAPOR ATOMIC FLUORESCENCE SPECTROMETRY.
Billy B. Potter. U.S. Environmental Protection Agency, Environmental
Monitoring Systems Laboratory, Cincinnati, Ohio; William H. McDaniel,
Jenny Scifres, Michael A. Wasko, U.S. Environmental Protection Agency,
Region 4, Athens, Georgia; Winslow J. Bashe and Miguel D. Castellanos,
Technology Applications, Inc., Cincinnati, Ohio.
ABSTRACT
The U.S. Environmental Protection Agency (USEPA), Environmental
Monitoring Systems Laboratory - Cincinnati (EMSL-Cincinnati) and Region
4, Environmental Services Division, Laboratory, Athens Ga., have
developed EPA Method 245.7 for the determination of total and dissolved
mercury found in water. The mercury method has an estimated method
detection limit (MDL) that ranges between 0.3 ppt to 5 ppt of mercury.
The MDL is made possible by digesting the sample using bromide/bromate
reagent followed by detection of elemental mercury by cold vapor atomic
fluorescence spectrometry at 253.7 nm.
INTRODUCTION
The METHOD 245.7, DETERMINATION OF MERCURY BY AUTOMATED COLD VAPOR,
ATOMIC FLUORESCENCE SPECTROMETRY is written in the Environmental
Monitoring Management Council (EMMC) method format. The EMMC format
consists of the following sections:
1.0 SCOPE AND APPLICATION
2.0 SUMMARY OF METHOD
3.0 DEFINITIONS
4.0 INTERFERENCES
5.0 SAFETY
6.0 APPARATUS, EQUIPMENT, LABORATORY AND CLEANING REQUIREMENTS
7.0 REAGENTS AND CONSUMABLE MATERIALS
8.0 SAMPLE COLLECTION, PRESERVATION, AND STORAGE
9.0 QUALITY CONTROL
10.0 CALIBRATION AND STANDARDIZATION
11.0 PROCEDURE
12.0 DATA ANALYSIS AND CALCULATIONS
13.0 METHOD PERFORMANCE
14.0 POLLUTION PREVENTION
15.0 WASTE MANAGEMENT
16.0 REFERENCES
17.0 TABLES, DIAGRAMS, FLOWCHARTS, AND VALIDATION DATA
Each section addresses the details of the method application, procedures
and quality control issues necessary for the proper execution of the
method.
344
-------
Method 245.7 describes procedures for the determination of mercury
(organic + inorganic) total recoverable or dissolved (filtered 0.45/0,
in drinking water, surface and ground water, sea and brackish water,
industrial and domestic wastewaters. The chemistry of sample digestion
is based on the brominating reagent produces bromine monochloride:
KBr03 + 2KBr + 6HC1 -* 3BrCl + 3KC1 + 3H20
In the presence of excess bromide ions and acid, the bromine
monochloride is then converted to free bromine1:
BrCl + excess KBr -»• Br2 + KC1
Mercury from inorganic mercury compounds are rapidly oxidized by bromine
and organomercury species are degraded by the oxidizing properties of
bromine releasing mercury (II)2-3 as follows:
HgR2 + Br2 -»• RHgBr + RBr
RHgBr + Br2 -*• HgBr2 + RBr
The excess Br2 reacts to oxidize mercury to form a complex. After the
oxidation reactions are complete, excess bromine is removed by the
addition of hydroxylamine hydrochloride.
The elemental mercury vapor is generated from the digested sample by
reduction with stannous (tin II) chloride in the presence of
hydrochloric acid*1. High purity argon gas is used to purge the mercury
vapor from a gas/liquid separator driving the equilibrium to the right
as follows:
Sn+2 + Hg+2 -*• (Art), Hg°t + Sn+U
The excess Sn"1"2-!-, HC14-, solution is discharged to a waste container and
the mercury vapor is carried by the argon flow to the mercury
concentrator or detector. The liquid containing spent reagents and
sample are flushed continuously from the gas/liquid separator to a waste
container. This waste contains tin and hydrochloric acid and does not
contain mercury. The elemental mercury vapor is then purged from
solution by a carrier stream of argon through a semi-permeable dryer
tube5 that removes water vapor.
For additional sensitivity mercury vapor may be concentrated on the
optional gold amalgam concentrator. The concentrator is then heated
rapidly to 450°C. The concentrated mercury vapor passes to the detector
and is integrated as a peak. If the concentrator is not used, the vapor
passes directly to the detector and is measured as a change in the rise
(height) from the baseline. The mercury vapor concentration is
determined by atomic fluorescence spectrometry at 253.7 run6-79
345
-------
The method was optimized using a statistically-based experimental design
or chemometric approach as described by Deming and Morgan (1987)10.
The chemometric experimental approach was applied to this mercury method
to speed the process of method evaluation. The chemometric approach is
dynamic (modifiable) and recursive (experiments may be repeated).
During the execution of the experiments an evaluation of each "phase" of
an experiment is required. When a modification of the experiment was
required, it was strongly supported by the statistical evidence. The
experimental design consisted of the following phases:
Phase 1 Familiarization Study.
Phase 2 Automated Instrument Optimization Study.
Phase 3 Automated Instrument Linearity Study.
Phase 4 Mercury Precision and Recovery Study.
Phase 5 Instrument Stability Study.
Phase 6 Initial Interference Study.
Phase 7 Sample Preservation Study.
Phase 8 Single Laboratory Validation Study.
Phase 9 Establish Instrument Control Charts.
Phase 10 Establish Clean Laboratory Protocol.
The automated instrument is generally configured as shown below. The
gold amalgam accessory for the instrument system is not shown in this
configuration. The gold amalgam accessory will be evaluated during the
ruggedness testing part of the method development and is not part of the
scope of this experimentation.
In the familiarization and optimization phase of the experiments, the
mercury analyzer was optimized for maximum sensitivity and/or signal-to-
noise ratio. The use of Simplex optimization was investigated using the
carrier gas and sheath gas flow rates as selected variables. A range
for optimized settings was found as described in Table 1. These settings
may be changed periodically to optimize the instrument. Small changes,
as long as they remain within the specified ranges, do not adversely
effect the instruments performance. However one setting was made and
procedures were held constant for the remaining experiments.
346
-------
PERISTALTIC
PUMP
AUTOSAMPLER
FLUORESCENCE
DETECTOR
CARBON
WASTE FILTER
CVent to Hood)
FIGURE 1: PSA AUTOMATED MERCURY FLUORESCENCE SYSTEM
347
-------
INSTRUMENT CONTROL SETTING AND ARGON GAS FLOW SETTINGS
Fluorescence Instrument
Parameters
Delay Time
Rise Time
Analysis Time
Memory Time
Argon Gas Control
Gas Regulator
Carrier Flow
Drier Tube Flow
Sheath Flow
PSA Merlin Series AFS
Range of Settings
5 to 15 seconds
20 to 30 seconds
30 seconds
60 seconds
Range of Settings
20 to 30 psi.
150 to 450 mL/minute
2.5 to 3.5 L/minute
150 to 250 mL/minute
The Method 245.7 procedure used is based on a method used by the
Yorkshire Water Authority (YWA) in the United Kingdom11. The method
procedure is simplified and summarized as follows:
1) Add 5 mL (1+1) hydrochloric acid and 1 mL 0.1N potassium
bromate/potassium bromide solution to a 50 mL conical vial.
2) Transfer of sample to conical vial filling to the 50 mL mark.
3) Allow samples to stand for at least 30 minutes before analysis.
4) Add 50 /iL hydroxylamine hydrochloride solution to each conical
vial.
5) Turn on the automated instrument/detector and allow to
stabilize.
6) The sample enters gas/liquid separator with SnCl2 to form
mercury vapor.
7) The vapor is analyzed by cold vapor atomic fluorescence
spectrometry.
DISCUSSION
The determination of total mercury by automated cold vapor atomic
fluorescence spectrometry has a linear range approximately 2 ng-Hg/L to
25 /ig-Hg/L. The MDLs as calculated are as follows:
348
-------
METHOD DETECTION LIMITS FOR MERCURY (ng Hg/L)
MATRIX
Reagent Water
Florida Marsh Water
Synthetic Sea Water
Sea Water
Lake Water
Waste Water
EPA\EMSL-Cin.
Glove Box
MDL
1.8
3.3
2.6
EPA\Region 4
Clean Room
MDL
0.31 to 1.0
S . E . Environ .
Research, FIU
MDL
0.27 to 0.59
1.4
0.33
0.40
The MDL's may then be used for enforcement of water quality-based
effluent limitations (WQBELs) by establishing the interim minimum level
(Interim ML) for mercury. The Interim ML is calculated when a method-
specific ML does not exist. It is calculated by multiplying the MDL by
3.18. The factor of 3.18 is derived from the ACS definition of level of
quantitation (LOQ) that is 10 standard deviation above the average blank
signal and is divided by the 3.14 (student t value) for the MDL, i.e.
3.18=10/3.14t for n=7. The calculated ML is then rounded up to 1, 2, 5,
10, 20, 50, etc. The Interim ML for mercury would then range from 5 to
20 ppt depending on the water matrix and the laboratories skill. The
interim ML for mercury concentrations would be 10 to 20 times higher
than most ambient concentrations found in natural waters. If the risk
assessment is used to establish an ML, it is conceivable that the ML
will fall below the ambient level for mercury and below the MDL for this
method.
ACKNOWLEDGEMENTS
The following individuals are acknowledged for their contributions to
this project: Professor Peter Stockwell, Paul Stockwell and Dr. Warren
Corns, (P.S. Analytical Ltd., Kent, UK) and Jim Coates (Questron
Corporation, Princeton, NJ) are thanked for their technical support.
Dr. Ron Jones (Florida International University, Miami, FL) is
gratefully acknowledged for providing the surface water sample from the
Everglades National Park, Florida and for providing method detection
limits.
REFERENCES
1. E. Schulek, K. Burger, Talanta, 1-2, 219, (1958).
2. B.J. Farey, L.A. Nelson, M.G. Rolph, Analyst, 103,656,(1978).
349
-------
3. L.A. Nelson, Anal. Chem., 51, 13, 2289,(1979)
4. W.R. Hatch, W.L. Ott, Anal. Chem. 40, 14, 2085, (1968).
5. W.T. Corns, L.E. Ebdon, S.J. Hill, P.B. Stockwell, Analyst, 177, 717,
(1992).
6. K.C. Thompson, G.D. Reynolds, Analyst, 96, 771,(1971).
7. K.C. Thompson, R.G. Godden, Analyst, 100, 544, (1975).
9. P.B. Stockwell, R.G. Godden, J. Anal. At. Spec, 4, 301,(1989).
10. S.N. Deming, S.L. Morgan; "Experimental Design: A Chemometric Approach",
Elsevier, Amsterdam, 1987.
11. Yorkshire Water Methods of Analysis, 5th Ed. 1988.
(ISBN 0 905057 23 6)
350
-------
53
DETERMINATION OF SELECTED METALS BY
PORTABLE X-RAY FLUORESCENCE (XRF) SPECTROSCOPY
Victoria Taylor. Region 9 Environmental Services Assistance Team, ICF Kaiser, San Francisco,
California, 94105
ABSTRACT
X-ray fluorescence (XRF) spectroscopy provides a means for profiling heavy metals contamination
quickly through the rapid screening of environmental samples. A study was conducted to estimate the
reliability of analytical results generated by XRF for use in site characterization. The purpose of this
study was to evaluate the effectiveness of XRF spectroscopy as a field screening tool for the
determination of metal and to present field samplers with a technique to reduce the number of fixed
laboratory sample analyses and to shorten the time required to obtain usable results. A portable XRF,
X-MET 920 (Outokumpu Electronics, Inc.), was used to analyze soil samples collected from two
Superfund sites for mercury, lead and zinc. The effects of sample preparation and instrument
standardization techniques on data reliability were tested by comparing XRF results to fixed laboratory
results generated for the same soil samples. The variable parameters evaluated were the physical
sample matrix and the type of materials used to prepare instrument calibration standards. Samples
were analyzed directly and after drying and fractionation to a uniform particle size. These studies were
designed to test the effects of the sample matrix on method performance. Two types of calibration
standards were used to evaluate the effects resulting from the use of different instrument calibration
materials: 1) field samples that were previously analyzed by a fixed laboratory to establish contaminant
levels, and 2) site background samples into which known quantities of target elements, both as salts and
aqueous solutions, were added. The results of this study show that XRF screening may be used to
expedite field operations by reducing the number of samples that require fixed laboratory analysis, and
to provide reliable concentration profiles for metals contamination in soil.
INTRODUCTION
XRF analysis is based on the interaction of a characteristic X-ray with elements in the sample. The
incident or primary X-ray from the probe source has sufficient energy to eject an inner shell electron
from the target element. The loss of the inner shell electron creates an opportunity for an outer shell,
higher energy, electron to fall into the inner shell vacancy, with loss of energy. This loss of energy is
emitted as an X-ray and is detected by the high resolution S(Li) detector in the X-Met 920 Surface
Analysis Probe. This process is presented schematically in Figure 1.
For this study, soil samples were obtained from two different site investigations. Twenty four samples,
collected as part of a remedial investigation to determine the extent of mercury contamination from mill
tailings piles, were analyzed for mercury by XRF and by an EPA Contract Laboratory Program (CLP)
laboratory by cold vapor atomic absorption spectroscopy (CVAAS). The samples from the mercury
investigation were used to perform initial tests to determine the acceptability of different techniques of
standardization. A second set of 37 soil samples was collected as part of a removal project for which
the target analytes were lead and zinc. For this project, the XRF was used in the field to analyze
samples during the removal activity. All of the samples from the removal project were analyzed by
two CLP laboratories using Inductively Coupled Plasma (ICP) Emission Spectroscopy and by the field
laboratory using XRF.
351
-------
EXPERIMENTAL
XRF Sample Preparation: Samples for XRF determination of lead and zinc were prepared by drying
and then sieving through a 100 mesh screen. The samples for the mercury investigation were prepared
by sieving without drying prior to both CVAAS and XRF analysis. The samples analyzed for lead and
zinc were submitted to one CLP laboratory sieved and dried, and to the second CLP laboratory as
collected, with the final concentration adjusted for sample percent moisture.
Calibration Sample Preparation: Three techniques were used to obtain calibration standards for
mercury XRF analysis. Background samples, which were found to contain no analytes above the XRF
limit of detection, were fortified with the elements of interest by the addition of either aqueous
solutions or elemental salts. The third type of calibration standard was obtained by selection of field
samples which contained an acceptable concentration range of target analytes as determined by fixed
laboratory analysis. These samples were treated as calibration standards and the ICP assay was
accepted as the true concentration. Calibration samples for the zinc and lead analyses were prepared
by fortification of a site background sample with aqueous standards.
Calibration: Individual sample curves were prepared for each metal using the procedures contained in
the Outokumpu software. After spectra for each calibration level were collected, a regression equation
was selected from the possible combinations from the terms contained in the multivariate regression
equation:
- _ -. * fl>
where fj = J.
§^:a
fj = Ij / BS
fj = Ij * Ij I BS
f,. = i / BS
Several terms may be selected to account for interelement interferences. Acceptable calibration models
were selected based on the correlation coefficient and the difference between the standard
concentrations determined from the regression line and the true values.
RESULTS AND DISCUSSION
Mercury Determination by XRF: Mercury calibration curves were prepared from a mill tailings pile
sample which was below the XRF limit of detection. Nine 3 gram aliquoits of this sample were
fortified with mercury from an aqueous atomic absorption standard so that the calibration samples
covered the range of 30 to 1600 ppm. A second set of calibration standards were prepared by
fortification of aliquoits of the same field sample with arsenic trioxide, resulting in calibration samples
with a concentration range of 100 to 15000 ppm of mercury. The third set of calibration standards
were obtained from actual field samples, analyzed by CVAAS. The CVAAS results were accepted as
the true value and the concentration range of this set of calibration samples was 51 to 1650 ppm. An
352
-------
unfortified aliquot of the sample was used along with the other fortified samples to construct the
calibration curve. The following correlation coefficients were obtained for the calibration models
constructed based on the mercury response obtained from each of the calibration sets:
Type of Calibration Sample
Liquid Fortified
As2O3 Fortified
Field Samples
Correlation Coefficient
0.998
1.000
0.981
Each standard preparation method appears to produce equivalent linearity. The regression equation
used for these calibrations contained one term, with the exception of the As2O3 fortified samples.
Because of the increased calibration range compared to the ranges obtained with the other calibration
techniques, a cross product term was used to account for the nonlinearity encountered at the higher
concentrations. The use of a multiterm calibration model did not appear to introduce any bias in the
XRF results compared to the fixed laboratory results. Table 1 presents the results of mercury
determination using different calibration methods in selected samples:
Table 1: Mercury Results Obtained Under Three XRF Calibration Methods and CVAAS
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
CVAAS
1950
150
960
790
41
0.57
110
650
260
0.22
620
310
77
1230
750
1480
124
110
840
51
XRF: Liquid
Fortification
Calibration
1970
210
1.2
1980
110
2.6
100
2820
380
16
450
300
69
2560
1750
830
334
180
640
72
XRF: Solid
Fortification Calibration
550
190
200
100
1280
440
45
630
88
1090
630
255
400
-----
XRF: Field Sample
Calibration
550
240
28
100
87
1330
0
350
70
900
260
n
4
3
4
2
4
2
4
4
3
4
4-
2
4
2
4
3
4
2
3
2
Coefficient of
Variation
49
7
41
21
35
32
6
41
9
106
20
0.5
10
18
29
15
15
12
18
8
353
-------
Sample
Number
21
22
23
24
CVAAS
9.8
800
650
17
XRF: Liquid
Fortification
Calibration
990
1600
1160
87
XRF: Solid
Fortification Calibration
1080
XRF: Field Sample
Calibration
560
850
n
3
4
2
2
Coefficient of
Variation
41
25
14
16
The results listed above are presented as a bar graph in Figure 2. There are two results which appear
questionable. In the case of samples 3 and 21, the CVAAS results and each of the XRF analyses differ
by an order of magnitude or more. The source of this difference may be sample nonhomogenaity, or
may be the result of an error in sample identification by one or both of the laboratories. Because of
this large discrepancy, the results for samples 3 and 21 were not included in the comparison. For each
of the calibration methods, the XRF results were plotted against the CVAAS results and the linear
regression equation calculated. These plots are presented in Figures 3 through 5. For the regression
plots, the log value of concentration was used to maintain a reasonable scale.
Lead and Zinc Determination by XRF: Table 2 presents the results for lead determined by XRF and
CLP with the coefficient of variation:
Table 2: Lead Results Obtained by XRF and CLP Analysis
Sample Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
XRF Lead ppm
6830
5260
2450
1200
430
730
2090
3920
1420
15000
3420
590
1020
1750
6920
290
300
3660
CLP1 Lead ppm
2450
3410
1080
1010
310
550
1600
1720
570
13000
2450
1620
730
1030
3300
120
100
3220
CLP2 Lead ppm
330
2460
900
260
190
100
900
1170
420
9140
6970
1060
500
530
4020
110
80
4420
3£CV
84
31
47
49
32
58
32
52
55
20
45
38
29
45
33
46
62
13
354
-------
Sample Number
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
XRF Lead ppm
1510
1100
1310
1510
1880
3360
5620
9000
6300
7630
1200
5422
1470
6740
480
150
1430
5850
5390
CLP1 Lead ppm
620
540
920
720
700
1750
3170
2770
3560
3440
600
4420
920
1620
300
70
830
2440
2470
CLP2 Lead ppm
580
500
1230
350
27
1080
2540
6140
2370
4610
370
3240
490
1520
230
40
580
2310
640
%CV
48
38
15
56
88
46
35
43
40
34
48
20
42
74
31
54
38
46
69
The lead results are presented as a bar plot in Figure 6. The Table 3 presents the nineteen comparable
zinc results obtained by XRF analysis and ICP analysis following EPA CLP Statement of Work (SOW)
protocols.
Table 3: Zinc Results Obtained by XRF and CLP Analysis
Sample
1
2
3
4
5
6
7
8
9
10
XRF Zinc ppm
160
180
630
79
170
73
120
65
2310
150
CLP1 Zinc ppm
100
200
550
230
170
98
96
190
1060
260
CLP2 Zinc ppm
79
130
480
190
110
110
86
150
770
180
%CV
32
16
11
39
20
16
16
39
48
23
355
-------
Sample
11
12
13
14
15
16
17
18
19
XRF Zinc ppra
71
73
480
110
260
110
410
1210
1250
CLP1 Zinc ppm
110
92
320
210
410
170
400
760
820
CLP2 Zinc ppm
100
81
570
140
340
57
320
861
310
%CV
18
10
23
25
18
41
10
20
48
The zinc results are presented as a bar graph in Figure 7. The 37 lead and 19 zinc results reported by
each of the laboratories were used to prepare linear regression plots to determine the level of
correlation. As in the case of the mercury regression plots, a log scale was used for the graphical
presentation. Figures 8 and 9 present the results of regression plots for lead and zinc. The extent of
correlation between data sets may be estimated by the correlation coefficient, where an r value of 1
represents perfect correlation. For mercury, each of the methods of standard preparation appear to
produce data which are correlated with the analysis performed by a CLP laboratory using CVAAS.
The samples analyzed for lead and zinc were analyzed by two different CLP laboratories. In the case
of the first CLP laboratory (CLP1), the samples were dried and sieved in a manner identical to the
XRF preparation method. In the case of the second CLP laboratory (CLP2), the samples were
analyzed on a as received basis and the final result adjusted for sample percent moisture. The
effectiveness of drying and sieving samples prior to XRF analysis reduces measurement variations due
to random X-ray scattering due to uneven sample particle size distribution. The fractionation of
samples does not appear to significantly effect the ICP results, since the CLP1 results are very similar
to the CLP2 results, although the correlation with the XRF results is higher for CLP1 than CLP2.
In the case of mercury analysis, no systematic bias was observed between the results obtained from
XRF or CVAAS, and no one method of calibration standard preparation appears to produce superior
correlation. In the case of lead and zinc analysis, the XRF results are biased high with respect to the
ICP results from both laboratories, most significantly with respect to the lead analyses. In the case of
both the regression plots for lead and zinc, the slope of the regression line is less than 1, confirming a
trend toward high bias on the part of the XRF.
Use of a X-Ray fluorescent spectrometer provides a means for site investigations to obtain high density
profiles of site contamination very rapidly. During the removal investigation a total of 125 samples
were analyzed over a period of three days and individual sample results were available within a few
hours from collection.
Based on the comparison of these data sets, the XRF technique provides information which is in good
agreement with fixed laboratory ICP results. Successful calibration samples may be prepared by
fortification of site background samples with liquid standards or solid reagents, or by analysis of
independently assayed field samples.
Acknowledgements: Special thanks to Beiyi Chen for the experimental work on the XRF.
356
-------
REFERENCES
1. Kuharic, C.A.; Cole, W.H. "An X-Ray Fluorescence Survey of Lead Contaminated Residential
Soils in Leadville, Colorado: A Case Study": Environmental Monitoring Systems Laboratory, Office
of Research and Development, U.S. Environmental Protection Agency, Las Vegas, NV. 1993,
EPA/600/R-93/073.
2. Helland, B.; McCall, W. Field Screening Methods for Hazardous Wastes and Toxic Chemicals,
Proceedings of the International Specialty Conference, Las Vegas, NV. 1993.
3. Outokumpu Instruments, X-Met 920, ver. 1.3, Reference Manual, August, 1992.
357
-------
Figure 1: Production of X-Rays
fluorescent
X-Ray
ejected electron (e1)
incident X-ray
Electron
Energy
inner shell
transition
Figure 2: Comparison of Mercury Results From CVAAS and
XRF
3,000
2,500
2,000
1,500
1,000
500
0
I
Sample Number
I CLP CVAAS • XRF-Liquid Fortification
I XRF- Solid Fortification • XRF-Field Sample Calib.
L
358
-------
Figure 3: XRF Mercury Results using Liquid Standards vs
CVAAS Results
u.
OL
X
n
3
i
E
Q.
O.
s
IU.UUU
3,000
1,000
300
100
30
10
3
1
i 9
€ •
i •
: •
*• **
*. . " »
-
: r= 0.73
i i i i
0.1 1 10 100 1,000
Log ppm Mercury by CVAAS
10,000
Figure 4: XRF Mercury Results using Salt-Fortified Standards
vs CVAAS Results
LL
X
JO
|
0>
S
E
Q.
Q.
O)
O
J,UUU
1,000
300
100
30
10
3
i
•_
•
• •
• •
• •
•
•
. •
-
.-
r=0.52
i i . i i . i . ^^^j , —
3 10 30 100 300 1,000 3,000
Log ppm Mercury by CVAAS
359
-------
Figure 5: XRF Mercury Results using Field Sample
Calibration Standards vs CVAAS Results
u.
OL
X
>,
.Q
3
i_
0)
S
E
Q.
Q.
0)
O
— 1
o.uuu
1,000
300
100
30
10
3
•i
r %
:
*
*
' . '
-
r
-
r=0.60
1 1 , 1 1 , 1 1
3 10 30 100 300 1,000 3,000
Log ppm Mercury by CVAAS
Figure 6: Comparison of Lead Results From Three
Determinations
d>
Q.
a
16,000
14,000
12,000
10,000
8,000
6,000
4,000
2,000
0
UkllL
LllL
1357 91113151719212325272931333537
Sample Number
• XRF i
ICLP1
• CLP2
360
-------
Figure 7: Zinc Results From Three Determinations
1 3 5 7 9 11 13 15 17 19
2 4 6 8 10 12 14 16 18
Sample Number
I XRF H CLP1 • CLP2
Figure 8: Lead Regression Plot
a.
_i
o
.Q
IQ
0)
_l
E
Q.
O>
0
_l
10,000
3,000
1,000
300
100
30
in
~ B _
: ffl • H :
.• ^r-
- • • t. •• •
• ^* * " :
: 9
« • -
1 ,,!,..,! 1 . . 1 , , , , 1
ou,uuu ^
10,000 rJ
o
3,000 -a
1,000 S
— '
300 |
a.
100 o,
o
^n J
"100 500 2,000 10,000
200 1,000 5,000 20,000
Log ppm Lead by XRF
CLP1:r=0.88 CLP2: r=0.79
361
-------
Figure 9: Zinc Regression Plot
IT
0.
o
.a
o
c
N
Q.
Q.
TO
0
1 u
1,000
500
200
100
50
on
•
; *
~ B
• •
" • • II
•
i "t r
'_ *
: , , , 1 1 1 , , , , 1 1 .
£.\J
CM
a.
10o
^
a
c U
5 c
N
3 1
„ a
2 TO
o
•1
50 100 200 500 1,000 2,000
Log ppm Zinc by XRF
5,000
CLP1:r=0.95 CLP2: r=0.81
362
-------
54
Improving Mercury Detection Limits Using a Dedicated Flow Injection System
S. Mclntosh, S. Sauerhoff, and Z. Grosser
The Perkin-Elmer Corporation', 50 Danbury Rd, Wilton, CT 06897-0259
Traditionally the determination of mercury in environmental samples has been performed using a cold
vapor atomic absorption technique. An atomic absorption instrument equipped with a vapor generation
accessory provides mercury detection limits ranging from 50-200 ng/L. This configuration has performed
satisfactorily for current EPA requirements. However, as regulatory and environmental pressures force
mercury detection limits lower, there is an increasing need for improved instrumentation.
A dedicated mercury system can be optimized for improved detection limits and sample throughput. The
Perkin-Elmer Flow Injection Mercury System (FIMS) was characterized and evaluated for the
determination of mercury in a variety of water samples.
Experimental
The FIMS optical system consists of a low pressure mercury source, an absorption cell widi removable
quartz windows and a solar blind detector with maximum sensitivity at 254 nm. The system is automated
by using flow injection techniques to add the HC1 acid carrier and SnCl2 reductant. The reaction mixture
then passes through a gas-liquid separator, which directs the mercury vapor to the absorption cell where the
measurement takes place. To minimize moisture transfer to the absorption cell the gas-liquid separator is
equipped with a PTFE membrane.
Samples were prepared as specified in EPA Method 245.1. The method has been modified to include flow
injection, which has been approved by the EPA under the Alternate Test Procedure (ATP) process. All
acids and reagents used for sample preparation and analysis were of "ultrapure" or "mercury-free" grade
To minimize contamination, all glassware and digestion vessels were cleaned and soaked for 24 hours in
1:1 nitric acid solution.
The flow injection (FI) program is shown in Figure 1. The first step, shown as the Prefill step, was used
only for the first reading in a series of replicates. This step ensures that the tubing leading from the
autosampler to the sample loop is adequately filled with solution. The next step, Step 1, is used for all
replicates. In this step the sample loop, located on the flow injection valve, is filled with solution. In the
final step, Step 2, the sample is transported from the sample loop by the HC1 carrier to the mixing
manifold, where the sample is merged with the reductant. At this point the instrument read step is activated
and the absorbance signal is measured.
AAWInLab Analyst
[Jig £ilil Jpola Analyses Qplloits ffilnUow
m \s ra
hi ts
MelhgdEdllnr.HG245.tti
Onmallon
FIAS Plliqrom
Sinn
HceJ( ffl/MKl _
" inn r («» i o
tnu r (ft i O
inn r O ' Q
a r
-------
Data for all determinations were collected using peak height.
Results and Discussion
Figure 2 shows a peak profile of a water sample spiked with 0.1 u,g/L Hg, demonstrating the excellent
signal to noise ratio of the system.
0.003 i
0.002
Abs.
0.001 -I
Determination of Hg in Water
Concentration: 0.1 pg/L
Sample volume: 0.5 ml
10
Time (s)
20
Figure 2. Low Level Water Spike
As part of this study, instrument and method detection limits were measured for typical EPA methods. The
instrument detection limit (IDL) for mercury was measured using a procedure outlined in the EPA
CERCLA statement of work ILM02. The instrument response for seven replicate analyses of a low-level
mercury solution (20 ng/L in this case), conducted on three nonconsecutive days was measured. The
standard deviation for each of these analyses was multiplied by the t-distribution value and resulted in a
valueof4ng/L.Hg.
The method detection limit (MDL) for mercury detection limit by EPA method 245.1 was determined
following the procedure outlined in the Code of Federal Regulations (40CFR), part 136. Seven individual
aqueous standards were carried through the complete sample dissolution procedure described in the
method. While the mercury contamination in the original reagents was only at ng/L levels it was the
limiting factor in the MDL. The MDL was established at 9 ng/L. Both the IDL and MDL represent a
significant improvement in detection limits over the conventional cold vapor system of 10-50 times.
Table I shows the performance of the method in determining mercury in a variety of water matrices. The
only certified reference material available is NIST 1641C, which was diluted to 0.2 ug/L, prior to analysis.
AJso examined were spiked samples of groundwater, drinking water, and SLRS-2 Riverine water, obtained
from the National Research Council of Canada. The recoveries demonstrate the F1MS ability to determine
trace levels of mercury.
364
-------
Table I
Analytical Results
Sample
Ground Water spike - 1 u.g/L Hg
Drinking Water spike - 1 ug/L Hg
NIST 1641c, diluted to 0.2 ug/L Hg
SLRS-2
SLRS-2 (Duplicate)
SLRS - spiked - 1 ug/L Hg
Hg Found (ug/L)
1.00 (±0.01)
1.00 (±0.02)
0.19 (±0.01)
0.13 (±0.01)
0.13 (±0.01)
1.17 (±0.02)
The system was characterized for a number of variables including the ability to vary detection limits and
throughput by varying the size of the sample loop. The sample loop volume can be varied between 40 uL
and greater than one mL. Larger sample volumes improve the detection limits and smaller volumes
increase the sample throughput. The optimum compromise for this analysis was found to be a sample
volume of 500 uL. The detection limits were as noted and the sample throughput was 120 determinations
per hour.
The stability of the system was evaluated and found to be approximately 4%RSD over the course of
several hours, for a solution containing 0.1 ug/L Hg.
One of the many advantages that flow injection has over the continuous flow technique includes reduced
carryover effects. With flow injection, the continuous rinsing action of the carrier stream reduces carry
over from samples containing higher concentrations. The analysis of a 100 ug/L Hg standard, followed
immediately by the analysis of the blank, shown in Figure 3, demonstrates the effectiveness of the rinsing
action.
• Flow Injection
EH Cont. Flow
100 0 0 0
Concentration
0
Figure 3. Carry-over Comparison
365
-------
Conclusion
The advantages of using a flow injection system with improved sensitivity, such as the FIMS, for the
determination of environmental mercury are many and include:
• Increased sample throughput for better lab productivity
• Reduced reagent consumption for less initial expense and expense for waste disposal
• Reduced sample consumption, important for critical small samples
• Reduced carry over for better accuracy with a variety of samples
• Improved performance providing ultra-trace Hg detection limits
The determination of mercury in environmental samples can be performed by the FIMS on a routine basis
at lower levels than previously possible with this technology. The limiting factor for Hg determinations in
environmental samples has become the integrity of the reagents rather than the detectability of the
instrumentation.
366
-------
55
RAPID MICRO DISTILLATION OF TOTAL CYANIDE USING LIGAND
DISPLACEMENT AND DETERMINATION BY FLOW INJECTON ANALYSIS
William R. Prokopy. Applications Chemist, Lachat Instruments, 6646 W. Mill Rd.,
Milwaukee, Wisconsin 53218, Maureen Stone, Lachat Instruments, 6645 W. Mill Rd.,
Milwaukee, Wisconsin 53218
ABSTRACT
In wastewater treatment procedures throughout various industries, cyanide must be
monitored as it pollutes the environment and is a significant health hazard. To
accurately monitor cyanide, a distillation procedure is required to dissociate metal
cyanide complexes and subsequent determination. The classical strong acid macro
reflux distillation (EPA Method 335.2) has been in use for many years but does not
eliminate interferences such as thiocyanate and sulfide, and is time consuming. Also,
the EPA approved cyanide determination step uses dihydrogen phosphate as a buffer
that does not have the capacity to effectively control changes in pH.
In this presentation we will introduce a micro steam distillation utilizing the weak acid
ligand displacement procedure and determination with Flow Injection Analysis. The
distillation is complete in forty minutes. The cyanide in the distillate is buffered with
an acetate buffer and reacted with chloramine-T where cyanogen chloride forms a
complex with 1,3 - dimethylbarbituric acid. Method support data including
interferences, precision, accuracy and method detection limit will be presented.
367
-------
56
ADVANCES IN QUALITY ASSURANCE FOR XRP DETERMINATION OF LEAD.
H. A Vincent, U. S. Environmental Protection Agency, EMSL-LV,
P.O. Box 93478, Las Vegas, Nevada 89193-3478, and J. E. Kilduff,
Lockheed Environmental Sciences and Technologies, 980 Kelly
Johnson Drive, Las Vegas, Nevada 89119.
ABSTRACT
The EPA's Environmental Monitoring Systems Laboratory at Las
Vegas (EMSL-LV), has been studying the application of X-ray
fluorescence analysis (XRF) to perform lead determinations
in a number of different types of sample media. XRF techniques
have some unique advantages in applications to solid samples but
do present some problems.
XRF lead determinations in paints, soils, and dusts can be
made without sample dissolution and often can be done with
minimal sample preparation. Prior XRF studies show difficulties
in sample-related biases, in false negative or positive
reporting, in improper sampling, in lack of traceability to
standards, and in calibration for nonhomogeneous samples. For
many EPA applications, data from many laboratories must be
comparable and must meet both regulatory and technical
requirements.
The XRF studies in this project have focused on examining
problems for both laboratory-based and portable XRF measurements
for lead. XRF experiments were done using external reference
materials, but focused primarily on gold. The gold reference
material was placed in the optical pathway of the excitation
source but anterior to the sample. Both characteristic L-series
and K-series lead X-ray lines were studied.
The results reveal some of the difficulties with current
data handling systems that correct XRF intensities for various
interferences. It was found that an external reference can have
value as part of a quality control system. The reporting of false
negatives for lead, when the characteristic lead L-series X-rays
are used, can be avoided. Similarly, when the lead K-series X-
rays are used, false positives can be avoided. Experiments at
EMSL-LV show that absorption values for external reference gold
X-rays by various sample constituents can be used to adjust lead
x-ray intensities affected by the presence of those constituents.
This kind of information helps in describing analytical models
for the different sample types.
368
-------
57
THE STABILITY OF CALIBRATION STANDARDS FOR ICP/AES ANALYSIS:
A TWO-YEAR STUDY*
Doris R. Huff. Scientific Associate, and Edmund A. Huff, Chemist, Analytical Chemistry
Laboratory, Chemical Technology Division, Argonne National Laboratory, 9700 S. Cass
Avenue, Argonne, Illinois 60439-4831
ABSTRACT
The stability of instrument calibration standards for Inductively Coupled Plasma/Atomic
Emission Spectrometric analysis was studied over a two-year period. Data were obtained as
functions of analyte concentration, acid type, and acidity. The impact of acid concentration on
signal-to-background ratios was also assessed. The results show that, with the appropriate
choice of inorganic acid preservatives, most analytes maintain their integrity over extended
periods; thus frequent standard preparations are not necessary to obtain valid analytical data.
This conclusion allows for more efficient use of commercially purchased reference materials,
which should reduce procurement costs and minimize chemical waste.
INTRODUCTION
Inductively Coupled Plasma/Atomic Emission Spectrometry (ICP/AES) is used extensively in
the Analytical Chemistry Laboratory (ACL) to characterize diverse analytical samples. This
method simultaneously measures the concentrations of multiple cations in solutions. The
working standards used for instrument calibration are generally in the 5-20 pg/mL range. They
are prepared from certified single or multi-element stock solutions by serial dilutions and are
preserved with inorganic acids. The accuracy and precision of analytical measurements are
greatly affected by uncertainties in standard stabilities as functions of preservative type, analyte
content, and acid concentration. Consequently, a systematic study was initiated to assess the
effect of these parameters on data quality over time. The results have implications regarding
standard procurement needs and waste minimization.
EXPERIMENTAL
APPARATUS
The ICP/AES measurements were performed on a spectrometer system that incorporated a 48-
channel polychromator and a computer-controlled scanning monochromator (Instruments S. A.,
Inc., Edison, NJ). Both instruments were focused on a single plasma excitation source. Table
1 lists the instrumentation, and Table 2 details the operating conditions for this two-year study.
Typical experimental detection limits (DLs) and their corresponding wavelengths are given in
Table 3.
*Work submitted by the U.S. Department of Energy under Contract W-31-109-Eng.38.
369
-------
REAGENTS
Instra-Analyzed HC1 and HNO3 (Baker Chemical Co.) were used in blank and standard
preparations. Water for dilutions (>18 Mohm-cm) was obtained from a Sybron/Barnstead ion
exchange system. Certified standard stock solutions were purchased from the National Institute
of Standards and Technology (NIST), SPEX Industries, and Baker Chemical Co. Test solutions
and calibration standards were prepared by making serial dilutions of the concentrates and
adjusting to the appropriate acidity. The information on the standard test solutions is
summarized in Table 4.
PROCEDURES
The plasma system was optimized with respect to maximum signal-to-background (S/B) ratio
as a function of the vertical viewing aperture. The instrument was calibrated at this position
by using the two-point calibration procedure with standards and blanks prepared in a 2% acid
(HC1 or HNO3) medium. Working standards were prepared on the day of the analytical run
and verified against previously used calibration solutions. An agreement of ±3% for this
comparative analysis was considered to provide a valid calibration curve for the subsequent
analytical run.
Since instrument calibration standards were prepared at a single acid concentration (2%),
information on the effect of acidity on signal intensity was needed for data assessment of the
diverse analytical matrices. Therefore, S/B ratios were determined as a function of HC1
concentration, using a 1-pg/mL multielement standard and a corresponding blank.
The stability of standards is also affected by the possible loss of analyte by diffusion and/or
evaporation through plastic containers. Since stock and working standards are stored in
polyethylene bottles, diffusion was studied by weight-loss measurements from high density
polyethylene (HDPE) and Teflon containers. Bottles were filled with acid solutions, and the
change in weight as a function of time was established gravimetrically.
RESULTS AND DISCUSSION
SIGNAL-TO-BACKGROUND RATIOS
One objective of this investigation was to determine the stability of standards as a function of
acidity. The analyte response as a function of this parameter is of interest because instrument
calibration in our laboratory is generally performed at a 2% acid concentration. Table 5
summarizes S/B ratios for 18 elements in HC1 solutions. The data indicate that the presence
of mineral acids decreases the emission intensities of atomic and ionic lines for most elements,
which is consistent with published results. (1) It has been postulated (cited publication) that
these findings are related to the physical state of the plasma at lower acidities (<1M) and to the
reduced rate of sample uptake at higher acid concentrations (>1M). A few, very sensitive
elements (i.e., Be, Ca, Mg, and Sr) exhibit an increase in S/B ratios, an observation inconsistent
with the above mechanism. The results show that for best accuracy, calibration standards,
370
-------
blanks, and samples should be matched with respect to acid concentration. Furthermore, for
the purpose of this study, only relative analyte changes should be considered in stability
assessments, since instrument calibration was performed at a single acid concentration.
DIFFUSION
Diffusion data for a nine-month period were presented in an earlier report of this study. (2)
Continued measurements (>two years) showed these losses to be a function of container type
and acid preservative. Solutions in 2% HNO3 showed the lowest weight loss in HDPE
containers (0.5%). Decreases in weight of 0.7% (5% HC1-1% HNO3) and 0.9% (2% HC1) are
considered to be acceptable for the prolonged storage of ICP/AES standards. Solutions
prepared in Teflon bottles maintained their weights best in 5% HC1-1% HNO3 (0.2% loss).
Both 2% HC1 and 2% HNO3 showed a diffusion-related decrease of 0.6% and 1.1%,
respectively. This controlled study did not address variables pertaining to repeated sampling
of bottles and diffusion effects from partially filled containers. The data do, however, suggest
that within the accuracy range of ICP/AES (3-10%), changes via these mechanisms would still
provide valid analytical data.
STABILITY DATA
Tables 6 and 7 summarize representative stability results in 2% nitric acid." The data were
obtained from 25 independent measurements over a two-year period. None of the listed
analytes, except Sn, show excessive fluctuations at any concentration in this preservative.
Similar data sets were obtained with all the other preservatives studied.
As expected, deviations from the mean decrease with increasing analyte content, consistent with
more favorable statistics at higher signal intensities. Representative data are shown graphically
in Figures 1 and 2. A slight, positive trend in the Cu concentration as a function of time
(Figure 1) is apparent; this is probably caused by diffusion losses from progressively smaller
sample volumes. Similar behavior was observed for other analytes. The Pb values (Figure 2)
remained constant, irrespective of analyte concentration, within the 3-10% stipulated range.
Standards of Nb and W were studied only in HC1 and HC1-HF acids, since HNO3 is known to
cause precipitation. Figure 3 shows the behavior of W as a function of time and acid
composition. A similar trend was noted for Nb. These results verify that HF is required to
stabilize these elements in calibration standards and ensure longer shelf lives.
At a low analyte concentration (0.2 pg/mL), elements with relatively high DLs (Figure 4) have
relative standard deviations (RSDs) exceeding 10%, a result that can be correlated with greater
signal fluctuations. For comparison, an analyte with a low DL (more favorable statistics) is
shown in Figure 5.
The unexpected, reasonably good stability of Sn in HNO3 can be explained by the presence
of HC1 in the stock standard used to prepare the test mixtures. Separate Sn standards prepared
in HNO3 alone deteriorated severely within one week, presumably through the precipitation of
371
-------
stannic acid. The stability of Ag standards was studied only in HNO3 and HC1-HN03
preservatives, with satisfactory results. The deterioration of Ag solutions can usually be
attributed to the introduction of chloride ions during washout procedures.
During the study, it was noted that the Sb and Si results were increasing with time. Closer
examination of the stock solutions(2% HC1) used to prepare working standards revealed that
precipitates were forming, causing the Si and Sb concentrations in solution to be lower. This
resulted in apparently higher values for Si and Sb in the test standards. New stock solutions
were used for the working standards in subsequent measurements, and the Sb and Si values
for the test matrices returned to an acceptable range. In retrospect, the stability of Si standards
should have been examined in the presence of HF, a medium known to promote complexation
reactions and thus avoid precipitation due to hydrolysis.
CONCLUSION
The results on the stability of calibration standards for ICP/AES analysis show that a suitable
acid preservative enables analyte solutions (5-20 pg/mL) to maintain their integrity over a
two-year period. For the wide range of elements studied, a mixture of 5% HC1-1% HNO3,
suggested by the U.S. EPA Contract Laboratory Program (3), appears to provide the best
compromise for acceptable shelf lives for most cations. However, for those situations where
the relatively high acidities are impractical or undesirable, 2% HC1 or 2% HNO3 stabilizes
analytes satisfactorily, provided any chemical incompatibilities (i.e., precipitation, hydrolysis)
are resolved.
The variations in S/B ratios as a function of acidity mandate that for best accuracy and
precision, calibration standards, blanks, and samples should be matrix-matched (i.e., same acid
type and concentration, same major constituents). However, for analytical data normally
reported in the accuracy range of 3-10%, approximate acid concentrations are sufficient for
satisfactory results.
Changes in analyte concentrations due to evaporation and diffusion through plastic containers
did not, over a reasonable time (one year), appear to make a significant contribution to
inaccuracies, compared to the more pronounced effects introduced by the excitation and sample
transport system. A trend toward increasing concentration was noted over the second year,
although the variation was within the expected accuracy range. Partially filled bottles, frequent
sampling, and storage conditions could have a considerably larger impact on standard
stabilities than those observed here.
This study clearly shows that the frequency of calibration standard preparations could be
reduced significantly for ICP/AES analyses at moderate accuracies (3-10%). A complementary
investigation at this laboratory on the stability of standards used for graphite furnace atomic
absorption (GFAA) analyses concluded that even lower analyte concentrations (<0.25 pg/mL)
are stable for at least a nine-month period. (4) Thus, mandated standard preparation
requirements for environmental sample analyses could be relaxed, which would result in lower
standard procurement costs, waste minimization, and better allocation of analytical effort.
372
-------
Table 1. Instrumentation
Spectrometers
Polychromator
Monochromator
Instruments S. A. Inc., Model J-Y48P 1-m Paschen-Runge with
2550 grooves/mm holographic concave grating, 20-pm entrance
and 50-pm exit slits; Hamamatsu R300 and R306 photo-
multiplier tubes, 48 channels.
Instruments S. A. Inc., Model J-Y38 1-m Czerny-Turner with
Spectra-Link controller, 2400 grooves/mm holographic plane
grating, variable entrance and exit slits; Hamamatsu R955
photomultiplier tube.
Power and Nebulizer System
Generator
Nebulizer
Torch
Spray Chamber
Computer System
Computer
Terminals
Software
Plasma-Therm Model HFP-2500, 27.12 MHz with 3-tum copper
load coil.
Instruments S. A., Inc. concentric with a Pt-Ir capillary tube.
Instruments S. A., Inc. demountable in a Mermet-Trassey
configuration.
Instruments S. A., Inc. double-barrel, constructed from Ryton.
Digital Equipment Corp. PDP-11/73 with 768K byte of
memory, three RL-02 disks.
Digital Equipment Corp. DEC Writer HI hard copy and DEC
VT340 video with graphics.
1. Instruments S. A., Inc.-supplied analytical program run by
the DEC RSX-11M operating system.
2. ANL-developed report generation and quality
assurance programs.
Table 2. Operating Conditions
Forward R.F. Power
Reflected Power
Observation Height
Integration Times
Polychromator
Monochromator
1.10 kW
<5 W
16mm above
load coil
10 sec
1 sec
Argon Flow Rates
Outer Gas
Intermediate Gas
Sheath Gas
Nebulizer Gas
Sample Uptake Rate
14 L/min
0.4 L/min
0.5 L/min
0.6 L/min
2.8 mL/min
373
-------
Table 3. Detection Limits
Element
Ag
Al
As
B
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Mg
Mn
Mo
Ni
Pb
Sb
Se
Si
Sn
Sr
Ti
Tl
V
Zn
Zr
Wavelength, nm
328.068
308.215
193.695
208.959
233.527
313.042
393.366
226.502
228.616
267.716
324.754
238.204
279.553
257.610
202.030
231.604
220.353
206.833
196.026
251.611
189.926
407.771
334.941
190.801
292.402
213.856
343.823
DL. \L\
1.9
38.0
52.0
11.0
1.3
0.4
1.3
3.4
36.0
4.6
2.1
2.1
1.1
1.5
14.0
9.3
90.0
54.0
58.0
14.0
30.0
0.5
2.2
24.0
2.9
0.5
2.7
Monochromator
Mb
W
309.418
224.875
8.9
7.9
Detection Limit (DL) = 3 x the standard deviation of the baseline noise.
374
-------
Table 4. Summary of Standard Solutions
File Name
QF01
QF02
QF03
QF04
QF05
QF06
QF07
QF08
QF09
QF10
QF11
QF12
QF13
QF14
QF15
QF16
QF17
QF18
QF19
QF20
QF21
QF22
QF23
QF24
QI01
QI02
QIC1
QIC2
QA01
QA02
QA03
QA04
QA05
QA06
QA07
QA08
QA09
QA10
Concentration,
Hg/mL
0.2
0.2
0.2
0.2
0.2
0.2
0.2
1.0
1.0
1.0
1.0
1.0
1.0
1.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
10.0
10.0
10.0
1:10 (HI.
l:10dil.
l:10dil.
1:10 da.
0.2
0.2
0.2
1.0
1.0
1.0
4.0
4.0
4.0
10.0
Elements
SpexAa
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
Spex A
ICV-lb
ICV-1
ICV-1
ICV-1
Ag
Ag
Ag
Ag
Ag
Ag
Ag
Ag
Ag
Ag
Acidity
1%HNO3 + 0.002% HC1
2% HNO3 + 0.002% HC1
5% HNO3 + 0.002% HC1
1% HC1
2% HC1
5% HC1
5% HC1 + 1% HNO3
1%HN03 + 0.01%HC1
2%HN03 + 0.01%HC1
5%HNO3 + 0.01%HC1
1% HC1
2% HC1
5% HC1
5% HC1 + 1% HNO3
1%HNO3 + 0.04%HC1
2% HNO3 + 0.04% HC1
5% HNO3 + 0.04% HC1
1% HC1
2% HC1
5% HC1
5%HC1+1%HNO3
2%HNO3 + 0.1%HC1
2% HC1
5%HC1+1%HNO3
2% HNO3
5%HC1+1%HNO3
2% HNO3
5%HC1+1%HNO3
1% HNO3
2% HNO3
5% HNO3
1% HNO3
2% HN03
5% HNO3
1%HNO3
2% HNO3
5% HNO3
2% HN03
a Spex A: Al,Ba,Be,Ca,Cd,Co,Cr,Cu,Fe,Mg,Mn,Ni,Pb,Sn,Sr,V,Zn,Zr
b -1: Ag,Al,Ba,Be,Ca,Cd,Co,Cr,Cu,Fe,Mg,Mn,Ni,Pb,V,Zn
375
-------
Table 4. Summary of Standard Solutions (cont'd)
File Name
QSil
QS12
QSi3
QSi4
QS15
QSi6
QSi7
QSi8
QS19
QAsl
QAs2
QAs3
QAs4
QAs5
QAs6
QSnl
QSn2
QSn3
QSn4
QSn5
QSn6
QW01
QW02
QW03
QW04
QW05
QW06
QNbl
QNb2
QNb3
QNM
QNb5
QNb6
Concentration,
Hg/mL
0.5; 1.0
0.5; 1.0
0.5; 1.0
2.5; 5.0
2.5; 5.0
2.5; 5.0
10.0; 20.0
10.0; 20.0
10.0; 20.0
5.0
5.0
5.0
20.0
20.0
20.0
5.0
5.0
5.0
20.0
20.0
20.0
1.0
1.0
5.0
5.0
20.0
20.0
1.0
1.0
5.0
5.0
20.0
20.0
Elements
B,Mo,Ti;Si
B, Mo, Ti; Si
B, Mo, Ti; Si
B,Mo,Ti;Si
B,Mo,Ti;Si
B, Mo, Ti; Si
B, Mo, Ti; Si
B, Mo, Ti; Si
B,Mo,Ti;Si
As, Sb, Se
As, Sb, Se
As, Sb, Se
As, Sb, Se
As, Sb, Se
As, Sb, Se
Sn
Sn
Sn
Sn
Sn
Sn
W
W
W
W
W
W
Nb
Mb
Nb
Nb
Nb
Nb
Acidity
2% HNO3
2% HC1
5%HC1+1%HNO3
2% HNO3
2% HC1
5% HC1 + 1% HNO3
2% HNO3
2% HC1
5%HC1+ 1%HNO3
2% HNO3
2% HC1
5%HC1+1%HNO3
2% HNO3
2% HC1
5%HC1+1%HNO3
2% HNO3
2% HC1
5%HC1+1%HN03
2% HNO3
2% HC1
5%HC1+ 1%HNO3
2% HC1
2% HC1 + 0.5% HF
2% HC1
2% HC1 + 0.5% HF
2% HC1
2% HC1 + 0.5% HF
2% HC1
2% HC1 + 0.5% HF
2% HC1
2% HC1 + 0.5% HF
2% HC1
2% HC1 + 0.5% HF
376
-------
Table 5. Signal-to-Background Ratio vs. Acidity
Signal-to-Background Ratio % Change
Element
Al
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Mg
Mn
Ni
Pb
Sn
Sr
V
Zn
Zr
1%HC1
1.63
6.50
83.4
14.58
7.75
2.99
4.36
5.38
5.02
30.1
17.75
2.69
1.57
1.68
63.8
4.23
18.97
4.21
2% HCI
1.64
6.44
84.5
15.47
7.65
2.98
4.30
5.29
4.86
29.9
17.57
2.71
1.54
1.65
64.8
4.21
18.90
4.21
5% HCI
1.70
6.35
88.4
19.13
7.60
2.96
4.28
5.31
4.80
38.7
17.45
2.70
1.56
1.61
68.9
4.19
19.17
4.26
10% HCI
1.60
6.20
90.7
18.86
7.43
2.94
4.25
5.23
4.69
41.2
17.05
2.65
1.53
1.67
70.3
4.12
18.33
4.19
20% HCI
1.60
6.04
91.2
18.08
7.26
2.84
4.14
5.22
4.62
38.4
16.64
2.59
1.51
1.60
71.2
4.03
17.43
4.14
1-5% HCI 1-20% HCI
+4.3 -1.8
-2.3 -7.1
+6.0 +9.3
+31.2 +24.0
-1.9 -6.3
-1.0 -3.4
-1.8 -5.0
-1.3 -3.0
-4.4 -8.0
+28.6 +27.6
-1.7 -6.3
-0.4 -3.7
-0.6 -3.8
-4.2 -4.8
+8.0 +11.6
-0.9 -4.7
+1.0 -8.1
+1.1 -1.7
-------
Table 6. Stability of Spex A Standard in 2% Nitric Acid over a Two-Year Period
Concentration 0200 ng/mL
Element
Ag
Al
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Mg
Mn
Ni
Pb
Sn
Sr
V
Zn
Zr
Average
Standard
Deviation
1.00
Average
4.00
10.00 jig/mL
Standard
Deviation
Average
Standard
Deviation
Average
Standard
Deviation
0.200
0.206
0.203
0.200
0.180
0.204
0.207
0.212
0.206
0.203
0.201
0.206
0.203
0.196
0.223
0.204
0.210
0.204
0.200
0.010
0.020
0.004
0.003
0.028
0.003
0.006
0.009
0.004
0.010
0.004
0.003
0.008
0.030
0.097
0.004
0.005
0.005
0.005
1.026
1.007
0.998
0.973
1.000
0.999
1.013
1.033
1.009
1.006
1.005
1.005
0.999
0.992
0.994
0.992
1.016
0.990
0.996
0.017
0.042
0.024
0.021
0.035
0.021
0.027
0.042
0.023
0.026
0.019
0.020
0.026
0.049
0.255
0.024
0.021
0.029
0.017
4.018
4.051
4.018
3.970
4.064
3.989
4.016
4.134
4.010
4.059
4.045
4.022
4.012
4.014
3.623
4.046
4.018
4.026
3.948
0.051
0.073
0.066
0.065
0.092
0.073
0.072
0.135
0.066
0.074
0.068
0.074
0.072
0.075
0.777
0.082
0.064
0.102
0.088
9.800
10.24
10.06
9.990
10.39
10.11
10.26
10.39
10.25
10.17
10.22
10.20
9.946
9.949
8.432
10.25
10.26
10.14
10.14
0.346
0.236
0.274
0.231
0.274
0.234
0.240 a.
0.439 £
0.201
0.274
0.214
0.230
0.236
0.232
0.649
0.294
0.217
0.303
0.160
-------
Table 7. Stability of Selected Elements in 2% Nitric Acid over a Two-Year Period
Concentration
Element
B
Mo
Ti
0.500 ng/mL
2.50
10.0
Average
0.496
0.496
0.511
Standard
Deviation
0.018
0.024
0.010
Average
2.418
2.439
2.517
Standard
Deviation
0.070
0.071
0.070
Average
9.855
10.06
10.25
Standard
Deviation
0.207
0.20
0.26
O)
r-
co
Concentration
Si
As
Sb
Se
1.00 \nglmL
1.023 0.058
5.00
20.0
5.057
5.201
5.344
4.906
0.217
0.148
0.360
0.141
20.95
20.75
21.26
20.21
0.66
0.48
1.43
0.37
-------
1.10
Figure 1. Copper Concentration as a Function of Time
I
1-05
•2 1.00
§
§
U 0.95
1% HNO3
2% HNO3
5% HNO3
1% HC1
2% HC1
5%HC1
1%HNO3 + 5%HC1
o
oo
co
0.90
10 20 30 40 50 60 70
Time (weeks)
80
90 100 110
Acid Medium
1%HNO3
2% HNO3
5% HNO3
1% HC1
2% HC1
5% HC1
1% HNO3 + 5% HC1
Rel. Standard
Deviation
1.80
2.28
1.75
1.85
1.96
1.90
1.73
-------
Figure 2. Concentration of Lead in 2% HC1 as a Function of
Time
4.00
-| 3.00
§
2.00
|
3 i.oo
0.00
0.2 ppm
1 ppm
4 ppm
10 20 30 40 50 60 70 80 90 100 110
Time (weeks)
Figure 3. Tungsten Concentration as a Function of Time
6.00
5.00
4.00
o
'I 3.00
•M
§
§ 2.00
u
1.00
0.00
2% HC1
2% HC1 + 0.5% HF
10 20 30 40 50 60
Time (weeks)
70 80
90
381
-------
v.
X"
Figure 4. Relative Standard Deviation for Aluminum as a
Function of Concentration
16 s
0.2 1 4
Standard Concentration, (ng/mL)
10
1%HNO3
HC1
1%HNO3
2% HNO3
5% HNO3
Figure 5. Relative Standard Deviation for Cadmium as a Function
of Concentration
0.2 1 4
Standard Concentration,
10
-• l%HNO3 + 5%
HC1
382
-------
'The entire data base could not be included in this presentation due to space limitations. This
information is available from the authors upon request.
LITERATURE CITED
(1) Yoshimura, J. E.; Suzuki, H.; Yamazake, S.; Toda, S. "Interference by Mineral Acids
in Inductively Coupled Plasma Atomic Emission Spectrometry," Analyst 115, 167,
1990.
(2) Huff, E. A.; Huff, D. R. "The Stability of Calibration Standards for ICP/AES
Analysis: Six-month Study," ANL/ACL-92/3, May 1992.
(3) USEPA Contract Laboratory Program Statement of Work for Inorganic Analysis, U.S.
Environmental Protection Agency Report ILM02.0, 1992.
(4) Bass, D. A.; TenKate, L. B. "Stability of Low Concentration Calibration Standards for
Graphite Furnace Atomic Absorption Spectrophotometry," ANL/ACL-93/3, November
1993.
383
-------
58
CLP-Type Analyses Using an Axial Plasma ICP-OES
Marc Paustian. Karen Barnes, and Zoe Grosser
The Perkin-Elmer Corporation, 761 Main Avenue, Norwalk, CT 06859-0080
Introduction
The Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) of 1981,
updated by the Superfund Amendments and Reauthorization Act (SARA) of 1986 gave the EPA
responsibility for the cleanup of hazardous waste disposal sites that have been abandoned. In order to
assess the contamination of these sites and monitor the cleanup efforts the EPA has created the Contract
Laboratory Program (CLP) to control the collection and disbursement of analytical data. Routine samples
are contracted out to commercial analytical laboratories to analyze using a fixed set of protocols specified
in the Statement of Work (SOW). The quality assurance and control measures specified in the SOW are
quite stringent to ensure that the data can stand alone in a Court of Law in the event that potentially
responsible parties are sued for cleanup costs.
Over the years this program has been in effect, the profitability of this kind of sample for the typical
environmental laboratory has declined. However, the quality control used by this program has become a
standard synonymous with high quality and many customers specify "CLP-like" analyses. Therefore the
pressure on the environmental laboratory has been to produce more data, with more QC, at a reasonable
cost to maintain a reasonable profit level.
One way to increase productivity, resulting in a lower sample cost, is to streamline testing by performing
the analyses required on fewer techniques. Advances in ICP-OES technology have allowed improvements
in detection limits of approximately an order of magnitude which permits the elements typically
determined by graphite furnace atomic absorption (GFAA), specifically arsenic, selenium, lead, and
thallium, to be determined by ICP-OES. Of the conventional inorganic analytes, only mercury must still
be determined using a separate technique. This provides substantial savings in time because a separate
sample preparation, required for graphite furnace, is eliminated. In addition, the time to do the analysis is
reduced because the ICP technique is faster and the number of QC checks required are fewer than
specified for graphite furnace.
This paper describes the performance of the ICP method 200.7 modified for the CLP program SOW
ILM03. The initial performance of the instrument is demonstrated through the documentation of the
linear ranges of each wavelength chosen and instrument detection limits. The conditions used to achieve
the best performance on a variety of samples are specified.
Experimental
The Perkin-Elmer Optima 3000 XL ICP-OES was used for the analysis of a variety of CLP samples. The
Optima 3000 XL ICP-OES is a simultaneous ICP with an echelle polychromator and Segmented-Array
Charge-coupled Detector (SCO). Simultaneous measurement of the background and analyte emission
allows for accurate correction of transient background fluctuations.
The instrument conditions used for the instrument detection limits (IDLs), linear range analysis, and
analytical results are shown in Table I.
384
-------
Table I
Instrumental Conditions
RF Power
Plasma Gas Flow
Auxiliary Gas Flow
Nebulizer Gas Flow
Solution Pump Rate
Nebulizer
Equilibration Time
1300 W
15 L/min
0.5 L/min
0.85 L/min
1.5 mL/min
GemTip Cross Flow
15 seconds
Calibration standards were prepared from PE Pure CLP standards. One standard and a calibration blank
were used for calibration. The standards were prepared in a HNC>3/HC1 as specified by Method 200.7-
CLP-M, to approximate the digested sample matrix.
Method 200.7-CLP-M, as published in the SOW ILM03, was followed for the initial demonstration of the
instrument performance and the analytical determination of the samples. The QC checks specified in the
method were automated through the use of QC Expert Software. QC Expert will monitor the analysis and
make real-time decisions about recalibration, rerunning the samples, or halting the analysis, at the
Analyst's direction. This removes the tedious manual calculation of the QC results from the Analyst and
makes automated QC charting possible.
Results and Discussion
The instrument detection limits and linear ranges were evaluated to characterize the instrument
capabilities. A selection of detection limits are shown in Table II, determined using area processing with
simultaneous background correction or using Multicomponent Spectral Fitting (MSF) a mathematical
algorithm combining Interfering Element Correction (IEC) capabilities with simultaneous background
correction. They are compared with the contract required detection limits (CRDLs) specified in the SOW
and, in each case, at least one wavelength is available which meets the criteria.
Table II
Optima 3000 XL Detection Limits
Element
As
As
Cd
Cd
Pb
Pb
Se
Se
Tl
Tl
Wavelength
188.979
193.696
214.438
226.502
216.999
220.353
196.026
203.985
190.800
276.787
EPA CRDL (us/L)
10
10
5
5
3
3
5
5
10
10
AREA - BGC
4.9
4.4
0.12
0.11
6.6
2.0
3.4
9.2
6.3
8.8
MSF
3.1
4.3
0.11
0.09
8.6
1.8
2.2
4.7
5.8
3.8
The linear ranges are shown for a selection of elements in Table III. The linear dynamic range of the
ICP-OES is generally more than five orders of magnitude, allowing the analysis of samples of widely
varying concentrations and the compensation of interferences at various concentrations through
interfering element corrections (lECs). The linear range was preserved in many cases, in addition to
lowering the detection limits.
385
-------
Table III
Linear Range for Selected Elements (
Element
As
Cd
Pb
Se
Sb
Tl
Fe
Mg
Al
Linear Range
50
>200
>200
>200
50
>200
475
20
475
jpm)
Stability is important for long runs without time-consuming recalibrations. Furthermore, lECs rely on a
stable system to perform within the required criteria without frequent remodeling. Figure 1 shows the
excellent recovery of the continuing calibration verification (CCV) standard over a long-term run. The
limits for the CCV are 90-110% and the elements tested (only a subset shown here) all fell within the
required range.
80
8 6:28
/I PM
8:07
PM
9:46
PM
11:23
PM
1:01
AM
2:08
AM
2:50
AM
4:31
AM
Time
Figure 1. Long Term Stability of the CCV
The results for a variety of CLP-type samples will be discussed. Implementation of the QC and the
resulting performance will be examined.
386
-------
59
DETERMINATION OF MERCURY BY ICP-MS
David. E. Dobb. J. T. Rowan, and D. Cardenas, Lockheed Environmental Systems &
Technologies Company, Las Vegas, NV 89119, and L.C. Butler, U.S. Environmental
Protection Agency, Las Vegas, NV 89119
ABSTRACT
Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is a desirable alternative to cold
vapor atomic absorption (CVAA) because of its multielement capability and it serves as a new
independent method for the analysis of mercury. CVAA is dependant on the successful
reduction of mercury to the metal to accomplish volatilization prior to detection. Since ICP-
MS is not dependent on the electrochemical state of species in solution, it is not prone to the
same interferences as CVAA. All oxidation states of mercury are measurable by ICP-MS.
The determination of mercury by ICP-MS has been limited by a lack of suitable sample
digestion techniques. The high levels of permanganate and persulfate in the CVAA (SW 846
Method 7471) digest are prohibitive for ICP-MS and open-beaker acid digestions have proven
inadequate because of mercury losses through volatilization. An aqua-regia digestion in a
closed vessel does not experience such losses and recoveries average 10 to 15% greater than
those achieved with the CVAA digestion using permanganate and persulfate. In addition to
good recoveries in a matrix that is suitable for ICP-MS, the closed-vessel procedure is
relatively simple and is as reproducible as CVAA.
Mercury analysis by ICP-MS has traditionally been plagued with long rinse-out times. Low
levels of mercury present less of a "memory" problem and longer integration times can be
used to compensate for low concentrations and yield the required sensitivity. A rinse solution
consisting of 2.5 mg/L Au in 6% v/v HN03 has been shown to be effective in dealing with
mercury retention in the sample introduction system, particularly if the analytical range is
maintained in a region from 0.15-25 /yg/L. The analytical range can be increased to 50 /yg/L
if the acid concentration of the rinse solution is increased, but erosion of the nickel sampling
cones may result.
The closed-vessel aqua-regia digest has been shown to be appropriate for the rest of the EPA
target metals as well. Holding times for mercury, however, are currently much shorter than
those for the other CLP metals. The use of gold in solution has been shown to stabilize
mercury for as long as the holding times of the other metals. With the stabilization of mercury
in solution, all of the EPA target metals can be determined with one digestion and one
instrument. This could allow significant savings and improved analytical efficiency over the
current practice of using three digestions and three analytical techniques.
NOTICE
Although the research described in this article has been funded wholly by the U.S.
Environmental Protection Agency through Contract 68-CO-0049 to Lockheed Environmental
Systems & Technologies Company, it has not been subjected to Agency review. Therefore,
it does not necessarily reflect the views of the Agency. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
387
-------
INTRODUCTION
The determination of mercury has traditionally been accomplished by cold vapor atomic
absorption (CVAA). The analysis relies on a series of electrochemical transitions, all of which
are subject to interferences. Mercury is detected as a metal vapor and so it is necessary to
reduce the mercury in the sample to the metallic form. The success of this step depends
largely on the digestion process and is in turn dependant on the nature of the samples
themselves. The reduction step is only effective for free mercury ions. Organo-mercury
compounds must be decomposed in order to liberate the mercury ions prior to the reduction
step. This requires the use of strong oxidizing agents and heating. The amounts of reagents
used depends on the nature of the sample and will vary accordingly. In addition, the presence
of oxidizing or reducing agents in the sample may cause premature reduction or prevent final
electrochemical reduction to the metal.
Mercury determinations by alternative methods are becoming increasingly feasible. Mercury
determinations employing relatively new technologies such as He MIP-AES1, photoacoustic
spectroscopy2, and ICP-MS with isotope dilution3'4 and flow injection analysis have been
reported. ICP-MS is a most promising alternative to CVAA for the determination of mercury.
Sensitivities are comparable, however, ICP-MS has the advantage that all forms of mercury
are measurable as long as they can be introduced to the instrument. In addition, ICP-MS can
also be used to determine all of the other EPA target metals at the same time as mercury is
being analyzed. ICP-MS would also serve as a truly independent method because it is not
dependent on the electrochemical states of species in solution and will not be prone to the
same interferences as CVAA.
The use of ICP-MS for routine mercury determinations has been limited by the lack of suitable
sample digestion techniques. The high levels of permanganate and persulfate in the CVAA
(SW 846 Method 7471) digest are prohibitive for ICP-MS and open-beaker acid digestions
have proven inadequate because of mercury losses through volatilization5. Another problem
that is particularly difficult to overcome in the case of mercury is the memory effect6'9' In
fact, slow rinse-out is traditionally the main reason for not using ICP-MS for mercury analysis.
Both sample preparation compatibility and memory problems will need to be overcome in order
to successfully apply ICP-MS to the routine determination of mercury.
This paper describes an investigation into the resolution of both of these problems. A closed-
vessel nitric/hydrochloric acid digestion was investigated as an alternative to the CVAA
digestion, and the use of a gold additive to the instrument rinse solution was examined as a
possible solution to the memory problem. An alternative sample introduction system was also
evaluated. The resolution of these issues will clear the way to using a single technique for
the digestion and analysis of all of the EPA target metals.
EXPERIMENTAL
A digestion method for mercury was developed to be compatible with ICP-MS
instrumentation, provide complete extraction of mercury from water or soil samples, and
retain mercury without volatilization or adsorption losses. The method described below uses
a nitric acid/hydrochloric acid mixture heated in a microwave vessel to satisfy these
requirements. Application of the method to a number of heavily contaminated soil samples
was performed.
388
-------
ft************************
Digestion Procedure for Total Mercury Analysis by ICP-MS
Prepare a solution of 1 part hydrochloric acid to 6 parts of nitric acid to 1 7 parts of water in
a cleaned container (1:6:17 acid mixture). In standard teflon-lined microwave digestion
vessels (equivalent to CEM model MDS-81 D, Matthews, NC) add 1 gram of sample weighed
to the nearest milligram. Add 24 ml_ of 1:6:17 acid mixture and, in the event of apparent
gaseous reactions, allow to de-gas for at least five minutes. Assemble the microwave vessel.
Weigh each container to the nearest hundredth of a gram. Using manufacturer's instructions,
attach a pressure controller sensor to a microwave vessel that contains an actual soil sample
(not a blank). Adjust pressure sensor to 150 psi, power setting to 100%, and timer to 30
minutes. After microwave timer has cycled and samples have cooled, reweigh samples to
check for leakage. Samples that have lost more than 1 gram should be rejected and
redigested. Once the weight loss check is completed, pour and rinse contents of the teflon
liner into a 100 ml_ volumetric. Add 0.25 ml_ of 1000 mg/L AuCI3 (as Au3+) as a preservative.
Bring to volume with ASTM Type II water. Allow particulates to settle or centrifuge before
ICP-MS analysis. If heavily contaminated samples are encountered, dilute the samples in 6%
HN03 solution until the Hg is in a 1 to 50 ppb range. The method is also suitable for digestion
of other metals of interest in environmental samples.
*#**#*#*#******#**#**#**
Following digestion, extracts were analyzed on a VG PlasmaQuad ICP-MS using the
parameters listed in Table 1.
Table 1. ICP-MS Analysis Parameters.
Parameter (VG PlasmaQuad PQ2 + )
RF Power
Nebulizer Gas
Auxiliary Gas
Coolant Gas
Nebulizer; Spray Chamber; Torch
Solution Uptake Rate; Rinse Soln.
Sampler, Skimmer Cones
Masses (peak jump dwell - //s)
Integration Method
Data Collection Parameters
Setting
1.2 KW
0.69 LPM
0.20 LPM
13 LPM
Hildebrand Grid; Chilled Scott Spray Chamber @ 4°C
made of glass; ICP (Fassel type) quartz torch
1 .2 mL/min; 2.5 ppm Au in 6% HN03
Nickel (1 mm orifice sampler, 0.7 mm orifice skimmer)
'59Tb (2560), 200Hg (40960), 202Hg (40960), 209Bi (2560)
Constant Area - 0.9 amu
Pulse Collector, 10 sweeps, 5 pts/peak, 5 dac-steps/pt
For comparison, the same samples were digested and analyzed by CVAA using EPA SW-846
Method 7471.
Memory effects were studied by exposing the ICP-MS instrument to a solution of 10 //g/L Hg
in 6% v/v HN03 for a normal sample analysis time of 3 minutes. Multiple integrations of 15
seconds each were begun, followed by introduction of a rinse solution of 6% HN03. Signal
counts per second for the rinse solution were measured versus washout times for about 20
389
-------
minutes or until the mercury signal reached baseline. The experiment was repeated on a rinse
solution containing 2.5 mg/L AuCL3 and 6% HN03. The washout experiments were also
performed on a Perkin Elmer Elan 5000 ICP-MS equipped with a demountable ICP torch with
a ceramic injector tip and a plastic spray chamber using conditions similar to those used on
the VG instrument. Data was reduced by summing counts per second for mercury isotopes
(200 and 202) then dividing by bismuth internal standard counts per second at mass 209.
The ratio was blank subtracted to give a net intensity ratio. The net intensity ratio was then
divided by the net intensity ratios obtained for the 10 //g/L Hg standard that caused the
memory, then multiplied by 100 to give a percent relative intensity for each of the 15 second
integrations. Relative intensities were then plotted versus washout time in minutes.
RESULTS AND DISCUSSION
Overcoming Washout Problems
The principal analytical problem in the determination of mercury by ICP-MS is the long wash-
out time that is often found for mercury. Two approaches were followed to limit the wash-
out time. The first approach was to modify the sample introduction hardware where the
problem is thought to originate. The second approach was to modify the rinse solution to
either prevent mercury from being retained or to free any retained mercury in the sample
introduction system. Both approaches were successful, as illustrated in Figure 1.
120
01
c
01
01
01
cc
-20-
CQrarriic Inj. rio Au rinse
-2
-i-
0
-t-
-i-
-1-
4 6 8 10
Washout Time - Min.
—i—
12
14
16
Figure 1. Washout Profiles for Mercury
390
-------
If nothing is done to restrict memory problems for mercury, washout curves like that shown
for the Quartz Injector - no Au rinse, may be seen. In fact, the memory signal can be as high
as the previous sample that caused the memory. Clearly, no useful analytical information can
be obtained in this situation and it is not feasible to wait for signals to return to their
background levels. By modifying the composition of the ICP torch injector tube, where
significant memory behavior can originate, instrument manufacturers have found that memory
can be greatly restricted. In experiments performed with a Perkin-Elmer SCIEX ELAN 5000,
modified with a plastic spray chamber and a ceramic injector tube, no discernable difference
was seen in the washout curves whether or not gold was used in the rinse solution. Overall,
the modified sample introduction system showed very little memory compared to the
traditional quartz components. Unfortunately, the ceramic modifications are not common
among ICP-MS installations, but they could obviate the need for rinse solution modifiers. The
main component of memory that is left when a ceramic injector tube is used is characteristic
of dissipation of the fog in the spray chamber from the previous sample.6
In contrast, there is a significant increase in carryover of mercury when a quartz injector tube
is used. It has been speculated that there is a chromatographic effect causing the retention
in which the hot silica injector tip acts much like a chromatography column.6'10 The
phenomenon has been observed for other metals as well, however, it is much less significant
for the other metals than it is for mercury. Unfortunately, most ICP-MS systems use quartz
torch injector tips. Therefore, additional methods of limiting memory effects had to be
studied. This led to modification of the rinse solution.
Modification of the rinse solution to rapidly strip retained mercury required a reagent that
would rapidly desorb mercury and keep it in solution. If compatible, the same reagent could
be added to the samples to prevent the mercury from being retained. Studies with gold held
the most promise. For the past 20 years, numerous studies11 have shown that mercury in
solution (as Hg(l)) is unstable in aqueous environmental samples, and in contact with materials
such as glass or plastic. If Hg(l) is not properly preserved, it can volatilize, adsorb to the
container walls, or diffuse through plastic. Adding gold ions to solution mitigates these
effects. This led to a separate study by the authors to simulate the equilibrium speciation of
Hg(l) preservation with saturated Au(lll)CI3 solution in 0.31 7M HN03 (2%v/v). At 25°C and
0.6 to 1.6 pH, the dominant oxidation-reduction reactions to form the more stable Hg(ll)
species are:
** AUC12 + 2C1~ +
and
Au(OH)3Cl + Hgl* *- Au(OH)2 + OH~ + Cl + 2Hg2+
The maximum Hg(l) concentration that may be preserved in an acidified sample (pH =0.6) is
controlled by the solubility of Au(OH)3(c), and is 785 ± 30 /vg/L of Au(lll) at 25°C.
Experiments conducted at NIST laboratories12 showed that spiking with 1 //g/mL of AuCI3
solution could restore adsorbed mercury to its original concentration in glass containers over
391
-------
a period of weeks. The MIST work also showed that 1.0//g/L mercury solutions stabilized
with 10 /yg/mL Au3+ (in 0.5N HN03) do not lose mercury to containers made of teflon, glass,
or polyethylene. In fact, solutions stabilized in this manner are stable for many years. With
this evidence it was decided that Au in acidic solution may be useful as a rinsing agent to help
avoid memory problems with mercury.
As shown in Figure 1, a rinse solution consisting of 2.5 mg/L Au3+ in 6% v/v HN03 is very
effective in dealing with mercury retention in a glass spray chamber/quartz torch sample
introduction system. The improvement is seen upon mercury exposures of up to 25 //g/L
mercury. The analytical range can be increased to 50 //g/L if the acid concentration of the
rinse solution is increased, but erosion of the nickel sampling cones may result. (The reader
should note that the quartz injector data was obtained on a VG Plasmaquad instrument that
was operated at maximum sensitivity, hence, more noise was observed). Using this rinse
solution with washout times of 3 minutes gives analytical results that are free of memory.
Au can also be added directly to the sample to prevent mercury from being retained in the first
place. This, in conjunction with Au in the rinse, will provide no opportunity for mercury to be
retained and cause memory. A side benefit is the fact that Au preserves mercury much more
effectively than nitric acid alone.13 Holding times for mercury are currently much shorter than
those for the other CLP metals. The use of gold in solution has been shown to stabilize
mercury in solution for as long as the holding times of the other metals (at least six months).
With the stabilization of mercury in solution and matching holding times with the other metals,
all of the EPA target metals can be determined with one digestion and one instrument. This
could allow significant savings and improved analytical efficiency over current practices which
require three digestions and three analytical techniques.
Performance of the Closed Vessel Digestion Method
Since the CVAA digestion technique yields a sample matrix that is untenable for ICP-MS, an
alternative digestion technique would be required before ICP-MS could be applied to the
routine determination of mercury. Open vessel procedures such as are typically used for ICP
and GFAA digestions have proven unsuitable for mercury because of losses through
volatilization and during the digestion. Digestion in closed vessels in a mixture of nitric and
hydrochloric acid does not suffer the losses experienced by open beaker techniques. Mercury
recoveries are comparable to those achieved by CVAA with Method 7471 on most samples.
On samples that are heavily contaminated with mercury (>1000 ppm), the closed-vessel
method actually performs better, averaging 10 to 15% greater extraction than that which can
be achieved by Method 7471. If adopted, the closed-vessel digestion would have a number
of advantages over current digestion methodology for mercury. The sample matrix is suitable
for ICP-MS, the procedure is relatively simple, it is quick, reproducibility is comparable to
CVAA, and the closed-vessel mixed-acid digest has been shown to be appropriate for the rest
of the EPA target metals as well.
A number of soil samples were obtained from a site heavily contaminated with a variety of
mercury compounds. The samples were digested using the microwave assisted mixed acid
method, then analyzed by ICP-MS using the aforementioned rinse solution. Results are shown
in Table 2. Comparison with CVAA analyses (by EPA Method 7471) of the same samples
indicated both methods agree up to 1000//g/g, but the ICP-MS determinations get slightly
larger totals on higher samples. This may be a limitation of the ability of the 7471 digestion
method to extract large quantities of mercury from a soil.
392
-------
Table 2. Comparison of total mercury results in heavily contaminated soils.
Soil
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Mercury in fjg/g
ICP-MS
27.8
442
64.7
339
281
23.8
217
157
1,670
73.5
2,090
96.4
1,060
294
3,300
301
2,130
247
2,630
CVAA
29.2
376
56.2
589
454
21.4
183
129
1,360
64.8
1,830
85.6
1,190
258
2,850
281
2,020
226
2,060
There are few available solid reference materials with certified values for mercury, probably
due to it's poor stability. However, accuracy can be evaluated by analyzing soils that are
accurately spiked with mercury compounds. Table 3 shows a typical spike recovery as well
as an indication of overall method precision.
Table 3. Results of ICP-MS analysis of the Laboratory Control Sample (LCS) for total Hg
ASTM Standard Soil
Mercury Spike
(mg/kg)
939
939
939
Mercury Found
(mg/kg)
901
1001
848
Mercury Found
(%)
96
107
90
393
-------
CONCLUSIONS
ICP-MS has been successfully used for analysis of heavily contaminated soil samples. This was
possible after development of a closed-vessel mixed-acid digestion method for mercury that is
compatible with ICP-MS. Washout problems with mercury were also successfully overcome.
Mercury retention in the sample introduction system can be mitigated by application of a rinse
solution of 2.5 mg/L Au3+ in 6% v/v HNO3, or by application of a ceramic ICP torch injector
tube\plastic spray chamber design. The combination of the digestion method and the
multielement capabilities of ICP-MS will allow analysis of all EPA target metals on a single
instrument at the same time. This could represent significant analytical cost savings for the
Agency.
REFERENCES
1. Fukushi, K.; Willie, S. N.; Sturgeon, R. E. Analytical Letters 1993 26(2), 325-340.
2. VanderNoot, V.A.; Lai, E.P.C. Anal. Chem. 1992 64(24), 3187-3190.
3. Beauchemin, D.; Siu, K.W.M.; Herman, S.S. Anal. Chem. 1988 60(23), 2587-2590.
4. Smith, Ralph G. Anal. Chem. 1993 65(18), 2485-2488.
5. Dumarey, R.; Van Ryckeghem, M.; Dams, R. J. Trace and Microprobe Techniques
1987 5(2,3), 229-242.
6. Dobb, David E.; Jenke, Dennis R. Appl Spectrosc. 1983 37(4), 379-384.
7. Wohlers, C.C. Jarrell-Ash Plasma Newsl. 1978 1, 22.
8. Olsen, K.W.; Haas W.J. Jr.; Fassel, V.A. Anal. Chem. 1977 49, 632.
9. Thelan, B. Analyst 1981 106, 54.
10. Kantor, T., Grofne, H.E., Nagyne, H.B., and Purgor, E., Magy. Kam. Foly., 1981, 87,
39.
11. Jenne, E.A.; Avotins, P. Journal of Environmental Quality 1975 4(4), 427-431.
12. Moody, J.R.; Paulsen, P.J.; Rains, T.C.; Rook, H.L.; Accuracy in Trace Analysis:
Sampling, Sample Handling, and Analysis, LaFleur, P.D., Ed.; National Bureau of
Standards Spec. Publ. 422; 1976; 267-273.
13. Dobb, David E.; Metcalf, Richard C.; Gerlach, Robert W.; Butler, Larry C.; to be
presented at Emerging Technologies in Hazardous Waste Management VI Atlanta GA,
September 1994
394
-------
60
Errors Eliminations and Quality Assurance Procedures for the Determination of
Germanium, Arsenic and Selenium in Biological Sample by Inductively Coupled
Plasma Mass Spectrometry
Fu-Hsiang Ko, Institute of Nuclear Science, National Tsing Hua University, Hsinchu,
30043, Taiwan; and Ti Yeh, Center for Measurement Standard, Industry Technology
Research Institute, Hsinchu, 30042, Taiwan
ABSTRACT
The matrix components in biological tissues are rather complicated, thereby making
the determination of trace elements in biological tissues quite difficult. In the past, the
sensitivity of instrumentation has been limited. Recently, inductively coupled plasma
mass spectrometry (ICP-MS) has shown highly promising potential for direct trace
element determinations. Recognizing those errors which arise from matrix effects is
essential. Those errors can then be hopefully understood, with the ultimate intentions
of controlling and eliminating them.
The types of errors in ICP-MS can be divided into three basic categories, i.e., gross
error, statistical error and systematic error. The systematic errors which arise in the
determination of germanium, arsenic and selenium in a biological sample by ICP-MS
occur due to additive or multiplicative interferences, i.e. the formation of polyatomic
ions, refractory oxide ions and doubly charged ions. In addition, isobar overlap, salt
build-up on the cone of interface, the process of sample introduction and transport, the
pathway of ion extraction in ion optics, and sample matrix induced ion suppression or
enhancement are also feasible origins of systematic errors. The optimal approach of
eliminating the systematic errors in ICP-MS measurement involves overcoming the
interferences due to spectroscopy or non-spectroscopy.
The origins of errors in ICP-MS measurement are clearly understood. The quality
assurance program is developed to ensure that the analytical result is reliable.
Experimental results have indicated in this study that although principles of quality
control and assurance are fairly specific, their interpretation and utilization are treated
considerably different in distinctive programs. This difference shown is obvious since
each organization tends to adjust its program towards its own operations and
requirements. What may be suitable for a large complex laboratory involved with
several disciplines or examines a variety of products may not be appropriate for a small
laboratory whose activities are limited to a few tests or a few products. However,
regardless of differences in laboratory complexity, conventional practices and
procedures require being incorporated into the quality assurance program.
INTRODUCTION
Inductively coupled plasma mass spectrometry (ICP-MS) is a powerful technique
employed for the measurement of trace elements; however, it is not always readily
applicable for all elements in every conceivable sample. The direct determination of As,
Se and Ge by ICP-MS in complex matrix is relatively difficult. Generally, Cl and Na
are either removed by chemical separation procedures (1-4) or treated with a high-
395
-------
resolution mass spectrometer (5).
Up to now, the standard method is limited (6) for using ICP-MS as a multi-element
detector. Vanhoe and co-workers (7-8) have published a method for the determination
of 11 ultra-trace elements in human serum by ICP-MS. Alves et. al. (4) added a small
dose (2%) of H2 to the aerosol gas flow that enhanced analyte signals by a factor of 2-
3 for V, Ni and As in seawater and urine reference materials by ICP-MS.
The primary aim of this work involves establishing the hyphenated system and, then,
applying it towards the determination of As, Se and Ge in urine samples. The matrix in
urine sample was separated by the system, and the interference was removed. A series
of analytical datas were listed and the suspected data was tested. The nomality and
randomness were also tested. Finally, the quality assurance program is designed to
ensure the analytical result is reliable.
EXPERIMENTAL
COLUMN PREPARATION AND REAGENTS: Anion, AG1-X8 (100-200 mesh,
Cl'form), exchange resin was purchased from BIO-RAD (USA) and the resin was
slurry packed into a 4.6 mm i.d. x 25 cm PEEK column (Alltech Associates, USA)
with a packing machine obtained from Alltech Associates. The self-prepared column
was successively rinsed with 100 ml of 10% HNO3 and 100 ml of water.
Unless specified otherwise, all reagents were of analytical or higher grade (E. Merck,
Darmstadt). Stock standard solutions (1000 ppm) of Se(VI) were prepared from
sodium arsenate (Aldrich, USA), and the other standard solutions of Ge(TV) and As(V)
were purchased from Aldrich and E. Merck, respectively. All buffers and eluents were
prepared in Millipore (MA, USA) Milli-RO 10 PLUS water system which was used
throughout the study. High purity nitric acid was obtained by sub-boiling distillation of
reagent grade acid from a quartz apparatus. The pH values of the solutions in the range
1-4 were adjusted with dilute nitric acid. In the pH range 4-7, the buffer contained
ammonium acetate and acetic acid, or ammonium hydroxide and ammonium acetate
for the range 7-12. The pH value of solutions in the range 12-13 was adjusted with
sodium hydroxide. All buffers were purified by passing through the column. After the
uptake (purification) step, the retained analytes were eluted with 10% of HNO3
solution
APPARATUS: The liquid chromatographic system consisted of a solvent delivery
pump (DIONEX, MODEL DQP-1, USA) equipped with a self-constructed 5-port
valve and the ELAN 5000 ICP-MS system (PERKJN-ELMER SCEEX, THORNHELL,
ONTARIO, CANADA), The operating parameters of the mass spectrometer are
summarized in Table 1. The schematic diagram of the hyphenated system is provided in
Figure 1.
SAMPLE PREPARATION: A freeze-dried human urine reference material [National
396
-------
Institute of Standards and Technology (NIST) Standard Reference Material (SRM)
2670 Low Level Toxic Metals in Human Urine] was reconstituted by an addition of 20
mL of pure water. The solutions were transferred into a 120 mL of moderate-pressure
closed Teflon PFA vessel and 10 mL H2O2 and 1 mL HNO3 were added. These
vessels were processed through microwave digestion (CEM, MDS-81D, USA), which
was monitored with pressure monitoring equipment. The pressure limit was set at 90
psig and the digestion program is gived in Table 2.
RESULTS AND DISCUSSION
MATRIX SEPARATION: In order to optimize the chromatographic conditions for
retention of the arsenate (AsO43~), selenate (SeO42~) and germanate (GeO32") in the
strong anion-exchange column, the recovery was monitored spectrophotometrically
while changing the pH value of the loading eluent. The results (9) show that the
efficiency of arsenate was retained quantitatively in the pH range 3.2-12.6. The
selenate was recovered at the 100% level in the wide pH range 1.6-12.6, however the
pH range is critical for germanate between 10.7-12.2. The typical performance of the
proposed hyphenated technique for matrix separation is shown in Figure 2(a). Figure
2(a) clearly indicated that the sodium cation was not retained by the anion-exchange
resin and eluted out first. During the loading step, the column effluent was directed to
waste in Figure 2(b) indicated that the sample pH was around 11.5 and all of the three
analytes were retained by this column. Next, in the cleaning step, water was pumped
through the column for 2 min in order to remove matrix completely. Next the 6-port
valve was switched to direct the column effluent into the nebulizer of ICP-MS. Finally,
in the washing step, the 5-port valve was switched to elute the germanate and arsenate.
Next the 5-port valve was switched again to elute selenate and chloride, respectively.
It should be expected that the polyatomic interference occurring from the chloride was
alleviated to minimum by chloride separation with ^2Ge+, ^^Ge+ and ^As+,
respectively. Experimental results indicated that the selective isotope of selenium
avoids spectroscopic interference as discussed later. A detailed evaluation of the
accuracy and precision about the proposed method has been provided in reference 9.
THE TYPES OF INTERFERENCES IN ICP-MS: Tschopel and Tolg (10) have
mentioned the reasons which produce a systematic error and described the available
methods to detect this error. The systematic error includes the spectroscopic and non-
spectroscopic interferences in ICP-MS. The interferences originated from ICP-MS
may lead to analytical errors as shown in Figure 3, which involve the additive or
multiplicative interferences, i.e. the formation of polyatomic ions, refractory oxide ions
and doubly charged ions. In addition, isobar overlap, salt build-up on the cone of
interface, the process of sample introduction and transport, the pathway of ion
extraction in ion optics, and sample matrix induced ion-suppression or enhancement
are also the feasible sources of errors. Notably, the most interferences were placed
under best control after sample digestion and instrumentation optimization.
Experimental results revealed that errors were strongly dependent on salt induced
397
-------
effects, salt clog the interface, isobar and polyatomic interferences for the arsenate,
germanate and selenate in urine samples. As mentioned above, the matrix was
separated and bypassed to the waste. Table 3 lists the polyatomic interference for Ge,
As and Se. The selenium mass 78 was selected to prevent the chloide interference. The
responses for the influence of various sodium chloride concentration are tabulated in
Table 4, The analyte concentration equivalents arising from various acid are displayed
in Table 5. The effect of ion induced suppression occurred where the concentration of
sodium chloride reached 1000 ppm, indicating that a small amount of sodium chloride
would have slightly promoted the signal.
OUTLIER, GOODNESS OF FIT AND RUN TEST: The proposed hyphenated
technique has been successfully applied towards urine, sample analysis. The replicate
determinations of arsenic in urine (normal) MIST SRM 2670 are 58.1, 60.0, 57.8, 55.2,
54.7, 60.3, 58.0, 62.9, 63.2, 59.8, 57.0, 65.7, 63.2, 57.5, 56.2, 59.8, 61.2, 58.8, 64.3,
64.0, 67.2, 70.9, 66.2, 57.6, 58.4 and 50.6 ppb, respectively. One approach of
assessing a suspect measurement involves making a comparison of the difference in the
analysis results. Table 6 indicates that the critical value 50.6 and 70.9 were accepted
under a 95% confidence level for various tests. Table 7 shows that the result of chi-
square test, Clearly reveal that the data series were normality under a 95% confidence
limit. Table 8 indicates that the data series were random under a 95% confidence level.
QUALITY ASSURANCE PROCEDURES: The origins of interference in ICP-MS
measurement have been clearly understood, Therefore, the quality assurance program
is developed in this study to ensure that the analytical result is reliable. The quality
control diagram shown in Figure 4 should be applied towards the determination of
trace elements in a biological sample. Garfield (11) has already mentioned that
although principles of quality control and assurance are fairly specific, their
interpretation and utilization are treated considerably different in distinctive programs.
Each organization tends to adjust its own program toward its specific operations and
requirements. For example, a large complex laboratory involved with several
disciplines or determines various samples may not be appropriate for a small laboratory
whose activities are limited to only a few samples. However, regardless of differences
in laboratory complexity, conventional practices and procedures require being
incorporated into the quality assurance program.
SUMMARY
The concentration of trace elements in the human body is extremely low; that is,
analysis of these elements would be rather difficult. The analytical results for a
biological tissue is beneficial for all human beings. Obtaining no analytical result is
generally more acceptable than obtaining the wrong one in the sense that the wrong
conclusion would be drawn on the basis of inaccurate results. The reliable data
obtained from ICP-MS correlates sufficiently with a good quality assurance program
(12) in any laboratory.
398
-------
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the financial support of the National Science
Council of Taiwan.
REFERENCES
(1) Beauchemin, D.; Berman, S. S. "Determination of Trace Metals in Reference
Water Standards by Inductively Coupled Plasma Mass Spectrometry with On-Line
Preconcentration", Anal. Chem., 1989, 61, 1857-1862.
(2) Plantz, M. R.; Fritz, J. S.; Smith, F. G.; Houk, R. S. "Separation of Trace Metal
Complexes for Analysis of Samples of High Salt Content by Inductively Coupled
Plasma Mass Spectrometry", Anal. Chem., 1989,61, 149-153.
(3) Goossens, J.; Moens, L.; Dams, R.; "Anion Exchange for the Determination of
Arsenic and Selenium by Inductively Coupled Plasma Mass Spectrometry", J. Anal. At.
Spectrom., 1993, 8, 921-926.
(4) Alves, L. C.; Allen, L. A.; Houk, R. S. "Measurement of Vanadium, Nickel, and
Arsenic in Seawater and Urine Reference Materials by Inductively Coupled Plasma
Mass Spectrometry with Cryogenic Desolvation", Anal. Chem., 1993, 65,2468-2471.
(5) Morita, M.; Ito, H.; Uehiro, T.; Otsuka, K. "High Resolution Mass Spectrometry
with Inductively Coupled Argon Plasma lonization Source", Anal. Sci., 1989, 5, 609-
610.
(6) Long, S. E.; Martin, T. D. "Determination of Trace Elements in Waters and Wastes
by Inductively Coupled Plasma- Mass Spectrometry", 1989, Method 200.8, U.S. EPA.
(7) Vanhoe, H.; Vandecasteele, C.; Versieck, J.; Dams, R. "Determination of Iron,
Cobalt, Copper, Zink, Rubidium, Molybdenum, and Cesium in Human Serum by
Inductively Coupled Plasma Mass Spectrometry", Anal. Chem. 1989, 61, 1851-1857.
(8) Vanhoe, Hans; Dams, Richard; Versieck, J. "Use of Inductively Coupled Plasma
Mass Spectrometry for the Determination of Ultra-trace Elements in Human Serum", J.
Anal. At. Spectrom., 1994, 9, 23-31.
(9) Ko, F.-H.; Yang, M.-H. "Matrix Separation for the Determination of Arsenic,
Germanium, and Selenium in Human Urine Samples by Inductively Coupled Plasma
Mass Spectrometry", in Preparation.
(10) Tschopel, P; Tolg, G. "Comments on the Accuracy of Analytical Results in ng-
and -pg Trace Analysis of the Elements", J. Trace and Microprobe Techniques, 1982,
1(1), 1-77.
(11) Garfield, F. M. "Quality Assurance Principles for Analytical Laboratories", 1985,
AOAC, Inc., USA, 1-4.
(12) Yeh, T. "Quality Assurance Program of Analytical Chemistry", 1993, National
Tsing Hua University, Chinese.
