.
ANNUAL
WASTE
. QUALITY
ASSURANCE
SYMPOSIUM
PROCEEDINGS
JULY 12-16, 1993
HYATT REGENCY CRYSTAL CITY
ARLINGTON, VA
-------
THE SYMPOSIUM IS MANAGED BY THE AMERICAN CHEMICAL SOCIETY
printed on recycled paper
-------
TABLE OF CONTENTS
Paper Page
Number Number
QUALITY ASSURANCE
1 Current Trends and Uses of Performance Evaluation Materials in the Superfund Contract Laboratory Program.
J. Barron 1
2 Improved Laboratory Performance Through Matrix-Analyte Specific Proficiency Testing. M. Caruso, K.W.
Jackson 11
3 Commerical Laboratory Position on National Laboratory Accreditation. J. Farrell, ffl, J.R. Hall, B.R. Hill,
M. Moore, J.L. Parr, J. Ploscyca, P. Sleevi, D. Speis, S. Tucker, C. Vinson 18
4 Cost Effective Project Planning: An Alternative Approach to Initial RCRA Facility Investigations. D. Loring,
S. Chapnick 29
5 Matrix Spikes and Surrogates, Do We Need Both? G. Robertson, M. Stapanian, F. Garner 37
6 Status of Performance Based Method. I. Deloatch 38
7 Quality Assurance Guidance for Environmental Sampling. N. Adolfo, A. Rosecrance 39
8 Automated Real-Time Project Level Laboratory Quality Assurance. A. Gladwell, C. Bruce Godfrey, K. O'Brien,
M. Schmick 47
9 Field Auditing Procedures for EPA Contractor Personnel. G. Janice, O.B. Douglass 56
10 The Field Chemical Data Acquisition Plan: A Data Quality Objectives Approach for the Use of Real Time Field
Analytical Data in Environmental Programs. P. Greenlaw, R.J. Bath, R.D. Spear, D. Lillian 65
11 A New Cost-Effective Strategy for Environmental Reporting Software. E.A. Haley, L.A. Richardson 66
12 Local Area Network Data Validation Software System. A. Davis, J. Vernon 71
13 Quality Assurance and Quality Control Measures for the Determination of Hexavalent Chromium from
Stationary Sources. T.P. Dux, K. Hall, G. Jungclaus, A. Williams 72
14 Development of Technology Performance Specifications for Volatile Organic Compounds. M.D. Erickson,
S.C. Carpenter, P.V. Doskey, A.D. Pflug, P.C. Lindahl, C. Purdy, W.E. Schutte 83
15 The QAPjP Qaugmire. M.A. Kuehl 89
16 Evaluation of Methods for Determinations of Instrument Detection Limits. B. Organ, D.E. Dobb, TJ. Meszaros,
J.C. Converse, R. McCallister 98
17 Technical Information and Management for the Pit 9 Project Idaho National Engineering Laboratory.
J. Owens, D. Forsberg, D. Macdonald 99
18 Quality Assurance Plans for Basic Research and New Concept Studies. E J. Poziomek, A. Cross-Smiecinski 100
19 Data Validation Guidance for Conventional Wet Chemistry Analyses. A. Rosecrance, L. Kibler 115
-------
20 Developing a Performance-Based Approach to Environmental Analytical Testing. B. Thomas, G.M. Mong,
M. McCulloch, S.K. Fadeff, D.S. Sklarew, R.G. Riley, S.C. Goheen, J. Poppiti 13°
21 Quality Assurance and Data Collection: Electronic Data Transfer. L. Tomczak, W. Lohner, E. Ray, J. Salesky,
H. Spitz J.C. Gore 136
22 An Integrated Approach to Maintaining Quality Assurance During the RI/FS Process. A. Tracy, W. O'Brien,
W. Mills 146
23 Effluent Emissions Monitoring at the DOE Hanford Site. L.W. Vance 155
24 Quality Assurance Audits of Laboratories. R.T. Winward 166
SAMPLING AND FIELD
26 Practical Suggestions to Improve the Quality of Field Work A Sampler's Perspective. T. Diebold 179
27 Evaluation of the Hydro Punch to Access Groundwater Contamination by Volatile Organics. C. Van Sciver,
E.Wallace 180
28 An Integrated Application of Field Screening to Environmental Site Investigations: A Case Study. W. Mills,
A. Tracy, T. Cline-Thomas 181
29 Divergence of Field and Lab Results in the Ponca City Investigation. S. Dutta 193
30 Field Screening for Hazardous Materials in Soil and Groundwater at a Maintenance Building. A.M. Larson,
H.R. Kleiser, P.M. Curl 206
31 Field and Laboratory Methods in Ecological Risk Assessments for Wetland and Terrestrial Habitats. G. Linder,
M. Bollman, C. Gillet, R. King, J. Nwosu, S. Ott, D. Wilborn, G. Henderson, J. DalSoglio, T. Pfleeger 210
32 Increasing the Sensitivity of Field Headspace Analysis for Volatile Organic Compounds. C. Van Sciver,
R. Fowler 222
ORGANICS
33 Regulatory Aspects of RCRA Analyses. B. Lesnik 236
34 Stepwise Development Process for Regulatory Analytical Methods. P. Marsden, S.F. Tsang, M. Roby 240
35 Recycling Program for Laboratory Organic Solvents - Methylene Chloride and Freon-113. C.A. Valkenburg,
W.T. Brown 245
36 The Technology and Performance of Several New Immunoassay Methods and a New Supporting Instrumentation
System. S.B. Friedman, J.P. Mapes, R.L. Allen, T.N. Stewart, W.B. Studebaker, P.P. McDonald, R.E. Almond,
T.A. Withers, S.P. Arrowood, D.P. Johnson 261
37 Screening of TCLP Extracts of Soil and Wastewater for 2,4-D by Immunoassay. M.C Hayes, S.W. Jourdan,
T.S. Lawruk, D.P. Herzog 262
38 A Rapid, On-Site Immunoassay for Detecting Petroleum Products in Ground Water. J.P. Mapes, P.P. McDonald,
R.E. Almond, S.B. Friedman, 276
39 Appendix IX Extractions by Accelerated One Step Extractor/Concentrator. M. Bruce, J. Carl, B. Killough,
T. Zine, D. Burkitt 283
40 Microscale Solvent Extraction Methods for Tarry Soils from Former Manufactured Gas Plant (MGP) Sites.
J. Schneider, I. Murarka, D. Mauro, M. Young, B. Taylor 29R
-------
41 Development of an Environmental Method for the Analysis of Volatile Organic Compounds in Soils by Static
Headspace GC/MS. T.J. Riga, G. O'Neil 313
42 In-situ Derivatization of Soil and Sediment Samples in an Supercritical Fluid Extraction Cell: The Extraction of
Phenols and Chlorinated Phenols. D.R. Gere, L.G. Randall, C. Knipe, W. Pipkin, H. B. Lee 326
43 Supercritical Fluid Extraction (SFE) of PCBs. M.Bruce 335
44 Determination of Organic Compounds in Ground Water by Liquid-Solid Extraction Followed by Supercritical
Fluid Elution and Capillary Column Gas Chromatography/Ion-Trap Mass Spectrometry. J.S. Ho, P.H. Tang,
J.W. Eichelberger, W.L. Budde 337
45 Comparison of Solid Phase Extraction with Salting-Out Solvent Extraction for Preconcentration of Nitroaromatic
and Nitramine Explosives from Water. T.F. Jenkins, P.H. Miyares, K.F. Myers, E.F. McCormick, A.B. Strong 338
46 Dinoseb Analysis in the Field and the Laboratory. D. Anderson, S.F. Tsang, T. Jackson, P. Marsden 347
47 Quantitation of Polychlorinated Biphenyls Using 19 Specific Congeners. P. Marsden, S.F. Tsang,
M. Kennedy, B. Lesnik 360
48 Direct Determination of TCLP Phenols and Herbicides by HPLC. R.L. Schenley, W.H. Griest, J.E. Caton 372
49 Approaching the Sensitivity of an Electron Capture Detector (ECD) for the Analysis of Pesticides by Using
GC/MS. L.C. Doherty, N. Low 373
50 Determination of Polynuclear Aromatic Hydrocarbons in Soil at 1 ug/kg Using GC/MS. B.N. Colby,
C.S. Parsons, J.S. Smith 379
51 Application of Lee Retention Indices to the Confirmation of Tentatively Identified Compounds from GC/MS
Analysis of Environmental Samples. P.H. Chen, W.S. Keeran, W.A. Van Ausdale, D.R. Schindler, D.F. Roberts 385
52 A Quick Performance-Based HPLC Method for the Analysis of Polynuclear Aromatic Hydrocarbons (PAHs).
M.W. Dong, J.M. DiBussolo 396
53 The Extraction and Analysis of Polychlorinated Biphenyls (PCBs) by SFE and GC/MS. Improvement of Net
Detection Levels. D.R. Gere, L.G. Randall, C.R. Knipe, W. Pipkin, L.C. Doherty 397
54 Environmental Analyses-PAHs in Solid Waste: Bridging the Automation Gap Between SFE and HPLC.
D.R. Gere, L.G. Randall C.R. Knipe, W. Pipkin 406
55 Analysis of PCB's in Soil, Sediments, and Other Matrices by Enzyme Immunoassay. R.O. Harrison,
R.E. Carlson, A.J. Weiss 419
56 Explosives Analysis of Atypical Matrices and Technical Enhancements to EPA Method 8330. D. Hooton,
M. Christopherson, T.P. Dux, G. Jungclaus, H. Randa, R. Sauter 430
57 Liquid Chromatography/Mass Spectrometry of High-Molecular-Weight Polycyclic Aromatic Hydrocarbons
Using the Particle Beam Interface. M. Roby, L.D. Betowski, C. Pace 438
58 Quantitation of Pentachlorophenol (PCP) by a Rapid Magnetic Particle-Based Solid-Phase ELISA in Water and
Soil. F.M. Rubio, T.S. Lawruk, C.S. Hottenstein, S.W. Jourdan, D. P. Herzog, Y. Zha 439
59 Chromatographic Optimization for the Analysis of an Expanded List of Volatile Organic Pollutants.
D.M. Shelow, M. J. Feeney 452
60 Application of SW-846 Methods to the Identification of Unusual Brominated Compounds in High-Concentration
Process Streams. D.V. Smith, L.P. Pollack, F.T. Varcoe 453
61 Results of Laboratory Tests of the Accelerated One-Step Liquid-Liquid Extractor. R.K. Smith, R.G. Owens, Jr.,
D.S. Geier, T.A. Shaw, D. Phinney, J. Carl 457
-------
62 A Matched Dual Capillary Column System for Pesticides. A. Spilkin, M.R. Hastings 466
63 Performance Data for the Analysis of Phenols, Nitroaromatics, Cyclic Ketones, Haloethers, and Chlorinated
Hydrocarbons. S.F. Tsang, N. Chau, P. Marsden, B. Lesnik 481
64 Rapid Confirmation of Nitroaromatics and Nitramines Using UV Diode Array Spectral Comparison.
B.A. Weichert, R.D. Baker, C.R. Campbell 482
65 The Effect of GPC Cleanup of Semivolatile Extracts on GC/MS Analysis. T. Willig, J.S. Kauffman 483
66 Evaluation of the Ensys PAH-RISc Test Kit. R.P. Swift, J.R. Leavell, C.W. Brandenburg 484
67 Results of Analytical Field Trials for PCBs Using an Immunoassay Technique. A. Weiss, A. Parsons 500
68 Important Factors in Enhancing Supercritical Fluid Extraction Efficiencies for Environmental Applications.
J.M. Levy, L.A. Dolata, M. Ravey 501
INORGANICS
69 Evaluation of a Rapid Steam-Distillation Procedure for the Extraction of Cyanide from Liquid and Solid Wastes.
E.M. Heithmar, R. Herman, M.R. Straka 503
70 A Method Evaluation Study for the Analysis for Hexavalent Chromium in Solid Samples Using a Modified
Alkaline Digestion Procedure and Colorimetric Determination. RJ. Vitale 505
71 The Analysis of Reducing Soils for Hexavalent Chromium. D.P. Miller 506
72 Factors Affecting Sample Throughput for an ICP-OES System with a Segmented-Array CCD Detector.
Z.A. Grosser, K.J. Fredeen, C. Anderau, D.A. Yates, K.M Barnes, T.J. Gluodenis, Jr. 516
73 Design, Performance, and Environmental Applications of an ICP Array Detector Spectrometer. D.D. Nygaard,
F. Bulman, M. Almeida 519
74 Variability in TCLP Metals Results from Stabilized Waste. G. Merewether, N. Shah 520
75 Analgamation CVAA to Improve the Detection Limit of Mercury in Environmental Samples. S. Sauerhoff,
Z.A. Grosser, S. Mclntosh 521
76 The Development of EPA Total Mercury Method Using Cold Vapor, Fluorescence Mercury Detection Systems.
The Mercury Method Will be Used for the Analyses of Mercury in Water, Wastewater, Sea Water and Related
Matrices. B.B. Potter, S.E. Long, J.A. Doster, R. Davis 524
77 Determination of Tetraethyl Lead in Groundwater. B.N. Colby, D.J. Bencivengo, C.L. Helms 539
78 Use of a Telephone Service and Database to Provide Guidance and Increase Public Involvement in OSW Methods.
D. Anderson, R. Carlston, S. Hartwell, G. Hansen, K. Kirkland 543
79 Extraction of Metal Ions from Solid Matrices by Complexation SFE. W.F. Beckert, Y. Liu, V. Lopez-Avila,
M. Alcaraz 544
80 Microwave Sample Preparation for Mercury Analysis via Cold Vapor Atomic Absorption Spectrometry.
W.G. Engelhart, S.E. Littau, E.T. Hasty 545
81 Study of EPA Method 300.0: Application to Hazardous Waste Analysis. H. Mehra, D. Gjerde 545
82 The Effects of Size Reduction Techniques on TCLP Analysis of Solidified Mixed Waste. R.D. Thiel 554
-------
AIR AND GROUNDWATER
83 Canister Analysis by Using Gas Chromatography/High Resolution Mass Spectrometer. J.P. Hsu, G. Miller,
J.C Pan 555
84 An Effective Moisture Control Solution for the GC/MS Determination of VOCs in Canister Samples Collected
at Superfund Sites. M.G. Winslow, D.F. Roberts, M.E. Keller 556
85 Selection Criteria for Groundwater Monitoring Well Construction Materials. J.R. Brown, L. Parker,
A.E. Johnson 566
86 Auto GC Design and Operation for Remote Unattended VOC Determinations. J.F. Ryan, I. Seeley 567
-------
QUALITY ASSURANCE
-------
•J CURRENT TRENDS AND USES OF PERFORMANCE EVALUATION MATERIALS IN THE
SUPERFUND CONTRACT LABORATORY PROGRAM
James W. Barren, Quality Assurance Coordinator, Analytical
Operations Branch Hazardous Site Evaluation Division,
Environmental Protection Agency,401 M. St., S.W.,
Washington, DC 20460
Good day, I'm Jim Barton with the Analytical Operations Branch (AOB) of the Hazardous Site Evaluation
Division. Our group manages some of the analytical services utilized by Superfund, including the contract
laboratory program commonly known as the CLP. This includes a comprehensive Quality Assurance
program. One facet of this program is ensuring an adequate variety of Performance Evaluation Materials
(PEMs), also referred to as Performance Evaluation Samples (PES). Historically, PES have been used by
government agencies for several purposes.
1. Multi-laboratory Validation Studies for validation
of analytical methodology.
2. In the form of Blind PES for Laboratory Certification.
3. In an on-going Quality Assurance (QA) Program.
The CLP uses PES for all three purposes. PES are prepared to support inter-laboratory studies to validate
CLP Statements of Work (SOWs). The Laboratories are required to participate in pre and post-award
performance sample studies to maintain contract status, the so-called "de facto certification program." We
consider number 3, the PES we provide for routine assessment of data, the most important part of our
program. I'd like to discuss some of the trends we see in the use of these materials in environmental work,
and some of the approaches we are taking in our program. We feel we have addressed on a smaller scale
many of the questions EPA will to consider as it looks toward a national laboratory certification program.
The Analytical Operations Branch has established and operates a Superfund Performance Evaluation Sample
Repository through it's Quality Assurance Technical Support Laboratory in Las Vegas. The Superfund
Performance Evaluation Sample Repository was established in response to general and specific requests
from Regions for an independent and relevant mechanism to monitor laboratory performance in the analysis
of environmental samples from known and suspected hazardous waste sites. A wide range of performance
evaluation samples is available at no cost to the Regions to support their needs for independent and external
monitoring of laboratory performance in the analysis of environmental samples in the Superfund Program.
A key part of AOB's program is that it does not normally distribute standards or Laboratory Control
Samples, it distributes PES materials to monitor analytical work where the government has a strong interest
in determining the quality of the data it is paying for. Some exceptions are LCS'es such as the Interference
Check Samples run to determine interferences in trace level metal analysis, and the multi-parameter LCS'es
used in some Statements of Work. (24, 26) Available performance evaluation sample types are water and
soils matrices. Performance evaluation samples for air analysis have been under development, but work
has stopped due to lack of regional interest. Analytes incorporated into these samples include volatile and
semi-volatile organic compounds, pesticides and PCBs, metals, anions, cyanide, and chlorinated dioxins
and furans. Stability studies are run are archived on PES to ensure values are still within the statistical
windows established for the particular PES. Regions select the type of performance evaluation samples
-------
most suited to their needs from a widely distributed catalogue.(20)
Through the Superfund Performance Evaluation Sample Repository, the Analytical Operations Branch also
provides related support services to the Regions. A PC-based tracking system PEACTOOLS has been
developed, and is distributed to the Regions for use in assessing the analytical results for performance
evaluation samples reported by laboratories. Using this system, Regional data reviewers evaluate
laboratory performance on a particular set of environmental samples as soon as the data package is
received. These evaluations also provide important documentation of data quality, when the PE sample
is related to the site. Through the Superfund Performance Evaluation Sample Repository, Regions can also
obtain assistance in establishing and operating their own performance sample repository which can serve
regional needs in monitoring analytical work done for the Superfund Program.
AOB gets support and oversight from ORD at Las Vegas for it's program. In so far as possible we attempt
to obtain real world materials from sites to prepare PEMs, however we also purchase neat materials and
prepared sets from commercial vendors.
We prepare PES for contract compliance, in our quarterly blind program (Qbs) and for routine monitoring
of data collection activities. In both types of samples, concentrations and recipes of the analytes support
CLP, and other Superfund work. Our quarterly blind samples are the typical single blind PES supplied
by most programs for certification. Some of our Qbs are solids, but several years ago we went from full-
volume liquid samples to ampuled samples, because of problems in the samples, and to cut costs. It costs
about 10 more to prepare full volume samples compared to amputated samples. However AOB was aware
of the pitfalls of amputated PES already being used by other EPA certification programs. The first
problem is the laboratories may run the ampuled sample concentrate undiluted, the second is calling around
to compare answers, and third expending much more time and effort on these types of samples, then the
laboratory would on a routine sample. Some parties have recently come out of the closet, admitting the
excessive time spent on this type of sample, and possibly running the ampule straight ( ). The problem
we found is the labs do not regard this as cheating.just an annoyance they must get through to stay
certified. Calling around is probably more of a problem with government laboratories, where there is a
more collegia! atmosphere, than the commercial sector, where loss of certification by some laboratories,
means more business for the successful labs.
What can be done about this problem? One answer is to give less importance to the do or die
"Certification PES." Another is to send out different PESs to labs in the same study. However as noted
we do not consider answer comparing a problem in the CLP, since we have companies with multiple
laboratories, with one laboratory in the group failing a QB, while the rest pass. In any event we do not
have sufficient Laboratories to use two sets, since our Qbs are scored on a statistical basis. A third
approach AOB has completed is to develop indicator compounds that can be added to the ampulated PES
samples.(l,2) This insures an aliquot will be removed from the ampule, diluted to the proper volume, and
run according to the Statement of Work the laboratory is trying to be certified for. To be useful as
indicator compounds, for contract compliance, analytes must possess these characteristics:
• Must be Stable in the test solution matrix
• Must be detectable by GC/MS when directly injected from ampule.
• Good Chromatography when directly injected and when run from normal extract
• Must be degraded or poorly recovered by the sample preparation technique used.
• Must be a clearly detectable difference in the results by direct injection and normal preparation.
-------
Most of the people in our branch our experienced environmental chemists. While a preliminary study to
pick candidate indicator compounds included VOCS, our experience we reasoned that probably no tries to
run the ampule for VOCs, since this could produce a worse answer, that running it properly. This leaves
Semi-volatiles, metals and pesticides. The first group picked were the semi-volatiles.
Chemical
1-Hexanamine
Cyclohexalamine
2-Propenamide
1,3-Benzenediol
Phthalic Anhydride
Nicotine
Di-n-hexylamine
Simazine
Norflurazon
CAS number
111-26-2
108-91-8
79-06-1
108-46-3
85-44-9
54-11-5
143-16-8
122-34-9
27314-13-2
The above are some of the materials we decided to test as indicators, that were selected from a larger
list. (2) Our reports are available to parties who wish to design their own Indicator program. Since we
now use continuous liquid -liquid extraction at a low Ph we narrowed the field quickly. Interestingly
enough the compound we may use is nicotine. This was accomplished since the CLP requires
identification of tentatively identified compounds (TICs). Since the laboratories did not identify all the
materials, it was necessary to run tape audits on the raw data. EPA, EMSL-LV helped us with tapes for
GC/MS Systems we did not have equipment for. There were other logistics involved, such as picking
which internal standard to us with which indicator. The next group we are going to work on are the
metals. The task will in effect reverse the work we did in developing interference check samples,(3,4))
by adding those interferents at levels that must be diluted to the proper volume, to be able to identify target
analytes. At present the Pesticides offer no easy solution, since they are done by Gas Chromatography.
The course we are following here is to build on work done by the Office of Drinking Water, in developing
method 525, and try to achieve low enough quantitation limits, so all Target Compound list Pesticides could
be run by GC/Mass Spectrometry. This would enable us to also develop indicator compounds for the
Pesticides.(34) Indicator compounds are not an absolute determinant of fraud, but they are one more tool
to ensure data integrity.
The CLP Quarterly Blind PE Sample Study data does have other uses besides contract compliance:
• CLP Laboratory Performance Measurement - This is the primary mission of the PE Study
Program. All labs are measured against an independently prepared research mixture.
• Surrogate Optimization- Compounds added to the PES are checked for correlation to the Target
Compound List, to find new surrogates which can give better analyte recovery information.
• Feasibility of Target Compound List additions - The QB studies have been used to determine
how well SOW perform with new analytes which have become candidates for addition to the Target
Compound List.
• Method Performance Reports - Because the QB samples are uniform samples which are sent to
all laboratories, inter-laboratory precision and accuracy is determined.
-------
• Extraction Method Comparison - QB sample results allowed a comparison between Separatory
funnel versus continuous extraction techniques for semi-volatiles, which are now used exclusively in the
CLP.
• Laboratory Performance Database The QB results are part of a laboratory performance
database to determine overall performance of a Contract Lab. (21)
Failure to analyze the Quarterly Blind sample correctly, results in suspension of the laboratory until a
remedial PES is successfully analyzed. When statistically derived precision windows are used, as in this
case, a certain percentage of laboratories in the study will fail. This does not mean the data previously
produced by the laboratory is in question, unless the PES results show a trend of unacceptable
performance.
In the near future the so-called national certification PES may not be that important to the CLP. CLP
Routine Analytical Statements of Work (SOW), now being utilized, or in the testing phase (see table 1.)
shows, are more specialized, and may only have one laboratory supplying that service. For these SOWs
Table I.
Performance Evaluation Sample Use.
Document/Reference number.
CLP Low Concentration
Water/22
CLP Quick-Turn-Around
Methods/24
CLP Dioxin Furan/23
CLP High Concentration
Inorganic/28
CLP High Concentration
Organic/29
CLP Low-Medium Inorganic/30
CLP GeoTechnical/26
CLP Water and Soil
Characterizat./25
CLP Quick Turnaround
Dioxin/31
CLP Air Toxics./27
Types and Recommended Frequencies of PES
LCS (Zero-blind)
(1)
by the Sample Delivery
Group (SDG)
Special, multi-param
by the SDG or use >
by the SDG
by the SDG
by the SDG
Special, multi-parameter
PES
by the SDG
inorg., by SDG
Single or Double Blind (2)
by the SDG
by the SDG
by the SDG
by the SDG
by the SDG
by the SDG
By the SDG, All Org. could be
double-blind, on all types of
samplers, i.e. Tedlar bags, PUFs.
Contract
Action
for Poor
Performa
nee
on PES
yes
yes
yes
yes
yes
yes
yes
yes
yes
(1) All liquid samples are ampuled, all soils homogenized
(2) Single-blinds, all liquid samples are ampuled, all soils homogenized
Double-blinds, all liquids are full volume, soils are not homogenized
-------
the project officer can have the laboratory run up to 5% PES.(35) He has the option of inserting them as
double-blind, that is appearing to be a routine sample. However these "by the case" PES have their own
set of problems as we are finding out with our Dioxin/Furan SOW. Our Dioxin/Furan PES were our best
set of PEMs, all made from materials and hot spots obtained from Dioxin sites, with real world
intereferents such as PCBs present. We've now gone through our entire set, and need to reblend the PES
to produce new concentration levels. Recent catalogues received from commercial vendors indicate they
are also addressing the problem of PES recognition. (32) Laboratories keep track of PES, and will compare
old results with the latest PES.
This last problem pits two schools of thought on PES, used on a daily basis, or by the case, to assess the
quality of environmental data. The first group seems to feel that PES should have extensive multi-
laboratory validation, producing somewhat tight windows for bias and precision, resulting in fairly static
recipes due to cost. The second, which AOB subscribes to, feels PES should be economical, readily
available, and in a wide variety of realistic materials, representing the matrices and analytes being
monitored. Recipes should be changed frequently. While the PES should evidence sufficient statistical
control to assess the analytes of concern, it isn't necessary to have NIST grade materials.
As noted above, AOB distributes a variety of multi-Media PES For both CLP and non-CLP work. Initially
PES are produced to support a proposed SOW. As is often stated PES should be similar to samples
collected from a waste site, contain interferences relevant to the samples being analyzed, contain analytes,
and concentrations approximate the range found in actual environmental samples. An initial workplan for
concentrations and recipes of analytes expected to be monitored under the proposed SOW is prepared. For
an interlaboratory study the PES are then made or purchased according to the workplan. These PES
materials will continue to be made an utilized if the SOW goes into effect. The categories of PES AOB
distributes or is working on include:
CATEGORIES OF PES DISTRIBUTED OR BEING DEVELOPED BY AOB
CLP Organic Low/Medium, Low Con., High Con, Target Compounds, and TICs
CLP Inorganic Low/Medium, Hicon, Target compounds
SITE-CHARACTERIZED PES, Analytes typical of the site
SITE CATEGORIZED (INDUSTRIALLY) PES, Analytes typical of that industry
DIOXIN/FURANS
QUICK TURN-AROUND PES, Really multi-parameter Laboratory Control Samples
GEOTECHNICAL PES, Really multi-parameter Laboratory Control Samples
FIELD ANALYTICAL METHODS
We don't plan to cover every type of PES that could be produced, for example a frequent Special
Analytical Services request are the traditional water quality parameters. We wouldn't stock these since they
have been available for a long time commercially, with good quality.
PESs are one of the few external QA/QC measures for independent monitoring of
laboratory performance and demonstration of data comparability among laboratories. When the PE is site
related, it documents the quality of analytical results generated in the Superfund program, for that site.
Under the Superfund Accelerated Cleanup Model (SACM). the Office of Emergency and Remedial
Response is attempting to reduce duplicative site investigations. However this may reduce the redundancy
of data making an external QC indicator such as PE samples more important. Another Superfund initiative
has been a change in the delivery of analytical services, with more analytical resources being placed in the
region. With the increase in the use of non-CLP laboratories, field analytical procedures, and mobile
laboratories, the availability of PES validated by fixed laboratory methods would allow for data
-------
comparability. (36)
When we first set up our PES repository at Las Vegas, the philosophy was to have good range of PES in
regard to matrices, analytes present, and concentration ranges. We set up a scoring program
PEACTOOLS, which would score PES on a go/no basis. This was to be a fairly static operation. This
is no longer viable, and we are now planning to shift recipes and concentrations on a regular basis, with
more emphasis on site characterized PES.
Double-blind PES like site characterized or site specific materials, are the optimal type of PES, if available.
They constitute a PES, which the laboratory feels is a routine sample, and does not put forth that extra
effort associated with single-blind PES. Some matrices lend themselves to double-blinds better than other
matrices. Solids are the most desirable matrix for double-blinds for Superfund, but they are also the most
difficult to produce. Attempting to manufacture a solid where there is some statistical reliability of the
concentrations of analytes invariably leads to homogenization, allowing the sample to be recognized as a
PES.
There does not appear to be great difficulty in preparing water soluble analytes, so metal samples and the
so-called Water Quality Parameters do not present a major problem. (7) More extensive stability studies
are needed for this matrix to determine "shelf life." Some work has been done in the regions with Volatile
Organic Compounds (VOCs). AOB is recommending the regions follow the work done by Oak Ridge
National Labs in preparing double-blind VOC's in water.(14) AOB's QATS laboratory has recently been
able to duplicate the results of work done by the US Army's CRREL group. This involves preparing
VOCs on soil.(15,16 ) Because of the shelf life of these materials we are recommending they be prepared
at the regional level, from materials in Regional PE repositories. One problem noted in preparing double-
blinds using site-characterized materials for metals, was particle size. This was predicted from other work,
but verified by analytical results. This has spun off another project to determine if sample sizes for metals
should be changed of be more flexible. This work will also have application to field analytical procedures,
where analytical results are also affected by particle size.
Air Samples lend themselves to double-blinds, as they are trapped in canisters, or bags or on solid matrices
such as Poly-Urethane Foam (PUFs) or Tenax. In practice, real samples are indistinguishable from the
PES. Again, a major need in this area is stability studies to determine shelf life. The only air parameter
we get requests for are double-blinds prepared on PUFs for pesticides and Dioxin. Except in very
specialized case we feel these types of PES should be prepared in the Regions.
The problem of double-blind solids PES is still difficult at this point. A sample that has been homogenized
by particle size is immediately recognizable as a PE. Samples which have been thoroughly mixed, but not
sized, require sample weights far above those specified in most methods to overcome the effects of particle
size on concentration. The type of double-blind that seems to hold the most promise at this time are site
characterized materials, mentioned below, that is materials that are obtained from a site, spiked and
homogenized, but not sized, and returned to the site, to be submitted blind as part of the regular sample
stream.
What has become apparent through workgroup meetings and conferences on PES and Site-characterized
materials, is that environmental personnel are not concerned if they get a "NIST Grade" materials for this
type of sample. The higher priority is being able to assess data quality in a timely manner at a reasonable
cost, and realistic time frames.
An impediment to producing double-blind aqueous semi-volatile organic PES is more administrative than
technical. As parties have commented, returning to one liter bottles instead of the two and four liter bottles
commonly used, with rinsing of the container to collect materials that may have plated out on the glass,
-------
would allow greater accuracy in determining this type of PES. As one commentator noted, this would also
correct an obvious negative bias in real samples. Double-blinds should approximate the environmental
sample in water color, and turbidity, or soil color, texture, particle size range and even % moisture.
Containers should be the same as being used by the sampling team, and obtained from them in advance.
The references, and anecdotal information indicate double-blind PES are being prepared and utilized on
a more frequent basis than in the past.
We have done a study comparing our Contract Required Quantitation limits for target compounds against
actual results in our analytical results database, and with values decided on in Records of decision, and
the PRO cancer and non-cancer values for those analytes.(5) This has become part of our Pesticide task
mentioned above, to see at what levels we can get good quantitation for important analytes, and has led
us to re-evaluate our whole PES library. While we will probably not change our CLP CRQLs, the PES
do support other Superfund work where lower quantitation levels are needed.(6 ) This latter work
assignment actually complements our efforts to convert our CLP pesticides to GC/MS.(34)
Whether single or double-blind samples, we try to prepare site-oriented materials. AOB doesn't prepare
site-specific materials per se, only when its a large enough project to a have a surplus as site-characterized
materials. Before we get cited for not explaining sites, our perception of these materials is:
"SITE PES"
SITE SPECIFIC PES: materials obtained from "hot spots" at a waste site,
additional typical contaminants being found at that site.
Site Characterized PES: may be site specific materials, but the surplus can
at sites with similar matrices.
possibly spiked
be used
with
SITE CATEGORIZED PES: Synthetic PES prepared using information obtained from CARD and
the NPL databases, according to the industrial category of that site.
We think the Site Categorized PES is one of the more interesting projects we have worked on is our effort
to develop industrially categorized site An other branches in our division has categorized waste sites on the
NPL by the industrial category they fall under, i.e electroplating. Searching our databases we are able to
identify the cases associated with those sites, and pull down the samples associated with those cases. From
our CARD database we can get the sample results for the types of contaminants and their concentration
ranges. This allows us to prepare synthetic single blind materials that are typical of waste sites, and in
the absence of real world materials allows a project manager some means to assess the quality of a
laboratories data. There are 45 categories, but we may find enough overlap in analytes lists between
related categories, that we would only develop one set of PES to use for both sites.
Typical Information From the Databases for PES Design
Industry
Manufacturing, Gen.
Electro-plating
Fabricated Metal
Products
Sites on the NPL
683
79
136
Number of Cases
286
32
53
Number of Cases in
CARD
59
10
14
-------
A project we have provided assistance to in Region X, seems to indicate, at least for long term projects,
site characterized materials can be prepared economically, and in a timely manner.(8) The advantage to
the National Program Office, is that one of the hardest parts of the task is to persuade field personnel to
fill up a few 55 gal barrels of dirt, from hot spots if possible, and ship them to the QATS Laboratory.
However when working on a set of site-characterized PES, one can insist on a surplus of materials as quid
pro quo.
A word of clarification is probably needed to differentiate site-characterized materials from blind spikes
or matrix spikes. In spiking, one simply squirts some analytes of known concentration onto or into the
sample. Site-characterized PES are made from the real world matrix, but a liquid recipe is mixed with a
small volume of the soil, dried and blended with a larger portion of soil. Homogeneity is verified,
concentration levels, and acceptance limits are validated by multi-laboratory analysis.
We developed approaches to preparing PES for Field Analytical Methods such as field X-ray Fluorescence,
sometime ago, but with the renewed interest in Field Analytical Methods, we are going to implement the
study. (36) A survey of the Regions indicated the most commonly requested PES was for VOCs in soil.
As noted above we feel this type of PES can now be produced with some assurance of statistical control
of the analytical results. Many of the types of specialized materials, such as wipe samples, are best
prepared locally.
To summarize, we think the trend in Performance Evaluation Materials is toward:
• Away from "national certification" PES to "by the case" PES.
• More emphasis on PES tailored to specific waste sites, Site Specific, Site Characterized materials.
• Lower concentration levels, to approach as closely as possible, health effects criteria.
• "Recipes" changed more frequently to avoid PE recognition.
• Increased use of Double-blind PES.
References:
(1) Work Assignment 1.7 Indicator Compounds for Evaluating Contract Compliance in Low/Medium
and Low Concentration Semi-volatile Methods,
Quality Assurance Technical Support Laboratory, March 15, 1993
(2) Work Assignment 1.7, "Evaluationof Organic Indicator Compounds for Full Organic Multi-Media
and Low Concentration PES Programs, Quality Assurance Technical Support Laboratory,
Oct. 3, 1991
(3) Technical Guidance Document 1.6 "Development of an Interference Check Sample for Graphite
Furnace Atomic Absorption Spectrophotometry," Quality Assurance Technical Support Laboratory,
Dec. 28, 1990
(4) Technical Guidance Document 1.6 "Development of an Interference Check Sample for Inductively
Coupled Plasma Optical Emission Spectrophotometry," Quality Assurance Technical Support
-------
Laboratory, June 26, 1990
(5) Review of the Contract Laboratory Organic Target Compound List, prepared for the Analytical
Operations Branch by VIAR and Co., June, 1992
(6) Work Assignment 1.12 Re-Evaluation of Concentration Levels and Recipes Used In The
Performance Evaluation Sample Inventory, Quality Assurance Technical Support Laboratory,
Oct. 3, 1992
(7) Carter, Mark, et. al. "How Good Is Your Lab?", Environmental Laboratory,
(8) Preliminary Study on INEL Site "Site Characterized Materials Study", On the Effect of Particle
Size on Inorganic Results.
(9) Work Assignment 1-9, Quality Assurance Technical Support Laboratory, Organic PE preparation,
verification, Dec. 11, 1992.
(14) Maskarinec, Micheal P., et al, "Preparation of Reference Water and Soil Samples for Performance
Evaluation of Volatile Organic Analysis," JAOAC, vol. 72, no. 5, 1989, pg. 823.
(15) Hewitt, Alan D., US Army CRREL, "Vapor Fortification: A Method to Prepare Quality
Assurance Soil Samples for the Analysis of Volatile Organic Compounds", Nat. Symposium
on Measuring and Interpreting VOCs in Soils, Las Vegas, NV, Jan. 12-14, 1993.
(16) Hewitt, Allen D., et al, " Comparison of Analytical Methods for Determination of Volatile
Organic Compounds in Soil", ES & T, Vol. 26, No. 10, 1992, pg. 1932.
(17) Proceedings of the Site-specific Soil QA Materials Workshop, Las Vegas, NV, April 13-16, 1992
(18) Memo from Roy Jones, Region X to Jim Barren, AOB concerning differences between PES,
MS/MSD and Site-characterized materials.
(19) Management Report, "Performance Evaluation Sample Repository and Related Services"
available from AOBs' Quality Assurance Technical Support Laboratory.
(20) Quality Assurance Technical Support Laboratory, "Performance Evaluation Sample Catalogue",
1993.
(21) Memo from Larry Butler, EMSL-LV to Jim Barron, QAC, AOB on uses of the QB Program,
Feb. 5, 1991.
(22) OLCO2.0 Statement of Work, Low Concentration Water for Organic Analysis.
(23) DFLM01.0 Statement of Work, Polychlorinated Dibenzo-p-Dioxins and Polychlorinated
Dibenzofurans, Multi-Media, Multi-Concentration.
(24) Draft Statement of Work, "Quick Turn-Around Analysis"
(25) WSCIN Statement of Work, Water and Soil Characterization in Multi-Media, Multi-Concentration,
Nov., 1990
(26) Draft Geotechnical Statement of Work, Aug. 1992
9
-------
(27) Statement of Work, Air Toxics
(28) Statement of Work for High Concentration Inorganic Analysis
(29) Statement of Work for High Concentration Organic Analysis
(30) ILM03.0 Statement of Work For Inorganic Analysis, Multi-Media, Multi-Concentration
(31) Statement of Work "Rapid Turnaround Dioxin Analysis, Multi-Media", 11/92 Rev.
(32) Environmental Quality Control Standards, 1993 Catalogue, Environmental Resource Associates.
(33) WA 1.3, Development of PEMs to Support Regional Quality Assurance Needs for CLP and
non-CLP Superfund Work in Solid Media Quality Assurance Technical Support Laboratory,
continuation of WA, Feb. 26, 1993
(34) WA 2.5, Evaluation, Improvement, Creation, and Standardization of the CLP Pesticide Procedures
in the Organic RAS SOWs. Improvement of Mass Spectrometry Quantitation Levels of Semi-
Volatile Target Compounds, Jan. 30, 1993.
(35) Analytical Operations Branch, EPA/540-R-93-062, Draft "Guidance For The Uniform Use Of
Performance Evaluation Materials In The Contract Laboratory Program And Related Superfund
Activities, April, 1993
(36) Development of Performance Evaluation Samples for Field Analytical Screening Methods, Quality
Assurance Technical Support Laboratory, Jan, 1991
10
-------
IMPROVED LABORATORY PERFORMANCE THROUGH MATRIX-ANALYTE
SPECIFIC PROFICIENCY TESTING
M. Carusof Technical Director,and K.W. Jackson, Director, Environmental
Laboratory Approval Program, Wadsworth Center for Laboratories and
Research, New York State Department of Health, P.O. Box 509 Albany, New
York 12201-0509
ABSTRACT
The New York State Dept. of Health's Environmental Laboratory Approval
Program has been challenging its participating laboratories with
proficiency tests since 1985. A review of the proficiency test data
demonstrates a distinct improvement in laboratory performance, as measured
by percent recovery, for analytes that have not been included in EPA's WP
series. For example, the recovery for Non-potable Water phenol has
improved from an average of 3.3% in 1987 to 45.1% in 1991. A similar
improvement can be demonstrated for phenol in soil where first round
testing yielded 30.7 % recovery and 1992 test data demonstrates an average
recovery of 61.0%. The production of Solid Waste proficiency test
samples, and laboratory performance in analyzing these samples is
discussed. The data show clearly that the traditional practice of only
challenging laboratories with sealed ampules containing analyte
concentrates does not realistically assess their ability to analyze real
solid waste samples.
INTRODUCTION
New York State's Environmental Laboratory Approval Program (ELAP) was
established in the Fall of 1985 as a successor to and expansion of the
state's Safe Drinking Water Act laboratory certification program.
Initially, ELAP provided accreditation in the categories of Potable Water
and Non-potable Water to approximately 600 laboratories. As originally
structured, ELAP produced and distributed its own proficiency test samples
for both chemistry and bacteriology. ELAP has since grown and now
accredits over 900 laboratories in 30 other states and six foreign
countries. Each of the certification categories, Potable Water, Non-
potable Water, Solid and Hazardous Waste, and Air and Emissions, is
proficiency tested twice yearly. In 1992, the program manufactured and
distributed over 40,000 samples. These consisted of concentrates in
sealed ampules for the water chemistry analytes, full-volume water
bacteriology samples, air filter strips, and full-volume solid waste
samples. A list of the program's proficiency test samples is given in
Table 1.
The Solid and Hazardous Waste and Air and Emissions categories were added
to the existing water categories early in 1987. Due to the rapid expansion
of the program's breadth, ELAP initially accredited laboratories in the
Solid and Hazardous Waste (SHW) category by requiring them to have and
maintain accreditation for the same analytes in the Non-potable Water
11
-------
category. This was based on the theory that both the Non-potable Water and
Solid Waste methods are essentially the same. Thus accreditation for SHW
could be granted by a combination of Non-potable Waster proficiency
testing and on-site inspection of SHW preliminary extraction procedures.
However, this did not take into account the reduced recoveries and errors
that can be introduced during the additional extraction or matrix
reduction processes required with a real-world solid waste sample (such as
soil, sediment etc.). This paper will examine the improvement in
laboratory performance that is seen through the use of proficiency test
samples that closely mimic the real samples that are being analyzed
routinely.
PRODUCTION OF SHW PROFICIENCY TEST SAMPLES
In 1988 ELAP started producing it own SHW proficiency test samples
starting with metals in sludge. In general, production of other analyte-
matrix samples (listed in Table 1) follows this same procedure.
Initially a bulk sample of apparently analyte-free sludge, sediment, silt,
sand or soil is collected and oven-dried at 105° C overnight. This sample
is then ground, sieved to 100 mesh and roller mixed. Sample aliquots are
now randomly collected to determine background concentrations of the
analytes of interest. The remaining dried sample is then weighed and
sufficient appropriate solvent is added to create a slurry. Using a
paddle blade mixer, the analytes of interest are spiked into the slurry at
concentrations calculated to produce the selected dry-weight targets. The
slurry is then air or oven dried (depending on the solvent used) and
roller mixed. The dried powder is then dispensed into screw cap bottles.
De-ionized water is next added to each bottle to produce a wet real-world
consistency. The bottles are then capped and sealed with stretch tape.
Several samples are collected from the beginning, middle and end of
production and analyzed for product homogeneity. Additional samples are
collected randomly and analyzed over time to determine product stability.
DISCUSSION
A question frequently asked of any laboratory certification program is:
"How effective are you?". One way of demonstrating effectiveness is to
show improvements in recoveries and interlaboratory precision. A cursory
review of statistical summaries over the years demonstrates that
improvement has been made in recovery and precision, especially for
analytes not included in other performance testing schemes. However,
because of the rapid growth of the Program and change in the nature of
laboratories approved, from basically facility-types to basically full-
service commercials, the improved statistics may be due to changes in
laboratory type rather than laboratory ability. An in-depth study of the
causes of apparent improved performance appeared appropriate.
12
-------
The study protocol called for the selection of a set of laboratories (n ~
50) that participated in proficiency testing both in 1986 and the current
rounds. As demonstrated in Table 2, for the analysis of water proficiency
test ampules, the laboratories that analyzed for purgeable aromatics in
1986 had an average recovery of 97.1%. In the current round these
laboratories had an average recovery of 100%, a marginal improvement. It
is believed that since purgeable aromatics are commonly found in other
performance evaluation schemes in which the laboratories may have
participated, these laboratories were approaching their peak performance
in 1986. Alternatively, if one examines laboratory performance, as
measured by recovery, for phthalate esters, polynuclear aromatics and
priority pollutant phenols a distinct improvement can be seen. Again,
comparison was made using a set of laboratories participating in both the
1986 and current rounds. Average recoveries for phthalate esters,
polynuclear aromatics and phenols went from 10.7% to 91.4%, 6.5% to 79.1%
and 5.8% to 62.4% respectively.
The data presented in Table 3 reflect the improved performance of a set of
laboratories (n = 50) that initially participated in proficiency testing
for metals or phenol in solid waste. The table also presents the same
group's performance in an aqueous matrix for the same analytes. These
laboratories were only able to produce an average recovery of 55.4% for
cadmium, nickel and lead in a solid waste, while achieving an average
recovery of 99.5% for these same metals in an aqueous matrix. However,
through continuing proficiency testing the initial set of laboratories was
able to improve their recovery of metals from solid waste to a current
average of 95.8%. Similar improvement can be noted for phenol, with
recovery going from 30.7% to 61.0%.
As illustrated in Table 4, improved recoveries were not the only benefits
of continued matrix-specific proficiency testing. Interlaboratory
precision for cadmium, nickel and lead markedly improved from an average
relative standard deviation of 27.9% to 8.2%.
An additional factor to be considered is that several laboratories that
initially applied for certification in the SHW category withdrew as soon
as they were challenged with the full-volume samples. An obvious
conclusion is that they doubted their ability to handle these samples and
pass the test. However, if the program had continued to rely on sealed
ampule concentrates to test these laboratories, they would have presumably
continued to participate and been granted certification for Solid and
Hazardous Waste analysis.
13
-------
SUMMARY
The data Leave Little doubt that the traditional protocol of challenging
laboratories with proficiency test samples consisting of concentrates in
sealed ampules does not accurately assess their ability to analyze real-
world solid waste samples. Considerable resources are being provided for
the identification and clean-up of solid waste sites, and the additional
expense of research, development and production of realistic proficiency
test samples would seem to be easily justified.
14
-------
Table 1
NYS ELAP: CURRENT PROFICIENCY TEST SAMPLE TYPES
Potable Water
Drinking Water Bacteriology:
Total Coliform (& E.coli)
Standard Plate Count
Analytes represenitive of:
Chlorinated Acids
Trihalomethanes
Microextractables
SDWA Metals
Methylcarbamate Pesticides
SDWA Minerals
Organohalide Pesticides
Volatile Aromatics
Volatile Halocarbons
Miscellaneous:
Asbestos
Semi-volatiles
Non-potable Water
Wastewater Bacteriology:
Total coliform
Fecal coliform
Analytes represenitive of:
Chlorinated Hydrocarbon
Pesticides
Chlorinated Hydrocarbons
Atrazine and Carbaryl
Dioxins
Demand
Haloethers
Metals I:
Ba, Cd, Ca, Cr, Cu, Fe, Pb,
Mg, Mn, Ni, K, Ag and Na
Metals II:
Al, Sb, As, Be, Hg, Se, V
and Zn
Metals III:
Co, Mo, Tl and Ti
Nitroaromatics and Isophorone
Minerals
Miscellaneous:
B, CN, Phenols, MBAS, TOG,
pH and Oil & Grease
Nitrosoamines
Nutrients
Organophosphate Pesticides
Polynuclear Aromatics
Polychlorinated Biphenyls
Phthalate Esters
Priority Pollutant Phenols
Purgeable Aromatics
Purgeable Halocarbons
Residue
Solid and Hazardous Waste
Analytes represenitive of:
Chlorinated Hydrocarbon
Pesticides
Metals I:
Ba, Cd, Cr, Pb, Ni and Ag
Metals II:
Sb, As, Hg and Se
Polynuclear Aromatic Hydrocarbons
Polychlorinated Biphenyls
Phthalate Esters
Priority Pollutant Phenols
Miscellaneous: Asbestos
Air and Emissions
Asbestos
Fibers
Formaldehyde
Lead
Nitrate
15
-------
Table 2
IMPROVED RECOVERIES THROUGH PROFICIENCY TESTING
Analyte
Benzene
Chlorobenzene
Bis (2-ethyl-
hexyl)
phthalate
Benzo butyl
phthalate
Dimethyl
phthalate
Phenol
Pentachloro-
phenol
Benzo (b)
fluoranthene
Dibenzo (a,h)
anthracene
1986
Target
59.4
127.3
161.3
69.9
32.2
85.8
64.5
107.0
94.7
42.1
81.9
243.0
75.2
120.0
148.0
63.4
38.6
76.4
Mean
59.4
119.7
158.0
67.3
5.3
8.1
6.8
8.6
7.4
5.0
3.54
5.3
7.0
9.0
4.7
4.0
3.5
8.1
% Rec.
100.0
94.0
98.0
96.3
16.5
9.4
10.5
8.0
7.8
11.9
4.3
2.2
9.2
7.5
3.2
6.3
9.0
10.6
Current
Target
31.2
50.2
29.8
47.7
33.5
83.8
28.4
68.7
59.9
75.2
105.0
75.1
79.6
99.5
44.1
118.0
30.3
48.6
Mean
31.4
49.9
30.1
47.5
25.9
67.3
28.4
69.3
56.9
71.5
47.7
33.7
64.0
78.6
43.09
97.3
20.7
32.0
% Rec.
100.5
99.4
100.9
99.5
77.3
80.3
100.0
101.0
95.0
95.0
45.4
44.8
80.4
79.0
99.5
82.5
68.3
65.9
16
-------
Table 3
MATRIX-SPECIFIC RECOVERY IMPROVEMENT
Analyte
Cadmium
Water
Solids
Nickel
Water
Solids
Lead
Water
Solids
Phenol
Solids
1988
Target
33.0
5.0
170.0
40. 0
220.0
113.0
183.0
Mean
33.2
3.6
168.8
17.6
217.1
56.2
56.2
% Rec.
100.6.
72.6
99.3
43.9
98.7
49.8
30.7
Current
Target
70.0
60.0
400.0
401.0
250.0
184.0
333.0
Mean
68.3
56.2
394.7
403.3
246.9
171.4
203.0
% Rec.
97.0
93.7
98.8
100.6
98.8
93.2
61.0
Table 4
INTERLABORATORY PRECISION IMPROVEMENT
Analyte
Lead
Nickel
Cadmium
Mean
56.2
17.6
3.6
1988
8
9.8
5.8
1.2
RSD
17.4
32.9
33.3
Current
Mean
314.0
398.0
56.2
6
28.1
28.8
4.7
RSD
8.9
7.2
8.4
17
-------
COMMERCIAL LABORATORY POSITION ON NATIONAL LABORATORY ACCREDITATION
John E. Farrell. Ill Vice President, Enseco, a division of Corning
Laboratory Services Inc., 2200 Cottontail Lane, Somerset, New Jersey
08873; Jack R. Hall, Technical Director, IT Corporation, 9000 Executive
Park Drive, Suite A110, Knoxville, Tennessee 37923; Barbara R. Hill,
Director of Administration, WMX Environmental Monitoring, 2100
Cleanwater Drive, Geneva, Illinois 60134; Marlene 0. Moore, President,
Advanced Systems, Inc., Post Office Box 8090, Newark, Delaware 19714;
Jerry L. Parr, Director of Quality Assurance and Technology, Enseco,
Rocky Mountain Analytical Laboratory, 4955 Yarrow Street, Arvada,
Colorado 80002; James Ploscyca, Director of Quality, Industrial
Environmental Analysts, Post Office Box 12846, Research Triangle Park,
North Carolina 27709; Peggy Sleevi, Director of Quality Assurance,
Enseco, 2612 Olde Stone Road, Midlothian, VA 23113; David N. Speis,
Director Quality Assurance and Technology, Environmental Testing and
Certification Corporation, 284 Raritan Center Parkway, Edison, New
Jersey 08818; E. Scott Tucker, Director of Analytical Laboratory,
Clemson Technical Center, 100 Technology Drive, Anderson, South
Carolina 29625; Craig 0. Vinson, Senior Laboratory Manager, CH2M Hill,
2567 Fairlane Drive, Montgomery, Alabama 36123
ABSTRACT
Environmental testing laboratories conduct sample analysis on soil,
air, water, and other materials that may contain potentially hazardous
substances. Services of the environmental testing industry are widely
used by the U. S. Government, state governments, and private industry
because almost every environmental decision is based on laboratory
data. Reliable data are critical to protect the public health and the
environment.
While the primary need for the data generated by these laboratories is
to comply with federal environmental regulations, there is no national
accreditation process to ensure uniform, rigorous standards.
Environmental laboratories currently are accredited under a system of
multiple state and private programs with varying quality assurance
criteria and excessive processing costs. Implementation of a national
accreditation program is a cost-effective method of ensuring data
quality.
International Association of Environmental Testing Laboratories, Inc.
(IAETL) has assembled a committee of its membership to discuss the
effect and feasibility of a national program. IAETL represents over
150 commercial laboratories with an estimated 45 to 50 percent of the
18
-------
commercial laboratory capacity. IAETL has adopted a position in
support of national certification. IAETL1s committee has reviewed
various scenarios and implications of a national program including the
recommendations of the Committee on National Accreditation of
Environmental Laboratories (CNEAL) Report to EPA. lAETL's position
supports the CNEAL recommendations and emphasizes the adoption of
certain critical components: data assessment, onsite assessment,
performance evaluation testing, a clearly defined process for both
laboratories and accrediting organizations, and uniform, well-defined
criteria.
The objective of our presentation is to share information on the
implications of a national program and the specific recommendations we
have developed on standards, organization, and enforcement to ensure
that the national program will be effective.
INTRODUCTION
In late 1989, the International Association of Environmental Testing
Laboratories (IAETL) recognized the critical need for national
accreditation. The unwieldy, expensive, and time-consuming
accreditation processes currently in place were adding considerable
cost to environmental lab testing with little or no benefit to the
users of laboratory services. IAETL formed a National Accreditation
Committee to address the issues.
The National Accreditation Committee reviewed current federal programs,
specifically the CLP program, USATHAMA, the Navy HAZWRAP, A2LA, AIHA,
DOE, DOD, and the Drinking Water program. Major categories common to
the programs were determined and a questionnaire was developed to
survey current state programs.
At this point, mailings on the subject revealed a wider interest group
than just environmental labs for a national accreditation program.
Customers of the environmental laboratories such as engineering firms,
various industries, and regulatory bodies also had a vested interest
because of concerns about environmental data quality and prices. A new
coalition was formed from the various interest groups called NELAC -
the National Environmental Laboratory Accreditation Coalition.
NELAC used information from its participants along with the previously
gathered information on federal and state programs on standard
environmental laboratory practices and accreditation programs
structures to formulate four options to be considered as possible
models for a national accreditation program. After thoroughly
examining and debating the various strengths and weaknesses of the four
models, NELAC presented a single, comprehensive option.
19
-------
The advantages of the national accreditation program as described in
the option developed by the IAETL/NELAC groups include:
• Self-sustaining structure
• Built-in controls to ensure continuing quality and uniformity of
the program.
• Structures already in place (state agencies) may be used in the
process.
• A mechanism for making necessary changes through the Advisory
Board exists.
• Regulatory control is retained.
• Third-party administration is provided for.
• Funding requirements from regulatory agencies is minimized.
• States would be reimbursed for their participation.
• Higher standards of quality could be applied consistently to
environmental data across the country.
• Duplication of efforts in accreditation could be minimized,
particularly for audits, performance evaluation (PE) samples,
and applications.
• QA requirements could be centralized within a structure similar
to existing programs within the EPA. There would be a dynamic
mechanism of input into these requirements through the Advisory
Board.
• The EPA can influence states to adopt reciprocity as part of the
acceptance criteria and through funding.
The key elements of this option were the accreditation process, the
accreditation criteria, data assessment, onsite assessment, performance
evaluation testing, accountability, and enforcement.
20
-------
ACCREDITATION PROCESS
The accreditation process concurred upon can be diagramed as follows:
EPA FEDERAL PROGRAM ADMINISTRATION
ADVISORY BOARD
ACCREDITATION COUNCIL
ACCREDITORS
ASSESSORS
ACCREDITED ENVIRONMENTAL LABORATORIES
The EPA FEDERAL PROGRAM ADMINISTRATION is managed by a high-level EPA
official who administers the Accreditation Council contracts.
The ADVISORY BOARD approves standards. Initially it would be made up
of individuals from federal and state agencies, laboratories and data
users appointed for two years from a volunteer list. In subsequent
years, board members would be elected from nominated individuals.
The ACCREDITATION COUNCIL is an EPA contractor hired with a minimum
contract of 3 to 5 years to administer the program, approve accrediting
agencies, and maintain records.
The ACCREDITORS are state agencies, independent third-party
organizations, and federal agencies that would accredit per the
national program criteria. This would allow existing infrastructure in
these organizations to remain intact and functional if they so desire.
The ASSESSORS are the trained individuals hired by the accreditors and
required to undergo competency testing to perform data and site
assessments.
21
-------
ACCREDITATION CRITERIA
The program must consist of clearly established, uniform criteria. The
laboratories must have detailed requirements that they are expected to
meet to be accredited. Using these criteria, the laboratory may design
the appropriate systems and the internal assessments to maintain these
systems. The criteria must be clearly defined so that they can be
applied uniformly from location to location by any accreditor. The
accreditation council, the accreditors, and the assessors also must
have uniform criteria by which the program is administered and by which
accreditors and assessors are selected, trained, perform laboratory
assessments, and document the accreditation process.
Accreditation Criteria for Laboratories
Uniform criteria for the laboratory will enhance a laboratory's ability
to generate data of known and documented quality. A system must be
established by which laboratories can be evaluated to determine if they
are capable of performing competently and that their QA/QC programs are
functioning to ensure the production of reliable and usable data. To
this end, the criteria for evaluation of laboratories must address the
following.
Organizational Structure and Staff
The organizational structure must be defined such that the technical
functions can be performed adequately. The structure must not put
undue pressure on staff to influence their judgment. The person
responsible for Quality Assurance (QA) must report to senior
management, and the QA role should be independent from the data
generation process.
Staff must have adequate education, training, and experience for their
assigned functions. Laboratory management should not rely on education
and experience alone to ensure that technical assignments are completed
properly. Therefore, the laboratory's training program must be
comprehensive and documented. All employees should have a clear
understanding of their job functions. Adequate supervision should be
provided to ensure that work is properly completed.
Quality Systems
The criteria for accreditation should dictate the minimum requirements
for documenting an internal quality system appropriate to the size of
the facility and the scope of the analyses performed. The quality
system should be well documented so that all staff understand its
function and requirements. The accreditation program should provide
enough direction to the laboratory to generate quality documents that
22
-------
describe the management's policies and values, the laboratory's
organization structure, responsibilities of management and staff,
general QA procedures, detailed QC requirements, and corrective action
processes for the systems described.
Facilities and Equipment
Minimum criteria for the physical plant are delineated to ensure that
the environment does not adversely affect test results, that adequate
security is available, that appropriate equipment is available and
employed for the offered testing, and that appropriate health and
safety programs are implemented.
Data Generation and Record Keeping
The accreditation criteria must address minimum requirements for
documentation of the procedures used in the laboratory by requiring
standard operating procedures, sufficient documentation of the
analytical process to reconstruct the events, and adequate procedures
for storage of these data for an appropriate period.
Accreditation Criteria for the Accrediting Council and Accreditors
The accrediting council, the accreditors, and the assessors must meet
criteria similar to those as described above. The accreditation
council (the administrative arm of the program/directed by the
advisory board) and the accreditors (agencies providing accreditation
under the direction of the accrediting council) must be clearly
defined. The organizational structures must be known and provide for
an environment in which the staff are free from compromising
influences. Assessors (those people performing the data and site
assessments hired by the accreditors) must have adequate education,
training, and experience for their assigned functions.
Additionally, adequate training must be provided and documented to
ensure that the assessors are competent to assess the laboratory and
can provide uniform assessments to all laboratories that they visit.
Each of these groups should abide by a documented quality program,
including corrective action processes. The process by which the
laboratory assessments take place, the reporting mechanism, and the
maintenance of these records must be clearly defined.
DATA ASSESSMENT
Data assessment must be conducted as part of the quality assurance
review of the laboratory operation. Data assessment can be conducted
in two parts: (1) determination of the data acceptability relative to
minimum data quality criteria established for nationally accredited
laboratories and (2) determination of the data availability at the
laboratory's site.
23
-------
Data can be reviewed in advance of the onsite visit to determine
problem areas that need to be investigated during the visit. The
periodic onsite review process will determine if the laboratory's data
assessment process includes data handling procedures, control criteria,
and policies for data corrective actions.
The data assessment criteria should be defined clearly by the Advisory
Board. These criteria will ensure that all data generated by a
nationally accredited laboratory will meet a minimum data quality
standard. Project- or regulatory-specific criteria still will be
needed for more stringent data requirements; however, a national
program will set minimum standards for data quality criteria.
Current data quality often is based on the contents of the data
deliverable package. For example, data deliverables presented in CLP
format are assumed to be of better quality than drinking water data
presented as a single reported concentration. The data user assumes
that quality control parameters for an application are within
acceptable limits if the data are presented in CLP data deliverable
format. The drinking water data assessment criteria are clearly stated
for every primary drinking water parameter, and each must be met before
reporting any data. The CLP data deliverables include documented data
assessment criteria as part of the report. The question of data
quality is not answered by how much of the data are presented and
verified in the report; the question of data quality is answered only
when the data are found to meet the acceptance criteria.
All data from a laboratory must be documented as to its quality. The
data assessment verifies that data both meet the minimum data criteria
and are well documented.
ONSITE ASSESSMENT
A single national program for accreditation of environmental
laboratories must include the key element of an independent, onsite
assessment as part of the accreditation process. This assessment,
along with the data assessment described previously, should form the
foundation to grant (or deny) accreditation to a laboratory. We have
used the word "assess" to connote that the primary function of the
assessment is to observe and evaluate, not to correct deficiencies or
make recommendations.
The onsite assessment should address all aspects of the facility,
staff, and systems to provide confidence that the laboratory is capable
of providing the services for which accreditation is requested.
24
-------
After the initial assessment, the accreditor should prepare a report
that lists the findings and documents any deficiencies that must be
corrected for the laboratory to obtain accreditation. The laboratory
should be provided with an opportunity to remedy any deficiencies. If
the assessment shows that accreditation is justified, the accrediting
council then should issue a certificate authorizing the laboratory to
provide data for the parameters included in the scope of accreditation.
This information should be maintained by the accrediting council for
dissemination to users of laboratory services. The accrediting council
should establish a procedure by which laboratories can seek
accreditation for additional tests.
The accreditation should be valid for 1 year. An annual reassessment
must be performed by the accreditor that should address any changes in
the laboratory based on information retrieval and the results on PE
samples. Where concern is warranted, an onsite assessment may be
performed. In any event, additional onsite assessments should be
performed at least once every 2 years.
The assessment should address all aspects of the laboratory's facility,
organization, quality systems, and record keeping. A critical element
of this process is to determine if the laboratory personnel have the
necessary education, training, technical knowledge, and experience to
perform their assigned functions. It is the responsibility of
laboratory management to ensure that laboratory personnel meet this
requirement. A separate accreditation of analysts would not improve
the process or add to the assurances already established by requiring
comprehensive in-house training and documentation programs. The
program could establish criteria relative to the expected education,
experience, and knowledge for analysts.
Assessors must have a thorough knowledge of the accreditation elements,
including detailed knowledge of the analytical procedures, and must act
in an impartial manner. The accreditation system must establish
procedures for ensuring the competency of assessors.
PERFORMANCE EVALUATION TESTING
Performance Evaluation or Proficiency Testing is a key element of the
program in confirming the laboratory's capability to perform the
methodology and identifying areas for improvement. A performance
evaluation program is therefore a critical requirement of national
laboratory accreditation.
25
-------
The accreditation program will require completion of an initial series
of PE samples representing the parameters and matrices in which the
laboratory is seeking accreditation. The results of the PE samples
will be compared to the certified values and the expected acceptable
range for that parameter and matrix.
Because PE samples are not available for all parameters in all matrix
types, the PE program must have a cross section of PE samples that
covers representative parameters. For example, a volatiles PE may
contain 5 to 8 compounds to represent the 20 to 30 volatiles normally
analyzed in wastewater methods. Matrices evaluation may be limited to
water, air, soil, and sludges for specific inorganics and organics,
with other matrices available for only select items (PCBs in oil or
mercury in fish for example). The areas of matrix accreditation must
be clear to all participants and potential users of the national
accreditation program. If, for instance, a PE for dioxins in fish does
not exist, does the program accredit a laboratory for dioxins in fish
if the laboratory is accredited in dioxins in water and soil; or should
the approach simply be that unless covered by PEs, particular
parameters in matrices will not be accredited.
The fee structure will depend on the type and number of PEs required by
the laboratory. As new PEs or PEs in different matrices become
available, previously accredited laboratories as well as new
applicants will be encouraged to participate in these studies.
Only laboratories generating acceptable data will be accredited for a
particular parameter or set of parameters for a matrix type by the
method certified by the laboratory. For a laboratory failing a
particular PE, the accrediting council will work with the laboratory to
evaluate the program areas and needed changes. Laboratories, at their
expense, have up to three chances in a 1-year period to perform a
single PE successfully. After initial accreditation approval, the
laboratory will be required to perform PEs annually to maintain
accreditation for the designated PE parameter(s) in the matrix
provided. Failing multiple PEs (two or more) by a laboratory will
initiate a reassessment by the accrediting council.
A well-organized, broad-based PE program will be the heart of the
national accreditation program to ensure continued proficiency of the
member laboratories.
26
-------
ACCOUNTABILITY AND ENFORCEMENT
A critical element of any accreditation program is accountability on
the part of the accreditors as well as the laboratories being
accredited. In addition, the accreditors must be in a position to
enforce any and all accreditation and/or revocation of accreditation
decisions that they make.
The accreditors will evaluate laboratories through a combination of
onsite assessments, data assessments, and performance on proficiency
test samples submitted to the laboratories. The onsite and data
assessment activities will be used to grant or revoke accreditation for
the laboratory. The accreditation must be specific for the areas of
testing performed by the laboratory, to incorporate both parameters and
matrix. As discussed previously, unacceptable performance in analysis
of PE samples can result in loss of accreditation for a particular
parameter. In this context, the accrediting council must be held
accountable for developing the following:
• The accrediting council is accountable for providing
requirements and standards for conducting onsite and data
assessments. The accrediting council also is responsible for
establishing a viable performance evaluation testing program.
• Once assessors are selected, the performance of each assessor
must be evaluated annually, at a minimum, and allow for more
frequent evaluation if their performance is questioned at any
time. This review should be based on a combination of their
accreditation record and evaluations from laboratories.
• A hearing process must be establishedstoMresolve disputes about
assessor performance and laboratory performance issues. This
process would provide all parties involved the opportunity to
express their points of view before 'affinal determination is
made.
• Assessors who have demonstrated unacceptable performance must
not be allowed to continue in that capacity once a determination
has been made. The accrediting council would maintain an active
list of acceptable accreditors and assessors.
• Laboratories that have demonstrated unacceptable performance
must not be allowed to maintain an accredited status. The
accrediting council would maintain an active list of accredited
laboratories, along with the areas of accreditation.
Enforcement of decisions about acceptable laboratories, accreditors,
and assessors would be controlled through the maintenance of the
approved listings. A real-time listing of approved assessors and
accredited laboratories could be maintained on a computerized bulletin
board system.
27
-------
CONCLUSION
Implementing a national accreditation program is a cost-effective
vehicle for improving and maintaining data quality. Such a national
program will help ensure that decision makers and the public receive
accurate and reliable data to make cost-effective environmental
decisions. Users of environmental testing services need an independent
and comprehensive verification that laboratories have the capability to
perform competently and that laboratory Quality Assurance/Quality
Control programs are functioning to ensure the production of reliable
and usable data. It is critically important that environmental
monitoring data are reliable, reproducible, accurate, and defensible to
support regulatory compliance and enforcement decisions made by the
public and private sector.
A single, comprehensive national program
environmental laboratories must establish:
for accreditation of
1. Uniform, rigorous requirements for environmental testing
laboratories to generate data of known and documented quality.
2. Uniform, rigorous standards under which
organizations can oversee the program.
accrediting
The accreditation process should be administered by a private/public
partnership with federal oversight. Procedures would be developed by a
governing board of representatives from the EPA and other interested
federal parties, state agencies, the laboratory community, and users of
environmental data.
A program designed and administered with these features
costly, redundant requirements while significantly
accuracy, reliability, and timeliness of decisions that affect the
quality of environmental cleanup and restoration efforts.
would eliminate
increasing the
IAETL supports and continues to work toward a single, comprehensive
national program that will provide an effective means of improving the
consistent generation of high quality data to comply with environmental
laws and to protect public health and safety.
REFERENCES
Jeanne Hankins, Final Report of the Committee on National Accreditation
of Environmental Laboratories. United States Environmental Protection
Agency, Washington, DC September, 1992.
28
-------
COST EFFECTIVE PROJECT PLANNING: AN ALTERNATIVE APPROACH
TO INITIAL RCRA FACILITY INVESTIGATIONS
D. Loring. Principal, Loring Environmental Associates, 10 Graham Terrace, West Roxbury,
Massachusetts 02132; and
S. Chapnick, Program Manager, Gradient Corporation, 44 Brattle Street, Cambridge,
Massachusetts 02138.
ABSTRACT
In a recent report by the General Accounting Office, the RCRA program is criticized for being
"excessively slow in nearly all areas, and charges EPA with failing to develop a viable RCRA
cleanup strategy."1 It is the conduct of the large-scale multimedia RCRA Facility Investigations
(RFIs) which this paper addresses. We propose an approach that combines stages in the RFI
process and redefines the current phased approach to RFIs. This alternative approach can reduce
the time and cost associated with the various RFI process stages and improve the responsiveness
of determining appropriate corrective measures.
Project planning is an essential part of designing large-scale investigative programs. Although
the initial planing process can be quite time consuming, it is vital in achieving data that are both
scientifically sound and legally defensible to support corrective action decisions. However, many
investigative programs have been delayed for years in the planning and negotiation stages. This
delay is inefficient, costly, and can result in a continuing risk to public health and the
environment.
A key element in the proposed approach is a Phase I scope that is defined through an iterative
process with the regulatory agency, owner, and engineering and laboratory contractors. The
scope is defined based on site-specific information (e.g., historical chemical and physical data,
hydrogeological information), applicable and relevant or appropriate requirements (ARARs),
health based criteria and other toxicological information. A site-specific sampling scheme and
target compound list is defined. For example, if groundwater is not a relevant exposure pathway
at the site, then groundwater samples would not need to be collected in the initial phase.
Recommendations are made to base methodology on both health-based criteria and background
levels of contaminants of potential concern. This approach allows for the controlled use of
alternative methodologies (i.e., field techniques) that may have higher detection limits than
standard methods but are acceptable to define the site conditions in reference to risk calculations.
The initial object of the process would be to reduce risk, not to return the site to a "pristine"
level.
This approach for RCRA sites is similar, in goals, to the Superfund Accelerated Cleanup Model
(SACM). The use of a simpler approach and a narrower focus to the Phase I RFI allows for
timelier risk reduction for people and the environment.
29
-------
INTRODUCTION
RFI guidance defines the RCRA Corrective Action Process as having numerous stages.
Generally, the approach to RCRA Facility Investigations (RFI) has included a large-scale
sampling and analysis program (as Phase I of a phased strategy) encompassing a broad range of
target compounds such as the full Appendix IX. As a result of the magnitude of the Phase I
investigation (in terms of number of parameters, analytical methods, detection limits, etc.), the
program details are so voluminous that the planning process can take years. Additionally, once
the RFI is underway, several other stages in the RCRA process follow, including: review of the
results by the regulatory agency based on health and environmental assessments;
recommendations for interim corrective measures, and/or a Corrective Measures Study (CMS);
implementation of the CMS; evaluation of the CMS; recommendations of appropriate corrective
measures (actions), and Corrective Measures Implementation(CMI).
In a recent report by the General Accounting Office, the RCRA program is criticized for being
"excessively slow in nearly all areas, and charges EPA with failing to develop a viable RCRA
cleanup strategy."1
It is the conduct of the large-scale multimedia RCRA Facility Investigations (RFIs) which this
paper addresses. We propose an approach that combines stages in the RFI process and redefines
the current phased approach to RFIs. This alternative approach can reduce the time and cost
associated with the various RFI process stages and improve the responsiveness of determining
appropriate corrective measures.
OVERVIEW OF THE RCRA CORRECTIVE ACTION PROCESS
The RFI is carried out under either a permit or enforcement order. RFIs can range from limited
investigations to large-scale investigations for a wide variety of target analytes and matrices. The
RCRA Corrective Action Process has numerous stages. A RCRA Facility Assessment (RFA)
should include a review of historical information at the site and a site inspection by the EPA, to
identify solid waste management units (SWMUs), collect existing information on contaminant
releases, and identify suspected releases needing further investigation. Depending on the results
of the RFA, it is possible to either proceed with a RCRA Facility Investigation (RFI) or, in cases
where public health or the environment is threatened, interim corrective measures may be
required.
The owner or operator then performs the RFI to either verify the contaminant release, if required,
or to characterize the nature, extent, and rate of migration of potential future releases. The
results of this stage of the process are submitted to the regulatory agency for review. Further
steps in the RCRA corrective action process include evaluation of the data generated from the
RFI against established health and environmental criteria. At this stage, if further action is
determined necessary, a corrective measures study (CMS) is performed which recommends
appropriate measures to correct the release. Interim Corrective measures can be taken either
prior to or during the CMS if a threat to health or the environment is identified during the RFI
stage.
30
-------
Finally, subsequent to the CMS, corrective measures are implemented (CMI). This includes
design, construction, operation, and monitoring of the corrective measures decided upon by the-
regulatory agency.
ITERATIVE PROCESS IN DESIGNING THE PHASE 1 SCOPE
Evaluation of a RCRA site is extremely complex, and requires the application of a wide variety
of scientific disciplines, and diverse technical and regulatory knowledge. During the course of
an RFI, many individual contractors are hired to participate in the various phases of the project.
The contractors can include engineers, field samplers, hydrogeologists, laboratory personnel,
toxicologists, risk assessment specialists, and quality assurance consultants.
However, during the initial planning stages in which the Sampling and Analysis Plan (SAP) and
Quality Assurance Project Plan (QAPP) are developed, the expertise of a number of these
individuals is not utilized. For many programs, there is minimal communication between
members of the "team." It is crucial to the success and expediency of the RFI process to
assemble the entire project team to develop the work plan for the specific site that takes into
account all of the necessary information required to generate valid data upon which to support
corrective action decisions.
Project team members with risk assessment or toxicological expertise can influence the scope of
the RI at the planning stage in order to avoid the low-level analysis of certain analytes that do not
pose a threat to human health or in order to ensure low-level detection limits in cases of
ecological risk. Additionally, these team members will help define the possible exposure
pathways and media so that the SAP can be formulated to minimize the numbers and types of
samples being collected and analyzed while maximizing the usable data generated. Also, this
planning process can avoid costly resampling efforts due to omission of a significant media (e.g.,
air) or parameter type. At this stage, human-health based risk calculations can be performed
using historical data, if available, as an estimate of the potential risks at the site. Health-based
cleanup levels can also be estimated from these risk calculations to help clearly define the
detection levels needed for the analytes of concern specific to the site.
Both the EPA and the site owner are in a position to require and facilitate working meetings
during which an appropriate target analyte list is developed and associated required method
detection limit are determined. This requires a commitment from both the regulatory agency
and the site owner. From this information, methodology choice (in particular field vs. laboratory
methods), data quality objectives, and a relevant sampling scope can be developed. Only from
these criteria can a valid SAP and QAPP be developed. In terms of cost-effectiveness and
timeliness it is crucial to evaluate field analytical methods (e.g., field gas chromatographic
methods for analysis of volatile organic compounds, X-Ray Fluorescence techniques for the field
analysis of metals, bioassay techniques for the field analysis of PCBs) as viable alternatives to
laboratory analyses. Field method technology has become more sophisticated and rigorous and
allows for the rapid analysis of constituents of concern on-site. In addition, laboratories can
develop fast, cost-effective methodology when presented with a site-specific TAL and associated
detection limits (e.g., "Soxtet" extraction, Solid Phase Extraction, GC-MS-SIM, GC-ITD).
Therefore, decisions that are based on chemical information can be made more quickly - speeding
up the RFI process.
31
-------
The current RFI guidelines are certainly not in conflict with this recommended process. In fact,
some of these activities are required. However, from experience (in both authoring and
reviewing numerous Work Plans, SAPs, and QAPPs in various EPA Regions and states) it is
evident that an organized, effective planning process rarely occurs. Much of the information
which does exist about the site history, physical site conditions, and historical data is not taken
into account during the planning phase or, if it is included in the Work Plan, is not included as
relevant information upon which to define specific QA/QC criteria in the QAPP This results
in a program where the target analyte list and selected methodologies are inconsistent with the
site history and expected contaminants and physical conditions at the site.
Coordination between the project team members needs to be maintained to generate a SAP and
QAPP that the field personnel and laboratory personnel can follow as a protocol and that is
sufficient to support usable data. However, the SAP and QAPP are often done as "boiler plate"
documents with only minor changes due to site-specific information. This results in a waste of
both cost and time when, for example, strict QA/QC criteria are required that are not relevant
to the type of samples being collected or the expected chemical contamination. For example, a
QAPP for a site in Louisiana required CLP-like quality control, DQOs, and detection limits for
sludge samples that were known to be highly contaminated with organics and metals. The
laboratory could not meet most of the QAPP required matrix QC or the detection limits stated
because of the highly complex and contaminated matrix. The complex matrix and the level of
contamination expected should have been taken into account during the generation of the QAPP
Less strict QC would not have compromised the usability of the data generated for this site due
to the highly contaminated nature of the samples — in other words, the relative extent of
contamination and the order-of-magnitude were the necessary information. Lack of coordination
between project team members caused an increase in cost (laboratories charge more for CLP-like
QC than for more standard analyses) and time (CLP-like QC required more time to analyze and
generate laboratory reports).
GUIDELINES FOR PHASE I OF RFI
The following steps are presented as guidelines for the coordinated project team approach in
generation of the project QAPP, with specific emphasis on the usability of the chemical data.
Similar team approaches would be required in generating the SAP and work plan and would bring
together the engineering and field personnel with the laboratory and toxicologists and risk
assessment specialists:
1. Determine Target Analyte List (TAL) based on Historical Site Activities & Data
The traditional approach in planning an RFI has been to analyze for the full Appendix IX list of
compounds. However, many of the analytes included in Appendix IX may not be appropriate
given the historical information which exists about operations at a particular facility and the
results of previous sampling and analysis. The project team members involved in the initial
planning need to have the site-specific information necessary to define the scope in terms of the
target analyte list. A site-specific target analyte list must then be developed using this
information. The expertise of both risk assessment and analytical chemistry personnel should be
involved in defining the site-specific TAL. Without this information and coordination of project
team members, the result may be inflated sampling and laboratory costs, and also higher costs
32
-------
to validate data which is not of any particular use in the final determination of required corrective
measures. In the Superfund Accelerated Cleanup Model (SACM), this problem is addressed:
"Many, if not most of these assessments start from scratch, — they do
not necessarily take into consideration the information and data generated
by the studies that preceded them. This happens not only because of the
obvious financial incentives to the contractor community and the human
inclination to distrust the work of others, but because each part of the
program is gathering data to respond to its particular perceived need."2
In some cases, the directives to produce redundant data come directly from the regulatory
authorities in an effort to restore the.site to "pristine" conditions. However, the focus must shift
from this effort to an effort of "worst first"3 as described in the SACM model.
2. Define the Sampling and Analysis Procedures Based on Site Conditions, Historical
Information, and Exposure Pathways
With the input of the engineering and risk assessment personnel, project costs and time-frames
may be reduced, for example, by deciding that a portion of the soil and/or groundwater samples
may be analyzed using field rather than laboratory methods. This would allow for less strict
QA/QC, rapid generation of chemical results, and contribute to a more timely progression to the
corrective measures process.
Often, the project personnel writing the QAPP are given instructions to "write a QAPP for the
full Appendix IX list for water and soil samples." Although, for example, hydrogeological
information about the site is used in defining the sampling plan, samples collected and analyzed
from a crucial part of the site (e.g., a known hot spot or chemical release) are not treated
differently than others in terms of defining DQOs, methods, quality control criteria, detection
limits, and validation requirements. The same DQOs and detection limits are generally applied
to all samples in this type of generic QAPP. By changing this approach, by generating a QAPP
specific to the real needs of the program, usable data will be generated in a more cost-efficient
approach.
3. Review SAP, QAPP, and Target Analyte List in Terms of Both Ecological and Human-
Health Risk Potential
Once the site-specific TAL is identified for the site, risk assessment estimates (preliminary
calculations based upon existing data from the site) must be taken into account to choose
appropriate detection limits required for each of the compounds on the list.
Volumes of historical data on which risk calculations may be based exist for most sites which are
in the RCRA corrective action process. However, risk assessment personnel are usually not
brought into the RFI process until after the Work Plan, SAP, and QAPP have been developed,
and after the analyses have been performed. The interactive process of the project team, to make
decisions on detection limit requirements and methodologies, must occur prior to the development
of these site documents that direct the activities for the RFI.
33
-------
The input from risk assessment project team members is crucial in choosing appropriate
methodologies, both conventional and alternative. This information is vital to adequately
determine required quantitation levels and DQOs specific to the project. For example, recent
RODs (1991) in both Region 1 and Region 5 have required targeted cleanup levels of 1 M§/L for
vinyl chloride and 1,000 ng/L for toluene.4 However, most QAPPs require a detection limit of
10 fig/L for both compounds and identical DQOs and QC criteria for both compounds. In this
scenario, the toluene results were generated by the laboratory under far more stringent QC
requirements than necessary to support the usability of the data in a cleanup level model.
Conversely, any not detected vinyl chloride results generated at the 10 /xg/L detection limit are
not usable because these nondetects are reported at an order-of-magnitude higher level than the
level of concern (cleanup level) for this compound.
If, at the planning stage, it is known that both toluene and vinyl chloride expected contaminants
at the site, and that toluene is potentially highly contaminated (i.e., in the ppm range) as
compared with vinyl chloride (i.e., in the low ppb range), methodology can be selected which
will give relevant reporting limits for both compounds based on the targeted cleanup levels, as
listed above.
However, if the historical information about toluene and vinyl chloride at the site were not
reviewed, and the planning not performed as a project team process, the routine approach would
be to define QAPP requirements to analyze for the Appendix IX Volatiles list by Method 8240.
This would result in not detected vinyl chloride results reported at a high sample detection limit
due to dilutions that would be required for high levels of toluene. With appropriate planning,
incorporating the expertise of an analytical chemist and risk assessment personnel, a GC
screening method for toluene could be defined for the project and a GC/MS method with strict
QC (e.g., EPA method 524.2) could be defined for vinyl chloride so that the low-level detection
limit needed for usability of the data could be achieved.
4. Establish A "Hot" List for Interim Corrective Measures
Interim Corrective Measures may be necessary at any point in the RFI process. During the RFI,
sources can be identified as hot-spots as those regions of the site with the highest concentrations
of contaminants. Once these hot spots have been identified, they should be evaluated against
health-based criteria or cleanup level criteria, as appropriate. Field screening analytical
techniques can be extremely useful in identification of hot spots and extent of contamination for
rapid response. Interim corrective measures are appropriate at this stage.
SUMMARY
One of the crucial differences in the strategy which is recommended is evaluation of the historical
data against both health-based criteria and estimated (or historical) cleanup goals at the initial
planning stages of the RFI and prior to the formulation of the Work Plan, SAP, and QAPP. This
evaluation should be the basis from which the target analyte list (TAL) is identified, the Sampling
and Analysis Plan (SAP) and Quality Assurance Project Plan (QAPP) are written, and the basis
from which the Data Quality Objectives (DQOs), including target quantitation limits and
methodologies, are chosen. Without this step, many of the current RFI activities are superfluous
in that methodology which is costly and time consuming is being used for analytes which are not
34
-------
harmful to human health at the low detection levels required by the methods or not present at the
site being investigated. This strains the budgets and timetables and results in taking resources
away from remediation.
EPA is presently in the process of establishing requirements for RCRA corrective action. These
rules are to address various parts of the RCRA corrective action process that are presently either
not defined or subject to misinterpretation in implementation. These regulations will be published
as Subpart S of 40 CFR Part 264 and are intended to address the following issues:
• Action levels based on health and environmental factors
• Methods for development of action levels
• Specific action levels for soil and surface water
• Adjustment of action levels based on additive toxicity
• Influence of quantitation limits on action levels
• Definition of "constituent"
• Definition of Solid Waste Management Unit (SWMU)
• Notification and reporting
• Semantics
The approach proposed in this paper was intended to address the issue of health-based action
levels by involving risk and toxicology personnel up-front in the planning process. A key
element in the proposed approach is a Phase I scope that is defined through an iterative process
with the regulatory agency, owner, and engineering and laboratory contractors. The scope is
defined based on site-specific information (e.g., historical chemical and physical data,
hydrogeological information), applicable and relevant or appropriate requirements (ARARs),
health-based criteria and other toxicological information. A site-specific sampling scheme and
target compound list is defined. For example, if groundwater is not a relevant exposure pathway
at the site, then groundwater samples would not need to be collected in the initial phase.
Recommendations are made to base methodology on both health-based criteria and background
levels of contaminants of potential concern. This approach allows for the controlled use of
alternative methodologies (i.e., field techniques) that may have higher detection limits than
standard methods but are acceptable to define the site conditions in reference to risk-based
calculations for cleanup levels. The object of this phase of the RFI process would be to reduce
risk, not to return the site to a "pristine" level.
This approach for RCRA sites is similar, in goals, to the Superfund Accelerated Cleanup Model
(SACM). The use of a simpler approach and a narrower focus to the Phase I RFI allows for
timelier risk reduction for people and the environment. Additionally, this approach focuses on
one of the four primary goals set by EPA's Administrator, Carol M. Browner, to make
permitting and other EPA decisions more prompt, fair, and definitive by using streamlined
reviews and greater reliance upon state and local agencies.
FOOTNOTES
1 Energy & Environmental Document Service Catalog First Quarter 1993.
35
-------
2 Superfund Accelerated Cleanup Model (SACM), US EPA Office of Emergency and
Remedial Response, Pub. 9203.1-01, March, 1992.
3 Ibid.
4 ROD Annual Report, April 1992, US EPA Office of Emergency & Remedial Response,
Publication 9355.6-05-1.
REFERENCES
RCRA Orientation Manual. 1990 edition, US EPA Office of Solid Waste, EPA/530-SW-90-036.
Interim Final RCRA Facility Investigation (RED Guidance - Volume I and II. May 1989, US
EPA Office of Solid Waste, EPA 530/SW-89-031.
36
-------
MATRIX SPIKES AND SURROGATES, DO WE NEED BOTH?
G. Robertson, U.S. Environmental Protection Agency, EMSL-LV, Las
Vegas, Nevada 89193, and M. Stapanian and F. Garner, Lockheed
Environmental Science and Technology Company, Las Vegas, Nevada
89119.
The users of measurement data in environmental chemistry require
specific knowledge of the quality of environmental measurements.
Since samples are known to vary considerably in analytical
difficulty, it is best to have estimates of precision and
accuracy for each analysis. Conventional approaches, such as
performing a replicate spiking study (e.g., standard additions)
on each sample would be prohibitively expensive for the analysis
of organic samples. By examining existing quality assurance
data, we investigated the relative merits of two lower-cost
alternatives. These include 1) spiking one sample from each
batch of samples of the same matrix with a subset of the analyte
list and performing duplicate analyses of this fortified sample,
and 2) spiking every sample with a known quantity of surrogate
compounds (substitute or isotopically labeled compounds). While
either approach will yield estimates of precision and accuracy,
we found better information in the .estimates derived from
surrogate analysis. Minor modification of the set of surrogate
compounds would further enhance the quality of these estimates.
Based on these results and on cost considerations, we recommend,
contingent upon the execution of a properly designed confirmatory
analytical experimental study, the use of surrogate recoveries in
the estimation of precision and accuracy in each semi-volatile
and pesticide analysis. The analytical design of the analysis
for volatile compounds will require a new approach to obtain
meaningful estimates of precision and accuracy.
NOTICE
Although the research described in this abstract has been funded wholly
(or in part) by the United States Environmental Protection Agency, through
Contract Number 68-C-CO-0049 to the Lockheed Environmental Systems and
Technologies Company, it has not been subjected to Agency review. Therefore,
it does not necessarily reflect the views of the Agency and no official
endorsement should be inferred.
37
-------
Abstract not Available
-------
QUALITY ASSURANCE FOR ENVIRONMENTAL SAMPLING
Nick C. Adolfo. QA/QC Coordinator, Core Laboratories, 1250 Gene Autry Way,
Anaheim, CA 92805; Ann Rosecrance, Corporate Quality Assurance Director, Analytical
Chemistry Division, Core Laboratories, 10205 Westheimer, Houston, TX 77042.
ABSTRACT
The concept of quality assurance (QA), in environmental sampling and analytical
activities, is most often applied to the environmental laboratory. However, to ensure that
data are both legally defensible and scientifically valid, the procedures, equipment, and
personnel for the sampling portion of the environmental measurement process must meet
requirements similar to those for the laboratory. They must also be reviewed with equal
criticism, because the potential for error which can result in inadequate, biased, or invalid
data is just as great during sampling operations as it is during laboratory analysis.
Though the QA requirements for field sampling are well documented in Environmental
Protection Agency (EPA) references, they are often contained in documents where
sampling is not the main focus, and are not always noted or understood. Therefore, much
of what is known or practiced in QA for field sampling is obtained the hard way, through
trial and error. It is essential that those performing the work have clear and specific
guidance on QA procedures for the sampling portion of the environmental measurement
process.
The QA activities associated with environmental sampling can be divided into three areas:
pre-sampling preparation, collection of samples including quality control samples, and
post-sampling preparation. This article will review these areas and provide
recommendations on the specific steps needed to produce legally defensible and
scientifically valid data.
PRESAMPLING ACTIVITIES
Presampling activities include both organizational and procedural planning. A project
coordinator and an independent QA coordinator must be selected. Also, to ensure that
the responsibilities for each different phase of the project are clear and that critical parts
of the investigation are not be overlooked, a project organization chart must be written
outlining the specific responsibilities of each project team member.
Once the project organization is clearly defined, a QA project plan (QAPP) must be
written, which often happens in tandem with the writing of a work plan. Sometimes, the
two plans are combined into one document. Required elements that must be included in
the QAPP are listed in Figure 1. In addition, the workplan or combination QAPP/work
plan should also include:
39
-------
FIGURE 1. ELEMENTS OF THE QUALITY ASSURANCE
PROJECT PLAN
1. Project Description and Purpose
2. Project Organization
3. Data Quality Objectives
4. Sampling Procedures
5. Sample Custody and Preservation
6. Calibration Procedures
7. Analytical Procedures
8. Data Reduction, Validation, and Reporting
9. Internal Quality Control
10. Systems and Performance Audits
11. Preventative Maintenance
12. Data Quality Assessment
13. Corrective Action
14. Quality Assurance Reporting
40
-------
o a site specific sampling plan which is based on geostatistical or historic
information about the site;
o standard operating procedures (SOPs) for sampling for every type of
matrix involved including, groundwater, wastewater, soils, sludges, air, and
composited samples;
o a summary of the tests to be performed based on historical data, the
regulation of concern, and the data quality objectives; and,
o the number of samples, analytical methods, bottles, preservatives, and
sample holding times for each sampling location;
The analytical laboratory should be selected prior to beginning the sampling phase of the
project. The selection should be based on the laboratory's reputation for quality, the
results of EPA performance evaluation (PE) sample studies, the analysis of PE samples
for the specific tests to be performed, and an on-site laboratory audit. The laboratory
should be kept fully informed of the anticipated sampling schedule, the number of
samples to be taken, and the time the samples will arrive at the laboratory. This allows
the laboratory to ensure that both holding and turnaround times can be met. Sending
samples to a laboratory that is not informed and not prepared to receive them can result
in the samples being given less attention or priority then samples which the laboratory
had already accepted for analysis.
Sample bottles, preservatives, chain-of-custody records, custody seals, and other
documentation should be acquired before going into the field. The sample shipping
method also should also be determined beforehand. Proper planning canjielp eliminate
common and detrimental sampling-related mistakes, such as the use of inappropriate
sample containers, not preserving samples adequately, and not shipping the samples to the
laboratory for several days because the laboratory or the method of shipment was
unknown.
Project team members should be thoroughly trained for all equipment and procedures they
will be expected to use during the sampling operation. They should also have been given
a 40-hour hazardous waste training course as required in the Code of Federal Regulations
(29 C.F.R. 1910.120) before entering potentially hazardous areas and collecting samples.
This training should be documented in individual training records for each project team
member. If the training documentation is procedure-specific and not project-specific,
these training records can be used for future projects that involve the same procedures.
A project-specific safety plan must be written and an initial project meeting should be
held to make certain that all project team members are familiar with the procedures to be
41
-------
used. The project safety plan should outline the specific chemical and physical hazards
expected to be encountered during the sampling operation, and provide measures for
protection against those hazards. It is advisable, especially for larger projects, to select
a project safety officer who is either an industrial safety or industrial hygiene professional
to develop and implement the safety plan. It is also advisable to conduct personnel
exposure monitoring and safety audits during the course of field operations.
The need to have quality considerations written into the formal plans before the actual
operation begins is documented in the EPA's Guidance on Remedial investigations Under
CERCLA: "Because the primary aim of the quality assurance/quality control program is
to ensure that the data are reliable, rather than to ensure that a poorly conducted program
is adequately documented, the QA/QC aspects should be planned in advance as an
integral part of the investigation."
COLLECTION OF SAMPLES
Quality considerations in the collection of field samples include the use of sampling
equipment and standard procedures, documentation of field activities, collection of field
QA samples, and performance of field QA audits.
Standard cleaning procedures for sampling equipment should be used between samples
to ensure that cross contamination does not occur. If historical data about the
concentration levels of analytes of interest in the samples to be taken is available, it is
often advisable to take the samples in order of increasing concentration to help prevent
contamination of the sampling equipment after a high level sample is taken.
SOPs for use of sampling equipment, collection of samples, and the handling of samples
after collection should be available in the field. Copies of the QAPP and/or workplan
also should be available to project members in the field as reference. The most best-
written QAPP and the most detailed SOPs can become ineffective if the sampling crew
is not provided with copies.
Bound field notebooks should be used to record such items as the general sequence of
events, the number of samples taken, any changes in sampling plans, geological
observations, atmospheric conditions, difficulties in obtaining samples, and sample dates,
times, locations, identifications, and descriptions (e.g., odors and colors). Any other
unusual circumstances also should be recorded. For projects involving drilling operations,
boring logs should be used.
Bottles, labels, preservatives, and chain-of-custody records should be prepared according
to specific instructions before any samples are placed in the containers. Bottles should
be washed and prepared using the appropriate EPA procedures, and they should be
labeled with a sample identification that is traceable to field notebooks or to a master
42
-------
sample-tracking log. The unique sample identifications, sampling location, date and time
of sampling, sampler's name, tests to be performed on the sample, and the laboratory to
which the sample is sent should be recorded in the field notebook or master tracking log
should be . This process can be computerized in projects where computer capabilities are
available; the sample information can be entered into the electronic logbook and sample
labels and chain-of-custody records can be generated electronically.
QA samples, including trip blanks, equipment blanks, duplicates, PE samples and
background samples, should be collected in the field. Consideration should also be given
to the collection of additional sample volumes for laboratory duplicates and matrix spike
analyses. To provide an independent check of the primary laboratory, a percentage of the
samples can also be split and sent to two or more different laboratories. A summary of
the types and recommended frequencies for field QA samples is presented in Table 1.
On-site sampling operation audits should be conducted and documented by the project QA
coordinator to verify conformance with the QAPP and project SOPs, and to ensure that
the field notebooks, chain-of-custody, and other documentation are complete. The
representativeness of field and QA samples should also be monitored.
POST-SAMPLING ACTIVITIES
The first post-sampling activities include the decontamination of clothing, sampling
equipment and the outside of sample containers. The decontamination procedures are
outlined in the 40-hour hazardous waste training course. Care should be taken to ensure
that the decontamination of the outside of sample containers does not affect sample
results. For example, it would not be advisable to use benzene, toluene, or other solvents
to clean the outside of a sample container holding sludge that will be analyzed for volatile
organic compounds. All waste from the decontamination, including disposable clothing,
should be disposed of using approved hazardous waste procedures.
Before leaving the sampling location, field logbooks and other documents should be
rechecked to ensure that the information recorded is complete. Information that can be
retrieved easily at this time may not be recoverable once the sampling crew has left the
site, or once the site is altered in anyway after the sampling event (such as by earth
moving equipment).
Samples that require cooling as part of the preservation should be placed in cold storage
at 4°C as soon as possible after sample collection. It is important that containers be
prepared adequately prior to shipment to ensure that the temperature is maintained until
the samples arrive at the laboratory. Insulated coolers should be filled with sample
containers and ice so as to prevent breakage of samples during shipment. Individual
containers can be placed in plastic bags to avoid sample cross-contamination. Custody
seals should be placed on the cap or lid of each container and on the lid of the shipping
43
-------
TABLE 1. RECOMMENDED TYPES AND FREQUENCIES FOR
FIELD QUALITY ASSURANCE SAMPLES
TYPE
PURPOSE
RECOMMENDED
FREQUENCY
Field (Equipment
Blank
Trip Blank
Decontamination
Blanks
Replicates
Splits
Field PE Samples
Measure background
level in sampling
equipment and
containers
Measures background
level from sample
transport and
containers
Measures background
level from field
decontamination
procedures
Measure variability
in results from
representative field
locations or
sampling procedures
Measure variability
in results between
laboratories
Measure accuracy of
overall measurement
system
5% or one per
sampling event
One per sample
shipment
One per each
decontamination
procedure
5% or one per
batch
5% or one per
sampling event
One set per
project
44
-------
container to ensure that samples are not tampered with during shipment.