399
-------
Table 1 LC-ICP-MS operating conditions
Plasma conditions
R. f. power/W 1050
Plasma gas flow rate/1 min" 1 15
Auxiliary gas flow rate/1 min" 1 0.9
Nebulizer gas flow rate/1 min'l 0.9
Mass spectrometer settings
Bessel box lens/V . 10.95
Bessel box plate lens/V ' -73.8
Photon stop lens/V -10.05
Einzel lenses 1 and 3/V 1.57
LC conditions
Column Strong anion exchange
Eluent flow rate/ml min'1 1.72
Table 2 Microwave digestion parameters for urine samples
Step
1
2
3
4
5
6
7
8
Time (min)
10
10
10
10
10
30
30
30
Power (%)
30
50
50
30
30
20
16
16
Pressure
4
21
81
78
84
88
85
86
(psig)
400
-------
Table 3 Polyatomic interferences for As, Se and Ge, respectively
Element Natural abundance(X) Possible polyatomic interference
Ge-70
Ge-72
Ge-73
Ge-74
Ge-76
As-75
Se-74
Se-76
Se-77
Se-78
Se-80
,Se-82
20.5
27.4
7,8
36.5
7.8
100
0.9
9.0
7.6
23.5
49.6
9.4
70Zn, 35C135C1, ^Ar^S, "Ar3^, 54Cr160
35C137C1, 40Ar32S
40Ar33S) 36Ar37Clf 38^5^
74Se, 37C137C1, "Ar^S
76Se, 36Ar40Ar, 40Ar36S
40Ar35Cl,. 38Ar37Cl
74Ge, 37C137C1, ""Ar^S
76Ge, "Ar^Ar, «Ar*S, 31P31PMN
«Ar"Cl. "Ar^Ar1!!
78Kr, 38Ar40Ar, '«p«p>«o
80Kr, 40Ar40Ar, «P"PI80
82Kr, 1ZC35C135C1
Table 4 The effect of various NaCl concentrations on the signal of As, Se and Ge
CNaCOppn 0 15 20 60 100 500 1000 2500 7500 10000
Ge72 100 109 115 109 112 98 89 80 70 69
Ge73 100 108 116 109 113 98 89 80 71 70
Se78 100 113 114 114 109 90 77 63 48 46
Se82 100 111 116 114 110 89 76 63 50 44
As75 100 112 113 113 111 98 90 78 64 62
n=20.
401
-------
Table 5 Analyte concentration equivalents (ppb) arising from interference at various concentration
o
N>
mass
70
72
73
74
75
76
77
78
82
mass
70
72
73
74
75
76
77
78
82
mass
70
72
73
various nitrogen concentration (ppm)
90.4
0
0
0
0
0
0
0
0
0
110
0
0
0
0
0
0
0
0
0
170
0
0
0
0
0
0
0
0
0
226
0
0
0
0
0
0
0
0
0
440
0
0
0
0
0
1
2
0
0
1760
0
0
0 .
0
0
1
3
0
0
7040
0
0
0
0
0
1
12
0
0
various sulfur concentration
33.3
0
0
0
0
0
0
0
0
-
56.7
0
0
0
0
0
0
0
0
-
133
0
0
0
0
0
0
0
0
-
250
0
0
0
0
0
0
0
0
-
333
0
0
0
0
0
0
0
0
2
395
0
0
0
.0
0
0
0
0
2
1579
0
1
0
0
0
0
0
1
6
6314
0
2
0
0
1
0
0
2
23
25256
0
7
0
0
5
0
0
4
100
various chlorine concentration
76
_
0
0
100
1
o
0
400
0
0
0
750
0
0
0
760
0
0
0
1000
0
0
0
3040
5
2
1
6080
15
7
2
12160
44
20
6
/I
75
•J£
/O
77
78
82
mass
70
72
73
74
fje
/J
76
77
19
/O
82
mass
70
72
73
74
7«
/ J
76
77
70
to
82
000000012
1 4 5 11 - 70 150 274
0000- 0456
3 6 17 23 55 348 762 1390
0000003 11 16
00000039 19
various phosphorus concentration
14 29 115 230 575 1150 2300 4598
00011222
00011 1 1 1
0 0 0.2 3 1 4 4
0 0 0 1 ' 1 1 1 1
0001 1 123
00012278
0 3 2 11 11 13 16 20
0004559 10
0 0 3 9 11 13 23 37
various carbon concentration
11 88 446 892 1785 3569
101111
001110
001111
000000
On i i n rt
U 1 1 U U
000100
1 0
11 i A n n
1 1 u U 0
I 1 I -0 0 0
* The background counts at mass 80 is too high due to 40Ar40Ar+
-------
Table 6 Various outlier tests for verifying suspected data
Data series: 50.6< 54.7< 55.2< 56.2< 57.0< 57.5< 57.6< 57.8< 58.0< 58.1< 58.4<
58.8< 59.8= 59.8< 60.0< 60.3< 61.2< 62.9< 63.2 = 63.2 < 64.CK 64.3< 65.7< 66.2<
67.2< 70.9
X±S = 60.3±4.5
(a) The huge error test ( M= I suspect-mean l/s, M>4 then M is outlier)
M(70.9) = (70.9-60.3)74.5= 2.4< 4
M(50.6) = (60.3-50.6)74.5= 2.2< 4
All data are accepted.
(b) The Dixon test (Under 95% confidence limit)
rmin = r50 6 = (X4-X1)/(XL.2-X1) = (56.2-50.6)7(66.2-50.6) = 0.359O.4
rmax = r70.9 = (XL-XL-3)/(XL-X3) = (70.9-65.7)770.9-55.2) = 0.33 K0.4
All data are accepted.
(c) The Grubbs test (Under 95% confidence limit)
Tmin = T50.6 = (X-X^/S = (60.3-50.6)74.5 = 2.156<2.663
Tmax = T70.9 = (Xn-X)7S = (70.9-60.3)74.5 = 2.356<2.663
All data are accepted.
Table 7 Normal distribution test for data series
Measurement Observed Theoretical probability Theoretical
ppb O[ P ej
< 56.38
56.38-58.97
58.97-61.63
61.63-64.22
> 64.22
4
8
5
4
5
0.192
0.192
0.231
0.192
0.192
5
5
6
5
5
0.2
1.8
0.2
0.2
0
Sum 26 1.0 26 2.4
X2(0.95,2) = 5.99 > 2.4, The data series is normal distribution.
403
-------
Table 8 Randomness test for data series
Data sequencers. 1 60.0 57.8 55.2 54.7 60.3 58.0 62.9 63.2
+ + - + +
59.8 57.0 65.7 63.2 57.5 56.2 59.8 61.2 58.8 64.3 64.0
+ + + - + +
67.2 70.9 66.2 57.6 58.4 50.6
The median is (59.8+59.8)72 = 59.8
ni = 12, n2 = 14, r = 13 (Under 95% confidence limit), rrj = 20, rL = 8 ,
then rrj=20 > r = 13 > q^ = 8, the data sequence is randomness.
404
-------
5-Port Valve
6-Port Valve
o
en
Pump2 _H2O
Figure 1 Schematic diagram of the hyphenated system
-------
20000 y
18000 -
16000 -
14000 -
12000 -•
10000 -
8000 --
6000 -
4000
2000 -
0 -t
(a)
23Na+
(OmniRange=7)
35C1+
(OmniRange=4)
40^350+
100
200
300
400
500
15000 T
13000
11000
9000
7000 4
5000
3000 4
1000
(b)
0
100
200
300
400
500
Figure 2.(a) Elution performance of 10 mg NaCl; 1-200 sec is pretreatment step at pH
11, 201-500 sec is eluted with 4% HNO3. (b) Chromatogram of 24 ng As, Se and Ge
in 10 mg NaCl matrix; 1-160 sec is pretreatment step at pHll, 161-320 sec is eluted
with pH1.25 HNC>3 solution and 321-500 sec is eluted with 4% HNC>3 solution.
406
-------
SPECTROSCOPIC
INTERFERENCES
ADDITIVE
INTERFERENCE
INTERFERENCES IN ICP-MS
NON-SPECTROSCOPIC
INTERFERENCES
MULTIPLICATIVE
INTERFERENCE
POLYATOMIC
OR ADDUCT
IONS
REFRACTORY
OXIDE IONS
SAMPLE
INTRODUCTION
AND TRANSPORT
ION EXTRACTION
EFFECT OP
IONIZATION IN
THE PLASMA
SAMPLE MATRIX INDUCED
SUPPRESSION IN TOE ION BEAM
ION THROUGHOUT IN THE
RESULTANT ION BEAM
SALT BUILD-UP ON THE CONE BY
HIGH TOTAL DISSOLVED SOLIDS
Figure 3 Various types of interferences in ICP-MS
-------
Figure 4 The diagram of quality control procedure in ICP-MS measurement
Sample pretreatment procedure
The determination of instrument]
detection limit (IDL) J
\/
Calibration curve (At least 5 concentrations, 1 blank)}
V
C The blank of calibration curve < IDL
V
(Correlation coefEciency > 0.995
V
Initial calibration (Bias < 10%)
During the period of analysis
Blank analysis
Sample blank analysis
Sample analysis
Spike analysis or serial dilution
Replicate analysis
Quality control sample analysis
Continuously sample blank correction and calibration curve check
The control of recovery and the
frequency of each action are
dependent on the demand of
each laboratory
>
f Data report J
408
-------
ORGANICS
-------
61
1BLWP19.94
IMMUNOASSAY METHODS: DEVELOPMENT AND IMPLEMENTATION
PROGRAM AT THE USEPA
by Barry Lesnik, USEPA, Office of Solid Waste, Methods
Section (5304), 401 M St., SW, Washington, DC 20460
Introduction and Background
Immunoassay technology has several attributes which make it a
useful tool for environmental monitoring, e.g. selectivity,
sensitivity, portability, arid rapid turnaround time. Immunoassay
kits can be tailored to target specific analytes or classes of
analytes, thus eliminating the need for cleanup methods in most
cases to remove interferences. They also have the capability of
detecting target analytes at very low levels, which are needed in
many environmental applications. The portability of immunoassay
test kits and speed of analysis allows for rapid analyses to be run
on a site in the field. This capability can be especially useful
in lowering the costs of cleanup projects because equipment does
not have to lay idle while awaiting the results of laboratory
analyses.
The USEPA has been looking at the potential use of immunoassay
technology for environmental monitoring for several years. The
early methods development efforts were unsuccessful because the
immunoassay chemistry utilized in the methods was not sufficiently
rugged for use on real world environmental matrices. The methods
performed well on clean water matrices arid spiked samples, but did
not perform effectively on natural environmental samples. Because
of this poor initial performance on real samples, EPA Program
Office interest in the technology declined.
In January, 1992, EnSys, Inc. demonstrated a viable
immunoassay test kit for pentachlorophenol in both soil and water
matrices to EPA's Office of Solid Waste (OSW). Since that time,
OSW has been working with several manufacturers to develop and
validate a whole battery of immunoassay test kits both for
individual analytes and for classes of analytes. Currently, OSW
has issued four immunoassay methods which can be used for analyses
performed under the Resource Conservation and Recovery Act (RCRA).
Two others are in the final stages of validation and several more
are in various stages of validation.
General Guidelines for Development of Screening Methods
The primary applicability that we, in the RCRA Program, see
for immunoassay methods is for quantitative screening purposes. By
quantitative screening, we mean setting a quantitative action level
(usually the regulatory action level), where a positive response
409
-------
means that the analyte is present at or above the action level, and
a negative response tells us that the analyte is either absent or
present below the level of regulatory concern. Analyses can be run
at multiple action levels giving a useful range of concentrations
for specific target analytes. For example, if we are mapping a
site contaminated with polychlorinated biphenyls (PCB) to determine
the extent to which it needs to be cleaned up, knowing where PCB
levels are <10 ppm, between 10 and 100 ppm, and >100 ppm can be
useful in planning and expediting the cleanup.
The OSW Methods Section distributes a letter, on request, to
potential developers of screening methods providing guidance on
what general validation criteria should be applied to a screening
method that will potentially be included in SW-846. While
screening procedures need not be fully quantitative, they should
measure the presence or absence of target analytes at or below
regulatory action levels. Therefore, initial demonstration of
method performance involves measuring the percentage of false
negatives and false positives generated using the procedure for a
single sample matrix. Data should be submitted for split samples
analyzed using the developer's technique and an appropriate SW-846
quantitative method. A candidate procedure should ideally produce
no false negatives and no more than 10% false positives.
Definition of a false negative is a negative response for a sample
that contains up to two times the stated detection level of the
target analyte(s). Definition of a false positive is a positive
response for a sample that contains analytes at one half the
detection level. Specific details on initial submission criteria
are included in the guidance letters, and I will not include them
in this article.
Other factors to be considered are interferences and matrix
effects. The effects of interferences, both positive and negative,
resulting from both target and non-target analytes should be
evaluated. Method performance and matrix effects in a variety of
RCRA matrices (e.g., soils, concentrated waste, ash, groundwater,
leachate, or wastewaters) should be demonstrated. It may not be
necessary or appropriate to spike all of the target analytes listed
within a chemical class for these initial method evaluations. If
field data is available, it would be valuable to be able to
compare the results obtained using the screening procedure-r.with
sample concentrations determined in a laboratory using SW-846
methods.
To summarize, the Methods Section does not require an
unreasonable body of data for the initial evaluation of new
techniques. Data will need to be submitted on the percentage of
false negatives, percentage of false positives, sensitivity to
method interferences, and matrix-specific performance data. In
addition to these data, the developer should also provide a
description of the procedure and a copy of any instructions
provided with the test kits.
410
-------
Validation Criteria for Immunoassay Methods
In addition to the guidelines for developing screening methods
in general, OSW, based on its own experience, has generated some
validation criteria specifically applicable to immunoassay methods.
These validation criteria are required to be submitted to OSW for
review for all immunoassay test kits, whether the kits are to be
the basis for a new method or as an alternative kit being added to
existing methods. The data needed for validation of immunoassay
methods that will be included directly in the method is as follows:
1) Cross Reactivity with similar analytes,
2) Cross Reactivity with dissimilar analytes which may be
reasonably expected to be found at waste sites,
3) False Negative/False Positive Rates,
4) Extraction efficiency (for soil test kits),
5) Performance data on spiked samples in environmental
matrices validated against standard SW-846 analytical
methods, and
6) Performance data on actual environmental field samples
validated against standard SW-846 analytical methods.
Since interferences can be a major problem in environmental
analyses, it is important to demonstrate that the analytes of
concern can be identified in the presence of similar analytes or
dissimilar analytes which may be present in environmental samples.
In many instances, substantial cross reactivity with other analytes
is a desirable situation. Examples of desirable cross reactivity
include sensitivity to esters of 2,4-dichlorophenoxyacetic acid
(2,4-D) as well as the 2,4-D, and for other 3-, 4-, and 5-membered
polynuclear aromatic hydrocarbons (PAH) when testing for
phenanthrene in a PAH screening method.
The false negative/false positive rate for a particular
immunoassay kit is very important. OSW screening methods are
designed to generate 0% false negatives and up to 10% false
positives at the regulatory action level. Slightly higher false
positive rates are tolerable, e.g. up to 25%. High false positive
rates, i.e. >25%, negate the cost effectiveness of the technique
because of the excessive numbers of confirmatory tests that would
need to be performed. High false negative rates, i.e. >5% at the
regulatory action level eliminate the potential use of the method
for regulatory purposes.
The extraction efficiency data is important for setting the
appropriate action level for a soil analysis. Recoveries are the
primary determining factor for making sure that the analyte of
411
-------
concern can be detected at the regulatory action level and for
minimizing false negative/false positive rates.
The performance data generated from environmental samples
spiked with the target analytes gives a good indication as to
whether or not an immunoassay method will work. However, the
performance generated in the field on real environmental samples is
the key determining factor on whether or not the immunoassay method
is sufficiently rugged to be included in SW-846 as an analytical
method.
Additional data that OSW requests, but does not include in the
method and treats as confidential business information (CBI),
includes dosage curves and the manufacturer's internal validation
and quality control criteria. The slope of the dosage curve can be
a good indication of whether or not an immunoassay method will
exhibit a high rate of false positives. Manufacturing quality
control and validation information gives a good indication as to
continued test kit availability.
Up to this time, all of the immunoassay test kits (10-15) that
the OSW has evaluated have been extensively tested and validated by
the manufacturers. EPA validation has primarily consisted of
confirmation of the manufacturers' results and performing some
additional testing on well-characterized environmental samples,
which are more easily available to EPA Regional laboratories.
Overview of the Regulatory Approval Process
As I have mentioned in other articles, RCRA methods are
published through an official regulatory notice printed in the
Federal Register incorporating them by reference into SW-846.
These methods must be issued as regulations, because any method in
the manual could be required for potential use under the few
sections of the RCRA regulations where the use of SW-846 methods is
mandatory. For the majority of RCRA applications, "any reliable
analytical method" may be used, i.e., a method which will determine
the analytes of concern in the matrix of concern at the regulatory
level of concern at the necessary confidence level.
I will now very briefly explain the regulatory process by
which methods are incorporated into SW-846, and the status and
applicable uses of methods at the various stages of the process.
The regulatory process for analytical methods briefly consists of
the following phases:
1) Technical Workgroup Review,
2) Proposed Regulation for Public Comment through a Federal
Register Notice,
412
-------
3) Response to Public Comment, and
4) Promulgated Final Regulation through a Federal Register
Notice.
The first phase in the method approval process is the only EPA
technical review during the approval process. All subsequent
reviews by EPA are administrative in nature. Public comments may
be either technical or administrative, and all must be addressed.
The most critical phase of the review process for methods is
the Technical Workgroup Review. The SW-846 Technical Workgroups
are made up of EPA chemists from across the Agency representing
Program Offices, ORD, Regional Laboratories, and Enforcement. The
methods are carefully reviewed to determine whether they
demonstrate appropriate performance and applicability to address
RCRA requirements. After Workgroup approval, the Methods Section
addresses any Workgroup comments on a method, both technical and
editorial, and prepares a "Draft" method. "Draft" methods are
available from the Methods Section on request, and can be used for
any RCRA application for which they are appropriate, and for which
the use of SW-846 methods is not mandatory.
The individual "Draft" methods are then assembled as a package
and a Federal Register Notice is prepared for public comment
proposing that the methods be added to SW-846. The methods are now
considered "Proposed" methods, but still have the same regulatory
status as "Draft" methods, i.e., they can only be used for RCRA
analyses where the use of SW-846 methods is not mandatory.
The public comment period is normally 45-60 days. Comments
are compiled and addressed. Any appropriate technical and
editorial changes are made to the "Proposed" methods in response to
public comment. The final edited methods are again assembled as a
package and the methods are added to SW-846 through a second
Federal Register Notice. The methods are now "Promulgated" or
"Approved" methods and can be used for any appropriate RCRA
application including those where the use of SW-846 methods is
mandatory -
Status of the EPA Immunoassav Methods Development Program
Several EPA Program Offices are investigating the potential
applicability of immunoassay methods to their programs. However,
the OSW is the first EPA Program Office to formally incorporate
these methods into its methods program. OSW began evaluation of
its first immunoassay method (for pentachlorophenol) in January of
1992, followed by two others in rapid succession. As of September,
1993, OSW has completed validation of four immunoassay methods
utilizing five kits, and is in the final stages of validating two
413
-------
additional methods and several additional kits for existing
methods. For an EPA regulatory Program Office, this is very rapid
progress. The four validated methods are
Method 4010: Pentachlorophenol (PCP) in Water and Soils by
Immunoassay,
Method 4020: Polychlorinated Biphenyls (PCBs) in Soil by
Immunoassay,
Method 4030: Total Petroleum Hydrocarbons (TPH) in Soil by
Immunoassay, and
Method 4035: Soil Screening for Polynuclear Aromatic
Hydrocarbons (PAHs) by Immunoassay.
The two methods in the final stages of validation, i.e., field
testing are
Method 4015: 2,4-D in Water and Soils by Immunoassay and
Method 4031: Soil Screening for BTEX By Immunoassay.
Currently, Methods 4020, 4030, and 4035, are "Draft" methods.
Method 4010 was promulgated on January 4, 1994 as Update IIA to the
Third Edition of SW-846 (59 PR 458-69) and is an "approved" method.
Methods 4020, 4030, and 4035 are available on request from the
Methods Section Office, and can be used for RCRA analyses for which
the use of SW-84G methods is not mandatory. Method 4010 is also
available from the Methods Section Office, and can be used for any
RCRA application for which it is appropriate. The final validation
studies for Methods 4015 and 4031 are expected to be completed with
Workgroup approval in 1994.
OSW is working with several of the manufacturers to initiate
validation studies on kits for many additional analytes, primarily
pesticides, which are coming on the market. Some of the target
analytes scheduled for evaluation include Alachlor, Aldicarb,
Carbofuran, Cyclodienes, DDT, Lindane, and mercury as mercuric ion.
OSW is also working with the manufacturers to develop new test kits
by delineating groups of target analytes that would be useful and
appropriate to determine by immunoassay methods. These analytes
include dioxins and furans, halogenated volatile solvents, phenols,
benzidines, and other amines.
Some of the other EPA Program Offices are also looking at
immunoassay methods to address some of their analytical
requirements. OPPTS is considering using immunoassay methods in
its Pesticide Registration Program. OW is beginning to look at
using the technology in both the Drinking Water and Wastewater
Programs. However, they may have to revise some of their
regulations to allow for the use of "less than" values in reporting
414
-------
Minimum Contaminant Levels (MCL) or look toward developing
quantitative immunoassays.
Potential Environmental Applications for Immunoassay Methods
OSW decided to take a cautious approach to the introduction of
a new technology to the environmental field, with which most
analytical practitioners were unfamiliar, and limit the initial
applications of immunoassay methods to quantitative screening. We
were aware that the technique had been used in Clinical
Laboratories for many years in both screening and determinative
applications. Since Regulatory Agencies tend to be slow to accept
new and different approaches to analysis, anyway, we decided to
take a "walk before you run" approach to introducing the new
methodology to the people actually doing site assessments and
cleanups.
The two primary applications of immunoassay methods in the
RCRA Program are mapping of contamination at well-characterized
sites slated for cleanup and monitoring the effectiveness of
cleanup activities. Immunoassay lends itself very well to these
two particular applications. It is not particularly applicable to
the identification and characterization of unknown contaminants at
waste sites when compared to much more comprehensive techniques
such as gas chromatography/mass spectrometry (GC/MS). However, for
monitoring applications of known contaminants, its specificity,
sensitivity, and cost effectiveness are excellent.
Over the past year, the general acceptability and willingness
to use immunoassay methods within the EPA Regions for RCRA and
Superfund applications has increased exponentially. A significant
factor in this change of attitude, in addition to OSW s attempts to
educate users in the applicability of the technique, is the specter
of shrinking budgets. Field people who are charged with actually
doing cleanups are looking for more cost effective ways to do their
jobs with less available money. A technique, such as immunoassay
methodology, which can generate high-quality results in real-time,
and can keep the bulldozers rolling can contribute significantly to
reducing the costs of cleanups, and is being looked upon more
favorably-
The initial application of immunoassay technology in the RCRA
Program was for determining compliance at wood surface treating
facilities with PCP regulatory limits. The selectivity and
sensitivity of the immunoassay method easily met the regulatory
action limit of 0 .1 ppm. Use of the PCP immunoassay method (Method
4010) for compliance monitoring was encouraged by OSW and the
method was proposed for inclusion in SW-846 as a part of the Wood
Surface Treatment Rule.
The major applications for which immunoassay methods
are
415
-------
currently used in the RCRA Program are site mapping and monitoring
cleanups at sites contaminated with PCBs. Use of the PCB method
(Method 4020) has resulted in cost savings at many sites in several
Regions. The speed and low cost of the test allows for more
extensive mapping of contamination at a site, because many more
samples can be analyzed on site, thus generating a more detailed
map of the site. This results in lower cleanup costs, since the
cleanup efforts can be directed only at the places that need to be
cleaned up, instead of to a broader area. The design of the method
allows for rapid determination of whether or not the site cleanup
level has been met, thus reducing costs of cleanup in both time and
equipment. With the recent availability of the PAH method (Method
4035) , the technique is now beginning to be used on sites
contaminated with PAHs.
Another major application within OSWER is for mapping and
cleanup of sites contaminated with petroleum hydrocarbons from
leaking storage tanks for the Office of Underground Storage Tanks
(OUST). The TPH method (Method 4030) is effective for determining
gasoline, diesel, kerosene, and jet fuel at required cleanup
levels.
Additional analytes will be targeted as new methods are
developed and validated. Eventually, OSW intends to perform
quantitative analysis either using immunoassay methods for direct
quantitation or as concentration techniques using affinity
chromatography, with quantitation by existing techniques, e.g. HPLC
or GC/MS. The latter approach will be particularly effective for
doing analyses where multiple analytes within a class need to be
individually identified.
Barriers to Use of Immunoassay Methods
There have been some initial barriers to getting immunoassay
methods accepted for routine use in the environmental community.
These barriers have been both technical and cultural in nature.
The technical barriers include lack of knowledge about analytical
options; use of expensive time-consuming methodology when more
efficient methodology is available; poor planning of the initial
analytical scheme; failure to identify proper questions to be
answered resulting in generation of data inappropriate to address
the problem at hand. Cultural barriers include inappropriate or
excessive regulatory restrictions on 'use of new methods, e.g.
requiring the use of only promulgated methods for program
applications that do not have these requirements, and requiring the
use of expensive broad-scope methods, e.g., GC/MS, for limited
monitoring, applications for only a few known and well-characterized
analytes.
An additional issue of concern was whether the Regulatory
Program Offices could live with analytical values that were not a
416
-------
specified number, i.e., a less than value (usually the regulatory
action level) vs. a definite number (0.1 ppm) or a range of values
(>5 and <50) . We, in the RCRA Program, decided that we could
indeed use these values to answer the basic questions for which
these analyses were performed, i.e., Have we attained our cleanup
criteria? Where do we have to focus our cleanup efforts? We
decided that our normal operating procedures for confirming
quantitative screening results would be to use the standard
reference method to confirm positives and to spot check a certain
percentage (usually 10%) of negative results.
Other Program Offices in EPA, such as the OW may have some
restrictions in their current regulations which require them to
generate a definite analytical value. If this is indeed the case,
their focus would be on quantitative immunoassay methods rather
than screening methods.
OSW has initiated a major effort to train EPA permit writers,
enforcement people, and others who deal with analytical methods in
their jobs in the regulatory aspects of using RCRA methods.
Historically there have been problems where only promulgated
methods were allowed to be used in many applications under the RCRA
regulations where this was not a requirement. The Methods Section
is in the process of developing a formal training program for RCRA
personnel in the Regions and at Headquarters to make them aware as
to which methods are allowable and appropriate to use under the
RCRA regulations in both mandatory and non-mandatory applications,
and how to prepare efficient, cost effective sampling and analysis
plans.
State programs are a little more difficult. Since RCRA is a
Federal Program which has been passed down to most States to
administer, the State regulations can be more restrictive and tend
to vary greatly. Some States mandate the use of SW-846 methods for
all RCRA analytical applications within the State. Flexibility
within State Programs varies from allowing only the use of
promulgated methods to using any method that may be appropriate for
an application. Through dialogue with the EPA Regions and
Headquarters, some of the States are beginning to take an interest
in utilizing immunoassay methods. TPH analysis is the major focus
right now in State Programs, since it is not regulated at the
Federal level. Several States are beginning to adopt Method 4030
for use in their Underground Storage Tank (UST) Programs, e.g.
Georgia and California.
Summary and Conclusions
The OSW immunoassay methods program was initiated in January,
1992, with the evaluation of the screening method for PCP. Methods
for PCBs and TPH followed soon after. These three screening
methods were recommended for inclusion in SW-846 by the Technical
417
-------
Workgroup. A fourth method for PAHs, also received Workgroup
approval last summer. Two other methods are in the final stages of
validation, with several more beginning the validation process.
For a Regulatory Program, immunoassay methodology has advanced very
rapidly -
There was initially a general reluctance among the regulatory
and regulated community to use immunoassay methods, even for
applications for which they were appropriate. This was due to a
lack of knowledge about the technology and a belief among both the
regulators and regulated community that only promulgated SW-846
methods could be used for all RCRA applications.
The climate has changed considerably during the past year
regarding the use of screening methods in general, and immunoassay
methods in particular in the environmental community- We have
noticed a much greater willingness for EPA Regional and some State
regulators to allow for the use of immunoassay methods in their
RCRA Programs. Apparently, the dissemination of information about
the effective performance of immunoassay methods and a budget
crunch which drives both regulatory and remediation personnel to
look for more cost effective means to do their jobs have begun to
have an impact in the environmental community.
The future looks bright for the environmental application of
immunoassay methodology. Many other Federal Agencies with massive
cleanup problems, e.g. The Department of Energy (DOE) and The
Department of Defense (DOD), have become interested in the
technology for its overall utility in significantly cutting the
costs of cleanup operations. Within the next two years, we expect
to focus more on quantitative immunoassay methods. When the
environmental community has reached an appropriate comfort level
with using immunoassay screening methods, we will introduce the
quantitative methods. The EPA will continue its cooperative effort
with the immunoassay manufacturers and other methods developers to
develop the methods that are needed for its environmental programs.
418
-------
62
QUANTITATIVE ENVIRONMENTAL IMMUNOASSAY, THE NEXT STEP?
Dr. P.J. Marsden and S.F. Tsang, Environmental Health and Safety Group,
Science Applications International Corporation (SAIC), 4161 Campus Pt. Ct.,
San Diego, California 92121
ABSTRACT
Immunoassay or enzyme-linked immunosorbant assay (ELISA) is a powerful
tool for providing analytical support during environmental monitoring and
cleanup projects. Immunoassays can produce accurate, cost-effective and
sensitive measurements of specific contaminants. Immunoassay methods
currently promulgated for the RCRA program or being considered for SW-846
are suitable only for sample screening. Analysts must demonstrate the
suitability of quantitative immunoassay techniques before they can be used to
support environmental decisions. Demonstration of the suitability of quantitative
immunoassay will require rigorous comparisons between those techniques and
promulgated chromatographic methods. Refinements in immunoassay
methods will be required to ensure that measurements are quantitative
including improvements in extraction, measurement and QA procedures.
Data are presented in this paper from two studies that provide a side-by-side
comparison of immunoassay with GC/ECD analysis (Method 8081) for
multicomponent analytes. The first study involved the analysis of Toxaphene;
the second study involved the analysis of PCBs. In both studies, results are
provided for laboratory spiked and in situ contaminated soils. Soil was
contaminated with Toxaphene through pest control activities in New Mexico
(i.e., sheep dips for scabies). Comparison of the Toxaphene concentrations
measured using immunoassay and using Method 8081 provided an excellent
correlation over a concentration range of 0.5 to 400 ppm. Soil and waste
contaminated with PCBs were provided as reference materials by the Research
Technology Corporation of Laramie, WY.
The results of these studies illustrate the suitability of quantitative immunoassay.
The paper also presents a discussion of some of the current barriers to the
adoption of quantitative immunoassay as an environmental tool, including the
variety of RCRA matrices, the need to train staff to perform and interpret
immunoassays as well as some requirements for validating these procedures.
Measurement scientists must overcome these barriers before quantitative
immunoassay can be adopted for environmental monitoring.
INTRODUCTION
Immunoassay (enzyme-linked immunosorbent assay [ELISA]) is gaining
acceptance as a screening technique for measuring specific compounds in
environmental samples. Immunochemical analyses are suitable only for the
419
-------
analysis of specific target analytes for which specific antibodies have been
prepared. Therefore, immunoassay is not appropriate for initial site
characterization. However, once the nature of contamination at a site has been
established using conventional chemical analysis, immunoassay field
screening may be used to (1) reduce the number of expensive chromatographic
analyses required for site monitoring, (2) focus resources on the worst portion of
the site, (3) analyze more samples or (4) obtain analytical data more rapidly.
While immunoassay is usually considered as a field technique, the use of
immunoassay as a laboratory screening procedure could benefit our industry by
(1) reducing the need to analyze samples with less than detectable levels of
contamination and (2) reducing the need to dilute samples with very high levels
of contamination. SW-846 immunoassay methods have been proposed or
promulgated for pentachlorophenol (Method 4010), polychlorinated biphenyls
(PCBs, Method 4020) and petroleum hydrocarbons (Method 4030). Additional
techniques are being considered by the Office of Solid Waste Organic Methods
Workgroup for possible inclusion in future updates of SW-846. The Office of
Pesticide Programs is testing immunoassay techniques for the analysis of
Alachlor, Metalochlor, Atrazine and 2,4-D. The Office of Water is considering
the use of immunoassay for several target analytes including cyanazine.
Immunoassay differs from conventional chromatographic techniques in both the
measurement technique and the sample preparation procedures. Methanol
extraction is specified for immunoassay procedures rather than methylene
chloride/acetone extraction because methanol is more compatible with the
biochemicals (e.g., antibodies and proteins) required for immunoassay. The
use of different extraction solvents for immunoassay and GC analyses results in
two major concerns, (1) data obtained using the two techniques may not be fully
intercomparable and (2) a requirement for two different extraction procedures
may limit the acceptance of immunoassay as a laboratory screening procedure.
In contrast to environmental applications where they are applied only as
screening procedures, immunoassays are routinely used for quantitative clinical
analyses. This paper provides quantitative environmental data comparing
immunoassay with the chromatographic procedure, Method 8081. We will also
provide data that demonstrates that immunoassay kits are suitable for screening
laboratory extracts (methylene chloride/acetone) after exchange to methanol.
Finally, we will provide discussion of some of the barriers to adopting
quantitative immunoassay for environmental applications.
EXPERIMENTAL DESIGN
SAIC conducted a side-by-side comparison of analyses of chlorinated
multicomponent pollutants using immunoassay techniques and GC/ECD.
These studies were funded separately under contract to the Millipore
Corporation and EnSys, Inc. Samples included both soil spiked in our
laboratory and contaminated soils collected in the field. Toxaphene
420
-------
immunoassays were performed by Immunosystems in Maine using the Millipore
EnviroGard kit. PCB immunoassays were conducted by SAIC using the EnSys
PCB RIS0-™ kit in our San Diego laboratory according to Method 4020. All of
the GC/ECD analysis were conducted in the SAIC Methods Laboratory at San
Diego according to Method 8081. Soil samples were extracted by Soxhlet
extraction using methylene chloride/acetone (Method 3540) or by methanol
swirled in a polyethylene bottle with five stainless steel ball bearings using the
extraction procedure described in Method 4020.
Toxaphene
Initial studies were conducted using San Diego (low-organic) soil spiked with
Toxaphene at 0.25, 0.5, 1.0, 2.5, and 5 ppm. The Toxaphene standard was
obtained from Supelco (cat# 4-8700M, 2000 |ig/mL in methanol). Aliquots of
the spiked soils were extracted using Method 3540 (Soxhlet extraction) with a
solvent mixture of methylene chloride/acetone (1:1 v/v). Three replicate
extractions were performed on each level of fortification and each extract was
analyzed by Method 8081. Individual aliquots of each spiked soil were also
analyzed using the Millipore EnviroGard Toxaphene kit.
Millipore provided 31 real-world samples contaminated with Toxaphene from
New Mexico. The site had been used for sheep dips where Toxaphene had
been used to remove fleas, ticks and other exoparasites from the animals. The
soil samples analyzed by GC/ECD were extracted in a Soxhlet apparatus using
methylene chloride/acetone (Method 3540). Each extract was concentrated to a
volume between 1 ancj 10 ml_ using a Kudurna-Danish apparatus. Extracts
were exchanged to hexane under a stream of dry nitrogen and analyzed by
Method 8081. Two matrix spike/matrix spike duplicate pairs and blanks were
analyzed with the set of 31 samples.
SAIC also demonstrated the suitability of a back extraction technique for
preparing methanol extracts for GC/ECD analysis. This was accomplished by
preparing a Toxaphene standard in methanol and partitioning the target analyte
into hexane, a solvent more suitable for GC/ECD analysis. The solvent partition
was accomplished by adding 1 ml_ of the methanol sample to 5 ml_ of distilled
water, extracting the aqueous methanol two times with 5 ml_ hexane. The
resulting hexane extract was concentrated to 1.0 ml_ for GC/ECD analysis.
PCBs
Initial experiments were performed using laboratory spiked soils. San Diego
(low-organic) soil was spiked with Aroclor 1242 at 4 ng/g and 50 jig/g and with
Aroclor 1260 at 0.8 |ig/g, 10 |o.g/g, 50 ng/g using standards purchased from Ultra
Scientific. Spiking solutions were prepared in methanol, added to soil in a clean
jar and stirred. Homogeneity of the spiked soil was ensured by mixing each
spiked sample on a rock-tumbler roller for at least one hour.
421
-------
Five portions of each spiked soil were extracted using Method 3540 (Soxhlet)
with methylene chloride/acetone (1:1 v/v). The extracts were then split. One
aliquot from each extraction was blown to dryness, redissolved in methanol and
analyzed using the immunoassay procedure for PCBs in proposed Method
4020. The other methylene chloride/acetone aliquot was exchanged to hexane
and analyzed using GC/ECD (Method 8081). GC/ECD conditions are provided
in Appendix 1. Five replicate extracts and one blank were analyzed by each
method at each level of fortification in order to compare the accuracy and
precision of PCB measurements using immunoassay and GC/ECD.
One to seven days later, three portions of each of the spiked soils were
extracted using methanol in a polyethylene jar according the procedure
described in Method 4020. One aliquot from each extraction was analyzed
using immunoassay procedure. A second 1.0 mL aliquot of each methanol
extract was diluted with water and the PCBs were back extracted into hexane.
The hexane fractions were analyzed using GC/ECD (Method 8081). These
analyses provided a direct comparison of the measurement of PCBs extracted
with methanol using immunoassay and GC/ECD.
The performance of immunoassay was also compared with GC/ECD analysis
using five real-world samples contaminated with PCBs. Two of these samples
were reference materials provided by the Research Technology Corporation
(RTC of Laramie, WY); EnSys provided the other three real-world samples as
unknowns. The design of the experiments using the real-world samples was
similar to that used for spiked soils. Five portions of soil were extracted using
Method 3540 (Soxhlet) with methylene chloride acetone (1:1). The extracts
were split. One aliquot of each extract was blown to dryness, redissolved in
methanol and analyzed using the EnSys PCB RISc™ test kit procedure
(proposed Method 4020). The other aliquot from each extraction was
exchanged to hexane, subjected to sulfuric acid cleanup (proposed Method
3665) and analyzed using GC/ECD (Method 8081). Previous studies at SAIC
have confirmed that the sulfuric acid cleanup does not destroy PCBs and can
significantly improve the quantitation of PCBs. Five replicate Soxhlet
extractions and one blank were analyzed for each of the RTC samples.
However, only one Soxhlet extract was analyzed for each of the EnSys samples
because the amount of material was insufficient for replicate analyses.
One to seven days later, three portions of each of the RTC soils and one portion
of the EnSys soils were extracted using methanol. Those extracts were split.
One aliquot from each methanol extract was analyzed using the immunoassay.
The PCBs in a 1.0 mL aliquot of each of the methanol extracts was back
extracted into hexane. The hexane extracts were analyzed using Method 8081
in order to compare, the measurement of PCBs in samples using immunoassay
and GC/ECD.
422
-------
It was also demonstrated that immunoassay measurement is not affected by a
small amount of methylene chloride (1 %) that might remain after solvent
exchange by spiking aliquots of each of the methanol extracts with 1 percent
methylene chloride prior to some immunoassays.
RESULTS
Toxaphene Analyses
Soil samples spiked with Toxaphene and real-world samples contaminated
with Toxaphene were analyzed using GC/ECD and immunoassay. Three
aliquots of spiked samples were extracted using Method 3540 (Soxhlet) and
analyzed using Method 8081. An aliquot of each extract was sent to
ImmunoSystems at Maine and analyzed using the Millipore Toxaphene
EnviroGard Toxaphene kit. This kit has significant cross-reactivity with other
cyclodiene insecticides (e.g., Aldrin, Dieldrin, and Heptachlor), none of which
were detected in these samples. However, significant amounts of DDE were
observed in some samples using GC/MS. Measurement of Toxaphene was
accomplished using a GC/ECD calibrated with the Toxaphene standard; a
single response factor was calculated using the five major peaks present in that
analytical standard. Thirty one real world samples were analyzed by Millipore
Inc., and were also extracted and analyzed using Method 8081 in San Diego.
The Toxaphene concentrations measured using Soxhlet extraction (Method
3540) and GC/ECD (Method 8081) were compared with quantitative
immunoassay as part of the performance testing of the method. Results for the
analysis of spiked soil determined by SAIC San Diego and by ImmunoSystems
are provided in Table 1. Results for the analysis of New Mexico soil using both
methods are provided in Table 2. The correlation between GC/ECD analyses
and immunoassay results are represented graphically using log/log axes in
Figure 1
TABLE 1. RESULTS FOR SOIL SPIKED WITH TOXAPHENE
Spike in ppm (jig/g) Immunoassay (jig/g) GC/ECD range (jig/g)
0.25 0.27 0.16-0.20
0.50 0.66 0.30 0.37
1.0 1.02 0.77 1.0
2.5 2.8 2.5 3.1
5.0 6.7 5.4 -5.7
RSDs for triplicate GC/ECD analyses 2.7 12.4%
423
-------
TABLE 2. RESULTS FOR SOIL CONTAMINATED WITH TOXAPHENE
Sample number GC/ECD, fj,g/g Immunoassay, fig/g
28, 89 0.09 J 0.3
70,104 0.04 J 0.9
54, 89 0.04 J 0.8
103, 50 0.01 J 0.2
10,30 40 58
45,33 19.3 54.8
Nazalini soil #12 <0.5 0.2
0,33 <0.5 1.7
23,104 0.26 J 1.1
78,33 1.0 2.6
64, 5 0.14 J 2.1
53, 75 0.27 J 11
35, 75 27.2 3.8
17,75 0.14 J 0.9
65, 33 0.48 J 2.8
82,75 0.21 J 1.8
19,50 4.8 6.0
97,104 0.05 J 0.6
48,104 0.05 J 1.1
0,50 1.3 2.3
102,75 0,15 J 0.3
84, 50 0.06 J 0.9
25,33 88.3 130
0,75 0.5 1.9
12,40 pit 34.1 45.5
0,89 0.16 J 6.9
0,104 0.88 2.1
98, 89 0.41 3.4
104,33 0.30 J 0.7
76,89 0.10 J 5.8
40, 50 324 460
J = an estimated value. This is used to indicate the result is less than the lowest
calibration point but greater than the method detection limit.
Sample numbers were assigned by ImmunoSystems and represent a grid at
the site.
424
-------
FIGURE 1. COMPARISON OF IMMUNOASSAY AND GC/ECD ANALYSES
3
2.5-
2-
1.5-
o
Q.
o 0.5^
o-
-0.5-
-1
Toxaphene Concentrations 0 - 400 ppm
Comparison of the log concentration
determined using immunoassay (log
Millipore) with the log concentration
determined using GC/ECD.
''' i'''' i''' • i''
-2 -1.5 -1 -0.5
0 0.5 1
log GC/ECD
1.5 2 2.5 3
425
-------
The quality of the measurements of Toxaphene in soil was confirmed using
matrix spike and matrix spike duplicate analyses.
TABLE 3. MATRIX SPIKE RECOVERIES FROM NEW MEXICO SOIL
Sample ID
M16 70,104
M24, 0,75
2M2 (methanol ext.)
MS % recovery
115.9
112.0
40.0
MSD % recovery
115.6
106.9
37.0
% RPD
0.33
4.7
7.8
The performance of a hexane back extraction of Toxaphene from methanol
extracts was established experimentally. A 1 |ig/mL Toxaphene standard which
was analyzed by GC/ECD. The determinations gave a mean recovery of 87%
for five replicate extractions of the Toxaphene standard. The performance of the
back extraction is presented below.
TABLE 4. BACK EXTRACTION OF TOXAPHENE FROM METHANOL
cone.
%
rec
ext-1
0.83
82.5
ext -2
0.89
89.3
ext -3
0.86
86.2
ext -4
0.84
83.7
Mean rec.=
std =
%RSD =
ext- 5
0.95
94.7
87.3
4.9
5.6
PCB Analyses
Establishing the suitability of immunoassay as a laboratory screening
procedure requires a comparison of the measurement of PCBs using Soxhlet
extraction using methylene chloride/acetone (Method 3540) and immunoassay
(Method 4020). Direct injection of two different extracts is not an appropriate
means of comparison because the use of different injection solvents may
significantly affect GC/ECD results. Difficulties arising from differences in
injection solvents were minimized by using the hexane back extraction
426
-------
procedure described in this paper for three 1.0 ml portions of methanol spiked
with Aroclor 1242 at 1.0 |u,g/mL Those results are provided in Table 5.
TABLE 5. AROCLOR RECOVERY AFTER SOLVENT EXCHANGE
area
percent
recovery
1 |ig/g std
Aroclor 1242
392988
NA
extraction 1
418996
106.6
extraction 2
431546
109.8
extraction 3
430156
109.5
Extracts of soil spiked with RGBs (Aroclor 1242 and 1260) were analyzed using
both the PCB RISc™ kits and GC/ECD in order to compare unambiguously the
results of the different extraction and measurement techniques. These data are
reported in Table 6 for Aroclor 1242 and in Table 7 for Aroclor 1260. The first
column of each table gives the spiking level, column two gives the results for
immunoassay and GC/ECD measurements for PCBs in methylene
chloride/acetone extracts and column three gives the immunoassay and
GC/ECD measurements for PCBs in methanol extracts. These data
demonstrate that methanol extracts somewhat less than half of the PCBs than
does methylene chloride/acetone. These data also document the slightly
positive bias of the immunoassay kits for PCBs (e.g., 5 of 5 kits were greater
than 2 |ig/g when the GC/ECD results were 3.8 (ig/g, 2 of 3 kits were greater
than 2 (ig/g when the GC/ECD results were 1.5 |ig/g). These data demonstrate
the suitability of immunoassay as a conservative laboratory screening
procedure that is unlikely to produce false negatives.
TABLE 6. ANALYSIS OF AROCLOR 1242 IN SPIKED SOIL
Spike
Level
^g/g
4
50
DCM/acetone, |ig/g
Immunoassay
>2, <25
n = 5
>2, >25
n = 5
GC/ECD
3. 78 + .36
(recovery
94.4%)
45.6 ±.7
(recovery
91.1%)
MeOH, |ig/g
Immunoassay
>2, <25
n = 2
<2, <25
n = 1
>2, >25
n = 3
GC/ECD
1.5 ±.3
(recovery
37.2%)
22 ± 1 .4
(recovery
44%)
427
-------
TABLE 7. ANALYSIS OF AROCLOR 1260 IN SPIKED SOIL
Spike
Level
M-g/g
0.8
10
50
DCM/acetone, |ig/g
Immunoassay
>0.4, <5
n = 3
>0.4, >5
n = 2
>0.4, >5,
n = 5
>5, >25
n = 5
GC/ECD
0.78 ±.12
(recovery
98%)
10.3 ±1.3
(recovery
1 03%)
54.5 ±3.0
(recovery
1 09%)
MeOH, |ig/g
Immunoassay
>0.4, <5
n = 2
<0.4, <5
n = 1
>0.4, >5
n = 3
>5, >25
n = 3
GC/ECD
0.35 ±0.7
(recovery
44%)
4.9 ±.4
(recovery
49%)
15.2 ±2.5
(recovery
30%)
These results demonstrate that immunoassay can be used to produce
internally consistent Aroclor data. However, it is recommended that more than
one ELISA determination be made when samples are near the reporting limit.
The results for the immunoassay of the methylene chloride/acetone extracts
indicate that ELISA techniques may be used as laboratory screening or even
quantitative procedures for PCBs. Evaluation of these data generated near the
detection limits also support the OSW requirement that at least 10% of the non-
contaminated samples should be submitted for GC/ECD analyses. Each set of
immunochemical measurements included methanol extracts of the samples
spiked with 1 percent methylene chloride. This experiment was designed to
ensure that any residual methylene chloride remaining after solvent exchange
to methanol would not effect ELISA measurements. These data indicate that
immunoassay can tolerate low levels of methylene chloride.
SAIC performed replicate analyses of two RTC samples contaminated with
PCBs using immunoassay and GC/ECD. These results are presented in Tables
8,9 and 10. The values determined for Aroclors in the RTC samples are 40.5
|o,g/g Aroclor 1242 in RTC sample #0912 (Table 8)and 1.5 |ig/g Aroclor 1254.in
RTC sample #4179 (Table 9). RTC sample #4179 also contained high
concentrations of PAH. The EnSys samples contained Aroclor 1260 (Table 10).
The GC/ECD results are fully consistent with the immunoassay results for each
of these samples. The methylene chloride/acetone extraction gave somewhat
higher recoveries than methanol extraction. These higher recoveries are best
428
-------
evidenced by the fact that all of the immunoassays of the methylene
chloride/acetone extracts gave >25 |ig/g for RTC sample #0912 (40 jig/g) while
that concentration was detected in only 1 of 3 methanol extracts. Further
evidence is provided by the fact that all 5 of the methylene chloride/acetone
extracts of RTC sample 4179 (1.5 u,g/g) were assayed at >2 jig/g while 3
methanol extracts were assayed at <2 |ig/g. Both methanol and methylene
chloride/acetone extracts produced false positives by immunoassay when
measuring PCBs in EnSys sample #126. False negatives were not observed
during the analysis of any of these soil and waste samples.
These results demonstrate that measurements made using immunoassay and
GC/ECD analyses are generally intercomparable. These data also indicate that
immunoassay may be used to produce reliable environmental data. They
further demonstrate that methylene chloride/acetone extracts can be blown to
dryness and reconstituted in methanol as part of a laboratory screening
procedure using immunoassay. Use of this technique for screening
environmental samples would have reduced our project analytical costs and
produced no false negatives. The limited number of false positives would have
resulted in some additional GC/ECD analyses which seems quite reasonable
given our mission of generating data and methods that are used to monitor
pollution which poses a hazard to human and environmental health.
TABLES. ANALYSIS OF RTC SAMPLE #0912
Immunoassay results jig/g, Aroclor 1242 kit
DCM/acetone
>2, >25
n = 5
methanol
>2, >25, n = 1
>2, <25, n = 2
methanol +
1 % DCM
>2, >25, n = 3
GC/ECD results
jig/g, n = 5
mean + s.d. Aroclor
1242
40.5 ± 1.1
429
-------
TABLE 9. ANALYSIS OF RTC SAMPLE #4179
Immunoassay results ^ig/g, Aroclor 1260
DCM/acetone
>2, <25
n = 5
methanol
<2, <25, n = 3
methanol +
1 % DCM
<2, <25, n = 1
>2, <25, n = 2
GC/ECD results
jig/g, n = 5
mean + s.d. Aroclor
1254
1.5±.1
TABLE 10. ANALYSIS OF EnSys SAMPLES (SINGLE DETERMINATIONS)
Sample
number
126
254
148
Immunoassay results |ig/g, 1260 kit
DCM/acetone
>5, >25
>5, >25
>5, >25
methanol
>5, >25
>5, >25
>5, >25
methanol +
1% DCM
>5, >25
>5, >25
>5, >25
GC/ECD, Lig/g
10.5, 1260
49.6, 1260
37.8, 1254
DISCUSSION
Clinical immunoassay is not an exact model for the requirements of the RCRA
program . Unlike tissue, urine or feces, hazardous waste and environmental
samples can pose very different, site-specific matrix problems. Adopting
quantitative ELISA techniques as part of SW-846 will require proof that
immunoassay can be used to measure specific toxicants in a variety of sample
matrices. The authors believe that there are four major issues that must be
addressed before immunoassay is viewed as a quantitative technique for RCRA
matrices.
Intercomparability of data - Analysis of Toxaphene and PCBs using
immunoassay and GC/ECD analyses involves two different measurement
technologies and two different extraction techniques. The EPA requires
comparison of the performance of both methods using a variety of sample
matrices. The fact that methanol extraction is less effective than methylene
chloride/acetone will require specific method performance data to establish that
data obtained using immunoassay kits can be compared with GC/ECD results.
430
-------
Suitability for waste matrices Immunoassay has not been demonstrated as
suitable for the tremendous variety of waste matrices considered under the
RCRA program. Data on the suitability of immunoassay should be collected
over time as more investigators use immunoassay to measure toxicants in
waste and soil samples.
Performing immunoassays a number of the operations required for
immunoassay (e.g., use of the repeating pipettor) are not generally used in
environmental laboratories. Analysts in environmental laboratories will need to
be trained in biochemical techniques in order to use immunochemical
techniques for quantitative measurements during site monitoring or for
laboratory screening procedures.
Interpretation of data interpretation of immunochemical data is not
straightforward. Because the assays work by enzyme-linked visualization
techniques after competitive binding with the antibody, the amount of color
developed in a tube is inversely related to the amount of analyte in the sample.
More color means less analyte. Analysts and data reviewers will need to be
trained to interpret these data and to use dual wave length data for
immunochemical measurements. The measurement technique used in
immunoassay are very different from chromatography methods.
CONCLUSION
These studies demonstrate that immunoassay kits can be used for
measurement of multicomponent organochlorine pollutants in soil. Use of
immunoassay for the analysis of Toxaphene and PCBs may actually result in
less measurement uncertainty and interlaboratory differences in quantitation
than conventional GC/ECD analysis (Method 8081) because the GC/ECD
method requires significant analyst interpretation of peak patterns in order to
identify and quantitate the target analytes.
Immunoassay is biased to produce false positives for samples with
contaminants near the detection limit. This bias makes it unlikely that a
contaminated sample will test as a non-detect when it is contaminated above
the limit of detection. However, analysts should expect to see both positive and
negative results when contamination levels are near the method detection limits
for samples analyzed by immunoassay.
The studies reported here also demonstrate that methylene chloride/acetone
sample extracts can be blown to dryness and reconstituted in methanol prior to
an immunoassay laboratory screening procedure. Experiments established
that immunoassay is relatively insensitive to small amounts of methylene
chloride that might be left from incomplete blow down of sample extracts.
431
-------
APPENDIX
GC/ECD Conditions, Method 8081 for Toxaphene and PCBs:
Inj.temp. 200°C
Det. temp.(ECD) 320°C
Initial temp: 150°C
initial time: 0.5 min
rate: 5°C/min
final temp.: 275°C
final time: 8 min
carrier (N2) 5 ml/min
makeup (N2) 55 ml/min
432
-------
63
TOXAPHENE IN SOIL/TRICHLOROETHYLENE/RAPID DIOXIN
SCREENING BY IMMUNOASSY. T.S. Fan, B. Young, D. Grouse, H. Allen,
R.O Harrison, H. Shirkhan, R.E. Carlson
This presentation was a combination of T.S. Fan's Organic paper #s 84 and 88,
and R.O. Harrison's Organic paper #100.
433
-------
64
TNT AND RDX BY IMMUNOASSAY. G. B. Teaney, J.M. Melby, J.W Stave,
R. T. Hudak
This presentation was a combination of G. B. Teaney's Organic paper #s 90 and
94.
434
-------
65
COMMERCIAL LABORATORY POSITION ON THE USE OF SW-846
Jerry L. Parr, Chair, Technical Committee, International Association of
Environmental Testing Laboratories (IAETL), 1911 N. Fort Myer Drive, Arlington,
Virginia 22209
ABSTRACT
Test Methods for Evaluating Solid Wastes. Physical/Chemical Methods (SW-846) is
stated to provide test procedures and guidance that are recommended for use in
conducting the evaluations and measurements needed to comply with the RCRA
regulations. While several of the hazardous waste regulations under Subtitle C
of RCRA require that specific testing methods described in SW-846 be employed for
certain applications, the document states that any reliable analytical method may
be used to meet other requirements.
The International Association of Environmental Testing Laboratories (IAETL) has
been in existence for six years, has over 170 members in the U.S. and Canada, and
represents more than half the revenue generated in the environmental analytical
testing industry. lAETL's Technical Committee supports EPA's position on the use
of SW-846 as a viable guidance document.
The objective of our presentation is to share our concerns regarding the current
confusion on the use of SW-846. The presentation will address prescriptive use
versus guidance, varying levels of interpretation by laboratories, regulators,
and data users of SW-846, and areas of concern.
We intend to present ideas that will clarify the use of SW-846 as guidance and
speed the process for incorporation of new technology for cost-effective
monitoring.
435
-------
66
SURROGATE-BASED CORRECTIONS 0? DATA FOR METHOD 5032
Michael H. Hiatt and Carole Farr,
United States Environmental Protection Agency
EMSL-LV, Nevada 89119
ABSTRACT
As part of the Environmental Monitoring Systems Laboratory-Las
Vegas Quality Assurance Research Program, the quality assurance
necessary for analytical methods is investigated and optimized.
We are reporting on work that was conducted to identify compounds
which could be utilized as surrogates for a vacuum distillation
procedure (Method 5032). Recovery of analytes using Method 5032
is found to be related to specific physical properties of the
analytes. Therefore, surrogate compounds selected to assess
losses related to individual physical properties, are used to
predicate analyte recoveries. Such predicted recoveries can be
expected to exceed 95% of the experimentally measured recoveries.
This approach is shown to be applicable to a variety of samples.
436
-------
67
STATIC HEADSPACE ANALYSIS OF SOIL SAMPLES
J. Purcell, A. Tippler, G. McClure and J. Ryan, The Perkin-EImer Corporation, 761 Main
Avenue, Norwaik, CT 06859
The analysis of soli samples for volatile organic compounds (VOCs) can be a difficult
procedure. Use of SW-846 method 8260 requires that VOCs in samples be introduced to a gas
chromatograph either by purge-and-trap or direct injection. These sample introduction
techniques can cause significant problems with inter-sample analyte contamination and carry
over. In addition, the transfer of samples from collection container to dynamic purge vessel can
result in significant losses of VOCs.
A means of alleviating these problems has been developed. Using automated static headspace
instruments, it has been shown that soil samples can be analyzed for their volatile constituents
without fear of cross contamination and with a minimum of sample preparation. Laboratory
productivity is improved.
This static headspace VOC method is performed by adding 3 mL of chilled water directly to a
chilled soil sample. The headspace vial is then sealed and the sample sonicated for 15
minutes, following which it was transferred directly to the automated headspace analyzer
without further preparation.
While method performance can vary depending on the sample type, data will be presented for
the analysis of VOCs from a range of soil types showing that recovery and reproducibiiity
statistics compare favorably with those of method 8260.
437
-------
68
SOLIDS PHASE EXTRACTION OF TCLP LEACHATES
Craig G. Markell, 3M-New Products, St. Paul, Minnesota, Li Song, 3M-New Products,
St. Paul, Minnesota and Rich Pieper, 3M-New Products, St. Paul, Minnesota
ABSTRACT
A method has been developed to use solid phase extraction disks for TCLP leachates.
Results will be presented on pesticides, herbicides, acid extractables, and base neutral
extractables.
438
-------
69
Dianne A. Bennett. Barry Lesnik* and S. Mark Lee
Chemistry Laboratory Services Branch *Organics Method Program
California Department of Food and Agriculture USEPA Office of Solid Waste
Sacramento, CA 95832 Washington, D.C. 20460
SUPERCRITICAL FLUID EXTRACTION OF ORGANOCHLORINE PESTICIDE
RESIDUES FROM SOILS
An analytical method has been developed for the supercritical fluid extraction of
organochlorine pesticide residues from soils. The method development strategy will be outlined
to demonstrate how the optimum extraction conditions were established. Key experimental
parameters include fluid density, temperature, fluid composition, flow rate, extraction time, and
extraction mode (static/dynamic). The influence of the chemical and physical properties of the
analytes on the extraction method were also investigated, as well as the effects and consequences
of analyte-matrix interactions.
The results from experiments using fifteen organochlorine pesticides and three soil
matrices will be presented. These results include spike recovery, precision, method bias and
method detection levels. A comparison with Soxhlet extraction for soils containing "native"
organochlorine residues will also be presented.
439
-------
70
SUPERCRITICAL FLUID EXTRACTION (SPE) OF ORGANOCHLORINE
AND ORGANOPHOSPHATE PESTICIDES FROM SOILS
John L. Snyder, Senior Chemist, Lancaster Laboratories,
Inc., 2425 New Holland Pike, Lancaster, PA 17601
ABSTRACT
In this study, 12 pesticides representative of common
organochlorine and organophosphate pesticides were extracted
from spiked soils and real-world samples using a Suprex SFE-
50 instrument. The organochlorine pesticides included two
EPA CLP surrogates, Tetrachlorometaxylene (TCMX) and
Decachlorobiphenyl (DCB), and also Endrin, Endrin Aldehyde,
p,p' DDT, and Mirex. The organophosphate compounds were
Dichlorvos, Ronnel, Parathion, Methidathion,
Tetrachlorvinphos, and Diazinon.
Both carbon dioxide and carbon dioxide modified with 3%
methanol were used to extract the soils. The pesticides
were then quantified using either gas chromatography with
electron capture detection (GC-ECD) or with gas
chromatography mass spectrometry in the selective ion
monitoring mode (GC-MS-SIM).
Experiments were also performed to demonstrate the effect of
extraction conditions such as density and pressure,
temperature, and extraction volumes of supercritical fluid
on the recoveries of the pesticides. The effects of other
variables such as pH, moisture, and different types of soil
matrices on pesticide recovery were also investigated.
It was found that quantitative recovery of the more polar
pesticides such as Endrin Aldehyde and most of the
organophosphate pesticides could not be achieved in soil
matrices using only CO2. Quantitative recovery of all the
pesticides could, however, be achieved in all soils, when 3%
methanol was added to the C02. Density and pressure were
found to have a much greater effect on the pesticide
recovery than did temperature.
Lastly, a comparison was made between SFE and the classical
sonication and Soxhlet extraction techniques. A large batch
of top soil was fortified with each of the pesticides and
extracted repetitively using each extraction method. It was
found that the overall average recoveries of the 12
pesticides by sonication, soxhlet, and supercritical fluid
extraction were 94.7%, 93.1%, and 91.6%, respectively. SFE
440
-------
demonstrated the best precision of the three extraction
methods with the overall average relative standard deviation
for all the pesticide compounds being 2.94%. Supercritical
Fluid Extraction (SFE) was equivalent to the sonication
extraction in recovering organochlorine pesticides from
real-world soil samples.
INTRODUCTION
The sonication and Soxhlet extractions are promulgated by
the the United States Environmental Protection Agency
(USEPA) as an extraction method for removing organic
contaminants from soil samples (1,2). However, both the
sonication and Soxhlet extractions are undesirable because
they require large amounts of expensive solvents, many of
which are chlorinated and are hazardous to the environment
and humans. Regulations aimed at saving our ozone layer may
ban the production and sale of these solvents in the future
(3). After the sonication and Soxhlet extractions it is
also necessary to concentrate the analytes in the sample
extracts by evaporating the solvent in order for the
analytes to be detected by instrumental methods. This
further increases the likelihood of human exposure and air
pollution.
The advent of SFE in recent years offers an alternative to
the liquid-solid extraction for removing organic pollutants
from soil samples prior to analysis (4). In the
environmental area, supercritical fluid extraction (SFE) has
been used by numerous investigators to extract organic
contaminants from soils and other environmental samples.
Often SFE has been compared to the Soxhlet or sonication
extraction techniques. Most of the time, SFE has been
reported as being superior in extraction efficiency and
faster (5-12).
This study investigated the effect of varying SFE conditions
and soil matrix variables on the recoveries of the
pesticides listed in Table I. The SFE recoveries of these
organochlorine and organophosphate pesticides were also
compared with the recoveries of these pesticides from spiked
soils and contaminated native soils using the classical
sonication and Soxhlet extraction methods.
441
-------
EXPERIMENTAL
Reagents
1. Standards
The pesticide compounds used in this study were
obtained from Chemical Services Inc., West Chester, PA,
and the Environmental Protection Agency Repository,
Research Triangle Park, NC. Stock solutions of each
compound at 1 mg/mL were prepared in methanol.
Five levels of working calibration standards were
prepared in methyl tert-butyl ether (MTBE) by serial
dilution of an intermediate standard, which was
prepared from the 12 individual stock solutions and
also made in MTBE. The concentration ranges were as
follows: dichlorvos and diazinon 100 to 2100 ng/mL;
TCMX and DCB 6 to 120 ng/mL; ronnel 5 to 100 ng/mL;
parathion 15 to 310 ng/mL; methidathion 19 to
400 ng/mL; tetrachlorvinphos 7 to 150 ng/mL; endrin
and endrin aldehyde 8 to 180 ng/mL; p,p' DDT 7 to
160 ng/mL; mirex 15 to 310 ng/mL. Spiking solutions
were also prepared in acetone by dilution of the
intermediate standard.
2. Solvents
Only pesticide grade methylene chloride (MeCl2) ,
acetone (Ac), and methyl tert-butyl ether (MTBE) were
used in this study. These were obtained from Fisher
Scientific (Atlanta, GA).
3. Supercritical Fluids
SFC grade carbon dioxide and carbon dioxide containing
3% methanol (Scott Specialty Gases, Plumsteadville, PA)
were used for all supercritical fluid extractions.
Extraction Methods
1. Supercritical Fluid Extractions
All extractions were performed singly on a Suprex SFE
50 using 2-10 mL stainless steel extraction vessels.
After optimization experiments, all extractions were
performed at 350 atm pressure and 50°C using CO2
modified with 3% methanol. When 10 g of soil were
442
-------
extracted, a 10 minute static soak was performed prior
to a 40-50 minute dynamic extraction. Back pressure
was maintained and the flow controlled in the 10 mL
extraction cell using a 30 to 50 cm length of 50 ^m
i.d. fused silica tubing restrictor for the outlet.
The analytes were collected by bubbling the vented CO2
through 5 mL of methyl tert-butyl ether (MTBE)
contained in a 40 mL vial capped with a Teflon faced
septum. The restrictor was inserted through this
septum and the end was placed at the bottom of the vial
about 15 mm under the surface of the MTBE. The vial
was vented to atmosphere using a wide bore stainless
steel syringe needle pierced through the septum. No
attempt was made to control the temperature of the
collection fluid. Because of the cryogenic effect of
the expanding CO2, the vial was well below room
temperature during the collection process.
Occasionally the vial had to be immersed in warm water
to unplug the end of the capillary restrictor which
would freeze. Because some of the MTBE evaporated
during the extraction all final volumes of extract were
adjusted to 5 mL with additional MTBE.
2. Sonication and Soxhlet Extractions were conducted
according to the USEPA Methods 3540 and 3550 except
10 g of soil was used and the solvent volume was scaled
accordingly (1,2).
Gas Chromatographic Analysis of Extracts
The pesticides in the extracts were determined by gas
chromatography and an electron capture detector (GC-ECD).
Direct injection of 1 j«L of the extracts was performed using
an autosampler. The injection port was maintained at 250°C
and the BCD was maintained at 300°C. The temperature
program was started at 140°C, held for 1 minute, and then
ramped at 4°C/min to a final temperature of 290°C, and held
for 15 min. Helium was used as the carrier gas at 5 mL/min
and nitrogen was used as a make up gas to the BCD at
25 mL/min. Quantification was performed using a 5-point
calibration curve plotting peak area versus concentration.
443
-------
RESULTS AND DISCUSSION
Figure 1 shows plots of recoveries vs. density for the
organophosphate and organochlorine pesticides. None of the
organophosphate pesticides were recovered at 0.2 g/mL. At a
density of 0.4 g/mL significant increases in recoveries were
observed for ronnel, parathion, and dichlorvos. Very low
recoveries (10-20%) were found for some of the more polar
pesticides, namely diazinon, methidathion, and
tetrachlorvinphos at this density. Over the remainder of
the density range, the recoveries of ronnel, parathion, and
dichlorvos appear to remain consistent. Nearly quantitative
recoveries were achieved for ronnel and parathion, two of
the less polar organophosphate pesticides, at a density of
0.60 g/mL.
Similar trends can be noticed for the organochlorine
pesticides which are also shown in Figure 1. With the
exception of endrin aldehyde, it appears there is a
threshold density of about 0.4 g/mL above which good
recoveries can be obtained for the organochlorine
pesticides. Endrin aldehyde, which contains a polar oxygen
on one end of the molecule, showed poorest recoveries and
was only partially recovered (approx. 70%) at the maximum
density.
Curves plotting recoveries of the organophosphate and
organochlorine pesticides vs. density are shown in Figure 2.
The addition of the polar methanol modifier increased the
extraction efficiencies for most of the pesticides.
Quantitative recoveries were obtained for the polar
pesticides which CO2 alone was unable to remove from the
soil. It appears from these figures that above densities of
about 0.5 g/mL (or pressures > 100 atm) acceptable
recoveries (> 90%) can be obtained for most of the
pesticides.
Supercritical fluid extractions performed on fortified soil
at temperatures of 40°, 60°, 80°, 100°, and 120°C keeping
the density constant at 0.7 g/mL demonstrated that
temperature had little influence on the recoveries of the
pesticides. Decreases in the recoveries of dichlorvos and
endrin were observed at the elevated temperatures (100-
120°C).
Five soils (reagent sand, top soil, river sediment, clay,
and the top soil after treatment in a muffle furnace) were
extracted repetitively with C02 modified with 3% methanol
and pure CO2. When C02 modified with 3% methanol was used
as the extraction fluid, recoveries ranging from 60-110%
were obtained for all of the pesticides from all of the soil
444
-------
matrices. These recoveries are listed in Table II. The
overall average recovery of the 12 pesticides from the sand,
the furnaced top soil, the river sediment, the clay, and the
top soil was 94% (N=3 or 4). The precision for these
experiments yielded an overall average RSD of 5% for all the
pesticides in all five soils. RSDs for the pesticides in
the river sediment were generally higher and ranged from
6-39%, while in the other soils, all pesticide RSDs were
less than 9%. The poorer precision observed for dichlorvos
(RSD=39%) and low average recovery (63%) for the river
sediment (these were included in the overall averages given
above) were due to a background peak attributed to the
sediment matrix which eluted at the same retention time as
dichlorvos. The area of this peak was subtracted from the
area of the dichlorvos peak when recoveries were calculated.
The results of the extractions when the soils were extracted
with C02 alone using identical SFE conditions are shown in
Table III. In general, recoveries are lower and the
precision is poorer with pure CO2 than when methanol
modified CO, was used. The overall average recovery for 12
pesticides in all the soils was 72% and the overall average
precision was 23%. Carbon dioxide alone was effective in
removing the pesticides from sand only; the overall average
recovery was 96% and RSDs ranged from 4.3-17%. The sand
matrix does not adsorb the pesticides as tightly because of
the large particle size and the relatively inert silica
surface. Recoveries were lower with the other soil
matrices. The overall average pesticide recoveries for the
clay, furnaced top soil, top soil, and river sediment were
65%, 70%, 66%, and 65%, respectively. The precision in
extracting these soils with C02 yielded overall average RSDs
for the 12 pesticides of 21%, 29%, 30%, and 28% for the
furnaced top soil, river sediment, top soil, and clay,
respectively. The poorest precision (12-118%, RSD), as
shown in Table III, was observed for the polar pesticides,
diazinon, methidathion, tetrachlorvinphos, dichlorvos, and
endrin aldehyde. Diazinon was one the most difficult
pesticides to recover using pure CO2 (0-81%).
Tables IV and V show a comparison of recoveries of the
organochlorine and organophosphate pesticides, respectively,
from a batch of top soil specially spiked with the
pesticides and extracted repetitively using SFE and the
classical Soxhlet and sonication extractions. The top soil
was spiked in a large batch (400 g) by immersing the soil in
methylene chloride, adding the spiking solution, and slowly
evaporating the solvent. After the methylene chloride was
completely evaporated and the soil was tumbled to ensure
homogenuity, 10 g portions of the fortified top soil (FTS)
were used for each of the extraction methods.
445
-------
Good extraction efficiencies were observed for the three
extraction methods as measured by the overall average of the
mean recoveries of the 12 pesticide compounds. Sonication
was highest at 94.7% and was followed by Soxhlet at 93.1%
and SFE at 91.6%. The low recovery of Parathion obtained by
the Soxhlet extractions was deleted from this average. The
sonication extraction, however, gave statistically better
recoveries from a majority of the individual pesticides when
compared to SFE. When comparing the Soxhlet to SFE, the
majority of the individual pesticides had no significant
difference in recovery at the 95% confidence level.
In this study SFE was found to have the best overall
precision as indicated by an overall average RSD for the 12
pesticides of 2.94%. Sonication ranked second with an
average RSD of 4.47%. However, for many of the individual
pesticides, there was no significance in the difference in
the precisions at the 95% confidence level between
sonication and SFE. The Soxhlet extraction was found in
these experiments to be the least precise and the average
RSD for the 11 of the pesticides, excluding parathion which
was poorly recovered, was 7.42%.
Three real-world soil samples submitted to our laboratory
for pesticide analysis and found to contain significant
levels of organochlorine pesticides were repetitively
extracted by SFE and sonication. These soils are designated
Real-world Soil #1, Soil #2, and Soil #3. Comparisons of
the organochlorine pesticide recoveries found by SFE and the
sonication extraction from these three soils are presented
in Tables VI, VII, and VIII. No significant difference was
found between the sonication and SFE recoveries and the
precisions of the extraction methods.
CONCLUSION:
SFE has been shown to be a successful analytical technique
in extracting organochlorine and organophosphate from a
variety of spiked and native soils. In practical terms, the
SFE method was faster and easier to perform than either the
sonication or Soxhlet extraction. Less solvent was consumed
by SFE. Approximately 400 to 500 mL of solvent was consumed
per Soxhlet extraction and about 150 to 200 mL per
sonication extraction. SFE required only 5 to 10 mL of
solvent. Because of the small amount of solvent necessary,
SFE required no solvent concentration step using the Kuderna
Danish apparatus. During this process the chance for
analyte loss because of evaporation, breakdown or reaction
of the compound is greatly increased.
446
-------
120
100-
80-
60-
40-
20-
Dichlorvos
Tetrachlorvinphos
Diazinon
Ronnel
Parathlon
Methidathion
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Density(g/mL)
120
100-
1 - 1 - 1 - -
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Density(g/mL)
Figure 1. Effect of density on the recoveries of
pesticides using pure CC>2 - Temperature
was kept constant at 50°C; 2 g of spiked
top soil; n=2.
447
-------
120
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Density(g/mL)
120
100-
80-
60-
40-
20-
1 I I
TCMX
Endrin
Endrin aldehyde
p, p'DDT
Mir ex
DCB
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Density(g/mL)
Figure 2. Effect of density on the recoveries of
pesticides using C02 with 3% methanol.
Temperature was kept constant at 50°C;
2 g of spiked top soil; n=2.
448
-------
Table I. Chemical Names and Formulas of Organochlorine and
Organophosphate Pesticides
Compound's Chemical
Common Name Formula CAS No. M.W.
tetrachlorometaxylene (TCMX) C8H6C14 877-09-8 244
dichlorvos C4H7C12O4P 62-73-7 221
diazinon C12H21N2O3PS 333-41-5 304
ronnel C8H8C13O3PS 299-84-3 322
parathion C10HUNO3PS 56-38-2 291
methidathion C6HUN2O4PS3 950-37-8 302
tetrachlorvinphos C10H9C14O4P 22248-79-9 366
endrin C^HgCleO 72-20-8 381
endrin aldehyde C12H8C16O 7421-93-4 381
p,p' DDT C14H9C15 50-29-3 355
mirex C10C112 2385-85-5 545
decachlorobiphenyl (DCB) C12H10C112 2051-24-3 499
Table II. Recovery of Organochlorine and Organophosphate
Pesticides from Spiked Soils Using Carbon Dioxide
with 3% Methanol
sand furnaced river clay top soil
top soil sediment
amotnt n=4 n=4 n=4 n=4 n=4
dichlorvos
TCMX
diazinon
ronnel
parathion
methidathion
tetrachlorvinphos
endrin
endrin aldehyde
p,p' DDT
mirex
DCB
added
(ng/g)
1040
60
1030
51
155
199
72
91
86
77
152
60
X
recv
81
87
101
90
88
81
81
86
85
84
87
88
RSD,
X
3.6
6.0
3.4
2.9
5.6
2.9
3.6
2.1
1.6
1.5
1.2
1.8
X
recv
93
94
93
99
94
95
97
102
95
96
103
105
RSD.
X
7.0
4.1
8.5
4.3
6.8
5.9
6.1
5.8
5.4
5.8
4.4
3.6
X
recv
63
88
92
86
86
90
85
91
83
84
86
88
RSD,
X
39
8.9
18
18
10
11
11
6.2
5.8
6.4
6.8
9.5
X
recv
91
102
113
109
97
94
89
100
93
94
96
97
RSD,
X
3.2
2.9
4.1
2.8
2.1
2.5
4.6
2.5
2.9
2.0
2.8
2.7
X
recv
101
105
90
109
104
108
111
111
102
106
107
104
RSD,
X
2.5
1.3
4.7
0.3
1.5
3.9
3.1
3.1
1.1
1.1
2.0
0.6
449
-------
Table III.
Recovery of Organochlorine and Organophosphate
Pesticides from Spiked Soils Using Carbon Dioxide
Only
dichlorvos
TCHX
diazinon
ronne I
parathion
methidathion
tetrachlorvinphos
endrin
endrin aldehyde
p,p' DDT
mi rex
DCB
amotnt
added
(ng/g)
1040
60
1030
51
155
199
72
91
86
77
152
60
sand
n=4
X
recv
93
102
81
110
103
77
79
104
95
102
106
105
RSD,
X
12
4.3
16
4.6
3.5
17
16
6.3
5.9
4.5
5.8
4.9
furnaced
top soil
n=4
X
recv
57
95
5
100
77
35
42
98
43
95
99
99
RSD,
X
19
14
59
13
15
30
25
12
20
12
15
14
river
sediment
n=4
X
recv
56
89
22
88
79
25
34
84
50
77
85
90
RSD,
X
62
2.4
85
3.8
13
68
65
5.7
25
10
3.8
7.8
clay
n=4
X
recv
43
109
0
107
88
4
10
100
25
86
101
103
RSD,
X
94
5.2
--
1.1
16
91
56
2.1
24
15
3.3
4.7
top soil
rM
X
recv
62
95
3
102
87
21
35
86
41
82
88
89
RSD,
X
26
7.3
118
8.1
5.3
88
50
8.2
26
4.3
7.5
8.3
SFE conditions for Tables II and III - 350 atm; 50°C; 2.0 g
soil contained in 2 mL vessel; 10-minute static; 10-minute
dynamic
450
-------
Table IV. Comparison of Extraction Methods for Organochlorine
Pesticides: Soxhlet vs. Sonication vs. SFE
Compound
TCMX
endrin
endrin
aldehyde
p,p' DDT
mirex
DCB
Amount.
Added
(ng/g)
30
45
43
38
76
30
Method
sx
SP
SFE
SX
SP
SFE
SX
SP
SFE
SX
SP
SFE
SX
SP
SFE
SX
SP
SFE
Trials
N
8
7
9
8
7
9
8
7
9
8
7
9
8
7
9
8
7
9
Found
Avg.
(ng/g)
24
22
22
44
49
44
35
30
36
33
31
38
71
80
74
29
31
29
Std.
Dev.
2.1
0.79
1.2
2.6
0.5
1.5
1.3
0.95
1.2
4.3
0.57
0.52
8.1
1.2
2.2
1.1
0.57
0.50
RSD
%
8.9
3.2
5.4
6.0
1.1
3.5
3.7
3.1
3.3
13
1.8
1.4
11
1.5
3.0
3.7
1.8
1.7
% Rec.
Avg.
78.0
81.0
74.0
97.0
108.0
97.0
81.0
71.0
84.0
87.0
81.0
99.0
94.0
105.0
97.0
98.0
104.0
96.0
451
-------
Table V. Comparison of Extraction Methods for Organophosphate
Pesticides: Soxhlet vs. Sonication vs. SFE
Amount.
Added
Compound
dichlorvos 520
diazinon
ronnel
parathion
methada-
thion
515
25
78
100
tetrachlo- 36
vinphos
Method
sx
SP
SFE
SX
SP
SFE
SX
SP
SFE
SX
SP
SFE
SX
SP
SFE
SX
SP
SFE
Trials
N
8
6
5
8
6
9
8
6
9
8
6
9
8
6
9
8
6
8
Found
Avg.
(ng/g)
333
375
318
479
493
433
26
27
25
22
77
73
108
100
100
42
41
39
Std.
Dev.
s
28
42
13
24
27
13
1.4
2.4
1.2
20
1.2
0.88
5.7
4.8
3.4
4.7
1.3
2.5
RSD
%
8.4
11
4.0
5.1
5.4
3.0
5.2
9.0
5.0
92
1.6
1.2
5.3
4.8
3.2
11
3.2
6.3
% Rec
Avg.
64
72
61
93
96
84
104
106
98
28
99
94
108
100
106
117
113
109
Table VI. Quantification of Organochlorine Pesticides in
Soil Sample #1 - SFE vs. Sonication
Supercritical Fluid
TCMX
p,p' DDE
p,p' ODD
p,p' DDT
DCS
n=4
Amount
Avg.
(ng/g)
95%
44
43
453
97%
RSD
%
4.1
7.1
8.6
5.0
2.1
Sonication
n=3
Amount
Avg . RSD
(ng/g) %
93%
36
39
431
102%
12
15
8
17
2
.4
.6
TCMX and DCB added as surrogates prior to extractions
results expressed as % recovered.
452
-------
Table VII.
Quantification of Organochlorine Pesticides in
Soil Sample #2 - SFE vs. Sonication
TCMX
heptachlor
aldrin
dieldrin
alpha chlordane
gamma chlordane
endrin
endosulfan II
endrin aldehyde
endrin ketone
DCB
Supercritical
n=6
Amount
Avg.
(ng/g)
76%
14
81
327
429
80
1418
54
29
2601
90%
Fluid
RSD
%
14
40
9.
9.
8.
9.
4.
3.
84
14
5.
Sonication
n=3
Amount
Avg . RSD
(ng/g) %
7
3
4
0
6
0
9
68%
16
114
344
483
88
1524
58
45
2675
132%
11
17
6.3
2.4
4.7
6.5
2.5
6.2
3.3
9.3
6.9
TCMX and DCB added as surrogates prior to extractions
results expressed as % recovered.
Table VIII.
Quantification of Organochlorine Pesticides in
Soil Sample #3 - SFE vs. Sonication
Supercritical
n=6
Amount
Avg.
(ng/g)
80%
32
23
273
17
60
93%
Fluid
RSD
%
1.3
19
14
24
36
12
15
Sonication
n=3
Amount
Avg . RSD
(ng/g) %
96%
33
20
312
31
88
112%
4
4
17
7
14
14
8
.8
.7
.7
.3
TCMX
aldrin
alpha chlordane
endrin
endrin aldehyde
endrin ketone
DCB
TCMX and DCB added as surrogates prior to extractions
results expressed as % recovered.
453
-------
REFERENCES:
1. Test Methods for Evaluating Solid Waste, EPA SW-846,
Method 3540, 3rd edition, 1990.
2. Test Methods for Evaluating Solid Waste, EPA SW-846,
Method 3550, 3rd edition, 1990.
3. Clean Air Act Amendments 1990.
4. Hawthorne, S. B. Anal. Chem. 1990, 62, 633A.
5. Schantz, M. M.; Chesler, S. N. J. Chromatogr. 1986, 363,
397.
6. Hawthorne, S. B.; Miller, D. J. Anal. Chem. 1987, 59,
1705.
7. McNally, M. E.; Wheeler, J. R. J. Chromatogr. 1988, 447,
53.
8. Onuska, F. I.; Terry, K. A. J. High Resolut. Chromatogr.
1989, 12, 357.
9. Alexandrou, N.; Pawliszyn, J. Anal. Chem. 1989, 61, 2770,
10. Campbell, R. M.; Richards, M. LCGC 1991, 9, 358.
11. Snyder, J. L.: Grob, R. L.; McNally, M. E.: Oostdyk, T.S.
Anal. Chem. 1992, 64, 1940.
12. Snyder, J. L.: Grob, R. L.; McNally, M. E.: Oostdyk, T.S.
J. Chromatogr. Science 1993, 31, 183.
454
-------
71
SFE IN A PRODUCTION ENVIRONMENTAL LAB
Mark L. Bruce. Enseco, 4101 Shuffel Dr. NW, North Canton, OH 44720 and
Michael Millie, Enseco, 2544 Industrial Blvd., West Sacramento, CA 95691.
ABSTRACT
High temperature and high flow rate extractions have produced good analyte recovery
from dry and wet clay matrices for TPHs and PCBs. Rugged methods combined with
automated SFE systems should be able to reduce extraction cost and turn-around time for
solid environmental samples.
INTRODUCTION
Supercritical Fluid Extraction is making in-roads to the production environmental
laboratory. Researchers and regulatory agencies have expended much effort producing
extraction methods for environmental applications. The process of moving SFE
technology from the research lab to the production environmental lab has been slowed by
a combination of requirements unique to this laboratory segment.
Solid environmental matrices range from amorphous clays to zeolites. Soil samples are
most the common, but the composition is highly variable. Many analytes are covered by
federal and state regulations. For example, the US EPA 40 CFR Appendix IX list from
7/91 has about 150 semivolatile compounds that must be extracted by SFE if it is to
completely replace existing sonication and Soxhlet methods. Since environmental
samples are often heterogeneous, the question of sub-sample representativeness must be
addressed. SFE is easiest with small samples while statistical concerns recommend large
samples. Highly automated instruments are required to meet through-put expectations.
Working in the environmental field requires heightened awareness to health & safety and
environmental issues. Using reagents and solvents that are or may be perceived as
detrimental is discouraged. Lastly, all methods require federal, regional or state approval
before they can be used.
The goal of method development for a production environmental lab is to establish a
single set of extraction conditions that work for as wide a range of solid matrices as
practical. The method should emphasize high through-put and low cost at client defined
data quality objectives.
INSTRUMENTATION, EQUIPMENT AND SUPPLIES
Supercritical Fluid Extractor
Suprex, PrepMaster and AP44
1,5,10 mL extraction vessel
Gas Chromatograph
Hewlett Packard 5890
Electron Capture Detector
Column RTX-5, 30 m, 0.32 mm ID, 0.5 |i DF
Infrared Spectrophotometer
Perkin-Elmer, 710
Buck Scientific Oil in Water Analyzer
10 mm, 3 mL quartz cell
Solvents
Hexane, EM Science
455
-------
CO2, SFC grade with 1500 PSIA Helium headspace with dip tube, Scott Specialty
Tetrachloroethene, J.T. Baker - Oil & Grease
EXPERIMENTAL
There are several supercritical fluid extraction parameters to optimize. They are
extraction time (static & dynamic), CO2 pressure, CO2 flow rate, extraction temperature,
analyte trap, modifier amount & type and amount of sample. Solid matrices such as
loam, humus and various clays (kaolin, Fuller's earth, montmorillonite) were examined.
Fractional factorial experiment designs were used to investigate and optimize these
parameters as well as measure method ruggedness. Also, the issues of matrix variability,
automation, environmental pollution and regulatory acceptance have been considered.
Total Petroleum Hydrocarbon analyses were performed using draft Method 8440. PCB
analyses were performed by GC-ECD with modified Method 8081.
RESULTS AND DISCUSSION
TPHs
Draft Method 3560 states that a small amount of water in the sample improves analyte
recovery. However, high moisture content may cause two types of problems. First, if
water is liberated from the sample during the extraction, some common SFE restrictors
may plug which then stops the extraction. Second, water reduces TPH recovery for some
matrices under the SFE conditions specified in the draft method (Dec 1992). There are
two approaches to handling water. 1) Mix: an absorbent into the sample, allow sufficient
interaction time and extract below the temperature where the drying agent releases water.
2) Use a restrictor that tolerates water and extract at a high enough temperature to dry the
sample during the extraction process.
Use of a drying agent is the primary suggestion in 3560. The drying agent quickly
absorbs free water so that the restrictor is protected. However, there can be sufficient
water remaining on the surface of the matrix that TPH recovery is reduced unless the
drying agent is allowed to interact with the sample for a long time. The EPA has
determined that overnight drying is sufficient when using magnesium sulfate
monohydrate as the drying agent. The results of our time study are consistent with this
recommendation. Figure 1 shows how TPH recovery improves as the interaction time
between the sample and Hydromatrix (diatomaceous earth) increases. For this difficult
sample Hydromatrix is not sufficiently active to achieve acceptable results in a reasonable
time frame. Using a more active drying agent such as magnesium sulfate accelerates the
process but after three hours (the minimum allowed in 3560) TPH recovery is still low.
The last bar in Figure 1 shows that the motor oil analyte is completely recovered from the
sample when water content was near 0%. Thus, the draft 3560 (Dec 1992) conditions
work well with dry samples but require an overnight drying step for wet samples. As a
production lab overnight drying is not a desirable method requirement. We sought other
ways to address the analyte recovery problems from wet samples.
Extractions of dry motor oil spiked kaolin show significant extraction time reduction by
elevating the extraction temperature from 80 to 150°C. Flow rate and pressure changes
do not affect extraction efficiency when varied from 1.2 to 6 ml/min and 340 to 500 atm.
When motor oil spiked Fuller's earth is extracted wet (50% moisture) the temperature
effect is still strong but flow rate becomes equally important. Higher flows improve
extraction efficiency. Parameters were varied as follows, pressure 340 450 atm, CO2
flow rate 3.5 7 mL/min and extraction temp 120 150°C. Figure 2 shows the results
456
-------
from the factorial design experiments. Raising the temperature from 120°C to 150°C
improved recovery about 25%. Raising the flow rate from 3.5 to 7 rmVmin increased
recovery about 25% as well. These results are summarized in Figure 2. The higher
temperature and flow rate removed water from the sample quickly, presumably this
allowed the supercritical CO2 greater access to the analytes on the clay surfaces. These
high temperature and high flow rate extractions have produced good analyte recovery
from wet clay matrix spiked samples. Table 1 summarizes this data. The wet sample
(3g) was mixed with 1.5g of Hydromatrix and extracted immediately. The Hydromatrix
soaks up any free water and disperses the sample to prevent it from turning into a brick
during the course of the extraction. The moisture from the sample is blown out of the
extraction vessel through the restrictor and analyte trap to waste. The restrictor and
analyte trap can tolerate up to 5 g of water.
Motor oil spiked Fuller's Earth
50% moisture
120T
80 ..
0)
O 60
0)
OC 40
340 atm, 80°C
1-2 mL/min CO2
30 mln dynamic
3.5 6.5 4 days
Hydromatrix
4 days
Dry
MgS04
Drying time (hrs)
Figure 1 Drying Agent Time Study.
30-r
-1O-1-
2 & 3 factor interactions
Figure 2. Factorial design results for wet kaolin.
457
-------
Table 1 TPH Recovery from Wet Clays - High Temperature and Flow Extraction.
Matrix
Fuller's Earth
Montmorillonite
KSF
Montmorillonite
K10
Analyte
Diesel
Motor Oil
Diesel
Motor Oil
Diesel
Motor Oil
% Recovery
84
99
95
80
91
94
340 atm, 150°C, 6.5 mL/min CC>2, 25 min dynamic, Hydromatrix drying agent
PCBs
A similar water effect has been seen when extracting polychlorinated biphenyls (PCBs).
Increasing temperature improves PCB recovery from dry clays. Increasing flow rate and
temperature improves PCB recovery from wet clays but eventually further increases in
flow rate have no effect. Figure 3 shows the factorial results for dry kaolin using the
following SFE conditions; pressure 380 450 atm, CO2 flow rate 4 7 ml/min and
extraction temp 120 - 150°C. Increasing the temperature from 120°C to 150°C improved
PCB recovery about 17%. The other factors showed no significant effect.
Water was added to the same spiked kaolin (50% moisture). The sample was extracted
under less severe conditions; pressure 350 450 atm, CO2 flow rate 2-4 ml/min and
extraction temp 80 150°C. Under these conditions (Figure 4) both flow rate and
temperature significantly effect analyte recovery. Increasing the flow rate further in the
next factorial experiment shows no recovery increase (Figure 5). The extractions
conditions were; pressure 380 - 450 atm, CO2 flow rate 4-7 ml/min and extraction temp
120 - 150°C.
o
O)
c
CO
o
Q)
O
O
0)
DC
c£
20-,
15-
10-
5 -
0 -
-5-
-10-
Pressure
Figure 3 Extraction of Dry Kaolin Spiked with PCB 1254.
458
-------
V
O)
c
TO
£
O
>
w
V
>
O
u
0)
CC
-•5
ff^
20-
15-
10-
5 -
0
-5-
-10-
-15-
-20-
Pressure
Flow
rate
Temp
P&F&T
Figure 4 Extraction of Wet Kaolin Spiked with PCB 1254.
20 -r
P&F&T
-20 -L
Figure 5 High Flow Extraction of Wet Kaolin Spiked with PCB 1254.
Real samples with native PCB analytes were extracted with the following conditions;
pressure 350 - 450 atm, CO2 flow rate 2 - 4 ml/min and extraction temp 80 150°C. A
humus and cement flakes sample with 4.5 ppm PCB 1260 was studied. The sample
results are shown in Figure 6. The standard deviation of replicate runs was 15%. Since
no calculated effects were much larger than 15%, there were no significant effects or
factors. The %RSD for all extractions was 15%. Thus, SFE was very rugged for this
sample but the precision was limited by sample homogeneity.
The next sample was sandy soil with 29.2 ppm PCB 1242. The standard deviation of
replicate runs was 4.5%. Since no calculated effects were much larger than 4.5%, there
were no significant effects or factors. The %RSD for all extractions was 5.8%. Thus,
SFE was rugged for this sample. See Figure 7 for these results.
459
-------
The last sample was a standard reference material from National Research Council
Canada. It was marine sediment HS-1 with 0.02 ppm PCB 1254. The standard deviation
of replicate runs was 19.9%. Since no calculated effects were larger than 19.9%, there
were no significant effects or factors. The concentration of the extract was near the GC
instrument detection limit where precision is poor. Further concentration of the extract
should produce better precision. The %RSD for all extractions was 16.8%. These results
are summarized in Figure 8.
o>
O>
C
CO
O
0)
O
O
0)
DC
20 -r
15 -- -
10 --
55 -10--
-15 --
Figure 6 Humus & cement flakes, 4.5 ppm PCB 1260.
Figure 7 Sandy soil, 29.2 ppm PCB 1242.
460
-------
0>
O>
C
a
u
a>
o
o
0)
cc
P&F&T >
3
00
Figure 8. Factorial design results for HS-1.
CONCLUSION
TPHs and PCBs can be efficiently recovered from spiked clay (dry & wet) as well as
many real samples. The use of high extraction temperatures and high flow rates has
reduced the need for organic co-solvents (modifiers). Extraction is typically complete in
10-30 minutes. The use of automated SFE systems should reduce labor and solvent costs
while reducing extraction turn-around time.
ACKNOWLEDGMENT
The authors would like to thank the following people that provided valuable support
during this project: Suprex Ray Houck, Glenn Williams, Lori Dalatta, Jerry Wisser,
Doug Koebler. JT Baker- Paul Buis. Enseco - Carolynne Roach, Rhonda Kuster, Todd
Benenati, Tami Stephens, Paul Winkler.
461
-------
72
PREEXTRACTION HOLDING TIMES FOR NITROAROMATICS AND NITRAMINES IN SOILS
T.F. Jenkins. Research Chemist, U.S. Army Cold Regions Research and Engineering Laboratory, Hanover,
New Hampshire 03755-1290; C.L. Grant, Professor Emeritus, Chemistry Department, University of New
Hampshire, Durham, New Hampshire 03824; and K.F. Myers and E.F. McCormick, Chemists, U.S. Army
Engineer Waterways Experiment Station, Environmental Laboratory, Vicksburg, Mississippi 39180.
ABSTRACT
Studies were conducted to investigate the maximum acceptable preextraction analytical holding times
(MHTs) for nitroaromatic and nitramine explosives in soils. Initial experiments were conducted using three
soils fortified with two nitramines, HMX (octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine) and RDX (hexa-
hydro-l,3,5-trinitro-l,3,5-triazine), and four nitroaromatics, TNB (1,3,5-trinitrobenzene), TNT (2,4,6-
trinitrotoluene), 2,4-DNT (2,4-dinitrotoluene) and tetryl (N-methyl-N,2,4,6-tetranitroaniline). Fortification
was accomplished using aqueous solutions, and spiked concentrations were in the low (Og/g range. Addi-
tional studies were conducted with field-contaminated soils from the Crane Naval Surface Warfare Center.
These soils had been field-contaminated with HMX, RDX, TNT, and TNB. Fortified samples were held for
periods up to eight weeks in the dark at room temperature (22°C), under refrigeration (2°C), or frozen
(-15°C). Field-contaminated samples were held in the dark up to eight weeks under refrigeration.
Whether fortified or field-contaminated, the two nitramines (HMX and RDX) were stable over the eight-
week test period at all storage temperatures. For the fortified soils, however, nitroaromatics were rea-
sonably stable when frozen, degraded rapidly at room temperature, and more slowly under refrigeration.
TNB and tetryl were the least stable of the four nitroaromatics tested, with losses ranging from 67% to 99%
when stored under refrigeration for seven days. In contrast, TNT and TNB, at very similar concentrations to
those in the fortified samples, were quite stable under refrigeration for four field-contaminated soils. When
these field-contaminated soils were subsequently fortified with TNT and TNB, rapid degradation under
refrigeration was again observed for the added nitroaromatics. We conclude that studies using fortified soils
can produce very different estimates of MHTs compared to those using field-contaminated soils.
INTRODUCTION
Several years ago, the U.S. Army Cold Regions Research and Engineering Laboratory (CRREL) developed
a laboratory method for the determination of nitroaromatic and nitramine explosives in soil (1). This
method was collaboratively tested (2) and subsequently given preliminary acceptance by the EPA as
SW846 Method 8330 (3). One criterion that was not experimentally evaluated during this method develop-
ment process was an acceptable preextraction sample holding time. Lacking available experimental data,
the EPA method established a preextraction hold time of seven days for soil (3). This holding time was cho-
sen to be consistent with those for other organics in a soil matrix and for contractual compliance.
Subsequently, a study was conducted at Oak Ridge National Laboratory to provide the data necessary to
recommend appropriate maximum preextraction holding times (MHTs) for soils contaminated with nitro-
aromatic and nitramine explosives (4). In their study, soils were fortified with two nitroaromatics (TNT and
2,4-DNT) and two nitramines (RDX and HMX) and stored at either room temperature (+20°C), refrigerator
temperature (+4°C), or freezer temperature for periods up to a year. Soil fortification was accomplished
using an acetonitrile solution of the analytes (2 mL in 2 g of soil) and the acetonitrile was not removed
prior to the onset of the study. While the effect of this large amount of acetonitrile on the soil biota is un-
known, storage of soils under acetonitrile does not mimic the manner in which normal soil samples are
stored prior to analysis for nitroaromatics and nitramines. In fact, acetonitrile is the extraction solvent of
462
-------
choice for analysis of soils for these analytes (1,3). While the study was carefully done and the statistical
treatment of the data extensive, we feel the Oak Ridge study suffers a flaw and the resulting MHT estimates
may not be appropriate for customary soil sample storage procedures. Nevertheless, based on their results,
they recommend storage of RDX, HMX, and 2,4-DNT contaminated soils at 4°C (refrigerator) with a MHT
of six weeks, and TNT contaminated soils at -20°C (freezer), also with a MHT of six weeks.