Unless circumstances prevent it, the samples should be shipped the same day they are
taken, and should be received the following day at the laboratory. This helps to ensure
that the samples reach the laboratory at the proper temperature, and that the laboratory
has ample time to analyze the samples within the holding and turnaround times. If
samples are not shipped immediately after sample collection, they should be put in secure
storage areas that are accessible only by appropriate project personnel. Samples should
never be left unattended. Before samples are shipped, it is advisable to record the airbill
or shipping bill number on the chain-of-custody record. Chain-of-Custody, which must
be completed before the sample is shipped, are placed in a sealed plastic bag inside of
the shipping container. The chain-of-custody record should never be shipped separately.
CORRECTIVE ACTION AND DATA EVALUATION
Corrective action should be taken immediately for any deficiencies noted by a field
sampling operations audit or field data evaluation. However, unlike laboratory QC
samples, which are used as an ongoing control of the measurement process, field QA
samples can only be used for assessing the quality of the data after they have been
collected. Since this feedback is not immediate, the use of standard field QA procedures
is the best way to minimize errors that can result in inadequate, biased, of invalid data.
After the project is completed, the field documentation, results of field QA samples, and
reports of sampling operations audits should be reviewed, along with the laboratory
documentation and results, to determine if the requirements originally outlined in the
QAPP have been met. A summary of this review should be included in any report
written about the study, and where appropriate, the data should be qualified to indicate
any
limitations there may be in fulfilling the original purpose of the study.
ACKNOWLEDGEMENTS
This paper was previously published in the May/June 1993 issue of Environmental Testing
and Analysis.
REFERENCES
1. U.S. EPA. Data Quality Objectives for Remedial Response Activities. PB 88-
131370. March 1987
2. U. S. EPA. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods.
SW-846, 3rd edition.
3. U.S. EPA. Preparation of Soil Sampling Protocol Techniques and Strategies. PB
45
-------
83-206979. May 1983
4. U.S. EPA. Guidance on Remedial Investigations Under CERCLA. EPA/540/G-
85/002. June 1985
5. U.S. EPA. Interim Guidelines and Specifications for Preparing Quality Assurance
Project Plans. QAMS-005/80. December 29, 1980.
6. U.S. EPA. RCRA Ground-Water Monitoring Enforcement Guidance. August
1985.
7. U.S. EPA. Field Sampler Training Course Manual. PB86-217635. May 1986.
8. Kontopanos, K. N. and Williams, E. S. Quality Control in Field Sampling
Methods. Waste Testing and Quality Assurance: Third Volume. ASTM STP
1075. American Society for Testing and Materials. 1991.
46
-------
AUTOMATED REAL-TIME PROJECT LEVEL LABORATORY QUALITY
ASSURANCE
A. Gladwell. QA Director, Curtis & Tompkins, Ltd., 2495 Da Vinci, Irvine,
California, 92714; Kathy O'Brien, QA Director, Mark Schmick, Systems Manager,
Dr. C. Bruce Godfrey, President, Curtis & Tompkins, Ltd., 2323 Fifth Street,
Berkeley, California, 94710
ABSTRACT
Automated, real-time laboratory data management systems can be used in
conjunction with Data Quality Objectives (DQO's) based on client needs, objectives
and historical site information to design a relational database management system
that provides control beyond the typical analysis and quality control database. With
appropriate client/laboratory communication and interaction, such a system will
provide clients with a complete project results database.
Curtis & Tompkins has designed a real-time laboratory data management system
that acquires data directly from instruments and automatically checks data for batch
quality control and calibration as well as consistency with historical project data.
This system provides maximum automation without compromising data security
yielding improved efficiency while minimizing response time in correcting errors.
The system provides Real-Time Quality Control (RT-QC) at these levels:
•Data Generation
Calibration
Batch Quality Control
Completeness
•Monitoring
Current sampling event trends and anomalies
Historical trends and anomalies
RCRA statistical data reduction
Analytical applications of this system can also be automatically performed.
Examples of this are anion/cation balances, ICP interfering element checks, BTEX
ratios (found vs. theoretical), and carbon balance (BOD/COD/TOC).
47
-------
Such a system has benefits not only for the laboratory but for the client as well,
including:
• Real-time feedback on data anomalies,
• Reduced costs of QC compliance and data review,
• Automatic statistical data reduction and graphical reporting, and
• Timely delivery of complete and defensible monitoring data.
This system results in unprecedented improvements in efficiency, accuracy, and
process control to the data generation phase of monitoring projects. Consultants
using these data can be more confident in data quality and historical consistency.
They are free to concentrate on higher tasks such as trends determination and
environmental risk assessment.
INTRODUCTION
Why Real-Time Quality Control
There are many sources of data associated with environmental projects including
sampling, geological, geographical, analytical, meteorological, analytical QA/QC and
industrial hygiene, project data quality objectives and available historical data. Any
of the elements available to laboratory personnel can be entered into the database
management system and used in producing the client or project results database.
Using the Real-Time Quality Control (RT-QC) application of this system, we are
able to provide a detailed review of summary and raw laboratory data including the
evaluation of measurable data quality parameters such as holding times, instrument
calibration and batch quality control (including blank, laboratory control sample,
surrogate, matrix spike/matrix spike duplicate and sample/sample duplicate
information). The data are assessed to ensure that the raw data do indeed reflect
what has been presented in the summary format (such as CLP-type forms). The
overall goal being a complete report on the overall usability of the laboratory data
specific to a particular job or project. The benefits of this automated screening
system are tremendous.
• Diskette deliverables are evaluated for compliance with the measurable
parameters
• Data may be checked quickly and easily to ensure compliance with
project-specific requirements
• A comprehensive and thorough data compliance report is generated
• Highly trained, experienced personnel are not required to perform this
operation
48
-------
There are parameters which cannot be checked using this system. These include
pattern recognition of chromatograms (e.g.,petroleum hydrocarbons and PCB's),
mass spectral interpretation of TIC identity, ICP and AA spectral interferences, and
masking interferences in the case of matrix interference (e.g., with surrogates).
Therefore, the system cannot perform all functions and does have some drawbacks
including:
• Complete data validation cannot be performed
• Many laboratories cannot produce diskette deliverables
• Many data users do not understand the technology
By utilizing the system fully, a long-term monitoring application can be applied which
will evaluate historical data in relation to the most recent sampling event and
provide trending data or anomalous data points.
Designing the System
Certainly, one of the core concepts driving C&T's LIMS project is the recognition
that the screening of laboratory data is most effective and appropriate when
performed at the point of data generation. However, this is only the first of many
decisions in the LIMS development process. Concerns other than data quality
control impact the design, development and implementation of RT-QC. The
generation of technically and legally defensible data requires the collection of all the
data, whether or not it is to be reported to the client, and the future accessibility of
this data must be assured. A client may not want a data validation package today,
but they may need it in the future and all the information necessary to generate
the package must be retrievable. Laboratory size, capital available for
improvements, software and hardware systems currently in use, and the complexity
of client requirements faced by the laboratory were our major considerations. C&T's
experience is used as an example to illustrate some of these points.
System Requirements
RT-QC requires the availability of LIMS throughout the lab at every instrument and
workstation where data are generated. C&T has invested heavily in computer
hardware and software to support our vision of easy data access throughout the lab.
(See Figure 1). C&T's RT-QC system builds on top of the data sharing backbone,
but only a few components are really critical for a basic implementation of Real-
Time QC.
49
-------
Instrument Interfacing: there must be a means of gathering analytical data
electronically and "piping"it directly into the IIMS. The majority of instruments at
C&T are connected to LIMS via serial, or "RS-232" lines. Other instruments (gas
chromatographs, for example) that are controlled by PC's send their data into LIMS
over the local area network.
Relational Database Management Software (RDBMS): in our opinion a relational
database is crucial to flexibly manage the large data feeds generated by a typical lab.
C&T uses Oracle but selection of the RDBMS should be based on the actual
database structures required by the laboratory, size of the facility, and the dollars
available to allocate to such a system.
Multi-processing Operating System: the basic LIMS operating system must be
capable of processing many streams of data at once, in real-time. A robust, multi-
user operating system such as UNIX or Novell Netware should be considered.
Software Toolkit: it is most prudent to begin by examining the LJMS presently in
place in the laboratory to determine it's ability to proceed with development of RT-
QC. C&T started with a LIMS package from Automated Compliance Systems,
which provides much out-of-the-box LIMS functionality, as well as a toolJkit for
customizing the software according to the needs of a particular lab. C&T RT-QC
system was built in-house using this toolkit.
People: RT-QC development requires a high level of understanding of RDBMS
technology in order to design the database structures required to efficiently screen
data and provide immediate feedback.
When selecting a programmer (either as an in-house hire or a contractor) particular
attention should be paid to how the person will fit within the culture of the
laboratory. The programmer must be capable of drawing information out of
chemists and technicians regarding the processes they are running and introducing
new technology and concepts.
To maximize the success of any LIMS project, you must get buy in from the
analytical staff to answer the question, "What's in it for me?" They must be involved
with the project from the initial planning stages and should clearly see the benefits
of its complete implementation. Adequate training of all staff will require time and
resources and must be included by management in any development schedule.
50
-------
The C&T Scenario
C&T has completely automated the procedure for completing the screening of EPA
CLP metals data. We selected this process for testing our RT-QC ideas because:
no data system was in place, which eliminated the need to determine how to
integrate it into RT-QC; the CLP graphite furnace requirements are complex and
confusing which often result in errors requiring rework (and extensive analyst
training); the short duration of the analysis means that immediate feedback will
result in timely correction and prevent generation of unacceptable data; and we had
an immediate need to improve the laboratory's ability to deliver the product in a
timely manner.
RT-QC has completely automated the process of assuring that data which are in the
final results database meet project-specific requirements (often requiring
implementation of CLP SOW). The data are handled electronically from initial
generation, and deliverables are automatically generated to paper or diskette.
Diskette results invariably match the paper report because they are generated from
the same database. (See Figure 2.)
From the graphite furnace analysts's standpoint, RT-QC provides the answer to the
question, can I go on? It does this by reviewing the analytical result against all the
criteria specified in the CLP SOW; calibrations, blanks, QC samples, post digestion
spikes, duplicate burns, and all the other criteria for acceptable data. For example,
suppose we've taken a GFAA reading whose injection RSD is > 20%, the SOW
specified limit. About 5 seconds after the reading is taken, LJMS parses the data
file and reports back to the analyst that they will need to complete another analysis
of the sample (along with the reason why). The analyst completes the next run, to
which the LIMS responds again: this time, noting that the data have been marked
with the appropriate CLP RSD flag. The analyst then continues the analytical
sequence using subsequent feedback in the same manner.
At this point, the data are ready to report. A full-screen reporting utility is used to
queue up and generate client-ready reports in proper CLP format. Note that
nowhere in this process were data typed in by hand. Note also that no runs had to
be repeated unnecessarily; under RT-QC, the analyst knows immediately when QC
criteria are out of specification, and can adjust or abort the run as needed. And
only compliant data were stored in the final results database.
51
-------
SUMMARY
Laboratory Benefits
The benefits of RT-QC and associated data screening and evaluation are reduced
turnaround time, increased productivity (both through rapid feedback and reducing
rework), and improved data quality. Real-Time Quality Control uses automation to
complete the tasks which people find tedious and error prone thereby producing
these benefits.
Although RT-QC is rapid, thorough and many times more complete than traditional
manual method of "after the fact" compliance screening, it cannot replace aspects of
laboratory data evaluation which require "intelligent" review such as spectral and
chromatographic interpretations. These activities require analytical judgement based
on experience which is most efficiently provided by a chemist. In an indirect manner
RT-QC gives the chemist the time necessary to do thorough chromatographic
interpretation, turning this limitation into an overall advantage.
Client Benefits
Regulatory requirements continue to drive the quality improvement process in
environmental laboratories. Commercial environmental laboratories have another
incentive we have to do it faster, cheaper, and better to remain competitive. This
scenario provides a bright and challenging future for RT-QC.
The CLP example presented here is a practical place for a commercial laboratory
to begin developing RT-QC, but many possibilities exist. Projects which involve
more stable, consistent sample types will encounter fewer obstacles in the
development stage. It may be advantageous to begin by providing immediate
feedback concerning compliance with NPDES specified methods. The laboratory
could easily extend the advantages of RT-QC by developing client, site, and sample
specific requirements. Data could be submitted for internal laboratory QC
requirements designed in the same manner as the example provided above but
tailored to the laboratory's own QA/QC plan. Data which meet laboratory
requirements could then be fed into a statistics software package and screened for
trending against historic data, regulatory limits, or client supplied data quality
objectives. Reanalysis of outliers could be initiated immediately (literally within
seconds of generating the outlier) which would reduce the possibility of reporting an
error, such as a false positive, to a client. Client reports could automatically indicate
if regulatory limits were exceeded. The possibilities for improving service to the
client is tremendous.
52
-------
RT-QC and the Future
At C&T we believe that RT-QC is a necessary step in meeting the challenges of the
'90s for commercial environmental laboratories. In the future as more clients
develop project-specific data quality objectives, require electronic deliverables, and
push for ever more rapid turnaround times laboratories must respond by automating
the data screening process. The water pollution control community will continue to
feel detection limit pressure with ever increasing possibilities of false positives and
associated costs. Data trending against historic data points will become increasingly
important to your clients. Extensions of RT-QC will allow transfer of results directly
into statistical analysis software and provide immediate feedback to the analyst
regarding the status of the data point. Such advantages are going to be key in
meeting the challenges of the future.
53
-------
FIGURE 1.
LIMS NETWORK BLOCK DIAGRAM
INORGANICS LAB
UNIX
RDBMS
Figure 1: Complex multi-operating system
computer network used to collect and evaluate
data from instruments and data systems.
54
-------
FIGURE 2.
REAL TIME QUALITY CONTROL
DATA GENERATION
ENTER QC
SPECIFICATIONS
LIMS
RDBMS
QCDATA
SPECIFICATION
TABLES
"REALTIME" ^,
QC SUMMAR7**
EVALUATE RESULTS
AGAINST SPECIFICATIONS
(IDXL)
Figure 2: Process flow chart for real time QC. Data evaluated
against QC specifications enetered into structures on LIMS
file server are summarized directly to the analysts at their
workstation within seconds of data generation.
55
-------
FIELD AUDITING PROCEDURES FOR EPA CONTRACTOR PERSONNEL
Greg Janiec. Quality Assurance Officer, and Owen B. Douglass Jr., Ph.D., CIH, Health
and Safety Officer, Zone I Technical Assistance Team Contract, Roy F. Weston, Inc.,
(WESTONo) West Chester, Pennsylvania 19380
ABSTRACT
Traditionally, the major focus of quality assurance auditing in the hazardous waste
remediation field has been directed at laboratory monitoring and performance. Field
sampling and data collection activities have been identified as significant components in
evaluating data validity and error. A comprehensive field auditing program should
ensure that those who perform hazardous waste and environmental sampling are doing
so in accordance with established Standard Operating Procedures (SOPs), Quality
Assurance Objectives and site-specific scopes of work.
A procedure has been developed for performing internal audits of contractor performance
at CERCLA Removal Actions within EPA Regions I through IV. This procedure is
implemented through the use of a brief but thorough Site Audit Survey which directs the
auditor's attention to all critical phases of hazardous waste site operations support. The
prime focus areas include sampling methods including field screening (hazard
categorization) techniques, use and calibration of direct-reading field monitoring
instrumentation, site documentation, site data management, compliance with health and
safety requirements, and adherence to specific contract requirements.
INTRODUCTION
For the Zone I Technical Assistance Team (TAT) contract, WESTON and its
subcontractors employ over 200 multi-disciplinary environmental professionals in 11
office locations. These professionals provide technical and logistical support to the U.S.
Environmental Protection Agency (EPA) at CERCLA emergency response and hazardous
waste sites. The scope of work includes site assessment and evaluation, multi-media
environmental sampling (including biota), treatment system design and evaluation, air
monitoring, and general technical and engineering support within EPA Regions I through
IV, ERT and Headquarters. All work assignments must be performed at the highest
levels of professional quality and in accordance with the safety and health standards
established by OSHA. Periodic visits from the regional and program management staff
are necessary to ensure the quality of field work and client deliverables, and to identify
procedural refinements to augment the quality of the TAT services. Audits of field/site
operations are confined to the TAT contractor's scope of work and do not encompass any
aspect of the EPA's decision-making process or management techniques.
56
-------
DISCUSSION
Detailed Program Goals
The site audit format was designed and implemented as a positive and constructive means
to 1) ensure adherence to EPA, TAT and WESTON quality assurance policies and
procedures; 2) evaluate field screening and sampling capability and proficiency of field
personnel; 3) identify regional and/or program-wide deficiencies or trends; 4) exchange
innovative field techniques; and 5) obtain direct input from EPA On-Scene Coordinators
and Site Managers regarding TAT contractor performance.
Program Preparation
The TAT auditing procedure required a strong commitment of both program and
corporate resources. Initially, auditing objectives and protocols had to be developed and
implemented which could evaluate and present an accurate picture of the proficiency and
capability of personnel assigned to hazardous waste site operations and related support
activities. WESTON chose to evaluate all aspects of field work, including technical,
health and safety and administrative aspects. WESTON has embraced the concept of
Total Quality Management (TQM) and therefore structured its auditing procedures to
evaluate each component of a project and examine its performance and relationship to
meeting the client's expectations.
Subsequently, qualified auditors were identified and trained. All participants were
required to complete a 40-hour Hazardous Waste Site Worker Health and Safety Training
Program required by OSHA in 29 CFR 1910.120. Completion of this training phase
permits the auditors to freely move about hazardous waste sites to observe and evaluate
worker compliance with specific sections of the site health and safety and QA sampling
plans. Consistent with all other field personnel, the auditors are enrolled in a corporate
medical monitoring program which evaluates and certifies each participant's physical
ability to participate in site activities including the ability to use respiratory protective
devices. Medical and respirator certifications enable the auditors to participate in site
entry, sampling, and monitoring operations located in the hot zone.
Auditing Procedures
The centerpiece of the current site auditing program is the Site Audit Survey Form
(Figures 1-1, 1-2 and 1-3). This form provides a framework for the site auditor
including a brief listing of areas which should be examined during the auditing
procedure. The following protocols are used as guidance when performing a site audit:
57
-------
ZONE I
TECHNICAL ASSISTANCE TEAM (TAT)
SITE AUDIT SURVEY
SITE:
LOCATION:
SITE PHONE:
OSC:
REGION:
SITE LEAD:
TATM ONSITE:
in
09
EVALUATION CRITERIA
SCOPE OF WORK
TDD w present in file
TAT is performing tasks requested
SITE DOCUMENTATION
Site log is in proper format
Site files are organized
POLREPs are factual and regular
Personnel sign-in sheets are present
Site entry logs are maintained
Photographs are labeled and orderly
Off-site TAT costs are documented
COST TRACKING
TAT costs are current
Project costs are current
RCMS is being used
EQUIPMENT USAGE
TAT equipment is controlled
TAT personnel are trained in operation and
maintenance of monitoring equipment
Instrumentation calibration and maintenance is
documented
COMMENTS
Y=Yes
N=No
M=Most of the Time
O=Occasionally N=Never N/A=Not Applicable
-------
EVALUATION CRITERIA COMMENTS
SITE MANAGEMENT (TAT)
Site lead is designated
Individual duties ate assigned
Morning safety meetings are held
TECHNICAL INPUT
TAT is monitoring removal
TAT assisted with scope of work
TAT performed spill containment
Innovative treatment technologies were evaluated
Disposal options were evaluated
SAMPLING AND MONITORING
Sampling plan is current
Applicable SOPs are present and followed
Sampling equipment is decontaminated properly
Field screening is used appropriately
Air monitoring is conducted as per SOPs
Analytical Data is present
Validation conducted by
Funding source utilized
HEALTH AND SAFETY
Safety plan is complete to date and signed
Directions to the hospital are present
Emergency phone numbers are posted
Site Emergency Plan exists
Safety Plan has been implemented
Site is controlled
Decontamination systems are in place and
effective
Overall safety precautions are in effect
C:\WP31\DOCSl»OSC_DOC«IKBaTAT.aTB.AXID
Y=Yes N=No M= Most of the Time O= Occasionally N=Never N/A=Not Applicable
1~Z
-------
Page 3 of 3
OVERALL PERFORMANCE COMMENTS
Describe Site Conditions:
Describe Activities Observed:
o>
o
EPA Performance Comments:
Other Comments:
TAT AUDITORS: DATE:
DATE:
FIGURE 1-3
-------
1) If an audit is to be conducted by the TAT Zone Program Management Office
(ZPMO), a member of the ZPMO contacts the appropriate TAT regional office manager
to schedule a mutually agreeable date to perform the site audit. The office manager may
designate himself/herself or a senior member of the regional management team to
accompany the ZPMO representative to the site on the established date.
2) If the audit is to be performed solely by the regional TAT office, the office manager
or selected senior regional management member is responsible for the completion of the
audit and corresponding reporting procedure.
3) The EPA On-Scene Coordinator or Site Manager is notified by .the TAT office
manager of the pending audit at least 24 hours in advance. At this time, the EPA site
manager may be requested not to inform the on-site contractor personnel of the upcoming
audit.
4) Upon arriving on site, the auditor identifies himself/herself to the EPA and contractor
personnel, produces an identification and health and safety certification card, completes
the site sign-in sheet, reads/acknowledges the site safety plan, and briefs appropriate
personnel on the purpose and procedures associated with the audit process.
S) The auditor evaluates site activities and records in a manner to minimize the
interference with normal site activities. The auditor uses the Site Audit Survey Form to
evaluate the following major areas as applicable to the site scope of work:
• Work assignment documentation
• Field and personal log books including supporting documentation
• Site QA sampling plan
• Site health and safety plan including amendments
• Equipment (monitoring instrumentation) use proficiency of field personnel
• Field calibration and maintenance logs for all field instrumentation
• Storage and maintenance of equipment,
• Site reports and client work product deliverables
6) A hot zone entry may be made by the auditor to observe site operations (e.g., soil
sampling) and ensure adherence to relevant sections of the site scope of work, QA
sampling plan and established EPA and TAT standard operating procedures. If properly
trained, the auditor may participate in the respective site activity as a team member to
more fully appreciate the actual working conditions and related operational and weather-
related problems.
7) During a normal break period or after completion of the daily site activities, the
auditor conducts a briefing with the contractor site lead to identify areas in need of
61
-------
improvement; establishes a corrective actions plan to improve site operations; and
reviews relevant program policies or procedures.
8) At the EPA Site Manager's convenience, the auditor solicits a candid evaluation of
general contractor performance from the EPA manager. The auditor briefs the EPA Site
Manager regarding the identified items to improve site operations and any technical
advice.
9) The auditor subsequently drafts a report using a standardized outline (Figure 1-4)
identifying the audit findings and corrective suggestions. This is sent to the EPA Site
Manager and TAT regional manager for review and comment. After a ten day review
period, the auditor initiates the final site report. These reports are then distributed to the
TAT Zone Program Manager (ZPM), regional EPA Deputy Project Officer (DPO), and
EPA Project Officer (PO) in Washington, DC.
Audit Frequency
The target frequency for regional field audits is three per quarter per each region. The
ZPMO goal is to conduct three audits per quarter across the Zone.
SUMMARY
The current TAT site audit program has been in effect since June of 1989. Informal site
auditing was performed using non-standardized formats prior to this time. A concerted
effort was made to keep the Audit Survey Form brief despite the tendency to make
continued revisions and lengthen the document. Finalized audit reports are typically
from two to three pages in length. The simplicity of this auditing program is one of the
primary causes for its longevity and acceptance by contractor and client personnel.
Although the paperwork and personnel time commitments are minimal, this audit
program has demonstrated its ability to maintain and improve the delivery of high quality
services on a consistent basis. In terms of cost/benefit analysis, the return is significantly
greater than the cost.
This auditing program has achieved numerous positive effects within the Zone I TAT
contract including 1) the on-site evaluation of program field personnel; 2) the candid and
timely feedback regarding contractor performance from the on-site EPA client; 3) the
evaluation of the effectiveness and practicality of current standard operating procedures;
and 4) the identification of regional or program-wide opportunities to improve training
programs, purchase new field monitoring/support equipment, or revise standard operating
procedures to accommodate new technologies.
Unanticipated benefits regarding this project are substantial. The on-site nature of this
audit program requires the "hands-on" involvement of management, quality assurance and
62
-------
Figure 1-4
SITE AUDIT REPORT OUTLINE
A. Introduction
1. Site Name
2. Location
3. Date of Audit
4. Auditors
5. Time Spent on site
B. Observations of Activities
1. Review assignment scope of work
2. Documentation
3. Equipment calibration, maintenance and use
4. Site management
5. Technical support
6. Health and Safety
C. Recommendations for Improvement
63
-------
health and safety personnel. These individuals, many of whom had not performed actual
site work for one or two years, regained their "real world" perspective on field work.
This experience is extremely valuable to ensure that new or revised policies and
procedures have a positive impact on field operations.
The technology transfer benefit of this auditing program also proved to be very
beneficial. The auditors from the ZPMO were able to observe state-of-the-art techniques
and innovative approaches to resolving unique site problems in a real world situation.
Additionally, the auditors were able to participate in the assessment of the success or
shortcomings of each application. Bimonthly program management meetings served as
the forum to present these first hand observations to all the regional TAT office managers
to augment their capabilities to serve their clients.
Another unplanned benefit was a marked improvement1 in the morale of field personnel
including EPA Site Managers. The on-site presence of senior regional or program
management auditors illustrated a commitment to field work and the front-line personnel.
The site audit visit proved to be an excellent opportunity for field personnel to candidly
interact with management to discuss issues and ideas in a non-office setting.
In conclusion, WESTON's audit program has been extremely effective during the
execution of the Zone I TAT contract. It is our belief that this program can serve as a
model for other hazardous waste site operations contracts.
64
-------
10 THE FIELD CHEMICAL DATA ACQUISITION PLAN:
A DATA QUALITY OBJECTIVES
APPROACH FOR THE USE OF REAL-TIME FIELD ANALYTICAL DATA IN
ENVIRONMENTAL PROGRAMS
Raymond J. Bath Ph.D., Waste Policy Institute, Germantown, MD 20874; Richard D. Spear Ph.D. United States
Environmental Protection Agency, Region 2, Edison, NJ 08837; Daniel Lillian Ph.D, United States Department of
Energy, Environmental Restoration and Waste Management, Germantown, MD 20874; and Pamela Greenlaw, Lockheed
Environmental Systems & Technologies Co. Edison, NJ 08837
ABSTRACT:
The acquisition and use of data from on site real-time sample analyses has conceptually always
been the preferred approach for environmental programs. An examination of some field programs
has shown that a fUll capability field analytical service yields a variety of project benefits
including major decreases in project time; substantial cost savings in analytical services; increases
in data quality; and decreases in health and safety concerns for sampling crews and nearby
residents. Specialized vehicles for field analyses, e.g., vans and other mobile facilities, have been
developed and manufactured. A variety of high quality portable and transportable analytical
instruments have been developed to analyze soil, water, and air. Formalized research and
developmental analytical programs exist in sensor technologies for industry and government. The
question then arises, why are such field analytical services not routinely used? A possible
explanation is in the lack of an up front planning guide for the acquisition of field chemical data.
While the current data quality objective (DQO) guidelines present the up front planning framework
for the entire remediation work plan process, the sampling and analysis plan has not folly explored
the use of real-time field data. This oversight has restricted field analytical services. With the
addition of a field chemical data acquisition plan to the work plan process, real-time field analytical
data acquisition can assist in strearnling the investigation and remediation process. This paper
examines the state of environmental field analytical services today for the collection of real-time
data and presents an approach using the current DQO methodology for preparing and
implementation of a field data acquisition plan as part of the sampling and analysis plan.
This paper is the opinion of the authors. It does not reflect the current position/policy of the USEPA or USDOE. Any
manufacturers noted are not recommendations of equipment but are examples of current equipment trends.
65
-------
•J 1 A NEW COST EFFECTIVE STRATEGY
FOR ENVIRONMENTAL REPORTING SOFTWARE
Leigh A. Richardson, Director of Technical Services, and E. Alan Haley. Director of Marketing,
TELECATION, INC., 19423 N. Turkey Creek, Suite C, Morrison, CO 80465 (303) 697-8080
ABSTRACT
The prices for CLP reporting software have skyrocketed in the past year, reflecting the costs incurred by
software vendors in attempting to stay current with the changing requirements and screening standards
for CLP deliverables. The higher prices have put a new burden on environmental analysis laboratories,
who are required to produce CLP and CLP-like computer generated reports. In an attempt to control the
costs of environmental reporting software and reduce this burden on laboratories, TELECATION, INC.
has spent the last year developing a new "software system", which takes advantage of efficient
programming strategy to produce a variety of environmental deliverables at far less cost. In addition to
the cost savings, the reporting options are more flexible, allowing both vendor and user to readily make
sophisticated changes to address special variations in requirements. This approach has been used to
generate reports modeled after the requirements of both CLP and RCRA/SW846, and other protocols are
being examined. This paper will describe the techniques used to deliver both cost effectiveness and
versatile software application for environmental reporting.
INTRODUCTION
The cost of computer software is controlled roughly by the ratio of the complexity of development to the
market size. The market size for laboratory software applications is infinitesimally small compared to the
general business software market place. This, along with the fact that many of the requirements of
laboratory software make it inherently complex to develop, serve to inflate the final price of software
designed specifically for laboratory use. The size of the market is a factor which cannot be controlled by
a software vendor. However, the complexity of development can be controlled by applying large scale
efficiencies in the process of software development.
Most laboratory software applications fall into the general categories of:
DATABASE APPLICATION - for data storage, retrieval, report generation, or;
SPREADSHEET APPLICATION - for calculations, data relationships, and graph generation
Since these are readily available tools, laboratories have always had the option of addressing their
software needs by using one or more of the numerous, low cost generic database and spreadsheet
products on the market and configuring the tools around each required application. Many laboratories,
however, are not staffed with personnel who have the background, training, or time to configure and/or
program the general purpose tools into a final application solution. Further, for laboratories who do
have the programming resources, the additional costs of applying these resources to basic laboratory
software needs largely offset the original cost savings realized in the general purpose software purchase.
66
-------
A software vender who develops turnkey software solutions specifically for the laboratory market place
must also incur these initial development costs. However, unlike the laboratory developing their own
software, the software vendor can amortize the initial development cost over the entire client base for the
product, providing a turnkey solution to a very specific task at a lower cost than it can be developed
in-house, even if the programming resources exist. Even so, if each software application is developed
from the "ground up" as a stand-alone product, the initial development cost for that product must be
absorbed, only by the purchasers of that individual product. The next application will undergo a similar
level of development, adding to the cost of that product, which again must be ultimately borne by the
purchasers of that single product. The challenge for the software vendor seeking to reduce the prices of
finished laboratory products is to address the repetitive nature of overhead development costs associated
with each individual product.
A STRATEGY FOR PROGRAMMING EFFICIENCY
Approximately one year ago, TELECATION, INC., a company specializing in PC-based software for
the laboratory, initiated a study into ways to dramatically increase the efficiencies of application
programming with a corresponding reduction in development time and final software price. In doing so,
TELECATION systematically analyzed the software needs of the laboratory community, looking for the
software functions common to many different laboratory applications. The software company then
embarked on a project to develop a base product which would contain most, if not all, of the software
functions needed to implement almost any laboratory application. Additionally, the base product was
specified to contain a variety of tools to give the user control over such things as database design, report
definition, formula definition, etc. The goal for this product, in and unto itself, was that it should be the
single-most powerful and flexible tool ever designed to specifically address laboratory software
applications.
This product was to be different from the generic database and spreadsheet products on the market, in
that it was to directly address specific laboratory needs. The system was designed to address general
purpose laboratory data handling with little or no configuration by the final user of the product. Special
laboratory-specific functions, like analytical quality control procedures, instrument interfacing, even
export to LIMS, were part of the basic product.
To allow for extension of the capabilities of this product beyond the basic design, tools were provided for
extensive tailoring for any specific task. A typical computer-oriented chemist could use the tools to
automate very specific laboratory tasks, which could be routinely executed by anyone in the laboratory.
While user configurability of the system was to be a major feature of the software product,
self-configuration of software solutions is, as previously stated, not necessarily in the best interest of
every laboratory. Therefore, TELECATION expanded its strategy for the "universal" laboratory
software tool to include its own, turnkeyed applications, running on the base software product. By using
the configuration tools designed into the basic system, TELECATION did not have to design, code, test,
validate and implement the basic functions, for which these processes had been previously conducted on
the base product. In other words, the development overhead incurred in the base product, did not have
to be repeated for each and every software solution operating on the basic system. TELECATION's goal
of providing complete and specific software applications for the laboratory at breakthrough prices was,
therefore, realized by implementing a systematic strategy for future software development, which
dramatically reduces the product development overhead.
A block diagram of the concept is illustrated in Figure 1.
67
-------
INSTRUMENT
DATA
KEYBOARD
ENTRIES
Input
Basic laboratory tool
"DATA MANAGER 2000"
Output
1
GRAPHS
CHARTS
1
REPORTS
1
USER
APPLICATIONS
1
"PLUG-IN"
LABORATORY
APPLICATION
"2000 Series"
1
LIMS
Figure 1. An Efficient Strategy for Laboratory Software Development
PLUG-IN APPLICATIONS FOR A GENERIC LABORATORY TOOL
The "base product" referred to above is called "DATA MANAGER 2000". The add-on applications,
each of which takes advantage of the software development which went into DATA MANAGER 2000,
are collectively referred to as "2000 Series Applications." All "2000 Series" applications literally "plug
in" to DATA MANAGER 2000 through a series of program "hooks" designed to send and/or receive
data from special data handling modules. The magnitude of effort which is required for each "2000
Series" application varies from literally a few hours for a simple application-specific database (like the
Chemical Inventory Database) to many weeks for a more complicated application (CLP-like reporting),
which may involve the automation of many relationships, calculations, and report formats. In all cases,
the prices for "2000 Series" applications are far less than software products developed from the ground
up with a specific purpose in mind.
A side by side comparison of software prices can be made in the case of CLP report generation software.
TELECATION, INC. offers software products (the "ENVIROFORMS" ® product line) designed
specifically to automate as much as possible the deliverable data requirements of the USEPA's Contract
Laboratory Program (CLP). While the "2000 Series" products do not address all EPA-specific
requirements with the same rigor as the "ENVIROFORMS" products, they provide an environmental
deliverable which will meet the needs of most laboratories at a cost savings of approximately 85%
compared to the CLP-specific alternative. Indeed, some of the price difference is associated with the
development overhead associated with meeting EPA details in the "ENVIROFORMS" products, but
much of the difference accrues from the programming efficiencies associated with the "2000 Series"
68
-------
concept. (NOTE: TELECATION, INC. continues to offer and update the "ENVIROFORMS" software
products for strict CLP applications, as well as the "2000 Series" products for those desiring to produce
CLP-like deliverable at reduced prices.)
Table I lists the "2000 Series" applications currently available or under development. Given the
company's extensive history with environmental software and the variety of environmental reporting
protocols which have been promulgated, it is not surprising that many of these packages address
environmental issues. However, of the current and planned applications, many address other laboratory
tasks, as well. Specifically, a series of ISO 9000 and "Good Automated Laboratory Practices'1 (GALP)
oriented tools are currently available and many more are under development.
Table I
EXAMPLES OF "2000 SERIES" PLUG-IN SOFTWARE APPLICATIONS
o CLP-like reporting for Inorganics, VOA, SNA, PEST
o RCRA/SW846 reporting for Inorganics, VOA, BNA, PEST
o Chemical Inventory System
o Hazardous Waste Disposal Tracking System
o Detection Limit Calculator & Reporter
0 SOP Filing System
o Personnel Training Log
o R-Bar Chart
o Standards/Reference Material Log
o Computer Hardware/Revision Log
0 Instrument Usage, Calibration & Maintenance Log
o Raw Data Archive Log
o Multi-Phasic TCLP/Inorganic Reporting System
0 Storm Water Discharge Monitoring
69
-------
ADDITIONAL BENEFITS OF THE STRATEGY
It was previously stated that market size, in addition to development complexity, plays a significant role
in software pricing. In a very real sense, the product strategy described above increases the market size
while reducing development complexity. Since each plug-in application uses the same base product, the
development costs of the base product can be amortized over the market for all compatible products.
which is significantly larger than for any single product. In this way, the development strategy is
controlling both factors which affect software prices.
Price is not the only benefit from the strategy, however. Personnel training is a major concern and
expense for a laboratory. In today's laboratory, a major portion of the training dollar goes toward
training software operators. Since all applications in the "2000 Series" product line operate through a
common base product, the learning process is greatly simplified. Once a user learns to use one
application, they already have the background to use any other. The strategy thereby maximizes a
laboratory's return on personnel training, by applying the experience gained with one application to all
other system-compatible applications.
SUMMARY
Software prices for laboratory applications have traditionally been high due the small market size and
complexity of development. Software prices can be significantly reduced through a software
development strategy based on plug-in, application-specific routines which depend on the functions of a
common base product for operation. In this way, the development effort which went into the base
product can be used over and over again, thereby reducing the effort required to develop each compatible
product. Additionally, the costs of the basic product can be amortized over the market size for aH
compatible products, which is significantly larger than that for any single product.
70
-------
^ 2 Local Area Network Data Validation Software System
A. Davis. Quality Assurance Manager, BCM Engineers Inc., One Plymouth Meeting,
Plymouth Meeting, Pennsylvania, 19462, and J, Vernon, Approach, Inc., 100 Summit
Lake Drive Valhalla, New York, 19595.
ABSTRACT
The extensive data validation procedures required tjo review Contract Laboratory Program
(CLP) data packages generated for both regulatory and private parties environmental
investigations are extremely labor intensive and usually do not consider previous
investigation data. To address these concerns, we have developed DataVal, a data
validation software that reduces the review effort and enables the validators to access the
full investigative database. DataVal translates CLP diskette deliverable formats (DDF)
and automates validation procedures. DataVal enables the validation specialists to
concentrate their review efforts in the interpretive areas of the data package. Using this
software's capabilities to access the database generated over the entire course of a large
investigation, the validation specialists and data users can better interpret the viability and
usability of the analytical program results.
The DataVal software is a Client/Server application implemented on Windows™ based
personal computers accessing a Relational Data Base Management System (RDBMS) that
is a part of a Local Area Network (LAN). The Client component was developed using
Visual Basic™ and the RDBMS component chosen was SQL Server. The software itself
is modular in design, permitting the validators the flexibility to choose a "path" through
the validation protocols. These "paths" allow assessment of complex data relationships
and hierarchical review. The modular software design lets the validator set separate
criteria for regional data validation modifications or set criteria for previous or future data
validation requirements. The ability to access the previously validated database in the
SQL Server allows validators to evaluate individual data sets in relation to the historical
data.
Our company's project management system allows the transfer of laboratory information
management system (LEVIS) data directly into a larger project management historical
database. DataVal is being integrated into this project management system. Upon
completion of data validation, the data qualifiers are electronically transferred into the
project management historical database, ensuring consistency of data qualification. The
integration of DataVal into the project management data system combines a powerful data
interpretation tool with a high capacity database management system. The data users are
supplied with a a project database that is better evaluated, cost effective, and highly
accurate.
71
-------
4 Q QUALITY ASSURANCE AND QUALITY CONTROL
10 MEASURES FOR THE DETERMINATION
OF HEXAVALENT CHROMIUM FROM STATIONARY SOURCES
Thomas Dux. Quality Assurance Coordinator,
K. Hall, G. Jungclaus, and A. Williams,
Midwest Research Institute, 425 Volker Blvd,
Kansas City, Missouri 64110
ABSTRACT
Midwest Research Institute has been collecting and analyzing
stack gas samples for hexavalent chromium for the last 4
years utilizing the method from the Methods Manual for
Compliance with BIF Regulations (EPA/530-SW-91-010). MRI
has enhanced the methodology with the addition of matrix
spikes, audit samples, reporting limit procedures, and
validation of the integrity of sample preservation. This
paper discusses these enhancements and associated data
quality objectives plus the results from quality control
samples.
In addition to performing hexavalent chromium analyses, MRI
has also reviewed numerous trial burn plans and trial burn
results under contract to the EPA. Through this experience,
MRI has found that some firms conducting trial burns do not
fully characterize the sample results in terms of the
quality of the data. Problems include the contamination of
collection reagent with hexavalent chromium, the lack of key
information on the integrity of the collection media
following sample collection and the lack of precision and
accuracy data for the analysis results. These difficulties
have resulted in trial burn plans which were deficient and
trial burn results which could not be validated. This paper
also addresses these issues.
1.0 INTRODUCTION AND BACKGROUND
1.1 Introduction
Midwest Research Institute (MRI) has been collecting and
analyzing stack gas samples for hexavalent chromium for the
last 4 years utilizing the method from the Methods Manual
for Compliance with BIF Regulations (EPA/530-SW-91-010). In
addition to performing hexavalent chromium analyses, MRI has
also reviewed numerous trial burn plans and trial burn
results under contract to the EPA. Through this experience,
MRI has found that the method as written does not fully
characterize the quality of sample results in regards to the
72
-------
accuracy of the determinations, the detection limit and the
validation of sampling activities. These difficulties have
resulted in trial burn plans which were deficient and/or
trial burn results which could not be validated due to a
lack of information.
1.2 Background
Besides the typical Method 5 quality control (QC) procedures
(e.g. calibrations, leak checks, etc.) the method for
hexavalent chromium has 5 major quality control procedures:
1. Sampling Train Validation: Hexavalent chromium is an
analyte which is very subject to reduction since it is
a good oxidizer. To prevent loss of Cr+6 the pH of
sampling train media (aqueous potassium hydroxide) must
be maintained above 8.5, the train must be purged with
nitrogen following sampling and the contents of the
impingers must be filtered immediately following
recovery.
2. Instrument Calibration; The instrument must be
calibrated with at least four standards over one order
of magnitude. A curve is run before and after samples
and the relative percent difference (RPD) of the
instrument response for the two analyses must be less
than 10 %. In addition, the calibration standards must
exhibit a relative accuracy of 93 to 107 % when
calculated versus the linear regression curve.
3. Duplicate Analysis; Each sample must be analyzed in
duplicate and the RPD of the instrument response must
be less than 10 %. If RPD is not less than 10 %, the
samples are reanalyzed until this criteria can be met.
4. Audit Sample; An EPA audit sample must be analyzed as
an independent check on the accuracy of the
determinations.
5. Blanks; Field blanks must be analyzed and can be used
to correct emission data for blank contributions.
The method does not address the following three quality
control subjects:
1. Sample Holding Times; Holding times are not given in
the method. The only applicable RCRA holding time
currently available indicates a 24 hour holding time
(Chapter 2, SW-846, Revision 2, 11/90), but does not
73
-------
require preservation at a pH above 8.5.
2. Accuracy: The method does not require any
determination of the accuracy of the analysis as it
pertains to the stack gas matrix (i.e. spikes).
3. Detection Limits; The method does not address
detection limits or give a procedure for determining
them.
The remainder of this paper will discuss some results for
the method specified QC procedures and MRI's additional QC
measures.
2.0 QUALITY CONTROL PROCEDURES AND RESULTS
2.1 Holding Times and Validation of Sampling
As previously indicated, holding times are not specified in
the BIF method. The most current available RCRA holding
time is 24 hours (Chapter 2 SW-846). However, most firms
use a 14 or 28 day holding time with preservation requiring
a pH greater than 8.5 and a temperature of 4 °C. The
EPA/ASTM method for hexavalent chromium in water indicates a
holding time of 24 hours with preservation at pH 9 to 9.5
and a temperature of 4 °C ("Determination of Dissolved
Hexavalent Chromium in Drinking Water, Ground Water and
Industrial Wastewater Effluents by Ion Chromatography:
Collaborative Study" K. Edgell, et. al., Bionetics
Corporation, April 6, 1992.) This document also indicates
that dilute standards should be prepared daily but that
stock standard solutions are stable for at least 3 months.
In planning trial burns or stack testing, it is best to be
conservative and use the minimum acceptable holding time.
Most EPA regions and states will approve a 14 to 28 day
holding time with preservation at a pH greater than 8.5 and
a temperature of 4 °C. The MRI field sampling and analysis
team uses a 14 day holding time and preservation as
indicated. However, when sampling is not performed by MRI,
the laboratory meets a 14 day holding time from laboratory
receipt.
During reviews of trial burn plans and reports, it has been
noted that some researchers are not aware of holding times
and preservation considerations for these samples. The
following table gives results for the same samples, from the
same stack, analyzed 6 and 19 days from receipt.
74
-------
Table 1 * Results for Hexavalent Chromium Determinations in
Stack Samples Analyzed 6 and 19 Days from
Laboratory Receipt
Sample
1
2
3
4
5
pH
14
14
14
13
13
Result
(ug/L) 6
Days From
Receipt
7.8
10.6
3.7
2.3
15.3
Result
(ug/L) 19
Days From
Receipt
3.4
3.0
2.0
0.74
8.1
% Loss of
Analyte
From 6 to
19 Days
56
72
46
68
47
These results are only indicative of hexavalent chromium
stability in impingers from that specific stack. It is
interesting to note that the average loss is 58 % with a
standard deviation of 12 indicating good agreement for the
stability data despite the varying concentrations of the
samples. These data clearly indicate that holding times are
an issue and sample results can drop within the time frame
accepted by regulatory agencies.
The BIF method has three validation steps which are specific
to Cr+6 (as opposed to Method 5 requirements). These are a
nitrogen purge of the train following sampling, the
filtering of the samples following train takedown and each
impinger must have a pH greater than 8.5 at the end of an
analysis run. Many Quality Assurance Project Plans (QAPjP)
for trial burns give descriptions of the sampling procedures
but do not address these guality control measures. They
should be discussed in the QAPjP and compliance must be
documented during sampling to validate the collection and
preservation of the samples.
Of critical importance is the pH of the impinger solutions.
Many trial burn plans and reports do not address this topic.
Some field sampling crews have adopted the practice of
testing the stack before the trial burn and increasing the
normality of the potassium hydroxide solution if needed for
highly acidic stack gases to maintain a high pH. In
addition, a significant number of trial burn reports and
associated raw data do not have the pH documented in the
field at the conclusion of sampling or upon receipt at the
laboratory. Without evidence of a pH greater than 8.5, the
75
-------
hexavalent chromium results cannot be validated.
Some permit applicants challenge the rejection of hexavalent
chromium results in the absence of a pH check, and insist
that the check was done even if it was not documented. MRI
tracks the pH of samples received for hexavalent chromium
analysis and in the past year about 10 % of the samples had
a pH less than 8.5. The pH ranged from 5 to 7, and only one
of these low pH samples had Cr*6 above the detection limit.
(All samples collected by MRI personnel met the pH
requirement.) Without documentation of sample preservation,
the sample results can not be validated.
MRI measures the pH of all Cr*6 samples before analysis and
reports the pH with the sample results in a manner similar
to inorganic, volatile and cyanide samples which also
require pH adjustment for preservation. Accreditation
organizations, customers and regulators should demand that
sampling firms document the pH of the impingers in the field
and the pH should also be verified by the laboratory.
Laboratories should not report Cr*6 results without also
reporting the pH of the sample.
2.2 Precision Determinations
The method requires that each sample be analyzed in
duplicate and that the RPD of the instrument response must
be less than 10 %. If RPD is not less than 10 %, the
samples are reanalyzed until this criterion can be met.
There are three reasons why this is not a very practical way
to handle precision. First, this requirement does not take
into account the decrease in precision as the response
approaches the instrument detection limit. Second, without
a technical reason, the first analyses that do not meet 10 %
RPD cannot be validly rejected and should be included in an
average result. Third, this requirement assumes manual
injection allowing for a repeat of analysis until the
criteria is met. This is labor intensive and not practical
given the availability of autosamplers.
MRI usually uses an autosampler for unattended, overnight
analysis and injects all samples in duplicate. The
precision criteria is rarely exceeded (<5 % of all
analyses). In the cases where there are problems, it is
usually due to the response being at the detection limit or
there is significant interference from the stack gas matrix
which is evident in the chromatogram (e.g. extraneous
peaks.) If manual injections are made, the RPD is tracked
and if samples do not meet the criterion it is analyzed a
76
-------
third time and the average of all three analyses is
reported.
In reviewing trial burn data for various EPA regions and
states, MRI has noted that precision data are often not
reported with trial burn results. Some laboratory reports
do not address precision at all and others indicate only
single analysis of each sample. These samples do not have
any preparation beside filtering; therefore duplicate
analysis is the only measurement of precision. The
chromatograms are relatively short (<10 minutes) and
duplicate analysis is not a burden. Permit writers need to
be aware of this issue in reviewing trial burn plans and
results.
2.3 Instrument Calibration
The method requires that the instrument be calibrated with
at least four standards over a concentration range of one
order of magnitude. All standards are run before and after
samples and the relative percent difference (RPD) of the
instrument response for the two analyses must be less than
10 %. In addition, the average of the calibration
standards, analyzed before and after samples, must exhibit a
relative accuracy of 93 to 107 % when calculated versus the
linear regression curve. The calibration curve is verified
with an EPA audit sample.
This method of calibration is not amenable to automation and
routine analysis of large groups of samples. Since the
standards are run before and after samples, the success or
failure of calibration cannot be determined until after
samples are analyzed. This not very practical. Instrument
drift is usually negligible, however the relative accuracy
of 93 to 107 % is sometimes hard to achieve with standards
being run before and after samples. In addition, a curve of
one order of magnitude covers a relatively small
concentration range. Calibration criteria can be met with a
25-fold range and with a slight relaxation of criteria (i.e.
90 to 110 % relative accuracy) can be extended over a 100-
fold range.
Currently MRI's clients want the method followed exactly as
written to provide defensible data without any method
modifications. However, the method could be updated and
made more suitable for routine analysis by incorporating the
following changes:
Do not limit the calibration range.
77
-------
Require a curve at the beginning but not at the end.
Initial calibration criteria should be 90 to 110 %
relative accuracy.
Require continuing calibration standards which must
meet a 10 % drift criteria (measured using
instrument response.)
These changes will increase laboratory efficiency and make
the analysis method more like other SW-846 methods. It is
hard to control instrument calibration when the calibration
occurs before and after sample analysis. If criteria are
not met; all samples must be reanalyzed, however the
modified calibration criteria will require reanalysis for
only those samples analyzed after the last successful
continuing calibration. Of course, using the average of the
curve run before and after samples provides marginally
better data, but if the calibration criteria are set
appropriately, data quality will not be affected.
These changes will not significantly affect data quality-
When put in the context of the data quality objectives of
the trial burn and other measurements of equal importance
(GC/MS criteria are significantly wider) these changes will
not affect regulatory decisions.
2.4 EPA Audit Samples
The method requires an analysis of an EPA audit sample and
results must be within 90 to 110 % of the true value. The
purpose of this sample is to serve as an independent check
on the accuracy of the analysis; however it is not currently
available from the EPA. Therefore, some method of
verification of the accuracy of sample analysis is needed.
The analysis method does not require an independent standard
for calibration verification, thus there is no verification
of acceptable calibration. For VOST and SVOST there are EPA
audit samples and for metals analysis there is a requirement
for calibration verification with an independent standard.
At MRI, we have QA personnel prepare an independent audit
sample that is submitted for analysis with each batch of
samples. The sample is prepared from NIST reference
material in dilute potassium hydroxide to mimic the
collection reagent and provide stability. The concentration
of this sample is not known by analysis personnel. The
results of the analysis must be within 90 to 110 % of its
true value. The results for about 6 months of audit samples
is presented below in Table 2.
78
-------
Table 2 * Accuracy of MRI Hexavalent Chromium Internal Audit
Samples
Audit sample True
Concentration (ug/L)
10.2
11.5
9.5
13.7
9.1
6.9
8.8
12
8.8
8.6
Average (standard deviation)
% Accuracy
102
110
98
101
101
98
110
97
101
100
102 (5)
Currently most firms do not provide this independent check
on the accuracy of sample data; however, a similar check can
be done by field samplers by submitting a reference
solution. The preparation of such a standard requires
purchasing reference material and the time takes about 1.5
hours (solution standards are stable for at least 3 months).
Permit writers and trial burn reviewers should be aware that
some independent check on the accuracy is needed.
2.5 Spike Sample Analysis
The analytical method does not require spikes of samples to
determine the accuracy of the analysis given the stack gas
matrix. However SW-846 requires spike samples, and the
previously mentioned study using the ASTM and EPA methods
for hexavalent chromium in water also require matrix spikes.
MRI requires a determination of the accuracy and precision
of every analysis to be evaluated in terms of the sample
matrix. MRI does at least one matrix spike and sometimes a
duplicate matrix spike for a set of samples from an
incinerator, boiler or industrial furnace.
To prepare spikes, an aliquot of the sample is fortified
with a very small amount of a high concentration standard to
79
-------
achieve a concentration level in the middle of our
calibration curve. This takes about 15 minutes of a
laboratory technician's time. Our objective is for recovery
to be between 80 and 120 %; however other objectives for
trial burns, such as GC/MS surrogate recovery, are as wide
as 50 to 150 % recovery. The data for the past year is
summarized in Table 3.
Table 3 * Recovery of Stack Gas Impingers Spiked with
Hexavalent Chromium
Initial Sample
Concentration
(ug/L)1
13.1
6.39
5.96
7.08
7.29
3.81
<1.0
<1.0
<1.0
1.93
Overall Average
(standard
deviation)
Average % Recovery
from Matrix Spike
Duplicates
101
109
104
40
61
85
92
92
89
94
84 (22)
Relative Percent
Difference Between
Duplicate
NA2
16
4
4
2
3
7
3
NA
NA
NA
1 Spike level was about 10 ug/L for all samples
2 NA = Not applicable; a single spike was prepared instead
of duplicate spikes.
The spike recovery data indicates that most samples give a
recovery well within the 80 to 120 % range; and all but one
sample was within 50 to 150 % recovery- The precision of
the results is excellent. Even though recovery is generally
good, these data indicate that recovery is variable and
without any spike determinations, the accuracy of the
emission measurement is unknown. Usually samples with lower
recovery show poor peak symmetry or extraneous peaks in the
chromatogram, and these difficulties are consistent for
80
-------
duplicate analyses of the matrix spike and between matrix
spike duplicates. This indicates a truly different matrix
from calibration standards and supports the claim that low
recovery is due to the sample matrix and not to laboratory
operations. In addition, good results for the audit sample
provides evidence that the recovery problem is due to matrix
effects and not a problem with calibration.
What is the impact of this low recovery? That depends upon
the emission results. For example, the apparent problem
indicated by the low recovery (40 to 65 %) in the previous
table did not translate into a regulatory problem. Stack
emissions for hexavalent chromium were very low and if
emission rates were adjusted for the low recovery, risk
assessment indicated acceptable incinerator performance.
However, in cases where the emission results are borderline,
recovery data are needed to support the regulatory decision.
2.6 Detection Limit Determinations
The method does not specifically define detection limits nor
require their derivation. MRI has done detection limit
determinations using multiple analyses of low standards (0.5
ug/L) and determined a detection based upon precision of
about 0.2 ug/L. This 0.2 ug/L level is only slightly above
the 5 times signal to noise ratio for most samples.
In analyzing calibration curves, MRI has determined that
there exists a small amount of Cr*6 as background in our
standards. The amount of this background level is between
0.2 and 0.4 ug/L and is evident as a positive intercept in
the calibration curve by linear regression. This background
level is due to the potassium hydroxide (0.1N) used to
prepare calibration standards, since 0.1N potassium
hydroxide has an observable Cr*6 peak and straight deionized
water does not have any observable Cr*6. However, this
background is insignificant when compared to stack gas
samples and the levels of regulatory interest. In addition,
this background level is lower than or equal to that found
in field blanks of collection reagent.
Due to this background, MRI has adopted a calibration curve
that extends down to 1 ug/L and uses this level as a
reporting limit. Since regulatory use of the data often
employs the detection limit to prove low or no risk to the
environment when the analyte is not detected, the use of the
lowest calibration standard provides a conservative
reporting limit. However, in reviewing trial burn plans and
reports, MRI has noted that many firms doing these analyses
81
-------
present detection limits without describing the derivation
of the limit. In most cases this is not a difficulty due to
the presence of hexavalent chromium in the samples, but in
cases where the analyte is not detected, and emissions rates
are based upon the detection limit, the detection limit must
be supported with a technical justification.
2.7 Blank Correction
The method allows for correction of emission data by
subtracting the amount found in the blank; however, a
statistical justification for blank correction is not
required. For example, to blank correct VOST results, the
permit applicant must prove that the blanks are
statistically different from the sample results. In
reviewing trial burn reports, the blank correction option is
often employed; however, its use is neither statistically
justified nor is its impact upon the regulatory decision
discussed. Often this blank correction comes from the
analysis of a single field blank. In one case, the blank
correction lowered emission levels by 25 to 50 %.
The general policy for blank correction in trial burns
presented in Handbook of QA/QC Procedures in Hazardous Waste
Incineration (EPA/625/6-89/023, January 1990) is to provide
a statistical justification of the blank correction and to
present emission data both corrected and uncorrected. Given
the high toxicity of hexavalent chromium, this policy should
be incorporated in future revisions of the method. In
addition, permit writers have considerable latitude in
requiring trial burn information, and they should accept
blank corrected emissions only with clear justification and
the emission rates presented as corrected and uncorrected.
3.0 CONCLUSION
The hexavalent chromium analysis method is rigorous and
produces relatively precise and accurate data suitable for
regulatory decision making. However, it does have a few
short-comings in QA/QC procedures for holding times,
accuracy, detection limits and handling blank data. In
addition, calibration and duplicate analysis requirements
could be modified to allow a more cost effective automation
of the analysis. It is hoped that the topics brought up in
this paper will be used when the method is updated and that
the enhancements will be required by trial burn plan
reviewers to monitor and demonstrate the quality of the
sample results.
82
-------
14 DEVELOPMENT OF TECHNOLOGY PERFORMANCE
SPECIFICATIONS FOR VOLATILE ORGANIC COMPOUNDS
C. Purdy and W.E. Schutte, Office of Environmental Development and Waste
Management, U.S. Department of Energy, Germantown, Maryland; M.D. Erickson
(Environmental Research),, S.C.Carpenter (Chemical Technology), P.V. Doskey
(Environmental Research), P.C. Lindahl (Chemical Technology), and A.D. Pflug
(Chemical Technology), Argonne National Laboratory, Argonne, Illinois.
INTRODUCTION
The Office of Technology Development (OTD) within the Office of Environmental
Restoration and Waste Management of the Department of Energy has a mission to deliver
needed and usable technologies to its customers. The primary customers are individuals
and organizations performing environmental characterization and remediation, waste
cleanup, and pollution prevention at DOE sites. DOE faces a monumental task in cleaning
up the dozen or so major sites and hundreds of smaller sites that were or are used to
produce the U. S. nuclear weapons arsenal and to develop nuclear technologies for national
defense and for peaceful puposes (1,2). Contaminants and waste materials include the
radionuclides associated with nuclear weapons, such as plutonium and tritium, and more
common pollutants and wastes of industrial activity such as chromium, chlorinated
solvents, and polychlorinated biphenyls (PCBs). Quite frequently hazardous wastes
regulated by the Environmental Protection Agency are co-mingled with radioactive wastes
regulated by the Nuclear Regulatory Commission to yield a "mixed waste," which
increases the cleanup challenges from several perspectives.
To help OTD and its investigators meet DOE's cleanup goal, technology performance
specifications are being implemented for research and development and DT&E projects.
Technology performance specifications or "performance goals" describe, quantitatively
where possible, the technology development needs being addressed. These specifications
are used to establish milestones, evaluate the status of ongoing projects, and determine the
success of completed projects. All too common within proposals, statements of work, and
project milestones is a vague statement that the technology being developed will be "better,
faster, safer, cheaper." A comparison to available technologies and quantitative definitions
of the anticipated improvements are frequently lacking in proposals. Program managers
and customers have little indication of whether they are funding a needed improvement in
measurement technologies until the technology is actually developed, commercialized, and
implemented.
This pilot project has focused on chemical measurement technologies for volatile organic
compounds (VOCs) in soil and ground water. The project team has developed a model for
the development and implementation of performance specifications and has applied this
model to VOCs. The customers and their needs define the problems to be solved, a review
of relevant available technologies defines the baseline from which development must
progress, and quantitative performance specifications are identified on the basis of unmet
monitoring needs. These general performance specifications are then translated into
project-specific performance specifications in consultation with the principal investigator.
PERFORMANCE SPECIFICATION APPROACH
The general approach is to document what currently exists or is nearing completion and
compare that baseline to the customers' needs to identify the unmet requirements. These
83
-------
unmet requirements then are the basis for the technology development needs by the
investigators, to be addressed. The process needs to be quantitative where appropriate, to
steer project goals away from vague generalities like "faster" and toward specifics like
"reduction of labor requirements for Step A from 4 hr to 0.5 hr."
Performance specifications constitute a program management tool and must fit within an
overall program management structure. During the planning phase of a program,
performance specifications are needed to identify the research or demonstrations requiring
support. During the execution phase, performance specifications can help in the evaluation
of both proposals and ongoing projects. The success of completed projects can be
evaluated against the established performance specifications to provide a basis for
considerations about further funding, technology transfer, or implementation.
For an individual project, technology performance specifications will be used to establish
milestones, evaluate the status of ongoing projects, and determine the success of
completed projects. Preliminary performance specifications will be required in proposals
and will be evaluated as part of the proposal evaluation.
I. PERFORMANCE SPECIFICATION PROCESS
A. Baseline Technologies for an Environmental Research or
Waste Management Need
1. State the need of the customer. What problems need to be solved? What
are the boundaries of the need? (Limit the scope to a specific need to keep
the process focused). What are the current deficiencies? What are needs
versus desires? This information should be supplied in consultation with
customers where possible; interactions need to both quantify the need and
distinguish truly needed features and performance specifications from
desired, but not required, properties.
Discuss the potential impact of improved technologies. Where possible,
quantitatively estimate what can be done better, faster, safer, or cheaper.
This preliminary statement, prior to other work, is refined in Section I.
2. Describe the relevant technologies that can be applied to the customer's
need. The focus is on the customers' need, not the technique, so address
all viable alternatives, not only directly competing alternatives. (For
example, characterization by gas chromatography/mass spectrometry,
describe all VOC characterization technologies, not just the competing
related technologies). In addition to available technologies, address
current worldwide technology development and assess the probabilities of
addressing unmet needs and the schedule for availability of the technology.
For upcoming technologies, present only a brief overview of technologies
that are still at the basic research stage, will not be available within the
customer's time frame, or have a low probability of successful
implementation. If no technologies currently exist to solve this problem
state that
The text needs to include descriptions of technical alternatives, their status,
their availability (with a list of vendors), and any barriers to
implementation.
84
-------
3. Describe how the technology will be applied. What are the conditions of
the operating environment (environmental, space, health and safety,
operational constraints)? What are the time requirements (setup time,
measurement time, time to data availability, need for 24-hour operation.)
What are the user attributes (education/training, number of operators).
4. Tabulate quantitative performance specifications for current technologies
(including technologies under development that appear to be on a path
toward successful and timely implementation and availability to DOE; i.e.,
don't fund new research and development if something is already in the
pipeline). Since it is possible to consider specifications in dozens of areas
(technical performance, cost considerations, applicability, availability,
etc.), restrict the effort by limiting the list to the ten most important
specifications, cost should be one of the ten.
5. Document all of the above and provide statements about the quality of the
information (e.g., average of five vendors' specifications with similar
applications to DOE needs versus educated guess with no firm
information).
B. Customer Requirements
Revise and refine the customer need statement developed in Section LA
Address the following issues in more detail:
1. Identify problems that need to be solved. See Section LA for more
details.
2. Identify the customer(s). Name sites, sub organizations, individuals,
operable units, etc. where available.
3. Define quantitative requirements (e.g., regulatory drivers for detection
limits).
4. Define functional and operating requirements .
5. Assess the priorities of the need (e.g., critical versus desired for
convenience).
6. Assess the urgency of the need (e.g., a start date for projects, by which
commitments to technologies must be made).
7. Document all of the above.
C. Unmet Requirements
Compare baseline technologies versus customer needs and list unmet
requirements.
85
-------
D. Technology Development Needs
1. Discuss the unmet requirements that can realistically be met by the
proposed technology; discuss the quantitative improvements that can be
anticipated.
2. Tabulate quantitative performance specifications for the proposed
technology (i.e., improvements over baseline).
3. Assess the quality of estimates (e.g., proven in other applications; within
theoretical limits of techniques; best scientific judgment). Provide or
reference documentation where available.
4. Recommend priorities where appropriate, for addressing requirements.
These priorities should be based on the customers' priorities, the
probability of success, and the time frame for completion. Discuss all
considerations in the prioritization.
5. Identify expertise (organization, name, address, and phone) that can be
tapped to confirm the validity of the effort.
After these steps, the needs baseline is established, and the performance
specifications can be used as a program management tool to establish milestones,
evaluate the status of ongoing projects, and determine the success of completed
projects.
II. PERFORMANCE SPECIFICATION PROCEDURE
The development of performance specifications is a program management function.
To minimize bias in their preparation, the specifications must be prepared by a disinterested
individual(s); on the other hand, persons preparing performance specifications need good
technical skills and current awareness in the area. These skills are precisely those needed
by potential investigators, so conflicts are to be expected. After internal management
review, the draft should be critiqued by selected principal investigators both to provide
additional technical input and to educate the principal investigators on the process.
HI. USE OF PERFORMANCE SPECIFICATIONS
1 The following guidelines should be applied to the use of performance
specifications: The principal investigator should use the performance
specifications as a goal in developing work plans, quality assurance project
plans, and experimental plans and in assessing the progress of work.
2. The program manager should use the performance specifications as a
criterion for evaluating progress on the project. This will be especially
critical at reviews where decisions on continued funding are made.
The program manager should also refer to the performance specifications
in communications about the projected value of the technology to potential
users.
86
-------
3. Customers can use the performance specifications to assess the
applicability of the technology to their needs and can plan to transfer the
technology to their site application as it nears completion of development.
4. All involved parties (principal investigator, program manager, and
customer) must remember that performance specifications are subject to
change as new information becomes available (new research results,
change in programmatic goals or changes in the customer's needs). All
involved parties must initiate change whenever new information becomes
available that will change the basis for the performance specifications.
Changes in performance specifications can constitute a change in the statement
of work and thus necessitate a contractual change and possibly adjustments in
cost and schedule.
IV. APPLICATION TO VOC MEASUREMENT
Reviews of currently available technologies and ongoing research projects were prepared
on the following topics:
Gas Chromatograhy
Gas Chromatography/Mass Spectrometry (GC/MS)
Electrochemical sensors
Raman spectroscopy
Infrared spectroscopy
Fiber optics (ultraviolet, fluorescence, etc.)
As an example, the GC/MS review found the following:
Two manufacturers currently market field GC/MS instrumentation,
Bruker and Viking. Both systems are capable of operation under either line or
battery power. Both are relatively insensitive to environmental conditions, but
neither is light enough to be carried by a person. VOCs can be introduced to the
system by liquid injection, thermal desorption, direct air sampling (bypass GC),
air sampling by trapping on sorbent thermal desorption and purging and
trapping.
An older version of the Bruker instrument has been extensively used in the
field for VOC analysis with results generally both more accurate and more
precise than those from parallel fixed-base laboratory analysis, because loss
of analyte during storage and transport is minimized. Substantial cost
savings have been documented with field VOC analysis using field-portable
GC/MS.
Current research on instrumentation focuses on developing a customized
purge-and-trap/temperature-programmable GC for the instrument (Leibman
et al. Los Alamos National Laboratory); development of a microprocessor to
control the sampling system and software designed to integrate the sampling
and ion trap functions (Hemberger et al. Los Alamos National Laboratory);
and development of direct-sampling ion trap mass spectrometry technology
by Guerin and co-workers at Oak Ridge National Laboratory. On the last
project, no gas chromatography is provided for separation of compounds;
87
-------
the mass spectral discrimination is judged sufficient Sample introduction
systems have been developed for air and water. The Los Alamos and Oak
Ridge instruments have been field tested and are in various stages of
commercialization. Additional work is being conducted on development of a
person-portable (backpack) instrument by Meuzelaar and co-workers at the
University of Utah.
Work is in progress to integrate the initial findings of the reviews and the customer needs to
determine the unmet needs and thus the technology development requirements.
V. CONCLUSION
Performance specifications formalize an informal process to screen potential areas of
research and match technology development with customers' needs. Although the steps
outlined here may appear daunting, the process provides a structure for program
management. The benefits include early development of a linkage between ongoing
research and the customer's needs and also minimization of replication of research.
Much work remains on this project as we attempt to implement performance specifications
on technology development for VOC measurement. Application of this methodology to
other characterization technologies and remediation technologies is contemplated.
ACKNOWLEDGMENT
Work supported by the U.S. Department Of Energy, Assistant Secretary for Environmental
Restoration and Waste Management, Office of Technology Development, under contract
W-31-109-Eng.-38.
BIBLIOGRAPHY
(1) U.S. DOE Environmental Restoration and Waste Management Five-Year Plan, Fiscal
Years 1993-1997, DOE/5-0090P, August 1991.
(2) U.S. Congress, Office of Technology Assessments, Complex Cleanup: The
Environmental Legacy of Nuclear Weapons Production, OTA-O-484 (Washington, DC:
U.S. Government Printing Office, February 1991).
88
-------
•J5 THE QAPjP QUAGMIRE
Marcia A. Kuehl. President/Owner, M.A. Kuehl Company, 3470
Charlevoix Court, Green Bay, Wisconsin 54311
ABSTRACT
The Quality Assurance Project Plan (QAPjP) has become
quicksand for environmental investigations. As a regulator,
I reviewed them, as a contractor I wrote them, as a laboratory
director I tried to comply with them. As a taxpayer I am
appalled by how what was originally intended to be a document
used to ensure that data of adequate quality is obtained has
degenerated into a pound of paper that takes too long to
write, review, and revise. This paper will present the sticky
points in the writing of QAPjPs and how good science is
sacrificed for ease of QAPjP approval. The substance of
regulatory agency comments and concerns about QAPjP contents
and how they can be addressed will be relayed via horror
stories. The viewpoint of the laboratory and it's often lack
of involvement in the QAPjP will be offered. The new industry
of data validation will be reviewed and it's impact on the
QAPjP assessed. Suggestions for a creative overhaul of the
QAPjP elements themselves and ideas for streamlining the
development and approval processes will be proposed.
INTRODUCTION
December 29, 1980 marked the birth of the term Quality
Assurance Project Plan (QAPjP). The EPA Office of Monitoring
Systems and Quality Assurance created QAMS-005/80 "Interim
Guidelines and Specifications for Preparing Quality Assurance
Project Plans". And, like so many other "guidance" documents,
QAMS-005/80 soon became defacto regulation. All QAPjPs, if
they are to be approved, must follow these specifications.
Earlier, the EPA Administrator in May, 1979 required that "QA"
must be done, smarting from the lessons learned at Love Canal.
QAMS-005/80 prescribed that the primary way that QA was going
to be assured for environmental measurements was through this
QAPjP document. QAMS-005/80 painstakingly describes how to
number the pages but glosses over how specifically to relay
precision, accuracy, representativeness, comparability and
completeness (PARCC) parameters and the mysterious process of
data assessment. The QAPjP was originally intended to address
both the laboratory and field activities as part of the
Sampling and Analysis Plan (SAP). The field activities have
since been segregated in a stand alone document, the Field
Sampling Plan (FSP). Unfortunately, the PARCC parameters are
not carried over into the FSP as they relate to the field
investigation. Subsequent documents covering Data Quality
Objectives (EPA-540/G-87-003) and Category IV QAPjPs (EPA-
89
-------
600/8-91-006) were developed to provide additional aid. All
of these documents are intended to help in the production of
an "approvable" QAPjP so that the investigation can begin.
The quest for the approval signatures requires some
floundering in the quagmire.
WRITING THE QAPiP
Despite the fact that the format of the QAPjP is relatively
rigid, some creative writing is required to get it through the
approval process. Avoiding controversial points often results
in good science being sacrificed for ease of approval.
Specific target analytes or indicators are avoided because
they are non-routine and often the project team settles for
CLP Contract Required Quantitation Limits (CRQL) because it is
easier than finding/validating a method with a lower detection
limit. Use of an already approved QAPjP as a template with the
names changed (project, organizational chart, laboratory) is
helpful, but this practice perpetuates the QAPjP as a one-
size-fits-all pantyhose exercise that is not tailored to the
specific project needs.
The major issues that delay approval of the QAPjP are
identifying non-CLP laboratories to conduct the analyses, and
including non-routine analytes and/or methods. Any way that
these issues can be presented to soften their deviation from
the usual CLP boilerplate, such as citing other EPA accepted
references and internal guidance, can aid in quicker approval.
To CLP or not to CLP
The Contract Laboratory Program (CLP), first implemented in
1980, was never intended to be the defacto "certification" for
environmental laboratories that it has evolved into today.
This "certification" perception was a major finding of the
1991 EPA management review of the CLP. The perception is that
if a laboratory is participating in the CLP, it has jumped
through enough documentation, facilities audits, and QC hoops
to "guarantee" error free data. This perception is not only
unfair to those laboratories that for economic reasons choose
not to pursue CLP contracts, it is untrue. Granted, the data
package documentation that has taken 13 years to refine has
withstood legal and independent data validator scrutiny,
however non-CLP laboratories are also capable of providing the
same defensible data packages to meet requirements.
It appears that when a non-CLP laboratory is named in a QAPjP,
the degree of scrutiny by the approving agency increases. An
on-site audit is triggered and often blind performance
evaluation (PE) samples will be sent prior to the on-site
audit. It is ironic that despite this increased surveillance
90
-------
of non-CLP labs cited in QAPjPs, only CLP labs have ever been
suspended or disbarred by the EPA. The QAPjP author can help
head off an unnecessary round of comments by supplying the
following information in the first draft:
- Lab certifications/state approvals
- Dates of independent on-site audits
- Blind PE studies lab participates in
- Experience of lab with Level IV data packages
Offers to provide written audit reports from other state
and/or federal agencies and results of the last blind PE study
should be made at the pre-QAPjP meeting.
To SAS or not to SAS
Active CLP labs are not immune to scrutiny, especially if non-
traditional methods (those not included in the CLP Statement
of Work) or lower detection limits are needed for the project.
These conditions require a Special Analytical Services (SAS)
request if the project is under the auspices of an EPA
contract. The SAS necessitates that the author supply a fully
documented Standard Operating Procedure (SOP) with QC
specifications for number, type, and limits of internal QC
samples. This is usually a Catch 22 proposition, as the
analysis is "special" because there is no SOP, let alone
standardized limits. Even if the QAPjP is not for a federal
lead project, the proposed laboratory must supply the details
of what will constitute the analysis method and acceptance
criteria.
It is imperative that the lab method SOP submitted for non-
traditional analyses contain the following items, or an
additional submittal will likely be requested:
- Method specific QC (frequency and limits)
- Basis for limit determination
- How detection limit was determined
- Supporting data for detection limit
One of the most frustrating issues in QAPjP authoring is
centered around the SAS. Often, reviewers with little or no
laboratory experience will be tasked with determining the
appropriateness of the proposed SOP submitted. The shortage of
bench chemists who desire to switch gears to review paperwork,
and the dynamic nature of environmental analytical chemistry
contribute to the lack of QAPjP reviewers with hands-on lab
experience. The most successful tactic to deal with this is
to ask to have access to the agency reviewer and agency lab
staff set up a conference with the proposed lab's analytical
team and draft an acceptable SOP together.
91
-------
OAPiP REVIEW COMMENTS
Written comments are received from the agency reviewers and
assessed by the project team for whether the QAPjP would need
a substantive change affecting project scope, budget and/or
schedule. Over the course of 13 years of authoring QAPjPs and
responding to comments, I have compiled some requested
revisions that illustrate the concept of losing the forest for
the trees by focusing on minute technical details that do not
make the project data more defensible:
The Case of the 5 Point Curve
The reviewer insisted on an initial 5 point curve for an
indicator analyte that was expected at high concentrations.
The lab SOP detailed a 4 point curve and an independent check
standard at a fifth concentration analyzed every 20 samples.
Heated 2 hour conference calls over the linearity of the curve
near the detection limit and the semantics of the SW846
terminology of "minimum calibration" requirements against
agency "requirements" ensued while leachate from the site ran
into the river. The gap between this PRP QAPjP submittal and
start of the field work was 15 months. The lab rolled over
and rewrote their SOP to include an additional calibration
curve concentration near the detection limit and dedicated a
project "babysitter" to ensure that the analysts added the
extra curve concentration. The detected concentrations in the
samples were so high that a 3 point curve would have easily
sufficed as the range component of the calibration function
was the driving issue for the project data, not the
sensitivity.
The Case of the Best Professional Judgement Holding Time
A SAS request was prepared using a USGS reference method. As
no holding time was established by the USGS in their method,
the agency reviewer disregarded the 28 day holding time
suggested for the same analyte in an EPA method, and used
"best professional judgement" to require that the sample be
analyzed within 48 hours of collection. The process for SMO to
solicit labs that would even bid on the work took months.
After the project was finally underway, the analysis could not
be consistently accomplished within the 48 hour constraints,
and resampling was done. Out of intellectual curiosity,
samples were analyzed again within 28 days and results were
not significantly different. The mythical holding time cost
the project 3 extra days field time at an estimated cost of
$6,000 and delayed the project start by 2 months.
92
-------
Top Secret SOPs
Some laboratories feel that their analytical SOPs are so
revolutionary and innovative that they consider them
confidential, proprietary business information. Evidently
there are many complex nuances to these analyses that another
lab learning a helpful hint would affect competitive pricing.
As proprietary information, these labs will only submit their
SOPs directly to the reviewing agency and not to the QAPjP
author. The perception is that there is a lucrative black
market for lab SOPs and the QAPjP author is selling them on
the street corner to other labs too lazy to write their own,
let alone follow the reference method. What is ironic is that
for any federal lead project, a citizen can request all
pertinent laboratory data and SOPs under the Freedom of
Information Act. Being caught between the reviewing agency
questioning the details of an SOP one is not allowed to see,
yet being held responsible for revising the SOP and using to
assess the project data for compliance is a dilemma for the QA
professional.
OAPiP COMMENT HALL OF FAME
QAPjPs have been revised to satisfy the following
"substantive" comments:
1. References to data or locations should be changed to read
"in" Table X/Figure Y.
2. In the organizational chart, the line from the Project QA
Officer to the Project Manager should be a dotted line,
and the agency QA Officer box should be equal in size and
location to the agency Project Manager's box.
3. The procedure for calibration and maintenance of the field
trailer thermometer/barometer should be included.
4. Sample container cleaning protocols are not adequate to
ensure contaminant-free bottles. (The SOP attached to the
QAPjP was the agency's Scope of Work from it's currently
awarded bottle contract, only retyped.)
Once a revision to a QAPjP is made and resubmitted for review,
it is not unusual to have a different reviewer assigned to it
and the "do not pass go, do not collect $200" review process
starts all over again.
THE MISSING LINK
More often than not, the lab that must follow the QAPjP is
left out of it's creation entirely, or consulted at the last
93
-------
minute before submittal. An approval signature from the lab's
QA and managerial personnel was never included in QAMS-005/80.
For those QAPjPs involving the federal CLP lab pool, this
would be burdensome, but for PRP and state lead projects,
review and approval involvement is a necessity, especially if
the QAPjP author is from an independent engineering firm.
This review/approval also is necessary for the lab to
appropriately price the analyses and advise whether they have
the capacity to produce the deliverables in the turnaround
time allotted in the project schedule.
Once a candidate lab is selected and preliminary contact is
made, the QA representative from the lab should be involved in
project scoping and DQO negotiations. Attendance at the pre-
QAPjP meeting is usually invaluable. This time is not
perceived as profitable as lab analyses in the bottom line,
but the angst this upfront planning will save later is well
worth it.
Being placed in the middle of philosophical arguments between
the agency chemist/reviewer and the lab on issues like the
real utility of matrix spikes and lab duplicate control
statistics is not a pleasant experience. Trying to relay
chemical "technospeak" is frustrating and direct contact
between the factions with the QAPjP author as mediator is the
most efficient way to resolve these issues to get the QAPjP
written and submitted.
DATA VALIDATION
It is not enough to specify the methods, QC, calibration,
reporting forms, analytical standards, and data package
contents/order. The resultant data package must often be
assessed by an "independent" data reviewer to look for
analytical errors or take issue with "best professional
judgement" calls. The new industry of data validation boomed
after the reported discovery of fraud in several laboratories,
yet it is doubtful that without an on-site inspection by a
skilled computer guru and chemist with lab experience, that
the anomalies would have been caught in data packages during
independent validation.
Data validation itself is an inexact science. Most validators
defer to the National Functional Guidelines, but even these
documents allow for interpretive dancing by the validator in
the areas of the effects of out of compliance calibration
compounds, matrix spikes and field precision data. One of the
most disconcerting positions to be in is to be the laboratory
QA fall guy when the data validators assess the data using
criteria you were never made aware of - and nonpayment of the
analytical bill is threatened by the QAPjP author. The
94
-------
criteria to be used in the validation process should be stated
in the QAPjP, along with who/what firm will accomplish it.
The validator's authority, responsibility, and communication
protocols should be clearly defined to control requests to the
lab for resubmissions and additional information. Data
validators should be involved early in the project planning to
become educated in the project goals and expected analytical
challenges. Validating analytical data in a vacuum without
regard to balance between the intended use of the data and
definitive precision and accuracy of each number is only an
exercise in accounting.
OAPiP MAKEOVER
The 16 element QAPjp contains at least 9 sections that have
evolved into boilerplate. The meat of the QAPjP is the table
of proposed samples, analytes and methods generally found in
Section 5.0. Listed below is each section of the current
QAPjP format and suggested changes to each to streamline the
writing and approval processes:
Section 1.0 Title Page: Add provisions for signatures for
Laboratory Director, Laboratory QA Manager, and data
validation firm.
Section 2.0 Table of Contents: Change to reflect
consolidated sections as described below.
Section 3.0 Project Description: Expend effort to organize
clearly the available historical data and analytical methods
used, indicate historical data quality assessment and issues.
Section 4.0 Project Organization and Responsibility: Add the
authority and responsibility of the data validator including
reporting, lines and methods of communication and the process
for interaction with the laboratory.
Section 5.0 QA Objectives for Measurement Data in Terms of
Precision, Accuracy, Representativeness, Comparability and
Completeness: Reduce this section to the table of samples,
analytes, methods and internal/external QC samples and limits.
Highlight any special detection limits and/or method
restrictions. Delete canned text boilerplate definitions of
PARCC. The comparability descriptions and objectives are not
quantifiable, and it remains to be seen what magic limit of
percent completeness constitutes an acceptable level. The
whole issue of field data quality objectives needs to be
recognized and defined for the project in this section.
Section 6.0 Sampling Procedures: This section usually defers
to a separate site specific sampling plan, which is generally
95
-------
reviewed separately-. The sampling procedures, especially the
proposed sample collection containers, volumes, and
preservatives should be reviewed by the lab to ensure that the
specifications are compatible with the lab analyses.
Section 7.0 Custody: No earth shattering revolutions in
chain-of-custody protocols has transpired over the last 13
years. This section is filler boilerplate and can be handled
by reference to the sample collection plan and laboratory
custody SOP-
Section 8.0 Calibration Procedures and Frequency: This
section usually refers to the specific lab methods and should
be deleted and contained in section 9.0.
Section 9.0 Analytical Methods: Lab specific SOPs should be
attached or the CLP SOW cited. Any known project specific
deviations or contingencies should be mentioned.
Section 10.0 Data Reduction, Validation and Reporting: The
data validation firm should supply their proposed SOP for
validation and aid in the preparation of this section. The
usual rehash of how to calculate RPD, % recovery, etc. is
unnecessary filler.
Section 11.0 Internal QC Checks: This section is not needed
as it should be included in the table of
samples/analytes/methods/QC in section 5.0. Any lab specific
terms used should have been defined in footnotes or text in
section 5.0.
Section 12.0 Performance and Systems Audits: This section is
more canned boilerplate unless a site specific field and/or
lab audit is proposed. Delete this section and include in
section 4.0.
Section 13.0 Preventive Maintenance: Delete this section as
the analytical method or field SOP should cover it. Any field
meter maintenance should be included in the field sampling
plan.
Section 14.0 Specific Routine Procedures to Assess Data
Precision, Accuracy, and Completeness: This section repeats
the table in section 5.0 and generally contains pages of
definitions and verbiage on how to calculate items such as
mean and standard deviation. Back in 1980, QAMS-005/80 noted
that "Recommended guidelines and procedures to assess data
precision, accuracy and completeness are being developed."
We're still waiting.
Section 15.0 Corrective Action: Who can stop/restart lab
96
-------
analyses under what conditions is a major area of concern to
project budgets and schedule. Section 4.0 could include the
flow of action to out of control situations.
Section 16.0 QA Reports to Management: Usually an elegantly
written section that says little, it should contain the
specifics of the project QA summary and who will write it.
SUMMARY
In the current system, for less than a year approval time from
QAPjP submittal to approval, authors should use an active,
upstanding CLP lab to do only routine CLP SOW regulated
analyses on nice clean soil and water samples. The utility of
full Level IV data packages being reviewed by a third party
data validator firm should be evaluated for the project.
An overhaul of the QAPjP purpose and process is needed. An
injection of the lab personnel and data validators early in
the project planning and evidence of their commitment by
signatory approval of the QAPjP is critical to the success of
the project. The perceived dismal record of Superfund
investigations completed and Record of Decisions signed and
cleanup accomplished is testimony to this need for change. A
task force consisting of lab personnel, engineering firms,
data validators and risk assessors should be initiated by the
EPA Quality Assurance Management Staff. As Congress now calls
for Superfund to be "fixed", the logical place to start is
it's infrastructure, the QAPjP.
REFERENCES
U.S. EPA, December 29, 1980. Interim Guidelines and
Specifications for Preparing Quality Assurance Project Plans,
QAMS-005/80.
U.S. EPA, March 1987. Data Quality Objectives for Remedial
Response Activities; Development Process, EPA 540/G-87-003.
U.S. EPA, February 1991. Preparation Aids for the Development
of Category IV Quality Assurance Project Plans. EPA-600/8-91-
006.
U.S. EPA, Draft December 1990, Revised June, 1991. National
Functional Guidelines for Organic Data Review.
U.S. EPA, October 1989 Revision. Laboratory Data Validation
Functional Guidelines for Evaluating Inorganics Analysis.
U.S. EPA, October 1991. "Management Review, Contract
Laboratory Program".
97
-------
16
EVALUATION OF METHODS FOR DETERMINATION OF
INSTRUMENT DETECTION LIMITS
Meszaros, T.J., Technical Support Supervisor, Dobb, D.E., Scientific
Supervisor, Organ. B.D.. Senior Scientist, Converse, J.C., Data Audit
Technician, Lockheed Environmental Systems & Technologies Company, Las
Vegas, Nevada; Russell McCallister, U.S. Environmental Protection Agency,
(EPA) Environmental Monitoring Systems Laboratory, Las Vegas, Nevada.
ABSTRACT
Instrument detection limits (IDLs) that are used in the EPA Contract
Laboratory Program (CLP) are defined in the current Statement of Work
(SOW), and are reported by CLP laboratories each quarter. The SOW
requires that CLP laboratories perform seven analyses for each analyte
(except cyanide) on three nonconsecutive days; determine the standard
deviation for the analyses on each of those days; and sum the three
standard deviations. This procedure is labor-intensive, requiring a full
week before results are obtained. Questions have been raised regarding
applicability of the results over an entire quarter. In an effort to
obtain more current IDL information, the authors evaluated the daily IDLs
obtained over a 3-month period and compared their results to the contract
required detection limit (CRDL) and the quarterly IDL. Using data
submitted by CLP laboratories, IDLs were calculated using continuing
calibration blanks (CCB) and standards at two times the CRDL (CRA for
atomic absorption analytes and CRI for inductively coupled plasma
analytes). The daily method calculates IDLs from existing QA/QC data and
analyses dedicated to just IDL determinations are not necessary. We
compare the results from the analyses using the alternate daily methods to
results from the analyses using the quarterly method defined in the SOW.
These data may be useful for determining the applicability of the SOW
method of IDL calculation.
Notice: Although the research described in this article has been funded
wholly or in part by the United States Environmental Protection Agency
through contract number 68-CO-0049 to Lockheed Environmental Systems and
Technologies Company, it has not been subjected to Agency review and
therefore does not necessarily reflect the views of the Agency and no
official endorsement should be inferred.
98
-------
17 TECHNICAL INFORMATION MANAGEMENT FOR THE PIT 9 PROJECT
IDAHO NATIONAL ENGINEERING LABORATORY
J- Owens, Senior Scientist, Chemical Services, EG&G Idaho, Idaho Falls, ID 83415,
D. Forsberg. Senior Engineering Specialist, Buried Waste Program - Pit 9 Project,
EG&G Idaho, Idaho Falls, ID 83415,
D. Macdonald. Program Manager Buried Waste Program, Department of Energy Idaho
Falls Field Office, Idaho Falls, ID 83415.
ABSTRACT
This paper describes the development of an access controlled data system that
provides group-wide accessibility of laboratory and field information in support
of the Pit 9 cleanup at the Idaho National Engineering Laboratory. Pit 9 was
selected for remedial action in accordance with the Comprehensive Environmental
Response, Compensation, and Liability Act (CERCLA) pursuant to the Idaho National
Engineering Laboratory Federal Facility Agreement and Consent Order. Pit 9
contains approximately 150,000 ft3 of buried waste, of which approximately
110,000 ft3 of transuranic waste was generated at the Rocky Flats Plant.
The Pit 9 remedial action consists of three phases: (1) Proof-of-Process Test
(POP), (2) Limited Production Test (LPT), and (3) Final remediation. Data
produced during POP will be used to evaluate subcontractor's performance as part
of the procurement process, and to provide the Environmental Protection Agency
(EPA) and the Idaho Department of Health and Welfare (IDHW) with data justifying
use of the proposed process for remediation of Pit 9. It is in the best interest
of the subcontractors and EG&G Idaho that POP data be identified as reliable,
accurate, and verifiable by the U.S. Department of Energy, EPA, and IDHW. The
LPT phase would demonstrate that all integrated systems function as proposed to
give a high degree of confidence that all systems are reliable before full-scale
remediation would begin. The data generated during full remediation will be used
as evidence that applicable or relevant and appropriate requirements described
in the Pit 9 record of decision had been met.
EG&G Idaho feels the credibility and legitimacy of data produced during the
remedial action can be greatly improved by independent data management. EG&G
Idaho will provide another layer of data legitimacy through electronic links to
subcontractor's laboratory instruments and by using an independent laboratory
information management system (LIMS). EG&G Idaho will independently store the
laboratory instrument signals, the signal processing software, and the quality
assurance records/documentation associated with sampling and/or analysis to an
optical disk for future analysis, if required, to verify objectives have been
met.
Data and information that supports the 3 phases of the Pit 9 remedial action
will be received in many forms such as photos, images, graphs, video-text, audio-
text, field notebooks, and electronic media. The system was developed on a
VAX/WMS using Standard Query Language (SQL) Multimedia for rdB/VMS which is a
solution for manipulating, storing, and retrieving objects from client
applications. Storing multimedia objects in a database provides data integrity.
concurrent access, security, standard storage format, and the ability to share
data across applications. The SQL/services provide the ability to interface to
the following environments: VMS, MS-DOS, Microsoft Windows, and Macintosh
systems. The system will provide for storage on optical media using Write Once
Read Many (WORM) technology to provide long term storage, and will provide the
necessary tools for rapid manipulation of the data to controlled user community
long term. Ultimately, the purpose of this system is to provide a secure,
historical data repository for multimedia technical information for the Pit 9
project.
99
-------
1 p QUALITY ASSURANCE PLANS FOR BASIC RESEARCH
' ° AND NEW CONCEPT STUDIES
EJ. Poziomek, project manager, and A. Cross-Smiecinski, quality assurance officer, Harry
Reid Center for Environmental Studies1, University of Nevada, Las Vegas, 4505 S.
Maryland Pkwy., Box 454009, Las Vegas, Nevada, 89154-4009.
ABSTRACT
Writing quality assurance project plans (QAPjPs) for basic research and new concept studies
presents a major challenge to scientists and engineers. Due to the nature of basic research
studies, few functions are followed repetitively in exactly the same manner, at least in the
initial stages of the study and sometimes throughout the study. This often precludes the use
of standard methods, operating procedures, the ability to designate appropriate corrective
action, even planning complementary quality control checks, and other topics usually
addressed in a QAPjP. Recently we were challenged with writing just such a QAPjP for a
set of studies. It was clear that we needed to maximize the flexibility of EPA QA
requirements, and so we chose to try the Category IV guidelines prepared by Guy Simes of
the Risk Reduction Engineering Laboratory in Cincinnati. Because the studies all fell into
the realm of field screening methods involving proofs of principle and new concepts, we
addressed them together in the first version of the QAPjP. We soon found that many of
the QA topics in those guidelines were still difficult to address. We resolved that the
QAPjP be written with statements of commitment that, as progress is made in the research
studies, the "holes" in the QAPjP would be filled and eventually, as appropriate, the studies
would be broken out into individual QAPjPs. We also put a major emphasis on
recordkeeping in order to ensure that each experimental step, observation, thought, and
action be recorded in such a way that the research can be thoroughly understood, reported,
and repeated. The purpose of our poster presentation is to provide a detailed explanation
of how we prepared the QAPjP discussed and to give examples of how other research
studies can address QAPjP requirements.
INTRODUCTION
The Harry Reid Center for Environmental Studies (HRC) has a number of basic research
and new concept studies, the sponsors of which include the U.S. Environmental Protection
Agency (EPA) and the Advanced Research Projects Agency (ARPA). These require
preparation of Quality Assurance Project Plans (QAPjPs). If the QA requirements are not
1 Although the information in this paper has been funded wholly or in part by the U.S.
Environmental Protection Agency under Cooperative Agreement No. CR 818353 with the
University of Nevada-Las Vegas, it does not necessarily reflect the views of the Agency, and no
official endorsement should be inferred. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use. Partial support was also provided by the
Advanced Research Projects Agency under Contract DAAD05-92-C-0027.
100
-------
specified, HRC still has a policy of the need to define data quality objectives (DQOs) for
its research and new concepts studies. In the process of trying to meet standard QAPjP
requirements (U.S. EPA, 1983) to these types of basic research and new concept studies,
it was discovered that many questions and objectives could not be established beforehand;
they could not necessarily be anticipated. The QAPjP requirements were difficult and in
some instances not possible to apply to research studies. In searching for Agency-accepted
QAPjP guidance, some of the clearest, most concise, reasonable guidance devised was
discovered in the form of the graded approach (Simes, 1991) written by Guy Simes at the
U.S. EPA Risk Reduction Engineering Laboratory. This groundbreaking set of guidance
documents introduced the concepts that 1) not all projects can meet quality needs in the
same manner, and 2) QAPjP requirements are more flexible and therefore more supportive
of study objectives than often realized.
This paper describes HRC's adaptation of the graded approach of QAPjP requirements to
basic proof-of-principle and new concept research studies. It is meant to provide a basic
template to assist investigators in writing QAPjPs.
SUMMARY
The graded approach to composing QAPjPs has four levels of requirements based on the
category of study or project represented. Level I requirements are the most rigorous
because the projects they address are critical to Agency goals and must withstand legal
challenge (Simes, 1991). In comparison, Level IV requirements, the least rigorous of the
four levels, address studies whose results are used to assess suppositions (Simes, 1991). A
comparison of QAPjP requirements is listed.
Standard QAPjP requirements Level IV QAPjP requirements
(U.S. EPA, 1983) (Simes, 1991)
Project Description Project Description
Organization and Responsibilities
QA Objectives QA Objectives
Site Selection and Sampling Procedures
Sample Custody
Calibration Procedures and Frequency
Analytical Procedures Sampling and Analytical
Data Reduction, Validation, and Reporting
Internal QC Checks Approach to QA
Performance and Systems Audits
Preventive Maintenance
Calculation of Data Quality Indicators
Corrective Action
Quality Control Reports to Management
References References
Other Items Other Items
Difficulty in specifying QA Objectives, parts of the Project Description, and a complete
Approach to QA/QC presents problems due to the unavailability of data in new concept
101
-------
and basic research studies. The solution to this problem lies in the commitment of the
research team to document requirements as answers are obtained and to state intents within
the QAPjP. The attached EXHIBIT is a condensed version of how HRC met the Level IV
QAPjP requirements for one set of basic research studies and commits to the requirements
for which there currently are no answers. Hopefully, research investigators can benefit from
the information and our experience.
102
-------
EXHIBIT
For an explanation of what is needed for each requirement of a Level IV QAPjP, refer
directly to Simes (1989) for guidance. Presented here is a description of how HRC
composed the QAPjP entitled "Proof-of-Principle and New Concepts Studies in Chemical
Sensors, Biosensors, and In Situ Methods for Field Screening." The italicized print indicates
guidance and notes for consideration in composing the missing portions of the QAPjP.
SECTION 1.0
PROJECT DESCRIPTION
1.1 GENERAL OVERVIEW
Field screening methods for hazardous waste site investigations need to be rapid and low
cost to support on-site monitoring and characterization activities. A variety of in situ and
field portable analytical methods are needed for the analysis of contaminants in all types of
environmental media. Needs for environmental sensors also go hand-in-hand with needs
for sampling devices and methods. This project is meant to address these requirements.
There are major barriers which, if overcome, would provide new capabilities in selectivity
and sensitivity of devices for field screening. This applies to both point and remote sensing
applications. For example, the major technology barrier to the development of fiber optic
and other sensors based on chemical/biochemical reactions and/or molecular association
effects is the proper selection of the sensor coating materials. In most cases there is also
a challenge to marry the chemistry to the particular device being used. There is a need to
develop small and low power systems while maintaining selectivity and sensitivity. There
are a variety of opportunities to exploit including the use of solid phase sensor chemical and
biochemical coatings, self-indicating films, solid-state extraction disks in combination with
spectroscopy, ion mobility spectrometry as an in situ well monitor, spectroelectrochemistry,
and technology integration, to mention a few.
1.1.1 The Process
There are two parts to this project. The first involves performing applied research on field
screening and measurement methods relevant to the type of pollutants most frequently
encountered at hazardous waste sites. This is an investment for the long term and is meant
to address the needs for basic information for proof-of-concepts. The second part is for the
short term and involves integration of existing technologies in the search for new concepts.
The strategy also includes reviewing emerging research and technology, and transferring
appropriate aspects to development of methods for field screening.
Meeting field screening needs presents a major challenge because of the many scenarios
which could be encountered involving any combination of air, water, or soil. In addition,
the possible analytes, specifically their number and nature, can present a major challenge
103
-------
even to the most experienced chemist. Organic chemicals alone - volatile, semi-volatile, and
non-volatile - could range from several to hundreds in a particular sample.
Field screening systems, whether they use microelectronics, electrodes, or fiber optics as
transducers, i.e., the physical probe, require chemistry to be worked out for the specific
chemical or class of chemicals to be monitored and measured. Many times not only are
there no guidelines for practitioners for a particular application, but there is also no
information base on the fundamental chemistry. For example, major barriers still remaining
to be solved in recognition coating selection and evaluation include:
• Lack of an information base on reactions and molecular association effects
in vapor-solid and liquid-solid phases which can be drawn upon for
development of sensor coatings.
• Lack of guidelines on choosing selective coatings for use in conjunction with
sensors.
• Lack of standard or reference methods for screening and evaluating
candidate coatings in sensor applications.
1.1.2 Statement of Project Objectives
The general objective of this project is to design and develop new field screening methods,
including those based on in situ processes. Priority is given to systems which offer simplicity,
selectivity, sensitivity, and low cost. One focus will be to overcome the technology barrier
posed by nonselective recognition coatings in sensor monitoring and measurement needs.
The output can be applied to a variety of sensors utilizing molecular recognition events
based on reactions or molecular association effects. Another focus will be to integrate
existing technologies toward the development of new concepts in field screening. Listed
below are specific objectives for some of the studies under this project taken from research
and study plans. New and updated project objectives will be documented in revisions of this
QAPjP and/or in study-specific reports/plans as the objectives develop.
Potential Use of Ultrasound in Chemical Monitoring--(Research Plan, M.S. Thesis in
Environmental Chemistry, University of Nevada, Las Vegas: Grazyna Orzechowska,
November 1992)
The major objective of this study is to examine the potential of combining sonication
with existing multiparameter monitors such as chloride specific electrodes in detecting
organochlorine pollutants in water. This is an example of technology integration.
The approach in using sonication is obviously applicable to compounds other than
organochlorine ones; these could contain inorganic elements (other halides, phosphorus,
nitrogen, and sulfur). Anions specific to the inorganic components would be produced in
sonication. Changes in anion concentrations before and after sonication would be used in
monitoring for the pollutants. The choice of organochlorine compounds was made because
these compounds are the most common pollutants found at hazardous waste sites. Success
104
-------
with compounds such as trichloroethylene, chloroform, etc., will serve as proof-of-principle
and form a base for expanding the research to other pollutant classes.
Surface Acoustic Wave (SAW) Biosensor for Detection of Environmental Pollutants: Kim
Rogers, 1992
Objectives of this study are to: 1) demonstrate the feasibility of a receptor- or
antibody-based biosensor which can detect a single compound or several classes of
compounds of environmental concern; and 2) develop protocols which will explore and
define the applicability (i.e., limits of detection, sample matrix and preparation
requirements, etc.) of this biosensor for field assays. In this case the recognition coating
may be a chemical one or a biologically-derived material.
1.2 EXPERIMENTAL DESIGN
Experimental designs for many of the studies under this project are under
development and will be further documented in revisions of this QAPjP and/or in study-
specific reports/plans (as the Project Manager deems necessary) and as the designs are
finalized. Listed here for the benefit of the reader/reviewer are current approaches for the
specific studies addressed in the project objectives section.
Potential Use of Ultrasound in Chemical Monitoring-(Research Plan, M.S. Thesis in
Environmental Chemistry, University of Nevada, Las Vegas: Grazyna Orzechowska,
November 1992)
As mentioned in the project objectives section, the study is meant to examine the
potential of combining sonication with existing multiparameter monitors in detecting
organochlorine pollutants in water. Success will establish a proof-of-principle for obtaining
more detailed data and for using the concept with other organic pollutant classes. There
are a number of factors which influence the effectiveness of sonication in a chemical
reacting system. These will be taken into account in designing the experiments. Examples
follow.
Choice of the solvent—Factors such as molecular cohesion and surface tension of the
solvent can affect cavitation. There is no solvent choice in the proposed research since the
application is in water. This will be the solvent.
Reaction temperature—It is common practice in sonochemistry to use the lowest
possible reaction temperature consistent with reasonable reaction time. Keeping
temperature low (thus reducing solvent vapor pressure) minimizes "cushioning" effects which
could reduce the temperatures and pressures in cavitation. The temperature will be kept
close to room temperature using a cooling bath.
Irradiation frequencies-Frequencies between 20 and 50 kHz are normally used in
sonochemistry. A Branson Sonifier - Cell Disrupter model 101-063-200 with 20 kHz and an
output power of 400 W will be used for this work.
Choice of disruptor horns-Disruptor horns (probes) transmit ultrasonic energy into
a solution. The dimensions of the horn and the ultrasonic output power setting determine
105
-------
the amount of amplitude (tip movement) and degree of ultrasonic activity in the liquid.
Generally, the smaller the tip diameter, the higher the amplitude. Larger tip diameters have
less amplitude but can accommodate larger volumes.
Three horns have been chosen for this work, a 3/4" diameter high gain horn with
solid tip, a high intensity 1" diameter cup horn, and a tapered 1/8" diameter microtip horn.
This will allow some flexibility in designing the experiments. The tapered microtip horn, for
example, will have an amplitude three and a half times greater at its end than a standard
horn. It can be used with small solution volumes (1-2 mL).
Other sonication equipment--A stainless steel continuous flow attachment with a
water cooling jacket will be available for continuous processing of test solutions.
Though ultrasound is above the audible range of the human ear, mechanical noise
occurs when liquids are treated ultrasonically. A soundproof box will be available to reduce
this noise.
Surface Acoustic Wave (SAW) Biosensor for Detection of Environmental Pollutants--(2%
Initiative Study Description: Kim Rogers, 1992)
A number of basic parameters will be considered in the research approach to
develop sensor recognition coatings. These include choices of the coatings and the test beds.
Choice of candidate recognition coatings-Several considerations have to be taken
into account when choosing a recognition coating for sensor applications. First, one must
decide whether a reversible or irreversible effect is to be sought, then the chemistry of the
coating has to be considered, for example:
• chemical reaction vs. molecular association mechanisms
• covalent bond formation vs. weak Van der Waals interaction
• hydrogen bond vs. charge-transfer vs. other donor-acceptor mechanism
• bimolecular vs. catalytic reaction mechanisms
The strategy for biosensors is to seek coatings based on known reaction chemistry
that occurs in solution. However, a different approach may be needed for gas-solid and
vapor-solid configurations in sensor reaction chemistry. Factors that may affect sensor
performance include:
coating adhesion to the surface
mass vs. mechanical sensor coating mechanism
surface vs. bulk coating sensor effects
gain or loss in coating mass with analyte sorption
boiling point of the coating material
crystalline solids vs. amorphous
Factors such as solubility and molecular weight will be considered in selection of the
candidate coating material. Though the physical and chemical properties of a coating may
seem ideal, the material may not form a film conducive to sensor applications. Some
experiments may need to be designed by trial and error.
106
-------
Choice of test bed (physical transducer)--Test beds for examining the candidate
recognition coatings will utilize SAW probes and quartz crystal microbalance (QCM)
sensors. These devices monitor mass changes.
The equipment for measuring SAW sensor effects include a 158 MHz electronic
system, a 4-channel SAW data acquisition system, and a VG-400 automatic vapor generation
system. This sensor is able to detect low ppb levels with a selectivity of 1,000:1 or more.
The QCM is 12 MHz and somewhat less sensitive. However, it is useful for preliminary
experiments.
Temperature of the reaction-The test system will be maintained at constant
temperature, close to ambient conditions for the initial experiments. Different temperatures
using the same coating may be considered to increase selectivity.
Analytes to be tested—In the biosensor research organic analytes such as PCBs will
be chosen. The results of the study with PCBs will be extended to other organic analytes.
The biosensor study will address two main challenges: first, to immobilize an active receptor
or antibodies to piezoelectric devices and, second, to develop methodology which will allow
for sensitive and reversible detection of environmental pollutants.
1.3 SCHEDULE
The technology integration activity utilizing sonication is scheduled to run for two
years with a goal to demonstrate proof-of-principle.
The biosensor effort is also a two-year effort with a goal in the first year to construct
a flow cell and to develop assay protocols for detection of the target compounds. At the end
of two years, it is expected that standard assay protocols will be established and detection
limits determined for target analytes. Matrix effects and sample preparation requirements
for "real world" samples will be determined.
1.4 PROJECT ORGANIZATION AND RESPONSIBILITIES
This subsection, since it is self-explanatory, is eliminated to save space in these
proceedings.
SECTION 2.0
QUALITY ASSURANCE OBJECTIVES
2.1 QA OBJECTIVES
Project objectives have been discussed in Section 1.0. Because the research
performed under this project is basic proof of principles and new concepts investigations,
specific QA objectives are not specified at this time. Initially the data will be evaluated
qualitatively or semi-quantitatively. The data will be used to make decisions whether to
proceed further in evaluating the field screening methods developed by this project and to
evaluate proofs and concepts. A future goal of the project is the development of complete
107
-------
QA objectives for specific activities once proofs-of-principles have been determined and
concept validation is initiated involving environmentally-related measurement data.
2.2 QUANTITATIVE QA OBJECTIVES: PRECISION, ACCURACY, METHOD
DETECTION LIMIT, AND COMPLETENESS
See Section 2.1. Specific quantitative QA objectives are not specified at this time.
As they are developed, they will be included in reports and/or revisions of this document.
Listed below in italics are self-notes of intention and guidance to be considered when the
answers are available. Much of this guidance has been taken verbatim from a Simes
document (Simes, 1989, 1991).
2.2.1 Precision
It is believed that it is possible to estimate precision from duplicate determinations when
the project gets to the stage where measurements are being made.
2.2.2 Accuracy
Likewise accuracy may be determined from measurements of a known. For example,
if the field screening method is applied to an SRM and gets an 80% recovery, this is an
indication of accuracy.
2.2.3 Method Detection Limit
2.2.4 Completeness (Simes, 1991)
In addition, it must be explained how the QA objectives are to be interpreted in a
statistical sense. Often they are interpreted in a sense that all data must fall within certain
windows; for this type of specification, any data that fail to satisfy QA objectives are rejected and
corrective action is taken. However, other interpretations are possible. For example, the
requirements may be satisfied if the average recovery is within the objectives. This is where we
describe how tabulated QA objectives will be interpreted.
2.3 QUALITATIVE QA OBJECTIVES: COMPARABILITY AND
REPRESENTATIVENESS
See Section 2.1.
2.3.1 Comparability
This is the degree to which one data set can be compared to another. For example, to
evaluate a cleanup process, analyses of the feed and discharge streams must be comparable.
Another example, to perform a nationwide environmental survey, methods used at different
locations must be comparable, probably by the use of consistent methods and traceability of
standards to the same or a reliable source (Simes, 1991),
108
-------
/ think that for this project, comparability would be achieved by use of the same
evaluation methods for each test phase or mechanism. For example, comparability would be
achieved by using the same analytical methods to evaluate each sensor coating.
2.3.2 Representativeness
This is the degree to which a sample or group of samples is indicative of the population
being studied. An environmental sample is representative for a parameter of interest when the
average value obtained from a group of such samples tends towards the true value of that
parameter in the actual environment, as the number of representative samples is increased.
Representativeness is normally achieved by collecting a sufficiently large number of unbiased
samples (Simes, 1991).
Because this study does not involve a sampling exercise, I think that representativeness
is associated with the applicability to the "real world" in two ways: 1) testing a field screening
device on spiked water or sand is not as representative as testing it in a real field situation; and
2) approaching the true value of the parameter is gained by testing the device a sufficiently large
number of times in a real field situation.
2.4 OTHER QA OBJECTIVES
Mass changes for SAW and QCM sensors for "screening experiments" are sought to
be ± 10% based on frequency changes.
2.5 EFFECTS OF NOT MEETING QA OBJECTIVES
Not meeting QA objectives may give false indications as to the efficacy of candidate
recognition coatings being screened.
SECTION 3.0
SAMPLING AND ANALYTICAL PROCEDURES
3.1 SAMPLING PROCEDURES
This requirement does not apply for proof-of-principle and new-concept studies.
These will be developed as part of each activity.
3.2 PROCESS MEASUREMENTS
Manufacturers' instructions for any needed process instruments will be followed.
3.3 ANALYTICAL PROCEDURES AND CALIBRATION
109
-------
3.3.1 EPA-Approved or Other Validated Standard Methods
This requirement does not apply for proof-of-principle and new-concept studies.
Methods will be developed as part of each activity.
3.3.2 Nonstandard or Modified Methods
This requirement does not apply for proof-of-principle and new-concept studies.
Methods will be developed as part of each activity.
3.3.3 Calibration Procedure and Frequency
Manufacturers' instructions for instruments will be followed as recommended and
necessary. An example is the calibration protocol for the Microsensor Systems Inc. Vapor
Generator (Appendix A).
SECTION 4.0
APPROACH TO QA/QC
4.1 DATA RECORDKEEPING
Due to the very basic nature of the research performed for this project, few
functions will be followed repetitively in exactly the same manner, at least in the initial
stages of the studies. Therefore, it is important that each experimental step, observation,
thought, and action be recorded to document the entire study in such a way that it can be
thoroughly understood, reported, and repeated. Accurate and legible records of the
research are noted in permanently-bound scientific notebooks. The scientific notebooks are
prepared, controlled, and reviewed using the HRC standard operating procedure, scientific
notebooks (Appendix A).
4.2 CALCULATION OF RESULTS (Simes, 1991)
This section is primarily reserved to show how analytical results will be manipulated to
prove or disprove a hypothesis. TJiis section should provide formulas and summarize any
statistical procedures for reducing the data, including units and definitions of terms. Also define
procedures that will be employed for determining outliers or flagging data.
If mass balance calculations are required, provide exact formulas relating the mass
balance to the individual measurements that will be made. Specify data reporting requirements
or plans at this point. Indicate units, matrices, and wet or dry intentions, etc. List deliverables
and data deliverables for the report. Will the data package include raw data? What type of QC
data will be reported?
110
-------
The final report should also include a summary of the original QA objectives and a
statement of whether they were met or further developments in this arena. If project/QA
objectives were not met, explain the effect on the study or project.
4.2 INTERNAL QUALITY CONTROL (Simes, 1991)
Describe the nature and frequency of all QC methods. The QC procedures should relate
to the QA/project objectives. They should be designed to support the objectives. Examples of
QC checks to consider follow:
• Samples
Collocated, split, replicate
• Spikes
Spiked and duplicate spiked samples
Spiked blanks
Surrogate spikes and internal standards
• Blanks
Sampling, field, trip, method, reagent, instrument
Zero and span gases
• Others
Standard reference materials
Mass tuning for mass analysis
Confirmation by second column (for GC) or instrument
Control charts
Independent check standard
Determinations for detection limits
Calibration standards
Proficiency testing of analysts
Any additional checks required by special needs of study
The use of ambiguous terms is a common problem in this section. The terms
"duplicate" and "replicate" are examples. Explain the exact point in the study where replication
occurs. Does "replicate" refer to samples collected simultaneously or sequentially in the field;
to samples collected at the same sampling point but at different times; to samples that are split
upon receipt in the lab; or to samples collected under yet another splitting protocol?
Similarly, there are numerous types of blanks; the exact procedure for preparing these
blanks must also be described.
4.3 CALCULATION OF DATA QUALITY INDICATORS
111
-------
4.3.1 Precision - (Simes, 1989)
If calculated from duplicate measurements:
RPD =
l - C2) x 100
(C, - C2)/2
RPD = relative percent difference,
Cj = larger of the two observed values, and
C = smaller of the two observed values.
If calculated from three or more replicates, use relative standard deviation (RSD)
rather than RPD:
RSD = (s/y) x 100 ;
RSD = relative standard deviation,
s = standard deviation, and
y = mean of replicate analyses.
When s is defined as follows:
y
n
s =
2
(y,-y)
n-1
standard deviation,
measured value of the i-th replicate,
mean of replicate measurements, and
number of replicates.
4.3.2 Accuracy - (Simes, 1989)
For measurements where spikes are used:
%R = 100 x
S-U
112
-------
%R = percent recovery,
S = measured concentration in spiked aliquot,
U = measured concentration in unspiked aliquot, and
C^ = actual concentration of spike added.
For situations where a standard reference material (SRM) is used instead of, or in
addition to, spikes:
%R = 100 x
C
srm
%R= percent recovery,
Cm= measured concentration of SRM, and
Csrm= actual concentration of SRM.
4.3.3 Completeness (sampling and analytical) - (Simes, 1989)
\V~\
%C = 100 x \ — \ ;
\T\
%C = percent completeness,
V = number of measurements judged valid, and
T = total number of measurements.
(statistical) - (Simes, 1989)
Defined as follows for all measurements:
\
%C = 100 x \ — \ ;
%C = percent completeness,
V = number of measurements judged valid, and
n = total number of measurements necessary to achieve a specified
statistical level of confidence in decision making
4.3.4 Method Detection Limit
Defined as follows for all measurements:
MDL=tn.1> !.„=„.„) xS
113
-------
MDL = method detection limit
S = standard deviation of the replicate
measurements
Vi i-«=099) = Students' t-value appropriate to a 99% confidence level
and a standard deviation estimate with n-1 degrees of
freedom.
SECTION 5.0
REFERENCES
APPENDIX A
PROCEDURES
ACKNOWLEDGEMENTS
Discussion with William H. Engelmann, Kim R. Rogers, and Eric N. Koglin of EPA,
Environmental Monitoring Systems Laboratory-Las Vegas, Grazyna Orzechowska,
University of Nevada-Las Vegas, and James Petrowski, Naval Explosive Ordnance Disposal
Technology Center, Indian Head, Maryland, is gratefully acknowledged.
REFERENCES
1. Simes, Guy F. Preparation Aids for the Development of Category IV Quality
Assurance Project Plans. EPA/600/8-91/006. U.S. Environmental Protection
Agency, Cincinnati, OH, 1991.
2. Simes, Guy F. Preparing Perfect Project Plans. A Pocket guide for the Preparation
of Quality Assurance Project Plans. EPA/600/9-89/087. U.S. Environmental
Protection Agency, Cincinnati, OH, 1989.
3. Interim Guidelines and Specifications for Preparing Quality Assurance Project Plans
QAMS-005/80. EPA-600/4-83-004. U.S. Environmental Protection Agency,
Washington, DC, 1983.
114
-------
19
DATA VALIDATION GUIDANCE FOR CONVENTIONAL WET CHEMISTRY ANALYSES
Ann Rosecrance. Corporate Quality Assurance Director, Analytical Chemistry Division, Core
Laboratories, and LaDonna Kibler, Quality Assurance Administrator, Analytical Chemistry
Division, Core Laboratories, 10205 Westheimer, Houston, Texas 77042
Abstract: Data validation is the process of determining the compliance of analytical data with
established method criteria and project specifications. The validation of environmental data
assesses the quality of the data generated for environmental measurements and determines if the
applicable analytical method requirements and project data quality objectives were met. This
paper provides guidance for the validation of data for conventional parameters obtained by wet
chemistry analysis methods. Summaries of method specified quality control (QC) criteria for
select wet chemistry methods are provided, in addition to general recommended QC guidelines.
Examples of data review checklists to use in conducting and documenting data validation are
included. A standard approach is presented for the review and validation of data for
environmental samples analyzed for conventional parameters by wet chemistry analysis methods.
INTRODUCTION
Analytical methods from the U.S.EPA, Standard Methods, and other sources are available for
the determination of conventional parameters in various sample matrices by wet chemistry and
instrumental techniques. Each method and associated QC sections define specific requirements
associated with application of the method; additional requirements may be further specified by
the governing regulatory agencies or defined in the quality assurance project plan and associated
data quality objectives. Laboratory analysts and data reviewers need to be familiar with the
requirements of the analytical methods that are routinely used in order to ensure that the
appropriate procedures are followed and that the required criteria are achieved. While the
available methods for organic and inorganic environmental analyses provide detailed and
extensive specifications on QC and other technical requirements, the wet chemistry methods do
not always include the same level of detail or specifications. Relevant guidance for wet
chemistry analyses is needed in order to validate data for conventional parameters.
Data validation activities determine if analytical data are in compliance with the analytical
method requirements and project specifications. Data validation procedures developed by EPA
for specific programs are used as standards for data validation.1'2'3 A previous publication
provided method comparisons and data validation guidance for EPA organic and inorganic
analysis methods for the determination of volatiles, semivolatiles, pesticides/PCBs, and metals
in ambient air, drinking water, wastewater, solid waste, and hazardous waste.4 In this paper,
an overview of QC requirements and recommended guidelines for select wet chemistry methods
are provided. A standard approach to data validation is presented that can be applied to the
validation of data for conventional parameters obtained from wet chemistry analysis methods.
The wet chemistry analyses covered include methods from EPA Methods for Chemical Analysis
of Water and Wastes, EPA SW-846 Solid Waste methods, EPA Contract Laboratory Program
(CLP) Statement of Work (SOW) for Inorganic Analysis, and Standard Methods for the
Examination of Water and Wastewater where EPA methods are not specified. The conventional
115
-------
parameters covered include inorganic and organic non-metallics analyzed by colorimetric,
gravimetric, titrimetric, turbidimetric, spectrophotometric, and other instrumental techniques.
DATA VALIDATION DEFINITION
Data validation has been defined by EPA as a systematic process, consisting of data editing,
screening, checking, auditing, verification, certification, and review in order to provide
assurance that data are adequate for their intended use.5 Data validation can be considered a
question and answer process to determine if the data meet both the analytical method
requirements and the associated project specifications. The four major questions to assess in
validating data are the following: (1) were the required technical and QC analyses performed,
(2) were they included at the required frequency, (3) were the required acceptance criteria met,
and (4) were the project specifications or data quality objectives met.
DATA VALIDATION PROCESS
The data validation process follows a step by step procedure for reviewing the data for
completeness, correctness, and acceptability. Specific procedures for data validation may be
found in EPA data validation documents or project specified data validation procedures.
Presented here is a generalized procedure for the data validation process as applied to wet
chemistry analyses. After defining the project requirements and obtaining the data package, the
first step in the data validation process should be a data completeness check to determine if all
of the required data and documentation are present. If any of the required data or documentation
are missing, then the needed information should be obtained, if possible, before conducting the
data validation. The next recommended step in the process is a review of the QC data associated
with the analysis of the sample batch, referred to here as method QC. Next is a review of the
sample data and related QC information for each sample, followed by an overall review of all
of the data and associated documentation. Any data that does not meet the method requirements
or project specifications at any of these steps should be qualified appropriately. The final step
should be the preparation and distribution of a data validation report. A recommended sequence
for the data validation process that includes each of these steps is provided in Figure 1.
DATA COMPLETENESS CHECK
In the first step of the data validation process, the data package should be reviewed to determine
if all of the required data and associated documentation are present. This check should include
a review of the analysis specifications to ensure that the required method, target analytes and
their reporting limits, and data quality objectives are defined. The next step should be a review
of the documentation on sample custody and condition in order to determine if custody was
maintained and if the sample was received and maintained in the proper condition. The data
report, QC results, and associated raw data for each analytical parameter should be reviewed to
ensure that all required information is present. Documentation for sample preparation, standards
preparation, standards and sample analysis, data interpretation and calculations, and
correspondence should be reviewed to ensure that all of the required information is present. An
example checklist to use for performing the data completeness check is provided in Figure 2.
116
-------
DATA VALIDATION APPROACH
The recommended approach for performing data validation is to first summarize the analytical
method requirements, associated reference QC requirements, and the project specifications.
Summaries of the QC and other technical requirements are prepared for reviewing and validating
data for each target analyte. An example of a summary chart comparing the QC requirements
for select cyanide analysis methods is provided in Table 1. The data validation process proceeds
by following closely to the sequence of the analysis and the procedures established by EPA for
data validation. Data generated from any of the specified methods are reviewed for compliance
with each requirement and applicable criteria using general or method specific data review
checklists. Examples of a general checklist for wet chemistry analyses and a method specific
checklist for cyanide analysis by the EPA CLP SOW for Inorganic Analysis are provided in
Figures 3 and 4, respectively. Using summary charts that provide the required criteria and
checklists that record compliance with the applicable criteria, a standard approach to data
validation can be performed effectively and efficiently for wet chemistry analysis methods.
QUALITY CONTROL REVIEW
The types of analyses that are subjected to data validation are method specified quality control
and sample specific quality control. Method quality control refers to the analyses that are
necessary for initiating sample analyses and that are common to the sample analysis batch. This
includes instrument calibration, calibration verification standards, blanks, laboratory control
standards, spikes, and duplicates. Also included in method QC are other analyses that are
necessary to assess the field and laboratory procedures related to the sample data. This includes
container certifications, field blanks, field replicates, detection limit determinations, precision
and accuracy determinations, and performance evaluation analyses. Sample quality control
refers to the criteria that are specific to each sample. This includes holding times, sample
preparation and analysis, and the identification and quantitation of target analytes.
The following section provides a summary of items to review during the validation of wet
chemistry data. Comparisons of the QC requirements for cyanide analysis methods and
summaries of the method specified QC requirements for other wet chemistry analyses are
provided in Tables 1 and 2, respectively. Because QC requirements for wet chemistry analyses
are often undefined or not specified, general QC guidelines are provided in Table 3 that may be
considered when method or project specified criteria are not available. These general QC
guidelines are not intended as a replacement for specific requirements defined in the reference
method or project specifications, but are presented as guidelines for performing data validation
when QC requirements are not specified. Further details on QC requirements are provided in
the reference methods and EPA or project specific data validation procedures.
Instrument Calibration. Instrument calibration data are evaluated to ensure that the analysis and
the associated instrument were set up properly and that acceptable quantitative and qualitative
data were generated at the beginning of the analysis. Initial calibration data are reviewed for
the analysis of the required analytes, at the required number of levels and concentrations, at the
required frequency, and within the required or recommended acceptance criteria (i.e., a linear
calibration curve with a correlation coefficient of > 0.995).
117
-------
Calibration Verification. Calibration verification analyses are evaluated to ensure that acceptable
quantitative and qualitative data were generated during the course of and at the end of the
analytical run. Initial and continuing calibration verification data are reviewed for the analysis
of the required analytes, from the correct source, at the required concentrations, at the required
frequency, and within the required acceptance criteria (i.e., % recovery).
Blanks. Data for calibration blanks and preparation blanks are evaluated to check the
background from the analysis and preparation procedures, respectively. Blank data are reviewed
for the analysis of the correct source of material, at the required frequency, and within the
acceptable background levels for the target analytes.
Laboratory Control Standards CLCS). LCS analyses are evaluated to check the overall
performance of the laboratory, including both sample preparation and analysis procedures, on
each method utilized. LCS data are reviewed for the analysis of the correct type of standard,
at the required frequency, and within the required acceptance criteria (i.e., % recovery).
Spikes. Data for spikes are evaluated to check if the analytical performance was within the
accuracy specifications that have been established for the method. Blank spikes determine the
laboratory performance for recovery of analytes in blank matrices and matrix spikes assess the
effect of the sample matrix on analyte recovery. Spike data are reviewed for analysis of the
correct type of spike (blank or matrix spike and pre-preparation or post-preparation), at the
required frequency, with the correct analytes at the required concentration, and within the
required acceptance criteria (i.e., % recovery).
Duplicates. Data for duplicates are evaluated to check if the analytical performance was within
the precision specifications that have been established for the method. Duplicate sample data
are reviewed for analysis of the correct type of duplicate (sample duplicate or matrix spike
duplicate), at the required frequency, and within the required acceptance criteria for relative
percent difference or other specified duplicate evaluation criteria.
Other Quality Control. Other quality control refers to additional analyses associated with
evaluation of the acceptability of the sampling and analysis procedures. Quality control
measures associated with sampling include the evaluation of field blanks and replicate samples
to determine background contamination and sampling precision, respectively. Other laboratory
quality control measures include instrument detection limit determinations and performance
evaluation analyses. Data for each quality control analysis are reviewed to determine if the
required analyses were performed at the required frequency and if the results were within the
required acceptance criteria.
Sample Quality Control. Sample data are evaluated to ensure that holding times were met, that
preparation and analysis procedures were performed correctly, that the analytes were reported
correctly (both identification and quantitation), and that reported values were within the
calibration range or linear range. The adherence to holding times should be reviewed for each
parameter as holding times are variable and depend upon the analyte and method specification.
Raw data for all samples should be reviewed and recalculated to determine if the reported results
correlate with the raw data.
118
-------
DATA VALIDATION DOCUMENTATION
Data validation procedures should be documented on standardized forms such as the example
data review checklists provided in Figures 3 and 4. The checklist or other form should report
the adherence or lack of adherence to each of the method requirements or project specifications.
The agreement of the raw data and the data report should be determined and documented. Any
major or minor deficiencies identified should be documented in a data validation report
describing each deficiency and its potential impact on the sample results. Qualifiers used for
data in question should be in accordance with the project specifications and they should be
clearly defined. Examples of qualifiers used in EPA data validation procedures are: (R), the
results are rejected due to serious deficiencies in quality control criteria; (J), the associated
numerical value is an estimated quantity because certain quality control criteria were not met;
(N), presumptive evidence of presence of material; (U) the material was analyzed for but not
detected; and (UJ), a combination of U and J.1 The data validation report should include an
overall assessment of the data, in addition to any recommendations for further action.
SUMMARY
Data validation is an integral part of the environmental data generation process and in order to
be efficient and effective, the data validation process must be versatile and widely applicable.
With the large number of analytical methods that are available for environmental sample
analyses, laboratory analysts and data reviewers must be familiar with the QC requirements for
each method and project. The information provided in this paper summarizes the QC
requirements for several wet chemistry analysis methods and provides general QC guidelines
when method specifications are not available. Data validation guidance, including example
checklists for performing and documenting data validation, is provided so the data reviewer can
use a standard approach when validating data from various analytical methods. This guidance
provides an effective and efficient approach for addressing the specific requirements of each
method utilized and for determining if the applicable QC specifications were met. This
information is not intended as a replacement for the analytical methods or EPA data validation
procedures, but is a guide on QC requirements and data validation for wet chemistry analyses.
REFERENCES
1. U.S.EPA Quality Assurance/Quality Control Guidance for Removal Activities. Sampling QA/OC Plan
and Data Validation Procedures. EPA 540-G-90 004, April 1990.
2. U.S.EPA Laboratory Data Validation Functional Guidelines for Evaluating Inorganics Analyses. 1988.
3. U.S.EPA Laboratory Data Validation Functional Guidelines for Evaluating Organics Analyses. 1991.
4. A.E.Rosecrance, "Data Validation Guidance for EPA Organic and Inorganic Analytical Methods",
Proceedings of the EPA Eighth Annual Waste Testing and Quality Assurance Symposium. Arlington,
VA, July 1992.
5. U.S.EPA Interim Guidelines and Specifications for Preparing Quality Assurance Project Plans.
QAMS-005-80, December 1980.
119
-------
Figure 1. DATA VALIDATION PROCESS
Obtain Project
Specifications and Data
Quality Objectives
Obtain Data Validation
Requirements
J
/ wuicuil vtua r«iui\aye rr
1
Perform Data
Completeness Check
- Use Checklist
N
Perform Method QC Review
- Use Checklist
Was Method
Criteria Met?
Perform Sample QC Review
Use Checklist
Qualify Data
Was Sample
Criteria Met?
Perform Overall Evaluation
-^ ^-\_
Were Project Specifications ^"~-~^( . .
and Data Quality Objectives Mef ^
^^ ^^^^
^^T^^
Yes
*
/ Prepare and Distribute
7
Qualify Data
Data Validation Report /
120
-------
Figure 2. Data Package Checklist
REVIEW ITEM
ANALYSIS SPECIFICATIONS
Project Specifications
Method Reference
Target Analytes and DL/QL
Specifications
SAMPLE CONDITION / CUSTODY
Chain of Custody Records
Sample Receipt/Log-ln Records
DATA REPORT
Case Narrative
Client ID/Lab ID
Parameters
Matrices
Dates of Preparation and Analysis
Method Reference
Required Forms
Sample Results/DLs
QC Results
RAW DATA
Organic
Metals
Wet Chemistry
Radiochemistry
Other
LOGS /RECORDS
Sample Preparation
Sample Analysis/Run Logs
Standards Preparation
Calculations
Correspondence
INCLUDED
NOT INCLUDED
NA
COMMENTS
OTHER
121
-------
Figure 3. Wet Chemistry Data Review Checklist
REQUIREMENT
A. CALIBRATION CURVE
1. Frequency: Daily?
2. Levels: Blank +• £ 3 or
standards?
3. Range: Appropriate?
4. Correlation coefficient 2 0.995?
B. LABORATORY CONTROL STANDARD
1 . Frequency: Each batch?
2. Source: Independent?
3. Reference: Certified?
4. % Recovery: 90- 110% or ?
C. BLANK
1 . Frequency: Each batch?
2. Matrix: Matrix specific?
3. Preparation: Entire procedure?
4. Analytes concentration:
-------
Figure 4. DATA REVIEW CHECKLIST FOR CYANIDE ANALYSIS DATA BY EPA CLP SOW
REQUIREMENT
A. CALIBRATION
1 . Frequency: Daily?
2. Levels/Range: Blank + 3 stds (1 at CRDL)?
3. 'Correlation coefficient: £ 0.995?
B. CONTINUING CALIBRATION
1. Frequency: Beginning/end/every 10 samples-2 hrs?
2. Level: Mid-range for CCV?
3. % Recovery: 85-115%?
C. OTHER STANDARDS (DISTILLED)
1 . Frequency: 1 per batch?
2. Level: Mid-range?
3. %Recovery: 85-115%?
D. CALIBRATION BLANK
1 . Frequency: Beginning/end/every 1 0 samples-2 hrs?
2. Analyte concentration:
-------
Table I. Comparison of QC Requirements for EPA Cyanide Analysis Methods
Requirement
Water and Waste-
water Analysis
Method 335.2
RCRA Solid
Waste Analysis
Methods 9010/901OA
Sunerfund Hazardous
Waste Analysis
CLP SOW ILM03.0
Method Detection Limit
Titration: 1mg/L
Colorimetric: 0.02 mg/L
Titration: 0.1 mg/L
Colorimetric: 0.02 mg/L
CRDL: 10ng/L
Holding Time
14 days (24 hours when 14 days
sulfide is present)
12 days from sample receipt
Initial Calibration'1'
Frequency
6 standards and a blank
Daily
6 standards and a blank
Daily
3 standards and a blank
(one standard at the CRDL)
Daily
Calibration Verification121
Frequency
Criteria
Other Standards (Distilled)
Frequency
Criteria
Calibration Blanks
Frequency
Criteria
Preparation Blanks
Frequency
Criteria
Laboratory Control Standard
Frequency
Criteria
Matrix Spike Samples
Frequency
Criteria
Duplicate Samples
Frequency
Criteria
NS
NS
NS
High and low standard
1 each per batch
90-1 10% Recovery
Colorimetric: 1 per batch
Use in initial calibration
Titration: 1 per batch
Colorimetric: Not specified
Titration: Use in calculation
Colorimetric: Not specified
NS
NS
NS
1 per batch to check
distillation efficiency
NS
NS
NS
Mid-range standard
Every 15 samples
85- 11 5% Recovery
High and low standard
1 each per batch
90-110% Recovery
Colorimetric: 1 per batch
Use in initial calibration
Titration: 1 per batch
Colorimetric: Not specified
Titration: Use in calculation
Colorimetric: Not specified
Independent check standard
1 per batch
85-115% Recovery
Matrix spike and matrix
spike duplicate per batch
NS
1 matrix spike duplicate
per batch
NS
CCV: Mid-range standard
Beginning, end, and every
10 samples or 2 hours
85-115% Recovery
Mid-level standard
1 per batch
85-1 15% Recovery
Colorimetric: Beginning, end and
every 10 samples or 2 hours
£CRDL
Titration: 1 per batch
Colorimetric: 1 per batch
Titration: Use in calculation
Colorimetric: <, CRDL
Distilled independent standard (ICV)
1 per batch
85-1 15% Recovery
1 per matrix per concentration
level per batch
75-1 25% Recovery
1 per matrix per concentration
level per batch
< 20% RPD for values > 5x CRDL
Other Method Criteria
Verify sample pH £ 12;
Check for oxidizing agents
and sulfides
Verify sample pH > 12;
Check for oxidizing agents
and sulfides
Verify sample pH £ 12;
Check for oxidizing agents
and sulfides
(1) Calibration standards must be distilled tor EPA Methods 335.2 and9010/9010A when sulfides are present In the samples.
(2) CLP SOW specifies that the initial calibration verification standard (ICV) be distilled and analyzed as the laboratory control standard (LCS).
For complete Information, refer to the method reference.
Acronym Definitions
CCV = continuing calibration verification
CRDL = contract required detection limit
ICV = initial calibration verification
LCS = laboratory control standard
NS = not specified
RPD = relative percent difference
124
-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES
PARAMETER (METHOD NO.)
Acidity (305.1)'
Alkalinity (310.1)'
Anions (300.0)'
by ion chromatography
BODandCBOD (405.1)'
by electrode (5210 B)2
Boron (4500-B)2
by colorimetry
Bromide (320.1)'
by titration
Carbon (TOO (415.1)'
by combustion or oxidation
Chloride (325.3)'
by Hg(N03)j titration
Chloride (9252)3
by Hg(N03)2 titration
Chloride (4500E-CI- B)2
by AgN03 titration
COD (HACH 8000)'
by colorimetry
COD (Mid-level) (410. 1)1
by titration
COD (Low Level) (410.2)'
by titration
Color (110.2)'
by platinum-cobalt
Corrosivity (1110)3
toward steel
CALIBRATION CURVE
Frequency
pH Calib.
Daily/2
pH Calib.
Daily/2
Daily
Daily
Daily
Daily
Daily/
Range
Acceptance
Criteria
pH Meter
Specification
pH Meter
Specification
± 10% of
True Value
DO Meter
Specification
Linear
Instrument
Specification
Instrument
Specification
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Acceptance
Criteria
< Detection
Limit
£ 0.2 mg/L
Zero
Instrument
Blank
Correcting
< Detection
Limit
Blank
Correcting
Blank
Correcting
Blank
Correcting
Zero
Instrument
Blank
Correcting
Blank
Correcting
< Detection
Limit
Blank
Correcting
SPIKES
Frequency
10% or
1 /Batch
5% or
1 /Batch
5% or
1 /Batch
5% or
1 /Batch
Acceptance
Criteria
Not
Specified
80-120%
Not
Specified
80-120%
DUPLICATES
Frequency
10% or
1 /Batch
5% or
1 /Batch
10% or
1 /Batch
5% or
1 /Batch
5% or
1 /Batch
Acceptance
Criteria
See Table
1020:l2
See Table
1020:l2
Not
Specified
See Table
1020:l2
Not
Specified
OTHER
Type
pH Check
pH Check
GIG Std.;
Seed
Control;
LCS
LCS
LCS
LCS
Frequency
1 /Batch
1 /Batch
2/Batch;
2/Batch;
Daily
1 /Batch
Every 15
Samples
1 /Batch
Acceptance
Criteria
± 0.05 of
True Value
± 0.05 of
True Value
198 ± 30.5
mg/L;
0.6-1.0
mg/L;
90-110%
90-110%
Not Specified
90-110%
-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES (continued)
PARAMETER (METHOD NO.)
Phenol (9065)'
by spectrophotometry
Phenol (420.1)'
by spectrophotometry
Phenol (420.2)'
by automated 4-AAP
Phos. Total «. Ortho (365.2)'
ascorbic acid, single reagent
Pho«. Total & Ortho (365.3)'
ascorbic acid, two reagent
Sulfate (375.4)'
by turbidimetry
Sulfida (376.2)'
by colorimetry
Sulfite (377.1)'
by titration
Surfactants, Anionic (425.1)'
as MBAS (5540C)2
Total Petroleum (418.1)'
Hydrocarbons (TPH) by IR
Turbidity (180.1)'
CALIBRATION CURVE
Frequency
Daily
Daily
Daily
Daily
Daily
Daily
Daily
Daily
Daily
Daily
Acceptance
Criteria
Linear
Linear
Linear
± 2% of
True Value
Linear
Linear
Linear
Linear
Linear
Instrument
Specification
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Solvent
Blank:
1 /Batch
Acceptance
Criteria
< Detection
Limit
Zero
Instrument
Establish
Baseline
Zero
Instrument
Zero
Instrument
Zero
Instrument
Use in
Calibration
Blank
Correcting
Use in
Calibration
Zero
Instrument
SPIKES
Frequency
10% or
1 /Batch
5% or
1 /Batch
Acceptance
Criteria
Not
Specified
80-120%
DUPLICATES
Frequency
5% or
1 /Batch
Acceptance
Criteria
See Table
1020:l2
OTHER
Type
LCS
LCS
LCS
Check
Standard
Frequency
Every 1 5
Samples
Every 3 or
4 Samples
1 /Batch
1 /Batch
per Range
Acceptance
Criteria
Not Specified
Not Specified
90-110%
Instrument
Specification
ro
O)
CCV Continuing Calibration Verification
G/G Glucose/Glutamic Acid Standard
LCS Laboratory Control Standard
RPD Relative Percent Difference
Methods for Chemical Analysis of Waster and Wastes, EPA-600/4-79-020, U.S. EPA, Cincinnati, Revised March 1983.
' Standard Methods for the Examination of Watar and Wastewater, APHA-AWWA-WPCF, 17th. Edition, 1989.
' Test Methods for Evaluating Solid Waste. Physical/Chemical Methods. SW-846, U.S. EPA, Washington D.C., Third Edition, September 1986.
' Gibbs, C. R., "Introduction to Chemical Oxygen Demand", HACH Company, Technical Information Series-Booklet No. 8, 1992.
1 Statement of Work for Inorganics Analysis. Multi-Media Multi-Concentration. U.S. EPA Contract Laboratory Program, ILM03.0.
' O93-8O. "Ta*t Methods for Rash Point by Pansky-Martans Closed Tester", American Society for Testing and Materials, Philadelphia.
-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES (continued)
PARAMETER (METHOD NO.)
HaxavalentCr (7197)3
by chelation and FLAA
Iodide (345.1 )'
by titration
Nitrogen - Ammonia (350. 1)1
by automated colorimetry
Nitrogen - Ammonia (350.2)'
by colorimetry
Nitrogen - Ammonia (350.2)'
by ion electrode (350. 3)1
Nitrogen - Nitrate (352.1)'
by colorimetry
Nitrogen - Nitrite (354.1)'
by colorimetry
Nitrogen - Nitrate & Nitrite
by hydrazine reduction (353.1)'
Nitrogen - Nitrate & Nitrite
by auto. Cd reduction (353.2)'
Nitrogen - TKN (351.3)'
by titration
Oil and Grease (413.2)'
by IR
Oil and Grease (413.1)'
by gravimetry
pH in Soil (9045)3
pH in Water (150.1)'
CALIBRATION CURVE
Frequency
Daily
Daily
Daily
Daily
Daily
Daily
Daily
Daily
Daily
pH Calib.
Daily/2
pH Calib.
Daily/2
Acceptance
Criteria
Linear
Linear
Linear
Instrument
Specification
Linear
Linear
Linear
Linear
Linear
pH Meter
Specification
± 0.05 of
True Value
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Solvent
Blank:
1 /Batch
Solvent
Blank:
1 /Batch
Acceptance
Criteria
< Detection
Limit
< Detection
Limit
Establish
Baseline
Zero
Instrument
< Detection
Limit
Zero
Instrument
Zero
Instrument
Establish
Baseline
Establish
Baseline
Blank
Correcting
Zero
Instrument
Blank
Correcting
SPIKES
Frequency
10% or
1 /Batch
Acceptance
Criteria
85-115%
DUPLICATES
Frequency
5% or
1 /Batch
5% or
1 /Batch
Every
Sample
Acceptance
Criteria
Not
Specified
Not
Specified
± 0.01 pH
units
OTHER
Type
LCS;
Method of
Standard
Addition
Distilled
Standards
Distilled
Standards
Efficiency
Check
Efficiency
Check
Distilled
Standards
Check
Standard
pH Check
Frequency
15%;
All EP Tox
Extracts
High&
Low/Batch
High&
Low/Batch
1 /Batch
1 /Batch
HighS,
Low/Batch
1 /Batch
1 /Batch
Acceptance
Criteria
Not
Specified;
Linear
Curve
Comparable
to Undistilled
Standards
Comparable
to Undistilled
Standards
100%
Reduction
100%
Reduction
Comparable
to Undistilled
Standards
Not Specified
± 0.05 of
True Value
-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES (continued)
PARAMETER (METHOD NO.)
Cyanide, Total (335.2)'
by titration
Cyanide, Total (335.2 CLP-M)'
by titrntion
Cyanide, Total (335.2)'
by colorimetry
Cyanide, Total (335.2 CLP-M)*
by colorimetry
Cyanide, Total (335.3)'
by automated UV
Cyanide, Total (335.2 CLP-M)'
by semi-automation
Rash Point (1010)' (D93)'
by Pensky-Martan Closed Cup
Fluoride (340.2)'
by ion electrode
Fluoride (340.3)'
by automated colorimetry
Halides (TOX) (9020)'
Hardness (130.2)'
by titration
Hexavalent Cr (218.4)'
by chelation and FLAA
CALIBRATION CURVE
Frequency
Daily
Daily
Daily
Daily
Daily
Daily
Daily in
Duplicate
Daily
Acceptance
Criteria
Linear
Linear
Linear
Linear
Instrument
Specification
Linear
± 3% of
True Value
Linear
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Acceptance
Criteria
Blank
Correcting
Blank
Correcting
Use in
Calibration
S Contract
Required
Detection
Limit
Use In
Calibration
s Contract
Required
Detection
Limit
< Detection
Limit
Establish
Baseline
< Detection
Limit
< Detection
Limit
< Detection
Limit
SPIKES
Frequency
1 /Batch
5% or
1 /Batch
1 /Batch
5% or
1 /Batch
5% or
1 /Batch
10% or
1 /Batch
Acceptance
Criteria
Distillation
Check
75-125%
Distillation
Check
75-125%
75-125%
Not
Specified
DUPLICATES
Frequency
5% or
1 /Batch
5% or
1 /Batch
5% or
1 /Batch
Every
Sample
Aspirate
Samples in
Duplicate
Acceptance
Criteria
:S 20%RPD
S 20%RPD
S 20%RPD
Not
Specified
Report
Average
OTHER
Type
Distilled
Standards
LCS;
Distilled
Standard
Distilled
Standards
LCS;
CCV;
Distilled
Standard
LCS;
CCV;
Distilled
Standard
p-Xylene
Standard
LCS;
Nitrate
Blank;
Efficiency
Check
Standard
Frequency
High &
Low/Batch
1 /Batch;
1 /Batch
Highi
Low/Batch
1 /Batch;
10%;
1 /Batch
1 /Batch;
10%;
1 /Batch
1 /Batch
15%;
Every 8
Samples;
1 /Batch
Acceptance
Criteria
90-110%
85-115%;
85-115%
90-110%
85-115%;
85-115%;
85-115%
85-115%;
85-115%;
85-115%
81 ± 2°F
Not
Specified;
Blank
Correcting;
±5% of True
Value
-------
Table III. General QC Guidelines for Wet Chemistry Parameters*
Calibration
QC Check Sample/LCS
Blank
Spike
Duplicate
Type
Multiple standards + blank
Independent source
Preparation blank
Matrix spike, pre-digestion
Sample duplicate,
pre-digestion
Freqency
Daily
Batch
Batch
5-10%
5-10%
Criteria
r > 0.995
90- 110% recovery
£DL
75 - 125% recovery
£ 20% RPD for values > 5 x DL;
± 1 x DL for values < 5 x DL
1 Recommended guidelines for QC analyses where method QC requirements are not available.
Refer to the method reference lor method specific requirements.
Acronym Definitions
DL = detection limit
LCS = laboratory control standard
RPD = relative percent difference
129
-------
Of) DEVELOPING A PERFORMANCE-BASED APPROACH TO ENVIRONMENTAL
ANALYTICAL TESTING
Berta L. Thomas. Dr. Robert G. Riley, Gary M. Mong, Margaret McCulloch, Sandra K.
Fadeff, Dr. Deborah S. Sklarew, and Dr. Steven C. Goheen. Pacific Northwest
Laboratory^), Richland, Washington 99352; Dr. James S. Poppiti, Laboratory
Management Division, U.S. Department of Energy, Germantown, Maryland 20874
ABSTRACT
Documented analytical methods are prescriptive. However, they can be generated,
selected, or modified based upon performance criteria. Freedom to select and use methods
based on performance criteria allows laboratories flexibility to generate data that meet well-
defined data quality objectives (DQOs) cost effectively.
To support the U.S. Department of Energy Waste Management and Environmental
Restoration (DOE-EM) programs, we suggest using DQOs and a performance-based
approach for selecting methods to analyze environmental and waste samples. Over the
long-term, a performance based approach is viewed as a generally cost-effective approach
for producing analytical data of the needed quality. Adopting such an approach will
encourage development of innovative and efficient analytical technologies that address the
complexities of waste and environmental characterization at DOE sites while ensuring
comparable data are produced among the many laboratories supporting DOE-EM programs.
A primary mission of DOE is the environmental restoration of its waste sites.
Accomplishing this mission will require that a vast number of environmental and waste
samples be analyzed for the presence of inorganic, organic, and radionuclide constituents.
Many of the needed analyses cannot be performed according to existing published methods
(e.g., ASTM or EPA) because of constraints that accompany sample characterization in
radiological environments. Selecting and applying analytical methods within the
framework of a performance-based program would improve effectiveness of DOE
analytical support programs.
A guidance document, DOE Methods for Evaluating Environmental and Waste
Management Samples (DOE Methods), is being developed in support of DOE-EM
programs. This document contains general guidance for characterization of contaminants in
radioactive waste and environmental samples as well as recommended analytical methods
necessary to meet the broad range of analytical characterization needs existing across the
DOE complex. DOE Methods and other recognized documents (e.g., EPA's SW-846)
should be consulted when selecting methods to perform analyses in support of DOE-EM
programs; however, these methods may be modified as long as the modified methods are
demonstrated to meet DQOs. In addition, alternate sources of methods may be used if
those methods are cost-effective and again, are demonstrated to meet project DQOs.
(a)Pacific Northwest Laboratory is operated for the U.S. Department of Energy by Battelle
Memorial Institute under Contract DE-AC06-76RLO 1830.
130
-------
INTRODUCTION
This paper presents concepts associated with implementing a performance-based approach
to waste and environmental analytical testing within United States Department of Energy
(DOE) programs. The approach employs a data quality objectives (DQO) process that leads
to the establishment of project specific DQOs, rather than imposing prescriptive analytical
method requirements on analytical laboratories to meet program goals. For the purposes of
this paper, we define DQOs as statements that define the analytical-testing-method-
performance requirements for a program based upon a consensus that is reached by the
regulatory community, DOE Program Management, and analytical laboratory management
and staff (EPA 1986).
This paper describes 1) the basic elements of the approach, 2) strengths versus perceived
weaknesses in applying a performance-based approach to analytical testing, and 3)
arguments that implementing such an approach over the long-term will reduce costs and
improve analytical data quality. The goal is to stimulate responsible debate regarding the
adoption of a performance-based approach to analytical testing.
BACKGROUND
Currently, analytical testing in support of environmental monitoring, restoration, and waste
management programs is based upon required laboratory implementation of standard,
usually promulgated, analytical methods. Thus, laboratory personnel responsible for
supplying data in support of these programs are required to adhere strictly to analytical
methods specified in program specific plans. This approach is prescriptive in nature;
laboratory personnel are given very little flexibility to perform analytical testing according
to methods they believe would best meet project specific DQOs or to deal with problems
associated with highly complex sample matrices. Many problems associated with
characterization of DOE environmental and waste sample matrices are not addressed by
promulgated methods.
As a framework for the performance-based approach, the DQO process converges the
thoughts of interested parties (i.e., program managers, analysts, regulators, clients) toward
a consensus opinion regarding project goals, DQOs to meet project goals, and analytical
methods to be used. A decision to use promulgated methods, unpromulgated methods, or
some combination of both to meet DQOs will be a product of mis DQO process.
ELEMENTS OF A PERFORMANCE-BASED APPROACH
Once an analytical method has been selected through the DQO process, responsibility for
documenting validation of the method lies with the organization that will be implementing
the method, the analytical laboratory. Validation is the process of establishing the
suitability of a method for authorized use in the laboratory (Dux, 1990). Guidance can be
found (Dux 1990; ASTM 1986) relating to the validation process. One aspect of the
validation process consists of demonstrating method performance using relevant sample
matrices and addressing sample logistics (e.g., manipulation of samples containing high
levels of radioactivity in a hot cell or glovebox). The following quality control
documentation should accompany validation of an analytical method.
• The method's bias and precision should be documented by analyzing spiked and
replicate samples of the matrix of interest. When it is not possible to perform this
131
-------
test on a portion of the actual sample matrix, the test should be performed using a
matrix that exhibits as many of the physical and chemical characteristics of the
actual sample matrix as possible. When applicable, test data can be used to develop
method warning and control limits.
• Analysis of method and matrix blanks should be documented.
• The working range of the method should be documented. Parameters for which
calculated data should be provided include minimum detection limit, limit of
detection, and limit of quantitation.
• The ruggedness of the method should be documented. Youden presents
information relating to ruggedness testing (Youden and Steiner, 1975).
• Method interferences and limitations as they apply to the matrix in question should
be identified and documented.
• The method should be controlled so that changes to the method subsequent to its
selection and validation are documented and receive concurrence from the
appropriate regulators and DOE project management
• Method comparability tests should be documented. Where applicable, the analytical
laboratory should consider documenting comparability of the selected method with
a standard, published method to obtain data credibility with peers. "Round robin"
interlaboratory comparison studies can be helpful in demonstrating that a method
produces comparable data. Periodic requalification of validated methods should be
performed by laboratories using those methods.
STRENGTHS OF A PERFORMANCE-BASED APPROACH-COST AND DATA
QUALITY BENEFITS
The most important strengths of a performance-based approach to analytical testing of
waste and environmental samples fall into two broad categories: improved cost and
improved data quality over time.
Cost-A performance-based approach allows methods that are tailored to provide
data of the quality needed to meet program objectives (part of the DQO process
decision) to be selected. The performance-based approach makes a larger pool of
analytical methods available for consideration during the DQO process. Design of
the most effective analysis program may be based upon cost, as long as assurances
are made that DQOs will be met. For example, analytical data generated from
standard methods are often based upon minimum detection limits (MDLs), limits of
detection (LODs), or limits of quantitation (LOQs) that are, in general, lower than
those required to meet project DQOs. In contrast, the performance-based approach
allows the flexibility to adapt or chose alternative analytical methods that provide
data of only the quality needed, while minimizing additional costs associated with
methods that produce data of higher quality than project DQOs require. In
evaluating cost, factors to consider include time required to perform the method,
equipment and facilities required to assure worker safety, training of analysts, and
waste minimization and disposal.
132
-------
In applying the DQO process within the performance-based approach, analytical
methods can be selected that are tailored to the resources available (e.g., personnel
and equipment) rather than requiring purchase of additional equipment, making
expensive modifications to existing facilities, or hiring specialized personnel to
perform the analytical testing according to standard/promulgated methods.
• Data Quality-A performance-based approach to analytical testing offers a means to
ensure enhanced data quality by selection/application of the best available analytical
methodology. In doing so, weaknesses associated with standard methods can be
avoided. Currently, methods accepted for use in support of regulatory
requirements present a very conservative approach to laboratory analysis. Little
room for adoption of innovative approaches exists in the current prescriptive
environment. This prescriptive approach has the effect of stifling creativity in
analytical testing methodology.
Standard methods, such as those published in SW-846, were not developed to
accommodate the variety of complex sample matrices encountered in DOE programs
and frequently cannot be easily adapted to such matrices. In a program requiring
prescriptive methods, the process of data production can over shadow the data upon
which decisions are based. In this situation, the process can become more
important than the product. This can be dangerous because data quality may be
compromised to comply with standard methods. Such situations may be avoided
using a performance-based approach.
PERCEIVED WEAKNESSES OF A PERFORMANCE-BASED APPROACH
The following are perceived as weaknesses of the performance-based approach:
• Approach Acceptance-The performance-based approach is currently untested.
Historically, promulgated methods are adopted to secure regulatory "buy-in" by
offering some assurance that data of high quality are produced and are defensible in
the event that they are introduced into litigation. Because of entrenched habits, it
will take time for regulators and the regulated community to accept a performance-
based approach to analytical testing. After the performance-based approach
achieves success, the rigid, conservative approach to method selection will yield to
the more flexible performance-based approach.
• Cost of Documentation—Implementing a performance-based approach to
environmental and waste analysis could require a greater "up front" investment of
time and dollars to gain acceptance. Method validation by a laboratory could
require that more quality control data be generated and documented in the near-term
than might be required to support standard/promulgated methods. These added "up
front" costs could create cash flow problems for small laboratories. However, the
QC data produced will be more pertinent to program DQOs and to the matrix of
interest. Data will, therefore, be more meaningful.
• Training Costs—The flexibility built into a performance-based approach places
greater responsibility for management of unpromulgated or modified promulgated
methods (i.e., selection, validation and application) on laboratory personnel. The
additional technical responsibility could lead to the need for more education and
training for laboratory analysts and managers to assure they are able to make sound
133
-------
scientific judgments during the method selection and application process. This
increased training for laboratory analysts and managers should, however, improve
the quality of the data.
Similarly, auditors providing program oversight might need more extensive training
to effectively audit programs developed according to a performance-based
approach. The emphasis in auditing prescriptive-based programs is placed on
monitoring for compliance to program requirements. "Monitoring for compliance
type auditing" is generally thought to be much simpler (i.e., requiring less extensive
training) than monitoring a performance-based program for its effectiveness.
However, this change in emphasis of the auditing process should produce value
added over time.
Contractual Costs-More planning could be required for acquiring and monitoring
subcontracts for analytical services in programs developed according to a
performance-based approach. The flexibility inherent in the performance-based
approach could require DOE to spend more time preparing a request for proposal
and the subcontracted organization to spend more time preparing the proposal than
might be required using the more traditional approach that requires the use of
standard/promulgated analytical methods. However, the resulting proposal is more
likely to better meet the needs of the program.
CONCLUSION
A primary mission of DOE is the environmental restoration of its waste sites. Decisions
regarding site remediation require the use of legally defensible data of high quality.
Because projected costs associated with remediation efforts are so high, DOE, the
regulatory community, and analytical laboratory staff need to consider how best to reduce
analytical costs while still providing sound data to serve as the bases for decisions.
The DOE is endeavoring to create an environment where continuous improvement is
fostered. This continuous improvement atmosphere extends to organizations outside DOE
that have been contracted to provide analytical testing support for DOE programs. A
performance-based approach to analytical testing allows the flexibility to implement the
continuous improvement philosophy by allowing the DQO process to drive the selection of
sampling and analytical methods.
Standard/promulgated methods currently available for analytical testing do not address the
problems associated with measuring analytes in complex waste and environmental matrices
found at DOE sites, nor do they provide the data quality required for sound site remediation
and waste management decisions.
To bridge this gap, the Laboratory Management Division of DOE is supporting
development of DOE Methods for Evaluating Environmental and Waste Management
Samples (DOE Methods). DOE Methods contains guidance and methods that are not
currently provided by existing guidance manuals for DOE and DOE contractor laboratories
to use for radioactive waste and environmental sampling and analysis at DOE sites. DOE
Methods recommends that analytical methods contained within the document be
considered, through the DQO process, for use in support of DOE programs. However,
encouragement is given to use alternate methods if they can be demonstrated to provide a
134
-------
cost effective alternative to methods published in DOE Methods or other recognized
methods documents (Goheen et al. 1992)
This paper describes the elements of a performance-based approach to analytical testing and
the strengths and weaknesses of the approach compared to a strictly prescriptive approach.
Readers of this paper are encouraged to provide constructive comments on the
performance-based approach that incorporates the DQO process. All feedback will be
given serious consideration as this subject is addressed in forthcoming revisions of DOE
Methods. All comments should be addressed to Dr. Steven C. Goheen, at Pacific
Northwest Laboratory, P.O. Box 999, MS P8-08, Richland, WA 99352.
ACKNOWLEDGMENT
This work is funded by the Laboratory Management Division of the Department of
Energy's Office of Technology Development.
REFERENCES
American Society for Testing and Materials (ASTM) Standards. 1986. "ASTM Standard
Guide for Qualification of Measurement Methods by a Laboratory Within the Nuclear
Industry." ASTM C 1068-86. Annual book, Vol 12.01.
Dux, J. P. 1990. Handbook of Quality Assurance for the Analytical Chemistry
Laboratory, Second Edition. Van Nostrand Reinhold, New York 10003, p. 62.
Goheen, S. C., M. McCulloch, S. K. Fadeff, D. S. Sklarew, B. L. Thomas, R. M. Bean,
G. K. Ruebsamen, S. M. Anantatmula, and R. G. Riley. 1992. "A Guidance and
Methods Compendium: DOE Methods for Evaluating Environmental and Waste
Management Samples." Proceedings, 8th Annual Waste Testing & Quality Assurance
Symposium, pp 105-124.
U.S. Environmental Protection Agency (EPA). 1986. Development of Data Quality
Objectives: Description of Stages I and II. Quality Assurance Management Staff.
Washington, D.C.
Youden, W. J., and E. H. Steiner. 1975. Statistical Manual of the AOAC (Association
of Official Analytical Chemists). Washington, D.C.
135
-------
21 QUALITY ASSURANCE AND DATA COLLECTION: ELECTRONIC DATA TRANSFER
L. M. Tomczak Senior Staff Scientist, J. C. Gore, Technologist II, W. G. Lohner,
Engineer I, E. C. Ray, Meteorologist, and J. A. Selasky, Technologist II, Fernald
Environmental Restoration Management Corporation, Radiological Environmental
Monitoring group, P.O. Box 398704, Cincinnati, Ohio 45239-8704 and H. B. Spitz
PhD, University of Cincinnati, College of Nuclear Engineering, Cincinnati, Ohio
45242-0072
ABSTRACT
The Radiological Environmental Monitoring (REM) group at the Fernald
Environmental Management Project is involved in an Electronic Data Transfer
practice that will result in the improved quality assurance of collected data.
This practice focuses on electronic data transfer from the recording instrument
to reduce the manpower normally required for manual data entry and improve the
quality of the data transferred.
The application of this practice can enhance any data collection program where
instruments with electronic memories and a signal output are utilized.
Organizations employing this practice can strengthen the quality and efficiency
of their data collection program. The use of these practices can assist in
complying with Quality Assurance requirements under ASME NQA-1, RCRA, CERCLA, and
DOE Order activities.
Data from Pylon AB-5 instrumentation is typically configured to print data to a
tape. The REM group has developed a process to electronically transfer stored
data. The data are sent from the Pylon AB-5 field instrument to a Hewlett-
Packard portable hand computer, model HP95LX. Data are recorded and stored on
a 128 K-byte RAM card and later transferred to a PC database as an electronic
file for analysis. The advantage of this system is twofold: (1) Data entry
errors are eliminated and (2) considerable data collection and entry time is
eliminated. Checks can then be conducted for data validity between recorded
intervals due to light leaks etc. and the detection of outliers.
This paper/presentation will discuss the interface and connector components that
allow this transfer of data from the Pylon to the PC to take place and the
process to perform that activity.
INTRODUCTION
Fifteen continuous radon monitors have been installed at the DOE Fernald
Environmental Monitoring Project (FEMP) at both on-site and off-site locations
to measure what, if any, contribution to the natural radon background is made
by sources of radon located on the Fernald project. On-site outdoor locations
for each of the samplers have been selected to provide representative
measurements of radon that are either close to the sources of emission (K-65
silos that contain concentrated radium byproducts) or near buildings and areas
where production occurred. Several monitors have measured radon at the FEMP
fenceline and from two background locations a minimum of 13 miles from the
project including one located in the opposite direction of the prevailing
wind.
Each monitor prints the results of hourly measurements and also stores the
136
-------
data i.n the instrument computer memory. Each result consists of four fields
of data: sequence number, hour (24 hour clock), instrument response
(counts/hour), and radon concentration (pCi/1). Data is retrieved from each
instrument by environmental data collection personnel on a weekly basis.
Data collection personnel initial each data paper tape and reset the
instrument sequence number to initiate another run. An identification number,
calibration factors, and other data constants are printed on the data tape
whenever the instrument is reset. Data are also collected/transferred
electronically onto a hand held computer at each location. The instrument
response is also conducted using a check source on a weekly/monthly basis per
established protocol.
The radon collection instruments are housed in environmental enclosures to
protect them from the direct adverse effects of the environment but are not
heated or sealed in any manner. They operate 24-hours a day and are expected
to function properly in all types of weather and generate 168 hourly readings
each week. The instruments sample radon in the ambient air based upon the
diffusion principle and, thus, have no moving parts or pumps. Other than
weekly inspection by the data collection personnel to retrieve data, reset the
instrument, and insure that the instrument is still operating, no other
operator attention is provided or required. Instrument re-calibration is
scheduled on an annual basis contingent upon the satisfactory instrument
source check results.
OVERVIEW
Collection of data, albeit an important function, comprises only half of the
environmental monitoring activity. Sometimes it can be just as cumbersome to
appropriately deal with the data that has been collected, as it can be to
collect the data. The Radiological Environmental Monito-ing group at the DOE
FEMP facility has identified a solution, to at least take some of the dread
and monotony out of the voluminous environmental data reduction and reporting
activity, without sacrificing quality.
OBJECTIVE
Obtaining a workable electronic file for generating reports from data tapes
requires considerable labor intense quality assurance activities. These steps
are depicted in Figure 1. To relieve the burden of work related to the data
collection and reporting, a method has been devised to transfer the data
electronically. These steps are depicted in Figure 2.
There are essentially two steps in the electronic data transfer process. The
first step is the transfer of an electronic data file from the Pylon AB-5 unit
to RAM card on a handheld computer in the field. The second step is the
transfer of field data from the RAM card to an electronic data file back at
the office on the hard drive of a PC.
This data is then transferred via disk along with the hard copy data tape of
the Pylon AB-5 data to a data previewer, and then to an independent group that
produces the environmental reports. This activity should eliminate data
errors and data entry form.
137
-------
MATERIALS NEEDED
The following is a list of equipment that is needed to complete the electronic
data transfer from the AB-5 Pylon to a personal IBM disk operating system
computer.
. Pylon AB-5 Radon Monitor
. Pylon Model PPT-1 Printer
. Pylon Model CI-55 Computer Interface
. CPRD (radon detector) or a 300A Lucas cell
. DB15 Female A/B Switch Box
. DB15 3 foot M.M. Transfer Cable
. DB15 3 foot M.F. Transfer Cable
. DB25/9 M.M. Gender Changer
. DB15 M.M. Gender Changer
. HP95LX Computer
. HP F1001A Connectivity Pack
. HP F1002A 128 K-byte RAM card
. 128 K-byte RAM card drive linked to PC
METHODOLOGY
Data collection personnel have procedures for performing instrument checks and
retrieving data from each of the instruments. They observe instrument
operation and determine that the data output tape appears normal. They also
initial the data tape, electronically transfer the data, and reset the
instrument. Should the instrument appear to be malfunctioning, they will
notify management and initiate a service request from the instrumentation
group. Periodically, after electronically transferring and collecting the
data, the data collection personnel perform an instrument response check by
removing the radon sampler and attaching a check source. Following the
performance check, the data collection personnel re-attach the radon sampler
to the monitor, and reset the instrument.
The following are the steps that are needed to transfer the electronic data
file from the Pylon AB-5 unit to a personal IBM disk operating system
computer. Each step of transfer will be presented in detail in the Appendix.
QUALITY ASSURANCE OF TRANSFERRED DATA
To ensure that the data are successfully transferred from the Pylon AB-5 to
the RAM card, a quality control check is performed each week. During the
weekly inspection one data tape from those collected is randomly selected and
10% of the data points are randomly chosen and compared to the same data
points on the electronic file. Once each quarter, one data tape is randomly
selected and all data tape data points are compared with the electronic file
data points.
SUMMARY
Using the electronic data transfer scheme described in this paper, radon data
can be easily converted into an electronic file for further processing. Once
the data are transferred to a disk, the data can be previewed by the
individual assigned the task of quality assurance and data validation. It
138
-------
will be necessary to parse the data using a Lotus function to be able to work
with each of the columns. This data translation process would also be
required if other software programs (any software capable of handling ASCII)
were used in the transfer process. A check of data will then need to be
conducted to confirm that the data was correctly transferred and that all data
points were valid observations.
By utilizing the electronic transfer of data, considerable time can be saved
in both the areas of manual data entry, and the checks that are required to
ensure that the data is correct. Activities that had previously taken more
than a day can be performed in much less than a day without sacrificing data
quality.
REFERENCES
HP 95LX User's Guide, Hewlett-Packard Co., 1991, Corvallis, OR
139
-------
Data Collection
Personnel
QC/ Invalid Data
Discrepancy Control
Review Data for
Completeness/
Readability
Perform Double
Key Entry
Yes
Perform
Maintenance/ Repair
s Data
Complete
and
Readable?
Generate
Reports
t
w
Statistical
Analysis
FIGURE 1
Manual Data Collection and Data Entry
140
-------
Attach HP
Connection Pack
to AB-5 Unit
Initiate HP
BSLX for Data
Transfer
Place A/B Switch
to "B" Position
Return A/B
Switch to "A"
Position
Disconnect HP
Connection Pack
Import Raw Data
Files into Lotus
123
Save Data Files
in Lotus 1 23
Copy Files to
RAM Card in HP
9BLK
Remove RAM
Card from HP
95LX
Insert RAM Card
into PC RAM
Drive
Review Data
FIGURE 2
Data Transfer from AB-5 to HP 95LX
Data Transfer from HP 95LX to PC
141
-------
APPENDIX
TRANSFERRINS DATA FROM THE PYLON MONITOR TO THE HP95LX COMPUTER
1) Hook up the HPF100A1 Connectivity Pack to the HP95LX and the CI-55
Computer Interface
Connect the CI-55 to a DB-15 M.M. Transfer Cable.
2) Prepare the HP95LX for use
Turn on the HP95LX
Press: ON/OFF key
Press: BLUE FILER key
Select PRN (DIR)
Press: CURSOR DOWN key
Note: Delete all PRN files to start off the day. This will
ensure that sufficient memory will be available on the
unit to collect the data.
Press: F9 key to tag the old file to be deleted
Press: F3 key
Y key (to delete)
MENU key
Press: Q key to quit
Press: COMM key
Press: MENU key
Press S key for SETTINGS selection
Press: U key for USE selection
Press:Cursor Arrow —> key over to File (PYL.DCF)
Press: ENTER key
Check the following setup information:
Port:
Inter.
Baud
Stop
Parity
Char
Dial:
Type
1
3
1
none
8
Puls
Press: Q key for QUIT selection
Hold CTRL key and press F-5 key simultaneously
Backspace to DAT and delete by using the <— (Break) key
Type PRN\name.prn
Press: ENTER key
142
-------
Note: 8 characters max. name
Press: MENU key
Press: C key for CONNECT selection
(at this point set up the pylon to send data)
3) Transferring data from the pylon unit to the HP95LX:
Pylon Monitor
Press: START/STOP key to stop current run
Switch Box
Place A/B switch to B position
Pylon AB-5 unit
Press: RECALL key
Press: PROG STEP key, this will show the most current run.
If it is necessary to collect a different run use the PUMP key to
change the first digit and the START/STOP key to change the second
digit. After the RUN is selected press the PROG/STEP key 3 times.
Press and hold: STATUS key
Display will show the run, cycle, and interval.
Press: PROG/STEP key while holding the STATUS key
This will start the data transfer.
Release: PROG/STEP and STATUS key.
When the data transfer is complete the Pylon AB-5 display will
display the word "ready" .
Switch Box
Reposition A/B switch to the A position
Disconnect CI-55 Computer Interface cable from transfer
cable on switch box.
HP96LX Computer
Press: MENU key to exit program
Press: Q key for QUIT selection
4) Checking data transfer
Press: FILER key
Select the PRN (dir) ENTER key
(check if file and data exists)
Press: MENU key
Press: Q key for QUIT selection
Turn off instrument
Press: ON/OFF key
Note: Prepare for next station. At the next location it is not
necessary to repeat step #2, but step #1, the hookup is
still applicable.
143
-------
DATA TRANSFER FROM THE INTERNAL DRIVE OF THE HP95LX TO THE RAM CARD
1) Transfer data from the HP95LX capture file to the HP95LX Lotus 123
Directory on the RAM Card
Turn on the HP95LX
Press: OFF/ON key
Press: LOTUS 123 key
Press: MENU key
Press: F key for FILE selection
I key for IMPORT selection
T key for TEXT selection
Arrow to —> _PRN\ enter
Search for file of interest
Press: ENTER key
2) Saving the file
Press: MENU key
F key for FILE selection
S key for SAVE selection
Press ESC key two times
Type in A:\file name
Press: ENTER key
Press: R key for REPLACE selection if necessary to write over
previous data
Press: ENTER key
Repeat this step to transfer the files to the LOTUS (dir Drive A
RAM card)
Transfer data from the 128 K-byte RAM Card to the PC via the RAM Card
Drive
Press: MENU key
Press: Q for QUIT selection
Press: Y for YES selection
Turn off the HP95LX
Press: ON/OFF key
144
-------
THE FOLLOWING ARE DISCLAIMERS FOR ENVIRONMENTAL PAPERS/REPORTS:
THIS ABSTRACT/PAPER/REPORT 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.
NOTICE FOR OTHER THAN ENVIRONMENTAL REPORTS (INTERNAL SPECIALS, SPECIALS,
SUBCONTRACTOR, TOPICAL REPORTS):
THIS REPORT WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED BY AN AGENCY OF THE
UNITED STATES GOVERNMENT. NEITHER THE UNITED STATES GOVERNMENT OR ANY AGENCY
THEREOF, NOR ANY OF THEIR EMPLOYEES, NOR ANY OF ITS CONTRACTORS, SUBCONTRACTORS
NOR THEIR EMPLOYEES MAKES ANY WARRANTY, EXPRESS OR IMPLIED, OR ASSUMES ANY LEGAL
LIABILITY OR RESPONSIBILITY FOR THE ACCURACY, COMPLETENESS, OR USEFULNESS OF ANY
INFORMATION, APPARATUS, PRODUCT, OR PROCESS DISCLOSED, OR REPRESENTS THAT ITS USE
WOULD NOT INFRINGE PRIVATELY OWNED RIGHTS. REFERENCE HEREIN TO ANY SPECIFIC
COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY TRADE NAME, MANUFACTURER OR OTHERWISE,
DOES NOT NECESSARILY 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.
DOE-FERMCO PRIME CONTRACT NO. IS DE-ACOR-920R21972.
145
-------
22 AN INTEGRATED APPROACH TO MAINTAINING QUALITY
ASSURANCE DURING THE RI/FS PROCESS
Alex Tracy, William O'Brien, William Mills
Woodward-Clyde Federal Services, Rockville MD
Abstract
Recently much attention has been given to the applications and requirements of
Laboratory Information Management Systems (LIMS), including maintenance of data
integrity, implementation of automated scripts for the processing and reporting of data and
data security. However, data management and tracking during the field investigation
phase of a project have not received the same scrutiny. The case study described in this
paper focuses on two potential problem areas in the data collection/reduction process: the
standardization and documentation of sample tracking, numbering, and the collection
process; and the automation of electronic data transfer, analysis and review.
Introduction
Interest in the project began when an unusual number of cases of Hodgkin's
disease were reported in the area of a former army hospital. Public concern over a
biomedical waste incinerator operated at the army hospital prompted the U.S. Army Corps
of Engineers (USACOE) to launch a site investigation. The purpose of the site
investigation was to provide a baseline risk assessment, a contamination assessment and an
ecological risk assessment for the areas of concern. Due in part to the high risk of
litigation, Woodward-Clyde Federal Services (WCFS) augmented its existing sample
tracking program and implemented strict sample handling and documentation procedures.
This system was designed to track all samples from collection through the reporting and
data analysis phases with a high degree of confidence. All samples were systematically
numbered and logged upon collection. The analytical laboratories supplied documentation
detailing sample receipt and status in the analytical process at specified time intervals.
Both analytical results and supporting QA/QC information were uploaded into a WCFS
database using a data-parsing program that converts ASCII text files into database format.
Once the data from the analytical laboratories was uploaded into the WCFS
database, programs were developed to compile statistical information in support of the on-
going human and ecological risk assessments. Programs were also developed to review
and qualify data, and provide output suitable for the final report. By training the field
team and maintaining strict documentation, samples could be accurately tracked from the
field through the lab until results were reported. By utilizing data-parsing programs along
with batch processing to automatically upload data to the database system, WCFS
eliminated manual entry of data without having to write complex computer programs that
could not offer the flexibility required by the data analysis. A diagram that outlines the
146
-------
Data Flow Diagram
Field Information on Samples
Sample Log-books
COCs
Lab Info on Samples
Sampling database from lab
Data from lab
COC Check to ensure
data completeness
Sampling Database
Data extraction using
Monarch
O <
IP
3 ^^
JQ o.
Paradox
O pj
£?. O.
° £|
OQ
r-t-
0
•a
^.
S-
v
a.
en
O
ft>
Quality Control
Sample log-books
1 Ecological Risk
Evaluate holding
times
1
1
Risk Assessment
Stat Tables
Data Distribution
Graph. Representation
Assignment of CT &
RME Values
Calculate Risk
Method Blank
Contamination
1
Surrogate
Recoveries
Contamination Assessment
Hits Tables
Background Assessment
fD
I
o>
en
O
a.
c
o
s.
en'
T3
en
O
el-
s'
O>
-------
Roughly 520,000 records were manipulated in support of this project, and every
milestone for deliverables was met on time. Through documentation and automation,
WCFS was able to maintain the integrity of the data during the course of the project. This
presentation will discuss the systematized methods used to manage data and their impact
on the successful completion of the project.
Systematic Numbering Procedure
In order to glean as much information as possible from the sampling event, all
sample numbers were coded to provide information about the date of sampling, the sample
matrix, whether the sample was a QC sample, etc. To make the numbering system as
simple as possible, the date was used as a key element of the sample number. The site
identification or spatial location were not part of the coded sample number because all of
the sites would have to be selected and coded for sampling well in advance of the sampling
event. The basic format of the sample number was:
AAYYMMDD-##XX
where AA refers to the installation or site code
YYMMDD is the date (in the format year-month-day)
## is the number of the sample collected for that day
and XX is an extension that tells the matrix being sampled whether the sample is a
QC sample.
This method of coding information into the sample numbers provided field
personnel with a fairly simple method of numbering the environmental samples in a
systematic fashion. It also provided the staff members checking the sampling data with
additional information that allowed for further QC of the dataset.
Field Sampling Documentation
Sample documentation was another key element in the quality control of the RI:
chains of custody were used to document sample transport, but a sample log-book, along
with extended documentation provided from the laboratory, made it easier to track and
QC samples from collection to analysis and invoicing. Chains of custody were
sequentially numbered in the same fashion as bank checks. If a COC was voided, it was
retained to ensure that all COCs were accounted for.
The sample log-book contained the coded field sample number, the sample's spatial
location and depth, the analyses requested, the COC number used to document sample
transport and the initials of the individuals performing the sampling. The labs were
required to confirm receipt of samples by a written statement listing the WCFS field
sample number, the corresponding lab sample number, the analyses requested for that
148
-------
sample and a copy of the chain of custody for those samples. Every two weeks the labs
were required to submit a status report listing all samples still within the analytical process
at the laboratory as well as their status within the lab (logged, prepped, analysis complete,
etc.)
At major milestones in the project the lab was required to submit an ASCII file
correlating WCFS field sample numbers with lab sample numbers. This allowed a check
of the WCFS sample log-book and all of the information provided by the lab, data upload
and analysis could begin.
Data Extraction from Diskette Deliverables
Contract Laboratory Protocol (CLP) was used as the primary analytical
methodology for the RI/FS (herbicides and dioxins were analyzed using SW-846
methods). This protocol is characterized by very specific quality control, documentation,
and deliverables. The CLP diskette contains a wealth of information seldom used in its
entirety. However, by using data-parsing and pattern-scanning programs, WCFS was able
to extract both the analytical results and all of the supporting quality control information
that allowed WCFS to perform an automated data review.
Templates to extract data from diskette deliverables were developed using
Monarch, a data-parsing program. This program allows the user to read ASCII files,
define traps that key on selection criteria to determine which records to extract, name and
format all fields for export into a database format. Templates could be developed in
roughly a half-day, and then all data could be extracted without any data entry. Gawk, a
pattern-scanning language originally developed for the UNIX operating system, was used
to create the batch files that would run Monarch. It was used to extract the information
from the diskettes for roughly 150 separate sample delivery groups. This totaled nearly
520,000 records for six different analytical methodologies.
Quality Control
Master database files were created for each of the six analytical methods. This was
done by appending individual data files, and the files were then checked against the
information contained in the sample log-book to ensure that all samples were accounted
for and that the proper analysis had been performed and reported. A percentage of the
results in the database were checked against hard copies supplied by the lab and no errors
were found. For the entire project, only three errors were found in the entire data set, and
all were attributable to data entry errors at the lab.
Once the content of the database files had been checked to ensure accurate upload
had been performed, results were qualified by the chemistry staff. Queries were created to
evaluate hold times, method blank contamination, and surrogate recoveries for all samples.
149
-------
Sample results whose hold times had been exceeded were qualified, and method blanks
were correlated with their associated samples to perform a blank evaluation.
Under CLP methodology, the "B" flag is used to denote a compound which has
been detected both in an environmental sample and its associated blanks. If there is a
blank contamination, the EPA has issued guidance to qualify those results: for compounds
that are considered common lab contaminants to be considered as detects (not artifacts of
the sampling/analysis process), the concentration in the environmental sample must exceed
ten times the concentration found in any associated blank for those compounds. For all
other compounds (those not considered common lab contaminants), the concentration in
the environmental sample must be five times greater than the concentration found in any
associated blank. The table shown below lists the compounds found in method blanks:
Lab Contaminants Found in Method Blanks
Compounds
(Units are ug/L)
Volatiles
2-Hexanone
Acetone
Methy Ethyl Ketone
Methylene Chloride
Semivolatiles
3-Nitroaniline
4-Nitroaniline
Carbazole
Di-n-Butylphthalate
Di-n-Octyl Phthalate
Diethylphthalate
Phenol
bis(2-ethylhexyl) pthalate
Avg.
Value
2.56
7.26
4.03
1.91
40.00
46.00
34.00
40.50
82.00
3.00
1200.00
203.09
Minimum
Value
1.00
3.00
1.00
1.00
40.00
46.00
34.00
1.00
82.00
3.00
1200.00
1.00
Maximum
Value
5.00
18.00
19.00
14.00
40.00
46.00
34.00
80.00
82.00
3.00
1200.00
580.00
No. of
Detects
9
46
33
35
1
1
1
10
1
1
1
11
No. of
Samples
90
90
90
90
66
66
66
66
66
66
66
66
Roughly 293 detects in environmental samples had the "B" flag as a qualifier: after
blank assessment and qualification, only two detects with the "B" flag were considered
actual detects and not artifacts of the sampling/analysis process.
Surrogates were analyzed using Shewart charting techniques commonly used for
process control. Samples were grouped by matrix, analysis date, and sample delivery
group to check for trends resulting from either matrix effects, or instrument or operator
error. The standard deviation was calculated for each group of samples and warning and
control limits were calculated by multiplying the standard deviation by a factor of two or
three, respectively. The data for those samples with unacceptable surrogate recoveries
150
-------
was reviewed, the chromatograms were analyzed and if necessary, the results were
qualified. Some Shewart charts for sediment samples are shown on the following page:
Surrogate Recovery for Sediment Samples: Volatiles
130
70 -1-
• Recovery '
n UWL
— * LWL
— * UCL
—* LCL
Surrogate Recovery for Sediment: Semivolatiles
• Recovery '
Q UWL
— * LWL
— o UCL
— * LCL
Both blind and designated duplicates were submitted to the lab. They were
evaluated for the precision associated with both the sampling and preparation (sub-
sampling) methods. For samples with detects above the contract-required detection or
quantitation limits, the relative percent difference between a sample and its duplicate was
typically below 20%. Matrix spike and matrix spike duplicate results were also evaluated
and were found to have similar precision. Accuracy of the methods was evaluated using
151
-------
matrix spike/matrix spike duplicate results, laboratory control samples, and surrogate
recoveries.
Risk Assessment Calculations
In support of the risk assessment, samples were first grouped by exposure area for
each group of receptors. The samples were further subdivided by matrix. Contaminants
of concern were determined using automated queries to calculate minimum, maximum,
minimum non-detect, maximum non-detect, average, standard deviation, and number of
detects vs. number of samples for every chemical analysis by exposure area and media.
Data was log-transformed and the process was repeated. Both normal and log-
transformed data were evaluated with respect to background samples to determine if the
data set being analyzed was statistically distinct from the background results. Based on
that statistical information and frequency of detection, chemicals to be included in the risk
assessment were selected.
Plots similar to the one shown below were used as a graphic means to determine if
the background samples were part of an statistically distinct population as compared to
on-site samples. The graph shows the average plus or minus the standard error of the
mean (standard error of the mean^hree times standard deviation) for arsenic (AS),
beryllium (BE), cadmium (CD), chromium (CR), nickel (NT) and selenium (SE). On-site
results for a particular element are designated by that element's symbol, while background
results are shown as the symbol plus "-BG." Based on this technique, none of the
elements shown have statistically distinct populations between on-site and background
samples. It must be emphasized, however, that this test was merely one of several
techniques used to evaluate the presence or absence of on-site contamination.
Evaluation of Inorganic Analytes with Respect to Background
60 i
50
40
30
20
10
0
-10
c
c
ft
]
, I , ? , - , 1 , I ,
]
D
1
' 1 ' • ' ' • ' • '
S AS-BG BE BE-BG CD CD-BG CR CR,
C
1 1
c
' ' i
l
BG Nl NI-BG
C-BG" refer* to a value obtained from background samples)
Avg-Std Error of Mean ° Avg + Std Error of Mean • Average
152
-------
Once the chemicals of concern were selected, data distribution was evaluated to
determine the proper treatment of the data for risk assessment purposes. For this site,
calculations using log-transformed data proved to be more conservative, so the log-
transformed version of the data was employed for all further risk assessment calculations.
A number of scenarios were established based primarily on the number of samples
of a particular matrix for an exposure area. Those scenarios are listed below:
Scenario Summary for Risk Assessment Data
1 0 or more samples per matrix per exposure area
Less than 10 same
No detects per mat
UCL > max.
UCL > avg., UCL < max.
es per matrix per exposure area
avg. < max.
max < avg.
rix per exposure area
Values used for RA
CT
UCL
Average
Average
Maximum
Omitted
RME
UCL
UCL
Maximum
Maximum
Omitted
Note:
Maximum indicates the maximum detected value
Average is the average of all samples
UCL is the log-normally distributed UCL95 value
Given the scenarios presented above, the data was uploaded into a spreadsheet
application that had the compatibility to easily convert files from the database application.
Three fields were added to each record; they included the name of the scenario for which
the data belonged, and the Central Tendency (CT) and Reasonable Maximum Exposure
(RME) values that should be used with that scenario. The CT and RME values were
modified by uptake factors for the dermal and ingestion portions of the risk assessment.
For well and surface water, the CT and RME values were modified by a value Kp, which
is the permeability constant for dermal uptake of compounds from water. The CT and
RME were multiplied by an absorption factor for dermal uptake of compounds in direct
contact with the skin for all sediment, surface soil, and soil boring data. These
calculations were again automated through the use of macros to only calculate if the
situation was correct for that scenario. The final tables contained either the calculated
value or an "NA" if the data should not be modified by an uptake factor.
153
-------
Given the now reduced number of chemicals of concern with detects, the data was
uploaded into the Woodward-Clyde Risk Assessment application, Assess, which was used
to calculate risk. Following all calculations, the hard-copy reports generated were
checked for validity against the spreadsheet tables and the source database files. The
spreadsheet macros proved helpful in classifying the data into the correct scenario and
then choosing the values to be used in the risk assessment.
Conclusion
Methods can be easily employed during the field investigation phase of a project to
ensure levels of data integrity that are comparable to a Laboratory Information
Management System. The first phase, which deals primarily with the sampling event,
requires thorough planning, appropriate documentation and training of all field personnel.
Planning provides a forum for all project staff to arrive at a workable solution and
documentation provides a means of ensuring that sufficient information is gathered
regardless of the individual performing the task. Typically, field personnel are typically
given only the minimum training required by their job: proper training and complete
involvement in the project allows the staff to realize the intended uses of the data. This
translates into more accurate data collection.
The second phase of data management should be designed to minimize any manual
entry of data and should also allow sufficient information to be uploaded into a database
to permit automated review and analysis of all data. By using Monarch, a software that
will extract of information from ASCII files, it was possible to make use of all of the
information contained in a CLP diskette. By using a relational database, it was possible to
easily perform the ad hoc queries required to track all samples from the time of collection
and to provide output for analysis. This project made use of a variety of software
packages, including spreadsheets and risk assessment programs to minimize the chance for
human errors. These elements of WCFS1 program allowed the staff to complete the RI/FS
on-time and on-budget.
154
-------
23 EFFLUENT EMISSIONS MONITORING AT THE DOE HANFORD SITE
LaDell W. Vance. Principal Quality Engineer
Environmental Management Systems Integration
Westinghouse Hanford Company
P.O. Box 1970
Richland, Washington 99352
ABSTRACT
There are numerous regulatory requirements controlling the effluent
emissions monitoring at a U.S. Department of Energy site. This paper
defines how these regulatory effluent emissions monitoring requirements
and the Quality Assurance oversight of these requirements were implemented
by Westinghouse Hanford Company, the operations contractor, at the DOE
Hanford Site.
INTRODUCTION
Westinghouse Hanford Company (WHC) struggled with the implementation of
DOE and other regulatory effluent monitoring Quality Assurance (QA)
requirements. There was little guidance or direction provided on how this
should be done. By using the QA organization from the earliest planning
stages, a more effective compliant process has evolved.
These initial monitoring requirements were prescribed by the
implementation of DOE Order 5400.1 which specified that the sites
implement plans to monitor regulated hazardous and radioactive effluents
by November 1, 1991. Included in this order were the environmental QA
requirements to be used in this effort, but there was little direction
given for their implementation.
The DOE environmental QA requirements are very similar to U.S.
Environmental Protection Agency (EPA) Quality Assurance Management Staff
(QAMS) requirements that WHC had been using for well sampling and analysis
activities. These plans had been effective in monitoring well water so
WHC decided to use these as guidance for preparing plans for the liquid
effluents.
The radioactive air requirements from 40CFR61, Appendix B, Method 144,
Section 4, include the DOE requirements plus specific items related
directly to radioactive air measurements, items that would normally be
considered Quality Control (QC) activities. It was decided to prepare a
QA Program Plan (QAPP) in accordance with the EPA QAMS requirements and
include the stack specific measurement requirements as appendices to this
document.
155
-------
These effluent monitoring activities are currently being performed with
added enhancements as the effort continues.
HANFORD SITE
The Hanford Site covers approximately 560 mi2 (1,450 km2) of semiarid land
that is owned by the U.S. Government and managed by the U.S. Department of
Energy, Richland Operations Office (RL). In early 1943, the U.S. Army
Corps of Engineers selected the Hanford Site as the location of reactor,
chemical separation, and related activities for the production and
purification of special nuclear materials and other nuclear activities.
Since 1988, the mission of the Hanford Site has focused on environmental
remediation and restoration.
Activities on the Hanford Site are centralized in numerically designated
areas (see Figure 1). The reactors are located along the Columbia River
in the 100 Areas. The reactor fuel processing units are in the 200 Areas,
which are on a plateau approximately 7 mi (11 km) from the Columbia River.
The 300 Area, located adjacent to and north of the City of Richland,
contains the reactor fuel manufacturing plants and the research and
development laboratories. The 400 Area, 5 mi (8 km) northwest of the 300
Area, contains the Fast Flux Test Facility that is used for testing liquid
metal reactor systems. Other areas and buildings designated on the map
are for administrative purposes and do not produce regulated effluents.
Hanford Site facilities are no longer producing nuclear materials for
defense purposes, rather operations are continuing for cleanup purposes.
The air and liquid effluent streams are being generated during maintenance
or shutdown of existing facilities.
EFFLUENT MONITORING REGULATIONS AND QUALITY ASSURANCE REQUIREMENTS
Figure 2 depicts the roles of the DOE, EPA, Washington State Department of
Ecology (Ecology), Washington State Department of Health (DOH), and the
sources of the environmental QA requirements specified in the regulations.
This paper is limited to major effluent monitoring efforts of the
operating facilities that require regulatory control. The effluent
monitoring environmental QA regulations are identified by the double-lined
boxes on Figure 2. The RCRA/CERCLA efforts required for cleanup, waste
analysis, and efforts specified only for DOE purposes are not included.
A major factor in almost all environmentally regulated activities at the
Hanford Site is the Hanford Federal Facility Agreement and Consent Order
more commonly referred to as the Tri-Party Agreement (Ecology, et.al.
1992). This agreement among DOE, EPA, and Ecology specifies requirements
and schedules for most environmentally regulated monitoring, operating,
and construction activities. As noted on Figure 2, the Tri-Party
Agreement relies on the EPA QAMS documents for effluent environmental QA
requirements.
156
-------
The DOE effluent monitoring QA requirements, DOE Order 5400.1, General
Environmental Protection Program, Chapter 4, Section 10, do not provide
direction on implementation. WHC decided to use the QAMS requirements for
all liquid effluent monitoring activities. These requirements specify
similar requirements while providing implementation guidance.
The other sources of QA effluent monitoring requirements found were:
National Emission Standards for Hazardous Air Pollutants (NESHAP),
40CFR61, Appendix B, Method 114, Section 4, QA Methods
National Pollutant Discharge Elimination System (NPDES),
40CFR122.44, Establishing Permit Conditions
Toxic Substance Control Act (TSCA), 40CFR763.121, Appendix A,
Quality Control Procedures, and Appendix B, Detailed Procedures for
Asbestos Sampling and Analysis.
There is only one effluent stream, associated with the N Reactor, that has
an NPDES permit. Since all monitoring activities are specified and
controlled by this permit, the N Reactor stream will not be further
addressed in this paper. The TSCA controls are also not addressed in this
paper.
The remainder of this paper will discuss how the regulated liquid and
radioactive air streams were identified and the environmental QA controls
were applied.
INITIATION OF EFFLUENT MONITORING PROGRAM
Regulated effluent monitoring was instigated in response to DOE Order
5400.1, which required the preparation of effluent monitoring plans by
November 1, 1991. The Order states that
"An implementation plan shall be prepared for each facility or
group of facilities, the purpose of which is to ... comply
with environmental regulations and DOE policies."
WHC responded to this by preparing a management plan for facility effluent
monitoring plan activities to direct the identification and implementation
of these monitoring efforts. The management plan specified the regulatory
criteria that would require monitoring, the facilities that would need to
be addressed, the basic information needed in a Facility Effluent
Monitoring Plan (FEMP), and the schedule for completing the
determinations.
The criteria, as noted in the management plan are taken from the following
regulatory requirements:
(1) For Radioactive Air Emissions, 40CFR61.93, "Radioactive emission
measurements ... shall be made at all release points which have a
157
-------
potential to discharge radionuclides into the air ... in excess of
1% of the standard.' As the EPA allowable dosage is 10 mrem/yr all
facilities that could cause radiation doses in excess of 0.1
mrem/yr to the maximally exposed member of the public must be
monitored.
(2) For discharges of nonradioactive materials, monitoring is required
of all facilities with the potential to release quantities of
hazardous materials exceeding the reportable quantities listed in
40CFR302.4, "Designation of Hazardous Substances."
(3) All potential radioactive discharges to the ground must be
monitored.
The plan preparation required that all facilities be reviewed to determine
if any of the above requirements had been exceeded and thus require
monitoring.
FACILITY EFFLUENT MONITORING PLAN DETERMINATIONS
Each facility had to be thoroughly examined to identify each effluent, and
an analysis was done on each stream to see if any of the above criteria
were exceeded. The determinations of the streams and the preparation of
the FEMPs took over one year to complete.
To put this in perspective, there were 17 geographical areas that included
approximately 110 facilities. Of these, 15 facilities, with 22 hazardous
and/or radioactive liquid streams and 7 radioactive emissions stacks,
exceeded the criteria. Discussions are ongoing with the Washington
Department of Health that may add other radioactive stacks to this
program. There is also another ongoing program (Air Emissions Inventory)
to identify the hazardous air emissions.
FACILITY EFFLUENT MONITORING PLANS
Each facility with a regulated stream(s) was required to prepare a FEMP to
address the requirements specified in the management plan. To expedite
this effort and to ensure consistency between facilities, a guide for
preparing Hanford Site facility effluent monitoring plans was prepared.
This document expanded on the management plan requirements with guidance
on what to include in each of the following sections including:
Purpose and scope
Applicable regulations
Facility/process description, source terms, release pathways
Characterization of the airborne and liquid discharges to be
monitored, providing potential effluent concentrations
Characterization of effluent points, providing design criteria and
technical specifications of monitoring/sampling systems
158
-------
Historical monitoring/sampling data and a comparison of normal and
upset conditions
Sample analysis and laboratory procedures
Notification and reporting requirements
Interface with operational reporting requirements
QA Plan
Internal and external plan review requirements
Compliance assessment describing conclusions reached, if monitoring
was required and will system need upgrading.
The fifteen FEMPs were prepared to these criteria with the appropriate
reviews and approvals.
QUALITY ASSURANCE INVOLVEMENT
The QA organization was invited to join in this effort shortly after it
was started when the developers encountered problems trying to implement
the DOE Order 5400.1 QA requirements. The QA representative was familiar
with the QAMS document with its defined method of implementation and this
was used to implement the DOE requirements. The QA organization also
reviewed the implementation documents to ensure that these were reviewed
and controlled in accordance with DOE Order 5700.6C, Quality Assurance.
The facility effluent monitoring QA Project Plan that was prepared
described the existing facility procedures and how they would be used to
implement the QA requirements. However, this effort was not successful as
the existing procedures did not have the needed custody or control checks
required for regulatory environmental QA controls. In addition, each
facility had their own procedures with little correlation between them.
There was a need for some standard operating procedures (SOP) with the
needed QA/QC controls.
The FEMP documentation provided the background information that was needed
to prepare for monitoring the facilities, but other programs were used for
QA verification. This was done through the implementation of two other
regulatory requirements, the Liquid Effluent Characterization Program and
National Emission Standards for Hazardous Air Pollutants (NESHAP).