To our knowledge there is no standard procedure for the estimation of MHTs for organic contaminants in
soil. The Oak Ridge study used a modified version of an ASTM method developed for MHT estimation for
water samples (5). According to this protocol, MHT is defined as the "maximum period of time during
which a properly preserved sample can be stored before such degradation of the constituent of interest
occurs or change in sample matrix occurs that the systematic error exceeds the 99% confidence interval
(not to exceed 15%) of the test about the mean concentration found at zero time." The zero time mean con-
centration and standard deviation are estimated from an appropriate number of samples (usually ten) an-
alyzed immediately after collection. If an analyte concentration is less than one order of magnitude higher
than the criterion of detection, a bulk sample is fortified and the zero time mean and standard deviation are
redetermined. Concentrations are then measured after various time intervals using a number of replicates
based on the percent relative standard deviation (RSD) of the zero time results. The average concentration
found at each analysis point is plotted versus time on linear graph paper and the "best graphical fit" to the
data points is drawn. A MHT is the point where the "best fit" line intersects the two-sided 99% confidence
interval about the zero time mean. Note that the number of replicates used in the confidence interval
calculation is the number used for each time interval measurement rather than the ten replicates used to
estimate the zero time mean. While the ASTM protocol allows use of field-contaminated waters for MHT
estimation, common practice for soils is the use of fortification, as was done in the Oak Ridge study. Ex-
periments on the stability of 1,2-dibromoethane (EDB) (6) and polyaromatic hydrocarbons (PAHs) (7) in
soils casts doubt on the ability of fortification to mimic analyte-soil interactions where contamination has
been in place for extended periods. For nitroaromatics and nitramines this is particularly germane since
these contaminants have been in place in some cases for over 50 years.
In the following study, we reexamine the issue of MHTs for soils contaminated with nitroaromatic and nit-
ramine explosives with emphasis on (a) avoidance of organic solvent addition during soil fortification and
(b) comparison of stability of fortified soils to field-contaminated soils.
EXPERIMENTAL
Chemicals
All standards and test solutions were prepared from Standard Analytical Reference Materials (SARMs) ob-
tained from the U.S. Army Environmental Center (USAEC), Aberdeen Proving Ground, Maryland. Aque-
ous standards and test solutions were prepared in reagent grade water obtained from a Milli-Q Type I Re-
agent Grade Water System (Millipore Corp.). Methanol used in the preparation of HPLC eluent and
acetonitrile used for soil extraction were HPLC Grade from Alltech and Baker, respectively. Eluent was
prepared by combining equal volumes of methanol and water and vacuum filtering through a nylon mem-
brane (0.45 nm) to degas and remove particulate matter.
Analyte spiking solutions
All analyte spiking solutions were prepared in water. SARMs for 2,4,6-trinitrotoluene (TNT), 2,4-
dinitrotoluene (2,4-DNT), 1,3,5-trinitrobenzene (TNB), l,3,5-hexahydro-l,3,5-trinitrotriazine (RDX),
l,3,5,7-octahydro-l,3,5,7-tetranitrotetrazocine (HMX), and N-methyl-N,2,4,6-tetranitroaniline (tetryl) were
placed in individual brown glass jugs, reagent grade water was added, and the contents were stirred at room
463
-------
Table 1. Physical and chemical properties of test soils.
Soil
Windsor
Charlton
Fort Edwards
Crane A
Crane B
Crane C
PH
6.2
6.0
8.4
8.6
8.4
7.9
% Clay
30
20
70
12
1
9
Property
Total Organic
Carbon (%)
1.1
1.8
0.5
0.5
0.9
1.5
Cation Exchange
Capacity, meq/lOOg
3.5
7.3
>150.0
15.7
3.2
9.5
temperature for a week. The solutions were then filtered through 0.45 \Jm nylon membranes into clean,
brown glass jugs. No solvents other than water were used in the preparation of these solutions.
The concentration of analyte in each aqueous spike solution was determined against standards prepared in
methanol or acetonitrile (1,3) diluted 1:1 with reagent grade water prior to analysis. A multianalyte spiking
solution was prepared by combining appropriate volumes of all of the individual, aqueous-based, analyte
solutions with the exception of tetryl and filtering through a 0.45 ujn nylon membrane.
Soils
Blank test soils were obtained locally from Vermont (Windsor), New Hampshire (Charlton), and New York
(Fort Edwards). These soils were air-dried, ground with a mortar and pestle, and passed through a 30-mesh
sieve (590 mm). Some physical and chemical properties of these soils are presented in Table 1. Replicate
5.0 ± 0.1 g subsamples of each blank soil were placed in individual 20-mL glass scintillation vials.
Three field-contaminated soil samples (designated A, B, and C) from the Rockeye site at the Naval Surface
Warfare Center, Crane, Indiana site were selected from a large group of samples on the basis of field
screening for explosives residues. These soils were not dried or sieved, but they were homogenized by hand
and rocks or other large particles were removed. They were stored in glass jars at 4°C in the dark. These
soils contained measurable concentrations of HMX, RDX, TNT, TNB, two isomeric microbiological trans-
formation products of TNT, 2-amino-4,6-dinitrotoluene (2-AmDNT) and 4-amino-2,6-dinitrotoluene
(4-AmDNT) and 3,5-dinitroaniline (3,5-DNA), which is a microbiological transformation product of TNB
(8, 9). Selected properties of these soils are shown in Table 1.
Soil wetting and analyte spiking
Prior to the onset of the experiment, previously air-dried test soils were rewetted. Because the water hold-
ing capacity for each soil varied, the amount of water added to each soil was varied such that after spike ad-
ditions were also made, there was no evidence of freestanding water. For the three initially blank soils, 0.20
mL of reagent grade water was added to the Windsor sandy loam and 1.00 mL was added to the Fort Ed-
wards Clay and Charlton silty loam. For the field-contaminated soil from Crane, 0.50 mL of reagent grade
water was added. After water addition, all soils were allowed to stand at room temperature in the dark for
three days to allow microbiological activity to reestablish (4).
Fortification of the three blank soils was made by adding 1.00 mL of a combined aqueous spiking solution
to each vial. An identical study was conducted separately for tetryl since the transformation products of
tetryl interfered analytically with the other nitroaromatics of interest. Except for the soils designated as day
464
-------
zero exposure and those to be stored frozen, the spiked soils were immediately placed in the appropriate
storage temperature in the dark. The zero day samples and the samples to be frozen were permitted to stand
for two hours after fortification to allow time for the analytes to interact with the soils prior to either extrac-
tion or freezing. The vials containing the field-contaminated soil were treated and stored in an identical
manner as described above except that no fortification was made. Initial analyte concentrations in the field-
contaminated soil ranged from 0.3 to 1.6 J4g/g.
Soil respiration
To ensure that the rewetted, fortified soils had regained microbial activity, three vials of each soil were
placed in separate 250-mL Erlenmeyer flasks enclosed with a two-hole rubber stopper. Air was drawn
through two aqueous NaOH scrubbers, through an Erlenmeyer containing a given soil, and into a CO2
collection tube containing standard aqueous NaOH. The CO2 evolved from the soils was collected as car-
bonate over a period of two weeks and the carbonate level determined by back titration with 0.5 N HC1.
The results indicated that resumption of microbial activity had occurred in the rewetted soils.
Soil holding time test parameters—fortified soils
Three storage conditions were examined: room temperature (22 ± 2°C), refrigerator (2 ± 2°C), and freezer
storage (-15 + 2°C). Portions stored under these conditions were extracted after 0, 3, 7, 14, 28, and 56 days
of storage and the analyte concentrations determined. Because of expected variability among replicates,
triplicates were analyzed for each temperature /time combination.
Soil holding time test—field-contaminated soils
Experiments with the three Crane soils were initiated within two days of sampling. No spiking took place
in the initial portion of this study. Seven 5.00-g subsamples of each field moist soil were immediately
extracted and analyzed. A bulk portion of each soil was weighed, air-dried overnight, and reweighed to
obtain percent moisture. The dried samples were ground and seven 3.00-g subsamples were extracted and
analyzed. This procedure was repeated with a second bulk portion of each soil after eight weeks storage
(moist) at 4 ± 2°C in the dark.
Since the results from this study differed greatly from those with fortified soils, an additional study was
conducted with the residual portion of these three Crane soils. In this experiment, 42 2.00-g subsamples of
moist soil were transferred to 20-mL vials for each of the three Crane soils. Additional samples were used
for the determination of percentage moisture after overnight air drying. Twenty-eight of the subsamples for
each soil were fortified with an aqueous solution of TNB and TNT of known concentration. The volume of
spiked solution varied somewhat from soil to soil in order to prevent free-standing solution, but all sub-
samples for a given soil received the same volume. Reagent grade water was added to the remaining 14
vials for each soil in amounts that matched the respective volumes of spike solution and served as controls.
One set of seven spiked subsamples and one set of water-treated subsamples from each soil were extracted
and analyzed the same day they were prepared. All other subsamples were stored at 4 ± 2°C in the dark un-
til removed for analysis.
Soil extraction and analysis
Soils were extracted and analyzed as described in SW846 Method 8330 (1, 3) with two differences. First
the soils were not air-dried prior to extraction, because it was judged that the time required to dry the soil in
the vials at room temperature could result in analyte loss and confound the effect of the holding time tem-
peratures. Second, a 5-gram portion of soil was used for the fortified samples instead of the usual sample
465
-------
size of two grams. This was necessary since the solubility of HMX and RDX in the aqueous spiking solu-
tion is limited (4 mg/L and 42 mg/L, respectively) as was the moisture holding capacity of the test soils.
Thus to obtain sufficiently high extract concentrations of these analytes without exceeding the moisture
holding capacity of the soils, larger soil samples were required.
Data analysis—fortified soils
The mean and standard deviation for each set of triplicate measurements were calculated. Suspect
individual measurements were flagged on the basis of extreme values of the % RSD (> 50%) and
inconsistencies in the overall pattern for that compound. Each suspect value was checked for possible com-
putation or transcription errors. Twelve individual extreme values were arbitrarily excluded because they
produced large distortions of both means and standard deviations. In no case was more than one datum
excluded from a triplicate set. These exclusions amounted to less than 1% of the values.
A modified version of the ASTM procedure was used to estimate MHTs where appropriate. To gain
degrees of freedom (d.f.) and to fairly represent precision for the entire experiment, pooled standard
deviations were calculated for the six sets of triplicates for each soil/storage condition where rapid degrada-
tion was absent. This produced more d.f. for the standard deviation than the nine that would have been ob-
tained if we had been able to run ten replicates on day zero as suggested by ASTM. Where a 99% con-
fidence interval exceeded ± 15% of the day zero mean, the limits were set at ± 15% as specified in the
ASTM procedure. This procedure worked well for the fortified soils where standard deviations were small,
and the results should be very comparable to the standard ASTM procedure.
Data analysis—field-contaminated soils
Analysis of variance (ANOVA) was used to compare mean concentrations for the results for the field-
contaminated soils. Rapid degradation of spiked TNB and TNT is depicted graphically; no statistical analy-
sis was required.
RESULTS AND DISCUSSION
Behavior of analytes in fortified soils as a function of holding time
The mean concentrations of the five fortified analytes and transformation products in Windsor soil are pre-
sented in Table 2 as a function of holding time and storage condition. Of the six fortified analytes, TNB
shows the most rapid rate of degradation. The rate of loss for TNB in Charlton and Fort Edwards soils is
even greater (8). For all three soils TNB concentration rapidly decreases at room temperature with only an
average of 6.5% remaining in these soils after three days. For refrigerator storage, the rate of disappearance
of TNB is slower than at room temperature, but even so, only an average of 15.3% remains after seven
days. Further reduction of TNB occurs by 14 days and by 28 days the concentration of TNB is below its
detection limit. This disappearance is accompanied by the appearance of an increased level of 3,5-DNA,
the expected initial microbiological transformation (reduction) product (9). These changes are shown for
Windsor soil stored at refrigerator temperature in Figure 1. On a molar basis, a maximum of 36%, 47%,
and 15% of the TNB lost could be accounted for as 3,5-DNA for the Windsor, Charlton and Fort Edwards
soils, respectively. It is also interesting to note the slow decrease in 3,5-DNA concentration in all three soils
once the TNB precursor is gone. Clearly this is a very dynamic system even under refrigeration. In contrast
to the rapid degradation found at room temperature and under refrigeration, TNB is quite stable in the
frozen soils (Table 2).
466
-------
Table 2. Concentrations of analytes and transformation products as a function of holding time and
storage condition, Windsor sandy loam.
Mean concentration (fig/g) ± standard deviation (ng/g)
Holdinv time1
00 Davp
Compound
HMX
RDX
TNB
3,5-DNA
TNT
4-Am-DNT
2-Am-DNT
2,4-DNT
Tetryl
Storage
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room. temp.
Refrigerator
Freezer
Room temp.
Refrigerator
Freezer
X
0.373
0.373
0.373
1.500
1.500
1.500
0.914
0.914
0.914
0.969
0.969
0.969
0.850
0.850
0.850
5.48
5.48
5.48
s
0.003
0.003
0.003
0.007
0.007
0.007
0.004
0.004
0.004
0.005
0.005
0.005
0.002
0.002
0.002
0.74
0.74
0.74
03 Dai/s
X
0.353
0.360
0.354
1.355
1.374
1.368
0.013
0.598
0.885
0.277
0.116
0.465
0.861
0.926
0.109
0.025
0.004
0.037
0.010
0.741
0.802
0.799
1.63
4.15
4.44
s
0.007
0.004
0.003
0.019
0.006
0.008
0.023
0.020
0.008
0.007
0.005
0.030
0.005
0.006
0.005
0.002
0.003
0.002
0.004
0.016
0.004
0.006
0.72
0.09
0.78
07 Dfli/s
X
0.385
0.375
0.377
1.608
1.612
1.600
0.300
0.946
0.274
0.220
0.067
0.777
0.975
0.202
0.041
0.074
0.016
0.716
0.837
0.863
0.66
2.94
5.10
s
0.012
0.004
0.011
0.005
0.009
0.023
0.030
0.037
0.016
0.013
0.010
0.013
0.026
0.006
0.006
0.004
0.004
0.021
0.007
0.020
0.42
1.01
0.87
1A Dfli/s 28 Dflt/s 56 Davs
X
0.377
0.381
0.377
1.568
1.590
1.575
0.090
0.952
0.238
0.283
0.637
0.978
0.215
0.067
0.088
0.031
0.626
0.828
0.856
s
0.001
0.009
0.005
0.003
0.015
0.004
0.027
0.001
0.007
0.003
0.043
0.003
0.005
0.010
0.002
0.002
0.006
0.013
0.005
X
0.392
0.379
0.399
1.622
1.597
1.633
0.937
0.191
0.277
0.014
0.309
0.980
0.211
0.124
0.092
0.051
0.573
0.772
0.857
s
0.009
0.004
0.018
0.015
0.006
0.025
0.054
0.008
0.007
0.024
0.026
0.024
0.010
0.002
0.000
0.002
0.004
0.005
0.017
X
0.349
0.350
0.354
1.572
1.570
1.586
0.013
0.949
0.127
0.255
0.028
0.086
0.954
0.169
0.169
0.079
0.065
0.004
0.419
0.675
0.808
s
0.008
0.009
0.008
0.016
0.046
0.010
0.001
0.010
0.009
0.022
0.001
0.017
0.010
0.004
0.003
0.003
0.018
0.007
0.020
0.015
0.007
The behavior of TNT in these fortified soils parallels that of TNB except that the rate of disappearance is
reduced (Table 2). The expected transformation products, 2- and 4-AmDNT (9), are observed to increase as
TNT concentrations decline. The rapid loss of TNT for the room temperature storage condition parallels
that observed in the Oak Ridge study (4) and elsewhere for low concentration spiked soils (11, 12). Our re-
sults are quite different from those found at Oak Ridge for refrigerated storage, however. We found that for
seven days of storage, the concentration of TNT remaining was only 80, 72, and 0%, respectively for Wind-
sor, Charlton and Fort Edwards, while the Oak Ridge study found no significant TNT loss until after day
seven for the three soils studied.
When soils were frozen, our modified ASTM criterion showed no significant change for TNT in Windsor
or Charlton soils during the 56-day test period. With Fort Edwards soil the TNT concentration reached the
lower 99% confidence interval (15% change) after about 20 days. However, the total decrease was still
only 16.1% after 56 days.
The behavior of tetryl in these fortified soils parallels that of TNB and TNT. For the Windsor soil, a large
reduction in the concentration of tetryl can be seen after only three days at room temperature (Table 2). In
467
-------
'
TNB Concentration
3,5-DNA
0 20 40
Refrigerator Holding Time (days)
Figure 1. Stability of fortified TNB in Windsor soil.
addition, three apparent transformation products are also evident. After seven days under refrigeration, the
tetryl concentration is reduced by about 46% and the same transformation products are seen. Loss of tetryl
in the Charlton soil is much more rapid than found for Windsor, both at room temperature and under re-
frigeration. The same three transformation products observed in the extracts of the Windsor soil are again
evident. After seven days under refrigeration, the initial concentration of tetryl is reduced by 97%. The rate
of degradation of tetryl in the Fort Edwards clay is even faster than found for the Charlton soil, resulting in
over 99% loss of tetryl after seven days of refrigeration. The relative amounts of the various transformation
products observed for this soil appear to be different than observed for the Windsor and Charlton soils. One
of the reasons for the different behavior for this soil is its pH, which is 8.4 compared to 6.2 for Windsor and
6.0 for Charlton (Table 1). It has been observed elsewhere that hydrolysis products of tetryl were different
at acidic and basic pH and that hydrolysis proceeds much more rapidly under basic conditions (13). Perhaps
the very rapid loss of tetryl for Fort Edwards clay is partially due to hydrolysis rather than biodegradation.
A more thorough discussion of the transformation products of tetryl is presented elsewhere (14). No
statistically significant loss of tetryl was observed for freezer storage at —15° C over the seven-day study
period for the three soils studied.
468
-------
The stability of 2,4-DNT in these fortified soils is much greater than TNB, TNT, or tetryl (Table 2). At
room temperature an average of 68.4% remained after three days. This increased stability of 2,4-DNT rel-
ative to TNT at room temperature agrees with the results from Oak Ridge (4). Under refrigeration an aver-
age of 86% remained after seven days of storage. A slow rate of loss continued throughout the study and by
14 days, peaks corresponding to the expected reduction products, 2-amino-4-nitrotoluene and 4-amino-2-
nitrotoluene, appeared (9). With freezer storage, 2,4-DNT was quite stable. For the Windsor soil, our mod-
ified ASTM criterion was exceeded on the low side after 42 days. Once again, this occurred because of a
very small pooled standard deviation, and at 56 days the concentration decrease was only 4.9%. Charlton
soil showed no significant change during the 56-day test period. However, the Fort Edwards soil produced
a significant 2,4-DNT decrease after 30 days and the loss after 56 days was 20.3%. Still, the overall mean
recovery relative to day zero for the three soils was 94.2% and the mean for the 56-day time was 90.8%.
The stability of HMX and RDX in these three fortified soils is much greater than TNB, TNT, or 2,4-DNT.
This agrees with the results obtained at Oak Ridge (4) and elsewhere (15, 16), where it was demonstrated
that RDX does not biodegrade under aerobic conditions. Regardless of storage conditions, no loss of HMX
or RDX was observed over the entire 56-day study.
Table 3. Drying and storage effects on explosives concentrations in field-contaminated soils.
(Day 0 = moist, unground; day 1 = overnight air-dry, ground; day 56 = 8 weeks at 4°C in dark at field
moisture, then overnight air-dry and ground.)
Mean Concentration, ng/g dry weight basis
Soil A
DayO
Day 1
Day 56
LSD
RSD (Means')
HMX
12.2
12.4
13.3
NS
8.0%
RDX
3.32
3.51
3.74
NS
4.4%
TNB
0.12n
0.13 !
0.15-
0.025
5.1%
3,5-DNA
0.16— ,
0.17 i
0.12— '
0.032
6.0%
TNT
0.64
0.87
0.82
NS
10.2%
4-AmDNT
0.71
0.73
0.73
N
3.5%
2-AmDNT
0.54
0.57
0.52
NS
3.4%
SoilB
DayO
Day 1
Day 56
LSD
RSD (Means)
SoilC
DayO
Day 1
Day 56
LSD
RSD (Means)
1.69
1.48
1.19
NS
1.7%
0.10
0.11
0.12
NS
7.2%
0.71
0.61
0.82
NS
7.8%
NQ
NQ
NQ
—
—
0.069
0.055-
0.082
0.019
7.6%
NQ
0.047
0.036
NS
14.2%
0.15
0.12
0.15
NS
6.2%
0.077
0.16
0.058-'
0.060
28.5%
0.46
0.40
0.48
NS
11.19
0.12—
0.47-
0.35—
0.22
19.1%
0.38-
0.31
0.44-
0.05
3.7%
0.18
0.17
0.18
NS
6.7%
0.17—,
0.16- !
0.21 —
0.03
4.3%
0.18
0.20
0.20
NS
6.9%
LSD = Least significant difference for means at significance level of 5%. Means that differ are
identified by lines connecting them.
NS = No significant difference in the three means at 5% significance.
NQ = Not possible to quantify due to very low concentration.
%RSD (means) = Syp(lOO)/X^n where Syp is the pooled standard deviation from the three sets of seven
replicates per mean (n = 7), and X is the overall mean of the three sets of seven rep-
licates.
469
-------
Holding time behavior—field-contaminated soils
Field-contaminated soils from the Crane-Rockeye site (A, B, and C) were analyzed initially, both with and
without drying (days 0 and 1) and again after eight weeks of storage (moist) at 4°C in the dark (day 56).
None of these samples are believed to have received significant recent explosives contamination; i.e., the
contamination occurred several years ago. Results are summarized in Table 3. Analysis of variance (ANO-
VA) was conducted on each set of replicates for each soil and each analyte. Where means were significant-
ly different (5% significance level), least significant differences (LSDs) were computed. Specific differ-
ences in means were identified in Table 3 by lines connecting them.
Neither HMX nor RDX showed any significant differences. In general, precision of the measurements was
very satisfactory as shown by the percent relative standard deviations of the means a. Absence of sig-
nificant differences in the means confirmed nitramine stability observed previously with fortified soils.
The nitroaromatics did produce a few significant differences but, in general, these differences were small in
magnitude, were not consistent, and were of no practical importance. A more thorough discussion of the
differences observed is presented elsewhere (17). The major observation here is the vastly improved stabil-
ity of the nitroaromatics in these field-contaminated soils relative to the three fortified soils discussed pre-
viously.
The question at this point was whether the improved stability of the nitroaromatics was a reflection of dif-
ferences in binding for nitroaromatics in aged field-contaminated soils, compared to fortified soil, or just
differences in the specific soils studied. To address this question, we spiked TNT and TNB into 42 aliquots
of each of the three field-contaminated soils from Crane-Rockeye discussed above and followed the con-
1.2
^0.8
o
I
o
o
O
CO
0.4
-V
I ^
Fortified
i Field Contaminated
Field
Contaminated
I
55
65 75
Days after Sampling
85
Figure 2. Refrigerator storage effects on TNB for Crane Soil B.
470
-------
3 —
•
2 —
ion
o
O
VI 1
II
cO
T '
t 1
A 1 1
> 55
1 ' 1 1
J Fortified
i Field Contaminated
\ Fortified TNT
N~- U "
r T
[ ? Field Contaminated TNT
1 1 1 1
65 75 8
1 —
Days after Sampling
Figure 3. Refrigerator storage effects on TNT for Crane Soil B.
centration changes during storage at 4°C in the dark. Control samples were prepared by adding water to 14
aliquots of each soil in amounts equal to the volumes of aqueous spiking solution used. One set of seven al-
iquots of each spiked soil and one set of seven aliquots of each water-treated soil were analyzed on the
same day they were prepared. The remaining aliquots were immediately placed in storage at 4°C in the
dark. Because degradation was found to be slower in soil A than in B and C, the former experiment went
21 days while the latter two lasted 14 days. The second set of water-treated samples for A was analyzed on
day 21 while the B and C controls were analyzed on day eight. Results are presented for TNB and TNT in
Crane soil B (Figures 2 and 3).
In general, analyte concentrations in the water-treated control samples showed either no change (within
experimental error) or a small increase after storage (Figures 2 and 3).
For the fortified, field-contaminated samples from Crane, results were very different from the controls. For
all three soils, the concentrations of TNB and TNT declined rapidly as observed for the three fortified soils
discussed earlier (Figures 2 and 3). The disappearance was again accompanied by increases in the
concentrations of the expected transformation products. These results afford unambiguous evidence of the
dramatic differences in the stability of explosives in fortified and field-contaminated soils and cast doubt on
the ability of fortified soils to mimic soil-analyte interactions in field-contaminated soils where analytes
have been in contact for extended periods.
471
-------
CONCLUSIONS
The major observation in this study is the large stability differences for nitroaromatics in fortified and field-
contaminated soils. Even with avoidance of the addition of unnatural solvents during fortification, fortified
soils do not appear to accurately mimic soils contaminated in the field over extended periods of time. This
result has implications that extend well beyond the experimental estimation of MHTs. Spiking analytes into
various matrices is used by analytical and environmental scientists for many purposes such as (a) es-
timating the accuracy of analytical procedures via recovery determinations, (b) generating quality control
data, and (c) testing the efficacy of remediation treatments. To date there has been only limited ac-
knowledgment (1, 6, 7) that fortification as currently practiced often fails to yield an accurate repre-
sentation of results to be expected with real samples. We believe that this topic deserves much greater at-
tention in the future.
The results for the nitramines are more consistent. Whether fortified or field-contaminated, RDX and HMX
appear stable for at least 56 days when refrigerated. This agrees with the results from the Oak Ridge study
(4). Since nitramines and nitroaromatics are often co-contaminants, MHTs will seldom, if ever, be
established for nitramines alone. Nevertheless, the MHT estimates for nitramines may be useful in
determining the useability of results for samples analyzed outside normal holding times.
Results from fortified soils appear most applicable to freshly contaminated soils such as one might find
near the front of a moving groundwater plume. If MHTs are to be established using worst case estimates,
the results for the nitroaromatics in the fortified soils in this study may be useful. These results indicate that
refrigeration is insufficient to stabilize TNB, TNT, and tetryl. If samples are immediately frozen, however,
the nitroaromatics will remain acceptably stable for at least eight weeks.
ACKNOWLEDGMENTS
The authors acknowledge D.C. Leggett and M.E. Walsh of the U.S. Army Cold Regions Research and
Engineering Laboratory (CRREL) for useful comments and suggestions on the manuscript. P.N. Thorne
(CRREL) and S.M. Golden of the Science and Technology Corporation are acknowledged for assistance in
several of the experiments described. In addition, J. Scott Miller and Donald E. Parker of AScI Corpora-
tion, Vicksburg, Mississippi, are acknowledged for their assistance in soil extraction and analysis. The au-
thors also acknowledge Patricia Schumacher, Lawrence Perry, and Ginger Boitnott (CRREL) for assistance
in preparation of analytical standards, providing test soils, and measurement of soil respiration, re-
spectively. Funding for this research was provided jointly by the U.S. Army Environmental Center, Aber-
deen Proving Ground, Maryland, Martin H. Stutz, Project Monitor, and the U.S. Army Engineer Water-
ways Experiment Station, Vicksburg, Mississippi, Ann B. Strong, Project Monitor.
LITERATURE CITED
1. Jenkins, T.F., Walsh, M.E., Schumacher, P.W., Miyares, P.H., Bauer, C.F., and Grant, C.L. Journal
oftheAOAC, 1989, 72, 890-899.
2. Bauer, C.F., Koza, S.M., and Jenkins, T.F. Journal of the AOAC, 1990, 73, 541-552.
3. EPA (1992) Nitroaromatics and Nitramines by HPLC. Second Update SW846 Method 8330.
4. Maskarinec, M.P., Bayne, C.K., Johnson, L.H., Holladay S.K., Jenkins, R.A., and Tomkins, B.A.
"Stability of Explosives in Environmental Water and Soil Samples," Oak Ridge National Laboratory
Report ORNL/TM-11770, 1991, Oak Ridge, Tennessee 37831.
5. ASTM, "Standard Practice for Estimation of Holding Time for Water Samples Containing Organic
and Inorganic Constituents," Method D4841-88, 1986.
472
-------
6. Steinberg, S.M., Pignatello, J.J., and Sawhney, B.L. Environ. Sci. Technol. 1987, 21, 1201-1208.
7. Burford, M.D., Hawthorne, S.B., and Miller, DJ. Anal. Chem. 1993, 65, 1497-1505.
8. Grant, C.L., Jenkins, T.F., and Golden, S.M. "Experimental Assessment of the Analytical Holding
Times for Nitroaromatic and Nitramine Explosives in Soi/," U.S. Army Cold Regions Research and
Engineering Laboratory, Hanover, New Hampshire 03755, CRREL Special Report 93-11, 1985.
9. McCormick, N.G., Feeherry, F.E., and Levinson, H.S. Applied and Environmental Microbiology,
1976,31,949-958.
10. Walsh, M.E. "Transformation Products of Nitroaromatics and Nitramines: Literature Review and
Recommendations for Analytical Method Development," U.S. Army Cold Regions Research and
Engineering Laboratory, Hanover, New Hampshire 03755, CRREL Special Report 90-2, 1990.
11. Pennington, J.C. and Patrick, W.H. Journal of Environmental Quality, 1990, 19, 559-567.
12. Cragin, J.H., Leggett, D.C., Foley, B.T., and Schumacher, P.W. "TNT, RDX, and HMX explosives
in soils and sediments: Analysis techniques and drying losses," U.S. Army Cold Regions Research
and Engineering Laboratory, Hanover, New Hampshire 03755, CRREL Report 85-15, 1985.
13. Kayser, E.G., Burlinson, N.E., and Rosenblatt, D.H. "Kinetics of hydrolysis and products of hydroly-
sis and photolysis of tetryl," Naval Surface Weapons Center Report NSWC TR 84-68, 1984, Silver
Spring, Maryland 20910.
14. Jenkins, T.F. and Walsh, M.E. Journal of Chromatography A, 1994, 662, 178-184.
15. Harvey, S.D., Fellows, R.J., Cataldo, D.A., and Bean, R.M. Environmental Toxicology and Chem-
istry, 1991, 10, 845-855.
16. Spanggord, R.J., Mill, T., Chou, T., Mabey, W., Smith, J., and Lee, S. "Environmental fate studies
on certain munition wastewater constituents-Lab studies," Menlo Park, California: SRI International.
ADA099256, 1980.
17. Grant, C.L., Jenkins, T.F., Myers, K.F., and McCormick, E.F. Environ. Sci. and Technol. (in press).
473
-------
73
Wide-Spread and Systematic Errors in the Analysis for
FCBs in Soils
David Eugene Kimbrough, Public Health Chemist II, Rustum
Chin, Public Health Chemist III, Janice Wakakuwa, Supervising
Chemist, California Environmental Protection Agency,
Department of Toxic Substances Control, Hazardous Materials
Laboratory - Southern California, 1449 W. Temple Street, Los
Angeles California 90026-5698.
Abstract
The analysis of PCBs in soils is one of the most widely
performed tests among environmental laboratories. It is a
very difficult test and is subject to a great deal of inter-
laboratory variance and bias. A new inter-laboratory study,
using five soils spiked at four different concentrations with
Aroclor 1260, was conducted with two groups of laboratories
numbering 20 and 129 respectively. The results of this study
are compared with the results of the other studies but show
that the bias is concentration dependent. A number of
definite patterns of recovery are noted indicating that the
large variance and bias was not due to random errors but to
wide-spread systematic errors.
Statistically significant differences were found between
laboratories depending on the type of extraction equipment,
solvents, and clean-up procedures that were used. Results
from laboratories using Soxhlet extraction showed
significantly more accurate results than did sonication,
especially at higher concentrations but with equal precision.
Results from laboratories using non-polar solvents showed
significantly lower accuracy than more polar solvents with
equal precision. Results from laboratories using Florisil§
column clean-up showed signifcantly more accurate and precise
results at lower concentrations than laboratories using no
clean-up procedure.
Other significant variables that affect the accuracy of
the analysis of PCBs in soils is the linear dynamic range,
calibration range, and detector drift of the gas
chromatographic instrument. There were three types of
errors: one was calibrating the instrument within the linear
dynamic but analyzing samples above and below this range.
The second was calibrating the instrument outside the linear
dynamic range. The third was to allow excessive drift in the
detector. Each of these errors produced a charateristic
pattern of biased results. The net result was a large inter-
laboratory variance and biases.
474
-------
74
An Evaluation of Gas Chromatography/Ion Trap Mass
Spectrometry for Analysis of Environmental
Organochlorine Pesticides
Gary Robertson, U.S. EPA, Environmental Systems
Monitoring Laboratory, Las Vegas, NV 89193
James Barren, U.S. EPA, Office of Solid Waste and
Emergency Response, Washington, D.C. 20460
The U.S. EPA is continually making efforts to
improve the quality of analytical data and supporting
documentation used for making decisions about
environmental contamination. A research project
evaluating the use of Gas Chromatography/Ion Trap Mass
Spectrometry (GC/ITMS) for the analysis of
organochlorine pesticides is being conducted by the
Environmental Monitoring Systems Laboratory-Las Vegas
and the Analytical Operations Branch of the Office of
Solid Waste and Emergency Response. The adoption of a
mass spectrometric method for the detection of
pesticides would provide the same assurances of
identification and quantitation as the Contract
Laboratory Program (CLP) analysis for semivolatile
compounds. This would reduce the costs involved in
data review, and provide the data user a more reliable
product.
The research has concentrated on the CLP list of
organochlorine pesticides. Instrumental operating
conditions and detection limits for pesticides on the
GC/ITMS have been established. Minor modifications in
the sample concentration step of the CLP method have
been developed to increase detectability. The effects
of interferences have been evaluated using samples
containing both synthetic and native interferences.
Results of these studies, which show GC/ITMS to be a
promising technique, will be discussed, and a
comparison will be made to current CLP quantitation
limits. Planned additional research will also be
discussed.
475
-------
75
RAPID AND COST-EFFECTIVE ANALYSIS OF 2,3,7,8-TCDD USING THE "DIOXIN
RISC®" IMMUNOASSAY TEST KIT
R. Allen, T. Stewart, D. Reynolds, S. Friedman, EnSys, Inc.
Research Triangle Park, North Carolina 27709
ABSTRACT
Current methods for detecting low levels of Dioxin are expensive
and laborious. We have developed a rapid and sensitive enzyme
immunoassay, DIOXIN RISc®, for the detection of 2,3,7,8-TCDD in a
variety of matrices. The immunoassay is designed to incorporate
sample processing protocols which are routinely used in research
and analytical laboratories. However, many of the steps used in
processing dioxin samples can be eliminated because of the inherent
specificity and affinity of the immunochemical reagents. The test
easily detects ppt to ppq levels of 2,3,7,8-TCDD in samples which
have been extracted and taken to dryness. The immunoassay shows
less than 0.01% cross-reactivity with PCBs, PAHs, Chlorophenols and
Chlorinated Aromatic Pesticides. The immunoassay recognizes 2,3,7-
TriCDD (20%) and 2,3,7,8-TCDF (5%) but does not significantly
cross-react with other Dioxin and Furan Congeners. DIOXIN RISc®
offers a simple, rapid, reliable and cost-effective alternative to
current methods for dioxin analysis.
476
-------
76
AUTOMATED SOXHLET EXTRACTION AND CONCENTRATION OF
SEMTVOLATILE ANALYTES IN SOIL SAMPLES
E. E. Conrad, K. P. Kelly,
ABC Instruments (a division of Laboratory Automation, Inc.), Columbia, MO 65202
INTRODUCTION
US EPA Draft SW-846 Method 3541 describes an automated Soxhlet technique which
reduces solvent consumption and speeds up sample processing relative to traditional
Soxhlet extraction, such as SW-846 Method 3540. Compared to the traditional method,
this newer technique uses half the solvent or less and requires under three hours to
complete compared to 16 hours or longer. Some laboratories have avoided the Soxhlet
extraction because of the long processing times required under promulgated Soxhlet
methods. With faster and more automated Soxhlet extraction available, the advantages
of decreased solvent consumption and reduced labor investment make Soxhlet extraction
a more attractive choice when compared with other techniques such as ultrasonication
extraction (e.g. SW-846 Method 3550).
The automated technique is similar to traditional Soxhlet extraction, but extraction occurs
in two distinct stages with this procedure. First there is an initial boiling time, during
which the sample is immersed in boiling extraction solvent. The second stage resembles
traditional Soxhlet-type extraction, with condensate dripping through the sample thimble;
however, the stream of extraction solvent is continuous rather than soaking the thimble
with batches of condensed solvent. The technique used in these experiments follows the
chemistry of Method 3541, but it provides a higher degree of automation by eliminating
the need for operator intervention during extraction. Between the two extraction stages
and following the extraction, solvent volume is automatically reduced in increments by
diverting measured portions of the solvent condensate to a collection tank rather than
allowing it to return to the extraction vessel (see Figure). This allows automated
concentration of the extract to a small volume. Experiments are reported here using
spiked samples to investigate automated Soxhlet extraction and evaporation recoveries of
semivolatile analytes.
An additional advantage of the technique used for this work is automated collection of
nearly all solvent used during the sample preparation process. Without operator
intervention required, the system diverts condensed solvent vapors to a collection tank
where the spent solvent is available for recycling or appropriate disposal. This reduces
emissions of hazardous chemicals into the atmosphere.
477
-------
Collecting device
with defined volume
Air pressure pulse
Cooling wafer ouflef
Cooling wafer inlet
Solvent tank
478
-------
PROCEDURES
The extraction instrument (ABC Instruments Soxtherm) was programmed with parameters
appropriate to Method 3541. Heater temperature was 160° C for samples extracted with
a 1:1 mixture of acetone and hexane or methylene chloride. Operation of the instrument
was unattended and run time was slightly over two hours, including evaporation. Ten
grams of sand or prepared soil mixture (clay soil and sand) was spiked with analytes at
one or more levels using solutions in acetone or methanol. Samples were contained in
preextracted single-thickness paper thimbles (33 mm diameter). A stainless steel holder
suspended the extraction thimble and sample in the glass extraction beaker.
Solvent reduction parameters were chosen to produce final volumes that facilitated further
sample processing. Concentrated extracts were removed from the extraction beakers and
the concentrate was further evaporated using the microSnyder technique and/or a nitrogen
blowdown to produce final volumes as low as 1 mL. Hexane solvent exchange was
employed for mixtures that were analyzed using electron capture detection (BCD).
Analytes in processed samples were quantitated using gas chromatography and results
were compared to chromatograms obtained from the spiking solutions used.
EXPERIMENTAL PROGRAM
Various subgroups of semivolatile analytes were selected for spiking. Since the method
includes an evaporation step, analytes which would reflect losses during reduction of
concentrates to small volumes were included. Adequate control of the final concentrate
volume is necessary to obtain good recoveries of such analytes. The instrument uses
gaskets located between the glass extraction beaker and the upper portion of the
extraction apparatus to ensure operation with minimal solvent leakage. Some gaskets are
prone to cause sample contamination, thus more than one gasket type was investigated.
479
-------
77
EVALUATION OF A QUANTITATIVE IMMUNOASSAY FIELD
SCREENING METHOD FOR DETERMINATION OF
PENTACHLOROPHENOL IN SOIL AND WATER
Marv C. Haves. Scott W. Jourdan and David P. Herzog, Ohmicron Environmental
Diagnostics, 375 Pheasant Run, Newtown, Pennsylvania 18940
ABSTRACT
Analysis of environmental samples for pentachlorophenol (PCP) using EPA approved
extraction and measurement methods is typically costly, time consuming and requires
specialized equipment. Immunoassay technologies have the potential to quickly and
inexpensively screen for PCP in both soil and water. In the summer of 1993, a
demonstration and evaluation of immunoassay screening methods for determination of
PCP in soil and water was conducted under the EPA EMSL, Las Vegas, Nevada,
Superfimd Innovative Technology Evaluation (SITE) Program by PRC Environmental
Management, Inc. The Pentachlorophenol RaPID Assay®, a quantitative immunoassay
for PCP, utilizing a three point calibration curve, was one of the methods evaluated. The
accuracy and precision of the RaPID Assay on water and soil samples from two
contaminated sites is compared to the same parameters for a conventional confirmatory
laboratory using EPA approved analytical methods.
INTRODUCTION
Pentachlorophenol (PCP) is a wood preservative which has been used extensively by the
wood treating industry since the 1950's. PCP is a regulated chemical, is included in the
EPA Extremely Hazardous Substance List, and is reported in the EPA Toxic Substance
Control Act Inventory. PCP is included as a target compound of many EPA approved
analytical methods. These analytical methods use solvent extraction and gas
chromatography for separating PCP from other compounds in the sample. Analyzing
samples for PCP using these methods is typically costly and time consuming.
The Environmental Monitoring Systems Laboratory (EMSL) of the US EPA identified the
need for effective, accurate, low cost screening technologies that could provide near real-
time analytical data. Immunoassay technologies were selected for a demonstration, each
with the potential to quickly and inexpensively screen PCP in both water and soil samples.
Administrative and technical support for the demonstration was contracted to PRC
Environmental Management, Inc. of Kansas City, KS. (PRC).
The purpose of the demonstration was to evaluate each of the field screening technologies
for accuracy and precision in detecting high and low levels of PCP in soil and water
samples. Each technology was also evaluated for the length of time required for analysis,
ease of use, portability, and operating costs. The accuracy and precision of each
480
-------
technology was compared to results obtained by a confirmatory laboratory using a GC/MS
EPA analytical methods (Method 8270). Statistical comparisons were used to determine
the highest data quality level that each technology could attain in field applications. For
the purposes of this demonstration, three primary data quality levels were defined. (1)
Level 1 data quality provides only an indication of contamination. This data is not
necessarily analyte specific. (2) Level 2 data quality provides analyte specific data. To
provide an accuracy check, verification analysis for at least 10 percent of the samples by
an EPA-approved method is necessary. (3) Level 3 data quality provides formal or
confirmatory analysis. The method is considered analyte specific and generally involves
second-method confirmation on 100 percent of critical samples.
The PCP RaPID Assay, and the RaPID Prep Soil Collection and Extraction Kits
(Ohmicron Environmental Diagnostics) were evaluated as one of the immunoassay
methods. The demonstration took place in August of 1993 at a Superfund site. Recently,
the results of this evaluation were presented in a Draft Technical Evaluation Report
prepared by PRC and submitted to the US EPA. A summary of the RaPID Assay results
is presented here.
DESCRIPTION OF DEMONSTRATION
At the time of the demonstration, soil and water samples were collected from two PCP
contaminated sites, the former Koppers site near Morrisville, North Carolina and the
Winona Post site, Winona, Missouri. A previous investigation found levels of PCP
ranging from non-detect to several parts per thousand in water and soil at both sites. For
the demonstration, 53 soil and 5 ground water samples were collected from the Koppers
site and 45 soil and 5 surface water samples were collected from the Winona Post site.
To evaluate the field screening technologies under field conditions, the demonstration of
immunoassay tests were conducted at the Koppers site. The immunoassays were
performed on-site in a portable trailer. Samples collected at the Winona Post site were
sent overnight to the Koppers site for the demonstration. Samples collected from each
site were mixed thoroughly for homogeneity and then split for analysis by each
immunoassay technology and by the confirmatory lab, the US EPA Region 7 Laboratory.
Analysis of the demonstration samples was completed within 9 days of sample collection
for each of the immunoassay technologies. Turnaround times for samples analyzed by the
confirmatory lab ranged from 14 to 30 days.
Each immunoassay method was demonstrated by a designated PRC operator. A one day
training session was given by an Ohmicron technical representative for the operator to
familiarize him with the PCP RaPID Assay and the RaPID Prep Soil Collection and
Extraction Kits.
481
-------
STATISTICAL ANALYSIS OF DATA
This demonstration consisted of comparisons of various groups of data. Non-detectable
concentrations were eliminated from the comparison. Outlier tests were then applied to
eliminate the six poorest correlating non-zero data points. Each data group was then
analyzed in a similar fashion. For the immunoassay methods, two data sets were created,
one for soil samples and the other for water samples. In addition, each water and soil data
set was composed to two subsets, one produced by the samples taken from Koppers site
and one produced from the samples collected at Winona Post site. A third subdivision
involved the grouping of the site-specific data sets into results greater than 100 ppm or
less than 100 ppm PCP The 100 ppm level represented a regulatory or action level for
the demonstration
The data generated by quantitative immunoassay methods was subjected to linear
regression analysis to compare concentration results of the immunoassay to those given by
the confirmatory lab, considered the "true" concentration of PCP for each sample.
Another statistical method used for assessing intermethod accuracy was the Wilcoxon
Signed Rank Test. This test can be used to evaluate whether two sets of data differ
significantly. Both the Wilcoxon Signed Rank Test and linear regression analysis were
used to determine the level of quality data (e.g. Level 1, 2 or 3) produced by each of the
quantitative immunoassays
Calculating linear regression makes it possible to determine whether two sets of data are
reasonably related, if so, how closely. When a linear regression is calculated, results are
expressed in an equation (y = mx + b) that can also be visually expressed as a line drawn
through an x-y plot of the datasets. The y-intercept (b), the slope of the line (m), and the
correlation coefficient (r2) determined by linear regression were used to assess quality of
data generated by the immunoassays. The r2 expresses the mathematical relationship
between the two data sets. If r2 is one, then the two data sets are directly related. Lower
r2 values indicates less of a relationship. To meet Level 3 requirements, r2 was required to
be 0.85 to 1 and the slope and y-intercept had to be statistically the same as their ideal
values. If the r2 was between 0.75 and 1, and the slope and/or the y-intercept was not
equal to the ideal value, the technology was considered inaccurate but capable of
producing Level 2 quality data. Data placed in the Level 1 category had r2 values less
than 0.75, the data was not statistically similar to the confirmatory lab, based on
parametric testing, or the results failed to meet the developer's performance specifications.
Field soil and water duplicate samples were analyzed during the demonstration to compare
the precision of each immunoassay technology to the precision of the confirmatory lab.
The Dunnett's Test and Wilcoxon Rank Sum Test were used to determine intermethod
precision for quantitative immunoassays.
482
-------
IMMUNOASSAY METHOD
The RaPID Prep™ PCP Sample Extraction Kit was used to process all soil samples. The
PCP RaPID Assay produced quantitative results for water and diluted soil extracts. Water
samples are pipeted directly into the immunoassay reaction tube. Soil samples are
extracted in a special device with a methanol solvent for five minutes prior to filtration,
dilution and subsequent introduction into the reaction tube. This immunoassay produces
quantitative results by comparing the optical density (O.D.) of the colored reaction
byproduct of unknown soil extracts and waters to the color produced by PCP calibrators
in a standard curve run simultaneously (Figure 1). Absorbance readings are transformed
to PCP concentrations using a log-logit algorithm programed into the RPA-I™ RaPID
Photometric Analyzer. The detection range for water is 0.06 to 10 ppb. The detection
limit for soil 100 ppb to 10 ppm. Samples giving concentrations above the upper
detection limit were diluted until the O.D. of the diluted unknown fell within the O.D.
range of the kit calibrators. The additional dilution factors were then applied to the PCP
result.
SUMMARY OF RESULTS
A summary of intermethod accuracy, as determined by linear regression analysis is
presented in Table 1. The combined soil data set (Figure 2) and the combined water data
sets (Figure 3), both fell into the Level 2 data quality category. Separation of these data
sets by site and by concentration produced some data groupings of Level 1 data quality
category. Lower r on these data groupings suggest a possible concentration and/or site
effects on the two methods. Possibilities for these differences include: 1) technique
variability in handling or preparing extracts for each of the technologies; 2) extraction
efficiency between the immunoassay method and the more vigorous EPA extraction
method (especially at concentrations >1000 ppm); and 3) sample size differences used in
analysis by immunoassay and the confirmation lab. By grouping, the assay produced
Level 2 quality data for water samples at the Koppers site and soil samples at the Winona
Post site. The assay produced Level 1 quality data for water samples at the Winona Post
site and soil samples at the Koppers site.
INTERMETHOD ACCURACY ON FIELD SOIL SAMPLES
The RaPID Prep PCP Sample Extraction Kit package insert indicates that the kit system is
expected to give less than 100% quantitative recovery on soil samples at the short
extraction time called for in the procedure. In house studies on three soils types including
clays and loams fortified with PCP to final concentrations of 1.0, 1.5 and 5 ppm gave
spike recovery results that ranged from 63 to 83%. Greater analyte recovery can be
achieved if extraction time is extended from five minutes up to thirty minutes. If RaPID
Assay PCP concentrations are used as screening results for recognition of 100 ppm as
shown in Figure 4, a 70 ppm cutoff can be imposed to correct for the diminished
483
-------
extraction efficiency with the short extraction time. When this is done, the false negative rate
drops to 6 in 84 (7%) from 7 in 84 (8%) while maintaining the same false positive rate (4%).
A typical user of Ohmicron's kits would be advised to make a constant additional dilution
prior to analysis if a regulatory or action level of 100 ppm was being targeted. This would
reduce the frequency of additional sample dilution and re-assays and improve precision and
accuracy by placing the decision point in the middle of the immunoassay calibration curve.
Users are instructed to make simple calculations for factors affecting extraction efficiency as
well as analytical confidence (precision) as shown in Figures 5 and 6. When an analytical
confidence factor of 0.7 is applied to the data in Figure 2 to minimize false negatives at the
100 ppm level, the false negative rate drops to 3 in 84 (4%) while the false positive rate
changes by just 1 in 84 to 5% in this data set. This treatment is shown in Figure 4.
The ability of the RaPID Prep system to recover 5 ppm spikes in a variety of soil samples was
excellent as shown in Table 2. This was the case for these blank field soils notwithstanding
the somewhat lower recoveries of PCP from some randomly selected clay and loam samples
tested in our laboratories. Unfortunately the reference method was not used to demonstrate
recovery on these same samples so no comparison can be made for this classic test of method
accuracy.
The precision of Ohmicron's kits on field soil and water was similar to that of the
confirmatory lab's precision as demonstrated on the performance of both methods on blind
duplicates of the same sample in Table 3. A recently completed AOAC Collaborative study of
the Ohmicron Atrazine RaPID Assay kit with drinking and surface waters showed the
immunoassay to possess precision equivalent to EPA Method 507 for drinking water analysis
at similar atrazine concentrations.
CONCLUSION
Overall, the accuracy and precision of Ohmicron kits was very good despite the fact that soil
data sets were not corrected for extraction efficiencies. Intermethod accuracy on the total soil
and water data sets, placed the Ohmicron kits in Level 2 data quality category. Technologies
that produce Level 2 quality data can be used to guide field work and sampling efforts. These
technologies will provide data, or can be corrected to provide data, which corresponds to
confirmatory results. Since many of the data groupings produced by Ohmicron kits were
found to be linear, the results could be corrected mathematically. If 10 or 20 percent of the
soil samples are sent to a confirmatory lab for comparison of analyte concentrations, then the
results from the other 80 to 90 percent can be corrected for any bias and reported with
analytical confidence. This approach could result in significant savings in analytical costs.
This information is presented with the permission of Mr. Larry Jack, EPA, EMSL, Las Vegas,
NV, sponsor of the Superfund SITE Demonstration.
484
-------
Figure 1. Pentachlorophenol RaPID Assay calibration curve
100
80
m 60
40
20
0.1
Pentachlorophenol (ppb)
10
485
-------
Figure 2. Total soil data set
10000
r
100
0.01
0.01
1 100
EPA Confirmatory Lab Result (ppm)
10000
486
-------
Figure 3. Total water data set.
100
s
§ 0.01
0.0001
o
cP
0
o
8
0.0001
o
o
0.01 1
EPA Confirmatory Lab Result (ppm)
100
487
-------
Figure 4. Total soil data set showing cutoffs for
extraction efficiency and analytical confidence factors.
10000
I
a
ca
Pi
PH
0.01
0.01
1 100
EPA Confirmatory Lab Result (ppm)
10000
488
-------
FIGURE 5. CALCULATION OF A CUTOFF CONCENTRATION
WORKSHEET
Required Detection Limit
Cross-Reactivity Factor
(from Table 1)
Extraction Recovery Factor
Analytical Confidence Factor
(from Table 2)
Cutoff Concentration
Ax B x C x D
(B)
.(Q
Example -
To optimize the PCB RaPID Assay system for detection Aroclor 1260 at 10 ppm in soil:
Required Detection Limit 10 ppm (A)
Cross-Reactivity Factor
(from Table 1)
Extraction Recovery Factor
Analytical Confidence Factor
(from Table 2)
0.81 (B)
0.85 (Q
0.80 (D)
Cutoff Concentration
5 ppm (E)
A Detection Limit of 10 ppm of Arochlor 1260 is
assumed for this example. See Detection Limit in
text above.
Since the PCB to be detected in this example is
Aroclor 1260, the Cross-Reactivity Factor
obtained from Table 1 of the PCB RaPID Assay
Performance Characteristics is 0.81.
The Extraction Recovery Factor of 0.85 is
obtained from the RaPID Prep Sample Extraction
kit package insert. Preferably, this factor would be
obtained from a spiked matrix determination.
By examining Table 2 for PCB in Soil it can be
observed that an Analytical Confidence Factor of
0.8 is estimated to yield 96.1% negative results at a
PCB concentration of 5 ppm Arochlor 1260 (0.5 x
Detection Limit[\0 ppm Arochlor 1260]). In this
example it is judged that a 3.9% "false positive"
rate would be acceptable for samples at 5 ppm. It
is also noted that at 10 ppm (Detection Limit) the
estimated rate of positive results is 88% while at 20
ppm (2.0 x Detection Limit) the incidence for false
negative results is estimated to be <0.1%.
Performing the calculation AxBxCxD the result
is 5.5 ppm. In this example the Cutoff
Concentration used for the PCB assay result (after
correction for dilution) was rounded down to 5
ppm to be conservative.
489
-------
FIGURE 6. ESTIMATING REQUIRED ANALYTICAL
CONFIDENCE LIMITS
PENTACHLOROPHENOL RaPID ASSAY CHARACTERISTICS
Table 1 - Cross- Reactivity Factors for Pentachlorophenol RaPED Assay
Compound
Pentachlorophenol
2,3,5,6-
Tetrachlorophenol
2,3,4,6-
Tetrachlorophenol
Cross-
Reactivity
Factor
1.00
0.54
0.15
Table 2 - Analytical Confidence Factor Data
Pentachlorophenol in Water
Analytical
Confidence
Factor
0,9
0.8
07
0.6
0.5
@ 0.5 X Detection
Limit
%
Negative
99,4
97.0
894
73.4
50.0
%
Positive
0..6
3.0
10.6
26.6
50.0
@ 1.0 X Detection
Limit
%
Negative
26,6
10.6
3.0
0.6
<0.1
%
Positive
73,4
89.4
97,0
99.4
>99.9
@ 2.0 X Detection
Limit
%
Negative
99.9
>99,9
>99,9
>99.9
>99.9
Pentachlorophenol in Soil
Analytical
Confidence
Factor
0.9
0.8
0.7
0.6
0.5
@ 0.5 X Detection
Limit
%
Negative
95.9
90.4
80.8
66.7
50.0
%
Positive
4.1
9.6
19.2
33.3 ,
50.0
@ 1.0 X Detection
Limit
%
Negative
33.3
19.2
9.6
4.1
1.5
%
Positive
66.7
80,8
90.4
95.9
98.5
@ 2.0 X Detection
Limit
%
Negative
99.9
>99.9
>99.9
>99.9
>99.9
Note - The shaded data achieves false positive rates of around 10% or less at 0.5 x Detection
Limit and less than 5% false negative rates at 2.0 x Detection Limit.
490
-------
TABLE 1
SUMMARY OF INTERMETHOD REGRESSION ANALYSIS
All Data
AllData<100ppm
AllData>100ppm
Koppers All Data
Koppers <100 ppm
Koppers >100 ppm
Winona All Data
Winona<100 ppm
Winona >100 ppm
N
84
46
33
48
32
13
33
15
20
r2
.81
.76
.68
.65
.61
.60
.90
.69
.75
Y-int
28
.55
71
-3.4
1.5
-110
34
-3.8
98
Slope
.43
.71
.43
.37
.57
.39
.51
1.2
.46
N = Number of data points
r2 = Coefficient of correlation
Y-int. = Y-axis intercept of the regression line
491
-------
TABLE 2
PCP RAPID ASSAY
SOIL MATRIX SPIKE SAMPLE RESULTS
Sample
No.
ARD26020
ARD26029
ARD26039
ARD26051
ARD26086
ARD26090
ARD26095
Amount Found in
Orig. Sample
(ppm)
0.13
ND
ND
ND
1.86
ND
ND
Amount Added to
Matrix Spike
Sample (ppm)
5.0
5.0
5.0
5.0
5.0
5.0
5.0
DETN1
(ppm)
3.60
4.57
5.50
5.20
8.46
5.87
6.31
% Recovery
70
91
110
104
123
117
126
DETN2
(ppm)
4.21
4.65
6.23
4.74
8.68
4.95
4.94
% Recovery
82
93
125
95
126
99
99
Average % Recovery of 5 ppm spike (n = 14) = 104%
492
-------
TABLE 3
SOIL FIELD DUPLICATE SAMPLE RESULTS
PCP RAPID ASSAY SYSTEM AND EPA METHODS 8270A AND 8151A
Sample No.
ARD26001
ARD26011
ARD26020
ARD26030
ARD26040
ARD26048
ARD26050
ARD26055
ARD26058
ARD26059
ARD26073
ARD26074
ARD26086
ARD26087
AVERAGE
PCP Rapid Assay
Original
Sample
Result
(ppm)
1.61
55
0.13
8.3
19
11,800
1.17
748
2.92
1,800
123
649
1.86
29
Field Duplicate
Sample Result
(ppm)
2.53
64
ND
8.6
10
11,700
1.46
670
2.16
2,200
125
690
7.21
20.2
Relative Percent
Difference
(%)
44
15
NA
3
62
1
22
11
30
20
2
6
118
36
28 (n=13)
EPA Methods
Original
Sample
Result
(ppm)
4.42
106
0.10
28.6
400
26,100
2.16
3,135
3.53
9,600
74.8
836
6.59
34
Field Duplicate
Sample Result
(ppm)
4.18
112
0.09
29.0
34.4
30,260
1.25
3,003
9.13
10,260
78.2
1,520
688
51.8
Relative Percent
Difference
(%)
6
6
11
1
168
15
53
4
88
7
4
58
4
41
33 (n=14)
CO
co
-------
78
Screening for Low Level Chlorinated Pesticides Using Solid Phase Extraction
Lydia Nolan, Chemist, Robert Shirey, Chemist, Research and Development, Supelco, Inc.,
Supelco Park, Bellefonte, Pennsylvania 16823
ABSTRACT
The US Environmental Protection Agency has continued in its efforts to develop
and implement new methods designed to assist in the determination and identification of
potentially hazardous abandoned waste disposal sites. The Quick Turnaround Methods
(QTM) procedures are being developed as screening methods to allow quick decisions to
be made about the extent and nature of hazardous contamination at a given site. The main
emphasis of these methods is to provide rapid, on site sample preparation and analysis.
For this reason, solid phase extraction (SPE) has been the main focus for semivolatile
sample preparation of aqueous samples.
This study will present results of extractions using an octylsilane modified, silica-
based SPE tube that provides reproducible results as well as the extremely low
background levels required for analysis of these compounds. The polymerically bonded
phase provides sufficient capacity to prevent breakthrough of analytes up to the 250 ng
per 100 mL sample level. Five lots of tubes were tested with twenty target analytes and
statistically analyzed for variability. Recoveries of spiked reagent water samples ranged
from 91% to 110%, with variability below 10% except for endosulfan II and endrin
ketone. Results also indicate that the pesticides method requires an exceptionally inert
sampling system as well, since sample concentrations are as low as 10 ng per 100 mL of
water. The effect of sampling system interferences on the reliability of extraction data will
also be demonstrated. The study analysis shows that these tubes can be used, with a low
background sampling system, to meet the proposed extraction performance criteria for this
method.
INTRODUCTION
The main emphasis of the CLP QTM methods is to provide rapid, on-site sample
preparation and analysis. For this reason, SPE has remained a major focus for semivolatile
sample preparation of aqueous hazardous waste samples. An octylsilane-modified, silica
based SPE cartridge has been named in the draft procedure released 3/92. Octadecylsilane
modified, silica-based cartridges were approved for the extraction of organochlorine
pesticides from drinking water in EPA Method 508 [1] and various published studies have
cited their use for organochlorine pesticides in other matrices [2,3,4]. Performance of a
polymerically bonded octylsilane phase cartridge, as specified in the draft procedure, was
evaluated in this study.
494
-------
Sample volumes are limited to 100 mL, and can be introduced directly into the
SPE tube from the sample collection bottle using a novel Teflon® line sampling system and
a vacuum source. The tubes are first conditioned to wet the packing and solvate the
phase, maximizing interaction of hydrophobic analytes with the stationary phase. The
sample pH is adjusted and methanol added to maintain conditioning of the adsorbent bed.
The sample is introduced to the conditioned tube at a rate of 5 mL /minute. Sampling is
completed in 20 minutes and the packing is dried using direct nitrogen purge. The
extraction solvent is allowed to soak into the dried bed and eluted into a glass receiving
vial. The eluates are dried to a constant 1 mL volume. Extracted samples are ready for
analysis using capillary GC with an electron capture detector.
Extractions to analyze background levels of various systems were also run. All
cartridge device extractions were performed using a vacuum manifold with disposable
Teflon valve liners, reducing extraneous peaks and any possible carryover from sample to
sample. Standard polypropylene cartridges with polyethylene frits were extracted and
compared to glass cartridges with Teflon frits and Teflon-lined polypropylene tubes with
Teflon frits.
The device with the lowest acceptable background levels was used to extract
spiked reagent water samples. Recovery results of 5 different bonded lots of material are
reported. Effects of sample pH, drying time and analyte capacity on the extraction
efficiency of the cartridge were examined. Optimum conditions were then used for the
extraction of spiked wastewater and hazardous waste samples.
EXPERIMENTAL
Materials:
Chlorinated pesticides standards were obtained as a custom mix and concentrated
system monitoring compound solution from Supelco, Inc. Working solutions were diluted
as specified in the draft EPA method procedure. Solid phase extraction tubes contained
ENVI™-8 silica based packing material and were extracted using a DL model Visiprep
vacuum manifold with disposable Teflon liners. Wastewater and hazardous waste,
prescreened field samples, were delivered to the tubes using a Large Volume Sampler
system.
Instrumentation:
Samples and standards were analyzed using an HP 5890 Series II capillary gas
chromatograph with BCD detector and an HP 7673 autosampler. The capillary column
was a PTE™-5 QTM, 15m x 0.53mm ID, 0.5um film. Temperature program and
conditions as described in the draft EPA method procedure. A cold, on-column injector
was not used.
495
-------
RESULTS AND DISCUSSION
The minimum background level for the analytical system was 84 ng/mL. This was
determined by spiking the system monitoring compound into 15 mL of extraction solvent
solution, reducing the volume with nitrogen purge to 1 mL and comparing all peak areas
during the run to the standard peak response. Background analyses of Teflon lined tubes
with Teflon frits exhibited high levels of interferences. Pretreatment with heat or solvent
rinsing did not produce sufficient reduction of those interferences to allow use of these
tubes for the extraction study (see chromatograms in Figure 1). The average background
level for glass tubes with Teflon frits was 109 ng/mL, with no interferences coeluting at
the same retention times as quantitated peaks. Polypropylene tubes with Teflon frits had
slightly higher backgrounds of 191 ng/mL however, the presence of more coeluting
interferences would make low level quantitation unreliable. Results are summarized in
Figure 2.
The extraction procedure recommended in the draft EPA method was used to
determine lot-to-lot reproducibility for five bonding lots of ENVI-8 adsorbent packing.
Results are summarized in Table 1. Although overall results were acceptable, anomalies in
the recoveries of key compounds were observed during these extractions. Recovery of
endrin was occasionally very low, with elevated recoveries of the aldehyde and ketone
forms. Assuming pH could be a factor in the conversion of these analytes, extractions
were done using a single lot, adjusting the sample to three different pH levels. This factor
had little effect on these compounds (Table 2). Suspecting volatility as another possible
factor, eluate drying times were also tested. Table 3 represents a side-by-side comparison
using two different lots, with reduction in overall recoveries of all analytes when excessive
drying occurs. Extraction volumes were reduced to 10 mL to minimize drying times.
Using the low background tube extraction system and reduced extraction volume,
wastewater and aqueous hazardous waste samples were prepared. Recoveries are
reported in Table 4. As expected, recoveries were lower and more variable than with
reagent water samples. Increasing the bed weight did not significantly increase the
recovery rates or decrease the co-extractants responsible for the variability.
CONCLUSIONS
SPE tubes with glass cartridges and Teflon frits, packed with ENVT-8 phase
material, meet the acceptance criteria outlined in the draft EPA QTM CLP pesticide
method released 3/92. The low background extraction system will reduce laboratory
interferences and cross contamination. SPE tubes used with a large volume sampler will
minimize hands-on sample preparation time, essential for on-site and in-lab screening
procedures. The optimized extraction procedure was acceptable for the screening of
spiked wastewater and aqueous hazardous wastes examined in this study; however, high
levels of coextractants remain in the eluates making very low level analysis difficult.
Further study, using other solid phase technologies, are necessary to improve reliability at
496
-------
the low levels specified in the draft method, especially for complex hazardous waste field
samples.
497
-------
FIGURE 1
Sample : CLP QTM Pesticides Standard 10 ng/mL
Sample : Extracted Tube Blank, Teflon -lined / Tefloo frit
1.
2.
3.
4.
3.
t.
7.
6.
9.
It.
11.
J.2.
13.
14.
13.
16.
17.
If.
19.
29.
21.
Alphm-BHC
B0tm-BHC
B»mmu-BHC
Ofltu-BHC
Hfptmohlor
Aldrln
Hoptmohlor mponldr
Bmmmu-Chlordmnm
Kodomulfun I
Alphm-ChlorOmn*
p.p'-DDf
fndrln
fndofultmn II
f,p'-DDO
Eadrl.n mldrhydf
Cndofultmn mulfmt
p.p'-DDT
Kndrln k»ton»
.00 5.00 10DO 15.00 20.00
Retention time In minutes
Sample : Extracted Tube Blank. Polypropylene, ENVI-B
0.00 5.00 10DO 15.00 20DO 25DO
Retention time In minutes
Sample : Extracted Tuba Blank. Glaaa/Teflon frit; ENVI-6
00 5.00 10UO 15.00
Retention time In minutes
—I—
20.00
0.00
—I—
5.00
10.00
15.00
20.00
25DO
498
-------
FIGURE 2
Background Levels
2500 T
2283
Standard 10 Teflon lined Polypropylene Glass/Teflon
ng/mL
499
-------
TABLE 1
Extraction of CLP QTM Pesticides
Recovery Study
Analytes
a-BHC
P-BHC
7-BHC
8-BHC
Heptachlor
Aldrin
Heptachlor epoxide
•y-Chlordane
Endosulfan I
oc-Chlordane
Dieldrin
p,p' - DDE
Endrin
Endosulfan II
p,p' - ODD
Endrin aldehyde
Endosulfan sulfate
p,p' - DDT
Endrin ketone
Methoxychlor
Decachlorobiphenyl
(SMC)
SPE Capacity Level
High Calibration Standard
Percent
recovery
91.4
100.1
94.0
92.9
95.0
94.9
96.8
97.4
94.6
96.4
98.2
99.9
97.9
98.2
104.4
101.0
104.6
107.9
110.4
108.4
101.0
Std dev
7.0
10.1
6.1
7.2
5.7
4.8
4.2
4.5
4.7
4.6
3.8
4.2
13.2
3.8
5.4
5.5
3.9
4.9
14.3
3.2
7.0
Relative
Std dev
7.7
10.1
6.4
7.7
6.0
5.1
4.4
4.6
4.9
4.7
3.9
4.2
13.5
3.8
5.2
5.5
3.7
4.5
12.9
2.9
6.9
SPE Contamination Level
Low Calibration Standard
Percent
recovery
84.6
88.2
85.1
86.6
84.7
96.3
85.3
84.8
90.2
88.6
80.1
77.8
90.0
98.6
coelute
96.2
95.3
104.6
93.4
72.9
Std dev
9.2
10.8
7.3
6.4
13.9
5.6
6.3
8.5
7.4
6.0
7.7
20.1
8.0
10.9
coelute
7.9
13.4
29.9
13.1
14.0
Relative
Std dev
10.9
12.2
8.5
7.3
16.4
5.8
7.3
10.1
8.2
6.8
9.6
25.9
8.9
11.1
coelute
8.3
14.1
28.6
14.0
19.3
n=10
results are average of 5 bonding lots
n=10
500
-------
TABLE 2
Extraction of CLP QTM Pesticides
Effect of pH
High Calibration Standard Recovery
Analytes
250 ng/ 100 mL
a-BHC
P-BHC
•y-BHC
6-BHC
Heptachlor
Aldrin
Heptachlor epoxide
•y-Chlordane
Endosulfan I
a-Chlordane
Dieldrin
p,p' - DDE
Endrin
Endosulfan II
p,p' - ODD
Endrin aldehyde
Endosulfan sulfate
p,p'-DDT
Endrin ketone
Methoxychlor
Decachlorobiphenyl
(SMC)
Percent recovery
pH=2
74.3
105.4
75.9
79.0
75.6
86.4
81.6
86.0
70.4
82.6
81.8
88.6
68.6
76.5
105.2
74.4
104.9
106.2
120.3
99.9
100.5
PH=5
77.0
100.8
78.7
82.0
77.4
83.7
80.6
82.2
70.1
78.7
80.1
84.4
48.0
74.9
81.8
90.3
95.9
93.8
135.3
93.4
90.6
pH=7
82.4
101.8
82.9
85.8
85.4
91.1
87.8
89.3
76.5
86.0
88.3
92.2
51.5
80.9
95.9
94.6
99.4
104.9
150.1
104.8
96.0
501
-------
TABLE 3
Extraction of CLP QTM Pesticides
Effect of Drying
High Calibration Standard Recovery
Analytes
250 ng/ 100 mL
a-BHC
P-BHC
•y-BHC
8-BHC
Heptachlor
Aldrin
Heptachlor epoxide
'y-Chlordane
Endosulfan I
a-Chlordane
Dieldrin
p,p' - DDE
Endrin
Endosulfan II
p,p' - ODD
Endrin aldehyde
Endosulfan sulfate
p,p' - DDT
Endrin ketone
Methoxychlor
Decachlorobiphenyl
(SMC)
Percent recovery
dry to
ImL
100.3
122.1
103.5
98.2
103.7
103.7
105.0
106.6
106.7
105.4
108.2
107.4
114.6
108.3
114.1
105.8
113.5
116.4
106.9
116.8
116.0
complete
dry
73.4
103.6
79.9
84.7
69.6
67.0
74.9
72.3
72.8
72.6
75.9
71.2
79.1
84.7
123.1
coelute
93.9
76.0
83.5
85.9
97.7
average % difference
percent
difference
27
15
23
14
33
35
29
32
32
31
30
34
31
22
-16
coelute
17
35
22
26
16
26
502
-------
TABLE 4
Extraction of CLP QTM Pesticides
Effect of Sample Matrix
Analytes
100 ng/ 100 mL
a-BHC
p-BHC
-y-BHC
8-BHC
Heptachlor
Aldrin
Heptachlor epoxide
^-Chlordane
Endosulfan I
oc-Chlordane
Dieldrin
p,p' - DDE
Endrin
Endosulfan II
p,p' - ODD
Endrin aldehyde
Endosulfan sulfate
p,p'-DDT
Endrin ketone
Methoxychlor
Decachlorobiphenyl
(SMC)
Percent recovery
deionized
water
92.4
57.0
90.3
90.3
89.7
87.3
94.7
94.2
92.7
95.3
90.0
94.1
69.0
93.0
88.9
93.8
90.3
87.4
97.2
83.9
72.1
waste
water
74.9
48.5
107.4
111.4
77.5
78.2
85.7
171.2
94.0
91.6
94.5
94.7
89.5
90.2
97.0
86.5
91.1
90.7
98.2
98.2
117.7
hazardous
waste
80.9
79.9
332.3
120.1
91.5
192.7
90.9
116.1
99.5
135.8
127.7
86.2
136.8
110.1
115.7
75.9
94.9
96.7
112.6
104.4
123.4
n=3
n=3
n=3
503
-------
LITERATURE CITED
[1] Eichelberger, J.W.; Munch, J.W.; Shoemaker, J.A. Determination of Organic
Compounds in Drinking Water by Liquid-Solid Extraction and Capillary Column Gas
Chromatography/MassSpectrometry, Revision 1.0, February 1994
[2] Prapamontol, T.; Stevenson, D. Rapid Method for the Determination of
Organochlorine Pesticides in Milk, Journal of Chromatography, 552, pp 249-257 (1991).
[3] Bagnati, R.; Benfenati, E.; Davoli, E.; Fanelli, R. Screening of 21 Pesticides in Water
by Single Extraction with CIS Silica Bonded Phase Columns andHRGC-MS,
Chemosphere, 77, No. 1, pp 59-65 (1988).
[4] Miyahara, M; Suzuki, T.; Saito, Y. Multiresidue Method for Some Pesticides in
Lanolin by Capillary Gas Chromatography with Detection by Electron Capture, Flame
Photometric, Mass Spectrometric, and Atomic Emission Techniques, Journal of
Agriculture and Food Chemistry, 40, pp 64-69 (1992).
504
-------
79
CONTAMINANT DEGRADATION STUDY
AT THE HANFORD SITE 1100-EM-l OPERABLE UNIT
R.A. Bechtold, Principal Scientist, Westinghouse Hanford Company, Richland,
Washington 99352 : K. M. Angelos. Senior Environmental Scientist, Colder
Associates Inc., Redmond, Washington.
ABSTRACT
A study of volatile organic contaminant degradation conducted by Westinghouse
Hanford Company (WHC) on environmental groundwater samples is presented. WHC
is the U.S. Department of Energy's (DOE) Operations and Engineering Contractor for
the Hanford Site. A contaminant degradation study was conducted from December
1990 through December 1991 to evaluate the degradation of trichloroethene (TCE) in
ground water in the vicinity of Horn Rapids Landfill. The Horn Rapids Landfill is part
of the 1100-EM-l Operable Unit which is one of four operable units within the 1100
Area of the Hanford Site. The 1100 Area was placed on the National Priorities List
(NPL) in July 1989. This paper describes work which was part of the Remedial
Investigation Phase 2 Supplemental Work Plan for the 1100-EM-l Operable Unit
(DOE-RL 1990a). The objective of this study was to evaluate the degradation of TCE
present in the ground water in the vicinity of the Horn Rapids Landfill. Contaminant
degradation was studied in ground water from monitoring wells known to have
detectable TCE.
The 1100 Area is the central warehousing, vehicle maintenance, and
trasnportation operations center for the Hanford Site. This area was designated an NPL
site in July, 1989, and is divided into four operable units. The first equipment
maintenance operable unit, 1100-EM-l, was assigned the highest RI/FS priority within
both the 1100 Area and the Hanford Site as a whole. The Horn Rapids Landfill is a
solid waste facility used primarily for the disposal of office and construction waste and
the burning of classified documents; asbestos, sewage sludge, fly ash, and potentially,
drums of unidentified organic liquids (presumably carbon tetrachloride) were also
disposed at this site.
Samples of ground water were collected from three groundwater monitoring
wells (MW-11, MW-12, and MW-15) located in the vicinity of Horn Rapids Landfill.
The samples were collected by WHC personnel into 40 mL volatile organic analysis
(VOA) bottles in sufficient quantity to allow for repeated analysis of the samples at
frequencies of 1, 2, 4, 8 and 12 months from the date of collection. The VOA bottles
were not preserved with mineral acid as is customary for volatile organic samples so as
not to inihibit degradation of the organic compounds. The samples were stored in the
dark at room temperature (approximately 25°C) throughout the duration of the study.
505
-------
Sample analysis was conducted in accordance with the U.S. EPA Contract Laboratory
Program, Statement of Work for Organics Analyses, (EPA 1988a) with the exception
that the laboratory reported the following compounds to a lower quantitation limit as
opposed to the usual quantitation limit of 5 or 10 (J.g/L: vinyl chloride, 1,1-
dichloroethene, 1,2-dichloroethane, and trichloroethene.
Results of the study showed that TCE was degraded by as much as 50% over
the life of the study. However, within the first three months of the study, contaminant
concentrations remained stable and little or no degradation was observed. Further work
is being conducted on comparison of degradation rates to sample holding times in order
to develop rationale for extending sample holding times.
INTRODUCTION
This report presents the results of a study conducted from December 1990
through December 1991 to evaluate the degradation of TCE in the ground water in the
vicinity of Horn Rapids Landfill. The Horn Rapids Landfill is part of the 1100-EM-l
Operable Unit which is one of four operable units within the 1100 Area of the Hanford
Site. The 1100 Area was placed on the National Priorities List (NPL) in July 1989.
This report is part of the work described in the draft Remedial Investigation Phase 2
Supplemental Work Plan for the 1100-EM-l Operable Unit (DOE-RL, 1990a).
The objective of the study was to evaluate the degradation of TCE present in the
ground water in the vicinity of the Horn Rapids Landfill. Contaminant degradation
was studied in ground water from monitoring wells known to have detectable TCE in
low, medium and high concentrations.
STUDY AREA BACKGROUND AND PHYSICAL SETTING
The 1100 Area is the central warehousing, vehicle maintenance, and
transportation operations center for the Hanford Site. This area was designated an NPL
site in July, 1989, and is divided into four operable units. The first equipment
maintenance operable unit, 1100-EM-l, was assigned the highest RI/FS priority within
both the 1100 Area and the Hanford Site as a whole.
A detailed description of the regional and physical characteristics of the
operable unit may be found in the Phase I Remedial Investigation Report (DOE-RL,
1990b). The Horn Rapids Landfill is a solid waste facility used primarily for the
disposal of office and construction waste and the burning of classified documents;
asbestos, sewage sludge, fly ash, and potentially, drums of unidentified organic liquids
(presumably carbon tetrachloride) were also disposed at this site. Figure 1 shows the
location of the 1100-EM-l Operable Unit and Horn Rapids Landfill.
506
-------
SAMPLING AND ANALYSIS
Samples of ground water were collected from three groundwater monitoring
wells (MW-11, MW-12, and MW-15) located in the vicinity of Horn Rapids Landfill.
The well locations are identified in Figure 2. The samples were collected by WHC
personnel into 40 mL volatile organic analysis (VGA) bottles in sufficient quantity to
allow for repeated analysis of the samples at frequencies of 1, 2, 4, 8 and 12 months
from the date of collection. The VOA bottles were not preserved with mineral acid as
is customary for volatile organic samples so as not to inihibit degradation of the organic
compounds. The samples were stored in the dark at room temperature (approximately
25°C) throughout the duration of the study.
Following sample collection, the samples were cooled and placed into a
shipping container for shipment under chain-of-custody to the analytical laboratory.
The laboratory used for analysis of the samples was Pacific Northwest Environmental
Laboratories Inc. (PNELI) of Redmond, Washington.
Sample analysis was conducted in accordance with the U.S. EPA Contract
Laboratory Program, Statement of Work for Organics Analyses, (EPA 1988a) with the
exception that the laboratory reported the following compounds to a lower quantitation
limit as opposed to the usual quantitation limit of 5 or 10 ng/L: vinyl chloride, 1,1-
dichloroethene, 1,2-dichloroethane, and trichloroethene.
RESULTS OF THE STUDY
This section presents the results of the contaminant degradation study for each
well sampled. Results for each well were compared beginning with the fourth round of
1100-EM-l groundwater monitoring data (December 5, 1990) and completing with the
12th month round of analysis data (December 4, 1991). Table 1 presents a summary of
the compounds detected and the resulting contaminant concentrations for each of the
wells sampled.
Monitoring Well MW-11
For the six sets of analysis data, 1,1,1-trichloroethane (TCA) was detected in
four of the analyses. During the December 28, 1990 analysis and the December 4,
1991 analysis, TCA was not detected at a quantitation limit of 5 jig/L. Trichloroethene
(TCE) was detected during all six analyses beginning at a concentration of 3 |ig/L and
then dropping to a concentration of 2 ng/L, where it remained through the completion
of the study (see Figure 3). The degradation products, vinyl chloride and
dichloroethylene were not detected in any of the samples.
507
-------
Monitoring Well MW-12
TCA was detected at concentrations of 2 and 1 |^g/L during the first five sets of
analyses, and during the final set of analysis, was not detected at a quantitation limit of
5 |ig/L. TCE was detected during all six analyses, beginning at a concentration of
74pig/L and dropping to a concentration of 36 ng/L during the final analysis set, a
reduction of 51 % (see Figure 4). The degradation products vinyl chloride and
dichloroethylene were not detected in any of the samples.
Monitoring Well MW-15
TCA was detected in all but one analysis at concentrations of 1 p.g/L. TCE was
detected at its highest concentration at 61 ng/L during the 1/28/91 analysis and at the
conclusion of the test, was detected at a concentration of 45 ng/L, a reduction of 26%
(see Figure 5). The degradation products vinyl chloride and dichloroethylene were not
detected in any of the samples.
Discussion
TCE may be degraded in the environment by a variety of processes including
hydrolysis, oxidation, reduction, dehydrohalogenation, volatilization, and
biodegradation, however the major process effecting degradation is considered to be
biodegradation (Olsen, R.L., et. al., 1990). Biodegradation of TCE generally results
in the formation of the halogenated by-products, dichloroethene and vinyl chloride
neither of which were detected during this study. In addition, neither dichloroethene
nor vinyl chloride have been been detected during routine groundwater monitoring of
the wells sampled for this study (DOE, 1990b).
The results of analysis on wells MW-12 and MW-15 indicate that TCE
concentrations are reduced as much as 51 % over the course of a year under the test
conditions established for this study, yet no degradation products were detected.
Possible explanations or refutations for the reduced TCE concentrations are
summarized as follows:
• Hydrolysis may be a likely reaction pathway since this would result in
the formation of intermediate degradation products (alcohols and free
chlorine) which would not be detected using the analytical procedure
used in this study. The half-life of TCE undergoing hydrolysis
degradation is estimated at approximately 11 months which is similar to
the half-life of 12 months observed for TCE in the samples analyzed for
well MW-12
508
-------
• TCE may be undergoing oxidation, however for this process to occur
oxidizing agents such a peroxides, ozone or chlorine are usually
necessary to promote the reaction and these compounds do not occur
naturally in groundwater. Because of this oxidation would not be
considered the likely process contributing to reduction of the TCE,
• Reduction reactions acting on the TCE present would most likely result
in the formation of lesser chlorinated compounds such as dichloroethene
which was not detected during this study,
• Biodegradation may be occuring however the common anaerobic
reaction products would be dichloroethylene and vinyl chloride which
were not detected during this study. In addition, biodegradation half-
lifes for TCE are estimated to range from 33 to 230 days (Olsen, R.L.,
et. al. 1990). A TCE half-life of 365 days was observed for MW-12
during this study,
• Dehydrohalogenation is a possible reaction pathway, however only the
chlorinated ethane compounds (1,1,1-TCA) are susceptible to this
process since it requires the removal of chlorine and hydrogen atoms
from the molecule and the formation of an ethene compound
(dichloroethene, DCE). Since DCE was not detected in this study this
reaction process is not likely contributing to the reduction of TCE, and
• Volatilization of TCE through the septum seal in the sample containers
may be causing the reduction in concentrations, however, laboratory
studies on the stability of TCE under similar storage conditions indicate
that losses due to volatilization through the container septa are
insignificant (Maskarinec, M.P, et. al., 1990).
SUMMARY AND CONCLUSIONS
A contaminant degradation study was conducted at the 1100-EM-l Operable
Unit beginning in December 1990 and completed in December 1991. Three
groundwater wells were sampled for volatile organic compounds and the samples were
stored unpreserved, in the dark at room temperature and analyzed over the course of a
year to determine if degradation of TCE is occurring by analysis for it's degradation
products, dichloroethylene and vinyl chloride. Results of the study are summarized
below:
509
-------
In groundwater monitoring well MW-11, TCE was detected initially at a
concentration of 3 (ig/L and at the completion of the study, at a
concentration of 2
• In groundwater monitoring well MW-12, TCE was detected initially at a
concentration of 74 fig/L and at the completion of the study, at a
concentration of 36 [ig/L, a reduction of 51 %.
• In groundwater monitoring well MW-15, TCE was detected at a
concentration of 61 (ig/L and at the completion of the study, at a
concentration of 45 ng/L, a reduction of 26%.
• The degradation products, dichloroethylene and vinyl chloride were not
detected in any of the samples analyzed as part of this study,
• The half-life of TCE during this study was 12 months as indicated by the
results of analysis on well MW-12, (see Section 4.2).
Results of the study indicate that the half-life of TCE present in the groundwater
and analyzed under the storage conditions imposed by this study is approximately one-
year as indicated by the results of the analysis for the samples collected from well MW-
12. Rapid degradation of the TCE does not appear to be occuring based on the absence
of it's degradation products. The reduction in TCE that was observed in this study is
likely due to natural hydrolysis that would result in the formation of alcohols and free
chlorine which would not be detected by the analytical procedure employed by this
procedure.
REFERENCES
Bleyler, 1988, Laboratory Data Validation Functional Guidelines for Evaluating
Organics Anaylsis, U.S. Environmental Protection Agency, Washington, D.C.
DOE-RL, 1990a, Remedial Investigation Phase 2 Supplemental Work Plan for the
Hanford Site 1100-EM-l Operable Unit, U.S. Department of Energy, Richland
Operations Office, Richland, Washington.
DOE-RL 1990b, Phase I Remedial Investigation Report for the Hanford Site 1100-EM-
1 Operable Unit, U.S. Department of Energy, Richland Operations Office, Richland,
Washington.
510
-------
Maskarinec, M.P.; Johnson, L.H.; Holladay, S.K.; Moody, R.L; Bayne, C.K.;
Stability of Volatile Organic Compounds in Environmental Water Samples during
Transport and Storage, Environ. Sci. Technol., 1990, Vol. 24, 1665-1670.
Olsen, R.L.; Davis, A; Predicting the Fate and Transport of Organic Compounds in
Groundwater, Hazardous Materials Consultant, May/June 1990, pp. 39-64;
July/August 1990, pp. 18-37.
511
-------
«—™i ',y
**' I
"" ""
;1JN-1100-5
.* ;1100-4
* —..h 100-1
:'£^; ,,,-'|Ephemeral Pbol
*<..£*"""!; \ fc?
Base map adapted from USGS 1978
1100-EM-1 Operable Subunit
location and designation
1000 Meters
3000 Feet
1100-EM-1 Operable Unit
903-1221/24647/3-25-92
Figure 1. 1100-EM-l Operable Unit
512
-------
r ^
Base map adapted from USGS
Existing Monitoring Well Location
Soil Gas Testing Location
2/2/91 9031221\31716
Figure 2. Location of Groundwater Monitoring Wells
-------
en
3.2
3
2.8
2.6
2.4
2.2
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
D
-
-
-
-
- D D D D D
-
+ + + +
I I i I l I I i I | I
11/23/90 03/03/91 06/11/91 09/19/91 12/28/91
D TCE,jig/L + l,l,l-TCA,jig/L
903 1221/28104/3-25-92
Figure 3. Degradation Study Results, Groundwater Monitoring Well MW-11
-------
80
O\f
70
60
50
40
30
20
10
o
a
a
D O
-
a
-
_
D
-
-
-
, + 4- ,+ ,-*-, , +, , , . ,
11/23/90 03/03/91 06/11/91 09/19/91 12/28/91
D TCE,jig/L + 1,1,1-TCA.ng/L
903 1221/28105/3-25-92
Figure 4. Degradation Study Results, Groundwater Monitoring Well MW-12
-------
en
o>
70
60
50
40
30
20
10
0
D
D D
a
-
a
a
1+,1-f 1-1-1 i +i i i + i
11/23/90 03/03/91 06/11/91 09/19/91 12/28/91
D TCE,jig/L + 1,1,1-TCA, jig/L
903 1221/28106/3-25-92
Figure 5. Degradation Study Results, Groundwater Monitoring Well MW-15
-------
Table 1. TCE Degradation Study Results
WELL:
DATE SAMPLED:
DATE ANALYZED:
1,1,1-Trichloroethane
Trichloroethene
WELL:
DATE SAMPLED:
DATE ANALYZED:
1,1,1-Trichloroethane
Trichloroethene
WELL:
DATE SAMPLED:
DATE ANALYZED:
1,1,1-Trichloroethane
Trichloroethene
MW-12
1V26/90
12/05/90
2J
74
12/28/90
U
68
1/28/91
2J
68
3/25/91
2J
70
7/23/91
2J
55
12/4/91
5U
36
MW-15
11/27/90
12/05/90
1J
59
12/28/90
5U
58
1/28/91
U
61
3/25/91
U
54
7/23/91
U
43
1Z/4/91
U
45
MW-11
11/25/90
12/05/90
1J
3
12/28/90
5U
2
V28/91
1J
2
3/25/91
U
2
7/23/91
1J
2
12/4/91
5U
2
Ul
U - Indicates the compound was analyzed for but not detected. The value reported is the sample quantitation limit
J - Indicates the concentration reported is less than the contract required quantitation limit (CRQL) but greater than the
instrument detection limit.
-------
80
THE USE OF AUTOMATED SUPERCRITICAL FLUID EXTRACTION/GC-MS FOR
THE QUANTITATIVE DETERMINATION OF PAHS IN SOIL
ABSTRACT
Dr. Joseph M. Lew, Vice President of Technology and Customer Support; Lori A.
Dolata, Technical Specialist; Robert M. Ravey, Application Chemist: and Anita
Cardamone, Application Chemist, Suprex Corporation, 125 William Pitt Way,
Pittsburgh, PA 15238
Supercritical fluid extraction (SFE) has a broad range of applicability, especially with
regards to environmental problems. SFE has achieved a significant amount of attention
due to the benefits of eliminating toxic, liquid solvent usage, reduction in sample
preparation time and an increase in the overall analytical reliability of determinations.
SFE/GC-MS is a powerful technique to accurately analyze and quantitate environmental
analytes. The off-line transfer of SFE effluents to collection vials adds a considerable
amount of flexibility in characterizing complex matrices since a full complement of
analytical tools can be used (ie, GC, LC, IR, NMR and UV). Moreover, the advantages
of SFE can be further augmented by the development of automation for greater sample
throughput which can be especially useful for the volume of samples associated with
environmental applications.
This paper will discuss the use of automated SFE methodologies for the determination
polynuclear aromatic hydrocarbons (PAH) in soil. Details of method development
strategies will be presented. This discussion will focus on the experimental verification
of optimized SFE extraction and collection variables to achieve efficient and quantitative
extractions of the target analytes in the soil with sequential replicates.
INTRODUCTION
Currently, polynuclear aromatic hydrocarbons (PAH) are extracted from environmental
matrices (e.g. soils, sediments) using Soxhlet extraction or sonication sample preparation
techniques. Over the past few years, supercritical fluid extraction (SFE) has surfaced as
an alternative sample preparation technique offering the distinct advantages of greatly
reduced sample preparation times, comparable extraction efficiencies, nontoxic (CO2)
solvent use, and an increase in the overall reliability of the analytical technique (1,2).
As SFE matures as a sample preparation technique, optimized methods will be developed
for specific sample matrices and target analytes. In method development strategies,
various stages of the entire SFE process need to be investigated. This includes three
basic stages: pre-extraction strategies, extraction strategies and collection strategies. In
pre-extraction strategies, various sample manipulation techniques such as grinding,
freeze-milling or adsorbent addition can be employed to make the sample matrix more
appropriate (more surface area or immobilized water) for SFE. Extraction strategies
involve optimizing pressures, temperatures, sample size, durations, static/dynamic modes
and modifier type and concentration. Collection strategies include off-line or on-line
modes, restrictor flow rates, collection temperatures, desorption temperatures, adsorbent
518
-------
type, and type of wash solvent. Moreover, these strategies need to be optimized for
heterogeneous sample matrices to achieve the ultimate goal of the highest possible
efficiency in the shortest time with high precision.
EXPERIMENTAL
A stand-alone manual supercritical fluid extraction system (PrepMaster®, Suprex;
Pittsburgh, PA), which has been described earlier (2,3), and a sequentially automated
supercritical fluid extraction system (AutoPrep-44™, Suprex; Pittsburgh, PA) were used
for all of the extraction experiments. Extractions were accomplished with various sizes
of extraction vessels (0.5 to 10 mL) using operating conditions that are listed in the text.
An off-line collection module, (AccuTrap®, Suprex) was used to perform the cryogenic
solid phase trap collection (4-6) for both the manual and automated systems. This
collection module included the following components: a VariFlow™ (Suprex) automati-
cally variable restrictor (kept at 50°C) which controlled the CO2 flow rates (from 1 to 7
mL/minute compressed), a temperature-controlled solid phase trap packed with unibeads
and CIS modified silica for analyte trapping, a liquid pump for delivering an appropriate
wash solvent for analyte desorption, and a solenoid valve for delivering a stream of
nitrogen to purge the adsorbent trap and connecting tubing after desorpton. For this
work 1.5 mL of methylene chloride with a flow rate of 2 mL/minute was sufficient to
quantitatively wash the adsorbed PAH out from the trap into a GC autosampler vial.
After off-line collection, all of the extracted analytes were analyzed using a AutoSystem
gas chromatograph (Perkin Elmer, Norwalk, CT) and a Q-Mass benchtop mass
spectrometer (Perkin Elmer). A Varian 3400 gas chromatograph (Varian; Sunnyvale,
CA) equipped with a capillary split/splitless injection port and flame ionization detection
(FID) was used for GC/FID characterization. All of the soil and sediment sample
matrices were naturally incurred and characterized by Soxhlet or sonication methods
before GC/MS analysis according to Environmental Protection Agency's (EPA) method
8270. The GC conditions for all of the PAH determinations were 60°C (2 min.)
programmed to 310°C at 7°C/min. utilizing a 25 m x 0.2 mm i.d. methyl silicone (DB-1)
column (J&W Scientific; Folsom, CA).
RESULTS AND DISCUSSIONS
One of the first SFE operational parameters that was investigated for PAH extraction was
extraction pressure. Conventionally, most researchers tend to refer to this parameter first
when developing a SFE method because of its relation to controlling the solubilizing
characteristics of supercritical CO2- For a PAH contaminated soil, the SFE conditions
were 75°C, 40 minutes (5 static/35 dynamic), with a compressed flow of 2.5 ml/minute
for a 600 mg sample. The effluent was collected in a solid phase trap (C18/unibeads) at
-30°C. Table 1 (7) shows the GC/MS PAH results for this soil at pressures of 250 atm,
350 atm and 450 atm (keeping all other extraction and collection strategies constant).
The concentration levels were determined using calibrated external standards and were
compared to the acceptance range for each PAH as outlined by EPA method 8270. The
highest pressure of 450 atm provided the best agreement with the EPA acceptance range
for all of the PAH from two to five fused aromatic rings. The lower molecular weight
PAHs (ie, two and three fused aromatic rings) were in fact within the acceptance range
519
-------
for all three of the extraction pressures. To fall within the acceptance ranges, the higher
molecular weight PAH (ie, four and five fused aromatic rings) required higher solubiliz-
ing powers. This particular soil sample was contaminated with PAHs at higher levels,
which were not apparently tightly associated with the matrix surface. Using the
optimized extraction pressure, the levels of precision were determined for the various
PAH for 42 replicate runs which are listed in Table 2 using the automated SFE system.
The relative standard deviation (RSD) varied from 3.6 to 7.6% for 2 to 4 fused ring
PAH. It is also important to note that the determined RSDs represent total system preci-
sion including the SFE sample preparation, the off-line cryogenic solid phase PAH
trapping, and the GC/FID analytical determination. No additional sample cleanup or
work-up was required after SFE and before the GC injection.
Table 1: Off-Line SFE/GC-MS of PAH Contaminated Soil - Effect of Extraction
Pressure (7).
EPA Method 8270
Acceptance
Compound
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)Anthracene
Chrysene
Benzo(b,k)Fluoranthene
Benzo(a)Pyrene
Range
(PPM)
24.2-40.6
14.7-23.5
527-737
414-570
1270-1966
373-471
1060-1500
744-1322
214-290
271-323
130-174
80.1-114.3
Concentration Levels (PPM)
250 atm
23
20
566
445
1682
357
1028
703
74
74
<1.0
<1.0
350 atm
23
*
601
471
1978
439
1459
1153
235
251
107
64
450 atm
25
22
614
458
1911
400
1571
1269
284
314
155
89
Table 2: Automated SFE/GC-FID
Determination of PAH in Soil
Concentration (PPM)
Analyte
Acenaphthene
Anthracene
Pyrene
Benzo(a)Anthracene
Chrysene
Based upon total sample
Run
#1
628
379
858
239
422
runs
Run
#12
631
340
856
256
434
Run
#21
636
345
872
249
431
of 42 replicates
Run
#26
641
422
864
246
444
Run
#31
738
393
870
265
437
Run
#40
726
327
866
250
437
and Soxhlet/GC-MS
Run
#42 Mean %
629 633
404 373
818
237
436
target
858
240
434
values.
RSD
5.8
7.6
4.6
3.6
4.9
520
-------
CONCLUSION
SFE technology has undergone a significant evolution over the past two years, particu-
larly in the areas of restrictor technology, solid phase trapping, wrench-free extraction
vessel design and sequential extraction vessel feeding. These advances have allowed
easier and more reliable analytical use of SFE as a sample preparation tool for routine
environmentally related applications.
REFERENCES
1. Hawthorne, S.B., Anal. Chem. 62: 633A-42A (1990).
2. Levy, J.M., Rosselli, A.C., Storozynsky E., Ravey, R.M., Dolata, L.A., Ashraf-
Khorassani, M., LC-GC Magazine, 10: No. 5 386-91 (1992).
3. Levy,J.M., Am. Lab. 8:25-32(1991).
4. Ashraf-Khorassani, M., Houck, R.M., Levy, J.M., J. Chromatogr. Sci. 30(9): 361-66
(1992).
5. Levy, J.M., Ravey, R.M., Houck, R.K., Ashraf-Khorassani, M., Fresenius Z of Anal.
Chem. 344:517-20(1992).