LIQUID EFFLUENT CHARACTERIZATION PROGRAM
At the time of the signing of the Tri-Party Agreement, the DOE agreed to
monitor and analyze 33 specified liquid streams, called the Liquid
Effluent Study (LES), in accordance with the EPA QAMS QA criteria. There
had been ongoing monitoring of these streams since 1985 in response to
congressional requests and the regulators requested that this be continued
and that EPA QA requirements be applied to the activities. These were the
same streams as those identified in the FEMP determinations; except
however, as the site continues to shut down facilities, some of the
streams are no longer in use.
159
-------
QUALITY ASSURANCE OVERSIGHT OF LIQUID EFFLUENT MONITORING ACTIVITIES
The initial LES sampling and analysis activity was done shortly after the
signing of the Tri-Party Agreement in 1989 without QA support and with no
QA/QC protocols. Since information could not be verified or validated it
was not acceptable to Ecology. After this was completed, the QA
organization was asked to perform a post-completion surveillance of the
sampling and analysis activities. The intent was to verify the
implementation of the required EPA QAMS requirements. After finding this
uncontrolled situation, the QA organization was invited to be an integral
part in setting up the QA Program to ensure the validity of future data.
It was agreed initially that a QA Project Plan (QAPjP) would be prepared
in accordance with the requirements of EPA QAMS-005/80. This was later
expanded to include management control and changed to a QAPP. There was
also consensus that all sampling and analysis activities could not be
included in a single plan. Those activities that were common to all
streams were included in the sitewide QAPP. Individual Sampling and
Analysis Plans (SAP) were prepared for stream-specific information.
The QAPP provides the general framework for sample collection, laboratory
analysis, and data reporting with the common QA requirements for the
characterization of liquid effluent streams. These common QA requirements
include program description, Data Quality Objectives (DQO), program
organization, sample control, laboratory analysis, and data processing.
The SAPs are wastestream-specific documents that describe how QAPP
requirements are implemented. The SAPs include a discussion of stream-
specific DQOs, a description of the liquid effluent, and a justification
for sample location and frequency. The SAPs also identify any stream-
specific exceptions to the QAPP.
An effluent monitoring organization was responsible for identifying the
regulatory requirements, preparing the QAPP, and specifying the
implementing requirements and criteria to include in the SAPs. The SAPs
format and content were based on Guidance for Conducting Remedial
Investigations and Feasibility Studies under CERCLA (EPA/540/G-89/004,
October 1988) for a field sampling plan (FSP).
The SAPs are prepared by each facility with input from the common sampling
organization. This coordination of sampling and analysis preparation
provides the control needed to achieve environmental QA implementation.
The environmental and facility QA groups work together to perform the
stream surveillances. The environmental QA group reviewed and approved
the QAPP and SAPs. The QA group's familiarity with these documents allows
them to prepare the stream surveillance plans. The facility QA personnel
are familiar with the facility and thus better able to witness the field
sampling activity. Environmental QA is responsible for the surveillance,
but both groups coordinate and approve the surveillance findings.
160
-------
RADIONUCLIDE AIR MONITORING
The Clean Air Act was amended in December 1989 to include 40CFR61, Subpart
H, National Emission Standards for Emissions of RadionucTides Other Than
Radon from Department of Energy Facilities. This subpart requires that a
QAPP for radionuclide air monitoring be prepared in accordance with the
requirements of 40CFR61, National Emission Standards for Hazardous Air
Pollutants (NESHAP), Appendix B, Method 114, Section 4, QA Methods.
The Washington Department of Health was satisfied that the seven stacks
required monitoring, but did not feel that the FEMPs adequately covered
the continuous monitoring requirements. The DOH required that regulated
stacks and analytical laboratories be Point by Point (PBP) compared to
40CFR61, Appendix B, Method 114 to be certain that all requirements were
satisfied. This major PBP exercise compared the existing air monitoring
equipment and program with that specified in the requirements, including
Section 4, QA Methods.
It was decided to use the QAPP as the binding document to control the
NESHAPs activities. All stack monitoring sampling and analytical
requirements are included in the document.
NESHAPs QUALITY ASSURANCE PROGRAM PLAN (QAPP)
There was a concern as to how to address the QA requirements of Method 114
for all the various activities at the stacks. It was initially proposed
that an overall QAPP be prepared with separate QAPjPs for each stack. It
was finally agreed that a QAPP with the stack PBPs appended would address
all QA program requirements.
The QAPP was prepared in accordance with the guidance provided by EPA
QAMS-004/80 and is organized as shown on Figure 3, NESHAPs Table of
Contents. The major responsibilities and actions needed to control the
NESHAPs program are defined in Sections 3.0, 6.0, and the appendices.
Section 3.0, Quality Assurance Management, defines the various oversight
organizations that are responsible to verify that the NESHAPs activities
are being properly done. The body of the QAPP describes these various
organizations and the procedures they use to perform these oversight
activities. WHC is organized with a central QA group, Environmental
Services Quality Assurance. This central group reviews and approves
documentation and provides guidance to facility QA organizations, who
assist with field overviews. The same relationship exists as in the
liquid effluent monitoring where actions are combined, but the central
environmental QA group is responsible for the surveillances.
Section 6.0, Radioactive Air Emissions Measurement Quality Assurance
Project Implementation, specifies the organizations that are performing
the NESHAPs activities. This section describes organizational
responsibilities and interfaces. Included are responsibilities for
specifying NESHAPs requirements, sampling and sample transport, laboratory
161
-------
analysis, flow data information, systems calibration, regulatory reports,
procedures preparation, and data verification. Section 6.0 was difficult
to prepare but defining these responsibilities and organizing the
activities was crucial to properly completing the NESHAPs activities. In
preparing this document it was determined that some activities were not
being properly addressed. These were corrected and included in the
document.
The PBP comparisons for the seven stacks, appended to the QAPP, provide
the total overall program control needed for NESHAPs activities. Items
that are specific to a stack are included in the PBP, but standard
operating procedures are included, or referenced, in the body of the QAPP.
There is an added advantage to this in that if additional stacks have to
be registered a stack-specific PBP can be prepared and appended to the
existing QAPP. There will be no need to prepare a QA Plan for each added
stack.
The QAPP provides the needed control of the NESHAPs activities but allows
the flexibility to add regulated stacks as needed.
SUMMARY
The preparation and implementation of the QAPP was crucial to properly
organizing the activities and obtaining verifiable data.
Both liquid and air monitoring started out without Environmental Services
QA help and the responsible facilities were at a loss as to how to
implement these new unique sampling and analysis requirements. The
Environmental QA person was able to define what information was needed for
the QAPPs and help the responsible organizations in their preparation and
implementation.
It is important to have the implementing organizations prepare the plans.
In the case of the FEMP activities, outside consultants prepared most of
the documentation without facility input. The facilities were not aware
of the commitments made and were hesitant to implement these. As noted in
the document, the LES program is used to verify the control of liquid
effluents.
In both the liquid and radioactive air monitoring existing facility
procedures were initially used; however, it became imperative to prepare
SOPs. In the FEMP program the existing procedures were not definitive
enough nor did they have the needed QC controls. In the NESHAPs program
it was determined that it would be easier to have common procedures and
these have been prepared.
In closure, someone familiar with Environmental QA requirements, but not
part of the assessment organization, should be a working partner with the
implementing organizations in preparation and implementation of
environmental QA requirements. This relationship will allow for
compliance and will ensure that the data gathered is verifiable.
162
-------
Radioactive
Mixed Waste
Storage Facility
3000 Area
H930S018.2
Figure 1
163
-------
40 CFR 763.121
App. A and B
o
i 40 CFR 61 App. B
1 Method 114
RCRA/CERCLA
Closure Plans
Ch.4,
Section 10
Ch.4,
Section 10
CAA
CWA
DOE
DOH
Ecology
EPA
FEMP
LES
NESHAP
NPDES
OEMP
QAMS
TPA
TSCA
WAP
EPA Authorizes State Agencies to Regulate
= Clean Air Act
= Clean Water Act
= Department of Energy
= Washington State Department of Health
= Washington State Department of Ecology
= Environmental Protection Agency
= Facility Effluent Monitoring Plan
= Liquid Effluent Study
= National Emission Standards for Hazardous Air Pollutants
= National Pollutant Discharge Elimination System
= Operational Environmental Monitoring Plan
= EPA Quality Assurance Management System
= Tri-Party Agreement H93osoi8.i
= Toxic Substances Control Act
= Waste Analysis Plan
Figure 2
-------
Figure 3. National Emission Standards for Hazardous Air
Pollutants Table of Contents.
CONTENTS
1.0 INTRODUCTION 1
2.0 QUALITY ASSURANCE POLICY STATEMENT 1
3.0 QUALITY ASSURANCE MANAGEMENT 1
3.1 ENVIRONMENTAL ASSURANCE 3
3.1.1 Environmental Protection 3
3.1.2 Environmental Compliance Verification 3
3.2 QUALITY ASSURANCE 3
3.2.1 Environmental Services Quality Assurance 3
3.2.2 Plant Quality Assurance Engineers . 4
3.2.3 Audit Program Administration 4
4.0 DOCUMENT CONTROL AND RECORDS 5
4.1 ENVIRONMENTAL ASSURANCE . 5
4.1.1 Environmental Protection 5
4.1.2 Environmental Compliance Verification 5.
4.2 QUALITY ASSURANCE 5
4.2.1 Environmental Services Quality Assurance 5
4.2.2 Plant Quality Assurance 5
4.2.3 Audit Program Administration 5
5.0 PERSONNEL QUALIFICATIONS 6
6.0 RADIOACTIVE AIR EMISSIONS MEASUREMENT QUALITY ASSURANCE
PROJECT IMPLEMENTATION 6
6.1 INTRODUCTION 6
6.2 ORGANIZATION AND RESPONSIBILITIES 7
6.2.1 Environmental Protection 7
6.2.2 Facilities 7
6.2.3 Occupational Health and Safety 10
6.2.4 Central Support Services 10
6.2.5 Laboratories . 11
6:2.6 Regulatory Analysis 12
6.2.7 Facility Compliance 12
6.2.8 Other Support Contractors 12
7.0 PERFORMANCE AND SYSTEM AUDITS 12
8.0 CORRECTIVE ACTION 13
9.0 QUALITY ASSURANCE REPORTS 13
10.0 REFERENCES 13
APPENDICES
A METHOD 114 COMPARISON FOR STACK 291-A-1 A-1
B METHOD 114 COMPARISON FOR STACK 291-8-1 B-1
C METHOD 114 COMPARISON FOR STACK 291-Z-1 C-
D METHOD 114 COMPARISON FOR STACK 296-A-22 D-
E METHOD 114 COMPARISON FOR STACK 296-A-40 E-
F METHOD 114 COMPARISON FOR STACK 340-NT-EX F-
G METHOD 114 COMPARISON FOR 222 LABORATORY G-
H METHOD 114 COMPARISON FOR 325 LABORATORY H-
165
-------
24 QUALITY ASSURANCE AUDITS OF LABORATORIES
R. Terry Hinward. Senior Quality Engineer
Environmental Services Quality Assurance
Westlnghouse Hanford Company
P.O. Box 1970
Richland, Washington 99352
ABSTRACT
Experience gained through audits of numerous laboratories across the
United States has helped to focus on some common problems found in
laboratories. Various regulatory requirements and programs help to
complicate the audit process. Many laboratories are almost overwhelmed by
the requirements, and the quantity and diversity of audits imposed upon
them. Occasionally auditors add to this confusion as they sometimes
arbitrarily require the laboratories to follow set protocols of
documentation instead of focusing on the end products of the systems in
place. Such actions can result in the laboratories developing multiple
parallel systems to produce the same end results. This process of
multiple systems can be very costly and a source of numerous errors within
the laboratory. This presentation focuses on common pitfalls found within
Laboratory Quality Assurance/Quality Control Systems and in the process of
auditing such systems and their performance outcomes.
INTRODUCTION AND BACKGROUND
Requirements from numerous regulatory or governmental agencies seem never
to stop growing and expanding in both scope and complexity. Guidance for
interpretation and application of these requirements is not always clear
or maybe even available. This can add more confusion as to exactly what
is required and how to approach the various situations.
To compound these requirements, numerous companies, associations and
boards publish or provide guidance and or requirements to be followed.
Throughout the years, constant change and additions to requirements has
produced a tremendous burden on both the laboratories and those auditing
the laboratories. Even the terminology and approaches used to evaluate
the laboratories has become very diversified and confusing. At present
the author evaluates some laboratories within the DOE System and several
Commercial Laboratories. Samples sent to these laboratories by
Westinghouse Hanford Company may be analyzed under several different
regulations.
166
-------
Examples of these regulations are
the Safe Drinking Water Act (SDWA)
the Comprehensive Environmental Response, Compensation and
Liability Act (CERCLA) and it's Superfund Amendments and
Reauthorization Act (SARA)
the Resource Conservation and Recovery Act (RCRA) and it's
Hazardous and Solid Waste Amendments (HSWA)
the Clean Air Act (CAA) and the National Emission Standards
for Hazardous Air Pollutants (NESHAP)
the Clean Water Act (CWA) and the National Pollutant Discharge
Elimination System (NPDES)
the Water Quality Act of 1987 (WQA)
the Toxic Substances Control Act (TSCA)
the Federal Insecticide, Fungicide and Rodenticide Act
(FIFRA), and
the Solid Waste Act (SWA).
Sample types include:
radiochemical
mixed waste
hazardous waste
water quality
air quality, and
can consist of several matrixes.
These matrixes include:
water
air
solid waste
animal tissue, and
vegetation.
These samples may or may not be subject to the Hanford Federal Facility
Agreement and Consent Order of May 1989. This agreement is between the
Washington State Department of Ecology, the United States Environmental
Protection Agency and the United States Department of Energy. It is
commonly referred to as the Hanford Tri-Party Agreement. Additional
requirements from this agreement are contained in most of the Statements
of Work (SOWs) to contract laboratories. It is extremely difficult to
provide all exacting guidance in this fashion due to the enormous
diversity of work.
A discussion of lessons learned and some common pitfalls will be
presented. It is not the intent to single out laboratories or types of
laboratories but to discuss things found throughout several laboratories
in the nation. No attempt will be made to quantitate the extent of such
problems or to statistically treat this subject.
167
-------
AUDIT. SURVEILLANCE. ASSESSMENT OR INSPECTION?
In evaluating these laboratories each of these terms has been used in
different notifications to laboratories about upcoming activities.
Several laboratories have expressed concern about what to expect based on
what term is used. EPA in general does different types of audits (system
and performance as an example), whereas the NRC calls similar activities
inspections. The DOE sometimes breaks these activities into
surveillances, inspections, assessments and audits. In this paper, the
term audit will be used to cover all these activities. Each agency can
readily define what is meant by the terms it uses but often fails to do so
in a clear way to the laboratories it associates with. What is important
to the laboratory is to know the extent of the activity (audit); time
required at the site; requirements for availability of personnel, the
facility and records; what topics and areas will be covered; the focus of
the audit in some detail.
The author's experience has shown that sharing the topics to be covered by
the audit checklists with the laboratory, a few weeks before the audit has
in general been very successful. This enables the laboratory to review
requirements (often finding areas of miscommunication or incomplete
understanding), allows time for the laboratory management to perform a
focused self assessment (management audit), allows the laboratory time to
gather the needed evidence and yes, even time to get everything caught up.
The philosophy used here is a win-win approach.
A mistake auditors can make is to equate the success of an audit with the
greater the number of findings, the better the audit. Large numbers of
findings often result from poor communication and produce a lose-lose
situation. If a company can correct mistakes before the audit, much time
is saved in getting the laboratory into compliance. The audit trail is
still there to show when the mistakes were corrected and the auditor
should require the laboratory to still assess what potential impacts were
to the data, and to correct or flag the affected data if needed. Much can
be learned about the attitudes, philosophy and general operations of
laboratories by this approach. When other purposes are paramount,
checklist topics can be withheld from the laboratory.
REQUIREMENTS
Great care needs to be taken to identify and evaluate based on the
established requirements. When auditing a laboratory's work done under
specific regulatory requirements, those requirements need to be complied
with unless a written waiver has been obtained from the applicable
regulatory body(ies). Auditors sometime impose certain formats or
requirements which are not based on the regulations or the contract with
the laboratory. These are not enforceable and can be a source of both
contention and loss of credibility.
Sometimes, laboratories sign contracts in which additional requirements
have been imposed by the client. These laboratories sometimes try to hide
168
-------
behind the fact that the area of noncompliance is not regulatory and
therefore is not important. Any such written agreement may be just as
important as the regulatory requirements to the client.
GIVE ME A CHECKLIST!
A critical mistake some companies make is to develop checklists to be used
on audits and then send just anyone trained in auditing to perform the
audit. First, of all, it is very difficult to produce a generic checklist
that can be used in various laboratories. Second, the available of
commercially generated checklists is decreasing as the industry reacts to
suits or threatened suits from companies. These companies have used a
generic checklist and have missed critical items that have resulted in
heavy fines or loss of money from poor investments.
The most significant single point to be emphasized in this paper follows.
The author fully believes that an auditor should be recognized as
technically competent in the area of the audit being dealt with. This
means that teams will often be used to perform the audit. The checklist
is only a guide and a reference point and extensive knowledge of the
subject can not be substituted for by making the checklist more extensive.
Knowledge gained through audits is vital to contracts and strong
consideration should be given to who will conduct it. Auditors
inexperienced in a given area tend to do one of two things. Either they
don't act in areas of concern because they aren't sure of the situation,
or they have multiple findings which are not valid. This misuses the
laboratory's time and can cause needless confusion and conflict. Everyone
makes mistakes but technically competent auditors tend to make fewer and
less critical ones.
COMPLIANCE VERSUS PERFORMANCE
Significant confusion exists in both laboratories and with auditors as to
what is the governing consideration between good performance and
compliance to requirements. This is not an easy question to answer and
the correct answer depends upon individual situations.
Some agencies have been more traditionally performance oriented whereas
others have been more compliance oriented. All have incorporated both
elements but have emphasized one to a greater degree. For example, EPA
has Performance Evaluation (PE) and Round Robin Sample Programs in which
it can be determined in a given matrix if the laboratory can produce
results that meet certain accuracy, precision and detection requirements.
Good scores on these samples are mandatory to receive work or to continue
to receive work. On the other hand other agencies often have checklists
based on compliance to very specific parts or steps of procedures and may
or may not have quantitative means of assessing the final outcome.
The author, given the regulatory bodies that will assess the laboratory
results in this case, takes the stance that both of these considerations
169
-------
must be fully satisfied. Until all regulatory bodies who overview this
work agree as to acceptance of a different approach, both considerations
must be fully satisfied. An example is presented in the following
discussion of the potential different approaches.
PROCEDURAL COMPLIANCE
A good example of the conflict between compliance versus performance comes
in analytical procedural compliance. Some companies and auditors contend
that regulatory procedures must always be exactly adhered to; while others
say that, if the performance can be shown to be equivalent, a
significantly modified procedure is acceptable. The author's experience
has been that most companies take it for granted that all laboratories
follow the approved procedures.
While performing audits, quite the opposite has been found in many cases.
While reviewing RCRA water pH procedures of some 22 laboratories, the
author found on his first visit that 21 out of 22 procedures were
noncompliant in following the national standard procedure. A procedure
found in SW-846 and one found in another approved Standard Method were the
two procedures allowed to be used in these contracts. Both required that
separate aliquots of the water sample be repeatedly measured until the pH
measurement agreement was within 0.1 or 0.05 pH unit respectively. Twenty
one of 22 laboratories were performing sample measurements on a single
aliquot. Some laboratories had other nonconformances to the procedures
also. To the author's knowledge all 22 laboratories's pH procedures are
now compliant in this regard.
The point that is emphasized is that pH is a very straight forward
measurement and if there is this much noncompliance to national procedures
with the pH analysis, what about all the other analytical procedures.
There are several significant noncompliances to regulatory procedures in
well know laboratories today. Laboratories have the responsibility to
report all potentially quality affecting modifications found in their
procedures that are reported to be "equivalent to" or reformatted national
standard procedures. Auditors can not afford to take quick looks at
procedures if their contract requires compliance. Companies need to know
if strict procedural compliance is mandatory on the samples submitted to
laboratories and, if so, hold laboratories to the responsibility of being
compliant.
The pH measurement example has been attempted to be refuted with the
argument that through Performance Evaluation (PE) samples and other
evidence that there is a sound scientific basis that the correct answer
was obtained despite the nonconformance. This argument stands it's ground
if performance alone is an acceptable end product. When the potential for
litigation exists concerning the conclusions drawn based on these sample
analyses, the question of procedural compliance becomes overpowering. The
bottom line is to know the requirements for the samples in question and to
have written waivers if necessary. In most cases such waivers are very
hard to obtain and involve significant cost and time.
170
-------
TRAINING
This Is an area in which most laboratories have trouble understanding what
is required and in presenting proper evidence. Training requirements will
vary based on company, regulations and situations encountered. Some
companies require less strict training documentation than those for which
these audits were performed. The requirement categories presented here
are:
Qualification requirements including educational and training
Job or position descriptions
Evidence of training to policies and procedures
On the job training or other performance indicators
Documentation as to the required frequency of training for
specific tasks
• Evidence of training and approval before sample analysis
Common deficiencies found include:
• Not having specific job descriptions
• Not having signed qualification statements listing specific
qualifications required and demonstrated
• Procedural or policy reading lists that do not require
training to revisions
• Training to revisions up to several months after put in use
• No indication if this item is a one time requirement or what
the requalification period is
• Copies of test scores or QA sample results with no indication
in writing what a passing score is and what happens if a
person does not pass, or
• No record of what actions have been performed by the analyst
before they are signed off as qualified.
A common mistake of auditors which causes laboratories time and financial
burdens is to insist where or in what format this evidence is kept. What
is needed is for the laboratory to be able to produce the documentation.
DOCUMENT CONTROL AND CHANGE CONTROL
The major aspects of document and change control that appear to be lacking
in some laboratories will be discussed here. Laboratory personnel need to
work to controlled copies of procedures. Controlled copies means that a
system of accountability has issued and assigned specific procedures that:
are identifiable as to review and signature responsibility
are traceable to issue and effective date
provide clear indication as to what types of samples, matrixes
and analytes are covered
reference standard procedures this procedure meets
clearly indicate what QA/QC accompanies the procedure and that
the copy is identifiable as a controlled copy.
171
-------
The laboratory should develop a system to assure that when new revisions
of procedures are issued, training documentation is in place before the
procedure is used on client samples; that all copies of old revisions are
removed from the work place; that the master copy of both old and new
procedures are securely stored and maintained.
RECORDS MANAGEMENT AND SECURITY
Records must be stored in a certain manner to meet requirements. These are
not universal requirements and potential laboratory shortcomings are
offered as things to consider. Laboratories were required, by the
contracts associated with these observations, to store records in the
equivalent of an one hour lockable fire safe or cabinet. These containers
needed to be either locked at all times when not in actual sight and use
by the records custodian (or delegated representative) or they needed to
be in a locked restricted access room to which only identified persons had
access. Storage in two separate locations that meet other specific
requirements could substitute for the fire safe requirement.
Even though several laboratories signed contracts with these requirements,
it took several months to have most come into full compliance.
Inexperience as well as incomplete checklists significantly contributed to
the delay in compliance.
CORRECTIVE ACTION AND NQNCONFORMANCES
Laboratories should have systems to identify nonconformances. They also
need systems to identify, track, respond to, and correct conditions or
items identified for corrective action. Nonconformances are conditions in
which material or action does not match the requirements of procedures,
contracts or policies. If such nonconformances are identified to be
potentially quality affecting or violate the terms of a contract, they
need to enter the corrective action system. The most common problems
encountered in the audits of these laboratories ranged from not having
systems in place to having systems that only partially addressed
corrective actions.
Conditions were often found not to be formally addressed in the first and
sometimes in follow up visits to laboratories, such as
loss of temperature control
missed holding times
missed reporting times
poor results on round robin or PE samples
loss of document or change control
loss of chain of custody or internal chain of custody control
loss of or mishandled samples
training noncompliances
procedural nonconformances
procurement noncompliances
using standards or chemicals past their expiration date
172
-------
using noncompliant standards (or lack of documentation on
standards)
incomplete labeling of (standards, reagents or samples)
problems with balances
not maintaining proper calibration of instruments
problems with water systems
data calculation or reporting problems
lack of proper control charting where applicable
not meeting laboratory data specifications or contractual data
quality objectives, and
• improper documentation of (standard, reagent or sample) make
up.
The other common situation was to find that follow up and closure of
identified conditions (especially internally identified conditions) were
either not being done or were not being documented. Few laboratories had
developed trending systems and a major weakness was that the effects of
deficient conditions on past work were not being evaluated adequately.
INTERNAL CHAIN OF CUSTODY
The contracts audited against, required documented internal chain of
custody. This meant that at any given time that the samples could be
accounted for and could be traced to where they were. Personnel had to
sign out the samples from storage areas, maintain custody requirements
while they used the samples and signed unused portions back into storage.
Several laboratories initially had difficulty in complying with this
requirement. Once systems were developed the level of compliance
dramatically increased.
PE AND ROUND ROBIN PROGRAMS
Most laboratories audited were participating in more than one PE and/or
round robin program. Most laboratories fared satisfactorily to well in
the programs. The major problem identified in this area was that many of
the laboratories were not taking advantage of the information obtained
from these studies. First, when passing scores were obtained, those areas
in which the results were outside the control limits were not followed up
on to identify whether a systems problem existed for that analyte.
Second, samples that could have had incorrect results reported for a
certain analyte as indicated by poor PE results were not checked to see if
they were affected.
TEMPERATURE CONTROL
Refrigerator and freezer storage requirements for samples and standards
were checked in the audits. Standard storage requirement differences for
CERCLA and RCRA programs were not accounted for in many laboratories. The
requirements for semi volatile organic compounds (semiVOAs) illustrates
several points. The two cited programs require some of the standards to
173
-------
be stored at different temperature ranges than for the other program.
Most of the same standards for semiVOAs can be used for both programs.
The dilemma for the laboratory is to maintain two separate sets of
standards in two different freezers or store the standards under the most
stringent requirements and meet the intent of both programs. Here is
where experience and knowledge are essential for the auditor. In the case
of semiVOAs the critical factor is to prevent volatilization and escape of
the analyte and no ill affects will be presented to the standards stored
at the slightly lower temperature range of the more stringent program.
The author's personal opinion and policy is to not present a finding if
the laboratory takes this stance. It can be argued that this is not
strict compliance and a slight possibility exists that the practice is
quality affecting but the author stands by his opinion until directed
otherwise.
The narrow storage range for inorganic samples is much more critical. If
one stores at lower temperatures the possibility of freezing quickly
increases. The freezing process can not be tolerated. Several
laboratories were found to have informal temperature documentation
programs during the first visit to them. Problem areas included the
following:
• measurements done in air, not liquid, which caused frequent
excursions from the acceptable temperature range,
• frequent excursions from acceptable ranges with no corrective
action,
• no documentation as to which samples were associated with
which refrigerators or freezers,
• incorrect postings on refrigerators and/or freezers as to
temperature requirements,
• no procedure or policy as to how to handle temperature
excursions,
• use of non-calibrated thermometers,
• use of calibrated thermometers passed their expiration date,
• not measuring and/or properly documenting measurements and not
documenting temperatures of incoming samples.
Formal documentation and compliance improved after the first visit and
with each subsequent visit. The most persistent problem remaining is the
formal documentation of problems and addressing types of potential impacts
to specific samples.
INTERNAL QA/QC PROGRAMS
The extent of the need for this program varies significantly with the
needs of individual companies. Specific requirements were given to these
laboratories in the regulations and with the contracts governing the work
discussed here. Several laboratories did not meet the intent and
requirements of this program until after the author's first audit. Some
174
-------
then came into partial compliance and it was after the next audit that
more progress was made.
Internal QA/QC is as much for the benefit of the laboratory as it is for
the customer's benefit. Here is where a good idea of how the laboratory
systems are functioning is obtained on a very frequent basis. Good
programs are very cost effective and help avoid a lot of repeat work.
Laboratories were encouraged to take full advantage of opportunities
provided by internal QA/QC programs but were audited to the requirements
set forth. Several root causes of types of deficiencies found were traced
to not having an adequate internal QA/QC program.
INTERNAL LABORATORY AUDITS (SURVEILLANCES ETC.)
Internal audits of two types will be described. EPA considers audits as
either system audits or performance audits. By system audit, it is meant
to assess if all portions of a needed system or activity were developed
and were auditable. Being auditable is defined here as being able to be
traced to meeting requirements and in being adequately documented. Often
the activities involve looking at documentation and handling much
paperwork. Visual inspections of processes can also be included.
The term Performance Audit is commonly used in two different ways. First,
this term is used to describe assessments of activities as they actually
happen. The performance of processes or individuals is measured against
procedural or contract requirements, in the real time mode. In this case,
a system audit could also contain a performance portion. The second use,
and the more common CERCLA or RCRA definition is to measure the
performance outcomes against the requirements. For example is the
required detection limit, accuracy, precision, holding times and
completeness being obtained.
Either type of audit can range in scope from informal to formal, from
minutes to several days or even weeks, from focused to a narrow scope to
being very broad. The number of auditors can vary significantly also.
The scope for external audits is even more variable. For example,
external performance audits involving more than one laboratory may focus
on shared Performance Evaluation (PE) and/or Round Robin samples, and on
the results of split samples performed by all involved laboratories.
Few laboratories were meeting the internal audit requirements before this
audit program was started. These requirements had been imposed by both
the regulations and the contracts but were not fully understood by the
laboratories and in most cases were being handled informally. The
planning, scheduling, completion percentage, quality, quantity and scope
of internal audit activity dramatically increased in most laboratories as
audits continued.
175
-------
LABORATORY MANAGEMENT ASSESSMENTS
These are activities that have met with varied requirements and success.
The actual requirements are somewhat vague and as such a wide range of
laboratory management assessment practices have been experienced.
The principal reasons for this process is to keep management in touch with
what is actually happening in the laboratory and to provide an independent
evaluation of the quality of work being done. Actually seeing and
evaluating laboratory operations on a real time basis helps management to
stop noncompliances and shortcut processes that may creep into
laboratories. It is surprising how managers were sometimes not aware of
current conditions in portions of their laboratories.
QUALITY CONTROL CHARTS
CERCLA type analyses do not specifically require control charts. Here the
limits are preset and specific instructions are given as to what to do
when the quality control limits are exceeded. Control charts can be used
in CERCLA work for the purpose of trending data. For RCRA and most other
work, control charts are necessary.
The three most common types of control charts are standard recovery charts
that indicate the laboratory bias for that analyte, spike recovery control
charts that demonstrate accuracy for a certain matrix for a specific
analyte and duplicate or replicate control charts that demonstrate the
precision for a certain analyte. Control charts need to be kept up at or
near real time. Their use, besides defining working accuracy and
precision limits, is for warning of marginal or unsatisfactory conditions
needing immediate attention and which might lead to formal corrective
action. They are also used for trending conditions. The most common
problems encountered in this area were:
not having control charts
not having all three types of charts
not keeping the charts on a real time or near real time basis
not acting or documenting actions when limits were exceeded
not tying noncompliances to specific samples, and
not recognizing that seven consecutive data points on the same
side of the mean constituted an out of control situation.
LABORATORY DATA SPECIFICATIONS VERSUS DATA QUALITY OBJECTIVES
Each laboratory needs to determine and provide to it's customers the
laboratory data specifications. This helps the customer evaluate if the
laboratory can meet the customer's needs. Accuracy, Precision, Detection
Limits, Instrument Detection Limits, Quantitation limits (these are called
by other names also), and other types of laboratory quality control
parameters are these specifications.
176
-------
Each customer needs to identify what it's data quality objectives (DQOs)
are that apply to the laboratory analysis. Examples of these include:
precision
accuracy
representati veness
completeness and
comparability (the PARCC parameters).
Specific method requirements, detection limits, instrument detection
limits, quantitation limits and other types of laboratory quality control
parameters are also specified.
The auditor needs to determine if the above specifications match. In the
observations discussed in this paper, it was not possible to properly
evaluate this information as it was not sufficiently available in most
cases. It appears that some clean up projects are struggling to first
develop adequate DQOs and then to properly pass those on to the
laboratories. It also appears that the laboratories have in general not
defined their data specification capability in enough detail to be fully
auditable.
RADIOCHEMICAL PROCEDURES
The regulations and requirements for radiochemical and mixed waste samples
are the most complex and incomplete of any involved. There are a few
methods or procedures that have national recognition. It is this
auditor's experience that compliance with these is less than with other
procedures.
Several types of analysis have individual procedures that have been
developed by various laboratories for their own use. These procedures
often employ different separation and/or analysis techniques and are not
easily comparable. In the contracts involved in these observations,
multiple radiochemical laboratories were used.
A PE sample program is very difficult to develop for many of these
analytes and the unavailability of this tool has severely handicapped
evaluating the comparability of the results from these laboratories.
Extensive work is needed in this area. A final observation regarding
radiochemical laboratories is that in general they appear less focused on
procedural compliance and on adapting to changes than do hazardous waste
laboratories.
SUMMARY AND CONCLUSIONS
Requirements that laboratories must comply with are extremely complex and
variable. These same requirements impose much difficulty to those who do
audits (various types of assessments) of the laboratories. Clear
direction as to how to interpret and impose these requirements is not
always obtainable. There remains significant controversy even within
177
-------
regulatory bodies as to how to apply some of these requirements. This
situation adds stronger argument to the position that auditors need to be
well informed both technically and with the quality assurance process.
Information obtained through performing audits of a variety of types of
DOE System and Commercial Laboratories has been shared. No attempt has
been made to identify these laboratories or to quantitate the problems
involved. Examples of some types of problems encountered have been
presented. It must also be recognized that this is a living system with
constant change and audit programs must adapt to the changes as must the
laboratories.
Repeat rounds of audits have been accomplished and compliance to
requirements has been evaluated. It is concluded, that when requirements
and failures to meet these requirements were documented in the audit
process, the amount of time to come into compliance decreased. Increased
frequency of audits from yearly to twice a year also sped up the
compliance process. The amount of improvement has not been quantitated.
If improvement in performance results can be tied to the audit process was
not evaluated.
178
-------
SAMPLING/FIELD
-------
PRACTICAL SUGGESTIONS TO IMPROVE THE QUALITY OF FIELD WORK
- SAMPLER'S PERSPECTIVE
T. Diebold, U.S. Environmental Protection Agency, Region IX,
San Francisco, CA 94105.
ABSTRACT
Before you attend another meeting on field sampling and
quality assurance, consider the sampler's perspective on the
work they perform.
Several conferences have been held recently to discuss the
quality of environmental monitoring data. The discussions
have expanded beyond an initial focus on laboratory analysis
to include debates on the quality of field sample collection
activities. Increasing emphasis is being put on assessing the
errors associated with sample collection, which may be carried
through the entire site evaluation process. In these
discussions, samplers can provide a valuable perspective on
improving the quality of field work. To date, their potential
to contribute in this area has not been fully realized.
In this presentation, some of the root causes of problems in
sample collection activities will be discussed from a
sampler's point of view. Examples from actual field sampling
projects will be used to illustrate some of these problems,
which can occur in every stage of a project from planning to
sample handling.
Common pitfalls to be avoided in the planning and scheduling
stages will be highlighted, and workable solutions to these
problems will be proposed. Other practical suggestions to
improve the quality of field work will be presented in the
form of "Simple Things You Can Do To Succeed in Sampling".
Sampler training and certification will be discussed, and the
benefits of cross-training of field and laboratory personnel
will be emphasized. Lessons learned in the laboratory that
can be applied to the field will be shared. The application
of Total Quality Management principles to field work will also
be discussed.
Notice: Although the information discussed in this paper has
been funded wholly or in part by the United States
Environmental Protection Agency, it has not been subjected to
Agency review and does not necessarily reflect the views of
the Agency and no official endorsement should be inferred.
179
-------
07 EVALUATION OF THE HYDROPUNCH**1 TO ASSESS GROUNDWATER
£ / r«r»ijrp»»f TWfcTTniJ T*V VOT.aTTT.T3 ORRANTCS
CONTAMINATION BY VOLATILE ORGANICS
Charles Van Sciver
Erica Wallace
New Jersey Department of Environmental
Protection and Energy
Environmental Measurements Section
25 Arctic Parkway
Trenton, New Jersey 08625
ABSTRACT
Hydrogeologic investigations have traditionally been
performed with the use of monitor wells. Their disadvantages
include the high cost to install, develop, sample and
dispose of generated cuttings, development and purge water.
Data is not available from a well until it has been
installed, which can lead to numerous misplaced wells,
providing useless data. This can lead to an investigation
requiring months to complete, as the process is repeated,
until wells are properly located to yield the required data.
The HydroPuncht111 II is a groundwater sampling tool developed
for rapid and cost effective collection of groundwater
samples from any depth in the saturated zone without the
installation of a monitor well. The tool was evaluated to
determine a cost effective decontamination procedure to
prevent cross contamination between sample locations and if
representative groundwater samples could be obtained when
operating in the "hydrocarbon" sampling mode.
Decontamination procedures were tested on a HydroPunchtm
unit exposed to pure gasoline. After decontamination blank
water was run through the tool and analyzed. The procedure
which provided non-detects for the target analytes was: a
soap and water wash (laboratory grade glassware detergent)
followed by a steam wash and distilled/deionized water
rinse.
Samples of groundwater were collected from the HydroPunchtm
II in the hydrocarbon sample mode and from stainless steel
well points. The units were placed so the screen of each
tool would sample the same cross sectional interval within a
5 foot radius of each other. Twenty two paired samples were
collected and analyzed by EPA method 602 plus xylenes.
Statistical results using a paired T-Test and the analysis
of plots generated from logs of the sample concentrations
indicate no significant differences between the tools and
good correlation of data. This evaluation demonstrates the
HydroPunchtin II in the hydrocarbon mode will provide a
representative sample of volatile organics in groundwater.
180
-------
28 AN INTEGRATED APPLICATION OF FIELD SCREENING TO
ENVIRONMENTAL SITE INVESTIGATIONS: A CASE STUDY
Tina Cline-Thomas, William Mills, Alex Tracy
Woodward-Clyde Federal Services, Rockville MD
Abstract
A base in the Washington D.C. area was slated to undergo facility expansion. This
expansion was to include construction of a commissary and parking lot, along with movement
of an existing sports field and a playground to new areas. Immediately prior to the start of the
construction, information became available which indicated that a landfill had existed in the
general area slated for construction activities. The exact location and extent of the landfill were
not known. Woodward-Clyde Federal Services (WCFS) was retained by the Baltimore District
of the U. S. Army Corps of Engineers (COE) to perform an investigation of the area to
determine the extent of the landfill and the health risk to construction workers and playground
users. Due to the construction schedule, the project work had to be completed in an 8-week
time period. Normally this type of project would require up to 6 months. In order to meet both
time and budgetary constraints, an intensive field sampling effort was undertaken in conjunction
with the use of field screening methods. Several field screening methodologies were employed
to more fully characterize the site during the time between field sampling and lab analysis: 22
metals by portable X-Ray Fluorescence Spectrometry (XRF) and Polychlorinated Biphenyls
(PCB) and Aromatic Hydrocarbons (AH) by immunoassay. The field screening was used to:
• direct field sampling efforts by delineating contaminated areas
• prioritize samples for laboratory analysis
• provide the laboratory with information on the expected range of contaminants
In total, approximately 130 samples were screened in the field and approximately 30 of those
samples had results verified by lab analysis. The metals results for XRF and lab analysis
generally corresponded to each other, provided samples were thoroughly homogenized.
Although some false positives were observed by field screening for PCB and AH, no false
negatives were observed. This presentation will discuss time and budgetary savings, QA/QC
procedures and comparability of the field screening and lab results.
Introduction
Woodward-Clyde Federal Services (WCFS) was charged with the task of clearing a site for
construction activities within 8 weeks while ensuring that sufficient samples had been taken to
characterize the site and that those sample results were accurate. In addition to the historical
information which indicated the presence of a landfill whose size and contents were unknown,
there was the possibility that the landfill area contained burn pits where PCB transformer
carcasses had been disposed. Some preliminary work indicated elevated levels of PCBs,
miscellaneous other organics and metals in the areas where the present playground was located
and where the athletic field was to be relocated (Figure 1).
181
-------
Figure 1: Site Map Showing Proposed Construction
As a result of health and safety concerns for construction workers and future residents a site
clearance was undertaken. Since the construction contract had already been awarded and the
government would face penalties if construction was delayed the time frame available for the
investigation was very short.
Lab analysis takes 3-4 weeks at normal turnaround times. Faster turnaround times are
possible at premium rates but even the fastest analysis requires 24-48 hours before results can
be reported. The approach that was developed for this project involved field screening and lab
analysis with five day turnaround. Lab analysis was required to verify field screening results
and to provide the quantitative information required for the health-based risk assessment
samples. Field screening was performed on soil samples for PCB and aromatic hydrocarbons
(the Petrorisc" immunoassay which was chosen is designed to provide total petroleum
hydrocarbon data but is very sensitive to bi and tri-cyclic polyaromatic hydrocarbons) using
immunoassay technology, and 22 metals using a field portable X-Ray fluorescence (XRF)
spectrometer. The field screening methods were used to prioritize samples for lab analysis,
provide information to the lab on the approximate concentration range expected to minimize
reanalysis and provide extent of contamination information for the areas being investigated.
182
-------
Methods
Immunoassay technologies are well established within the medical lab industry where they have
been used provide rapid, accurate test results for medical professionals. In recent years this
technology has also started to emerge in the environmental analysis field. The two parameters
analyzed by immunoassays for this project were PCB, using the Envirogard"" immunoassay by
Millipore and aromatic hydrocarbons using Ensys' Petrorisc11" immunoassay kits. Soil samples
were analyzed for 22 metals using the Spectrace 9000 field-portable XRF. The Spectrace 9000""
XRF uses a mercuric iodide (HgI2) detector along with a fundamental parameters algorithm to
qualitatively and quantitatively identify the metals.
A lab facility was set up on base for sample log-in and analysis. All samples were labeled,
logged into a sample tracking system on and screened at this location. A portable computer was
used for sample tracking as well as to store both the results and the spectra produced by the
XRF. All immunoassay data, including balance calibration, extraction weight and the
absorbencies of both the samples and standards were recorded in bound lab notebooks.
Immunoassays-General
As both immunoassay kits used methanol as their extraction solvent and immunoassay tests
are quite specific for their target compound(s), one extraction was performed on each sample
and the extract was refrigerated in a labeled screw-top vial. This ensured any re-analysis would
be performed on the same aliquot of each environmental sample. The analysis reagents were
added according to each manufacturer's instructions1'2'3'4 and all samples were run immediately
following calibration of the test kit. A spectrophotometer set to 450 nm was used to record the
absorbencies of both the standards and samples.
Aromatic Hydrocarbons (AH)
For the Petrorisc"11 immunoassays, sample absorbencies were determined relative to a low
standard (0.7 ppm m-Xylene, which is equivalent to 100 ppm gasoline) which served as the
threshold of detection. Two aliquots of the methanol extract were analyzed relative to this
standard: the first represented the sample without any dilution and the second was the same
extract at a ten-fold dilution. In this manner, approximate concentrations of petroleum
constituents can be determined with relative ease. While the petroleum kits were calibrated
using m-Xylene, they were sensitive to a variety of compounds found in petroleum products
including bi- and tri-cyclic aromatics.1'2 WCFS utilized the Petrorisc"" kits' sensitivity to
aromatic hydrocarbons to indicate burn areas where these aromatic hydrocarbons remained as
products of incomplete combustion. Because the Petrorisc"" kits are sensitive to a variety of
compounds, the immunoassay results correlated well with the hot spots as defined by lab
analysis. Data on the correlation between the two methods is shown in Table 1.
183
-------
Table 1: Comparison of Lab and Field Values for Petrorisc1
Sample Number
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Sample 9
Sample 10
Sample 1 1
Sample 12
Sample 13
Sample 14
Sample 15
Sample 16
Sample 17
Sample 18
Sample 19
Sample 20
Sample 21
Sample 22
Sample 23
Sample 24
Sample 25
Sample 26
Sample 27
Sample 28
Sample 29
Pelrorisc™ Value
ND
ND
ND
ND
ND
Detect
Detect
ND
Detect
Delect
Detect
ND
ND
ND
ND
ND
ND
Delect
ND
ND
ND
ND
ND
Detect
ND
ND
ND
ND
ND
Sum of PAH Values*
ND
ND
ND
ND
ND
ND**
> 1 ppm
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
> 1 ppm
ND
ND
ND
ND
ND
> 1 ppm
ND
ND
ND
ND
ND
*Sum PAH = Sum of all detects for compounds listed in SW-846 Method 8100.
**Dilulion at lab prevented proper quanlilalion.
184
-------
The protocol for performing analysis dictated that the difference between duplicate standards
(Delta Std) could not exceed 0.2 absorbance units (a.u.) or the calibration would be considered
invalid and the samples would be re-analyzed. Although the precision data for the petroleum
kits was acceptable, the use of a repeat pipettor rather than dropper bottles would have
improved the precision. (All figures showing precision data are scaled to equal size, so a visual
comparison may be made.) Ensys will supply the reagents either in dropper bottles or in bulk
(for use with a pipettor), but for this project the dropper bottles were used. A Shewart plot of
Delta Std for aromatic hydrocarbon analysis is shown in Figure 2.
Figure 2: Precision Data for Petrorisc"" Immunoassay
1 2 3-4 3 G 78 9 10
_a_DeKo Std. _^_Ux»r tornlngL1.1t
X upper control Llnrt Q Lower control Linn
3 24 23 26 27 ZB
Low«r MirnlnQ Limit
Fob/chlorinated Biphenyls (PCS)
For the PCB kits3'4 (Envirogard*"1), the concentration of the two calibration standards (2 and
10 ppm Aroclor 1248) were used and a linear dose-response was assumed between those two
points in order to calculate an approximate Aroclor 1248 concentration. As Aroclor 1260 was
the PCB found at the site, Aroclor 1248 concentrations were converted to Aroclor 1260
concentrations using relative response data for the two Aroclors provided by Millipore3'4. The
PCB kits were used to delineate the volume of the burn pits being investigated and the samples
taken for lab analysis in those areas was directed by the field screening results. Correlation data
is presented in Table 2.
185
-------
Table 2: Comparison of Lab and Field Values for PCB
Sample Number
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Sample 9
Sample 10
Sample 1 1
Sample 12
Sample 13
Sample 14
Sample 15
Sample 16
Sample 17
Sample 18
Sample 19
Sample 20
Sample 21
Sample 22
Sample 23
Sample 2-4
Sample 25
Sample 26
Sample 27
Sample 28
Sample 29
Sample 30
Sample 3 1
Sample 32
Immunoassay Value
ND
ND
ND
ND
Delect
ND
ND
ND
ND
Delect
ND
ND
ND
ND
Delect
Delect
ND
Delecl
ND
ND
Detect
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
PCBs by Melhod 8080
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND*
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
186
-------
Sample Number
Sample 33
Sample 34
Sample 35
Sample 36
Sample 37
Sample 38
Immunoassay Value
ND
ND
ND
ND
Detect
ND
PCBs by Method 8080
ND
ND
ND
ND
ND
ND
'Dilution at lab prevented proper quantitation
Reagents were added using an Eppendorf repeat pipettor, which allowed for rapid analysis
with good accuracy and precision. Typically a total of twenty tubes were analyzed per run:
assuming analysis performed in duplicate, three standards and seven samples could be analyzed
in one run. A Shewart plot (Delta Std) for both the low and high standards is provided in
Figures 3 and 4.
Figure 3: Precision Data for PCB Low Std.
D.2
-0.2
'
-
^ "~^A / \/
-----^^-^t^*^**^
m Delta 9td. + Ibper Vtarolno Ll»u jfc-- Low»r Warning Limit
X upper control Limit _^_ LOWW control Liim
187
-------
Figure 4: Precision Data for PCB High Std.
control LFmU
fcrning L)»lt
Control U>1T
Vbrnlng Unit:
Field-Portable XRF
All samples were first air dried and sieved through a 10-mesh sieve and sample descriptions
were recorded for all samples. If a significant amount of material would not pass through the
sieve, the material remaining in the sieve was retained and labeled as the exclusion products of
that environmental sample. The dried samples were then placed in 32mm sample cups and
covered with Mylar film.5'6 Each sample cup was labeled with the field sample number and
retained for re-analysis, if necessary. At the end of each day, both the results and spectra were
downloaded to a laptop computer for storage and data processing.
A mid-range standard reference material (SRM) and a quartz blank were run daily prior to
any samples, after every 10 samples and at the end of the analytical run (10% frequency). This
standard was an environmental sample which had been certified using traditional wet prep
techniques followed by GFAA and ICP analysis. Both the SRM and blank were used to confirm
instrument stability during the project. The standard was less homogeneous than was initially
assumed at the start of the project: approximately a week after the XRF screening had begun,
particles where discovered in the sample cup which would not have passed through a 10-mesh
sieve. As there was a week of data on the standard it was not re-prepped. Particle-size effects
from these large particles were believed to introduce some variability in the standard as
illustrated in the Shewart plots for the standard. The samples were believed to be more
homogeneous because all samples were dried and sieved prior to analysis. A plot of the Pb
results (Pb was one of the elements which had poor precision relative to most of the analytes)
for the standard over the course of the project is shown in Figure 5 and a table of the accuracy
and precision data is shown in Table 3.
188
-------
Figure 5: Precision Data for XRF Std (Pb)
Conoentralon
per Control Limit
- 4>p*r wrnlng Liruft
.Lower Control Llrift
1H 19 ZO 21 22 23 24
. Lower Warning Lfmit
Table 3: Accuracy and Precision Data for XRF Standard Reference Material
Average (mg/kg)
True Value1
Percent Recovery
Std. Dev.
Relative Std. Dev.
Cr
184170
160287
114.9%
2263
1.2%
Ni
16014
13105
122.2%
423
2.6%
Cu
2787
2946
94.6%
144
5.2%
Pb
115
141
81.6%
19
16.9%
Cd
387
292
132.6%
65
16.9%
'All true values obtained by acid digestion followed by GFAA/1CP analysis
Conclusion
Method-Specific: Immunoassay
The staff of two chemists performing analysis in the field lab was able to screen approximately
20 samples per day for metals, PCBs and aromatic hydrocarbons. The use of a repeat pipettor
is recommended both to speed immunoassay analysis and to achieve better precision and
accuracy. The correlation between lab and field data was good, but the difference in detection
limits and sample heterogeneity sometimes make it difficult to directly compare immunoassay
and lab data. However, the regions indicated as contaminated by field screening correlated very
well with the areas indicated as contaminated by lab analysis, historical data, and PID/OVA
results of samples taken in the field.
189
-------
Method-Specific: XRF
The use of an independent standard which was certified by traditional metals techniques
(GFAA/ICP) gave the data produced by the XRF an additional level of confidence. The
instrument showed good stability over the course of the project and 16.9% was the worst
relative standard deviation for any of the certified analytes (Cr, Ni, Cu, Pb, and Cd). Had the
standard reference material been completely homogeneous, the standard deviation would
probably have been considerably lower: Shewart plots showed only a few values which were
close to the control limit (Average+/-[3*Std. Dev.]). If these values were considered outliers
and removed, the standard deviation decreased markedly.XRF analysis of the standard reference
material correlated very well with its certified values: the average percent recovery (defined as
[XRF value/True Value]*100) was 109.2% with a high of 132.6% for Cd and a low of 81.6%
for Pb.
OA/OC Issues
Field screening can provide either Level I or Level II data9: for this project the field
screening data was regarded as Level I data and the laboratory analysis was used to make all
final decisions regarding site contamination. Specific guidelines for producing Level II data may
vary from site to site, and the sampling and analysis program must address the problems of
sample heterogeneity, matrix effects, interfering compounds, and sample contamination as a
result of improper handling or preparation7.
Although field screening can present additional challenges to the field team, there are many
instances where the additional data produced from the lower-cost field screening tests can
significantly reduce the sampling error in site investigations. Analytical error (bias and
variability introduced in the lab) typically accounts for only 15% of the total error introduced
in the site investigation process. The remaining 85% of the error in site investigations results
from insufficient samples or samples which do not accurately represent the contamination at the
site8. Field screening allows for rapid analysis following sample collection, which reduces
problems in sample handling, preservation and transport, and gives the field team the flexibility
to employ an iterative sampling strategy to fully characterize the contamination.
Effect on Sampling & Lab Analysis
Through the integrated use of field screening WCFS completed the site clearance on-time,
better delineated the extent of contamination and helped to direct the activities of the field crew.
In addition to the data provided by the lab, historical information such as aerial photos were used
to identify the area occupied by the former landfill (Figure 6). While XRF has been used
successfully in site investigations in the past, new advances in detector technology will provide
field teams with an instrument which is both portable and sensitive. At the time of this project
(April-May 1992), none of the immunoassay techniques had been recognized as methods by
EPA. Subsequently EPA has granted SW-846 third update numbers of 4010
(pentachlorophenol), 4020 (PCB), and 4030 (Total Petroleum Hydrocarbons) for immunoassay
screening techniques.
190
-------
In keeping with the DQO development process defined by EPA9, the project should be
planned with field screening in mind from the outset, and a chemist familiar with the technology
to be employed should be involved during the planning stage. The overall effectiveness of field
screening will depend on project-specific needs. It is recommended that the actual screening
analysis be carried out by, or under the supervision of, a qualified chemist to minimize
resampling and reanalysis and to ensure that results are not used inappropriately. Both
immunoassay and XRF are mature screening technologies, which, when used properly can be
very cost-effective tools in the site investigation process.
Cost savings were realized in two ways on this project by utilizing field screening; 1) a reduced
number of samples were analyzed by the laboratory since the field screening did not result in
very many false negatives, and 2) contamination information was transferred to the project
manager in a much faster manner than for a normal project where one has to wait 2 4 weeks
for a lab result. This meant that the project could proceed with decisions in a faster manner.
Figure 6: Site Contamination as Delineated by Field Screening and Lab Results
woodwer a- CI y a
191
-------
References
'PETRO RIScm User's Guide, Ensys Inc, 1992.
2"Soil Screening for Petroleum Hydrocarbons by Immunoassay," Draft Method 4030, USEPA
SW-846 Third Update, July 1992.
3Envirogard Tests Kits User's Guide, Millipore Corporation, 1992.
""Soil Screening for Polychlorinated Biphenyls by Immunoassay," Draft Method 4020, USEPA
SW-846 Third Update, July 1992.
5Spectrace 9000 User's Guide. TN Technologies
6Donald E. Leyden, Fundamentals of X-Ray Spectrometry as Applied to Energy Dispersive
Techniques: Tracer Xray, 1984.
7Kevin J. Nesbitt, "Application and QA/QC Guidance USEPA SW-846 Immunoassay-Based
Field Methods 4010, 4020 & 4030;" Ensys Inc, 1992.
8Francis Pittard, Principles of Environmental Sampling: A Short Course Presented Prior to the
8th Annual Waste Testing & Quality Assurance Symposium, July 11-12, 1992.
9USEPA, Data Quality Objectives for Remedial Response Activities. EPA/540/G-87/003, March
1987.
192
-------
DIVERGENCE OF FIELD AND LAB RESULTS IN
THE PONCA CITY INVESTIGATION
Subijoy Dutta. Environmental Engineer, U.S. EPA, OSW/PSPD/CAPB, 5303W;
401 M St. SW, Washington, D.C. 20460.
ABSTRACT
This case study focuses on a Benzene-Toluene-Xylene (BTX) pollution of soils and water
in a small hydrogeologically secluded pocket of Ponca City, Oklahoma. Despite
complaint by the residents, State and Federal officials could not find significant levels
of contamination at the site and recommended no action at this site. At the request of the
State treasurer, an independent hydrogeological investigation was conducted by the author
identifying high levels of BTX in the perched water table in an isolated pocket.
During an independent investigation, conducted by the author, very useful help was
rendered by the community in data collection. Samples from the storm water runoff at
the site produced flame for a very short duration when lighted. Other field tests, such as,
oil emulsion/layering of the samples proved that the contamination was of significant
level. Laboratory results, however, did not show very high levels of contamination. The
hydrogeological information from the site uncovered the mystery behind this inordinately
high levels of BTX in a pocket of the perched water table.
The feasibility of a commercial hydrocarbon recovery from the spillage has been looked
into as a cleanup option. However, CONOCO has rehabilitated some of the residents
from the site and started a cleanup action at this site in 1990 with full community
cooperation. The divergence of the field and laboratory results during initial
investigation at this site caused major problems and dissentment between the residents
and the State. The results from this study call for emphasizing the need to rely upon
field observations to locate and identify contamination and hazards associated with it.
INTRODUCTION
Ponca City is located in the north central part of Oklahoma. The CONOCO refinery
borders southwest part of Ponca City. The site location is shown on Figure 1. The area
of investigation for this assessment encompassed the boundaries of the CONOCO refinery
in Ponca City and the Circle Drive area of south Ponca City adjacent to the eastern
boundary of the refinery. The study area is bounded by South Avenue on the north, by
Waverly Street on the west, by a county road on the south, and by State Highway 77
(14th Street) on the east. The horizontal extent of the study area is shown on Figure 2.
Refinery area designations and the watershed boundary of the study area are also shown
in this Figure. The Circle Drive portion of the study area extends from the eastern
Refinery boundary formed by South Third Street, South Fourth Street, and Seneca Drive
193
-------
Figure 1 The Study Area in Ponca City, Oklahoma
-------
.
~^ • ' 'I >*^
i: • • • . -\Jr«
Figure 2 The Watershed Boundary and Sampling Locations in the Study Area
195
-------
eastward to the 14th Street.
The vertical extent of the study area extends to the base of the uppermost aquifer
underlying the study area. This alluvial aquifer ranges in depth from approximately 30
feet to 60 feet and is highly variable across the area thinning eastward towards the
Arkansas River.
Ponca city refinery was developed in early 1900's. After sixty years there were reports
of contamination of ground water. Hydrocarbon in ground water was observed in the
1950's and 1960's in the Circle Drive residential area of Ponca City adjacent to the east
boundary of the Refinery. In 1968 the Oklahoma Corporation Commission received
reports from two houses who had water with high hydrocarbon odor in their basements.
Significant hydrocarbon accumulations have not occurred since the late-1960's when
CONOCO, Cities Service, and Ponca City jointly installed and operated a hydrocarbon
recovery well network across the East and South Refinery areas and the Circle Drive
area of Ponca City. CONOCO has taken some measures to prevent and contain any
potential releases which might migrate off-site.
Ground waters containing dilute concentrations of volatile organic petroleum
hydrocarbons were detected seeping into Hoover Ditch from adjacent stream banks in
late-1986 and early-1987. This seepage was a direct result of heavy rainfalls in excess
of 20 inches which occurred in October and November 1986. The heavy rainfalls and
accompanying recharge caused ground water elevations to rise to surface discharge points
in the Circle Drive area of Ponca City. Due to serious public outcry, Mr. Ellis
Edwards, state treasurer at that time, asked the author to conduct an independent
hydrogeological investigation of the ponca city pollution problem.
The objectives of the independent hydrogeological investigation, conducted in August,
1988, were to:
(1) Define the sources and extent of hydrocarbon contamination in the area;
and
(2) Develop recommendations for remedial actions for controlling and
remediating the hydrocarbon contamination problem, taking all viable
alternatives into consideration.
The scope of the investigation included:
Hi Study of the geological setting of the study area.
HI Physical and chemical testing of water and water bearing
formations/soils.
196
-------
H Recommendation of possible steps to reduce severity of the problem.
H Recommendation of recovery of hydrocarbons from the contaminated
water.
Nature and Extent of the Problem:
The initial assessment of the ground water and hydrocarbon conditions underlying
the CONOCO's Ponca City Refinery and Circle Drive area indicates that there are two
general types of hydrocarbon accumulations in the subsurface [1]:
(1) Free product - petroleum hydrocarbon mounds; and
(2) Low concentration, soil-absorbed hydrocarbon residuals from historical leaks
and spills.
The initial study [1] determined that there were possibly four potential free
product hydrocarbon mounds which may exist under the CONOCO Refinery.
Hydrocarbons in these areas are refined products, such as gasoline, kerosine, or heating
oils, and can be chemically identified as such. Hydrocarbons observed in ground water
seepage in the Circle Drive area are dissolved residuals and cannot be chemically
characterized as a unique petroleum product. This hydrocarbon appears to be
significantly older, as is indicated by the low concentrations of benzene, toluene, and
xylene.
A major limitation of the study by Downs [1] was the la~k of adequate ground
water observation points, as well as concerns for the construction and completion history
of old hydrocarbon monitoring wells from the late-1950's and 1960's. The 1987 field
program for filling the information gaps identified in the initial study was implemented
in the summer and fall of 1987. During this program, fifty-six (56) new cone
penetrometer soil borings were constructed to develop detailed information on geological
maps for the site. In addition, twenty-nine (29) new ground water/hydrocarbon
monitoring wells were installed to develop new data on geology, ground water elevations,
and potential hydrocarbon accumulations in the study area.
The hydrogeologic study [2] was focused on the quality and quantity of water in
the surface and groundwater systems in the study area. The estimates were based upon
the available information. A description of the baseline hydrologic conditions and a
prediction of the effects on these baseline conditions in the study and adjacent area due
to the pollution problem were amalgamated to form an integral part of the study.
The conclusion from the study [2] are based upon the available database and field
measurements. Inadequate information on groundwater quality and quantity in the
197
-------
adjacent area is a limiting factor in arriving at unequivocal conclusions concerning the
groundwater of the adjacent area of the study.
The divergence of field observations and the laboratory results for the study area
caused major problem in the hazard assessment for the site.
SUMMARY
The principal purpose of the study undertaken by the author was to assess the
hydrogeological impact of an underground spillage problem on a tract of land located in
Section 34-26N-2E of Indian Meridian in Kay County, Oklahoma. This pollution
problem has surfaced near the Circle Drive area of Ponca City, Oklahoma.
The scope of the study also included investigation of the existing climatic and
hydrogeological conditions prevailing in and around the study area.
Climate:
A humid climate prevails in the study area and its vicinity. Annual precipitation
on the general area based on a 10-year average is found to be 32.34 inches. Seasonal
precipitation is found to vary from 17% to 31% of 10-year average annual rainfall.
Spring is the wettest season when 31% of the precipitation occurs from April through
June. During the Fall season from October through December, minimum precipitation
is observed to occur.
The total pan evaporation based on a 10-year average data is determined to be
56.2 inches. Some of the local meteorological parameters having effect on surface
runoff, air quality, particulate transport, and intensity of photochemical smog were
studied for a 10-year period.
Wind velocity varies considerably and can cause adverse effect due to increase
local concentration of dust and other particulates. Prevailing wind direction is from south
to north. The 10-year annual average wind velocity for the study area is 10.4 miles per
hour per month with a standard deviation of 1.12 and a variance of 1.25 during the 12
month period. There are periods of prolonged wind movements up to 15-20 miles per
hour.
The 10-year mean of annual temperature is found to be 60 degrees F. The
minimum mean monthly temperature occurred in January, and the maximum mean
monthly temperature was reached in the month of July. Drastic variations within a
certain month are not uncommon.
The total pan evaporation is highest during June, July, and August, ranging from
198
-------
8.23 inches to 9.35 inches per day. Mean annual evaporation is 56.3 inches.
Geology:
Geology of the Study Area [2]:
The pivotal area of study is section 34-T26N-R2E within the south and eastern
city limits of Ponca City, Oklahoma. The area is between the Conoco Oil Refinery Plant
on the west and the Arkansas River on the east. The general study area consist of nine
sections being sections 2,3 and 4 T25N, R2E and sections 26, 27, 28, 33, 34, and 35
T26N, R2E in Kay County, Oklahoma. The area is covered on the Ponca City
Quadrangle Oklahoma, 7.5 minute series, U.S. 6.5 Topographic map.
Topography:
The surface elevation varies from the highest point of 1020 feet in middle of
section 27 to around 900 feet in the stream bed of the eastward flowing Arkansas River
found in SW/4 of section 2, T25N, R2E. This is a north to south differences of 120
feet. Elevation differences from west to east of approximately 110 feet. In the area the
surface slopes south and east and all surface water runoff is drained into the Arkansas
River.
The study area sits atop of a portion of a northeast-southwest trending anticline.
This anticline is expressed in the surface topography where the surface rocks dip slightly
either eastward or to the west depending on the location of outcroppings. The anticlinal
axis bisects section 34 slightly east of its northeast corner to the east of the southwest
corner of section 34, T26N,R2E. The outcropping of strata from a regional aspect is
characterized by broad expanses of strata of benches or alternating parallel escapements
with abrupt slopes facing eastward. Older strata outcrops toward the east. The regional
dip is to the west less than 50 feet per mile.
Stratigraphy and Lithology:
The surface rocks are sedimentary. With the exception of the terrace sands, the
surface and the upper subsurface formations are of lower permian age. They embrace a
approximately 300 feet of sediments and consists of four limestone formations alternating
with same number of shale bud. Locally thin sandstones are present in the shale
formations. The Permian and subsurface Pennsylvanian beds have been tilted gently
westward, and subsequent erosion has beveled the section, exposing progressively older
beds to the east. Resting unconformably on the truncated edges of the Permian sediments
are local deposits of unconsolidated Quaternary alluvium and terrace materials that have
been deposited by the Arkansas River and its tributaries as it slowly worked its way
westward down the dip slopes of the limestone formations. The key beds of the study
199
-------
area are listed in a descending order.
QUATERNARY- Alluvium and terrace deposits.
PERMIAN- System
LEONARD SERIES: Summer Group, Wellington shale
WOLFCAMP SERIES: Chase group of formations
Alluvium and Terrace Deposits:
Along both sides of the Arkansas River a considerable thickness of unconsolidated
sands, slit, clay, and gravel occur overlaying the several Permian formations. Many of
the alluvial and terrace deposits contain sand and gravel layers that are highly porous and
permeable, and these deposits typically contain groundwater resources. Recharge areas
for groundwater resources in alluvium and terrace deposits themselves become almost all
ground water contained in these deposits as a result of downward percolation of water
from the land surface.
Remedial Investigation:
As noted earlier, ground water seepage and the minor concentrations of organic
and inorganic constituents that have been adsorbed and transported by rising ground
water levels have discharged into various areas of Circle Drive. This includes natural
ground water discharge into Hoover Ditch and surrounding ditches and natural ground
water discharge into basements and foundation areas which intersect the rising ground
water table. The manifestation of natural ground water seepage in the Circle Drive area
has caused citizens to become concerned about both the occurrence of ground water
seepage and what this ground water might be transporting. The hydrogeologic
investigation conducted by the author involved soil and water sampling from the study
area. Sampling locations are shown on Figure 2. Water samples were collected from
locations A and B and soil borings were conducted at location #1, #2, and #3.
Testing of Soil & Water:
Chemical analyses were carried out on soil and water samples from the study
area.
A columnar section of the alluvial formation from the study area is shown in
Figure 3.
200
-------
ZONE # DEPTH LITHOLOGIC DESCRIPTION
A. 6-8'
B. 9-45"
C. 46-80"
Topsoil
Grey, Sandy, Silty
Loam (Highly Porous)
Reddish Brown Clay
(Very tight)
Figure 3 Columnar Section of the Alluvial Formation from the Study Area
Experiments were run to calculate porosity and permeability.
A highly porous zone (#B) was present on top of a tight clay Zone (#C). Table 1 Shows
the experimental results.
201
-------
TABLE 1 EXPERIMENTAL RESULTS OF SOIL SAMPLES
Porosity Dispersion
Coefficient
ZONE # (%) Permeability (sq. cm/sec.)
(mD) Sample#/Remarks
B 45 2000 0.037 # (1-2)7(2-2)
C 52 0.022 0.00 # (3-2B)
The grey alluvium soil containing 19% of moisture showed a porosity of 45% and
a permeability of 2000 millidarcy. This zone was followed at the bottom by a tight clay
having a permeability of 0.022 millidarcy. The Zone C thus served as a confining layer
for the fluids to move through Zone B of in the study area.
The dispersion coefficient of the porous zone, as estimated roughly [3] , was
found to be 0.037 sq. cm/sec, and the dispersion coefficient of the underlying clay zone
was found to be 0.0 sq. cm./sec.
The result of the hydrogeologic investigation of the study area clearly identified
a zone of high permeability sand channel at about 3-10 ft. from the ground which carried
all the recharged water from the watershed which was contaminated with hydrocarbon.
The very tight clay underneath the sand channel served as a confining layer.
Field Observations:
During an initial field visit in July 1988 the basements of two houses in the study
area were visited. Very strong hydrocarbon odor was present in both the basements and
oily liquids were visible. An oily surface discharge, oozing out of the ground, near the
Mclnley School was also noticed. Hydrocarbon odor was also present at this location.
The worst situation was observed where a 24" pipe was discharging high volume of
water, with very strong odor of hydrocarbon, from underneath a road into the storm
drainage system. The source of this water is from a subsurface system from the south
side of the site. The discharge at this point could be visibly identified to carry
hydrocarbon vapors.
202
-------
A 1/2 gallon jar was filled up with this water and was taken to a dark room with
the lid tightly closed. An open flame was introduced at the top of the jar upon immediate
opening of the lid. Spontaneous blue flame was observed for a very short time at the
mouth of the jar. This field test confirmed the presence of volatile hydrocarbon in the
perched water table of the study area.
Lab Results:
Samples from the same discharge point (A) was analyzed in the lab which showed
a flash point of 220° F and 100 ppm Benzene, 200 ppm Xylene, and 300 ppm of heavy
hydrocarbons.
Samples from the basement of the house (B) showed 1.49 ppm toluene, and
< .0002 ppm both benzene, and xylene. Faint smell of hydrocarbon was noted by the
lab. These lab results were very similar to the earlier laboratory results as furnished in
the following paragraphs.
An earlier report on Ponca City water analysis which was done on January 20,
1988 showed that the basement water had total hydrocarbon levels which ranged from
121 to 4,725 part per billion. The concurrent benzene levels in basement water ranged
from 32 to 2,000 parts per billion. This water was not used for drinking or bathing.
The results showed that there was volatilization potential for the various hydrocarbons
contained in the basement water. The air samples taken at the time of water sampling
were below 0.5 ppm for benzene using the dragger tube calorimetric technique.
The analysis by the Oklahoma State Department of Health Environmental
Laboratory of basement water samples taken from 6 homes identified by the citizens as
the "worst possible cases" in the Circle Drive Area of Ponca City indicated total
hydrocarbon levels from 121 to 4,725 parts per billion. Of those total hydrocarbons, the
concurrent benzene levels in basement waters ranged from 32 to 2,000 parts per billion.
These samples in and of themselves only address the potential for human exposure
through direct contact with the basement water. This water was not used for drinking
or bathing. These results do show that there was volatilization potential for the various
hydrocarbons contained in the basement water. Entry surveillance air samples taken at
the time of water sampling were below 0.5 ppm for benzene using the Dragger tube
calorimetric technique.
The analysis by the Oklahoma State Department of Health Environmental
Laboratory of drinking water samples taken at 8 sites in the Circle Drive neighborhood
and adjacent areas indicated 1 sample with a level of 4 parts per billion of total
hydrocarbon. The remainder of the drinking water samples were below the total
hydrocarbon detection limit of less than 2 ppb. Additionally, none of the specific
203
-------
hydrocarbons including benzene, xylene, or toluene were found in drinking water
samples at the 1 ppb detection limit for specific hydrocarbons.
The drastic difference between the field observation and the laboratory results in
this study could not be accounted towards any noticeable sampling or other procedural
errors. High volatilization of the samples might have caused the discrepancy.
Conclusions and Recommendations:
The dispersion characteristic of the strata from the subject study area were
investigated and the results show that the estimated dispersion coefficient of the highly
permeable zone is very high compared to the dispersion coefficient of the underlying
tight clay layer.
In this study the calculations of porosity and permeability of different layers show
that a highly permeable zone lies between two clay layers in the affected area. This
channeling effect of the highly permeable sand is partially responsible and may be
considered as a natural calamity causing the seepage of the contaminated water into the
basements of local residents. The recovery of free hydrocarbons from this porous zone
could be done by using a pumping system. This will largely reduce the hydrocarbon
percentage of the contaminated groundwater.
The chemical test report indicates that the concentration of Benzene, Toluene, and
Xylene (BTX) is less then the standards set by Suggested No Adverse Response Levels
(SNARL). However, when this discharge seeps through the basements of local residents,
it creates uncomfortable and unhealthy living condition. The field observation from this
study strongly supports the need for a remedial action at this site. On the contrary, the
laboratory results do not call for such action. This divergence in the field and lab result
is highly controversial, and more so in a residential area. The author recommends that
an emphasis should be put on the field observations in such cases so that some remedial
measures could be initiated at these sites. Also, it is recommended that some refinement
in sampling and analytical procedures might be able to converge the differences between
the field observations and laboratory results to an acceptable degree.
204
-------
REFERENCES
1. Downs, C. E., "Ground Water Hydrocarbon Assessment Ponca City Refinery And
Vicinity Conoco Inc. ", Report on Hydrogeologic Study conducted in Feb. 1988.
2. Dutta, S., "Hydrogeological Investigation of the Pollution Problem in Section 34
-T26N-R2E. in Kay County. Oklahoma." , Phase I & Phase II Report, July 5,
and August 16, 1988.
3. Menzie, D. E., et. al., " Dispersivity as an Oil Reservoir Rock Characteristics ",
Annual Report for Period Oct. 1, 1986 to Sept. 30, 1987, submitted to U.S. DOE
for publication, Dec. 1987.
205
-------
FIELD SCREENING FOR HAZARDOUS MATERIALS
IN SOIL AND GROUNDWATER AT A MAINTENANCE BUILDING
Anita M. Larson. Senior Hydrogeologist, Harding Lawson Associates, 2400 ARCO Tower, 707 17th
Street, Denver, Colorado 80202; Patricia M. Curl, Senior Hydrogeologist, Harding Lawson Associates,
3247 Ramos Circle, Sacramento, California 95827; Harry R. Kleiser, Project Manager, Installation
Restoration Division, U. S. Army Environmental Center, ENAEC-IR-B, Building E4480, Aberdeen
Proving Ground, Maryland 21010.
ABSTRACT
A combination of field screening techniques were used to assess the nature and horizontal extent of
hazardous materials in soil and groundwater at a maintenance building area at the Sierra Army Depot,
California. Soil-gas surveys and groundwater sampling directly from a hollow-stem auger were employed
to study a maintenance building area for which little information regarding past waste management handling
and disposal practices was available. Although one maintenance building was identified as the potential
source area for the generation of hazardous materials, eight other maintenance buildings and eleven
warehouses in the area were discovered to be additional potential source areas of hazardous materials.
A series of field screening tools were used to design a cost-effective remedial investigation program and
a groundwater monitoring network at this site. Since the use of solvents was documented, an extensive
soil-gas survey program was designed to identify the potential horizontal extent of solvents in the soil and
groundwater. A grid of 300 soil-gas sampling locations over 40 acres was established. The soil-boring
and groundwater sampling program was focused in areas where soil-gas contaminant concentrations were
highest. Since the groundwater gradient at the site was unknown, a series of piezometers were installed
to serve as both groundwater level monitoring locations and as one-time groundwater sampling points.
Additional one-time groundwater sampling points were determined on the basis of the groundwater gradient
and on a preliminary estimate of the horizontal extent of any detected contamination, thereby allowing for
monitoring wells to be installed down-gradient of detected contamination.
INTRODUCTION
Harding Lawson Associates (HLA) was contracted by the U. S. Army Environmental Center (USAEC),
formerly the U. S. Army Toxic and Hazardous Materials Agency, to perform a Remedial Investigation and
Feasibility Study (RI/FS) for nine of the twenty-two sites identified in the Sierra Army Depot (SIAD)
Federal Facilities Site Remediation Agreement (FFSRA) between the State of California-Environmental
Protection Agency (Cal-EPA) and SIAD. As a result of the Preliminary Assessment performed by the U.S.
Army Hygiene Agency in 1987, the scope of work for the remedial investigation of these nine sites was
referenced in the FFSRA. On the basis of initial site reconnaissance, the Field Sampling Design Plan
(FSDP) and other associated work plans for the RI/FS were written with a two-stage field investigation to
meet the scope of work requirements of the FFSRA and the overall objectives of the RI/FS in a cost-
effective manner.
To facilitate this two-staged investigation, routine communication with Cal-EPA was required for FFSRA
deadlines to be met and cost-savings to be realized. Upon completion of the first stage of the field
investigation, sites were eliminated from further study while the scope of work for other sites was
expanded. Data from the first stage of the RI allowed for data gaps to be identified and the scope of the
206
-------
second stage of the RI to be determined. Through documented negotiations with Cal-EPA, the second
stage of the RI was performed prior to the issue of a RI report. This allowed for the writing of an RI
report that drew conclusions more meaningful than "that more data was needed." The staging of the field
investigation and continuous communication with the Cal-EPA allowed for a flexible field screening
program to be implemented resulting in significant cost savings for the field investigation at the Building
210 Area which is the topic of this paper.
SUMMARY
Stage one of the field program at the Building 210 Area included an extensive soil-gas survey and limited
soil-boring sampling. Soil samples were obtained from areas where there were high detections of
trichloroethene (TCE), freon-11, and trichloroethane (TCA) in the soil-gas. The results of soil samples
from these areas revealed either very low or no detections of these compounds indicating that the
groundwater was the likely source of the detections in the soil-gas although the depth to groundwater is in
excess of 85 feet below the ground surface.
The stage two investigation designed for the Building 210 Area was a combination of soil-gas surveying1
and Hydropunch®2 sampling in conjunction with the installation of piezometers and groundwater
monitoring wells. The objective of the stage two soil-gas survey was to reach an area west of the site
where the soil-gas results were non-detections. It was important for an area of non-detections in the soil-
gas data to be identified to verify that the source of TCE in the soil-gas was not from another site west of
the Building 210 Area since groundwater flow direction was unknown and groundwater contamination
approximately one-half mile from the Building 210 Area was documented.
The objective of the Hydropunch® sampling was to identify areas in the groundwater of both high
contaminant levels and no contamination for the installation of groundwater monitoring wells. In addition,
piezometers were installed prior to the Hydropunch® sampling to determine the local groundwater gradient.
Piezometer locations were selected for their usefulness in determining the groundwater gradient and for
verifying whether areas of suspected contamination were contaminated. Stage one soil-gas and soil boring
data and stage two groundwater flow data from the piezometers were employed to determine optimal
Hydropunch® locations.
Due to weather and geologic conditions, the field program was implemented differently than planned. Due
to record snowfalls, the soil-gas survey and drilling efforts became out-of-phase and a local laboratory was
contracted to do 24-hour turnaround organic screening via Environmental Protection Agency analytical
method 624 (EPA Method 624) for the Hydropunch® groundwater samples. Another modification to the
program was required when it was discovered at the first boring that the sands at the site were too dense
for the Hydropunch® tool to penetrate. At that time it was determined that groundwater samples would
be collected directly from the hollow-stem auger.
The hollow-stem auger was advanced five feet below the water table and a stainless steel bailer was used
to obtain groundwater samples. There was concern that the aquifer would be disturbed and that the
groundwater samples would not be representative. Table 1 presents a comparison of TCE in groundwater
samples collected in the hollow-stem auger to samples obtained in colocated wells. The analytical results
in Table 1 are reported in micrograms per liter (ug/1); analytical methods used were EPA Method 624 for
the screening (bailed) samples and USATHAMA Method UM 20 for the well samples. The detection limit
for Method UM 20 is 0.50 ug/1.
207
-------
Table 1 Comparison of TCE Detections in Groundwater
Sample Identification
B21-3-MW
B21-4-MW
B21-5-PZ
B21-6-MW
Screening Results (ug/1)
3400
<1
210
<1
Well Results (ug/1)
3496
0.75
180
<0.50
Figure 1 presents the locations where groundwater screening samples (noted by "HP" in the sample
identification) were taken and where piezometers and monitoring wells were installed. Figure 1 also
demonstrates how the screening data was used to outline potential contaminant plumes by contouring
screening data concentrations. Table 1 and Figure 1 demonstrate one of the successes of this screening
program, the close correlation between the groundwater screening data and the monitoring well data. One
of the problems encountered implementing this program was the discovery of an extremely flat local and
regional groundwater gradient and the project team's delayed ability to determine the local gradient due
to fluctuations in the water table that were not stabilized until the wells and piezometers were developed.
Once the groundwater gradient was determined, it was confirmed that contaminants had not only migrated
to the northeast, as early field data supported, but that contaminants had also migrated to the southeast as
the last field screening data set had indicated. Continued water level monitoring data showed that a
groundwater mound, which was not anticipated, was present at this site.
In conclusion, the field screening techniques employed at the Building 210 Area met or exceeded their
intended objectives although unanticipated groundwater conditions did not allow for all necessary field
investigation activities required at this site to be performed during stages one and two.
1. Soil-gas surveying is the collection and analysis of air from the vadose zone for the analysis of
target analyses.
2. Hydropunch® sampling is a technique where a stainless steel probe is either driven into the soil
or into an existing soil boring to collect groundwater.
208
-------
Explanation
Isoconcentration contours In microgn
Building, permanent
Building, semipermanent
Pavement, roads, and parking
Earth road
Trail
Railroad
EnvMWNTwnUI Serwai
Fence
— •• — • Reservation boundary
A Stage 2 sampling location
A Proposed sampling location
ug/l Micrograms per liter
Prepared for:
U.S. Army Environmental Center
Aberdeen Proving Ground, Maryland
Sierra Army Depot, Lassen County, California
Figure 1
Building 210 Area -
Trichloroethene Concentrations Detected in
Groundwater and Proposed Sampling
Locations
209
-------
31 FIELD AND LABORATORY METHODS IN ECOLOGICAL RISK
ASSESSMENTS FOR WETLAND AND TERRESTRIAL HABITATS
Greg Linder. Mike Bollman, Chris Gillett, Ricky King, Julius Nwosu, Suean Ott,
David Wilborn, ManTech Environmental Technology; Gray Henderson, University of
Missouri; Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis OR,
97333; Julie DalSoglio, US EPA/Region 8, Helena, MT, 59626, and Thomas
Pfleeger, US EPA, Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis
OR, 97333.
ABSTRACT
Field surveys are frequently included as part of an ecological assessment for hazardous
waste sites. In addition, biological evaluations in both field and laboratory have
increasingly been considered critical components in the ecological risk assessment process
for Superfund. Depending upon habitat type, field methods have been developed which
lend themselves directly to the Superfund ecological risk assessment process. Freshwater
wetlands, for example, are frequently impacted by various anthropogenic chemicals, and
field methods have been developed to help focus chemical and biological tests that have
been identified for laboratory studies. Here, we compare and contrast the field methods
used in evaluating wetlands that occurred in markedly different habitats and were
impacted by two different, but relatively simple contaminant sources. Both field studies
were completed as part of the baseline ecological risk assessment process. One site,
located in western Montana on the Clark Fork River, was Milltown Reservoir wetlands,
and had been impacted by heavy metals and sedimentation as a result of nearly ninety
years of upstream mining activities. The other site occurred in the shrub-steppe of
eastern Oregon, which had been a storage and disposal area for over 25,000 barrels of
chlorophenoxy herbicides and chlorophenols. In both case studies, assessment methods
had to be applied in lacustrine wetlands as well as the surrounding upland habitats. As
part of the ecological assessment for each site, a variety of field methods (e.g., terrestrial
and aquatic tests] were critical to the evaluation, yet despite differences in contaminant
sources, each field site considered ecological effects potentially associated with:
• Preliminary food-web contamination;
• Phytotoxicity and contaminant uptake by plants;
• Adverse biological effects expressed by soil macroinvertebrates;
• Adverse biological effects expressed by sediment macroinvertebrates; and
Adverse biological effects expressed by terrestrial and wetland vertebrates.
210
-------
In conjunction with chemical analyses of soil, sediment, and water samples, these
biological and ecological evaluations yielded an integrated evaluation of ecological effects
and exposure for the baseline ecological risk assessments. Overall, the field activities at
each site illustrate the role that preliminary field screening plays in the ecological
assessment process.
INTRODUCTION
Ecological risk assessment has recently gained a more prominent role in hazardous waste
site evaluations (US EPA 1991; Suter 1993), and this increased role has been reflected
in technical guidance to assure that reliable, yet cost-effective methods are used in the
process (Warren-Hicks, et al 1989; US EPA 1991; Under, etal. 1992). Field methods
for chemical analysis, particularly for screening purposes, have been developed and are
relatively routine in their application in waste site assessment (Fribush and Fisk 1991).
But, biological methods are not nearly as widely used, despite biological tests and survey
methods being readily available, well developed and standardized (Warren-Hicks, et al.
1989; ASTM 1992). As previously summarized, both chemically-based and toxicity-
based approaches may make significant contributions to ecological assessments for
hazardous waste sites (Parkhurst, et al. 1989). When used together, chemical and
biological test and survey methods can directly evaluate ecological endpoints pertinent to
risk assessment. From an ecotoxicological perspective, ecological effects and exposure
assessments are complex interrelated functions which yield estimates of hazards, and
potentially risks, associated with environmental contaminants in various matrices sampled
at a waste site. Here, using two wetland habitats as examples, we illustrate an integrated
approach to ecological risk assessment, and in particular we focus on field and laboratory
activities that occurred early in the ecological risk assessment process.
Montana Field Site
Milltown Reservoir is on the Clark Fork River in western Montana, six miles east of
Missoula, Montana. The reservoir was formed in 1907 following the construction of a
hydroelectric facility located on the Clark Fork River immediately downstream from its
confluence with the Blackfoot River. Since construction of the dam, a wetland habitat
has been created. Because of the upstream copper mining activities in the Clark Fork
River watershed, Milltown Reservoir has accumulated a large volume of sediment
enriched with heavy metals and metalloids, including arsenic, cadmium, copper, lead and
zinc. The Milltown Reservoir wetland was initially identified under CERCLA
(Comprehensive Environmental Response, Compensation, and Liability Act) in 1981 after
community well-water samples were found to have arsenic levels that ranged from 0.22
to 0.51 mg/L; the EPA recommendation for potable water supplies suggested that arsenic
not exceed 0.05 mg/L. Within an ecological context, however, the impact of the
contaminated wetland soils and sediments on the indigenous wildlife and vegetation
characteristic of the site was unclear.
211
-------
Oregon Field Site
Alkali Lake, Oregon is located in south central Oregon (elevation ca 1300m). When
seasonally recharged or during wet years, Alkali Lake is characteristic of many Great
Basin lakes, being relatively shallow and highly alkaline. During dry years, the lake
recedes and the exposed lake bed presents extensive evaporative deposits of sodium
carbonate. Between 1969 and 1971 over 25,000 drums containing herbicide formulations
(primarily, 2,4-dichlorophenoxyacetic acid [2,4-D], MCPA [4-methyl-2-
chlorophenoxyacetic acid], and related chlorophenoxy herbicides) were hauled from a
Portland, Oregon production facility to a storage facility located immediately west of
Alkali Lake. In November, 1976 the storage facility was converted to a disposal site, and
the drums were placed into shallow (1 m), unlined trenches, crushed, and buried on the
playa near Alkali Lake. Unfortunately, in addition to releasing the materials to the soil,
groundwater contamination resulted from these disposal activities, since a relatively
shallow aquifer (1 to 3 m below ground level) existed in the closed-basin around the
Alkali Lake playa. Prior to these preliminary studies, the ecological effects associated
with this contaminated soil and aquifer were poorly understood.
METHODS
As part of the ecological assessment at each site, a variety of terrestrial and aquatic
laboratory tests or field survey methods were critical to each wetland evaluation (Table
1). While this overview must limit detailed descriptions of laboratory and field methods,
those methods applied at each site reflected site-specific contingencies whenever possible.
For example, amphibian work at the Montana site used standard methods (ASTM 1992),
but also tested surrogate species (bull frog, Rana catesbeiana) that were more relevant to
ecological interpretation (Linder, et al. 1991). In contrast, if amphibian work were
identified as critical to the high desert site, an alternative surrogate (spadefoot toad,
Scaphiopus intermontanus) would be appropriate. Despite differences in contaminant
sources, each assessment considered ecological effects potentially associated with:
• phytotoxicity and contaminant uptake by plants;
• adverse biological effects expressed by soil macroinvertebrates;
• adverse biological effects expressed by sediment macroinvertebrates;
• adverse biological effects expressed by terrestrial and wetland vertebrates;
and
• preliminary food-chain contamination.
212
-------
Table 1. Summary of field survey and test methods used at study sites (from Linder, et
al. 1994).
Ecological assessment component
Phytotoxicity and contaminant
uptake by terrestrial and
wetland plants, and field
surveys for in situ effects and
vegetation communities
to
_L
CO
Candidate field survey or test
methods
Standard laboratory phytotoxicity
tests for seed germination, root
elongation, and plant vigor
Field tests for seed germination and
plant vigor
Field surveys for vegetation
communities
Field and laboratory measurements of
soil physicochemical properties
Methods potentially applicable to
Oregon field site, Alkali Lake
disposal area
Seed germination, root elongation,
and plant vigor tests with lettuce
Alternative site-specific tests
species, including alkaline
tolerant plants and native species
such as Sacrobatus vermiculatus
(greasewood)
Field testing limited in
preliminary studies, and seed
collections gathered for laboratory
testing
Field surveys for evaluating in
situ effects (e.g., plant health)
and vegetation mapping for baseline
evaluations
Soil texture, organic matter, pH,
cation exchange capacity, base
saturation, extractable acidity,
extractable nutrients, H-P,
electrical conductivity, water
retention, bulk density,
infiltration capacity, hydraulic
conductivity
Methods potentially applicable to
Montana field site, Milltown
Reservoir wetlands
Lettuce seed germination and root
elongation; plant vigor tests
with lettuce and Hydrilla
verticUlata
Plant vigor tests using
alternative species, including
native species (Potamogeton
pectinatus. or sago pondweed)
In-field seed germination tests
with standard laboratory test
species and plant vigor tests
using sago pondweed
Wetlands delineation and survey
of facultative and obligate
wetland plants
Soil texture, organic matter, pH,
cation exchange capacity, base
saturation, extractable acidity,
extractable nutrients, N-P,
electrical conductivity, water
retention, bulk density,
infiltration capacity, hydraulic
conductivity
-------
Ecological assessment component
Biological effects in soil
macroinvertebrates, and field
surveys of in situ effects and
soil communities
Candidate field survey or test
methods
Standard laboratory tests using
earthworms, nematodes, ants, or
isopods
Field tests using earthworms, ants,
or isopods
Field surveys for soil coranunities
Methods potentially applicable to
Oregon field site, Alkali Lake
disposal area
Laboratory tests with nematodes,
harvester ants or isopods; soil
connunity tests (raicrobial or
nematodes) using soil cores
Field tests using harvester ants or
isopods
Field collections of frequently
captured surface- and subsurface-
dwelling invertebrates and
vegetation-inhabiting invertebrates
Methods potentially applicable to
Montana field site, Mllltown
Reservoir wetlands
Laboratory tests using nematodes
or earthworms; soil community
tests (microbial or nematode)
using soil cores
In-field testing with earthworms
Field collections of frequently
captured surface-dwelling and
vegetation-inhabiting
invertebrates
IS}
Biological effects in sediment
raacroinvertebrates, and field
surveys of in situ effects and
benthic communities
Standard laboratory tests using
benthic raacroinvertebrates
Field tests using benthic
macroinvertbrates
Laboratory tests using alkaline
tolerant sediment- or water column-
dwelling invertebrate (e.g., brine
shrimp)
Ho field tests currently available
for alkaline sediments or surface
water
Laboratory tests using Hyallela
azteca and Chironomus riparius
In situ testing with aquatic
macroinvertebrates such as
Daphnia maona
Field surveys for benthic coranunities
residues evaluation
Field survey collections of
sediment-dwelling larval
invertebrates and brine shrimp
community and contaminant
Field collections of benthic
communities for sediment quality
triad analysis
-------
Ecological assessment component
Biological effects in
terrestrial and wetland
vertebrates, and field surveys
of in situ effects and wetland
communities
Preliminary food-chain
contamination evaluation
to
tn
Candidate field survey or test
methods
Standard laboratory tests using small
mammals, birds, or amphibians
Field tests using terrestrial
vertebrates
Field surveys for terrestrial
vertebrate communities
Designed feeding trials using
pertinent test species (e.g.,
terrestrial or wetland vertebrates)
Methods potentially applicable to
Oregon field site, Alkali Lake
disposal area
Candidate methods include
laboratory tests using small
mammals, birds, or amphibians
Candidate field methods using small
mammals, birds, and amphibians
Field surveys for small manraals and
other terrestrial vertebrates;
collection for analysis of gut
contents
Candidate studies address sediment-
dwelling benthic invertebrate and
shorebird food-chain contamination
evaluation
Methods potentially applicable to
Montana field site, Militant
Reservoir wetlands
Laboratory testing with
amphibians (or terrestrial
vertebrates)
In situ testing with amphibians
Terrestrial vertebrate surveys
and collections for contaminant
residues analysis
Candidate studies address
vegetation consumption by small
Baranalian herbivores
-------
Overall, the field activities for the ecological assessment at each site illustrate the role that
preliminary scoping activities play in the ecological assessment process.
Synopsis of Biological Methods at Montana Field Site
In the preliminary ecological survey for Milltown Reservoir wetlands, terrestrial
vegetation assessments were accomplished through a wetlands delineation (Federal
Interagency Committee 1989), and standard plant tests for soil contamination evaluations
(seed germination, root elongation, and plant vigor; see Warren-Hicks, et at. 1989;
Linder, et al. 1992). In emergent zone habitats, plant and animal tests were conducted
in laboratory and field (Warren-Hicks, et al. 1989; Linder, et al. 1992; ASTM 1992).
In situ tests were conducted using amphibians, and submerged aquatic vegetation was also
evaluated in emergent zone habitats (Linder, et al. 1992). In upland areas earthworm and
seed germination tests were completed on-site and in the laboratory. Root elongation
tests using groundwater were also completed. Additionally, surface soils were sampled
for metals analysis (US EPA 1986) and characterizations of soil quality (Page, et al.
1982). In order to determine food source contamination for site-specific receptors like
small mammals, samples of native vegetation and terrestrial invertebrates were collected
and analyzed for target analytes (cadmium, copper, zinc, lead, and arsenic). For aquatic
food-chains, samples of aquatic biota, including fish, aquatic insects, and plants, were
collected and analyzed for target analytes. Sediment samples were also collected from
emergent zone habitats and analyzed to characterize the extent of metal contamination in
depositional areas in the wetland.
Synopsis of Biological Methods at Oregon Field Site
As part of the preliminary ecological survey for Alkali Lake, terrestrial plants and
invertebrates were sampled in various seasons. Aquatic invertebrates were also sampled
at various times of the year, and representative species were identified and analyzed for
tissue residues of target analytes (2,4-D and chlorophenols). Vegetation measurements
taken during the survey determined cover and qualitative evaluations of plant health. In
conjunction with survey samples for vegetation, terrestrial invertebrates were sampled
using a combination of pit-fall traps and net surveys; for terrestrial vertebrates, small
mammals were trapped and gut contents were analyzed to evaluate food sources. In
addition to small mammal surveys, habitat use by birds was also noted; shore birds were
surveyed and seasonal patterns in habitat use were recorded. With these preliminary field
results target species were identified, and conceptual food-chain models were developed
for risk assessments for terrestrial vertebrates. Vegetation patterns were also evaluated
for their future role in long-term site monitoring.
216
-------
SUMMARY
Montana field site. Milltown Reservoir wetlands represents a single operable unit
within the Clark Fork River valley, but the wetlands must be viewed as a whole if they
are to be considered within ecological contexts. Nonetheless, these preliminary studies
have been incorporated into the baseline risk assessment being developed for the
wetlands, and suggested future work in the continuing ecological risk assessment process.
For example, soils and sediments at Milltown wetlands may be characterized as having
greater metal loads than the surrounding watersheds, although the historic and current
release and movement of metals may be quite dissimilar. At Milltown wetlands, the
measures of exposure-point concentrations suggest that metals, and specifically target
analytes ~ arsenic, cadmium, copper, lead, and zinc -- occur at greatest concentrations
in deposition zones located throughout Milltown wetlands. On the basis of total metal
concentrations, it may be said that soils and sediments are both metal-enriched, though
biological assessments suggest that only a small, and variable, fraction of that metal is
biologically available and associated with subtle biological effects. Clearly, these
deposition areas may warrant additional study, or at a minimum should be monitored to
ensure that adverse effects do not develop with time.
As potential sources of metal, the soils and sediments, as well as the associated water
column, must be considered in light of the potential biological and ecological receptors
inhabiting Milltown wetlands. While diverse in their selection, the biological tests
applied during the field and laboratory efforts at Milltown wetlands fall short of
anticipating all the receptors potentially at risk. For example, while terrestrial vertebrates
such as small mammals have been considered, the comparative-toxicity data base is
clearly incomplete. Evaluations of vegetation completed at Milltown wetlands may have
missed soil microbial effects. In summary, an understanding of the potential effects of
metals on communities and populations at Milltown wetlands must be developed by
inference, based upon these quantitative estimates of effects made during the ecological
assessment.
The biological assessments at Milltown wetlands suggest that little, or no acute toxicity
or adverse biological effects are occurring now. Consistently, and regardless of the field-
or laboratory-test methods used, biological assessments at Milltown wetlands could not
characterize acute toxicity; however, evidence of subtle biological effects was noted in
samples collected from depositional areas. Future remediation plans should weigh the
potential biological and ecological impacts associated with remediation efforts against of
the current impacts associated with elevated metals in soils and sediments. While the
current subtle effects and potential effects should not be understated in the risk-
assessment, currently any widespread physical alteration of wetland habitats may not be
justified on a technical basis. Future site monitoring should address problems that may
occur with remediation at Milltown wetlands and at upstream operable units, and in
particular should consider the long-term effects of metal exposures on vegetation in
217
-------
depositional areas. Also, indirect effects associated with future management plans (e.g.,
habitat alterations associated with "no action" or sediment dredging alternatives) should
be considered as a potential consequence of the metals that presently occur in the soils
and sediments at Milltown wetlands.
The data gathered from laboratory and field work at Milltown wetlands has reduced the
uncertainty associated with the baseline ecological risk assessment, and has suggested that
biological and ecological effects, if present in the wetland, were subtle. The uncertainty
associated with historic data or data collected in the current assessment at Milltown
wetlands must be stressed, e.g, were surrogate test species (in either laboratory or field)
sufficient for the biological assessment? Questions like these must be considered
proximate sources of uncertainty in the ecological risk-assessment process at Milltown
wetlands, and in any management decision derived in part from these studies. At
Milltown wetlands:
• The preliminary survey of linear food-chain contamination showed
evidence of metal bioaccumulation in small mammalian herbivores.