6 Levy, J.M., Houck, R.K., Am. Lab, 5: 36R-36Y (1993).
7. Levy, J.M., Dolata, L.A., Ravey, R.M., J. Chromatogr. Sci. 31: 349-52 (1993).
521
-------
81
ROUND-ROBIN STUDY OF PERFORMANCE EVALUATION MATERIALS
FOR THE ANALYSIS OF VOLATILE ORGANIC COMPOUNDS
IN SOIL: PRELIMINARY ASSESSMENT
Alan D. Hewitt and Thomas F. Jenkins, U.S. Army Cold Regions Research and Engineering Laboratory,
Hanover, New Hampshire 03755-1290; Clarence L. Grant, University of New Hampshire, Durham, New
Hampshire 03824; and Martin H. Stutz, U.S. Army Environmental Center, Aberdeen Proving Ground,
Maryland 21010
ABSTRACT
A round-robin study of the analysis of soil subsamples vapor-fortified with volatile organic compounds
(VOCs) was recently performed by twelve laboratories. Vapor fortification has been proposed as a method
of spiking soils with VOCs so that they can be used as performance evaluation materials. Each laboratory
was sent two sets of three different vapor-fortified soil subsamples containing trans- 1,2-dichloroethylene
(TDCE), trichloroethylene (TCE), benzene (Ben), and toluene (Tol). Analyte concentration estimates for
these vapor-fortified soils were obtained using SW846 Method 8240. Preliminary analysis of the results
showed a range of relative standard deviations from 9 to 22%, with an average of less than 15%. These
results confirm that vapor-fortification treatment, followed by confinement in sealed glass ampoules, is a
precise means of preparing and storing VOC-contaminated soil subsamples that can be used in quality
assurance programs.
INTRODUCTION
The wide use and subsequent improper disposal of petroleum products and chlorinated solvents has made
volatile organic compounds (VOCs) one of our most ubiquitous environmental hazardous waste problems
(1). Despite the large number of vadose zone soil samples routinely characterized for VOCs, no per-
formance evaluation materials now exist for this matrix (2). Currently the accuracy of soil VOC analyses
relies on solution spike and recovery tests. One common practice is to introduce dilute methanol (MeOH)
solutions of the analytes of interest into samples after they have been placed in the purge chamber of a
purge-and-trap system. This method is of limited utility because (a) it evaluates only the determinative step,
(b) it allows no time for natural sorptive processes to occur, (c) it involves a carrier solvent, thereby
affecting sorptive interactions, and (d) it does not simulate the manner in which soils become contaminated
in the field.
The accuracy of laboratory estimates of analyte concentrations in environmental samples is initially
dependent on analytical calibration. Thereafter, accuracy is monitored during routine analysis of real
samples by reference to results on accompanying quality assurance (QA) and quality control (QC) samples.
For this system to work effectively, QA and QC reference samples must be available in a stable form that
effectively mimics real samples and with accurately known concentrations. For VOCs in soils, preparation
of such reference materials is very difficult (3,4).
Vapor equilibration offers a means of fortifying soils that overcomes many of the shortfalls of previous
methods (5). This method of soil spiking simulates one process by which vadose zone soils become con-
taminated. It involves an extended exposure period, it is precise both within and among treatment batches,
and it has shown analyte concentration stability for greater than a 60-day holding period for several soils.
Moreover, vapor-fortified soils can be used to evaluate both extraction efficiency and determinative
accuracy (6, 7).
522
-------
This treatment method exposes desiccated soils to the vapors of chosen analytes. During exposure, in-
dividual soil subsamples are contained in unsealed 1-mL glass ampoules inside a closed desiccator. After
treatment, the ampoules are removed from the desiccator and quickly heat-sealed to prevent volatilization
losses, which have often compromised previous methods. When used, the sample is analyzed after breaking
the ampoule in a closed vessel containing the extraction solvent (i.e., MeOH, water). Performance evalua-
tion materials for the analysis of VOCs in soil, prepared by vapor fortification and sealed in glass
ampoules, offer a method by which soil subsamples may be precisely fabricated, distributed, prepared and
analyzed without analyte volatilization losses (8, 9).
In this study, 12 laboratories participated in a collaborative evaluation of this vapor-fortification procedure.
Duplicate subsamples of three different soils were vapor-fortified with trans-1,2-dichloroethylene (TDCE),
trichloroethylene (TCE), benzene (Ben), and toluene (Tol). Each laboratory was directed to extract the six
soil subsamples and analyze them within 30 days of receipt using the EPA's SW846, Method 8240 for
high-level (> 1 Mg/g) VOCs in soil (10). The twelve laboratories that participated in the study are listed in
Appendix A. The design and a preliminary assessment of the results of this round-robin study will be pre-
sented.
PREPARATION AND DISTRIBUTION OF TEST MATERIALS
Three soil matrices were used in this study. Two of these soils serve as reference matrices for the U.S.
Army Environmental Center: a marine sediment from Tampa Bay, Florida (TB), and a composite soil from
the Rocky Mountain Arsenal (RMA) in Denver, Colorado. The third soil was from Point Barrow, Alaska
(PBA). Clay and organic carbon concentration for these soils are shown in Table 1.
These soils were air-dried, sieved through a 30-mesh screen, and mixed thoroughly before subsampling.
Forty subsamples of each soil type were placed into 1.0-mL glass ampoules using a stainless steel spatula
and plastic funnel. PBA subsamples were weighed to 1.50 ± 0.01 g, while the RMA soil and TB sediment
subsamples were weighed to 2.00 ± 0.01 g.
Vapor fortification was performed after placing twenty ampoules filled with one of the three soil types into
a 5.6-L desiccator with a dish of anhydrous CaSO4 for two days. Following desiccation, the CaSO4 was re-
placed with an open 60-mL glass bottle containing 50 mL of a solution prepared by combining 25 mL of a
MeOH stock solution (0.60 g Tol, 0.59 g TCE, 0.50 g TDCE, and 0.35 g Ben diluted to 100 mL) with 25
mL of tetraethylene glycol dimethyl ether (tetraglyme). After seven days of vapor fortification, a desiccator
was opened and a 5-mm-diameter glass bead was rapidly placed on top of each ampoule, forming a tem-
porary cap. Then, as quickly as possible, each ampoule was put in a metal tension clamp and the neck was
heat-sealed using a propane plumber's torch. A sharp-pointed tip was created when sealing the ampoules to
facilitate ease of breaking when preparing for analysis. Details on the development of this procedure and its
performance have been documented elsewhere (5-9).
Table 1. Characteristics of soils.
Characteristic TB RMA PBA
% Organic Carbon 0.31 0.05 6.7
%Clay <5 NA 20.1
523
-------
The sequence in which the individual ampoules were sealed after removal from the dessicator was re-
corded. Sixteen sample sets were prepared by randomly selecting a single fortified soil subsample from
each of the six treatment batches. This design allows us to examine batch-to-batch precision, as duplicate
samples were prepared in different dessicators. The design also permits evaluation of intra- and inter-
laboratory analysis precision.
Each sample package contained:
• Eight VOA vials (one extra).
• Seven sealed glass ampoules containing soil [three pairs (1 & 2) of vapor-fortified soils labeled A,
B, and C, and one blank soil labeled D for practicing the ampoule-breaking operation].
• A QA ampoule (VOA-2, Ultra Scientific) with MeOH as a solvent and containing certified VOC
analyte concentrations.
• Handling and analysis instructions and reporting forms.
SOIL SUBSAMPLES AND QA SAMPLE HANDLING AND ANALYSIS
It was imperative that the procedures used by each laboratory be as consistent as possible to minimize
systematic errors. For this reason, very specific handling and analysis instructions were provided. Prior to
analysis, the laboratories were requested to calibrate their instrument for the analysis of TDCE, TCE, Ben,
and Tol over a range of at least 1 to 100 Hg/L. Analysis of both the QA standard and soil subsamples were
to be performed within a single day of operation, and within one month of sample receipt. Between receipt
and analysis, the entire sample set was to be refrigerated (4°C). All analyses were performed by PT/GG/MS
following the general guidelines of SW846, Me > 1 |0g/g VOC. The soil subsamples were first dispersed in
MeOH, then a portion of the methanolic solution was combined with =5 mL of water for PT/GC/MS analy-
sis.
On the day of analysis, the analyst was instructed to make a 1/20 dilution of the QA standard provided
using the same MeOH that was to be used for extracting the soil subsamples. A 100-(JL aliquot of the 1/20
dilution was added to 4.90 mL of water in the purge chamber of a PT/GC/MS. The results of this analysis
should be within ± 20% of the certified concentrations for TCE, Ben, and Tol. If not, the analyst was
requested to reanalyze the QA standard and/or consider recalibration prior to continuing.
To prepare a vapor-fortified soils for analysis, 20.0 mL of MeOH was added to one of the supplied VOA
vials. An inverted (tip pointing toward the bottom) ampoule was placed into the solution and the vial quick-
ly capped. The ampoule tip was then broken by shaking and the soil dispersed for approximately two
minutes by hand agitation. A practice soil subsample ("D") was included to allow the analyst a chance to
determine how hard to shake the VOA vial when breaking the ampoule. We recommended that the analyst
wear rubber gloves and that the strength of shaking be increased slowly, and not increased further once the
ampoule was broken. Our experience has shown that only a portion of the tip has to be broken in order for
the soil to be released for extraction. However, if, after two minutes of shaking, soil remains trapped in the
ampoule, shaking should be continued until complete dispersion results. Transfer of the methanolic extracts
was performed after the soil had settled (approximately 20 minutes) by opening the VOA vial and with-
drawing the specified aliquot with the appropriate syringe. Subsample pairs 1 and 2 were to be dis-
tinguished throughout.
The ampoules marked "A" and "B" contained vapor-fortified TB sediment and RMA soil, respectively. A
50.0-|jL aliquot of the 20.0-mL MeOH extraction solution was transferred with a 100-|oL syringe to a
5.00-mL syringe containing 4.90 mL of water. In order to keep a constant volume of MeOH added to the
purge-and-trap vessel, 50.0 \JtL of analyte-free MeOH was also added with a fresh syringe.
524
-------
Ampoules marked "C" contained vapor-fortified PBA soil. A 10.0-|jL aliquot of the 20.0-mL MeOH ex-
traction solution was transferred with a 25-|oL syringe to a 5.00-mL syringe containing 4.90 mL of water. In
order to keep a constant volume of MeOH added to the purge-and-trap vessel, 90.0 pL of analyte-free
MeOH was also added with a fresh syringe.
Each laboratory was asked to provide results of their (a) instrumental calibration for TDCE, TCE, Ben, and
Tol, (b) concentration estimates for TCE, Ben, and Tol in the supplied QA standard, along with results for
any other analytes detected, regardless of whether the instrument was calibrated for them, and (c) con-
centration estimates for TDCE, TCE, Ben, Tol, and any other detected analytes, for each of the six fortified
soil subsamples. All analyte concentrations were reported in (og/mL. We converted these results to soil con-
centrations using the masses of the subsamples. In addition, each laboratory was asked to report the date of
analysis, the model and manufacturer of their mass spectrometer, purge-and-trap instrument, purge-and-
trap column, gas chromatograph, and column.
RESULTS
Table 2 shows the results obtained for the analysis of the QA standard supplied with each sample set. One
laboratory failed to report their determined values. The mean values obtained for TCE, Ben, and Tol by the
other eleven laboratories were within 4% of the certified concentrations with relative standard deviations of
less than 10%.
Table 3 shows a summary of the results obtained for the analysis of three duplicate fortified soil sub-
samples. Unlike the QA standard, the analysts were not informed of the concentrations present in these soil
subsamples. When tabulating these results, some of the data were excluded based on procedural problems
identified by the analyst or by the authors. In addition, a few widely discordant values were deleted. No
more than four out of twenty-four values were excluded from a set, and in some case no values were ex-
cluded. The relative standard deviations (RSDs) obtained for these analytes are remarkably tight, ranging
from 9 to 22% with an average of less than 15%. It is important to remember that these means and standard
deviations include batch preparation variations for duplicates plus within- and between-laboratory analysis
errors.
Table 2. Preliminary assessment of the interlaboratory results
for the certified QA standard (Volatiles Mix 2, Ultra Scientific).
Benzene (50.2 |ag/mL certified)
Mean 51.9
StdDev. 3.8 (n = ll)
RSD* 7.4%
Tricholoethylene (50.1 |Og/mL certified)
Mean 51.9
StdDev. 3.8 (n=ll)
RSD 7.2%
Toluene (50.1 |Og/mL certified)
Mean 50.8
StdDev. 4.5 (n=ll)
RSD 8.8%
* Relative standard deviation
525
-------
Table 3. Preliminary assessment of the interlaboratory results for vapor-fortified soil subsamples.
Soil "A" Tampa Bay sediments (|0g/g)
TDCE Ben TCE Tol
Mean 6.0 7.2 8.6 10.3
StdDev. 1.2 1.6 1.0 0.96
RSD 20% 11% 12% 9%
n 20 21 21 21
Soil "B" Rocky Mt. Arsenal soil (|Jg/g)
TDCE Ben TCE Tol
Mean 11.6 14.0 15.4 19.7
StdDev. 2.6 1.7 1.7 2.6
RSD 22% 12% 11% 13%
n 24 24 24 24
Soil "C" Pt. Barrow Alaska soil (|Jg/g)
Mean
Std Dev.
RSD
n
TDCE
41.0
7.0
17%
22
Ben
40.7
5.2
13%
22
TCE
56.2
7.6
14%
20
Tol
67.2
9.6
14%
20
It is anticipated that, upon thorough statistical analysis of these data, even tighter relative standard devia-
tions around the consensus values will be obtained because there is evidence of systematic calibration error.
Furthermore, although not shown here, intra-laboratory results for the subsample duplicates were observed
to show mean relative standard deviations that averaged two-thirds to one-half of the inter-laboratory
results. This finding, coupled with the close agreement of the batch means, supports the concept that there
is little variation between treatment batches for a given soil.
DISCUSSION
Having reliable QA/QC soil subsamples for the analysis of VOCs will not only improve our ability to mon-
itor off-site laboratory performance, but also will allow us to evaluate alternative methods of analysis. This
is particularly important for VOCs in soil because of the difficulties encountered when sampling, handling,
and storing vadose zone subsamples (11-13). Moreover, performance evaluation materials prepared and
handled by the techniques described here provide data on both extraction and determinative operations
(14).
This round-robin study has confirmed that vapor-fortification treatment and glass ampoule confinement is a
precise means to produce and store soil subsamples contaminated with VOCs for both QA and QC pur-
poses. The subsample concentrations achieved by this method are soil-specific. Once sealed in glass am-
poules, analyte concentrations are stable during transportation and preparation for analysis. These features
make vapor-fortified soils ideal for assessing the precision and accuracy of sample preparation (extraction)
and analysis (determination) procedures.
526
-------
REFERENCES
(1) Plumb, R.H., Jr., and Pitchford, A.M. "Volatile organic scans: Implications for ground water mon-
itoring." Presented at the National Water Well Association/American Petroleum Institute Conference
on Petroleum Hydrocarbons and Organic Chemicals in Ground Water, Houston, Texas, November
13-15, 1985.
(2) Zarrabi, K., Cross-Smiecinski, A.J., and Starks, T. "A review of existing soil quality assurance
materials." In: Second International Symposium, Field Screening Methods for Hazardous Waste and
Toxic Chemicals, Las Vegas, Nevada, p. 235-252, February 12-14, 1991.
(3) Maskarinec, M.P., Johnson, L.H., and Bayne, C.K. Journal of the Association of Official Analytical
Chemists, 72: 823-827, 1989.
(4) Minnich, M. and Zimmer, J. (In press) Preparation and analysis of fortified dry soils for volatile
organic compound performance evaluation materials. Proc. of the International Symp. on Volatile
Organic Compounds (VOCs) in the Environment. Montreal, Quebec, Canada, April 17-19, 1994.
(5) Hewitt, A.D. "Feasibility study of preparing performance evaluation soils for analyzing volatile
organic compounds." U.S. Army Cold Regions Research and Engineering Laboratory, Special
Report 93-5, 1993.
(6) Hewitt, A.D. (In press) Preparation of spiked soils for volatile organic compound analysis by vapor
fortification. Journal of the Association of Official Analytical Chemists.
(7) Hewitt, A.D. "Vapor-fortified QA/QC soil subsamples for the analysis of volatile organic com-
pounds." American Environmental Laboratory, March 1994.
(8) Hewitt, A.D. (In Press) Vapor fortification: A method to prepare performance evaluation standards
for the analysis of volatile organic compounds in soil. Proc. of the 2nd International Conf. On-site
Analysis, Houston, Texas, January 24-25, 1994.
(9) Hewitt, A.D. (In Press) A new approach to evaluating pre-analytical holding times for the
determination of volatile organic compounds in soil. Proc. of the International Symp. on Volatile
Organic Compounds (VOCs) in the Environment. Montreal, Quebec, Canada. April 17-19, 1994.
(10) U.S. Environmental Protection Agency Test Methods for Evaluating Solid Waste, vol. IB, 1986.
(11) Urban, M.J., Smith, J.S., Schultz, E.K., and Dickinson, R.K. "Volatile organic analysis for a soil sed-
iment or waste sample." In: 5th Annual Waste Testing & Quality Assurance Symp., U.S. Environ-
mental Protection Agency, Washington, DC, pp. II-87-II-101, 1989.
(12) Hewitt, A.D. Journal of the Association of Official Analytical Chemists, 77, 458-463, 1994.
(13) Hewitt, A.D. (In Press) "Comparison of collection and handling practices for soils to be analyzed for
volatile organic compounds." Proc. of the International Symp. on Volatile Organic Compounds
(VOCs) in the Environment. Montreal, Quebec, Canada, April 17-19, 1994.
(14) Hewitt, A.D., Miyares, P.H., Leggett, D.C., and Jenkins, T.F. Environmental Science and Tech-
nology, 26, 1932-1938, 1992.
527
-------
ACKNOWLEDGMENTS
Funding for this work was provided by the U.S. Army Environmental Center, Martin H. Stutz, Project
Monitor. The authors thanks all the participants listed in Appendix A. In addition, the authors acknowledge
Marianne Walsh and Philip Thorne of CRREL for reviewing this manuscript.
This publication reflects the views of the authors and does not suggest or reflect policy, practices,
programs, or doctrine of the U.S. Army or of the Government of the United States.
528
-------
APPENDIX A: PARTICIPANTS IN ROUND-ROBIN STUDY
William Saner
U.S. Army Corps of Engineers
New England Division Laboratory
Hubbardson, Massachusetts 01452
Lawrence S. Hall
Data Chem Laboratories
960 LeVoy Drive
Salt Lake City, Utah 84123
Roger Rowan, Assistant Mass Spectrometrist; Don D. Gay, Director
Midwest Research Institute
425 Volker Bl
Kansas City, Missouri 64110
Thomas G. Leuschen, Analyst; Dave E. Splichal, Supervisor
U.S. Army Corps of Engineers
Missouri River Division Laboratory
420 South 18th Street
Omaha, Nebraska 68102
Bobby Jones, Analyst; Richard A. Karn, Team Leader; Karen F. Myers, QA/QC Officer; Ann B. Strong,
Group Chief; Linda K. Stevenson, Technician
U.S. Army Engineer Waterways Experiment Station
3909 Halls Ferry Road
Vicksburg, Mississippi 39180
Kerry R. Nusekabel, Analyst; John Addams, Supervisor
U.S. Army Corps of Engineers
Ohio River Division Laboratory
11275 Sebring Drive
Cincinnati, Ohio 24534
Gregory G. Lamb, Analyst; Mike Winslow, Supervisor
ESE Laboratory
P.O. Box ESE
Gainsville, Florida 32602
Gretchen Loshbaugh, Analyst; Steve Donvon, Supervisor;
Rust Geotech
PO Box 1400
Grand Junction, Colorado 81502
Steven Heller, Analyst; Suresh Srivasta and Joseph Montanaro, EPA Contacts; Joseph Montanaro, Data
Reviewer
EPA Region 1 Service Division
60 Westview Street
Lexington, Massachusetts 02173
529
-------
David Sherman, Analyst; Mike Urban, Supervisor
Enviro-Tech Research
777 New Durham Road
Edison, New Jersey 08817
Don Dale and Matthew Monagle, Analysts; Chris Leibman , Supervisor; Peggy Gautier, QA & DM Section
Leader
MS K484
Los Alamos National Lab
Los Alamos, New Mexico 07545
Chad Pidgeon
U.S. Army Cold Regions Research and Engineering Laboratory
72 Lyme Road
Hanover, New Hampshire 03755-1290
530
-------
82
ENVIRONMENTAL ANALYSIS USING HPLC WITH ON-LINE PHOTODIODE
ARRAY/MASS SPECTROMETER DETECTION
J.P. Romano, M.P. Balogh, R.C. Cotter, E.S. Bouvier, and S. Oehrie, Waters Industrial
Division of Millipore Corporation, Milford, MA 01757
ABSTRACT
High Performance Liquid Chromatography (HPLC) is an excellent analytical tool for the
analysis of environmentally significant compounds in water matrices. The analytes are
measured by photodiode array (PDA) detection in order to provide more meaningful
spectral data for environmental monitoring. Mass spectrometer (MS) detection provides
information that leads to positive compound identification of these pollutants.
In order to achieve low limits of detection, solid phase extraction combined either off-line
or on-line with the chromatographic separation is employed. A good example of this is in
the case of the European Economic Community (EEC) Drinking Water Directive which
sets a limit of 0.1 (ig/1 for individual pesticides.
The scope of this study will include the analysis of phenoxyacid herbicides, pesticides,
nitroaromatic, and nitramine explosives. A fully integrated HPLC system that includes
PDA/MS detection is described. This work also demonstrates the advanced capabilities of
a PDA detector that provides excellent sensitivity plus high resolution spectral data for peak
identification and peak purity for structurally and spectrally similar compounds
encountered when performing environmental analysis. When used simultaneously with
MS detection, confirmational information is obtained providing positive compound
identification.
531
-------
83
MICROWAVE-ASSISTED EXTRACTION OF ORGANIC COMPOUNDS FROM SOILS AND
SEDIMENTS
W.F. Beckert, U.S. Environmental Protection Agency, EMSL-LV,- Las
Vegas, Nevada 89119, and V. Lopez-Avila, R. Young, R. Kim, J.
Benedicto, and P= Ho, Midwest Research Institute, California
Operations, Mountain View, California 94043.
ABSTRACT
As part of an ongoing evaluation of novel sample-preparation
techniques by the U.S.. Environmental Protection Agency (EPA), ••
-------
technique is 10 min for- extraction and 40 min for extract
cooling, centrifugation, and extract concentration) and reduced
solvent use (30 mL in the MAE versus 300 mL in the Soxhlet
extraction). Up to 12 samples can be extracted simultaneously
with one microwave oven, resulting in high sample throughput.
NOTICE
The U.S. Environmental Protection Agency (EPA), through its
Office of Research and Development (ORD), had this abstract
prepared for a proposed oral or poster presentation. It does not
necessarily reflect the views of the EPA or ORD. Readers should
note the existence of a patent (Pare, J.R.J., et al., US Patent
5,002,784, March 1991) describing the use of microwave-assisted
extraction for biological materials.
533
-------
84
DETECTION OF TOXAPHENE IN SOIL BY IMMUNOASSAY
Titan S. Fan and Barbara Young, Millipore Corporation, Bedford MA 01730; Daniel G. Grouse,
Roy F. Weston, NJ 08837-3679; and Harry Allen, Environmental Protection Agency, NJ.08837-
3679
ABSTRACT
From the mid-1940s until the mid-1980s, toxaphene was one of the most heavily utilized
agricultural chemicals world-wide. Past studies performed on the effects of toxaphene have
implicated this chemical as a human carcinogen and mutagen. While use this chemical has virtually
been eliminated today because of its long half-life, undegraded toxaphene deposits in soil may still
be available to living organisms.
The EnviroGard Toxaphene immunoassay test kit is a user-friendly, inexpensive screening test.
This test kit was used effectively to detect toxaphene in two Navajo Nation Superfund sites:
Nazlini and Whippoorwill. Toxaphene had been used in a sheep dip solution to protect sheep from
insects at these sites. There are 136 such sites in The Navajo Nation, and 20 sites will be
remediated in 1994. Thirty soil samples were collected from the Nazlini site and analyzed by both
the EnviroGard Toxaphene test kit and by EPA Method 8081. The soil samples were split and
extracted with methylene chloride/acetone using a Soxhlet apparatus (SW-846 method 3540) for
GC/ECD analysis and with 90% MeOH for 2 minutes with an EnviroGard soil extraction kit for
immunoassay analysis. The correlation of these methods was excellent. The R value was 0.996.
INTRODUCTION
Toxaphene is a complex mixture of polychlorocamphenes. It was used on army worms, cutworms
and grasshoppers in cotton, corn and small grains. It was also used widely on cattle and sheep as
dips for scabies. Toxaphene is carcinogenic in laboratory animals, and accumulates in the animals
tissues. This insecticide was banned in November, 1982. One of the major concerns was that
toxaphene had become widespread in the Great Lakes and the Mississippi Delta regions, both with
large human populations that consumed fish from the contaminated water. During the 1960s and
1970s toxaphene was used in the United States in larger quantities than any other insecticide. In
the three years before its ban over 30,000 tones of toxaphene was applied in the southern states.
Cleaning up the toxaphene contaminated sites has become an important task for the Environmental
Protection Agency. To clean up these sites, the first step is to determine the size of the
contaminated area and the contamination level. An easy to use field test kit would be very useful to
detect the level of toxaphene in the soil. The second step is to remediate the contaminated soil.
This field kit would also be useful for monitoring the progress of remediation. The EnviroGard
Toxaphene Test Kit was used successfully in the field at two bioremediation sites in the Navajo
Nation The results are reported in this paper.
METHODS
534
-------
Description of the Immunoassay:
Polyclonal antibodies were produced in rabbits immunized with a proprietary derivative of
Cyclodiene chemically bound to a carrier protein. The antibody was coated on the inner wells of
polystyrene tubes. When toxaphene is present in the sample, it competes with the toxaphene-
enzyme conjugate for a limited number of antibody binding sites. Since there are the same number
of antibody binding sites on every test tube and each test tube receives the same number of
toxaphene-enzyme conjugate molecules, a sample that contains a low concentration of toxaphene
allows the antibody to bind many toxaphene-enzyme conjugate molecules. Therefore, a low
concentration of toxaphene produces a dark blue solution. Conversely, a high concentration of
toxaphene allows fewer toxaphene-enzyme conjugate molecules to be bound by the antibodies,
resulting in a lighter blue solution.
The protocol for the commercially available EnviroGard Toxaphene in soil test kit (part # ENVR
000 30, Millipore Corp., Bedford, MA) is as follows:
A) Soil sample extraction:
A field-portable, fast methanol extraction of soil samples is utilized with this immunoassay. Five
grams of a well homogenized soil sample are weighed with a portable balance and transferred into
a disposable extraction bottle containing three stainless-steel mixing balls. Ten ml of 90%
methanol in water are added and the bottle is shaken vigorously by hand for 2 minutes. The soil is
allowed to settle for at least one minute, then a filtration cap is placed on the extraction bottle. After
pumping air into the bottle with a syringe, the bottle is inverted over a glass collection vial. The
filtered extract is allowed to drip into the vial, which is then tightly capped and stored under
refrigeration, in the dark, until analyzed.
B) Assay protocol:
1. Add 250 ul of assay diluent to each test tube. Add 50 ul of each calibrator ( 0.5, 2.0 and 10
ppm) and sample extracts to the corresponding test tubes. Let the test tubes incubate for 15 min
before adding 200 ul of conjugate to each tube.
2. Shake the test tube rack and incubate for 5 minutes. Vigorously shake out the test tube contents
into a sink or suitable container. Wash the tubes four time with laboratory-grade water, removing
any unbound compounds.
3. A clear solution (500 ul) of chromogenic substrate (3,3,5,5-tetramethylbenzidine hrdrochloride)
is then added to the tubes and incubated for 5 minutes.
4. Assay results may be instrumentally recorded by adding 500 ul of IN HC1 stop solution to each
tube and reading the optical density (OD) at 450 nanometers(nm) minus a 600 nm reference
wavelength.
Immunoassay Validation Procedures:
535
-------
1. Matrix effects/Determination of the Lower Limit of Detection: For these tests, eight control soil
samples were subjected to triplicate extraction and triplicate assays of each extract, in order to
determine the varibility associated with toxaphene-free soils. The mean color development due to
these samples minus 2 standard deviations was determined and defined to equal the loweset
concentration of toxaphene in soil detectable by the immunoassay.
2. False positive/False Negative Testing: Four soil samples were spiked with toxaphene at 0.4
ppm and four at 4.0 ppm, and compared to a 2 ppm toxaphene calibrator in triplicate assays. The
rate of occurrence of "false positives" and "False negatives" was determined.
3. Assay of Toxaphene-Spiked Soil Samples: Three toxaphene-free soils were spiked at two levels
with toxaphene, in triplicate, and assayed along with three toxaphene calibrators. The rate of
correct interpretation of the samples was determined.
4. Cross-Reactivity Studies: A large number of compounds structurally related to toxaphene were
tested for their reactivity in this immunoassay. The compounds were prepared in methanol and run
as samples in the EIA.
5. Tests with Field Samples: Thirty contaminated soil samples from Navajo Superfund Nazlini site
were tested by EnviroGard Toxaphene test kits and split samples were extracted according to
Method 3540 (Soxhlet extraction) and analyzed by SW-846 method (8081) with GC/ECD. The
extent of agreement between the EIA and GC was assessed.
RESULTS AND DISCUSSION
Validation Procedures:
1. Repeated testing of toxaphene-free soils resulted in an overall mean %Bo of 88.1, with a
standard deviation (SD) of 4.3. The mean minus 2 SD equals 79.5% Bo. Reading 79.5%Bo off
the graph of toxaphene gives an approximate lower limit of detection of 250 ppb toxaphene in soil.
2. An exercise was conducted to determine the rate of false negative and false positive assay calls
when soils were spiked to 0.4 or 4 ppm toxaphene and compared to 2 ppm calibrator. Any sample
with an OD greater than the OD of the 2 ppm calibrator was called "<2 ppm". Any sample with an
OD equal to or less than the OD of the 2 ppm calibrator was called'>2 ppm". Out of 48 samples
run, there were no false positive calls and zero false negative calls.
3. The 6 results of three soils spiked with 1.0 and 10 ppm toxaphene showed a mean recovery of
98%.
4. Table 1 summarizes the results of cross-reactivity testing conducted with the Toxaphene EIA.
The extent of cross-reactivity is displayed in two ways: The concentration of compound that
corresponds to the lower limit of detection (LLD) in the assay, and the concentration of the
compound required to inhibit 50% of the color developed by negative control (50%Bo).
536
-------
Chlordane, endrin, endosulfan endosulfan II, dieldrin, heptachlor aldrin have cross reactivities at
ppb levels while toxaphene, gamma-BHC, alpha-BHC and delta-BHC have ppm levels.
S.Tests conducted with thirty field samples showed very good agreement between EIA and GC as
shown in Figure 1.
CONCLUSIONS:
All of the validation procedures run within our laboratory have demonstrated the ability of the
toxaphene in soil test kit to effectively screen for toxaphene in soil samples. The combination of its
ease of use and low cost per sample can allow for increased rates of sampling, potentially resulting
in improved characterization and mapping of polluted sites as demonstrated at the Nazlini site. The
use of complex, expensive chromatographic techniques may then be used for quality control and
confirmation.
ACKNOLEDGEMENTS:
Thanks to Stanley Edison, Eugene Esplain, Joseph Tom Morris from Navajo Superfund Office,
Darwin W. Lowery from USD A for assistance to obtain Nazlini soil samples.
LITERATURE CITED
1) Bidleman, T.F. M.T. Zaranski, and M.D. Walla. "Toxaphene: Usage, Areal transport and
Deposition" 1988. In Toxic Contamination in Large Lakes, N.W. Schmidtke, Lewis Publishers.pp
257-284.
537
-------
Toxaphene in Soil
en
w
CD
EnviroGard
ppm
1 \J \J \y -
-
100 .
"
-
-
1
~
-
0.1
c
n
H L-3
D
C
i i i 1 1 in
n
]
DD g ^
n
u
n
D
i i 1 1 1 M i
0.01 0.1 1
n
ib
i i 1 1 1 1 u
c
V
1 1 1 1 1 1 II
n
1 1 1 I 1 1 1 1
10 100 1000
GC/ECD ppm
-------
01
oo
CD
Cross-Reactivity Studies
Compound
Toxaphene
Chlordane
Endrin
Endosulfan
Endosulfan II
Dieldrin
Heptachlor
Aldrin
Gamma-BHC
Alpha-BHC
Delta-BHC
Cone, in Soil
Gamma-BHC =
LLD
0.2 ppm
14 ppb
6 ppb
6 ppb
6 ppb
6 ppb
6 ppb
20 ppb
0.6 ppm
2 ppm
2 ppm
Lindane
IC50
2.8 ppm
76 ppb
22 ppb
36 ppb
28 ppb
42 ppb
34 ppb
1 1 6 ppb
4.6 ppm
19 ppm
40 ppm
-------
85
DETECTION OF DDT AND ITS METABOLITES IN SOIL BY ENZYME
IMMUNOASSAY
Karen A. Larkin. Research Scientist, Jonathan J. Matt, Research Scientist, and Bruce S.
Ferguson, Chief Scientific Officer, ImmunoSystems, Subsidiary of Millipore Corporation,
4 Washington Avenue, Scarborough, Maine 04074; Helen L. Beasley, Experimental
Scientist, Division of Plant Industry, CSIRO, PO Box 7, North Ryde, NSW 2113,
Australia; John H. Skerritt, Principal Research Scientist, Division of Plant Industry,
CSIRO, GPO Box 1600, Canberra, ACT 2601, Australia
ABSTRACT
DDT (l,l,l-trichloro-2,2-bis (4-chlorophenyl) ethane) is a highly persistent,
environmentally toxic organochlorine insecticide. It has been very heavily applied for
many decades in countries around the world. The combination of its liberal use and its
long-term soil persistence has resulted in the existence of many soil sites contaminated with
DDT and its major metabolites, DDD and DDE. Conventional analytical techniques for
detection of these compounds in soil are costly and time-consuming. Therefore, an enzyme
immunoassay (EIA) has been developed to detect DDT and its metabolites in soil. Rabbit
polyclonal antibodies prepared against a DDT derivative are pre-coated to polystyrene test
tubes. Soil extracts and DDT-enzyme conjugate are added to the tubes and compete for
antibody binding sites on the tubes. Excess conjugate is washed away and a substrate
solution is added; subsequently, the amount of color developed in the tube is inversely
proportional to the concentration of DDT in the sample. This field-portable screening test
utilizes a rapid and simple methanol extraction of soil and can detect DDT and its
metabolites at levels as low as 0.1 ppm. Assay results can be obtained within 30 minutes.
EIA validation data and correlation with a gas-chromatographic method will be presented.
INTRODUCTION
The potency of DDT as an insecticide was discovered in 1939 by Dr. Paul Muller, working
at what is now Ciba-Geigy AG in Switzerland (1). DDT became highly valued for its low
cost, its potential to help control insect-borne disease, and for its persistence after
application, and was very heavily used world-wide, from 1942 to 1972. Although there is
no solid evidence of lasting harm to humans due to routine exposure to DDT, other ill
effects of the compound eventually became apparent. DDT accumulates in the fat of living
organisms and becomes concentrated in animals at the top of the food chain. This proved
detrimental to species of predatory birds, in which thinning eggshells and other
reproductive problems were directly attributed to DDT and certain of its metabolites. DDT
was also found to be acutely toxic to fishes. Consequently, the use of DDT is currently
either banned or severely restricted in many parts of the world. (2) Because DDT and its
metabolites are so persistent in soil, with half-lives estimated to be 15 to 20 years (3),
residues are still very much a concern, even in areas where the pesticide is no longer used.
Most commonly, DDT and other organochlorine pesticides are quantitated by costly gas-
chromatographic (GC) methods involving either electron capture detection (BCD) or mass-
spectroscopy. These methods must be preceded by time consuming extraction and clean-
up procedures which utilize large volumes of solvents. The need exists for a fast, cost-
effective screening method for "total DDT" (the sum of p,p'-DD1, p,p'-DDD, and p,p'-
DDE) in soil.
540
-------
There have been a number of studies published recently utilizing enzyme immunoassay to
detect various pesticides in soil, including atrazine (4,5), chlordane (6), and metolachlor
(7). Antibodies for the detection of DDT have been reported previously (8-11). The
objective of the following work was to couple a fast soil extraction protocol with a simple
immunoassay test, together capable of providing accurate estimations of total DDT isomers
in soil.
METHODS
Description of the Immunoassay:
Polyclonal antibodies were produced in rabbits immunized with a proprietary derivative of
DDT chemically bound to a carrier protein. The antibody was coated on the inner walls of
polystyrene tubes at various concentrations and titered against serial dilutions of a DDT-
derivative conjugated to horseradish peroxidase enzyme, until the optimum combination
was obtained. Various immunoassay protocols were tested, with the final choice being a
simultaneous competitive format. This finished immunoassay possesses the desired
reactivity with p,p'-DDT, p,p'-DDD, and p,p'-DDE, and can detect all three compounds
with sensitivities in the range of 0.1 ppm in soil.
The protocol for the commercially available DDT in Soil Test Kit (part #ENVR 000 31,
Millipore Corp., Bedford, MA) is as follows:
1. A sample extract or calibrator (25 pL) containing DDT is added to an antibody-coated
test-tube containing 500 jiL of an assay diluent. 100 |iL of DDT-enzyme conjugate is
added and the tubes are mixed and incubated at ambient temperature for 15 minutes.
2. The tubes are then washed with tap or laboratory-grade water, removing any unbound
compounds.
3. A clear solution (500 |J,L) of chromogenic substrate (3,3,5,5-tetramethylbenzidine
hydrochloride) is then added to the tubes and incubated at ambient temperature for 10
minutes. In the presence of bound DDT-enzyme conjugate, the clear substrate is converted
to a blue color. Since there are the same number of antibody binding sites on every test
tube and each test tube receives the same number of DDT-enzyme conjugate molecules, a
sample that contains a low concentration of DDT allows the antibody to bind many DDT-
enzyme conjugate molecules. Therefore, a low sample concentration of DDT produces a
dark blue solution, while a high sample concentration of DDT allows fewer DDT-enzyme
conjugate molecules to be bound by the antibodies, resulting in a lighter blue solution.
4. Assay results may be instrumentally recorded by adding 500 |iL of a 1 N HC1 stop
solution to each tube and reading the optical density (OD) at 450 nanometers (nm) minus a
600 nm reference wavelength. (Alternatively, a battery-operated differential photometer
may be used in field situations.) Results are interpreted by comparing the OD of samples to
the OD of p,p'-DDT calibrators provided in the test kit.
Soil Sample Extraction:
A field-portable, fast methanol extraction of soil samples is utilized with this immunoassay.
Five grams of a well-homogenized soil sample are weighed with a portable balance and'
transferred into a disposable extraction bottle containing three stainless-steel mixing balls.
Five mL of methanol are added and the bottle is shaken vigorously by hand for 2 minutes.
The soil is allowed to settle for at least 1 minute, then a filtration cap is placed on the
541
-------
extraction bottle. After pumping air into the bottle with a syringe, the bottle is inverted over
a glass collection vial. The filtered extract is allowed to drip into the vial, which is then
tightly capped and stored under refrigeration, in the dark, until analyzed.
Immunoassay Validation Procedures:
1. Matrix Effects/Determination of the Lower Limit of Detection: For these tests, eight
control soil samples were subjected to triplicate extraction's and triplicate assays of each
extract, in order to determine the variability associated with DDT-free soils. The mean
color development due to these samples minus 2 standard deviations was determined to
equal the lowest concentration of p,p'-DDT in soil detectable by the immunoassay.
2. False Positive/False Negative Testing: Eight soil samples were spiked with p,p'-DDT
at 0.2 and 2.0 ppm, and compared to a 1 ppm p,p'-DDT calibrator in triplicate assays. The
rate of occurrence of "false positives" and "false negatives" was determined.
3. Assay of DDT-Spiked Soil Samples: Three DDT-free soils were spiked at two levels
withp,p'-DDT, in triplicate, and assayed along with three p,p'-DDT calibrators. The rate
of correct interpretation of the samples was determined.
4. Cross-Reactivity Studies: A large number of compounds structurally related to DDT
were tested for their reactivity in this immunoassay. The compounds were prepared in
methanol and run as sample extracts in the El A.
5. Tests With Samples Containing Incurred Residues of DDT: Two separate field site
studies were conducted. In each of the below cases, extent of agreement between the EIA
and GC was assessed.
Forty-seven soil samples from a site in Kansas were extracted using the immunoassay
protocol. These methanol extracts were analyzed by both immunoassay and in-house gas-
chromatography:
A Perkin-Elmer (PE, Norwalk,CT) AutoSystem GC with a PE-608 column (30 m,
0.53 mm i.d., 0.80 (im film thickness, methyl phenyl cyanopropyl silicone) was
used, under the following conditions:
oven: 190 °C to 260°Cat 10°/min., held for 3 min.
• injector: packed, converted to megabore capillary, using direct flash
vaporization, held at 200°C.
• injection size: 1 |iL
• detector: ECD held at 350°C, with argon/methane make-up
• carrier gas: helium, at 7 mL/min.
Thirty-two soil samples from a site in Florida were each split into two aliquots. One
aliquot was extracted and analyzed by the EIA protocols; the other aliquot was sent to an
independent laboratory (APPL Laboratories, Inc., Fresno, CA) where it was extracted by
EPA Method 3550 and analyzed by GC with EPA Method 8080.
RESULTS
Standard Curve: The useful range of the DDT immunoassay was determined by
running p,p'-DDT over a wide span of concentrations. Data were normalized against the
negative control (neat methanol) run in each assay by calculating the %Bo of each sample.
Percent Bo equals the OD of the sample or calibrator divided by the OD of the negative
542
-------
control, multiplied by 100. The graph in Figure 1 shows the standard curve forp,/?'-DDT
soil extracts, as well as those for p,p'-DDD and p,p'-DDE. Because the p,p'-DDT curve is
most linear from 0.1 to 10 ppm, calibrators of 0.1, 1, and 10 ppm were used during the
EIA validation process. Calibrators for the commercially available DDT in Soil Test Kit
were chosen to be 0.2, 1, and 10 ppm, to avoid potential false positive interpretations due
to soil matrix effects. The range of the EIA can be easily expanded by diluting sample
extracts in methanol prior to running them in the immunoassay. For example, samples
known to contain 100 ppm or more can be characterized by diluting the sample extracts
1:100 in methanol prior to assay. Due to this 1:100 dilution, the above kit calibrators
would then represent 20, 100, and 1000 ppm, and the soil samples can be described in
relation to these levels.
Caution must be used with soil samples which contain greater than 1000 ppm DDT; the
limited solubility of the compound in methanol may adversely affect extraction efficiency.
In this case, it is recommended that the volume of methanol used to extract 5 grams of soil
be increased to 20 mL from 5 mL. This is a 1:4 dilution factor which must be factored into
the sample interpretation. For example, for a site with a clean-up action level of 2000 ppm,
soils should be extracted with 20 mL of methanol and extracts should be diluted 1:500 in
methanol. Due to this 1:2000 dilution (1:4 during extraction and 1:500 of the extract) the
kit calibrators would represent 400, 2000 and 20,000 ppm total DDT in soil.
Soil Extraction: The soil extraction method used with the DDT in Soil EIA is not
expected to be 100% efficient for all samples. Preliminary trials showed negligible
differences in total DDT recovery from selected contaminated samples when the extraction
times were increased from 2 minutes to 24 hours. Because the objective of this method
development was to create a fast, field portable screening test, the 2 minute extraction time
was deemed sufficient. As noted above, highly contaminated samples may suffer
decreased extraction efficiencies due solely to the problem of solubility. Increasing the
volume of extraction solvent helps to solve this problem.
Validation Procedures:
1. Repeated testing of DDT-free soils resulted in an overall mean %Bo of 93.4, with a
standard deviation (SD) of 6.0. The mean minus 2 SD is equal to 81.4% Bo. Reading
81.4% Bo off the graph of p,p'-DDT in Figure 1 gives an approximate lower limit of
detection of 0.044 ppm p,p'-DDT in soil.
2. An exercise was conducted to determine the rate of false negative and false positive
assay calls when soils were spiked to 0.2 or 2 ppm p,p'-DDT and compared to a 1 ppm
calibrator. Any sample with an OD greater than the OD of the 1 ppm calibrator was called
"<1 ppm". Any sample with an OD equal to or less than the OD of the 1 ppm calibrator
was called "> 1 ppm". Out of 48 samples run, zero false positive calls resulted (i.e. all of
the 0.2 ppm spikes were called "< 1 ppm"). Six false negative calls resulted (i.e. 6 of the 2
ppm spikes were called "< 1 ppm"). This rate of "false negatives" was unacceptable. To
correct the situation, calibrators in the commercial EIA kit contain p,p'-DDT at
concentrations that are 75% of that stated on the calibrator labels. Calculations from the
data generated in this section of the validation indicated that this would have eliminated all
the false negative calls in the data set without creating any false positive calls.
3. The analysis of soils spiked with 0.25 and 2.5 ppm p,p'-DDT showed 100% correct
calls of all runs of all replicate spikes.
543
-------
4. Table 1 summarizes the results of cross-reactivity testing conducted with the DDT EIA.
The extent of cross-reactivity is displayed in three ways: the concentration of compound in
soil that corresponds to the lower limit of detection in the assay (81.4% Bo), the
concentration of the compound in soil required to inhibit one-half of the color developed by
negative control tubes (50% Bo), and the % cross-reactivity.
This data shows that the immunoassay is well equipped to measure the total of the p,p'-
isomers of DDT and its metabolites. The o,p'- forms of DDT and DDE are recognized very
poorly by the antibody, while o,p'-DDD is somewhat more reactive. Because the
reactivities of all these components are not identical, and because at least three of them are
normally found in contaminated soils in varying proportions, this EIA cannot be used
accurately as a quantitative measuring device, but instead is intended as an effective
screening tool.
Chloropropylate and DDA are far better recognized by this polyclonal antibody than is
DDT. This should not be of concern to most soil analysts, as DDA (the major metabolite of
DDT in living organisms) is not prevalent in soil and chloropropylate is a discontinued
acaracide with limited persistence.
Chlorobenzilate and dicofol can be detected with approximately the same sensitivity as the
p.p'-DDT isomers, making the EIA useful to analysts looking for one of these compounds
rather than, or in addition to, DDT.
The remainder of the chemicals on the list are mildly reactive, and demonstrate the affinity
of this polyclonal antibody for compounds containing a terminal phenyl group with a
chlorine in the para- position.
5. Tests conducted with two groups of soil samples containing incurred residues of DDT
showed very good agreement between EIA and GC, both on split soil extracts and on split
soil samples.
Table 2 summarizes the data on the Kansas site soils, for which GC data was generated on
the same extracts used in the EIA. There were 4 disagreements between the two
techniques. Three of these involved samples containing less than 0.2 ppm total DDT. Of
the four disagreements, three were called more positive by EIA than by GC, which is the
preferred type of error. These samples may well contain a cross-reacting compound which
was not tested for in the GC protocol. The fact that some of the samples which were called
"<0.1 ppm" by GC were called ">0.1 ppm" by the EIA prompted the change of the lowest
kit calibrator in the commercial product from 0.1 to 0.2 ppm.
Table 3 summarizes the data from the Florida site soils, which was a test of the extraction
protocol as well as the immunoassay, because soil samples were split and extracted by
either the simple EIA protocol or by official EPA methods in an independent analytical
laboratory. Please note that these data were generated after the lowest calibrator in the
commercial kit was changed to 0.2 ppm from 0.1 ppm. There were a total of five
disagreements in this data set, all of which involved the EIA calling positive samples more
positively than did the GC. This indicates that the EIA extraction protocol is sufficient for
these low-level DDT soil samples, and that a slight trend for over-estimation of DDT
content exists. However, all of the GC "not detectable's" (less than 0.05 ppm total DDT
isomers) were correctly interpreted as containing less than the 0.2 ppm kit calibrator.
544
-------
CONCLUSIONS
All of the validation procedures run within our laboratory have demonstrated the ability of
the DDT in Soil Test Kit to effectively screen for total DDT in soil samples. The
undesirable rate of false negatives seen in the second validation parameter have been
corrected in the commercial test kit by lowering the concentration of p,p'-DDT in the
calibrators to 75% of their nominal value. The potential for false positives due to unknown
matrix effects from field-site soils was decreased by raising the lowest kit calibrator level to
0.2 ppm.
This DDT in Soil immunoassay possesses great applicability as a field screening tool. Its
0.2 ppm detection limit should prove more than adequate for measuring tolerable levels of
soil contamination under most federal or state guidelines and regulations. The combination
of its ease of use and low cost per sample can allow for increased rates of sampling,
potentially resulting in improved characterization and mapping of polluted sites. The use of
complex, expensive chromatographic techniques may then be limited to quality control and
confirmations.
REFERENCES
(1) Ecobichon, D. J., Introduction, in Pesticides and Neurological Diseases, ed. D. J.
Ecobichon and R. M. Joy, pp. 1-14, CRC Press, Boca Raton, FL, 1982.
(2) Murphy, S. D., Pesticides, in Casarett and Doull's Toxicology: The Basic Science of
Poisons, ed. J. Doull, C. D. Klaassen, M. O. Amdur, pp. 357-408, Macmillan, New
York, 1980.
(3) Martijn, A., Bakker, H. and R. H. Schreuder, Soil Persistence of DDT, Dieldrin, and
Lindane Over a Long Period, Bull. Environ. Contam. Toxicol., 51: 178-184, 1993.
(4) Bushway, R. J., Perkins, B., Savage, S. A., Lekousi, S. J. and B. S. Ferguson,
Determination of Atrazine Residues in Water and Soil by Enzyme Immunoassay, Bull.
Environ. Contam. Toxicol.. 40: 647-654,1988.
(5) Goh, K. S., Hernandez, J., Powell, S. J. and C. D. Greene, Atrazine Soil Residue
Analysis by Enzyme Immunoassay, Bull. Environ. Contam. Toxicol.. 45: 208-214, 1990.
( 6) Bushway, R. J., Pask, W. M., King, J., Perkins, B. And B. S. Ferguson,
Determination of Chlordane in Soil by Enzyme Immunoassay, Proceedings Field Screening
Methods for Hazardous Waste Site Investigations. U.S. EPA, Las Vegas, 1988.
(7) Schlaeppi, J. M., Moser, H. and K. Ramsteiner, Determination of Metolachlor by
Competitive Enzyme Immunoassay Using a Specific Monoclonal Antibody, J. Agric. Food
Chem.. 39: 1533-1536, 1991.
(8) Haas, G. J. and E. J. Guardia, Production of Antibodies Against Insecticide Protein,
Proc. Soc. Exp. Biol. Med.. 129: 546-551, 1968.
(9) Centeno, E. R., Johnson, W. J. and A. H. Sehon, Antibodies to Common Pesticides
DDT and Malathion. Int. Archs. Allergy Appl. Immun.. 37: 1-13, 1970.
545
-------
(10) Furuya, K. and S. Urasawa, Gas-Liquid Chromatographic Demonstration of the
Specificity of Rabbit IgG Antibody to the Pesticide DDT and Its Metabolites, Molecular
Immunol.. 18: 95-102, 1981.
(11) Burgisser, D., Frey, S., Gutte, B. and S. Klauser, Preparation and Characterization
of Polyclonal and Monoclonal Antibodies Against the Insecticide DDT, Biochem. Biophys,
Res. Comm.. 166: 1228-1236, 1990.
546
-------
0
0.001
FIGURE 1
EIA REACTIVITY OF p,p' DDT, DDD, AND DDE
*-p,p'-DDT ___
B-p,p':DDD~~
A-p,p'-DDE
% Bo
100 1000
Pesticide in Soil (ppm)
547
-------
TABLE 1
CROSS-REACTIVITY STUDIES
Compound
p,p'-DDT
p,p'-DDD
p,p'-DDE
o,p'-DDT
o,p'-DDD
o,p'-DDE
DBA
Chloropropylate
Chlorobenzilate
Dicofol
Chloroxuron
Monolinuron
Thiobencarb
Tebuconazole
Neburon
Tetradifon
Diclofop
PPM
LLD
0.04
0.01
0.18
4
0.4
3
0.002
0.007
0.03
0.14
24
25
5
7
17
1.2
70
in Soil
IC50
1.25
0.3
3.6
93
11
93
0.04
0.08
0.35
2
220
710
52
95
280
14
> 1000
% C.R.*
100
417
35
1.3
11
1.3
3125
1562
357
63
0.6
0.2
2
1.3
0.4
9
<().!
The following are not detected at 100 ppm:
2,4-D
Dicamba
Chlordane
Lindane
Diflubenzuron
Chlortoluron
Chlorbromuron
4-chlorophenoxyacetic acid
2,4-Dichloronitrobenzene
Mecoprop
Diuron
MCPB
MCPA acid
Linuron
*% C.R. = % cross-reactivity = the IC50 of p,p'-DDT divided by the IC50 of
the cross-reactant, x 100.
543
-------
TABLE 2
KANSAS FIELD SITE RESULTS
Samnle ID
61701
61702
61703
61704
61706
61707
61708
61709
60801
60802
60803
60804
60807
60808
60809
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
GC
Total DDT
(Dom)
18
12
11
30
0.44
0.47
0.40
0.33
41
52
40
79
0.18
0.23
0.18
1140
2755
245
850
1832
1185
14,375
5025
460
217
536
308
1316
1415
1204
1370
435
4800
755
1760
115
579
129
115
2904
1512
< 0.1
< 0.1
< 0.1
<0.1
0.14
0.15
Immunoassay
Interpretation
(pom ranees)
> 10<50
> 10<50
> 10 < 100
> 10<50
> 0.1 < 1
> 0.1 < 1
> 0.1 < 1
> 0.1 < 1
> 10 < 100
> 10 < 100
> 10 < 100
> 10 < 100
>0.1 < 1
>0.1 < 1
>0.1 < 1
> 500 < 5000
> 500 < 5000
> 50 < 500
> 100 < 1000
> 500 < 5000
> 1000 < 5000
> 5000 < 50,000
> 5000 < 50,000
> 100 < 1000
> 50 < 500
> 200 < 2000
> 50 < 500
> 200 < 2000
> 1000 < 10,000
> 200 < 2000
> 1000 < 10,000
> 50 < 500
> 5000 < 50,000
> 100 < 1000
> 1000< 10,000
> 20 < 200
> 100 < 1000
> 30 < 300
> 100 < 1000
> 300 < 3000
> 200 < 2000
<().!
> 0.1 < 1
<0.1
>0.1 < 1
<().!
>0.1 < 1
Agreement?
(YES/NO)
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
NO
YES
YES
YES
YES
YES
YES
YES
YES
YES
NO
YES
NO
NO
YES
549
-------
TABLE 3
FLORIDA FIELD SITE RESULTS
Sample ID
2
3
4
5
6
7
8
9
10
13
14
15
17
20
21
22
23
24
25
26
27
28
28-17D
29
30
31
32
33
34
35
36
41F
GC
Total DDT
(ppm)
3.6
0.55
2.3
ND*
0.15
0.3
0.1
0.8
0.23
0.79
0.58
0.35
ND
0.18
0.06
ND
ND
1.2
0.12
ND
ND
0.16
0.18
0.69
0.73
0.68
ND
0.32
0.23
0.52
1
ND
Immunoassay
Interpretation
(ppm ranges)
>10
> 0.2 < 1
>1<10
<0.2
> 0.2 < 1
> 0.2 < 1
<0.2
> 0.2 < 1
> 0.2 < 1
> 0.2 < 1
> 0.2 < 1
> 0.2 < 1
<0.2
<0.2
<0.2
<0.2
<0.2
>1<10
<0.2
<0.2
<0.2
<0.2
> 0.2 < 1
> 0.2 < 1
>1<10
>1<10
<0.2
> 0.2 < 1
> 0.2 < 1
> 0.2 < 1
> 0.2 < 1
<0.2
Agreement?
(YES/NO)
NO
YES
YES
YES
NO
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
NO
YES
NO
NO
YES
YES
YES
YES
YES
YES
*ND = not detectable (< 0.05 ppm total DDT isomers).
550
-------
DETERMINATION OF POLYNUCLEAR AROMATIC HYDROCARBONS
(PAHs) IN SOIL BY A MAGNETIC PARTICLE-BASED ENZYME
IMMUNOASSAY.
Fernando M. Rubio. Timothy S. Lawruk, Charles E. Lachman and David P. Herzog,
Ohmicron Environmental Diagnostics, 375 Pheasant Run, Newtown, Pennsylvania 18940;
James R. Flecker, North Dakota State University, P.O. Box 5516, Fargo, North Dakota
58105
ABSTRACT
Use of immunoassays as field-screening methods to detect environmental contaminants
has increased dramatically over the past few years. Immunochemical assays are sensitive,
rapid, reliable, cost-effective and can be used for lab or field analysis. A magnetic particle-
based immunoassay system has been developed for the quantitation of Polynuclear
Aromatic Hydrocarbons (PAHs) in soil. Paramagnetic particles used as the solid-phase,
allow for the precise addition of antibody and rapid reaction kinetics. The magnetic
particle-based immunoassay is ideally suited for on-site investigation and remediation
processes to delineate PAHs contamination. This system includes easy-to-use materials
for collection, extraction, filtration and dilution of soil samples prior to analysis by
immunoassay. The method detects PAHs, including anthracene, chrysene, fluoranthene,
phenanthrene, pyrene and benzo[a]pyrene, at sub-parts per million levels in soil. The
assay procedure and detailed performance characteristics including precision, spike
recovery and correlation with U.S. EPA methods are discussed.
INTRODUCTION
Polycyclic or polynuclear hydrocarbons (PAHs) are a group of compounds composed of
two or more fused aromatic rings. The U.S. EPA has selected 16 unsubstituted PAHs as
Consent Decree priority pollutants for regulatory purposes. Some of the four, five and
six-ring PAHs such as chrysene, benzo[a]pyrene and indeno[l,2,3-cd]pyrene are
considered to be possible or probable human carcinogens (U.S. EPA, 1985). The two and
three-ring PAHs such as naphthalene, anthracene, and phenanthrene are non-carcinogenic
and found as a component of certain grades of fossil fuels.
PAHs are introduced into the environment as a product of natural and fossil fuel
combustion. Volcanic eruptions and forest fires are among the major sources of naturally
produced PAHs. However, activities attributed to fossil fuel combustion sources, such as
automobiles, coking plants, asphalt production, and manufacturing facilities that use fossil
fuels, have dramatically increased the quantity of PAHs in the environment. Wood
551
-------
preserving sites that use creosote as a preservative, petrochemical waste disposal sites,
and leaking underground fuel storage tanks, have also contributed to the widespread
contamination of PAHs in the environment.
The large number of sites contaminated with PAHs in soil and groundwater and the
reenactment of key environmental legislation (Safe Drinking Water Act, Superfund
Amendment Reauthorization Act) has led federal and state agencies to mandate their
clean-up. The federal and state agencies have set various regulatory levels for PAHs in
soil, however the usual concentrations of interest are 1 ppm and 10 ppm. The analysis of
PAHs contamination in environmental samples is typically performed by GC/MS or HPLC
methods which are accurate and precise but can be time-consuming and expensive. This
poster describes a magnetic-particle solid-phase immunoassay method in soil samples.
Immunoassays have the advantage of being rapid and less expensive than GC/MS or
HPLC, as well as field-portable.
The principles of enzyme linked immunosorbent assays (ELISA) have been described
(Hammock and Mumma, 1980). Magnetic particle-based ELISA's have previously been
described and applied to the detection of pesticide residues (Itak et al, 1993; Lawruk et al,
1993; Itak et al, 1992; Lawruk et al, 1992; Rubio et al, 1991). These ELISA's eliminate
the imprecision problems that may be associated with antibody coated plates and tubes
(Harrison et al, 1989; Engvall, 1980) through the covalent coupling of antibody to the
magnetic particle solid-phase. The uniform dispersion of particles throughout the reaction
mixture allows for rapid reaction kinetics and precise addition of antibody. The PAHs
magnetic-based ELISA described in this paper combines antibodies specific for PAHs with
enzyme labeled PAHs. The presence of PAHs in a sample is visualized through a
colorimetric enzymatic reaction and results are obtained by comparing the color in sample
tubes to those of calibrators.
MATERIALS AND METHODS
Amine terminated superparamagnetic particles of approximately 1 um diameter were
obtained from Advanced Magnetics, Inc. (Cambridge, MA). Glutaraldehyde (Sigma
Chemical, St. Louis, MO). Rabbit anti-PAHs serum and PAH-HRP conjugate (Ohmicron,
Newtown, PA). Hydrogen peroxide and TMB (Kirkegaard & Perry Labs, Gaithersburg,
MD). PAHs and related compounds, as well as non-related cross-reactants (Chem
Service, West Chester, PA).
The anti-PAHs coupled magnetic particles were prepared by glutaraldehyde activation
(Weston and Avrameas, 1971). The unbound glutaraldehyde was removed from the
particles by magnetic separation and washing four times with 2-(N-morpholino) ethane
sulfonic acid (MES) buffer. The PAHs antiserum and the activated particles were
incubated overnight at room temperature with agitation. The unreacted glutaraldehyde
552
-------
was quenched with glycine buffer and the covalently coupled anti-PAHs particles were
washed and diluted with a Tris-saline/gel preserved buffer.
The various PAHs compounds used during cross-reactivity studies were diluted in DMF
to obtain a stock concentration of 1 mg/mL. The stock was further diluted in PAH diluent
to obtain concentrations of 10 , 1, 0.1, 0.01, 0.001, and 0.0001 ppm. The creosote sample
was diluted in methanol to obtain a stock concentration of 1 mg/mL, the stock was further
diluted as listed previously. After dilution, the diluted compounds were analyzed as
samples in the assay.
When analyzing soil samples, a simple extraction is performed prior to analysis: 10 g of
soil and 20 mL of a methanolic solution are added to a soil collector (Figure 1). The
collector was shaken vigorously for 1 minute and the mixture allowed to sit at least five
minutes. The cap of the soil collector was then replaced with a filter cap and the extract
collected in a small glass vial. The filtered extract was then diluted 1:50 in PAH zero
standard and assayed.
Diluted soil extract sample (250 uL) and horseradish peroxidase (HRP) labeled PAHs
(250 uL) were incubated for 30 minutes with the antibody coupled solid-phase (500 uL)
(step 1). A magnetic field was applied to the magnetic solid-phase to facilitate washing
and removal of unbound PAHs-HRP and eliminate any potential interfering substances
(step 2). The enzyme substrate (hydrogen peroxide) and TMB chromogen (3,3',5,5'-
tetramethyl benzidine) were then added and incubated for 20 minutes (step 3). The
reaction was stopped with the addition of acid and the final colored product was analyzed
using the RPA-I RaPID Analyzer™ by determining the absorbance at 450 nm. The
observed absorbance results were compared to a linear regression line using a log-logit
standard curve prepared from calibrators containing 0, 2.0, 10.0, and 50.0 ppb of
phenanthrene. If the assay is performed in the field (on-site), a battery powered
photometer such as the RPA-III™ can be used.
RESULTS AND DTSCUSSTON
Figure 2 illustrates the mean standard curve for the PAHs calibrators collected over 30
runs; error bars represent two standard deviations (SD). This figure shows the typical
response of the assay and the reproducibility of the standard curve from run-to-run. The
displacement at the 2.0 ppb level is significant (81.3% B/Bo, where B/Bo is the
absorbance at 450 nm observed for a sample or standard divided by the absorbance at the
zero standard). The assay sensitivity in diluent based on 90% B/Bo (Midgley et al, 1969)
is 0.70 ppb. When analyzing soils with PAH Sample Extraction Kit, the assay has a range
of 0.20 ppm to 5 ppm as a result of sample dilution.
553
-------
In a precision study conducted using diluent buffer fortified with phenanthrene at 4
concentrations, the samples were assayed 5 times in singlicate per assay on five different
days. The results are shown in Table 1. Coefficients of variation (%CV) within and
between day (Bookbinder and Panosian, 1986) were less than 10% and 15% respectively.
In another precision study, ten samples of two soils were weighed on a balance or
measured by packed volume in the soil collector. The samples were then extracted and
diluted (as described in the Methods Section), followed by assaying in duplicate in one
assay. Results are shown in Table 2. The overall coefficient of variation for PAHs
measurement using components of the Soil Collection and the PAHs Soil Extraction Kit
with analysis by the PAHs RaPID Assay® was determined to be less than 18% in both
cases.
Table 3 and Figure 3, summarize the cross-reactivity data of the PAHs RaPID Assay for
various polynuclear aromatic hydrocarbons and petroleum products. The percent cross-
reactivity was determined as the amount of analog required to achieve 50% B/Bo. The
broad specificity of the antibody used, allows for the detection of a majority of the PAHs.
Many non-structurally related organic compounds demonstrated no reactivity at
concentrations up to 10,000 ppb (data not shown).
Table 4 summarize the accuracy of the PAHs RaPID Assay in soil samples. Thirteen
different soil types were fortified with phenanthrene at 1 ppm. The samples were
extracted and diluted as described above, followed by analysis in the immunoassay. The
average recovery of phenanthrene in the samples was 108% with one sample (Alkali Lake)
given higher recoveries; the reason for the higher recovery on that sample is currently
under investigation. To demonstrate the detection of other PAHs in soil, anthracene,
benzo[a]pyrene, chrysene, fluoranthene, and pyrene, were spiked into four soils at 1 ppm;
recoveries of those PAHs (data not shown) agreed closely with the predicted response
based on the previously reported cross-reactivity data for the assay.
Correlation of twenty five samples, including both field contaminated soils and analytically
spiked soils analyzed by the ELISA method (y) and HPLC EPA Method 8310 (x) is
illustrated in Figure 4. The regression analysis yields a correlation of 0.931 and a slope of
2.02 between methods.
SUMMARY
This work describes a magnetic particle-based ELISA for the detection of PAHs and its
performance characteristics using soil samples. The assay compares favorably to HPLC
determinations, is faster, and eliminates the need for expensive instrumentation and solvent
disposal. The ELISA exhibits good precision and accuracy which can provide consistent
monitoring of environmental samples. Using this ELISA, fifty (50) results from soil
554
-------
monitoring of environmental samples. Using this ELISA, fifty (50) results from soil
samples can be obtained in less than two hours without the variability encountered with
antibody coated tubes and microtiter plates (e.g. coating variability, antibody leaching,
etc.). This system is ideally suited for adaptation to on-site monitoring of PAHs in water,
soil, and solid waste samples.
555
-------
REFERENCES
Bookbinder, M.J.; Panosian, K.J. Correct and Incorrect Estimation of Within-day and
Between-day Variation. Clin.Chem. 1986, 32, 1734-1737.
Engvall, B. Enzyme Immunoassay ELISA and EMIT. In Methods in Enzymology: Van
Vunakis, H.; Langone, J.J., Eds.; Academic Press: New York, 1980; pp 419-439.
Hammock, B.D.; Mumma, R.O. Potential of Immunochemical Technology for Pesticide
Analysis. In Pesticide Identification at the Residue Level, Gould, R.F., Ed.; ACS
Symposium Series, Vol. 136; American Chemical Society: Washington, DC, 1980; pp
321-352.
Harrison, R.O.; Braun, A.L.; Gee, S.J.; O'Brien, D.J.; Hammock, B.D. Evaluation of an
Enzyme-Linked Immunosorbent Assay (ELISA) for the Direct Analysis of Molinate
(Odram®) in Rice Field Water. Food & Agricultural Immunology 1989, 7, 37-51.
Itak, J.A.; Olson, E.G.; Fleeker, J.R.; Herzog, D.P. Validation of a Paramagentic Particle-
Based ELISA for the Quantitative Determination of Carbaryl in Water. Bulletin of
Environmental Contamination and Toxicology 1993, 57, in press.
Itak, J.A., Selisker, M.Y., Herzog, D.P. Development and Evaluation of a Magnetic
Particle Based Enzyme Immunoassay for Aldicarb, Aldicarb Sulfone and Aldicarb
Sulfoxide. Chemosphere 1992, 24, 11-21.
Lawruk, T.S.; Lachman, C.E.; Jourdan, S.W.; Fleeker, J.R.; Herzog, D.P.; Rubio, P.M.
Quantification of Cyanazine in Water and Soil by a Magnetic Particle-Based ELISA. J.
Agric. FoodChem. 1993, 41(5), 747-752.
Lawruk, T.S., Hottenstein, C.S.; Herzog, D.P.; Rubio, F.M. Quantification of Alachlor in
Water by a Novel Magnetic Particle-Based ELISA. Bulletin of Environmental
Contamination and Toxicology 1992, 48, 643-650.
Midgley, A.R.; Niswender, G.D.; Rebar, R.W. Principles for the Assessment of Reliability
of Radioimmunoassay Methods (Precision, Accuracy, Sensitivity, Specificity). Acta
Endocrinologica. 1969, 63 163-179.
Rubio, P.M.; Itak, J.A.; Scutellaro, A.M.; Selisker, M.Y.; Herzog, D.P.; Performance
Characteristics of a Novel Magnetic Particle-Based Enzyme-Linked Immunosorbent Assay
for the Quantitative Analysis of Atrazine and Related Triazines in Water Samples. Food &
Agricultural Immunology 1991, 3, 113-125.
556
-------
USEPA (1985), Evaluation and Estimation of Potential Carcinogenic Risks of Polynuclear
Aromatic Hydrocarbons; Carcinogen Assessment Group, Office of Health and
Environmental Assessment. Office of Research and Development. U.S. Environmental
Protection Agency: Washington, DC.
USEPA (1991), National Primary Drinking Water Regulations; Final Rule, Federal
Register, 40 CFR Parts 141-143, Vol. 56, No. 20, Jan. 30,1991.
Weston, PD, Avrameas, S.; Proteins coupled to polyacrylamide beads using
glutaraldehyde. Biochem. Biophys. Res. Comnnm. 1971,45, 1574-1580.
557
-------
Table 1
Precision of PAHs Measurement in Diluent
Pool Number
Replicates
Days
N
Mean (ppb)
% CV (within)
% CV (between)
Pooll
5
5
25
5.48
9.2
12.5
Pool 2
5
5
25
8.67
7.2
14.5
Pool 3
5
5
25
21.98
5.6
13.7
Pool 4
5
5
25
42.08
5.5
10.9
Table 2
Precision of PAHs Measurement in Soil
Soil:
Sample Collection Method
Replicates
Mean (ppm)
% CV (total)
Wisconsin
Joshua Tree
weight
10
1.57
17.6
volume
10
1.18
17.8
weight
10
1.43
14.3
volume
10
1.26
14.4
558
-------
Table 4
Phenanthrene Recovery From Different Soil Types
Soil
Sample
Alkali Lake
Beardon
Holland
Joshua Tree
Levittown
Muck
Munin
Piscataway
Sagamore
Sharkey
Tennessee
Wiscosin
Virginia
Mean
SD
%CV
Neat
(ppm)
0.50
0.42
2.30
0.07
4.70
2.90
0.17
2.69
4.40
0.68
0.30
0.72
1.54
Total
(ppm)
2.19
1.45
3.22
1.28
5.82
3.94
1.23
3.67
5.56
1.92
1.29
1.80
2.64
Recovered
(ppm)
1.69
1.03
0.92
1.21
1.12
1.04
1.06
0.98
1.16
1.24
0.99
1.08
1.10
1.12
0.19
17.1
Recovery
£%}
169
103
92
121
112
104
106
98
116
124
99
108
110
108
Soil Type
loamy sand
clay loam
silt loam
sand
silt loam
organic potting
clay loam
sandy loam
silty clay loam
clay loam
sandy loam
loam
loamy sand
559
-------
Table 3
Specificity (Cross-Reactivity)
Compound
Phenanthrene
Anthracene
Fluoranthene
Chrysene
Pyrene
Benzo(a)Pyrene
Fluorene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
l,12Benzoperylene
1,2:5,6 Dibenzanthracene
Indeno (l,2,3-c,d)pyrene
Naphthalene
Acenaphthalene
Acenaphthylene
1.2 Benzanthracene
1 -Methylnaphthalene
2-Methylnaphthalene
Aroclor 1242
Aroclor 1248
Aroclor 1254
Aroclor 1260
Benzene
Toluene
Phthalate
Pentachlorophenol
Biphenyl
CCA
Creosote
Fuel Oil #6
Gulf Diesel Fuel
Sunoco Home Heating Fuel
Kerosine
Jet A Fuel
Regular Gasoline
Premium Gasoline
90% B/Bo
LDD (nnb)
0.70
0.54
0.32
0.40
0.70
0.50
1.65
0.91
0.77
14.7
25.7
0.78
65
12.9
10
0.77
28.2
28.2
37.5
41.0
> 10000
> 10000
>10000
> 10000
> 10000
340
15.9
> 10000
1.1
5
19.6
12.8
1250
> 10000
1000
597
50% B/Bo
ED50 (ppb)
16.5
12.5
4.7
7.8
15.1
6.9
35.2
54.2
524
>1000
>1000
27.2
>1000
688
447
28.4
1330
802
1450
5330
> 10000
>10000
> 10000
>10000
> 10000
> 10000
703
>10000
16.6
53.7
497
292
> 10000
> 10000
>10000
> 10000
% Cross
Reactivity
100
132
351
212
214
239
47
30
3
<2
<2
61
<2
2
4
58
1.2
2.1
1.8
0.4
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
2.8
<0.2
117
31
3
6
<0.1
<0.1
<0.1
<0.1
560
-------
Figure 1. Diagram of soil collector used to collect and extract soil samples.
Figure 2. PAHs RaPID Assay dose response curve. Each point represents the mean
of 30 determinations. Vertical bars indicate +/- 2 SD about the mean.
Figure 3. Specificity of the PAHs RaPID Assay against selected polynuclear
aromatic hydrocarbons and creosote.
561
-------
SOIL SAMPLE COLLECTION DEVICE
01
O)
ro
Soil Collector
BOTTOM -J
1
( J
"•
i
\
i ur
-<
«JVy
1 CVV
^1
•
.
oa (^
T1
Luer Cap
Plunger Rod Plunger
Luer Cone
Figure 1. Diagram of soil collector used to collect and extract soil samples.
-------
97
95
90 H
80
o
59 70
CO
| 60
o5 50
Q.
40
30
20 H
10
7
10
Phenanthrene (ppb)
100
Figure 2. PAHs RaPID Assay dose response curve. Each point represents the mean
of 30 determinations. Vertical bars indicate +/- 2 SD about the mean.
563
-------
CD
97
95
90 -
80 -
70 -
60 -
50
40
30
20 H
10
7
10
PAH (ppb)
Phenanthrene
Anthracene
Chrysene
Fluoranthene
Pyrene
Benzo(a)pyrene
Creosote
100
Figure 3. Specificity of the PAHs RaPID Assay against selected polynuclear
aromatic hydrocarbons and creosote.
564
-------
0 5 10 15 20 25
PAH HPLC Method 8310 (ppm)
30
Figure 4. Correlation between PAH's concentrations as determined by the ELISA
and HPLC Method 8310 in soil samples, n = 25, r = 0.931, y = 2.02x + 1.55.
565
-------
87
LABORATORY EVALUATION OF IMMUNOASSAY PCB TESTS
Mark L. Bruce and Raymond M. Risden, Enseco, 4101 Shuffel Dr. NW, North Canton,
OH 44720, Randolphe P. Swenson Jr., Enseco, 5251 DTC Pkwy Suite 415, Englewood,
CO 80111, Kevin R. Carter, EnSys Environmental Products Inc., Box 14063, Research
Triangle Park, NC 27709.
ABSTRACT
Immunoassay (IA) results were generally a factor of two higher than the GC-ECD results
for the same PCB extract for the 41 samples studied. The manual assay procedure used
for this study should be accurate enough to predict proper dilution levels for final GC-
ECD analysis. The manual solvent exchange and assay requires less than 4 minutes of
labor per sample. Equipment cost for the assay is minimal. The preliminary cost
estimates indicate that immunoassay should be considered as a viable alternative to gas
chromatography for internal PCB screening.
INTRODUCTION
Environmental concerns and economic forces are exerting pressure on the environmental
analysis market to reduce turn-around times and costs. At the same time, some data users
have realized that traditional QA/QC requirements are not necessary for all
environmental decisions. One area ripe for streamlining time and cost is laboratory
sample screening.
Several companies have adapted immunoassay (IA) to test environmental pollutants in
order to bring lA's advantages to environmental analysis. IA has been widely used in the
health sciences field where it has proven to be a low cost, fast turn-around, high capacity
analysis technology.
These same characteristics make it very attractive for environmental testing.
Immunoassay does not provide data that is identical to the traditional GC, LC and
GC/MS tests. The specificity of IA should make it less susceptible to the interferences
that limit chromatographic analyses. However, the biochemical nature of IA makes it
sensitive to new interferences. The QA/QC data normally available from IA includes
replicates and matrix spikes, but not internal standards or surrogates. This QA/QC
reduction may not be acceptable to all data users. The IA response for multicomponent
analytes such as PCBs, TPHs and PAHs changes as the composition of the mixture
varies. Thus, the analytical accuracy for an unknown mixture will be reduced. The assay
does not identify the components in the mixture. Some data users may view this reduced
amount of information as lower quality data, but it should be more than adequate for
many environmental decisions.
The current project focused on shortening sample turn-around time and reducing cost in
the analysis of PCBs. We examined the IA PCB test as a replacement for GC in
screening samples to determine the proper extract dilution for final analysis. This would
be most useful to a lab that currently performs a high percentage of dilutions. This use
for IA would be relatively fast to implement since no external approval would be
required. LA could also be used for the final analysis under certain circumstances. In this
case IA would replace a corresponding GC test. This use would produce the greatest time
and cost savings. However, it would also require regulatory and client approval.
566
-------
We examined the PCB results for about 10,000 samples analyzed at Enseco-North
Canton. Most frequently no Aroclors were detected. About 16% of the time dilutions
were required to bring the Aroclor concentration into the calibration range of the GC-
ECD. Also, some samples reported as non-detect were diluted because of interferences
Other Enseco facilities estimate 20-30% dilution rates. GC instrument maintenance tends
to increase if highly concentrated samples are injected. Thus quick, effective and
inexpensive screening is needed to preserve the health of the GC used for final analysis.
0.4
0.2
o
o
o
o"
o
o
o
o
o"
o
o
Initial Extract Concentration (|ug/mL)
Figure 1. Distribution of Aroclor Concentrations.
The frequency of various Aroclor types and the IA sensitivity differences indicated that
calibration with Aroclor 1248 was the best compromise. Samples analyzed in North
Canton contained Aroclors 1242 and 1254 most frequently (Figure 2). According to
EnSys, immunoassay response (sensitivity) with their kit varies as the Aroclor mixture
changes. Table 1 has been normalized to 1248.
Table 1. Aroclor Response Differences
Aroclor Relative Response t
1016
1232
1242
1248
1254
1260
0.25
0.25
0.5
1.0
2.5
2.5
EXPERIMENTAL
t Higher number means more sensitive
All samples were extracted with 3500 series SW-846 sample preparation techniques. The
extracts were cleaned with 3600 series clean-up techniques prior to gas chromatographic
analysis by method 8080 or 8081. These same extracts were then solvent exchanged
from hexane to methanol and analyzed with immunoassay supplies developed by EnSys
567
-------
for their field PCB kit. The EnSys PCB RIS£® test is a chromogenic enzyme-linked
immunosorbent assay (ELISA). The monoclonal antibody was produced to respond (i.e.,
bind) to a particular subset of the pentachlorobiphenyl congeners. As the
pentachlorobiphenyl composition varies from one Aroclor mixture to the next, the
sensitivity of the IA method changes. This is analogous to quantitating a PCB
chromatogram using only the pentachlorobiphenyl peaks. Those Aroclors that have a
high percentage of pentachlorobiphenyl have the best sensitivity.
60%
(0
_ 50% - -
I 40% - -
•3 30%--
C 20% - -
o
g 10%-.
Q.
0% -
0.2
0.2
CD
O
CM
CM
Aroclor
Figure 2. Frequency of Various Aroclors
The EnSys PCB kit uses an antibody immobilized on polystyrene tubes. This technique
is a semicompetitive immunoassay where the antibodies are exposed to sample analytes.
Later, enzyme conjugate is added to the same solution. This procedure summarized
below was based on EnSys experience and chosen as starting point to determine the best
compromise between analysis speed, accuracy, precision and cost. This process is
completely manual using standard syringes and repeating single channel pipettes.
1) Set up buffer and antibody tubes in rows of 6. Do not exceed 5 rows.
2) Transfer 30 |iL of standard, blank or sample extract into the buffer tube. Mix.
3) Transfer buffer tube contents to the antibody tubes. Mix and wait 10 minutes.
4) Dispense enzyme conjugate, mix and wait 5 minutes.
5) Wash antibody tubes thoroughly with dilute buffered detergent solution.
6) Add substrate A to antibody tubes.
7) Add substrate B to antibody tubes, mix and wait 2.5 minutes.
8) Add stop solution to antibody tubes.
9) Measure and record absorbance at 450 nm.
The sample extracts were exchanged to methanol (Fisher Optima grade or Burdick &
Jackson Purge and Trap grade) by two different procedures. Early sample batches were
exchanged from hexane to methanol using K-D concentration/exchange followed by
nitrogen blowdown as described in SW-846 3500 series methods. Later extracts used a
568
-------
simplified exchange procedure which was acceptable for high boiling analytes such as
PCBs. Hexane extract (200 [iL) was evaporated to dryness in a microvial using a gentle
stream of nitrogen. The PCBs were then redissolved in 200 |iL of methanol. Dilutions
were prepared from this exchanged extract as needed. The samples were assayed singly
in the same batch as replicate standards and blanks.
The standard SW-846 solvent exchange was time and labor consuming. The simplified
exchange was adopted to reduce solvent requirements, accelerate the process and reduce
labor. The evaporate to dryness exchange procedure showed no significant PCB losses
when several 1248 standards were evaporated to dryness and redissolved in methanol.
RESULTS AND DISCUSSION
Calibration
The routine immunoassay calibration procedure consisted of replicate analyses of 3
Aroclor 1248 standards and a blank. Absorbance of the standard was plotted against the
logio of the standard concentration. This semilogarithmic calibration produced a straight
line with a negative slope. Many traditional immunological calibration procedures plot
% B/Bo versus logio (concentration). The absorbance of the standard (or sample) is
divided by the absorbance of the blank and expressed as a percentage (% B/Bo). Since
this scaling process does not affect the final calculated result, we decided to use the
original absorbance readings, which are more familiar to environmental lab personnel.
A duplicate six point calibration from 20 to 400 ng/mL was constructed for Aroclor 1248.
The semilog calibration was linear throughout the majority of the range with only a small
amount of curvature at the extremes. The calibration equation was;
absorbance = 3.22 - 1.11 • log (concentration)
The correlation coefficient (r) of the least squares line was 0.976. Subsequent
calibrations consisted of duplicate measurements at three concentration levels (20, 70,
200 ng/mL). The calibration line equations were;
absorbance = 3.19 - 1.04 • log (concentration) r = 0.993
absorbance = 2.74 - 0.92 • log (concentration) r = 0.979
absorbance = 2.52 - 0.89 • log (concentration) r = 0.972
The average deviation between replicate assays of standards was 17%.
Samples
Soil sample extracts (41) were solvent exchanged as described above. Each extract was
assayed once in conjunction with duplicate assays of three calibration standards and a
blank. The results from the immunoassay are compared with the corresponding GC-ECD
results in Table 2. Figure 4 graphs the same data in a log-log plot. The diagonal line
represents 1 to 1 correspondence between the IA and GC results. The shaded region of
the graph covers IA/GC ratios between 2:1 and 1:2. Typically the IA concentration
estimate was about two times higher than that determined by GC-ECD. The high 1254
IA concentration was expected because of the sensitivity difference between the calibrant
(1248) and sample analyte (1254). Refer to Table 1. The 1248 IA results were higher
than expected in some extracts. This could have been caused by differences in the
pentachlorobiphenyl composition of the calibrant and native Aroclor in the samples.
Also, there may be interferences in the sample extracts which resulted in the occasional
positive bias. Small errors in the extract solvent exchange process may account for part
569
-------
of the difference, as well. Four extracts which had no PCBs detected by GC-ECD were
analyzed by IA. These waste dilution extracts contained many chromatographic
interferences. The IA results were all below the GC reporting limit. Thus, the IA showed
no tendency toward false positives for these samples.
2 -r
1.5 --
0)
o
c
CO
•e 1 +
o
a>
(0
0.5 --
4-
4-
1.00
1.50
2.50
3.00
2.00
log (ng/mL)
Figure 3. Aroclor 1248 Calibration.
Despite the differences between the GC-ECD and immunoassay results, the IA data were
generally accurate enough to predict the proper dilution level for final GC-ECD analysis.
Since the IA data were sometimes much higher than the GC data this would occasionally
lead to over diluting the extract for final analysis. Sporadic over dilution is preferable to
under dilution since GC reliability and maintenance should not be effected.
This manual immunoassay procedure was sensitive to analyst technique. Raw data often
did not compare well from person to person. However, if the calibrants were prepared in
the same batch as the samples by the same person then acceptable accuracy was achieved.
Data reproducibility was very dependent on the consistency of the analyst. Partial or full
automation should improve precision significantly.
Screening a batch of 16 samples by GC-ECD required about 8 hours including 45
minutes of labor to prepare the extracts, interpret the data and calculate the results. IA
required about one hour to completely process a batch of 16 samples. Since this form of
IA was completely manual the labor time was also one hour. The capital requirements
were much lower for IA since the only instrument required was a small
spectrophotometer. A detailed cost analysis has not been completed yet, but the initial
estimates indicate that IA should be quicker and less costly than GC for extract screening.
570
-------
Table 2. Comparison of GC-ECD and Immunoassay Data.
Extract #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Aroclor
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1248
1254
1254
1254
1254
1254
1254
1254
1254
1254
1254
1254
1254
1254
1254
NDbyGC
NDbyGC
NDbyGC
NDbyGC
GC cone
(mg/kg or L)
34
1.4
14.0
1.4
2.3
94.0
3.0
1400.0
2.3
20.6
29.7
2.2
7.0
948.3
1.3
0.4
528.8
5.5
27.7
5.6
3167.8
981.4
19.9
20.0
14.0
7.7
1.6
15.0
7.1
8.1
9.8
1.0
1.3
3.9
14.0
0.00024
0.00022
<0.5
<0.5
<0.5
<0.5
IA cone
(mg/kg or L)
80
4.3
30.2
2.9
2.6
103.9
7.9
1391.6
1.9
11.8
31.2
2.4
12.1
2905.1
6.5
1.3
1777.8
7.6
66.6
6.3
4181.4
1483.6
7.8
60.0
57.4
22.6
2.5
20.2
13.3
13.0
31.2
0.9
1.6
6.3
17.1
0.00065
0.00048
0.4
0.1
0.1
0.1
Average Ratio =
IA/GC
ratio
2.4
3.1
2.2
2.1
1.1
1.1
2.6
1.0
0.8
0.6
1.1
1.1
1.7
3.1
5.0
3.0
3.4
1.4
2.4
1.1
1.3
1.5
0.4
3.0
4.1
2.9
1.5
1.3
1.9
1.6
3.2
0.9
1.2
1.6
1.2
2.7
2.2
2.0
571
-------
O)
E
L-
o
O)
O)
E
\—'
re
re
o
c
I
E
10,000 -r-
1,000 --
100 --
IA/GC=2
IA/GC=1
IA/GC=0.5
+ PCB-1248
PCB-1254
GC-ND
0.001 ••
0.0001 --
0.0001 0.001 0.1 1 10 100 1,000 10,000
GC-ECD (mg/kg or mg/L)
Figure 4 Comparison of Immunoassay Results to GC-ECD Results.
CONCLUSION
Immunoassay results were generally within a factor of two of the GC-ECD results for the
same extract. This indicates that IA should be accurate enough to predict proper dilution
levels for final GC-ECD analysis. The manual exchange and assay requires less than 4
minutes of labor per sample. Equipment cost for the assay is minimal. The preliminary
cost estimates indicate that immunoassay should be considered as a viable alternative to
gas chromatography for internal PCB screening.
ACKNOWLEDGMENT
The authors would like to thank the following people that provided valuable support
during this project: EnSys Rhonda Mudd, Karen McKenzie, Alan Staple, MetPath
Roger Juselius, Donna Puskar, Maureen Humes, Enseco Paul Winkler, Dennis Flynn,
Dan Segal, John Flaherty, Brad Belding, Mike Orbanosky, Richard Burrows, Phil
Dufresne, Lou Mancini, Paul Sharkey, Kelly Evans and Colleen Davis.
572
-------
DETERMINATION OF TRICHLOROETHYLENE BY IMMUNOASSAY
Titan S. Fan, ImmunoSystems Inc. Millipore Corporation, Scarborough, ME 04074;
Barbara Young, Millipore Corporation, Bedford, MA 01730; Robert E. Carlson,
ECOCHEM Research Inc. Chaska, MN 55318-1043.
ABSTRACT
A competitive enzyme immunoassay (EIA) has been developed for the determination of
trichloroethylene CTCE). TCE derivatives were attached to carrier proteins, and these
immunogens were used to raise anti-TCE antisera in rabbits. The antisera were used in
combination with enzyme labelled heterologous TCE derivatives to develop an EIA that
is specific for TCE. Selected antisera and enzyme conjugates have been incorporated
into an assay that is similar to our previous PCB, PAH, and Petroleum Fuels test kits for
the field analysis of soil, water , and other matrices. Preliminary studies indicate good
specificity for TCE. The assay has a working sensitivity range of ppb to ppm, which is
suitable for the analysis of TCE in soil and water.
573
-------
89
IMPORTANT FACTORS IN ENHANCING SUPERCRITICAL FLUID EXTRACTION
EFFICIENCIES FOR ENVIRONMENTAL APPLICATIONS
Dr. Joseph M. Lew, Vice President of Technology and Customer Support; Lori A.
Dolata, Technical Specialist; Robert M. Ravey, Application Chemist; Victor Danielson,
Application Chemist; Anita Cardamone, Application Chemist; Suprex Corporation, 125
William Pitt Way, Pittsburgh, PA 15238
ABSTRACT
Supercritical fluid extraction (SFE) has a broad range of applicability, especially with
regards to environmental matrices. SFE has achieved a significant amount of attention
due to the benefits of eliminating toxic, liquid solvent usage, reduction in sample
preparation time and an increase in the overall analytical reliability of determinations.
SFE/GC-MS, GC-FID and IR are powerful techniques to accurately analyze and quanti-
tate environmental analytes in soils and other solid wastes. The off-line transfer of SFE
effluents to collection vials adds a considerable amount of flexibility in characterizing
complex matrices since a full complement of analytical tools can be used (ie, GC, LC,
IR, NMR and UV). Moreover, the advantages of SFE can be further augmented by the
use of sequential automation for greater sample throughput for large volume sample
preparation, statistical replicate analysis, and optimization of SFE operational
parameters, which can be especially important for environmental applications.
Examples will be presented in this paper showing the use of SFE/GC-MS, GC-FID and
IR methodologies for the quantitative determination of different target analytes in
environmental matrices, such as polynuclear aromatic hydrocarbons (PAH), and total
petroleum hydrocarbons (TPH) in soil.
INTRODUCTION
Before most chemical analyses can be performed, some type of extraction is usually
required to remove suspected environmental contaminants from their particular environ-
ment matrix (ie, soil) and often this is a time-consuming and costly step. One of the
most common extraction methods used in the US Environmental Protection Agency is
SW 846 which often requires long extraction times (up to 24 hours for Soxhlet methods),
large quantities of toxic organic solvents, (ie, Freon, methylene chloride) and lengthy
solvent concentration or clean-up procedures. Historically, sample preparation involves
as much as 75% of the total analysis time. Although much information has been
published on SFE (1-15), little guidance is available for analysts who want to incorporate
this powerful sample preparation technology into quantitative analytical methods. This
includes all segments of the extraction process (ie, analyte removal, analyte transfer, and
analyte collection). The fact that SFE with carbon dioxide (CO2) has been touted as the
solution to the organic solvent waste disposal problem, and that SFE is fast makes this
methodology attractive for the analytical environmental laboratory of the 21st century.
The successful collection of extracted analytes after SFE affords tremendous analytical
flexibility in terms of the ability to collect into media that are suitable for subsequent
574
-------
chromatographic or spectroscopic characterization. In some cases this can be done
without any additional liquid solvent where the supercritical CO2 vaporizes and only
collected target analytes remain.
Supercritical fluids possess favorable physical properties that are intermediate between
those of the gas and liquid states. These physical properties (high diffusion coefficients,
low viscosities, high densities and zero surface tension) result in the rapid and efficient
mass transfer of target analytes from the sample matrix into the extraction solvent.
Liquids have higher surface tension, which often inhibit the timely penetration of liquid
solvent molecules into the pores of a heterogeneous sample matrix (ie, soils). For this
reason, Soxhlet extractions of soils for example, often need to be performed for many
hours (8-24 hours) to allow for the physical penetration of a liquid solvent (ie, methylene
chloride) into a typical porous matrix (ie, soil). Supercritical fluids, on the other hand,
have no surface tension limitations which allow for a rapid penetration (minutes and not
hours) of a sample matrix to achieve equal to or very often greater than the extraction
efficiencies that a liquid solvent could achieve.
EXPERIMENTAL
A stand-alone manual supercritical fluid extraction system (PrepMaster®, Suprex;
Pittsburgh, PA), which has been described earlier (15), and a sequentially automated
supercritical fluid extraction system (AutoPrep-44™, Suprex; Pittsburgh, PA) were used
for all of the extraction experiments. Extractions were accomplished with various sizes
of extraction vessels (0.5 to 10 mL) using operating conditions that are listed in the text.
An off-line collection module, (AccuTrap®, Suprex) was used to perform the cryogenic
solid phase trap collection (6,7) for both the manual and automated systems. This
collection module included the following components: a VariFlow™ (Suprex) automati-
cally variable restrictor (kept at 50°C) which controlled the CC>2 flow rates (from 1 to 7
mL/minute compressed), a temperature-controlled solid phase trap packed with unibeads
and CIS modified silica for analyte trapping, a liquid pump for delivering an appropriate
wash solvent for analyte desorption, and a solenoid valve for delivering a stream of
nitrogen to purge the adsorbent trap and connecting tubing after desorpton.
RESULTS AND DISCUSSIONS
SFE has been successfully applied for the extraction of environmental target analytes in
soil, sand, silt, clay, fly ash, sludge, petroleum waster, river sediment, urban dust, diesel
exhaust particulate, tar pitch, fish, grain and vegetable matrices. The target analytes that
have been successfully extracted include the following general categories:
Polynuclear aromatic hydrocarbons (PAH)
Total petroleum hydrocarbons (TPH)
Pesticides
Polychlorinated biphenyls (PCBs)
Polychlorinated dibenzofurans (PCDFs)
Polychlorinated dibenzo-p-dioxins (PCDDs)
575
-------
In order to perform successful extractions of these envrionmental categories from "wet"
soils (>1% water), various approaches as shown in Table 1 have been employed to either
remove or immobilize water.
Table 1: Approaches to Performing SFE of Wet Soils
Adsorbent use
Pre-heating, freeze-drying, or air drying
Pre-extraction at low densities
Secondary effluent adsorbent filter
Many naturally incurred soils contain water which historically has been problematic in
SFE due to the freezing of water in the restrictor tip after undergoing expansive cooling.
This ultimately has resulted in plugging of the restrictor or have caused problems in
maintaining extraction flow rates after vessel frits became blocked. Dehydration
techniques have been time consuming and could promote the loss of volatile or semi-
volatile analytes. Adsorbents have been used successfully to retain water such as hydro-
matrix (diatomaceous earth), magnesium sulfate, sodium sulfate, calcium chloride,
molecular sieve and silica inside the extraction vessel (16). Moreover, new developments
in restrictor technology have further increased the likelihood of extracting wet soils.
Specifically, utilizing the Variflow™ automatically variable restrictor with or without
adsorbents (hydromatrix), soils with TPH (total petroleum hydrocarbons) and PAH
(polynuclear aromatic hydrocarbons) contamination and 20 to 50% water content have
been efficiently extracted using the sequentially automated SFE without flow inhibition
problems. Historically, these sample types would cause restrictor plugging for
conventional linear, fixed restrictors due to the freezing of the water at the point of
decompression. Moreover, results were difficult to quantitatively replicate between lab
sites since the restrictor operation was dependent on the operator's dexterity or level of
experience. Analytical gas chromatographic results highlighting PAH and TPH
contaminated native wet soil extractions are shown in Figures 1-3. Each of these results
demonstrates the ability to load the sample as received in the extraction vessel without
any sample pretreatment when utilizing the automatic variable restrictor. Thermally,
both the restrictor and the solid phase trap were maintained at 100°C to keep the water as
a vapor after decompression. The excess water was collected in an additional wash vial.
Moreover, for each of these soils, analytical precision of < 5% RSD was demonstrated
when performing replicate extractions on the automated SFE (16 replicates for the PAH
contaminated soil and 24 replicates for the TPH contaminated clay).
Recent results have indicated that specific solid phases for off-line trapping can distinctly
affect overall SFE efficiencies. Table 2, for example, shows the comparison between
CIS modified silica and silanized glass bead solid phase traps used for the off-line SFE
collection of extracted TPH from soils. When compared to the reference value after IR
analysis, the CIS trap provided the closest agreement for the effective collection and
desorption of the diesel range organics.
576
-------
Table 2: Automated SFE/IR of TPH in Soils*
(CIS Modified Silica vs. Glass Beads Solid Phase Traps)
Replicates
1
2
3
4
5
6
7
8
9
10
Concentration (ppm)
CJ8 Glass Beads
408
441
451
435
472
455
435
432
464
477
Average: 447
Standard Deviation: 19.9
% RSD: 4.45% 5.
*EPA Certified soil without fatty acids. Target value of 450 ppm (diesel
organics).
280
293
291
272
111
278
287
251
255
262
275
13.8
02%
range
The SFE operating conditions (modified from EPA draft method 3560) that were used
for this study, are summarized in Table 3.
Table 3: Operating Conditions for the Automated SFE of TPH in Soils
Extraction Conditions
Pressure: 450 atm
Temperature: 120°C
Modes: 5 min static, 30 min Dynamic
Restrictor: VariFlow at 50°C for <10% F^O, 100°C for
>10%H20
Restrictor Flow: 3.0 ml/min Compressed CO2
Extraction Vessel: 3 ml vessel
Adsorbent/Dispersent: Hydromatrix
Collection Conditions:
Collection Temperature: -10°C (<10% H20), 100°C (>10% H20)
Desorption Temperature: 40°C
Wash Solvent: Freon, 3 mis
A further assessment of method precision using automated SFE with IR analysis was
done with twenty four replicate extractions on an EPA certified soil. The results are
listed in Table 4. Using IR analysis, relative standard deviations averaged at 13.3% with
an average concentration of 228 ppm (compared to the 230 ppm target value).
577
-------
Table 4: Automated SFE/IR of TPH in Soils*
Replicates
1
2
3
4
5
6
7
8
9
10
11
12
*EPA Certified
organics).
Cone, (ppm)
293
256
208
255
256
266
241
245
276
246
248
248
Standard
soil without fatty acids. Target
Replicates Cone, (ppm)
13
14
15
16
17
18
19
20
21
22
23
24
Average:
Deviation:
% RSD:
value of 230 ppm (diesel
225
211
195
183
203
182
211
201
202
205
197
207
228
30.3
13.3%
range
CONCLUSION
Out of the numerous application areas for using SFE as a sample preparation technique,
the extraction of environmental pollutants out of soils is perhaps one of the most straight-
forward. This is partially due to the inherent porous nature and loose texture of soils and
the fact that supercritical fluid technology has evolved to the extent that instrumental
enhancements have provided increased capabilities for researchers to use in method
development. SFE technology today provides environmental chemists with the
opportunity to utilize new and automated sample preparation techniques to open up
virtually new frontiers. This includes the use of environmentally friendly, politically
acceptable solvents (ie, CO2), the decrease of sample turnaround times, the elimination
of toxic chemical exposure, and the elimination of solvent disposal costs. Moreover,
SFE can improve the overall reliability of analytical results and establish new standards
for efficient monitoring of the environment. With the key development of variable
restriction and with automation, SFE has reached a new pinnacle for economic
application of this technique.
REFERENCES
1. Hawthorne, S.B., Miller, D.J., J. Chromatogr. Sci., 24, 258-64 (1986).
2. Hawthorne, S.B., Krieger, M.S., Miller D.J., Anal. Chem., 61, 736-40 (1989).
3. Hawthorne, S.B., Krieger, M.S., Miller D.J., Anal. Chem., 60, 472-77 (1988).
4. Levy, J.M., Rosselli A.C, Chromatographia., 28, issue 11/12 (1989).
578
-------
5. Wright, B.W., Frye, S.R., McMinn, D.G., Smith, R.D.. Anal. Chem., 59, 640-44
(1987).
6. Levy, J.M., Houck, R.K., American Laboratory, 36R-36Y (1993).
7. Levy, J.M., Dolata, L.A., Ravey, R.M., J. of Chromatogr. Sci., 31, 349-52 (1993).
8. Levy, J.M., Dolata, L.A., Ravey, R.M., Storozynsky, E. Holowczak, K.A., J. of
High Resolution Chromatography, 16, 368-71 (1993).
9. Ashraf-Khorassani, M., Houck, R.K., Levy, J.M., J. of Chromatogr. Sci., 30, 361-
66 (1992).
10. Hawthorne, S.B., Miller, D.J., Krieger, M.S., J. Chromatogr. Sci., 27, 347-54
(1989).
11. Nielen, M.W.F., Sanderson, J.T., Frei, R.W., Brinkman, U.A.T.. J. Chromatogr.,
474, 388-95 (1989).
12. Hawthorne, S.B., Miller, D.J., Langenfeld, J.J., J. Chromatogr. Sci., 28, 2-8
(1990).
13. Sugiyama, K., Saito, M., Hondo, T., Senda, M., J. of Chromatogr., 332, 107-16
(1985).
14. Hawthorne, S.B., Anal. Chem., 62, (11), 633A-42A (1990).
15. Levy, J.M., Am. Lab., 23-32(1991).
16. Burford, M.D., Hawthorne, S.B., Miller, D.J., J. of Chromatogr., 657, 413-
27.(1993)
579
-------
CH2CI2
/
Sample Pretreatment:
Varlflow:
Trap Temperature:
Adsorbent Trap:
Sample Size:
Replicates:
Chrysene % BSD:
None
3 ml/minute
100'C
100'C
CIS/glass beads
3 grams
16
4.2
Inject
Figure 1: Automated Off-Line SFE/GC-FID Characterization of PAH Contaminated
Soil (20% water)
/
Sample Pretreatment:
Variflow:
Trap Temperature:
Adsorbent Trap:
Duration:
TPH Concentration:
None
3 ml/minute
100'C
100'C
CIS/glass beads
35 minutes
330 ppm
Inject
Figure 2: GC-FID Characterization of Automated Off-Line SFE of TPH Contaminated
Soil (43% water).
580
-------
•a
r>
is
oo
CD
•t
Inject
"W ':
r>: °
MJ
Sample Pretreatment:
Varlflow:
Trap Temperature:
Adsorbent Trap:
Sample Size:
Replicates:
C14 % RSD:
None
3 ml/minute
100'C
100'C
CIS/glass beads
1 gram
24
3.9
Figure 3: Automated Off-Line SFE/GC-FID Characterization of TPH Contaminated
Clay (36% water).
581
-------
90
Applications And Performance Of The D TECH™ TNT
Environmental Testing System
George B. Teaney. James M. Melby, and James W. Stave
Strategic Diagnostics Incorporated
128 Sandy Dr.
Newark, DE 19713
ABSTRACT
Field screening technologies have gained increased attention in recent years.
Immunoassay based field screening systems, designed to detect small molecular weight
priority pollutants, are relatively new to the environmental market, however their core
technology has been proven in the medical diagnostics industry since the early 1950's. In
order to be effective, field screening systems should be quick, easy to use, reliable, and
inexpensive. Immunoassay based field screening systems can provide all of these qualities
as well as a high degree of sensitivity and specificity towards die analyte of interest.
The D TECH TNT Environmental Testing System is an immunoassay based, self
contained, field screening system which employs a competitive enzyme immunoassay
format. Free TNT in a water or soil sample competes with an enzyme linked TNT analog
for binding sites on Anti-TNT antibody coated latex particles. Unbound materials are
separated from the latex by filtration and wash step, and an enzyme substrate is added to
the particles. The color that develops is inversely proportional to the concentration of TNT
in the assay. This color is determined using either the color card provided with the kit, or
the DETECHTOR™ reflectometer. This test has demonstrated 5 ppb and 0.2 ppm
sensitivity in waters and soils, respectively. All EPA SW-846 Method 8330 analytes were
tested for cross-reactivity in this assay. Four (4) of these compounds were found to be
marginally (3%) to moderately (30%) cross-reactive. The assay shows little interference
from non-explosive organic pollutants as well as nitrates and nitrites. Furthermore, the
assay's performance has been unaltered when tested in a wide range of soil and water
matrices. Field trial data show a low occurrence of false positives and false negatives, and
good correlation to Method 8330. The cost effectiveness of this system was demonstrated
during a 334 sample field trial in the western U.S.
INTRODUCTION
Field screening technologies have gained increased attention in recent years.
Immunoassay based field screening systems, designed to detect small molecular weight
priority pollutants, are relatively new to the environmental market, however their core
technology has been proven in the medical diagnostics industry since the early 1950's. In
order to be effective, field screening systems should be quick, easy to use, reliable, and
inexpensive. Immunoassay based field screening systems can provide all of these qualities
as well as a high degree of sensitivity and specificity towards the analyte of interest.
The D TECH TNT Environmental Testing System is an immunoassay based, self
contained, field screening system. This portable assay has been designed to quickly and
effectively identify TNT contaminated soils and waters during site characterization studies
and throughout the remediation process.
582
-------
This paper describes many of the performance characteristics of the D TECH TNT assay
and details some of its applications in the field.
METHODS
The D TECH TNT Environmental Testing System employs a competitive enzyme
immunoassay format (Figure 1). Free TNT in a water or soil sample competes with an
enzyme linked TNT analog for binding sites on Anti-TNT antibody coated latex particles.
A reference reaction is run simultaneously to the test reaction in which a known amount of
TNT reference material competes with the enzyme conjugate for antibody binding sites.
This reference reaction sets the minimum detection limit (MDL) of the assay and helps to
control for environmental influences on the assay. Unbound materials are separated from
the latex in the 2 reactions by a filtration and wash step, and an enzyme substrate is added
to the particles. The color that develops from the bound enzyme is inversely proportional
to the concentration of TNT in the assay. This color is determined using either the color
card provided with the kit, or the DETECHTOR™ reflectometer.
Water samples are simply diluted into an assay buffer and filtered into the assay. Sample
preparation for soil analysis involves pulling a 4.5 g sample using a volumetric coring
device. The soil sample is extracted in 100 % acetone for 3 minutes and the soil extract is
diluted into assay buffer. This diluted soil extract is then treated as if it were a water
sample.
RESULTS AND DISCUSSION
The assay can detect TNT concentrations as low as 5 ppb in water samples and 0.2 ppm in
soils. The assay dose response curve indicates an assay range of 5 ppb to 60 ppb in water
samples (Figure 2) and 0.2 ppm to 2.0 ppm in soils (data not shown). The test has been
evaluated for cross-reactivity with all 13 of the compounds listed in EPA SW-846 Method
8330^ (Table 1). A compound is considered cross-reactive if it can be detected at a
concentration 100 times the MDL of the target analyte. The four compounds, Tetryl,
1,3,5-trinitrobenzene, 2-amino-4,6-dinitrotoluene, and 2,4-dinitrotoluene, have been
defined as cross-reactants in the TNT kit. The degrees of cross-reactivity for these
compounds are 30%, 25%,17%, and 4% respectively (Figure 3). Because these cross-
reactants are common to sites contaminated with TNT, D TECH sample results are reported
as TNT equivalents.
Twenty three potential co-contaminants were screened as possible interfering substances in
the TNT assay (Table 2). All of the compounds listed were tested at a level equivalent to
100 ppm in soil (1 ppm in water samples). None of the 23 compounds tested yielded a
positive result in the assay at the concentration tested.
Sample matrix effects on assay sensitivity were assessed at the false positive (FP) and false
negative (FN) thresholds. As defined by the EPA2, a FP is a positive result from a sample
containing less than one half of the method detection limit (2.5 ppb in water). Similarly, a
FN is a negative result from a sample containing greater than twice the method detection
limit (10 ppb in water). Thirty chemically diverse soil and water types were spiked with
TNT at one half and twice the assay detection limit and were run in the assay (Figures 4
and 5). Samples yielding a % relative reflectance (%RR) greater than zero are determined
to be greater than the assay detection limit. Likewise, samples yielding a %RR less than
one are determined as less than the assay detection limit. None of the water and soil
583
-------
samples spiked at the FN threshold were determined to be negative, where as 2 of the water
samples and 2 of the soil samples spiked at the FP threshold were seen as positive (6.6%).
The application of these tests in the field has been demonstrated at several sites across the
United States. Results from a TNT field trial at Joliet Army Ammunition Plant reported
good agreement between the D TECH method and Method 8330 (Table 3). During this
study, 42 soil samples were screened on site by the D TECH method and shipped to an
independent laboratory for Method 8330 confirmation. Results from this study reported no
false negatives and 1 false positive (2 %). When confirmed by Method 8330, 98% of the
samples showed the same trends as Method 8330 with 74 % of the sample results in direct
agreement between methods.
Results from a larger scale TNT study (334 samples) has shown these tests to be accurate
as well as cost and time effective (Figure 6). During this study, the D TECH TNT kit was
used to screen 334 samples on site at a cost of approximately $30 per test. Approximately
50 samples were tested by a person on a given day. Four tests could be executed in
approximately 20 minutes. The total number of samples run in a day was limited by the
speed with which they were collected from the field. All soil samples were submitted for
Method 8330 confirmation at a cost of $300 per test. Ninety one percent of the 334
samples screened on site by the D TECH method were in agreement with Method 8330
results. Ten samples were reported as false positive (3%) the D TECH method reported
108 of the samples contained TNT concentrations greater than the project action limit (2
ppm). If the D TECH results had been used to determine which samples were greater than
the project action limit, and only those samples were submitted for Method 8330
confirmation, a net savings of $57,000 would have resulted.
CONCLUSIONS
This environmental EIA has demonstrated a high degree of sensitivity and specificity for
the analyte of interest. False positive and false negative rates are well within acceptable
limits in soil and water matrices. This assay is field tested and shows good correlation with
EPA SW-846 Method 8330 while demonstrating the time and cost effective attributes
required for field screening systems.
REFERENCES
1. Draft Method 8330-1 Nitroaromatics and Nitramines by High Performance
Liquid Chromatography (HPLC). Manual SW-846, U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response, Washington,
DC. 1992.
2 Barry Lesnik, U.S. Environmental Protection Agency, Office of Solid Waste,
Methods Section (OS-331), 401 M St., SW, Washington, DC. 1993.
584
-------
Figure 1. Competitive Enzyme Immunoassay
O
-/-s "
A-4 W
White Latex Particle with
Conjugated Antibody
Membrane Filter
585
-------
8
o
CD
'i
i
0 5 10 15 20 25 30 35 40 45
TNT Concentration in Water (ppb)
Figure 2. The D TECH TNT water kit standard curve
i
50
I
55
60
65
o
£
8
(U
100-
75-
50-
25-
0-
10
100
1000
Analyte Concentration in Water (ppb)
Figure 3. The TNT assay sensitivity to cross-reactive analogs
586
-------
0.5X MDL
2XMDL
o
a
u
tf
3
Water Sample ID
Figure 4. Water sample matrix effects on the TNT assay sensitivity
(U
O
(U
(U
100
90'
80'
70'
60-
50-
40-
30-
20-
10-
0-
-10-
-20-
0.5X MDL
2XMDL
j Jjf j J J J J| • Jj JJ •• J J
I I I I I I I I I I I I I I I
Soil Sample ID
PQ
Figure 5. Soil sample matrix effects on the TNT assay sensitivity
587
-------
Figure 6. TNT field study cost analysis
• Analytical Costs without Field Screening:
334 Samples @ $300 per HPLC Sample = $100,020
• Analytical Costs with Field Screening:
334 Samples @ $30 per D TECH Sample = $10,020
108 Samples (5) $300 per HPLC Sample -= $32.400
Total Costs =$42,420
Savings = $57,600
588
-------
Table 1. Compounds tested for cross-reactivity in the D TECH TNT assay.
If detected at 500 ppb or less, the compound was determined to be cross-
reactive.^
Analytes of Interest Detected
M8330 Compounds
1,3-Dinitrobenzene No
2,4-Dinitrotoluene Yes
2,6-Dinitrotoluene No
HMX (octahydro-l,3,5,7-tetranitro-l,3,5-triazine) No
RDX (hexahydro-l,3,5-trinitro-l,355-triazine) No
Nitrobenzene No
2-Nitrotoluene No
3-Nitrotoluene No
4-Nitrotoluene No
Tetryl (Methyl-2,4,6-trinitrophenylnitramine) Yes
2,4,6-Trinitrotoluene Yes
1,3,5-Trinitrobenzene Yes
2-Amino-4,6-dinitrotoluene Yes
4-Amino-2,6-dinitrotoluene No
Others
Nitrate No
Nitrite No
589
-------
Table 2. Potential interfering organic co-contaminants tested in the TNT
assay
Compound Concentration of analyte in a water
sample required to yield a positive test
result (ppm)
Atrazine > 100
Aroclor 1254 > 100
Acetone > 100
Toluene > 100
Ethylbenzene > 100
Xylene > 100
Benzene > 100
Methanol > 100
Benzo(a)pyrene > 100
Acenaphthene > 100
Acenaphthalene > 100
1,2-Benzanthracene > 100
Benzo(k)fluoranthene > 100
Benzo(ghi)perylene > 100
Benzo(b)fluoranthene > 100
Chrysene > 100
Dibenz(ah)anthracene > 100
Fluoranthene > 100
Fluorene > 100
Indeno(123-cd)pyrene > 100
Naphthalene > 100
Pyrene > 100
Phenanthrene > 100
590
-------
Table 3. TNT field trial results from Joliet Army
Ammunition Plant
Sample
ID
61-1
61-10
61-11
61-12
61-13
61-14
61-15
61-16
61-17
61-18
61-19
61-2
61-20
61-21
61-22
61-23
61-24
61-25
61-26
61-27
61-28
61-29
61-3
61-30
61-4
61-5
61-6
61-7
61-8
61-9
TET-1
TET-2
TET-3
TL-1
TL-2
TL-3
TL-4
TL-5
TL-6
TL-7
TL-8
TL-9
DTECH
Result
(ppm)
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
> 1.5
<0.2
0.5-1.0
<0.2
<0.2
1.0-1.5
<0.2
<0.2
0.2-0.5
<0.2
<0.2
1.0-1.5
<0.2
> 1.5
0.5-1.0
> 1.5
<0.2
0.5-1.0
0.2-0.5
0.5-1.0
<0.2
<0.2
0.2-0.5
> 1.5
> 1.5
0.2-0.5
> 1.5
0.2-0.5
0.2-0.5
0.5-1.0
0.2-0.5
Method 8330
Result
(ppm)
<0.09
<0.09
<0.09
<0.09
<0.09
<0.09
<0.09
<0.09
<0.09
<0.09
<0.09
>3.0
<0.09
2.44
<0.09
<0.09
1.4
<0.09
<0.09
0.27
<0.09
<0.09
1.3
<0.09
1.1
1.0
>3.0
<0.09
1.0
0.54
<0.09
<0.09
<0.09
0.99
1.2
>3.0
0.66
>3.0
0.66
0.71
1.46
0.92
Correlation
Between
Methods
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
UE
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
OE
YES
YES
YES
YES
YES
FP
YES
YES
UE
OE
YES
UE
YES
UE
UE
UE
UE
OE = Over Estimate
UE = Under Estimate
FP = False Positive
591
-------
91
SCREENING WASTE BY FT-IR
David L. Bumgarner, Director, Analytical Services, Leslie S. Bucina, Laboratory
Manager, Kemron Environmental Services, 109 Starlite Park, Marietta, Ohio, 45750.
ABSTRACT
Infrared (IR) spectrophotometry has been widely used in industrial and academic
laboratories as a powerful tool in the identification of unknown compounds, especially in
combination with nuclear magnetic resonance (NMR), mass spectrometry, and physical
properties such as melting and boiling points. In recent years its utility for qualitative and
quantitative determinations has been enhanced significantly with the introduction of
Fourier Transform Infrared (FT-IR) Spectrophotometry . However, its implementation in
environmental analysis has been limited to a few special applications, specifically EPA
Method 418.1 for the determination of petroleum hydrocarbons in water and soil in
support of NPDES permit monitoring and investigations of leaking underground storage
tanks. For the past three years we have used IR for identification of major components
in unknown waste samples. Our results, which will be summarized in this paper, indicate
the technique is useful as both a screening tool and, in some cases, the primary analytical
procedure. Our paper will provide details for sample preparation, analysis, example
spectra, and the use of commercial and user compiled libraries for both organic and
inorganic liquids and solids. Our examples will illustrate specific laboratory analyses in
which FT-IR was the primary tool used to identify such products and wastes as high
purity industrial solvents, paint wastes, antifreeze, aqueous wastes, and inorganic cyanide
compounds. It is also a useful screening procedure for identification of hydrocarbon
mixtures such as gasoline, diesel and jet fuels, lubricating oils, and solvent mixtures. IR
spectra of these samples will be included along with discussion to document the
analytical conclusions of such projects. The data will demonstrate the utility of FT-IR
as a quick and cost-effective method for identification of major constituents in an
unknown, often limiting the need for subsequent lengthy and expensive analyses.
592
-------
INTRODUCTION
Background
More and more, environmental laboratories are being required to provide fast, cost-
effective methods for analytical screening of samples of unknown composition. These
unknowns may originate from various regulatory compliance and site investigation
activities, and typically consist of bulk liquids and solids collected from unlabeled drums,
tanks, or pits with little or no history to verify the contents. Without a systematic
approach to the screening of such waste, the laboratory will often embark on a costly
analytical project which might include comprehensive analyses for inorganics, metals,
volatile and semivolatile organics, pesticides, and herbicides. A screening technique which
can narrow the list of possible components is necessary to minimize both cost of analysis
and turnaround time. Many clients request extensive waste analyses including full TCLP
on unknowns that are later found to be solvents or other organic liquids. Matrix
interference on these samples can result in detection limits higher than regulatory limits
which are either interpreted as useless, or potentially may cause the waste to be
designated RCRA hazardous by default. A screening analysis can be used to quickly
categorize the waste as a pure product, or common commercial mixture, often ruling out
TCLP. In this context we would like to discuss the use of IR, and FT-IR in particular, in
the determination of unknown waste profiles (major constituent analysis). The techniques
and some actual laboratory examples will be presented in this paper.
Equipment and Supplies
Hardware: Infrared Spectrophotometer: Perkin Elmer Model 1600 FT-IR
Hewlett-Packard Color Pro Plotter
Potassium Bromide Windows (KBr plates) (38.5 x 19.5 x 4 mm)
Demountable Cell Mount
Potassium Bromide Mini-Press
Beta Gas Cell (10 cm)
Precision Scientific Convection Oven Model EG
Reagents: Acetone Resi-analyzed reagent for residue analysis
Water - lonpure Type I or equivalent
Hexane Resi-analyzed reagent for residue analysis
Potassium Bromide - Infrared quality
593
-------
Sample Preparation
Sample preparation for liquid samples is extremely simple requiring only disposable
pipets and two potassium bromide (KBr) plates which serve as a cell. A small drop of
the sample is placed on one of the plates and a thin film is formed by pressing the two
plates together. The plates are placed in the cell holder and the spectrum determined by
FT-IR. The complete sample preparation/analysis process takes less than one minute.
Cleaning of the KBr plates consists of rinsing with resi-analyzed acetone. Etching from
water and other liquids is inevitable, however, this is eliminated by rubbing of the plates
on fine grit sandpaper. Solids are prepared by the KBr mini-press pellet procedure.
Highly concentrated solvents are analyzed utilizing their vapor phase by gas cell
methodology in which the gas cell contents is evacuated and replenished with the vapor
phase of the sample through the transfer of sample headspace using a gaslight syringe.
FT-IR Analysis - Method Overview
Both commercial and user prepared spectral reference libraries are necessary to maximize
the information obtained from the IR spectrum of an unknown material. Some general
conclusions can usually be drawn very quickly regarding the general composition for
certain classes of liquid wastes. It is often readily apparent that the waste is either of
aqueous or petroleum origin based the characteristic simple spectra for water and
hydrocarbons. Further investigation will often suggest the presence of ketones, alcohols,
glycols, or chlorinated solvents. An instrument with library search capabilities can be very
helpful in identification of specific liquid unknowns when the purity is high, although the
information provided by the spectra on complex mixtures must be verified by additional
analytical techniques.
RESULTS AND DISCUSSION
Sample 1
This transparent, colorless, liquid was miscible in water and acetone but immiscible in
hexane. The FT-IR spectrum was determined using the dual KBr plate thin-film
technique, and is presented in Figure 1A. The library spectrum for deionized water is
presented in Figure IB and it is obviously a close match. The liquid was presumed to be
an aqueous waste, with no major organic constituents probable at one per cent or greater.
The water content was later confirmed by the Karl Fischer titration method to be greater
than 99 %. Based on the clients knowledge of the waste profiles at the base, the
laboratory was instructed to limit additional testing to the eight TCLP metals on a total
basis.
594
-------
Sample 2
A state regulatory agency collected this white, unidentified solid while conducting an
investigation at an abandoned site. Although unlabeled, history and other evidence at the
site suggested the possibility the material could be cyanide salts. The agency requested
that the sample be tested for total cyanide by SW-846 Method 9010. Due to concern for
analyst safety and apparatus contamination, a screening analysis by FT-IR was proposed
to determine if the solid contained high levels of cyanide. The KBr pellet technique was
employed in this case and the spectrum presented in Figure 2A was generated. The
absorption band at 2079 cm"1 is characteristic of cyanide salts attributed to the C-N bond.
The laboratory used solid potassium cyanide and a KBr pellet to produce the spectrum
labeled as Figure 2B. Due to the excellent match, we presumed the white powder to be
a cyanide salt and the cyanide analyst was able to proceed on that assumption. A known
mass of the sample was prepared in the manner of a cyanide standard stock solution and
serially diluted into the expected linear range of the calibration curve. The analysis by
Method 9010 confirmed the IR data, giving a result of 440,000 mg/kg CN (44% cyanide).
Sample 3
This transparent liquid with solvent-like odor was miscible in acetone and hexane but
immiscible in water. The FT-IR spectrum (Figure 3A) was very similar to the library
spectrum for methyl ethyl ketone (Figure 3B). Client had requested analysis by EPA
Method 8240 to check for solvents. The preliminary FT-IR analysis provided the GC/MS
analyst with enough information about the sample to prepare the proper dilution of the
waste and confirm the screening data. Although the KBr plate method (thin film) was
employed for this sample, the gas/vapor cells are often used in similar situations to
identify solvents of relatively high purity.
Sample 4
This transparent liquid with familiar odor was received as an unknown from the airforce
base. The sample was hexane and acetone miscible but water immiscible. A thin film
was prepared between two KBr plates and the FT-IR spectrum presented in Figure 4A
was generated. A library search gave several possibilities with the best match being
unleaded gasoline presented in Figure 4B. Confirmation by direct injection into a
GC/FID produced a fingerprint with a carbon range of C4 C10, which was indicative of
gasoline range organics. No additional testing was indicated or required.
595
-------
Sample 5
This was an amber, clear liquid with solvent like odor which was miscible in hexane but
immiscible in water or acetone. The thin film KBr plate technique yielded the FT-IR
spectrum shown in Figure 5A. This spectrum is typical of many petroleum distillate
unknowns received at our laboratory for identification, classification, or possible
candidates for fuels blending. Note the similarity of the spectrum to that of Figure 5B-
a laboratory specific entry in our library for a "brush cleaner", a product similar to an oil-
base paint stripper. Confirmation by GC/FID produced a fingerprint and carbon range not
unlike that of our standard brush cleaner. Although, the FT-IR spectrum is typical of
many petroleum distillates, the absorption band at approximately 1700 cm"1 is typical of
carbonyl compounds. Because of this additional data, it was determined that analysis by
GC/MS Method 8240 might be needed to identify and quantify any other major
components. GC/MS analysis confirmed the presence of 4-methyl-2-pentanone, 2-
butanane, toluene, and total xylenes all at percent levels.
Sample 6
A clear, colorless liquid, this sample was miscible in both water and acetone, but
immiscible in hexane. The thin-film KBr plate technique yielded the spectrum labeled
as Figure 6A. The spectrum shows characteristics of -OH from water and alcohols (3400
cm"1) and the absorption band at 2900 3000 cm"1 is typical of the C-H stretch. The IR
spectrum was similar to that of short chain alcohols and glycols including methanol and
ethylene glycol. It was determined that further analysis by SW-846 Method 8015
(GC/FID) would be necessary to identify and quantitate the major constituents. Based on
retention time data for a series of alcohols, it was determined that the sample was high
purity methanol. The IR spectrum for methanol is presented as Figure 6B, however; note
the similarity to ethylene glycol (Figure 7B).
Sample 7
Another typical waste originating from an airforce base, this green liquid was miscible
in water and acetone but immiscible in hexane. The thin film KBr plate technique was
used to produce the FT-IR spectrum presented in Figure 7A. The bands at 3345, 2944,
and 2832 cm"1 are typical of neat alcohols and glycols, or aqueous solutions of the same.
In this particular sample the particular fluorescent green color is often indicative that we
are dealing with an ethylene glycol based antifreeze mixture. Although confirmation may
not have been necessary based on the color, sample origin, and the similarity to the
spectrum for ethylene glycol (Figure 7B), the composition was confirmed both
qualitatively and quantitatively by GC/FID analysis to be ethylene glycol.
596
-------
Sample 8
A transparent liquid sample with a solvent-like odor was collected from an unknown
waste drum for characterization. Solubility tests were performed which revealed the
miscibility of the sample in both hexane and acetone as well as the immiscibility of the
sample in water. A FT-IR spectrum was obtained using the thin film dual KBr plate
technique, presented in Figure 8A. The IR spectrum when library searched closely
resembled the laboratory Freon 113 standard which contains characteristic bands within
the fingerprint region (Figure 8B). The confirmation for the halogenated solvent was
performed by use of a volatile GC/MS analysis. The Freon 113 was qualified by a
positive spectral identification and quantified at a concentration of greater than 99%.
Sample 9
This sample which was a yellow liquid which was soluble in hexane and insoluble in
water and acetone was submitted from a vehicle maintenance facility. The FT-IR
spectrum was achieved using the thin film KBr plate technique (Refer to Figure 9A). The
resulting spectrum which was characteristic of ° hydrocarbon was library searched with
a tentative identification of a lubricating oil (See Figure 9B). The hydrocarbon sample
was then fingerprinted by GC using a modified SW-846 Method 8015. The hydrocarbon
fingerprint in the chromatogram had a carbon range of C14-C32 which is similar to a lube
oil carbon range.
Sample 10
This unknown was a transparent liquid sample which exhibited a distinct fuel odor. The
FT-IR analysis using the thin film KBr plate technique gave a characteristic hydrocarbon
spectrum as shown in Figure 10A. The fuel library when searched gave a tentative match
of diesel fuel (Figure 10B). The sample was then analyzed for diesel range organics by
a modified SW-846 Method 8015 to confirm the carbon range and to provide a
quantitation. The diesel pattern had a carbon range of Q-C^ and a total area quantitation
of approximately 67%.
597
-------
SUMMARY
From the examples discussed in this paper, we have shown that FT-IR can be a fast and
effective tool in the identification of unknown wastes. Basic infrared sample preparation
techniques which employ the thin film plate technique, KBr pellets, mulls, disposable
porous cells, and the gas cell are used to prepare the sample for the FT/IR screening.
This preparation and screening with resultant spectra, which provide enough information
to characterize wastes as aqueous, petroleum distillates, pure solvents, solvent mixtures,
and common commercial products, can be performed in just a few minutes. Although
some samples can be positively identified by FT-IR alone, the main utility of the spectra
is to guide the laboratory in the in the selection of subsequent confirmatory techniques.
Armed with the information from the FT-IR analysis and with, or without, knowledge of
waste, several costly and time consuming analytical methods (i.e. TCLP) can be reduced
or eliminated entirely.
ACKNOWLEDGMENTS
The authors would like to thank the management and technical staff at the Occupational
and Environmental Health (OEHL) Laboratory and Brooks Airforce Base, San Antonio,
for their technical assistance in implementing these procedures. Most of the original work
with the IR screening procedures presented were developed at OEHL's Armstrong
Laboratory. We would also like to thank KEMRON employees Chad Barnes, Rodney
Campbell, and Cheryl Koelsch for their help in preparing this paper.
598
-------
0.00
94/04/29 1-4:56
X: 4 scans. 4.0cm-l. sbex
Sample 1
FIGURE 1A
4000 3500 3000 3500 MOO
BU0007: 0 scans. O.Ocm-1. apod none, not 1600
01 Hater
FIGURE IB
5S9
-------
0.00
4000 3500 3000 2500 2000
1500
SOO
94/05/04 1229
X: 4 scans. 4.0ca~l. abex
Sample 2
FIGURE 2A
100.00-
jrr
o.oo
—i 1 1 1 1—
4000 3500 3000 2500 2000
1500
1 I—
1000 C«H 600
BUOOB4: 0 scans, O.Ocn-1. apod none, not 1BOO
CYANIDE KBT
FIGURE 2B
600
-------
100.00-
XT
0.00
—I 1 1 1 1—
4000 3500 3000 2500 2000
1500
1 1—
1000 c«-" 500
94/05/04 10:01
X: 4 scans. 4.0co-l. abex
sample 3
FIGURE 3A
100.00-
rr
0.00-" , 1 1 r
4000 3500 3000 2500 2000
1
1500
1000
500
BU0074: 0 scans. O.Oco-1. apod none, not 1600
MEK (BHKS)
FIGURE 3B
601
-------
PG1KIN EUfPI
0.00
4000
3500
3000
2500
2000
cm-1 500
94/04/29 15: 31
X: 4 scans. 4.0cm-l. abex
Sample 4
FIGURE 4A
0.00
4000
3000
2500
8000
1SOO
1000
C«H 500
94/04/29 15:34
Y: 0 scans. O.Ocn-1. apod none. abex. not 1600
GASOLINE
FIGURE 4B
602
-------
100.00-
XT
0.00
—I—
•4000
3500
2500
2000
1500
T~
500
94/04/29 15: 12
X: 4 scans. 4.0cm-l. abex
Sample 5
FIGURE 5A
PE3K/N BJffl
0.00
1000
500
8U0063: 0 scans. O.Ocn-1, apod none, not 1EOO
Brush Cleaner
FIGURE SB
603
-------
100.00-
jrr
o.oo-
4000 3500 3000 2500 2000
1000
CB-" 500
94/04/11 OS 49
0413102: 1 scan. 4.0cm-l. abex
Sample 6
FIGURE 6A
100.00-
XT
0.00-
-t-
1 1 1 1—
4000 3500 3000 2500 2000
1
1500
1
1000
r—
500
BU0005: 0 scans. O.Ocm-1. apod none, not 1600
Methanol
FIGURE 6B
604
-------
PGJKIN B-tf*
100.00-1
XT
0.00
(\
4000
3500
3000
3500
2000
1500
1000
cm-' 500
94/04/28 13:30
0441601: 4 scans. 4.0cn-l. abex
Sample 7
FIGURE 7A
PSWtlN ELfff
100.00-1
XT
0.00-
—I 1 1 1 1—
4000 3500 3000 2500 3000
cmH 500
1500
1000
BU0051: 0 scans. O.Ocm-1. apod none, not 1600
Etnylene Slycol
FIGURE 7B
605
-------
3500
3000
2500
2000
cm-' 500
94/02/14 13:46
0224201: 4 scans.
Sample a
4.0cm-l. abex
FIGURE 8A
100.00-
XT
*"" «
o.oo-1 1 1 1 r
1
1500
4000 3500 3000 2500 2000
1000
r~
500
BU0038: 0 scans. O.Ocm-1. apod none, not 1600
Freon-113
FIGURE 8B
606
-------
0.00
94/03/04 13: 13
0308101: l scan. 4.0cm-1. abex
Sample 9
FIGURE 9 A
0.00
LU2S5A: 0 scans. O.Ocn-1, apod none, not 1600
SHELL TELLUS LUBRICANT
FIGURE 9B
607
-------
0.00
4000 3500 3000 2500 2000
1500
1000
CBH 500
94/04/11 09:28
0412601: 1 scan. 4.0cm-l. abex
Sample 10
FIGURE 10A
100.00-
XT
0.00
1 1 1 1 1
4000 3500 3000 2500 2000
1500
1000
800
BUOOB2: 0 scans, O.Ocra-1. apod none, not 1600
dlesel fuel
FIGURE 10B
608
-------
An Immunoassay for the Detection of Benzene in Water
S. Friedman, T. Stewart, and R. Allen
EnSys Environmental Products, Inc.
Research Triangle Park, NC 27709
Benzene is a toxic component of gasoline and other refined petroleum products. It is
classified as both a carcinogen and a mutagen and presents a serious health risk to the
general population. The contamination of groundwater is facilitated by its water solubility
and occurs during the transport, processing, and storage of refined petroleum products.
Total benzene usage in the United States has been estimated to be approximately 11 billion
gallons per year, and 238,000 people are believed to be occupationally exposed to its toxic
effects. Benzene testing is usually performed with a laboratory by GC using EPA Method
8020 or 8021 A.
The antigen-antibody binding reaction of an immunoassay is a function of conformational
complimentary of the reactants (i.e. lock and key). The equilibrium constant of the
complex limits the sensitivity attainable by the assay, and is a function of cooperative, nori-
covalent, interactions that exist between the ligand and side chain chemistry of the
antibody binding pocket. Larger ligands, having numerous sites of interaction with the
antibody, produce the most sensitive irnmunoassay methods.
Benzene is, from an immunochemical perspective, a small and unremarkable molecule. Its
aromatic ring configuration is a common constituent of many other compounds found in
the environment. An antibody and immunoassay that recognizes benzene will often
recognize other irrelevant compounds found in the same sample. The molecule's small size
and chemistry also limits the potential association constant and assay sensitivity that can
reasonably be expected. Current regulatory statutes for drinking water in most states,
however, necessitates a method that can reliably detect benzene at approximately l
We have developed an immunoassay that can detect benzene at 5 ppb in a water sample.
The test can be completed on location within approximately thirty minutes. The method
includes a sample processing component that is used to collect, concentrate, and modify
benzene in the sample. Extracted samples are subsequently analyzed using a monoclonal
antibody-based ELISA that provides a chromogenic indication of contamination. The
cross-reactivity of the test to a variety of associated aliphatic and aromatic compounds
was found to be <1% for the majority of compounds tested. Analysis of BTEX
compounds indicated approximately 10% cross reactivity for toluene, <1% for xylene
isomers, and <1% for ethylbenzene. The technical strategy, protocol, and performance
characteristics of the Benzene-RIS£® immunoassay method will be presented.
609
-------
93
CHARACTERIZATION OF MULTIPHASE PARTITIONING
H. M. Liljestrand, Y. Shimizu, I. Lo, Y. D. Lee, and J. K. Koo, Civil Engineering ECJ 8.6, The
University of Texas, Austin, Texas 78712.
ABSTRACT
The characterization of partitioning of pollutants among the various phases in complex mixtures is
a common and challenging question. For neutral hydrophobic pollutants at low concentrations, the
batch equilibrium partitioning between gas, aqueous, oily, and solid natural organic carbon phases
is linear and depends upon the relative amounts of each phase as well as the partition coefficients
between phases. By varying the total mass of mixture while holding total mass of tracer species
and total volume constant in a headspace analysis for chemical species of a range of octanol-water
coefficients, one can determine the product of the linear partitioning coefficient to a phase and the
mass fraction of that phase, analogous to Kp = (%OC) Koc in aqueous-solid natural organic
carbon systems. Ideally for n different phases, n different tracer species are used, with one species
preferentially partitioning into each different phase. In practice, the number of species introduced
is at least 3X the number of phases. For samples with a high content of oily phase, the headspace
partial pressure of a homologous series of perfluorocarbons is used with the mass balance to
determine the oil content and partitioning. A comparison of this process of introducing tracers and
measuring the headspace fugacity after sorption equilibrium with the alternative approach of
quantifying the amount of each phase and determining the desorption equilibrium by sequential
extraction measurements indicates the former is more rapid both in terms of the time to reach
equilibrium and the number of tests.
610
-------
94
Validation of an Enzyme Immunoassay Based Field Screening System for the
Detection of RDX in Soil and Water Samples
George B. Teanev anri Robert T. Hudak
Strategic Diagnostics Incorporated
128 Sandy Dr.
Newark, DE 19713
ABSTRACT
RDX, hexahydro-l,3,5-trinitro-l,3,5-triazine, js a secondary military explosive commonly found
in soils and ground water at munitions demilitarization sites. This compound is very mobile in
soils, and is often used to indicate the leading edge of the explosives plume in soils and ground
water. The development of a field screening method specific to RDX would greatly benefit the
explosives remediation effort. The use of inexpensive field screening methods can reduce the
total cost and improve the efficiency of site surveys and remediation projects. Field screening
techniques can also circumvent the long turnaround time of laboratory analysis by providing
reliable on site results.
The portable RDX enzyme immunoassay (EIA) is a quick, cost effective, highly specific, and
user friendly field screen option. The components of this EIA include RDX specific antibodies
(Ab) covalently linked to small latex particles, an RDX analog which is covalently linked to
alkaline phosphatase, and the free RDX in the water sample. The free RDX competes with the
enzyme linked analog for the Ab binding sites. The latex particles are then collected on a filter
device, washed, and an enzyme substrate is added. The amount of color produced is inversely
proportional to the concentration of free RDX in the water or soil sample, and can be determined
using a hand held reflectometer, or a color card.
This paper presents much of the work undertaken to validate this field screening method. This
assay has demonstrated 5 ppb and 0.5 ppm sensitivity in water and soil samples respectively,
with minimal cross-reactivity to EPA SW-846 Method 8330 analytes and other common organic
pollutants. This assay has displayed minimal false negative and false positive rates in soil and
water matrices, and incorporates a 3 minute soil extraction step with average recoveries of
90%. The RDX assay has been proven effective and reliable as field trials report 90% and 96%
correlation to Method 8330. False positive rates during field trials have been reported as
approximately 4 %, where as no false negative results have been reported.
INTRODUCTION
RDX, hexahydro-l,3,5-trinitro-l,3,5-triazine, is a secondary military explosive commonly found
in soils and ground water at munitions demilitarization sites. This compound is very mobile in
soils, and is often used to indicate the leading edge of the explosives plume in soil and ground
water. As such, characterization and remediation of these sites are essential to insure ground
water quality in the surrounding region.
Initial efforts in any remediation project will generally focus on site characterization. The most
commonly used analytical method for explosives determination is EPA SW-846 Method 83301.
This laboratory method is very robust, however it is also time consuming, and costly. When
surveying the boundaries of the contaminant plume, a large percentage of the samples analyzed
611
-------
are free of contamination. These sample can be more efficiently identified using field screening
techniques, sending only known positive samples to be confirmed by the laboratory procedure.
The development of a field screening method specific to RDX would greatly benefit the
explosives remediation effort providing the method is accurate, reliable, cost effective and user
friendly. The use of inexpensive field screening methods can reduce the total cost and improve
the efficiency of site surveys and remediation projects. Field screening techniques can also
circumvent the long turnaround time of laboratory analysis by providing reliable on site results.
Immunoassay based field screening techniques have gained increased popularity in recent years.
Immuno-technology is rooted in medical diagnostics industry and has been used since the late
1950's. This technology is based on the specificity and sensitivity of the antibody-antigen
interaction. Antibodies are proteins produced in response to a foreign substance (antigen) in the
body. These proteins recognize 3-dimensional molecular structure and physically bind the
compound which elicited their response.
This paper discusses the development of an immunoassay based field screening system for the
detection of RDX.
METHODS
The assay format of this test is a competitive enzyme immunoassay (EIA). Briefly, the
components of this EIA include RDX specific antibodies (Ab) covalently linked to small latex
particles, an RDX analog which is covalently coupled to alkaline phosphatase, RDX reference
material, and the free RDX in the water sample. The water sample is added to a test vial
containing lyophilized Ab-Latex and enzyme conjugate. At the same time a rehydration buffer is
added to the reference vial, suspending the lyophilized Ab-Latex, enzyme conjugate and
reference material within. The free RDX in the water sample or the reference material competes
with the enzyme linked analog for the Ab binding sites. The latex particles from the test and
reference vials are then collected in a duel well filter device, and washed to remove any unbound
material. An enzyme substrate is added, and the reference color allowed to develop. The
amount of color produced in the test well is inversely proportional to the concentration of free
RDX in the water or soil sample, and can be determined relative to the reference well using a
hand held reflectometer, or a color card. The reflectometer measures the reflectance in each well
and displays its results in units of percent relative reflectance (%RR), the reflectance of the test
relative to that of the reference. The %RR can be used with the translation table supplied in the
D TECH™ kit to estimate RDX concentration.
Sample preparation of soil or water matrices is minimal. Soil samples must go through a quick
and simple extraction in 100 % acetone followed by a 2 step dilution and filtration step prior to
testing. A water sample requires a quick 1 step dilution and filtration prior to testing. Unless
noted below, all extraction, dilution and filtration steps were carried out as described in the D
TECH product literature.
Using the test as described above, we investigated the sensitivity, and specificity of the assay.
Attributes of the test such as the false positive / false negative rates were determined both
theoretically and using spiked and real world samples. Organic co-contaminants were evaluated
as possible interfering substances. Soil extraction efficiencies were evaluated as was
performance under field conditions.
612
-------
RESULTS AND DISCUSSION
Assay sensitivity was determined for water and soil systems by developing standard curves
(Figure 1) while assessing the error associated with one half and twice the minimum detection
limit (MDL) (Table 1). The assay sensitivity has consistently been demonstrated at 5 ppb in
water and 0.5 ppm in soil samples with an over all assay range of 5 ppb to 60 ppb and 0.5 ppm
to 6.0 ppm respectively. As described by the EPA a false positive is a sample which yields a
positive test result but contains less than half of the method detection limit X Similarly, a false
negative is a sample which yields a negative test result but contains more than twice the method
detection limit. Using these criteria, and a 96% confidence interval (2 standard deviations), 5
ppb is clearly discernible from zero, the false positive (2.5 ppb) and the false negative (10 ppb)
thresholds in water. The false positive threshold (2.5 ppb) is clearly discernible from the MDL
by 2 standard deviations(SD) while the false negative threshold (10 ppb) is discernible by 3 SD
(99%).
Assay specificity was evaluated by testing a host of explosives commonly associated with RDX
remediation sites. All compounds described in Method 8330 as well as others were tested for
cross-reactivity in the assay (Table 2). These compounds were initially tested at a concentration
100 times the method detection limit (5 ppb) in the assay. One analyte, octahydro-1,3,5,7-
tetranitro-l,3,5-triazine (HMX), was found to be marginally cross-reactive (~2% relative to
RDX), while all other compounds tested were undetected (Figure 2). The MDL of HMX in a
water sample is approximately 150 ppb and the IC5Q (concentration yielding 50 % inhibition in
the assay) approximates 900 ppb. In the real world, sites containing RDX contamination are also
likely to have HMX. Due to this likelihood, the results of the field screen are expressed as RDX
equivalents.
Twenty three potential organic co-contaminants were evaluated in the RDX assay at a
concentration of 500 ppb (100 times the MDL of the assay). These compounds represent a cross
section of herbicides, organic solvents and polyaromatic hydrocarbons likely to be found at
explosives sites (Table 3). When tested at 500 ppb, none of these compounds tested positive in
the assay.
Water and soil matrix effects were evaluated by testing 22 different RDX-free water and soil
samples collected from across the country (Table 4). The water samples encompass a wide range
of ground and surface water chemistries. Likewise, the soils tested reflect a wide range of soil
types, textures, and chemistries. Water and soil extracts were spiked at 0.5X and 2X the assay
detection limit to determine the effect of sample matrix on the assay. All treatments were run in
triplicate including non-matrix controls (Figures 3 and 4). All water samples spiked at 10 ppb
were reported to contain RDX at levels greater than 5ppb (%RR > 0). Similarly, with the
exception of sample 8, all of the waters spiked at 2.5 ppb were reported as < 5ppb (%RR < 0).
The water type in question (sample 8) was collected from the Pacific Ocean, where all other
samples were collected from fresh surface water or ground water sources. As a result, the use of
a salt water matrix in this system is not recommended. Soil matrix results show similar trends
(Figure 4). In all 22 soil treatments, the assay false negative threshold (1.0 ppm) were reported
as > 0.5 ppm, while samples containing the false negative threshold (0.25 ppm) were reported as
< 0.5 ppm (no false negative or false positive samples reported).
Soil spike and recovery studies were conducted using 10 chemically and physically diverse soils
(Table 5). Soil samples were spiked with RDX concentrations of 1 ppm and 6 ppm and each
treatment was replicated 6 times. All samples were run as described in the product users guide
and the relative reflectance data was converted to discrete RDX concentrations using the assay
standard curve. Extraction efficiencies across all treatments ranged from 53 % to 114 %
recovery (Table 6). Recoveries of the 6 ppm spike appears to be highly efficient where as
613
-------
recoveries of the lower concentration averaged 86 %. There was no apparent correlation
between the soil characteristics in table 5 and the poor recovery in soil 101 spiked at 1 ppm.
Considering most of the analytical laboratories accept a matrix spike recovery of ± 30 %, the
field screen extraction procedure is effective across the soils tested.
Thirty one (31) real world soil samples were collected from proprietary sites in the north western
United States. The field moist samples were mixed thoroughly and split into 2 sub-samples.
One sub-sample from each of the 31 soils was sent to Data Chem Laboratories, Salt Lake City,
Utah, for Method 8330 analysis. The second set of sub-samples were tested using the D TECH
RDX immunoassay. The samples were extracted using the standard D TECH procedure with the
following exception; in order to lower the sensitivity of the assay, the soil extract dilution step
was modified. This lowered the working range of the kit from 0.5 - 6.0 ppm to 0.4 - 4.0 ppm.
The results from the D TECH assay were compared with the results of Method 8330 to assess
the correlation between methods (Figure 5). No false positive or false negative samples were
reported by D TECH. In order to obtain a more accurate assessment of the correlation in the
quantitative range of the kit, samples reported as less than 0.4 ppm by both methods were
omitted from the analysis. Similarly, samples greater than 4.0 ppm by both methods were
omitted from the analysis. Regression analysis of this modified data indicate a slope of 1.02
with a coefficient of correlation reported as 90% (Figure 5). When all of the data were included
in the analysis the coefficient of correlation increased to 94% and the y-intercept was reported as
0.03 (data not shown). In either case there is good agreement between the laboratory based
method and the field screening method.
The results of a recent field trial report exceptional agreement between D TECH and Method
8330. Sixty six samples were extracted and tested in duplicate by D TECH and were then sent to
Hercules Environmental Testing Laboratories, Magna, Utah, for Method 8330 confirmation.
No false negatives were reported by D TECH during this trial. However, 3 samples (4%) were
reported as false positive. Linear regression analysis of the D TECH and Method 8330 data was
performed as described above. Samples less than 0.5 ppm by both methods and greater than 6.0
ppm by both methods were omitted from the data set. The results from this analysis report a
slope of 1.1 and a 96% coefficient of correlation (Figure 6).
A collaborative field trial was conducted with the Roy F. Weston Company in Lionville, PA.
During this study 30 soil samples from random explosives sites were pulled from the Weston
archives. Using the historical analytical data, samples were selected so that 10 were below a
concentration of 1.3 ppm and the remaining 20 were greater than 1.3 ppm. In cases where the
historical data yielded RDX concentrations greater than 6 ppm, the sample extract was diluted in
an appropriate volume of acetone to bring the sample within the working range of the assay. The
samples were tested using the immunoassay and results confirmed by SW-846 Method 8330. All
assays were run by Weston personnel. No false positive or false negative samples were reported
in this study. Using the assay standard curve, the relative reflectance data and the additional
dilution factor were transformed into discrete RDX concentrations. These data were regressed
against Method 8330 results omitting the samples below the detectable limit of both methods.
The regression analysis reported a slope of 0.92 and a correlation coefficient of 91% (Figure 7)
In the field, soil samples were tested in less than 20 minutes (15 minutes for a water sample).
and several tests can be run simultaneously. Previous experience with this system has shown
that 1 person can run, on average, 50 tests per day. The water test requires the user to follow 7
easy steps and all components required to run a test are included in the kit. The cost of this
system is approximately one fifth to one tenth the cost of more rigorous analytical procedures.
614
-------
CONCLUSIONS
The D TECH RDX field screen system has demonstrated a high degree of sensitivity and
specificity while demonstrating a low false positive (<, 4 %) and false negative (< 1%) potential.
The assay has been designed to minimize sample matrix effects in water and soil. The low cost
per test coupled with the ease of use and quick results facilitate a much wiser use of the
analytical dollar. Using such a system greatly reduces the number of negative samples sent to
the laboratory for analytical confirmation.
The D TECH method has shown greater than 90 % correlation with EPA SW-846 Method
8330, while exhibiting the ease, time and cost effective attributes required for field screening
systems.
ACKNOWLEDGMENTS
The authors would like to thank Robin Dolan for her outstanding effort at the bench, Dr. Larry
Motyka and Dr. James Stave for their work in developing the RDX antibody, and Dr. James
Melby for his guidance throughout the project.
REFERENCES
1. Draft Method 8330-1 - Nitroaromatics and Nitramines by High Performance Liquid
Chromatography (HPLC). Manual SW-846, U.S. Environmental Protection Agency, Office of
Solid Waste and Emergency Response, Washington, DC. 1992.
2 Barry Lesnik, U.S. Environmental Protection Agency, Office of Solid Waste, Methods
Section (OS-331), 401 M St., SW, Washington, DC. 1993.
615
-------
0
0)
1
RDX Concentration in Water (ppb)
Figure 1. The D TECH RDX water kit standard curve
100-
75-
501
(25)
25-
0.0
0.1 1.0 10.0 100.0
Analyte Concentration in Water (ppb)
1000.0
M
10000.0
Figure 2. Cross-reactivity characteristics of HMX in the D TECH RDX assay
616
-------
o
o
1
I
80-
70-
60-
50-
40-
30-
20-
10-
-10-
-20-
n = 3
error bars = 1 standard deviation
1/2XMDL
2XMDL
T - T T T
si Ea fa T A in
yyyypuyuyyup
¥
Water Sample ID
Figure 3. The effect of different water types on the sensitivity of the D TECH RDX assay.
g
o
8
o
-------
I
VI
ffi
Q
y=1.010x + 0.140 r = 0.897
n = 18
o.oo O
o.oo
4.00 4.40 4.80
EPA SW846 Method 8330 Result (ppm)
Figure 5. Method correlation from an analysis of 18 real world RDX soil samples
•-
1
ffi
8
Q
y=1.095x +0.228 r = 0.959
n = 49
0 0.5
1 1.5
EPA SW-846 Method 8330 Result (ppm)
618
-------
1250-
1000-
?
a
\^s
'S 750-
ffi
g 500-
Q
250-
n-l
y = 0.918x- 16.153 r = 0.906
n = 20
D
^
^^
^^
^^
^^^^ U
^^
^Af^^ Q
o
a
8
p~
EPA SW-846 Method 8330 Result (ppm)
Figure 7. Method correlation from an analysis of 20 RDX contaminated soils
conducted by Roy F. Weston Company
619
-------
Table 1. RDX Assay sensitivity and range in deionized water
RDX Concentration Concentration MeanRR Standard
(ppb) Significance (%) Deviation
0
2.5
5
10
30
60
n = 10
FP = false positive
FN = false negative
RR = relative reflectance
True Negative
FP Threshold
MDL
FN Threshold
Mid-Range
High-Range
-10
-4
4
18
56
77
3
2
4
3
4
5
Table 2. Compounds tested for cross-reactivity in the D TECH RDX assay.
If detected at 500 ppb or less, the compound was determined to be cross-
reactive.
Analytes of Interest Detected
M8330 Compounds
1,3-Dinitrobenzene No
2,4-Dinitrotoluene No
2,6-Dinitrotoluene No
HMX (octahydro-l,3,5,7-tetranitro-l,3,5-triazine) Yes
RDX (hexahydro-l,3,5-trinitro-l,3,5-triazine) Yes
Nitrobenzene No
2-Nitrotoluene No
3-Nitrotoluene No
4-Nitrotoluene No
Tetryl (Methyl-2,4,6-trinitrophenyhiitramine) No
2,4,6-Trinitrotoluene No
1,3,5-Trinitrobenzene No
2-Amino-4,6-dinitrotoluene No
4-Amino-2,6-dinitrotoluene No
Others
NG (Nitroglycerine) No
PETN (Pentaerythritoltetranitrate) No
620
-------
Table 3. Potential interfering organic co-contaminants tested in the RDX
assay
Compound Concentration of analyte in a water sample
required to yield a positive test
result (ppb)
Atrazine > 500
Aroclor 1254 > 500
Acetone > 500
Toluene > 500
Ethylbenzene > 500
Xylene > 500
Benzene > 500
Methanol > 500
Benzo(a)pyrene > 500
Acenaphthene > 500
Acenaphthalene > 500
1,2-Benzanthracene > 500
Benzo(k)fluoranthene > 500
Benzo(ghi)perylene > 500
Benzo(b)fluoranthene > 500
Chrysene > 500
Dibenz(ah)anthracene > 500
Huoranthene > 500
Fluorene > 500
Indeno(123-cd)pyrene > 500
Naphthalene > 500
Pyrene > 500
Phenanthrene > 500
621
-------
Table 4. Soil and water types used in D TECH RDX assay matrix studies
Soil/Water ID Soil Type Water type
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Low organic clay loam
Sassafras sandy loam
Cecil soil sandy clay loam
Davidson clay loam
Shontik-Casa Grande clay loam
Trix sandy clay loam
Trix-Casa Grande clay loam
Yolo loam
Capay silry clay
Sycamore silt loam
Dennis silt loam
Luray silty clay loam
Wooster silt loam
Vienna loam
Opal clay
Raub silt loam
Rockfield silt loam
Cisne
Muscatine
Avonberg
Matapeake silt loam
Evesboro low OM samd
Non-Soil Control
Adamsville, RI
Buttermilk Falls, PA
Hudson River, PA
Germantown, PA
Houston, TX
Houston, TX
Ontario, CA
Pacific Ocean
Dartmouth, MA
Newark, DE
#641
#643
#645
#654
#659
#843
#848
#850
Georgetown, DE
Georgetown, DE
Smith Island, MD
Newark, DE
DI Water Control
622
-------
Table 5. Chemical and physical characteristics of soils used in the RDX
spike and recovery studies
Soil
ID
101
106
108
109
110
116
117
123
126
128
Sand
34
93
31
37
88
14
52
11
15
22
Silt
46
4
15
12
6
35
30
27
46
68
Clay
20
3
54
51
6
51
18
62
39
10
pH
6.0
5.5
6.6
5.4
4.9
6.8
7.6
6.7
5.2
6.2
C*T
1.5
0.1
2.2
2.3
12.1
2.2
2.8
3.6
5.2
4.4
Soluable
Salts
(mmho/cm)
0.12
0.34
0.09
0.12
0.33
0.02
0.37
0.45
0.38
0.75
Fe-Oxide
(mg/kg)
5170
740
27875
25250
252
6870
5605
9010
4711
4286
Al-Oxide
(mg/kg)
1129
334
3265
1724
1086
595
332
725
939
593
CECb
(meq/lOOg)
9.1
2.0
8.3
13.5
24.8
34.1
23.8
36.0
41.2
29.5
a OM = Organic Matter content as derived by loss on ignition
b CEC = Total Cation Exchange Capacity at pH 7.0
Table 6. RDX soil spike and recovery study
Soil
ID
101
106
108
109
110
116
117
123
126
128
Non-Soil
Control
Average
101
106
108
109
110
116
117
123
126
128
Non-Soil
Control
Average
RDX
Spike
(ppm)
1
1
1
1
1
1
1
1
1
1
1
1
6
6
6
6
6
6
6
6
6
6
6
6
Mean RDX
Concentration
(ppm)
0.53
0.88
0.86
0.66
0.70
0.96
0.92
1.00
1.03
1.02
1.05
0.86
4.92
6.15
5.67
6.11
6.12
6.26
5.71
6.05
6.82
6.02
6.02
5.98
Standard
Deviation
0.19
0.13
0.23
0.22
0.14
0.12
0.42
0.45
0.25
0.18
0.13
0.23
0.54
0.84
1.09
0.93
0.46
1.21
0.72
0.80
0.33
0.62
0.83
0.75
Coefficient of
Variation
(%)
35
15
26
34
19
13
46
45
24
18
12
27
11
14
19
15
8
19
13
13
5
10
4
13
Recovery
(%)
53
88
86
66
70
96
92
100
103
102
105
86
82
103
95
102
102
104
95
101
114
100
100
100
623
-------
95
DIRECT ANALYSIS BY RAMAN SCATTERING: AN EMERGING TECHNOLOGY
FOR WASTE REMEDIATION
Charles K. Mann. Thomas J. Vickers, Department of Chemistry, Florida State University,
Tallahassee, FL 32306-3006
Abstract
The property of Raman spectroscopy of producing distinctive vibrational spectra by
measuring light scattered from a sample make it an attractive basis for direct analysis of
complex mixtures. Accordingly, an extensive study has been made with the aim of under-
standing the requirements for its routine application in chemical analysis. After examina-
tion of both dispersive and interferometer-based spectrometers, it became apparent that,
for chemical analysis, dispersive instruments using charge coupled device (CCD) detectors
and operating in the visible or near-infrared offer significant advantages in sensitivity.
We describe developments aimed at reducing Raman scattering measurements to auto-
mated routine practice. These involve several types of corrections of systematic errors that
occur. Detector sensitivity variation, wavelength dependence, read pattern signal and dark
signal must be compensated. In addition, corrections are made for wavelength calibration
error and for day-to-day variations in the sensitivity of system response. Procedures have
been worked out to support automated implementation of these corrections. There are two
stages. The first is a calibration of spectrometer wavelength response which is performed
on installation of a system. The second consists of several measurements that are made at
each measurement session to monitor and compensate for day-to-day changes in perform-
ance.
Limits of detection for quantitative analysis range from 0.005 to 5 percent, depending
upon the identity of the analyte. Limits for qualitative analysis are usually about an order
of magnitude higher than the quantitative limits. Examples are shown.
Raman measurements applied to several types of samples are illustrated. One group
consists of examples of measurements that were made on material that was removed from
the disposal tanks at the Hanford Site in Richland, WA. Another is a sequence of quanti-
tative measurements that were made on slurries of organics and inorganics. Another class
is taken from a study of analysis of paint films. A fourth type demonstrates the wide
dynamic range of the technique with a series of measurements in which a minor component
is extracted from the signal produced by a major component that is present in 1000-fold
excess.
Introduction
Raman spectroscopy is capable of dealing with certain types of samples that are difficult
to handle by most other techniques available to the analytical chemist. It works well with
solids, mixtures of solids and liquids and with nonvolatile solutes in solutions. Water does
not interfere with analyses, but it can be determined down to about the five-percent level.
Raman scattering produces a form of vibrational spectra. Accordingly, only substances
with some degree of covalent structure show Raman activity The selection rules differ
from those for infrared, however in compounds with more than a few atoms, both forms
generally produce bands at the same position, with differing intensities. From the perspec-
tive of chemical analysis, both types of spectra produce similar information. The spectra
are very distinctive. A molecular structure change as subtle as variation in the geometry
around a single carbon atom in an 800-Dalton molecule produces obvious changes in the
spectrum.
624
-------
In a mixture, it will ordinarily be true that any component that is present in excess of one
percent of the sample will make a visible contribution to the shape of the mixture
spectrum. The Raman effect is linear and it is possible to measure spectra under conditions
that preserve that linearity. Accordingly, the distinctive features of a one-percent constitu-
ent can usually be made the basis for a direct quantitative determination, irrespective of
the global composition of the sample. This works because the true limit of detection (LOD)
is ordinarily one to two orders of magnitude below one percent. With current technology
it varies from 0.005 to 5 percent, depending upon the analyte. In the usual case it is around
0.1 percent., with high LOD values being found for disordered polymeric materials, such
as water, the hydrous oxides and ammonium ion.
Unlike absorption spectra, Raman signals of both solids and liquids follow the same
relationships governing peak position and intensity. In general the peaks produced by
materials in the solid state are slightly more narrow than those for the same substance in
solution. Often they are slightly displaced along the abscissa. Therefore it is possible to
determine the solution and solid phase of a substance in a single sample.
The exact position and shape of peaks is a function of the chemical environment of the
sample, particularly the dielectric constant of a solvent or the counter ion of a solid salt.
Preparation of reference spectra should be done with that restriction in mind, but this does
make it possible to discriminate between different forms of salts. For example sodium,
potassium, barium and ammonium sulfates can be distinguished from each other and from
dissolved sulfate in a mixture. With attention to chemical environment, references pre-
pared in the laboratory can be used in subsequent field measurements.
The shape of spectral features of solids
is not a function of particle size. The
intensity of those features is affected
because particle size variations can al-
ter the average depth of penetration of
the exciting beam into the sample.
However, a substance in a strongly
Sample scattering sample produces the same
I spectrum that would be found in a
weakly scattering sample, providing
that the chemical environment remains
: the same.
Spectro-
meter
Zl
/- 1
Laser
Figure 1 Spectrometer block diagram.
Sample morphology affects the calibra-
tion of an analysis. Relevant factors are
particle size, color and dielectric con-
stant. Observations on paints has dem-
onstrated that dielectric constant is
considerably more important than the
other factors.
Description of the Instrument
The components of a typical system are illustrated in Fig. 1. The sample is illuminated by
a laser beam, usually brought in through an optical fiber. The laser is usually either green
light around 520 nm or near-infrared around 800 nm. The most commonly used one is the
argon-ion laser which has a convenient line at 514.5 nm. For field operation, the diode-
pumped doubled Nd:YAG laser offers a 532-nm line. It has the advantage of providing
625
-------
higher efficiency and smaller size than an argon-ion laser. Near infrared operation is
available by using stabilized external cavity diode lasers.
Light scattered by the sample is collected by a second fiber, in this case actually a group
of six symmetrically arranged around the excitation fiber to enhance collection efficiency.
The collected radiation is taken to the spectrometer for analysis.
For analytical applications, single stage spectrographs with 0.25...0.64-m focal lengths
are used. When these are fitted with 1000-element CCD detectors, the spectral window will
vary from around 1000...3000 cm"1 Spectral resolution of 8...10 cm"1 is appropriate for this
type of application. To achieve the stability that is required for routine mixture analysis,
it is necessary to control the temperature of the spectrograph to ±2°, so the use of a
temperature controlled enclosure is required for field operation. Remote control can be
arranged by using a telephone line to connect the host and remote computers.
Instrument Calibration
Emission spectroscopy, of which Raman scattering is an example, is not readily imple-
mented through two-channel design. A consequence is that operation of emission spec-
trometers has ordinarily been less convenient than absorption spectrophotometers. We
undertake to sidestep this disadvantage by computer control of the measurement, rather as
is done in FT spectrometry which is also inherently single channel.
The are several factors. Raman spectrometer outputs have conventionally had abscissa
measurements in wavelength. Conversion to wavenumber shifts by a process that is
invisible to the operator is desirable. This is accomplished by a one-time calibration of a
system at the time of installation.1 Some dispersive instruments do not record the entire
fundamental vibrational region in one spectral window. This is rarely much of a disadvan-
tage because one generally doesn't use the entire region in a single measurement. How-
ever, there is ordinarily no convenient method for relating signal intensity in one window
to that in another. This is overcome by the calibration procedure.
Current technology has developed sufficiently to make short term fluctuations in laser
intensity, detector sensitivity and amplifier response small enough to be ignored in most
applications. Once conditions are established in a measurement session, they will nor-
mally not show appreciable change. However, that is not true of day-to-day fluctuations
and single channel instruments do not provide the compensation that is achieved by
ratioing in two-channel design. This is overcome by the calibration procedure, based on a
chemical reference standard. CCD detectors exhibit both wavelength dependence and
pixel-to-pixel sensitivity variations. These are overcome by daily measurement of the
response of a standard lamp.
Vibrational spectra are highly distinctive, making it possible to do qualitative and quan-
titative analysis for a minor component of a mixture without having to be concerned for
spectral interference, so long as LOD criteria are met. However, this assumes that meas-
urements are not corrupted by abscissa error or by ordinate nonlinearity. In practice the
most important abscissa errors originate from calibration errors, rather than from non-
linear abscissa response. Accordingly, each-session calibration that is keyed to neon
emission line references can hold abscissa errors to less than 0.1 cm"1, which is good
enough to support mixture analysis.2
Several calibration measurements are therefore required at each measurement session. The
calibration process has been automated by designing calibration signals that can be used
on a day-to-day basis without requiring operator attention other than to observe a warning
signal generated by the host computer. To do this it is necessary to measure the calibration
626
-------
Computer-controlled
micropositioner
signals, under the control of the controlling
computer. The apparatus currently in use is il-
lustrated in Fig. 2.
Direct Analysis of Solid Waste Samples
As an illustration of the ability of Raman meas-
urements do deal with the kinds of complex sam-
ples that are encountered in waste remediation
work, we present some spectra that have been
recorded on waste material that was recovered
from the disposal tanks on the Hanford Site in
Richland, WA.
This material is housed in tanks with capacities
of as much as one million gallons. It consists of
residues from plutonium and uranium recovery
processes that have been in use since the 1940's.
The actual waste took the form of acidic solu-
tions, with organic phases at certain times.
These were neutralized with NaOH and concentrated by precipitation and evaporation to
leave material which is a basic mixture of solids and liquids. Although the amount and
identities of the original tank charges are known, the contents have been mixed during
operation of the tank farms with the result that the composition of material in any
particular tank can not be obtained from records.
As a part of the waste remediation process, the contents of some of the tanks are being
sampled by taking cores. These are brought to a laboratory and examined by chemical
analysis. The spectra that we present were obtained by direct measurements on some of
these cores by remote controlled fiber optic Raman sampling in a hot cell.3 Measurements
were made with one-minute to ten-minute integration times.
Entrance
Tungsten
Neon Lamp
Figure 2 Fiber optic signal switching.
800
1200
Delta Wavenumbers
1600
Figure 3 Nitrate region for waste tank material. The prominent peak is nitrate. No
correction for fluorescence.
627
-------
600
1000 1400
Delta Wavenumbers
1800
Figure 4 Nitrate region. A...C, three locations on the same core, showing variations in
composition. Curve D is a solid NaNOs reference spectrum. Ten minute exposures.
These are preliminary results, made to determine whether it is possible to observe signals
from individual compounds from this type of sample. No attempt has been made to identify
the peaks, other than those of nitrate, which is the major component.
An example is shown in Fig. 3. This is the spectral region that contains the nitrate
response. It was made with a two-minute integration. The silica response is strong in this
region and that signal component has been subtracted from the data. The curved back-
ground is mainly caused by fluorescence. Several peaks are evident, of which the strongest
is the 1067-cm"1 nitrate response.
In Fig. 4 we show measurements in the same spectral window, but at three locations on a
core from a different tank. These are 10-minute exposures. A reference spectrum of NaNOs
1600
2000
Delta Wavenumbers
2400
Figure 5 Cyanide region for a core. Both measurements made at approximately the same
position on the core, but on different days. Curve A, one minute, curve B, 10 minute
integration.
628
-------
2400
2800
Delta Wavenumbers
3200
Figure 6 CH region for the same core represented in Fig. 5. A. One minute integration.
B. Ten minute integration. Spectra recorded on separate days.
is shown for comparison. These spectra have had both the silica and fluorescence signal
components removed. Most of the peaks are common to all three regions. However,
varying peak ratios show that the the core is not is not homogeneous. One of the uses for
Raman measurements in this effort is to provide preliminary screening to guide the
sampling operation that is used to prepare samples for wet chemical analysis of core
material.
Figure 5 shows the CN region of a core. The spectra were recorded on different days but
at approximately the same location on the core. This gives some perspective on the
repeatability of the measurements. In this figure, the upper curve is a one-minute integra-
tion, while the lower represents a 10-minute exposure. The S/N is pretty much the same
for both. This reflects difficulty in positioning the probe in the sample. The operation was
carried out using remote manipulators with the sample several feet away from the operator.
Figure 6 shows two measurements on the same core that was used for the measurements
of Fig. 5. These were also made on different days, with one- and ten-minute integration
times. The spectrometer was moved to show the CH region. Again this demonstrates that
the visible features are reproducible spectral peaks. In addition, note that there is some
overlap between Fig. 5 and Fig. 6. Peaks that occur in the common region appear in the
expected positions.
Analyses of Slurry Samples
Raman scattering differs from other optical vibrational spectroscopy in that sample parti-
cle size has no direct effect on the shape of the spectrum. Accordingly, it does not make
a qualitative difference if the sample is changed from a solid to the same material
dispersed as a slurry. Similarly, the presence of the solid particles in the slurry does not
affect the shape of the spectral features of the liquid components. Changing from a clear
solution to a slurry does affect the intensities of the responses because it alters the
effective depth of penetration of the laser beam into the sample.
The upshot of this is that reference spectra that are made on the individual sample
components can be used in analysis of slurries, providing that calibration shifts caused by
variations in beam penetration are taken into account. As an illustration of this capability
629
-------
A.
1400
we present some results from a study of the
heterogeneous catalytic conversion of
quadricyclane to norbornadiene.4 The reac-
tants are dissolved in chloroform. The reac-
tion is catalyzed by dispersed solid copper
sulfate.
In Fig. 7 the spectra of the individual compo-
nents are shown. The region of principal in-
terest is from 850 to 1150 cm"1 In Fig. 8 the
spectra of the reactant and product are shown
along with spectra of the reaction mixture
taken at the start and near the end of the
reaction. This is a complex system. The
Raman signals due to the probe and the sus-
pended CuSO4 powder almost hide the
strongest features of the reactant and prod-
uct. The chloroform line at 1216 cm"1, which
we wish to use as an internal standard, over-
laps significant features of CuSO4 and the
two organic compounds. This was dealt with
by using least-squares fitting for quantifica-
tion because it handles spectral overlap effi-
1000 1200
Delta Wavenumbers
Figure 7 A. Chloroform. B. 0.1 M Quadri- ciently. The fit values were used in a stand-
cyclane in CHCh. C. 0.1 M Norbornadiene ard addition procedure for which one of the
in CHCh. D. CuSCu powder. E. Raman calibration curves is shown in Fig. 9. As is
Spectrum
probe.
of silica from the fiber optic
pointed out in the paper cited, a plot of the
responses of the reactant and product vs.
time shows the expected decrease of one and
growth of the other.
900
Delta Wavenumbers
mMoles Norbornadiene
Q HIIVJLUICS rNuiuuiuauiciic A
Figure 8 A. Quadricyclane. B. Starting Added
reaction mixture. C. Ending reaction mix- Figure 9 Standard addition curve.
ture. D. Norbornadiene
Aging of Paint Films
Since Raman measurements are based upon examination of scattered light, it is possible
to do chemical analysis on the surfaces of totally opaque samples. As an example we cite
some results from an accelerated aging test that was run on a water based paint.5 After a
normal drying time, paint films were exposed for varying periods to UV, primarily in the
275...350 nm range. Although there is no perceptible change in color during the period of
630
-------
these tests, measurable changes in spectra of the pigments occurred in less than three hours
oi exposure.
The paint was blue. It contained a proprietary film material, together with four principal
pigment materials: titanium dioxide, carbon black, copper phthalocyanine and carbazole
dioxazme violet. The film produced a relatively weak spectrum which did not have much
effect on these measurements. TiO2 produces strong peaks which are outside the window
of interest and carbon black produces only a weak and featureless response Accordingly
the spectra were dominated by the two organic pigments. This is illustrated in Fig 10 with
the spectra of the individual pigments and that of the dried paint. The paint film appears
to be the sum of the spectra of the pigments. Actually, if they are removed, one is left with
a recognizable spectrum of the binding polymer.
Figure 11 shows the result of an exposure test. There are readily apparent changes in the
relative intensities of several peaks. Two pairs are marked on the figure. The intensity of
the 1529-cnT peak of copper phthalocyanine, marked c on the figure, increases relative
to the 1341-cm" peak of the same compound, a, and also relative to the 1391-cm"1 peak
of carbazole dioxazine violet, b. Peak c represents the vibration of a functional group that
is photochemically unreactive. In fact, its absolute intensity varied little during the period
of the test and it was used as an internal standard.
Measurement of these peaks demonstrated that both of the pigments are degraded by UV
exposure. The carbazole violet response decayed more rapidly than the phthalocyanine
response.
Dynamic Range
Chemical analysis of mixtures generally involves detecting and measuring minor compo-
nents in the presence of major components which may not be of interest. The process
usually boils down to exclusion of the major component signal in order to make the minor
component visible. There are a great many ways of doing this. When the phenomenon is
1529 cm'1
1000
1200 1400 1600
Delta Wavenumbers
Figure 10 A. Copper Phthalocyanine. Figure 11 Effect of exposure to UV. A.
B.Carbazole Dioxazine Violet. C. Dried Unexposed. B. After 3 hours. C. 25 hours.
paint film. D. 86 hours. E.100 hours.
631
-------
800 1000 1200
Delta Wavenumbers
Figure 12 1 M NaNOs
800 1000 1200
Delta Wavenumbers
Figure 13 0.1 M (NH4)2SO4 in 1 M NaNO3
800 1000 1200
Delta Wavenumbers
Figure 14 0.01 M (NH4)2SO4 in 1
800
Figure 15
NaNO3
1000 1200
Delta Wavenumbers
0.001 M (NH4)2SO4 in 1 M
800
1000
Delta Wavenumbers
1200
800
1000
Delta Wavenumbers
1200
Figure 16 O.I M (NH4)2SO4 in Water
Figure 17 0.1 M (NH4)2SO4 in 1 M NaNO3
800 1000 1200
Delta Wavenumbers
Figure 18 0.01 M (NH4)2SO4 in 1 M
NaNOs
800 1000 1200
Delta Wavenumbers
Figure 19 0.001 M (NH4)2SO4 in 1 M
NaNOs. Sulfate peak marked.
632
-------
linear, it can be done by simple subtraction. In the case of Raman scattering, the phenome-
non is linear, providing that the chemical environment in the sample is held constant.
Limits on the dynamic range are set by the apparatus, at the present time primarily by
detector response.
To demonstrate this we present measurements on solutions that are 1 M in NaNO3 and 0.1,
0.01 and 0.001M in sulfate. A nitrate reference and the sulfate containing samples are
shown in Fig. 12... 15. In these, the Raman spectrum of silica from the fiber optic sample
is appreciable and has been subtracted from the sets. Figures 17... 19 are Fig. 13... 15 after
subtraction of the nitrate signal. Figure 16 shows 0.1 M sulfate for comparison.
After removing 1 M nitrate from one tenth to one thousandth that concentration of sulfate,
the sulfate is visible in all cases. In Fig. 19, scale expansion is sufficient to reveal two
artifacts which were produced by the small degree of nonlinear response that is present in
this system. This indicates that the practical dynamic range is around 1000:1 with current
equipment.
Summary
In this paper we have demonstrated that Raman scattering is capable of supporting direct
chemical analysis on a variety of types of samples. The examples have been solid mixtures,
slurries of solids suspended in liquids and opaque surface layers. In addition we demon-
strate a dynamic range of around three orders of magnitude. When applied to quantitative
determination of target substances, the limit of detection will range from 0.005...5 per-
cent, with most falling around 0.1 percent. Measurements are made without sample prepa-
ration and normally require from one to ten minutes. Complete systems can be obtained
for around $80,000 at the present, may weigh around fifty pounds and can be housed in a
ten cubic foot enclosure. Remote operation is feasible. This is a technology that is
sufficiently mature to be used in routine applications.
Literature Cited
1 C.-H. Tseng, J.F Ford, C.K. Mann and T.J. Vickers, "Wavelength Calibration of a
Multichannel Spectrometer," Appl. Spectrosc. 47, 1808-1813 (1993).
2. C. Shen, T.J. Vickers, and C.K. Mann, "Abscissa Error Detection and Correction in
Raman Spectroscopy,M/>/7/. Spectrosc. 46, Ill-Ill (1992).
3 D.R. Lombardi, C. Wang, B. Sun, A.W Fountain, III, T. J. Vickers, C. K. Mann, F
R. Reich, J. G. Douglas, B.A. Crawford and F. L. Kohlasch, "Quantitative and Qualitative
Analysis of Some Inorganic Compounds by Raman Spectroscopy, Appl. Spectrosc. In
Press, 48, (1994).
4. J. F Ford, C. K. Mann and T. J. Vickers, "Monitoring the Heterogeneously Cata-
lyzed Conversion of Quadricyclane to Norbornadiene,^/?/?/. Spectrosc, In Press 48 (1994).
5. B. Sun, T.J. Vickers and C. K. Mann, Unpublished Work.
633
-------
96
INTEGRATED SAMPLE AUTOMATION FOR
CONCENTRATION, EXTRACTION, AND CLEANUP
MARK CAVA Zymark Corporation, Zymark Center, Hopkinton, MA 01748
The maturing environmental services industry is characterized by a
commodity situation where it is difficult for buyers to distinguish sellers.
Oversupply, a new national administration, and a flat period for government
contracts have forced lab industry consolidation. To compete, management
must lower costs and improve margins. The easiest way to do this is to
apply sample automation to standard operating procedures. Implementing
sample automation technology requires change. To make change easy
companies let their most important assets, their best people, drive sample
automation projects. These companies achieve growth by automating to
obtain excellent QC, lowest cost, and fast delivery. For these leading labs,
automation is a solution for improved competitiveness and profitability in a
changing market.
Innovative concentration workstations have revolutionized laboratories by
providing consistent results, low cost per sample, and twice the throughput.
Payback periods of less than 3 months are normal for the automated
concentration of trace organics. Related to sample concentration is the
increasing use of solid phase extraction(SPE) as the medium for
simultaneous extraction and concentration.
A unique automated SPE workstation takes the advantages of solid phase
extraction, low solvent costs and extraction selectivity, and combines them
with the efficiency and control of automation. Workstation extraction
provides improved consistency for large volume SPE, fast turnaround, and
payback within 6 to 12 months. Even when extraction and concentration is
done, many crude isolates or extracts are too laden with high molecular
weight compounds or coellutant interferences. This adds cleanup as a
necessary step in sample processing.
Companies using multimethod robotic workstations for sample cleanup see
more than 60% annual return on investment, 60 % improved consistency,
and a dramatic 60% improvement in sample turnaround. Payback periods
for mutimethod cleanup automation ranges from 6 months to 18 months.
Workstation automation of sample concentration, extraction, and cleanup for
drinking water, groundwater, waste water, and soil will be discussed.
Results are for compounds regulated under RCRA, NPDES, and SDWA.
Sample automation for concentration, extraction and cleanup will be
examined. PCB, GPC (8080 and 8270 compounds), and Florisil Pesticide
cleanup (a GPC companion for 8080 compounds) will be discussed. The
automation methods, GC results, HPLC results and payback are examined.
634
-------
97
ANALYSIS OF ASH RESIDUES FOR VOLATILE ORGANIC
CONSTITUENTS BY USING A MODIFIED SW-846 METHOD 8260
Richard P. Grese. Sr. Chemist, Tennessee Eastman Division, Marsha D. Edwards, Sr.
Technician, Tennessee Eastman Division, Eastman Chemical Company, Kingsport Tennessee
37662
ABSTRACT
Application of SW-846 Methods 8240 or 8260 for volatile organic analysis (VGA) of some ash
residues is limited because of suppression of spiked analytes by the adsorptive matrix. The
suppression of internal standard recoveries produces large biases in the quantitation of analyte
concentrations. In addition, suppression of spiked analytes prevents the use of normal QA/QC
criteria for validation of the data. A slightly modified Method 8260 is proposed to minimize the
biases and to improve the data quality of the VOC analysis of ash residues. The modification of
Method 8260 described here allows most laboratories to improve the quality of he data without
extensive development of a new method or without resorting to a more expensive approach such
as "isotope-dilution" GC/MS.
In this work, the adsorptive behavior of specific analytes for a given matrix were determined, and
analytes were assigned to internal standards with similar adsorptivity. Five internal standards and
five surrogates were recommended for the analysis of fly ash samples. The modified method was
used to evaluate fly ash from two sources by determining spike recovery acceptance criteria and
biases for each analyte studied. Of the analytes studied, recoveries (based on the new internal
standard assignments) were generally between 85 to 120%. Matrix-induced decomposition of
1,1,2,2-tetrachloroethane by one type of boiler fly ash is reported.
INTRODUCTION
Regulations permitting the burning of hazardous waste in a boiler or industrial furnace require
that the resulting residue (e.g. fly ash) be analyzed for contaminants that are in the waste or that
may be generated by products of incomplete combustion. The residue may be excluded from the
definition of a hazardous waste if the analysis demonstrates that the hazardous waste did not
significantly affect the residue. The regulations also require that the analyses be in conformance
with procedures in SW-846.1 The analyses of ash residues for trace level volatile organic
constituents (VOC's) represent one of the most difficult challenges to the environmental
laboratory.
SW-846 methodology for volatile organic analysis (VGA) of solid wastes cannot be reliably used
for the analysis of some ash residues, especially those containing a large percentage of activated
carbon. The adsorptive capacity of these matrices for volatile organic compounds effectively
suppress the recovery of spiked analytes and thus limit the quality and usefulness of the purge-
and-trap analysis. The suppression of spiked analytes in effect prevents the use of normal
QA/QC criteria for validation of the data. The data reviewer or regulator must be
knowledgeable of the matrix interactions and how they affect data quality prior to using the final
data.
635
-------
In the VOC analysis, the analyte must partition initially from the solid matrix into the water prior
to being purged out of the water and onto the trap. If the analyte adsorbs more strongly to the
solid matrix than the water, the analyte will not be available for purging onto the trap and
recovery is poor. The adsorptive behavior of an analyte is thought to be related to various
physical and chemical properties of both the ash and the analytes, and may be quite different for
ash samples collected from different sources. The poor recovery of spiked analytes appears to
be directly proportional to boiling point, and has been suggested to be a result of the adsorptive
potential and the adsorptive surface area of the ash, the temperature, physical characteristics of
the analyte, and the amount of time during which the analytes are in contact with the ash matrix.2
Prior attempts by this laboratory to increase partitioning of volatile analytes into the water have
had limited success.3 Procedural changes include: use of smaller sample size, increased heating
of sample during purging, and addition of a matrix modifier to the water to compete with active
binding sites on the fly ash.
The poor recovery of internal standards from the ash matrix is an especially serious problem as
it may produce a large bias in the data when the adsorptive behavior of a specific analyte is
different to that of its assigned internal standard. This bias can be either positive or negative
depending on the measured amount of the internal standard. The actual recovery of an analyte
from the matrix, as determined by comparison to external standards, is also not an attractive
quantitation alternative as the low recoveries of some compounds will under-estimate the
contamination of the fly ash. External standard quantitation methods, in addition, provide no
correction for changes of instrument sensitivity during the analysis.
The low "actual" recovery of analytes (0 to 10% for some compounds), however, may not be
critical for quantitation if an internal standard method is used which corrects for the bias. This
is essentially the approach of an "isotope-dilution" GC/MS method in which isotopically-labelled
standards of every analyte are added to each sample, thus continually correcting the calibration
for each sample's matrix effects. Ideally, a method employing isotope-dilution GC/MS
methodology is an attractive approach for analyzing volatile organics in an adsorptive matrix such
as fly ash. However, the cost of procuring isotopically-labelled standards for all analytes inhibits
most laboratories from this approach.
The goal of this research was to develop a method, only slightly modified from SW-846 Method
8260, that improves the quality of the analysis of boiler fly ash. A simple way to minimize the
biases resulting from poor recoveries is to assign internal standards based on adsorptivity to the
matrix rather than GC retention time as is usually done. The new method was developed by first
determining the adsorptive behavior of specific analytes for a given matrix, and then assigning
the analytes to internal standards with similar adsorptivity. Fly ash generated by two boiler types
were included to investigate how differences in the matrix affect the data generated by the
modified method.
EXPERIMENTAL
Reagents and Procedures. The deionized water used in these experiments was purified by
passing through a Millipore (Bedford, MA) Milli-Q filtering system and activated carbon prior
to use. All GC/MS standards, unless otherwise noted, were obtained from Supelco (Bellefonte,
PA) as certified standard solutions. Fluorobenzene and pentafluorobenzene were obtained as neat
636
-------
Three compounds need additional comment in Sample-B. Trichloroethene (164%),
tetrachloroethene (164%) and 1,1,2,2-tetrachloroethane (16%) gave percent recoveries quite
different from those from Sample-A. It is known the 1,1,2,2-tetrachloroethane is converted to
trichloroethene by a based-catalyzed dehydrohalogenation reaction.4 Similarly, tetrachloroethene
may be formed by loss of hydrogen from 1,1,2,2-tetrachloroethane. Ten replicate samples of
Sample-B were prepared and loaded onto the autosampler and analyzed sequentially. Figure 2
shows the recoveries of the three analytes overtime as well as the total (combined) recoveries.
This data points to decomposition of the 1,1,2,2-tetrachloroethane as the possible source of the
increased amounts of the two other analytes. Calcium hydroxide present in the fly ash is thought
to be responsible for the decomposition of 1,1,2,2-tetrachloroethane; addition of HC1 may be
useful to prevent this reaction.
CONCLUSIONS
Analysis of fly ash with SW-846 Methods 8240 or 8260 is difficult because of adsorption of
analytes onto the ash. The poor recovery of internal standards is especially troubling due to the
biases it creates in the quantitation of analytes. The modification of Method 8260 described here
allows most laboratories to improve the quality of the data without extensive development of a
new method. It also allows the laboratory to develop data quality objectives for recovery of
matrix spikes and surrogates. Once these objectives are determined, a more reasonable and
accurate assessment of volatile organic constituents in the ash can be made. A disadvantage of
this approach is that most laboratories can not devote the time to develop data quality objectives
for every adsorptive matrix. However, if a large number of analyses of a specific matrix type
is anticipated, development of a specific method may be a worthwhile endeavor. We have
developed a method on a "worst case" matrix, speculating that the method would be useful for
less adsorptive matrices as well.
By definition, an "internal standard" should not be affected by method or matrix interferences.
Therefore, strictly speaking, the modified method described here should not be termed an
"internal standard method." We have, however, used the terminology here to contrast the
quantitation method in which external standards are used. Our approach has been to select
"internal standards" which respond similarly to the matrix as the analytes assigned to them. In
this regard, the approach is more like an "isotope-dilution method" without the use of labelled
analytes. Realistically, as long as purge-and-trap is used for extraction of fly ash there continues
to be opportunities for improvement in this method.
Future work in this laboratory will include the investigation of method detection limits (MDL's)
and the application of this method to other ash matrices.
ACKNOWLEDGEMENTS
The authors are grateful to Mr. Craig Hoyme for the preliminary work with analysis of fly ash,
and to Dr. Darrel Wilder for his support of this research.
637
-------
TABLE 1
GC/MS INSTRUMENTAL PARAMETERS
Purge-And-Trap
Instrument: Tekmar LSC2000
AutoSampler: Tekmar ALS2016 with Tekmar Sample Heater
Trap: Supelco VOCARB 4000 adsorbent trap, 30.5 cm x 0.125" OD x 0.105" ID
(Carbopack C, Carbopack B, Carboxen 1000, Carboxen 1001)
Method 2:
Sample Temp: 40°C
Purge Preheat: 3 min.
Purge Time: 11 min
Desorb Preheat: 245°C
Desorb Time: 2 min
Desorb Temp: 250°C
Bake Time: 4 min
Bake Temp: 260°C
Transfer Line Temp: 80°C
Purge Flow: 40 mL/min
Gas Chromatograph
Instrument: Hewlett-Packard 5890 Series II
Column: Rtx-5, 60 m, 0.25 mm ID, 1.00 /x df (Restek)
Oven Temperature: 35°C for 2 min, then 35-200°C @ 10°/min, 6 min hold
Oven Temperature (for BFB) : 100°C for 0.5 min, then 100-220°C @ 10°/min
Injector: Split 25:1 (EPC constant flow)
Injector Temperature: 200°C
Transfer Line Temperature: 280°C
GC/MS Interface: Direct
Carrier Gas: He (1 mL/min)
Mass Spectrometer
Instrument: Hewlett-Packard 5972MSD
Scan Limits (Low): 35 amu
Scan Limits (High): 300 amu
Scans/Second: 2
638
-------
TABLE 2
PERCENT RECOVERY FROM FLYASH
Spiked at 4 mg/kg (n=5)
External Standard Calculation
No.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
COMPOUND Sample-A Sample-B
lodomethane
1 ,4-Dioxane
Isobutanol
Methylene chloride
Acrylonitrile
Bromochloromethane
Carbon tetrachloride
Chloroform
Acetone
1 ,2-Dichloroethane-d4
Bromomethane
Dibromomethane
2-Butanone (MEK)
1,1-Dichloroethene
1,1,1-Trichloroethane
Methacrylonitrile
Bromodichloromethane
2-Bromo-1 -chloropropane
Trichloroethene
Benzene-d6
Benzene
Crotonaldehyde
Fluorobenzene
Bromoform
1 ,4-Difluorobenzene
Tetrachloroethene
1 , 1 ,2,2-Tetrachloroethane
Pentafluorobenzene
Toluene-d8
Toluene
cis-1 ,4-Dichloro-2-butene
Ethylbenzene-d10
Chlorobenzene-d5
Chlorobenzene
4-Bromofluorobenzene
1 ,2-Dichlorobenzene
123
106
105
105
97
94
90
90
89
88
87
87
86
84
83
81
80
61
59
55
54
48
38
36
30
25
24
19
19
18
14
8
7
7
2
0
136
104
93
71
89
102
102
106
77
98
99
101
104
103
101
98
99
93
145
91
90
79
85
76
77
119
17
71
69
72
51
54
53
55
29
15
639
-------
o>
-&.
o
o
SAMPLE-A FLY ASH
EXTERNAL STANDARD CALCULATIONS
PERCENT RECOVERY
140
Group 1 (75-100%)
Group 2 (45-75%)
Group 3 (25-45%)
15 20 25
COMPOUND NUMBER
-------
TABLE 3
ASSIGNMENT OF INTERNAL STANDARDS
AND SURROGATES
Percent Recovery of Fly Ash
(External Standard Calculations)
IS
ss
ss
IS
ss
IS
IS
ss
ss
IS
No.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
COMPOUND
lodomethane
1 ,4-Dioxane
Isobutanol
Methylene chloride
Acrylonitrile
Bromochloromethane
Carbon tetrachloride
Chloroform
Acetone
1 ,2-Dichloroethane-d4
Bromomethane
Dibromomethane
2-Butanone (MEK)
1,1-Dichloroethene
1 ,1 ,1-Trichloroethane
Methacrylonitrile
Bromodichloromethane
2-Bromo-1 -chloropropane
Trichloroethene
Benzene-d6
Benzene
Crotonaldehyde
Fluorobenzene
Bromoform
1 ,4-Difluorobenzene
Tetrachloroethene
1 ,1 ,2,2-Tetrachloroethane
Pentafluorobenzene
Toluene-d8
Toluene
cis-1 ,4-Dichloro-2-butene
Ethylbenzene-d10
Chlorobenzene-d5
Chlorobenzene
4-Bromofluorobenzene
1,2-Dichlorobenzene
Sample-A
123
106
105
105
97
94
90
90
89
88
87
87
86
84
83
81
80
61
59
55
54
48
38
36
30
25
24
19
19
18
14
8
7
7
2
0
Range
75-100%
45-75%
25-45%
1 0-25%
0-1 0%
641
-------
TABLE 4
PERCENT RECOVERY FROM FLYASH SAMPLE-A
Spiked at 4 mg/kg (n=10)
Internal Standard Method
Analytes
Surrogates
COMPOUND
Bromomethane
Acetone
1,1-Dichloroethene
Acrylonitrile
lodomethane
Methylene chloride
2-Butanone (MEK)
Methacrylonitrile
Chloroform
Isobutanol
1 ,1 ,1-Trichloroethane
Crotonaldehyde
Benzene
Carbon tetrachloride
Trichloroethene
Dibromomethane
1 ,4-Dioxane
Bromodichloromethane
Toluene
Tetrachloroethene
Chlorobenzene
Bromoform
cis-1 ,4-Dichloro-2-butene
1 ,1 ,2,2-Tetrachloroethane
4-Bromofluorobenzene
1 ,2-Dichlorobenzene
1 ,2-Dichloroethane-d4
2-Bromo-1 -chloropropane
Fluorobenzene
Toluene-d8
Ethylbenzene-d10
% Rec.
95
84
89
101
148
118
90
89
96
110
89
91
98
85
108
95
108
85
100
80
94
70
105
89
24
6
94
116
73
104
105
s.d.
7
10
7
6
12
43
5
4
5
15
7
5
2
7
3
3
9
5
6
4
5
6
29
10
4
2
2
4
2
6
8
Range
75
54
69
82
113
-10
77
76
81
66
69
75
91
66
99
84
81
70
83
69
80
52
19
60
12
-1
88
105
66
87
81
(3 sigma)
114
114
109
120
183
246
104
102
111
154
108
107
104
105
117
105
134
100
116
92
108
88
190
119
35
12
99
126
80
121
128
642
-------
TABLE 5
QUANTITATION BASED ON INTERNAL STANDARDS
CONCENTRATION COMPARISON (SAMPLE-A)
AVERAGE % RECOVERY (N=3)
COMPOUND
Bromomethane
Acetone
1,1-Dichloroethene
Acrylonitrile
lodomethane
Methylene chloride
2-Butanone (MEK)
Methacrylonitrile
Chloroform
Isobutanol
1,1,1-Trichloroethane
Crotonaldehyde
Benzene
Carbon tetrachloride
Trichloroethene
Dibromomethane
1 ,4-Dioxane
Bromodichloromethane
Toluene
Tetrachloroethene
Chlorobenzene
Bromoform
cis-1 ,4-Dichloro-2-butene
1 ,1 ,2,2-Tetrachloroethane
4-Bromofluorobenzene
1 ,2-Dichlorobenzene
1 ,2-Dichloroethane-d4
2-Bromo-1 -chloropropane
Fluorobenzene
Toluene-d8
Ethylbenzene-d1 0
2 PPM
91
121
80
111
135
198
97
89
93
132
80
103
97
68
118
93
114
76
97
152
92
188
66
227
12
0
93
138
152
101
100
4 PPM
90
93
84
105
140
149
91
90
92
120
83
90
99
81
109
95
113
81
97
132
93
140
119
159
21
5
94
119
139
103
101
5 PPM
88
106
84
115
141
177
101
97
94
128
85
104
99
84
106
96
125
83
95
121
94
120
85
137
22
5
95
114
133
101
100
QA
75
54
69
82
113
-10
77
76
81
66
69
75
91
66
99
84
81
70
83
69
80
52
19
60
12
-1
88
105
66
87
81
Range
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
(1)
114
114
109
120
183
246
104
102
111
154
108
107
104
105
117
105
134
100
116
92
108
88
190
119
35
12
99
126
80
121
128
(1) Acceptance Criteria from Table 4.
643
-------
TABLE 6
PERCENT RECOVERY FROM FLYASH SAMPLE-B
Spiked at 4 mg/kg (n=10)
Internal Standard Method
Analytes
Surrogates
COMPOUND
Bromomethane
Acetone
1,1-Dichloroethene
Acrylonitrile
lodomethane
Methylene chloride
2-Butanone (MEK)
Methacrylonitrile
Chloroform
Isobutanol
1 ,1 ,1 -Trichloroethane
Crotonaldehyde
Benzene
Carbon tetrachloride
Trichloroethene
Dibromomethane
1 ,4-Dioxane
Bromodichloromethane
Toluene
Tetrachloroethene
Chlorobenzene
Bromoform
cis-1 ,4-Dichloro-2-butene
1 ,1 ,2,2-Tetrachloroethane
4-Bromofluorobenzene
1,2-Dichlorobenzene
1 ,2-Dichloroethane-d4
2-Bromo-1 -chloropropane
Fluorobenzene
Toluene-d8
Ethylbenzene-d10
% Rec.
94
77
100
87
127
95
108
100
103
96
97
93
99
99
164
98
103
96
101
155
102
85
64
16
58
32
95
102
100
97
102
s.d.