Bioconcentration of metals was also occurring in emergent vegetation in
some reaches of the deposition zones within Milltown wetlands, but
biomagnification of metals in herbivores did not appear to be a problem
for the endpoints considered. Field surveys at Milltown wetlands did not
contradict these assessments, but the sparse comparative-toxicity data
available, particularly for chronic endpoints, must be considered as a
source of uncertainty.
• Vegetation tests, especially those used in evaluating water collected from
the rhizosphere, suggested that no acute effects were associated with
groundwater and surface water, but subtle growth-related effects were
occasionally observed in samples collected from deposition zones at
Milltown wetlands. These subtle indications were noted in both laboratory
tests using emergent vegetation and in standard root-elongation tests.
Again, field surveys found no sign of altered vegetation patterns and
reduced cover was evident only in those areas that had previously been
physically manipulated.
• Earthworm evaluations in both field and laboratory were consistently
negative for acute toxicity, suggesting that soil macroinvertebrates may not
be at great risk as long as the current soil conditions exist. (This is
provided that earthworms are good sentinel species for assessing soil
health.) Soil microbial communities, however, were not adequately
described and should be evaluated when methods are available.
218
-------
• Preliminary studies using amphibians suggested that subtle biological
effects may be found at Milltown wetlands, but qualitative field surveys
did not support any conclusions that those effects would be prominent nor
adverse.
• Preliminary studies using bacterial tests yielded results consistent with the
balance of biological-tests used at Milltown wetlands, but may not be
representative of the soil community there.
• Plant-uptake studies suggest that garden crops bioconcentrate metals.
Bioconcentration of metals in plants differs depending upon the plant
species and anatomical feature considered (e.g., root versus leaf).
• Characterization and metal analysis of Milltown wetlands soils, sediments,
and biota indicated that metals have accumulated in various environmental
matrices, and that Milltown wetlands is spatially quite heterogeneous with
respect to metal deposition. Within-sample unit variation was relatively
less than across Milltown wetlands variation, and metal deposition was
consistent within topographic settings. Soils within sample units were
relatively homogeneous.
Oregon field site. While the present work was designed for data collection and survey,
its role within an ecological risk assessment for Alkali Lake has become essential. The
critical early phases (problem formulation, and pilot investigations for exposure and
ecological effects assessment) of the ecological risk assessment process have been
addressed, and existing data needs have been identified, which should allow more
adequate characterizations of uncertainty during the risk assessment process. Baseline
exposure and ecological effects assessments can now be developed and based upon site-
specific data, and any future data needs may be clearly identified. For example, within
the ecological effects assessment, field surveys and literature reviews have been
considered, but by design little toxicity assessment has been completed. Future work
could be developed to address this data gap, and additional comparative literature
regarding contaminant effects could then be gathered for evaluating site-specific toxicity
data. Whether future work regards these specific data needs, the baseline ecological
assessment at Alkali Lake provides empirical information upon which decisions can be
based.
Numerous technical recommendations can be identified as a result of these baseline data
collections. While the studies at Alkali Lake were preliminary and not developed as a
risk assessment, the information should be considered in developing management plans
for Alkali Lake, including any future ecological risk assessment. On the basis of
historical information and the completed field reconnaissance, the following may be
relevant to future management decisions regarding Alkali Lake:
219
-------
• With the baseline information gathered during these preliminary studies,
efforts should be continued to evaluate vegetation pattern changes at Alkali
Lake. The biological information subsequently generated in such a
monitoring program could potentially benefit future evaluations of the
effects of groundwater quality on vegetation. These should be completed
using, for example, plant vigor tests modified to use representative
species, e.g., Sacrobatus vermiculatus.
• Periodic soil sampling should be included in a monitoring program to
evaluate soil contamination and the changing spatial patterns of the
groundwater plume.
• Additional aquatic invertebrate sampling should be considered in
conjunction with an evaluation of shorebird diet and foraging activities at
Alkali Lake, particularly during the nesting season (e.g., spring and early
summer field activity). Modified standard tests (ASTM 1992) using
nestling shorebird exposures may be beneficial to evaluations of risks to
this terrestrial vertebrate.
Future work must reflect management and policy related issues relevant to site managers,
and field and laboratory efforts can directly address issues regarding risk characterization
and uncertainty. Ecological risk assessments then could be developed more soundly and
site management more adequately implemented.
ACKNOWLEDGEMENTS
Developing field and laboratory methods for ecological risk assessment has been
supported in part by U.S. Army Biomedical Research and Development Laboratory, Fort
Detrick, Maryland. Portions of these proceedings have also been reported at the Third
International Symposium on Field Screening Methods for Hazardous Waste Sites held in
Las Vegas, Nevada (February 1993), and will be considered in more detail in Linder, et
al. (1994).
REFERENCES
ASTM. 1992. Annual book of ASTM standards, Volume 11.04, American Society tor
Testing and Materials (ASTM), Philadelphia, PA.
Federal Interagency Committee for Wetland Delineation. 1989. Federal manual for
identifying and delineating jurisdictional wetlands. U.S. Army Corps of Engineers, U.S.
Environmental Protection Agency, U.S. Fish and Wildlife Service, and U.S.D.A. Soil
Conservation Service, Washington, D.C. Cooperative technical publication, 76pp.
220
-------
Fribush, H. and J. Fisk. 1991. Field analytical methods for Superfund. In Proceedings
of Second International Symposium on Field Screening Methods for Hazardous Waste and
Toxic Chemicals, Las Vegas, Nevada, pp. 25-30.
Linder, G., J. Wyant, R. Meganck, and B. Williams. 1991. Evaluating amphibian
responses in wetlands impacted by mining activities in the western United States, in Fifth
Biennial Symposium on Issues and Technology in the Management of Impacted Wildlife,
Thome Ecological Institute, Boulder, CO. pp. 17-25.
Linder, G., E. Ingham, C.J. Brandt, and G. Henderson. 1992. Evaluation of terrestrial
indicators for use in ecological assessments at hazardous waste sites, EPA/600/R-92/183,
U.S. Environmental Protection Agency, Environmental Research Laboratory, Corvallis,
OR.
Linder, G., Bollman, M., Gillett, C., King, R., Nwosu, J., Ott, S., Wilborn, D.,
Henderson, G., Pfleeger, T., Darrow, T., and Lightfoot, D. 1994. The role of
preliminary field and laboratory studies for ecological risk assessments in wetland and
terrestrial habitats: two case studies. la Environmental Toxicology and Risk Assessment -
Third Volume, ASTM STP 1218, Jane S. Hughes, Gregory R. Biddinger, and Eugene
Mones, Eds., American Society for Testing and Materials, Philadelphia. In Press.
Page, A.L., R.H. Miller, and D.R. Keeney, Eds. 1982. Methods of soil analysis. Part
2: Chemical and microbiological properties, Second Edition. American Society of
Agronomy Monograph 9. Madison, WI.
Parkhurst, B., G. Linder, K. McBee, G. Bitton, B. Dutka, and C. Hendricks. 1989.
Toxicity tests. In W. Warren-Hicks, B. Parkhurst, and S. Baker, Jr. (eds.). Ecological
assessment of hazardous waste sites. EPA/600/3-89/013, U.S. Environmental Protection
Agency, Environmental Research Laboratory, Corvallis, OR, pp. 6-1 - 6-66.
US EPA. 1986. Testing methods for evaluating solid waste. U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response, Washington, D.C.,
SW-846.
US EPA. 1991. Ecological assessment of Superfund sites: An overview; ECO Update,
Vol. 1, Number 2. U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency Response, Office of Emergency and Remedial Response, Hazardous Site
Evaluation Division (OS-230), Washington, D.C., Publication 9345.0-051.
Suter, II, G. 1993. Ecological risk assessment. Lewis Publishers, Inc. Chelsea, MI.
Warren-Hicks, W., B. Parkhurst, and S. Baker, Jr., Eds. 1989. Ecological assessment
of hazardous waste sites. EPA/600/3-89/013. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR.
221
-------
INCREASING THE SENSITIVITY OF FIELD HEADSPACE
ANALYSIS FOR VOLATILE ORGANIC COMPOUNDS
Charles Van Sciver
Robert Fowler
New Jersey Department of Enviromental
Protection and Energy
Environmental Measurements Section
25 Arctic Parkway
Trenton, New Jersey 08625
ABSTRACT
The use of portable gas chromatographs for on onsite
analysis of volatile organics using the headspace analytical
technique has rapidly increased. The increased use of these
methods is in response to the significant savings in both
rapid turnaround time of results and savings of analytical
costs as compared to conventional laboratory analysis. This
and other field screening analytical methods have proven to
be a valuble asset toward achievement of more rapid, cost
effective strategies for performing site investigations.
This study evaluated the ability of several methods to
increase the sensitivity of headspace analysis. The methods
included the addition of salt, heating and agitation. A
Photovac 10S50 field portable gas chromatograph equipped
with a photoionization detector (PID) was used for the
analysis of the samples. The target analytes were
trichloroethylene (TCE), perchloroethylene (PCE), benzene,
toluene, ethylbenzene, meta and ortho xylene. Three vials
were prepared, one without salt, one with 5g NaoSO4 and one
with lOg Na2S04. The vials were spiked to yield a
concentration of 50 ppb for each target analyte. Triplicate
analysis was performed on each spiked sample. The vials were
then placed in a hot water bath at 65°C and analyzed, then
shaken and analyzed. The percent increase in response was
determined for each method and compound. Response increases
of BTEX, TCE and PCE were greatest with the combination of
adding salt and heat to the sample. Individually, the
addition of heat to the sample produced the a greater
response compared to the addition of salt alone, although
both produced responses greater than 100% for all compounds,
except for benzene, compared to the reference solution. This
evaluation demonstrates several methods which can be used
seperately or in conjunction to increase the sensitivity of
headspace analysis under field conditions.
INTRODUCTION
Soil and ground water contamination by volatile organic
compounds (VOC) is frequently encountered at industrial and
222
-------
hazardous waste sites. The physical properties of VOCs
allow for their rapid movement in the vadose zone to the
groundwater. Once in the ground water, VOCs may rapidly
impact potable wells causing immediate health concerns to
receptors.
The concern for the rapid assessment of groundwater
contamination has in turn created a need for rapid methods
to assess the extent of a problem. The use of field
screening analytical methods for the analysis of volatile
organic compounds in groundwater offers a means for
real-time decision making during groundwater contaminant
studies.
Compound specific and quanitative field static headspace
analytical techniques l\ave been developed for gas
chromatographs (GC) I1/2/3). The static headspace method is
based upon the partitoning of volatile organics between the
aqueous and vapor phases when a sample is enclosed in sealed
container. An aliquot of the headspace is then withdrawn and
analyzed by the GC equipped with a photoionization detector
(PID).
THEORY
The principle of all quantitative headspace analysis is
based on the thermodynamic conditions of the phases. These
conditions require that equilibrium be established between
the liquid and gas phase. The concentration of a VOC in the
headspace of a liquid under static conditions is based upon
contaminant concentration in the liquid and its distribution
coefficient or Henery's Law.
C1 = KG11
Where:
C1 = Initial Concentration
K = Distribution Coefficient
cH = Final Concentration
The distribution coefficient is a measure of the
partitioning of a compound between the liquid and gas
phases. The sensitivity of the static headspace method will
vary between VOCs due to their extremely wide range of K
values. Therefore, to increase the sensitivity of the
static headspace technique, the value of the distribution
coefficient must be decreased. This will increase the mass
of VOCs driven into the headspace.
The techniques used to accomplish this under field
conditions include the addition of salt or "salting out"
which reduces the solubility of a compound and drives it
into the headspace. The use of heat and agitation applied to
a sample can also drive additional amounts of volatile
223
-------
compounds into the gas phase, thus becoming availible for
detection.
OBJECTIVE
The objective of this study was to evaluate the ability of
various methods to increase the sensitivity of headspace
analysis for selected compounds. The methods included the
addition of sodium sulfate (Na2SO4), heating the sample to
65°C, and physical agitation.
The target analytes were common aromatics, benzene,
toluene, ethylbenzene, meta-xylene, and ortho-xylene (BTEX).
Also targeted were two common chlorinated aliphatic
hydrocarbons, trichloroethylene (TCE) and perchloroethylene
(PCE).
METHODOLOGY
Two groups of 40 ml. screw cap septum vials were prepared.
Each group had one vial without salt (number 1), one with 5g
of Na2SO4 (number 2) and the third with lOg of Na2SO4
(number 3) . The solutions were prepared by filling each
vial with laboratory grade ASTM Type II water with no
headspace. The target analytes were added to each vial by
piercing the septum with a syringe and injecting the
compounds in sufficient quanity to yield a concentrationof
50 parts per billion (ppb) for each target analyte. The
vials were shaken intermittantly for ten minutes to ensure
the compounds were mixed thoroughly-
The headspace was created by withdrawing 10 ml of solution
from the sealed vial with a 10 ml syringe. To allow air
entry, a separate syringe needle with an attached charcaol
tube to prevent the entry of any air contaminants, was
placed through the septum and into the vial.
Each of the three vials in Group 1 were spiked with benzene,
toluene, ethylbenzene, meta-xylene and ortho-xylene
resulting in an approximate concentration of 50 ppb for each
compound. Each of the vials in Group II were spiked with
TCE and PCE resulting in a concentration of approximately 50
ppb. All vials were allowed to equilibrate at room
temperature (20°C) for one hour.
A Photovac 10S50 field gas chromatograph equipped with a
10.6 ev photoionization detector (PID) and a CPSil 5 wide
bore capillary column under isothermal conditions of 40°C
was used to perform all analyses. All samples were manually
injected with gas tight syringes. Syringes were purged
between samples with ultra high purity air (Ultra Zero). A
224
-------
one point calibration was performed using sample number 1
(without Na2SO4).
Each vial was analyzed in triplicate under ambient
temperature (20°C) static headspace conditions. Each vial
was then placed inverted in a hot water bath at 65°C for one
hour for temperature equilibration. Each vial was again
analyzed in triplicate, returning the vial to the hot water
bath after each injection. Care was used to minimize
agitation of the vial during handling. Sample Nos. 1 and 3
from each group were removed from the hot water bath, shaken
for 60 seconds and immediately analyzed. The results for
each sample were averaged and compared with Sample 1 of each
group.
RESULTS
The results are included in Tables I and II which give the
percent increases in response due to the addition of
Na2SO4, heat, agitation and a combination of all as compared
to sample number 1.
DISCUSSION
The results revealed the sensitivity for BTEX and selected
chlorinated aliphatic hydrocarbons increased with the
addition of heat, salt and agitation.
For TCE and PCE the addition of lOg of NaoSO4 resulted in a
larger response compared to the vial containing 5g.
Additional literature states mineralization over 20% would
only result in slight increases in sensitivity, and would
lead to excessive waste of the reagent.I1)
Heating the sample to 65°C resulted in sensitivity increases
which were comparable to adding lOg of Na2SO4. Heated
headspace for PCE gave a 14% greater response to the PID
compared to the addition lOg Na2SO4.
When heat was applied to sample 3A, the increase in response
was greater than the combined responses of heat and
mineralization. This was most pronounced for TCE with a
488% increase in response.
Benzene and toluene were not as responsive to the
mineralization as TCE and PCE. They were more responsive to
the addition of heat, which gave on an average, a 1.5-2 fold
increase in response compared to the lOg mineralization of
the sample. The addition of heat and mineralization for
the BTEX samples gave an increase in response greater than
the sum of the seperate increased responses for both heat
and mineralization. This result was similar to the response
225
-------
TABLE I
PERCENT INCREASE IN RESPONSE TO BTEX
Sample #
Benzene
Toluene
Ethyl -
Benzene
Meta-
Xylene
Ortho-
Xylene
2
Salt
5g
70.4%
103.9%
112.3%
127.1%
148.8%
3
Salt
lOg
99%
124%
110%
131%
198%
1
HWB
138%
223%
287%
312%
277%
2
Salt
HWB
242%
342%
419%
458%
634%
3
Salt
HWB
345%
391%
4,44%
483%
752%
1
HWB-SHK
149%
250%
347%
372%
413%
3
Salt
HWB-SHK
534%
570%
699%
758%
1255%
Salt - Sodium Sulfate.
HWB - Sample placed in hot water bath @ 65°C.
SHK - Sample shaken for one minute prior to analysis.
226
-------
TABLE II
PERCENT INCREASE IN RESPONSE OF TCE AND PCE
Sample #
TCE
PCE
2A
Salt
5g
125%
124%
3A
Salt
lOg
171%
137%
1A
HWB
161%
151%
2A
Salt
HWB
252%
228%
3A
Salt
HWB
488%
377%
1A
HWB-SHK
225%
273%
3A
Salt
HWB-SHK
947%
572%
Salt - Sodium Sulfate.
HWB - Sample placed in hot water bath @ 65°C.
SHK - Sample shaken for one minute prior to analysis.
227
-------
of the chlorinated hydrocarbons under the same conditions.
The greatest increases in response were with the later
eluting and heavier compounds. Ortho-Xylene had the
greatest increase in response (752%).
To further increase the response of the select compounds to
the PID, the vials were shaken vigorously for one minute and
an aliquot of the headspace immediately injected. This
physical action moves additional VOAs out of solution into
the headspace. Only one sample was injected, but responses
over ambient temperature headspace for the BTEX compounds
ranged from a 534% increase for benzene to a 1255% increase
for ortho-xylene. Similar increases in responses were noted
for TCE and PCE with 947% and 572% respectively.
Figures I and II show the comparison of chromatographs of
the BTEX, TCE and PCE samples which were at ambient
temperature, mineralized, heated, and agitated.
There was one problem encountered with the mineralization
technique. When adding the water sample to the vial with
Na2SO4, a bubble results w 1-3 mis. This is due to air in
the pore spaces of the Na2S04. When 10 mis of water are
removed to create the headspace varying volumes of headspace
are created which can affect the comparability of results
between samples. When the water sample is added to the vial
slowly and the vial is gently agitated to saturate the
Na2SO4, a loss of VOAs will result.
To eliminate the problem of varying headspace, all vials
should be marked at a volume of 30 mis or graduated vials
can be used. When the headspace is created, a volume of
water is removed to the 30 ml mark. Therefore all vials
will have an equal volume of headspace.
Caution should be taken when using heated headspace and
shaking the vials with analysis by GC. The GC column should
have a packing which is hydrophobic to eliminate moisture
buildup in the column. Excessive moisture can clog the
column and possibly damage the detector.
SUMMARY
This limited study has shown that increased analytical
sensitivity can be obtained when employing any number of the
techniques tested. The greater sensitivity to the PID can be
used to detect lower concentrations of compounds in water
with less sensitive instrument gains or less sensitive
detectors such as an FID. This also reduces the problems of
syringe contamination since at less sensitive settings, any
228
-------
FIGURE I
CHROMATOGRAMS FOR BTEX
PHOTOUAC
PHOTOUAC
PHOTOUAC
STPRT
STOP B 60B.0
SffllPLE LIBRflRT 1 PEC 3B 1332 13= 1
WflLTSIS » 4 uofl TEST
INTERWflL TEnP 28 2SB UL
GPIN 2 1-STP
CDnPOUNP MRTE PEflK R.T. PRERXPFn
STOP 8 600. a
SffllPLE LIBRPRT 1 JBN 7 1333
BMflLTSIS « IB UOfl TEST
INTERNfiL TEHP 23 2SB UL
GBIN 2 2-STP
conpourip
STOP a 575.0
SM1PLE LIBRflRT 1 Jffll 7 1333 13JM
flNPLTSIS a 13 UDfl TEST
INTERNAL TEnP 23 23B UL
GfilN 2 3-STP
NflnE PECK f«.T. RREP/PPn
PEBK R.T. PREP/PPM
UNKNOUN
BEN2EHE
TOLUENE
UNKND'JN
ETHTLBEN2ENE
n-p XTLENE
o XTLEHE
i
"2
•3
1
5
6
7
25.
88.
J86.
317.
33B.
123.
5BB.
8
3
1
3
8
6
1
82.3
16.
«.
311
1B.
33.
11.
55
16
.8
61
11
B8
HiUS
PPB
PPB
nUS
PPB
PPB
FFB
UNKNOUN
TOLUENE
ETHTLBENZENE
n-p XTLENE
0 XTLENE
J
2
3
1
5
6
26.7 21.7 nUS
83.7 71.12 PPB
186.3 83. 12 PPB
331.5 82.37 PfB
121.1 8B.B6 PPB
5B2.2 31.72 PPB
UNKNDUN
BENZENE
TOLUENE
ETHTLBENZENE
n-P XYLENE
0 XTLENE
J -26. B 10. 3
-------
FIGURE I CONTED1
CHROMATOGRAMS FOR BTEX
IPHOTOURC
PHOTOUAC
IPHOTOUflC
SIRRI.,
STOf * S7S.B
sunruE LIBRPRT i
17 UOO TEST
IMTERWL TEW 23
Oflf 2
STOP 8 S25.0
SW1PLE LIBRflpT 1 JW< 7 1333 H =23
PMOLTSIS » 13 UOP TEST
IHTERWL TEHP 23 23B UL
WlIM 2 2-STD NJB
STOP 8 575.0
SWIPLE LIBPWT 1
BMflLTSIS o 21
INTERNAL TEHP 28
SPIN 2
J«N 7 1333
UOft TEST
23B UL
3-STD HUB
BENUfE
UNM-iuUfl
TDLUtW
PEfx R.r.
26
83
1 11
IB*
262
318.5
conpouNO
PEPX P.T.
CDWDUMO NW.E PEW F.T.
n-P XYLENE
D XTLEKE
73.7 nUS
I IB.2 PfB
31. 2 «*>S
112.1 PPB
23. 3 P>US
2. 1 US
TOLU£M£
33J.6 1«5.7 PPB
173.6 jei.2 PfB
SB I . 3 136. 8 PPB
n-P XTl_EH
0 XTLENE
•26. I 15.6 fUS
83.7 155.0 PPB
111.3 37.3 *US
J85.7 131.5 PFE
76?.3 31.5 pUE
33B.6 •JIS.l PPB
123.6 217.2 PPB
5BI.3 311.8 FPB
TOLUENE
P-P
0
75.8 16.8 r<-'S
83.I 201.7 PCS
111.J 50,B rUS
JP6.1 113.7 PPH
262.3 26.5 <•<-'!.
33B.B 223.B PPB
Itt.f 217. 3 rfB
jfc 1 . 3 311. I PPB
SAMPLE 1
HWB
SAMPLE 2
5g Na2SO4/ HWB
SAMPLE 3
10g Na2SO4/ HWB
230
-------
FIGURE I CONTED'
CHROMATOGRAMS FOR BTEX
PHOTOUAC
ETPRI
STOP a 575.0
SAHPLE LIBRARY 1 JAN 7 1333 16=
ANALYSIS * 26 UOA TEST
INTERNAL TEflF 23 250 UL
GAIN 2 1-STD HUB-SHAKE
cnnPOUND NAPE PERK (5.T.
UNKNCUN
BENZENE
UMKNOUN
TOLUENE
UNKNOUN
EWYLBENiENE
n-P XYLENE
0 XYLENE
26.B 17. 0 fiOS
83.5 112.7 FFB
111.3 33.7 mVJS
186.1 152.1 FPB
•262.3 11.5 <"US
313.1 1.8 KS
331.S 183.1 FPS
121.1 183.2 FFE
5B1.3 217.6 PPB
PH01
rounc
STPRT
STDP 8 575.0
SW1PLE LIBRPRT 1 JflM 7 1333 16:26
BMPLTSIS » 28 UDfl TEST
INTERNAL TEPP 23 25B UU
GAIM 2 3-STD HUB-SHAKE
NWIE PEflK I?.T.
UNKNOUN
BEN1EWE
UNKfiDUM
TDLUENE
UNKMOUIM
n-P XYLEME
•26,1 10.2 mUS
83.3 287.2 FPB
J11.3 38.2 piOS
187,3 231.S FFB
262.8 61.8 jnUS
331.5 316.5 FFB
121.1 333.0 PFB
8 502,2 57^.6
SAMPLE 1
HWB-SHK
SAMPLE 3
109 Na2SO4, HWB-SHK
231
-------
FIGURE II
CHROMATOGRAMS FOR TCE AND PCE
': or * 3eo. o
innpLE LIBRKRT -> JRT» B 1333 >i--5i
flNflLTSIs » 37 UDP TEST
INTERWL TEHP 27 25B UL
GpJN 2 IP STP
COI^ L'ij;f.5 :07.7 ffil
"J 270. -3 135.5 PPS
PHOTOURC
6JDP » 30B.0
SfmPLE LIBRflRT 3 JPM 8 133U I3'.2l
flnftLTSIS » 3l UOP TEST
IMTEIWL TEtV 28 230 UL
GOIH 2 2P STD HUB
conpou^o rwiE FEW P.T. PREP^PPH
UNKr^DUrv J 27.6 117.2 rnUS
TCE 7 10B.3 182.7 FPB
FCE 3 722.i 171.3 FP8
PHOTOUAC
"C1 V"iT
STOP • 3B2.0
SWVLE LJBKPRT 3 JWC 8 1333 11 'II
PMftLTSIS « SB UOB TEST
INTERNOL TEW 28 23« UL
anil* 2 3p STD HUB
CDHPOU^D
PEW R.'.
TCE
PCE
28. e )33. 3
33.3 17.6
SAMPLE 1
HWB
SAMPLE 2
5g Na2SO4/ HWB
SAMPLE 3
10g Na2SO4/ HWB
232
-------
FIGURE II CONTED*
CHROMATOGRAMS FOR TOE AND PCE
PHOTOUflC
wi,..r-
STOP « 3B0.a
SPW1.E UBIWT 3
flNPLTSIS « S5
IHTEKWU- TEnP 28
Gf»IM 2
JBH 8 1333 13=53
UOB TEST
2SB UL
1« STP HUB SMKM
MPHJ PEF'K f. •
UTIKMOUM
TCE
PCE
1 7^.5 66.1 enUS
2 33.5 163.B PPB
3 228.I 133.3 PPB
PHOTOUAC
-s-r
STOP i 30B.O
SfinPLE LIBRflRT 3 JflU 8 1333 11=33
BHPLTSIS « S3 UDft TEST
INTEPMW. TEMP 28 25B UL
GAIN 2 3P STD HUB SHKN
CDnPDUND NPHE
UNKND1JN
TCE
P£PX 15. T.
36.6 112.1 FFB
156,8 It.5 nUS
213.3 3J53.1 PPB
SAMPLE 1
HWB-SHK
SAMPLE 3
10g Ma2SO4/ HWB-SHK
233
-------
small amount of contamination under normal conditions is not
detected. In situations where on site field analysis of
water samples is implemented in the absence of line power to
heat the samples, mineralization and shaking the vial can be
utilized to increase the analytical sensitivity
dramatically.
Further evaluation of these techniques is needed to assess
their application to other compounds, detectors, headspace
volume ratio, Na2SO4 concentrations and other types of
mineral salts.
ACKNOWLEDGEMENTS
The authors would like to thank Ken Glasser and William
Lowry for their assistance and technical review of this
paper.
234
-------
REFERENCES
1. loffe, B.V.; Vitenberg, AG.
Head-Space Analysis and Related Methods in Gas
Chromatography. Wiley - Interscience p62
2. Dietz, E.A.; Singley, K.F.
Determination of Chlorinated Hydrocarbons in Water
by Headspace Gas Chromatography. Analytical
Chemistry, Vol. 51, No. 11, pp. 1809-1814 Sept.
1979.
3. Kepner, R.E., Maarse, H., Strating, J.
Gas Chromatographic Head Space Techniques for the
Quanitative Determination of Volatile Components
in Multicomponent Aqueous Solutions. Analytical
Chemistry, Vol. 36, No. 1, pp. 77-82. Jan. 1964.
4. Buchmiller, R.C. Screening of Ground Water
Samples for Volatile Organic Compounds Using A
Portable Gas Chromatograph. Ground Water
Monitoring Review, Vol. IX No. 2 pp. 126-130.
235
-------
ORGANICS
-------
33
REGULATORY ASPECTS OF RCRA ANALYSES
by Barry Lesnik, USEPA, Office of Solid Waste, Methods
Section (OS-331), 401 M St., SW, Washington, DC 20460
What is RCRA?:
RCRA is the Resource Conservation and Recovery Act of 1976 which regulates
the management of hazardous waste under Subtitle C. Administration of the
program has been passed down to the states via authorization of State
Programs in most States.
How is a waste classified as hazardous waste under RCRA Subtitle C?
A waste is classified as hazardous under RCRA Subtitle C if it exhibits
any of the characteristics of ignitability. corrosivity, reactivity, or
toxicity, or if it contains specific constituents which are listed as
hazardous.
What is SW-846?
SW-846 is the compendium of analytical and test methods approved by EPA's
Office of Solid Waste (OSW) for use in determining regulatory compliance
under the Resource Conservation and Recovery Act (RCRA).
What are the drivers that determine which environmental analyses must be
performed?
Specific regulations under RCRA Subtitle C are the determining factors as
to which analyses need to be performed, and which specific target analytes
need to be identified in order to be in compliance. Examples of these
specific regulations include:
1) monitoring of leachates from hazardous waste landfills (Appendix
IX),
2) determination of whether a waste is hazardous by characteristic,
3) compliance with boiler, incinerator, furnace (BIF) rules,
4) compliance with land disposal restrictions (LDRs),
5) permit compliance for surface impoundments, storage facilities,
etc., and
6) corrective action.
236
-------
Are SW-846 methods required to be used for all analyses performed under RCRA
Subtitle C?
Except for the specific cases listed below, where use of SW-846 methods is
mandatory, SW-846 functions as a guidance document setting forth
acceptable, although not required, methods to be implemented by the user,
as appropriate, in responding to RCRA-related sampling and analysis
requirements. The RCRA applications listed in 40 CFR Parts 260 through
270 where the use of SW-846 methods is mandatory are the following:
(1) § 260.22(d)(l)(i) - Submission of data in support of petitions to
exclude a waste produced at a particular facility (i.e.. delisting
petitions) ;
(2) § 261.22(a)(l) and (2) Evaluation of waste against the corrosivity
characteristic;
(3) § 261.24(a) Leaching procedure for evaluation of waste against the
toxicity characteristic;
(4) §§ 264.190(a), 264.314(c), 265.190(a), and 265.314(d) Evaluation
of waste to determine if free liquid is a component of the waste;
(5) § 266.112(b)(1) Certain analyses in support of exclusion from the
definition of a hazardous waste of a residue which was derived from
burning hazardous waste in boilers and industrial furnaces;
(6) § 268.32(i) Evaluation of a waste to determine if it is a liquid
for purposes of certain land disposal prohibitions;
(7) §§ 268.40(a), 268.41(a), and 268.43(a) - Leaching procedure for
evaluation of waste to determine compliance with Land Disposal
treatment standards;
(8) §§ 270.19(c)(l)(iii) and (iv), and 270.62(b)(2)(i)(C) and (D) -
Analysis and approximate quantification of the hazardous
constituents identified in the waste prior to conducting a trial
burn in support of an application for a hazardous waste incineration
permit; and
(9) §§ 270.22(a)(2)(ii)(B) and 270.66(c)(2)(i) and (ii) - Analysis
conducted in support of a destruction and removal efficiency (ORE)
trial burn waiver for boilers and industrial furnaces burning low
risk wastes, and analysis and approximate quantitation conducted for
a trial burn in support of an application for a permit to burn
hazardous waste in a boiler and industrial furnace.
In Hazardous Waste Programs in RCRA-authorized States, the States can
require the use of SW-846 methods. A number of States have regulations
that require the use of SW-846 methods for hazardous waste analysis under
their RCRA Programs. Some of these States require the use of Second
Edition methods, while others require the use of Third Edition methods.
237
-------
State policies on the use of draft methods vary from blanket approval of
some methods to methods review on a case by case basis. In addition to
the mandatory applications specified above, EPA Regions can request the
use of specific Second Edition, Third Edition, or draft SW-846 methods for
other regulatory applications.
What is the current version of SW-846 that must be used for mandatory
applications?
The use of the SW-846 Second Edition, including Updates, is still
mandatory for these applications. Also, the Final Rule replacing the
Second Edition with the Third Edition of SW-846 (including Update 1) for
these mandatory applications is passing through the final stages of the
regulatory process. We expect to promulgate this Final Rule early in
1993.
What is an "EPA-approved" method?
This is a term that has been bantered about quite a bit recently,
sometimes rather loosely, particularly with respect to some of the new
analytical technologies. From the RCRA point of view, "EPA Approved"
means that a method has been incorporated by reference in a Final Rule
which has been published as a Federal Register Notice (FRN) either into
SW-846 or directly into the RCRA regulations. In short "EPA Approved"
methods are promulgated methods which can be used without special
permission for RCRA applications where the use of SW-846 methods is
mandatory. Therefore, until a method is promulgated by an FRN, it is not
an approved method for these mandatory applications, no matter where it
may be in the regulatory process.
EPA Regions may grant a general Regional approval to certain draft
analytical methods for use within a Region for specific applications. In
addition to this general Regional approval, Regions and States may grant
facility-specific approval for the use of draft methods for permit-
specific applications. Examples of facility-specific approval include the
use of Method 8330 by some Regions for the analysis of explosives residues
by High Performance Liquid Chromatography and Method 8290 for the analysis
of dioxins by High Resolution GC/MS. Both Method 8330 and Method 8290 are
currently draft methods which will be proposed in the Second Update.
How does a method become "EPA-Approved"?
This is a very brief overview of the EPA regulatory process, through which
SW-846 regulatory packages must pass. SW-846 methods are not published in
the Federal Register, but are incorporated by reference in the appropriate
RCRA regulations. The process roughly follows the following steps and
takes about 18 to 30 months to complete:
1) After methods are approved by the SW-846 Technical Workgroups,
238
-------
technical comments are addressed, and a regulatory package is
prepared. This regulatory package consists of a proposed rule, a
preamble to the proposed rule, and the necessary transmittal
documentation to pass it through the system.
2) After completion of the internal EPA review, the regulatory package
is sent to the Office of Management and Budget (OMB) for its review
and approval.
3) After OMB approval, the proposed rule is signed by the EPA
Administrator, is published in the Federal Register, and becomes
available for public comment. (The proposed methods are provided to
SW-846 subscribers by the Government Printing Office for comment.)
4) After the public comments are addressed, the regulatory package goes
through the same internal EPA and OMB reviews a second time, and the
rule is promulgated through a second Federal Register notice.
Can "draft" methods be used for analyses performed under RCRA Subtitle C?
Yes, since most analyses performed for compliance with RCRA Subtitle C
regulations fall outside of the scope of the mandatory applications of SW-
846 methods, for which promulgated methods must be used. "EPA Approval"
of methods is a factor only in the mandatory applications of SW-846
methods under RCRA, or in State Programs where use of SW-846 methods is
mandated. For all other non-mandatory applications under RCRA, draft
methods can, and, in many instances where they provide improved data,
should be used.
Do SW-846 methods have to be performed exactly as written?
Monitoring requirements under RCRA Subtitle C specify only that the
analyst must demonstrate that he can determine the analytes of concern in
the matrix of concern at the regulatory level of concern. Since SW-846
methods are written as guidance for a wide variety of matrices, it is up
to the individual analyst to optimize a particular method to his specific
needs. Allowable modifications include adjustment of sample size or
injection volumes, dilution or concentration of the sample, and
modification or replacement of equipment. These method changes must be
documented, as I stated previously the analyst must demonstrate that his
method can meet the previously-stated analytical requirements.
However, methods are inflexible in some aspects, e.g. generation and use
of an appropriate calibration curve, determination of recoveries and
precision, and demonstration of applicable detection limits in the matrix
of concern.
Philosophically, unlike other EPA Programs, RCRA specifies "what" needs to
be determined, and leaves the "how" up to the analyst.
239
-------
34
STEPWISE DEVELOPMENT PROCESS FOR REGULATORY ANALYTICAL METHODS
Paul Marsden. Methods Project Manager, Siu-Fai Tsang and Mark Roby,
Science Applications International Corporation (SAIC), 92121
ABSTRACT
It is unusual for an EPA Program Office to fund the development of new or
untested technology for regulation analysis. In fact, most methods
development projects are better described as the translation of a
procedure practiced in a limited number of laboratories to a written SW-
846 method that may be performed by many analysts. Although the same
analytical techniques employed in commercial analyses may be employed
during method development, there are significant differences in the
approach to method development compared to commercial analysis. For
example, there is no single approach to method development; scientists
must make professional judgements during the conduct of a study. The
quality assurance program is often a reflection of the analyst's
commitment to measurement quality and familiarity with the analytical
system rather than a set of defined acceptance criteria. This
presentation describes a ten step process for method development and
method validation conducted for the OSW: (1) Identification of
need/definition of performance objectives, (2) Delineation of QC
procedures, (3) Selection of analytical approach, (4) Determination of
instrument sensitivity, (5) Method optimization, (6) Generation of
performance data for a clean matrix, (7) Documentation of interferences,
(8) Demonstration of matrix suitability, (9) Calculation of quantitation
and detection limits, and (10) Workgroup review. These ten points will be
illustrated using examples of method development efforts conducted for the
OSW. Mechanisms that facilitate laboratories and manufacturers
participation in the method development effort will also be discussed.
INTRODUCTION
The Methods Section of the Office of Solid Waste (OSW) has responsibility
for developing and promulgating reliable analytical methods for the
determination of chemicals on Appendix 8 or for waste characteristics
(e.g., toxicity, flammability and corrosivity). These methods are
designed for use in a variety of laboratories and have embedded quality
control (QC) requirements that facilitate the intercomparison of data.
Each method must be evaluated for suitability for the use in environmental
or waste matrices. Central to the validation process is single laboratory
testing, which establishes the quality of data that can be obtained using
a method and its suitability for the soil matrix. Validation of potential
SW-846 methods is an important element in OSW's quality assurance program.
Following method validation, a SW-846 method and method validation report
is reviewed by an OSW workgroup.
240
-------
DEFINITION OF NEED
Each method development/evaluation project begins with a definition of
need. This need is identified by the EPA Program Office responsible for
the regulation driving the measurement. The statement of need generally
includes the target analyte list, potential sample matrices, target
detection levels, and QC requirements. At a minimum, method QC
requirements must be compatible with SW-846 Chapter 1 as well as any other
applicable general methods (e.g., Method 8000 Chromatography). The
definition of need is developed into the project data quality objectives
and the target performance criteria.
ANALYTICAL APPROACH
The analytical approach for a method development or evaluation project is
based on measurement sensitivity, stability/repeatability of the
measurement, and bias. In addition, potential SW-846 procedures are
evaluated on the basis of several qualitative criteria including the cost
requirements for analytical hardware/equipment, any additional facilities
requirements, ease of use/operator skill, cost per determination, and
analyses per day.
In many cases, the analytical approach has been used for a similar
application. Program Offices method studies are generally initiated after
a technology has been demonstrated as useful for the analysis of
environmental or waste matrices. Some of the sources for this information
are provided below, along with the period of time that might be required
to initiate a method evaluation study.
SOURCE
Instrument Vendors
EPA Regions and States
DOE National Laboratories
ORD
USATHAMA/USDA
Commercial Laboratories
Chemical Literature
Available
now
X
X
X
X
X
X
6-12 months
X
X
X
X
X
X
2-5 years
X
X
X
X
Program Offices often re-test methods obtained from other sources in order
to establish that the sensitivity, precision, ruggedness or matrix
suitability are appropriate for their regulatory needs.
241
-------
INITIAL EVALUATION
Initial evaluation of a method generally starts with establishing that the
technique is operating under statistical control. The detector linearity
and dynamic range are then established for determinative methods using a
calibration using at least five concentrations of standards. The analyst
will determine whether there are interferences in a clean control matrix
(i.e., no false positives) for sample preparation methods.
Critical measures of data performance will be established during single
laboratory testing. These include the accuracy, precision, and limits of
detection/quantitation that can be achieved in sample matrices of concern
to the RCRA program. In addition, single laboratory testing will
establish the probability of obtaining false positives and false negatives
with the technique. Initial laboratory testing is generally conducted
using soil and water spiked with known concentrations of target and
interfering analytes.
OPTIMIZATION AND RUGGEDNESS TESTING
In many cases, additional method optimization may be conducted to provide
the target sensitivity, accuracy, precision or matrix suitability required
by the Program Office. This optimization is accomplished by systematic
alteration of method conditions in order to achieve the performance
targets.
Ruggedness testing may also be conducted on candidate methods if they
perform erratically or when it is important to specify control limits for
critical method parameters. AOAC-type ruggedness testing assumes that
each method parameter varied in the test is independent. Limited
factorial designs (e.g. Planket-Herman) are more appropriate when
dependent method variables are tested. In either approach, previously
identified method parameters are varied between two conditions and
statistical evaluations of the results are used to interpret the
performance changes.
PERFORMANCE TESTING
Method performance is first established using a spiked clean soil.
Aliquots of one soil type prepared at three concentrations: one half the
target detection limit, two times the target detection limit and 10 times
the target detection limit. Method bias (accuracy) is determined at each
concentration by calculating the mean recovery of the spiked (or
characterized) analytes for the seven replicates. Method repeatability is
established by running the method over several weeks.
242
-------
x 100%
X = Mean value for the seven replicate determinations
X = Spiked or characterized concentration
Method precision is determined by calculating the percent relative
standard deviation of the spiked analyte recoveries for the seven
replicates at each concentration.
precision = —- x 100%
X
X = Mean value for the seven replicate determinations
o = Standard deviation for the seven replicates
After demonstrating suitable performance for the analysis of a clean
matrix, performance testing of more challenging matrices of interest to
the Program Office is conducted. Most current SW-846 methods provide data
on three different types of RCRA matrices. Those samples may be
characterized reference materials or spiked matrices containing known
amounts of target analytes. Analyses of these samples also provides
information on the effects of matrix interference. However, it would be
nearly impossible to generate data on all matrix/analyte interactions. As
a result, Chapter 1 of SW-846 provides guidance on evaluating the
suitability of methods for specific samples.
WORKGROUP REVIEW
Once methods are evaluated, they are submitted for approval by the Program
Office. In the case of OSW, workgroups meet to discuss organic,
inorganic, and miscellaneous methods as well as quality assurance issues.
Workgroup members are provided methods in SW-846 format, which includes:
•Scope and Application including analytes and matrices for which the
method is recommended
•a method summary (scope and principles)
•a list of required apparatus and reagents
•instructions for the preparation of standards
•sufficient detail and clarity in any sample preparation procedures
•detailed analytical procedure
•Calibration procedure
243
-------
•Method interferences
•calculation procedures
The workgroup is also provided performance data, including:
•the accuracy and precision of recoveries
•detection and quantitation limits
•Example output
•Analytical QC data and control charts
•results of confirmatory analysis
•Results of ruggedness testing, if performed
•Discussion of selectivity and sensitivity of the method
•Discussion of limitations of the method
The workgroup reviews the method narrative and the performance review.
They have the options of approving the method, rejecting the method or
requesting additional information.
CONCLUSION
The process of method development and evaluation does not seem to be a
growing activity in our industry. This is unfortunate given the
analytical challenges that we face in producing scientifically credible
data that is useful to decision makers. Chemists need to educate clients
about what is the most appropriate type of analysis. Regulators need to
show greater flexibility in allowing alternates to standard methods.
244
-------
35 RECYCLING PROGRAM FOR LABORATORY ORGANIC SOLVENTS
Methylene Chloride and Freon-113
Cornelius A. Valkenburg, Senior Scientist, and William T. Brown, Laboratory Manager,
Energy Laboratories Inc., 1107 South Broadway, Billings, Montana 59107
ABSTRACT
Due to hazardous waste disposal costs and associated environmental concerns, programs for
hazardous waste reduction and recycling of laboratory chemicals are of increasing importance.
A large fraction of the hazardous wastes generated by environmental testing laboratories
originate from the solvents used to extract organic hazardous constituents from aqueous, soil,
and waste samples. These solvents include methylene chloride (MeCI2), MeCI2/acetone
mixtures, diethyl-ether, hexane, and Freon-113. Glassware rinsing and chromatographic
extract clean-up techniques such as gel permeation chromatography also produce significant
quantities of waste solvents.
Minimization of solvent use is attained by selection/recommendation of the EPA method,
methods development and modification, and instrument and glassware selections. Modified
procedures must meet or exceed all Quality Assurance/Quality Control (QA/QC) parameters
specified by the method. Continuous liquid-liquid extraction techniques for EPA method 8270
are currently developed to use less than 200 mL of MeCI2 per sample. Similar solvent use
reductions have been applied to pesticide/PCB and herbicide extractions. Solid phase
extraction techniques such as those developed for the USEPA drinking water program use
much less solvent and are recommended to clients when appropriate to the sample matrix and
the desired analytical objective. A solvent recycling program can minimize solvent
consumption by processing waste solvents for re-use within the laboratory. An on-site solvent
recycling program using spinning band distillation to purify waste solvents is presented for
MeCI2 and Freon-113. Detailed work with MeCI2 and preliminary results for Freon-113 are
discussed.
Our solvent recycling program begins with the recovery, collection, and segregation of waste
solvents. Recovered solvents are segregated according to the type of laboratory operation.
Recovered MeCI2 is divided into GPC waste, extractor waste, general waste, and recovered
condensate from solvent concentrations. Each type of waste receives different handling and
processing prior to purification by spinning band distillation.
Purification of MeCI2 by spinning band distillation produces 3 or more fractions. Handling and
quality testing of each collected fraction are determined by the original segregated waste
source. Quality testing includes specific gravity, GC/FID, GC/MS, GC/ECD, and FT/IR.
Recycled solvent is returned to suitable laboratory operations based on testing results.
Recycled MeCI2 is demonstrated to be suitable for re-use in sample preparation for semivolatile
analysis by EPA Method 8270. Estimates of labor and material costs demonstrate the cost
effectiveness of the recycling program.
INTRODUCTION
In EPA pollution prevention strategies, source reduction is the preferred method for reducing
waste. However, for those wastes that cannot be reduced at the source, recycling is
considered the next best alternative.
245
-------
A large portion of the hazardous wastes generated by analytical laboratories orgiginates from
use of organic solvents in laboratory operations. Methylene chloride, Freon-113, methyl
alcohol, hexane, acetone, and xylene, constitute the bulk of the solvents used by
environmental testing laboratories.
Laboratory procedures which generate organic solvent waste include: glassware cleaning and
rinsing, sample extraction and preparation, and liquid chromatographic techniques used for
extract clean-up and/or analysis. Collected condensate from evaporative extract concentration
techniques are a source of additional waste solvent. Many laboratories vent solvent vapors
from extract concentration procedures directly to the atmosphere. Recycling solvents
encourages the condensation and recovery of these solvent emissions.
Land disposal of solvents was banned by the EPA in 1986. Laboratories typically handle their
solvent wastes by off-site incineration. Halogenated solvent wastes are normally segregated
from non-halogenated as they require more costly incineration at facilities specifically designed
to handle these materials.
Laboratory solvent wastes are well suited for recycling programs using distillation technology.
Solvent recycling can be performed either on-site or off-site. On-site recycling refers to the
laboratory performing all aspects of the recycling program. Off-site recycling is recycling
performed by a commercial solvent recycling facility. After off-site recycling, processed
material is either returned to the original generator using a custom or toll processing, or is sold
on the open market. In either on-site or off-site recycling programs, appropriate solvent
collection and segregation, is required in order to produce optimum recycled solvent quality.
Advantages to recycling on-site include: generally favorable economics; more control over
waste management (feedback loops for appropriate solvent segregation); reduced
transportation risks; and reduced reporting requirements (currently RCRA treatment permits are
not required to recycle on-site). Analytical testing laboratories are well suited to an on-site
recycling program. They can perform the quality testing necessary for their recycled solvents
and can utilize the recycled material appropriately, based on the quality testing performance
data. Analytical testing laboratories also should have the technical expertise to handle these
hazardous wastes. Disadvantages to on-site recycling include the initial capital costs, safety
hazards associated with handling solvent wastes, permitting requirements imposed by some
states, and the additional labor, technical, and managerial concerns.
Advantages to off-site recycling are the reduced technical and managerial demands, and in
limited cases-improved economics for small solvent volumes. Currently, economical off-site
recycling of laboratory solvent waste is limited to certain geographic areas primarily due to
hazardous waste transportation costs. Off-site recycling commonly recycles non-halogenated
solvents (flammables) as a fuel source.
Energy Laboratories Inc (ELI)., of Billings, Montana is committed to reducing the generation of
hazardous wastes and the consumption of solvents. This paper describes the ELI recycling
program which is fully operational for methylene chloride, the solvent used in largest volumes.
Preliminary results with Freon-113 recycling efforts are also discussed.
246
-------
EXPERIMENTAL
Materials
Table 1 provides a list of materials and their uses in the recycling laboratory.
Methods
MeCI2 Waste Processing
The handling of methylene chloride waste streams is illustrated in Figure 1. The MeCI2 waste
stream processing flow chart lists the procedures associated with recycling four categories of
MeCI2 waste. Throughout the solvent collection and recycling process, segregation of waste
and distillate fractions is maintained.
Methylene chloride wastes are segregated according to the types of contaminants expected.
Wastes with percent levels of other solvent contaminations are handled separately. Generally,
higher boiling trace contaminants present few problems for purification by distillation.
However, waste with high levels of higher boiling contaminants, such as phthalates, may
require pre-distillation, or double-distillation in the high-purity distilling system. Solvent
contaminated with high levels of low boiling compounds (other solvents), particularly those
with similar boiling points, or which form azeotropes with MeCI2, require specialized handling.
Methyl alcohol and water are known to form azeotropes with MeCI2.
Phase separation is performed by decanting or by transfer using separatory funnels. Bulk
filtration (if required) uses sediment settling, decanting, and vacuum filtration through 10 inch
diameter Whatman #4 filter paper. Pre-distillation is used to remove non-volatile soluble
residues. Pre-distillation is performed with custom manufactured glassware which allows the
teflon spinning band distilling column of the ABC Integrity 2000 system to be by-passed.
Automation of pre-distillation is achieved by using the electronic control systems of the ABC
Integrity 2000. Solvent drying is done by mixing calcium chloride pellets (5-10 g/L) with the
pre-distilled waste solvent for 24 hours. In-line vacuum solvent filtration (see Table 1,
Materials list) is used to separate the drying agent from the solvent. In-line filtration is favored
since solvent vapor emmisions are controlled and recoverable.
Distillation is performed using the ABC Integrity 2000 spinning band distillation system.
System operating parameters are per the manufacture's recommendations with modifications
made relative to waste characteristics. The system is operated with manual switching
between the collection of the forecut and the main fraction. Handling of forecut and still-
bottom fractions is discussed later.
Collect fractions are preserved with cyclohexene to a level of 0.01% (v/v). The preservative
prevents the accumulation of hydrochloric acid (HCI) in the methylene chloride. Cyclohexene
or amylene are commonly used for this purpose. These compounds do not prevent the
formation of acid in methylene chloride, but undergo an addition reaction with HCI across the
double bond. Proper solvent storage minimizes decomposition.
GC/FID, GC/MS, and other screening procedures, are used to determine the collected fraction
quality. Based on screening results, the material is segregated for re-distillation or appropriate
use. Segregated material is composited with material of similar quality to generate larger lots
of material. Composited material is further tested and distributed for appropriate use.
247
-------
Freon-113 Waste Processing
A flowchart for recycling Freon-113 Waste Streams is shown in Figure 2. The procedures for
handling Freon-113 are comparable to those for methylene chloride. Recycled Freon-113
quality is characterized according to intended use. Recycled IR grade Freon-113 (IRF113) has
no significant absorbance at the C-H stretching wavelength 2950 (cm"1) used to quantitate
Total Petroleum Hydrocarbons by IR as per EPA Method 418.1. EPA method 413.2 for oil and
grease by IR also has similar requirements. Trace levels of contaminants may be acceptable
in IRF113 since quantitation by IR is performed by measuring the differences between the
absorbance of the extracted sample against the reference solvent. Levels of acceptable purity
have not yet been determined. Non-IR grade recycled Freon-113 (NIRF113) is contaminated
with part per hundred or less (determined by GC/FID) of volatile solvents and exhibits a
significant interfering IR absorbance.
Complete segregation of recyclable IRF113 from recyclable NIRF113 is critical. Certain
compounds/solvents cannot be removed effectively from Freon-113 by spinning-band
distillation. These compounds (yet to be identified) either azeotrope with, or have similar
boiling points to Freon-113. These co-distilling compounds are considered contaminates when
their concentration and/or interfering absorbance is significant. Co-distilling/interfering
compounds may persist regardless of the number of re-distillations, and, if mixed with IRF113
would irreversibly contaminate the IRF113. Unleaded gasoline is known to contain non-
removable compounds of this type.
NIRF113 is suitable for use in determining oil and grease by gravimetry since the contaminants
are determined by the mass of extracted residues after evaporation. NIRF113 has insignificant
residue after drying. Evaporated solvent condensate collected from procedures using NIRF113
is suitable for recycling only as NIRF113.
Distillate Fractions Handling
Handling of distillate fractions for MeCI2 and Freon-113 is shown in Figure 3. Segregation of
all materials is maintained by batch and fraction type. Pre-distillations are performed
automatically using the modified ABC Integrity 2000 system. Distillations are performed using
the ABC Integrity 2000 teflon spinning band distillation system. This distillation system has
been tested by the manufacturer to have 50 theoretical plates or more with cyclohexane and
methylcyclohexane.
High contaminant levels, regardless of boiling point, can effect the quality of any separation.
Experience has shown that increasing the volume of the forecut and still-bottom fractions can
improve the quality of the collected fraction. The number of collect fractions can be subdivided
manually if desired.
The amount of forecut collected varies according to the waste type and system operating
conditions. Azeotropes and lighter boiling materials are collected in the forecut. Water forms
azeotropes with both MeCI2 and Freon-113, and is always collected in the forecut. The solvent
drying step prior to distillation reduces the amount of forecut needed to remove trace water
from a Freon-113 or MeCI2 solvent waste. Often, free water is observed as an immiscible layer
in the initial portion of the forecut. Head temperature and observation of moisture condensate
at the head of the distilling column are used to determine the presence or absence of water.
Forecut is not complete until all moisture is removed. Collected forecuts should be dried and
re-distilled with other similar materials. Due to possible percent levels of other solvents in a
248
-------
re-distilled forecut, recycled forecut may be suitable only for glassware washing and other
similar laboratory operations. Specialized handling techniques can be used to remove certain
solvents prior to re-distillation.
Depending on the quality of starting material, still-bottoms (raffinate) from pre-distillations and
distillations are either disposed of as hazardous waste or are retained and combined for further
recycling with other similar still-bottom material. The amount of forecut collected when
reprocessing them should be minimal.
QA/QC Procedures
Testing procedures are selected to characterize material suitability based on intended use of
the recycled product. Recycled MeCI2 and MeCI2 waste solvents were prepared for GC/MS
analysis using the solvent concentration procedure normally used for EPA method 3520/8270
analysis. J.T. Baker "Capillary Analyzed" MeCI2 is used as the control solvent. Recycled
solvents are concentrated from 200 ml to a 1 mL final volume in a Turbovap prior to analysis.
Solvent concentrates are analyzed according to EPA method 8270 on a Hewlett Packard 5970
MSD. The selected list of 8270 analytes tested for represents most of the analytes from the
CLP Semivolatile Target Compound List. A five point calibration ranging from 20 //g/ml to 150
/yg/mL was used. Internal standards fortification levels were 40 fjg/ml. The gas
chromatograph was operated in the splitless injection mode with a 0.5 /A. injection on a 30
meter 0.25 mm i.d., 25 //m film thickness, XTI-5 capillary column. The GC was temperature
programmed: 40° C hold 4 minutes, 20° C/min. to 120° C., hold 1 minute, 10° C/min to
250° C hold 2 minutes, 10° C/min to 310° C hold 6 minutes.
Additional QA/QC testing for MeCI2 such as density, acidity, solvent purity by GC/FID and
GC/ECD (concentrated then hexane solvent exchanged) are performed, but not reported here.
Freon-113 is tested for impurities by GC/FID, and also by Infrared (IR) according to EPA
Method 418.1.
RESULTS AND DISCUSSION
Analytical data which demonstrate the effectiveness of recycling by distillation are shown in
Figures 4 through 8. Chromatograms of solvent method blanks for these analyses are given in
Figures 7 and 8. The material presented illustrates the results from recycling a highly
contaminated MeCI2 solvent waste. Figures 4, 5, 6, 7, and 8 are EPA method 8270 GC/MS
reconstructed ion chromatograms (RIC) of the primary starting material, first distillation, second
distillation, solvent method blank for Figures 4 and 5, and solvent method blank for Figure 6,
respectively. Individual RICs are normalized to the largest peak present. Internal standards
are at the same level in all RICs presented (40 ^g/ml) and can be used to compare results
between the RICs shown. Internal standards were added after concentration and thus
concentration factors need to be accounted for. Target and non-target analyte GG/MS
identifications and quantitations are discussed during the presentation.
Figure 4 is the RIC of the original starting material concentrated 20:1. The source of this
material was general laboratory MeCI2 waste. This waste required phase separation and bulk
filtration to remove water and sediments. The resulting filtered product had a strong yellow
tint and smelled of gasoline. The filtered material was pre-distilled to remove soluble non-
volatile residues. The 1.5 L of still-bottom from the pre-distillation was disposed of as
hazardous waste.
249
-------
Figure 5 is the RIC of the collected fraction of the first distillation. The collected material was
concentrated 200:1 for analysis by GC/MS. Total batch size was 19 L to give: 2 L of forecut;
15.5 L of collected fraction; and 1.5 L of still-bottom. No analytes on the EPA method 8270
list were detected above method reporting limits. However, the material was not considered
suitable for use in EPA Method 8270 or other laboratory operations due a series of early eluting
peaks observed in the RIC. The collected fraction was re-distilled and the forecut and still-
bottom were recycled.
Figure 6 is the RIC of a second distillation which contained 15 L of the material shown in
Figure 6, and 4 L of a similar material also requiring a second distillation. The distillate
collected was concentrated 200:1 for GC/MS analysis. The distillation produced: 1.8 L of
forecut; 13.5 L of collect; and 3.5 L of still-bottom. The forecut and still-bottom were
recycled. The collected material is considered suitable for use in EPA Method 8270 and was
also determined to be suitable for use in EPA method 8080 (data not shown).
Figures 7 and 8 are the solvent method blanks associated with the solvent concentration and
analysis. Differences in GC/MS instrument sensitivity and the chromatographic peak shapes
reflect analyses performed over two different time periods.
SUMMARY
Solvent recycling for analytical testing laboratories is environmentally responsible and
economically viable. Recycling not only reduces disposal costs, but provides a reusable
product at potentially less cost than the virgin solvent. Estimates of labor and material costs
are given during the presentation.
The material presented here evaluates the potential for solvent recycling using purification by
spinning band distillation. The purity of the recycled MeCI2 solvent is evaluated relative to
probable intended use. Quality assurance and quality control requirements for EPA method
8270, and other similar procedures, specify that target analyte cannot exist in the method
blanks at concentrations above the reporting limits. Therefore, solvents used in sample
preparations for EPA Method 8270 must be free of target analytes and interferences after
equivalent solvent concentration techniques. American Chemical Society (ACS) specifications
for solvent purity are also recommended guidelines for determining solvent quality. The quality
of the recycled MeCI2 product produced in the second distillation discussed here met these
requirements. Some recycled materials not suited for use in sample extraction may be suitable
for use in other laboratory operations such as glassware cleaning and rinsing.
Methylene chloride is particularly well suited for recycling by distillation since it is primarily
used in analysis procedures for semivolatile compounds. Trace levels of other solvent
impurities do not interfere in the analysis of semivolatile compounds. EPA Method 418.1 using
Freon-113 extraction and analysis by IR is affected by trace levels of solvent impurities.
Because trace levels of other solvents can interfere in IR analysis, recycling by distillation is
more difficult and requires very careful waste solvent segregation.
250
-------
ACKNOWLEDGEMENTS
The authors would like to thank the following people for their contributions: Gary Warren,
Jack Hanewald, Jim Rasmussen, TJ Parker, Dan Dalbey, Bruce Williams, John Standish, and
other chemists and staff of Energy Laboratories for their support and on-going cooperation in
solvent collection, segregation, and testing; Jim Stunkel, Kevin Kelly, and Mark Orr of ABC
laboratories for their technical assistance; Judy Gebhart of ICF-QATS for her editorial
suggestions; and Roger Rourk Jr. and Paul Van Trieste of B/R Instruments for the pre-distilling
glassware.
251
-------
TABLE 1
SOLVENT RECYCLING LABORATORY - MATERIALS and EQUIPMENT LIST
FACILITIES
Solvent preparation area
Distilling apparatus area
Fume Hood or equivalent vapor control system
Bulk solvent storage area
Two, 110 volt 20 amp circuits
Deep well sinks
APPARATUS
ABC Integrity 2000 automated spinning band
distillation system - 19 Liter option
Custom Glassware to fit ABC Integrity 2000
19 Liter pot flask (ABC)
2 Liter separatory funnels
5 inch powder funnels (glass)
Oilless solvent resistant vacuum pump
(Cole Farmer teflon coated diaphragm pump)
Cryotrap
Ice bath
4 Liter vacuum flask
10 inch diameter buchner funnel
Whatman #4 filter paper, 24 cm
Alltech In-line filter degasser
Whatman #41 filter paper, 4.7cm
Anhydrous calcium chloride pellets
Cyclohexene
Hydrometers
Solvent storage containers
(Clean empty 4 Liter amber glass solvent jugs)
(Clean empty 100# Freon-113 Cans)
Sampling containers
(40 ml VOA vials, 4 Liter amber glass jugs)
Hazardous waste storage containers
Analytical testing laboratory
MISCELLANEOUS APPARATUS
Aluminum foil
Tools
Teflon tape
Nitrogen gas supply (clean and dry)
Drum cart
Cleaning brushes
Glassware cleaning detergents
1:1 Sulfuric acid
Methyl alcohol
Transfer pipets (5, 15 and 25 mL)
Safety equipment
Teflon tubing
Teflon squeeze bottle
Rubber stoppers
FUNCTION
Site for solvent handling
Site for distillations
Solvent vapor removal
Storage of solvents
System power
Glassware washing
FUNCTION
Distillation and Pre-distillation
Pre-distillation
Pre-distillation
Phase separation
Solvent transferring
Vacuum filtration and solvent transferring
Vacuum filtration vapor recovery
Vacuum filtration vapor recovery
Bulk filtration
Bulk filtration
Bulk vacuum filtration
In-line vacuum filtration and transferring
In-line vacuum solvent filtration
Solvent drying
Methylene chloride solvent preservative
Density determinations
Solvent fraction storage
Sampling
Storage of un-recyclable solvent wastes
Materials testing
FUNCTION
Clean work surface, vapor control
Setup and maintenance
Fittings and sealing solvent containers
Evacuating solvent vapors from containers
Drum handling
Glassware cleaning
Glassware cleaning
Glassware cleaning
Glassware cleaning
Sample collection
Safety
Solvent transfer
Solvent handling
Vacuum system connections
252
-------
Figure 1
METHYLENE CHLORIDE WASTE STREAMS
Processing Flowchart
1 GENERAL &
EXTRACTOR
Contaminants include:
Solvents
Semivolatiles
Water
Particulates
1
1
SEGREGATION
1
PHASE SEPARATION
|
BULK FILTRATION
1
PRE-DISTILL
*1) Collect Fraction
2) Still Bottom
|
SOLVENT DRYING
I
I
WASTE CATEGORY
GPC
Contaminants include:
Semivolatiles
Trace Solvents
| SEGREGATION
PRE-DISTILL
(as indicated)
SOLVENT DRYING
(as indicated)
-» Pll TPATinM ,-.
CONDENSATES FROM
TURBOVAP or
KUDERNA DANISH
Contaminants include:
Solvents
Methyl Alcohol
Hexane
Acetone
Water
| TESTING
SEGREGATION J
| PHASE SEPARATION
L SOLVENT DRYING
DISTILLATION
1) Forecut
*2) Collect Fraction
3) Still Bottom
PRESERVATION
SCREENING TESTS
(pass )
(fail)
SEGREGATION
COMPOSITING
COMPOSITE TESTING
DISTRIBUTION
•Collect Fraction is processed towards final product. See Figure 3 "Distillations" for
handling procedures of other fractions.
253
-------
Figure 2
FREON-113 WASTE STREAMS
Processing Flowchart
(NIRF113)
NON-IR GRADE
WASTE CATEGORY
NON-IR or IR GRADE
(IRF113)
IR GRADE
Contaminants include:
>2000 ppm TPH
Unleaded gasoline
Recyled Non-IR Freon
Water
Participates
T
Contaminants include:
200-2000 ppm TPH
Diesel Range Organics
Water
Particulates
TESTING
SEGREGATION |
I
PHASE SEPARATION
BULK FILTRATION
PRE-DISTILL
*1) Collect Fraction
2) Still Bottom
SOLVENT DRYING
FILTRATION
TESTING
DISTILLATION
1) Forecut
*2) Collect Fraction
3) Still Bottom
TESTING
(Pass)
SEGREGATION
COMPOSITING
COMPOSITE TESTING
Contaminants include:
<200 ppm TPH
Water
Particulates
I
(Fail)
| DISTRIBUTION |
''Collect Fraction is processed towards final product. See Figure 3 "Distillations" for
handling procedures of other fractions.
254
-------
in
01
Figure 3
DISTILLATIONS
FRACTIONS HANDUNG
i. PRE-DISTILLATrqN]
I
I , , I ,
COLLECT (80-95%) | fSTILL-BOTTOM (5-20%]
(Handle as per solvent |
processing flow chart) |
I I , I
FILTRATION H I WASTE
I ,
PRE-DISTILLATION |
I
DISTILLATION
! ,
i
i
I , ,__
COLLECT (60-80%) | | FORECUT (10-20%) | | STILL-BOTTOM (10-20%)
(Handle as per solvent | |
processing flow chart) | j
PHASE SEPERATE
WASTE
SOLVENT TESTING
DRYING AGENT
FILTRATION
FILTRATION
WASTE
PRE-DISTILLATION
DISTILLATION
-------
to
s
Figure 4. Reconstructed Total Ion Chromatogram of General Methylene Chloride Waste concentrated 20 : 1 and analyzed per
EPA Method 8270. Internal Standards at 40ug/L
-------
ro
tn
6.4-
6.2:
t.O-
5.8-
5.6-
5.4-
5.2-
5.0-
4.8-
4.6-
4.4-
4.2-
4.0-
3.9-
3.6-
33.4-
S3.2-
X
^3.0-
2.8-
2.6:
2.4:
2.2:
2.0:
1.6:
1.4-
1.2:
1.0:
0.8-
0.6-
0.4-
0.0 ,
4
i
li
t
1
1
L
i
I
6
i
U
f
B
S. g ^
O> LJ U1
5 ^ o
1 xx r~t
£ ™ I
1 1 I
I i ,J
-
Wj [__
\1
1
J W ^
co c^' J5
o
-p
§
1
V
I ?
1
1
,
1
^.]
10 12 14 16 18 20 22 24 26 28 30
Hln
32
34
36
38
Figure 5. Reconstructed Total Ion Chromatogram from first distillation of General Methylene Chloride Waste concentrated 200:1
and analyzed per EPA Method 8270. Internal standards at 40ug/L
-------
1.6
1.5
to
01
09
1.3-
1.2-
1.1
1.0
0.9
So.8-
X
*s
>-
0.6-
0.5
0.4
0.3
0.2
0.1
0.0-,
4
10
12
00
16
18
20
Hin
22
24
Figure 6. Reconstructed Total !on Chtomatograms from second distillation of General Methylene Chloride Waste concentrated
200:1 and analyzed per EPA Method 8270. Internal standards at 40ug/L
-------
8.8-
8.5-
8.2-;
8.0-
7.8-
7.5-
7.2-
7.0-
6.8-
6.5-;
6.2-:
6.0-!
5.8-:
5.5-:
5.2-i
5.0-:
4.8-:
iN
34.2-:
^4.0-:
3.8-1
3.5-:
3.2^
J.o-i
2.8-1
2.5-:
2.2-;
2.0^
1.8-:
1.5-i
1.2-!
i.o-:
O.B-:
0.5-
0.2-
-S -*
ro ^ « i-
^ ^ tn
•^f r^. ^
•p ^ "^
1 i' 1
I 1 ]
i \ T
r-t
F
L..JI .. I
cj in
•^) OJ
(T» hO
T3J \fl
s
ID
fe
t-
>-i
1
r-j
a
TJ
k
c
j>
i
\i .. —
A Q i/ * I, i^ li.J»t..^.LL | | , M — r i. . — - - - —
4 6 8 10 12 14 16 18 20 22 24 26 28 30
Hln
32
36
38
Figure 7. Reconstructed Total Ion Chromatograms of Solvent Method Blank for Figures 4 and 5.
Solvent concentrated 200:1 and analyzed per EPA Method 8270. Internal standards at 40ug/L
-------
2.4-i
2.3i
2.2-:
2.1-i
2.0-i
1.8-;
1.7-i
1.6-:
1.5-;
1.4-
i.o-:
0.9-:
0.8-;
0.7-i
0.6H
0.5-:
0.4-;
0.3-i
0.2-i
0.1-'
14
16
18
20
Hln
22
24
36
Figure 8. Reconstructed Total Ion Chromatograms of Solvent Method Blank for Figure 6.
Solvent concentrated 200:1 and analyzed per EPA Method 8270. Internal standards at 40ug/L
-------
3R The Technology and Performance of Several New Immunoassay Methods and the
RISc 2000 Instrumentation System.
Stephen B. Friedman, James P. Mapes, Randy L. Allen, Thomas N. Stewart, William B.
Studebaker, Patricia P McDonald, Richard E. Almond, Tracy A. Withers, Shannon P.
Arrowood, and Darrel P. Johnson
EnSys Inc.
Data generation is often the rate-limiting element that influences the overall
efficiency of the environmental clean-up process. New immunoassay methods are
providing a faster and less expensive way to detect the pathogenic compounds that
are present in the environment.
We have developed a number of immunoassay methods that expedite site mapping,
remediation and monitoring activities by providing reliable results immediately
and for less cost. The immunoassay methods for Pentachlorophenol, PCB's and
Total Petroleum Hydrocarbon detection have been previously reviewed and have
been accepted for use by the EPA.
Additional methods, and a RISc 2000 instrumentation system, have been developed
and evaluated. These new methods will '.steel PAH's, explosives \,.e. TNT, ^Dx),
and petroleum contamination in a variety of matrixes. The PAH and Petroleum
Contamination tests use monoclonal antibody reagents in a chromogenic enzyme
immunoassay format. The tests for explosives use the chromogenic chemistry of
the recently accepted EPA method 8510. All of the tests permit the simultaneous
testing of multiple samples at multiple detection levels and provide results
within 30 minutes. The PAH, Explosives and Petroleum Test for Water all retain
our conservative approach to environmental testing and have demonstrated a >95%
confidence in their ability to detect contamination when used in accordance with
the recommended protocols.
The EnSys RISc 2000 Instrumentation System has been developed to support the
expanding library of immunoassay methods that are currently available or that
will be introduced. The instrument rapidly analyzes the kinetics of the
immunoassay's chromogenic reaction and uses this data in the calculation of
results. The RISc 2000 instrument will read and record sample results at the rate
of one every 10 seconds. System characteristics include faster results, additional
quality control, and a printed report that summarizes the data obtained and the
interpretation of results. This caper will review the applications, chemistry and
performance characteristics of the methods and instrumentation developed.
261
-------
37 SCREENING OF TCLP EXTRACTS OF SOIL AND WASTEWATER
FOR 2,4-D BY IMMUNOASSAY
M.C. Haves. S.W. Jourdan, T.S. Lawruk and D.P. Herzog, OHMICRON, 375 Pheasant
Run, Newtown, Pennsylvania 18940
ABSTRACT
A commercially available enzyme immunoassay kit has been adapted to screen TCLP
extracts of environmental samples for the regulatory action levels of the commonly used
herbicide, 2,4-Dichlorophenoxyacetic Acid (2,4-D). The immunoassay was originally
designed to quantitate very low concentrations of 2,4-D in water and includes calibrators
at 1, 10 and 50 ppb. A simple one-step 1:1000 dilution of a TCLP extract brings the
detection range of the assay in line with the 2,4-D toxicity characteristic established by the
EPA at 10 ppm. The method which combines the standard TCLP extraction with dilution
and subsequent analysis by a magnetic particle-based ELISA has been evaluated for
sensitivity, matrix effects, accuracy and susceptibility to interferences. The results of the
evaluation studies are described here.
INTRODUCTION
The Methods Section of the EPA Office of Solid Waste has stated a need for'more rapid,
less expensive field screening procedures that do not compromise the accuracy of
pollutant evaluation. Availability of reliable, sensitive, inexpensive screening of large
sample loads of potentially contaminated samples coupled with subsequent quantitative
analysis of positives by established instrumental methods is an approach that addresses this
need. The Methods Section of OSW has provided an outline of the type and quantity of
work needed to demonstrate whether a proposed procedure is suitable for inclusion in
SW-846 as a screening method. Using this outline as an experimental model, a
demonstration of the performance characteristics of an immunoassay screening method for
2,4-dichlorophenoxyacetic acid (2,4-D) has been completed and is presented here.
Currently approved SW-846 methods for analysis of 2,4-D in soil or water samples
require the performance of difficult solvent extractions and chemical derivatization
reactions by highly trained chemists and the use of expensive analytical instruments
(Methods #8150 and #8151). The use of the Toxicity Characteristic Leaching Procedure
(TCLP/Method 1311) as the method of choice for preparing extracts of liquid and solid
wastes for subsequent analysis of the original fourteen priority pollutants, including 2,4-D,
heavily influenced the design of the study. All wastewaters and soils used in
demonstrating the capability of the immunoassay method were subjected to the Method
1311 extraction procedure prior to analysis. The 2,4-D RaPID Assay® kit, a magnetic
particle-based immunoassay for 2,4-D developed and manufactured by Ohmicron, has
been commercially available since 1992 for the determination of part per billion levels of
262
-------
the herbicide in water. This study shows that the immunoassay can be readily adapted for
use as a screening method for analysis of TCLP extracts for the presence of 2,4-D at the
regulatory level of 10 ppm.
MATERIALS AND METHODS
Equipment and reagents: 2,4-D RaPID Assay kits, 12 x 75 mm disposable plastic test
tubes, 2,4-D TCLP Sample Diluent, precision pipets, magnetic separator, RPA-I
photometer and vortex mixer are available from Ohmicron, Newtown, PA. 2,4-D powder
for gravimetric preparation of spiking stock solutions was obtained from Reidel-de Haen
(Hanover, Germany). All stock solutions were prepared in acetonitrile (pesticide grade).
The accuracy of preparation of the stock solutions was assessed by comparison with an
EPA certified standard solution (5000 ug/mL in acetonitrile) purchased from NSI
Environmental Solutions, Research Triangle Park, NC.
Environmental Samples. Several water samples were studied for matrix effect. They
included municipal tap water from Newtown, PA and surface water ("runoff") collected
from a stream that flows through eastern Pennsylvania farmland. Wastewaters ("effluent")
were collected from two separate drainage or discharge pipes in an agricultural community
in south central New York State. Two soil types were evaluated: a Sassafras sandy loam
(<5% organic matter) from New Jersey and a muck soil (46% organic matter) also from
New Jersey.
Procedures. TCLP extractions were performed on soils and wastewaters as described in
Method 1311. Spikes were made volumetrically into the final extract after filtration. The
detailed TCLP extract screening procedure is published and available from Ohmicron as an
"Application Procedure." All immunoassay results are converted to ppm 2,4-D in the
waste extract for evaluation versus the regulatory level. Detailed directions on use of the
immunoassay kit are found in the package insert.
Reference Method. Selected blank and spiked matrices were split and sent to a local
reference lab for analysis by SW-846 Method #8150 for chlorinated herbicides.
RESULTS AND DISCUSSION
Establishing a cutoff concentration for the screening procedure. In order to use the
quantitative results produced by the immunoassay in a qualitative way, a cutoff was
established that could reliably discriminate the regulatory action level (10 ppm) from one-
half that level (5 ppm). The TCLP extraction buffer (sodium acetate/acetic acid pH 4.9)
was diluted with sample diluent and spiked with 2,4-D at 5, 7.5, 10 and 15 ppm. The
spiking stock solutions were verified for accuracy before use by comparison with an EPA
certified reference preparation. Results of this comparison are given in Table 1. Thirty
replicates of each solution were then tested in three immunoassay batch runs. Figure 1
shows that use of a cutoff concentration of 7.5 ppm provides perfect discrimination
263
-------
between 10 ppm and 5 ppm. This cutoff concentration falls near the mid-point on the
calibration curve where precision is optimal. In fact, the distribution of results observed
for a 7.5 ppm solution overlapped neither the 5 nor the 10 ppm distributions. In a similar
manner the TCLP extract of a 2,4-D-free organic soil was spiked and tested by the
immunoassay. In the case of the soil extract, ten replicates were tested at each level of
interest in one assay run. A similar pattern of results distribution was observed when a
TCLP extract of this muck type soil was used as the background matrix. Figure 2
demonstrates no overlap of the regulatory level with half its concentration. When the
definitions of "positive" as greater than 7.5 ppm and "negative" as less than 7.5 ppm are
applied to these data (Figure 3), the utility of the method as a screening procedure can be
seen. A clear distinction is observed for the buffer and the soil extract matrices between 5
and 10 ppm.
Cross reactivity with chlorophenoxy compounds and other pesticides. The compounds
listed in Table 2 were added to the immunoassay system until a response equal to 10 ppm
2,4-D was observed. When an effect equivalent to 10 ppm 2,4-D could not be found at a
concentration of 5,000 ppm, cross-reactants were tested once at 10,000 ppm. Table 2
shows that most of the 2,4-D esters tested reacted with more potency in the assay than
2,4-D itself. However, these derivatized compounds are not ordinarily significant for
environmental samples. Many commonly used pesticides that are structurally unrelated to
2,4-D showed no reactivity in the screening procedure up to 10,000 ppm. Structurally
related compounds of regulatory interest, 2,4,5-T and 2,4,5-TP (Silvex), reacted much
more weakly than 2,4-D. Greater than ten times more 2,4,5-T and approximately 135
times more Silvex would be required to produce a positive result with the screening
method. In other words, a sample load of 130 ppm 2,4,5-T or 1375 ppm Silvex must be
present to produce a result equivalent to a 10 ppm concentration of 2,4-D.
Silvex was studied in greater detail because it is considered a priority pollutant and a
regulatory action level for the TCLP Toxicity Characteristic has been published for this
compound. Increasing amounts of Silvex were mixed with a 5 ppm 2,4-D solution in
order to evaluate the amount of cross-reacting Silvex required to produce a response
above the 7.5 ppm cutoff. Results shown in Table 3 indicate that between 100 and 200
ppm of Silvex would be required to elevate a marginal 2,4-D sample concentration (5 ppm
spike) into the positive range. The regulatory action level for Silvex is 1 ppm. This
concentration of Silvex by itself or in combination with negative or even borderline levels
of 2,4-D in the immunoassay will not influence the baseline 2,4-D result whatsoever.
Matrix specific performance of the screening method. A variety of water and soil matrix
types were studied to determine their interference, if any, with the immunoassay screening
method. All matrices except the municipal water were first treated as solid or liquid waste
according to the procedures given in Method 1311 (TCLP Procedure). Prior to spiking,
each extract was diluted, tested in the immunoassay and determined to contain a non-
detectable concentration of 2,4-D. Aliquots of the extract were then spiked volumetrically
with 5, 10 and 15 ppm 2,4-D prior to 1:1000 dilution and subsequent immunoassay.
264
-------
Municipal drinking water was not extracted or filtered but was spiked directly with 35, 70
and 140 ppb 2,4-D prior to 1:7 dilution and immunoassay. This dilution brought the
drinking water concentrations into the same range of the calibration curve for analysis that
was used for the waste extract testing. These lower spiking levels were chosen because
the USEPA has proposed a MCLG of 2,4-D in drinking water of 70 ppb. Each of the 28
blank and spiked matrices were tested five times by the screening method.
Table 4 shows that the proposed screening method accurately discriminated negative and
positive regulatory levels in all matrices tested. All results on matrices spiked at one-half
the regulatory level (5 ppm) were less than 5.9 ppm 2,4-D showing no false positives
(results >7.5 ppm). Results on matrices spiked at the 10 ppm regulatory level ranged from
a low of 10.2 ppm to a high of 14.5 ppm with no false negatives (results <7.5 ppm).
Samples spiked above the regulatory level at 15 ppm gave quantitative results from 15.5
to 19.1 ppm 2,4-D, again showing no negatives. The slight positive bias in quantitative
results at the 10 and 15 ppm spike levels may be due to the fit of the calibration curve in
this concentration range or may have been introduced in the spiking or dilution steps. The
nature of the matrix appears to have no influence on the results. The results on the spiked
drinking water demonstrates that the 2,4-D immunoassay can be applied in a similar
manner (with a lesser dilution) to screening of drinking water matrices. Although the five
replicates run here do not show statistical discrimination of the concentrations of
regulatory interest demonstrated above, the drinking water results were calculated from
the same portion of the 2,4-D calibration curve as for the waste extracts so the precision
of the measurements should be identical to the buffer system shown in Figure 1.
Undiluted drinking water from a variety of sources has been shown to produce no
interference in the 2,4-D RaPID Assay (Lawruk, et al, 1993). Excellent discrimination of
negative from positive results on drinking water samples in the 35 to 70 ppb range can
therefore be expected.
Correlation with a quantitative SW-846 reference method. Selected samples from the
spiking study were submitted for analysis by the SW-846 Method #8150 for chlorinated
herbicides. The results of the analyses are given in Table 5. 100% method agreement was
seen at 1/2 the regulatory level and at 1.5 times the regulatory level - an important goal for
new screening methods as stated by EPA/OSW. The quantitative immunoassay results
agreed with the spiked 2,4-D concentrations better than the GC results did. However, in
cases where the spike level was frankly positive (15 ppm) or negative (5 ppm), the
interpretations of the two methods were identical. The qualitative status agreement
between the screening method and the reference method was more variable when the
spiking concentration was at or near the cutoff of both methods (10 ppm). This was true
in part because we are using different criteria to make the status judgment (7.5 ppm cutoff
for the immunoassay vs. 10 ppm cutoff for the reference method). In addition, the GC
method is expected to show a normal distribution of results around 10 ppm in repeated
analyses. If the GC result is used as the criteria for correct assessment of the sample
status, then only three results (#7, 19 and 23) show opposite status interpretations. In all
265
-------
three cases, the immunoassay would give a positive when GC was negative (<10 ppm).
This indicates that there may be a greater tendency toward false positive interpretation
than false negative interpretation when the screening method is evaluated in this way. The
correlation of the quantitative results of the two methods on eleven spiked samples in the
5 to 15 ppm range shown in Figure 4 confirms this observation. Regression analysis of
this limited data set predicts a 10 ppm GC concentration would be read at about 13 ppm
by immunoassay. The apparent higher concentration predicted by immunoassay could
result from the fit of the calibration curve in the 10 to 20 ppb range giving results that are
slightly higher than expected or lower results by the reference method could be due to
some loss of analyte during sample extraction, concentration and or derivatization steps.
A larger sample correlation study across a broader concentration range would provide a
better characterization of the correlation of these two methods, but quantitative
comparison of methods was not an important objective of this study. In another study
done with 56 water samples in the 10 to 500 part per billion range collected from various
locations across the United States, the 2,4-D RaPK) Assay was shown to correlate very
well (r = 0.970) with a GC/MS method (Lawruk et al, 1993).
SUMMARY
A magnetic particle based immunoassay for 2,4-D has been successfully adapted for use as
a screening method with TCLP extracts of some commonly encountered forms of liquid
and solid waste. The method is inexpensive compared to GC methods, is easy to perform
and can provide results on as many as 50 samples in less than an hour. The quantitative
accuracy of the immunoassay results in the ppm range has been demonstrated by
comparison of selected spiked samples with the currently accepted SW-846 method. A
screening cutoff concentration has been established that discriminates between the 10 ppm
regulatory level and half that level with no false negatives or false positives based on
spiking studies. Cross-reactivity with environmentally relevant compounds has been
shown to be insignificant and masking agents are of little consequence because of the
thousand fold dilution made prior to the immunoassay. This test should now be applied to
field testing situations where it is expected to be a valuable tool for producing rapid results
for site mapping of contaminated soils, for effluent monitoring in agricultural applications
and in remediation activities.
REFERENCES
Lawruk, T.S., Hottenstein, C.S., Flecker, J.R., Hall, J.C., Herzog, D.P., Rubio, P.M.,
Quantitation of 2,4-D and Related Chlorophenoxy Herbicides by a Magnetic Particle-
Based ELISA, 1993, (manuscript submitted for publication).
266
-------
Table 1. Calibration of spiking solution with EPA reference standard. Two hundred
fifty milligrams of 2,4-D (free acid) was dissolved in acetonitrile in a 25 mL volumetric
flask. Aliquots from 0.5 mL to 3.5 mL were diluted in one liter volumetric flasks to
prepare solutions containing 5, 10, 15 and 35 ppm of 2,4-D. A 2,4-D EPA reference
standard (NSI Environmental Solutions, Research Triangle Park, NC) with a certified
concentration of 5000 jig/mL (in acetonitrile) was diluted volumetrically in a similar
manner in order to produce solutions at the same concentrations. lOOjiL aliquots of each
solution were then diluted in 100 mL of TCLP Diluent prior to immunoassay.
concentration of 2.4-D (ppm)
source of 2.4-D predicted measured % difference
spiking solution 5 5.05 101
10 11.33 113
15 17.40 116
35 37.30 106
reference standard 5 5.05 101
10 10.70 107
15 17.14 114
35 38.91 111
267
-------
TABLE 2. Cross-reactivity of chtorophenoxy compounds and structurally unrelated
pesticides in the 2,4-D RaPID Assay. The compounds shown below were prepared in
the TCLP Sample Diluent and added to the immunoassay in increasing amounts until a
result equal to that seen in a sample containing 10 ppm 2,4-D was seen.
Compound
2,4-D
2,4-D propylene glycol ester
2,4-D ethyl ester
2,4-D isopropyl ester
2,4-D methyl ester
2,4-D sec-butyl ester
2,4-D butyl ester
2,4-D butoxyethyl ester
2,4,5-T methyl ester
2,4-D isooctyl ester
2,4-D butoxy-propylene ester
2,4-DB
MCPA
2,4,5-T
Silvex methyl ester
4-chlorophenoxyacetic acid
MCPB
Silvex (2,4,5-TP)
Dichlorophenol
Dichlorprop
Triclopyr
MCPP
Mecoprop
Pentachlorophenol
Picloram
Alachlor
Aldicarb
Aldicarb sulfate
Aldicarb sulfoxide
Atrazine
Benomyl
Butylate
Captan
Captofol
Carbaryl
Carbofuran
Dicamba
1,3-Dichloropropene
Dinoseb
Metolachlor
Metribuzin
Simazine
Terbufos
Thiabendazol
Concentration that
causes a 10 ppm 2.4-D result
10 ppm
0.52 ppm
0.54 ppm
0.96 ppm
1.09 ppm
1.40 ppm
1.60 ppm
2.00 ppm
12.0 ppm
20.0 ppm
20.6 ppm
95 ppm
110 ppm
130 ppm
665 ppm
815 ppm
980 ppm
1375 ppm
2380 ppm
5000 ppm
> 10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
> 10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
> 10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
> 10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
> 10,000 ppm
>10,000 ppm
>10,000 ppm
>10,000 ppm
268
-------
Table 3. Effect of silvex on 2,4-D results near the positive/negative cutoff point.
Solutions of silvex alone and silvex/2,4-D mixtures were prepared in TCLP buffer to
demonstrate the potential effect of a structurally similar, environmentally significant cross-
reactant on the immunoassay results. Silvex alone produced non-detectable 2,4-D results
in concentrations as high as 100 ppm. When increasing amounts of silvex
with a 5 ppm 2,4-D solution no
change in the status
of the result was seen
were mixed
until 200 ppm
of silvex was present in the mixture.
Silvex concentration 2.4-D
(ppm)
0
0.5
1.0
2.0
100
200
0
0.5
1.0
2.0
100
200
nd = non-detectable
concentration
(ppm)
0
0
0
0
0
0
5.0
5.0
5.0
5.0
5.0
5.0
2.4-D RaPID
Assay result
(ppm)
nd
nd
nd
nd
nd
0.8
4.46
4.99
4.48
5.03
7.09
9.17
Interpretation
negative
negative
negative
negative
negative
negative
negative
negative
negative
negative
negative
positive
269
-------
Table 4. 2,4-D spiking results on TCLP extracts of environmental waste matrices.
2,4-D was spiked into buffer, water or a TCLP extract of the matrices shown after the final filtration step. Each matrix was
diluted 1:1000 and tested by immunoassay five times. Each result was then compared with the 7.5 ppm cutoff to
determine if its status was positive or negative.
2.4-D concentration bv Immunoassav
Matrix/Spike
1 TCLP buffer
2 TCLP buffer + 15 ppm
3 TCLP buffer + 10 ppm
4 TCLP buffer + 5 ppm
5 Sandy soil extract
6 Sandy extract + 15 ppm
7 Sandy extract + 10 ppm
8 Sandy extract + 5 ppm
9 Organic soil extract
10 Organic extract+15 ppm
11 Organic extract + 10 ppm
12 Organic extract + 5 ppm
13 Effluent #1
14 Effluent #1 + 15 ppm
15 Effluent #1 + 10 ppm
16 Effluent #1 + 5 ppm
17 Effluent #2
18 Effluent #2 +15 ppm
19 Effluent #2 + 10 ppm
20 Effluent #2 + 5 ppm
21 Runoff
22 Runoff+15 ppm
23 Runoff+10 ppm
24 Runoff + 5 ppm
25 Municipal water
26 Municipal water + 140 ppb
27 Municipal water +70 ppb
28 Municipal water + 35 ppb
Determination #
cone, units
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppb
ppb
ppb
ppb
1
nd
15.5
10.6
5.0
nd
16.0
11.1
5.8
nd
16.1
11.2
5.2
nd
19.1
11.7
5.1
nd
16.3
12.1
5.1
nd
16.7
11.5
5.1
nd
162
75
38
2
nd
16.9
12.0
4.7
nd
15.8
12.4
5.9
nd
17.4
11.2
5.1
nd
18.2
12.6
5.7
nd
18.1
12.5
5.8
nd
19.0
11.2
5.4
nd
168
71
42
3
nd
16.3
10.2
5.0
nd
17.5
11.8
6.4
nd
16.3
11.5
5.0
nd
18.9
14.5
5.1
nd
17.0
10.4
4.5
nd
18.8
11.1
5.7
nd
155
75
37
4
nd
16.6
10.9
4.9
nd
16.5
10.8
4.9
nd
16.5
10.4
4.5
nd
17.9
11.1
4.8
nd
17.9
10.2
5.0
nd
17.5
11.8
5.2
nd
156
63
40
5
nd
17.9
10.4
4.3
nd
16.0
12.7
5.5
nd
17.5
11.0
5.1
nd
18.8
13
5.6
nd
15.8
12.6
4.9
nd
19.3
11.4
5.8
nd
165
78
33
mean
.
16.6
10.8
4.8
-
16.4
11.8
5.7
.
16.8
11.1
5.0
18.6
12.6
5.3
-
17.0
11.6
5.1
-
18.2
11.4
5.4
-
161
72
38
% POS
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
N/A
N/A
N/A
N/A
%Ntt
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
N/A
N/A
N/A
N/A
N/A = not applicable to wastewater regulatory limit
270
-------
Table 5. Comparison of screening results with quantitative 2,4-D results obtained with SW-846 Method #8150.
Selected samples described in Table 4 were split and sent to a local reference lab for analysis. The mean of five immunoassay results and the
status of that mean with reference to the 7.5 ppm cutoff is shown. Quantitative GC results by method #8150 are given. Some samples were
submitted in duplicate. The status of the GC result is reported as positive if greater than or equal to the 10 ppm regulatory limit, negative if less
than 10 ppm. Agreement of the screening result with the GC result is assessed .
immunoassay
LD.#
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
Marrix/SDtke
TCLP buffer
TCLP buffer +15 ppm
TCLP buffer +10 ppm
TCLP buffer + 5 ppm
Sandy soil extract
Sandy extract +15 ppm
Sandy extract +10 ppm
Sandy extract + 5 ppm
Organic soil extract
Organic extract +15 ppm
Organic extract + 10 ppm
Organic extract + 5 ppm
Effluent #1
Effluent #1 + 15 ppm
Effluent #1 + 10 ppm
Effluent #1 +5 ppm
Effluent #2
Effluent #2 +15 ppm
Effluent #2 + 10 ppm
Effluent #2 + 5 ppm
Runoff
Runoff + 1 5 ppm
Runoff + 1 0 ppm
Runoff + 5 ppm
Municipal water
Municipal water + 140ppb
Municipal water +70 ppb
Municipal water + 35 ppb
cone, units
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppb
ppb
ppb
ppb
mean
nd
16.6
10.8
4.8
nd
16.4
11.8
5.7
nd
16.8
11.1
5.0
nd
18.6
12.6
5.3
nd
17.0
11.6
5.1
nd
18.2
11.4
5.4
nd
161
72
38
immunoassay
status
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
N/A
N/A
N/A
N/A
2,4-D result by
Method #8150
nd
13.0
11.0
5.6
nd
*
5.9,52
*
nd
*
10.0, 9.5
*
•
*
11.0, 7.8
3.6
*
11.0
8.8, 9.5
*
nd
*
9.7, 8.6
5.5
nd
*
58, 59
*
GC status
negative
positive
positive
negative
negative
•
neg, neg
*
negative
*
pos, neg
*
*
*
pos, neg
negative
*
positive
neg, neg
*
negative
*
neg, neg
neg
N/A
N/A
N/A
N/A
Agreement of
screen and GC">
yes
yes
yes
ye*
yes
•
no
•
yes
*
equivocal
*
•
•
equivocal
yea
*
yes
no
•
yes
•
no
V"
N/A
N/A
N/A
N/A
nd = non-detectable
N/A = not applicable to wastewater regulatory limit
• = no analysis with method #8150
271
-------
15
Cutoff
0)
c
a
a:
_c
.3
0)
0)
">
Negative
Positive
1/XJ5.0 ppm Spike
l\\|7.5 ppm Spike
H010.0 ppm Spike
15.0 ppm Spike
10
15
20
ppm 2,4—D by Immunoassay
Figure 1. Demonstration of the sensitivity of the selected cutoff for 2,4-D in the
TCLP buffer matrix. 2,4-D was spiked into TCLP extraction buffer at concentrations of
5, 7.5, 10 and 15 ppm. Each spiked solution was then diluted 1:1000 in Ohmicron 2,4-D
TCLP Sample Diluent. Thirty immunoassay determinations were then made on each
diluted solution. The frequency of occurence of the results is plotted against the range of
2,4-D concentrations measured.
272
-------
8
Cutoff
CD
cn
c
o
V)
0>
O
=**=
6 -
4 -
2 -
0
5.0 ppm Spike
7.5 ppm Spike
10.0 ppm Spike
15.0 ppm Spike
10 15
ppm 2,4—D by Immunoassay
20
Figure 2. Demonstration of the sensitivity of the selected cutoff for 2,4-D in the
TCLP extract of an organic soil. 2,4-D was spiked into the TCLP extract of an organic
soil at concentrations of 5, 7.5, 10 and 15 ppm. Each spiked solution was then diluted
1:1000 in Ohmicron 2,4-D TCLP Sample Diluent. Ten immunoassay determinations were
made on each diluted solution. The frequency of occurence of the results is plotted
against the range of 2,4-D concentrations measured.
273
-------
3 100
E
Q.
Q.
in 80
Q)
O
Q.
E
o
o
tn
en
0)
0)
tn
O
Q_
60
40
20
0
O TCLP Buffer
V TCLP Extract of
Organic Soil
5 10 15
ppm 2,4-D Spiked into Matrix
20
Figure 3. Break point of the 2,4-D screening test in buffer and soil extract. Spiked
buffer or soil extracts described in Figs. 1 and 2 were graded for positivity or negativity
based on the 7.5 ppm cutoff. In both matrices the 10 ppm regulatory level is
distinguishable from 5 ppm (one-half the regulatory level) 100 % of the time.
274
-------
20
15
O
0
Q.
Q.
v«x
Q
i
^J-
C\j"
r = .836
slope = 0.589
intercept = 2.05
10
0
0
5 10 15
2,4-D (ppm) by immunoassay
20
Figure 4. Quantitative correlation of 2,4-D immunoassay with method #8150. TCLP
extracts of various matrices that were spiked at 5, 10 or 15 ppm 2,4-D were split and
submitted for quantitative analysis by method #8150 for chlorinated herbicides (GC
detection.) The GC result (average of duplicates in some cases) was plotted against the
corresponding immunoassay result (average of five replicates). The correlation
coefficient, slope and intercept was calculated by regression analysis for results on these
eleven samples.
275
-------
PETRO RIS5® WATER - A RAPID, ON-SITE IMMUNOASSAY FOR
DETECTING PETROLEUM PRODUCTS IN GROUND WATER.
Patricia P. McDonald, Richard E. Almond, James P. Mapes and Stephen B. Friedman
EnSys Inc. P.O. Box 14063, Research Triangle Park, NC 27709
ABSTRACT
There is a recognized need for innovative and cost-effective field methods for identifying
and assessing contamination of ground water with petroleum products. A rapid, on-site
enzyme immunoassay was developed for the detection of low levels of gasoline, diesel
fuel and other petroleum components in contaminated ground water samples. The test
sensitivity for gasoline is 165 ppb, for diesel is 245 ppb and for m-xylene is 100 ppb.
The test can also be used to measure other petroleum fuels such as fuel oil and jet fuel.
The immunoassay was developed as a complete kit to be used on-site. The test is rapid,
taking less than 30 minutes to complete the immunoassay.
Validation studies established that the test is specific for refined petroleum hydrocarbon
(PHC) constituents. Ground water samples which were PHC-free by GC analysis tested
negative in the immunoassay. Minimal matrix interferences were observed from ground
water obtained from different regions of the United States. The evaluation of field
samples demonstrated good correlation between the results obtained by the immunoassay
and by standard analytical methods for petroleum hydrocarbons. The test serves as an
accurate field based alternative to traditional lab methods for screening contaminated
waste sites.
INTRODUCTION
The contamination of ground water by petroleum hydrocarbons frequently occurs during
the processing and storage of refined petroleum products. A rapid procedure for
evaluating contaminated sites has been developed using competitive immunoassay
methodology. Immunoassay technology has been successfully applied to environmental
testing (1-6). To make it easier and less expensive to test for petroleum fuels in water,
we have developed and validated an on-site immunoassay for the detection of gasoline at
165 ppb and diesel at 245 ppb. In many states, regulations concerning water
contamination focus on the BTEX components. This test was designed to specifically
detect the significant components in gasoline and diesel contaminated samples. The
balanced response to both aliphatic and aromatic compounds gives the Petro RIS5® Water
Test wide applicability to testing petroleum fuels at low detection levels.