4
7
2
7
9
24
14
8
2
13
3
12
1
4
6
3
11
3
2
3
1
5
12
9
4
5
2
1
1
1
2
Range
83
55
93
65
101
24
66
77
97
58
89
57
96
88
145
89
71
88
95
147
100
70
28
-10
46
16
88
99
97
93
97
(3 sigma)
106
99
107
109
152
166
150
122
110
134
106
128
102
110
183
107
134
104
108
163
104
100
101
41
69
48
102
106
102
101
106
644
-------
O)
Ji.
O1
3
Q
SAMPLE-B FLY ASH
1,1,2,2-Tetrachloroethane Reaction
200
PERCENT RECOVERY
150 -
100
50
0
0
Q-
-e-
468
SAMPLE NUMBER (TIME)
10
Trichloroethene
1,1,2,2-TCE
Tetrachloroethene
Total (Combined)
12
-------
REFERENCES
1. U.S. Environmental Protection Agency "Test Methods for Evaluating Solid Waste -
Physical/Chemical Methods". Office of Solid Waste and Emergency Response. July
1992, SW-846, Third Edition, Final Update I. Method 8260, Revision 0.
2. O'Quinn, C.M.; Roudebush, W.; Kuehn, J.D.; Shmookler, M.; Thomas, F.; Hoffman,
K.D. " A Multi-Laboratory Determination of Method Detection Limits and Practical
Quantitation Limits for EPA Regulated Volatile Organics in Incinerator Ash."
Proceeding of the Sixth Annual Waste Testing and Quality Assurance Symposium,
Volume I, p. 1-302. (1990).
3. C.A. Hoyme; Grese, R.P., unpublished results.
4. March, J., "Advanced Organic Chemistry": McGraw-Hill: New York NY. 1977; 935-
936.
646
-------
98
PCB FIELD EXPERIENCE USING THE D TECH™ PCB IMMUNOASSAY FIELD SCREENING TEST.
James M. Melby, Development Group Leader, Brian Finlin, Scientist, Strategic
Diagnostics Incorporated, Newark, DE, 19713, Mary Knowles, Project Manager, and
Tracy Young, Environmental Scientist, Jaycor Environmental, Vienna, VA 22182.
ABSTRACT
PCB testing by immunoassay is gaining acceptance by environmental contractors
through validation testing against accepted analytical methods. Though proven as a
valid quantitative methodology in medical diagnostics for over 20 years, this
technology has only recently been accepted by the EPA as a valid field screening
method for PCB. The D TECH™ PCB field screening test kit incorporates immunoassay
technology into a user friendly, quick, accurate, and field portable system. Using
this technology, over 100 samples were tested in a site evaluation conducted to
establish the usefulness, convenience, and cost effectiveness of the D TECH system.
The procedure consists of the following 8 steps: 1. sample extraction 2. sample
filtration 3. sample dilution 4. reaction of sample with immunoreagents 5.
application of reaction mixture to filtration/ detection device 6. wash addition 7.
color reagent addition 8. read result. This test detects all of the most commonly
found aroclors (1254, 1242, 1248, 1260, 1262, 1268) equally, while the detection of
aroclor 1232 is 5-fold less than 1254, and the detection of 1016 and 1221 is about 10-
fold less than 1254. A sampling grid design was set up for a 1 acre site that was
suspected to have PCB contamination due to its history of use as a storage site for
transformers. A total of 117 samples were collected at 50 foot intervals in most cases,
attempting to evenly cover the entire site. Most samples were collected at both 1 and
2 foot depths. The samples were tested on site using an EPA-accepted test along with
the D TECH method, and 38 selected samples were sent to two analytical laboratories
for SW846 method 8080 GC analysis. Results indicated very low level contamination
(<1 ppm) at randomly located sites with all three methods, and most samples (88) were
less than the minimum detectable limit for all methods. The EPA-accepted test had
13% false positives and no false negatives when compared to GC and the D TECH
method had 9 % false positives and no false negatives. The cost for the D TECH test is
$30.00 per sample, which for this study amounted to $3,510.00 for the screening
method. Sending 10% of the field screened samples in addition to those indicated as
positive by the screening method (a total of 22 samples) for GC analysis cost an
additional $3,520.00 at $160.00 per sample. Therefore the total cost for the screening
event was $7,030.00. GC analysis of all 117 samples would have cost $18,720.00 at
$160.00 per sample. In addition, the field screening offers same day results whereas
the GC results typically take from 4-6 weeks. A savings of over $11,000.00 in addition
to the time savings offered by using the D TECH PCB field screening test on site
demonstrate the significant advantages of this method.
647
-------
99
IDENTIFICATION OF HIGH MOLECULAR WEIGHT BIOGENIC
N-ALKANES, N-ALKANOLS, N-ALKANALS, AND PLANT STEROLS
IN ENVERONMENTAL SAMPLES AND DETERMINATION OF
LEE RETENTION INDICES BY GC/MS EQUIPPED
WITH ELECTRONIC PRESSURE CONTROL
Paul H. Chen, William A. Van Ausdale, Athanasios Tomaras, and Dwight F. Roberts
Analytical Services Division, Environmental Science & Engineering, Inc.
P.O. Box 1703, Gainesville, FL 32602
ABSTRACT
Soil samples were analyzed by GC/MS according to EPA methods for semivolatile organics. High
molecular weight odd carbon number n-alkanes (C^ to C^), even carbon number 1-alkanols (C^ to C32)
and n-alkanals (C^, to C32), plant sterols, and pentacyclic triterpenes were found to be present in some
of the soil samples. The constituents and the relative quantities of alkanes, alcohols, and aldehydes are
similar to that found in plant waxes reported in the literature. This indicates that these compounds are
from plant material in the soil. It is important to know that these high molecular eight compounds are
biogenic not anthropogenic. The major plant sterols found are /3-sitosterol, stigmasterol and campesterol.
These sterols are 24 a-alkylsterols with a ring double bond between C-5 and C-6 positions. It has been
reported in the literature that vascular plants principally contain these sterols. Lee retention indices (RI)
were determined for the compounds identified in this study. In order to shorten the run time, GC
equipped with electronic pressure control was used for analysis. The RIs provide the elution order and
position for sterols and also for the homologs of n-alkanes, n-alkanols, and n-alkanals which are very
useful for compound confirmation from GC/MS analysis.
INTRODUCTION
In the EPA methods for the analysis of semivolatile organics by GC/MS, samples are generally analyzed
for the target compound list (TCL) components by an automated data system and for non-TCL
components by a library search of a published mass spectral data base. The non-TCL components are
reported as tentatively identified compounds (TIC) or unknowns. In our analysis of thousands of
environmental samples for non-TCL components, it was not unusual that we found n-nonacosane and n-
hentriacontane, and to a lesser extent, n-heptacosane and n-tritriacontane, in the soil samples. These high
molecular weight n-alkanes are not from crude oil or petroleum products because they are all odd carbon
number whereas n-alkanes in the petroleum product should not show any odd-to-even carbon
predominance. In addition to these high molecular weight n-alkanes, we often found high molecular
weight n-alkanols, n-alkanals, plant sterols, and pentacyclic triterpenes in some of these samples. These
classes of chemicals have been reported in leaf cuticles (1), leaf surfaces (2), road dust (3), and ambient
aerosols (4,5). Presumably many soil samples we have analyzed contained plant materials which
contributed to these chemicals in the samples. The objectives of this study were to identify these high
molecular weight compounds by GC/MS and to determine their Lee retention indices (RI) by GC/MS
equipped with electronic pressure control. The Lee RI can provide a means of confirmation of these
compounds along with mass spectral data.
Lee retention indices, based on a series of four polycyclic aromatic hydrocarbons (PAHs) as retention
index standards, have been reported by Lee et al. (6) and Vassilaros et al. (7) for a large number of
polycyclic aromatic compounds. These Lee RIs are determined using capillary columns GC operated
648
-------
under temperature programming conditions. Rostad and Pereira (8) reported Lee RIs determined by
GC/MS for a large number of PAHs and other organic compounds of environmental interest. We
reported last year in this symposium that high reproducibilities of Lee retention indices determined under
different GC conditions can only be achieved if compounds of interest and the RI standards such as
chrysene and benzo(g,h,i)perylene are eluted during the temperature ramping period. Using the
conventional GCs which are normally equipped with the constant pressure control, the column oven
temperature programming rate has to be decreased to about 5°C/min in order to have benzo(g,h,i)perylene
elute during the temperature ramping period. This will result in a long run time (about 60 min) which
is not practical in a routine analysis. Since the compounds of interest in this study are all high molecular
weight compounds and mostly elute between chrysene and benzo(g,h,i)perylene, we chose to run these
samples by GC equipped with electronic pressure control (EPC) which maintains a constant column flow
(9). Using EPC we can shorten the run time to 27 minutes and still achieve the elution of
benzo(g,h,i)perylene in the temperature ramping period.
EXPERIMENTAL SECTION
Sample Preparation. Soil samples were extracted with methylene chloride in Soxhlet extractors according
to EPA SW-846 Method 3540/8270 (10). Before extraction, each sample was spiked with 1.0 mL of
surrogate spiking solution which contains 100 /ig/ml each of acid surrogates and 50 /*g/ml each of
base/neutral surrogates. Methylene chloride extract was concentrated to 1 ml with Kuderna-Danish
concentrator and analyzed by GC/MS. Prior to GC/MS analysis, the extract was spiked with 20 ^1 of
an internal standard mixture which contains 2 jig//il each of l,4-dichlorobenzene-d4, naphthalene-d8,
acenaphthene-d10, phenanthrene-d10, chrysene-d12, and perylene-d12.
GC/MS Analysis. Sample extracts and standards were analyzed on HP5972 MSD connected to an
HP5890 Series II GC which was equipped with an EPC split/splitless injection port. The column used
was a 30m X 0.25mm i.d. DB-5MS (25 fim coating) fused silica capillary column (J & W Scientific,
Folsom, CA). A starting temperature of 50°C was used. The pressure was pulsed by ramping from 18
psi to 98.9 psi at a rate of 99 psi/sec, held for 0.3 min, then ramped back to 18 psi at 99 psi/sec. The
column flow at 18 psi and 50°C was 2.2 ml/min which was maintained at a constant flow for the
remainder of the run. The column temperature was held at 50°C for 2 min then programmed at 20°C/min
to 120°C and then to 320°C at 10°C/min and held isothermal at this final temperature for 3 min. Under
this programming benzo(g,h,i)perylene (the last RI standard) elutes during the temperature ramping
period. The mass spectrometer was scanned from 35 to 500 amu at a rate of 2 scans/sec. A forward
library search was performed for non-TCL compounds on a Wiley/NBS data base which contains 139,000
different spectra (11). Compounds were tentatively identified by library search or by elucidation of the
compound structure from its mass spectrum if no match was found in the library. The tentatively
identified compounds were confirmed if possible by agreement of the mass spectra and retention times
between the sample components and the authentic compounds. The retention time of the tentatively
identified compound or the standard was used for the calculation of Lee RI according to the equation
described in the following paragraph.
Calculation of Lee Retention Indices. The Lee retention indices are calculated according to the following
equation:
RI = 100 (Tx - Tz) + 100 * Z
Tz+i -Tz
649
-------
Where: Tx is the retention time of compound of interest
Tz is the retention time of the preceding RI standard
Tz+i is the retention time of the following RI standard
Z is the number of rings in the preceding RI standard
The retention index standards used are naphthalene (RI=200.00), phenanthrene (RI=300.00), chrysene
(RI=400.00), and benzo(g,h,i)perylene (RI=500.00). When these compounds are not found in the
sample, their retention times are calculated by adding the differences in the retention times between the
first three RI standards and their corresponding deuterated internal standards in the daily calibration
standard to the retention times of the corresponding internal standards in the sample.
RESULTS AND DISCUSSION
A total ion chromatogram of a soil sample that contains high molecular weight biogenic compounds is
shown in Figure 1. This chromatogram does not show all the biogenic compounds we have identified
in this paper. However, it does show five classes of compounds from biogenic source discussed in this
paper. These five classes of compounds are n-alkanes, n-alkanols, n-alkanals, plant sterols, and
pentacyclic triterpenes. The presence and distribution of these compounds are somewhat different from
sample to sample presumably due to the presence of different plant materials in the soil sample.
N-Alkanes. The alkanes found are all odd carbon number n-alkanes from Qj to C,3. Nonacosane
and hentriacontane (C3i) were found most frequently and most abundantly. Heptacosane and
tritriacontane were also found in the sample. These odd carbon number alkanes are not from petroleum
products, they are from plant material in the soil which contains plant waxes. Epicuticular plant waxes
are known to have n-alkanes of the odd-to-even carbon number predominance. The distribution of the
odd carbon number C^ to C33 n-alkanes found in the soil samples is similar to the distribution reported
in the literature for the leaf surfaces abrasion products (2), road dust (3), and ambient aerosols (4,5).
N-Alkanols. The alcohols found in the soil samples are all even carbon number 1-alkanols from C^ to
C32 generally with C^, C^, and C^ the most predominant. Similar predominancy of these even carbon
number n-alkanols have been reported by other scientists in the leaf surface abrasion products (2) and
road dust (3). This indicates that these alcohols are from plant material in the soil which contains
epicuticular plant waxes. Coelutions of C^ alkane with C& alcohol, C27 alkane with C^ alcohol, and CN
alkane with C^ alcohol in the DB-5 column make the quantitation of these alcohols and alkanes somewhat
difficult. Figure 2 shows the mass spectra of C& alkane, Cw alcohol, and C^ aldehyde. As shown in
this figure, C^ alkane has high intensity peaks at m/z 57, 71, and 85 with ratios of 57/55, 71/69, and
85/83 calculated from the tabulated form of spectrum to be 3.7, 3.6, and 3.6, respectively. On the other
hand, the ratios of these three pairs of peaks for C^ alcohol are 1.0, 0.93, and 0.47, respectively.
Furthermore, C^ alcohol has high intensity peaks at m/z 83, 97, and 111 with ratios of 83/85, 97/99,
and 111/113 equal to 2.1, 5.0, and 6.1, respectively. Similar ratios should be observed for other high
molecular weight alkanes and alcohols. Using this information, one can estimate the percentage of the
alcohol and the alkane in the coeluting peak.
N-Alkanals. The aldehydes found in the samples are even carbon number n-alkanals from C^ to C32.
The most abundant aldehydes are generally C^, C^, C^, and C32. This distribution is similar to that
reported in the literature (2,3). Again these aldehydes are likely from vascular plant waxes in the soil.
As shown in Figure 2, the mass spectrum of n-alkanal normally shows strong peaks at m/z 82 and 96.
Other less intense even mass peaks are also present in the spectrum. For high molecular weight n-
650
-------
alkanal, the molecular ion is usually absent. However, [M-H2O] and [M-H2O-C2H4] ions are occasionally
present. Since no authentic C^, to Qj aldehydes are available for confirmation, chemical ionization (CI)
spectra were obtained for one of the samples. The CI spectra gave [M+H] and [M+C2H5] ions for the
aldehyde which confirm the suspected aldehyde.
Plant Sterols. As shown in Figure 1, a few common plant sterols were found in the soil sample which
contains plant wax constituents. 0-Sitosterol (stigmast-5-en-3-ol or 24a-ethylcholesterol) was found most
frequently and most abundantly, next came stigmasterol (stigmata-5,22-dien-3-ol) and stigmast-4-en-3-one.
Stigmast-4-en-3-cone is not a common steroidal ketone in the plant, it is probably from the oxidation of
0-sitosterol to the corresponding 5-en-3-ketone, then isomerization to a more stable 4-en-3-ketone which
is stigmast-4-en-3-one. The oxidation and isomerization may occur by the microbial enzymes in the soil.
Compesterol (ergost-5-en-3-ol), a 24-a-methylcholesterol, was also found in the soil sample. Less
commonly, a trace amount of cholesterol was sometimes found to be present in the sample. The mass
spectra of four common plant sterols found in the soil are shown in Figure 3. The plant sterols found
in the soil sample are dominated by 24a-alkylsterols and with a ring double bond between C-5 and C-6
positions. The plants principally with these sterols have been designated as category I-A (12). Most
vascular plants belong to category I-A (12).
Pentacyclic triterpenes. Several pentacyclic triterpenes have been found in the soil samples. We are only
able to obtain two of the authentic compounds, i.e., lupen-3-one and friedelin, to confirm their presence
in the samples. The mass spectra of lupen-3-one (lup-20(29)-en-3-one) and friedelin (D:A-friedooleanan
or friedelin-3-one) are shown in Figure 4. Other pentacyclic triterpenes found in the soil samples but not
confirmed by the authentic compound include D-friedoolean-14-en-3-one (taraxerone with M. Wt. of 424)
and taraxerol methyl ether (M. Wt. 440). Oleanolic acid and orsolic acid were found in the fine
paniculate abrasion products from leaves by Rogge et al. (2), but were not found by us probably because
they are not amendable to analysis by GC/MS without derivatization of the cabroxyl group.
Other Classes of Compounds. Hexadecanoic acid, and to a lesser extent, octadecanoic acid, were often
found in the soil samples. 9-Hexadecenoic acid, 9-octadecenoic acid, lower alkanoic acids (Cj2-C16) and
higher alkanoic acids (C^ to C^) were sometimes found in the soil samples. Part of the lower alkanoic
acids are probably from microbial lipids. The free fatty acids found in the soil are even carbon number
n-alkanoic acids which are likely from the plants material present in the soil. Other biogenic compounds
found to be present in the soil samples include dehydroabietic acid, vitamin E, and squalene.
Dehydrobietic acid is a resin acid which is known to be present in conifers.
Lee Retention Indices. Table 1 shows the comparison of Lee RI of PAHs determined by the conventional
constant pressure GC and constant flow GC. Constant flow GC is achieved by using EPC injection port.
As shown in Table 1, Lee RI values determined by constant pressure and constant flow agree very well.
This indicates that Lee RI data determined by the conventional GC method such as those reported by Lee
et al. (6), Vassilaros et al. (7) and Rostad and Pereira (8), can be applied to RIs determined by GC/MS
equipped with EPC. We reported last year in this symposium that high reproducibilities of Lee RI
determined under different GC conditions can only be achieved if compounds of interest and the RI
standards such as chrysene and benzo(g,h,i) perylene are eluted during the temperature ramping period.
Since the compounds of interest in this study are all high molecular weight compounds,in order to shorten
the analysis time and still achieve the Lee RI values that can be reproduced by other labs, GC/MS
equipped with EPC was used for analysis.
651
-------
Lee Retention Indices of Biogenic Compounds. Lee RIs of biogenic compounds found in the soil are
listed in Table 2. These RI values can not be reproduced under different GC conditions with as good
reproducibility as that shown in Table 1. This is because in Table 1 the compounds of interest and the
RI standards are all PAHs, the changes in their chromatographic retention behavior under different GC
conditions should be similar. The RIs listed in Table 2 provide the elution order and position for plant
sterols and also for the homologs of n-alkanes, n-alkanols, n-alkanals, and n-alkanoic acids which are
very useful for compound confirmation from GC/MS analysis.
ACKNOWLEDGMENTS
We thank P. Dumas and his department for extracting the samples and S. Keeran, D. Schindler, and H.
Adcock for their assistance in GC/MS analysis. We also thank Dr. Dave Powell of University of Florida
for obtaining the CI mass spectra for us.
REFERENCES
1. Riederer, M., Schneider, G., Planta. 1990. 180, 154.
2. Rogge, W. F., Hildemann, L. M., Mazurek, M.A., Cass, G. R., Simoneit, B. R. T., Environ.
Sci. Technol.. 1993. 27, 2700.
3. Rogge, W. F., Hildemann, L. M., Mazurek, M. A., Cass, G. R., Simoneit, B. R. T., Environ.
Scie. Technol.. 1993. 27, 1892.
4. Simoneit, B. R. T., Mazurek, M. A., Atmos. Environ.. 1982. 16, 2139.
5. Mazurek, M. A., Cass, G. R., Simoneit, B. R. T., Environ. Sci. Technol.. 1991. 25, 684.
6. Lee, M. L., Vassilaros, D. L., White, C. M., Novotny, M., Anal. Chem.. 1979. 51, 768.
7. Vassilaros, D. M., Kong, R. C., Later, D. W., Lee, M. L., J. of Chromatog.. 1982. 252, 1.
8. Rostad, C. E., Pereira, W. E., J. High Resolution Chrom. and Chrom. Commun.. 1986. 9, 328.
9. Hermann, B. W. et al., J. High Resolution Chrom., 1990. 13, 361.
10. USEPA Test Methods for Evaluating Solid Waste: Physical/Chemical Methods, SW-846, 3rd
Edition, 1986.
11. McLafferty, F. W., Stauffer, D. B. The Wilev/NBS Registry of Mass Spectral Data. Wiley: New
York, 1989.
12. Nes, W. R., McKean, M. L., Biochemistry of Steroids and Other isopentenoids, University Park
Press, Baltimore, 1977.
652
-------
TABLE 1. COMPARISON OF LEE RETENTION INDICES OF PAH'S DETER-
MINED BY CONSTANT PRESSURE AND CONSTANT FLOW
Compound
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Indeno(l ,2,3-cd)pyrene
Dibenzo(a,h)anthracene
Benzo(g,h,i)perylene
Constant
Pressure
RI
300
301.59
344.83
352.72
398.53
400
442.34
443.05
454.06
492.36
493.42
500
Constant
Flow
RI
300
301.83
344.43
352.60
398.72
400
442.24
443.37
454.02
492.09
493.22
500
653
-------
TABLE 2. LEE RETENTION INDICES FOR BIOGENIC n-ALKANES,
1-ALKANOLS, n-ALKANALS, PLANT STEROLS, n-ALKANOIC
ACIDS, AND PENTACYCLIC TRITERPENES
Compound
RI
Pentacosane 402.40
Heptacosane 430.84
Nonacosane 456.71
Hentriacontane 480.80
Tritriacontane 503.44
Docosanol 402.41
Tetracosanol 430.85
Hexacosanol 457.26
Octacosanol 482.07
Triacontanol 505.21
Dotriacontanol 528.82
Docosanal 393.81
Tetracosanal 421.90
Hexacosanal 448.83
Octacosanal 473.88
Triacontanal 497.50
Dotriacontanal 520.10
Dodecanoic acid 264.11
Tetradecanoic acid 293.86
Hexadecanoic acid 326.41
Octadecanoic acid 356.78
Eicosanoic acid 385.02
Docosanoic acid 412.52
Tetracosanoic acid 439.96
Cholesterol (Cholest-5-en-3|3-ol) 484.43
Campesterol (24a-methylcholesterol) 496.58
Stigmasterol (A5'22-24a-ethylcholesterol) 498.71
P-Sitosterol (24a-ethylcholesterol) 506.48
Stigmast-4-en-3-one 520.89
Lupen-3-one 514.53
Friedelan-3-one (Friedelin) 534.44
Vitamin E 484.26
Squalene 449.71
654
-------
1800-,
16001
14001
-
o 12001
8
g 1000l
c
•§
§ 8001
6001
400-
200-
-
-
o
o
c
£.
CM g
7 « 1
<0 T- <
Q> i OS
£• go
5 t,-:
Q. 1 -c
I
«
O
r4-r
\
o
IT ^.
1 1
~" 5
1 s
o w
8 c
<
o 4
1 3
i I
i t 2
jo £-0
? 3 1
s *
8 i a | i
fl fl i 9 1 o
JJLjJuLJ u^Juu
J
i-A-^,
I o
-^
<
"
L°
/
1 e
8
<
o g
1 N
3 |
fiX
A /i
1 ft
vL \ .A
p
=— Stigmaster
-Alkane
\.
a
w
3
w
CO.
f
/
(D
c
P 9
£-- co
2 <=
0 "?
9 ,
r
*? -1-
r-
m '
! j
CM (J
^•"
^
to
g>
c/D
y,.
vu\r
H ' u v
Vuy \Jy v
1 1 j F 1 1 1 | 1 1 1 1 |
d
5)
5
£
(1)
9
=• Stigmast-3
-riedelin
y 7
W /LA
* v ^^
i r ~"i i
19.0
20.0
21.0
22.0
23.0
Time (min.)
24.0
25.0
26.0
27.0
Figure 1. Total ion chromatogram of a soil sample which contains high molecular weight
n-alkanes, n-alkanols, n-alkanals, plant sterols, and pentacyclic triterpenes.
655
-------
100,
8
c
CO
T3
C
D
< 50.
CD
"CD
DC
0
100
CD
o
c
CO
•o
c
D
rf 50.
^>
CD
"CD
DC
0
100
CD
o
c
CO
C
$ 50.
CD
'is
CD
DC
0
40
40
I
I
I
40
(
13
W
|
I |
e
,
30
57
6
50
57
6
I
1
L
i""
50
71
80
f
9
I
|
80
8
9
I
I
80
85
I, ,
1C
)3
i
1C
2
9
1C
A
n-Nonacosane
99
M+
I
I |J || Y ! 183 | 211 225 408
1 '"'I '"1 1 1 1 1 1 1 1 ""I i — i — | | r..l,....|
0 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
B
)7 1 -Octacosanol
111
125
I 139153 (M-H20)+
I If ll ' ™ I f.11 3P2
0 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
C
n-Octacosanal
6
(M - H2O - C2H4)+
/
/
109 / (ft/] - \-\yO) +
1 123 ' 2 /
| ll n 152 180 250 292 334 362 390
I I I I 'I I I"" I I"" I I I I "I I
0 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
Figure 2. Mass spectra of n-nonacosane (A), 1-octacosanol (B), and n-octacosanal (C).
656
-------
100., 43
1 WW_
co
o
c
CO
73
3
£ 50.
.1
To
V
a:
0
A
p-Sitosterol
57
,1 ll
11
TTJT ..,.. .!....,...-!.
An en an
1
,,
i 213
M+
I i ill I III Y3 , T T T "
llilLliJillLtlklil.,,.!,,,! ,,. J,. . I, L i , .
|,
""i"1'1'! i
r H
"1
40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
100.,
100^
I'l 11 T i T [ 1111 M 111111111111111 iTirrr i rpri'i'i iri'l'i'i nriTi 11 j'lyn im'|'i i I'T'I'I i T I [Vt'i'i i [• i'i in; iTI 1111 I'T [Ti i't i'1'i'i T 111 n'h 111 ] 111111111111 n 11 T 11 ft n 11 111 n'l 111 r; i n r i i'm-|-f 1111
60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
124 m/Z
O
o
CO
c
z>
50.
JO
CD
CC
43
55
I I
67 1
1 ll
|
9
||
C
Stigmast-4-en-3-on<
229
f 107 147 M+
III J I III 289 412
i . ll In 187 ,l 370 ,Q, 4,1Z
I III I HI I II II II I I T,,,,L.
40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
l'.| | l'| | | | |iiii|iii'i'|.iiT|'iiii| |ii'ii|'mi|ni.| |iiii|iiii|iHli|iiii|iiii|iiii|iiii|iifl|ini|n
40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
Figure 3. Mass spectra of four common plant sterols found in soil samples: (3-sitosterol (A),
Stigmasterol (B), stigmast-4-en-3-one (C), and campesterol (D).
657
-------
100
CD
O
c
CO
TJ
C
^
.a
< 50.
CD
ro
CD
cc
o
1
1
1
'"II'PIIII
A
69 /I
'I'"
Friedelin
35
1
. 1
123
|
163
i 205
I , 1?S I 218 273
1 HI || J Ml2!6 *
,1.11 „ 1 302 341
, i i||i ,ii|q
1 l|l|f l|l l|l|| 1 1 l| 'I II l< 1 1
40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
100^
Q)
O
CO
CD
DC
50.
0
I
6
I
I
7
I
|
|
81
I
95
I I
09
1;
I
J
>1
4
13
3
149
I
161
I,'
11
1
75
I I
2
39
I
|
B
Lupen-3-one
| 2|8 ^ 313 M^
„,„ 424
4°9 I
I 1 I' '!• \- If, ' f l| '!• ' T H 1
40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420
m/z
Figure 4. Mass spectra of two common pentacyclic triterpenes found in soil samples:
friedelin (A) and lupen-3-one (B).
658
-------
100
RAPID DIOXIN SCREENING BY ENZYME IMMUNOASSAY
Robert O. Harrison. ImmunoSystems, Inc., Scarborough, ME 04074
Robert E. Carlson, Ecochem Research Inc., Chaska, MN 55318
Hamid Shirkhan, Fluid Management Systems, Inc., Watertown MA 02172
Theresa Keimig, Midwest Research Institute, Kansas City MO 64110
Wayman E. Turner, Centers for Disease Control, Atlanta, GA 30341
ABSTRACT
A system has been developed for rapid screening of 2,3,7,8-TetraChloroD.ibenzo-p-
Dioxin (TCDD). The system uses a competitive inhibition finzyme immunoAssay (EIA)
based on a mouse monoclonal antibody which is specific for TCDD and related
congeners. Sample preparation can be performed with a programmable automated
extraction and cleanup system which uses disposable Teflon clad columns. The
extraction and cleanup system has been extensively validated by GC-MS for a
variety of sample types. The sample preparation system allows immunoassay
analysis of soil, serum, water, and other matrices by taking each sample type to the
same sample preparation endpoint. Concentration factors and endpoint conditions
are completely flexible and programmable. Immunoassay analysis is performed by
the addition of a prepared sample extract in organic solvent to an antibody coated
microwell containing an aqueous sample diluent. This is mixed and incubated for 30
minutes to allow the immobilized antibody to capture analyte from the sample. The
liquid is then removed and the well is washed to remove unbound materials. The
well is then incubated with a competitor-HRP conjugate capable of binding
specifically to the antibody sites not occupied by TCDD. After 30 minutes, the
unbound conjugate is washed away and enzyme substrate is added for color
development, the color generated is directly related to the amount of competitor-HRP
bound in the second step, which is inversely related to the amount of analyte bound
in the first step. After 30 minutes, a stop solution is added and the developed color
is read on a microplate reader. The total time required for the EIA analysis of a
prepared extract is less than 2 hours. Sensitivity for TCDD is better than 0.1
ng/well, allowing sensitive analysis of a variety of environmental matrices.
Preliminary results indicate that it is possible to detect 10 ppb 2378-TCDD in soil by
direct analysis of crude soil extracts. Work is being directed toward simplification
of the extraction procedure and improvement of the interface with the automated
sample cleanup system. The data presented here demonstrate that this system should
be useful for TCDD screening in many situations for a variety of matrices. The
system offers significant improvements in speed, sample throughput, and cost
compared to GC-MS.
INTRODUCTION
Reagents and Standards
The EIA for PCDD/F's uses the mouse monoclonal antibody DD3, which has been
described previously (1). Competitors which bind specifically to the dioxin binding
site of DD3 were conjugated to horseradish peroxidase (HRP) to make conjugates
659
-------
which can be captured by the immobilized DD3 antibody. The competitor-HRP
conjugates were tested for PCDD/F sensitivity and solvent and matrix tolerance.
Standard preparation was as follows, based in part on prior work by Sherry et al.
(2). PCDD/F standards in toluene or nonane were diluted in the same solvent in
silanized glass vials. A small volume of DMSO was added to each vial, then the
toluene or nonane was evaporated under a nitrogen stream. The DMSO was diluted
with an equal volume of methanol and the standard was mixed vigorously, then
sonicated for 15 minutes. Analysis was as described in the EIA procedure below,
by adding the DMSO/methanol solution to aqueous diluent in the EIA well.
PCDD/F EIA Procedure and Interpretation of Results
Following is the procedure for the EIA analysis of PCDD/F's: 1) mouse antibodies
which recognize the dioxin structure are immobilized on the walls of plastic
microwells; 2) PCDD/F's in solvent, in the form of either standards or samples
prepared as described above, are mixed with Assay Diluent in the wells, allowing
PCDD/F's to bind to the immobilized antibodies; 3) unbound sample is washed
away with water; 4) competitor-HRP conjugate is added and allowed to compete
with the captured analyte for the limited PCDD/F binding sites on the immobilized
antibodies; 5) unbound conjugate is washed away with Water, leaving an amount of
conjugate on the immobilized antibodies inversely related to the amount of PCDD/F's
that were present in the sample; 6) enzyme substrate is added to the wells for
color development by the bound enzyme. Tne intensity of color is proportional to the
amount of bound enzyme and is inversely related to the amount of PCDD/F's present
in the sample. Therefore, more color means less PCDD/F's. Total run time is
approximately 2 hours per test and up to 40 samples can be run in a single batch.
The optical density (OD) of each standard and sample well is measured and sample
PCDD/F concentrations are calculated based on the standard curve.
RESULTS AND DISCUSSION
Method Sensitivity
The use of heterologous haptens or competitors for improving immunoassay
sensitivity has been described previously in detail (3) and has been exploited here
using a well studied anti-dioxin antibody. Three heterologous competitors were
compared to the homologous competitor. The results of a single experiment (Figure
1) snow that the improvement in sensitivity obtained was greater than an order of
magnitude for one competitor and approximately one order of magnitude for the other
two competitors. Conjugates Ic and 2a were selected for further characterization.
Sensitivity to 2378-TCDD in all subsequent experiments with both Ic and 2a has
been better than 100 pg/well.
Test Specificity
The specificity of DD3 antibody has been described previously (1) and is primarily
directed toward selected tetra- and pentachlorodibenzodioxins, with reduced
recognition of the corresponding furans. This recognition profile corresponds roughly
660
-------
to the I-TEF values given in reference 4. Competitive inhibition tests were performed
using a complex mixture of dioxins and furans (Table 1) to assess the possibility of a
change in specificity due to the use of heterologous competitors. This mixture
contains compounds comprising the full range of recognition by DD3 antibody in the
system used by Stanker et al. (1), as well as TEF values covering three orders of
magnitude. The results shown in Figure 2 suggest that the specificity of DD3
antibody with competitor Ic does not differ significantly from the previously
established pattern. Similar results were seen for competitor 2a.
Interface of EIA with Automated Sample Cleanup System
The system described is capable of analyzing samples in a variety of solvents, but
has been designed to accommodate any sample that can be exchanged from a
volatile hydropnobic solvent into a non-volatile hydrophilic solvent. This allows the
analysis of any sample prepared by standard methods such as the FMS Dioxin-
Prep™ System for Automated Sample Cleanup (5). Preliminary experiments indicate
no interference in the EIA using a fully cleaned extract from the FMS system.
Ultimate method sensitivity is therefore determined primarily by sample size,
concentration factor, and interference from the concentrated matrix.
Soil Spiking and Extraction
The possibility of a rapid extraction and analysis for PCDD/F's in soil was
investigated as follows. Aliquots of 5 g of soil were weighed into silanized glass
extraction vials and air dried overnight. Soils were spiked by adding a toluene
solution of 2378-TCDD directly to the soil surface at multiple sites. After mixing the
soil and air drying 30 minutes, 3 steel BB's and 5 ml of solvent were added to each
vial. Methylene chloride and toluene samples were prewetted with a minimum
amount of acetone before adding the other solvent. Vials were sealed with Teflon
lined caps and soil samples were extracted by shaking for thirty minutes at 300 rpm
on an orbital shaker. The extracts were clarified by centrifugation for 15 minutes at
1-2000g. An aliquot of each extract was removed to a silanized vial, DMSO was
added, and the volatile solvents were removed by evaporation under a nitrogen
stream. Methanol was added to each sample and further handling and analysis
followed the procedure described above for standards. The data or Table 2 show
that 2378-TCDD spikes may be recovered and detected with a relatively simple
procedure, but that not all solvents will give adequate results. Toluene appears to
extract soil components which give strong false positive interference in the EIA,
while DMSO appears to give inadequate recovery. Neither methylene chloride nor
hexane:acetone gave significant false positive interference and both gave
acceptable recovery values. These results will form the basis of further experiments
directed toward a rapid soil extraction procedure for low level analysis of
PCDD/F's in soil.
661
-------
CONCLUSIONS
1. The test is capable of analyzing for PCDD/F's in less than 2 hours from prepared
extracts, using very little specialized equipment.
2. The heterologous competitor strategy employed here demonstrates significantly
improved sensitivity.
3. The specificity of the test appears to parallel the previously established profile for
DD3 antibody.
4. The design or the EIA accommodates significant variations in sample preparation,
allowing the analysis of PCDD/F's in many matrices.
5. Ongoing work with this kit includes improvement of the rapid soil extraction
procedure and validation for a variety of sample matrices.
Table 1. Composition of Native Standard and Toxic Equivalent Concentrations
Congener pg/ul Used TEF* TEC**
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3, 7,8, 9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD
Total
* TEF = toxic equivalency factor,
4
4
20
20
20
20
20
20
20
20
20
20
20
20
20
40
40
348
from reference 4
0.1
1
0.05
0.5
0.5
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.01
0.01
0.01
0.001
0.001
TCC _ TCrM
0.4
4
1
10
10
2
2
2
2
2
2
2
0.2
0.2
0.2
0.04
0.04
40.08
662
-------
Table 2. Recovery of 2378-TCDD from Soil by Four Extraction Methods. One soil
was spiked at 10 ng/g with 2378-TCDD and recovery was compared to a standard
evaporated onto a bare vial and recovered with DMSO. Eacn extraction method
and the no soil control included an unspiked sample for evaluation of the matrix
effect. Each value is the mean for three replicate wells in one EIA run.
Extraction Solvent Matrix Effect* Recovery of Standard**
DMSO 92 36
toluene 56 167
dichloromethane 108 91
hexane:acetone (1:1) 120 94
* optical density (OD) of unspiked soil as a percent of the no soil/no TCDD
control OD in the EIA
** percent recovery of 10 ng/ml soil spike relative to spike recovered from vial
with no soil
REFERENCES
1. Stanker, L.H., et al. "Monoclonal Antibodies for Dioxin: Antibody Characterization
and Assay Development"; Toxicoloav (19871 45:229-243.
2. Sherry, J.P., et al., "Use of DMSO as Solubilization Agent in the Detection of
2378-TCDD by Radioimmunoassay"; Chemosphere (1990) 20:1409-1416.
3. Harrison, R.O., et. al., "Hapten Synthesis for Pesticide Immunoassay
Development". Chapter 2 in Immunoassays for Trace Chemical Analysis,
Monitoring Toxic Cnemicals in Humans, Food, and the Environment; ACS
Symposium Series Vol. 451; Vanderlaan, M., Stanker, L.H., Watkins, B.E. and
Roberts, D.W., Eds., American Chemical Society: Washington, DC., 1990.
4. Kutz, F.W., et al., "International Toxicity Equivalency Factor (I-TEF) Method of Risk
Assessment for Complex Mixtures of Dioxin and Related Compounds";
Chemosphere (1990) 20:751-757.
5. Turner, W.E., et al., "An Evaluation of the FMS Dioxin-Prep™ System for
Automated Sample Cleanup Adapted to Human Serum"; Dioxin 1992 Conference,
Volume 8.
663
-------
•4 -\ f\ _,
1 1 U
100 *
P
'w
£ a ft
CD o U
•o
I 70 -
cL
o
o 60
o _ _
CD
•S A n _
to mj
O>
-------
80-r
7 n
c n
"55
c c n
Q
•c
~K
C
1
c
g
c
+•
n
s
"2
+•
I
I
Q n
o U
»
>
In
U
1
Figure 2
A--
r~~~-
—
A--
— - — .
' V
~-~,
\
As
IM
s native standard ( as mass)
v
\
\
^
\
s
\
\\
native standard (as TEC)
\
\
\
\
J2378-TCDD
v^~
\
\
\
\
\
V
\
\
?
\
0.0001 0.001 0.01 0.1 1 10
ng/well 2378-TCDD equivalent
"--•
4
100
Comparison of EIA Standard Curves of 2378-TCDD and a 17 Congener
Native Standard Mixture. HRP conjugate 1c and DD3 antibody were used
to detect both 2378-TCDD and native standard (Table 1). The native
standard response is expressed both as actual total mass and as toxic
equivalent concentration (TEC) according to Table 1 .
665
-------
101
EXTRACTION OF ENVIRONMENTAL ANALYTES USING A CARBON MEMBRANE
Steve Miller and Gary L. Nixon, Alltech Associates, 2701 Carolean Industrial Drive,
State College PA 16801
ABSTRACT
Membrane extraction disks are increasingly being used in solid-phase extraction of
aqueous environmental samples to take advantage of their ability to separate
pollutants from an aqueous matrix faster and with a lower solvent usage than
alternative techniques. Current methods generally cite membrane disks which employ
a reverse phase adsorbent (usually C18). We have developed a membrane using
graphitized carbon black as the adsorbent, combining the advantages of the
membrane extraction with the efficency of carbon . By embedding graphitized carbon
black in an inert support, we are able to demonstrate improved extraction of polar
analytes while maintaining high recoveries of non-polar compounds. These results
demonstrate that carbon membrane extraction provides a means to apply this
technology to a wider range of analytes than was previously possible.
666
-------
102
A PRELIMINARY INVESTIGATION OF RETENTION TIMES AND
ANALYTE RECOVERIES WHEN USING HIGH EFFICIENCY
GEL PERMEATION CHROMATOGRAPHY CLEANUP
E. E. Conrad, N, L. Schwartz, K. P. Kelly, ABC Instruments (a division of
Laboratory Automation, Inc.), Columbia, MO 65202
INTRODUCTION
Gel Permeation Chromatography (GPC) is widely used to clean up extracts that are
contaminated with various higher molecular weight coextractives. Data published by
other workers (T. Willig, J. Kauffman, 9th Annual WT & QA Conference, 1993,
Paper #65) demonstrate that GPC cleanup of sample extracts protects analytical
systems and preserves data quality by preventing the entrance of matrix coextractives
into sensitive equipment. High efficiency GPC employs smaller gel particles which
provide higher GPC resolution. Packing the small particle resin into short, stainless
steel columns decreases processing time and reduces solvent consumption relative to
larger, traditional, glass-barrelled GPC columns.
Establishing that retention properties observed during GPC cleanup are the same for
high efficiency versus traditional columns will allow analytical laboratories to reduce
waste generation and save money by decreasing solvent consumption while increasing
sample processing speed. Results reported here compare retention times and analyte
recoveries for some semivolatile target analytes chromatographed on traditional or
high efficiency columns and assess column matrix handling capacity.
PROCEDURES
Spiked solvent aliquots (5 mL) were injected and eluted through an Envirosep-ABC
high-efficiency GPC cleanup column set (23 mm i.d. x 410 mm bed length, 5 mL per
minute flow rate) by using an ABC Instruments Autoprep 1000 GPC. A UV detector
(254 nm) was used in-line when retention times were measured. The calibration
solution used was prepared per the directions provided within proposed EPA Method
3640A. Analyte recoveries were calculated by GC analysis. GC response for the
collected fraction, which was evaporated to a final volume of 5 mL, was compared
to GC response for the same analytes using the standard solution injected in the GPC.
Matrix loading capability was assessed by injecting 5 mL of solvent containing some
amount of non-volatile matrix coextractive, such as corn oil or potting soil extract.
The dumped eluent fraction was collected, transferred to a tared aluminum pan and
evaporated to dryness in a fume hood. The residue was weighed and the result was
compared with the amount that had been injected.
667
-------
EXPERIMENTAL PROGRAM
Representative subsets of analytes were chosen from target lists given in semivolatile
organic analysis methods. GPC cleanup is well known and has little tendency to
capture or destroy target analytes. While high-efficiency columns differ slightly from
GPC columns suggested in EPA methods, the differences are of a physical rather than
a chemical nature, therefore lower recoveries are not expected. However, since
analyte retention time is based primarily on physical factors, relative retention times
of analytes may shift. The analyst must be able to recover analytes from the collected
fraction using the calibration procedure that is provided in the method (or a workable
adaptation of it). Therefore, a principal goal of these experiments was determining
elution order for a number of common target analytes with high-efficiency columns
and comparing results obtained when using traditional (glass-barreled, low-pressure)
columns. In particular, analytes having elution times which are retarded (relative to
elution times based solely on size exclusion behavior) by adsorptive effects and which
may elute near the elution times of perylene and sulfur were examined in this study.
Another difference between traditional and high-efficiency columns is the capacity to
handle matrix loading without excessive loss of chromatographic resolution. Some
column capacity is given up when switching to high efficiency columns in order to
achieve benefits of reduced solvent usage and faster sample processing. Compared
to using traditional columns, high-efficiency columns may use up to 50% less solvent
because samples are processed at up to twice the speed.
Matrix handling capacity was rated by measuring the amount of non-volatile soluble
material removed from a sample at several matrix loading levels. The column and
instrument were calibrated using the recommended EPA procedure, in which 125 mg
of corn oil is loaded onto the column and dump time is adjusted to remove at least
85% of the oil while retaining at least 85% of to-(2-ethylhexyl) phthalate (DEHP).
Then a methylene chloride solution containing corn oil or extract of potting soil was
injected in the system. Each dumped fraction was collected and evaporated to
determine the amount of non-volatile material removed in the dumped fraction.
NOTE: Although high efficiency columns provide greatest resolution when small
injection volumes are used, in practice the 5 mL injection has been retained so that
viscosity effects or high solute loadings of concentrated solutions do not degrade
chromatographic performance or damage columns.
Capacity in this context is defined as ability to process samples while maintaining the
degree of coextractive removal equivalent to the specification (85%) given in the
method calibration instructions. Although results show that high-efficiency columns
have less matrix handling capacity than traditional columns, many environmental
samples can be processed without exceeding capacity.
668
-------
103
DETERMINATION OF POLYCYCLIC AROMATIC
HYDROCARBONS IN SOIL BY ENZYME IMMUNOASSAY
Robert O. Harrison. ImmunoSystems, Inc., Scarborough, Maine 04074
Robert E. Carlson, Ecochem Research Inc., Chaska, Minnesota 55318
ABSTRACT
A competitive inhibition Enzyme ImmunoAssay (EIA) has been developed for the
determination of £olycyclic Aromatic Hydrocarbons (PAH's) in soil. PAH derivatives
were attached to carrier proteins ana these conjugates were used to immunize
rabbits. The resulting anti-PAH antibodies were used to develop a class-specific
field-portable PAH EIA, parallel to the earlier development of PCB and other test kits
for field analysis of soil, water, and other matrices. The PAH EIA has sensitivity
better than 20 ppb for pyrene, which is the most prevalent PAH compound in
environmental samples. The test is formatted as a semiquantitative soil test with 95%
confidence of detection when the total PAH level is 1 ppm or greater. Decisions at
other levels can be made by use of different dilution protocols. The test detects the
key PAH indicator compounds pyrene, fluorantnene, benzo(a)pyrene, and
phenanthrene at less than 1 ppm (each compound alone) in soil using a
semiquantitative protocol. This protocol also detects 10 other PAHs at less than 8
ppm (each compound alone . The test does not recognize PCB's,
pentachlorophenol, BTEX, motor oi, or hydraulic fluid, but some positive interference
was observed with petroleum fuels such as diesel, fuel oils, Bunker C and crude oil,
which have PAH as a significant component. The data described here demonstrate
the usefulness of this EIA for screening PAH in soil in many field and laboratory
situations. The lab and field data developed during the validation of the test have
been submitted to the EPA Office of Solid Waste for evaluation as screening method
4035 of the SW846 compendium of solid waste methods.
INTRODUCTION
Reagent and Method Development
The development of the EIA for PAH's followed these steps: 1) PAH derivatives
were synthesized for conjugation to proteins; 2) one of these PAH derivatives was
conjugated to a carrier protein and the resulting conjugate was used to immunize
animals, which then produced antibodies recognizing both the PAH derivative and
PAH's; 3) a PAH derivative was conjugated to horseradish peroxidase (HRP) to
make a conjugate which can be capturea by anti-PAH antibodies; 4) the PAH-HRP
conjugate was used to screen and select antibodies; 5) the selected system was
optimized for PAH sensitivity and solvent and matrix tolerance, then characterized for
specificity; 6) the sample preparation method used previously for PCB in soil
analysis was applied and validated; 7) the method was validated in the lab by
verifying sensitivity, false positive/false negative rates and spike recovery and
669
-------
testing the effect of pH and water content; 8) the method was validated using field
samples.
Sample Preparation
Soil sample preparation is summarized as follows: Weigh 5 g soil on portable
balance and place in polyethylene extraction bottle; extract soil by adding 5 ml of
methanol ana shaking vigorously for two minutes. Filter extract and collect for
storage or immediate EIA analysis. Analyze extract as described in the EIA
procedure below, diluting the extract in methanol for decisions other than 1 ppm. All
components required for this method are commercially available in kit form, including
materials, protocol, and instructions for analysis of diluted extracts.
PAH EIA Procedure and Interpretation of Results
The procedure for the analysis of samples containing PAH's is as follows: 1) rabbit
antibodies which recognize the PAH structure are immobilized on the walls of plastic
test tubes; 2) PAH's in solvent, in the form of either calibrators or samples prepared
as described above, are mixed with Assay Diluent in the tubes, allowing PAH's to
compete with the PAH-enzyme conjugate for the limited PAH binding sites on the
immobilized antibodies; 3) unbound sample and conjugate are washed away with
tap water. The amount of conjugate retained by trie immobilized antiboaies is
inversely related to the amount or PAH present in the sample; 4) enzyme substrate
is added to the tubes for color development by the bouna enzyme. The intensity of
color is proportional to the amount of bound enzyme and is inversely related to the
amount of PAH present in the sample. Therefore, more color means less PAH.
Total run time is less than 30 minutes per test. The EIA is formatted as a
semiquantitative test only. The optical density (OD) of each sample tube is
compared to the OD of the calibrators. If a sample OD is greater than a calibrator
OD, then the sample has less PAH in it than that calibrator. If a sample OD is less
than a calibrator OD, then the sample may have more PAH in it than that calibrator.
A slight false positive bias is designed into the test to guarantee the false negative
rate is less than 5%.
RESULTS AND DISCUSSION
Test Specificity
The test response to the target PAH compounds of EPA methods 8100 and 8310 is
shown in Table 1. Other compounds and mixtures of compounds expected to be
found in conjunction with PAH contamination are also listed in Table 1. The test has
measurable, but low crossreactivity for several of these, mostly fuel oils. These
have a significant aromatic component, primarily naphthalene, alkylnaphthalenes, and
other low molecular weight PAH's. Other materials such as mineral oil, new motor
oil, and hydraulic fluid ao not contain significant aromatic components and therefore
are not crossreactive in the EIA. The PAH's most strongly recognized by the EIA are
pyrene, fluoranthene, phenanthrene, and benzo(a)pyrene. Eckel et al. (1) have
shown that this group of four compounds together is the most effective inaicator
group for prediction of soil contamination by other PAH's. Thus, the specificity of the
670
-------
EIA for these PAH's allows screening for soil contamination with a high level of
confidence in the results.
Method Sensitivity
Method sensitivity was determined by assaying 8 different soils which did not
contain PAH greater than 1 ppm. Each of these soils was extracted in triplicate and
each extract was assayed in three different assays. The mean and the standard
deviation of the resulting response values (percent of negative control) were
calculated and the sensitivity was estimated at two standard deviations below the
mean [mean - 2SD = 79.5% - 2x6.4% = 67%]. Based on this technique and the
average assay response, the method sensitivity is 0.25 ppm total PAH at a 95%
confidence level.
False Positive and False Negative Rates
False positive and false negative rates were estimated by fortifying eight soil
samples at 20% and 200% of the 1 ppm action level. Each sample was extracted
and each extract was tested in three different assays. Results obtained by
comparison to the 1 ppm PAH calibrator were correct 88% of the time, with 4%
false negatives and 8% false positives.
Effect of Water Content
The effect of water content of the soil samples was determined by assaying three
different untreated soil samples which had subsequently had water added to a final
concentration of 30% (w/w). Aliquots of these samples were then fortified with
PAH. Both the fortified and unfortified samples were extracted and each of these
extracts were assayed three times. It was determined that water in soil up to 30%
had no significant effect on the method.
Effect of pH
The effect of the pH of the soil extract was determined by adjusting the pH of three
soil samples. Soil samples were adjusted to pH 2 to 4 using 6N HCI and pH 10 to
12 using 6N NaOH. These soil samples were then fortified with PAH and the
unfortified and fortified samples were extracted. Each extract was assayed three
times. It was determined that soil samples within the pH range tested had no
detectable effect on the performance of the method.
Spike and Recovery
For the purpose of this experiment, quantitative results were obtained using a pyrene
standard curve. Three different soil samples were fortified at two levels, 0.2 and
0.8 ppm. The spike solution was a mixture of 3 PAH's equivalent in expected assay
response to 86 and 340 ppb pyrene. Three fortified samples of each soil were
extracted and each extract was assayed three times. Recovery values were
calculated based on the expected assay response to the mixture of the three spiked
compounds. Recovery for individual determinations ranged from 48% to 105%.
671
-------
Average recovery by individual extract ranged from 68% to 83%. Overall average
recovery for all samples was 76%.
Field Trial- Correlation with GC-MS Results
A field trial was conducted at a creosote contaminated site using the PAH EIA.
Samples were split for analysis by two EIA operators and also for GC-MS if the EIA
indicated an acceptable concentration for the correlation study. EIA Operator 1 was
experienced, while Operator 2 was trained specifically for this field trial. For this
work, quantitative EIA results were obtained using a negative control and a three
point pyrene standard curve. Duplicate tubes were run for all negative controls, but
all EIA results are based on single tubes for both standards and samples. EIA results
from Operator 2 were scored semiquantitatively and correlated to GC-MS results for
42 samples. There was agreement for 83.3% of the samples tested, with 2 false
negative results (4.8%) ana 5 false positive results (11.9%). The semiquantitative
correlation data shown in Figure 1 demonstrate that the method is comparable in
accuracy to Method 8270.
Field Trial- Precision of GC-MS and EIA
Four sets of field duplicates were run by both enzyme immunoassay (EIA) operators
and the GC-MS lab. Precision results were calculated as coefficients of variation
for total PAH concentration for each pair of field duplicates. The results shown in
Table 2 demonstrate that the method is comparable or superior in precision to Method
8270.
Field Trial- Correlation Between EIA Operators
EIA results from both operators were scored semiquantitatively and correlated for 98
samples. The semiquantitative correlation data shown in Figure 2 demonstrate that
the test performs well even when performed by a newly trained operator in a field
situation.
CONCLUSIONS
1. The test is capable of analyzing for PAH's in soil in the field in less than 30
minutes, using no specialized equipment.
2. Screening of soils containing PAH's can be performed at multiple levels from 1 to
10,000 ppm with 95% confidence of detection of contaminated samples.
3. The use of the same extraction protocol as for other kits such as PCB and BTEX
allows analysis of multiple analytes in a single sample extract.
4. The design of the EIA accommodates significant variations in sample type and
protocol, allowing the analysis of PAH's in other matrices.
5. Method 4035 is acceptable for field or laboratory use. The appropriate level of
quality assurance snould accompany the application of this method for
documentation of data quality.
6. Ongoing work with this kit includes sediment analysis and water analysis. Field
testing and validation are proceeding for these applications.
672
-------
Table 1. Cross Reactivity of Different Compounds in the PAH EIA
Compound
Acenaphthene
Acenaphthylene
Anthracene
Benzo(a)anthracene
Concentration (ppm) Required for
Positive Interpretation at 1 ppm
3.7
2.4
7.6
4.9
Benzo(b)fluoranthene
Benzojkjfluoranthene
Benzo(ghi)perylene
Benzolalpyrene
Chrysene
Dibenz(ah)anthracene
Fluorene
Fluoranthene
lndeno(l 23cd)pyrene
Naphthalene
Phenanthrene
Pvrene
Creosote
Diesel Fuel
Home Heating Oil
#2 fuel oil
#6 fuel oil
Bunker C oil
K-1 kerosene
crude oil
Gasoline
BTEX
pentachlorophenol
new motor oil
hydraulic oil
mineral oil
Biphenyl
Aroclor 1242
Aroclor 1248
Aroclor 1254
Aroclor 1260
2.7
6.2
5.3
0.8
4.3
356
3.4
0.3
6.5
40
0.9
0.2
3.5
75
80
150
150
125
500
800
1000
>300
>1000
>1000
>1000
>1000
250
>200
>200
>200
>200
673
-------
Table 2. Precision Within Laboratory and EIA Operators
Coefficients of variation for total PAH concentration for one 8270 lab and both EIA
operators for four pairs of field duplicates. EIA Operator 1 was experienced, while
Operator 2 had been trained specifically for this field trial.
%CV for total PAH concentration
Sample Pair
1
2
3
4
8270 Lab
(GC-MS)
28
96
77
47
EIA
Operator 1
2
93
49
7
EIA
Operator 2
12
47
0
20
Means 62 20 38
REFERENCE
1. Eckel, W.P., Jacob, T.A., Fisk, J.F.; "Co-occurrence Patterns of Polycyclic
Aromatic Hydrocarbons in Soils at Hazardous Waste Sites"; presentation at Data
Analysis and Interpretation for Environmental Surveillance Conference, Lexington
KY, February 1990.
674
-------
Figure 1. Correlation of Semiquantitative EIA and 8270
total n = 42
* This sample contained several pebble-like lumps which were not included in
the subsamples analyzed by immunoassay. Subsequent analysis indicated
these lumps to be coal or coal-like material which contained more than 10%
total PAH by weight by EIA.
10,000
.QCM
-------
Figure 2. Correlation of Semiquantitative EIA by Two Operators
total n = 98
10,000
>V 1,000
-------
104
IMMUNOASSAY DETECTION OF POLYCYCLIC AROMATIC HYDROCARBONS
SIMPLIFIES FIELD ANALYSIS OF SOIL AND WATER
Michael C. Mullenix, Scientist, Robert T. Hudak, Scientist, and James W. Stave,
Director of Research and Development, Strategic Diagnostics Incorporated, 128 Sandy
Drive, Newark, Delaware 19713
ABSTRACT
The carcinogenic and mutagenic nature of polycyclic aromatic hydrocarbons (PAH) has led
state and federal agencies to regulate their levels in the environment. Until now, site
assessment of PAH levels was limited by the high cost and long turnaround time associated
with the analytical laboratory techniques for measuring PAH. An immunoassay for the
semi-quantitative detection of PAH was developed as a part of the D TECH' environmental
detection systems line of products. This PAH immunoassay is designed to simplify the
process of site assessment by providing a rapid, reliable, and low cost alternative to
analytical laboratory techniques. In the immunoassay, PAH in a water sample or extract
from a soil sample inhibits the binding of a latex-immobilized anti-PAH antibody to an
alkaline phosphatase labeled PAH analog. This test detects the majority of the compounds
on the EPA's list of 16 PAH priority pollutants and its specificity is directed toward the 3,4
and 5 ring compounds. The assay detects the carcinogen benzo(a)pyrene and does not
detect BTEX, PCB, or other priority pollutants. Contaminated field samples of PAH were
collected and the PAH concentrations were determined by both the immunoassay and a
commercial laboratory using the EPA SW-846 method 8270. The immunoassay results
correlated wth method 8270 and displayed a sensitivity range of 0.3 to 10 ppm PAH in soil
and 8 to 500 ppb PAH in water. The low number of false positive and false negative
results confirmed the specificity of the immunoassay for PAHs. The use of immunoassay
simplifies field analysis by providing the advantages of a rapid onsite analysis, minimum
sample preparation, and lower cost per result. An additional advantage is that only minimal
training is required to run the D TECH immunoassay. Inclusion of this easy to use test in
PAH field screening protocols is a cost effective way to simplify site assessment.
INTRODUCTION
Polycyclic aromatic hydrocarbons (PAH) are aromatic compounds consisting of from two
to six fused carbon rings. Many of the four, five, and six ring PAHs are highly
carcinogenic causing them to be routinely measured environmental pollutants (1). PAH
pollution enters the environment from both petrogenic and pyrolytic sources. Petrogenic
PAH waste sources such as diesel fuel and crude oil are formed during low temperature
processes. Pyrolytic PAH waste sources such as coal tar and creosote are the result of high
temperature fossil fuel combustion. The abundance of carcinogenic and mutagenic PAHs
entering the environment through various industrial processes, has led state and federal
agencies to regulate their levels in both soil and water. The EPA lists 16 PAHs as priority
pollutants and regulates their levels in the environment..
In general, site assessment of PAH contamination is carried out by collecting soil and water
samples and sending them to an analytical laboratory for chemical analysis by high
performance liquid chromatography (HPLC) or gas chromatography-mass
spectraphotometry (GC-MS). This costly and time consuming approach prompted the
development of field analytical techniques that can be carried out on site thus reducing the
677
-------
need for laboratory analysis (2). Using field analysis techniques provides the advantages
of faster turnaround time, more accurate detection the semivolatile PAHs that may be lost
due to sample handling, and a significant cost savings.
A recent advance in field analysis occurred with the introduction of immunoassays for
detecting and and measuring PAH. Immunoassays offer several advantages to commercial
laboratory analytical techniques: portability, cost, turnaround time and ease of use. PAH
immunoassays are designed for field screening applications allowing multiple samples to be
run simultaneously. When used as a screening procedure only positive results must be
verified by an analytical method. This limits the number of samples sent off for lab
analysis by instrumental methods. This approach saves both analysis time and the overall
analysis cost of the traditional analytical techniques.
In this study an immunoassay developed as a part of the D TECH environmental detection
systems line of products was used to determine PAH concentrations in both soil and water.
The assay is shown to crossreact with the majority of compounds on the 16 PAH priority
pollutants. Spike and recovery studies in both soil and water demonstrate the assay's
reproducibility and accuracy. The results of immunoassay analysis of real world samples
and standard reference materials are compared to the SW-846 method 8270 (GC/MS). We
conclude that immunoassay directed field screening simplifies and allows a more complete
site assessment for PAH contamination.
MATERIALS AND METHODS
Immunoassay format: PAH in soil samples is extracted with isopropanol by shaking a
small soil sample for three minutes (figure 1). After extraction the sample is diluted twice
in a pH 8.5 aqueous buffer. The diluted sample is filtered directly into the immunoreagent
test reaction vial containing lyophilized latex-antibody and a PAH derivative-alkaline
phosphatase conjugate. An identical reference vial is prepared by adding buffer to a similar
reaction vial that contains the same lyophilized components with the addition of a small
amount of analyte. The reference vial sets the maximum color level and incubation time of
the test under field conditions. In addition, it serves as a procedural control since color
development occurs only if the appropriate reagents are used in the correct order. The
mixtures are resuspended and incubated for five minutes. After incubation, the reaction
mixture is poured onto the appropriate test or reference well on a membrane filter device
and allowed to drain. The filter will retain the antibody-latex and any alkaline phosphatase
conjugate bound to the latex. Increasing concentrations of analyte will inhibit antibody-
latex binding to the alkaline phosphatase conjugate. The test or reference wells are then
washed with a buffered detergent solution. Next the alkaline phosphatase substrate BCIP
is added to each of the wells and incubated until color development in the reference well
reaches the endpoint. As the concentration of free PAH increases in the sample the amount
of blue color in the test well decreases. The percent of reference reflectance of each test
well is determined using a hand held reflectometer or by color comprison to color card
reference card. The results are reported in percent relative reflectance. The PAH
immunoassay for water follows a similar assay format using only a single dilution step into
the aqueous buffer.
Cross reactivity determinations: The ability of the immunoassay to detect individual
PAH compounds was examined by determining the percent cross reactivity of the
678
-------
individual PAHs at their minimum detection limit. All PAH solutions were prepared in
aqueous buffer from individual commercially available PAH standards.
Spike and recovery studies: Negative soil and water samples were spiked at various
concentrations with a mixture of commercially obtained PAH standards. Immunoassay
results were interpreted from a previously generated standard curve.
Immunoassay correlation with EPA SW-846 method 8270 (GC/MS): Real
world soil samples containing PAH and PAH standard reference materials were collected or
ordered from commercial sources. The real world samples were analyzed by immunoassay
and by method 8270 (GC/MS) at a commercial laboratory.
679
-------
RESULTS
Immunoassay cross reactivity: To be used as an effective field screening assay, the
immunoassay must cross react with as many of the 16 PAH priority pollutants as possible.
The percent cross reactivity of the individual PAHs is shown in figure 2. The results
indicate the assay detects the 3,4, and 5 carbon ring PAHs best. Ten of the PAHs
demonstrate cross reactivity above 10 percent and include the most carcinogenic PAHs
such as benzo(a)pyrene and benzo(b)fluoranthene. Three PAH not detected above 10
percent cross reactivity are naphthalene, acenaphthylene and acenaphthene. These three
PAHs are less carcinogenic and vary widely in concentration in various sources of PAH
waste. In addition, the immunoassay was shown not to crossreact with BTEX, PCBs or
other priority pollutants (data not shown).
Soil spike and recovery: The D TECH soil immunoassay has a sensitivity range of
0.3 to 10 ppm of PAH. To demonstrate the linearity of the dose response, negative soil
samples were spiked with various levels of PAH and analyzed using the immunoassay.
The results were plotted verses their corresponding spike concentrations and fitted to a
linear curve (figure 3). The dose response remains linear throughout the assay sensitivity
range, the R^ = 0.987 indicates the high degree of sample correlation and reproducibility
in the assay.
Water spike and recovery: The D TECH water immunoassay has a sensitivity of 8.0
to 500 ppb of PAH in water. The higher sensitivity can be attributed to a smaller dilution
factor in the assay. The water immunoassay, unlike the soil assay, does not require a large
dilution to compensate for the effects of an extraction solvent on assay performance.
Immunoassay results were plotted verses the their corresponding spike concentrations and
fitted to a linear curve (figure 4). The dose response remains linear over the assay
sensitivity range . the R^ = 0.97 indicates a high degree of sample correlation and
reproducibility in the assay.
Immunoassay correlation with EPA SW-846 method 8270: Real world samples
collected at sites with both petrogenic and and pyrolytic sources of PAH contamination
were analyzed by immunoassay. PAH concentrations in the real world samples were
determined by a commercial analytical laboratory using method 8270 (GC/MS).
Immunoassay results were compared to method 8270 results (tables 1, and 2). The
immunoassay results correlated with the method 8270 results in all cases.
680
-------
SUMMARY
The advantages immunoassays offer simplify site assessment in the following ways: they
are portable and allow onsite testing, they are fast and allow results to be obtained in nearly
real time, multiple samples can be analyzed simultaneously, they are less expensive than
analytical analysis, and their format is often simple enough to be used by personnel with
only minimal technical training (3). The development of an immunoassay capable of
detecting PAH has added an important new weapon the effort of regulating levels of these
common and highly carcinogenic compounds in our environment
This study demonstrates the ability of a PAH immunoassay to detect and crossreact with a
wide range of PAH priority pollutants. Spike and recovery experiments with soil and
water samples served to underline the accuracy and reproducibility of the PAH
immunoassay. PAH immunoassay results were also shown to correlate well with both real
world samples and standard reference materials analyzed by EPA SW 846 method 8270
(GC/MS).
Immunoassays are gaining widespread acceptance and applications in the field of pollution
monitoring. The low cost, speed, specificity, and accuracy inherent to immunoassay
technology makes them especially suited to field screening (3). Immunoassay screening
methods for the detection of various priority pollutants have reached both "proposed" and
"draft" status for inclusion into EPA SW 846 method.
LITERATURE CITED
1. World Health Organization and International Agency for Research on Cancer, IARC
Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 32,
Polynuclear Aromatic Compounds, Part I: Chemical, Environmental and Experimental
Data, World Health Organization and International Agency for Research on Cancer, Lyon
1983.
2. Abraham, B. M.; Liu, T-Y.; Robbat, A."Data Comparison Study Between Field and
Laboratory Detection of Polychlorinated Biphenyls and Polycyclic Aromatic Hydrocarbons
at Superfund Sites." 1993. Hazardous Waste and Hazardous Materials 10(4):461-473.
3. Stevenson, R. "Immunoassays Offer Good Specificity and Accuracy in Real Time."
March 1994 Amer. Env. Lab. 26-27.
681
-------
Figure 1. PAH Immunoassay
Soil Sampler
Extraction Jar
PAH in Son (ppm)
>io 0.1-1
-------
Figure 2. PAH Assay Cross Reactivity
£»
•i—i
>
CO
CO
O
Vi
u
co
co
CD
-------
Figure 3
Soil Spike and Recovery Sample Correlation
10
O
a
3
S
ffi
O
S
o.i
R2 = 0.987
0.1
D
co
CO
10
ppm of PAH Spiked
-------
Figure 4
Water Spike and Recovery Sample Correlation
lOOOn
C/3
CO
cd
O
6
M
100-
10-
R2 = 0.97
n >^n
in
CO
CD
10
100
ppb of PAH Spiked
1000
-------
Table 1
Immunoassay Correlation with GC/MS Method 8270
Results are in ppm of PAH
Certified Reference Immunoassay Results
Material*
CRM 103-100
Priority
PollutnT™/CLP
Semivolatiles in Soil
SRM 1941a
0.90
3.00
0.37
Method 8270
1.10
1.90
0.55
* Certified reference materials are diluted to within assay sensitivity range
by homogenization in PAH negative soil.
686
-------
105
EXTRACTION CONDITIONS IN SUPERCRITICAL CARBON DIOXIDE RESULTING IN
PARTIAL BREAKDOWN OF DDT FROM CONTAMINATED SOILS
ALICIA L. WILSON. Environmental Protection Agency/Oak Ridge Institute for
Science and Eduction, 26 W. Martin Luther King Drive, Cincinnati, OH
45268; and Barry M. Auatern, Environmental Protection Agency, RREL,
26 West Martin Luther King Drive, Cincinnati, OH 45268.
ABSTRACT
Under some extraction conditions in supercritical carbon dioxide DDT has
been observed to break down to DDE and other similar compounds. When DDT
was extracted from soil by supercritical fluid extraction (SFE) using
carbon dioxide with methanol cosolvent, recoveries could be comparable to
those obtained with Soxhlet extraction by EPA Method 3540B. Eleven
samples were extracted by SFE and the results were compared to four
samples of the same soil extracted by Soxhlet. 96% of the identified
compounds recovered by Soxhlet were recovered by SFE. Only 93% of the 4,4
DDT extracted by Soxhlet was extracted by SFE. Some of this may have been
due to analytical error and differences in the soil samples, but
decomposition of DDT to DDE and other compounds were also observed in most
of the soil samples extracted by SFE. A significant increase in the
amount of DDE recovered was observed when supercritical carbon dioxide was
used without methanol for the initial extraction steps. Less
decomposition of DDT to DDE was observed when methanol cosolvent was used
to give a more rapid extraction than when carbon dioxide was used alone.
Carbon-13 labeled DDT and DDE were spiked onto soil that has been
contaminated with DDT and related compounds for more than 20 years to
determine whether spiked material was less likely to break down during
extraction. No breakdown of the spiked material was observed. A choice
of rapid extraction conditions appears to be desirable, not only to
increase the throughput of samples, but also to improve the quality of the
analytical results.
INTRODUCTION
DDT has been used extensively throughout the world and is still in use in
some countries, primarily for vector control. However, there have been
extensive reports of adverse effects on wildlife (1-3) as well as an
increasing concern about risks of cancers and reproductive problems in
humans (4,5). Even in countries where DDT is now banned, significant
levels of contamination remain in temperate climates due to the relatively
slow rate of degradation.
There is a need for a rapid and efficient means of analyzing soil in areas
where high levels of DDT still persist. Traditional extraction methods
such as Soxhlet and sonication are labor-intensive and require the use of
relatively large amounts of solvents with the consequent safety and
disposal costs. Supercritical fluid extraction using supercritical carbon
dioxide is relatively safe and should be relatively easy to automate as
the technology advances.
687
-------
Method development for SFE of soils is complicated by matrix effects and
the significant differences in adsorption of analytes to the wide range of
binding sites available on the soil. Aging of a contaminant such as DDT
in a non-homogeneous matrix such as soil results in more of the
contaminant being bound in relatively inaccessible pores on strongly
adsorbing portions of the soil. Since supercritical carbon dioxide
generally becomes a better solvent at increasing densities, extracting a
soil sample initially with carbon dioxide at a low density should only
remove the least tightly bound molecules of DDT. Repeatedly extracting
the same soil at successively greater densities of carbon dioxide should
give information about the proportion of DDT molecules which are most and
least easily extracted. In an attempt to develop a method for an aging
study, it was observed that the total amounts of DDT and DDE recovered
varied considerably depending on the conditions during the first few
extraction steps.
The degradation of DDT to DDE is commonly observed under a variety of
conditions, including those found at hazardous waste sites, so the
presence of DDE in DDT contaminated soils is unsurprising. Therefore,
conditions during SFE need to be chosen to avoid poor apparent recoveries
of DDT due to the breakdown of DDT during extraction of analytical
samples.
EXPERIMENTAL
The hexachlorobenzene was 98% technical grade from Aldrich. The 2,4 DDT
and 2,4 DDE were analytical grade from Supelco. Optima grade hexane and
HPLC grade methanol were from Fisher Scientific. SFC/SFE Grade carbon
dioxide from Air Products was used for extractions. Wright Brothers, Inc.
supplied the carbon dioxide was used for cooling in the extractor and the
prepurified nitrogen and high purity helium for the gas chromatograph.A
Hewlett-Packard Model 7680A SFE Module Supercritical Fluid Extractor was
used. A Hewlett-Packard Model 5890, Series II Gas Chromatograph with a
Supelco SPB-1 30 meter x 0.32 millimeter capillary column and a HP7673
GC/SFC injection tower. A Heat Systems, Inc. sonication bath was used for
cleaning extraction thimbles and syringes.
Soil which had been contaminated with DDT for more than 20 years was
obtained from a Superfund site. The soil refrigerated until it was air
dried in a hood at least a week prior to use. All soil was ground in a
mortar and pestle, and rocks larger than 2 mm were removed. After
grinding, the soil was stored at room temperature in a stoppered glass
bottle.
The Soxhlet extractions were done according to EPA Method 3540B using 1:1
(v/v) acetone/hexane as the solvent. Four grams of soil were extracted in
each Soxhlet extractor with about the same weight of sodium sulfate. The
extractions were run for 24 hours.
For the standard SFE extraction procedure two to four grams of soil were
weighed into a SFE extraction thimble and covered with a wad of glass
wool. One milliliter of methanol was then squirted onto the glass wool
from a syringe just before the sample was placed into the extractor.
688
-------
Supercritical carbon dioxide at a density of 0.75 g/ml was held in the
thimble for 15 minutes at a flow rate of 2 ml/minute. The thimble
temperature was 95°C and the extracted material was trapped at 65°C on
stainless steel beads. After the SFE extraction was completed, the trap
temperature was changed to 45°C and the trap was rinsed with hexane which
was collected in vials containing hexachlorobenzene as an internal
standard.
For the sequential extraction procedure one to four grams of soil were
weighed into the thimble. For the BCD data in Tables I through IV one
milliliter of methanol was added with a syringe. For the work done on the
GC/MS, 10% methanol was added with a modifier pump. The soil sample was
extracted for fifteen minutes at 95°C with carbon dioxide at a density of
0.25 g/ml. This was repeated with the same soil for step 2. For steps 3
and 4 the density was increased to 0.50 g/ml, and for steps 5 and 6 a
carbon dioxide density of 0.75 g/ml was used on the same soil sample.
Then the soil sample was extracted four times for steps 7 through 10 using
the standard extraction conditions described above.
The extracts for Tables I through IV were analyzed by GC using an ECD
detector. The initial temperature of 170°C was held for one minute then
ramped to 220°C at a rate of 10°C/min.; the temperature was then held at
220°C for 18 minutes. Then temperature was ramped to 230°C at 2°C/min.
and followed by a ramp of 10°C/min. to 300°C where it was held for 10
minutes.
The 13C-spiked samples were analyzed on a Finnigan Incos BOB GC/MS.
SUMMARY
DDT concentrations determined after sequential SFE are much lower than
values obtained from the same soil after Soxhlet extraction and the
standard SFE method (see Table I). However, the DDE values are much
higher after the sequential extraction suggesting that the DDT is
undergoing a dechlorination to form DDE. The total recovery of DDT and
DDE together after sequential extraction is less than would be expected
from the Soxhlet and standard SFE values. This may be due to degradation
of DDT to form other compounds for which standards were not available.
During the sequential extraction most of the DDT extracted is removed from
the soil during the first few extraction steps, while greater amounts of
DDE are recovered during later extraction steps (see Tables II, III and
IV). Unpublished data (6) indicate that less decompostion is observed
when methanol cosolvent is used than when supercritical carbon dioxide
alone is used, especially at lower densities of carbon dioxide. Methanol
cosolvent or higher densities of carbon dioxide which result in more rapid
extraction give better recoveries of DDT and less formation of DDE and
other related compounds. Spiked ISC-labeled DDT was extracted in the
first extraction step and no DDE formation was observed. Interactions of
the more tightly bound DDT with the soil may be involved in the formation
of DDE and other compounds during the extraction of aged soil.
689
-------
Since DDE is also hazardous, destruction of DDT in soil with supercritical
carbon dioxide would not provide adequate remediation under the conditions
used for this work, but supercritical carbon dioxide does appear to be
effective for the removal of DDT and related compounds from soil when
suitable extraction conditions are chosen. For analytical purposes, SFE
can give recoveries of DDT and related compounds that are comparable with
Soxhlet results without the use of large amounts of organic solvents.
ACKNOWLEDGEMENTS
We gratefully acknowledge the assistance of Radha Krishnan who provided
the DDT contaminated soil. We appreciate the funding from the United
States Environmental Protection Agency. This research was supported in
part by an appointment to the Research Participation Program at the Risk
Reduction Engineering Laboratory/U.S. Environmental Protection Agency
administered by the Oak Ridge Institute for Science and Education through
an interagency agreement between the U.S. Department of Energy and
RREL/EPA.
REFERENCES
1.) Ramesh, A., et.al. Characteristic Trend of Persistent Orgnochlorine
Contamination In Wildlife from a Tropical Agricultural Watershed,
South India; Arch. Environ. Contam. Toxicol., 23, 26-36 (1992).
2.) Hargrave, B.T., et.al. Organochlorine Pesticides and Polychlorinated
Biphenyls in the Arctic Ocean Food Web; Arch. Environ. Contam.
Toxocol., 22, 41-54 (1992).
3.) Rowan, David J. and Joseph B. Rasmussen. Why Don't Great Lakes Fish
Reflect Environmental Concentrations of Organic Contaminants?—An
Analysis of Between-Lake Variability in the Ecological Partitioning
of PCBs and DDT; J. Great Lakes Res., 18(4), 724-741 (1992).
4.) Wolff, Mary S., et.al. Blood Levels of Organochlorine Residues and
Risk of Breast Cancer; J. Natl. Cancer Inst., 85(8), 648-652 (1993).
5.) Hileman, Bette. Environmental Estrogens Linked to Reproductive
Abnormalities, Cancer; C&E News, 31 January 1994, 19-23.
6.) Wilson, Alicia L. Unpublished data.
690
-------
TABLE I COMPARISON OF SOXHLET EXTRACTION RESULTS WITH SFE RESULTS
BY STANDARD AND SEQUENTIAL PROCEDURES (in ug/g soil)
SFE
Compound
4,4 DDT
2,4 DDT
4,4 DDD
2,4 DDD
4,4 DDE
2,4 DDE
Soxhlet
1200
250
78
30
150
_28
Standard
1200
250
76
34
160
27
Sequential
250
84
37
27
650
200
TOTAL 1800 1700 1300
TABLE II DDT and DDE EXTRACTED USING SEQUENTIAL EXTRACTION STEPS
AT INCREASING DENSITIES FOR EACH SOIL SAMPLE (in ug/g Soil)
Step
1
2
3
4
5
6
7
8
9
10
C02
Density*
.25
.25
.50
.50
.75
.75
.75
.75
.75
.75
ml
MeOH
0
0
0
0
0
0
1.00
1.00
1.00
1.00
4.4 DDT
120
11
52
15
15
4
28
5
1
0.3
4.4 DDE
59
11
200
140
140
21
72
2
0.3
0.2
2.4 DDT
46
8
23
3
0
0
3
0
0
0
.4
.2
.2
.1
.1
2.4 DDE
9
3
120
37
13
3
24
2
0.5
0.2
TOTAL 250 650 84 200
Density in g/tnl
691
-------
TABLE III AVERAGE RATIO OF DDE TO DDT EXTRACTED DURING EACH OF FOUR
REPLICATE STEPS DONE TO EACH OF SIX SOIL SAMPLES
Step
1
2
3
4
C02
Density*
.75
.75
.75
.75
ml
MeOH
1.00
1.00
1.00
1.00
4,4 DDE
4,4 DDT
.14
.12
.16
.28
2,4 DDE
2,4 DDT
.11
.13
.15
.17
TABLE IV AVERAGE RATIO OF DDE TO DDT EXTRACTED AT EACH STEP OF A
SEQUENTIAL EXTRACTION OF EACH OF THREE SOIL SAMPLES.
Step
1
2
3
4
5
6
7
8
9
10
CO2
Density*
.25
.25
.50
.50
.75
.75
.75
.75
.75
.75
ml
MeOH
0
0
0
0
0
0
1.00
1.00
1.00
1.00
4,4 DDE
4,4 DDT
0.48
1.0
3.9
9.1
10
5.9
2.6
0.45
0.29
0.77
2,4 DDE
2,4 DDT
0.19
0.38
4.8
13
33
17
8.3
7.1
4.6
1.8
Density in g/ml
692
-------
W H Q « c« H
Q Q Q 9. Q Q
Q 0 Q g g Q
t t t » * *
a) N * N •» N *
100.0 -
188 -
«.9 -
258 -
RIC -
Tiie 16;
j\
•1 , /I -
I
i 1
48 28:88 23:28 26:48 38:88
b)
53.6 i
188 -(
100.0 ;
258 '
1
RIC -.
1
(
)]
!
_ \ -
13
C Labeled
Ions
Total Ions
13
'C Labeled
Ions
Total Ions
Tiie i6;40
20:00
23:28
26:48
38:88
Figure I a) GC/MS of Soil Spiked with Carbon-13 Labeled DDT
b) GC/MS of Soil Spiked with Carbon-13 Labeled DDT
693
-------
AIR AND GROUNDWATER
-------
106
ABS, FEP, FRE AND FRP MATERIALS: ABILITY TO WITHSTAND ATTACK
BY ORGANIC SOLVENTS AND SORPTION OF TRACE-LEVEL ORGANICS
L.V. Parker. U.S. Army Cold Regions Research and Engineering Laboratory,
Hanover, N.H. 03755; and T.A. Ranney, Science and Technology Corporation,
Hanover, N.H. 03755.
ABSTRACT
This paper examines the suitability of four polymeric materials [acrylonitrile
butadiene styrene (ABS), fluorinated ethylene propylene (FEP), fiberglass-rein-
forced epoxy (FRE) and fiberglass-reinforced plastic (FRP)] for potential use as well
casings in ground water monitoring wells. Specifically, two of the factors that
determine suitability were examined: the ability to withstand attack by organic
solvents and sorption of dissolved organic solutes by the well casings. These
materials are compared with two commonly used polymeric well casing materials:
polyvinyl chloride (PVC) and polytetrafluoroethylene (PTFE).
These six materials were exposed to 27 neat organic solvents, one neat organic acid,
and 25% solutions of hydrochloric acid and sodium hydroxide for up to 16 weeks.
PTFE and FEP were not degraded by any of the organic solvents, while ABS was
either dissolved or softened by almost all of the organic solvents and the neat organic
acid. FRE was attacked by two organic solvents and the neat organic acid, and FRP
was delaminated by eight organic solvents. PVC was dissolved by ten organic
solvents and softened/swollen by a number of others.
A six-week laboratory study compared sorption of low mg/L levels of a suite of
dissolved organics by these six materials. During this study ABS sorbed analytes
much more rapidly and to a greater extent than the other materials, and PVC and
FRE sorbed analytes the most slowly and to the least extent of the materials tested.
During this study we found an increasing number of unidentified peaks in the HPLC
chromatograms of some of our samples, indicating that organic contaminants were
leaching from these materials. By the end of the study (1000 hours), we had 11
additional peaks in the chromatograms of solutions exposed to ABS, 5 in those
exposed to FRP, and 1 in those exposed to FRE. There were no spurious peaks in any
of the chromatograms of solutions exposed to PVC, FEP and PTFE. Several of the
more volatile organic contaminants that leached were identified.
ABS does not appear to be a good material for well casings used to monitor organic
contaminants, while FRE looks quite promising.
694
-------
INTRODUCTION
Ideally any material used as either a well casing or screen in a ground water
monitoring well should retain sufficient strength once installed in the well, should
resist degradation by the environment, and should not affect analyte concentrations
in samples by leaching or sorbing organics or metals. Recent guidance by the U.S.
Environmental Protection Agency [1] acknowledged that none of the most com-
monly used well casing materials in ground water monitoring
[polytetrafluoroethylene (PTFE), polyvinyl chloride (PVC) or stainless steel] can be
used for all monitoring applications. PVC and especially PTFE are not strong
enough to be used in the deepest of wells, and PVC and stainless steel can both be
degraded by certain environments. PVC can be degraded by several neat organic
solvents or by high concentrations (near solubility) of aqueous solutions of these
solvents. Stainless steel will be corroded if any of the following conditions exist: low
pH, high dissolved oxygen and carbon dioxide levels, or the presence of high levels
of hydrogen sulfide, dissolved solids or chlorides [2, 3]. Previous studies by our
laboratory [4-6] and others [7-9] have also shown that PVC, PTFE and stainless steel
are not always inert with respect to sorption and leaching of analytes of interest.
Specifically PVC and PTFE sorb organics, and PVC and stainless steel sorb and leach
metals.
There are other materials that are being used or have been used for well casings or
sampling pipe and that perhaps could be used in situations where the previous
materials have proven unsatisfactory. Four such materials are acrylonitrile butadi-
ene styrene (ABS), fluorinated ethylene propylene (FEP), fiberglass-reinforced
epoxy (FRE) and fiberglass-reinforced plastic (FRP). ABS is a thermoplastic material
like PVC and is a terpolymer of acrylonitrile, butadiene and styrene. FEP is a
copolymer of tetrafluoroethylene and hexafluoropropylene, and because it is a
fluoropolymer it is similar to PTFE in its chemical and physical properties [10]. FRE
is constructed of 75% high-silica glass and 25% high-purity, closed-molecular
epoxy. It is manufactured from bisphenol-A-type epoxy resins cured with methyl
tetrahydrophthalic anhydride [11]. The FRP used in this study consisted of 70%
fiberglass and 30% polyester resin (by weight). This study evaluated two parameters
used to determine the suitability of materials for ground water monitoring: resis-
tance to chemical attack and sorption of organic solutes from aqueous solutions.
Information on the ability of these materials to withstand chemical attack is sketchy.
Most of the information we found was either provided by the manufacturer or taken
from the Cole-Parmer catalog [12]. FEP, like all fluoropolymers, is reported to have
excellent resistance to attack by corrosive reagents and dissolution by solvents. FRE
is reported by its manufacturer to be impervious to gasoline, hydrocarbon products
and most solvents and additives. While the Cole-Parmer catalog reports that
"epoxy" has good resistance to fuel oils, gasoline, jet fuel and kerosene, it also
reports that it is moderately affected by several ketones and is severely degraded by
695
-------
dichloroethane, dimethylf ormamide, benzaldehyde and others. Although the Cole-
Farmer catalog does not give any details on the type of "epoxy" tested, we suspect
that FRE casings will be attacked by the same organic solvents that are reported to
attack "epoxy." According to the same source [12], ABS is severely degraded by a
number of organic chemicals, including several ketones, chlorinated alkanes and
alkenes, and several hydrocarbons such as fuel oils, gasoline and kerosene. Again,
there is no detail on the type of ABS material tested. The manufacturer of the FRF
casings claims that their product is resistant to corrosion but makes no claims about
its resistance to solvents. Thus, it appears that between ABS, FEP and FRE, FEP will
be the most resistant polymer to degradation and ABS will be the least resistant.
We found only two studies on the sorption of organic solutes from aqueous
solutions by these materials. Gillham and O'Hannesin [8] conducted a study that
compared sorption of ppb-levels of six (mono) aromatic hydrocarbons by FRE, SS,
PTFE, polyethylene (PE) and rigid and flexible PVC (RPVC and FPVC, respectively).
They ranked the sorptiveness of the materials as (going from most sorptive to least):
FPVC > PE > PTFE > FRE > RPVC > SS. Jones and Miller [13] also compared sorption
of a number of organics from aqueous solution by SS, (rigid) PVC, ABS, FEP, PTFE
and polyvinylidene fluoride (PVDF). However, it is not clear what caused the losses
they observed since no biocide was added to prevent biological loss and there did
not appear to be any controls that losses could be compared with.
MATERIALS AND METHODS
Six materials were selected for these studies: ABS, FEP, FRE, FRP, PVC and PTFE.
PVC and PTFE were included so that comparisons could be made with these
commonly used materials. Five-cm- (2-in.-) diameter well casing or pipe were used
in these studies. For PVC, PTFE, FEP and FRE, we used well casings manufactured
specifically for ground water monitoring. We were not able to find a manufacturer
that made FEP well casings but did find one that made "pipe for sampling ground
water." Since the manufacturers of ABS well casings have gone out of business, we
purchased waste and vent pipe. Special care was taken to eliminate contamination
from grease or oil during the cutting process. All the cut pieces were washed with
detergent and deionized water as described by Ranney and arker [14,15]. All the
studies were conducted at room temperature.
Chemical Attack Study: Test coupons measuring approximately 1-cm2 were cut
from each type of material. The cutting process fractured some of the edges of the
two fiberglass materials, and therefore an effort was made to select only those
coupons showing little or no fracturing along the edges.
Each coupon was weighed and placed in a 22-mL borosilicate glass vial. Twenty-
eight neat organic compounds (including one organic acid), a 25% hydrochloric acid
solution, and a 25% sodium hydroxide solution were used in this study (Table 1).
696
-------
Table 1. Percent weight gain (or loss) of materials after 112 days of chemical exposure.*
Chemical PTFE FEP FRE FRP PVC ABS
Acetic acid (glacial)
Acetone
Benzaldehyde
Benzene
Benzyl alcohol
Bromochloromethane
N-butylamine
Carbon tetrachloride
Chlorobenzene
Chloroform
Cyclohexanone
1,2-dichlorobenzene
1,2-dichloroethane
trans-l,2-dichloroethylene
Diethylamine
Dimethylformamide
Gasoline (93 octane, unleaded)
Hexane
Hydrochloric acid (25% w/v)
Kerosene (K-l)
Methyl alcohol
Methyl ethyl ketone
Methylene chloride
Nitrobenzene
Sodium hydroxide (25% w/v)
Tetrachloroethylene
Tetrahydrofuran
Toluene
Trichloroethylene
o-xylene
0.4
0.3
0.0
0.4
0.0
0.7
0.2
0.6
0.3
1.0
0.0
0.2
0.4
1.4
0.5
0.0
0.3
0.4
0.0
0.0
0.0
0.3
0.9
0.1
0.0
0.9
0.3
0.2
1.3
0.1
0.3
0.2
0.0
0.3
0.0
0.6
0.1
0.4
0.3
0.8
0.0
0.1
0.3
1.2
0.3
0.1
0.2
0.2
0.0
0.0
0.0
0.2
0.8
0.0
0.1
0.6
0.3
0.2
1.1
0.1
Rt
2.7
0.3
0.0
0.1
26.2
R
0.0
0.2
7.3
-0.1
0.1
3.1
8.1
2.0
R
-0.1
-0.1
-4.7
0.0
7.7
3.0
15.6
0.4
0.2
0.0
3.3
0.0
0.3
-0.1
1.5
5.6
1.3
0.8
0.5
L
L
0.2
7.8
L
0.1
1.1
L
L
3.5
8.3
0.1
0.0
-5.0
0.2
1.9
4.8
L
1.0
1.5
0.5
L
0.9
L
0.2
0.4
157.8s
D
48.7s
0.1
D
D
0.1
159.8*
223.9s
D
217.7s
D
56.3"
31.8s
D
0.1
-0.1
0.3
0.0
0.4
D
D
D
0.1
1.7
D
51.4s
70.9s
65.7s
76.8s
D
D
D
D
D
D
317.2s
D
D
D
D
D
D
112.8s
D
61.9s
15.1
1.2
8.9
27.8
D
D
D
0.9
251.2s
D
D
D
D
* For most materials, degradation (R,L,D,s) occurred before the 112-day sampling time.
tR: resin came off coupon
L: fiberglass sheets delaminated
D: dissolved or so soft material broke up on handling
s: material swollen and/or softened
Most but not all of the test compounds were EPA priority pollutants. Five mL of the
test chemical was added to each vial, and the vial was closed with a Teflon-lined
screw cap. There were no replicates in this study. There were seven sampling times:
1,7,14,21,28,56 and 112 days. On the day of sampling, each coupon was removed
from the vial, blotted dry on a paper towel, and air dried for one minute prior to
being weighed, as described by Ranney and Parker [15]. Softening was determined
697
-------
by trying to indent the coupon using forceps. After the weights were taken and all
other observations were made, the coupon was returned to its vial and recapped.
Sorption of Organics Study: This experiment investigated sorption of 11 organic
solutes: cis-l,2-dichloroethylene (CDCE), trans-l,2-dichloroethylene (TDCE), ben-
zene (BENZ), m-nitrotoluene (MNT), trichloroethylene (TCE), chlorobenzene (CLB),
o-dichlorobenzene (ODCB), o-xylene (OXYL), p-dichlorobenzene (PDCB), m-xylene
(MXYL) and tetrachloroethylene (perchloroethylene or PCE). The test solutions
were prepared by adding each of the neat organics directly to well water in glass
volumetric flasks as described by Ranney and Parker [14]. Forty mg/L of HgCl2 was
added to the test solutions to prevent biological losses of the organics. Initial
concentrations of analytes varied from 1 to 2 mg/L except for BENZ, which had a
concentration of approximately 0.5 mg/L.
Two pieces of one type of casing material were placed in individual 40-mL borosili-
cate glass vials. The vials were filled with aqueous test solution so that there was no
headspace and capped with Teflon-lined plastic caps. Vials with test solution but no
casing material served as controls. These controls allowed us to correct for any
effects that might be due to the vials or caps. The ratio of material surface area to
solution volume was 0.79 cm2/mL, which is typical for a 2-in.-diameter well.
Separate vials were used for each sampling event so that the test solution could be
discarded after sampling. For each material and time, there were three replicates.
The vials were filled randomly in sets of seven, with each vial containing one of the
six materials or empty (the controls). The samples were kept in the dark for: 1 hr, 8
hr, 24 hr (1 day), 72 hr (3 days), 168 hr (1 week), 500 hr (3 weeks) and 1000 hr (6 weeks).
For analysis of each sample, a small aliquot of solution was transferred (using a glass
Pasteur pipet) to an autosampler vial (1.8 mL), which was (gently) filled so there was
no headspace and then capped. Teflon-backed silicone septa were used in the
autosampler vial caps. Analytical determinations of the organic solute concentra-
tions were by reversed-phase high-performance liquid chromatography (RP-HPLC).
A modular system was employed that consisted of a Spectra Physics SP 8810
isocratic pump, a Spectra Physics SP 8490 variable-wavelength UV detector set at
210 nm, a Spectra Physics SP 8875 autosampler with a 100-|oL injection loop, and a
Hewlett-Packard 3396 series II digital integrator. Separations were obtained on a 25-
cm x 4.6-mm (5 um) LC-18 column (Supelco) eluted with 62/38 (v/v) methanol/
water at 1.5 mL/min. The detector response was obtained from the digital integrator
operating in the peak height mode. Retention times of the analytes ranged from 4.0
to 16.3 min. Analytical details can be found in Ranney and Parker [14].
Leaching of Contaminants Studies: When we compared the chromatograms of
samples exposed to the casings with those for the control samples, we saw addi-
tional peaks in some of the samples. Thus, we decided to analyze the 1000-hour
samples using purge-and-trap GC/MS to determine the identity of at least some of
698
-------
these peaks. For these analyses, EPA method 8240 for volatile organics by GC/MS
[16] was used. The PT/GC/MS system consisted of a Tekmar LSC-2 liquid sample
concentrator, a Hewlett Packard 5890 series II gas chromatograph, and a Hewlett
Packard 5970 series mass selective detector. One sample for each type of material
plus a control sample were analyzed.
To confirm that the organics we had found in the test solutions resulted from
leaching from the casing materials, we placed two pieces of the cleaned casing
material (the same size as used previously) in 40-mL glass vials. These vials were
then filled with the well water that contained 40 mg/L of HgCl2 to prevent any
biological activity. These samples were analyzed after approximately 500 hours of
contact time, using the same method as described previously. We tested only those
materials that had leached contaminants in the previous study (ABS, FRE, FRP) and
a blank (water only); there were no replicates in this study.
RESULTS AND DISCUSSION
Chemical Attack Study: Table 1 shows which samples were degraded by chemical
exposure and the final percent weight gain for those samples that were not
destroyed. Although both fluoropolymers, PTFE and FEP, are generally recognized
as being inert to chemical attack, there were slight (1%) weight gains in samples
exposed to five chlorinated organic solvents by the end of the study. The weight
gains were slightly lower in the FEP samples than in the PTFE. There were no
apparent signs of softening, swelling or decrease in strength in any of these the FEP
or PTFE samples.
The FRE well casing material had a glossy external surface and a dull internal
surface. The external surface of these samples flaked when exposed to three
chemicals (N-butylamine, acetic acid and dimethylformamide), although these
particles did not subsequently dissolve. These samples appeared to remain strong,
and no further observations were made on them. Thirteen samples had weight gains
exceeding 1%; the samples exposed to bromochloromethane (26.2%) and methylene
chloride (15.6%) had the largest weight gains. The sample exposed to the
hy drocholoric acid solution had a slight weight loss (~5%). None of the FRE coupons
appeared to swell or soften, even the sample with the 26% weight gain. Some fraying
of the edges was observed on some coupons, but it is not clear whether this was due
to exposure to the chemicals or due to handling.
FRP was more severely degraded than the previous materials. Eight chemicals
delaminated it, i.e. the fiberglass sheets separated. This occurred within the first 1-
4 weeks and for one chemical within the first 24 hours. The samples that were
delaminated more slowly had weight gains of ~ 1-16% and showed signs of swelling
(i.e. liquid could be squeezed out of the material) prior to the sheets separating.
Eleven other chemicals caused weight gains of 1-10%, although there were no signs
699
-------
of swelling or softening. Again, some of the coupons showed frayed edges, although
this may have resulted from handling and not chemical exposure. As with FRE the
hydrochloric acid solution caused a slight loss in weight (5%).