This test has an advantage over traditional analytical methods because it is an inexpensive
means of evaluating samples on site. For this reason, some of the variability which is
276
-------
observed in PHC analysis, due to problems with sampling and handling of volatile
analytes, is reduced.
EXPERIMENTAL METHODS
This paper presents data that was obtained using the Petro R1S5® Water Immunoassay kit
(EnSys, Inc., cat no. 70410) and the procedure outlined in the kit instructions. The kit
contains all the components necessary to process the water sample and conduct the assay
in the field.
The test consists of 3 principal steps. The first step is to prepare the ground water
sample for use in the immunoassay. The sample is collected in a VOA vial containing
buffer salts which are dissolved in the sample by mixing. The sample is briefly stored
in ice to allow the sample to cool and the sediment to settle. The second step is to
perform the immunoassay. The sample is transferred to a tube containing an enzyme
conjugate pellet (in a stable lyophilized form). The reconstituted conjugate and sample
mixture is poured into an antibody-coated tube. After incubating for 10 minutes to allow
the competitive assay to proceed, the tube is washed and the color producing substrates
are added. The final step is the interpretation of results which compares the color
intensity of the sample to the kit standard which is run in parallel with the sample. A
lighter color for the sample indicates that the sample is contaminated.
RESULTS AND DISCUSSION
The Petro RIS5® Water Test was validated by evaluating several performance criteria
including the immunoassay sensitivity, specificity, and precision. The test was also
examined for correlation to a reference method using field samples and gasoline spikes.
The kit storage stability was evaluated at both ambient and elevated temperatures. All
of these performance characteristics are important to assure that the test is accurate and
field compatible.
The Petro RIS5® Water Test was configured to give less than 5% false negative results
for samples containing concentrations of petroleum fuels at or above the detection level.
This was accomplished by setting the m-xylene standard at 55 ppb. At this concentration,
95% of samples containing 100 ppb m-xylene, 165 ppb gasoline or 245 ppb diesel will
be detected as positive. Other compounds were tested in the assay to determine a cross-
reactivity profile. The concentrations of PHC needed to give a positive result in the assay
are listed in Table 1. The test recognizes several petroleum hydrocarbon constituents
with the greatest sensitivity to short chain aliphatic and small aromatic compounds.
The assay specificity was examined using two approaches. First, PHC-free water samples
were evaluated and found to be negative in the immunoassay compared to the standard.
Second, the matrix interference was examined in the assay using PHC-free water samples
277
-------
from different geographical locations across the United States. When the signals of the
PHC-free samples were compared to the signal from a PHC-free laboratory (deionized
> 18MQ with an organic removing filter) water control, the mean signal ratio (BSAMPLE
OD/ BLAB OD) was 0.96 ± 0.07. The absence of matrix effect would yield a ratio of 1.
This data demonstrates that there was minimal matrix interference in the immunoassay.
The precision of the test was evaluated by measuring intra-assay repeatability and inter-
assay reproducibility. Replicate determination of assay signals for samples containing 100
ppb m-xylene spikes were obtained. The intra-assay percent coefficient of variation was
< 8% which was equivalent to 12 ppb m-xylene. The inter-assay percent coefficient of
variation was < 12% which was equivalent to 19 ppb m-xylene.
The storage stability of the test kit was evaluated by testing the assay performance in real-
time and temperature-accelerated (storage at 37 °C) studies for 3 months (Figure 1). The
regression analysis of assay signal decay and sensitivity (BSTANDARD/BOPPM CONTROL) currently
suggests that long-term stability will be attained for at least 6 months when the kits are
stored at room temperature (about 22 °C).
Field trials were used to establish a correlation between the Petro RIS2® Water Test and
the reference USEPA method 5030/modified 8015. In these studies, there was an overall
correlation of 84% with individual site correlation ranging from 71 % to 100%, Table 2.
No false negative results were observed which is consistent with the test configuration.
CONCLUSION
This paper has presented an alternative method for quick field screening of petroleum
hydrocarbon contamination of ground water samples. We have validated a competitive
enzyme immunoassay for the detection of gasoline and diesel at low concentrations in
water. The rapid screening of multiple samples is a significant advantage to traditional
analytical methods.
REFERENCES
(1) Van Emon, J. M., & Lopez-Avila, V. (1992) Anal. Chem. 64, 79-88
(2) Van Emon, J. M. (1990) ACS Symposium Series 442, American Chem. Soc.,
Washington, D. C., p. 58-64
(3) Van Emon, J. M., R. W. Gerlach, R. J. White, & Silverstein, M. E. (1991) in
Field Screening Methods for Hazardous Wastes and Toxic Chemicals, p. 815-818
(4) Mapes, J. P., McKenzie, K. D., McClelland, L. R., Movassaghi, R., Reddy, R.
A., Allen, R. L., & Friedman, S. B. (1992) Bull. Environ. Contam. Toxicol. 49,
334-341
(5) Allen, R. L., Manning, W. B., McKenzie, K. D., Withers, T. A., Mapes, J. P.,
& Friedman, S. B. (1992) in Contaminated Soils, Diesel Fuel Contamination, P.
278
-------
T. Kostecki & E. J. Calabrese (Eds), Lewis Publishers, Chelsea, M, pp. 37-46
(6) Goh, K. S., Hernandez, J., Powell, S. J., Garretson, C., Troiano, J., Ray, M.,
& Greene, D.D. (1991) Bull. Environ. Contam. Toxicol. 46, 30-36
279
-------
Table 1. Petro RIS2® Water Test Sensitivity
Compound or Substance
Cone. Needed to Give
a Positive Result
(ppb)
Petroleum Fuel Products
Gasoline
Diesel fuel, #2
Jet A fuel
Jet fuel, JP-4
Kerosene
Fuel oil, #2
Formulated Petro Products
Mineral spirits
Aromatic Compounds
Toluene
Ethylbenzene
o-Xylene
m-Xylene
p-Xylene
Styrene
Naphthalene
Acenapththene
165
245
280
185
215
210
490
740
65
100
100
590
65
8
10
Methanol was spiked individually with the listed compound. Standard curves were
performed with each compound and compared to the m-xylene standard. The
concentration of cross-reactant necessary to give a positive result was calculated from the
relative ED^ (estimated dose at 50% binding) of the compounds and m-xylene.
280
-------
Table 2. Correlation of Petro RIS5* Water Test Field Trial Results with EPA GC-
FID Method (5030/modified 8015) Laboratory Results
Trial No.
1
2
3
4
Total
Correlation of
Immunoassay Results with
GC-FID Results
93 %•
100%«
71 %•
75 %•
84%
False Pos.
Results
(FP/Total)
1/14
0/10
4/14
3/11
7/49
Four field trials were conducted with the Petro RIS5® Water Test at former service
stations. Contaminated groundwater samples were collected in duplicate. One set of
samples was analyzed on-site and the other set was sent to the laboratory for analysis by
GC-FID. Gasoline field spikes were made in blank water at the time of collection.
• The laboratory underestimated the gasoline field spikes by > 25%.
281
-------
1.8
1.6
I
1.4
1.2
1
0.8
0.6
0.4
0
0.5
1.5 2 2.5
Time (months)
3.5
Figure 1. Storage stability of the Petro RIS5® Water Test
The test kit was stored at 37°C. At each of the indicated time points the
kits were allowed to cool to ambient temperature and were used for
analysis. Bo is the absorbance at 450 nm from an unspiked methanol
sample. Bstd is the absorbance obtained with the kit standard, and the
Bstd/Bo is the ratio of the two absorbance values. A change in the ratio
would indicate a change in the test sensitivity.
282
-------
39 APPENDIX IX EXTRACTIONS
BY
ACCELERATED ONE STEP™ EXTRACTOR / CONCENTRATOR
Mark Bruce. Director of Research and Development, Enseco-Wadsworth/ALERT Laboratories,
a Division of Corning Lab Services, 4101 Shuffel Dr. N.W., North Canton, Ohio 44720; James
Carl, Product Development Manager, Bruce Killough, Product Development Technician, Tony
Zine, Product Development Specialist, Science Products Division, Corning Incorporated, 673
County Rt. 64, Big Flats, New York 14814; David Burkirt, President, Burkitt Consultants, Inc.,
2 Captain's Way, Exeter, New Hampshire 03833.
ABSTRACT
The Accelerated One Step™ system shortens extraction time from 18 to 5.5 hours. The solvent
requirement is reduced from 500 to 100 mL. This new liquid-liquid extraction system was
validated by spiking all Appendix DC organochlorine pesticide and semivolatile (BNA) analytes
into 3 aqueous matrices at 2 concentration levels. Ground water, TCLP buffer #1 and waste
water were spiked near the expected method detection limits and at levels 20 to 50 times higher.
Organochlorine pesticide and BNA accuracy and precision are within the limits published in
SW-846. The average recovery for all Appendix EX pesticides and BNAs is 77% with 13%RSD.
The Accelerated One Step™ system reduces solvent requirements by eliminating the solvent
pool at the bottom of the extraction chamber, reducing the volume of the boiling flask and
minimizing solvent vapor losses. The extraction is accelerated by shortening the analyte
transfer time from extraction chamber to boiling flask and maintaining a high solvent boiling
rate. The membrane resists plugging by most common environmental samples. Only very oily
samples have impeded proper solvent cycling.
The Corning Accelerated One Step™ extractor/concentrator system is equivalent to
conventional continuous liquid-liquid extraction while using '/5 the solvent and '/3 the
extraction time. The accuracy, as measured by percent recovery, is equivalent to conventional
continuous liquid-liquid extraction. Precision, as measured by %RSD, is better with the
Accelerated One Step™ system. The extracts contain fewer interfering polar compounds, this
improves analysis (particularly for TCLP). The new system also reduces labor requirements and
increases lab extraction capacity. The Accelerated One Step™ system will contribute to more
timely and cost effective environmental analyses while reducing laboratory use, exposure and
disposal of hazardous solvents.
INTRODUCTION
Solvent reduction has become one of the battle cries of the Environmental Protection Agency
for the 1990's. The EPA has received considerable environmental and political pressure to
reduce the amount of pollution generated by analytical methods which are required to
demonstrate compliance with EPA regulations. The largest volume of analytical waste is from
the organic solvents required by current organic extraction methodologies. Both SW-846
methods, 3510 (separatory funnel) and 3520 (continuous liquid-liquid), are used for aqueous
samples and require large amounts (300-500 mL) of organic solvents such as methylene
chloride. In addition, continuous liquid-liquid extraction times are typically 18 hours, plus setup
and cleaning time.
Two technologies in particular are being touted by the EPA as the "solution" to the organic
solvent waste disposal problem. Solid Phase Extraction (SPE) and Supercritical Fluid
Extraction (SFE) offer much promise to reduce solvent use and shorten extraction time. The
largest impediment to full-scale environmental application of SPE and SFE is the lack of
283
-------
methodologies rugged enough to handle the great diversity of environmental matrices and target
analytes. Solid phase extraction has been successful for selected analytes in clean water
matrices such as drinking water. However, some chemical and mechanical problems remain.
Extraction of very water soluble analytes (such as phenol) from 1L samples has usually yielded
recoveries less than 20%. Particulates from waste water samples are prone to plug both SPE
cartridges and disks. Prefilters and larger disks have reduced the plugging problem but many
environmental samples are still too problematic to be handled routinely by SPE. Current SPE
technology cannot replace conventional liquid-liquid extraction for all analytes and matrices.
However, this does not mean that solvent reduction efforts must be postponed.
The Accelerated One-Step™ extractor/concentrator system is an excellent interim solvent
reduction solution that is available now. The solvent volume reduction from 500 mL to 100 mL
(or less) is very attractive. In addition, reducing the extraction time from 18 hr to 5.5 hr is also
beneficial. The extraction chemistry of the new One-Step™ system is the same as conventional
liquid-liquid extraction. Thus, the data are equivalent.
This new extractor was evaluated in several areas to test its viability for routine organic
extractions. Analyte, can the target compounds covered by the Appendix IX list be
quantitatively extracted? Matrix, can all usual water matrixes be extracted without mechanical
problems from emulsions or paniculate plugging? Accuracy & precision, are method bias and
reproducibility equal to or better than conventional liquid-liquid extraction? Ruggedness, is the
method and glassware durable enough to tolerate misuse and still produce acceptable results? Is
the extract "dry" enough that drying with sodium sulfate is no longer needed?
In short, the goal was to develop an apparatus which would extract as well as conventional
continuous liquid-liquid extraction for all common environmental water matrixes yet be safer,
faster and more cost effective.
EQUIPMENT AND SUPPLIES
Hardware
Accelerated One-Step extractor / concentrator systems (see Figure 1)
Neslab refrigerated circulator, CFT-75
Tecator heated circulators, 1046
Reagents
Methylene chloride, hexane, methanol, acetone, sulfuric acid, sodium hydroxide
RESULTS & DISCUSSION
Glassware
The accelerated One-Step™ apparatus differs from conventional continuous liquid-liquid
extraction (and the current One-Step™ extractor) in four key areas. 1) The solvent pool at the
bottom of the conventional extraction chamber has been eliminated. The solvent is returned
from the bottom of the extraction chamber to the distillation flask (or K-D) via gravity feed
rather than siphon action. A hydrophobic membrane is placed across the bottom of the
extraction chamber. Organic solvent is dripped through the sample in the conventional manner.
However, the solvent passes through the membrane at the bottom and runs back to the
distillation flask. No pool of solvent is required at the bottom of the extraction vessel for siphon
purposes. Thus, less solvent is required. The extraction time is also shorter since it is not
necessary to transfer analytes from the solvent pool to the distillation flask via the solvent pool
dilution process of a conventional continuous liquid-liquid extractor. Figure 1 shows the flow
of solvent. 2) The hydrophobic membrane effectively excludes water from the solvent thus
eliminating the need for a sodium sulfate drying step. 3) The solvent volume in the boiling flask
has also been reduced from 300 mL to 100 mL. 4) Evaporative solvent losses are reduced by
the shortened extraction time and well-sealed joints. Employee and environmental exposure to
solvent vapor are reduced. Also, initial solvent volume can be reduced, which saves on solventcost
284
-------
Solvent flow
indicated by
arrows —&
Condenser
tj;
Extraction Chamber
Pyrex®
Membrane —fc
Stopcock
t
Boiling Flask/ —^
Concentrator Tube
Snyder Column
u
Figure 1. The Accelerated One Step™ Extractor/Concentrator System.
285
-------
Extraction Method
Most extraction parameters are the same as described in SW-846 Method 3520. The slight
differences were glassware assembly and use, solvent volume, extraction time and acid / base
extraction order. The disposable hydrophobic membrane was sealed in place above the
stopcock. The stopcock was closed and 100 mL of methylene chloride was added to the
extraction chamber. The water sample was poured into the extraction chamber. Surrogate and
matrix spiking were performed normally, as well as pH adjustment. Once the boiling flask was
hot, the stopcock was opened allowing the solvent to run through the membrane into the boiling
flask. Solvent cycled through the extraction system. The solvent boiled off at a rate of 15
mL/min. The stopcock was closed to concentrate the extract after the extraction was complete
in 5.5 hours. The acid fraction of the semi volatile BNAs was extracted first, followed by the
base fraction. Acid and neutral analytes remained in the boiling flask during the basic
extraction. Nitrogen blow-down was performed in the combination boiling flask-concentrator
tube. The pesticide fraction extract was exchanged to hexane after the extraction was complete.
The hydrophobic membrane excludes water during the extraction so sodium sulfate drying was
not required.
METHOD VALIDATION
Three aqueous matrices were spiked at two concentration levels with Appendix IX
Organochlorine pesticide and BNA analytes. Analyte percent recovery (accuracy, bias) and
percent relative standard deviation of recovery (precision) were calculated using the results of
the appropriate analysis techniques. Each high level matrix spike was extracted in triplicate,
while there were 7 to 8 replicates of the low level spike. Method Detection Limits (MDL) and
Reliable Quantitation Limits (RQL) were calculated from the low level data. MDL = standard
deviation x 3.143 (for seven replicates). RQL = MDL x 4.
Ground water, TCLP buffer #1 and waste water were spiked with the analytes in Tables 1 and 2.
These tables contain most of the semivolatile analytes from Appendix IX. The low and high
concentration spike levels are also shown. SW-846 analysis methods 8270 and 8080/8081 were
used as appropriate. Also, several samples with "native" analytes were extracted. The results
were compared with data from the current liquid-liquid extraction method.
Test Matrices
Three aqueous matrices were studied; ground water, TCLP buffer #1 and industrial waste water.
The ground water was taken from a residential drinking water well, which had high
concentrations of calcium, magnesium and iron. TCLP buffer #1 is a sodium acetate - acetic
acid buffer solution used by the Toxicity Characteristic Leaching Procedure to simulate landfill
leaching. Its pH is 4.93 ± 0.05. Industrial waste waters have very diverse matrices. Therefore
no single waste water can be selected which completely represents all of the possibilities.
Nevertheless, a paint stripper effluent waste water sample was selected that represents many
common waste waters.
Matrix Spike Study
The three matrices were spiked with the compounds listed in Tables 1 and 2 just prior to
extraction. The BNA extractions were acidified first to extract the acid-neutral fraction. Next,
the sample was basified and the basic analytes extracted (note: the acid-neutral analytes
remained in the boiling flask). After the basic extraction was completed, the extract was
concentrated to 1 mL by closing the stopcock to perform the macro concentration followed by
N2 blowdown. Pesticide extraction was at a neutral pH as described by the current liquid-liquid
extraction methods. The high level matrix spikes were extracted first followed by the low levels
to test for analyte carryover. No carryover from one extraction to the next was observed.
286
-------
Mechanical Performance
The membrane's resistance to plugging was tested with very fine inorganic solids, dissolved
solids, dissolved organics and settleable organic muck. The membrane did not plug or restrict
proper solvent cycling under any of these conditions. The plugging effect of inorganic solids
was tested by adding 50 g of Kaolin to 1 L of tap water. Dissolved solids effects were tested
with TCLP buffer #1. Dissolved organics were examined with a pond water sample. The
settleable organic muck sample was prepared by taking 50 g of muck from the bottom of a pond
at a local hog farm.
Two types of contaminants have reduced membrane performance. High concentrations of
surfactants and very oily samples may generate enough emulsion to consume the entire 100 mL
of solvent.
VALIDATION STUDY RESULTS
The Accelerated One Step™ extractor/concentrator system showed equivalent accuracy and
better precision than conventional liquid-liquid extraction for semivolatile BNAs and
organochlorine pesticides. The average recovery for all compounds and spike levels was 85%
for pesticides and 69% for BNAs. The average %RSD was 10 and 16 for pesticides and BNAs,
respectively. Calculated MDLs were consistent with published EPA data. Since the chemistry
of the extraction has not changed, extraction efficiency should not change either. In short, those
compounds that extract well with current methodology are "well behaved" with the new system.
Those analytes that were "problem compounds" are still "problem compounds". Precision is
improved because sample extract handling is reduced. Space restrictions limit presentation of
all validation results. The full validation study results are available upon request.
Organochlorine Pesticides
Accuracy and precision were excellent for all ground water and full strength TCLP buffer
extractions. The high level waste water extractions were also good. As expected, the low level
waste water extractions were not as good as the others because of the high levels of non-target
compounds present in the sample. Pesticide recovery and %RSD data for each Appendix DC
compound are listed in Table 3. No data are available for endosulfan n because it coelutes with
other target analytes on both the RTX-5 and RTX-1701 GC columns used in this study. Non-
target compounds in the waste water prevented the analysis of several analytes at the low spike
level.
Figure 2 shows a frequency plot of recovery for the high level spikes and the data published in
method 8080B. The number of compounds with a particular recovery for each matrix is shown.
For example, 13 pesticides from the EPA data had recoveries between 80 and 90%. The shape
of the distribution plots show that overall analyte recovery is the same between the EPA data
using conventional liquid-liquid extraction and the new One Step system. The recovery
frequency plot in Figure 3 clearly shows that low level analyte recovery for ground water and
TCLP buffer compares well with the high level spikes. Recovery from low level waste water
spikes decreased and was more varied. The third replicate of the high level TCLP buffer spike
was invalidated because of a lab error. Thus, the recovery and %RSD for this data subset are
based on 2 replicates.
Figures 4 and 5 show the corresponding frequency distributions for %RSD. The EPA data show
typical %RSDs between 10-20%. Most %RSDs for the Accelerated One Step™ system were
less than 10%.
Table 4 compares overall average recovery and %RSD from this study with the method
proficiency data published in 8080B. All average recoveries were well within the range
specified in 8080B and compared well with the published EPA recoveries. All average %RSDs
287
-------
were lower than both the published EPA %RSD limit and the single analyst %RSD. This is
very encouraging since several extractionists participated in this study.
Pesticide MDL and RQL data are summarized in Tables 5 and 6 and Figure 6. All MDLs from
ground water and TCLP buffer are lower than the published reagent water MDLs from 8081
except for a-chlordane and methoxychlor. The difference for a-chlordane is not significant.
The higher MDL for methoxychlor is probably attributable to the higher spike level for this
compound in the commercially prepared mixtures. As expected, waste water MDLs are
elevated because of non-target compounds. This serves as a reminder that MDLs are matrix
dependent.
Semivolatile - BNAs
Accuracy and precision were excellent for the high level ground water and full strength TCLP
buffer extractions. The high level waste water extractions were also good. As expected,
accuracy and precision decreased in the low level extractions because of the lower spiking level.
BNA recovery and %RSD data for each Appendix IX compound are listed in Table 7.
Figure 7 shows a frequency plot of recovery for the high level spikes and the data published in
method 8270. The number of compounds with a particular recovery for each matrix is shown.
For example, 16 BNAs from the EPA data had recoveries between 80 and 90%. The shape of
the distributions show that overall analyte recovery is the same between the EPA data using
conventional liquid-liquid extraction and the Accelerated One Step™ system. The recovery
frequency plot in Figure 8 shows that analyte recovery for low level spikes decreased and was
more varied though quite acceptable.
Figures 9 and 10 show the corresponding frequency distributions for %RSD. The EPA data
show typical %RSDs between 10-30%. Most %RSDs for the new One Step system were less
than 10% for the high level spikes. Low level spike %RSDs were about 20%.
Table 8 compares overall average recovery and %RSD from this study with the single analyst
method proficiency data published in 8270. Over 95% of the average recoveries were within the
single analyst range specified in 8270 and compared well with the published EPA recoveries.
This is very good since multiple extractionists and GC/MS analysts performed this study. The
average recovery for 2 compounds did not fall within the EPA single analyst limits. The
average high level recovery for 4,6-dinitro-2-methylphenol was 74%, which is within the EPA
guidelines. The low pyrene recovery may be related to an analysis limitation in 8270. The
internal standard for pyrene does not consistently mimic pyrene's chromatographic performance.
The average pyrene recovery of 59% is within the overall performance range of 52-115%. All
average %RSDs (except 4 analytes) were lower than both the published EPA %RSD limit and
the single analyst %RSD. The 4 analytes which exceeded the single analyst %RSD window did
pass the %RSD limit criteria. Those compounds were: 2-chloronaphthalene, 4,6-dinitro-2-
methylphenol, 2,4-dinitrophenol, and pentachlorophenol. This indicates the Accelerated One
Step™ system has very good precision since several extractionists participated in this study.
BNA MDL and RQL data are summarized in Table 9 and Figure 11. All MDLs are consistent
with previous internal MDL studies. No EPA MDL data was available for comparison. As
expected, ground water MDLs were the best, demonstrating once again that MDLs are matrix
dependent.
Extract Cleanup
This new extraction system is chemically equivalent to conventional continuous liquid-liquid
extraction with 2 exceptions. First, extract drying with sodium sulfate is not necessary for
pesticides/PCBs or BNAs because the membrane effectively excludes water from the methylene
chloride extract. Second, full strength TCLP buffers can be extracted without transferring large
288
-------
quantities of acetic acid to the extract that interferes with GC and GC/MS analysis. Thus, the
RQLs are at least one order of magnitude lower than current Toxicity Characteristic regulatory
limits. The membrane does not exclude interferences such as diethylene glycol from the extract
CONCLUSION
Extraction with the Corning Accelerated One Step™ extractor/concentrator system is equivalent
to conventional continuous liquid-liquid extraction while using l/5 the solvent and l/j the
extraction time. The accuracy, as measured by percent recovery, is equivalent to conventional
continuous liquid-liquid extraction. Precision, as measured by %RSD, is better with the
Accelerated One Step™ system. The extracts contain fewer interfering polar compounds, this
improves analysis (particularly for TCLP). The membrane resists plugging for all common
environmental samples. The new system also reduces labor requirements and increases lab
extraction capacity. The Accelerated One Step™ system will contribute to more timely and cost
effective environmental analyses while reducing laboratory use, exposure and disposal of
hazardous solvents.
ACKNOWLEDGMENTS
Many people have contributed to the team effort which made this project a success.
Corning, Inc.: Tim Mitchell, Joe Speciale, Dave Patrick, Mike Oakley, Tim Hertlein, Pete Sisti
and Bob Sell.
Hoover Company: Tony Cassanta, Bob Sack.
Enseco-Erco Laboratory: Nancy Rothman, Peter Kane.
Enseco-Rocky Mountain Analytical: Diane Lowry.
Enseco-Wadsworth/ALERT Laboratories: Betty Winiarski, Dave Counts, Jennifer McCrory,
Jeff Guiler, Tom Hula, Mark Ulman, Don Kirstead, Angel Heist, Velma Carr, Doug Joy, Susie
Dempster, Laura Himes, Rita Tomayko, Chuck Jacobs, Ray Risden, Steve Jackson, Lou
Mancine, Tom Fausnight, Steve Sablar, Todd Benenati, Dorthy Leeson, Craig Hackett, Jim
Horton, Tammy Tokos, Carl Roth, Edward Brunner, Joe Grant, Ralph Byrd, Rick Danford, John
Gruber, Rhonda Kuster, Mike Sekel, Mark Nebiolo, Mike Paessun, Judy Schrock, Chris
Daugherty, Bob Scafate, Brain Haueter, Marv Stephens and Bob George.f
REFERENCES
SW-846, 3rd Edition, Method 3520, Revision 1, Dec., 1987 and Revision 2, Nov., 1990.
EPA = 80SOB data
GW = Ground Water
TC = TCLP Buffer »1
WW = Waste Water
%Recovery
20-i
GW a around Water
TC > TCLP Buffer »1
WW = Waste Water
%Recovery
Figure 2. High Level Pesticide Recovery.
Figure 3. Low Level Pesticide Recovery.
289
-------
30i
ww
TC
GW
6PA = 8080B data
GW = Ground Water
TC = TCLP Buffer «1
WW = Waste Water
50
40
10
%RSD
EPA
Figure 4. High Level Pesticide %RSD.
ww
GW = Ground Water
TC = TCLP Buffer *1
WW = Waste Water
%RSD
Figure 5. Low Level Pesticide %RSD.
20-1
WW
TC
EPA = 8081 data
GW = Ground Water
TC = TCLP Buffer #1
WW = Waste Water
GW
a1 Method
o.o3 Detection
0.01 Limit
EPA
40-|
EPA.6270
-------
Table 1. Organochlorine Pesticides and Spike Concentrations
Compound
Aldrin
a-BHC
P-BHC
5-BHC
Y-BHC
a-Chlordane
y-Chlordane
p-Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate (I & O)
Dieldrin
Endosulfan I
Endosulfan II
Endosulfan sulfate
Endrin
Endrin aldehyde
Endrin Ketone
Heptachlor
Heptachlor epoxide
Isodrin
Kepone
Methoxychlor
Methyl parathion
Parathion
Low (|ig/L)
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.10
0.10
0.10
0.10
1.0
0.10
0.05
0.10
0.10
0.10
0.10
0.10
0.05
0.05
0.05
0.5
0.5
0.05
0.10
High (|ig/L)
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5.0
5.0
5.0
5.0
50.
5.0
2.5
5.0
5.0
5.0
5.0
5.0
2.5
2.5
2.5
25.
25.
2.5
5.0
Surrogates
Decachlorobiphenyl
Dibutylchlorendate
Tetrachloro-m-xylene
0.20
0.10
10
5.0
5.0
291
-------
Table 2. Semivolatile Compounds (BNAs) and Spike Concentrations (|ig/L)
Compound
Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite
Benzo[a]anthracene
Benzofblfluoranthene
Benzofklfluoranthene
Benzo[g,h,i]perylene
Benzofajpyrene
Benzyl alcohol
Bis 2-chloroethoxy)methane
Bis 2-chloroethyl)ether
Bis 2-chloroisopropyl)ether
Bis 2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzvl phthalate
Carbazole
4-Chloroaniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Dimethoate
p-(Dimethylamino)azobenzene
7,12-Dimethylbenz[a]anthracene
Low Hiah*
10 500
10 500
10 500
20 1000
10 500
10 500
10 500
20 ns
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
20 1000
10 500
10 500
3,3'-Dimethylbenzidine (accidentally left out of spike)
a, a-Dimethyl-phenethylamine
2,4-Dimethylphenol
Dimethyl phthalate
1 ,3-Dinitrobenzene
4,6-Dinitro-2-methylphenol
2,4-Dinitrophenol
2.4-Dinitrotoluene
2,6-Dinitrotoluene
Di-n-octyl phthalate
Dinoseb
50 2500
10 500
10 500
20 1000
10 500
10 500
10 500
10 500
10 500
10 500
Diphenylamine (coelution problem, not spiked)
Disulfoton
Ethyl Methacrylate
Ethyl methanesulfonate
Famphur
Fluoranthene
Fluorene
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadiene
Hexachloroethane
10 500
20 1000
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
Compound
Hexachlorophene
Hexachloropropene
lndeno(1 ,2,3-cd)pyrene
Isophorone
Isosafrole
Methapyrilene
3-Methylcholanthrene
Methyl methanesulfonate
2-Methylnaphthalene
2-Methylphenol
Low
50
10
10
10
10
10
10
10
10
10
Hiah*
2500
500
500
500
500
500
500
500
500
500
3-Methylphenol (coelution problem, not spiked)
4-Methylphenol
Naphthalene
1,4-Naphthoquinone
1-Naphthylamine
2-Naphthylamine
2-Nitroaniline
3-Nitroaniline
4-Nitroaniline
Nitrobenzene
2-Nitrophenol
4-Nitrophenol
4-Nitroquinoline 1 -oxide
N-Nitrosodi-n-butylamine
N-Nitrosodiethylamine
N-Nitrosodimethylamine
N-Nitrosodiphenylamine
N-Nitrosodi-n-propylamine
N-Nitrosomethylethylamine
N-Nitrosomorpholine
N-Nitrosopiperidine
N-Nitrosopyrrolidine
5-Nitro-o-toluidine
Pentachlorobenzene
Pentachloroethane
Pentachloronitrobenzene
Pentachlorophenol
Phenacetin
Phenanthrene
Phenol
p-Phenylene diamine
Phorate
2-Picoline
Pronamide
Pyrene
Pyridine
Safrole
1 ,2,4,5-Tetrachlorobenzene
2,3,4,6-Tetrachlorophenol
Tetraethvl dithiopyrophosphate
o-Toluidine
1 ,2,4-Trichlorobenzene
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
O,O,O-Triethyl phosphorothioate
1 ,3,5-Trinitrobenzene
10
10
10
10
10
10
10
10
10
10
10
44
10
10
10
10
10
10
10
10
10
20
10
50
10
10
20
10
10
50
10
10
10
10
10
10
10
10
10
20
10
10
10
10
44
500
500
500
500
500
500
500
500
500
500
500
2200
500
500
500
500
500
500
500
500
500
1000
500
2500
500
500
1000
500
500
2500
500
500
500
500
500
500
500
500
500
1000
500
500
500
500
2200
Surrogates
Nitrobenzene-ds
2-Fluorobiphenyl
p-Terphenyl-d-|4
Phenol-ds
2-Fluorophenol
2,4,6-TribromoDhenol
50
50
50
100
100
100
500
500
500
1000
1000
1000
ns - not spiked
* High level spike concentration for ground water is 40% of the listed value and the high level spike concentration for waste
water is 50% of the listed value.
292
-------
Table 3. Pesticide Accuracy and Precision
Compound
Aldrin
a-BHC
S-BHC
Y-BHC
a-Chlordane
Y-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate II
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Ground Water
High
%R %RSD
97 5
86 7
91 6
90 8
82 7
99 5
98 5
94 6
89 6
92 5
97 6
70 5
74 5
98 5
97 5
91 4
97 5
91 6
93 6
Low
%R %RSD
92 9
85 6
95 7
89 8
83 6
105 7
102 8
101 10
91 7
92 7
101 9
84 7
84 6
100 7
102 7
95 9
100 7
106 7
97 8
TCLP Buffer
High
%R %RSD
90 5
87 11
82 5
89 6
78 6
91 3
90 6
80 3
79 4
83 5
89 4
62 5
59 7
92 7
93 4
91 4
91 4
69 34
85 7
Low
%R %RSD
93 10
92 12
88 11
94 8
80 12
99 5
93 7
105 12
88 4
83 11
94 15
84 12
69 17
92 10
93 9
92 11
96 9
91 12
90 11
Waste Water
High
%R %RSD
78 8
79 7
81 7
69 9
76 8
83 9
86 9
106 7
88 9
79 9
102 10
79 6
88 6
89 8
89 9
79 9
89 10
42 3
81 10
Low
%R %RSD
110 22
interference
65 7
64 22
51 10
74 22
45 14
interference
81 23
53 14
94 30
70 21
44 51
55 11
74 12
41 28
59 14
36 28
60 20
Table 4. Average Pesticide Accuracy and Precision vs EPA Data
Compound
Aldrin
a-BHC
P-BHC
S-BHC
y-BHC
a-Chlordane
Y-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate II
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Accelerated One-Step™
Extractor/Concentrator
Average Average
%R %RSD
93 10
86 9
84 7
83 10
75 8
92 9
86 8
97 8
86 9
80 9
96 12
75 9
70 15
88 8
91 8
82 11
89 8
73 15
85 10
8080B
%R %R
Range
81 42-122
84 37-134
81 17-147
81 19-140
82 32-127
82 45-119
82 45-119
84 31-141
85 30-145
93 25-160
90 36-146
97 45-153
89 26-144
89 30-147
85
%RSD Limit %RSD Single
Analyst
21 16
24 13
32 22
36 18
23 12
20 13
20 13
28 20
28 13
36 17
38 12
25 10
27 13
37 20
29 16
293
-------
Table 5. Pesticide Method Detection Limits
Compound
Aldrin
a-BHC
P-BHC
5-BHC
Y-BHC
a-Chlordane
y-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate H
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Ground Water
Low Level
MDL RQL
0.012 0.049
0.008 0.031
0.010 0.039
0.011 0.044
0.008 0.032
0.012 0.047
0.013 0.052
0.032 0.127
0.018 0.074
0.020 0.079
0.026 0.105
0.169 0.678
0.159 0.637
0.020 0.080
0.010 0.042
0.025 0.098
0.021 0.085
0.024 0.094
0.038 0.152
TCLP Buffer
Low Level
MDL RQL
0.014 0.057
0.016 0.066
0.014 0.057
0.012 0.046
0.014 0.057
0.007 0.029
0.010 0.039
0.038 0.152
0.010 0.041
0.027 0.107
0.044 0.174
0.301 1.206
0.343 1.373
0.028 0.112
0.013 0.050
0.030 0.120
0.027 0.110
0.032 0.130
0.053 0.213
Waste Water
Low Level
MDL RQL
0.036 0.143
interference
0.007 0.027
0.022 0.088
0.008 0.032
0.026 0.106
0.010 0.040
interference
0.056 0.225
0.022 0.089
0.084 0.337
0.435 1.739
0.676 2.703
0.018 0.071
0.015 0.060
0.036 0.142
0.027 0.110
0.027 0.109
0.083 0.332
Tabie 6. Average Pesticide Method Detection Limits vs. EPA Data
Compound
Aldrin
a-BHC
b-BHC
g-BHC
d-BHC
a-Chlordane
g-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate H
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Accelerated One-Step™
Extractor/Concentrator
Average Average
MDL RQL
0.021
0.012
0.010
0.010
0.015
0.015
0.011
0.035
0.028
0.023
0.051
0.302
0.393
0.022
0.013
0.030
0.025
0.028
0.055
0.083
0.048
0.041
0.040
0.059
0.061
0.044
0.140
0.113
0.091
0.205
1.207
1.571
0.088
0.051
0.120
0.101
0.111
0.221
8081
MDL
0.034
0.035
0.023
0.025
0.024
0.008
0.037
0.050
0.058
0.081
0.044
0.030
0.035
0.039
0.050
0.043
294
-------
Table 7. BNA Accuracy and Precision
Compound
Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite 1
Aramite 2
B enzo [ a] anthracene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[g,h,i]perylene
Benzo[a]pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzyl phthalate
Carbazole
4-Chloro aniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate 1
Diallate 2
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichloro benzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Average of all BNAs
Ground Water
High
%R %RSD
88 9
87 8
97 6
95 11
75 15
69 7
87 7
not spiked
not spiked
88 7
108 5
97 7
81 5
86 7
83 8
108 9
78 7
69 8
89 7
104 10
91 5
35 8
49 9
84 4
77 7
88 10
70 7
102 8
93 9
113 12
114 8
83 7
97 8
100 7.
69 4
67 4
69 4
109 19
74 11
101 10
100 6
86 9
Low
%R %RSD
70 7
60 6
67 9
36 12
41 30
72 11
57 6
92 19
70 26
57 9
45 14
60 8
35 13
33 8
58 13
91 7
72 9
86 15
23 17
77 14
42 20
25 10
62 9
40 19
55 7
97 6
45 8
90 10
71 10
72 16
61 9
29 16
78 6
60 21
56 13
55 13
60 9
25 33
45 12
49 20
72 13
55 17
TCLP Buffer
High
%R %RSD
91 4
89 5
81 5
115 3
140 10
92 8
76 4
not spiked
not spiked
80 4
86 7
79 5
77 2
87 5
100 1
113 4
80 5
118 5
94 1
87 4
92 5
43 5
87 8
71 3
110 5
131 5
70 7
104 4
83 4
83 4
78 4
80 4
96 4
97 4
65 9
60 8
65 7
117 2
83 5
93 0.4
106 7
84 8
Low
%R %RSD
74 11
66 8
67 17
30 39
57 28
75 20
61 7
101 17
87 21
62 11
56 13
66 26
48 20
40 19
59 21
94 17
68 25
85 18
156 49
82 6
47 16
28 9
71 19
46 20
55 8
100 13
45 31
93 10
69 16
80 14
66 10
40 22
82 7
81 16
55 24
50 27
56 28
23 43
47 18
55 21
84 12
59 25
Waste Water
High
%R %RSD
69 6
72 5
83 4
102 10
114 16
125 12
60 5
not spiked
not spiked
71 5
66 7
81 4
71 3
74 8
97 12
112 4
91 7
131 6
76 17
63 4
68 14
39 6
92 7
56 11
113 4
108 10
81 7
78 9
73 1
64 7
58 10
70 5
77 11
70 12
68 2
64 4
70 3
94 9
86 8
96 9
87 10
76 10
Low
%R %RSD
54 15
52 15
86 13
71 14
29 21
89 13
41 21
77 22
74 18
37 30
37 30
37 31
21 37
34 29
68 12
93 13
68 17
57 15
111 22
44 16
42 22
35 18
55 21
39 19
59 15
88 14
52 22
56 15
39 30
58 17
53 12
19 45
64 12
87 20
51 19
50 20
52 19
32 7
54 13
77 15
92 11
56 23
295
-------
Table 8. Average BNA Accuracy and Precision vs EPA Data
Compound
Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite 1
Aramite2
Benzo[a]anthracene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[g,h,i]perylene
Benzo[a]pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzyl phthalate
Carbazole
4-Chloroaniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate 1
Diallate 2
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Average of all BNAs
Accelerated One-Step™
Extractor/Concentrator
Average Average
%R %RSD
74 9
71 8
80 9
75 15
76 20
87 12
64 8
90 19
77 22
66 11
66 13
70 14
55 13
59 13
77 11
102 9
76 12
91 11
91 19
76 9
63 14
34 9
69 12
56 13
78 8
102 10
60 14
87 9
71 12
78 12
72 9
53 17
82 8
82 13
61 12
58 13
62 12
67 19
65 11
78 13
90 10
69 16
8270 Rev 2 1990
%R Range
%R
96 60-132
89 54-126
80 43-118
88 42-133
93 42-140
87 25-145
98 D-195
90 32-148
112 49-165
86 43-126
103 63-139
84 29-137
91 65-114
66 D-140
84 41-128
89 65-114
78 36-120
91 38-145
93 44-140
88 D-200
59 8-111
80 49-112
86 17-154
73 37-106
123 8-213
87 53-122
43 D-100
83
%RSD %RSD
Limit Sfrirje Analyst
28 15
40 24
32 21
28 15
39 22
32 19
59 29
39 22
35 16
55 35
46 24
41 26
23 13
23 18
37 23
13 7
29 18
33 20
48 28
70 30
17 13
31 20
42 25
32 24
71 28
26 15
27 28
36 21
296
-------
Table 9. BNA Method Detection Limits
Compound
Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite 1
Aramite 2
B enzo [a] anthracene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[g,h,i]perylene
Benzo[a]pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
B is(2-chloroisopropy 1 )ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzyl phthalate
Carbazole
4-Chloroaniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate 1
Diallate 2
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Average of all BNAs
Ground
Water
Low Level
MDL RQL
1.5 6.1
1.1 4.5
1.8 7.3
2.6 10.5
3.6 14.5
2.3 9.3
0.9 3.8
10.4 41.4
11.1 44.4
1.5 6.1
1.9 7.5
1.4 5.4
1.4 5.7
0.8 3.1
2.3 9.1
2.0 7.9
2.0 8.0
3.9 15.7
1.1 4.5
3.3 13.1
2.5 9.9
0.8 3.0
1.6 6.4
2.3 9.1
1.1 4.5
1.7 6.9
1.0 4.1
2.8 11.0
2.2 8.7
3.4 13.7
1.6 6.3
1.4 5.6
1.3 5.2
3.8 15.1
2.2 9.0
2.2 8.9
1.6 6.5
2.4 9.7
1.6 6.5
3.0 11.8
2.8 11.0
3.3 13.2
TCLP
Buffer
Low Level
MDL RQL
2.5 9.9
1.6 6.2
3.4 13.5
7.0 28.0
4.8 19.2
4.5 17.9
1.2 4.8
10.1 40.6
10.8 43.1
2.1 8.4
2.2 8.7
5.2 20.9
2.9 11.7
2.2 8.9
3.7 14.7
4.9 19.4
5.0 20.1
4.5 18.2
22.9 91.8
1.4 5.8
2.2 8.7
0.8 3.0
4.1 16.5
2.8 11.1
1.4 5.6
3.8 15.1
4.1 16.6
2.9 11.7
3.2 12.9
3.2 13.0
1.9 7.7
2.7 10.8
1.8 7.1
3.8 15.0
3.9 15.8
4.0 16.0
4.6 18.5
3.0 11.8
2.5 10.1
3.5 14.1
3.0 12.1
5.3 21.3
Waste
Water
Low Level
MDL RQL
2.5 9.9
2.3 9.2
3.3 13.1
6.0 23.9
1.8 7.2
3.3 13.3
2.6 10.2
10.0 40.0
7.8 31.3
3.3 13.3
3.3 13.1
3.5 13.8
2.3 9.3
3.0 11.9
2.4 9.4
3.7 14.7
3.5 14.2
2.6 10.4
7.4 29.6
2.1 8.3
2.8 11.2
1.9 7.7
3.5 14.0
2.3 9.1
2.6 10.5
3.8 15.3
3.4 13.6
2.6 10.5
3.5 14.0
2.9 11.6
1.9 7.4
2.5 10.0
2.3 9.4
5.3 21.3
2.8 11.4
3.1 12.2
2.9 11.7
0.7 2.6
2.1 8.5
3.6 14.3
2.9 11.7
4.1 16.3
Overall
Average
MDL RQL
2.2 8.6
1.7 6.6
2.8 11.3
5.2 20.8
3.4 13.6
3.4 13.5
1.6 6.3
10.2 40.7
9.9 39.6
2.3 9.3
2.4 9.8
3.3 13.4
2.2 8.9
2.0 8.0
2.8 11.1
3.5 14.0
3.5 14.1
3.7 14.8
10.5 42.0
2.3 9.1
2.5 10.0
1.1 4.6
3.1 12.3
2.4 9.8
1.7 6.9
3.1 12.4
2.9 11.4
2.8 11.1
3.0 11.9
3.2 12.8
1.8 7.2
2.2 8.8
1.8 7.2
4.3 17.2
3.0 12.0
3.1 12.4
3.1 12.2
2.0 8.1
2.1 8.4
3.3 13.4
2.9 11.6
4.5 18.0
297
-------
40 MICROSCALE SOLVENT EXTRACTION METHODS
FOR TARRY SOILS FROM FORMER
MANUFACTURED GAS PLANT (MGP) SITES
David Mauro (Vice President), June Schneider (Project Manager), Michael Young
(Senior Chemist), and Barbara Taylor (President), META Environmental, Inc., 49
Clarendon Street, Watertown, Massachusetts 02172.
Ishwar Murarka (Senior Program Manager), Land and Water Quality Studies,
Environment Division, Electric Power Research Institute, Palo Alto, California 94303.
ABSTRACT
This paper discusses two microscale solvent extraction (MSE) methods developed for the
Electric Power Research Institute (EPRI) for use at former Manufactured Gas Plant
(MGP) sites: one method is for the determination of monocyclic aromatic hydrocarbons
(MAHs), particularly benzene, and the other method is for the determination of
polycyclic aromatic hydrocarbons (PAHs).
The MAH MSE method was developed for a soil sampling and analysis project in which
it was important to be able to quickly and accurately determine the concentration of
benzene in soil samples in the field, and to use the soil concentrations to predict TCLP
benzene results. This method consists of a solvent extraction using methylene chloride
(DCM) that is carried out quickly over dry ice using refrigerator-chilled glassware and
tools. The extraction is followed by analysis using gas chromatography with flame
ionization detection (GC/FID).
The PAH MSE method was developed as a rapid method for use in the field to help aid
in the placement of borings or monitoring wells and for screening contaminant levels
during remedial activities. This method is a modification of EPA Methods 3550 and
8100. It consists of solvent extraction with DCM/acetone (1:1) followed by analysis
using GC/FID.
Both methods were compared with standard EPA laboratory methods in different studies.
The results of these comparison studies show that the MSE methods can provide
accurate, rapid, and cost-effective results that are comparable to those of standard
laboratory methods. Although the methods were developed primarily for field use at
former MGP sites, they can be used at most sites where soils are contaminated with
MAHs or PAHs, and can be used either in a field or standard laboratory setting.
298
-------
INTRODUCTION
In recent years there has been an increasing interest in potential environmental or human
health problems resulting from contaminants remaining at former manufactured gas plant
(MGP) sites (Moore, 1989). As a result of this concern, the Electric Power Research
Institute (EPRI) initiated a program to study several aspects of former MGP site
contamination problems. One portion of the research sponsored by EPRI has been the
development of analytical chemistry methods for the determination of the major
constituents in tarry materials derived from MGP processes.
In a recent study (META, 1993), tarry residues from eight former MGP sites across the
country were collected and analyzed. Although the tarry residues from former MGP
processes are known to be complex mixtures of hundreds of compounds (EEI, 1984),
concentrations as high as 328,000 mg/kg (or 32.8% of the total mass of the sample) of
19 polycyclic aromatic hydrocarbons (PAHs) and 25,200 mg/kg (or 2.52% of the mass
of the sample) of 7 monocyclic aromatic hydrocarbons (MAHs) were observed among
the eight samples tested. These high concentrations clearly show the likely abundance
of MAHs and PAHs in MGP-related tars and tar-contaminated soils.
As a result of the significance of the MAH and PAH content in residues from former
MGP sites, META Environmental, Inc. (META) developed, under EPRI sponsorship,
two methods for analysis of tarry MGP residues and soils. One method is for the
determination of MAHs and the other method is for the determination of PAHs in MGP
tars and soils. Both methods use microscale solvent extraction (MSB) methods followed
by analysis using gas chromatography with flame ionization detection (GC/FID), and
both methods are designed for easy adaptability for field use to provide accurate, rapid,
and cost-sensitive analysis of samples during site investigations or remedial activities.
MICROSCALE SOLVENT EXTRACTION (MSE) METHODS
Both of the MSE methods presented here were developed for EPRI for use at former
MGP sites, although the circumstances were different. The PAH MSE method was
developed first, for use in rapidly characterizing subsurface soils during site
characterization or remedial activities at MGP sites. Initially, the PAH results were used
to help define the horizontal and vertical extent of contamination in subsurface soils, as
well as to determine the most accurate placement of groundwater monitoring wells. The
PAH MSE method has gone through several revisions in the years since it was first
developed, and has been successfully used at approximately a dozen MGP sites across
the country.
The MAH MSE method was developed more recently, primarily for the accurate
determination of benzene in the field. Because benzene is a known constituent of MGP
tarry residues and is included among the Toxicity Characteristics Leaching Procedure
(TCLP) constituents, high concentrations of benzene in soils or residues will cause them
to be classified as characteristically hazardous when the TCLP benzene concentration is
299
-------
greater than 0.5 mg/L. If a relationship between the total benzene level and the TCLP
benzene level in MGP samples can be established, then monitoring for total benzene
levels in the field can be used to determine the probability that a given sample will fail
the TCLP test for benzene. Being able to predict TCLP failure rapidly in the field can
save considerable time and money during a variety of site activities.
Currently, there are several field methods for determining the concentration of benzene
in soil samples, including purge and trap, portable photoionization detector (PID), and
heated headspace methods. However, purge and trap analysis in the field is problematic,
and in addition there has been growing concern that standard purge and trap techniques
may significantly underestimate levels of volatile constituents, such as benzene, in soils
and other solid matrices (Sawhney, 1988 and Steinberg, 1987). Field portable PIDs are
not really appropriate because they monitor total volatile compounds and are not selective
for the compound benzene. Heated headspace has been shown to be significantly less
effective for soils that are high in organic content, such as soils containing MGP residues
(Voice, 1993). Thus, there was a need to develop an alternate method for analysis of
benzene in soils at former MGP sites.
As well as providing rapid turnaround of results either in a field or in-house laboratory
setting, the MSB methods are cost-effective, with savings usually in the range of 20 to
50% of standard laboratory analysis fees. In addition, the MSB methods are more
environmentally sound because of the reduced volumes of solvent used and solid and
solvent waste produced. Furthermore, the smaller sample size necessary for the analysis
significantly increases the options for available sampling techniques.
MAH MSE Method
In the MAH MSE method, all samples, solvents, glassware, and equipment are kept
chilled to minimize the amount of volatilization that occurs during the sample preparation
procedure. The MAH MSE method consists of the following steps:
1. approximately 2 grams of sample are placed in a 15 ml glass centrifuge
tube that is in a holder over dry ice;
2. approximately 2 grams of anhydrous sodium sulfate are added;
3. surrogate standard and 5 ml of methylene chloride (DCM) are added and
the tube is tightly capped;
4. the sample is shaken vigorously for at least 2 minutes, then centrifuged for
5 minutes; and
5. the extract is filtered through glass wool and sodium sulfate.
The extract is analyzed by gas chromatography using flame ionization detection
(GC/FID), and positive results are quantified using the external standard method of
calculation. Extraction efficiency is monitored by the recovery of the surrogate standard
300
-------
compounds fluorobenzene (FB) and difluorotoluene (DFT). The detection limit for this
method is 0.250 mg/kg (ppm).
PAH MSE Method
The PAH MSE method is a modification of EPA Methods 3550 and 8100 and consists
of the following steps:
1. approximately 2 grams of sample are placed into a 20 ml glass
scintillation vial along with 2 to 4 grams of anhydrous sodium sulfate;
2. surrogate standard and 10 ml of DCM/acetone (1:1) are added;
3. the sample is disrupted using an ultrasonic microtip probe for 2 minutes;
4. the extract is filtered through glass wool and sodium sulfate into a
Kuderna-Danish (K-D) concentrator tube;
5. the soil sample is extracted twice more by shaking briefly with 2 X 5 ml
of DCM/acetone; and
6. the extracts are combined and the volume reduced to 0.5 ml by K-D.
The final extract is analyzed by GC/FID and positive results are quantified using the
internal standard method of calculation. Extraction efficiency is monitored by the
recovery of the surrogate standard compound 2-fluorobiphenyl (2-FBP). The detection
limit for this method is 0.025 mg/kg.
RESULTS AND DISCUSSION
The results of several studies comparing the MSE methods to standard methods are
presented and discussed in the following paragraphs. It should be noted that graphical
representations of the data are sometimes presented in units of /xg/kg (ppb) rather than
the standard units of mg/kg (ppm) used throughout this report. The use of jig/kg units
in some figures is for convenience and visual simplicity only, and does not affect the
discussion of the results.
MAH MSE Method
The MAH MSE method was compared to standard EPA methods in two studies. In one
small study with 4 MGP soil samples, ranging in total MAH concentration from non-
detect to 1,500 mg/kg, the MAH MSE method showed significantly higher concentrations
for all MAHs than EPA Method 8020 (when concentrations were greater than the MAH
MSE detection limit of 0.250 mg/kg). The results of this small study are presented in
Table 1. As shown in this table, most of the Method 8020 results are 60 to 70% of the
values obtained using the MAH MSE method.
In a second, larger study, 92 MGP soils were collected from ten different sites around
the country. The samples were extracted and analyzed by META's laboratory using the
MAH MSE method, and by a second laboratory using Methods 5030 and 8240 for the
purge and trap extraction of volatile organic compounds followed by GC/MS analysis.
301
-------
MAH MSB results for benzene ranged from non-detect (0.250 mg/kg) to 2,500 mg/kg
while the Method 8240 benzene levels ranged from non-detect (0.005 mg/kg) to 1,100
mg/kg. The results are shown in Figure 1, which is a graphical representation of the
base ten log of the MAH MSB benzene results (in jtg/kg) plotted versus the base ten log
of the Method 8240 benzene results (in /*g/kg). Figure 1 also includes a 1:1 ratio line,
and any points located to the left of this line indicate results where the benzene value
obtained by the MAH MSB method was greater than the result obtained by Method 8240.
As can be seen in this figure, the majority (88%) of the MAH MSB benzene results were
greater than the Method 8240 benzene results. The correlation coefficient of the best fit
line is 0.912, indicating that the results are comparable, and visual inspection of the
graph indicates that the relationship is probably linear except for concentrations near or
below the MAH MSB detection limit. As can also be seen in this figure, the greatest
variability in the data occurs at the left hand side of the graph, when the MAH MSB
benzene results are in the range of 1.0 mg/kg (ppm). This is not surprising since those
measurements are made near the MAH MSB method detection limit of 0.250 mg/kg, and
greater variability near the detection limit is common.
The ratio of the MAH MSB benzene results to the Method 8240 results has an average
value of 34.7, indicating that measurements made using the MAH MSB method are an
average of 35 times greater than measurements made using Method 8240. However, the
standard deviation of the ratios is 103, indicating a large range of ratios, with a few very
high ratios skewing the mean upwards. The median ratio is 3.4, and is a better
indication of the actual difference observed between the methods.
Soil samples extracted by Method 5030 may be prepared in one of two ways, and the
large differences between the analytical results can be accounted for regardless of the
sample preparation technique. Low level soil samples are extracted by weighing 5 grams
of soil into a tube, adding 5 ml of water, and bubbling an inert gas through. Studies
performed by Steinberg (1987) and Sawhney (1988) indicate that low level purge and trap
analysis only removes volatile compounds that are located in the soil pore spaces, and
does not remove compounds that have diffused into the internal micropores of the soil.
Since the MAH MSB method is a solvent extraction technique, it should be able to
penetrate the soil particles and liberate the substances trapped inside, resulting in higher
observed volatile compounds levels. In Method 5030, high level samples are first
extracted by shaking 4 grams of soil with 10 ml of methanol, then spiking 5 ml of water
with an aliquot of the methanol and bubbling inert gas through. Some of the soils
analyzed for this project contained visible quantities of tarry residues, and would have
been prepared using the high level technique. MGP tar residues are not very soluble in
methanol, so the interaction between the volatile compounds and the extraction solvent
is significantly reduced, resulting in impaired extraction efficiency. It has also been
shown that methanol extraction of soils yields highly variable results and poor analytical
precision (Voice, 1993). In contrast, the MAH MSB method utilizes DCM as an
extraction solvent, which is capable of solubilizing a much greater proportion of the tar
302
-------
matrix constituents than methanol, again resulting in an increase in the observed levels
of volatile compounds.
The MAH MSB method for solvent extraction of soils for volatile components may be
problematic in that the sample preparation procedure must be carried out quickly and
carefully, preferably in a cool environment using chilled tools. However, the results of
this study clearly indicate that it is a feasible alternative, and yields data that are a better
representation of the true concentrations of volatile compounds in soil.
PAH MSE Results
The PAH MSE method was compared to standard EPA methods in two different studies.
The first study was conducted as an interlaboratory study, with splits of 73 MGP soil
samples extracted at META's laboratory by the PAH MSE method with analysis by
GC/FID, and at a second laboratory by the EPA Method 3550 for the sonication
extraction of soils with analysis by GC/MS (Method 8270). The total PAH results for
these soils ranged from non-detect to 100,000 mg/kg by Method 8270, and from non-
detect to 110,000 mg/kg by the PAH MSE method. Figure 2 is a graphical comparison
of the total PAH results, with the base ten log of the PAH MSE total PAH results plotted
versus the base ten log of the Method 8270 total PAH results. As can be seen in this
figure, the PAH MSE data compared very well with the Method 8270 data. The
relationship is nearly 1:1 (slope of best fit line = 1.01), indicating that the magnitude
of the individual results from the two methods are very similar. Also, the correlation
coefficient of the data is 0.996, again indicating that the results are very comparable, and
that the relationship is linear.
One of the major concerns regarding the use of microscale modifications of standard
methods has been the issue of sample size and the representativeness of samples
(Kratochvil and Taylor, 1981). It has been speculated that the decrease in sample size
(from 30 grams to 2 grams) may mean that a representative subsample would be difficult
to collect given the observed heterogeneity of soil samples. If this were truly a problem,
then the results of this study would have shown outlying points where non-representative
subsamples were analyzed. A wide range of soil types (from clay to coarse sand and
gravel) and concentration ranges (from visibly clean to visibly tarry) were examined
during this study, so any problems with non-homogeneous sample matrices should have
been readily apparent. The excellent correlation of the data points shown in Figure 2
indicates that such outliers were not present, so representative subsampling was not a
problem in this study.
The second study was also conducted as an interlaboratory study, with splits of 49 MGP
soil samples extracted at META's laboratory by the PAH MSE method with analysis by
GC/FID, and at a second laboratory by the EPA Method 3540 for the Soxhlet extraction
of soils followed by GC/MS analysis (Method 8270). The total PAH results ranged from
non-detect to 1,310 mg/kg by Method 8270, and from non-detect to 3,130 mg/kg by the
303
-------
PAH MSB method. Figure 3 shows a graphical comparison of the total PAH results,
which is similar in format to Figure 2. The results of this study were somewhat different
than those of the first study. Figure 3 includes a 1:1 ratio line, and any points located
to the left of this line indicate results where the total PAH value obtained using the PAH
MSB method was greater than the result obtained using Method 8270. As shown in this
figure, the majority (85 %) of the results are to the left of the 1:1 line, indicating that the
PAH MSB method yielded consistently higher total PAH results than Method 8270.
However, the correlation coefficient of the best fit line is 0.990 indicating that the results
are still comparable, and the relationship appears linear, as it did in the first study.
There are two main differences between the first and second studies which could account
for the different results. First, a different laboratory was used for the Method 8270
analyses in each study, whereas the same laboratory performed the PAH MSB method
for both studies. Second, in the first study the samples were extracted by Method 3550
for the sonication extraction of soils and in the second study the samples were extracted
by Method 3540 for the Soxhlet extraction of soils. Since the PAH MSB method is a
modification of Method 3550, it is very similar to that method. The excellent correlation
between the results in the first study may be a reflection of the similarity of the sample
preparation procedures employed by the two laboratories. In the same way, the
consistent under-recovery of analytes in the second study may be a reflection on the
dissimilarity of the sample preparation procedures, or it may simply be a result of some
kind of operational error by the second laboratory.
Operational factors that could result in a systematic bias in the sample results (i.e.,
consistent over- or under-recovery of analytes) include holding times (if exceeded),
storage conditions (if the refrigerator is too warm), extraction efficiency (as monitored
through the use of surrogate standards), and calibration of the analytical system (if old
or degraded standard solutions are used). Operational factors that could result in non-
systematic bias (i.e., only some sample results are affected) include sample matrix and
heterogeneity (reflected in matrix spike recovery and duplicate RPD results), holding
times (if only some samples exceed them), extraction efficiency (if only some surrogate
recoveries are outside criteria), and improper handling of samples (left at room
temperature, extract volume reduced too much). There was not enough information
available for the second study to allow full evaluation of the source of error, although
the fact that the PAH MSB and Method 8270 results correlate well certainly indicates a
systematic source of bias rather than a non-systematic source.
In an attempt to learn more about the differences between the PAH MSB data and the
Method 8270 data in the second study, the PAH results were summed according to the
number of rings in each compound, then graphed in a fashion similar to the total PAH
results. The graphs for the three to six ring compounds were very similar to the total
PAH graph (Figure 3), and are not presented here. However, the graph for the two ring
compounds was interesting, and is presented as Figure 4. Figure 4 illustrates that the
304
-------
difference between the results for the two ring compounds was even greater than the
difference between the total PAH results. Over 91 % of the values obtained by using the
PAH MSB method were greater than those obtained by using Method 8270.
Furthermore, the average ratio of the MSB to 8270 results was 5.5, indicating that the
MSB results were an average of five times greater than the 8270 results. (Whereas the
average ratio for total PAHs was only 2.3). The two ring compounds are the most
volatile of the PAH compounds, and as such are the most likely to be lost during the
sample preparation procedure. The Soxhlet extraction procedure involves continuous
heated solvent extraction of the soils for 16 to 24 hours. Soxhlet extraction is generally
believed to be a more thorough extraction technique for soils than sonication because of
the length of time involved. However, if not closely controlled it can lead to lower
recoveries of the more volatile components. Volatile components may also be lost during
the concentration procedure if the extract volume is reduced too much. It is probable
that some combination of these factors resulted in the lower PAH results, particularly the
two ring compound results, for the samples analyzed by Method 8270.
It should be noted that the PAH MSB analysis of the samples for the second comparison
study was complicated by the presence of large quantities of interfering compounds. The
peaks for those compounds sometimes coeluted with the target PAH peaks, making
quantitation difficult. At sites containing interfering compounds, it is possible to
introduce a cleanup or fractionation step to the PAH MSB method that would reduce or
remove the analytical interference. Another possibility would be to analyze the PAH
MSB extracts by GC/MS (if available) rather than GC/FID, which would also serve to
remove the analytical interference.
QUALITY CONTROL
Standard Quality Control (QC) measures were employed for the MSB methods in all of
the studies described in this paper. Parameters including surrogate standard recovery,
internal standard performance, matrix spike recovery, and duplicate sample performance
were investigated to evaluate the overall efficiency, precision, and accuracy of the
methods. In addition, instrument detection limits and linearity were monitored through
the analysis of initial and continuing calibration standards. Holding times were
established to preserve sample integrity and representativeness. As many aspects of the
standard EPA CLP Methods QA/QC program were emulated as was feasible for each
MSB method. Some of the QC summary results of the MSB and standard methods are
discussed here.
MSE Methods
Table 2 provides a summary of the QC data for the MSE methods. In this table, average
recovery values for matrix spike samples, and average RPD values for duplicate samples
refer to the average calculated for all compounds in all samples. The range given is for
the maximum and minimum calculated values, considering all individual compound
results in all samples.
305
-------
As can be seen in Table 2, all EPA CLP criteria were met overall, although some
individual measurements did exceed the stated criteria. Some fluctuation in QC results
is normal and expected, even under the best operating conditions. The results in Table
2 depict a method that is in control, and performing satisfactorily.
Standard Methods
Table 3 provides a summary of the QC data for the standard methods. QC results for
MAH MSB comparison study number 1 were not available for review, and are not
included in Table 3. Table 3 includes QC results only for the matrix spike compounds
that were directly comparable to the results of the MSB methods (i.e., MAHs and
PAHs). Similarly, recoveries were evaluated only for the surrogate standards associated
with the MAH and PAH compounds of interest.
As can be seen in Table 3, all EPA CLP criteria were met overall, although some
individual measurements did exceed the stated criteria. The results in Table 3 indicate
that the laboratory analytical system was in control during the analysis of the samples for
these studies.
SUMMARY
Two microscale solvent extraction (MSB) methods for the analysis of MAHs and PAHs
were introduced. Use of the methods is particularly suited to the analysis of soils
containing MGP-related residues, in support of investigatory or remedial activities. Also,
the methods are easily adaptable to a field laboratory setting, and are capable of
providing rapid turnaround of analytical results.
The MSB methods were compared to their EPA GC/MS counterparts, and the
comparisons were favorable. The MSB methods were shown to correlate well with
results obtained using standard methods. Furthermore, it was demonstrated that QC
results from the MSB methods were acceptable under the EPA criteria, and that
acceptable precision and accuracy were possible.
It was suggested that use of the MSB methods would greatly benefit site investigation and
remedial activities at former MGP sites, both in terms of reductions in costs and
generated wastes and the efficiency and accuracy of the sampling/remedial program.
Although the MSB methods have not been tested at other sites, it is speculated that their
use is applicable at any site containing similar types of organic contamination, provided
that significant analytical interferences are not present or are minimized appropriately.
306
-------
Table 1
MAH MSE Comparison Study No. 1
(in mg/kg)
Compound
Benzene
Toluene
Ethylbenzene
Total Xylenes
Total MAHs
Sample 1
MSE
ND1
ND
ND
ND
ND
8020
ND2
ND
ND
ND
ND
Sample 2
MSE
ND
ND
ND
ND
ND
8020
ND
0.002
ND
0.003
0.005
Sample 3
MSE
1.0
ND
0.94
2.8
4.7
8020
0.32
0.19
0.57
1.7
2.8
Sample 4
MSE
62
330
110
1,000
1,500
8020
42
220
77
630
970
The detection limit for the MAH MSE Method is 0.250 mg/kg.
The detection limit for Method 8020 is 0.0013 mg/kg.
307
-------
Figure 1
MAH Method Comparison Study No. 2
Benzene Results (in /yg/kg)
Log (MAH MSE)
8T
234
Log (Method 8240)
Data points
1:1 Line
Figure 2
PAH Method Comparison Study No. 1
Total PAH Results (in mg/kg)
Log (PAH MSE)
5
3
1
i '^^T
i
3
5
Log (Method 8270)
Data points
1:1 Line
308
-------
Figure 3
PAH Method Comparison Study No. 2
Total PAH Results (in mg/kg)
Log (PAH MSE)
4 T
Log (Method 8270)
• Data points
1:
1 Line
Figure 4
PAH Method Comparison Study No. 2
Two Ring Compound Results (in /sg/kg)
Log (PAH MSE)
8
0
234
Log (Method 8270)
Data points
1:1 Line
309
-------
Table 2
Summary of MSE Methods QC Results
Audit
Matrix Spike
%Recovery
Duplicate Sample
RPD1
Surrogate
Standard I2
% Recovery
Surrogate
Standard 23
% Recovery
Average
Range
Average
Range
Average
Range
Average
Range
EPA CLP
Criteria4
M: 66 - 139
P: 35 - 137
50%
M: 59 - 113
P: 43 - 116
59 - 113
MAHMSE
Study No. 1
84%
83-85
9%
6- 11
107%
93 - 131
101%
90- 119
Study No. 2
82%
41 - 105
37%
0- 120
85%
38- 119
82%
41 - 127
PAH MSE
Study No. 1
79%
37 - 120
38%
0.4 - 180
91%
52 - 123
NU
NU
Study No. 2
85%
45 118
34%
0-90
95%
56 - 141
NU
NU
Relative Percent Difference.
Surrogate standard 1 is fluorobenzene for MAH MSE, and 2-fluorobiphenyl for
PAH MSE.
Surrogate standard 2 is difluorotoluene for MAH MSE, and is Not Used (NU) for
PAH MSE.
EPA CLP criteria are taken from SOW OLM01, and are approximated using the
most stringent criteria for compounds that are chemically similar to those
analyzed for by the MSE methods. M and P refer to criteria for the MAH and
PAH MSE methods, respectively.
310
-------
Table 3
Summary of Standard Methods QC Results
Audit
Matrix Spike
Compound 1'
Matrix Spike
Compound 22
Duplicate Sample
RPD
Surrogate Standard
Compound 1'
Surrogate Standard
Compound 24
Ave %Rec
Range
Ave RPD
Range
Ave %Rec
Range
Ave RPD
Range
Average
Range
Ave %Rec
Range
Ave %Rec
Range
EPA CLP
Criteria5
V: 66 - 142
S: 31 - 137
V: 21
S: 19
V: 59 139
S: 35 - 142
V: 21
S: 36
50%
V: 84 138
S: 30 115
V: 59- 113
S: 18- 137
Method 8240
Study No. 2
96%
45 125
5%
2- 14
97%
46- 134
7%
0- 18
32%
0 100
105%
81 141
90%
62-112
Method 8270
Study No. 1
91%
42 - 141
13%
1 -85
145%
0-440
32%
1 - 100
12%
0-67
92%
42 139
119%
38 220
Study No. 2
94%
34 186
24%
0.8 - 83
125%
52 228
34%
1.7 77
10%
0-52
NA
NA
NA
NA
1 Matrix spike compound 1 is benzene for Method 8240, and acenaphthene for
Method 8270.
2 Matrix spike compound 2 is toluene for Method 8240, and pyrene for Method
8270.
3 Surrogate standard 1 is toluene-d8 for Method 8240, and 2-fluorobiphenyl for
Method 8270.
4 Surrogate standard 2 is bromofluorobenzene for Method 8240, and p-terphenyl-
d 14 for Method 8270.
5 EPA CLP criteria are taken from SOW OLM01. V and S refer to criteria for
Methods 8240 and 8270, respectively.
NA Information not available.
311
-------
ACKNOWLEDGEMENTS
The research work reported here was sponsored by the Electric Power Research Institute
under contracts RP2879-01 and RP2879-12, directed by Dr. Ishwar Murarka.
META would like to gratefully acknowledge the support and interest of the utility
company personnel who participated in this research effort, particularly Mr. Jean Pierre
Moreau and Dr. Edward Neuhauser of Niagara Mohawk Power Company. META
would also like to thank Atlantic Environmental Services, Inc. for their assistance.
REFERENCES
Moore, T., "Managing the Gaslight Legacy", EPRI Journal, July/August 1989,
pp. 22-31.
META Environmental, Inc., Chemical and Physical Characteristics of Tar
Samples from Selected Manufactured Gas Plant rMGP) Sites. EPRI Report, in
publication, 1993.
Edison Electric Institute (EEI), "Handbook on Manufactured Gas Plant Sites",
September 1984.
Sawhney, B.L., Pignatello, J.J., Steinberg, S.M., J. Environ. Qual., 1988, Vol.
17, pp. 149-152.
Steinberg, S.M., Pignatello, J.J., Sawhney, B.L., Environ. Sci. Technol., 1987,
Vol. 21, pp. 1201-1208.
Voice, T.C., Kolb, B., Environ. Sci. Technol., 1993, Vol. 27, pp. 709 - 713.
Kratichvil, B., Taylor, J., "Sampling for Chemical Analysis", Analytical
Chemistry, July 1981, Vol. 53, No. 8, pp. 924-938.
312
-------
41 DEVELOPMENT OF AN ENVIRONMENTAL METHOD FOR THE
ANALYSIS OF VOLATILE ORGANIC COMPOUNDS IN SOILS
BY STATIC HEADSPACE GC/MS
Tammv J. Riga. Applications Chemist, Greg O*Neil, Marketing Manager, Tekmar
Company, P.O. Box 429576 Cincinnati, Ohio 45242-9576
There are several methods available for analyzing volatile organic compounds in soils.
The purge and trap technique is widely applied to the analysis of these compounds in
soils. An analytical method is being developed using static headspace as an alternative
approach to purge and trap.
Static headspace is a discontinuous gas extraction where the analytes partition between
the sample and the headspace. Various soil types were analyzed using a GC/Saturn n
Mass Spectrometer and conditions were optimized for separation, precision, detection
limits and linearity. Standards were prepared by diffusing analytes through the soil (i.e.
under zero headspace conditions) in order to simulate plume contaminated soils in the
field. This was compared to the traditional technique of spiking the surface of the soil.
This method also studies the effects of pre-mixing of soils and the addition of organic
modifiers to increase the gas phase concentration within the headspace in order to in-
crease overall method sensitivity. Internal standards were added to a matrix modifying
solution for quantitation.
313
-------
INTRODUCTION
The most problematic concern of soils analysis is how to deal with the varying
characteristics of the soil matrix. EPA methodologies which focus on the analysis of
Volatile Organic Compounds (VOC's) in water have been developed around the Purge
and Trap technique (1). Variables such as moisture content, density, % organic matter,
amount of sand and silt, mineral content, and ionic interactions do not pose a major
concern in water analysis as they do with soils.
Soil characteristics need to be identified because each region of the country will have a
particular type of soil profile. These characteristics depend upon the native geology,
vegetation, the presence of the agriculture, and the qualities of the local climate. Soil
types may be grouped on the basis of their characteristics into two groups: Pedocals and
Pedalfers. Pedocals are distinguished by the accumulation of carbonates, calcium, or
magnesium in all or part of the soil. Pedalfers are distinguished by the absence of
carbonates and usually by an accumulation of iron and aluminum compounds (2). Once
the soils have been assigned to either of these two groups, they are further sub-typed into
hundreds of classes based upon ten distinct characteristics describing the soil and its
horizon (i.e. color, texture, structure, chemical composition including % organic matter,
% clay, % sand, type, and geology, etc.).
All of these characteristics indicate that soils exhibit highly variable matrices. The
reduction of variability in density, mineral content, and ionic interactions are some of the
problems addressed in this paper.
Two types of soils were analyzed in this study: topsoil, and sand. These soil types
represented extremes in % organic matter, mineral content, moisture, and densities. Each
soil type has a different density ranging from 0.5 gm/mL (topsoil) to 3.00 gm/mL (sand).
These densities cause the volume distribution within the headspace vial to vary. A way to
reduce this variability within the vial is to add a Matrix Modifying Solution (MMS). The
MMS is made up of a 100% saturated aqueous salt solution that is pH buffered below 2.
Utilizing a 100% aqueous salt solution eliminates the variability in mineral content and
helps to offset the sodication process. The sodication process is a natural occurrence in
areas of the country where precipitation is less than 20 inches annually. As the water
evaporates in the atmosphere, the salts are left behind in the soil. The process of
increasing the sodium saturation of the soils exchange process is called sodication. The
hydrolysis of the Na+ ions or Na2C03 compounds can cause a strong alkaline reaction
changing the pH of the soil also (3). With the addition of salt saturated MMS each sample
has the same Na* concentration prior to analysis. The MMS also increases the ionic
strength of the sample which leads to increased sensitivity in the headspace analysis.
Addition of acidified MMS to soil samples aids in the maintenance of constant pH
314
-------
conditions prior to analysis (resistance to strong alkaline reactions such as those
occurring in the sodication process). A low pH also helps to prevent
dehydrohalogenation of compounds and serves as a preservative against microbial
activity in the sample which could ultimately lead to degradation of analytes.
EXPERIMENTAL
The two types of soils analyzed in this study consisted of topsoil (very high organic
content) and sand (very low organic content). Each of these soil types has a different
density ranging from 0.5 gm/mL (topsoil) to 3.00 gm/mL (sand). These densities cause
the volume distribution within the headspace vial to vary. An average 2g sample of each
soil type used in this study occupied 4.0 mL (topsoil) and 0.67 mL (sand) volume within
the vial. The resulting phase ratios (V/Vm) produced by a 2g sample of each soil type
within a 22 mL vial was 4.5 (topsoil) and 31.8 (sand). This large difference in phase ratio
produces significant differences in headspace measurements (4).
To reduce this volume and phase ratio variability, 10 mL of matrix modifying solution
(MMS) was added to each 2g sample for a final volume of 14 mL (topsoil) and 10.67 mL
(sand). The resulting phase ratio differences were significantly reduced (0.57 for topsoil
and 1.06 for sand). The MMS is made up in 500 mL of reagent grade water. The solution
is then pH buffered below 2 with Phosphoric Acid (H3PO4) and saturated to 100% with
170g Sodium Chloride (NaCl).
The analytes chosen for this study were the 502 A/B standards purchased from
AccuStandard. A key factor in developing this method was finding an appropriate
internal standard/surrogate mixture so that the analytes could be tracked throughout the
entire chromatogram. The molecular weights, boiling points, and relative reaction
chemistries (polarities) were studied for each of the 502 A/B analytes. The Method 8260
mix was selected because it contained four internal standards and three surrogates with
similar properties of the 502 A/B analytes. Each internal standard was responsible for
tracking approximately 13 compounds in each section of the chromatogram (see Table 1
for listing).
Infernal Standards RT Surrogates RT
Pentafluorobenzene 17:50 Dibromofluoromethane 16:30
1,4-Difluorobenzene 19:53 Toluene-d8 24:19
Chlorobenzene-d5 27:08 4-Bromofluorobenzene 29:16
l,4-Dichlorobenzene-d4 31:37
Table 1. The relative retention times for the Method 8260 internal standards and
surrogates. Each internal standard is responsible for tracking approximately 13 analytes.
315
-------
Once the analyte list has been established, the next step in the development process is to
select the proper detector and column for the analysis. Due to the volatility of the
analytes, the detector should be very sensitive and the column should possess relatively
high retention characteristics to give good chromatographic separation. The following
GC conditions and 7000 Headspace Autosampler parameters used in this analysis are
listed below.
Tekmar 7000 Headspace Autosampler
Platen temperature:
Sample equilibration time:
Loop Size:
Vial size:
Pressurize time:
Pressure equilibration time:
Loop fill time:
Loop equilibration time:
Mix:
Mix power:
Injection time:
Line temperature:
Loop temperature:
Transfer line:
Transfer line back pressure:
Vial pressurization setting:
Vial needle flow rate:
85°C
10 min. (FET mode)
50 min. (samples)
250 uL (EFT)
22 mL
0.20 min.
0.08 min.
0.10 min.
0.08 min.
10.00 min.
8
1.00 min.
85°C
85°C
0.32 mm fused silica
15 psig
15 psig
40 mL/min.
Varian 3400 Gas Chromatograph/Saturn II Ion Trap Detector
Injector temperature:
Manifold temperature:
Transfer line temperature:
Oven temperature program:
Initial temperature:
Ramp 1 temperature:
Ramp 2 temperature:
Final temperature:
Electron multiplier:
Target value:
Scan range:
200°C
265°C
220°C
10°Chold7min.
4°C/min. to 100°C
10°C/min. to 200°C
200°C hold 4 min.
1400 mV
17,000
48-260 amu
316
-------
J&W DB-VRX (J&W Scientific, Folsum, California)
60 M x 0.32 mm ID x 1.8 udf
Carrier gas: Helium @ 1.0 mL/min.
Phase ratio: B=44 (High K= High retention)
The transfer line from the 7000 Headspace Autosampler was connected to the analytical
column via a zero dead volume union.
RESULTS and DISCUSSION
I. Gas Sampling and Injection (PET}
Full Evaporation Technique (FET) is a process by which an aliquot of methanolic
standard containing the volatile analytes of interest (usually 1-20 uL) is taken and placed
into a headspace vial. The vial is then sealed and the analytes are fully vaporized. This
technique is used to establish relative retention times and chromatographic separation.
Figure 1 is an example of chromatographic separation from an FET injection. All
compounds exhibit good chromatographic separation and response. The reproducibility
study showed that 95% of the analytes fell below 5% RSD.
195,000-
Figure 1: Chromatographic Separation from an FET injection
317
-------
II. Static Partitioning/Mass Distribution- 2 Phase Equilibration
Once the retention times and chromatography were established, the reproducibility of a
two phase gas/liquid headspace system was evaluated. For this experiment, 10 mL of the
MMS containing an analyte concentration of 100 ppb and 50 ppb internal standards/
surrogates was added to a headspace vial. The analytes in solution were equilibrated at
85°C for 50 minutes. This standard represents a baseline for 100% recovery and serves
as a comparison for soils analysis. It is a reference point to establish how much com-
pound is being absorbed when a soil matrix is analyzed. The chromatogram in Figure 2
shows good response and chromatographic resolution. Four analytes and two surrogates
were selected from each section of the chromatogram for reproducibility studies. Table 2
shows the %RSD's for 10 mL MMS based upon six replicate runs.
220,000
... I II ll
Figure 2: 10 mL MMS Spiked with 100 ppb 502 A/B Standard (IStd = 50 ppb)
Trichloroflouromethane
1,2,2-Trichloroethane
o-Xylene
1,2-Dichlorobenzene
Dibromoflouromethane (surr)
Toleuene-dg(surr)
RSD
2.4%
6.1
2.8
2.0
3.0
0.5
Table 2. Reproducibility of a two phase gas/liquid headspace system (10 mL MMS)
318
-------
HI. Static Partitioning/Mass Distribution- 3 Phase Equilibration
A. Spiked sand vs. Diffused sand
The next experiment looked at the addition of 2 g of soil to the headspace vials for analy-
sis. The addition of soil creates a three phase equilibrium system (solid, liquid, gas). The
analytes must migrate from the soil into the MMS, and then partition between the MMS
and the headspace. The first sample chosen for analysis was 2 g of sand. This represented
the best case scenario because it has low adsorptivity and very low organic content.
These characteristics indicate that sandy soils exhibit low matrix interferences.
Figure 3 illustrates an example of a 2 g spiked sand sample versus a 2 g diffused sand
sample. Spiked sand represents a technique where the analyst spikes the analytes, internal
standards, and surrogates directly into the sand. The 10 mL of MMS is then added to the
sand and the vial capped. In a diffused sand sample, the analytes, internal standards, and
surrogates are added directly to the MMS. The MMS (10 mL) is added to the 2 g soil
sample and mixed on a rotator for one hour. The purpose of the diffused technique is to
simulate real world plume contaminated samples.
75,000 I
509,000
Figure 3: Spiked Sand technique VS. Diffused Sand technique
319
-------
Table 3 shows the area counts and %RSD's comparing the two types of samples. The
same four analytes and surrogates selected in the above study were used as reference
points again. The area counts of the diffused sand sample are between six to twenty-two
times greater than the spiked sand. This is because the analytes in the spiked sand have
not been evenly diffused throughout the sample. The reproducibility over the average of
six replicate runs indicated that 87% of the analytes had RSD's less than 10% for both
samples.
Spiked Sand Diffused Sand
Area RSD Area RSD
Tricholoroflouromethane 15715 11.5 147974 2.5
1,1,2-Trichloroethane 7324 8.2 160146 2.0
o-Xylene 37728 3.6 236264 2.5
1,2-Dichlorobenzene 10905 5.0 82523 8.5
Dibromoflouromethane (surr) 5352 4.8 45694 5.1
Toluene-d- 17048 2.1 385076 2.2
5
Table 3. Area counts and % RSD's comparing spiked VS. diffused technique (sand)
B. Spiked topsoil vs. Diffused topsoil
The next sample consisted of 2 g of topsoil. This represented the worst case scenario
because it has high adsorptivity and very high organic content. These characteristics
indicate that topsoils exhibit a high degree of matrix interferences.
Figure 4 illustrates and example of a 2 g spiked topsoil versus a 2 g diffused topsoil. The
technique for spiking and diffusing the analytes into the soil was the same as that used in
the above sand study. When the analytes are spiked directly into the soil, the high organic
content of topsoil results in localized adsorption. The localized spike has a limited
surface area for analyte exchange and migration from the soil into the MMS. In a field
sample, where contaminants have "plumed" through the soil, the surface area for analyte
migration is much greater in comparison to a spiked sample.
320
-------
10,000
70,000
Spiked Topsoil
JJJL
Diffused Topsoil
Illl..
Figure 4: Spiked Topsoil technique VS. Diffused Topsoil technique
Table 4 shows the area counts and % RSD's for the comparison of the two types of
samples. The same four analytes and surrogates selected in the above study were used as
reference points again. The area counts of the diffused topsoil sample are once again
greater than that of the spiked topsoil. The reproducibility over the average of six repli-
cate runs indicated that 99% of the analytes in the diffused topsoil had RSD's less than
20%. Only 68% of the analytes in the spiked topsoil had RSD's less than 20%.
Spiked Topsoil
Area RSD
Diffused Topsoil
Area RSD
Tricholoroflouromethane
1,1,2-Trichloroe thane
o-Xylene
1,2-Dichlorobenzene
Dibromoflouromethane (surr)
Toluene-dg
Table 4. Area counts and % RSD's comparing spiked VS. diffused technique (topsoil)
22704
1896
4796
1112
2355
4025
14.6
9.3
35.1
16.9
3.6
3.1
145993
15998
33956
8058
15765
23333
7.5
5.0
16.2
3.4
7.9
4.8
321
-------
IV. Effects of Increasing Organic Content
Soil samples that contain a very high organic content tend to very adsorptive. The
analytes tend to adhere to organic sites in the soil which reduces the release of contami-
nants. The addition of the MMS helps to reduce some of this variability because the
analytes are able to migrate more rapidly from the soil to a liquid solvent (MMS).
Figure 5 is the resultant chromatogram illustrating a diffused sand sample to a diffused
topsoil sample.
509,000
70,000
Figure 5: Effects of increasing organic content.
Table 5 shows the difference in peak area counts when comparing soils that have varying
amounts of organic content. In both soil samples, Trichlorofluoromethane (gas) exhibits
Topsoil Sand
Area Area
Tricholoroflouromethane
1,1,2-Trichloroethane
o-Xylene
1,2-Dichlorobenzene
Dibromoflouromethane (SUIT)
Toluene-dg
145993
15998
33956
8058
15765
23333
147974
160146
236264
82523
45694
385076
Table 5. Area counts illustrating the effect of increasing organic content
322
-------
very similar area counts. This is because Trichlorofluoromethane is a relatively volatile
compound with a high diffusion coefficient. The other five compounds in table 5 show
between a seven to sixteen fold decrease in area counts with the topsoil sample. The
topsoil tends to hold onto the analytes, especially those heavier compounds found in the
middle and very end of the chromatogram.
V. Diffused Topsoil- 14 Day Time Study
Many EPA laboratories today have a period of up to 14 days before the sample has to be
run. The purpose of this next study was to examine the relative degradation of analytes
over a period of 14 days. Three replicates of samples for each day were prepared using 2g
of topsoil and 10 mL of MMS. The MMS was spiked with an analyte concentration of
100 ppb (internal standards and surrogates were at 50 ppb) and the samples were all
mixed for one hour to simulate the diffusion technique. Each day, for 14 days, three
replicates were pulled from a 4°C refrigerator and run by headspace.
Figure 6 is a graph of the area counts from early eluting compounds vs time
(Pentafluorobenzene as the internal standard). Four additional compounds were selected
from the front end. The graph shows that over a period of 14 days, there is some degrada-
tion of the analytes. The important point to notice is that the internal standard degraded at
the same rate as all the other analytes. This indicated that the internal standard was
properly tracking the analytes of interest. As long as the internal standard falls at the
same rate as the analytes, quantitive results may be expected over the time period tested.
Pentaflourobenzene (IS)
Chloromethane
1,1-Dichloroethene
1,2-Dlchloroethane
Figure 6: Area Counts of early eluting compounds VS. time
323
-------
Figure 7 is a graph of the area counts of late eluting compounds vs time (1,4-Dichloro-
benzene-d4 as the internal standard). Once again, all analytes are properly being tracked
and degrading at the same rate as the internal standard.
— 1,4-Dichlorobenzene-d4 (IS)
tert-Butylbenzene
p-lsopropyltoluene
1,2-Dlchlorobenzene
Naphthalene
Figure 7: Area counts of late eluting compounds VS. time
VI. Carryover
A common problem associated with soil analysis today is carryover from very highly
contaminated samples. Samples in the high ppm range can shut down an instrument and
halt sample processing due to high level contamination and carryover. The major advan-
tage of headspace analysis over purge and trap is that there is no carryover or contamina-
tion problems associated with this technique. The relative carryover percentage is less
than 0.2% for all analytes run after a 10 ppm sample. The Saturn Mass Spectrometer and
the column also showed no major effects of contamination.
CONCLUSION
The soils analysis method by static headspace is very clean, sensitive, and reproducible. It
also eliminates many variables found in current soil methodologies. Future work will
focus on practical field sampling techniques. One example of this is to use disposable
syringes that will deliver a volume of soil to a screw top headspace vial. The MMS with
the internal standards and surrogates will automatically be added to the sample in the
field and the vial sealed. This mode of field sampling is simple, less prone to error, and
eliminates losses associated with sample transfer prior to analysis.
324
-------
Another focus will be on the addition of a non-volatile organic modifier to the MMS in
order to increase the sensitivity of compounds analyzed from highly organic matrices.
Final detection limits and dynamic ranges will then follow.
REFERENCES
1. U.S. Environmental Protection Agency. Methods for the Determination of Organic
Compounds in Drinking Water, EPA-600/4-88-039; 1988.
2. Finkl, Charles W. Jr., 1982, Soil Classification, Benchmark Papers in Soil Science/1,
pp. 145-168
3. Tan, Kim H., 1982, Principles of Soil Chemistry, Dekker, New York, pp. 187-189.
4. B.V. loffe, A.G. Vitenberg, "Head-space Analysis and Related Methods in Gas Chro-
matography" 1984 John Wiley and Sons, Inc., pp. 30-37
325
-------
42 IN-SITU DERIVATIZATION OF SOIL AND SEDIMENT SAMPLES IN A
SUPERCRITICAL FLUID EXTRACTION CELL; THE EXTRACTION OF
PHENOLS AND CHLORINATED PHENOLS
DENNIS GERE. LENORE G. RANDALL, CHARLES KNIPE, WILLIAM PIPKIN
Hewlett-Packard, Little Falls Site, 2850 Centerville Rd., Wilmington, DE 19808
H. B. LEE, Environment Canada, P.O. Box 5050, Burlington, Ontario, Canada LTR 4A6
ABSTRACT
This manuscript will describe a series of experiments which allow the simultaneous derivatization of
analytes in solid and semi-solid samples containing phenolic compounds including the more intractable
chlorinated species. The derivatizing reagents acetic anhydride and triethyl amine were added by
admixing the separate reagents to the extraction cell before the cell was placed in the SFE apparatus.
Further experiments with respect to the reaction and extraction efficiency for methylphenols and
nitrophenols as well as the chlorophenols and permutations of mixed moieties is in progress, collectively
in the Little Falls and National Water Research Institute, Environment Canada labs. Preliminary results
suggest that dinitrophenols do not react under the conditions described previously, although
mononitrophenols appear to react as phenol and chlorinated phenols do. This work is continuing and will
be described in a later publication.
INTRODUCTION
The supercritical fluid extraction (SFE) of phenols and chlorinated phenols from soils and sludges is an
analyte/matrix pair receiving significant attention with respect to getting an acceptable, robust SFE method
developed and ultimately formalized as a registered EPA method. The group of H.B. Lee (1) has
published a method which employs in situ derivatization and post-SFE clean-up of the extract to remove
the unreacted derivatizing agent.
Experimental
SFE Conditions:
Hewlett-Packard Models 7680T and 7680A were used without modification.
SFE Method The following conditions are grouped according to function.
Modifiers/reagents added to the thimble; 100 uL acetic anhydride, 100 uL triethyl amine (sample size is
usually 1 gram).
1. Extraction — pressure, 5250 psi; extraction chamber temperature, 80 C; density, 0.80 g/mL;
extraction fluid composition, CO2; static equilibration time, 5 minutes; dynamic extraction time, 10
minutes; extraction fluid flow rate, 2.0 mL/min; resultant thimble-volumes-swept, 2.
2. Collection (during Extraction) - trap packing, Hypersil ODS; trap temperature, 15 C; nozzle (variable
restrictor) temperature, 45 C.
3. Reconstitution (of collected extracts) - rinse solvent, 50/50 (v/v) isooctane; collected fraction volume,
1.2 mL; trap temperature, 40 C; nozzle (variable restrictor) temperature, 45 C; rinse solvent flow rate,
1.0 mL/min; fraction destination, vial #n.
326
-------
GC Conditions:
Carrier gas = hydrogen, 15 psi head pressure .splitless injection, injection volume = 1.0 uL; column =
HP 5 fused silica capillary 0.200 mm x 25 meters; column temperature program , initial temperature
= 70 °C, initial program rate = 30 °C/mm (to 120°C) 2nd program rate = 2°C/min, final temperature
= 200 °C.
Results and Discussion
This is a study (FIGURE 1) which was started by the lab in Canada, and is being continued towards some
interesting possibilities. It was stated earlier that phenols and substituted phenols are reasonably polar
compounds, and are particularly difficult to extract out of solid matrices. Another problem that causes
them to be more intractable, or more difficult to extract and analyze, is that they often occur in solid
wastes in ratios of high-low concentrations. This is seen in a chromatogram in a later diagram (FIGURE
5). The phenols are usually distributed in a broad range of phenolic constituents—chloro, nitro, methyl
and every mix that might be possible.
There arc at least three possibilities to the extraction of phenolic compounds. From solid waste, one
approach would be to use carbon dioxide by itself. Carbon dioxide is a relatively non-polar solvent, and
for these relatively polar analyses, does not yield particularly good results in real samples. Another
approach might be to add a bquid to the CO2 stream-a modifier or additive. The approach which was
done by Bill Lee and his co-workers was actually to do an in-situ derivatization in the extraction vessel
and then extract the derivatized materials. We also will take a brief look at supercritical fluid
chromatography to suggust a third interesting approach for the extraction.
FIGURE 1. Outline of SFE extraction Study
Phenols and substituted phenols are polar
phenols occur in high/low concentration ratios
broad range of substituents
* chloro
* methyl
* nitro
* all permutations of above substituents mixed
3 Possible avenues of structuring the SFE method
CO ? only
in-situ denvauzation with CC>2 only
CO;> + modifier + additive ( as in packed column SFC)
Hard look al Canadian method and any limitations
Automation
PrepStation to automate manual cleanup steps
SFE- GC Bridge to speed throughput and minimize operator error
327
-------
Loading the Thimble for In-satu Derivatizaticsi
30 id Acetic Anhydride
50 li Standards
30 ul TrietMamine
FQter paper disk
200mgCelite
1 gram of solid sample
mntfomng Phenols
200mgCelite
FIGURE 2. Loading the extraction vessel with sample and reagents
328
-------
There are fewer manual steps as you automate, therefore, less hands-on time per method, and time is very
costJy in environmental laboratories. Improved accuracy, precision and the overall method become less
dependent upon the skill level and the particular approach of a given technician or operator.
Derivatization coupled with analysis allows just-in-time delivery of each sample for analysis when the
analytical device is finished with the last sample and is ready for the next one. Automated optimization of
reaction conditions will allow some interesting possibilities for unattended execution of designed
experiments. Samples in report out in a very timely manner, and unattended, you can set a set of eight
samples into the extractor and come back in the morning and find your results unattended.
Figure 2 illustrates the basic approach that's used in the in-situ derivatization. This is a schematic drawing
of the extraction vessel . The sample, reagents and other materials are placed in the extraction vessel
starting from the bottom. The end-piece is attached to the metal thimble first, then a filter paper disk is
placed on the face of the end-piece frit. Then, 200 mg, of an inert material such as celite is placed on top
of the filter paper disk, followed by a 1 gram sample of solid waste or sludge. A liquid aliquot
( 30 uL) of triethyl amine (TEA) is added to the sample via a micro syringe. This serves to raise the
effective pH to an optimum level for the reaction that will occur between the phenols and the acetic
anhydride. Another layer consisting of 200 mg of celite is then placed on top of the sample. Another Iquid
aliquot ( 30 uL) of acetic anhydride is metered onto the celite with a micro syringe. This acetic anhydride
is the derivatizing reagent to form acetyl derivatives of the various phenols. Finally, another disk of filter
paper is placed on top of this. The exit end-piece (fritted cap) is placed on the metal extraction vessel and
the entire sample vessel is placed in the carousel of the SFE.
The reconstitution solution resulting from the SFE step is removed from the extractor and a clean-up
procedure is carried out. This procedure is done manually (separate from the SFE equipment). Work is
underway to incorporate this into automatic PrepStation SPE manipulations to facilitate automating clean
up these steps. The reconstituted extract solution ( derivatized phenol acetates dissolved in hexane) is
mixed with 3 ml of a one percent solution of potassium carbonate and is agitated for one minute. That
potassium carbonate neutralizes the excess TEA reagent, which was added to obtain an effective optimum
pH for the derivatization reaction in the supercritical CO2. The extract solution is then passed through a
small bed of magnesium sulfate (this can be packed into a Pasteur pipet) to remove excess water. That
resulting filtrate is then passed through another short bed of 5% deactivated silica.The analytes are
eluted in a 10 ml solution, consisting of a hlmixture of petroleum ether/methylene chloride. This
solution is evaporated and re-exchanged to 5 ml of isooctane. A luL aliquot is injected into the gas
chromatograph. The SFE method used a temperature of 80 degrees at a constant density of 0.8 g/ml. with
a 5 minute static period for the derivatization and soaking of the sample. This was followed by a 10
minute period of dynamic extraction at a volume flow rate ( measured as liquid carbon dioxide at 4
degrees centigrade) of 2 ml per minute.
Results from Lee (1) and co-workers with a commercially available reference soil 130-100 , comparing
steam distillation to supercritical fluid extraction in-situ derivatization are seen in Figure 3.
FIGURE 3. Recovery and Precision from Reference Soil 130-100 ( Ref. 1)
Steam Distillation SFE Steam Distillation SFE
Recovery mg/kgm Recovery mg/kgm Precision mg/kgm Precision mg/kgm
2,3,5 Tri Cl 0.40 0.36 0.01 0.01
2,3,4,5 Tetrad 14.4 13.9 0.40 0.30
2,3,4,6 Tetrad 20.6 20.2 0.40 0.30
2,3,4,5 Tetrad 1.9 1.8 0.10 0.10
Pentachiorophenol 1,499.0 1,483.0 67.0 93.0
The recoveries for the supercritical fluid extraction correspond very well with the steam distillation for all
of the compounds. Steam distillation requires many additional manual manipulations—most are quite
329
-------
operator- dependent. In comparison, the SFE equipment used in this study is operated in the same
manner by all operators, because all steps are computer controlled and recalled from a stored method.
This table ( Figure 3.) shows the results of a representative phenol containing sample. This particular
sample is from a site in Canada adjacent to stored railroad ties (treated with creosote to promote long
lifetime of the wood). It is normally expected mat there will be a large concentration of many chlorinated
phenols. In addition, as mentioned above, there was a very small concentration of trichloro phenol and a
very high concentration of the pentachlorophenol.
In Figure 4, we see the chromatogram which yielded the information presented in the table (Figure
3.).The first peak annotated is 2-3-5 trichlorophenol, the next three annotated (small peaks) are tetrachlor
phenols and the large off-scale peak at the end of the chromatogram is pentachlorophenol. As noted
above, the small peak, the trichloro phenol, was found at a concentration of 0.4 mg/kgm and the largest
peak, pentachlorophenol was found at 1483 mg/kgm. This illustrates the large range of phenolic
compounds to be expected, greater than three orders of magnitude range.
FIGURE 4. GC/ECD Chromatogram of Reference Soil Containing Chlorinated Phenols
3 1
'
8:
.A...
'll
330
-------
FIGURE 5. Recovery from Spiked Sample - 6 replicates (Ref. 1)
1 Id
100
90
80
70
GO
50
-
n
•i .
2.G 3.
2.1
^
"i
M
-,
-L
—
-|
n
-i -,
•^
: - :
.
5 3,4 2.3.G 2,4,5 2.3,5,6 2,3,4.5
2.:! 2 1.»J 2,:i..r) 3, 1,5 2,3. ,(! 2,3, 1.5.G
• Rf-c'U'erj
0 5 mg/kgrr
Ket
5 0
oven
rag/kgm
Figure 5 shows the recovery of a number of the substituted chlorophenols ranging from the di-substituted
up to the penta chlorinated substituents. The solid (darkest) bar graphs represent recovey of spiked
levels of 0.5 mg/kgm ( parts per million on a weight-per-weight basis). The lighter patterned bar graphs
represent recoveries for the same compounds at 10 times higher concentration, approximately 5 mg per kg
level. It is relevant to note the significance of the recovery across this variety of chlorinated phenols at
the two different levels. As expected, there is a slightly lower recovery at the sub- part per million level
compared to the 5 part per million level, however, the average recovery for six replicates is greater than
87 % in the worst case and greater than 104 % in the best case, with an overall grand mean recovery
greater than 95 %.
331
-------
FIGURE 6. % Relative Standard Deviation from Spiked Sample - 6 replicates (Ref. 1)
10
2.6 3,5 3,4 2,3,6 2,4,5 2,3,5,6 2,3,4,5
2,1 2.3 2,4,6 2,3.5 3,1,5 2,3,4,6 pep Grand mean
|%rsd I
0.5 ing/kgm I
I %rsd
I 5.0 mg/kgni
In Figure 6 we see that in addition to good recovery, the SFE in-situ derivatiztion/extraction yields very
acceptable precision. The solid (darkest) bar graphs show that precision for the compounds at the 0.5
(parts per million ) mg/kg level ranges from 3% rsd to 8 % rsd. The (lighter) patterned bar graphs show
the precision at 5 mg/kgm (ppm). As one would normally expect, in trace analysis, the precision
improves significantly as the detected concentrations are 10 times greater. That is, the grand mean % rsd
for the 14 compounds at the 0.5 ppm level is 6.5 % while at the 5.0 ppm level it improves to 5.0%.
Because the derivatization method as described here does not derivatize dinitro phenols, we are examining
another possiblity for the broader set of phenols. We are following some experiments that our colleagues
at HP are doing with packed column SFC. When analyzing mixtures with packed column SFC, the
conditions in many respects are analogous to supercritical fluid extraction. If one can find experimental
data on packed column SFC, this may suggust good conditions for SFE with solid matrices. Jerry Deye
and Terry Berger of our laboratories at Hewlett-Packard in Wilmington, DE have an SFC chromatogram
performed the separation of many of the same phenols described in this study, as well as dinitro phenols
Figure 7 ).
332
-------
Figure 7
SFC of un-derivatized Phenols
EM MUM (MAM*
RTflw
1 014 2-Nfcphml
2(177
3QN
41.23
5143
I t.4f 2,4J)ttfa9ftml
71*
i in PM*
IU9
10 W
11 146
20 l
MoHl phlHi
ouM
IMfflpKlM
Cwwer
333
-------
The total elution time for the eleven phenols, is only three and a half minutes. That is a fast , high
resolution separation of the eleven phenol compounds. These include 2-nitro, the 4-6
dinitromethylphenol and phenol itself. This is a packed SFC DIOL-bonded silica column. The mobile
phase was supercritical CO2 with a mixture of trichloro acetic acid ( 0.05 %) in methanol . The
methanol/trichloro acetic acid mixture was then introduced as a 5% volume of the carbon dioxide. This
represents an additive in methanol, then added to CO2~it is actually a ternary mixture. The pressure
was very low, 2000 psi, at 37 degrees centigrade. The phenols are not derivatized. We are taking a
look at this simultaneously with the derivatization steps and the automation.
Summary
We are evaluating simplification of running the proposed method routinely by automating the post-SFE
fraction clean-up using automated ( HP PrepStation) instrumentation.
We have further looked at the reaction and extraction efficiency for methyl phenols and nitro phenols as
well as the chlorophenols described in the original study. We find the in-situ method very good for die
chlorinated phenols. For dinitrophenols, we are considering un-derivatized extraction with co-solvents and
additives to enhance the extraction and recovery. One or more of these approaches will be offered to the
USEPA SW846 committee in the form of "deliverables" (3) for consideration as methods for extracting
phenols and substituted phenols. This normally would result in a round robin study to test the ultimate
robustness in a wide variety of laboratories.
Acknowledgment
The authors would like to thank all of the people who are helping with these SFE phenol studies and
especially Barry Lesnik, Robert Hong-You, Tom Peart, Pat Castelli, Joel Cheng, William Pipkin and
many others who helped in their own way
References
1. H.B.Lee , T.E. Peart, R.L. Hong-You, J. Chromatogr. 605 (1), 109-113 (1992).
2. D.R. Gere, C.R. Knipe, P. Castelli, J. Hedrick, L.G. Randall, J. Orolin, H. Schulenberg-Schell, R.
Schuster, H.B. Lee, and L. Doherty "Bridging the Automation Gap between Sample Preparation and
Analysis: SFE, GC, GC/MSD and HPLC Applied to Environmental Samples" manuscript in press,
J.Chromatog. Sci.
3. Personal communication from Barry Lesnik, Office of Solid Waste, US Environmental Protection
Agency, 401 M Street, SW, Washington, DC 20460 USA.
334
-------
43 SUPERCRITICAL FLUID EXTRACTION (SFE) of PCBs
Mark L. Bruce. Director of Research and Development, Enseco-Wadsworth/ALERT
Laboratories, a Division of Corning Lab Services, 4101 Shuffel Dr. N.W., North Canton,
Ohio 44720
ABSTRACT
Supercritical fluid extraction parameters will be optimized using fractional factorial
experimental designs. Matrix modification with water will also be investigated.
Preliminary results with Aroclor 1248 indicate more rigorous extraction conditions than
those specified in Method 3560 may be necessary.
INTRODUCTION
Extraction of hydrocarbons from soil samples was one of the first applications of
analytical SFE with carbon dioxide. Similarly hydrocarbon extractions were the first SFE
method from the EPA Office of Solid Waste. The commercial instrumentation has
improved significantly in the past two years. The environmental lab can choose between
several SFE instruments specifically geared for the high through-put lab. The next step is
to extend the current SFE methodology to other environmental analytes. Non-polar
compounds such as PAHs and PCBs should be extractable under conditions similar to
hydrocarbons. SFE of TPHs will dramatically reduce the volume of Freon-113® used by
the Jab. Extending this technology to other analyte fractions will reduce the use of
methylene chloride. This will reduce both environmental and employee exposure to this
suspected carcinogen.
Work by Langenfeld, Hawthorne, Miller and Pawliszyn (1) shows good PAH and PCB
recovery using high temperature (200°C) extractions. Preliminary work at Enseco-
Wadsworth/ALERT indicates that the parameters of the TPH method may not be
sufficient to achieve good PCB recovery. Two approaches are possible, 1) use co-
solvents (modifiers) or 2) use higher temperature, pressure and flow. Modifiers have
both advantages and disadvantages. We will focus on the second option. If acceptable
PCB recovery can be obtained without modifiers for off-line extraction then the eventual
switch-over to on-line extraction will be simpler.
INSTRUMENTATION. EQUIPMENT AND SUPPLIES
Supercritical Fluid Extractor
Suprex, PrepMaster, AccuTrap
Restrictor, prototype variable restrictor (manual)
Suprex, AP44
5 rnL extraction vessel
Gas Chromatograph
Hewlett Packard 5890
Electron Capture Detector
Column RTX-5, 30 m, 0.32 mm ID, 0.5 [L DF
Solvents
Hexane, EM Science
CO2, SFC grade with 1500 PSIA Helium headspace with dip tube, Scott Specialty
335
-------
DISCUSSION
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 type, and amount. Solid matrices such as loam, ash and various
clays (kaolin, Fuller's earth, montmorillonite) will be examined. Fractional factorial
experiment designs will be used to investigate and optimize this large number of
variables.
Water affects organic extractions in several interesting ways. Too much water is
detrimental and reduces analyte recovery. However, too little water can also reduce
analyte recovery. Extracting TPHs from dry clays with Freon-113® was facilitated by
adding water to the sample (2). A similar effect was noted for supercritical fluid
extraction of TPHs in draft Method 3560 (3). However, too much water was a detriment.
Very wet clay samples must be partially dried before good TPH recovery can be achieved
(4,5). These water effects are inconsistent if one views water in the role of a CC>2
solubility modifier. Water as a polar solvent should only decrease the extraction
efficiency of non-polar analytes such as hydrocarbons. However, Hawthorne has noted at
several conferences (6, 7, 8) that co-solvents should also be viewed as matrix modifiers.
McNally observed physical changes in soil matrices when water was added (9). These
physical changes may explain in part how water affects the extraction process. This
raises the possibility of using water as an SFE modifier even for non-polar analytes. If
effective, water would be the least expensive and safest modifier.
Preliminary work with PCBs indicates that the SFE conditions specified in draft Method
3560, December 1992 may not be sufficient for good PCB recovery. Aroclor 1248
recovery from Kaolin was 51% using these parameters. The final goal is an extraction
method that yields acceptable PCB recovery in a short time frame under production
laboratory conditions.
SUMMARY
Fractional factorial experimental designs will be used to optimize the key variables
affecting PCB recovery from various solid matrices. The use of water as matrix modifier
will be investigated.
REFERENCES
(1) Analytical Chemistry 1993, 65, 338-344.
(2) American Environmental Laboratory, 1990,12,25-28.
(3) SW-846 draft Method 3560, December 1992,4-5.
(4) Interlaboratory Evaluation of an Off-line SFE/IR Method for Determination of
Petroleum Hydrocarbons in Solid Matrices Project Report, EPA/600/R-93/056 Apr, 1993.
(5) Proceedings 8th Annual Waste Testing & Quality Assurance Symp., 1992,427-436.
(6) PitCon, 1991
(7) 4th Internal. Symp. on Supercritical Fluid Chromatography and Extraction, 1992.
(8) American Petroleum Institute Workshop, Colorado Springs, Feb., 1992.
(9) Proceedings Internal. Symposium on Supercritical Fluid Chromatography and
Extraction, 1992, 35-36.
336
-------
44 Determination of Organic Compounds in Ground Water by Liquid-Solid Extraction
Followed by Supercritical Fluid Elution and Capillary Column Gas
Chromatography/Ion-Trap Mass Spectrometry
By J.S. Ho. P.M. Tang*, J.W. Eichelberger, and W.L. Budde
Environmental Monitoring Systems Laboratory, U.S. Environmental Protection
Agency, 26 W. Martin Luther King Dr. Cincinnati, Ohio 45268
* Technology Applications, Inc. c/o address above
Methylene chloride (MeCl2) has been widely used in RCRA/CERCLA monitoring
as an extraction solvent for extraction of organic contaminants in ground water
and solid waste samples. The testing protocol using methylene chloride liquid-
liquid extraction is wasteful in time for preparation of analytical samples and
also costly in disposing the waste solvent. Environmental Monitoring Systems
Laboratory-Cincinnati (EMSL-Cincinnati) of USEPA has initiated a research project
to develop alternative extraction methods to replace methylene chloride. One
alternative method for extraction of organic pollutants from water is a liquid-
solid extraction (LSE). This method uses solid sorbents (i.e. C18) to trap
analytes and uses solvents to elute the adsorbed analytes. The LSE method offers
low solvent consumption (i.e. 15-60*mL of MeCl2 compared to >400 mL of MeCl2 in
the liquid-liquid extraction) and less waste. The EMSL-Cincinnati is also
investigating a novel technique, analytical scale supercritical fluid extraction
(SFE), to replace the methylene chloride solvent elution. The SFE uses a fluid
(i.e. C02) that is non-toxic, non-combustible, chemically inert, and easy to
discard. For the LSE method, water samples are first passed through 47-mm
diameter Empore™ C18 disks. Then, different amounts of MeCl, are used to elute
the adsorbed analytes from the disks. An amount of 15 mL MeCT2 is used for clean
samples without particulates, while at least 60 mL of MeCl2 are required for
dirty samples containing particulates. These MeCl2 extracts need further
concentration before injecting into a capillary column gas chromatograph for
trace analysis. The use of SFE as an alternative elution technique to the
solvent elution of LSE is more attractive. After filtering the analytes from the
water by using the C18 disk, the adsorbed analytes are eluted with supercritical
CO, (about 30 mL) ana collected in a small amount of solvent (1-2 mL). The SFE
method combines the elution and the concentration procedures into one step. In
comparison with the solvent elution of LSE, the SFE reduces the solvent
consumption by 90% and generates much less waste. Detailed discussions of the
recovery efficiencies, relative standard deviation (RSD) and detection limits as
well as the effects of various solvent modifiers on the SFE extraction will be
presented at symposium. The disk liquid-solid extraction (LSE) coupled with the
supercritical fluid elution (SFE) provides clear advantages in terms of
extraction time, solvent used, and solvent disposal especially for samples
containing particulates like ground water sample.
337
-------
45 COMPARISON OF SOLID PHASE EXTRACTION WITH
SALTING-OUT SOLVENT EXTRACTION FOR PRECONCENTRATION OF
NITROAROMATIC AND NITRAMINE EXPLOSIVES FROM WATER
Thomas F. Jenkins and Paul H. Miyares, U.S. Army Cold Regions Research and Engineering
Laboratory, Hanover, New Hampshire 03755-1290, and Karen F. Myers, Erika F. McCormick,
and Ann B. Strong, U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi
39180
ABSTRACT
Residues of high explosives are one of the most significant pollution problems at DOD facili-
ties, Recently, the EPA has lowered the concentration at which these compounds are thought to
be harmful to human health. Because TNT, RDX, and HMX, as well as several manufacturing
impurities and environmental transformation products, are quite mobile in the soil and have
caused groundwater pollution, there is an increasing demand for low-level analysis of these
compounds in groundwater from installation boundary wells.
RDX and HMX are very polar, and normal liquid-liquid extraction/preconcentration techniques
result in poor recovery. Two innovative preconcentration techniques have been reported that
appear to offer improved recovery and adequate preconcentration: solid-phase extraction
(SPE), and salting-out solvent extraction (SOE). This paper compares cartridge-SPE, mem-
brane-SPE, and SOE using a series of reagent-grade water samples fortified with low concentra-
tions of 11 nitroaromatics and nitramines and a set of groundwater samples from an explosives-
contaminated DOD facility.
Results indicated that the three methods were comparable with respect to low-level detection
capability, which ranged from 0.05 to 0.30 (J-g/L. Percent recoveries were generally greater than
80% except for HMX and RDX by membrane-SPE, which were consistently lower. Interferences
were found in extracts from about half of the groundwater samples preconcentrated using the
two SPE procedures but were not found in any of the extracts from the SOE. These interferences
were traced to matrix interaction of the groundwater with the polymeric resins.
INTRODUCTION
Batch liquid-liquid extraction (LLE) has often been used to extract organic analytes from water
prior to analysis. The success of this approach depends upon favorable solvent/water partition
coefficients (K ) for the analytes of interest. Since many organic pollutants are relatively non-
polar, batch extraction with an immiscible organic solvent has been successful.
TNT (2,4,6-trinitrotoluene) and RDX (hexahydro-l,3,5-trinitro-l,3,5-triazine), along with
several of their manufacturing impurities and environmental degradation products, have been
observed in groundwater at a number of Army installations [1-5]. Health advisories have been
issued by the EPA for many of these compounds (Table 1), with drinking water criteria at sub-
part-per-billion concentrations. For this reason, the Army has developed analytical methods
to determine these compounds at trace levels in water. To detect these compounds at the sub-ppb
level, the first step has been preconcentration of the analytes using an extraction process.
338
-------
Table 1. Proposed drinking water criteria
for nitroaromatics and nitramines.
Proposed drinking
water limit
Compound
Reference
Classical LLE of these analytes has been report-
ed, but poor extraction efficiency has been found
for RDX and HMX. Octanol/ water partition co-
efficients (Kou.) for HMX and RDX are 1.15 and
7.24 [6], respectively, indicating the compounds
are quite polar and will be difficult to extract
from water using nonpolar solvents.
For this reason, several alternative procedures
were developed. The most successful of these
have been the use of solid-phase extraction
(SPE) [13, 14] and salting-out liquid-liquid ex-
traction (SOE) [15-17].
The objective of the work described here was to
obtain a direct comparison of the cartridge SPE
[13] and membrane SPE methods [14] with the
SOE method [17]. Aqueous test solutions of nitro-
aromatics and nitramines prepared in ground-
water without the use of organic solvents allow-
ed a realistic comparison at concentrations of
analytes below the proposed drinking water
criteria (Table 1). In addition, all three pro-
cedures were tested with a set of groundwater samples from a military toxic waste site known to
be contaminated with explosives.
HMX
RDX
TNT
2,4-DNT
2,6-DNT
1,3-DNB
400*
2.0*
2.0*
50"
0.1
40**
0.007
1.0*
171
18]
P]
[10]
[H]
[10]
[11]
[12]
EPA Lifetime Health Advisory Number
EPA number for increased cancer risk of
1.0x10-6.
EXPERIMENTAL
Salting-out solvent extraction/nonevaporative preconcentration procedure
This method was based on the salting-out solvent extraction procedure [17] in combination with
the nonevaporative preconcentration technique [18] reported elsewhere. A 251.3-g portion of
reagent-grade sodium chloride was added to a 1-L volumetric flask along with a 770-mL sam-
ple of water. A stir bar was added, and the contents were stirred at maximum speed (rpm) until
the salt was completely dissolved. A 164-mL aliquot of ACN was added and stirred for 15 min-
utes. The stirrer was turned off and the phases allowed to separate. The ACN phase (about 8
mL) was removed, and 10 mL of fresh ACN were added. The flask was again stirred for 15 min-
utes, followed by phase separation. The ACN was removed and combined with the initial
extract. The combined extract was placed in a 100-mL volumetric flask and 84 mL of salt water
(325 g NaCl per 1000 mL of water) was added, and the contents were stirred for 15 minutes.
After allowing the phases to separate, the ACN phase was carefully removed and placed in a
10-mL graduated cylinder. An additional 1.0-mL aliquot of ACN was then added to the volu-
metric flask and the contents were stirred. Again the phases were allowed to separate, and the
resulting ACN phase was added to the 10-mL graduated cylinder. The resulting extract, about
5-6 mL, was than diluted 1:1 with reagent-grade water prior to analysis [19].
339
-------
Cartridge solid-phase extraction
Porapak R (80/100 mesh) was obtained from Supelco and precleaned by Soxhlet extraction using
ACN [19]. After precleaning, the material was air-dried briefly and then oven-dried at 105°C
for 2 hr.
Empty 3-mL extraction cartridges were packed with 0.5 g of the precleaned Porapak R. The car-
tridges were placed on a Visiprep Solid-Phase Extraction Manifold (Supelco) and conditioned
by eluting with 30 mL of ACN followed by 50 mL of reagent-grade water. A 500-mL aliquot of
each water sample was pulled through a cartridge at about 10 mL/min.
The cartridges were eluted by passing a 5-mL aliquot of ACN through the cartridge at about 2
mL/min, and the eluate was collected in a 10-mL graduated cylinder. The resulting extract,
about 5 mL, was diluted 1:1 with reagent-grade water prior to analysis [19].
Membrane solid-phase extraction
The 47-mm Empore styrene-divinyl benzene disks were precleaned by soaking them in 50 mL of
ACN [19]. Each membrane was soaked for four 24-hour periods with fresh ACN. Each disk was
rinsed with ACN, then centered on a 47-mm vacuum filter apparatus and leached with a 20-mL
portion of ACN and a 50-mL aliquot of reagent-grade water. Just before the last of this water
was pulled through the membrane, the vacuum was stopped and the reservoir was filled with
sample. The vacuum was then turned on again and a 500-mL aliquot of a water sample was
pulled through the membrane. This took from 5 to 7 minutes, with resulting flow rates ranging
from 70 to 100 mL/min. Air was then drawn through the membrane for 1 minute to remove excess
water. A 5-mL aliquot of ACN was then used to extract the analytes from the disk, and the
extract was diluted 1:1 with reagent-grade water prior to analysis [19].
RP-HPLC analysis
All analyses were conducted using reversed-phase HPLC as described elsewhere [19,20].
Preparation of standards and samples
Since we felt it was important that test samples be completely free of organic solvents, we pre-
pared all test samples in a totally aqueous matrix. This was done by preparing individual
aqueous analyte stock solutions by placing a few hundred milligrams of Standard Analytical
Reference Material (SARM) of each specific analyte in individual 4-L brown glass bottles, fill-
ing the bottles with reagent-grade water, adding a stir bar, and stirring for two days at room
temperature.
Each solution was filtered through a 0.45-mm nylon-66 membrane (Supelco) into a clean brown
glass bottle. Aliquots of each solution were then analyzed against standards prepared in ace-
tonitrile to estimate the concentration of analyte in each aqueous stock solution.
340
-------
RESULTS AND DISCUSSION
Certified reporting limit test
Spiked test solutions were used to
conduct a four-day Certified Report-
ing Limit Test (CRL) as described
elsewhere [19]. This test allows a
comparison of low-concentration de-
tection capability, percent recovery,
interferences, and overall precision.
A discussion of the CRL and how re-
sults compare with the EPA Method
Detection Limit are presented else-
where [21]. Chromatograms for sam-
ples containing target analytes at
about 0.2 ug/L and preconcentrated
by the three procedures are present-
ed in Figure 1.
The CRLs obtained are shown in
Table 2. Overall, the CRLs for a
given analyte are quite similar for
all three preconcentration tech-
niques. None of the procedures is
consistently superior to the others in
low-concentration detection capabil-
ity. CRL values range from a low of
0.032 ug/L for DNB using cartridge
SPE to a high of 0.83 ug/L for tetryl
using membrane-SPE. All values for
HMX, RDX, TNT, 2,4-DNT, and 1,3-
DNB are below the proposed drink-
ing water limit for these compounds.
While 2,6-DNT was not tested, it is
very unlikely that a CRL as low as
the proposed value of 0.007 ug/L
would be obtained. The only CRL
value that appears to be out of line
is the 0.83 ug/L value of tetryl using
membrane-SPE. Inspection of the da-
ta indicates this high CRL is due to
low recovery on one of the four days.
Salting-out
2X Sample
Membrane SPE (Empore SDVB)
2X Sample
Cartridge SPE (Porapak R)
2X Sample
Retention Time (min)
Figure 1. LC-18 chromatograms for 2X sample
preconcentrated by salting-out, cartridge-SPE, and
membrane-SPE methods at CRREL.
The regression line obtained from the plot of found vs. spiked concentrations was also examined
for curvature using lack-of-fit testing. Linear relationships described the data at the 95%
confidence level. The slopes of these linear regression lines are measures of the overall percent
recoveries of these analytes using each preconcentration method. Recoveries (Table 3), in
general, are quite good (near a theoretical value of 100%). Measured recoveries for the salting-
out procedure range from 93-119%. Likewise, recoveries for cartridge-SPE and membrane-SPE
range from 83-133% and 81-116%, respectively.
341
-------
Table 2. Certified reporting limits for various Table 3. Overall percent recovery and relative
preconcentration techniques. standard deviation (RSD) from Certified Re-
porting Limit Test.
Analyte
CRL (fig/L)
Salting- Cartridge- Membrane-
out SPE SPE
% recovery* (% RSD)
Salting- Cartridge- Membrane-
Analyte
out
SPE
SPE
HMX
RDX
TNB
DNB
Tetryl
TNT
2-Am-DNT
2,4-DNT
o-NT
p-NT
m-NT
0.19
0.13
0.052
0.081
0.20
0.086
0.10
0.083
0.13
0.22
021
021
027
0.042
0.032
0.24
0.068
0.046
0.085
0.10
0.12
0.13
0.33
0.12
0.051
0.036
0.83
0.13
0.055
0.044
020
0.23*
0.37
HMX
RDX
TNB
DNB
Tetryl
TNT
2-Am-DNT
2,4-DNT
o-NT
p-NT
m-NT
106 (10.5)
106 (8.7)
119 (7.6)
102 (6.6)
93 (16.4)
105 (7.6)
102 (9.1)
101 (5.8)
102 (9.1)
96 (18.1)
97 (12.4)
107 (9.6)
116 (22.0)
133 (8.7)
115 (2:6)
83 (32.8)
111 (7.5)
113 (4.1)
109 (6.8)
107 (8.1)
104** (6.6)
100 (7.3)
81 (14.0)
116(11.1)
116 (10.3)
103 (6.5)
83 (46.4)
97 (10.5)
103 (8.9)
94 (6.6)
92 (15.6)
89t (18.0)
86 (17.2)
One outlier removed for this analyte/method
combination.
Slope of regression line of spiked concentration
vs. found concentration x 100.
Lack-of-fit test indicates data not adequately
described by linear relationship at the 95%
confidence level.
One outlier removed for this analyte/method
combination.
Comparison using groundwater samples
A further test was conducted using 58 groundwater samples from the Rockeye site at the Naval
Surface Warfare Center. All of the samples were analyzed by the direct method and after pre-
concentration as well using the three procedures (SOE, cartridge-SPE, and membrane-SPE).
Table 4 summarizes the results for samples where the concentrations of at least one nitroaro-
matic or nitramine analyte was high enough to be obtained using the direct method. Since the
concentrations obtained by the direct analysis are subject to fewer sources of error than those
obtained using preconcentration, we are treating these values as "true" values for purposes of
comparison. We can then compare the results from the various preconcentration techniques rela-
tive to these "true" values.
Examination of Table 4 indicates that all three preconcentration procedures did a fairly good
job of recovering these analytes (generally greater than 80%). The membrane-SPE method,
however, recovered less of HMX and RDX than the other two procedures. Recovery of HMX,
RDX, and TNT by the cartridge-SPE and salting-out methods was greater than 80% in all cases,
with a slightly better recovery for the cartridge-SPE method for HMX and RDX.
For about half of the samples, the chromatogram for the salting-out method is blank with
respect to target analytes and interferences, but the chromatograms for both the two solid-
phase methods, at the same attenuation, show large peaks at a number of retention times across
the entire chromatogram (Figure 2). Second-column confirmation indicated that none of these
peaks resulted from the presence of nitroaromatic or nitramine explosives. Their presence, how
342
-------
Table 4. Comparison of results for direct analysis of ground-
water samples from the Rockeye site at the Naval Surface
Warfare Center, Crane, Ind., with the three preconcentration
methods.
Concentration (ng/L)
Sample Method* HMX RDX TNB TNT 4A
ever, interfered with the abil-
ity to detect nitroaromatic and
nitramine analytes at concen-
trations well above the CRLs.
These large interference peaks
were not found when these
groundwater samples were
analyzed without preconcen-
tration and hence appear to be
introduced by the solid phases
themselves. Since the Porapak
R and the SDVB membranes
were cleaned separately and
all the Porapak R cartridges
were packed from material
cleaned in the same batch, we
do not believe these peaks
were a result of poorly cleaned
material. Rather it appears
that some component of these
samples interacted with the
solid phases to either degrade
the polymer or release contam-
inants from within the poly-
mer by either swelling or col-
lapsing the polymer matrix.
CONCLUSIONS AND
RECOMMENDATIONS
c,-.c j .. • j j * Membrane-SPE (SPE-M), cartridge-SPE (SPE-C), and salting-out
SOE and cartridge- and mem- (CQP)
brane-SPE were compared
with respect to their ability to
preconcentrate nitroaromatic and nitramine explosives from water prior to RP-HPLC analysis.
Both fortified reagent-grade water and contaminated groundwater samples were used in this
assessment.
Low detection capability and overall precision were comparable among the three procedures.
Recoveries of HMX and RDX are better using cartridge-SPE and SOE than membrane-SPE. Re-
covery of the nitroaromatics was acceptable for all three procedures.
At present the major problems associated with the use of the SPE procedures are the inadequacy
of the current cleaning procedures. While use of the Soxhlet procedure on a batch basis for the
Porapak R material appears to be adequate, cleaning must be accomplished on-site just before
use or the contamination reappears. The cleaning procedure we used for the membranes is cum-
bersome and wastes solvent.
Problems with interferences were encountered using both the cartridge-SPE and membrane-SPE
procedures for a number of actual groundwater samples. These interferences appeared to be iden
20649
20650
20662
20663
20667
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
151
98
156
161
119
60
107
98
26
19
17
22
281
153
214
232
318
199
356
319
135
121
147
138
82
64
85
66
160
176
138
154
94
89
109
90
618
488
666
558
33
32
34
38
9.0
7.6
92
10.3
42
51
34
46
21
22
26
26
19.2 284
19.5 317
19.6 328
18.6 320
9.6
112
122
13.7
65
75
78
78
166
216
239
217
343
-------
Sa»ing-out
Cartridge-SPE
Membrane-SPE
a 12
Retention Time (min)
16
20
Figure 2. LC-18 chromatograms for sample 20641 from
Rockeye site, Naval Surface Warfare Center, precon-
centrated by salting-out, cartridge-SPE, and membrane-
SPE methods at WES.
tical to the compounds released from
the solid phases during cleaning, but
were apparently released from the
SPE phases due to a matrix interac-
tion with a number of groundwater
samples. The nature of this inter-
action is still unclear. These com-
pounds would interfere with deter-
mination of nitroaromatic and nitra-
mine explosives even at reasonably
high concentrations.
The elimination of the need to use
evaporative preconcentration with
the salting-out procedure is a major
improvement. We believe the new
procedure will be more precise and
less subject to error in routine use
than the initial method, which uti-
lized a Kudena-Danish evaporator.
In summary, the SOE, cartridge-
SPE, and membrane-SPE preconcen-
tration techniques are all capable of
providing adequate analyte precon-
centration of nitroaromatics and ni-
tramines prior to RP-HPLC deter-
mination. Of the three, the SOE
method appears to be least prone to
interferences. The membrane-SPE
method requires the least sample-
processing time, but its recovery of
HMX is the poorest of the three
methods. The cartridge-SPE method
requires the least solvent per sam-
ple, but its routine use with the
currently available processing mani-
fold appears prone to problems with
cross contamination [19].
ACKNOWLEDGMENTS
The authors would like to acknowl-
edge Dr. C.L. Grant, Professor Emeri-
tus, University of New Hampshire,
and M.E. Walsh of the U.S. Army
Cold Regions Research and Engi-
neering Laboratory for useful com-
ments and suggestions on the manu-
script. In addition, Roy Wade of the
344
-------
U.S. Army Engineer Waterways Experiment Station (WES) is acknowledged for groundwater
samples collection, and S. Paige Pitts and Allyson H. Lynch of WES for assistance In
groundwater extractions. The authors also thank Dr. E.S.P. Bouvier of the Waters Chromatog-
raphy Division, Millipore, Inc., for useful discussions regarding the nature of the interferences
observed for the solid-phase extraction materials. Funding for this research was provided
jointly by the U.S. Army Environmental Center, Aberdeen Proving Ground, Maryland, Martin
H. Stutz, Project Monitor, and the U.S. Army Engineer Waterways Experiment Station,
Vicksburg, Mississippi, Ann B. Strong, Project Monitor.
LITERATURE CITED
[1] D.L. Pugh, Milan Army Ammunition Plant Survey, U.S. Army Toxic and Hazardous Mate-
rials Agency, Report DRXTH-FS-FR-82131, Aberdeen Proving Ground, MD, 1982.
[2] D.H. Rosenblatt, "Contaminated Soil Cleanup Objectives for Cornhusker Army Ammuni-
tion Plant," U,.S. Army Medical Bioengineering Research and Development Laboratory
Technical Report 8603, Fort Derrick, MD, 1986.
[3] R.F. Spaulding and J.W. Fulton, "Groundwater munition residues and nitrate near Grand
Island, Nebraska, USA," Journal of Contaminant Hydrology, 2: 139-153, 1988.
[4] M.P. Maskarinec, D.L. Manning, and R.W. Harvey, "Application of Solid Sorbent Collec-
tion Techniques and High Performance Liquid Chromatography with Electrochemical
Detection to the Analysis of Explosives in Water Samples," Oak Ridge National Labora-
tory, Report TM-10190, Oak Ridge, TN, 1986.
[5] D. Layton, B. Mallon, W. Mitchell, L. Hall, R. Fish, L. Perry, G. Snyder, K. Bogen, W.
Malloh, C. Ham, and P. Dowd, "Conventional Weapons Demilitarization: A Health and
Environmental Effects Data Base Assessment," Lawrence Livermore Laboratory, CA,
1987. AD-A220 588.
[6] T.F. Jenkins, "Development of an Analytical Method for the Determination of Extract-
able Nitroaromatics and Nitramines in Soils," University of New Hampshire, Durham,
Ph.D. dissertation, 1989.
[7] Environmental Protection Agency, Health Advisory for HMX, Criteria and Standards
Division, Office of Drinking Water, Washington, DC, November, 1988a.
[8] Environmental Protection Agency, Health Advisory for RDX, Criteria and Standards
Division, Office of Drinking Water, Washington, DC, November, 1988b.
[9] Environmental Protection Agency, Trinitrotoluene Health Advisory, Criteria and Stan-
dards Division, Office of Drinking Water, Washington, DC, January, 1989.
[10] Environmental Protection Agency, Dinitrotoluene Health Advisory, Criteria and Stan-
dards Division, Office of Drinking Water, Washington, DC, 1992.
[11] E.L. Ernier, "Water Quality Criteria for 2,4-dinitrotoluene and 2,6-dinitrotoluene," Oak
Ridge National Laboratory Report AD-ORNL 6312, Oak Ridge, TN, 1987.
[12] Environmental Protection Agency, Health Advisory for 1,3-dinitrobenzene, Criteria and
Standards Division, Office of Drinking Water, Washington, DC, January, 1991.
[13] M.G. Winslow, B.A. Weichert, and R.D. Baker, "Determination of Low-Level Explosive
Residues in Water by HPLC: Solid-phase Extraction vs. Salting-out Solvent Extraction,"
in Proceedings of the EPA 7th Annual Waste Testing and Quality Assurance Symposium,
July 8-12, 1991, Washington, D.C., 1991.
[14] D.F. Hagen, C.G. Markell, and G.A. Schmitt, "Membrane Approach to Solid-phase
Extraction," Analytica Chimica Acta, 236: 157-164, 1990.
[15] D.C. Leggett, T.F. Jenkins and P.H. Miyares, "Salting-out Solvent Extractions for Precon-
centration of Neutral Organic Solutes from Water," Analytical Chemistry, 62(13): 1355-
1356, 1990.
345
-------
[16] P.H. Miyares and T.F. Jenkins, "Salting-out Solvent Extraction Method for Determining
Low Levels of Nitroaromatics and Nitramines in Water," USA Cold Regions Research
and Engineering Laboratory, Special Report 90-30,1990.
[17] P.H. Miyares and T.F. Jenkins, "Improved Salting-out Extraction/Preconcentration Meth-
od for the Determination of Nitroaromatics and Nitramines in Water," USA Cold
Regions Research and Engineering Laboratory, Special Report 91-18,1991.
[18] T.F. Jenkins and P.H. Miyares, "Non-evaporative Preconcentration Technique for
Volatile and Semivolatile Solutes in Certain Polar Solvents," Analytical Chemistry,
63(13): 1341-1343,1991.
[19] T.F. Jenkins, P.H. Miyares, K.F. Myers, E.F. McCormick, and A.B. Strong, "Comparison of
Cartridge and Membrane Solid-phase Extraction with Salting-out Solvent Extraction for
Preconcentration of Nitroaromatic and Nitramine Explosives from Water," USA Cold
Regions Research and Engineering Laboratory, Special Report 92-25,1992.
[20] T.F. Jenkins, M.E. Walsh, P.W. Schumacher, P.H. Miyares, C.F. Bauer, and C.L. Grant,
"Liquid Chromatograph Method for Determination of Extractable Nitroaromatic and
Nitramine Residues in Soil," Journal of the Association of Official Analytical Chemists,
72(6): 890-899,1989.
[21] C.L. Grant, A.D. Hewitt, and T.F. Jenkins, "Experimental Comparison of EPA and USA-
THAMA Detection and Quantitation Capability Estimators," American Laboratory, pp.
15-33, February, 1991.
346
-------
DINOSEB ANALYSIS IN THE FIELD AND THE LABORATORY
D. Anderson. APPL, Fresno CA 93722; F. Tsang, T. Jackson, and P.
Marsden, SAIC, San Diego, CA 92121
ABSTRACT
Dinoseb, 6-sec-butyl-2,4-dinitro phenol, is a herbicide and a plant
desiccant. The multiple functional groups of Dinoseb provide both
opportunities and challenges to the analytical chemist. Dinoseb is
generally measured using Method 8150/8151; unfortunately, the method is
time consuming and requires the use of diazomethane (an explosive and a
carcinogen). In addition, several Dinoseb-specific difficulties are not
fully explained in Method 8150 (which was developed as a multianalyte
method for phenoxyacid herbicides). As a result, engineers and field
staff have had the experience of seeing the characteristic yellow color of
Dinoseb on site while the laboratory receiving samples reports non-
detects. The performance of alternate, more specific, techniques for
analyzing for Dinoseb are described, including a colorimetric field
screening procedure. This screening procedure produces a distinctive
yellow color on Florisil™ and has been used to measure Dinoseb
concentrations of 5 ppm in soil. Techniques for improving laboratory
analysis of Dinoseb such as HPLC are also discussed. Optimized sample
preparation procedures for Dinoseb are presented including derivatizing
samples using diazomethane and alternate extraction techniques allowed
under Method 8150. Finally, the conversion of Dinoseb to reduced (amino
substituted) analogs will be discussed. Dinoseb analyses offer a chemist
a tremendous opportunity to design a site-specific analytical program.
This presentation will provide guidance and performance data to those
interested in designing the most appropriate and cost-effective Dinoseb
monitoring program.
INTRODUCTION
Dinoseb, 2-(1-methylpropyl)-4,6-dinitrophenol is a plant desiccant and
defoliant. Prior to its ban by EPA, dinoseb was used on a variety of
crops including potatoes and cotton. Dinoseb residues are often visible
at a site because of its intense yellow color and its propensity to
crystalize out of soil. Standard analytical practice is to send samples
to a laboratory for herbicide (Method 8150) or GC/MS analysis (e.g.,
Method 8270). This can greatly increase the time to make a remediation
decision and does not necessarily improve the data available to the site
manager
This paper describes alternate, more rapid analytical techniques for
dinoseb. These include a field screening procedure, a smaller volume
347
-------
laboratory extraction procedure, direct GC (non-derivatized) analysis and
an HPLC procedure for Dinoseb. In addition, some guidance is offered in
extracting and derivatizing samples when Method 8150/8151 are employed.
The results were generated by Agricultural and Priority Pollutant
Laboratory of Fresno, CA and Science Applications International
Corporation in a number of studies for commercial and government clients.
Support for comparison of the swirling methanol extraction with Soxhlet
extraction (Method 3540) was provided by the Methods Section of the US EPA
Office of Solid Waste (OSW) and the Risk Reduction Engineering Laboratory
(RREL). The OSW study involved generation of performance data for 27
substituted phenols from spiked soils using both extraction techniques and
is reported in a poster presentation at this symposium. The RREL study
determined the recovery of spiked dinoseb and a surrogate (bromoxynil)
from soil and compared in situ contamination of Dinoseb using both
extraction techniques.
DINOSEB SCREENING PROCEDURE
Dinoseb has a bright yellow color that can be used to screen for the
compound. Dinoseb is extracted from soil using methylene chloride in a
scintillation vial. The sample is agitated and the solids are allowed to
settle and the methylene chloride is transferred to a second scintillation
containing a small amount of Florisil. The Florisil adsorbs the dinoseb
and presence of dinoseb is indicated by yellow color.
PREPARATION OF DINOSEB AS A PHENOL
Dinoseb data was developed during a performance study of several SW-846
methods including Method 8041 (phenols). Although a total of 27 phenols
were evaluated, data presented in this report is limited to mixture #1
which contains dinoseb.
Each of the phenols were spiked into a clean, sandy loam soil from San
Diego. Each 5 g aliquot was spiked separately, then extracted using a
Soxhlet apparatus (Method 3540) or the aqueous methanol leaching
procedure. Methylene chloride was used as the solvent in Method 3540;
acidic aqueous methanol (1:1 methanol/pH 2 water) was used to leach the
phenols from soil on a orbital shaker (150 rpm). Following extraction,
methylene chloride extracts (Method 3540) were exchanged to methanol so
that all sample extracts could be treated in the same fashion. . Sodium
chloride, distilled water and hydrochloric acid was added to each extract
so that it was primarily aqueous. Phenols were extracted using 3 X 3 mL
portions of methylene chloride. The methylene chloride was concentrated
under nitrogen and analyzed by GC/FID.
A subsequent study for RREL evaluated the use of HPLC for the analysis of
dinoseb. Reversed phase HPLC is an attractive complement to the methanol
348
-------
leach procedure because no solvent exchange is required prior to analysis.
Five replicates of sample spiked at 1, 5 and 25 /^g/mL were extracted using
pH 2 methanol to demonstrate recovery using the methanol procedure. The
suitability of spiking dinoseb into soil on the same day of analyses was
supported by comparison of recoveries 1 and 2 days after spiking. The
recovery of dinoseb was virtually the same when spiked samples were
extracted and analyzed 1 and 2 days after fortification.
The RREL study also included a comparison of Soxhlet extraction and the
methanol leaching procedure using a single sample contaminated with
dinoseb. Five 10.0 g aliquots of the sample were placed in an Erylenmyer
flask. The surrogate bromoxynil was spiked into each sample and
thoroughly mixed. A volume of 10 mL pH 2 methanol was added to the flask
to extract the phenols. Flasks containing the five replicate samples and
the lab blank were placed on a shaker table and swirled for 25 minutes at
150 RPM. The 10 mL methanol was pipetted off and an additional volume of
5 mL pH 2 methanol was added to the sample. The shaker table extraction
was repeated. The two portions of methanol were combined for HPLC
analysis. Five additional aliquots of the sample were extracted by the
Soxhlet procedure (Method 3540) using methylene chloride. The methylene
chloride was exchanged to methanol prior to analysis.
PREPARATION OF DINOSEB USING METHOD 8150
Dinoseb is an analyte for Method 8150/8151 (phenoxyacid herbicides). In
some cases, the acid/base partitions specified in the method may be
required to achieve adequate cleanup or to meet project DQOs. However,
analysts are cautioned that dinoseb may be lost when a 30 g sample is used
because of the increased sample handling and solvent transfers required.
Loss of dinoseb can be minimized by reducing the sample size, the number
of solvent transfers, and completing the preparation as quickly as is
prudent (i.e., there is no convenient place to stop Method 8150/8151 is
used for dinoseb) . Acid washing of all glassware will also improve
recovery of dinoseb.
GC ANALYSIS
After acquisition of the target chemicals for the OSW study, gas
chromatographic (GC) conditions were evaluated in order to achieve
separations of the target analytes and to establish nominal low level
calibration concentrations. GC analyses were performed on both
underivatized phenols and phenols derivatized using diazomethane. Data
for the underivatized phenols only are reported here as this report is
focused on rapid techniques for the analysis of dinoseb. The GC
conditions recommended for the direct analysis of substituted phenols,
including Dinoseb, are:
349
-------
GC CONDITIONS
Column: DB-5 30m x 0.53 mm id
Initial Temperature: 80°C
Program: Hold 1.5 minute
6<>C/min to 230
10oC/min to 275
Final Hold: 4.5 rain
Run time: 35 min
Detector: FID, 300°C
Carrier gas: Nitrogen, 6 mL/min
Hydrogen: 30mL/min
Total nitrogen: 30 mL/min (carrier and makeup)
Injector: 1/4 inch Packed w/ megabore liner, at 200°C
Initial five-point GC/FID calibration established the linear portion of
instrument response for each of the phenols including Dinoseb. The
concentrations of the calibration solutions, retention times (Rt) and
relative standard deviations of the individual response factors (RSD) for
the mixture containing Dinoseb are provided in Table 1. The later-eluting
analytes in mixture 1 generally had higher RSDs (18-21%) and more narrow
calibration ranges (10-20 fold) than the other phenols. The problems in
using GC to measure underivatized dinitro-, trichloro-, tetrachlor-, and
pentachlorophenols are evident in the IQL (Table 2) and recovery values
(Table 6 and 7).
Examination of the GC chromatograms of underivatized phenols reveals broad
and non-Gaussian peaks. The use of a temperature programmed injector
(Varian SPI) did narrow peaks somewhat, methylation is the only reliable
technique for producing narrow phenol peaks on GC. Therefore, if there is
significant co-extracted interferences or lower detection limits are
required for the GC analysis of dinoseb, sample extracts should be
derivatized. Diazomethane should be used for derivatization, boron
triflouride - methanol does not perform reliably with dinoseb.
350
-------
TABLE 1 GC/FID Calibration of Phenols Mix 1
Analyte
Phenol
2-Cresol
3-Cresol
2 , 4-Dimethylphenol
2 , 6-Dimethylphenol
2 , 3 -Dimethylphenol
3-Chlorophenol
4-Chloro-3-
methylphenol
2 , 3 , 5-Trichlorophenol
2 ,4 , 5-Trichlorophenol
2 , 5-Dinitrophenol
2 , 4-Dinitrophenol
2,3,5,6-
Tetrachlorophenol
4,6-dinitro-o-cresol
Dinoseb
Calibration
solutions /ig/mL
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
10,40,50,100,200
40,50,100,200,400
5,20,25,40,50
10,40,50,100,200
5,20,25,50,200
Rt) min
6.37
8.17
8.65
9.63
10.54
11.32
11.68
14.07
15.47
16.05
18.37
19.29
20.42
21.72
25.71
%RSD
2.5
3.4
4.9
3.1
5.8
6.1
12.3
14.8
18.2
18.3
15.3
20.8*
17.6
14.9
15.6
may require non-linear calibration
351
-------
TABLE 2 - Phenol GC/FID IQL - Mix 1
Analyte
Phenol
2-Cresol
3-Cresol
2 , 4 - D ime thy Ipheno 1
2 , 6-Dimethylphenol
2 , 3 - Dime thy Ipheno 1
3 - Chlorophenol
4-Chloro-3-methylphenol
2 , 3 , 5-Trichlorophenol
2,4, 5-Trichlorophenol
2, 5-Dinitrophenol
2 ,4-Dinitrophenol
2,3,5,6-
Tetrachlorophenol
4, 6-dinitro-o-cresol
Dinoseb
Rt , min
6.37
8.17
8.65
9.63
10.54
11.32
11.68
14.07
15.47
16.05
18.37
19.29
20.42
21.72
25.71
cone,
/jg/mL
10
10
10
10
10
10
10
10
10
10
20
20
10
20
10
std dev
(s)
0.35
0.36
0.47
0.36
0.46
0.47
0.69
0.82
0.96
1.03
3.63
2.82
1.96
3.18
1.56
IQL 10s
Mg/mL
3.5
3.6
4.7
3.6
4.6
4.7
6.9
8.2
9.6
10.3
36.3
28.2
19.6
31.8
15.6
352
-------
HPLC ANALYSIS
HPLC/UV is employed for the analysis of dinoseb. It is an attractive
technique for rapid laboratory analysis because of the low detection level
for dinoseb, the compatibility with the methanol leach procedure and the
relatively poor chromatographic performance of underivatized dinoseb using
GC/FID. A study was conducted for RREL to evaluate the practicality of
HPLC for the analysis of bioremediation samples. Dinoseb was analyzed
using a Hewlett-Packard model 1090A series II Liquid Chromatograph
equipped with a diode array UV/VIS detector and a computerized data
system. Dinoseb concentrations in water, soil, sediment and-treated soil
were determined by HPLC/UV with a PSP-Spherisorb ODS2 column (5 ftm, 250 x
4 nun i.d.). Dinoseb and a surrogate (bromoxynil) were detected by the
diode array UV/VIS detector at 254 nm.
A gradient solvent program was employed: 30% methanol and 70% 0.1% acetic
acid buffer for 2 minutes, and then the methanol was increased to 100%
over the next 6 minutes and held at that level for 2 minutes. The
methanol was then decreased to 30% over 4 minutes and held for 6 minutes.
The solvent flow rate was 0.5 mL/minute and the column temperature was at
40°C. The pH of the mobile phase was critical as the chromatographic
behavior of Dinoseb and bromoxynil are very dependent on the ionization of
these phenols; this is particularly true for the surrogate bromoxynil.
Total run time was 20 minutes.
Tables 3 and 4 provide six-point initial calibration data for dinoseb and
the surrogate bromoxynil. The tables provide the concentration of the
calibration solutions, the retention times of the target analytes, peak
area counts from the data system, response factors, and the percent RSD of
the response factors.
353
-------
TABLE 3 - Calibration of Dinoseb
Cone . /zg/mL
0.50
1.0
5.0
10.0
50.0
100.0
200.0
RT
14.47
14.48
14.47
14.48
14.52
14.55
14.55
Area Count
20.0
39.2
186
374
1862
3663
7199
Mean =
Std =
%RSD =
RF
0.02500
0.02551
0.02688
0.02674
0.02685
0.02730
0.02778
0.02658
0.00098
3.7
TABLE 4 HPLC Calibration of Bromoxynil
Cone . /ig/mL
0.50
1.0
5.0
10.0
50.0
100.0
200.0
RT
12.95
12.96
12.96
12.96
12.97
12.98
12.98
Area Count
21.8
44.3
221
447
2215
4394
8653
Mean =
Std =
%RSD =
RF
0.02294
0.02257
0.02262
0.02237
0.02257
0.02276
0.02311
0.02271
0.00025
1.1
354
-------
FIELD SCREENING RESULTS
The dinoseb screening procedure provides a rapid means of monitoring
dinoseb in soil. It can provide a detection level of 5 ppm and can be
used to establish the presence or absence of in situ dinoseb in the field.
The technique has been demonstrated to perform reliably at several sites
and has reduced the equipment costs associated with soil removal. A
normal practice for the removal of dinoseb at a contaminated site is to
"chase the color". That is, to remove soil until all of the yellow color
is gone. Unfortunately, this technique is quite subjective and is subject
to problems when dinoseb is bound within soil particles. A site may
appear to be clean at the end of the day but show color the next day. The
dinoseb screening procedure reduces the potential for mis-interpretation
of soil color by providing a reliable test for dinoseb extracted from
soil.
PHENOL STUDY
Dinoseb can be leached from soil using a variety of solvents including
methanol, methylene chloride and ether. Studies to date indicate that
leaching is as effective as the standard extraction procedures (Methods
3540 and 3550) However, the leaching procedure may need to be tailored
to specific soil types. Studies at SAIC have demonstrated that acidic
methanol provides the best recoveries from Idaho soil; while studies at
APPL demonstrate that an aqueous sodium sulfate/ether leach provides
better recoveries from California soil. Additional studies will required
to establish the ruggedness of leaching procedures for dinoseb and other
substituted phenols in additional sample matrices.
GC/FID analysis of underivatized phenols demonstrate that there is a
positive bias for the recoveries of the higher molecular weight species.
This was particularly evident for the samples extracted using Soxhlet
(Method 3540) which are reported in Table 5 (i.e. , a recovery of 200% was
reported for dinoseb). Perhaps the co-extracted interferences in soil
contributed to the quantitation difficulties observed using GC/FID
analysis.
The difficulties of analyzing underivatized phenols was also observed with
the alternate extraction procedures. The recoveries from the methanol
leaching procedure in Table 6 also reveal a consistent high bias for
dinoseb and other phenols. However, the recoveries of using the leaching
procedure are generally closer to 100% than those obtained with Soxhlet
extraction. This observation could support the supposition that a co-
extracted material is interfering with the GC analysis of phenols. In any
case, there appears to be the potential for significant difficulties when
dinoseb and other phenols are analyzed by GC/FID without derivatization.
355
-------
TABLE 5 Recovery of Underivatized Phenols by GC and Method 3540
Analyte
Phenol
2-Cresql
3-Cresol
2 ,4-Dimethylphenol
2 , 6-Dimethylphenol
2 , 3 -Dime thy Iphenol
3 - Chlorophenol
4-Chloro-3-
methylphenol
2,3, 5 -Trichlorophenol
2,4,5- Trichlorophenol
2 ,5-Dinitrophenol
2 ,4-Dinitrophenol
2,3,5,6-
Tetrachloro'phenol
4, 6-dinitro-o-cresol
Dinoseb
Spiking cone
/*g/g
20
20
20
20
20
20
20
20
20
20
40
40
20
40
20
Recovery
(%)
93
95
98
93
101
106
116
128
136
139
177
157
236
201
210
% RSD
16.9
13.6
10.3
11.5
8.1
7.1
6.7
3.8
4.1
3.0
5.1
7.3
3.5
3.8
4.9
n
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
356
-------
TABLE 6 Recovery of Phenols by Leaching and GC
Analyte
Phenol
2-Cresol
3-Cresol
2 , 4-Dimethylphenol
2 , 6-Dimethylphenol
2 , 3-Dimethylphenol
3 - Chlorophenol
4-Chloro-3-methylphenol
2,3, 5 -Trichlorophenol
2,4, 5 -Trichlorophenol
2 , 5-Dinitrophenol
2 , 4-Dinitrophenol
2,3,5 , 6-Tetrachlorophenol
4 , 6-dinitro-o-cresol
Dinoseb
Spiking
cone /ig/g
20
20
20
20
20
20
20
20
20
20
40
40
20
40
20
Recovery
(%)
94
97
99
89
82
101
111
119
127
134
123
131
204
166
169
% RSD
2.6
2.8
2.8
3.8
4.0
2.6
2.7
2.4
2.8
2.9
6.6
6.8
3.3
4.3
4.2
n
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
HPLC STUDY
HPLC was evaluated as an alternate procedure for the rapid analysis of
underivatized dinoseb. Use of HPLC analysis allows the analysis of the
methanol leachate without derivatization or solvent exchange. The strong
UV absorption of dinoseb provides an opportunity for a low detection
level. An initial evaluation of the methanol leaching procedure combined
with HPLC analysis was conducted by measuring the recovery of dinoseb and
proposed surrogate, bromoxynil, from spiked samples. The performance of
the method seems reliable based on the results provided in Table 7.
The methanol leach/HPLC technique was tested against Method 3540/HPLC by
extracting aliquots of the same contaminated samples using methanol
leaching (Table 8) and Soxhlet extraction (Table 9). The results obtained
for dinoseb using either extraction procedure are very similar; higher
recoveries for bromoxynil were obtained using Soxhlet extraction.
Additional work may be required to provide a more suitable surrogate.
357
-------
Table 7. 1 Jtg/g dinoseb spiked soil extracted with pH2 methanol
1 /zg/g spiked soil
RT
14.26
14.24
14.24
14.22
14.22
Area
37
38
36
31
32
Mean =
Std =
%RSD =
Cone .
1.1
1.1
1.1
0.9
0.9
1.0
0.091
8.9
% Rec.
108
111
106
91
94
102
9.1
8.9
5 ppra Bromoxynil
RT
11.51
11.49
11.48
11.45
11.47
Area
151
140
152
151
147
Mean =
Std =
%RSD =
Cone .
5.5
5.1
5.5
5.5
5.3
5.4
0.18
3.4
%Rec.
110
102
110
110
107
108
3.6
3.4
TABLE 8 Methanol Leach/HPLC - Five Replicates of a Contaminated Sample
sample #21 -Dinoseb cone. /ig/mL
RT
14.44
14.46
14.47
14.44
14.44
Area
1333
1758
1607
1722
1795
Mean =
Std =
%RSD =
Extract
Cone.
30.3
39.9
36.5
39.1
40.8
37.3
4.2
11.4
Sample
Cone.
45.4
59.9
54.7
58.7
61.1
56.0
6.4
11.4
5 ppm Bromoxynil recovery
RT
13.03
13.02
13.02
13.02
13.01
Area
116
194
198
148
146
Mean =
Std =
%RSD =
Cone .
2.3
3.9
4.0
3.0
3.0
3.2
0.7
21.8
%Rec.
46.9
78.5
80.1
59.9
59.0
64.9
14.1
21.8
358
-------
TABLE 9 Method 3540/HPLC Five Replicates of a Contaminated Sample
sample #21-Dinoseb cone. /ig/mL
RT
14.50
14.49
14.50
14.49
14.49
Area
1468
1912
1887
1530
1580
Mean =
Std =
%RSD -
Extract
Cone .
33.3
43.4
42.9
34.7
35.9
38.0
4.7
12.4
Sample
Cone.
50.0
65.1
64.3
52.1
53.8
57.1
7.1
12.4
5 ppm Bromoxynil recovery
RT
12.95
12.95
12.95
12.94
12.95
Area
111
266
267
239
256
Mean =
Std =
%RSD =
Cone .
5.6
5.4
5.4
4.8
5.2
5.3
0.5
5.5
%Rec.
112
108
108
96.7
104
106
5.8
5.5
CONCLUSIONS
In our experience the screening procedure described for Dinoseb can
provide reliable evidence that Dinoseb is present at a site. The
performance data presented here that a methanol leach can provide adequate
extraction of dinoseb from some soils particularly when combined with HPLC
analysis. The lower IQL and larger linear range of HPLC vs. GC/FID
indicate that HPLC may be a more appropriate technique than underivatized
phenol analysis for providing rapid laboratory measurements. Project
managers should consider the use of screening or simplified laboratory
procedures (HPLC) for the analysis of Dinoseb when Project DQOs do not
require the use of Methods 8150/8151 or 8270..
359
-------
QUANTITATION OF POLYCHLORINATED BIPHENYLS USING 19 SPECIFIC CONGENERS
Siu-Fai Tsang and Paul Marsden. SAIC, San Diego, CA 92121, Mike Kennedy,
Mesa College, San Diego 92111, and Barry Lesnik, Organic Methods Program
Manager, OSW/EPA 22403
ABSTRACT
Polychlorinated biphenyls (PCBs) are ubiquitous and persistent pollutants
that have been banned for over 20 years. Despite our long history in
measuring PCBs in the environment, the interpretation of analytical
results for PCBs are often a problem during monitoring activities. A
major analytical difficulty results from the fact that most PCB methods
specify the use Aroclor standards as the basis of quantitative analysis.
Aroclors are complex mixtures of chlorinated biphenyls (e.g., analysis of
an Aroclor 1260 standard using capillary GC/ECD should produce more than
50 peaks). While most of the PCBs found in the environment were
originally discharged as Aroclors, the PCBs found in sediments today have
been altered by environmental weathering processes. As a result, they may
not produce the same pattern of peaks as the original Aroclor contaminant.
In order to improve the reliability of PCB analyses, a mixture of 19
specific PCB congeners is proposed as an alternative calibration
standard. These individual congeners can be resolved on an SE-30 type
capillary column (e.g., DB-5), are available through commercial vendors,
and represent the major peaks of the individual Aroclors. Spiked soil and
reference materials have been analyzed using quantitation based on Aroclor
and congener calibration. Analysts have demonstrated that mercury and
sulfuric acid clean up do not effect the recovery of these congeners or
Aroclors.
INTRODUCTION
PCB's were produced in the United States under the trade name Aroclor by
Monsanto until they were banned in 1972. Aroclors are mixtures of many
individual chlorinated biphenyls (congeners) which were sold as products
on the basis of their weight percentage of chlorine. For example, each
molecule in Aroclor 1254 contains 12 carbons and an average chlorine
content (w/w) of the mixture was 54%. The exception to this system was
Aroclor 1016, which is similar to Aroclor 1242 except that 1016 contains
only 0.4% pentachloro congeners while 1242 contains 4.2%. The lack of
higher chlorinated congeners allows Aroclors 1016 to be combined with
Aroclor 1260 in a single calibration standard. All other Aroclors must be
analzed as individual standards.
PCBs are generally analyzed using gas chromatography with electron capture
detection (GC/ECD) (e.g., SW-846 Methods 8080/8081 or Method 608). These
360
-------
methods specify that analyte identification and quantitation be
accomplished by comparing the peak pattern observed in a sample with the
peak patterns from commercial Aroclor mixtures. This approach of patten
matching is appropriate when the pattern in the sample is similar to an
Aroclor standard and when the analyst is experienced in PCB measurements.
However, the practice of pattern matching can be difficult when a sample
contains a heavily weathered or an incompletely treated Aroclor. Use of
Aroclor pattern matching is also a questionable practice when the sample
contains more than one Aroclor or PCBs created as an industrial side
product (e.g., chlorination of a still keeper).
Capillary GC analysis increases the complexity of PCB quantitation based
on pattern matching while it provides opportunities to improve PCB
measurement based on congener analysis. The complexity has increased
because 40-100 peaks may be observed for an individual Aroclor using
capillary GC (Figure 1) in contrast to the seven to ten peaks observed
using packed column. Laboratories often utilize poorly documented
quantitation procedures for capillary data that are highly dependent on
the training of the analyst. Differences in quantitation practices
probably contribute to the interlaboratory variability observed for PCB
analyses (i.e., 36% of the 1991 WP PCB values were incorrect2). At the
same time, capillary GC provides a tool to resolve individual PCB
congeners which can be exploited to improve analyses3'*. This can be a
significant effort, chlorination of biphenyl (C12H10) produces 209 PCB
congeners. The IUPAC has adopted a numbering system for the PCB congeners
which is used in this paper.
While Method 8081B provides some guidance on the use of congeners for
quantitation of PCBs in Method 8081, no sample performance data are
provided. In addition, the recommended mixture of congeners needs to be
refined by adding higher chlorinated species. Calculation of
concentrations using a mixture of 12 congeners provides results very
similar to Aroclor-based quantitation for Aroclors 1016, 1232, 1242, and
1248. Unfortunately, calculation of concentrations using the congeners
produces values that are only 45-60% of the values obtained using Aroclors
1254 and 1260.5 Use of a longer list of congeners suggested by Draper of
California's Hazardous Materials Laboratory (HML) provides better
correlation (76-85%) for Aroclors 1254 and 1260.6 However, HML's list
indues forty congeners.
This study was designed to provide the Office of Solid Waste with
performance data for a congener-specif ic analysis of PCBs in solid
matrices. One of the considerations in designing this study is the
requirement for a five-point calibration of all target analytes. It is
also hoped that these data will assist the Methods Section in establishing
whether congener-specific analysis of PCBs has application for the RCRA
program.
361
-------
1 .
: cs
SOOO-j
6OOO-
]
4 O O O 4.
1 O
3O
1 O
30
2.
1 O<
co
- C5
-' SO
5OOO-
O
1 O
20
Figure 1. Chromatograms of Aroclors 1242, 1254 and 1260
362
-------
EXPERIMENTAL PROCEDURES
SELECTION OF CONGENERS During the first stage of this study, candidate
quantitation congeners were selected and chromatographic separations of
the compounds were established. Each congener solution was analyzed
individually to establish the retention time. The retention times of the
congeners were compared with the retention times of the major peaks of
Aroclors 1242, 1254 and 1260. These results were compared with Draper's
findings6. A group of 19 congeners were selected which represented the
major peaks in Aroclors. Table 1 lists the congeners and a chromatogram
of the standards is provided in Figure 2.
Comparison of the response factors generated in this study with the
relative response factors reported by Mullen* shows considerable
differences between the two data sets. The response factors generated in
this study were based on standards in hexane purchased from Ultra
Scientific. Mullen's relative response factors were obtained using
standards from S. Safe of Texas A&M. Because of the difficulties in
preparing pure congeners, it is recommended that additional analysis of
standards be completed to establish response factors for the selected
congeners. It is also recommended that relative response factors be
included in any specific-congener method as a quality control measure.
CHROMATOGRAPHIC CONDITIONS Some analysts recommend 60 m columns with
very slow temperature ramps (e.g., l°C/min, Mike Mullen) in order to
separate almost all the PCB congeners. This approach requires run times
from 90 to 120 minutes, results in short column lifetimes and allows only
6-7 analyses in a 12 hour shift. As the object of this study was to
develop a practical and routine analytical procedure, use of a 30 m DB-5
column seemed more appropriate for this procedure. The carrier gas was
nitrogen to demonstrate that these separations could be achieved in
commercial laboratories.
GC Conditions
Initial temperature
Temperature ramp
Final temperature
Run time
150°C, hold 0.5 min
5°C/min
275°C, hold 9.5 min
35 min
Carrier
Makeup
Injector
Detector
Column
6 mL/min, nitrogen
54 mL/min, nitrogen
packed, 200°C
electron capture, 300°C
30m X 0.53mm id, DB-5
363
-------
TABLE 1 - CONGENER MIX
Congener
2-Chloro
2,3-Dichloro
2,2' ,5-Trichloro
2,4' ,5-Trichloro
2,2' ,5,5'-Tetrachloro
2,2',3',5'-Tetrachloro
2 , 3 ' , 4 , 4 ' -Tetrachloro
2,2' ,4,5,5'-Pentachloro
2,2' ,3,4,5'-Pentachloro
2,3,3' ,4' ,6-Pentachloro
2,2' ,3,5,5' ,6-Hexachloro
2,2' ,4,4' ,5,5'-Hexachloro
2, 2', 3, 4, 4', 5' -Hexachloro
2,2' ,3 ,4, 5, 5 '-Hexachloro
2,2' ,3,4' ,5,5' ,6-Heptachloro
2,2' ,3,4,4' ,5'-, 6-Heptachloro
2,2',3,4,4',5,5' -Heptachloro
2,2' ,3,3' ,4,4' ,5 -Heptachloro
2, 2', 3,3', 4, 4', 5, 5', 6 -Nonachloro
Decachlorobiphenyl (surrogate)
IUPAC#
1
5
18
31
52
44
66
101
87
110
151
153
138
141
187
183
180
170
206
209
Rt
6.52
10.07
11.62
13.43
14.75
15.51
17.20
18.08
19.11
19.45
19.87
21.30
21.79
22.34
22.89
23.09
24.87
25.93
30.70
32.63
Column flow rate - 6 mL/min
364
-------
2 . Oeo-
30
FIGURE 2 Chromatogram of Specific Congeners
365
-------
Five-point calibrations were performed using the congener mixture
according to the criteria specified in Method 8000 of SW-846. The
calibration mixtures and the RSDs are reported in Table 2.
TABLE 2 - INITIAL CALIBRATION DATA
IUPAC*
1
5
18
31
52
44
66
101
87
110
151
153
138
141
187
183
180
170
206
209
(surrogate)
Rt
6.52
10.07
11.62
13.43
14.75
15.51
17.20
18.08
19.11
19.45
19.87
21.30
21.79
22.34
22.89
23.09
24.87
25.93
30.70
32.63
Calibration solutions ng/mL
25, 100, 250, 500, 1000
25, 100, 500, 1000, 5000
25, 100, 500, 1000, 2500
25, 100, 500, 1000, 5000
25, 100, 500, 1000, 2500
25, 100, 500, 1000, 5000
25, 100, 500, 1000, 5000
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 1250
25, 125, 250, 500, 1250
12.5, 50, 125, 250, 500
25, 50, 250, 500, 2500
25, 50, 250, 500, 2500
25, 125, 250, 500, 2500
25, 125, 250, 500, 2500
5, 25, 250, 500, 2500
5, 25, 250, 500, 2500
% RSD
14.1
4.6
10.5
3.1
9.1
4.2
6.1
6.4
9.4
7.0
4.8
9.2
13.4
7.8
5.9
5.3-
19.6
23.0*
5.6
7.3
*reduce calibration range
Column flow rate - 6 mL/min
Sulfuric acid and sulfur cleanups were evaluated to ensure that they did
not result in the loss of Aroclors or congeners. Although these
techniques are in common use in environmental laboratories, there is
366
-------
Little documentation that they can not destroy any PCBs7. A mixture of
congeners, Aroclor 1242, Aroclor 1254, and Aroclor 1260 were each
subjected to both cleanup procedures. Peak areas of the individual
congeners were compared before and after cleanup to determine congener
recoveries. The sum of the areas of six or seven peaks from each of the
Aroclor standards were compared before and after cleanup to determine
Aroclor recoveries. The results for the cleanup of Aroclor 1254 are
presented in Table 3 as illustration of the fact that neither cleanup
destroys PCBs.
TABLE 3 Cleanup of Aroclor 1254
Rt
14.26
15.01
16.65
18.92
20.82
21.82
total
Isomer #
52
44
66
110
153
138
Area of std
22824
9401
41318
35594
33339
34882
192790
Area after
cleanup
25170
10149
42677
36032
33964
34839
197521
% Recovery
103
Column flow 7 mL/min
RESULTS
Performance testing of congener quantitation was conducted using two EPA
freeze dried soil samples provided by the RT Corporation and two spiked
soils. PCBs were quantitated using the specific-congeners and Aroclor
standards. The EPA Soil Group 1 sample was contaminated with a reported
concentration of 44.1 mg/Kg of Aroclor 1242. The EPA Soil Group 2 sample
was contaminated with a reported concentration of 2.08 mg/Kg of Aroclor
1254. San Diego soil sample spiked with Aroclors 1254 and 1260 were
spiked at the 500 ppb level. A aliquot of 5 grams of each sample was
mixed with anhydrous sodium sulfate and extracted using a Soxhlet
apparatus (Method 3540) with methylene chloride as a solvent. A duplicate
determination of the spiked Aroclor 1254 was also prepared and analyzed.
The extracts were exchanged to hexane prior to analysis, no cleanup was
employed. Three of the soils were also extracted using a Dionex Model 703
supercritical fluid extractor with carbon dioxide at 100°C, 300 atm for 10
minutes static and 45 minutes dynamic. The restrictor was maintained at
150°C and the receiving vial at 4°C. The vial was filled with 15 ml of
hexane and decachlorobiphenyl as an internal standard. The recovery of
367
-------
internal standard was used to correct for the loss of solvent that
occurred during the extraction. Each extract was analyzed using a 30 m
DB-5 column using the conditions described above. Tables 4 and 5 present
the recoveries from the EPA soils determined using Soxhlet extraction
based on congener and Aroclor quantitation as well as supercritical fluid
extraction (SFE) using Aroclor quantitation. Table 6 provides the amount
of individual congeners determined in duplicate soils spiked with Aroclor
1254 as well as recoveries for Soxhlet and SFE. Table 7 provides recovery
data for soil spiked with Aroclor 1260.
In all cases, the concentrations based Aroclor standards were larger than
those obtained using congener analysis. The recoveries based on congeners
were lowest for the EPA soils in part because interference prevented
measurement of congener #66. It is worth noting that while the congener
values are also lower in the spiked samples, they are closer to the
nominal values than were obtained using Aroclor quantitation
TABLE 4 - Group 1 Soil (44.1 ppm Aroclor 1242)
Recovery
category
TOTAL (ppb)
TOTAL (%)
Recoveries
Quantities based
on congeners
using Soxhlet
22744 ppb
51.6%
Quantities
based on
Aroclors using
Soxhlet
49800 ppb
113%
SFE (by
Aroclor)
79.9%
column flow rate 6mL/min
SFE recovery = apparent recovery/(surrogate recovery*100)
TABLE 5 Group 2 Soil (2.08 ppm Aroclor 1254)
IUPAC*
TOTAL (ppb)
TOTAL (%)
Recoveries
Quantities based
on congeners
using Soxhlet
1363
65.5%
Quantities
based on
Aroclors using
Soxhlet
2202
106%
SFE (by
Aroclor)
73.2%
column flow rate 6mL/min
SFE recovery = apparent recovery/(surrogate recovery*100)
368
-------
TABLE 6 Duplicate Spiked Soil (500 ppb Aroclor 1254)
IUPAC*
52
44
66
101
87
110
151
153
138
141
187
183
180
170
206
209
TOTAL (ppb)
TOTAL (%)
Rt
14.83
15.59
17.24
18.15
19.53
19.96
21.45
21.88
22.41
24.96
25.04
32.63
Recoveries
Concentration
of congeners
by Soxhlet
36.5, 40.0
10.5, 11.2
82.6, 88.2
66.3, 71.4
107.5, 109.8
3.7, 4.6
59.4, 59.4
3.6, 3.9
54.0, 59.4
3.7, 4.4
3.2, 3.9
102, 101
431
86.2%
Aroclor
cone
X
X
X
X
X
X
661
132%
SFE by
(Aroclor)
X
X
X
X
X
X
218
83%
column flow rate 6mL/min
SFE recovery = apparent recovery/(surrogate recovery*100)
369
-------
TABLE 7 - Spiked Soil (500 ppb Aroclor 1260)
IUPAC*
66
101
87
110
151
153
138
141
187
183
180
170
206
209
TOTAL (ppb)
TOTAL (%)
Rt
17.25
18.17
19.44
19.96
21.41
21.88
22.41
22.97
23.17
24.95
26.03
30.85
32.63
Recoveries
Concentration of
congeners
by Soxhlet
15.7
11.7
-
82.8
33.2
70.4
18.9
40.8
34.1
9.4
85.9
18.8
2.4
99%
424.1
84.8%
Aroclor cone
X
X
X
X
X
X
687
137%
column flow rate omL/min
CONCLUSIONS
The approach of using a mixture of 19 specific congeners appears to
be a viable approach to measuring PCBs in the environment. Recoveries of
soils spiked with Aroclors 1254 and 1260 determined using the specific
congener mixture were between 80 and 90%. Aroclors in environmental
samples were recovered at 51-66% of the certified values.
While this approach is promising, it is not yet ready to be
distributed as formal method. At least one of the selected congeners
(#153) coelutes with another Aroclor congener. In addition, the correct
response factors for the congeners relative to the surrogate
decachlorobiphenyl needs to be established. Finally the list of congeners
370
-------
should be further evaluated in terms of common matrix interferences. For
example, isomer # 66 was buried beneath a chlorinated pesticide peak,
possibly DDE or DDT, in the EPA soil samples. Despite these
shortcommings, congener specific analysis could be applied to more
environmental and waste studies.
REFERENCES
1. Brinkman, U.A. and A. De Kok. In: Halogenated Biphenvls. Terphenyls.
Naphthalenes. Dibenzodioxins and Related Products. R.D. Kimbrough (ed.),
Elsevier. North Holland, Amsterdam (1980).
2. Rohrlich, T., "Waste Pile of Data on Pollution."Los Angeles Times.
p.l, September 13, 1992.
3. Alford-Stevens, A., T.A. Bellar, J.W. Eichelberger and W.L. Budde.
"Method 680. Determination of Pesticides and PCBs in Water and
Soil/Sediment by Gas Chromatography/Mass Spectrometry," EMSL-Cincinnati,
U.S. EPA, November 1985.
4. Mullin, M.D.; C.M. Pochini; S. McCrindle; M. Romkes; S.H. Safe; and
L.M. Safe, "High-Resolution PCB Analysis: Synthesis and Chromatographic
Properties of all 209 PCB Congeners," Environ. Sci. Technol.. 18, 466-476,
June 1984.
5. Marsden, P., "Evaluation of Methods for Analysis of Polychlorinated
Biphenyls (PCBs)" EPA/600/X-90/222, Environmental Monitoring Systems
Laboratory Las Vegas, 9/1990.
6. Draper, W.M. and D. Wijekoon. "A New and Improved Technique for
Speciation and Quantitation of Aroclors in Hazardous Wastes," Proceedings
of the Seventh Annual Waste Testing and Quality Assurance Symposium,
Office of Solid Waste and Emergency Response, U.S. EPA, Washington D.C.,
July 1990.
7. Copeland, G.B. and C.S.Gohmann. "Improved Method for Polychlorinated
Biphenyl Determination in Complex Matrices," Environ. Sci. Technol. 16(2):
121-124 (1982).
371
-------
AQ DIRECT DETERMINATION OF TCLP PHENOLS AND
H0 HERBICIDES BY HPLC
R. L. Schenley. J. E. Caton, and W. H. Griest
Analytical Chemistry Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37831-6120
A method for determining TCLP acidic semivolatile compounds (cresols, chlorophenols, and herbicides)
is being developed to permit radioactive mixed wastes and their leachates to be analyzed with a minimum
of operator exposure to sample radioactivity and very little laboratory waste generation versus current
GC-ECD methods. The HPLC method features direct injection of a 25 /tL aliquot of aqueous waste or
leachate (with filtration and acidification, if needed) onto a 150 mm X 4.1 mm ID (5 urn particles) PRP-1
polystyrene column. The mobile phase is a gradient of 0.1 M acetic acid in 10/90 (v/v) acetonitrile/water
and acetonitrile. Quantitation is by the method of external standards at wavelengths of 235 and 280 run
using a diode array UV detector. Compound confirmation is by UV spectrum or retention tune on an
ODS column. The detection limits meet TCLP Regulatory Limits, and run tunes are 16 min per sample,
including mobile phase reequilibration. Matrix spike recoveries at or below the Regulatory Limits in
TCLP Fluid No. 2 ranged from 88.6 to 128.2 %, but were much lower in alkaline, high-nitrate wastes
simulating nuclear wastes. Surrogate standard recoveries determined over the course of 12 months for
2,4,5-T spiked into 61 TCLP leachates at 5.0 to 7.5 mg/L averaged 96.7 ± 14.8%. A better
confirmation column phase and mass spectrometric confirmation will be added in the future.
Research sponsored by the Office of Technology Development, Laboratory Management Division under
U. S. Department of Energy contract DE-AC05-84OR21400 with Martin Marietta Energy Systems, Inc.
Th« uAfritttd mmmavt hM bun
authored by a omu»cior of th» U.S.
Govwmwnt mtar contract No. OE-
AC06-840R21400. AecorrJnoty. th« U.S.
GoMmmmt rettira • noraxcfcam.
royiltv-frM totraa to pubfah or reproduce
th* puMMwd form of ttn ctmutoiuon, or
•low ottwr* to do to. tor U.S. Govwnrram
purpoam."
372
-------
49 Approaching the Sensitivity of an Electron Capture Detector (BCD) for
the Analysis of Pesticides by using GC/MS
Linda C. Doherty and Norman Low, Scientific Instruments Division Hewlett-Packard
Co., 1601 California Ave., Palo Alto, CA 94304
ABSTRACT
Pesticides are analyzed routinely by GC/ECD with dual column confirmation. The ECD
has excellent sensitivity and can be used to determine some pesticides at levels in the
low femtogram range. Unfortunately, it is easily contaminated and lacks the linearity of
other detectors. Also the chromatography must be optimized, to eliminate any co-
eluting analytes.
We decided to approach the analysis of pesticides using a selected ion monitoring
(SIM) based GC/MS method. The MS tolerates dirty samples much more readily than
the ECD and is easier to clean. SIM has the benefit of giving up to 4 orders of
magnitude linearity, much better than an ECD. It also accommodates co-eluting
analytes by allowing for their separation based on mass. This speeds up method
development time.
The USEPA Method 60S1 pesticides were used for these experiments. (PCBs,
toxaphene and chlordane were excluded.) Initially a standard at the 100 pg level was
analyzed by electron impact (El) GC/MS in the scan mode (60-430 a.m.u.). One or two
significant ions were chosen for each pesticide from the mass spectrum. Because
pesticides have unique El fragmentation patterns, the specific ions chosen were
indicative of the pesticide. The SIM method was developed using those ions. The
pesticides were separated into elution groups to get the best results. Four groups of
ions were used. Standards were injected over a 30 fold concentration range to best
mimic the range required by the Contract Laboratory Program 3/90 Statement of
Work2 All pesticides were linear over this range. The hexachlorocyclohexanes
standards (BHCs) were analyzed from 0.1 pg to 1000 pg. Standards were then spiked
with reference gas oil and analyzed in both SIM and scan mode. Solutions were spiked
at both the 0.1 and 1.0% levels.
INTRODUCTION
Historically the ECD has been used for the identification and quantitation of chlorinated,
electrophilic pesticides. As mass spectrometers improve in sensitivity other approaches
to the analysis of pesticides need to be tried. Mass spectrometry affords the analyst
the ability to obtain a third dimension of information to confirm the presence or absence
of a certain pesticide. The following work explores the use of selected ion monitoring
(SIM) for the analysis of chlorinated pesticides.
373
-------
EXPERIMENTAL
The GC/MS system was composed of an HP 5890 Series II GC with Electronic
Pressure Control (EPC). EPC allows the user to control the pressure or flow of GC
carrier gas accurately. The autosampler was an HP 7673B ALS. The mass
spectrometer was the HP 5972A. The GC was installed with an HP-5MS column (30 m,
0.25 mm ID, 0.25 \w\ film). A single tapered, deactivated glass liner with glass wool
(HP P/N 5062-3587) was installed in the injection port.
250° C
85° C, 1.5 minute hold
30°C/minto190°C
3.6° C/min to 240° C
1.5 minutes
15 psi for 1.5 minutes, then
constant flow
280° C
2 stops (1 \L\, 5 nl syringe)
GC:
Injection port
Initial oven temperature
Purge on
EPC program
Transfer Line
Injection size
MS:
SIM (4 groups)
Group 1 start time 6.7 minutes
Group 1: m/z= 180.9, 108.95
Group 2 start time 8.7 minutes
Group 2: m/z= 100, 66, 240.8, 196.9
Group 3 start time 12.7 minutes
Group 3: m/z= 246, 79.05, 81, 67, 194.9
Group 4 start time 15 minutes
Group 4: m/z= 234.9, 271.8
max sensitivity autotune
EM = tune + 700 V
Group 1 included the four hexachlorocyclohexane pesticides (BHCs). Group 2 included
Heptachlor, Aldrin, Heptachlor Epoxide and Endosulfan I. Group 3 included 4,4'-DDE,
Dieldrin, Endrin, Endosulfan II, 4,4' ODD and Endrin Aldehyde. Group 4 included
Endosulfan Sulfate and 4,4'-DDT.
Some standards were prepared from a 100 ppb pesticide solution (HP Pesticide
Evaluation Sample P/N 5062-3589). Solutions for the standard curves were made with
a stock solution diluted from a Method 8080 standard (HP P/N 8500-6011) . All
SCAN
60-430 amu
A/D =2A2
Threshold = 0
max sensitivity autotune
EM = tune + 700 V
374
-------
solutions were prepared in iso-octane. ASTM Reference Gas Oil #1 (HP P/N 5060-
9086) was spiked into standards for the "dirty matrix" experiments.
RESULTS
Scan Mode:
All pesticides were analyzed, as standards, down to 10 pg injected. The data collected
in scan mode gave quality library searches at 25-50 pg depending on the pesticide.
The standard curves were linear from 25-1000 pg for all of the pesticides. Certain
pesticides gave three orders of magnitude linearity. The sensitivity level was
dependent on the pesticide. Some of the compounds fragmented easily. Figure 1
shows the mass spectrum of Endosulfan I versus DDE. DDE can be seen at much
lower concentrations than Endosulfan I. This is because the ion current for Endosulfan
I is spread over many masses and not concentrated into one or two masses.
Endosulfan I
.",V,Tt..lll,,.ll'r, ..;..;,,
300 320 340 380 380 400 420
150000
100000
50000
2-
178
105 |
87
82 J \ 1
i. i.fAtL
7 i«
11 1 I 233
IL. L .ill JL ,. i,l . I
4,4'DDE
31 a
I
281
I fc
100 120 140 160 180 200 zfo' 2^0 ' 2efo idti ' ido 320 340 3^0 ' '3j(o' ' 4yd VjV
Figure 1.
A comparison of the spectra for two pesticides to show the amount of
fragmentation. Notice the difference in maximum abundance counts.
The largest ion for Endosulfan is about SOX smaller than m/z= 246 for
DDE.
The most prominent ion (the base peak) in the spectrum also affects the ultimate
detection limit. The base peak is generally used as the quantitation ion. Figure 2
shows the spectrum of DDE versus Dieldrin. The base peak in DDE is m/z= 246 were
the base peak for Dieldrin is m/z= 79. The higher the mass of the ion, the less likely
that there will be co-eluting background at the same mass. In Figure 3 a chromatogram
of a standard at 100 pg is displayed.
375
-------
UUAfalttt
150000
100000 .
50000
0
200000 .
100000
0.
ntz->
62 j
J... I.
'bo" 'si
7
66
Vr "
60 8I
10
87
I
.jl Ajl
J 100
1
94
Vob
5
12
J,
Vi
D8
1-H
178
I
^ 21
L, A j. - l,ll r 1" ,
' ' V4b' ' 'ieb' ' 'ido ' '20^0 '
143 173 193 207
>' i4b ido' is'o 2ob
2«
Ho ' 24ld
237
•n Ti'lVV
220 240
™ 4, 4'-DDE
34*0' 3efd M'O 4clo' 42"o' '
Dieldrin
^277
1 L h 291 316 345 334
-(|l)j A - ,. ... ...
260 2SO 300 320 340 360 380 400 420
Figure 2. Spectra for two pesticides where the base peak differ in mass. The
higher the mass of the base peak, the less background interference.
Signal-to-noise improves at higher masses. Notice the abundance
counts for the base peak are similar.
100 pg 808 peehodes 15 pei ml; Run al
02 57 PMPSTcn Kfcn FebOI. 1093
V»]IW
Figure 3. The total ion chromatogram of a 100 pg pesticide standard analyzed in
the scan mode.
One benefit of mass spectrometry is the ability to separate co-eluting peaks without
extensive chromatographic method development. Figure 4 shows an example of this.
Dieldrin and DDE are co-eluting but still easily quantifiable.
376
-------
3
r, -i
DDE
Dieldrin
Figure 4. Separation of co-eluting DDE and Dieldrin by mass.
Standards were then spiked with reference gas oil #1 (RGO #1). A standard 100 pg
solution was spiked at the 0.1 % (1 |il to 1000 |il) and 1% levels (10 pJ to 1000 |il). DDE
and ODD gave high quality library searchable results at the 0.1% spiked level.
SIM
SIM experiments were run to get the lowest detection limits. SIM also gives better
quantitation reproducibility over scan. Since SIM allows for a higher rate of data
acquisition, more points are collected across a GC peak and the GC peak is better
defined and integrated. For dirty samples, SIM will reduce background noise,
increasing the sensitivity.
Standards were prepared and run from 5 pg to 1000 pg for all pesticides. The
hexachlorocyclohexane pesticides were analyzed down to 100 fg. Figure 5 shows the
calibration curve for Lindane. The RGO spiked standards were also analyzed in SIM.
Figure 6 gives a SIM/scan comparison for DDE in the presence of the RGO matrix.
Response = 1 76ft+OCB ' Ami - 8 OSe+003
COT Coef = 0 998 Cufw FT Lkwar
Figure 5.
The calibration curve for Lindane from 100 fg to 1000 pg injected
splitless.
377
-------
Abundance
1000-g
900-=
800^
700 \
600^
500 \
4001
3004
20QJ
P.P-DDE
SIM
800-;
700-E
600.;
500^
400^
300-j
200-1
JJ
Scan
Figure 6. A comparison between SIM and scan for the 0.1% spike standards. The
concentration for DDE is 100 pg.
SUMMARY
Results look favorable for substituting a mass spectrometer for an ECD. Detection
limits of MS approach those of an ECD and give a much less ambiguous confirmation
for the pesticides. By using one or two unique ions per pesticide for confirmation, false
positives and negatives are much easier to eliminate.
1USEPA Method 608, Document Number EPA-600/4-82-057, July 1982
2USEPA Contract Laboratory Program Statement of Work for Organics Analysis, Document
Number OLM01.8, August 1991
378
-------
50 DETERMINATION OF POLYNUCLEAR AROMATIC
HYDROCARBONS IN SOIL AT 1 ^g/kg USING GC/MS
Bruce N. Colbv and C. Steve Parsons, Pacific Analytical, 6349 Paseo Del Lago,
Carlsbad, California 92009, and James S. Smith, Trillium, 7A Graces Drive, Coatesville,
PA 19320
Abstract: A GC/MS method has been developed which is capable of determining PAHs
below 1 pig/kg in soil samples. It uses a 100 g initial sample, soxhlet extraction using
toluene, extensive sample cleanup including alumina column fractionation, and a 300 jiL
final extract volume. Measurements are made using selected ion monitoring techniques
on a quadrupole GC/MS.
INTRODUCTION
Because several polynuclear aromatic hydrocarbons (PAHs) are potent carcinogens, it is
appropriate to determine them at relatively low concentrations in environmental samples.
At present there are two accepted ways to measure PAHs in soils, Method 8270 and
Method 8310. Method 8270 is a reasonably robust GC/MS method with indicated PAH
detection limits of 660 |ig/kg. Method 8310 is an HPLC-fluorescence method with
indicated detection limits ranging from 9 to 1550 Hg/kg depending on the analyte.
Because of its generally lower detection limits, Method 8310 is frequently identified as
the method of choice for field studies. Unfortunately, Method 8310 is often unable to
provide these lower detection limit values due to interference from non-target materials
present in field samples. This is especially true in situations where alkyl-PAHs are
present.
The advent of more sensitive GC/MS instrumentation has made it possible to detect
PAHs at considerably lower levels than those identified in Method 8270. When this
more sensitive instrumentation is combined with extensive sample cleanup, it is possible
to generate PAH detection limits which are less than 1 Hg/kg.
METHOD SUMMARY
A 100 g aliquot of soil sample is fortified with the labeled analogs (Cambridge Isotopes)
of the target PAH analytes. Moisture is removed from the sample by boiling the sample
in toluene using a Barrett moisture trap attached to a boiling flask. The dehydrated
sample and toluene are then transferred to a soxhlet apparatus to generate the sample
extract. The extract is subjected to alumina column fractionation (SW-846 Method 3611,
Appendix B). Sulfur is removed using the tetrabutylammonium sulfite procedure if it is
present in substantial quantities.
The final extract is spiked with an internal standard and analyzed by GC/MS (VG
MD800) in the SIR data acquisition mode. Two mass descriptor groups are used which
379
-------
incorporated the quantitation mass for each analyte. The dwell time on each mass is 60
msec and the cycle time is 1 sec. The instrument is calibrated with solutions containing
0.01, 0.1, 1, 10 and 100 ng/^iL of each PAH plus 2 ng/^iL of each labeled (deuterated)
analog of each PAH. No labeled analog was available for indeno(l,2,3-cd)pyrene.
EXPERIMENTAL
Several experiments were undertaken during the course of the study summarized here.
They included a method detection limit (MDL) study, an analysis of the NIST SRM
1941, plus determinations on a variety of field samples many of which were
contaminated with coal tar derived materials. The MDL study was undertaken using
Ottawa Sand as the substrate. Seven aliquots were spiked to 2.0 |ig/kg with each PAH
plus the labeled analogs and then prepared and analyzed as described above.
In another experiment a 1 gram aliquot of the NIST SRM 1941 (river sediment) was
taken and processed as if it were a 10 gram sample. Calculated concentrations were
multiplied by a factor of ten to account for the different sample size.
A wide variety of field samples were prepared and analyzed using the above procedure.
These samples were associated with waste sites involving petroleum and coal tar related
materials contamination.
RESULTS AND DISCUSSION
The GC/MS instrument was readily calibrated for the lower molecular weight PAH
compounds from 0.01 to 10 ng/jiL. For the higher molecular weight PAH compounds
calibration was from 0.1 to 100 ng/(iL. The reason for this was peak shape. The lower
molecular weight PAHs resulted in very narrow peaks, so detector saturation was
encountered with the 100 ng/pL injections. The higher molecular weight PAHs did not
yield such narrow peaks and consequently did not saturate the detector, but for the same
reason they did not produce reliable peak areas for the 0.01 ng/pL injections. SICP
traces for the low point standards for chrysene and benzo(a)anthracene are shown in
Figure 1 and a typical calibration curve for benzo(a)pyrene is shown in Figure 2. After
these experiments were performed, an improved transfer line was fitted to the GC/MS. It
greatly improved peak shape for the higher molecular weight PAHs, but no calibration
curves have been generated since it's installation.
The MDL study proceeded without incident and produced the results shown in Table 1.
All of the MDL values were less than 1 |ig/kg. Percent recovery and analytical precision
values were also calculated from the MDL study data. These are also shown in Table 1.
Analysis of NIST SRM 1941 resulted in generally excellent agreement with the NIST
provided values (Table 2). There may be a slightly positive bias in the measurements but
it is not consistent. Also, the acenaphthylene value was quite different from the NIST
value. The rest, however, were reasonably close.
380
-------
Initial experiments were performed without benefit of sulfur removal but several field
samples as well as SRM 1941 contained substantial quantities of sulfur. With some of
the field samples there was so much sulfur that little else was evident on initial inspection
of the total ion current trace. While the sulfur did not seem to create measurement
problems through ionization supression, the removal step was added to assure that it
would not become a problem. Figure 3 shows a full scan chromatogram for a high sulfur
sample. The broad peak centered at scan 1124 is for S8. The baseline "hump" starting at
about scan 750 and continuing through 2500 is caused by the alkyl-PAHs which interfer
severely with Method 8310 measurements.
Finally, it is important to note that problems were encountered occasionally with
deuterium exchange on some of the labeled PAHs. This should not be surprising because
exchange is almost certainly the mechanism used to generate the labeled analogs. Due to
the potential for exchange, a secondary set of quantitative references was identified for
those compounds having analogs which exhibit exchange. Because this is only an
occasional problem and it is one that can be monitored and solved through the use of the
secondary references, it is not considered a serious problem. Further, Cambridge
Isotopes was contacted with regard to this problem and they have now initiated a
program to develop 13C-labeled PAHs which will be used in future experiments.
SUMMARY
By slightly increasing sample size, using extensive sample cleanup, slightly decreasing
final volume and making use of today's more sensitive GC/MS equipment, it is possible
to determine PAHs in soils at less than 1 ng/kg. Care must be taken to monitor for label
exchange if deuterated analogs are used for quantitation.
Figure 1 - SICP TRACES FOR 0.01 ng/jiL STANDARD
Pacific Analytical
Sanple: 8.B1 ngAtL std
feK15ft14
y.fS
Instrument: VG86
Chrysene-dl2
2638
Benz(a)anthracene-dl2
y.FS
96
Scn669
67B
675 688 685 698 693 7flB 785 718 715 72B
381
-------
Figure 2 - BENZO(A)PYRENE CALIBRATION CURVE
0.1
0)
CO
o
a
(0
0)
DC
1 10
Concentration (ng/uL)
100
Figure 3 - SOIL SAMPLE EXTRACT PRIOR TO SULFUR REMOVAL
Pacific Analytical
Sample: Soil Extract
Instrument: UGB4
382
-------
Table 1 - PRECISION. RECOVERY AND MDL RESULTS
Analyte
Acenaphthene
Acenaphthylene
Anthracene
Benz(a)anthracene
Benz(a)pyrene
Benz(b)fluoranthene
Benz(k)fluoranthene
Benz(ghi)perylene
Chrysene
Dibenz(a,h)anthracene
Fluoranthene
Fluorene
Indeno( 1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
Recovery
(%)
99
93
101
92
124
102
102
95
91
112
101
100
84
111
107
112
Precision
(%sd)
4.4
12.7
3.5
7.2
10.1
6.0
6.1
13.9
4.8
10.3
4.7
13.0
18.0
5.1
4.6
11.5
MDL
Gigftg)
0.27
0.74
0.22
0.42
0.79
0.39
0.39
0.83
0.27
0.73
0.30
0.82
0.95
0.36
0.31
0.81
383
-------
Table 2 - COMPARISON WITH NIST SRM 1941 RESULTS
Compound
Acenaphthene
Acenaphthylene
Anthracene
Benz(a)anthracene
Benz(a)pyrene
Benz(b)fluoranthene
Benz(k)fluoranthene
Benz(ghi)perylene
Chrysene
Fluoranthene
Fluorene
Indeno( 1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
PA
^g/kg
50
220
310
560
610
1100
1100
623
860
2000
110
570
1360
640
1600
NIST
Hg/kg
52
115
228
599
754
864
857
566
702
1401
104
559
1322
603
1238
%
Diff.
-4
63
30
-7
-21
24
25
10
20
35
6
2
3
6
26
384
-------
51 APPLICATION OF LEE RETENTION INDICES TO THE CONFIRMATION
OF TENTATIVELY IDENTIFIED COMPOUNDS FROM GC/MS
ANALYSIS OF ENVIRONMENTAL SAMPLES
Paul H Chen, W. Scott Keeran, William A. Van Ausdale, David R. Schindler, and Dwight F. Roberts
Analytical Services Division, Environmental Science & Engineering, Inc.
P.O. Box 1703, Gainesville, FL 32602
ABSTRACT
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. The non-TCL components identified and reported as
tentatively identified compounds (TIC) are rarely confirmed. The retention indices (RI) of TICs
may provide a means of confirmation. The purpose of this study was to investigate the
applicability of Lee retention indices to the confirmation of TICs for several classes of
compounds. The classes of compounds studied are PAHs, heterocyclic PAHs, n-alkanes, straight
chain primary alcohols, and fatty acid methyl esters. The RIs of these classes of compounds have
been reported by Rostad et al. and were used for comparison with our data. In addition, the RIs
for Appendix IX compounds, free fatty acids, and explosives and their related compounds were
also determined by us in this study. The results show that high reproducibilities between our RIs
and the published data and between our data determined by different GC conditions (normal and
slow ramps) were observed for PAHs and heterocyclic PAHs. However, poor reproducibility
was observed for other classes of compounds studied. It should be noted that high
reproducibilities of RIs determined under different GC conditions can only be achieved if
compounds of interest and RI standards are eluted during the temperature ramping period. The
high reproducibility for PAHs can be explained by the fact that the four Lee RI standards are all
PAHs, the change in their chromatographic retention behavior under different GC conditions
should be similar with other PAHs of interest. Though the reproducibility of RIs for non-PAH
compounds is not good for inter- and intra-laboratory comparisons, they provide the elution order
and position for isomers and homologs which are very useful for compound identification or
confirmation purpose.
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 forward library search of a published mass spectral database. The
results of TCL components analysis are generally accurate. However, non-TCL components
identified and reported as tentatively identified compounds (TIC) are rarely verified. The
retention indices (RI) of TICs may provide a means of confirmation for certain types of
compounds.