PVC appears to be much more readily degraded than the previous materials. By the
end of the study, ten chemicals dissolved or so softened PVC that the test piece could
not be weighed because it fell apart. Four of the chemicals had this effect within the
first day. Twelve other chemicals appeared to soften PVC, and for seven of those
chemicals, weight gains exceeded 100%. Squeezing the swollen coupons forced out
some of the liquid. Only 9 of the 30 chemicals used in this study had little or no effect
on PVC. These chemicals were glacial acetic acid, the acid and hydroxide solutions,
alcohols, hydrocarbons (gasoline, hexane and kerosene) and carbon tetrachloride.
ABS was by far the most readily degraded polymer tested. After only one day of
exposure, 19 of the 30 chemicals evaluated either dissolved ABS or softened it to the
point where it fell apart. Four other chemicals caused degradation (i.e. swelling
and/or softening) of the coupon on the first day. By the end of the study, only the
acid and hydroxide solutions had little effect (-1% weight gain). Clearly ABS is a
poor choice where exposure to neat organic solvents may be involved.
Generally where comparisons could be made, we had good agreement between our
results and those given by Cole-Parmer. This was especially true for the PTFE and
the ABS. (There were no listings for FEP or FRP.) The largest disparity is between
their ratings for "epoxy" and our findings for FRE. For FRE we would change the
ratings for 11 of the 30 chemicals tested. The differences between "epoxy" and FRE
most likely accounts for these differences.
We would rank the materials used in this study, from greatest to least resistance, as:
FEP = PTFE > FRE > FRP > PVC > ABS. This ranking should be used only as a general
guide, not as a rule. For chemicals that haven't been tested, we suggest testing them
with the casing material, especially ABS, FRE, FRP and PVC.
Sorption of Organics: Figures 1-4 show the losses of CDCE, TDCE, TCE and PCE
and are fairly typical for the losses we observed. The complete results are presented
elsewhere [14]. Generally we found that 1) ABS always sorbed analytes the most
rapidly and to the greatest extent of all the materials tested; 2) PVC and FRE sorbed
analytes the most slowly and to the least extent; and 3) neither PTFE, FEP nor FRP
performed consistently better than the others. The data are summarized in Table 2,
which shows the first sample time where a 10% loss in analyte concentrations were
observed. For several organics, 10% losses were observed in 8-24 hours for PTFE,
FEP and FRP, and in 1-8 hours for ABS. This was not the case for PVC and FRE. For
PVC the earliest a 10% loss was first observed was 500 hours, and for FRE the earliest
a 10% loss was observed was 72 hours. Losses of some compounds never reached
10%; this is especially true for FRE and PVC.
700
-------
200
400 600
Time (hrs)
800
Figure 1. Sorption of CDCE.
1000
Figure 2. Sorption of TDCE.
200 400 600 800 1000
Time (hrs)
1 1 1 1 I '
200
400 600
Time (hrs)
800
-| Figure 3. Sorption of TCE.
1000
701
-------
II I 1—I
Figure 4. Sorption of PCE.
200
400 600
Time (hrs)
800
1000
Our results generally agree well with those of Gillham and O'Hannesin [8] except
that they found that the rate and extent of sorption of the compounds they tested
[(mono)aromatic hydrocarbons] were always greater for FRE than for PVC. Gener-
ally we did not find this to be the case in our study. By the end of the study, we found
no significant difference between PVC and FRE samples exposed to two of the same
three compounds tested by Gillham and O'Hannesin [8]. Since both studies used a
constant surface-area/solution-volume ratio (which differed between the two
studies), we suspect that the reason their results differed some from ours is because
they tested materials other than well casings; they tested FRE tubing and PVC pipe.
The composition and densities of these materials may be quite different.
Leaching of Contaminants: When we examined the HPLC chromatograms, we saw
additional peaks in some of the samples when compared with the control samples.
By the end of the study, there were additional peaks in the chromatograms for the
ABS, FRE and FRP samples but not for the FEP, FIFE and PVC samples (Fig. 5).
These results agree reasonably well with what we found when we reviewed the
Table 2. Sampling time (hours) when material sorbed 10% or greater analyte.
Material CDCE TDCE TCE PCE BENZ CLB ODCB PDCB OXYL MXYL MNT
PVC
PTFE
FEP500
ABS
FRE
FRP
1000
168
24
8
NL
8
500
24
8
1
72t
8
1000
8
1
1
1000
8
1000
8
168
1
NL
8
NL*
168
24
8
NL
24
1000
24
8
1
1000
8
1000
24
8
1
1000
8
500
8
24
1
72
8
NL
24
8
1
NL
24
1000
8
NL
1
NL
8
NL
1000
8
NL
72
* NL: Never lost 10% by the end of the study.
t Subsequently, losses were only 7 and 4% at 168 and 500 hr, respectively.
702
-------
3
o
IWVJvV
CONTROL
I
PTFE
I
Tim6 (min)
12
16
Time (min)
12
16
FEP
PVC
8 12
Time (min)
16 0
8 12
Time (min)
16
Figure 5. HPLC Chromatograms for 1000-hour samples: a. Control, b. PTFE, c. FEP,
d. PVC.
literature on leaching of organic contaminants from these materials. Several studies
[17,18] have shown that PTFE leaches relatively few organic impurities. Presumably
FEP would behave similarly to PTFE. Leaching of organics from PVC has been found
to be considerably less problematic for rigid PVC, such as pipes and casings, than for
flexible tubing [17]. This is mainly because rigid PVC products contain almost no
plasticizers (<0.01%) [19].
703
-------
FRP
8 12
Time (min)
I i I
ABS
8 12
Time (min)
16
Figure 5. HPLC Chromatograms for 1000-hour samples: e. FRE, f. FRP, g. ABS.
The ABS samples appeared to leach the most contaminants since these samples had
the most additional peaks in their RP-HPLC chromatograms. The RP-HPLC chro-
matogram for the last (1000-hour) samples had 11 additional peaks (Fig. 5g).
However, even the one-hour samples had one extra peak, and the size of these peaks
increased as time continued [14]. The chromatograms for the FRP solutions had one
additional peak after 72 hours [14] and five additional peaks by the end of the study
704
-------
(Fig. 5f). There was only one additional peak in the chromatograms for the FRE
samples (Fig. 5e); this peak first appeared in the 72-hour samples [14]. With both of
the FRE and FRP samples, the size of the peaks increased as time continued.
'To determine the identity of at least some of these peaks, we analyzed the 1000-hour
samples by purge-and-trap GC/MS. When we ran the ABS sample, we observed six
peaks and were able to identify four of them. This sample contained acrylonitrile
and styrene (two of the three components of ABS), chloroform and ethylbenzene
(which is an intermediate in the production of styrene). The concentrations of these
compounds in this sample were quite low (<10 mg/L). The other peaks that we had
observed previously in the HPLC chromatograms apparently were due to the
presence of either nonvolatile or semivolatile organics or inorganic compounds (e.g.
metal salts). We only found one peak when we ran the FRP sample, and this was
determined to be toluene (which may be used as a solvent or degreaser in the
production of FRP). The concentration of the toluene was approximately 100 mg/
L. Again, the other four peaks that we observed previously in the HPLC chromato-
grams may be due to the presence of either nonvolatile or semivolatile organics or
inorganic compounds. The one peak we observed in the HPLC chromatograms for
the FRE samples is apparently not a volatile organic, since we did not observe any
peaks in the chromatograms for the purge-and-trap GC/MS analyses of these
samples. These results agree well with those of Cowgill [11], who tested intact
FRE well casings and a ground powder of these casings for leaching of any sub-
stance involved in its manufacture and EPA priority pollutants. Low levels of di-
ethylphthalate and bisphenol A leached from the powdered well casings
after three weeks but not from intact well casing. Cowgill noted that bisphenol
A is a component of manufacture, and diethylphthalate is a commonly used
plasticizer.
When we conducted a leaching study to confirm that the substances we found in the
previous samples had leached from the casing material, we found essentially the
same analytes in these samples as we did previously. We were able to identify two
of five peaks we found in the chromatograms of ABS leachate: ethylbenzene and
styrene. We did not find any spurious peaks in the FRE leachate sample. For the FRP
leachate samples, we found five peaks and were able to identify three of them. In
addition to finding toluene again, we also found 1,1,1-trichloroethane and
ethylbenzene. (This particular sample was run twice with similar results.) These
solvents may be used in either the manufacture of this product or cleaning some of
the equipment used in its manufacture.
With respect to leaching of contaminants, our results agree well with what is found
in the literature for FRE, FEP, PTFE and PVC. As expected, PEP and PTFE performed
similarly and did not appear to leach any contaminants. FRE appeared to leach only
one compound; most likely this is the same component (bisphenol A) Cowgill [11]
705
-------
observed leaching from ground FRE casing. Given that we used a waste and vent
ABS pipe rather than well casing manufactured for monitoring ground water, it is
not surprising that we found a number of contaminants leaching from this product.
CONCLUSIONS
Based on the results from these studies and others, it appears that FRE would be an
excellent material to be used for monitoring organics. It is relatively nonsorptive of
organic solutes from aqueous solutions, appears to leach few organic contaminants,
and is relatively inert to chemical attack (more so than PVC). FEP is inert to chemical
attack and does not appear to leach any organic contaminants. However, it is
relatively more sorptive of organic solutes than PVC and FRE. It does not appear to
offer any clear advantage or disadvantage over PTFE at this time. FRP leaches
several contaminants and is relatively sorptive of organics, but it is relatively
resistant to degradation by organic solvents. By far, ABS appears to be the poorest
choice of a material for monitoring organics. It is degraded by a large number of
organic solvents, sorbs organics very rapidly, and leaches a number of contami-
nants. However, while we feel this material would not be a good casing material, we
realize we tested waste and vent pipe rather than well casings and that quality well
casings may provide better performance, especially with respect to leaching. Since
ABS well casings are not available at this time, we do not see any point testing these
materials further. We also realize that those materials that initially sorb organics
rapidly (FEP, FRP and ABS) would eventually reach equilibrium and sorption
would then be less of a problem, depending upon how the sample was taken.
Although, desorption of sorbed analytes could also possibly be a problem for these
materials if ground water quality were to improve.
Our lab is currently evaluating FRE, FRP and FEP casings with respect to sorption
and leaching of metals. This information will allow us to better determine the over-
all suitability of these materials.
ACKNOWLEDGMENTS
We thank Martin H. Stutz, Project Monitor, and the U.S. Army Environmental
Center (USAEC, formerly USATHAMA), Aberdeen Proving Ground, Maryland, for
their support of this work. We also thank Dr. Thomas Jenkins and James Cragin,
CRREL, for their technical reviews of this manuscript.
This publication reflects the personal views of the authors and does not suggest or
reflect the policy, practices, programs or doctrine of the U.S. Army or the Govern-
ment of the United States. The contents of this report are not to be used for
advertising or promotion purposes. Citation of brand names does not constitute an
official endorsement or approval of the use of such commercial products.
706
-------
LITERATURE CITED
1. U.S. EPA, 1992. EPA RCRA Ground-Water Monitoring: Draft Technical Guid-
ance. EPA/530-R-93-001, Office of Solid Waste, U.S. Environmental Protection
Agency, Washington, D.C. 20460. NTIS #PB 93-139-350, National Technical
Information Service, Springfield, Virginia 22161 (703-487-4650).
2. Driscoll, F.G., 1986. Groundwater and Wells, Second Edition. Johnson Filtra-
tions Systems, Inc., St. Paul, Minnesota.
3. Aller, L., T.W. Bennett, G. Hackett, R.J. Petty, J.H. Lehr, H. Sedoris, D.M. Nielsen
and J.E. Denne, 1989. Handbook of Suggested Practices for the Design and
Installation of Ground-Water Monitoring Wells. National Water Well Associa-
tion, Dublin, Ohio.
4. Hewitt, A.D., 1989. Leaching of metal pollutants from four well casings used for
ground-water monitoring. Special Report 89-32, U.S. Army Cold Regions
Research and Engineering Laboratory, Hanover, N.H.
5. Parker, L.V., A.D. Hewitt and T.F. Jenkins, 1990. Influence of casing materials
on trace-level chemicals in well water. Ground Water Monitoring Review,
10(2): 146-156.
6. Hewitt, A.D., 1992. Potential of common well casing materials to influence
aqueous metal concentrations. Ground Water Monitoring Review, 12(2): 131-
135.
7. Reynolds, G.W. and R.W. Gillham, 1986. Adsorption of halogenated organic
compounds by polymer materials commonly used in groundwater monitors.
In Proceedings of Second Canadian/American Conference on Hydrogeology,
Hazardous Wastes in Ground Water: A Soluble Dilemma. National Water Well
Association, Dublin, Ohio, p. 125-132.
8. Gillham, R.W and S.F. O'Hannesin, 1990. Sorption of aromatic hydrocarbons by
materials used in construction of ground-water monitoring sampling wells. In
Ground Water and Vadose Zone Monitoring, ASTM STP 1053, D.M. Nielsen
and A.I. Johnson, Eds. American Society for Testing and Materials, Philadel-
phia, Pa., p. 108-122.
9. Reynolds, G.W., J.T. Hoff and R.W. Gillham, 1990. Sampling bias caused by
materials used to monitor halocarbons in groundwater. Environmental Science
and Technology, 24(1): 135-142.
707
-------
10. Sax, N.I. and R.J. Lewis Sr., 1987. Hawley's Condensed Chemical Dictionary,
Eleventh Edition. Van Nostrand Reinhold Company, New York, N.Y.
11. Cowgill, U.M., 1988. The chemical composition of leachate from a two-week
dwell-time study of PVC casing and a three-week dwell time study of fiberglass
reinforced epoxy well casing.In Ground-Water Contamination: Field Methods,
ASTM STP 963, A.G. Collins and A.I. Johnson, Eds. American Society for
Testing and Materials, Philadelphia, p. 172-184.
12. Cole-Farmer Instrument Company, 1992. Cole-Parmer Instrument Company
1993-1994 Catalogue. Cole-Parmer Instrument Company, Niles, HI., p. 1463-1471.
13. Jones, J.N. and G.D. Miller, 1988. Adsorption of selected organic contaminants
onto possible well casing materials. In Ground Water Contamination: Field
Methods, ASTM STP 963, A.G. Collins and A.I. Johnson, Eds. American Society
for Testing and Materials, Philadelphia, Pa., p. 185-198.
14. Ranney, T.A. and L.V. Parker, in press. Sorption of trace-level organics by ABS,
FEP, FRE, and FRP well casings. Special Report, U.S. Army Cold Regions
Research and Engineering Laboratory, Hanover, N.H.
15. Ranney, T.A. and L.V. Parker, in preparation. Susceptibility of ABS, FEP, FRE
and FRP to Chemical attack. Special Report, U.S. Army Cold Regions Research
and Engineering Laboratory, Hanover, N.H.
16. U.S. EPA, 1986. Test Methods for Evaluating Solid Waste. Volume IB: Labora-
tory Manual Physical/Chemical Methods. Third Edition. EPA number EPA/
SW 846, Office of Solid Waste and Emergency Response, U.S. Environmental
Protection Agency, Washington, D.C. 20460. NTIS# PB88-239223, part 2 of 4,
National Technical Information Service, Springfield, Virginia 22161 (703-487-
4650).
17. Curran, C.M and M.B. Tomson, 1983. Leaching of trace organics into water from
five common plastics. Ground water Monitoring Review, 3: 68-71.
18. Barcelona, M.J., J.A. Helfrich and E.E. Garske, 1985. Sampling tubing effects on
ground water samples. Analytical Chemistry, 57: 460-464.
19. Barcelona, M.J., J.P. Gibb and R.A. Miller, 1984. A Guide to the Selection of
Materials for Monitoring Well Construction and Ground Water Sampling. U.S.
Environmental Protection Agency Report No. EPA-600/2-84-024. U.S. Govern-
ment Printing Office, Washington, D.C.
708
-------
107
SAMPLING AND ANALYTICAL METHODS FOR HOUSE DUST AND DERMAL
EXPOSURE
J. P. Hsu, Paul Geno, Tapan Majumdar and David Camann, Southwest Research Institute, 6220
Culebra Road, San Antonio, Texas 78228
ABSTRACT
The indoor exposure to pesticides and polynuclear aromatics is becoming important issue. Pesticides
and polynuclear aromatic hydrocarbons are usually present in both house dust and air. There is a
strong correlation between the pesticide levels in house dust and air. Due to complex matrix of house
dust, it present a challenging problems for chemical analysis. The house dust is sampled from carpet
using a validated HVS3 (high Volume, Small, Surface Sampler). However, dermal sampling is also
quite important since infants and toddlers have more frequent floor contact since crawling around
floor or carpet and hand-to-mouth behavior. The dermal sampling is performed by using a PUF
roller, which is developed by Southwest Research Institute. The dust samples will be extracted,
clean-up and analyzed by GC/MS in single ion monitoring mode for the target compounds listed in
Table 1.
Table 1 Target Compound List
Alachlor Aldrin Atrazine Bendiocarb
Captan Carbaryl oc-Chlordane y-Chlordane
Chlorothalonil Chlorpyrifos Dacthal p,p'-DDE
p,p'-DDT Diazinon Dichlorvos Dicofol
Dieldrin Folpet Heptachlor Hexachlorobenzene
Lindane Malathion Methoxychlor cis/trans-PeTmethrin
Propoxur o-Phenylphenol Resmethrin
Benz(a)anthracene Benzo(g,h,i)perylene
Benzo(a)pyrene Benzo(b)fluoranthene
Benzo(k)fluoranthene Coronene
(-)Cotinine
709
-------
108
CO2 MANAGEMENT FOR TO-14 ANALYSES USING A CONTROLLED
DESORPTION TRAP (CDT)
R. Jesserand S. Reiss Graseby Nutech, RTF, NC 27703-9000
Large quantities of C02 injected onto capillary columns can complicate
the quantitation and identification of closely eluting components by GC or
GC/MS. Normally, the 330ppm C02 in ambient air presents no significant
problem for analyzing the trace Volatile Organic Compounds (VOCs) in
EPA Method TO-14. However, bioremediation approaches to
environmental cleanup can increase C02 levels by more than 100 times.
In addition, source sites and sealed systems can also contain extreme
levels of CO2- Consequently, the need has arisen for C02 management
in whole air methodology.
TO-14 suggests a Nafion® dryer for water management. Alternate
approaches must be employed when quantitating water soluble
compounds under TO-14 protocol. This paper will describe a modification
of a water management technique employed with TO-14 protocol for
water soluble compounds. Data will be presented showing that sample
recoveries and Relative Standard Deviations (RSDs) are good. This CDT
approach can lend itself to a wide variety of sample matrices.
710
-------
109
STATUS AND NEED FOR FUEL TOXICITY CHARACTERISTICS LEACHING
PROCEDURES (TCLP)
Ajmal M- Hias. Chemist, and George J. Medina. Chemist, U.S. Army Corps of Engineers,
North Pacific Division Laboratory, 1491 N.W. Graham Ave., Troutdale, Oregon, 97060.
ABSTRACT
Approximately fifty targeted analytes of volatiles, semi-volatiles and metals have been
included under the Toxicity Characteristics Leaching Procedure (TCLP) extraction
procedure. Studies on the extraction and leaching potential for petroleum hydrocarbon fuels
have not been formally investigated and published. It has been noted that an acidified
(HCL) TCLP extraction using semi-volatile organic procedures, followed by methylene
chloride extraction, has yielded substantial levels of hydrocarbon fuel but significantly less
than the conventional total soil extraction procedure using sonication or soxhlet techniques
with freon or methylene chloride solvents. Higher levels of fuels, such as kerosene or
diesel fuel No. 6 were found when water samples were subjected to TCLP condition
compared to the conventional EPA 3510 extraction without acidification. Advantages to
fuel leaching potential in soil (in particular landfill disposal and underground storage tank
sites) using this proposed TCLP procedure, coupled with the Corps of Engineers Fuel
Identification and Quantitation Method (FIQ) will be discussed.
INTRODUCTION
Currently, the U.S. EPA has relinquished the underground storage tank (UST)
authority and responsibility to the states for issuing permits, inspection and determining
mandatory cleanup of fuel/oil contamination in the environment. It is estimated that over a
million leaking USTs have been excavated. Evidence of leakage in the form of
contaminated soil have been found with up to 100 ppm of gasoline, and 200 ppm of diesel,
including lubricating oil. In certain states, this may trigger clean-up action procedures.
Because of the loosely regulated guidelines, the cost impact on small businesses and gas
stations (many of which are independently owned) has been so dramatic, many have been
forced to close in the State of Oregon alone. It is suggested that if the EPA, state or both
agencies would jointly regulate fuel contamination and remediation criteria on the basis of
TCLP, many of the storage sites may not need remediation or clean up. Small businesses,
which are going out of business due to non-compliance may not need to close, if regulated
under the proposed TCLP (1) guidelines.
Presently, no guideline for fuel TCLP is established but about one half of the
targeted analytes are components of fuel. If the sum of critical components of fuel
compounds are added, it is estimated that a maximum level of 10 ppm would not be
exceeded for fuels in the soils. This could serve as liberal but realistic guideline. Utilizing
711
-------
this "cut-off standard may save billions of dollars and an untold number of jobs from
unnecessary UST Tank removal, servicing or expensive remediation.
EXPERIMENTAL DATA
Fuel contaminated soil samples were extracted using EPA TCLP Method 1311. The TCLP
extracts were re-extracted with methylene chloride and analyzed for diesel range organics
DRO (2) and total recoverable petroleum hydrocarbons as diesel, TPH-D (3), with the use
of a gas chromatograph equiped with a flame ionization detector (FID). These two methods
are modified methods derived from EPA Method 8100 (4). The TCLP extracts were also
re-extracted with freon 13 (trichloro, trifluoro ethane) for TRPH analysis using EPA
Method 418.1 (5). For the purposes of control, a split of the soil sample was subjected to
TCLP and was extracted using methylene chloride. The other split was subjected to
sonication/extraction technique, Method 3550 (6) and subsequently analyzing for TRPH
using EPA Method 418.1.
Water samples were split. One of the splits was non-acidified and extracted, following the
protocols of EPA 3510 (7) and analyzed for various fuels using Army Corps of Engineers
Fuel Identification and Quantitation procedure, FIQ (8). The other split was acidified,
following TCLP protocols. The TCLP extract was re-extracted with methylene chloride
(EPA 3510) and also analyzed using the FIQ procedure. Details are presented in Table 3.
RESULTS AND DISCUSSION
Table 1 details the TPH-D results of fuel contaminated soil samples obtained from an
UST/landfill site using TCLP extraction and total extraction (EPA 3550). No TCLP TPH-
D were detected between the concentrations of 0.05 through 5.0 parts per million (ppm) in
the extract fraction subjected to TCLP and re-extraction with methylene chloride, indicating
no fuel leachability. TPH-D was found in the same samples which were directly extracted
with methylene chloride using EPA method 3550. The concentrations ranged between the
detection limit of 1 ppm through 2300 ppm.
Table 2 presents data comparison of soil samples subjected to TCLP/methylene chloride
re-extraction (EPA 3510) and classically extracted with using EPA Method 3550
(sonication), followed with GC/FID analysis for DRO. Inclusive in Table 2 are data of the
same samples, sonicated extracted and analyzed for TRPH (EPA 418.1), using an infra-red
spectrophotometer (IR). The average recovery of TCLP/methylene chloride extraction and
analyses for DRO is 0.805 ppm. The average recovery of straight soil sonication extraction
and analyses for DRO is 782 ppm. The average recovery of TRPH is 2165. The State of
Alaska's regulatory clean-up level III for DRO and TRPH are 1000 and 2000 ppm,
respectively. Both, the straight sonication/extraction analysis using a GC/FID for DRO and
IR method of analysis for TRPH provide data that are close to the State of Alaska's
712
-------
regulatory clean-up levels. Since no guidelines have been established for fuel TCLP, data
generated by TCLP for DRO may not trigger clean-up action.
Table 3 describes the result of fuels where water samples were subjected to TCLP
conditions. The TCLP extract was re-extracted with methylene chloride using EPA method
3510. Splits of the samples were also extracted without TCLP treatment. Both methylene
chloride extracts were analyzed using Army Corps of Engineers Fuel Identification and
Quantitation procedure (8). Recovery of gasoline was about one third less after TCLP than
the conventional extraction and analysis. However, concentration of kerosene and bunker C
fuel were greater after TCLP extraction, indicating, perhaps, that the acidified water matrix
yields more fuel, contrary to the non-acidified extracted water samples. Low gasoline
recoveries levels found in the TCLP treatment are probably due, in part, to volatilization
during the filtration process. Loss may be minimized if zero head space TCLP extraction is
employed.
CONCLUSION AND RECOMMENDATIONS
If we are to accept the premises that TCLP simulates field leaching trends, coupled with the
data presented, it is permissible to assume that TPH-D does not leach out in any appreciable
amounts. No leachable TPH-D was evident after the leachate was subjected to re-extraction
with methylene chloride (Table 1). Consequently, no remidiation would be warranted.
Following the normally prescribed methodologies, based on the state of Alaska's clean-up
levels, the location in question could conceivably require stringent monitoring and/or
remediation. Less than 1 ppm of TCLP DRO was found in the samples of Table 2,
indicating non-leachable fuel contamination.
With the exception of low boiling fuels, such as gasoline, water samples subjected to TCLP
conditions, meaning filtration and a lower pH, yielded higher recovery concentrations of
kerosene and heavier fuels (Table 3). The results suggest the need to modify the extraction
procedure currently being used (EPA 3510) in order to maximize fuel recoveries; in
particular at low levels.
Soil samples (as well as waste and water samples) subjected to a TCLP (semi volatiles)
extraction procedure will undergo loss of low boiling fuel compounds. To avoid the loss of
these low boiling compounds, the use of zero head space TCLP procedures are
recommended. Following zero head space extraction, the use of a purge and trap/GC-FID
analyses or the FIQ method of analysis with a GC-FID is recommended (see figure 1). The
FIQ method may be developed to classify and quantitate low boiling fuels such as gasoline,
certain jet and kerosene, mineral spirit, naphtha, etc. (Figure 1). The advantage of TCLP
extraction and application of the FIQ method is the potential for analysis of a broad range of
fuels, i.e., from gasoline through heavy lubricant oils.
713
-------
Without the implementation of more realistic criteria to evaluate clean-up levels for fuel
contaminated sites, many small businesses, such as local gasoline stations may have to
suffer from the expensive repair or remediation based on data that may erroneously provide
evidence of leachable fuels. Similarly, costly corrective action to landfills such as the
placement of dikes or liners may not be required. The presented data is far from being
conclusive. More controlled studies are required. However, the data clearly suggests that
there is room for the investigation into fuel migration in soil, the use of TCLP to evaluate
contaminated soil teachability and the implementation of more realistic clean-up level
criteria.
ACKNOWLEDGMENTS
The authors are grateful to Mr. Timothy J. Seeman, Director, U.S. Army Corps of
Engineers, North Pacific Division Laboratory. The authors further acknowledges the effort
and assistance of NET Pacific Inc., Santa Rosa California, for the production and
verification of some of this data for this paper.
REFERENCES
1. EPA Method 1311, SW-846, Volume IB, Final Update 1, 1992 40 CFR Part 261,
Appendix II, 1993
2. AK102, Diesel Range Organics, Alaska Department of Environmental Conservation,
Alaska, 1992.
3. TPH-as-diesel, Department of Ecology, Washington State, 1992.
4. EPA 8100, SW-846, Volume IB, 1986.
5. Total Recoverable Petroleum Hydrocarbons, EPA Method 418.1, EPA 600/4-79-029
Methods for Chemical Analysis of Water and Waste, 1983.
6. EPA Method 3550, Sonication Extraction, SW-846, Volume IB, Final Update 1,1992.
7. EPA Method 3510, Separatory Funnel, Liquid-Liquid Extraction, SW-846, Volume IB,
Final Update 1992.
8. Ilias A.M. Jaeger C. "Evaluation of Sampling Techniques for the Analysis of Volatiles
and Total Recoverable Petroleum Hydrocarbons (TRPH) by IR, GC and GC/MS Methods,"
Chapter 10, Volume III (Hydrocarbon Contaminated Solids), Lewis Publication, 1993.
714
-------
Figure 1
Flow Diagram for Soil
Fuel TCLP
and Analyses
Zero Head Space
Purge and Trap
GC-FID
TPH-G or TPH-D
Soil 100 Grams
EPA 1311 Extraction
Semi-Volatile Extraction
GC-FID
Method
FIQ
Fuels of Interest:
Gasoline, Mineral Spirits, etc.
Fuels of Interest:
Kerosene/Jet Fuels
Diesel Fuel No. 2
Through Lubricant Oil
715
-------
Table 3: Comparison of analysis data of fuel contaminated water using TCLP extraction procedures (EPA
1311) followed by re-extraction of the extract with methylene chloride and water extracted using EPA
Method 3510 utilizing a GC-FID.
CD
Fuel
Detected
Gasoline
Kerosene
Diesel Fuel No. 2
TCLP(EPA1311)
with EPA 35 10
RE-EXTRACTION
GC-FIQ
Result mg/1
440
310
780
WATER
EXTRACTED /
EPA 35 10
GC-FIQ
Result mg/1
660
280
620
-------
Table 1: Comparison of analysis data of fuel contaminated soil using TCLP extraction procedures (EPA
1311) followed by re-extraction of the extract with methylene chloride and soil extracted using EPA
Method 3550 to determine the leachability potential from the soil utilizing a GC-FID.
TCLP(EPA1311)
with EPA 3510
SOIL EXTRACTED
with EPA 3550
Sample
Description
sp-5
5-4
5-1
B-3
1-8
RE-EXTRACTION
Result
ND
ND
ND
ND
ND
Detection
limits
mg/1
0.05
0.05
0.05
0.5
5.0
Result
39.0
ND
44*
100
2300*
Detection
limits
mg/kg
1.0
1.0
1.0
1.0
5.0
* Positive result for petroleum hydrocarbon as diesel appears to be due to the presence of heavier
hydrocarbons rather than diesel.
-------
Table 2: Comparison and assessment of data for DRO/TRPH determination using three different types of
extraction and analyses procedures.
00
TCLP
(EPA 1311)
with EPA 3510
RE-
SOIL
EXTRACTED
with EPA 3550
SOIL
EXTRACTED
with EPA 3550
EXTRACTION
SAMPLE
DESCRIPTION
93UNKL01S
93UNKL02S
DRO
RESULT
0.880
0.730
DETECTION
LIMITS
mg/1
0.100
0.100
GC-FID
DRO
RESULT
675
890
DETECTION
LIMITS
mg/kg
5.0
5.0
418.1/IR
TRPH
RESULT
1320
3010
DETECTION
LIMITS
mg/kg
20
20
AVERAGE
0.805
782
2165
-------
110
ANALYSIS OF AIR CANISTER SAMPLES FOR POLAR AND NON-POLAR
VOLATILE COMPOUNDS USING MODIFIED TO-14
Paul E. Kester. Edmund T. Lewis, Alan T. Madden, Valerie J. Naughton, Tekmar Company, P.O.
Box 429576, Cincinnati, OH 45242-9576
ABSTRACT
i
Recent air studies in urban areas have revealed a suprising number of polar compounds Polar
compounds are a particular challenge because they can be highly unstable. Canisters permit the
effective sampling and analysis of both polar and non-polar VOCs in one analytical run. Canisters
present the additional challenge of analyzing this broad range of analytes in the same sample with
relatively high water concentrations.
A complete method for the analysis of the volatile range is necessary. Water removal is an
important aspect for successfully achieving this goal. A condensation trap is described which
selectively eliminates water without eliminating analytes of interest.
Approximately ninety (90) analytes, including selected CLP method compounds, are investigated.
Complete system performance is evaluated using different levels of relative humidity and sample
concentrations. Data to be presented includes response factors, relative standard deviations, and
calibration curves which exhibit the effectiveness of the moisture control system.
i
Ramamurthi, Mukund; Kelly, Thomas; and Spicer, Chester, "Temporal and Spatial Variability of
VOC Area Sources in Urban Air", Measurement of Toxic and Related Air Pollutants Conference,
Durham, NC, 1993.
719
-------
Ife
y&'j'j.
RADIATION
-------
111
RADIOCHEMICAL METHODS AND DETECTION LIMITS
Edmond J. Baratta. U.S. Food and Drug Administration, Winchester Engineering and
Analytical Center, Winchester, Massachusetts 01890
ABSTRACT
The methods for radiochemical analyses had been originally used to determine the
products of fission from the splitting of uranium and plutonium atoms. They were
also used to determine the properties and isotopes of the artificially produced
actinides. Later, radiochemical methods were used to determine the extent of
contamination from various sources. These included world-wide fallout from above
ground testing, emissions from nuclear power plants and other nuclear facilities.
New emphasis is now on decontamination of decommissioned nuclear power
plants and other facilities and sites. Traditionally, gamma spectrometry has been
used as one means of identifying these radionuclides. However, for the analysis of
"pure" beta-emitters, such as strontium-89 and 90, this is not possible. Also,
alpha-emitters from the actinides and some of their decay products require
radiochemical analyses. The levels of activity at which these radionuclides pose a
hazard are low, so that the use of gamma spectrometry is not feasible. In addition,
identification and quantification of various isotopes of the same element, such as
uranium, is not possible. The matrices that can be analyzed include air, water,
soil, food and other media.
The limits of detection that is required for the radionuclides of interest has
improved. This is due to the more sophisticated instrumentation and improved
methodology. This paper discusses the radiochemical methodology available and
the limits of detection that the U.S. Environmental Protection Agency has set from
its collaborative and inter-comparison studies over the years.
720
-------
INTRODUCTION
The methods for radiochemical analyses had been originally used to determine the
products of fission from the splitting of uranium and plutonium atoms (1234). They
were also used to determine the properties and isotopes of the artificially produced
actinides. Later, radiochemical methods were used to determine the extent of
contamination from various sources. These included world-wide fallout from above
ground testing, emissions from nuclear power plants and other nuclear facilities.
New emphasis is now on decontamination of decommissioned nuclear power
plants and other facilities and sites. Traditionally, gamma spectrometry has been
used as one means of identifying these radionuclides. However, for the analysis of
"pure" beta-emitters, such as strontium-89 and 90, this is not possible. Also,
alpha-emitters from the actinides and some of their decay products require
radiochemical analyses. The levels of activity at which these radionuclides pose a
hazard are low, so that the use of gamma spectrometry is not feasible. In addition,
identification and quantification of the various isotopes of the same element, such
as uranium, is not possible. The matrices that can be analyzed include air, water,
soil, food and other media.
Methods have been published by various societies'5671 in order to "standardize" the
procedures being used. This may be for regulatory purposes or other reasons.
Federal Government and International Agencies have produced Procedures Manuals
for use by these agencies to use which "standardize" the procedure being used
throughout their agencies and contractees18910111213141.
The limits of detection that is required for the radionuclides of interest has
improved over the years. This is in part due to the more sophisticated
instrumentation that is available and the improved methodologies. This paper will
discuss the radiochemical methodology available and the limits of detection and
errors associated with them. The U.S. Environmental Protection Agency has set
these detection limits and errors associated with the analyses, based on its
collaborative and inter-comparison studies over the years.
RADIOANALYTICAL CHEMISTRY
Basically, radioanalytical chemistry can be subdivided into three types of analyses.
Those gamma and x-ray emitting radionuclides that can be measured with little or
no sample preparation. Beta particles and alpha particles must, however, be
separated chemically.
721
-------
Also, the use of neutron activation analysis that has been used in the analysis of
specific radionuclides. The latter will not be discussed, due to the necessity of
having to have a "reactor" available for this type of analysis.
GAMMA EMITTERS:
In the earlier days, radiochemical separate was necessary even for gamma-
emitters. They were normally counted using their beta or alpha particles. They
could be detected using their gamma photons, however, the former methods were
more sensitive. With the advent of the sodium iodide crystals and the single
channel analyzers, the analysis of gamma emitters was made easier. In the very
late fifties and early sixties, the use of the solid state multi-channel analyzer and
the larger sodium iodide crystals, improved tremendously the analyses of gamma
emitters. Some radiochemistry was still
needed for the various radionuclides since multiple isotopes and gamma photons of
energies limited somewhat the use of gamma analyses by this method. Milk and
milk products, however, because of the "discrimination" of the cow made this
analysis fairly easy. The use of simultaneous equations made this analysis routine
for the following isotopes171: iodine-131, barium, lanthanum-140 and cesium-137.
Shorter-lived iodine isotopes did cause some interference, however, for other than
milk collected near test sites in the early hours of a weapons test, the problem was
minimal. The amount of cesium-134 produced didn't interfere as it did later in the
Chernobyl Accident. The first germanium (lithium-drifted) detectors and computer
assisted multi-channel analyzers made the analysis of gamma-emitters, faster and
easier. At the time of the Chernobyl Accident, the intrinsic high-efficiency
germanium detectors were available. Many more samples could be processed
more accurately and faster. The methods for peak analyses varied somewhat with
the systems being used. However, they are all basically similar. The IAEA
published a guidebook in 1989 describing this method'141. It made some specific
recommendations for the effective and accurate use of gamma analyses of
matrices of interest. These are as follows:
-Sample geometries must be selected for the matrices of interest including air
filters, water, vegetation, milk, fresh vegetables and other foods, and fresh water
and marine organisms.
-The geometries must be calibrated for the densities of the sample of interest as a
function of gamma ray energy. This involves the preparation of calibration curves
of gamma ray counting efficiency versus energy.
-In preparing the calibration curves, standard preparation of radionuclides from an
organization such as the National Institute of Standards and Technology (USA) or
other reliable sources must be used.
722
-------
-Calibration curves for unit density materials including water and meat can be made
by using known amounts of radioisotopes in aqueous solution in the sample
containers of interest. Samples of greater or less than unit density should be
radiolabelled with the appropriate radionuclide(s), and calibration curves should be
prepared on the labelled matrix.
-Radionuclide standards such as 137Cs and 60Co should be counted daily to ensure
that the gamma ray spectrometer is operating correctly.
The form in which the sample is presented to the gamma-ray detector depends on
the sample type, the available equipment, the radionuclides present and the levels
of activity. Some sample containers include nylon planchets, aluminum cans,
plastic "cottage cheese" containers and molded Marinelli beakers. Measuring
times vary with activity present, sample type, detection limits required, detector
efficiency and radionuclides of interest.
RADIOCHEMICAL METHODS
a.) Alpha-Emitters
The EPA lists various alpha-emitting radionuclides in its regulations regarding limits
for alpha emitters. They have promulgated new rules, but the present rules for
contaminations are radium-226, radium-228 and gross alpha in water is for
combined radium-226 and 228 - 5 pCi/L (0.185 Bq/L), gross alphas (including
radium-226, but excluding radon and uranium - 15 pCi/L (0.555 Bq/L). The
method recommended is either that appearing in the Standard Methods'5' or the
EPA Method'111. For samples containing solids greater than 500 mg/l the Standard
Method or the EPA Method1121 co-precipitation method is recommended. Should the
gross alpha limit exceed the 5.0 pCi/L limit, then analyses for radium-226 and
radium-228 must be performed. The gross alpha is a guide screening method
requiring a short period of time and man power. There are a variety of methods
available for the determination of the radium isotopes, these have been
summarized in the chapter1221 on "Analytical Methodology for Radium in Food and
Water" and in "Radon, Radium and Uranium in Drinking Water". These methods
are more time consuming. The methods use a barium carrier to separate the
radium isotopes from the solution. The radium may be precipitate and counted or,
in the case of radium-226, the ingrowth product radon-222 may be collected and
determined. A barium-133 tracer may also be used for the latter. Usually if the
radium-226 is determined by alpha counting, the ingrowth should be followed.
Two other alpha-emitters of importance are uranium and plutonium. The Standard
Methods151 includes both a method for total uranium and isotopes uranium. The
uranium proposed maximum contamination limit is 20 mg/L or 30 pCi/L (1.11
Bq/L). This analysis may be required when the gross alpha concentration exceeds
723
-------
15 pCi/L (0.555 Bq/L). The total uranium method requires only the alpha counting
of the separated uranium. The isotopic uranium method requires electroplating and
alpha counting using a solid state detector (normally a silicon detector). Some
laboratories are using a rare-earth fluoride co-precipitation method, which may be
used, if it can provide adequate resolution when using the alpha spectrometry
system.
Currently there are no "standard methods" for plutonium in drinking water or any
standard regulation. The proposed rule for the 40CFR Parts 141 and 142119'
proposes a limit of 6.2 E01 pCi/L (2.294 Bq/L). The current ICRP-30 limits the
ingestion of plutonium-239 to the Rad worker of an All of 5.4 E06 pCi. Per day,
for the general population, (1/100) compared with radium-226, it would be three
times that allowed for this isotope. The newer ICRP-61 (1990) reduces that to
approx 1/2 of the radium-226 allowable concentration. (Table A)
Currently there are several methods for plutonium in water19'101. These usually
require separation and addition of a tracer for the recovery. Electroplating and
detection by alpha spectrometry to determine the isotope content and tracer
recovery.
Radon is another alpha-emitting isotope that certainly requires discussion. The
present liquid scintillation method has been tested twice by the EPA. The first
study was inconclusive, at least at the level that EPA found to be of concern. The
second study has not been reported to date. The proposed regulations sets for
Radon-222 a limit for the Maximum Concentration Level at 300 pCi/L'191. There are
essentially two methods for determining radon-222 in water'20211. These are by
liquid scintillation counting and the Lucas Cell. The radon-222 in drinking water is
found only in groundwater supplies. Milny and Cathern1221 have reported that if the
water concentration of radon is 10,000 pCi/L, an average of 1 pCi/L is contributed
to the air, (a factor of 10,000 less). A radon concentration of 1000 pCi/L in water
would contribute to indoor air containing the amount roughly equal to the average
outdoor concentration which is about 0.1 pCi/L of air.
b.) Beta-Emitters
The EPA proposed drinking water standards for beta's and photon emitters is
limited to exposure of 4 mRem ede/yr (ede = effective dose commitment, for a 50
year period, considering 2 Liters/day intake). The 1976 Interim drinking water
standard for strontium-90 is 8 pCi/L (0.296 Bq/L), when other sources are not
considered. The other, being a long-lived beta emitter, is tritium. Presently, the
contamination level is set at 20,000 pCi/L (740 Bq/L).
724
-------
The strontium-90 and total strontium methods can be found in the official methods
books'5671. The methods essentially separate the radio-strontium using stable
strontium as a carrier by nitric acid separation. The radio-strontium may be
counted to determine the combined strontium-89 and strontium-90 (if the
strontium-89 is suspected to be present). The yttrium-90 is allowed to grow in
and separated by solvent extraction and/or precipitation. The yttrium-90 is then
counted in a low-background, low-level beta counter. The strontium-89 is
determined by calculations. The strontium-90 is calculated from the yttrium-90
content. Other methods'1516) have been used to directly determine the yttrium-90,
knowing that the strontium-89 is not present.
Tritium is determined by liquid scintillation counting. Normally it may be
determined directly, with little preparation. It may be concentrated by electrolysis19'
should the need arise for lower concentrations. The proposed EPA rules does not
require this concentration.
DETECTION LIMITS
Different conventions with differing terminology and mathematics have been used
to estimate the lower limit of detection (LLD) or the minimum detectable activity
(MDA)'91718>. To eliminate confusion and the production of noncomparable data, it
has been proposed that the Environmental Measurements Laboratory procedure19'
be used exclusively. The basis of this procedure is hypothesis testing. LLD is
defined as the smallest quantity of sample radioactivity that will yield a net count
for which there is a predetermined level of confidence that radioactivity is present.
Two errors may occur: Type I, in which a false conclusion is reached that
radioactivity is present, and Type II, with a false conclusion that radioactivity is
absent.
The LLD may be approximated as
LLD * (Ka + K,)S0
where:
Ka_ value for the upper percentile of the standardized normal variate
corresponding to the preselected risk of concluding falsely that activity
is present (a).
Kf = the corresponding value for the predetermined degree of confidence
for detecting presence of activity (1 -ft), and
S0 = estimated standard error of the net sample counting rate.
725
-------
For sample and background counting rates that are similar (as is expected at or
near the LLD) and for a and ft equal to 0.05, the smallest amount of radioactivity
that has a 95% probability of being detected is,
LLD35 = 4.66Sb
where:
S4 = standard deviation of the instrument background counting rate.
The LLD thus calculated is in units of counts per minute; to convert to
concentration use the appropriate factors of sample volume, counting efficiency,
etc. Note that the approximation LLD = 4.66 Sb can be used only for
determinations where Sb is known so that S0 = J2"S4 and there are no counting
interferences. Examples of appropriate determinations are tritium, gross alpha or
beta, or any single nuclide determination.
Where tracers are added to determine yield or more than one radionuclide is
counted in a sample, use the general form of the equation above, for which the
95% confidence level would be LLD = 3.29 S0.
The detection limits for various radionuclides have been determined by the EPA.
Also, the Laboratory Precision of these radionuclides have also been determined,
through various collaborative studies done by the EPA. These studies have been
accepted by the various National Societies in their Standard Methods'567I Books of
the past years. The methods in these Books are a result of these Collaborative
Studies that had been done for these Societies. The precision and accuracy of the
methods appearing in these Books are a result of these Studies. As early as 1966
this was proposed in an article in Health Laboratory Science1231. Table B is the
result of these studies. The "Activity Level" lists the "one standard deviation" for
a single determination. Activities (results) below these numbers are considered
"not detectable" on the LLD.
726
-------
REFERENCES:
1.) THE ACTINIDE ELEMENTS, Edited by Glen T. Seaborg and Joseph J. Katz,
McGraw-Hill Book Co., Inc., Parts 1 and 2 (1954)
2.) THE RADIOCHEMICAL STUDIES: THE FISSION PRODUCTS, Edited by
Charles D. Coryell and Nathan Sugarman, McGraw-Hill, Books 1, 2 and 3
(1951)
3.) BETA AND GAMMA-RAY SPECTROSCOPY, Edited by Kai Siegbahn, North-
Holland Publishing Co. (1955)
4.) Ibid Vol. 1 and 2 (1965)
5.) STANDARD METHODS FOR THE EXAMINATION OF WATER AND
WASTEWATER, 18th Ed. (Washington, D.C. American Public Health
Association, (1992)
6.) AMERICAN SOCIETY FOR TESTING MATERIALS, 1989 ANNUAL BOOK OF
ASTM STANDARDS, WATER AND ATMOSPHERIC ANALYSIS, Vol. 11.02,
Philadelphia, PA (1989) [ASTM, 1989]
7.) OFFICIAL METHODS OF ANALYSIS, 15th Ed., Vol. 1, Association of Official
Analytical Chemists, Washington, D.C. (1990)
8.) RADIOASSAY PROCEDURES FOR ENVIRONMENTAL SAMPLES, U.S. Public
Health Service Publication No. 999-RH-27 (Superintendent of Documents,
Washington, D.C.) (1967)
9.) ENVIRONMENTAL MEASUREMENTS LABORATORY: Procedures Manual
(U.S. Department of Energy - New York, NY) HASL-300 (Revised 1992)
10.) DOE METHODS FOR EVALUATING ENVIRONMENTAL AND WASTE
MANAGEMENT SAMPLES (NTIS, U.S. Department of Commerce,
Springfield, VA) DOE/EM-0089T (1993)
11.) RADIOCHEMICAL ANALYTICAL PROCEDURES FOR ANALYSIS OF
ENVIRONMENTAL SAMPLES, (U.S. Environmental Protection Agency-
Las Vegas, NV) EMSL-LV-0539-17 (1979)
12.) RADIOCHEMICAL PROCEDURES MANUAL, (U.S. Environmental Protection
Agency, Montgomery, AL) EPA 520/5-84-006 (1984)
727
-------
13.) METHODS OF RADIOCHEMICAL ANALYSIS, World Health Organization
Technical Report Series No. 173 (WHO, Geneva) (1959)
14.) MEASUREMENT OF RADIONUCLIDES IN FOOD AND THE ENVIRONMENT -
A GUIDE BOOK, IAEA - Technical Series No. 295 (IAEA-Vienna) (1989)
15.) Baratta, E.J. and T.C. Reavey, "Rapid Determination of Strontium-90 In
Tissue, Food, Biota And Other Environmental Media by Tributyl Phosphate,
J. Agri. Food Chem 17. (1960) pp. 1337-1339
16.) Volchok, H.L., J.E. Gaetjin, J.L. Kulk and W.R. Eckelmann,
"Determination Of 90Su And 140Ba In Bone, Dairy Products And Vegetation
and Soil". Ann NYAcal Sci 71(1957) pp. 293-304
17.) NATIONAL COUNCIL ON RADIATION PROTECTION AND MEASUREMENTS,
HANDBOOK OF RADIOACTIVITY MEASUREMENT PROCEDURES, NCRP
REPORT NO. 58 (Washington, D.C.) (1978)
18.) AMERICAN NATIONAL STANDARDS INSTITUTE, ANSI N13 10-1974
Inst. Electrical Electronics Engineers, Inc., (New York, N.Y.)(1974)
19.) Federal Register Vol. 56, No. 138, Part II National Primary Drinking Water
Regulations; Radionuclides; Proposed Rule (July 18, 1991) pp. 33050-
33127 [56FR 3305]
20.) Lucas, H.F., Jr. "A Fast and Accurate Survey Technique for Both Radon-222
and Radium-226", The Natural Radiation Environment, J.A.S. Adams and
W.M. Lourdes, Eds. (1964) pp. 315-29
21.) Prichard, H.M. and T.F. Gesell, "Rapid Measurements of 222Rn
Concentrations in Water Using a Commercial Liquid Scintillation Counter,
Health Physics Vol. 22, (1977) pp. 577-81
22.) RADON, RADIUM AND URANIUM IN DRINKING WATER, Edited by C.
Richard Cothern and Paul A. Rebens, Lewis Publishers (1990)
23.) Moeller, Dade W. "Standard Assay Technics - A Status Report,
Laboratory Health Science, Vol. 4, No. 3 (1967) pp. 146-152
728
-------
SUMMARY
The instrumentation used in the determination of radionuclides, especially gamma-
emitters has improved over the past decades. The detectors have gone from G-M
Tubes to high efficiency intrinsic germanium detectors. Computers now are used
to perform peak search, nuclear identification, efficiency calculations, alignment,
etc. They are highly efficient and usually user-friendly. Portable systems are now
in routine use, for on the spot determinations.
Many radiochemical procedures have been standardized and the instrumentation is
now highly sophisticated, usually computer assisted with the necessary user-
friendly software.
While the instrumentation has significantly improved, the lower limit of detection
has not been lowered by much. This is due primarily to natural background. The
standardization and improved instrumentation has improved to some extend the
precision and accuracy of methods.
729
-------
TABLE B. LABORATORY PRECISION: ONE STANDARD DEVIATION VALUES AND CONTROL LIMITS FOR VARIOUS ANALYSES
ANALYSIS
ACTIVITY LEVEL
ONE STANDARD DEVIATION
FOR SINGLE DETERMINATION
CONTROL LIMITS
AVG. OF 3 DET
Gamma emitters
Strontium-8 9
Strontium-90
Potassium
Gross Alpha
Gross Beta
Tritium
5 to 100 pCi/liter or kg
>100 pCi/liter or kg
5 to 100 pCi/liter or kg
>100 pCi/liter or kg
2 to 30 pCi/liter or kg
>30 pCi/liter or kg
aO.l g/liter or kg
s20 pCi/liter
>20 pCi/liter
slOO pCi/liter
<4,000 pCi/liter
a4,000 pCi/liter
5 pCi/liter
%5 of known value
5 pCi/liter
5% of known value
1.5 pCi/liter
5% of known value
5% of known value
5 pCi/liter
25% of known value
5 pCi/liter
5% of known value
ls{pCi/liter) = (170) (known)-0933
10% of known value
IJL ± 8.7 pCi/1
[i ± 0.087 \i
IJL ± 8.7 pCi/1
IJL ± 0.087 JJL
IJL ± 2.6 pCi/1
p. ± 0.087 LI
IJL ± 0.087 IJL
H ± 8.7 pCi/1
H ± 0.087 IJL
IJL ± 8.7 pCi/1
H ± 0.087 IJL
p ± 294 (M)'0933
Li ± 0.17 IJL
Radium-226,
Radium-228
Plutonium
Iodine-131
Uranium
aO.l pCi/liter
0.1 pCi/liter
gram or sample
s55 pCi/liter
>55 pCi/liter
s35 pCi/liter
>35 pCi/liter
15% of known value
10% of known value
6 pCi/liter
10% of known value
6 pCi/liter
15% of known value
IJL ± 0.26
IJL ± 0.17 LI
IJL ± 10.4 pCi/1
IJL ± 0.17 ft
IJL ± 10.4 pCi/1
IJL ± 0.26 LI
730
-------
TABLE A
Radiation Protection Guides (RPG) for Transient
Rates of Intake of Radionuclides Recommended for
the Average of Suitable Samples of the Population,
RADIONUCLIDE
Radium-226
RANGE I
0-2
intake/day (pCi/day)(l)
RANGE II RANGE III
2-20
20-200
EPA Drinking Water Standards(40 CFR 141.15) - 15 pCi/L
ICRP No. 2 (Basis for FRC Guidelines)
Frac. of Transfer from G.I. Tract to Blood
Radium-226
Plutonium-239
0.3
0.00003
ICRP No. 30
Radium-226
Plutonium-239
"Limits of Intake
for Rad Workers"
ALI-pCi
1.89 E06
5.4 E06
Per Day
(1/100)
p/Ci
51.8
148
ICRP No. 61 (1990)
Radium-226
Plutonium-239
2.43 E06
1.08 E06
66.5
29.6
(1) Range I requires only periodic surveillance;
Range II, quantative surveillance and routine control;
Range III, evaluation and additional controls.
731
-------
112
PROGRESS TOWARD MATURITY OF "DOE METHODS FOR
EVALUATING ENVIRONMENTAL AND WASTE MANAGEMENT
SAMPLES"
Steven C. Goheen. Senior Research Scientist, Margaret McCulloch, Technical Specialist,
Robert G. Riley, Senior Research Scientist, Sandra K. Fadeff, Research Scientist, Berta L.
Thomas, Project Manager, Gary M. Mong, Research Scientist, Georgia K. Ruebsamen,
Senior Clerk, Wayne C. Cosby, Editor, and Debbie S. Sklarew, Senior Research Scientist,
Pacific Northwest Laboratory, P. O. Box 999, Richland, Washington 99352
ABSTRACT
The document DOE Methods for Evaluating Environmental and Waste Management
Samples (DOE Methods) has been in circulation since October 1992. DOE Methods is a
living document, being updated twice each year. It contains both sampling and analytical
methods in support of U. S. Department of Energy/environmental restoration and waste
management (DOE/EM) activities. Guidance on how to carry out sampling and analysis
activities, focusing on EM needs, is also included in DOE Methods. This guidance applies
to all aspects of sampling and analysis for EM. Methods from traditional standard methods
documents often cannot provide needed characterization data because of radioactivity or
complexity of the matrix. The intent of DOE Methods is to provide an alternative source
of methods to meet this need. Efforts are underway to expand the use of DOE Methods
throughout all DOE/EM programs.
Copies of DOE Methods are available free of charge. The April 1994 update of the
document includes 42 methods, of which 13 are new. In October 1994, Revision 2 of
DOE Methods will be distributed. It will include additional guidance on how to plan
sampling and analysis activities and will also include several new methods.
DOE Methods is supported by the Laboratory Management Division of DOE. It is a
vehicle for transferring new technology for characterization capability within the
environmental restoration (ER) and/or waste management (WM) community. As DOE
Methods continues to evolve, its use and application will continue to grow.
INTRODUCTION
DOE Methods has been available since October 1992 (1). Since then it has grown in
circulation and size to become a standard for DOE and DOE contractor laboratories (history
of development to be presented at Spectrum '94 in a paper entitled "A DOE Manual: DOE
Methods For Evaluating Environmental and Waste Management Samples"). DOE
Methods also provides guidance on planning for sampling and analysis activities, and it
encourages the use of both sampling and analytical methods for ER and WM, selected
either from a variety of standard references (2-5) or from DOE Methods.
Two years ago at this conference, the first issue of DOE Methods was introduced. Last
year the performance approach to method selection and qualification was discussed. This
presentation is focused on the progress of the document toward maturity.
732
-------
Development of DOE Methods. Before 1991, all DOE sites independently developed
standard operating procedures for sampling and analyzing radioactive environmental and
waste matrices. In 1991, DOE decided that it was important to make these sampling and
analysis procedures available to all sites (in July 1991, the Laboratory Management Branch
of DOE's Office of Environmental Restoration and Waste Management issued a draft "five
year plan" for the Analytical Services Program; copies of the final version may be obtained
from Daniel Lillian, Mike Carter, or David Bottrell at the Office of Technology and Waste
Management, Trevion II, 12800 Middlebrook Road, Germantown, Maryland 20874).
This need was to be filled in two ways: 1) a database, containing key information, from
which standard operating procedures from all sites could be accessed, and 2) a document
similar to EPA's SW-846 (2). The DOE document has evolved into DOE Methods. The
DOE Procedures Database is available through Los Alamos National Laboratory (LANL)
staff.
The DOE Procedures Database contains many procedures from across the DOE complex
whose technical content is duplicated several times. For example, more than one hundred
procedures are available from the database for the analysis of uranium. Consolidating the
procedures would allow similar approaches to be represented in a single method.
Consolidated methods would allow the site chemists to indicate that their methods were
shared standard methods used by other laboratories across the DOE complex. This
standardization process could be useful in several ways:
• Methods accepted and applied complex-wide would be more easily accepted by
regulators.
• Uniformity would help keep results comparable and defensible.
• Guidance would be provided to contractors wanting to do DOE work.
Meeting these needs was the justification for publishing DOE Methods.
Needed Technology. DOE Methods addresses technology needs for sampling and
analyzing DOE/EM samples, particularly as these needs relate to the sampling and analysis
of radioactive and mixed waste. Characterization of radioactive and mixed waste is
currently being addressed at several DOE sites to help solve immediate problems. As
methods are developed to meet these needs, editors of DOE Methods are pursuing
documented solutions to the problems.
Distributing Information. DOE Methods is a tool for distributing technical
information to the DOE/EM characterization community. It provides guidance on quality
assurance, quality control, sampling, safety, and analysis. It also provides guidance on
how to modify existing methods or develop new methods. With each issue, DOE
Methods encourages comments from everyone on its distribution list. This provides a
mechanism for DOE Methods to become tailored to the needs of the EM community, thus
minimizing concerns that it may be outside the scope of any EM program. DOE Methods
primarily contains methods that fit the needs of DOE/EM programs.
DOE Methods is part of the DOE Methods Compendium Program, which includes
participants from the U.S. Environmental Protection Agency (EPA) and from DOE sites
733
-------
across the country. These participants work together to create a holistic approach to
solving the DOE/EM sampling and analysis problems. DOE Methods is one component of
this network. Staff at PNL and LANL work closely to consolidate methods. The EPA at
the National Air and Radiation Laboratory (NARL) and DOE-HQ are also involved in the
consolidation process. Oak Ridge National Laboratory, LANL, Argonne, INEL, and other
DOE labs provide methods generated specifically for inclusion in DOE Methods (see
earlier note on five year plan). Together, components of the DOE Methods Compendium
Program aim toward filling technology gaps that respond to EM needs as quickly as
possible.
DOE Methods is updated twice a year. Both guidance chapters and methods can undergo
revision during this updating cycle. The document is distributed in April and October of
each year. This frequent update cycle ensures that input from readers and new information
are incorporated quickly. All readers are encouraged to provide methods and/or comments
at any time. The publication schedule ensures that reader comments will be addressed and,
if appropriate, incorporated in less than 1 year from the time of submission.
The first issue of DOE Methods contained chapters on quality control, safety, waste
handling, and sampling methods; it also contained four analytical methods and an appendix
on method validation and selection. Since the first issue, nearly all chapters have been
modified. An index and a second appendix (guidance on a performance-based approach to
modified or new methods) have been added. A separate chapter on quality assurance was
included, and more than 40 methods are now part of this document. Over the past year, the
number of comments and corrections received has diminished dramatically, which appears
to indicate widespread acceptance of the contents.
The titles of methods and their distribution by method class are summarized in Table I.
Sampling methods include one addressing radioactive tank waste and two addressing vapor
samples from drums. Analytical methods include all major classes of analytes (organic,
inorganic, radiochemistry). Analytical methods generally include field screening, adapted
methods, and new methods. Adapted methods are those that have been modified, usually
from SW-846, to meet DOE requirements {e.g., as low as reasonably achievable
(ALARA)}. New methods reflect a new approach not described elsewhere. The initial
emphasis was on the inclusion of methods that were focused on high-level mixed waste.
As these needs are filled, additional needs will be addressed, including lower-level mixed
waste and environmental methods.
Sampling and analytical method formats are available (contact Margaret McCulloch at
Pacific Northwest Laboratory, P.O. Box 999, MS P8-08, Richland, Washington 99352) to
guide authors interested in submitting methods (6). These formats are similar to those of
SW-846, with some adjustments made to conform to DOE/EM needs. Methods that are
submitted undergo a peer review process. Many methods are distributed as draft methods
until they have successfully completed the peer review process, and quality control data
meet specifications (7). Once these criteria are met, the method becomes verified.
CONCLUSION
DOE Methods meets a need of EM programs by providing guidance and methods that are
unavailable elsewhere and are needed to support the DOE/EM mission. DOE Methods
734
-------
also makes the capabilities of the DOE complex more accessible, increasing the cost
effectiveness of characterization in support of EM activities at DOE sites.
DOE Methods contains sampling and analysis guidance as well as methods to address
DOE's characterization needs for EM programs.
ACKNOWLEDGMENTS
The DOE Methods Compendium Program is supported by the Laboratory Management
Division of the U.S. Department of Energy. The document DOE Methods for Evaluating
Environmental and Waste Management Samples is produced at Pacific Northwest
Laboratory. Pacific Northwest Laboratory is operated for the U.S. Department of Energy
by Battelle Memorial Institute under Contract DE-AC06-76RLO 1830.
REFERENCES
(1) Goheen, S. C., McCulloch, M., Thomas, B. L., Riley, R. G., Sklarew, D. S.,
Mong, G. M., and Fadeff, F. A. DOE Methods for Evaluating Environmental
and Waste Management Samples, PNL-7722, Rev. 0, Pacific Northwest
Laboratory, Richland, Washington, October 1992.
(2) U.S. Environmental Protection Agency, Test Methods for Evaluating Solid
Waste: Volume IB, Laboratory Manual Physical/Chemical Methods, 3rd
Edition, EPA/SW-846, Office of Solid Waste and Emergency Response,
Washington, DC. Available from National Technical Information Service,
Springfield, Virginia (1986).
(3) American Society for Testing and Materials, 1993 Annual Book of ASTM
Standards, Water (I), Vol. 11.01; Water (II), Vol. 11.02; Nuclear Energy (1), Vol.
12.01, Philadelphia, Pennsylvania (1993).
(4) Greenberg, A. E., Clesceri, L. S., and Eaton, A. D., Eds., Standard Methods for
the Examination of Water and Wastewater, 18th Ed., American Public Health
Association, Washington, DC, (1992).
(5) Chieco, N. A., Bogen, D. C., and Knutson, E. O., Eds., Environmental
Measurements Laboratory Procedures Manual, 27th Ed., HASL-300, U.S.
Department of Energy, Environmental Measurement Laboratory, New York, New
York (1990).
(6) Goheen, S. C., Riley, R. G., Fadeff, S. K., Sklarew, D. S., Mong, G. M.,
Cosby, W. C., McCulloch, M., Brug, W. P., and Poppiti, J. S. "Needed
Sampling and Analytical Methods for the Document: 'DOE Methods for Evaluating
Environmental and Waste Management Samples,'" proceedings of the Second
International Symposium on Mixed Waste, Baltimore, Maryland, August 17-20,
1993.
(7) Goheen, S. C., McCulloch, M., Thomas, B. L., Riley, R. G., Sklarew, D. S.,
Mong, G. M., and Fadeff, S. A. "Appendix A." DOE Methods for Evaluating
735
-------
Environmental and Waste Management Samples. DOE/EM-0089T, Pacific
Northwest Laboratory, Richland, Washington, April 1994.
736
-------
Table I. Methods Included in DOE Methods for Evaluating
Environmental and Waste Management Samples
Method Class
Title
Sampling
General Method for Sampling Liquids and Solids in Low-Level
Waste Storage Tanks
Sampling Headspace Gas for Volatile Organic Compounds Within
a TRU Waste Drum with a Sampling Manifold
Sampling Headspace Gas within a TRU Waste Drum with
SUMMA™ Canisters for Volatile Organic Compounds
Organic
Total Organic Chlorine in Oil, Field Test Kit Method
Immunoassay for Polychlorinated Biphenyls (PCBs) in Soils
A Photoacoustic Infrared Method for the Detection of Selected
Chlorinated Volatile Organic Chemicals (VOCs) in Water
Preparation and Cleanup of Hydrocarbon Containing Samples for
the Analysis of Volatile Organic Compounds
Remote Purge and Trap Gas Chromatography of Volatile
Organics in High-Level Radioactive Wastes
Ultrasonic Solvent Extraction for Volatile Organic Analysis of
Solid Radioactive Mixed Waste (RMW)
Purge and Trap in a Glovebox
PCBs in Aqueous Radioactive Mixed Wastes Using Solid Phase
Extraction Disks and Gas Chromatogrpahy-Electron Capture
Detection (GC-ECD)
Reduced-Scale Liquid-Liquid Extraction of Semivolatile Organic
Compounds
Ultrasonic Extraction in a Glovebox or Hot Cell
Major Nonhalogenated Volatile Organics in Radioactive Aqueous
Liquids Analyzed by Direct Aqueous Injection Gas
Chromatography (DAI-GC)
Analysis of PCBs as Aroclors in Solid Radioactive Mixed Wastes
737
-------
Table I. Contd.
Direct Analysis of Toxicity Characteristic Leach Procedure (TCLP)
Acidic Semivolatile Compounds in Radioactive Liquid Wastes or
Leachates using HPLC with UV Absorbance Detection
Inorganic
Immunoassay for Mercury in Soil
Solvent Extraction of Uranium and Thorium from Radioactive
Liquid Wastes
Cleanup of Transuranic Liquid Wastes using Extraction
Chromatography
Total CN by Microdistillation
An Indicator Strip-Based Test for Chromate Ions (CRO42~) in
Aqueous Samples
An Indicator Strip-Based Test for Lead in Water
An Indicator Strip-Based Test for Nitrate in Water and Soil
An Indicator Strip-Based Test for Nickel (Ni2+) in Aqueous
Samples
In situ Analysis of Gamma-Ray Emitting Radionuclides by
Borehole Logging
Iodine-129 Analysis in Aqueous Solutions
Nickel-59 and Nickel-63 Determination in Aqueous Samples
Separation of Niobium for Niobium-94 and Niobium-93m
Determination
Purification of Strontium in Water Before Strontium-89/Strontium-
90 Measurement
Determination of Total Radioactive Strontium in High-Level
Samples Using Extraction Chromatography
Determination of Strontium-90 in Dissolved Environmental
Samples Using Chelex-100
Determination of Strontium-90 in soil, Water, and Filter Samples
Determination of Selenium-79 in Aqueous Samples
Radiochemistry
738
-------
Table I. Contd.
Technetium-99 Analysis Using Extraction Chromatography
Waste Distillation from Soil and Aqueous Matrices Using a Lachat
Mirco-Dist™ System for Tritium Determination
Laboratory Method for Gross Alpha and Beta Activity
Determination
Rapid Determination of Gross Alpha, Gross Beta, and Gross
Tritium in Water Using a Liquid Scintillation Counter
Gross Gamma Screening for Environmental Matrices
Gamma-Ray Spectrometry
Liquid Scintillation Instrumentation Method
Method for Utilization of Alpha Track Detectors for
Characterization of Gross Alpha Emissions from Indoor Surfaces
Method for Utilization of Electret lonization Chambers for
Characterization of Gross-Alpha Emission from Indoor Surfaces
739
-------
113
QC DATA REVIEW FOR GAMMA RAY/LIQUID SCINTILLATION ANALYSIS
Robert Litman, Chemistry Support Supervisor, NAESCO, Seabrook Station
Seabrook , New Hampshire 03874
ABSTRACT
Independent review of laboratory quality control data can provide valuable insight to the
accuracy and precision of the analytical data produced, as well as provide important information
about laboratory operations. The tools used to review the data are initially statistical, and the
acceptance criteria for hese tools are based on laboratory history, or an 'industry' standard. The
reviewer should not be satisfied with 'the way it's always been', but instead should attempt to look
for the best way to improve laboratory performance. This may mean being innovative or just
inquiring, but in either case the objective should be improvement.
This paper presents the QC methodology we use at Seabrook Station for gamma ray and liquid
scintillation analysis. Specific cases of the utility of independent review are discussed.
INTRODUCTION
An important protocol for radiochemical laboratories to establish is not only how, but who
performs data review for radiochemical measurements. The parameters which effect the
measurement system may be as important to trend as the final QC analytical values. These other
parameters often may provide clues which can trace a problem to its source. Gamma ray
spectrometry and liquid scintillation analysis lend themselves to a myriad of parameters that can
be trended. A choice of what to routinely follow is the judgement of the independent reviewer.
The frequency of QC checks and data summary should be adjusted based on the sample loading,
number of analysts, number of different analyses and laboratory history. The data reviewer
should also review laboratory techniques.
DISCUSSION
Gamma Analysis
In our laboratory, we have four germanium detectors, each with 4 geometries. This is a potential
for 16 different QC checks. We have our geometry positions fixed relative to each other and, as
such, we feel a single QC check for the detector would serve all geometries for one detector.
The source used for QC is a solid, compact Eu source, whose true activity is within about
5% of the activity which is purchased. Since for this QC source we are more interested in the
trended activity, the approximation of its absolute activity is satisfactory. The long half life(13.5
a) is advantageous since it will transcend the repair/replacement of many components such as
pre-amps, amplifiers, cables , etc. A separate source which is very accurately known is used for
detector calibrations. This separate source does not only have different radionuclides from the
QC source, but is also purchased from a separate vendor. Figure 1 shows the gamma ray lines of
interest that we use for the QC source, which transcends the range of interest for us for gamma
ray analysis. Figure 2 shows a plot for the 122 KeV peak; the target value is plotted as a
constant, a horizontal line parallel to the X-axis. Our gamma ray analysis QC software decay
740
-------
corrects the activity to the standard date for the source. If the same data is plotted using the
current activitiy of the QC source(i.e., not decay corrected)and compare it to the theoretical
decay plot we would have a line with a negative slope for the true value. This method, used by
some laboratories puts the analyst at somewhat of a disadvantage because it is difficult to discern
any trends in the data at an angle. Feedback to the analyst on their work is important because it
can' short-circuit' problems in the analysis before they affect laboratory performance. Figure 3
shows a plot of our GO/NO GO check for the FWHM of the 964 KeV peak. This provides us
with a good indicator of the electronic response of the detector, without a rigorous trend
evaluation.
The messages we provide to the analyst for their recorded QC result are shown in Figure 4. Our
warning limit is 2 sigma, our control limit is 3 sigma and our outlier limit is 4 sigma(data points
wihich are outliers are not used in the data summary for statistical purposes). These messages
allow the analylst to participate in the data review process as the data is generated. Other
important parameters reported to the analyst are seen in Figure 5. We choose to use FWHM and
energy checks as a Go/No go value and not as a trend. The background is checked for
peaks(other than natural) and for potential contamination. This responsibility lies squarely with
the analyst for potential corrective actions. Our current software has an additonal feature which
prevents an analyst from analyzing a sampl if the QC check is unsatisfactory. They may take
remedial actions and re-count the QC but not a sample.
The X-axis for the data plots should run for 25-50 analyses, at which point a data summary
should be performed. A typical data summary is shown in Figure 6. The parameters of skewness
and kurtosis are used as tools to help evaluate the data trends, but they have no specific warning
messages. Figure 7 lists the statistical data tests that are performed on each data set.
The t-test compares the data set mean with the target, and uses a standard deciation based on the
historical data set which contains 100 values.
The F-test compares the variances of the current data set with the variance of the last 100 data
points.
The bias mean test compares the data set mean with the target value +0.65 sigma.
The Wilk-Shapiro test is performed on the current data set to determin if the data are normally
distributed.
A printout of the test results provides the reviewer with the information needed to evaluate the
analytical testing in the laboratory. Problem Resolution Reports(PRR) are issued when any of
the statistical evaluations fails the null hypothesis, and when an analyst reveives a message 4, 6,
or 8(see Figure 4). While the individual data entries are reviewed by the technician and
supervisor, the dat summaries are reviewed by an independent analyst, who does not participate
in the data generation. Figure 8 is a typical PRR.
The above data analyses are applied to each detector and each of the gamma rays used in the QC
analysis.
741
-------
Liquid Scintillation
Although we only perform tritium analysis in our laboratory, we apply the same statistical tests to
the data generated for this analysis. Figure 9 shows how the tests are applied. The QC
preparation involves pipetting a standard, and adding an aliquot of the cocktail as would be done
for the sample. It should be noted that on a quarterly basis the entire sample treatment (i.e.
including distillation) is performed on an interlaboratory cross check sample as part of the QC
process. Once again the tritium source used to generate the quench curve and the QC are of
separate origin.
Preparation of the QC in this manner will detect problems with
*glassware contamination
*cocktail decomposition
*pipetting technique
*loss of pipet calibration
*distillation carryover.
The routine checks performed also include
*total counts for a sealed unquenched standard
*background on the low energy channel
*background on full open energy channel
""comparison of the low energy to the full energy background to within 5%
* statistical tests on background and QC trend graphs
Figures 10 and 11 show two months of QC trend graphs. At the end of the first data set, the
mean activity was 1.95E-04 microcuries/ml(the target value for the QC is 1.98E-04). There
were some mechanical problems with the instrument for sample numbers 11 through 28. We had
also been getting lower than normal values for the low energy channel backgrounds(this seemed
to be related to the colder temperature in the count room during the winter months). The sample
analyses were still in control, but these problems masked an underlying long term trend in the
data which were not obvious to the supervisor or the analysts. When the printout for data set in
Figure 11 was reviewed it was clear that a bias existed for the data set; a PRR with the bias mean
low message was received after the statistical data evaluation. The bias limit for the data set was
1.92E-04 and the data set mean was 1.89E-04 microcuries/ml.
**An important point to note here was that only two of the data points in the current data set
were at or beyond the control limits(i.e. a normal data distribution) and we had just confirmed
that our quarterly QC interlaboratory check result was within the NRC acceptance
criteria(-7.0%). **
The investigation examined the following items:
a. Accuracy of the pipet calibration
b.Balance accuracy checks(used for the pipet calibration)
c.Repipet delivery volume(used for aliquoting the cocktail)
d. Operation of the instrument
742
-------
e.Expiration date of the cocktail
f. Contamination of the tritium standard with water(i.e. inadvertant dilution). This was
achieved through review of our source control tog sheets.
g.The quench curve and the original data for it.
h.The previous data set.
In the end it was items f, g, and h which solved the concern. The examination of the source log
sheets and the instrument quench curve showed that a new quench curve had been generated on
January 12(see data point 8 on Figure 10). Initially a low bias appeared to be in the works, but
then the sporadic instrument and low background concerns detracted from the significance of the
initaial drop in activity. The specific problem with the new quench curve was that the analyst
who prepared it forgot to decay correct the tritium to the present time, giving an apparent higher
activity for an equivalent count rate. In the QC procedure the decay correction is part of the
software program, so it appeared that the QC activity was low. When the decay for the tritium
was accounted for in the quench curve, the interlaboratory QC was within 1% of the stated value.
In this review process two important side issues were uncovered.
l.The quench curve procedure was missing a step to decay correct the standard
2.The source log sheets for the quench curve standard and the QC had both been labelled
"QC". This was of significance since the QC had been used to generate the quench
curve.
SUMMARY
The routine evaluation of QC data printouts as described herein takes approximately 15 minutes.
Often PRR's which accompany these ae easily resolved or the corrective actions verified during
that 15 minute review. Typically the analyst has performed the corrective actions and the
reviewer ensures that it was appropriate for the error message received. These actions include
gain adjustment, geometry adjustment, cleaning of the cocktail vial, low cryostat level, wrong
year or month used for decay correction. In the particular case discussed here for tritium,
approximately 5 hours were spent on the investigation; some of this time included interviews
with the analysts. Approximately three hours were needed to correct the problems.
The analyst and the supervisor are often 'too close' to minor QC problems to see them. It is
advantageous to have an independent resource examine the data and evaluate potential long term
trend problems. In our case the statistical treatment of our data provided the trigger to
investigate- the independent reviewer was able to look at the process without worrying about
sample schedules.
743
-------
FIGURE 1
152mEu t1/2=13.5a
Gamma 1 = 121.8KeV
Gamma 2 = 964.1 KeV
Gamma 3 = 1408 KeV
744
-------
QA filename
Parameter Name
Start/End Dates
Mean +- Std Dev
CAS$DISK:[SNS.QAF]ADC1_DCAL.QAF;22
NLACTVTY-121 (DECAY CORRECTED ACTIVITY 0121KEV)
10-FEB-1994 16:11:26 through 10-MAR-1994 00:00:00
1.54000 +- 3.922500E-02 (2.55 %)
1. 6
5!
01
o
-------
QA filename
Parameter Name
Start/End Dates
Lower/Upper Lmts:
: CASSDISK: [SNS.QAF]ADC1_DCAL . QAF; 22
: PSFUHM-1407 (PEAK FUHM - 1407)
: 10-FEB-1994 16:11:26 through 10-MAR-1994 00:00:00
1.50000 through 2.50000
2. 5
I
z
•r.
-------
FIGURE 4
Messages to Analyst After QC Data Input
1 .Data point is within one sigma of target
2.Data point is within warning limit
3.Data point is within control limit
4.Data point is fourth consecutive one beyond warning limit
5.Data point is beyond control limit
6. Second consecutive data point beyond control limit. STOP
analysis
7.Data point is beyond outlier limit
8. Data point is seventh consecutive point beyond the bias limit
747
-------
FIGURE 5
Other Parameters Checked As go/no go Values
FWHM <2.3 KeV @1408 and _< 1.3 KeV @ 122 KeV
Energy E^ ± 0.7 KeV
Background Trended like QC activity(daily short count)
748
-------
Date
Time
Countroom Quality Control Charts
CH-L-722
DETECTOR 664 QC DATA
13-JAN-93 to 09-FEB-93
Tech 121 121 Centroid 1299 1299 Centroid Energy Drift 121 Kev 121 964 Kev 964 1408 1408 BKGD BKGD
FWHM Flag Channel FWHM Flag Channel Drift Flag Activty Flag Activity Flag Activity Flag cnts Flag
1
2
3
4
5
6
7
8
9
10
11
12
13
^J
£ isr
16 V.
17
18
19
20
21
22
23
24
13-JAM-93
14-JAN-93
15-JAN-93
16-JAN-93
17- JAN -93
18-JAN-93
19-JAN-93
20-JAN-93
21-JAN-93
22-JAN-93
23-JAN-93
24-JAN-93
25-JAN-93
25-JAN-93
V 26-JAH-93
J 30-JAN-93
31-JAN-93
01-FEB-93
03-FEB-93
03-FEB-93
04-FEB-93
05-FEB-93
06-FEB-93
07-FEB-93
20:54:00
19:42:00
18:35:00
12:31:00
17:30:00
19:32:00
19:51:00
20:36:00
17:15:00
17:40:00
18:00:00
17:00:00
21:25:00
22:35:00
08:53:28
13:08:00
13:29:00
21:17:00
01:01:00
22:39:00
21:05:00
22:26:38
17:08:00
22:11:00
JAT
AAG
SDA
SDA
SDA
SDA
SDA
SDA
HAG
HAG
HAG
HAG
HAG
HAG
DAR
JDB
JDB
JDB
JDB
JDB
HJP
HJP
HJP
AAG
1.1
1.1
1.1
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
243.91
243.94
243.95
242.52
242.63
242.61
242.60
242.63
242.61
242.64
242.67
242.72
242.67
242.66
242.77
242.89
242.86
242.89
242.85
242.87
242.89
242.97
242.97
1.7
1.7
1.5
1.8
1.7
2.1
2.1
2.0
2.1
2.1
1.9
1.9
2.0
2.3
2.0
2.0
2.0
1.7
2.1
2.0
1.9
2.3
2.2
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
2615.17
2615.84
2615.27
2595.62
2596.66
2597.18
2596.56
2596.59
2596.52
2596.43
2596.90
2596.95
2597.47
2597.54
2598.65
2598.56
2598.70
2599.11
2598.84
2599.20
2598.93
2599.61
2599.57
0.10
0.42
0.25
-0.26
0.43
0.56
0.36
0.41
0.44
0.35
0.60
0.54
0.91
0.09
-0.17
-0.21
-0.14
0.22
0.08
0.12
0.18
0.44
0.36
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
FAIL
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
ok
0.107 1
0.108 1
0.107 1
0.107 1
0.106 1
0.106 1
0.106 1
0.107 1
0.107 1
0.106 1
0.108 1
0.106 1
0.106 1
0.108 1
0.107 1
0.107 1
0.105 1
0.106 1
0.105 1
0.106 1
0.108 1
0.107 1
0.107 1
0.106
0.104
0.105
0.106
0.104
0.103
0.109
0.105
0.103
0.106
0.104
0.108
0.105
0.106
0.107
0.106
0.103
0.104
0.105
0.104
0.104
0.105
0.103
1
1
1
1
1
2
2
1
2
1
1
2
1
1
1
1
2
1
1
1
1
1
2
0.101
0.107
0.109
0.107
0.105
0.106
0.107
0.108
0.104
0.110
0.108
0.107
0.109
0.109
0.105
0.106
0.105
0.104
0.103
0.106
0.108
0.107
0.105
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
780
809
820
865
773
813
813
818
810
836
816
823
793
793
803
840
8*. 9
799
801
808
811
775
768
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Flag messages:
1 = Result is inside the bias limits.
2 = Result is inside the warning limits.
3 = Result is inside the control limits.
4 = Possible trend outside warning limit developing.
Problem/Resolution Report is required.
Supervisor £_
Chemist
Date
Date
5 = Result is at or exceeds the control limits, a second
control check is required.
6 = Second control check failure, Problem/Resolution
Report is required.
7 = Result is at or exceeds the outlier limit.
-------
FIGURE 7
Statistical Evaluations
t-test
F-test
Bias Mean Test
Wilk-Shapiro Test of Normality
750
-------
Chemistry Department Problem/Resolution Report Form
CH-L89
I. Tech: System Generated Date: 23-APR-94 19:58:55
Instrument: LS 1800 Analyte: Activity Check
Target Value: 1.980e-04 uCi/m Result: 1.912e-04 uCi/ml
Upper Limit = 2.040e-04 uCi/m Lower Limit = 1.920e-04 uCi/m
ote : The mean for the Activity Check data set is less than the XBIAS'
limits of 1.920e-04 uCi/ml for this analysis. Please initiate
a Problem/Resolution Report as required by JS0999.001.
I. Supervisor:
A. Corrective action(s) recommended:
B. Are any procedural changes required? YES NO
If YES describe the revisions and list the date submitted.
C. Review ALL data since the last satisfactory QC check on this
analysis. SAT UNSAT If UNSAT describe actions taken and date,
751
-------
FIGURE 9
Liquid Scintillation
Low Energy Window Background
Full Energy Window Background
Sealed Source Counts(go/no go)
Preparation of Daily QC as a Sample
752
-------
FIGURE 10
Countroom Quality Control Charts
CH-L-725
LS 1800 Activity Check Results
07-JAN-94 to 12-FEB-94
a
co
2.206-04-
2.156-04-
2 . lOe— 04-
2.056-04-
U
9 2.006-04-
1
ro i.95e-04-
1
1.906-04-
1.856-04-
1.756-04-
(
^
*
*
*
* *
*
* * *
* * * *
it
*
*
* *
_.. .. .. .. .. .. .. .. .. .. £..- .. ±.. .. .. .. .. .. .. .. .. ..
*
) 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 3
Sample Number
* Activity Check Results
TARGET VALUE = 1.980e-04
BIAS LOW — 1.92O6— 04
WARNING LOW = 1.8606-04
WARNING HIGH = 2.100e-04
== CONTROL LOW = 1.800e-04
-. r r-ONTTVnT. HTflH — "5 1 fif1r>-n/l
-------
FIGURE 11
Countroom Quality Control Charts
CH-L-725
LS 1800 Activity Check Results
12-FEB-94 to 07-MAR-94
•vl
en
->
2.206-04-^
2.156-04-
. lue— 04
2.05e-04-
U
9 2.006-04-
1
m 1.95e-04-
1.906-04-
1.856-04-
. o ue u^
1.756-04
* *
*
* * * *
* * *
* *
* *
*
*
* * * * *
*
i i i i i i i i i i i i
3 2 4 6 8 10 12 14 16 18 20 22 24 26
Sample Number
* Activity Check Results
TARGET VALUE = 1.9806-04
RTTVQ HTPH — 9 nAOo — flA
T5TAC T ni*7 T QOH^i H A
D-L/\b LiUW — j-.y
-------
114
NATIONAL INSTITUTE of STANDARDS and TECHNOLOGY'S MEASUREMENT
QUALITY ASSURANCE PROGRAMS for IONIZING RADIATION
Kenneth G.W. Inn, Jimmy Humphreys and Jileen Shobe, National Institute of
Standards and Technology, Ionizing Radiation Division, Building 245, Room C229,
Gaithersburg, MD 20899.
ABSTRACT
The Ionizing Radiation Division of the National Institute of Standards and
Technology (NIST) has implemented a number of quality assurance programs to
provide a consistent basis for national and international ionizing radiation
measurement credibility and comparability. The programs cuts across a variety of
sectors that include: 1) personnel protection; 2) survey instrument calibration; 3)
medical diagnostics; 4) medical therapy; 5) radiopharmaceuticals; 6) nuclear power
plant radiochemistry; 7) regulatory agencies; 8) environmental; and 9)
radiobioassay. The four basic elements of the MQA programs are: 1) conformance
to fundamental consensus operational criteria; 2) documented in-house quality
assurance and control practices; 3) periodic performance evaluations using
appropriate testing materials and instruments; and 4) periodic on-site assessments
by technical experts. The periodic performance evaluations are particularly
important for the demonstration of measurement traceability to the national and
international physical standards. Traceability alone, however, must be augmented
by the other elements to provide the strongest rational for measurement assurance.
755
-------
115
DEVELOPMENT OF A MIXED-ANALYTE PERFORMANCE EVALUATION PROGRAM
FOR THE ENVIRONMENTAL RESTORATION AND WASTE MANAGEMENT OFFICE OF
THE UNITED STATES DEPARTMENT OF ENERGY
Dr. Stan Morton
Senior Scientist
United States Department of Energy
Radiological and Environmental Sciences Laboratory
850 Energy Drive
Idaho Falls, Idaho 83402
Telephone 208/526-2186
FAX 208/526-2548
The United States Department of Energy (DOE) has established the Office of Environmental
Restoration and Waste Management (EM) to oversee the program activities within the complex.
The Laboratory Management Division (EM-563) in the Office of Technology Development is
charged with assuring adequate analytical capabilities, quality assurance, and sample management
for the EM programs. To address these issues, EM-563 has developed an Analytical Services
Program consisting of: Analytical Support, Resource Planning, and Quality Assurance. The
Quality Assurance function includes the development, coordination and management of quality
assurance guidance documents, performance evaluation programs, and field and laboratory
assessment programs.
Increasing regulatory requirements demand that analyses performed by DOE quantify
nonradioactive compounds or elements in addition to the radioactive isotopes of interest.
Adequate quality assurance requires quality control maticos representative of the routine sample.
None of the major performance evaluation programs currently in use are designed to address
radioactive and nonradioactive constituents in a single sample matrix.
At the request of EM-563, a new performance evaluation program was developed by the
Radiological and Environmental Sciences Laboratory (RESL) that combines radioactive and
nonradioactive analytes into one analytical sample. The establishment of this program creates a
new category of performance evaluation called mixed-waste or more aptly the mixed-analyte
performance evaluation program (MAPEP). Participation requires analysis of only those
constituents that are a component of the facility's routine analytical work load. At the completion
of each round, a published report will present the participant's results using a scoring system
rather than a ranking system.
756
-------
The program will distribute samples on a semiannually basis. The samples, in various matrices,
will contain types and concentrations of analytes that are typical of those found in the complex.
The radioactive constituents will contain gamma, beta and alpha emitting isotopes at
concentrations ranging from one to 1000 Bq/1 or Bq/kg. The nonradioactive constituents will
consist of some or all of the metals listed in 40 CFR part 261, Appendix III. The concentration of
those metals listed in part 261.24, Table 1, will not exceed the regulatory levels as defined therein.
The program is designed to provide information on the quality of radiological and nonradiological
analytical techniques used by all laboratories on which DOE is relying for EM sample analyses.
The MAPEP will be a major part of the Integrated Performance Evaluation Program being
developed by EM.
757
-------
116
IMPLEMENTATION OF A FOURIER TRANSFORM INFRARED SPECTROPHOTOMETER
FOR THE DETERMINATION OF VOC'S IN WASTE DRUM HEAD SPACE
W.F. Bauer andM.J. Connolly, Idaho National Engineering Laboratory, EG&G Idaho Inc., Idaho
Falls Idaho 83415, D. Gravel and A. Rilling, Bomem (Hartman & Braun Inc.), Quebec, Quebec,
Canada G2E5S5, and F. Baudais, Applied Automation (Hartman & Braun Inc.), Willowbrook
Illinois 60521
ABSTRACT
The Conditional No-Migration Determination (NMD) for the Waste Isolation Pilot Plant (WIPP)
requires that a representative waste drum head space sample be collected for analysis. The NMD
also requires that the head space of all layers of confinement within the drum be sampled. This
level of sampling will be required until it can be demonstrated that existing process knowledge is
adequate for waste drum characterization. The Idaho National Engineering Laboratory has been
involved in the development and evaluation of methods for the analysis of RCRA (Resource
Conservation and Recovery Act) constituents in the gaseous head space of the waste drums.
Currently this analysis is performed by gas chromatography (GC) and gas chromatography/mass
spectrometry (GC/MS) of samples collected in SUMMA canisters. The ability to do "at-line"
analysis would significantly reduce the necessary sample handling and therefore the cost of each
VOC analysis. Fourier transform infrared spectroscopy (FTIR) was selected as a GC/MS
replacement for the analysis of volatile organic compounds (VOC's) since FTIR instrumentation is
relatively simple, durable and can easily be operated "at-line" A two phase study to assess the
feasibility of using FTIR for VOC analysis in waste drum head space was initiated. In the first
phase it was determined that FTIR could be used to identify and quantitate at least 25 gaseous
analytes simultaneously with detection limits for most analytes being within the 1-10 ppmv-m
range. The results from this first phase were used to develop the requirements for the second phase
of the study. This second phase of the study is currently under way and involves the evaluation of
a "turn-key" FTIR gas analysis system installed at a waste drum characterization facility.
The "turn-key" FTIR system was intended for use by technicians and as such is designed to operate
with minimal input from the technician. This FTIR system is directly interfaced to a preexisting
automated gas sampling system (GSS) used to collect SUMMA canister samples for laboratory
analyses. When sample volume permits, an additional aliquot of the gas sample is pulled into the
FTIR for analysis and the results compared to the GC-MS analysis results when they become
available. A single pass 20 cm cell is used in the FTIR system to accommodate the wide range of
concentrations expected in the waste drum head space while maintaining adequate detection limits.
The "turn-key" system is able to identify and quantitate 29 target VOC's and the C,-C3
hydrocarbons.
INTRODUCTION
The Idaho National Engineering Laboratory (INEL) is currently participating in the Waste
Isolation Pilot Plant (WIPP) Experimental Test Program (WETP). The goal of this program is to
collect data in support of a disposal decision for the WIPP In addition to the data required for
758
-------
WETP, the Environmental Protection Agency (EPA) has issued a Conditional No-Migration
Determination (NMD)1 for the WIPP which requires that a representative drum head space sample
be collected and that all layers of confinement be sampled until it can be demonstrated not to be
necessary. The uses of the data generated from these activities include verification of process
knowledge, further waste characterization, verification of gas generation and transport models, and
determining the suitability for transport to and acceptance in the WIPP.
The majority of the gas samples collected at the INEL will come from either the Drum Vent
Facility (DVF) or the Waste Characterization Area (WCA). At the DVF, waste drums will pass
through and gas samples collected at a rate of 1 every 5-8 minutes. At the WCA, only 2-3 waste
drums will be examined per week, however, multiple samples will be collected for each drum in
order to sample the drum head space and the head space of all the inner layers of confinement
within the drum.
The constituents of the waste drum head space are normally identified and quantitated by gas
chromatography (GC) and gas chromatography/mass spectrometry (GC/MS). The analysis of the
29 organic target compounds"1 typically costs on the order of $1000+ per sample and has a
relatively poor turn around time, typically at least 2 days. The analysis takes a minimum of ~4
hours. This time includes sampling into a SUMMA canister, transport of the sample to the
analytical laboratory, initial dilution of the sample, and the first GC analysis. The initial dilution
and analysis may need to be followed by an additional dilution and analysis, depending upon the
results of the initial analysis. For the routine analysis of drum head space on the thousands of
existing waste drums, the layers of confinement within the drum, and long-term monitoring of these
drums, the time and cost limitations of these GC methods are obvious and clearly prohibitive. An
alternative analysis technique that is rapid, reliable and easily operated "at-line" is needed. Fourier
transform infrared (FTIR) spectroscopy is a technique that can potentially answer this need.
With FTIR, relatively high resolution (0.5-2 cm"1) spectra can be collected in times as short as
seconds. Even with scan averaging to reduce the signal-to-noise ratio, spectra can be collected and
processed within minutes of sample collection. Qualitative identification is made from the IR
spectrum "fingerprint" of the compound and quantitation is performed using the Beer-Lambert law:
a(v) = a(v) c b
where a(v) is the absorbance of the sample at frequency v, a(v) is the molecular absorption
coefficient, c is the analyte concentration and b is the path length of the sample cell. The values of
a(v) are typically determined from a series of standards at known concentrations. When multiple
analytes are absorb at a given frequency, the total absorbance is a simple sum of the contributions
from each analyte.
Even though it is relatively easy to identify molecules by their infrared spectra, quantitation at the
part-per-million (ppmv) to part-per-billion (ppbv) level is difficult because the absorptivity
coefficients are typically quite low in the infrared region. In the gas phase, the poor detection
limits due to low absorptivities can be easily overcome by using longer path length cells.
Quantitation difficulties can be amplified when more than one analyte is present. Statistical
analysis tools such as classical least squares (CLS)3 and partial least squares (PLS) can be used to
759
-------
separate and quantitate the components of the spectrum. The target list of 29 organic analytes, as
well as some additional inorganic anaMes, appear to be well suited to determination by FTIR
spectroscopy. Analysis times on the order of minutes instead of hours should be easily achieved.
This study was originated and designed to assess the feasibility of utilizing FTIR spectroscopy to
determine the volatile components in the head space of waste drums containing transuranic waste
(TRU). The FTIR laboratory demonstration represented Phase I of a two phase feasibility study
aimed at demonstrating the applicability of FTIR in performing near real-time, at-line analysis of
drum head space volatile organic compounds (VOCs). Phase II consists of drum and inner layer of
confinement head space analysis by FTIR by a "turn-key" system installed "at-line" at the WCA
with subsequent comparison to results from SUMMA canister collection followed by GC/MS
analysis. The objective of the two phase feasibility study is to demonstrate to the U.S. EPA that
FTIR can be used to perform drum head space VOC analysis and yields comparable results to
SUMMA canister collection followed by GC/MS or meets established WIPP program requiments.
EXPERIMENTAL
Instrumentation: Bomcm furnished the FTIR based VOC system as specified by EG&G Idaho,
Inc. Because the "turn-key" FTIR system is located in a "suspect radiation contamination zone"
near where the waste daims are handled and mounted into the waste characterization hot cell at the
WCA, the hardware is enclosed in a NEMA 12 enclosure (48" tall x 36" wide x!6" deep) for
protection from damage and from contamination by radiation. Figure 1 is a schematic of the
components comprising the FTIR based VOC analysis system. A Bomem MB 100 series FTIR is
vertically mounted and equipped with a specially designed top plate. The optical bench is purged
with hydrocarbon and CO: free dry air which is vented into the NEMA 12 enclosure to help
maintain a slight positive pressure within the enclosure. A 20 cm gas cell with zinc selenide
windows coated with an anti reflection coating to reduce the refractive index and a DTGS detector
are mounted on the top plate. All sample transfer lines and the sample cell are maintained at
110°C. Transducers are mounted in the cell to record the temperature and pressure of each
sample. Because of the heat load supplied by the instrumentation and the other heated components,
the NEMA 12 enclosure needed to be cooled and is maintained at ~28°C with a closed cycle air
conditioner. Operation of the valves and the FTIR are controlled via RS232 and RS422,
respectively from a 486 based PC located >60 feet away at the WCA control center.
Sampling and analyses is initiated when the start signal is received from the GSS computer. At
this point the three way valve (VI) rotates toward the sample line and the cell and lines are
evacuated to <5 Torr by opening V4. The cell and lines are backfilled to -625 Torr with HC/C02
free dry air by closing V4 and opening V3. The cell and lines are reevacuated by closing V3 and
opening V4. Once evacuated to < 5 Torr, V4 is closed and the "ready evacuated" signal is sent to
the GSS computer which then opens VO and the lines and cell begin to fill with the sample. When
the pressure has stabilized, VI is rotated to isolate the cell and spectral acquisition is initiated.
Each spectrum is the result of 10 coadded scans.
Generation of library spectra: Because quantitative IR spectra of several of the VOC's of
interest were not available at all or not available at the conditions at which the sample spectra were
to be recorded, it was decided to record the quantitative spectra at the sampling conditions, i.e. at
760
-------
110°C with a nominal pressure of 640 Torr (near ambient pressure in eastern Idaho). A second
manifold which included a cold finger was constructed and connected to a second MB 100 series
spectrometer with an identical top plate. After inserting the sample of the neat analyte into the cold
finger and attaching it to the manifold, the sample was frozen with liquid nitrogen. The space over
the frozen sample in the cold finger, the manifold and heated gas cell were evacuated using a
mechanical and a turbo pump in series. Once a stable vacuum was reached, the cold finger
containing the sample was isolated by closing the valve, the liquid nitrogen was removed and the
sample heated to room temperature. The valve to the sample was then opened and the sample
allowed to "evaporate" into the evacuated manifold and cell until the desired partial pressure of the
analyte was reached (usually -0.64 Torr). The total pressure was then brought to 640 Torr with
nitrogen and the spectaim was acquired with 50 coadded scans at 1 cm"' resolution. Many of the
more polar and less volatile VOC's presented problems that were likely due to adsorption of the
analyte onto the walls of the cell and manifold. The addition of nitrogen to the sample to bring it to
640 Torr seemed to enhance the instability of the analyte partial pressure. For these cases
additional nitrogen was not added to bring the total pressure to 640 Torr. There were no apparent
differences in the spectral features of the analyte spectra recorded at low and high pressure.
Linearity of the samples created in this manifold system was verified by using carbon tetrachloride
to construct a calibration curve from 0 to 1000 ppmv.
Computer simulated spectra and CLS analysis: It would be extremely difficult and extremely
labor intensive to make and test in the laboratory enough gaseous samples containing the 29
analytes for a full evaluation of the potential and limitations of the FTIR technique. For this
reason, it was decided that computer simulated spectra would provide an alternative way of looking
at as many potential combinations of components as possible. Table I lists the 29 volatile organic
target compounds of interest. The maximum concentrations and the occurrence frequency
represent the results from ~93 vented drum and inner bag head space samples as determined by GC
and GC/MS analyses. The frequency and concentration ranges found in the table were used as
inputs for the generation of the unknown spectra used in the study. Also listed in Table I are the
upper concentrations noted in samples from 210 unvented as analyzed by gas phase mass
spectrometry5.
During the first phase of the study, simulated spectra were calculated from known spectra obtained
from the MDA spectral library (MDA Scientific, Inc., Lincolnshire, IL, 1991). In the second
phase, the new spectra that were acquired as described above were used. All calculations were
performed using Gauss - 386i Version 3.0 (Aptech Systems, Inc., Maple Valley, WA), a matrix
oriented statistical programming language. Spectra were linearly combined by randomly selecting
the components for the new spectrum at the occurrence frequencies found in Table I. The
concentrations for the components were also randomly determined within the range extending from
0 ppmv to the maximum found in Table I for the first phase. In the second phase, all
concentrations were assumed to range from 0-2000 ppmv with the exceptions being ethane at 1000
ppmv and propane at 500 ppmv. Water was added to each spectrum in concentrations ranging
from dry to saturated. Carbon dioxide was also added to each spectrum in concentrations ranging
from 0 to 10000 ppmv. Since there is some potential for NOX, N0: was also added to 10% of the
spectra in the range of 0-500 ppmv. Random noise of ±10"3 absorbance units (AU) was added to
each spectrum. Randomly fluctuating backgrounds consisting of a simple offset, a sloping line
with an offset, or a curved line (defined by x-log(x) where 0
-------
added as desired in either a positive or negative direction. Concentrations of the components in the
new spectrum were then determined using the CLS approach. For each set of conditions, 1000
spectra were generated and quantitated.
As already stated, the concentrations of the newly generated unknown spectra were determined by
CLS. This activity was performed in the both the "fingerprint region" of the spectrum using all
data between 671 and 1400 cm'1 and the "CH" stretching region extending from 2750 to 3225 cm'1
Water, C02, N0:, methane, ethane, propane, cyclohexane, an offset line and a sloping line were
included in the calibration set as interferences in the 671-1400 cm"' region. With the interferences,
a total of 37 components were in the calibration set for the 671-1400 cm'1 region. The primary
analytes in the 2750-3225 cm"' region were cyclohexane, methane, ethane and propane, however,
32 components were in the calibration set to accommodate potential interferences including offset
and sloping baselines. In both cases, the calibration sets were made from the pure component
spectra of the newly acquired library. Prior to forming the actual calibration matrices for each
region, each library spectrum was carefully examined and appropriate baseline corrections applied
using the baseline correction functions found in Spectra Calc (Galactic Industries).
The predicted concentrations of the computer simulated spectra were evaluated using the standard
error of the estimate (SEE):
I\~
predict — Ctruej
SEE =
where cpredlct and ctnte are the predicted and true concentration values for a component in the n
computer generated spectra. The SEE value is generated for each component and can be viewed as
the standard deviation of the predictions for each component given. Assuming this is a true
reflection of the standard deviation for the frequency of occurrence and concentration ranges for
the analytes, detection limits can be estimated as 3 x SEE. The detection limits calculated in this
fashion are likely to be most appropriate for the occurrence frequencies and concentration limits
used to calculate them. High SEE values for a given analyte indicate that the absorptivity for that
component is low and/or that there is a significant spectral interferences that limits the detectability
of that analyte.
PLS methods: PLS methods were generated using Galactic Industries PLSplus add-on package to
Lab Calc. Because of software limitations and the cumbersome calibration sets required for a
single method that would quantitate all 29 target VOC's and the CrC3 hydrocarbons, individual
PLS methods were used for each analyte in a selected region of the spectrum. Calibration sets
consisted of duplicate spectra of each analyte and potential interference (n=70). Wavelength
regions for each analyte were selected after evaluating the correlation spectra for that component.
Optimum numbers of factor for each analyte method were selected from the evaluation of the
predicted residual error sum of squares (PRESS) values determined using the diagnostic routines
provided in the PLSplus software. Evaluation of an unknown spectrum for the 32 analytes using
the 32 separate methods takes only -10 seconds. These PLS analysis methods are integrated into
the operating software of the FTIR system installed at the WCA. Complete analysis and reports
are provided within 5-6 minutes of sample introduction to the system
762
-------
RESULTS AND DISCUSSION
During phase I of the study, evaluation of FTIR spectroscopy for the determination of VOC's in
waste drum head space was performed using both computer simulated spectra with up to 25
components and actual spectra of 5 component mixed gas standards. The CLS methods used to
evaluate the computer generated spectra appeared to be quite adequate since the SEE values were
reasonably low and, therefore, the detection limits for most components were estimated at 1-10
ppmv-m. At this detection level, it was determined that reasonable detection limits could be
achieved in a somewhat shorter cell. Using a shorter sample cell would also maximize the working
range, a positive factor considering the very wide range of concentration noted in Table I for many
of the analytes.
The goal of phase II is to demonstrate that FTIR spectroscopy is a reliable tool for the analysis of
waste drum head space and can be operated "at-Iine" by facility technicians. The "turn-key" FTIR
system for the analysis of VOCs described above has been installed at the WCA and is currently in
operation. The use of the 20 cm path length cell in this system allows sufficient light throughput so
that a DTGS detector can be used instead of a liquid nitrogen cooled MCT detector. Since the
system is mounted in a radiation contamination area, the use of the DTGS detector eliminates the
need to transport dewars of liquid nitrogen in and out of this area. Operation of the system is quite
simple. After collecting a reference spectrum, no additional input is required by the operator. The
signal to start the sampling sequence comes from the GSS computer, spectral filenames are
assigned by date and time, and the PLS analysis results are sent to the printer. Total analysis
times, including a second spectrum of the diluted sample, are on the order of 5-6 minutes. So far
this system has proven to be reliable and the major efforts are now focused on collecting sufficient
data to evaluate FTIR spectroscopy for VOC analysis. Included in this effort is refining and
optimizing the quantitation algorithms.
Comparison of spectral analysis methods by evaluation of detection limits. Direct comparison
of the spectral analysis methods used in this work with those used elsewhere is difficult because of
the number and kinds of components. However, the detection limit values estimated in these
studies can be crudely compared to other previously published values. Detection limits for the CLS
methods were calculated from the SEE values of 1000 computer generated spectra with 7.5x10"*
noise and randomly selected backgrounds applied over the spectral range from 500-3300 cm"1
These conditions were selected after viewing the spectra obtained from the functioning system
installed at the WCA. Detection limits for the PLS methods were determined from 30 blank
spectra collected on the system used to generate the new calibration spectra. Table II compares the
detection limits estimated in this work with those extrapolated from the published in reference 6.
Overall, the detection limits for the CLS and PLS methods used in this work compare quite
favorably with those extrapolated from previously published values6 The detection limits in this
work are somewhat higher than the extrapolated values, but this is to be expected because of the
large numbers of components in the methods. Using CLS on the first derivative spectra obtained
by applying a 5 point derivative smooth7 is quite efficient in eliminating the various background
components and identifying the analytes actually present, however, it also has a tendency to
763
-------
dampen broad spectral features and therefore the method typically has poorer detection limits for
analytes with broad spectral features.
Detection limits for the PLS and CLS methods are essentially the same for most components
considering the differences in how they were determined. Some compounds do appear have
somewhat higher detection limits using the PLS methods. This can likely be attributed to a
combination of several factors including the fact that only two of the 70 spectra in the calibration
set actually contained the analyte of interest and that only narrow spectral ranges (<50 cm"1) were
used in the methods. The high detection limit for ethylbenzene using the PLS method is likely
attributable to its poor absorptivity coefficients, spectral interferences from carbon dioxide and as
many as 9 other VOC's in the window from 690-710 cm"1 Thirteen factors were required to
describe ethylbenzene in its PLS method. The CLS detection limit for ethylbenzene is also among
highest for that calculation method. In general, higher detection limits can be noted in the PLS
methods for analytes requiring a larger number of factors to describe a relatively narrow spectral
range (e.g. 25 cm"1).
Evaluation of FTIR based VOC analysis system installed at the WCA. Long term evaluation
of the FTIR methodologies for determining VOC's in waste drum head space will be the result of
two major tests. Repeated analysis of a routine reference standard will be used to evaluate long
term reproducibility and accuracy. FTIR analysis results of actual waste drum head space gases
will be directly compared to replicate samples collected in SUMMA canisters and analyzed by the
normal GC and GC/MS methods. Currently, no direct comparisons of the results from the GC
methods to the FTIR results have been made because the GC analysis is still pending. Several
analyses of the routine reference standard and of other standard mixtures have been made with the
FTIR system.
Table III is a summary of the results from the repeated analysis of the routine reference standard
over a period of several months. Precision is very good as relative standard deviations (RSDs)
range from 2-16%. The PLS and CLS methods produce comparable results with the possible
exception being for dichloromethane where the recovery is somewhat lower and the RSD is higher
for the PLS method. There appears to be a significant bias as most of the compounds are
recovered only in the 80-90% range. Methanol, however is only recovered at -35% and methane is
recovered at 93-95%. This reference gas mixture is a dry gas in a dry cylinder and must be
transported through -25 feet of tubing before it reaches any heated lines on the GSS. We believe
that methanol is adsorbed onto polar sites on the walls of the tubing and manifold and that once
adsorbed, it can help to reduce some of the remaining VOC's as well. Methane is very volatile and
not likely to be affected by this recovery problem. SUMMA canister samples are often stabilized
with water vapor to prevent this problem from occurring8 In order to test this concept, the routine
reference standard cylinder was replaced with a 6 L SUMMA canister pressurized to -25 psig with
a mix of 5 polar VOC's, including methanol, at -100 ppmv. Water was added to this standard at
up to 10000 ppmv. In this case methanol was recovered at 95-98% indicating that either the
addition of water to the reference gas standard will be required or the additional transfer line from
the reference gas cylinder to the GSS will need to be heated.
Examination of the data in Table I demonstrates that very rarely are there likely to be more than
four or five VOC's in a waste drum sample. Analysis of FTIR spectra with such a small number
764
-------
of analytes is done quite readily. If, however, a drum containing all or most of the potential
analytes each at a relatively low concentration but totaling to a significant total VOC
concentration, could the FTIR technique identify this situation. To answer this question, a mixture
containing 28 of the 29 VOCs was prepared in a 6 L SUMMA with up to 10000 ppmv water.
Concentrations ranged from 25-50 ppmv and were verified via the standard GC and GC/MS
methods. The 6 L SUMMA was then connected to the GSS in place of the routine reference
standard.
The FTIR analysis results of this mixture are presented in Table IV These results are quite
encouraging since both PLS and CLS identified most of the compounds present. The PLS methods
were able to identify -80% of the compounds it should have detected. The CLS method identified
-90% of the compounds it should have seen. In this particular instance, the big difference between
the CLS and PLS results is that many of the VOC concentrations were slightly closer to the
detection limits for the PLS methods and may have caused some identification and/or quantitation
problems for some of those analytes. For example, 1-butanol should have been present at -54
ppmv and the PLS detection limits were slightly less than 53 ppmv, however poor recovery was
also noted in the CLS method.
The results in Table IV also point out that some PLS methods may require additional method
refinements. In particular, there appears to be a spectral artifact that the PLS methods for ethane
and propane interpret as the analyte at significant concentrations. Spectral residuals are also high
for these methods. Dichloromethane was not identified at nearly twice the detection limit and 2-
butanone which uses a spectral region with significant interference from water is not detected at
relatively low concentrations even though it is readily picked up by the CLS method. 2-Butanone
is noted often in the spectra at concentration of 15-25 ppmv and is believed to be the result of some
contamination of the GSS that occurred when sampling of the ambient air was being performed
while some epoxy based paint was being applied at the WCA.
Also to be noted from Table IV is that methanol was recovered from this "wet" sample at >94%.
Recovery of other polar VOC's is also good except for 1-butanol and 4-methyl-2-pentanone.
These two compounds were also not recovered very well in the sample containing the five polar
VOC's. For 4-methyl-2-pentanone, it appears that using the absorption bands in the
2750-3225 cm"' as the primary quantitation region for this compound using CLS since the recovery
is >90%. Similar improvements were not noted for 1-butanol possibly indicating that a transport
problem still exists.
CONCLUSIONS
Even with the limited test data currently available, it appears that FTIR spectroscopy is a suitable
and cost effective method with which to analyze TRU waste drum head space at-line at both the
WCA and, in the near future, the DVF. Sampling and analysis can be completed within 5-6
minutes and the analysis results are immediately available. The PLS and CLS spectral analysis
methods are very comparable although some refinement of both methodologies is still required.
This The most significant errors in the analyses performed so far appear to be related to sample
transport and not the analytical technique.
765
-------
a
d
t
3
1 1 rj 1
l_ 1 _J
L I i
i ' ' r
J— ^
RS"2 ,/o
f~ CONTROLLER
1
u i_ J. .
i r
I
(READY EVACUATED) — .
— nr+'i '
(ACK.DATA ACOOI.) 1
1
1
r
1 —
1
1
J
FUR k SAMPLING
COMPUTER
DIGITAL
Figure 1. Schematic of FTIR based VOC analysis system.
SAUPLE OUT
NOTE: ALL PARTS UNOER
SIGN- I I
ARE ELECTRICALLY
HEATED
766
-------
Table I. Summary of data from previously sampled waste drums.
Compound
1,1,1-Trichloroethane
Toluene
Dichloromethane
1,1-Dichloroethene
Trichloroethene
Methanol
1,1-Dichloroethane
Cyclohexane
Benzene
1,2-DichIoroethane
Freon 1 13
m-Xylene
p-Xylene
Carbon Tetrachloride
1,3,5 -Trimethy Ibenzene
Acetone
Bromoform
1-Butanol
2-Butanone
Chlorobenzene
Chloroform
cis- 1 ,2-Dichloroethene
Ethyl Benzene
(di)Ethyl Ether
4-Methyl-2-Pentanone
1, 1,2,2-Tetrachloroethane
Tetrachloroethene
1,2,4-Trimethylbenzene
o-Xylene
Interferences/additional compounds
Water
Carbon Dioxide
Methane
Ethane
Propane
Nitrogen Dioxide
Occurrence
Frequency
%a
87
76
59
36
29
17
22
<5
<10
<10
<10
<5
<5
<10
<5
<15
<5
<5
<10
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
to consider and
100
100
<65b
<65b
<65b
b
<20
Maximum
Vented
Drum3
1400
110
72
13
540
670
33
900
2.6
10
900
4.4
4.4
6100
2.7
460
7
80
10
1.9
2.6
160
1.1
1.9
quantitate
6430
ppmv
Unvented
Drumb
74800
7100
3000
1700
1700
104000
40900
273000
<68000
<68000
<68000
264000
aData acquired from -93 previously vented drum and inner layers of confinement.
b
Data from the head space of 210 sealed drums.
767
-------
Table II. FTIR detection limits for VOC's in a 20 cm cell at a total pressure of 640 Torr
Compound
Acetone
Benzene
Bromoform
1-Butanol
2-Butanone
Carbon Tetrachloride
Chlorobenzene
Chloroform
Cyclohexane
cis- 1 2-Dichloroethene
11-Dichloroethene
11-Dichloroethane
12-Dichloroethane
Diethyl Ether
Ethane
Ethylbenzene
Freon-113
Methane
Methanol
Dichloromethane
4-Methy 1 -2 -Pentanone
Propane
1 122-Tetrachloroethane
Tetrachloroethene
Toluene
1 1 1 -Trichloroethane
Trichloroethene
1 24-Trimethylbenzene
135 -Trimethy Ibenzene
m-Xylene
o-Xylene
p-Xylene
Hansta
15.0
2.3
7.5
11.3
0.8
3.0
1.9
1.9
2.3
11.3
7.5
3.8
5.6
15.0
2.3
7.5
5.6
4.9
3.8
2.6
7.5
1.9
5.3
7.5
3.8
7.5
Detection Limits
CLS-lb
5.8
4.2
4.9
25.5
20.2
1.7
12.9
2.1
2.3
5.0
5.4
6.9
10.1
4.5
91.8
47.3
1.5
9.8
6.5
6.0
15.4
16.3
14.0
2.2
12.5
3.4
3.3
24.3
16.8
18.3
9.8
30.3
at 20 cm (ppmv)
CLS-2C
26.2
3.5
6.5
62.2
76.8
2.3
26.1
3.5
3.7
10.1
9.9
20.3
20.7
7.6
122.9
66.7
4.1
13.0
11.2
19.1
44.5
25.7
26.9
4.1
18.2
6.7
11.6
35.2
23.8
27.3
17.6
35.2
PLSd
10.0
3.4
9.2
62.7
4.0
1.6
30.6
7.4
2.7
7.9
8.1
19.0
90.8
9.4
35.4
114.4
5.9
11.8
10.8
27.9
22.4
26.5
51.2
20.6
10.4
10.5
6.2
15.3
15.9
21.2
20.4
23.8
b
'Extrapolated from data presented in reference 6 for a 100 m path length with 10"4 noise.
'Calculated from the errors produced by the application of CLS to 1000 simulated spectra with
7.5x10^ noise and random backgrounds.
'Calculated from the errors produced by the application of CLS to the first derivatives of 1000
simulated spectra with 7.5x10"* noise and random backgrounds.
Determined from the standard deviation resulting from the application of the PLS methods to 20
actual blank spectra.
768
-------
Table III. Analysis results from the repeated determination over a 2 month period of a
reference standard containing 9 VOC's and methane at -100 ppmv. Dry gas was transferred
through -25 feet of unheated tubing.
Concentrations (ppmv)
Compound
Carbon Tetrachloride
11-Dichloroethene
11-Dichloroethane
Freon- 113
Methane
Methanol
Dichloromethane
Toluene
111-Trichloroethane
Trichloroethene
True
96.3
98.9
96.9
95.5
99.5
96.2
97.6
96.7
97.3
100.3
PLS (n=9)
81.6±4.0
85.3 ±3.0
90.9 + 4.6
84.9 + 1.8
93.2 ±2.8
36.2 ±9.0
73.4 ±9.1
70+11
89.3 ±4.4
80.7 ±6.4
CLS (n=16)
85.0 ±3.0
88.4 ±4.8
93.1 ±8.3
80.8 ± 1.8
92.2 ±2.5
38 ± 15
84.6 ±3.4
85 + 12
88. 9 ±3.4
80.1 ±6.7
769
-------
Table IV. Analysis results from the determination of a prepared sample containing 28 VOC's.
Sampled from SUMMA canister which also contained -IQQQQ ppmv water.
Concentrations (ppmv)
Compound
Acetone
Benzene
Bromoform
1-Butanol
2-Butanone
Carbon Tetrachloride
Chlorobenzene
Chloroform
Cyclohexane
cis- 1 2-Dichloroethene
11-Dichloroethene
1 1-Dichloroethane
12-Dichloroethane
Diethyl Ether
Ethane
Ethylbenzene
Freon-113
Methane
Methanol
Dichloromethane
4-Methyl-2-Pentanone
Propane
1 122-Tetrachloroethane
Tetrachloroethene
Toluene
1 1 1-Trichloroethane
Trichloroethene
124-Trimethylbenzene
1 3 5 -Trimethy Ibenzene
m-Xylene
o-Xylene
p-Xylene
GC PLS-Run 1 PLS-Run 2 CLS-Run 1 CLS-Run 2
41.6
35.0
43. Oa
53.9
39.7
31.0
37.0
35.0
30.0
45.0
32.0
39.0
45.0
35.0
34.0
27.0
77.1
50.0
28.2
32.0
32.0
36.0
34.0
40.0
32.0
28.0
36.0
37.0
28.4
42.0
43.0
<52.6
<3.4b
29.3
<25.7
33.7
25.5
44.3
27.8
33.5
110.1
26.1
657.5b
<96.0
26.2
<9.9
72.4
<23.4
< 18.8
397. 6b
72.8
34.0
48.3
28.8
43.1
< 12.8
46.0
40.4
58.8
<20.0
32.6
33.5
44.5
<52.7
<3.4b
29.9
<25.8
36.0
27.4
50.9
29.2
40.4
<76.6
24.3
727.9b
<96.4
27.3
<9.9
74.0
<23.5
<26.6
427. 3b
63.9
34.9
52.6
30.1
49.3
< 12.9
42.4
38.9
54.1
<20.1
39.3
35.5
40.3
22.1
64. T
31.4
55.7
33.8
26.2
36.7
26.3
45.5
39.2
33.2
<77.1
<39.7
26.4
<8.2
72.7
47.2
< 12.9
< 13.7
41.8
32.5
53.4
33.7
40.0
24.2
42.0
45.4
21.8
<25.4
42.7
35.4
37.2
32.0
63. 2C
31.9
50.7
34.9
26.4
36.7
25.6
41.0
37.6
34.7
<77.4
<39.9
26.9
<8.3
74.7
48.0
< 13.0
< 13.7
36.6
32.3
43.1
35.6
39.1
<20.5
34.2
35.4
21.8
<25.5
'Potential problems with the GC analysis may have resulted in a low number.
bHigh spectral residuals were noted.
c2-Butanone is often noted at low concentrations and is either due to an unknown spectral
component or to residues remaining from prior painting activities in the area that may have
contaminated the system.
770
-------
ACKNOWLEDGMENT
This work was supported by the U.S. Department of Energy, Office of Waste Management, under
DOE Idaho Operations Contract DE-AC07-76ID01570.
771
-------
LITERATURE CITED
1. "Conditional No Migration Determination of the Department of Energy Waste Isolation Pilot
Plant", 55 FR 47700, Nov 14, 1990.
2. U.S. Department of Energy, "Quality Assurance Program Plan for the Waste Isolation Pilot
Plant Experimental-Waste Characterization Program," Revision 1, DOE/EM/48063-1, July
15, 1991.
3. D.M. Haaland. R.G. Easterling and D.A. Vopicka, "Multivariate Least-Squares Methods
Applied to the Quantitative Spectral Analysis of Multicomponent Samples," Appl. Spectrosc.,
1985,39,73-84.
4. P Geladi and B.R. Kowalski, "Partial Least-Squares Regression: A Tutorial," Anal. Chim.
Acta, 1986, 185, 1-17.
5. T.L. Clements, Jr. and D.E. Kudera, "TRU Waste Sampling Program: Volume I—Waste
Characterization," Idaho National Engineering Laboratory, Sept 1985, EGG-WM-6503.
6. Hanst and ST. Hanst, "Gas Analysis Manual for Analytical Chemists Vol. I, Infrared
Measurement of Gases and Vapors," Infrared Analysis, Inc., Anaheim, CA, 1990, pp 14-22.
7. A. Savitsky and M.J.E. Golay, "Smoothing and Differentiation of Data by Simplified Least
Squares Procedures," Anal. Chem., 1964, 36, 1627-1639.
8. A.R. Gholson, R.K.M. Jayauty, J.F Storm, "Evaluation of Aluminum Canister for the
Collection and Storage of Air Toxics", Anal. Chem., 1990, 62, 1899-1902.
772
-------
117
HANFORD ENVIRONMENTAL RESTORATION DATA VALIDATION PROCESS
M.R. Adams, Manager, Environmental Restoration Engineering, R. A. Bechtold,
Principal Scientist, Westinghouse Hanford Company, Richland, Washington : K. M.
Angelos, Senior Environmental Scientist, Colder Associates Inc., Redmond,
Washington.
ABSTRACT
Detailed procedures for validation of chemical and radiochemical data are used
to assure consistent application of validation principles and support a uniform database
of quality environmental data. During application of these procedures it was
determined that laboratory data packages were frequently missing certain types of
documentation causing delays in meeting critical milestones in the completion of
validation activities. A quality improvement team was assembled to address the
problems caused by missing documentation and streamline the entire process. The
result was the development of a separate data package verification procedure and
revisions to the data validation procedures. This has resulted in a system whereby
deficient data packages are immediately identified and corrected prior to validation and
revised validation procedures which more closely match the common analytical
reporting practices of laboratory service vendors.
INTRODUCTION
The Westinghouse Hanford Company (WHC) Environmental Restoration
Engineering (ERE) data validation process is applied to laboratory analyses of
environmental samples for Comprehensive Environmental Response, Compensation,
and Liability Act (CERCLA) related cleanup activities. WHC is the U.S. Department
of Energy's (DOE) Operations and Engineering Contractor for the Hanford Site.
Under the Hanford Federal Facility Agreement and Consent Order, the U.S.
Environmental Protection Agency (EPA), the Washington Department of Ecology, and
DOE committed to comply with CERCLA, as well as the Resource Conservation and
Recovery Act, and the State of Washington's Hazardous Waste Management Act. The
ERE function of WHC is conducting CERCLA hazardous substance response
investigations under this agreement which require that low-level and mixed waste
samples (up to 100 mrem/h) data be received from laboratories within 75 days of
collection (as annual average but not to exceed 90 days), validated within 21 days after
receipt, and forwarded to the regulatory agencies within an additional 15 days. The
tasks of managing sample collection and shipment to the laboratories, obtaining
acceptable laboratory deliverables, and providing validated analytical results for
decision making within prescribed time frames have been complex and challenging. To
support informed decision making and proper expenditure of funds for environmental
773
-------
cleanup, it is essential that data of known quantity and quality are obtained on a timely
basis. Delays in any step can have a cascading effect on ERE programs, resulting in
delays in meeting regulatory-imposed milestones for site restoration.
DATA VERIFICATION AND DATA PACKAGE COMPLETENESS
Traditionally, data package completeness checks and the verification of proper
laboratory reporting have been conducted by data validation personnel. During prior
application of data package completeness checks, the frequency of occurrence of
missing pieces of documentation was holding up the completion of validation activities.
To understand the reasons behind the missing data package items, a Pareto analysis was
conducted of over 50 data package submittals that contained missing documentation (as
identified by the validation procedures). Table I shows the results of a survey of
volatile organic data packages. As shown in Table I, the highest number of incidences
of missing documentation occurred with items that had been determined to be not
essential for completion of key validation processes. Requests for these missing
documentation items caused unnecessary delays in the completion of validation reports.
A quality improvement team was established with key members from ERE
management, data managers at Hanford, data validators, and the analytical laboratories'
personnel to address the problems caused by missing documentation and to streamline
the entire process. As a result, a separate data package verification procedure was
developed and implemented with the consensus of laboratories' personnel, data
managers, and validation personnel on the technical content of data packages and the
team provided a recommended approach. A diagram of the current verification-
validation process being followed for environmental samples at the Hanford Site is
shown in Fig. 1. Application of this new process has resulted in significant
improvements in the delivery of complete data packages. The data verification process
is finding significantly fewer deliverable omissions (a reduction of approximately 50%)
resulting in an overall improvement in the timely completion of data validation
activities.
Technical verification is the act of reviewing, inspecting, and checking
analytical laboratory data packages by a page-by-page review to determine and
document whether their contents conform to the specified package constituents required
for data validation.
Personnel currently performing verification are required to be trained with a 40-
hour technical training course. This course covers the identification of specific
deliverables from the individual laboratories as well as the use of the verification forms
and the processes for obtaining any missing information from that laboratory.
774
-------
Individualized chemistry checklists were prepared for metals, semivolatiles,
volatiles, pesticides/PCB, herbicides, and general/wet chemistry. Development of a
"standardized" checklist for general chemistry parameters (Fig. 2) proved to be the
most challenging due to the number and types of analytical methods used by the
laboratories. The resulting solution was a subdivision by types of analytical techniques.
The categories used currently are ion chromatography, colorimetric, gravimetric, ion
selective electrode, titrimetric, infrared spectroscopy, and other. Checklists for
radiochemistry were also prepared by subdivision according to the analytical technique.
Individualized checklists were prepared for gas proportional counting, alpha
spectroscopy, gamma spectroscopy, alpha-emitting isotopes by scintillation counting,
radium-226 by radon emanation, liquid scintillation counting, uranium by fluorometry,
uranium by kinetic phosphorimetry, and radioisotopes by inductively coupled
plasma/mass spectroscopy (ICP/MS).
During technical verification, the verifier completes checklists for every type of
analyte with each data package. Upon completion, the verifier prepares a daily
verification summary and missing information report (Fig. 3) for both chemical and
radiochemical data packages. This information is then compiled to track performance
of the improved verification process.
DATA VALIDATION PROCESS
Data validation is a formal process of reviewing a body of analytical data
against predefined criteria to assure the data are applicable for their intended use. The
process of data validation consists of the following steps:
• Editing and correcting of reported results.
• Verifying compliance with quality assurance (QA) requirements.
• Checking quality control (QC) values against defined limits.
• Applying qualifiers to analytical results for the purpose of defining the
limitations in the use of the reviewed data.
WHC ERE has developed detailed validation procedures for radiochemical (1)
and chemical (2) analytical data validation, which are in use by all personnel validating
data produced as a result of Hanford site CERCLA response and cleanup activities.
These procedures have been developed based on EPA functional guidelines and
standard laboratory reference methods. The chemical data validation procedures
generally follow the EPA Contract Laboratory Program (CLP) protocols (3,4) and SW-
846 (5) methodologies.
775
-------
For the Hanford Environmental Cleanup Program, data validation is conducted
using the WHC validation procedures in conjunction with the applicable project-specific
work plans, QA project plans, analytical method references, and contractor laboratory
statements of work. The final products consist of narrative reports, checklists,
summary reports, and electronic data deliverables. A flow diagram of the data
validation process used at the Hanford Site is given in Fig. 4.
As described under Data Verification and Data Package Completeness, in
addition to the improvement of data package completeness, a new and streamlined
radiochemical data validation procedure was prepared to improve the consistent
application of validation principles to nonstandard radioanalytical data. The
radiochemical procedures follow the EPA reference analytical methods and procedures
developed at other DOE contractor operated sites (6). The procedure addresses the
common requirements from radiochemical analyses, as well as those that are specific
for validation of data from the radioanalytical methods for gross alpha/beta, strontium-
90, alpha spectrometry, gamma spectroscopy, tritium, radium-226 by radon emanation,
fluorometric uranium, phosphorimetric uranium and ICP/MS.
The key aspects of this procedure are summarized in Table II. For consistency,
general validation control limits were established for key data quality indicators such
that variations in analytical procedure between laboratories would not result in rejection
of data sets as long as key QA requirements have been met. This procedure also
addresses requirements for the transmittal of electronic data in a format that is
compatible with the Hanford Environmental Information System (HEIS) database.
DATA VALIDATION DOCUMENTATION AND REPORTS
To provide a useful presentation of validation results without requiring the
completion of complex forms and checklists and the preparation of extensive technical
reports, a simple validation documentation and reporting format was developed.
Simple checklists are filled out at the completion of validation of a single data package
which may contain multiple analyses. These checklists are general enough to allow the
combining of several types of analyses into one checklist for documentation purposes.
Fig. 5 provides an example cover page for radiochemical analyses.
A simple reporting procedure is required in the latest revisions to the validation
procedures. The reporting format requires the summary by the validator of the
following key elements in a simple technical memorandum format:
• Introduction - identifying data package tracking number, samples,
matrices and analyses validated and the laboratory performing the
analysis.
776
-------
• Summary of Data Quality - summarizing the results of precision,
accuracy, result verification, detection limit compliance and
completeness in terms of the percentage of valid measurements versus
expected measurements
• Summary of Major Deficiencies an itemized listing of the major
QA/QC deficiencies identified during validation, which resulted in the
rejection of sample data.
• Summary of Minor Deficiencies - an itemized listing of the minor
QA/QC deficiencies identified during validation which resulted in
qualification of sample data, which, though qualified may still be
considered usable for decision making purposes.
ELECTRONIC DATA REPORTING
The HEIS database is accessible to various parties involved with environmental
restoration at the Hanford Site, including state and federal (EPA) regulators. Recent
improvements in the availability of electronic data summary software in use by
laboratories and subsequent upgrades of the HEIS utility programs have enabled the
easy transmittal of validated results and update of result qualifier flags within the HEIS
database. A standardized format for reporting validated data in electronic format has
been developed. This format known as the UPQUAL format enables the transmittal of
changed or qualified data only into the HEIS database resulting in faster updating of
results with fewer errors or omissions of necessary data. Table III provides the
structure of the UPQUAL format. Validation personnel can be provided with the
laboratory data in electronic format and can select only those results which require
correction and qualification for inclusion in the UPQUAL file, which is then
transmitted to the HEIS data managers for loading and verification. Future
improvements to the system will include the automation of certain aspects of technical
verification and data validation to further improve the timeliness, completeness and
consistency of data validation activities.
SUMMARY
Providing quality, timely, and valid environmental data for informed decision
making is essential for the successful completion of environmental restoration efforts at
Hanford. This paper presents the iterative processes involved in analytical data
package deliverable verification, the refinement of validation procedure requirements,
and the development of detailed and standard yet consistent, streamlined approaches to
the technical performance, documentation and reporting of these processes. As a
777
-------
result, a significant improvement in the completion schedule for delivery of validated
analytical data to site managers and regulatory authorities has been achieved.
REFERENCES
1. WHC, 1993, Data Validation Procedures for Radiochemical Analyses, WHC-SD-
EN-SPP-001, Rev. 1, Westinghouse Hanford Company, Richland, Washington.
2. WHC, 1993, Data Validation Procedures for Chemical Analyses, WHC-SD-EN-
SPP-002, Rev. 2, Westinghouse Hanford Company, Richland, Washington.
3. EPA 1990, USEPA Contract Laboratory Program Statement of Work for Organics
Analysis, Multi-Media, Multi-Concentration, Document Number OLM01.8, U.S.
Environmental Protection Agency, Washington, D.C.
4. EPA 1990, USEPA Contract Laboratory Program Statement of Work for Inorganics
Analysis, Multi-Media, Multi-Concentration, Document Number ILM02.0, U.S.
Environmental Protection Agency, Washington, D.C.
5. EPA 1992, Test Methods for Evaluating Solid Waste, Physical/Chemical Methods,
SW-846, 3rd Edition, Final Update I, U.S. Environmental Protection Agency,
Washington, D.C.
6. EG&G 1991, General Radiochemistry and Routine Analytical Services Protocol
(GRRASP), Version 2.1, July 1991, EG&G Rocky Flats, Environmental
Management Department, Rocky Flats Plant, Golden, Colorado.
778
-------
Figure 1. Data Validation Process Flow Diagram
Analytical
Data Package
Received by
WHC
)
r
UHC Records
Management
Activities
Performed
>
r
Technical
Verification of
Carpi eteness
Performed
JL
/-^Data^V
/ Package ^
\Conpl etc?/
Request for
Hissing
Documentation to
Laboratory
r~^ — -^
'QA/OC
Information ,-
plete?/
Final Data
Validation Sunnary
Prepared from
Multiple Data
Validation Packages
Narrative Surmary of
Validation Results
Evaluation of DQOs
Qualified (Annotated)
Laboratory Reports
Tabulated Data Summary
Electronic Deliverable
Data Validation
Package and
Original Data
Package Returned
to UHC
779
-------
Figure 2. General Chemistry Data Verification Checklist.
Package ID:.
GENERAL CHEMISTRY DATA VERIFICATION CHECKLIST - FORM A-7
Review the data package for completeness and check off the items below. If any data review
elements are missing, contact the laboratory for submittal of the omitted data.
Data Package item Present? Yes No N/A
[_] Anions by Ion Chromatography (Method 300.0)
Sample Results
Initial Calibration Data
Continuing Calibration Verification
Calibration Standard Concentrations
Blank Analysis Data or Summary Report Forms
Duplicate Sample Analysis Report Forms
Spike Sample Recovery Data
Laboratory Sample Control Data
Raw Data
Analytical Sequence
Ion Chromatograms
Quantitation Report
Additional Data
Moisture/% Solids Data Sheets
Sample Preparation Sheets (Soils, Other only)
D
Colormetric (Note: Identify by Name, Analyte and EPA Method).
Sample Results
Initial Calibration Data
Continuing Calibration Verification
Calibration Standard Concentrations
Blank Analysis Data or Summary Report Forms
Duplicate Sample Analysis Report Forms
Spike Sample Recovery Data
Laboratory Sample Control Data
Raw Data
Analytical Sequence
Laboratory Bench Sheets
Chart Recorder Printouts
Additional Data
Moisture/% Solids Data Sheets
Sample Preparation Sheets
GENNM092S93-C
730
-------
Figure 3. Daily Verification Summary and Missing Information Report.
Cf)
Daily Verification Summary and
Missing Information Report: Radchem
(DVS-M1R-R)
1 Verification BOA
Verifier
Date
2 Project(s)
Cognizant Engineer
Data packages verified
3 this date (list by number
or other identifier)
Gas
Court
n
i — i
n
4- n
n
5 •
Totsl .. .i i
Form-DVS-1
Revision 0, •
VIIR-R
1/18/93
or'nm
Analytical
packages
n n int
u
O
r— i r— i rn n r~ i 1
n c
4 Data packages and analytical groups with
Total
types contained in data
(shade -in)
ma-s
:
n
:
n
:
n
scInT Ra'22S LSC FIU°r' Liuer ICP/MS
n
i — i ._
n c
n =
c
n c
U U L
* * * * i__
" u n
~ n n
] n n c
i i i i i
j (, i ,,i
] n
_i i i i i i—
J U
I I
" n
] n
[ |
.
J U
missing information, (circle above)
5 % Data packages and analytical groups with missing info
%
6 Confirmation: Every data package item shaded
Verifiers initials
7 Confirmation: Every checkl
taxed to lab for 24 hour retur
Verifiers initials
in item 3
/•
rmation; Total 4 x
V. Total
*\
3°°J
has checklist attached.
st for data package items circled above has been
n of missing information to verifier.
8
9
WHC Distribution,
Verifier Signature
Notes:
Do not file daily report on days you do not complete
any verification.
MRA/M012193-A
781
-------
Figure 4. Data Validation Flow Diagram
Data Package*
Received by
Validaton
Log in Dab
Package* and
Verify R«uhi
Calibration*
Okay?
Blank*
Okay?
Accuracy
Okay?
Precision
Okay?
Holding Times I
Okay? I
Analyiil
Specific
Criteria Okay?
ReiutU/MDA
Transcription*
Calculations
Verified
Data Validation Package and
Analytical Data Package
Returned to WHC
782
-------
Figure 5. Radiochemical Data Validation Checklist.
Radiochemical Data Validation Checklist
(cover page)
VALIDATION LEVEL: ABODE
PROJECT:
VALIDATOR;
CASE:
DATA PACKAGE:
LAB:
DATE:
SDG:
ANALYSES PERFORMED
D Gross Alpha/Beta
O Gamma Spectroscopy
D
SAMPLES/MATRIX
D Strontlum-90
D Total Uranium
D
D Technetlum-99
D Radlum-226
a
Q Alpha Spectroscopy
LI Tritium
D
1. Completeness
Technical verification forms present?
Comments:
Q N/A
Yes No N/A
2. Initial Calibration
D
Instruments/detectors calibrated within
one year of sample analysis?
Initial calibration acceptable?
Standards NIST traceable?
Standards expired?
Comments:
Yes No N/A
Yes No N/A
Yes NO N;A
Yes No N/A
GEN\M0925»-B
783
-------
Table I. Summary of Missing Volatile Organic Data Package Items*
Data Package Item
Chemist Logbook Pages
Reduction Formulae
Instrument Time Logs
Sample Preparation Sheets
MS/MSD Report
MS/MSD Forms
Data Summary
RIC and Quantitation Reports for
MS/MSD
Calculation data for TIC
Moisture/Percent Solids Sheets
Internal Chain-of-Custody Forms
Other Data
Number of Times
Found Missing
48
47
45
15
13
9
9
6
5
5
2
1
Comments
Not essential
Not essential
Not essential
Required Item
Required Item
Required Item
Required Item
Required Item
Required Item
Required Item
(matrix-dependent)
Not essential
* A total of 50 data packages were evaluated.
MS/MSD = matrix spike/matrix spike duplicate.
RIC = reconstructed ion chromatogram.
TIC = tentatively identified compounds.
784
-------
Table II. Summary of Radiochemical Data Validation Requirements.
Data Validation Item
Initial demonstration of
instrument calibration
Calibration Checks
Background Checks
Method Blanks
Laboratory Control Standards
(LCS) or Blank Spikes
Chemical Recovery or Tracers
or Matrix Spikes
Laboratory Duplicates
Holding Times
Sample Result Calculations
MDA
Required Frequency of
Performance by Laboratory
Once for each project and at least
annually. Documentation required on a
periodic basis or whenever calibration is
updated on an instrument specific basis.
Each sample batch or weekly (for long
count analyses)
Each sample batch or weekly (for long
count analyses)
One in 20 samples or each sample batch,
whichever is more frequent
One in 20 samples or each sample batch,
whichever is more frequent
Each sample, blank, LCS, blank spike,
matrix spike and duplicate as applicable to
the method of analysis
One in 20 samples or each sample batch,
whichever is more frequent
Not applicable
Not applicable, calculation formulae must
be provided at least one time during
project startup and when revision is made
to analytical procedure
MDA must be reported for each
sample and analyte
Validation Criteria
Calibration must meet minimum
performance criteria that are method
specific. If criteria are not met or
documentation is unavailable, all
associated data must be rejected.
The most recent calibration check must be
within the laboratory-generated control
limits or ±3 standard deviations from the
mean check value.
The most recent background check must
be within the laboratory-generated control
limiU or ±3 standard deviations from the
mean background value.
The method blank must contain less than
the minimum detectable activities (MDA)
of any target analyte. Sample results that
are less than 1 Ox the blank results must be
qualified as undetected.
LCS or blank spike results must be within
the control limits of 50% to 150%. If
exceeded, associated results must be
rejected.
Recovery of chemical carrier or
radioactive tracer must be within the
control limits of 20 to 100% (115% for
tracers). If exceeded, associated
undetected results are rejected.
Relative percent difference between
duplicate results or MDA values (if analyte
undetected) must be 520% for water
samples and £35% for solid samples.
All analytes with the exception of short
half-life parameters must be analyzed
within 180 days of collection. Short half-
life parameters must be analyzed within
five half-lives.
If calculation is required based on the
specified level of data validation.
Sample aliquots must be at least large
enough to meet the required detection
limit or the MDA concentration must
be less than or equal to the required
detection limit.
785
-------
Table m. UPQUAL Input File Format
Field Name
Sample Number
Form Type
Form Number
Lab Code
Constituent ID
Media
Value Reported
Qualifier
Counting Error
Analysis Units
Column
1-12
13-16
17-19
20-25
26-35
36-38
39-48
49-54
55-64
65-72
Valid Values
HEIS Sample Number
CLP - Contract Laboratory
Program
LAS - Laboratory Analysis
System
NCLP- Non-CLP
1A, IB, 1C, ID, IE, IF,
I, R, NA
Abbreviated Lab Code
(assigned by WHC)
Element Symbol,
CAS Number,
Element Symbol and
Isotope Number
BI = biota
GW = groundwater
SS = surface soil
GS = geologic soil
SW = surface water
constituent concentration
validated result qualifier
radionuclide result
counting error
reporting units for
constituent (i.e. jig/L,
/zg/Kg, mg/L, mg/Kg, %,
ppm, ppb, pimhos/cm3,
pCi/L, pCi/g)
786
-------
GENERAL
-------
118
CORROSITEX®, a New Solution for Determining Corrosivity of Products and Waste.
Virginia C. Gordon, Soheila Mirhashemi, and Rosalind Wei. InVitro International, 16632 Millikan
Avenue, Irvine, CA 92714
ABSTRACT
A new in vitro method, CORROSITEX®, has been developed to determine the corrosive potential of chemicals,
formulations and waste. This in vitro method assigns corrosive materials into United Nation(UN)Packing
Groups I, n or IE which is required by the U.S. Department of Transportation (DOT) as of October 1993.
Corrosivity has been defined in a variety of manners. The DOT, previous to issuing an administrative exemption
for use of CORROSITEX (DOT E - 10904), defined Corrosivity by the time it took for a substance cause tissue
destruction on the backs of six albino rabbits (49 CFR 173 Appendix A)1. Consistent with UN Guidelines,
Packing Groups were determined by the following time criteria:
Packing Group I: Skin tissue destroyed after a 3 minute exposure
Packing Group II: Skin tissue destroyed after a 60 minute exposure
Packing Group ffl: Skin tissue destroyed after a 240 minute exposure
Noncorrosive: No skin tissue destruction within 240 minutes
The Office of Solid Wastes, a part of the Environmental Protection Agency (EPA), currently defines Corrosivity
as a substance which has a pH of below 2.0 or above 12.50 (40 CFR §261.22)2. Substances with a pH above 2.0
or below 12.50 are considered to be noncorrosive under these guidelines.
Several limitations are inherent in both the DOT and EPA definitions. Specifically, in vivo rabbit tests show a
significant variability from test to test Reproducibility is a challenge and is caused by such factors as variable
hair growth, age of the animal and lab-to-lab technique. The use of pH, on the other hand, is limited to aqueous
solutions and cannot be used for solids. In addition, a study of in vivo Corrosivity results and corresponding pH
demonstrates very limited correlation in determining corrosive versus noncorrosive substances.
CORROSITEX, as an alternative to either methodology, is a precise test that determines not only corrosivity, but
assigns UN Packing Groups as well. This in vitro assay consists of two compartments - a Dermal Biobarrier and
a Chemical Detection System (CDS). Test samples are placed on the Dermal Biobarrier. When the sample
destroys or penetrates the Biobarrier, the sample is detected by the CDS which produces a simple color or other
physical change. The time required to observe a color change is recorded and used to assign the Packing Group.
In a study of 85 commercially-available chemicals listed on the DOT Hazardous Materials Table (49 CFR
172.101), CORROSITEX assigned 80 to the same or safer Packing Group designated by the table. Additionally,
inter-laboratory studies were also conducted using a number of these chemicals and resulted in a lab-to-lab
reproducibility of about 95%. More than 4,000 substances have now been tested by CORROSITEX throughout a
variety of industries including hazardous waste. These industries have found the new methodology to be
accurate, easy-to-use and cost effective for determining corrosivity and the degree of Corrosivity for both liquid
(including non-aqueous) and solid test materials.
CORROSITEX is accepted by both the United States Department of Transportation and the Occupational Health
and Safety Administration, as well as Canada, Germany and Switzerland, for defining biological corrosivity.
787
-------
INTRODUCTION
In Vivo Corrosivity Testing
The Department of Transportation (DOT) uses an in vivo rabbit dermal test to evaluate the dermal
corrosivity of substances. Corrosion is considered to have occurred if the substance in contact with the
rabbit skin causes destruction or irreversible alteration of the tissue on at least two of six rabbits tested1.
Tissue destruction is defined, at any of the readings, as ulceration or necrosis of the skin. Tissue
destruction does not include merely sloughing of the epidermis, erythema, edema, or fissuring.
Assignment of Packing Groups for corrosive chemicals and formulations is based on the United Nations
printed special recommendations for Class 8 or corrosive chemicals3. The distinctions among chemicals
in Packing Groups I, II and III are as follows:
Packing Group I: Skin tissue destroyed after a 3 minute exposure
Packing Group n: Skin tissue destroyed after a 60 minute exposure
Packing Group HI: Skin tissue destroyed after a 240 minute exposure
Noncorrosive: No skin tissue destruction within 240 minutes
This in vivo test method has inherent limitations which are reflected in the test protocol of requiring only
two out of six rabbits to be positive in order to assign corrosivity. It is difficult to achieve reproducible
results with this test. Such factors as variable hair growth, age of the animal, the time of year the test is
run, and variations in lab-to-lab techniques, all add to the variability in test results. In addition tissue
ulceration and necrosis are subjective measures.
Utilization of pH
EPA regulations (40 C.F.R. §261.22) define a substance to be corrosive if it" is aqueous and has a pH less
than or equal to 2 or greater than or equal to 12.50". One of the most important limitations of this
definition is that there are many substances that have a pH greater than 2 or less than 12.5 that have been
proven to be corrosive by in vivo testing. Figure 1 gives the pH value and the Packing Group assignment
for 147 substances, as determined either by in vivo testing or as assigned by the Hazardous Materials
Table found in 49 C.F.R. §172.10. Sixty-nine percent of the substances whose pH was greater than 2 or
less than 12.5 were found to be corrosive. These would have been identified as noncorrosive using the
EPA definition of corrosivity thereby creating a substantial risk to transporters, workers and the
environment due to their misclassification. In addition, ten percent of the substances that did fall into a
corrosive category, according to the EPA definition, were shown by in vivo experiments to be
noncorrosive. This would mean that these substances would be treated as corrosives, even though they are
not, thereby incurring additional paperwork and expenses that would not be required under a more
accurate classification method..
Not only do pH values not accurately predict whether a substance is corrosive or not, it is impossible to
assign a Packing Group (PG) using a pH value. This would mean that severe, extremely harmful,
corrosives would not be differentiated from mild, less harmful, substances. For fifty-two substances that
had a pH of less than or equal to 2.0, four were assigned to PG I, thirty-five to PG II, eight to PG III, and
five were found to be noncorrosive. The same was also true for the thirty substances whose pH was great
than or equal to pH 12.5; two were assigned to PG I, fifteen to PG II, three to PG III and ten were found
to be noncorrosive. Narrowing a pH range down does not help in the assignment of Packing Groups. In
the pH range of 0 to 0.9, for each sample, substances were assigned to all three Packing Groups and two
were even found to be noncorrosive. Clearly, pH cannot be used to distinguish between severely harmful,
and less harmful substances.
788
-------
Figure 1. Relationship Between pH and Packing Groups
NC
PG
PGM
PGI 6
on n conn en o
a n 11 in 11 on
m
H
H 1 1 1
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
789
-------
Table 1. Relationship between pH and Packing Group Assignment
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Chemical
Fluorosulfonic acid
Nitric acid
Selenic acid
Trifluoroacetic acid
Benzyl chloroformate
Sulfur monochloride
COR 003
COR 001
Bromoacetyl bromide
Dichlorophenylphosphine
Phenyl trichlorosilane
Phosphorous pentachloride
Phosphorous tribromide
Sulfuric acid
Fumaryl chloride
Octyltrichlorosilane
Antimony trichloride
Hydrogen bromide (hydrobromic acid)
Mercaptoacetic acid
Octadecyltrichlorosilane
Antimony tribromide
Boron trifluoride-dihydrate
Valeryl chloride
Dichloroacetyl chloride
Dodecyl trichlorosilane
Hydrobenzene sulfonic acid
Dichloroacetic acid
Anisoyl chloride, ortho-
COR014
Trichloroacetic acid
Iodine monochloride
Ammonium hydrogen sulfate
Potassium hydrogen sulfate
Phenylacetyl chloride
Boron trifluoride-acetic acid complex
Aluminum bromide, anhydrous
Fluoboric acid
Bromoacetic acid
Chloroacetic acid
Formic Acid
Sulfurous acid
Acetic anhydride
Acetyl bromide
Ferrous chloride tetrahydrate
Acrylic acid
Acetic acid, glacial
Dimethylcarbamyl chloride
Aluminum chloride
Trichlorotoluene
Chromium (III) fluoride
Sodium hydrogen fluoride
Ammonium hydrogen difluoride
Thiophosphoryl chloride
pH o 10 %
Solution
0.00
0.00
0.00
0.75
2.54
5.20
13.54
13.56
0.00
0.00
0.00
0.00
0.00
0.00
0.05
0.10
0.30
0.30
0.30
0.30
0.35
0.41
0.45
0.46
0.50
0.55
0.64
0.72
0.72
0.74
0.75
0.78
0.85
0.92
0.95
.22
.30
.41
.44
.55
.78
.99
2.00
2.05
2.07
2.30
2.31
2.92
3.32
3.90
5.16
5.22
5.81
Pkg Group
DOT/In vivo
I
I
I
I
I
I
I
I
II
II
II
II
II
II
ii
n
n
n
II
II
n
n
n
n
ii
ii
II
ii
II
II
II
ii
n
n
II
II
II
II
II
II
II
II
n
ii
II
ii
n
n
II
II
II
II
ii
790
-------
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
Dimethylbenzylamine, N,N-
COR008
COR 036
COR 010
Dimethylcyclohexylamine 2,3-
Lithium hydroxide, monohydrate
Triethylenetetramine
Diethylenetriamine
Ethylenediamine
COR 037
Cyclohexylamine
COR 034
COR 033
COR 019
COR Oil
COR 009
COR 007
COR 016
COR 002
COR 005
Tetramethylammonium hydroxide pentahydrate
COR 015
Sodium hydroxide, solid
Sodium hydroxide
COR 017
Potassium hydroxide
Cleaner 1
Sulfamic acid
Sodium hydrogen sulfate
Phosphoric acid
Maleic anhydride
Maleic acid
Cyanuric chloride
Benzene sulfonyl chloride
Butyric acid
Crotonic acid
Propionic acid
Copper (II) chloride
Ferric Chloride
Hexanoic acid
Butyric anhydride
Hydroxylamine sulfate
Dicyclohexylamine
Tributylamine
Aminoethoxy) ethanol, 2-{2-
Sodium hypochlorite w/ 5% chlorine
Piperazine, 1(2-AE)
Ethanolamine
Tetraethylenepentamine
Ethylhexylamine, 2-
Diaminopropane, 1,2-
Diethylaminopropylamine, 3-
COR031
Cleaner 2
COR 020
COR 013
Oxalic acid, 10% (in 50% EtOH)
Dichloroacetic acid, 3.1%
Maleic acid, 3.8%
Formic acid, 7%
Citric acid, 14.6%
COR 039
10.70
10.81
11.00
11.17
11.79
11.80
11.91
12.01
12.13
12.20
12.34
12.50
12.60
12.80
12.85
12.99
13.19
13.29
13.34
13.35
13.61
13.64
13.81
13.81
13.85
14.00
0.51
0.65
0.75
0.85
1.05
1.30
1.72
1.80
2.15
2.30
2.68
2.99
3.00
3.00
3.08
3.58
9.57
10.70
11.30
11.65
11.78
11.82
11.85
11.98
12.06
12.17
12.40
12.83
12.90
13.34
0.81
0.98
1.25
1.70
1.73
2.80
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
III
NC
NC
NC
NC
NC
NC
791
-------
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
Adamquat BZ 80
COR 012
J0087
J0074
J0077
J0537
Bowl cleaner, blue
Bowl cleaner, green
Diethylamine-3-propionitrile
Heptanoic acid
Tnethanolamine, 100%
Laundry detergent
J0097
Window Cleaner 1
Window Cleaner 2
Dishwashing soap
COR 018
Butylamine, 7% (in EtOH/EG, 1:1)
Diethylamine, 6%
Cleaner3
COR 006
Cleaner 4
COR 032
Pyrrolidine, 5.30%
Calcium carbonate
Cleaner 5
Cleaner 6
Cleaner?
Sodium hydroxide, 0.36%
Cleaner 8
COR 004
Tnethanolamine, 80%
5.06
7.25
7.89
7.96
8.17
8.65
9.07
9.16
9.40
9.80
10.07
10.48
10.51
10.76
10.79
10.87
11.16
11.80
12.01
12.18
12.23
12.26
12.50
12.53
12.56
12.64
12.67
12.70
12.74
12.97
13.01
13.35
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
792
-------
In Vitro Testing
Compared to in vivo methods, in vitro methods are quantitative and reproducible, overcoming two major
scientific limitations of the in vivo methods. In vitro methods can screen multiple samples and
concentrations quantitatively making them useful in all stages of the manufacturing and formulating
process. These tools evaluate the safety of a product from its raw materials, through production and final
safety evaluation. Therefore, in vitro methods are now the foundation of broader safety evaluation
programs in many major industries. Today, some companies actually perform many more tests utilizing
in vitro methods than prior to 1988 when they used in vivo methods.
CORROSITEX is a new, reliable, reproducible and cost-effective in vitro test method that accurately
determines the corrosive potential of chemical formulations and waste.
MATERIALS AND METHODS
The CORROSITEX system consists of a pre-qualification test, a screening test, and a corrosivity
classification assay. The steps used to develop a tiered CORROSITEX in vitro approach are straight
forward (see Figure 3).
Figure 2. The CORROSITEX Test Method.
1. Evaluate test sample for compatibility with a prequalification test.
2. If the test sample is nonqualified in this in vitro method, test as described in 49 CFR
173.136.
3. If test sample is qualified, perform the Screening Test and assign Categories for test
samples.
4. Evaluate test sample with CORROSITEX and assign Packing Group I, II, or III or
Noncorrosive, based on test sample screening and time in CORROSITEX test.
Prequalification Test
Prior to performing the CORROSITEX test, the sample is pre-qualified to assure that the substance being
tested is compatible with CORROSITEX. The sample is placed directly in a small amount of CDS fluid.
If any detectable change occurs in the CDS, the sample is qualified and can be run in CORROSITEX.
Screening Test
After running in excess of 4,000 mixtures, pure chemicals and waste in CORROSITEX, it has been
determined that behavioral differences exist between substances that are strong acids or bases and those
that are weak acids or bases. Therefore, a screening test has been developed to distinguish between these
categories.
Using this test, samples are classified into categories. Packing Groups are assigned according to category
and the results from the CORROSITEX assay. Test samples are classified according to changes produced
in two well-defined classification tests ~ one designed to test for acids and another designed to test bases.
The four different categories are defined as follows: (see Figure 3).
793
-------
Figure 3. Categorization of Test Samples in Screening Test
• Category AI substances produce a distinct color change when they are added to
the acid test.
• Category BI substances produce a distinct color change when they are added to
the base test.
• Category A2 substances produce little or no color changes when added to the acid
test
• Category B2 substances produce little or no color changes when added to the base
test.
CORROSITEX Assay
CORROSITEX consists of two compartments — a Dermal Biobarrier and a Chemical Detection System
(CDS). The dermal biobarrier has been developed using relevant target macromolecules. Test chemicals
and formulations, including solids and liquids, are applied directly to the dermal biobarrier. (See Figure
4.) When the chemical destroys the full thickness of this biobarrier, it is detected by the CDS which
produces a simple color change. This color change is visually observed and the time required for the color
change to occur is recorded in order to assign a Packing Group. If no color change occurs, the chemical
may be noncorrosive.
Figure 4. Biobarrier and Chemical Detection System of CORROSITEX
Movement of test sample as
it destroys biobarrier
Test Sample
Release of
Chemical
Change in CDS upon
release of chemical
Biobarrier
Chemical Detection
System (CDS)
794
-------
1. Dermal Biobarrier
The Dermal Biobarrier has been developed using specifically formulated solubilized proteins. This
biobarrier is prepared by coating a support with a mixture of diluent and solubilized protein. The
macromolecules are gelled onto a cellulose support within a circular disc delivery system. The biobarrier
is then and stored at 4°C. The shelf life of the biobarrier is two weeks under these storage conditions.
Prior to mixing with the diluent, the shelf life of the materials is two years. Mixing may take place in a
laboratory setting or in the field.
2. Chemical Detection System
The Chemical Detection System (CDS) consists of multiple chemical detectors including specific ions,
indicators and other detectors which have demonstrated responsiveness with numerous classes of
chemicals. Examples of some of the Chemical and Product Classes identified by the CDS are given in
Table 2. A simple, visually detectable color change of the CDS occurs when the biobarrier is altered or
destroyed due to the chemical exposure. If no color change occurs, the sample is not compatible with the
COPvROSITEX assay.
Table 2. Chemical and Product Classes Identified by CORROSITEX Chemical Detection System.
primary alkyl amines
secondary alkyl amines
tertiary alkylamines
alkylamine salts
ethoxylates
n-oxides
organic chemicals
oxidizing agents
reducing agents
metal salts
halides
acids
bases
acid salts
acid esters
anhydrides
halogen derivatives
silanes
organic solvents
formulated amines
acrylates
sterols
quaternary salts
polymers
biocides
insect powders
pesticide classes
fertilizer classes
propellants
metal working fluids
organophosphates
fuel additives
heavy duty cleaners
window cleaners
strippers
light duty cleaners
powdered cleaners
anionic surfactants
cationic surfactants
nonionic surfactants
amphoteric surfactants
surfactant blends
mercaptans
hydrocarbons
795
-------
3. Packing Group Assignments.
In the COKROSITEX system Packing Group assignments are made by taking into account the category
that is assigned to a sample by the Screening test, and by the time it takes to detect a color change in the
CDS.
Packing Group I, II or III are assigned to test samples from Category AI and BI which produce a
detectable color change in the CDS between zero and three minutes, greater than 3 minutes to 60 minutes,
or greater than 60 minutes to 240 minutes, respectively. If no color change occurs in 240 minutes, the
chemical is classified as Noncorrosive (see Table 3).
Table 3. CORROSITEX Assignment of Corrosive Packing Groups for Category AI and B! Samples
Time (h:mm:ss)
0:00:00-0:03:00
X):03:00 - 1:00:00
>1:00:00- 4:00:00
>4:00:00
Abbreviation
I
II
III
NC
Classification
Corrosive Packing Group I
Corrosive Packing Group II
Corrosive Packing Group III
Noncorrosive
Packing Group I, II, or III is assigned to test samples from A2 and B2 which produce a detectable color
change in the CDS between zero and three minutes, greater than 3 minutes to 30 minutes, or greater than
30 minutes to 45 minutes, respectively. If no color change occurs in 45 minutes, the chemical is classified
as noncorrosive (see Table 4).
Table 4. CORROSITEX Assignment of Corrosive Packing Groups for Class A2 and B2 Samples
Time (h:mm:ss)
0:00:00-0:03:00
>0:03 :00 -0:30:00
>0:30:00 - 0:45:00
>0:45:00
Abbreviation
I
II
III
NC
Classification
Corrosive Packing Group I
Corrosive Packing Group II
Corrosive Packing Group III
Noncorrosive
796
-------
RESULTS
In order to assure that CORROSITEX would give the same Packing Groups assignment to pure chemicals
as the DOT Table 172.101, all commercially available corrosive chemicals listed on that table were tested
in CORROSITEX. A summary of the results is presented in Table 5 and the details are given in Table 6.
Corrosive classifications were assigned for the chemicals based on Tables 3 and 4. In vivo Packing
Groups were obtained from industrial laboratories or the DOT Hazardous Materials Table 172.101 of CFR
49. When actual in vivo data was available, it was the preferential data used for comparison. Using the
pre-screen test, CORROSITEX correctly distinguished 100% of the chemicals to be corrosive or
noncorrosive and also assigned correct Packing Groups with 84% concordance, having 9 overestimates
and 5 underestimates.
Table 5. Summary of the 85 DOT Corrosive Chemicals Study using New Categorization Test
Total Tested
Correct Corrosive/Noncorrosive Classification (concordance)
Correct Packing Group assignment (concordance)
Higher Packing Group Assignment (overestimates)
Lower Packing Group Assignment (underestimates)
Number of
Chemicals
85
85
71
9
5
New Study
With
Screening Test
100%
84%
10%
6%
Original
Study Without
Screening Test
97.7%
79%
16%
5%
797
-------
Table 6. Corrosive Chemicals of Hazardous Material Table (CFR 49. 172.101 U.S. Department of Transportation).
Overestimates
Underestimates
71
9
5
84%
10%
6%
Chemical
Benzyl Chloroformate
Fluorosulfonic Acid
Nitric Acid
Selenic Acid
Sulfur Monochloride
Trifluoroacetic Acid
Acetic Acid, Glacial
Acetic Anhydride
Acetyl Bromide
Acrylic Acid
Aluminum Bromide, anhydrous
Aluminum Chloride
Ammonium Hydrogen Fluoride
Ammonium Hydrogen Sulfate
Anisoyl Chloride, ortho-
Antimony Tribromide
Antimony Trichloride
Boron Fluoride-dihydrate
Boron Trifluoride-acetic acid complex
Bromoacetic acid
Bromoacetyl bromide
Chloroacetic acid
Chromium (ffi) Fluoride
Cyclohexylamine
Dichloroacetic acid
Dichloroacetyl Chloride
Dichlorophenylphosphine
U.N.
#
1739
1777
2031
1905
1828
2699
2789
1715
1716
2218
1725
1726
1727
2506
1729
NA 1549
1733
2851
1742
1938
2513
1751
1756
2357
1764
1765
2798
Chemical
Class
acid ester
acid
acid
acid
halogen derivative
acid
acid
anhydride
halogen derivative
acid
metal salt
metal salt
halogen derivative
acid
halogen derivative
metal halide
metal halide
reducing agent
reducing agent
acid
halogen derivative
acid
metal salt
amrne
acid
halogen derivative
acid
Cone.
95%
Pure
90%
95%
98%
99%
99+%
Pure
99%
99%
98+%
Pure
98%
neat
97%
99%
100%
96%
98%
99+%
98+%
99+%
97%
99%
99+%
99+%
97%
pHof
10%
2.54
0
0
0
5.2
0.75
2.3
1.99
-2.0
2.07
1.22
2.92
5.22
0.78
0.72
035
0.3
0.41
0.95
1.41
0
1.44
3.9
12.34
0.64
0.46
0
CORROSITEX
Time
Mean
>4 hours
0:00:13
0:00:34
0:01:41
0:05:54
0:04:30
0:29:31
0:47:00
0:00:42
0:29:00
0:10:30
0:16:30
0:26:30
0:13:00
0:10:20
0:22:00
0:04:15
0:01:15
0:03:34
0:09:30
0:02:05
0:04:30
3:00:00
0:31:00
0:08:00
0:05:00
0:02:00
Std
0:00:02
0:00:05
0:00:07
0:00:28
0:00:12
0:02:58
0:02:29
0:00:27
0:00:13
0:00:10
0:01:53
0:00:03
0:02:28
0:00:19
0:00:07
0:00:00
0:00:07
0:00:13
0:00:31
0:00:08
0:01:49
0:15:00
0:12:04
0:01:12
0:01:01
0:00:04
Cate-
gory
A,
Ai
A!
Ai
A,
A,
Ai
A!
A,
A,
A!
A!
A!
A,
A,
A,
A,
A,
A!
A!
A,
A!
A2
Bi
A!
A,
A,
Packing
Group
NC
I
I
I
n
n
n
n
i
n
n
n
n
n
n
n
n
i
n
n
i
n
NC
n
n
n
i
DOT
Table
I
I
I
I
I
I
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
In Vivo
Group
NC
I
I
I
n
NC
n
Concordance
1
1
1
1
under
under
1
1
over
i
1
1
1
1
1
1
1
1
1
1
over
1
1
1
1
1
over
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
CD
co
-------
CD
CD
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
Chemical
Diethylene triamine
Dimethylbenzylamine
Dimethylcarbamyl Chloride
Dimethylcyclohexylamine 2,3-
Dodecyl trichlorosilane
Ethylenediamine
Ferrous Chloride Tetrahydrate
Fluoboric Acid
Formic Acid
Fumaryl Chloride
Hydrobenzene Sulfom'c Acid
Hydrogen Bromide
Iodine Monochloride
Lithium Hydroxide, Monohydrate
Mercaptoacetic Acid
Octadecyltrichlorosilane
Octyltrichlorosilane
Phenyl acetyl chloride
Phenyl trichlorosilane
Phosphorus pentachloride
Phosphorus tribromide
Potassium Hvdrogen Sulfate
Potassium Hydroxide
Sodium Hydrogen Fluoride
Sodium Hydroxide, solid
Sulfuric Acid
Sulfurous Acid
Tetramethyiammonium hydroxide pentahydrate
Thiophosphoryl Chloride
Trichloroacetic acid
Trichlorotoluene
Triethylene Tetramine
Valeryl Chloride
1 (2-AE)piperazine
U.N.
n
2079
2619
2262
2264
1771
1604
NA 1759
1775
1779
1780
1803
1788
1792
2680
1940
1800
1801
2577
1804
1806
1808
2509
1813
2439
1823
1830
1833
1835
1837
1839
2226
2259
2502
2815
Chemical
Class
armne
amine
halogen derivative
amine
silane
amine
metal salt
acid
acid
halogen derivative
acid
acid
halogen derivative
base
acid
silane
silane
halogen derivative
halogen derivative
halogen derivative
halogen derivative
salt
base
halogen derivative
base
acid
acid
base
halogen derivative
acid
solvent
amine
halogen derivative
Cone.
99%
99+%
98%
99+%
98%
neat
pure
48wf%
96%
95%
65v\rt-%
48%
98%
98%
97%
95%
97%
98%
98%
98%
97%
35-37%
Pellets
99%
Pellets
100%
neat
99%
98%
99+%
99%
60%
98%
99%
pHof
10%
12.01
10.7
2.31
11.79
0.5
12.13
2.05
1.3
1.55
0.05
0.55
0.3
0.75
11.8
0.3
0.3
0.1
0.92
0
0
0
0.85
14
5.16
13.81
0
1.78
13.61
5.81
0.74
3.32
11.91
0.45
11.78
CORROSITEX
Time
Mean
0:34:00
1:27:00
0:17:00
1:25:00
0:11:35
0:19:00
0:35:00
0:02:24
0:06:30
0:18:50
0:13:00
0:02:39
0:03:12
0:20:00
0:11:37
0:17:00
0:07:30
0:13:20
0:07:00
0:00:18
0:01:00
0:21:00
0:06:50
1:36:00
0:12:00
0:01:30
0:18:00
0:11:30
0:10:08
0:11:00
> 4 hours
0:49:00
0:10:40
0:44:15
Std
0:00:22
0:00:00
0:00:18
0:02:00
0:00:09
0:00:30
0:10:00
0:00:10
0:00:36
0:00:12
0:00:35
0:00:18
0:00:03
0:00:24
0:02:27
0:00:18
0:00:40
0:00:31
0:00:18
0:00:01
0:00:11
0:00:16
0:00:05
0:00:00
0:02:00
0:00:04
0:03:24
0:00:45
0:00:23
0:00:21
0:01:19
0:00:09
0:05:11
Cate-
gory
Bi
B2
A2
Bi
A,
Bi
A2
A,
A,
A,
Ai
Ai
A,
Bi
A,
A,
A,
A,
A,
Ai
A,
A,
Bi
A2
Bi
A,
A,
B,
A!
A,
A2
Bi
Ai
Bi
Packing
Group
n
NC
n
m
n
n
m
i
n
n
n
i
n
n
n
n
n
n
n
i
i
n
n
NC
n
i
n
n
n
n
NC
n
n
n
DOT
Table
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
n
ra
In Vivo
Group
m
n
m
m
i
i
NC
I
NC
n
n
Concordance
1
under
1
under
1
1
1
over
1
1
1
over
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
-------
co
o
o
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
Chemical
2,2-Amino ethoxyethanol
Benzene sulfonyl chloride
Butyric Acid
Butyric Anhydride
Coppei(II) Chloride
Crotonic Acid
Cyanuric Chloride
Diaminopropane
Dicyclohexylamine
Diethylaminopropylamine
Ethanolamine
Ethylhexylamine
Fenic Chloride
Hexanoic Acid
Hydroxylamine Sulfate
Maleic Acid
Maleic Anhydride
Phosphoric Acid
Propionic Acid
Sodium Hydrogen Sulfate
Sodium Hypochlorite w / 5% available chlorine
Sulfamic Acid
Tetraethylenepentamine
Tributylamine
U.N.
#
3055
2225
2820
2739
2802
2823
2670
2258
2565
2684
2491
2276
1773
2829
2865
NA2215
2215
1805
1848
1821
1791
2967
2320
2542
Chemical
Class
halogen derivative
acid
anhydride
metal salt
acid
halogen derivative
solvent
anune
amine
amine
amine
metal salt
acid
salt
acid
anhydride
acid
acid
acid salt
oxidizer
acid
anune
amine
Cone.
98%
neat
99%
99.2%
97%
99+%
99%
99+%
99%
99+%
99+%
98%
98%
99.5%
97+%
99%
99%
85%
99+%
pure
5% Cl
99+%
neat
99+%
pHof
10%
11.3
1.8
2.15
3.08
2.99
2.3
1.72
12.06
9.57
12.17
11.82
11.98
3
3
3.58
1.3
1.05
0.85
2.68
0.75
11.65
0.65
11.85
10.7
CORROSITEX
Time
Mean
0:31:00
3:30:00
0:55:40
2:30:00
0:42:41
1:22:25
3:40:00
0:21:36
3:30:00
0:61:00
0:21:41
2:45:00
0:21:18
2:29:00
3:30:00
0:15:33
0:34:34
0:15:00
0:41:00
0:14:04
> 4 hours
0:20:55
1:01:00
> 4 hours
Std
0:04:09
0:14:22
0:04:20
0:02:35
0:06:20
0:00:00
0:04:05
0:00:14
0:02:17
0:00:40
0:00:00
0:14:21
0:00:59
0:15:00
0:00:00
0:00:00
0:02:19
0:00:57
0:05:15
0:01:06
0:01:14
0:00:00
Cate-
gory
B,
Ai
Ai
Ai
A,
A,
A,
B,
Bi
Bi
Bi
B!
Ai
A,
A,
A,
Ai
A,
A,
A,
B2
A,
B,
B2
Packing
Group
n
m
n
m
m
m
m
i n
m
m
n
m
n
m
m
n
n
n
n
n
NC
n
m
NC
DOT
Table
m
m
m
m
in
m
m
m
m
m
m
m
in
m
in
m
m
m
m
m
m
m
m
m
In Vivo
Group
n
n
i
n
n
n
n
NC
NC
Concordance
1
1
1
1
1
1
1
under
1
1
1
1
1
1
1
over
over
1
1
over
1
over
1
1
-------
In the studies indicated below, addition, 776 samples were analyzed in the CORROSITEX system. A
summary of the data is presented in Table 7. The samples include dilutions of pure chemicals,
petrochemicals, surfactants, industrial cleaners, waste, etc. Excellent concordance to in vivo data was
observed.
Table 7. Summary of CORROSITEX Results with 776 Test Materials
1
2
3
4
5
6
7
8
Study
General Industrial
Pure Surfactants
Mixtures and Dilutions
ECVAM Prevalidation/ CL
ECVAM Prevalidation/ IVI
A French Industry
A Swiss Industry
European Union Reference
Chemicals
Number of
Samples
104
63
395
50
50
45
19
50
No. of
Qualified
Samples with
In Vivo Data
58
34
207
50
35
19
16
47
Concordanc
e to In Vivo
88%
97%
78%
87%
83%
89%
94%
89%
% Under
3%
3%
6%
13%
17%
11%
0%
2%
% Over
9%
0%
16%
0%
0%
0%
6%
8%
CONCLUSION
Traditional in vivo corrosivity testing has demonstrated a number of limitations. The in vivo rabbit test,
previously prescribed by the DOT, lacks in accuracy, repeatability and humaneness. It is costly and time
consuming to industry overall. These limitations, among others, prompted the DOT to issue an
administrative exemption (DOT E-10904) for the use of CORROSITEX as an alternative to the rabbit test
for assigning UN Packing Groups to corrosive materials. OSHA subsequently followed the DOT's lead
and is now accepting CORROSITEX results for workplace safety.
In vitro methods, such as CORROSITEX, can be used for a wide variety of applications in a diverse
number of industries. Large companies, such as Chemron, Aldrich, Westinghouse, Texas Instruments,
and General Electric, are among more than 200 organizations using this new in vitro assay for products,
formulations and/or hazardous waste characterization. The accuracy of the test has shown to be far greater
for determining the corrosivity of substances compared to using either the in vivo or pH methodologies.
CORROSITEX is accepted by both the United States Department of Transportation and the Occupational Health
and Safety Administration, as well as Canada, Germany and Switzerland, for defining biological corrosivity.
InVitro International is currently in the process of working with other governmental agencies to gain
acceptance for CORROSITEX as well as seeking an expansion of the current DOT exemption. These
steps and acceptances will provide companies and others worldwide with the ability to more quickly, cost
effectively, and humanely ensure public safety.
REFERENCES
1. Code of Federal Regulations, Transportation Title 49, Part 173, Appendix A. Method of Testing
Corrosion to the Skin (1991).
2. Code of Federal Regulations, Protection of Environment Title 40, Part 261.22.
3. United Nations (1973). Transportation of Dangerous Goods. Orange Book. Special
Recommendations Relating to Class 8 p. 173.
801
-------
119
REACTIVITY CHARACTERIZATION
BY DIFFERENTIAL SCANNING CALORIMETRY
Marcela L. Siao. Chemist, Laboratory Services Division, and Joe H. Lowry, Senior
Science Advisor, Laboratory Services Division, National Enforcement Investigations
Center, US Environmental Protection Agency, Denver, CO 80225
ABSTRACT
Differential scanning calorimetry (DSC) was evaluated for the characterization of materials
for RCRA reactivity. Substances with National Fire Protection Association (NFPA)
reactivity ratings and actual waste specimens were analyzed for exothermic decomposition
energy under a number of DSC experimental conditions. The results indicate that the
exothermic decomposition energy observed with DSC using high pressure capsules and
scanning to 400 °C is plausible as a regulatory method when combined with a threshold of
1000 Joules/gram of exothermic energy. The above conditions can discern some
substances with NFPA reactivity ratings of 3 and 4.
INTRODUCTION
The evaluation of the thermal reaction hazard of wastes is of interest to the Agency.
Presently, the Agency's hazardous waste regulations provide for the identification of
RCRA reactivity characteristic wastes by addressing the thermal reaction hazard, but the
regulations do not provide analytical methods or thresholds. Differential thermal analysis
and differential scanning calorimetry are used in the chemical industry in quantitating heats
of reaction and studying reaction kinetics as an important aspect of the safety of
manufacturing, transporting, storing and processing chemicals. Standardization of
thermoanalytical practices and methodology has been undertaken (1-7). Such
standardization is a desirable for regulatory test methods. Adaptation of such methodology
could fill the Agency's regulatory gap. To this end, Science Applications International
Corporation in cooperation with Perkm Elmer produced a DSC method for use in reactivity
characterization for the USEPA in 1992 (8). This method used 60 microliter stainless steel
Vitron™ O-ring sealed capsules that withstand an internal pressure of 150 psi, collected the
thermogram from 20 to 250 °C, and recommended a reactivity threshold limit of 1000
Joules/gram of exothermic energy.
The wording of the RCRA reactivity characteristic in the regulations is quite similar to that
in the National Fire Protection Association (NFPA) code for reactivity (9). Given the
RCRA ties to NFPA, a desirable characteristic of a method and a respective threshold for
identifying wastes as thermal hazards, would be consistency with the NFPA reactivity
rating. Moreover the method should provide quantitative data and be simple to conduct.
The DSC method described below addresses these criteria and is a realistic option for
regulatory testing for the RCRA characteristic of reactivity.
802
-------
EXPERIMENTAL
Differential scanning calorimetry was performed on a Perkin Elmer DSC 7 thermal analysis
system. The software supplied by the manufacturer was used for the analysis of DSC
thermograms. Commercial samples of indium, tin and lead (99.999% purity) were used as
calibration standards. The melting transitions of indium, tin and lead at 156.6, 231.9 and
327.4 °C have enthalpies of 28.4, 59.6 and 23.2 Joules/gram, respectively. Nitrogen was
flushed through the DSC cell at the rate of about 80 mlVmin. The heating block was cooled
constantly with an ice bath. Stainless steel, Vitron™ O-Ring sealed capsules with a
volume of 60 microliters that can withstand an internal pressure of 350 psi (24
atmospheres) were used for the preliminary studies where thermograms were scanned to
300 °C. Stainless steel, copper sealed capsules with a volume of 30 microliters that can
withstand an internal pressure of 2175 psi (150 atmospheres) were used for experiments
where thermograms were scanned to 400 °C. Evaluation of the two capsules is part of the
study and is discussed below. Heating rates of 10 or 20 °C/min were employed.
RESULTS AND DISCUSSION
Sample Holders
The stainless steel, Vitron™ O-ring sealed capsules; stainless steel, copper sealed capsules;
and glass capillaries have can withstand maximum temperatures of 300,400, and 500 °C;
pressures of 24, 150, and greater than 204 atmospheres; and have maximum volumes of
60, 30, and 4 microliters, respectively. The glass capillary system is reported to be able
withstand greater than 204 atmospheres and possibly as high 680 atmospheres (10). In
consideration of the volumes and pressure limitations of the capsules, and assuming all
water molecules are in the gas phase, the maximum water volumes were calculated for each
capsule at different temperatures and are given in Table 1.
Table 1 Sample Holder Water Capacity
Sample Holder
SS Vitron O-ring
SS Copper ring
Glass Capillary
Holder
Vol,uL
60
30
4
P, alms
24
150
204
680
Temp, °C
250
300
250
400
250
400
500
250
400
500
Water
Vol, uL
0.605
0.551
1.89
1.47
0.342
0.266
0.232
1.14
0.887
0.772
As indicated by the calculated water sample size, all sample capsules have very limited
sample capacity. The nominal sample size is about one milligram regardless of the capsule
type. This very small sample size seems inadequate when considering the heterogeneity
observed in most hazardous waste materials. However, it has been our experience, that
most wastes encountered that are reactive are usually fairly pure substances or simple
803
-------
mixtures and formulations. Thus limited sample size should not restrict this test with
respect to ensuring a representative subsample.
The stainless steel, Vitron™ O-ring sealed capsules were found to be inadequate because
of the temperature and pressure limitations. The thermal reaction energy produced by
nitrocellulose could not be contained consistently, even at quite small sample sizes.
Furthermore, the maximum temperature of 300 °C is not high enough to observe the
thermal reaction energy of many compounds (11).
The glass capillary is capable of higher temperature and pressure, but has drawbacks, such
as the smaller sample size compared to the stainless steel, copper sealed capsule.
Moreover, two matched silver capillary sample holders must be fabricated and a minitorch
with a liquid nitrogen cold ringer must be assembled for sealing the glass capillaries.
Furthermore, placement of the sample inside the tiny capillary may be tedious. The glass
capillaries are, however, more suitable for handling volatile substances and are more inert
than the stainless steel capsules. Corrosion by acids on the stainless steel may present a
problem.
Of the two sample holders employed, the stainless steel, copper sealed capsules were the
most rugged. They can be cleaned for repeated use while the Vitron™ O-ring capsules can
not. Furthermore, they contained the most thermally reactive compounds tested even at
sample weights as high five milligrams.
Inorganic Compounds
Table 2 presents DSC results and provides NFPA codings for some common inorganic
chemicals. The decomposition energy of a functional group is not systematic (11). For
example, ammonium nitrate and nickel nitrate had large exothermic decomposition energies
while sodium nitrate and potassium nitrate did not. Some inorganic chemicals recognized
as oxidizers show an exothermic decomposition, e.g. hypochlorites, chlorates, peroxides,
perchlorates, and persulfates. Large exothermic decomposition energies are observed for
salts of ammonium with oxygen containing acids, e.g. ammonium nitrate and ammonium
perchlorate.
Table 2 DSC Thermal Reaction Energy for Some Inorganic Compounds
Substance
Perchloric Acid
Hydroxylamine
Ammonium Perchlorate
Potassium Perchlorate
Hydrogen Peroxide
Sodium Peroxide
Calcium Hypochlorite
Potassium Chlorate
NFPA
Code
3
3
4
1
3
1
2
1
Energy
J/g
-5660
-1030
-3570
-236
-291
-289
^19
-74
Substance
Potassium Hydroxide
Sodium Hydroxide
Ammonium Persulfate
Potassium Persulfate
Ammonium Nitrate
Nickel Nitrate
Sodium Nitrate
Potassium Nitrate
NFPA
Code
1
1
3
Energy
J/g
234
217
-295
-352
-1590
-1550
167
62
804
-------
Ideally, a DSC based reactivity method and threshold should be able to discern zero coded
compounds as non-reactive. With respect to the NFPA coding and basing reactivity
characterization simply on the observation of exothermic decomposition energy, false
positive identification would be obtained for ammonium persulfate, potassium persulfate,
and nickel nitrate. False negatives would be obtained for potassium hydroxide and sodium
hydroxide. In summary five of the sixteen inorganic compounds tested would be
misclassified with respect to the NFPA coding system. A threshold limit of 1000
Joules/gram of exothermic energy would classify nickel nitrate as reactive and only four of
the eleven reactive compounds would be classified correctly.
Organic Compounds
A large number of organic compounds are capable of exothermic decomposition. The DSC
decomposition energies of 32 organic compounds or substances and their respective NFPA
codings are given in Table 3. The relationship between functional group and exothermic
decomposition is more systematic than for inorganic compounds (11). Relatively low
energies were observed for some common "non-reactive" carbon, hydrogen and oxygen
containing compounds, e.g. cellulose, cornstarch, glacactose, glucose, and pyruvic acid.
Large energies were observed for many peroxides and nitro- containing compounds.
Table 3 DSC Thermal Reaction Energy for Some Organic Compounds
Substance
Cornstarch
Cellulose
Galactose
Glucose
Pyruvic Acid
Sucrose
Urea
p-Nitrotoluene
Ally Bromide
Maleic Anhydride
Benzyl Chloride
Acetic Anhydride
Acrylic Acid
NFPA
Code
0
1
1
1
1
2
Energy
J/g
-306
-166
^95
^06
-535
298
252
188
-710
-54
-148
-167
243
Substance
Nitroethane
Nitromethane
Propyl Nitrate
Nitrocellulose
p-Nitroaniline
2,4-Dinitroaniline
3,4-Dinitroaniline
2,4-Dinitrotoluene
o-Nitrotoluene
m-Nitro toluene
Trinitrotoluene
o-Dinitrobenzene
NFPA
Code
3
3
3
3
3
3
3
3
4
4
4
4
l-Chloro-2,4-Dinitrobenzene 4
o-Nitrophenol
p-Nitrophenol
Dibenzoyl Peroxide
t-Butyl Hydroperoxide
Cumene Hydroperoxide
t-Butyl Peroxybenzoate
4
4
4
4
4
4
Energy
J/g
-23
399
-495
-1730
-1850
-912
-1710
-844
-1360
268
-2540
136
-56
-3515
-1834
-1480
-800
-1040
-1590
The wide range of energies observed for compounds with NFPA codings of greater than
zero, precludes identification of all the compounds as reactive through these test results.
Use of a threshold limit of 1000 Joules/gram of exothermic energy would classify only ten
805
-------
of the twenty four "reactive" compounds properly, but no "non-reactives" would be
misclassified.
It should be noted that the weighing of nitroethane, nitromethane, and propyl nitrate was
quite difficult due to the volatility of these compounds and the energies are suspect
Waste Specimens
Table 2 presents DSC results for five waste materials that had been determined to be
reactive previously through constituent identification using variety of analytical instrumental
techniques. Use of a threshold limit of 1000 Joules/gram of exothermic energy would
classify these solid wastes as reactivity characteristic wastes. PETN (pentaerythritol
tetranitrate), RDX (cyclotrimethylenetrinitramine), and TNT (trinitrotoluene),
"noninitiating" explosives, are interestingly considered relatively insensitive to heat (12).
The rocket propellant was a mixture of potassium perchlorate with butyl rubber.
Table 2 DSC Thermal Reaction Energy for Some Wastes
Energy
Substance J/g
PETN -1690
RDX -3750
TNT -2540
Rocket Propellant -1300
Gunpowder -3690
CONCLUSIONS
Many substances with NFPA codings greater than zero are not discernible from zero coded
or uncoded substances by DSC thermal reaction energy. Some substances with NFPA
reactivity codings of 3 or 4 have DSC thermal reaction energies greater than 1000
Joules/gram of exothermic energy. The exothermic decomposition energy observed with
DSC using high pressure capsules and scanning to 400 °C is plausible as a regulatory
method when combined with a threshold of 1000 Joules/gram of exothermic energy. This
regulatory testing scheme would only identify some of the substances classified as thermal
reaction hazards by the NFPA.
REFERENCES
(1) ASTM E476-87, Standard Test Method for Thermal Instability of Confined
Condensed Phase Systems (Confinement Test), In Annual Book of ASTM
Standards, American Society for Testing and Materials, Philadelphia, PA.
(2) ASTM E793-85, Standard Test Method for Heats of Fusion and Crystallization by
Differential Scanning Calorimetry, In Annual Book of ASTM Standards, American
Society for Testing and Materials, Philadelphia, PA.
806
-------
(3) ASTM E967-92, Standard Practice for Temperature Calibration of Differential
Scanning Calorimeters and Differential Thermal Analyzers, In Annual Book of
ASTM Standards, American Society for Testing and Materials, Philadelphia, PA.
(4) ASTM E537-86, Standard Test Method for Assessing the Thermal Stability of
Chemicals by Methods of Differential Thermal Analysis, In Annual Book of ASTM
Standards, American Society for Testing and Materials, Philadelphia, PA.
(5) ASTM E698-79, Standard Test Method for Arrhenius Kinetic Constants for
Thermally Unstable Materials, In Annual Book of ASTM Standards, American
Society for Testing and Materials, Philadelphia, PA.
(6) ASTM E472-86, Standard Practice for Reporting Thermoanafytical Data, In Annual
Book of ASTM Standards, American Society for Testing and Materials,
Philadelphia, PA.
(7) ASTM E473-85, Standard Definitions of Terms Relating to Thermal Analysis, In
Annual Book of ASTM Standards, American Society for Testing and Materials,
Philadelphia, PA.
(8) O.M. Fordham, Personal communication, 1992.
(9) National Fire Protection Association, Fire Protection Guide to Hazardous Materials,
10th edition, 1991, National Fire Protection Association, Quincy, MA.
(10) L. Whiting, M. Labean, and S. Eadie, Thermochimica Acta, 1988, 136,231-245.
(11) T. Grewer, Thermochimica Acta, 1991, 187, 133-149.
(12) R. Saferstein, Criminalistics; An Introduction to Forsenic Science, 4th edition,
1990, Prentice-Hall Inc., Englewoods Cliffs, NJ.
807
-------
120
Overview of NIST SRM Activities for Inorganic Environmental Studies
Jean S. Kane
Standard Reference Materials (SRMs) are used to assess the validity
of analytical methods for environmental studies. This use assures
that decisions regarding environmental contamination and hazards
are based on data of demonstrable accuracy- The National Institute
of Standards and Technology first began issuing environmental
reference materials almost twenty years ago. Both the number of
environmental SRMs available, and the number of units per year
sold, have been continuously increasing ever since. This paper will
review inorganic environmental SRM certifications completed in the
last two years, and will also present those now in process, which
should be completed in the near future.
808
-------
AUTHOR INDEX
-------
AUTHOR INDEX
Acheson, E 6
Adams, M.R 117
Allen, H 63,84
Allen, R 75,92
Anderson, D.A 3
Angelos, K.M 79,117
Austern, B.M 105
Balogh, M.P. 82
Baratta, E 111
Barclay, D 41
Barnes, K 58
Barren, j 74
Bashe, W.) 52
Bath, R.J 22
Baudai, F. 116
Bauer, W. 116
Beasley, H.L 85
Bechtold, R.A 79,117
Beckert, W.F. 83
Benedicto, J 83
Bennett, D 69
Bennett, J.T. 10
Bettencourt, B 53
Birri, J 37
Bottrell, D 14
Bouvier, E 82
Boyer, D.M 20
Bredt, O.P. 7
Brockhoff, C.A 46,50
Brown, K.W. 40
Bruce, M.L 71,87
Bucina, L.S 91
Bumgarner, D.L 91
Bursey, J 26
Butler, LC 20,43
Camann, D 107
Camp, L 49
Campbell, B 5
Caradamone, A 80,89
Cardenas, D 40,59
Carlsen, T.M 8
Carlson, L 39
Carlson, R.E 63,88,100,103
Carter, K.R 87
Castellanos, M.D 52
Cava, M 96
Chacko, M.V. 9
Chen, P.H 99
Chessmore, R.B 13
Chin, R 73
Christenberry, R 49
Clayton, A 48
Connolly, J.M 3,13
Connolly, M.) 10,116
Conrad, E. E 76,102
Cosby, W.C 112
Cosulich, W.F 36
Cotter, R.C 82
Creed, J.T. 46, 50
Crouse, D 63,88
Crowder, C.A 10
Cypher, R.L 35
Danahy, R 5
Danielson, V. 89
Dobb, D.E 40,43,59
Dolata, LA 80,89
Doster, J.A 52
Dugan, T. 5
Edmonds, R 6
Edwards, M.D 97
Einerson, j.j 13
Eldridge, M.C 17
Eng, D 9
Fadeff, S.K 112
Fan, T.S 63,84,88
Farr, C 66
Ferguson, B 85
Ferguson, J.D 41
Finch, S 27,31
Finlin, B 98
Flax, P. 19
Flecker, J.R 86
Foster, R 44
Friedman, S 75,92
Garlington, S.E 1
Garner, F.C 43
Gaskill, A 4
Geno, P. 107
Gobeli, D 25
Goheen, S.C 112
Goldberg, M 48
Gomez, R.T. 32
Gordon, V.C 118
Grant, C.L 72,81
Gravel, D 116
Grese, R.P. 97
Grosser, Z 54,58
Harrison, R.0 63,100,103
Hayes, M 77
Heithmar, E. M 40
Herzog, D.P. 77,86
Hewitt, A.D 81
Hiatt, M 66
Hillman, D.C 20
Ho, P. 83
Holmquist, B 39
Hosaka, T.Y. 7
Hsu, ].P. 107
Hudak, R.T. 64,94,104
Huff, D.R 57
Huff, E.A 57
Humphreys, J 114
Ilias, A.M 11,109
Inn, K.G.W. 114
Ippolitto, T. 19
Jackson, M 26
James, B 38
Jassie, L.B 41
Jenkins, T.F. 72,81
Jesser, R.A 108
Johnson L 26
Johnson, M 14
Jourdan, S 77
Kanaan, M 9
Kane, J.S 120
Kassakhian, G.H 47
Keimig, T. 100
Kelly, K.P. 76,102
Kester, P.E 110
Kilduff, J.E 56
Kim, R 83
Kimbrough, D.E 18,21,28,42,51,73
King, E.E 41
Kircher, C. C 1
Kiszka, V.R 8
Knowles, M 98
Ko, F.H 60
Kolb, S.J 12
Koo, J.K 93
Kranz, L 15
Kuhl-Klinger, K.J 7
Kumar, S 2
Lachman, C.E 86
Laing, G.A 43
Lane, B.E 17
Langone, F.A 36
Lark, D.T. 29
-------
Larkin, K.A 85
Lawurk, T.S 86
Lazarus, L 19
Lee, S.M 69
Lee, Y.D 93
Legge, P.J 1
Lesnik, B 61,69
Levy, J.M 80,89
Lewis, E.T. 110
Liem, F 34
Liljestrand, H.M 93
Litman, R 113
Litow, R 9
Lo, 1 93
Lopez-Avila, V. 83
Loring, D.A 32
Lowry, ).H 119
Lu,J 30
Luciano, V.j 44
Luka, K 4
Lynn, T.B 27
Madden, AT. 110
Mann, C.K 95
Markell, C 68
Marsden, P. 62
Martin, T.D 46,50
Mathews, K 29
Matt, | 85
McClure, G 67
McCormick, E.F. 72
McCulloch, M 112
McDaniel, W.H 52
McCaughey, J 26
Mclntosh, S 54
Meachum, T.R 13
Medina, G.] 11,109
Melby, J.M 64,90,98
Merrill, R 26
Miille, M 71
Miller, D 15
Miller, E.L 40
Miller, S 101
Minnich, LE 17
Moghadami, F 22
Mong, G.M 112
Morton, S 115
Mozer, W 33
Mullenix, M 104
Mussoline, G 38
Myers, K.F 72
Nagourney, S.J 37
Naughton, V. 110
Newberry, B 14
Nixon, G.L 101
Nolan, L 78
Novick, P. 9
Oakley, CM 1
Oehrle, S 82
Olsen, K.B 30
Ottmar, L] 43
Pacheco, S.J 47
Parker, L 106
Parr, J.L 65
Patel, J 51
Patten, B.K 29
Paustian, M 58
Petura, J 38
Philips, T.A 29
Pool, K.N 7
Potter, B 48,52
Prokopy, W 55
Purcell, J 67
Ranney, T.A 106
Ravey, R.M 80,89
Reiss, S 108
Reynolds, D 75
Rice, A.D 7
Riddell, M 39
Riley, R.G 112
Rilling, A 116
Risden, R.M 87
Roberts, D.F. 99
Robertson, G.L 2,74
Robinson, M.M 35
Romano, ).P. 82
Rowan, J.T. 20,43,59
Rubio, F.M 86
Ruebsamen, G.K 112
Russell, R.P. 36
Ryan, J 67
Sacramone, L 27
Sailer, S.J 3,10,13
Sauerhoff, S 54
Sauter, A.D 16
Schwartz, N.L 102
Schweitzer, C 39
Scifres, J 52
Setiadji, R 30
Shimizu, Y. 93
Shirkhan, H 63,100
Shobe, J 114
Siao, M.L 119
Siegelman, F.L 34
Sklarew, D.S 112
Skerritt, J.H 85
Snyder, J.L 70
Stacey, C 7
Stave, J.W. 64,90,104
Stewart, T. 75,92
Stone, M 55
Stutz, M 81
Suffet, I.M 28
Sutton, C 5
Swenson Jr., R.P. 87
Tatro, M.E 49
Taylor, L.H 7
Taylor, V. 53
Teaney, G.B 64,90,94
Teng, C.Y. 29
Thakkar, J 15
Thielke, R 4
Thomas, B.L 112
Tippler, A 67
Tomaras, A 99
Tomlinson, F.K 5
Tsang, F. 62
Turner, W.E 100
Uhlfelder, M.M 35
Van Ausdale, W.A 99
Vikers, T.) 95
Vincent, H.A 56
Vitale, R.J 38
Vu, L 37
Wagoner, D 26
Wakakuwa, J 18,21,42,51,73
Walka, R.M 36
Wang,) 30
Wasko, M.A 52
Watson, J.T. 36
Wei, R 118
Wells, J 25
Wells, R.P 3
White, D.K 45
Wilson, A.L 105
Winters, W.I 7
Wolf, R.E 45
Worthington, J.C 29
Wright, K.A 22,27
Wylie, D 39
Yeh, T. 60
Young, B 63,84,88
Young, R 83
Young, T. 98
------- |