385
-------
Kovats retention indices (1), based on a homologous series of n-alkanes under isothermal
conditions, have been widely available for many organic compounds. Lee retention indices (RI),
based on a series of four polycyclic aromatic hydrocarbons (PAHs) as retention index standards,
have been reported by Lee et al. (2) and Vassilaros et al. (3) for a large number of polycyclic
aromatic compounds. These Lee RIs are determined using capillary columns GC operated under
temperature programming conditions. Rostad and Pereira (4) reported Lee RIs determined by
GC/MS for a large number of PAHs and other organic compounds of environmental interest.
The purposes of this presentation are: [1] to compare the Lee RIs determined in our lab with
those published by Rostad et al. and other workers for several classes of compounds, [2] to
compare the Lee RIs determined in our lab run at different GC conditions for several classes of
compounds, and [3] to explain the differences between our RI values and published values and
their applicability for compound confirmation. The classes of compounds that are reported by
Rostad et al. and our laboratories are PAHs (including N, S, and O-heterocycles), n-alkanes,
primary alcohols, and fatty acids methyl esters. The RI values of additional compounds that will
be reported by us in this study are Appendix IX compounds, free fatty acids, and explosives and
their related compounds.
EXPERIMENTAL SECTION
Sample Preparation. Water samples were extracted at pH > 11 and then at pH < 2 with methylene
chloride according to EPA Method 625 (5). Soil samples were extracted with methylene chloride
in Soxhlet extractors according to EPA SW-846 Method 3540/8270 (6). Before extraction, each
sample was spiked with 1.0 mL of surrogate spiking solution which contains 100 /ng/mL each of
acid surrogates and 50 ng/mL each of base/neutral surrogates. Methylene chloride extract was
concentrated to 1 mL with Kuderna-Danish concentrator and analyzed by GC/MS.
GC/MS Analysis. Samples and standards were analyzed on an HP 59^0 MSD and an HP 5988
GC/MS system. For samples run on the HP 5970, the column used was a 25m x 0.2mm i.d.
HP-5 (0.33 /im coating) fused silica capillary column. For the normal ramp, the column
temperature was held isothermal at 40°C for one minute and then programmed at 10°C/min to
310°C and held isothermal at this final temperature for 8-10 minutes. Under this temperature
programming, chrysene elutes during the temperature ramping period, but benzo(g,h,i)perylene
elutes during the isothermal period. In order to have benzo(g,h,i)perylene (the last RI standard)
elute during the temperature ramping period, a slow ramp was also used. In the slow ramp, the
oven temperature was programmed at 4°C/min instead of 10°C/min for the normal ramp. For
samples run on the HP 5988 GC/MS, the column used was a 30m x 0.25mm i.d. DB-5MS (0.50
urn coating) fused silica capillary column (J&W Scientific, Folsom, CA). For the normal ramp,
the column temperature was held isothermal at 40°C for 4 minutes and then programmed at 10°C
per minute to 295°C, and held isothermal at this final temperature for 10 minutes. For the slow
ramp, the oven temperature was programmed at 4°C/min to 315°C and held at this final
temperature for 3 minutes. The mass spectrometer was scanned from 35 to 500 amu per half
second. The extract or the standard was spiked with a mixture of six internal standards (which
contains naphthalene-d8, phenanthrene-dlO, and chrysene-d!2 among other compounds) before
GC/MS analysis according to CLP protocol (7). A forward library search was performed for
386
-------
non-TCL compounds on a Wiley/NBS data base which contains 139,000 different spectra (8).
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 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 Ox - Tz) + 100 * Z
Tz+i Tz
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. Determination of RIs for certain classes or categories of compounds (Appendix IX,
fatty acid methyl esters, primary alcohols, and some PAHs) were made from standard mixtures,
while others were made from actual sample extracts and, if possible, confirmed with standards.
RESULTS AND DISCUSSION
Retention indices (RI) of PAHs determined in this study for both normal and slow ramps are
listed in Table 1. RIs reported by Rostad et al. (4) are also shown in Table 1 for comparison.
When RIs are not available from Rostad for certain PAHs, the values from Vassilaros et al. (3)
are listed for comparison. When RI values are larger than 400 (which means the compounds
elute after chrysene), the values from Vassilaros, or in some cases, from Lee et al. (2), are also
listed in Table 1 for comparison.
The reason that the Rostad's RIs instead of Vassilaros' and Lee's RIs are primarily used for
comparison is that the former's instrumental conditions are very similar to ours. Rostad's RIs
were determined by GC/MS using DB-5 column with He as the carrier gas. Vassilaros' and
Lee's RIs were determined by GC using coated silica SE-52 column with H2 as the carrier gas
and glass capillary SE-52 column with He as the carrier gas, respectively.
Table 1 shows that the RI values obtained by different GC conditions (normal and slow ramps)
in our lab agree very well. Our RI values also agree very well with those reported by Rostad
for PAHs which elute before chrysene. For PAHs which elute after chrysene, large deviations
387
-------
are found between our values and Rostad's values. This discrepancy is likely due to the fact that
under the GC conditions we used for slow ramp, the last RI standard [benzo(g,h,i)perylene] is
eluted during the temperature ramping period, while this probably was not so for Rostad's work.
It should be noted that high reproducibility of the RI determined under different GC conditions
can only be achieved for those compounds having their upper bracketing RI standard elute during
the temperature ramping period. This was stressed by Lee but not by Rostad. For the GC
conditions used by both Lee and Vassilaros, the last RI standard (picene) was eluted during the
temperature programming period. Retention time of benzo(g,h,i)perylene was not given in
Rostad's paper. However, based on the GC conditions and the lower values of RI obtained for
compounds bracketed by chrysene and benzo(g,h,i)perylene, it appears that benzo(g,h,i)perylene
was eluted in the isothermal period. For the PAHs which elute after chrysene, our RI values
agree very well with those reported by Vassilaros or Lee with an exception of
dibenz(a,h)anthracene.
Retention indices of heterocyclic PAHs determined by this study and those reported by Rostad,
Vassilaros, and Lee are listed in Table 2. As shown in Table 2, our values agree very well with
those reported by the previous workers. High reproducibility within our lab run at different GC
conditions and good agreement between our data and the published data for PAHs and
heterocyclic PAHs can be explained by the fact that the RI standards used are all PAHs, the
changes in their chromatographic retention behavior under different GC conditions should be
similar with other PAHs of interest. The results of this study indicate that Lee retention indices
can provide a high degree of confirmation for PAHs and heterocyclic PAHs such as N-, S-, and
0-heterocyclic PAHs.
Table 3 lists the retention indices of n-alkanes determined in this study for both normal and slow
ramps. RIs reported by Rostad are also listed in Table 3 for comparison. As shown in Table
3, poor reproducibility was observed not only for RIs determined by different labs (ours and
Rostad's), but also for RIs measured at different GC conditions in our lab. This can be explained
by the fact that the changes in chromatographic behaviors are different for n-alkanes and the RI
standards (which are PAH) under different GC conditions. It should be noted that highly
reproducible RIs can be obtained if the samples are run on the same column and at the same
conditions. The RIs given in Table 3 can be used for identification of an individual n-alkane in
a series of homologous n-alkanes if the GC conditions are not drastically different from those
used in this work.
In addition to PAHs and n-alkanes, other classes of compounds were also investigated. They
include straight chain primary alcohols and fatty acid methyl esters. The RIs values determined
by us and Rostad for these classes of compounds are listed in Tables 4 and 5. Tables 4 and 5
show that RIs of primary alcohols and fatty acid methyl esters run at different GC conditions are
not very reproducible. The explanation described above for n-alkanes can be applied here also.
Again, the RIs given in these tables can be used for identification of an individual compound in
a series of homologs if the GC conditions used are not drastically different from those used in
this study.
388
-------
Lee RIs of Appendix DC compounds, n-fatty acids, and explosives and their related compounds
were also determined. They are shown in Table 6 to 8. The RIs for Appendix IX compounds
can be used to help the confirmation of these compounds in the samples if their standards are not
analyzed. Poor reproducibility was observed for most of the Appendix IX compounds, while
good reproducibility was observed for those which had the chromatographic retention behaviors
similar to those of PAHs.
ACKNOWLEDGMENTS
We thank P. Dumas and his department for extracting the samples and C. Diaz for her assistance
in GC/MS analysis.
REFERENCES
1. Kovats, E. Helv. Chim. Acta. 1958. 44, 1915.
2. Lee, M.L., Vassilaros, D.L., White, C.M., Novotny, M, Anal. Chem.. 1979. 51, 768.
3. Vassilaros, D.M., Kong, R.C., Later, D.W., Lee, M.L., J. of Chromatog.. 1982. 252, 1.
4. Rostad, C.E., Pereira, W.E., J. High Resolution Chrom. and Chrom. Commun.. 1986. 9,
328.
5. EPA 40 CFR Part 136 Fed. Regist.. 1984. 49 (No. 209).
6. USEPA Test Methods for Evaluating Solid Waste: Physical/Chemical Methods, SW-846,
3rd Edition, 1986.
7. USEPA Contract Laboratory Program Statement of Work for Organic Analysis, Multi-media,
Multi-Concentration, 1988.
8. McLafferty, F.W., Stauffer, D.B. The Wilev/NBS Registry of Mass Spectral Data. Wiley:
New York, 1989.
389
-------
TABLE 1. LEE RETENTION INDICIES OF POLYCYCLIC AROMTATIC HYDROCARBONS
CAS
Compound
Rostad RI
RI from this Study
Normal Slow
Ramp[l] Ramp[2]
496117 IH-Indene, 2,3-dihydro-
95136 IH-Indene
91178 Naphthalene, decahydro-
119642 Naphthalene, 1,2,3,4-tetrahydro-
447530 Naphthalene, 1,2-dihydro-
91203 Naphthalene
91576 Naphthalene, 2-methyl-
90120 Naphthalene, 1-methyl-
92524 l,l'-Biphenyl
939275 Naphthalene, 2-ethyl-
1127760 Naphthalene, 1-ethyl-
581420 Naphthalene, 2,6-dimethyl-
575417 Naphthalene, 1,3-dimethyl-
575439 Naphthalene, 1,6-dimethyl-
Naphthalene, 2-ethenyl-
581408 Naphthalene, 2,3-dimethyl-
571584 Naphthalene, 1,4-dimethyl-
571619 Naphthalene, 1,5-dimethyl-
208968 Acenaphthylene
573988 Naphthalene, 1,2-dimethyl-
Naphthalene, l-(2-propenyl)-
83329 Acenaphthene
644086 1 , 1'-Biphenyl, 4-methyl-
829265 Naphthalene, 2,3,6-trimethyl-
2245387 Naphthalene, 1,6,7-trimethyl-
86737 9H-Fluorene
1730376 9H-Fluorene, 1-methyl-
85018 Phenanthrene
120127 Anthracene
605027 Naphthalene, 1-phenyl-
Phenanthrene, 3-methyl-
Phenanthrene, 2-methyl-
613127 Anthracene, 2-methyl-
4H-Cyclopenta(def)phenanlhrene
Phenanthrene, 4-methyl-
832699 Phenanthrene, 1-methyl-
779022 Anthracene, 9-methyl-
35465715 Naphthalene, 2-phenyl-
Anthracene, 9-ethenyl-
206440 Fluoranthene
129000 Pyrene
168.87
170.83
173.31
195.47
195.85
200.00
221.57
224.53
236.59
239.27
239.88
240.89
243.76
244.06
247.04
247.15
247.62
248.75
249.67
254.98
256.12
264.99
267.43
270.77
289.49
300.00
301.75
312.74
319.19 [3]
319.93 [3]
321.47
321.77 [3]
322.81 [3]
323.79
329.52
330.73
344.68
352.77
169.04
171.05
173.01
195.21
200.00
221.51
224.42
236.34
239.06
239.30
240.42
243.42
243.96
245.81
247.04
246.55
247.00
248.88
249.53
253.94
254.87
264.98
267.94
270.57
289.27
300.00
301.84
319.08
322.52
330.85
334.13
344.78
352.54
200.00
220.98
224.16
245.39
248.48
253.53
254.36
270.28
300.00
301.59
312.91
318.88
319.76
321.16
322.42
322.87
323.67
330.76
334.63
344.83
352.72
(Continued)
390
-------
TABLE 1. LEE RETENTION INDICIES OF PAHs (Continued)
RI from this Study
CAS Compound
Benzo(a)fluorene
243174 Benzo(b)fluorene
Pyrene, 2-methyl-
Pyrene, 1 -methyl -
604535 l.l'-Binaphthalene
Benzo(ghi)fluoranthene
Cyclopenta(cd)pyrene
56553 Benz(a)anthracene
217594 Triphenylene
Benzo(c)phenanthrene
218019 Chrysene
1 -Phenylphenanthrene
Chrysene, 5-methyl-
612782 2,2'-Binaphthalene
1 2-Methylbenz(a)anthracene
7-Methylbenz(a)anthracene
Benzo(b)fluoranthene
Benzo(j)fluoranthene
207089 Benzo(k)fluoranthene
7, 12-Dimethylbenz(a)anthracene
Benzo(a)fluoranthene
Benzo(e)pyrene
50328 Benzo(a)pyrene
198550 Perylene
56495 3-Methylcholanthrene
193395 Indeno(l,2,3-cd)pyrene
215587 Benzo(b)triphenylene
53703 Dibenz(a,h)anthracene
Picene
191242 Benzo(ghi)perylene
Dibenzo(a,l)pyrene
Dibenzo(a,e)pyrene
Dibenzo(a,i)pyrene
Dibenzo(a,h)pyrene
[1] 40°C for 1 min., 40° to 310°C @ 10°C/min
[2] 40°C for 1 min., 40° to 310°C @ 4°C/min
[3] Value from Vassilaros
[4] Value from Lee
[5] Compound eluted during isothermal period
Rostad RI
366.72 [3]
369.17
369.40 [3]
373.45
385.23
389.92 [3]
396.55 [3]
398.77
399.45
391.24
400.00
421.66 [4]
420.20 [3]
421.81 423.91 [4]
439.51 443.13 [3]
443.13 [3]
440.04 444.02 [3]
452.29 [3]
448.69 454.02 [3]
451.27 457.17 [3]
462.09 468.44 [4]
489.49 493.24 [3]
490.13
491.01 496.20 [3]
500.00 [3]
500.00 500.29 [3]
Normal
Rampfl]
366.39
368.66
370.13
373.23
390.83
398.17
398.51
400.00
Slow
Ramp[2]
366.48
368.93
369.91
373.27
390.82
398.41
398.53
399.94
390.89
400.00
421.17
420.40
424.01
421.24
425.90
442.34
442.45
443.05
443.90
446.30
452.22
454.06
457.02
468.88
492.36
493.42
500.00
540.01 [5]
554.76 [5]
560.34 [5]
563.59 [5]
391
-------
TABLE 2. LEE RETENTION INDICIES FOR
CAS
230273
260946
85029
86748
4265252
132649
11095435
132650
239350
[1] Value from
Compound
Quinoline
Benzo(h)quinoline
Acridine
Benzo(f)quinoline
7-Azafluoranthene
9H-Carbazole
2-Azafluoranthene
Benz(c)acridine
9, 10-Anthracenedione
Benzofiiran, 2-methyl-
Dibenzofiiran
9-Fluorenone
Benzothiophene
Dibenzothiophene
Benzo(b)naphtho(2,l-d)thiophene
Vassilaros [2] Value from Lee
TABLE 3. LEE RETENTION INDICIES FOR
CAS
111842
124185
1120214
112403
629505
629594
629629
544763
629787
593453
629925
112958
629947
629970
638675
646311
629992
630013
593497
630024
630035
638686
630046
544854
630057
Compound
Nonane
Decane
Undecane
Dodecane
Tridecane
Tetradecane
Pentadecane
Hexadecane
Heptadecane
Octadecane
Nonadecane
Eicosane
Heneicosane
Docosane
Tricosane
Tetracosane
Pentacosane
Hexacosane
Heptacosane
Octacosane
Nonacosane
Triacontane
Hentriacontane
Dotriacontane
Tritriacontane
2,6, 10-Trimethylpentadecane
Pristane
Phytane
HETROCYCLIC PAHS
Rostad RI
210.32
301.50
303.99
307.30
350.50 [1]
309.22
347.39 [1]
393.41 392
330.53 [2]
184.50
259.75
293.88
201.84
296.03
390.12 389
n-ALKANES
Rostad RI
140.88
161.47
182.68
204.09
222.87
240.29
256.75
272.17
286.86
300.96
317.20
332.62
347.42
361.53
375.03
387.99
400.45
413.20
425.51
437.68
448.93
460.36
471.96
484.94
499.88
.60 [1]
.37 [1]
RIfrom
Normal
Ramp
138.27
159.66
180.44
200.12
218.83
236.21
252.61
268.29
282.99
296.84
312.43
327.63
342.11
356.14
369.44
382.35
394.59
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
275.46
283.84
298.42
RI from this Study
Normal
Ramp
210.37
306.86
350.37
308.63
347.15
392.06
330.26
259.74
293.64
295.85
389.64
this Study
Slow
Ramp
144.03
163.22
183.23
202.83
221.55
239.35
256.12
272.02
287.19
301.71
318.12
333.60
348.41
362.61
376.29
389.39
402.21
415.93
429.45
442.42
454.90
467.01
478.65
489.99
500.99
279.10
288.01
303.08
Slow
Ramp
309.24
392.11
[1] Compound eluted during isothermal period
392
-------
TABLE 4. LEE RETENTION INDICIES FOR SOME PRIMARY ALCOHOLS
RI from this Study
CAS
111706
111875
143088
112301
112425
112538
112721
36653824
112925
629969
661198
Compound
1-Heptanol
1-Octanol
1-Nonanol
1-Decanol
1-Undecanol
1-Dodecanol
1-Tridecanol
1-Tetradecanol
1-Pentadecanol
1-Hexadecanol
1-Octadecanol
1-Eicosanol
1-Docosanol
Rostad RI
155.86
176.31
196.63
216.09
234.41
251.68
283.29
314.03
345.48
374.56
400.57
Normal
Ramp
154.52
175.63
195.98
215.20
233.42
250.63
266.83
282.29
296.98
312.09
342.84
371.04
397.16
Slow
Ramp
157.77
177.62
197.48
216.62
234.77
252.06
268.36
283.90
298.56
315.14
346.29
374.80
401.29
TABLE 5. LEE RETENTION INDICIES OF FATTY ACID METHYL ESTERS
RI from this Study
CAS
111115
110429
111820
1731880
124107
7132641
1 12390
1731926
112618
1731948
1120281
Compound
Octanoic acid, methyl ester
Decanoic acid, methyl ester
Dodecanoic acid, methyl ester
Tridecanoic acid, methyl ester
Tetradecanoic acid, methyl ester
Pentadecanoic acid, methyl ester
Hexadecanoic acid, methyl ester
Heptadecanoic acid, methyl ester
Octadecanoic acid, methyl ester
Nonadecanoic acid, methyl ester
Eicosanoic acid, methyl ester
Rostad RI
187.83
226.04
260.05
275.62
290.32
305.23
321.34
336.74
351.38
365.43
378.87
Normal
Ramp
186.78
224.56
258.35
288.78
-
319.52
-
349.48
-
376.90
Slow
Ramp
188.24
226.16
260.37
291.18
-
322.66
352.84
380.64
393
-------
TABLE 6. LEE RETENTION INDICIES OF APPENDIX IX COMPOUNDS
RI from this Study
Compound
Pyridine
2-Picoline
N-Nitrosomethylethylamine
Methylmethanesulfonate
N-Nitrosodiethylamine
p-Benzoquinone
Ethylmethanesulfonate
Nitrosopyrrolidine
N-Nitrosomorpholine
o-Toluidine
3-Methylphenol
Acetophenone
N-Nitrosopiperidine
a,a-Dimethylphenethylamine
2,6-Dichlorophenol
HexacMoropropene
Resorcinol
N-Nitroso-di-n-butylamine
1 ,4-Benzenediamine
Safrole
Isosafrole
1 ,2,4,5-Tetrachlorobenzene
1 -Chloronaphthalene
1 ,4-Naphthoquinone
m-Dinitrobenzene
Pentachlorobenzene
1-Naphthylamine
2-Naphthylamine
2,3,4,6-TetracWorophenol
5-Nitro-o-toluidine
Diphenylamine/N-Nitrosodiphenylamine
Diphenylhydrazine/Azobenzene
1 ,3,5-Trinitrobenzene
Diallate
Phenacetin
Dimethoate
4-Aminobiphenyl
Pentachlorom'trobenzene
Pronamide
4-Nitroquinoline- 1 -oxide
Methapyrilene
Aramite
p-Dimethylaminoazobenzene
Chlorobenzilate
3,3'-Dimethylbenzidine
2-Acetylaminofluorene
7, 12-Dimethylbenz(a)anthracene
Hexachlorophene
3-Methylcholanthrene
Dibenz(a,j)acridine
1625-C [1]
112.18
116.23
125.97
135.55
146.43
178.41
177.76
175.81
188.31
207.47
215.58
219.97
230.20
228.25
241.72
260.55
263.47
265.58
276.62
276.62
288.47
294.81
299.19
339.60
430.13
453.37
Normal
Ramp
108.15
121.17
124.45
131.02
139.17
143.19
146.35
175.18
175.79
176.52
175.79
175.06
184.19
195.74
204.68
203.65
213.50
213.63
212.86
218.98
234.79
227.13
236.62
241.61
247.45
260.10
262.17
264.23
264.80
272.14
274.57
275.43
284.47
285.04
285.77
290.75
293.92
297.93
297.57
330.31
334.85
363.69
366.03
372.66
381.39
391.13
NA[2]
NA[2]
Slow
Ramp
121.47
130.12
132.65
137.85
144.23
147.33
150.77
176.52
177.39
177.86
178.68
176.57
185.84
195.47
203.40
203.60
216.79
215.35
219.88
235.68
227.19
236.15
241.61
248.35
260.09
262.20
264.62
265.40
272.81
276.00
276.42
287.18
287.02
288.72
292.33
295.26
298.71
300.49
330.82
337.56
364.71
368.20
371,32
380.70
390.32
443.82
468.69
487.56
[1] U.S.E.P.A. Industrial Technology Division Method 1625-C
[2] Compound eluted during isothermal period
394
-------
TABLE 7. LEE RETENTION INDICIES FOR N-FATTY ACIDS
RI from this Study
Compound
n-Heptanoic acid
n-Octanoic acid
n-Nonanoic acid
n-Decanoic acid
n-Undecanoic acid
n-Dodecanoic acid
n-Tridecanoic acid
n-Tetradecanoic acid
n-Pentadecanoic acid
n-Hexadecanoic acid
Hexadecenoic acid
Octadecenoic acid
n-Octadecanoic acid
Normal
Ramp
176.75
198.43
217.59
234.82
249.52
264.58
278.19
293.98
308.82
324.00
321.82
350.07
352.98
TABLE 8. LEE RETENTION INDICIES OF EXPLOSIVES AND
THEIR RELATED COMPOUNDS
RI from this Study
Normal
Compound Ramp
Nitrobenzene 180.05
o-Nitrotoluene 194.65
m-Nitrotoluene 202.07
p-Nitrotoluene 205.47
1,3,5,7-Tetraazatri-
cyclo[3.3.1.1(3.7)]decane 206.78
1,3-Dinitrobenzene 247.64
2,6-Dinitrotoluene 250.87
2,4-Dinitrotoluene 262.61
1,3,5-Trinitrobenzene 285.91
2,4,6-Trinitrotoluene 287.66
4-Amino-2,6-dinitrotoluene 321.91
2-Amino-4,6-dinitrotoluene 330.74
2,4,6-Trinitrobenzenamine 337.56
3,5-Dinitrobenzenamine 321.42
395
-------
52 A QUICK PERFORMANCE-BASED HPLC METHOD FOR THE ANALYSIS OF
POLYNUCLEAR AROMATIC HYDROCARBONS (PAHs)
M.W. Dong and J. M. DiBussolo, The Perkin-Elmer Corporation, 761 Main Avenue,
Norwalk, CT 06959-0250
ABSTRACT
A quick HPLC method for the analysis of polynuclear aromatic hydrocarbons (PAHs) in
multimedia samples is described. The method is based on rapid extraction, direct injection
of the diluted or concentrated sample extracts, fast HPLC separation, and selective
programmed fluorescence detection. The elimination of sample cleanup reduces the assay
time to one hour from sample extraction to data report. A 16-min PAH separation was
developed using a new HPLC column packed with 3-um polymeric CIS column
materials. Method sensitivity was found to be 20 ppb levels for soil samples and low ppt
levels for water samples. UV absorbance detection was used for the quantitation of
acenaphthylene and for peak confirmation. The method was applied to soil, sediment,
water, waste oil, and air particulate matter, and yielded excellent agreement with certified
values from standard reference materials. This method is useful for routine monitoring of
PAHs and for rapid sample screening in particular. Method validation parameters in
terms of precision, accuracy, sensitivity, selectivity, linearity, range, and ruggedness are
documented and compared to those from official methods. The proposed method
provides a'quick turnaround method alternative to EPA Methods 550, 550.1, 610, 8310
and TO-13. Advantages of this method include, cost-effectiveness, reduced solvent
usage, and good recovery of all PAHs including volatile ones. Method limitations and
precautions are also discussed.
396
-------
53 The Extraction and Analysis of Polychlorinated Biphenyls (PCB's) by SFE and
GC/MS. Improvement of Net Detection Levels
D. R. GERE. L. G. RANDALL , C. R. KNIPE, W. PIPKIN
Hewlett-Packard, Uttle Falls Site, 2850 Centerville Rd. , Wilmington, DE 19808
L. C. DOHERTY, Hewlett-Packard Company, Scientific Instruments Division, 1601
California Ave., Palo Alto, California 94304
Abstract
Contaminants in soils, sediments, and solid wastes is are problem today for the general public,
government labs and private industrial labs. This paper examines applying supercritical fluid extraction
(SFE) systems to the detection and analysis of PCBs from an oil matrices, solid wastes, sludges, and
other materials. In this work contaminants are extracted using SFE and analyzed by GC and GC/MSD.
The PCB from oil matrix work was done to explore whether SFE would decrease coextractant
interferences with the GC/MSD and GC/ECD analysis relative to established manual procedures -
thereby, affording lower detection levels and improved quantitation.
Introduction
The detection and analysis of PCB's in soils, sediments, fish tissue and solid wastes is a common problem
today for the general public, government labs and private industrial labs. One of the necessary steps
involves sampling to assess what chemicals and products are present, and to what quantitative extent.
Supercritical fluid extraction, (SFE) is an expedient and cost effect means of removing chemicals from the
solid waste samples as well as transformer oils, food, eggs and many other materials prior to analysis.
Experimental
SFE: A Hewlett-Packard Model 7680T was used without modification. The conditions for the
experiments are described below.
1. Extraction - pressure, 1218 psi ( 84 bar ); extraction chamber temperature, 40 ° C; density, 0.35
g/mL; extraction fluid composition, CC^; static equilibration time, 2 minutes; dynamic extraction time,
25 minutes; extraction fluid flow rate, 3.0 mL/min; resultant thimble-volumes-swept, 28; 75 ml bquid
CC>2 ( @ 4 ° C ) flowed during dynamic extraction.
2. Collection (during Extraction) - trap packing, Hypersil ODS; trap temperature, 40 ° C; nozzle
(variable restrictor) temperature, 45 ° C.
3. Reconstitution (of collected extracts) — rinse solvent, isooctane; collected fraction volume, 1.0 mL;
trap temperature, 40 C; nozzle (variable restrictor) temperature, 45 C; rinse solvent flow rate, 2.0
mL/min; fraction destination, vial #n.
GC : A Hewlett-Packard Model 5890 Series II GC/ECD with splitless injection, and Hewlett-Packard
HP-5 (30 m x 530 um, 0.88 urn film thickness) column was used. The chromatographic system included
the HP 7673 B injector.
A Hewlett-Packard Model 5890 Series II GC with Electronic Pressure Control (EPC), a
397
-------
Hewlett-Packard 7673 Automatic Liquid Sampler with Auto 250/320 urn for automated on-column
injection into a 250 urn ID column, and Hewlett-Packard Models 5972 and 5971 MSD were used. A
Hewlett-Packard HP-5MS column (30 m x 250 urn, 0.25 urn film thickness) with a 5 m retention gap
(250 urn ID) was used; it is a low-bleed column designed specifically for operation with the MSD.
Deactivated glass press-fit connectors were used to connect the retention gap to the column and polyimide
resin was applied to either end of the retention gap for a firm seal.
Materials and Reagents
Solvents. For the PCB's, the supercritical fluid extraction to remove PCB's from the oil samples was
done at the Little Falls site and subsequently the extracts were sent to the Scientific Instruments Division
for analysis. Consequently, Pesticide grade iso-octane from Burdick and Jackson was used as the SFE
reconstitution solvent at the Little Falls site; at the Scientific Instruments Division (SID), iso-octane and
hexane, both OmniSorv grade from EM Science, were used in preparing standards. Also at the SID
laboratory, ethyl acetate, distilled in glass from Burdick and Jackson, was used in the manual solid phase
extraction (SPE) clean-up of transformer oil.
Extraction fluid. Carbon dioxide, SFE/SFC ECD grade from Air Products, Allentown, PA, was used as
the bulk extraction fluid in all the supercritical fluid extractions.
Samples.
The sample of Aroclor 1260 in transformer oil was obtained from Environmental Resource Associates
(ERA), Arvada, Colorado. The Aroclor in the sample had a certified concentration value of 49.2 mg/kg.
Sample aliquots of 100 uL were used for each conventional manual clean-up experiment carried out at the
Scientific Instruments Division site using alumina solid phase extraction cartridges: Extract-Clean ESP
obtained from Alltech (Stock No. 400825, Alumina-N, 500 mg in 19 mL reservoir). (The 'ESP'
cartridges, with teflon reservoirs, are provided especially for mass spectral analysis.) Additionally, in
some of uie SFE work done at the Little Falls site, the oil was dispersed on activated alumina and/or
Chromosorb W-H.P. (80/100 mesh, Manville Products corp., Denver, CO). The alumina was obtained
from Supelco, Inc., Bellefonte, PA (Alumina F-l, Catalog No. 2-0284, 80/100 mesh). The procedure
outlined by Bellar and Lichtenberg ( 1 ) for activating the alumina was followed prior to its use. Where
cited, filter paper disks, cored out of Whatman Qualitative filter paper, Catalog No. 1003-055, are placed
at bom ends of the sample.
Results and Discussion
PCB's are present throughout our environment, and cause concern from many perspectives ( 2 ). The
analysis of samples containing PCB's is usually complicated by two factors (at least). PCB's are present
in trace quantities, and the matrices are usually complex. Usually the matrices include many other
compounds similar in polarity and chemical size. It was observed during mis study (which is still in
progress) that PCB's have solubility characteristics in carbon dioxide that are very similar to transformer
oil components and to a lesser extent similar to fats or lipids. PCB's are very often present in the same
mixture with such things as transformer oils, petroleum products and fats.
In (his paper, we will present data from the selective fractionation of PCB's via supercritical fluid
extraction (SFE), and the subsequent detection and analysis with the HP 7680 T SFE and the HP 5972 A
MSD . The initial study involves the optimization of the SFE extraction conditions for PCB's in
Transformer oil. Further work involved other matrices such as solid wastes and fish tissue. The SFE
cleanup is compared to conventional cleanup such as the use of alumina cartridges and alumina admixing.
Comparisons are drawn for sensitivity using both the SIM mode and Scanning modes with the SFE
398
-------
cleanup. Preliminary conclusions indicate that SFE improves detection limits, and eliminates the
interferences of transformer oil. Also, it appears that the SIM mode works well with such samples, and
thus scanning would not be necessary. The obvious advantages of the MSD over the use of the BCD with
GC will be discussed and summarized.
For some period of time, we have been selectively extracting (fractionating ) cholesterol from the fats
contained in various food products via SFE ( 3 ) as well as selectively extracting Vitamin A (beta
carotene) from cod liver oil ( Figures 1 and 2 ). This experiment was usually optimized for maximum
selectivity and minimal time by judicious choice of temperature and density (pressure) ( Figure 3 ).
A characteristic of supercritical fluid extraction is its amenability to fractionation: adjusting the solvent
power of the extraction fluid (through density, temperature, composition) can lead to selective solvation of
groups of compounds from a sample. Another approach is to superimpose selectivity on the supercritical
fluid extraction process by mixing a chemically active sorbent with the sample so that some of the
components extracted from the sample are preferentially bound to the sorbent — remaining in the
extraction chamber — while others are extracted and collected. In either approach the net result is
somewhat analogous to what is produced by applying large-scale column chromatography for clean-up
and fractionation of complex extracts; however, the whole process is easily automated within an SFE
method. Both approaches have been applied by other research groups to samples which are relevant to the
problem at hand — e.g., the nearly selective extraction of PCB's relative to fats from sea gull eggs by
David et al ( 4 ) by employing an appropriately low density and the selective extraction of chlorinated
pesticides from chicken fat by mixing the sample to be extracted with alumina by King et al (16).
Similarly, in our work within Hewlett-Packard (), we have been mapping extraction conditions at which
fats are largely extracted and those at which fats are largely not extracted (while other components like
cholesterol and Vitamin A are). These regions are outlined in Figure 8. Note that the "Fat Band"
demarcates the division into the regions — at conditions above the band (particularly higher
pressures/densities) large quantities of fats coextract with other components, and at conditions below the
band (lower pressures/densities) much smaller amounts of fats coextract. Within the band, minor to bulk
quantities of fat can be extracted; the level seems to be heavily influenced by the matrix. The designations
of the regions outlined in Figure 9 are not technically rigorous but that diagram is a very useful empirical
tool to use in designing an SFE strategy — as demonstrated by this work on selectively extracting PCB's
from an oil matrix.
One of the first published demonstrations of such experimental fractionation by supercritical carbon
dioxide was by Frank and Sandra ( 4 ) in a paper where they describe the fat free extraction of PCB's
from sea gull eggs. They selectively extracted PCB's away from the associated fats. At one combination
of density and temperature the PCB's were extracted along with the fats (* 35 % by weight). However by
reducing the temperature ( constant density ) it was possible to obtain the PCB extract with only a 3 96
fat content.
We continued this type of exploration into the SFE of PCB's in transformer oil. Extraction of transformer
oil sample was first done at conditions roughly represented by point number " 1" in Figure 3 after having
mixed the oil sample with alumina. This experiment was not successful, possibly due to poor
reproducibility of the procedure for activating the alumina, which is empirical at best, involving many
complex variables such as the origin of the alumina, particle diameter and other parameters beyond the
scope of this study. This was followed by SFE with no alumina sorbent, and extracting the oil sample at
the milder conditions, point "2" in Figure 3. Point 2 was chosen since a nearly fat-free extract of
cholesterol had been obtained in routine experiments and demonstrations from fatty matrices with
similar conditions. This is illustrated in Figure 1 where cholesterol was selectively extracted away from the
trigrycerides ( as well as mono and di glycerides ).
A comparative liquid solid cleanup of the PCB's from the oil matrix with an acid clean-up followed by
extraction with ethyl acetate was carried out. The extracts from the s SFE were compared to the ethyl
399
-------
acetate liquid solid extracts. Aggressive SFE conditions (point 1 in Figure 3), did not accomplish a
significantly better clean-up of die sample compared to the liquid-solid clean-up. However, at the milder
SFE conditions (point 2 in Figure 3), the resultant SFE extracts are noticeably less complex than those
produced by the manual clean-up. This is shown by Figures 4 and 5. The results for the
octachlorobiphenyl homolog series were quite similar to those of the heptachlorobiphenyl homolog series.
Notice mat mere are still some matrix interferences contributing to the background for the
hexachlorobiphenyls. This leads men to carrying out the SFE experiment with conditions indicated as zone
3 in Figure 3. One major goal will accomplished then if the SFE can selectively remove the PCB's from
a transformer oil matrix and effectively lower the mass spectrometer detector minimum sensing limits to
approach those of the clean standards. In this work an external standard curve was generated from 10 pg
to 500 pg. A pattern for a standard solution of Aroclor 1260 could be recognized down to 5 pg injected.
Quantitation was done by choosing the three largest peaks in each homolog series and summing their areas
to give a total for each level of chlorination. The work demonstrated that a S-ion SIM mode was fifteen
times more sensitive man detection in scanning mode while giving as much pattern information as is
needed for identification ( 5 ). This sensitivity approaches that achieved by an ECD. The MSD, used in
the electron impact (El) mode, does not have a problem with response factor variation for the different
PCB congeners. Co-eluting electrophilic peaks mat make it difficult to do ECD quantitation (e.g., DDT
and DDE) will not affect MSD quantitation because they are not of the same mass ( 8 ).
References
1. T.A. Bellar and J.J. Lichtenberg, "The Determination of Polychlorinated Biphenyls in Transformer
Fluid and Waste Oils," United States Environmental Protection Agency, Publication No. EPA-600/4-81-
045, September (1982).
2. Personal communication from Barry Lesnik, Office of Solid Waste, US Environmental Protection
Agency, 401 M Street, SW, Washington, DC 20460.
3. D.R. Gere, " Selective extraction of Cholesterol from Fats in Butter Cookies and other Baked Food
Products" presentation at Midwest AOAC annual meeting, Champaign, 111., June 1992.
4. F. David, M. Verschuere, and P. Sandra, Fresenius J. Anal. Chem. 344. 479-485 (1992).
5. D.R. Gere, C.R. Knipe, P. Castelli, J. Hedrick, L.G. Randall, J. Orolin, H. Schulenberg-Schell, R.
Schuster, H.B. Lee, and L. Doherty"Bridging the Automation Gap between Sample Preparation and
Analysis: SFE, GC, GC/MSD and HPLC Applied to Environmental Samples" manuscript in press,
J.Chronaatog. Sci.
6. I.E. France, J.W. King, and J.M. Snyder, J. Ag. Food Chem. 39, 1871-74 (1991).
7. W. Pipkin, D.R. Gere, L. Frank, and M. Fogelman, "Optimization of Supercritical Fluid Extraction
Conditions for the Extraction of Fats from Food Products," Paper No. 737, The Pittsburgh
Conference.Georgia World Congress Center, Atlanta Georgia, March 8-12, 1993.
8. M.D. Erickson, Analytical Chemistry of PCBs. Butterworth Publishers, Stoneham, MA, 1986.
400
-------
Figure 1. Selective SFE of
Cholesterol from Fats
D
(E
E
10;
9;
i
8:
7:
6
5
4
3:
•
2:
1:
LC fl 210,25 560,40
LQ
of NLE-CHOL.D
of NLE-VITfl.D
Triglycerides/Fats
holesterol Peak
4 6
Time (min.)
8
401
-------
Figure 2. Selective SFE of
Beta Carotene from Cod Liver Oil
Triglycerides/Fats
Vitamin A Peak
4 6
Time (mlh.)
402
-------
Figure 3. Density-Temperature Table
for setting SFE Selectivity
40 50 60 70
Temperature (C)
80 90 fOO IfO 120 130 140
1.00
0.95
0.90
0.85
0.80
0.75
j 0.70
E 0.65
0.60
0.50
§0.45
0 0.40
0.35
0.30
0.25
0.20
0.15
526 618 711 804
383 1463 544 644
281 155] 420 489
211 269 32TU04
•164] 213 264 314
175 218 261
223
097 990 1083
680 764 665
516 626 695
447 500 565
365p416 467 595
305 348 3921436 510
690
260 297 334 3721 425
227 259 f290l 322
93
91 109 1
89 104 122
284
354
311
375
338 365
299 322
266 286
87: #2
84: #3
81
77
IV
60
75
63
flJB 120
101 111
93 100
82
67 71
121
108 116
94 99
76 80
105 110 116
91
He head
space line
403
-------
Figure 4. Comparison of SFE to
Ethyl Acetate Extractoion for Hexachloro
biphenyls in Aroclor 1260
Hexachlorobiphenyl pattern for
Aroclor 1260
**
m
Jill
4
11
50 pg standard
Extracted ion m/F
360
I) M W
tu
HffP 7 T'T'T
U M
JiuU-~*..
SFE of an oi sampte
Extracted ion mft=
360
n't u
OS sample elutedwth
Extracted ion m/z=
300
fll Ml
404
-------
Figure 5. Comparison of SFE to
Ethyl Acetate Extraction for
Heptachloro biphenyls in Aroclor 1260
Heptachiorobiphenyi pattern for
Aroclor 1260
I
. _ .
M | 'I M | M I I .1 ' •
60 pfl standard
Extracted ton mfe= 394
II II tl n't •»
SFEofanoaumpli
Extracted ton mlSF 394
T-'I i-flrrtw*
i j
« Hi u ni
09 sample ekitedwiihEthy Acetate
Extracted ion mftF 394
- . .. ,f * ,f—*~—m— ~
*r i i i 'i i r i i i i i i i i i i » • i « i i
u u t> MI ni
MI
405
-------
54 Environmental Analysis - PAH's in Solid Waste: Bridging
the Automation Gap between SFE and HPLC
D.R. Gere/ C.R. Knipe, William Pipkin and L.G. Randall
Hewlett-Packard Company, Little Falls Site, 2850 Centerville Road,
Wilmington, DE, 19808, USA
ABSTRACT
The detection and analysis of contaminants in soil, sediments, and solid wastes
are a common problem today for the general public, government labs and private
industrial labs ( 1). This work with polyaromatic hydrocarbons is being
undertaken as the first step towards deriving an EPA SFE method similar to the
recent draft method 3560 for total petroleum hydrocarbons (TPH) ( 2 ) . This
present study has been in progress for some time at Hewlett-Packard. The study
focused on the extraction and analysis of PAH's (polyaromatic hydrocarbons,
sometimes called polynuclear aromatic hydrocarbons). There are two
significantly different parts of the study. The primary element of the study
has been to provide the experimental data to validate a tentative alternative
to the liquid solid extraction of solid waste PAH's by SFE. It is relevant and
helpful to slightly modify the standard EPA method 8310 for a couple of
reasons. One simple need is to frequently use an internal standard, and often
times a surrogate as relative measure of the extraction efficiency. Second,
it was helpful to operate the HPLC column at a higher linear velocity than that
called out in the EPA method 8310. This modification allowed increased speed
and resolution. Also, with regard to the analysis part of the overall method,
we made an adaptation of the tentatively optimized (final) SFE steps to better
accommodate the analysis by GC and GC/MS with capillary GC columns.
INTRODUCTION
A considerable effort was directed towards the optimization of the
supercritical fluid extraction steps. The primary parameters that were the
temperature, the density or the pressure and the combination of density and
temperature. Another parameter studied in detail was the use of modifiers or
co-solvents added to the flowing extracting stream of carbon dioxide. Liquid
modifiers can be added in an almost infinite number of permutations of types
and amounts.
Figure 1 is an outline of the study yielding the data in this manuscript. A
brief initial study re-iterated some of the HPLC method steps included in EPA
Method 8310. These iterations resulted in some minor changes to optimize the
time and resolution for sample throughput and robustness.
A second study (in much greater detail) reviewed the SFE extraction in light of
previous studies and experimental apparatus. The optimization in this study
involved two major areas in order to provide recoveries greater than 95 % for
all 16 PAH's described in EPA Method 8310. The first part involved use of sub
ambient solid trapping to ensure the recovery of the relatively volatile
compounds. Second part of the study involved the optimization of the co-
solvents or modifiers useful in recovery of the relatively large and more
intractable (from solid waste)
PAH compounds.
Experimental
406
-------
Materials and Reagents
Reconstitution solvents.
The reconstitution solvents dispensed by the SFE instruments varied according
to the particular analyte class being extracted. Furthermore, these same
solvents were used to create the internal and external standard solutions. For
the polyaromatic hydrocarbons, a 50/50 v/v mixture of
acetonitrile/tetrahydrofuran was used: both were HPLC grade from Burdick and
Jackson.
External calibration standards.
For polyaromatic hydrocarbons, a standard mixture from Supelco, Inc. in
Bellefonte PA, Polynuclear Aromatic Hydrocarbons Mixture 610-M (Catalog No. 4-
8743), or alternately, Hewlett-Packard standard mix # 85-6035was used. The
Supelco standards are a mixture of 16 PAH's at levels ranging from 100 to 2000
ug/mL in a 50/50 mixture of methanol/methylene chloride. The HP standard
contains the same 16 PAH's at the 500 ug/mL level in acetonitrile. These
standard samples were used in the automatic preparation of more concentrations
by serial dilution corresponding to 1/2, to 1/50 the original concentration,
using a 50/50 mixture of acetonitrile/THF as the diluent.
Extraction fluid.
Carbon dioxide, SFE/SFC ECD grade from Air Products, Allentown, PA, was used
as the bulk extraction fluid in all the supercritical fluid extractions. In
the PAH applications, modifiers were added to the bulk C02 extraction fluid.
In the PAH work done at the Little Falls laboratory, these were methanol,
water, and methylene chloride (all, HPLC grade from Burdick and Jackson,
Muskegan, Michigan), forming extraction fluid mixtures of 95/1/4 (v/v/v)
CO2/methanol/water in one method and 95/1/4 (v/v/v) CO2/methanol/methylene
chloride in the other.
Mobile phase solvents for liquid chromatography.
Acetonitrile and water were used; their source and purity are noted above
elsewhere.
Samples.
For the work done on the PAHs from soils, spiked samples were prepared in our
laboratory and also were purchased from Environmental Resource Associates
(ERA), Arvada, Colorado. The following outlines the chemicals and procedure
used at our Little Falls facility. First, soil from a local yard was prepared
by blending 200 g soil with an equal volume of dry ice prepared from SFE/SFC
ECD grade C02 (see above) in a blender for 5 minutes, manually removing the
remaining large rocks and sticks, and then sieving the cold mixture. This
process produces a finely divided soil with its moisture content still intact (
3 ). A PAH standard mixture, Supelpreme-HC PAH Mix (Catalog No. 4-8905),
obtained from Supelco, Inc. (Bellefonte, PA) and having 16 PAH's at levels of
2000 ug/mL was diluted by a factor of 100 using a 50/50 mixture of
acetonitrile/tetrahydrofuran (see above). One-hundred milliliters of the
resultant stock solution was mixed with 100 g of the finely divided soil. The
mixture was gently stirred for 3 hours until the acetonitrile and
407
-------
tetrahydrofuran had evaporated. This resulted in a spike mix of 20 mg/kg of
each PAH in the soil.
Supercritical fluid extraction.
For SFE, Hewlett-Packard Models 7680T and 7680A were used without modification.
In applications requiring modifier addition to the bulk CQ2, Hewlett-Packard
Model 1050 pumps (either isocratic or quaternary models) were used. The pumps
were coupled to the SFE instrument as prescribed by the manufacturer. A
contributed software program controlled the SFE/modifier pump system. Figure 2
depicts a block diagram of the combination of the SFE with a modifier pump with
the modular HPLC to create the "bridge" system used in an automatic sequence to
maximize overall sample throughput and to minimize operator, intervention, thus
providing reduced overall system error (improved sample % relative standard
deviation or % RSD).
Liquid chromatography.
A Hewlett-Packard Model 1050 LC equipped with an MWD diode array detector, 100
vial autosampler and quaternary gradient pump was used without modification.
The separation was carried out on a Vydac C18 (25 cm x 4.6 mm, 5 um particle
size) column.
Analytical Procedure,Extraction
A quantity of three grams of soil was extracted each time in applying the
method to the certified spiked sample for the assay of PAH's from the soil.
When screening with real samples of high concentration, sometimes only 1 gram
of soil was used, admixed with an equal amount of inert material such as sand
or Chromosorb W-H.P. Two similar SFE methods have been developed — one for
analysis by HPLC and one for analysis by GC. In both the HPLC- and GC-focused
SFE methods, there were three extraction steps with a single fraction produced
for the SFE recovery of PAH's in the final method. The purpose of Step 1
(Figure 5 ) was to collect more volatile PAH's. The purpose of Step 2 ( Figure
6 ) was to collect involatile PAH's. The purpose of Step 3 ( Figure 7 ) was to
sweep modifier from the instrument and raffinate before de pressurization and
to reconstitute the involatile PAH's in the same vial containing the more
volatile PAH's from Step 1.
PAH SFE Method for subsequent HPLC analysis.
The following conditions for Step 1 are grouped according to function.
1. Extraction — pressure, 1749 psi; extraction chamber temperature, 120 C;
density, 0.22 g/mL; extraction fluid composition, CO2; static equilibration
time, 2 minutes; dynamic extraction time, 10 minutes; extraction fluid flow
rate, 2.0 mL/min; resultant thimble-volumes-swept, 12. 2. Collection (during
Extraction) — trap packing, Hypersil ODS; trap temperature, 5 C; nozzle
(variable restrictor) temperature, 55 C. 3. Reconstitution (of collected
extracts) — rinse solvent, 50/50 (v/v) THF/acetonitrile; collected fraction
volume, 0.8 mL; trap temperature, 60 C; nozzle (variable restrictor)
temperature, 45 C; rinse solvent flow rate, 1.0 mL/min; fraction destination,
vial #n.
The following conditions for Step 2 are grouped according to function.
408
-------
1. Extraction — pressure, 4899 psi; density, 0.63 g/mL; extraction chamber
temperature, 120 C; extraction fluid composition, 95/1/4 (v/v/v)
C02/methanol/water; static equilibration time, 1 minutes; dynamic extraction
time, 30 minutes; extraction fluid total flow rate, 4.0 mL/min; resultant
thimble-volumes-swept, 25. 2. Collection (during Extraction) — trap packing,
Hypersil ODS; trap temperature, 80 C; nozzle (variable restrictor) temperature,
45 C. 3. Reconstitution (of collected extracts) — none.
The following conditions for Step 3 are grouped according to function.
1. Extraction — pressure, 4899 psi; density, 0.63 g/mL; extraction chamber
temperature, 120 C; extraction fluid composition, CO2; static equilibration
time, 5 minutes; dynamic extraction time, 10 minutes; CO2 flow rate, 4.0
mL/min; resultant thimble-volumes-swept, 8. 2. Collection (during Extraction)
— trap packing, Hypersil ODS; trap temperature, 80 C; nozzle (variable
restrictor) temperature, 45 C. 3. Reconstitution (of collected extracts) —
rinse solvent, 50/50 (v/v) THF/acetonitrile; collected fraction volume, 0.8 mL;
trap temperature, 80 C; nozzle (variable restrictor) temperature, 45 C; rinse
solvent flow rate, 1.0 mL/min; fraction destination, vial #n.
Method for subsequent GC analysis.
The following conditions for Step 1 are grouped according to function.
1. Extraction — pressure, 1749 psi; extraction chamber temperature, 120 C;
density, 0.22 g/mL; extraction fluid composition, CO2; static equilibration
time, 2 minutes; dynamic extraction time, 10 minutes; extraction fluid flow
rate, 2.0 mL/min; resultant thimble-volumes-swept, 12. 2. Collection (during
Extraction) — trap packing, Hypersil ODS; trap temperature, 5 C; nozzle
(variable restrictor) temperature, 55 C. 3. Reconstitution (of collected
extracts) -- rinse solvent, 1/1 (v/v) methanol/methylene chloride; collected
fraction volume, 0.8 mL; trap temperature, 30 C; nozzle (variable restrictor)
temperature, 45 C; rinse solvent flow rate, 1.0 mL/min; fraction destination,
vial #n.
The following conditions for Step 2 are grouped according to function.
1. Extraction — pressure, 4899 psi; density, 0.63 g/mL; extraction chamber
temperature, 120 C; extraction fluid composition, 95/1/4 (v/v/v)
C02/methanol/methylene chloride; static equilibration time, 1 minutes; dynamic
extraction time, 30 minutes; extraction fluid total flow rate, 4.0 mL/min;
resultant thimble-volumes-swept, 25. 2. Collection (during Extraction) —
trap packing, Hypersil ODS; trap temperature, 80 C; nozzle (variable
restrictor) temperature, 45 C. 3. Reconstitution (of collected extracts)
none.
The following conditions for Step 3 are grouped according to function.
1. Extraction — pressure, 4899 psi; density, 0.63 g/mL; extraction chamber
temperature, 120 C; extraction fluid composition, C02; static equilibration
time, 5 minutes; dynamic extraction time, 10 minutes; CO2 flow rate, 4.0
mL/min; resultant thimble-volumes-swept, 8. 2. Collection (during Extraction)
— trap packing, Hypersil ODS; trap temperature, 80 C; nozzle (variable
restrictor) temperature, 45 C. 3. Reconstitution (of collected extracts) —
rinse solvent, 1/1 (v/v) methanol/methylene chloride; collected fraction
volume, 0.8 mL; trap temperature, 30 C; nozzle (variable restrictor)
409
-------
temperature, 45 C; rinse solvent flow rate, 1.0 mL/min; fraction destination,
vial ttn.
Liquid Chromatography
The liquid chromatographic conditions for PAH's were the following: an
injection volume of 10 uL, a column oven temperature of 30 C, a
water/acetonitrile mobile phase which was programmed isocratic 40 % water for 3
minutes followed by 3 different ramps to increase the acetonitrile
concentration to a final value of 100 % for 3 minutes.
Results and Discussion
The extraction from solid samples of polyaromatic hydrocarbons (PAH's) is an
application which has been receiving significant attention with respect to
getting an acceptable, robust SFE method developed and ultimately formalized as
an EPA draft method.
EPA Method 8310 describes the HPLC analysis to be used for PAH's. In this
work, this method was modified slightly for the consideration of analysis time,
and resolution of the 16 prescribed compounds. In order to provide
chromatographic resolution of both the internal ( biphenyl or bromo benzene )
and surrogate (m-quaterphenyl) standards, it was found to be necessary to
change the programming of the mobile phase and to operate at higher linear
velocities. A representative liquid chromatographic separation is presented in
Figure 3.
In developing and optimizing the SFE method, personnel from the Hewlett-Packard
laboratory in Waldbronn, Germany and the Inland Consultants laboratory in
Chicago, Illinois, USA have collaborated with the Hewlett-Packard Little Falls
site laboratory. Additionally, the group of Dr. H.B. Lee of the National Water
Research Institute, Environment Canada in Ontario, Canada further modified the
method outlined here specifically to use with gas chromatography and GC/MS as
the analytical techniques ( 4 ) . Over the last six months, our work has been
directed towards optimizing the supercritical fluid extraction method itself —
including studying the effect of temperature, density (or pressure), and, most
importantly, composition of the extraction fluid. Throughout the course of
developing the method, we utilized three different ways of exposing the sample
in an SFE system to a modifier/CO2 solution: addition of liquid modifier into
thimbles containing the samples (e.g., 1 mL liquid to 10 g soil sample), the
use of premixed tanks of liquids plus CO2 (e.g., 5 weight % methanol in CO2),
and the use of an external pump to meter modifier into the CO2 stream ahead of
the extraction chamber. The method described here can be carried out with
either premixed tanks or the external modifier pump. Without some automated
means of wetting the sample directly with the modifier part way through the
method, the approach of addition into the thimble often provides lower average
recoveries for the whole range of PAH's studied.
The "PAH class of compounds" ( listed in Table 1 ) includes a range of
compounds that is broad with respect to volatility and molecular weight. The
SFE method development described here focused on two (overlapping) groups of
PAH's, with the result that each group is extracted, concentrated, and
reconstituted best by different SFE conditions. We defined these as different
substeps making up a full method. These groups are considered to include the
relatively volatile, smaller PAH's and the less volatile,larger size ( and
molecular weight) PAH's. In Figure 4 the individual compounds are sorted
according to size, going from smaller compounds at the left hand side to larger
410
-------
ones. The recoveries are low at both ends — the most volatile and the least
volatile (largest molecules) -- with a maximum in recoveries for intermediate
PAH's. In Figure 4 have plotted earlier data from a USEPA report (where the
apparatus did not have the luxury of such items as sub-ambient solid trapping
and modifier pumps ) and added the composite data of the three laboratories
participating in this methods development work. The data are outlined in Table
2.
*
I
2
3
4
5
6
7
8
Compound
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
1
9
10
11
12
13
14
15
16
Compound
Benzo (a) anthracene
Chrysene
Benzo (b) fluoranthene
Benzo (k) fluoranthene
Benzo (a) pyrene
Dibenzo (a, h) anthracene
Benzo (g,h, i) perylene
Indeno (1,2, 3-cd) pyrene
Table 2. A Comparison of Recoveries of PAH's
obtained in Prior Work and in This Work
after Method Optimization (% Recoveries)
PAH
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
EPA
Method- 1
(11)
40.0
53.1
80.5
71.4
98.2
65.7
79.3
77.2
61.8
60.0
47.2
47.2
30.1
EPA
Method-
2(11)
50.3
56.4
87.1
77.3
90.8
68.2
83.0
67.1
58.6
53.5
38.7
38.7
22.6
Grand Mean
Mean of
EPA
Reports
45.0
54.5
83.8
74.2
94.5
66.9
81.1
72.1
60.2
56.8
43.0
43.0
26.5
61.7
This
Study -
Little
Falls
103.4
84.5
78.8
119.9
143.3
50.0
104 ..0
66.0
99.4
114.8
95.9
95.1
75.3
99.2
89.3
88.4
This
Study --
Waldbronn
124.0
72.0
117.0
127.0
1000.0
106.0
40.0
156.0
96.0
104'. 0
79.0
77.0
53.0
46.0
112.0
This
Study --
Inland
Env
80.0
97.0
105.0
95.0
Mean of
This
Study
102.4
84.5
75.4
118.4
132.6
75.0
105.0
67.8
127.8
105.3
100.0
88.0
87.0
76.1
67.7
100.2
94.6
411
-------
In this study, the PAHs, have been divided into two groups—the early
eluting ones (in the chromatography) referred to as the relatively
volatile PAH's and the remaining PAH's which are larger in size and
molecular weight (and more intractable for extraction in solid wastes).
The recoveries at either end of the PAH list are improved. What we
believe has been happening in prior work ( 5,6 }is that conditions
appropriate for extracting and collecting intermediate and large PAH's
result in the volatilization of the more volatile compounds — e.g.,
temperatures are too high and extraction times are too long thereby
resulting in too much expanded CO2 sweeping over and through the
collected extracts. Conversely, those same extraction conditions were
not aggressive enough to extract the larger PAH's — particularly, with
respect to temperature and composition. These problems were overcome by
going to a multiple-step procedure using an SFE system which I. has a
solid trapping system capable of thermal operation at sub-ambient
temperatures for volatile collection and high temperatures for operation
with modifiers and 2. also has the capability of automatically changing
the extraction fluid composition from one step to another step.
The details were outlined in the Experimental section. Briefly, part of
the PAH's were extracted and recovered in the first step using pure CO2
at moderately low density and temperature and with cold trapping on an
ODS trap. These were reconstituted into an autosampler vial with just
0.80 mL collected fraction volume. The rest of the PAH's were removed by
a subsequent step using a mixture of CO2 with water and methanol as the
extraction fluid, higher operating temperature and density in the
extraction region, and a higher temperature in the trapping region with
the ODS. The PAH's are not reconstituted directly after the second step.
A short third step with pure CO2 (but with all other conditions as in the
second step) is used to purge the system of modifier before de
pressurization. The analytes recovered in the second step (and possibly,
any moved during the beginning of the third step) are reconstituted in
the same autosampler vial containing the first fraction — with another
0.8 mL collected fraction volume. Therefore, all recovered analytes are
merged automatically into a single fraction to be analyzed.
The data in Table 2 are plotted in Figure 4. The bar graphs show the
percent recovery for the 16 EPA PAH compounds. For instance, this 1st
compound is naphthalene, the last one is indeno (1,2,3-cd) pyrene and
compound number 6 is anthracene.
The solid bar graphs start at a relatively low percentage and increase
and then decrease again, we Two things, a solid trap which can be
operated at sub ambient temperature during the extraction step and the
use of modifiers, have allowed 2 significant advantages. One allows the
recovery of the relatively volatile PAH compounds. The clear bar graphs
represent the mean of the two labs within HP, the private environmental
lab in Skokie , Illinois ( 7 ) and the environmental lab in Burlington,
Ontario Canada. It can be seen that the recovery is significantly
improved for all 16 PAH compounds. This is the resultant of the two
separate phenomena. First, the volatiles are extracted under mild
conditions and rinsed to the reconstitution vial ( from the solvent
trap) before the later step with more aggressive extraction conditions.
Secondly, these more aggressive conditions with the modifier extract the
larger less soluble ( in CO2 only ) . During the extraction step these
412
-------
compounds are precipitated on the solid trap ( which is now at a
substantially elevated temperature ) and then rinsed into the same
vial. The grand mean recovery for the 16 compounds ( multiple sets of
experiments in four different labs in three different countries ! } is
96'. As a test of our hypothesis about the volatiles and the modifier
aiding the large PAHs, we then carried out experiments which divided the
3 steps such that the reconstituted solutes from step one were analyzed,
and then the extracts from steps 2 and 3 (combined) were injected into
the HPLC and analyzed. In Figure 8, the results from step one are shown
as solid bar graphs, while the result from the combination of steps 2 and
3 are shown as the open or clear bar graphs.
This data would appear to support the hypothesis very well.
Three are three different groups of compounds that are detected by the
HPLC UV detector, with essentially 3 different relative sensitivities.
The first sub group is represented by naphthalene. At the wavelength of
detection called out by the EPA HPLC method 8310 ( 254 nanometers), the
minimum detectable quantity (of the net SFE and HP analytical method) at
a signal to noise ratio of 4:1 is 0.100 mg per kg (ppm weight/weight
basis) for naphthalene. The most favorable compounds are represented by
anthracene naphthalene. Anthracene detected at 254 nm, is being sensed
virtually the lambda maximum, and so now the minimum detectable quantity,
at the same signal to noise ration is 0.005 mg/kgm. Indeno(1,2,3-
cd)pyrene represents the most typical type of the PAH compounds--has a
minimum detectable quantity of 0.030 mg per kg. All of the PAH
compounds together meet the EPA solid waste criteria for sensitivity.
Finally, it is relevant to mention an interesting observation seen by
the lab at the Canadian Center for Inland Waterways in Burlington,
Ontario, a branch of Environment Canada , H.B. (Bill) Lee and his
coworkers. Their experiments unlike the experiments of other three
labs collaborating, in this study (the Wilmington, DE HP site, the HP
Germany site, and Inland Environmental Consultants in Chicago) involved
the use of gas chromatography ( GC/MS) for the analysis step. This
brought some concern about the 4^ water as one of the co-solvents. If
any of the water remains in the reconstitution solution there might be a
detrimental effect on the GC fused silica capillary column and the
injection procedure. This lab modified the original SFE method from le
methanol, 4* water to 1- methanol, 4* methylene chloride (with all other
conditions the same) . This is work which we duplicated then in our
Wilmington laboratories and we found no significant difference in the
overall recovery. This work is described in a separate publication .
In summary , this particular method appears robust enough to be
considered for a Round Robin study such as occurred with the EPA method
3560 TPH last year.
Acknowledgment
Many people have contributed to this work by sharing ideas and results of
work in progress. They include Barry Lesnik, Tom Peart, Robert Hong-You,
Pat Sandra, Frank David, Jerry King, Viorica Lopez-Avila, Lee Altmayer,
Werner Beckert, Steven Pyle, Robert Marsden, and Philip Wylie.
413
-------
References
l.D.R. Gere, C.R. Knipe, P. Castelli, J. Hedrick, L.G. Randall, J.
Orolin, H. Schulenberg-Schell, R.
Schuster, H.B. Lee, and L. Doherty"Bridging the Automation Gap between
Sample Preparation and Analysis: SFE, GC, GC/MSD and HPLC Applied to
Environmental Samples" manuscript in press, J.Chromatog. Sci.
2. Personal communication from Dr. Barry Lesnik, Office of Solid Waste,
US Environmental Protection Agency, 401 M Street, SW, Washington, DC
20460 USA.
3. Dr. Mary Ellen MacNally, Agricultural Products, Experimental
Station, E402/3328B, E.I. du Pont de Nemours and Company,
Wilmington, DE USA.
4. H.B. Lee, T.E. Peart, R.L. Hong-You, and D.R.Gere, "Supercritical
Carbon Dioxide Extraction of Polycyclic Aromatic Hydrocarbons from
Sediments," manuscript submitted for publication.
5. W. Beckert, "An Overview of the EPA's Supercritical Fluid Extraction
(SFE) Methods Development Program," ACS Symposium: Supercritical Fluids
in Analytical Chemistry sponsored by the Division of Analytical
Chemistry at the 201st National meeting of the American Chemical Society,
Atlanta, Georgia, April 14-19, 1991.
6. V. Lopez-Avila, N.S. Dodhiwala, and J. Benedicto, Evaluation of
Various Supercritical Fluid Extraction Systems for Extracting Organics
from Environmental Samples, Final Report for Work Assignment 1-1, EPA
Contract 68-C1-0029, Environmental Monitoring Systems Laboratory, Office
of Research and Development, U.S. Environmental Protection Agency,
Las Vegas, Nevada 89119, February, 1992.
7. Personal communication from John Orolin,Inland Consultants, Inc.,
Environmental and Engineering Services, 3921 Howard St., Skokie, IL,
60076, USA .
414
-------
Figure 1 :
Outline of Study
Verify the HPLC method
- speed of analyRis.Reaolution
- Internal Standard (biphenyl or bromo
bonzeno)
Surrogate (Quaterphenyl)
Optimize SFE parameters
Temperature
-• Density (pressure)
Modifier
- Volatiles .Large mw PAH'n
Basic Parameters
• Modifier
— add to thimble
- Pre-mixed tanks
"• On-line modifier pump
Recovery
Reproducibility
Figure 2 : SFE-HPLC Bridge with Modifer
Pump
jr¥5
LC
Mod/fie*
Pump
415
-------
Figure 3 : EPA PAH Compounds HPLC
Elle graphics integration Qeports Abort Help
13] 003-01 OLD: MWD A. Stu-254.16 Rcl-390.20
Figure 4 : Comparing HP Method to Method Without
Modifier and Without Solid Subambient Trapping
100
20
0
—
—
..._
•;
\
; .
,
._..
,
,
,
—
im
T "•;;
_J -—51—
1 1
, 1 ; IV";;, , , ,
,
m MI
l AH
120
100
80
60
JO
1 AM
56789
t.SEPA PAH Compounds (increasing size)
• EPA mean HP mean
416
-------
Figure 5 : SFE Conditions Step 1
Sample - 3.0 g Soil
HP7680T SFE + HP1050 Pump: Step 1
Extiaction "Pressure 1749 psi
• Temperature 720 C
• Composition CO2
• Static Equilib.Time 2.0 mm
" Dynamic Extn. Time 10.0 mitt
* Flow Rate 2.0 mL/min
< i,-[t:- ?!...•!•: "Trap OD5
* Temperatures 5 C.trap; 45 C nozzle
Reconstitution • Solvent Fraction Vol. THF/ACN, 0.8 mL
• Temperatures eOC, trap; 45 C noz
• Flow Rate 2.0 mL/min
Figure 6 : SFE Conditions Step 2
Sample - 3.0 q Soil
HP7680T SFE + HP1050 Pump: Step 2
Kxtrn.'timi -Pressure 4950 psi
* Temperature 120 C
« Composition CO2 + 1 % MeOH +4% H2O
* Static Equilib.Time 2.0 min
• Dynamic Extn. Time 40.0 min
• Flow Rate 2.0 mL/min
Collection "TraP ODS
= Temperatures 80 C,trap; 45 C nozzle
Reconstitution : THIS STEP DOES NOT REQUIRE A RINSE
417
-------
Figure 7 : SFE Conditions Step 3
Sample - 3.0 g Soil
HP7680T SFE + HP1050 Pump: Step 3
Extraction
» Pressure
" Temperature
» Composition
» Static EquilibTime
« Dynamic Extn. Time
» Flow Rate
» Trap
« Temperatures
•4950 psi
120 C
CO2
0.5 min
10.0 mm
2.0 mL/min
ODS
80 C,trap; 45 C nozzle
Reconstitution. Solvent, Fraction Vol. THF/ACN, 0.8 mL
• Temperatures eOC, trap; 45 C noz
• Flow Rate 2.0 mL/min
Figure 8 : Separating the Variables: Recovery of the
Volatile Fraction of PAH's vs. Larger PAH's
40 • 1 ^ • '' • | i • ' ' •
i! i i! 1
I
E . --tfiHy-
1 *
-Lflwi n i
i I
^ 1
1
1
j
H-l
i
1
i
j
j
1
1
1
1
JL
ffi
8 9 10 II 12 13 14 IS 16
\ \ Recovery [ |X Recovery
iStep) I—I Steps 2& 3
418
-------
55 ANALYSIS OF PCB'S IN SOIL, SEDIMENTS, AND
OTHER MATRICES BY ENZYME IMMUNOASSAY
Robert O. Harrison, ImmunoSystems, Inc., Scarborough, Maine 04074
Robert E. Carlson, Ecochem Research Inc., Chaska, Minnesota 55318
Alan J. Weiss, Millipore Corp., Bedford, Massachusetts 01730
ABSTRACT
A competitive inhibition E_nzyme ImmunoAssay (EIA) for the determination
of polychlorinated biphenyls (PCB's) is commercially available as the
EnviroGard™ PCB Kit. Soil sample preparation can be performed in the
field using disposable kit components. The test can analyze PCB's in the
field in less than 30 minutes total, using no specialized equipment. Test
specificity is restricted to PCB's, primarily Aroclors 1016, 1242, 1248,
1254, and 1260. The kit can be used to screen unidentified Aroclors at
4 levels from 1 to 50 ppm in soil. This method has been reviewed by the
EPA Office of Solid Waste and will be proposed for inclusion in the
4000 series of screening methods in the next SW846 update.
Quantitative data for soils in this concentration range can be easily
obtained by standardization of the kit with the same Aroclor as in the
samples (if known). Quantitative EIA analyses of river sediments were
completed the same day in a portable lab on site, with a tenfold increase
in sample throughput and a tenfold decrease in cost per sample compared
to GC-ECD. Data for several methods and matrices are presented.
These results demonstrate that EIA kits can effectively analyze for PCB's
in many situations at a fraction of the cost and in a fraction of the time of
standard methods.
INTRODUCTION
Reagent and Method Development
The development of the EIA for PCB's followed these steps: 1) PCB
derivatives were synthesized for conjugation to proteins; 2) one or these
PCB derivatives was conjugated to a carrier protein and the resulting
conjugate was used to immunize animals, which then produced antibodies
recognizing both the PCB derivative and PCB's; 3) a PCB derivative
was conjugated to horseradish peroxidase (HRP) to make a conjugate
which can be captured by anti-PCB antibodies; 4) the PCB-HRP
conjugate was used to screen and select antibodies; 5) the selected
system was optimized for PCB sensitivity and solvent and matrix
tolerance, then characterized for specificity; 6) sample preparation
419
-------
methods were developed for specific sample types; 7) these methods
were validated using Held samples.
PCB EIA Procedure
Figure 1 schematically illustrates the procedure used for the analysis of
samples containing RGB's. In summary: 1) rabbit antibodies which
recognize the PCB structure are immobilized on the walls of plastic test
tubes; 2) PCB's in solvent, in the form of either standards, calibrators, or
samples prepared as described below, are mixed with Assay Diluent in
tubes, allowing PCB's to be captured by the immobilized antibodies
[incubation 1). PCB's are specifically retained on the solid phase when
the rest of the sample is washed away (wash /); 3) PCB-enzyme
conjugate is added to tubes and bound in the same manner as in step 2
[incubation 2}. The unbound conjugate is washed away [wash 2} and the
amount retained by the immobilized antibody is inversely proportional to
the amount of PCB bound in step 2; 4) enzyme substrate and chromogen
are added to the tubes for color development by the bound enzyme
[incubation 3). The intensity of color is proportional to the amount of
captured enzyme and is inversely proportional to the amount of PCB
bound in step 2. Therefore, more color means less PCB. Incubations
are typically, but not always, 5 minutes. Washes use tap water.
Sample and calibrator volumes vary with the sample type and protocol,
but are generally 5, 25, or 100 u,l. Incubation 1 volume is 500 uJ.
Sample Preparation
Sample preparation methods are given below for several matrices.
Where EIA analysis is indicated, the extract is used as described in step
2 of the EIA Procedure above (incubation 1 of Figure 1).
Soil sample preparation is summarized as follows: Weigh 5 g soil
on portable balance and place in polypropylene extraction bottle, extract
soil by adding 5 ml of methanol and shaking vigorously for two minutes.
Filter extract and collect for storage or immediate EIA analysis. Analyze
extract as described in step 2 of the EIA Procedure above. All
components required for this method are commercially available in kit
form.
This study used two methods of sediment sample preparation for
EIA analysis. They are summarized as follows: Method 1) homogenize
wet sediment, weigh 4 g onto solvent rinsed Al foil, add 10-15 g of
anhydrous sodium sulfate to form a friable mixture, extract using 20 ml
isopropanol with 2 minutes of sonication and 15 minutes of shaking, let
settle briefly or centrifuge, then filter supernatant. Some extracts were
prepared by variations of this method, all of which are similar in speed,
simplicity, and ease of field use. Analyze extract as described in step 2
of the EIA Procedure above. Method 2) homogenize, Soxhlet extract (48
hours) using 1:1 acetone:hexane, exchange to cyclohexane (for GC-ECD
420
-------
or storage), exchange to isopropanol for EIA. Analyze extract as
described in step 2 or the EIA Procedure above.
Surface wipe sample preparation is summarized as follows: Mark
surface with template, remove wipe from methanol in extraction bottle,
wipe area to be analyzed, place wipe back in extraction jar and extract
by vigorous shaking for 1 minute. Analyze extract as described in step 2
of the EIA Procedure above. All components required for this method are
commercially available in kit form.
Water sample ^reparation is summarized as follows: condition
glass barrel SPE column with isopropanol and reagent water, extract
sample by drawing through column, air dry column, elute with
isopropanol. Analyze eluate as described in step 2 of the EIA Procedure
above. All components required for this method are commercially
available.
RESULTS AND DISCUSSION
Test Specificity
The test response to Aroclors 1016, 1242, 1254, and 1260 is within
twofold of me response for Aroclor 1248. The broad specificity of the
EIA for the common moderately chlorinated Aroclors allows either
screening or Quantitative analysis. This specificity is relatively
independent of trie slight differences in protocol among the sample types
described below. Biphenyl and several chlorinated ring compounds
were also tested for crossreactivity in the EIA. All of these compounds
demonstrated less than 0.5% crossreactivity compared to Aroclor 1248:
3,3'-dichlorobenzidine, 1,2-dichlorobenzene, 1,3-dichlorobenzene, 1,4-
dichlorobenzene, 1,2,4-trichlorobenzene, biphenyl, 2,4-dichlorophenol,
2,5-dichlorophenol, 2,4,5-trichlorophenol, 2,4,6-trichlorophenol, and
pentachloropnenol. These results mean that more than 200 ppm of any of
these compounds would be required to give the same test response as 1
ppm of Aroclor 1248. Also, more fhan 10,000 ppm of any of these
compounds would be required to give the same test response as 50 ppm
of Aroclor 1248.
Soil Screening
Soil screening results obtained using the kit with calibrators at 5, 10, and
50 ppm are shown in Table 1. me respective actual concentrations of
these calibrator solutions are 3, 5, and 22 ppm Aroclor 1248. These
values were chosen to provide 99% confidence that positive samples
will be detected regardless of the Aroclor present, variations in extraction
efficiency, or method imprecision for either GC or EIA. This calibration
strategy biases the test slightly toward false positive results, providing
greater certainty that contaminated samples will be screened out
effectively. For a population of samples linearly distributed in
421
-------
concentration from 1 to 100 ppm, the calculated false positive rate is
less than 6%, while the calculated false negative rate is less than 0.1%
for all three decision points. Several samples in Table 1 illustrate this
point. For example, sample 3 is identified as > 50 ppm, but is less than
50 ppm by GC (29.6 ppm). This result is expected because the actual
concentration of the sample is higher than the actual 22 ppm of the 50
ppm calibrator. Table 1 shows that most of the false positive results
occur in the concentration range between the middle and low calibrators,
due precisely to the bias designed into the test. Prudent site screening
demands that samples in this range be identified for follow-up GC
analysis to ensure accurate decision making. The data of Table 1 show
that the slight inherent bias of the test accomplishes this effectively. This
method has been reviewed by the EPA Office of Solid Waste and will be
proposed for inclusion in the 4000 series of screening methods in the
next SW846 update. Another version of this kit has been developed for
screening at I and 5 ppm in soil. Test protocol and performance
characteristics are similar to the first kit.
Quantitative Soil Analysis
If the Aroclor present in the samples is known, quantitative data can
easily be obtained over a wide range of concentrations. This is done by
substituting for the calibrators a standard curve of the known Aroclor.
Concentrations of the Aroclor in the samples are determined by
comparison to the standard curve. Expansion of the range of quantitation
is possible by diluting high concentration extracts and by using a more
sensitive EIA protocol for levels near or below 1 ppm. Table 2 illustrates
results of the latter procedure for a set of soil samples containing less
than 5 ppm, compared to the GC results from 8080 analysis. Because
Aroclor 1248 is recognized slightly better than Aroclor 1260,
standardization on Aroclor 1260 will yield slightly higher values (as
demonstrated in Table 2). Correction for this difference in crossreactivity
between Aroclors 1248 and 1260 would approximately equalize the
slopes of the two regression lines in Table 2. Other projects using this kit
have obtained quantitative data from field samples covering a range 5
decades of concentration, from 0.1 to more than 10,000 ppm. Method
precision was tested in a 9 day validation study at 3 sites using 4 field
samples over a concentration range of 5 to 35 ppm, resulting in a 27%
concentration coefficient of variation within site and across day and kit
lot.
Quantitative Sediment Analysis
The PCB EIA kit was used for quantitative sediment analysis in two
different situations. A comparison of EIA and GC-ECD was made using
sediment samples from a typical Great Lakes industrial contamination site,
prepared for EIA by metnod 2. For GC-ECD analysis, these samples
422
-------
were further cleaned by gel permeation chromatography and other
techniques, including sulfur removal. Correlation between the methods
was good (Figure 2) over two decades of concentration, from 0.2 to 20
ppm. This result demonstrates that the EIA is able to measure RGB's at
low levels in sediments when properly presented to the antibodies,
despite the presence of sulfur compounds and heavy metals in these
sediment extracts. Sediments which have undergone significant
dechlorination may yield inaccurate results because of decreased
recognition of less chlorinated congeners by the antibodies; such was not
the case here. Precision in this study was good; the pooled
concentration coefficient of variation (%Cv) was 13% for 7 pairs of
duplicates, 2 sets of 6 replicates, and 1 set of 7 replicates.
Based on the above success of method 2/EIA for quantitative
sediment analysis, method 1 was developed for simpler sample
preparation ana EIA analysis in the field. Quantitative sediment analysis
by method 1/EIA was used at other Great Lakes industrial contamination
sites. Sediment cores were transported to a lab trailer on shore for
sampling. Same day analysis was performed using sample preparation
method 1/EIA. Performance data for the EIA are summarized in tables 3
and 4. Correlation of EIA results with GC-ECD results was determined
for a set of 41 samples containing 0.1 to 12 ppm from 3 sites; the
correlation coefficient was 0.84 and the regression equation was y =
0.63x + 0.47 ppm. Timely on site analysis of total PCB in sediments
enabled researchers to guide field sampling to ensure appropriate
samples were taken for GC-ECD analysis. In addition, project managers
were able to conform confidently to dredging guidelines and to guide
disposal of dredge spoil. Further simplification of sediment extraction
method 1 is currently under investigation.
Other Matrices
Analyses of PCB's on surfaces and in water are also possible using the
same EIA kit as for the previously described applications, but with slightly
modified protocols and appropriate sample preparation methods.
Decisions about PCB contaminated surface remediation are often limited
by the turnaround time for wipe sample analysis by GC. The PCB EIA kit
has also been formatted as a wipe test kit for screening use in such
situations, analogous to the previously described soil screening kit. Data
from 64 field samples from three sites were used to set actual
concentrations for the 10 and 100 u.g/100 cm2 calibrators. The wipe kit
design includes the same false positive bias as described above for the
soil screening kit, for exactly the same reasons. Decisions are made in
the same fashion as soil screening, as less than (or not less than) the
decision level. For a population of samples linearly distributed in
concentration from 1 to 100 u.g/100 cm2, the calculated false positive
423
-------
rate is 6 to 7%, while the calculated false negative rate is less than 0.7%
for both 10 and 100 ^ig/100 cm2 decision points.
PCB's in water adsorb strongly to even small amounts of
particulate material. Therefore, analysis of total PCB in water requires
efficient recovery of this adsorbed material. This is accomplished by the
use of high throughput CIS columns which efficiently capture particulate
material. These columns allow the analysis of slightly to moderately
turbid field water samples with minimal reduction in flow rate.
Isopropanol is used to elute the PCB's captured by the CIS from the
dissolved phase and simultaneously extract the adsorbed PCB's from the
trapped particles. EIA analysis is performed directly on the eluate without
volume reduction. Data on the recovery of PCB spiked into water are
shown in Table 5. These results demonstrate the ability of the test to
detect 0.1 ppb total PCB in a small (500 ml) sample in approximately
one hour using a simple sample concentration method. Preliminary data
indicate a detection limit below 25 ppt in 4000 ml of water, and
improvement may be possible with modest modifications of the protocol.
The ultimate sensitivity of this method will depend on several factors,
including sample volume, particle load, particle chemistry, etc.
CONCLUSIONS
1. Test specificity is restricted to PCB's, primarily Aroclors 1016, 1242,
1248, 1254, and 1260.
2. The test is capable of analyzing for PCB's in soil in the field in less
than 30 minutes, using no specialized equipment.
3. Screening of soils containing unidentified Aroclors can be performed at
50, 10, 5, or 1 ppm, with 99% confidence of detection of
contaminated samples.
4. Quantitation from 0.1 to 10,000 ppm is possible for soil samples
containing an identified Aroclor.
5. Screening of surfaces contaminated with unidentified Aroclors can be
performed at 10 and 100|ig/100 cm2 with 99% confidence of
detection.
6. The design of the EIA accommodates significant variations in sample
type and protocol, allowing the analysis of PCB's in many matrices.
7. Ongoing work with this kit includes improved sediment analysis, water
analysis, oil analysis, and biological tissue analysis. Field testing and
validation are proceeding for alf these applications.
ACKNOWLEDGEMENT
The initial phase of development of this PCB immunoassay was partially
supported by the US EPA through a sub-contract to ECOCHEM from Mid-
Pacific Environmental Laboratories, Inc.
424
-------
Table 1. Screening of 32 soils using the EnviroGard™ PCB EIA Kit. Color was
read by differential photometer and interpretation of EIA result was based on
comparison of optical density value to calibrator optical density value. The
correctness of the EIA interpretation is indicated for each calibrator level as
follows: normal = correct; italicized = false positive, bold = false negative.
Soil
Sample Tvoe*
2/
23
12
6
1
10
22
8
3
18
14
9
21
16
4
20
2
29
15
25
30
27
32
13
19
26
1 1
28
31
5
7
17
N/A
N/A
Loam/Clay
Sand
Clav
Clay
N/A
Clay
N/A
Sand
Clay
Loam/Sand
Clay
Clay
N/A
Clay
N/A
N/A
Clay
Loam
N/A
N/A
Loam
Clay
Clav
Clay
Loam/Clay
N/A
Clay
N/A
Loam/Clay
Clay
EnviroGard™ EIA Result
Comparison to Calibrators EnviroGard™
Aroclor <5 ppm <10 ppm <50 ppm EIA
Content Calibrator Calibrator Calibrator Interpretation
1254
1254/60
1260
1260
1260
1260
1242/54/60
1260
1254
1260
1260
1260
1242/54/60
1260
1254/60
1260
1254
1254/60
1260
1248/54
1254/60
1248/54
1248/54
1260
1260
1248/54
1260
1248/54
1254/60
1254
1260
1260
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
No
No
No
Yes
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
>50 ppm
<50 ppm**
>50 ppm
>50 ppm
>50 ppm
>50 ppm
<50 ppm
<50 ppm
>50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<10 ppm
<50 ppm
<10 ppm
<1 0 ppm
<50 ppm
<50 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
GC
Result
(ppm PCB
by EPA
8080)
1012
164
73
68
60
56
36.6
32
29.6
24
15
13
13
1 1
6.7
5
4.7
4.6
4
4
3.9
3
2.7
2
1.4
0.3
0.2
0.1
0.1
<0.1
<0.1
<0.1
N/A = not available
* Further work indicated false negative response was due to sample matrix effect.
Subsequent analysis indicated the presence of an interfering substance that prevented
PCB binding to the antibody. This false negative at the highest calibrator level was not
due to cross-reactivity or incomplete PCB extraction.
425
-------
Table 2. Correlation of quantitative soil analysis by EIA and GC.
Sample
A
B
E
1
H
D
SSI
G
C
SS2
F
SS5
GC Result
oom bv 8080
0
0
0
0.2
0.3
0.6
0.6
0.6
1.0
1.1
1.4
3.1
Lab A
ppm by EIA
«0.15
«0.15
«0.15
0.3
1.0
0.4
0.7
1.0
1.2
2.3
1.6
5.0
Lab B
ppm bv EIA
«0.1
«0.1
«0.1
0.29
0.46
0.52
0.74
1.05
0.92
1.8
1.45
3.6
Aroclor used for standard curve 1260
Regression equation data (vs. GC)
r2
slope
_ Y intercept
0.93
1 .56
-0.03
1248
0.96
1.15
0.05
Table 3. Performance of sample preparation method 1 with EIA for quantitative
sediment analysis.
Parameter
Detection Limit
Precision
Performance Comment
60 ppb Determined by repeat analysis of system blanks
(n = 7); 95% confidence
Concentration %CV, based on duplicate analyses
of extracts of field sediment samples. [%CV =
{(mean/standard deviation) xlOOJ]
intraassay 7.7 %CV n = 48 pairs of duplicates
interassay 7.9 %CV n = 52 pairs of duplicates
Spike recovery 90% of nominal Based on detection of Aroclor 1248 spiked into
clean soil, n = 8
Recovery
73±9%
Recovery of total PCB from dried SRM sediment,
relative to GC-ECD result, n = 6
426
-------
Table 4. Comparison of sediment analysis by PCB EIA and GC-ECD.
Method 1/EIA Method 2VGC-ECD
Cost per Sample $15 $150-300
Sample Throughput 10 per 2.5 hours 1 per 2.5 hours
* GC-ECD analysis requires Method 2 sample preparation procedure followed by gel
permeation chromatography and sulfur removal.
Table 5. Recovery of Aroclor 1248 Spikes from Tap Water Samples. This water
contained 0.5 mg/liter particulate material, determined gravimetrically using a 0.45
Jim filter. (* reagent water blank subtracted)
ng Spike nn[ Final ng/ml n Recovery (mean ± SD)
50 500 0.1 3 73±29%
250 500 0.5 3 1 14±50%
1000 500 2.0 3 90±40%
100 4000 0.025 1 125%*
FIGURE LEGENDS
Figure 1. Principles of the Test. Schematic depiction of the EIA for PCB's.
Figure 2. Correlation of GC and EIA Results for Sediment Analysis. Sediments were
analyzed by method 2 and EIA as described in the methods section.
427
-------
ittiation 1:
v =: .. -. % % ••
Dilution of sample or calibrator is incubated in
tube containing immobilized antibodies.
= PCB
= Interfering Material
= Anti-PCB Antibody
I**'.' •
Sample matrix is washed away,
leaving only RGB's bound to antibodies.
9&ubation2:
& - -
PCB-HRP binds to free anti-PCB sites
on immobilized antibodies.
E = HRP
(Horse Radish
Peroxidase Enzyme)
Unbound PCB-HRP is washed away, leaving an
amount of enzyme inversely proportional to the
PCB concentration in Incubation 1.
Colorless substrate and chromogen are
converted to blue color in proportion to
amount of bound enzyme. Less color means
mre£CB. Stop solution inactivates the HRP,
changes color to yellow, and stabilizes color.
S = Substrate
C = Chromogen
428
-------
100
10
EIA
ppm
y = 1.23x - 0.12
R = 0.97
n = 33
2
GC ppm
100
429
-------
56
EXPLOSIVES ANALYSIS OF ATYPICAL MATRICES AND TECHNICAL ENHANCEMENTS TO
U.S.EPA METHOD 8330
Dennis Hooton. Senior Chemist
M. Christopher son, T. Dux, G. Jungclaus, H. Randa, R. Sauter
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
ABSTRACT
The U.S. EPA Method 8330 from SW-846 is used for the determination of
nitroaromatics and nitramines (including HMX, RDX, TNT, etc.) in soil
and water samples by high performance liquid chromatography (HPLC).
This method is being used for remedial investigation programs at former
ammunition facilities throughout the United States. Variations of this
method have been used by Midwest Research Institute (MRI) as the basis
for analyzing more exotic matrices not addressed by the SW-846 method,
including: surface wipe samples, concrete chips, plant biota, and TCLP
extracts.
In order to improve data consistency, technical enhancements have been
incorporated into method applications to improve overall performance and
quality. These include: "pre-mixing" soil for better homogenization
and analytical subsampling, performing multiple extractions for trace-
level analysis of water samples using the "salting-out" technique (U.S.
Army Corps of Engineers draft procedure), and enhancing chromatographic
analysis with the addition of an internal standard for better
qualitative identification (relative retention time) and quantization
(relative response factors).
INTRODUCTION
Since 1985, MRI has performed explosives analyses for remedial
investigation of sites throughout the United States using the EPA Method
8330 "Nitroaromatics and Nitramines by High Performance Liquid
Chromatography (HPLC)." Although the basic method has remained the
same, MRI has added several enhancements to the procedure which has
improved overall method performance. In addition to analyzing soil and
water samples, the method has been adapted for use with other matrices,
including: surface wipes, concrete solids, TCLP extracts, and plant
biota. This paper presents an overview of how this method has performed
in these different applications and describes some techniques which have
improved overall performance.
430
-------
MATRICES
Method 8330 is a very versatile method which can be adapted to a variety
of matrices. However, special applications require special techniques
to achieve the most reliable results. The following paragraphs describe
how this method has been applied to various matrices.
Soil and Sediment
Method 8330 specifies that soil and sediment samples be initially dried
in air at room temperature or colder and protected from direct sunlight.
Then the sample is ground in a solvent-rinsed (acetonitrile) mortar.
MRI has found that preparing a "pre-mix" of the soil before actually
taking analytical subsamples helps to assure sample homogeneity and
precision of the method. The expanded procedure also uses the air-
drying technique, which is then followed by grinding the soil/sediment
using a solvent-rinsed laboratory burr mill to quickly reduce the
particle size of the soil. Nearly all of the sample will pass through a
30-mesh screen. This step is important for several reasons: first, it
allows much larger samples or an entire sample to be easily prepared for
analysis, reducing the concern of inadequate composite mixing during
field sample collection, and second, it provides a sample texture more
conducive to further homogenization. The ground sample is collected in
a glass jar and rotary tumbled to further mix the sample prior to
analytical subsampling.
Table 1 shows a comparison of Method 8330 precision results for the
analysis of contaminated soils. Data provided in the SW-846 methods'
"Intralaboratory Precision of Method for Field-Contaminated Soil
Samples," was based on a study where the method was tested by six
different laboratories. These data are compared to field-contaminated
samples which were analyzed by one laboratory (MRI) three separate times
over an approximate one-month period. Although these data are not
directly comparable because of the significant differences in analyte
concentrations, the data do indicate that comparable or better precision
can be achieved at trace contaminant levels using pre-mixed soil
samples.
Concrete Chips
The technique used for soil/sediment preparation has been successfully
used on hard clay soils and also very hard samples which would not
normally be amenable to standard grinding procedures (i.e., mortar and
pestle). For example, this technique was used on concrete chip samples
collected from cinder block walls to determine if residual explosives
remained in the concrete. The same technique, described above for the
soil samples, was used to reduce the concrete chips to a powdery texture
prior to analysis by Method 8330. Although none of the Method 8330
analyte list were found in these samples, method recovery and precision
431
-------
were determined using spiked concrete samples. Table 2 shows precision
and accuracy for this matrix; spiked soil samples, which were extracted
and analyzed concurrently with the concrete samples, are also shown for
comparison.
Table 1. Comparison of SW-846 Interlaboratory Precision Values
with "Pre-mixed" Results for Contaminated Soils
Analyte
Tetryl
RDX
TNT
TNB
SW-846 data
Concentration
(Mg/g)
2.3
104
877
7
669
3
72
Precision
(RSD)a
18
17
8
18
10
8
12
Pre-mixed sample data
Concentration
(Mg/g)
2.2
1.7
3.4
12.1
Precision
(RSD)a
8
7
9
1
a Relative standard deviation.
Table 2. Matrix Spike Duplicate Recovery and Precision
for Concrete and Soil Samples
Matrix
Concrete
Sample No. 1
Concrete
Sample No. 2
Soil
Spike
level
(units)
2.5
Mg/g
2.5
Mg/g
2.5
Mg/g
Average recovery (range)
HMX
97
(±4)
99
(±4)
102
(±D
RDX
97
(±8)
102
(±4)
98
(±D
TNB
78
(±47)
106
(±4)
108
(±3)
Tetryl
37
(±1)
83
(±4)
94
(±10)
TNT
29
(±26)
95
(±3)
100
(±1)
24 DNT
69
(±27)
105
(±4)
108
(±D
Water
The current version of Method 8330 (November 1990), is applicable to
relatively high-level determinations of explosives in water samples.
This method consists primarily of a "dilute-and-shoot" approach which is
effective and straight-forward for samples with analyte concentrations
ranging from approximately 4 fig/L (for DNB) to 44 /ig/L (for tetryl) .
Trace-level analysis of samples with analyte concentrations as low as
0.1 /zg/L require extraction and concentration of the resulting extract.
MRI uses a technique based on a U.S. Army Corps" of Engineer's (USAGE)
432
-------
draft protocol in which 400 mL of a water sample is saturated with
sodium chloride salt, followed by liquid-liquid extraction of the
explosive compounds into acetonitrile. This "salting-out" technique has
been successfully used for many years to analyze groundwater and surface
water samples from federal remediation sites.
MRI has modified and enhanced the original draft method in several ways
to improve overall performance or to reduce common reagent interferences
exhibited primarily at the trace levels. These changes include per-
forming a second acetonitrile extraction of the salt-saturated aqueous
sample to recover any residual explosives not collected in the first
extraction and pre-cleaning the salt reagents (sodium chloride and
sodium sulfate) by solvent extraction and heat. The results of these
changes are shown in the following table (Table 3) where recovery and
precision are shown for spiked control samples at trace level concen-
trations. The method has proven to be reliable and reproducible over
hundreds of water samples.
Table 3. Summary of Accuracy and Precision
Data for Spiked Water Samples8-13
Analyte
HMX
RDX
2,4-DNT
TNT
TNB
Tetryl
Average
recovery
(%)
96
83
53
78
85
76
Precision0
12
20
11
8
8
4
No. of
determinations
5
5
5
5
5
5
Method performance was measured by spiking clean water
samples with separate extraction batches.
Spike levels for the target analytes in water were
1.25 Mg/L for all analytes except tetryl at 5
s = Standard deviation of the average.
TCLP Extracts
The "salting-out" extraction technique has also been successfully used
for TCLP extracts in evaluating the leachability of explosive compounds
from soil. In one study, Method 8330 was used to extract highly
contaminated soil samples. The same soil samples underwent the TCLP
extraction procedure and a portion of the TCLP extract was analyzed
using the "salting-out" procedure.
433
-------
These results show that the method adequately' characterized both the
original soil and the resultant leachate for explosives contaminants.
The extract concentrations were consistent with the original soil
contamination, producing 25% to 100% of the total possible compound that
would be expected based on the original soil analysis (assuming a factor
of 50 to convert /zg/g soil concentrations to /zg/L leachate
concentrations). These results are shown in Table 4 below.
Table 4. Analysis Results for RCRA Characterization
Soils and TCLP Extracts
Matrix
TCLP prep, blank
Soil
TCLP ext.
Soil
TCLP ext.
Soil
TCLP ext.
Soil
.TCLP ext.
Soil
TCLP ext.
Soil
TCLP ext.
QUANTITATION LIMITS:
Matrix:
Soil
TCLP extract
Units
/ig/L
fg/8
PS/L
/ig/8
pg/L
/*g/g
MS/L
Mg/S
M8/L
/*g/8
/*8/L
/*8/S
PS/L
Units
**g/8
MS/L
Analyte concentration
HMX
< QL
0.53
< QL
0.33
19
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
RDX
< QL
3.0
180
3.3
150
0.34
4.8
0.97
25
0.91
20
0.19
< QL
TUB
< QL
41
170
4.9
21
4.3
57
43
1200
28
860
< QL
< QL
Tetryl
< QL
< QL
140
< QL
< QL
< QL
< QL
< QL
76
< QL
65
< QL
< QL
TNT
< QL
2600
70000
180
3700
89
3000
49
1500
31
990
22
630
2,6-DNT
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
2,4-DNT
< QL
8.3
130
2.5
40
3.0
74
0.96
25
0.90
22
< QL
< QL
Concentration
0.15
1.1
0.15
1.1
0.50
3.3
1.0
6.7
0.15
1.1
0.50
3.3
0.15
1.1
Surface Wipes
Surface wipes were analyzed using an adaptation of the soil/sediment
procedure in which surface areas were wiped with acetonitrile-saturated
gauze. The wipe samples were simply treated as a solid sample by adding
extracting solvent to the collection jar and processing the sample the
same as the Method 8330 procedure for soils. Since no explosives were
found in the field samples, a spiked control was used to demonstrate
recovery as shown in Table 5.
434
-------
Table 5. Control Spike Recovery For Surface
Wipe (Gauze) Sample
Analyte
HMX
RDX
TNB
Tetryl
TNT
2,4-DNT
Percent recovery
(*)
75
77
80
87
80
75
Spiked concentrations were 5 /ig per gauze sample
(representing a 100 cm2 surface area).
Plant Biota
Biota samples present a more unusual sample matrix in that the explosive
compounds may be chemically bound to the plant structure. As expected,
the extraction technique is quite different than would normally be used
in the absence of biological transformation of these contaminants. For
example, one technique used for plant studies is to acid-hydrolize the
plant sample, then extract the mixture with ether to recover the
explosive compounds, followed by a solvent exchange and a column
chromatography step. Method 8330 analysis parameters were used to
measure explosive residues in the plant material.
Method performance for plants is presented in Table 6 below. Although
chemical recoveries are lower than other matrices, these are reflective
of the more compelx extraction procedure (which is not part of
Method 8330) rather than the Method 8330 analysis procedure used to
analyze the extracts. This is illustrated by the fact that similar low
recoveries were found for the control samples (spiked reagent taken
through the extraction procedure) reported with the biota data. It
should also be noted that the plant extracts did not produce any unusual
or problematic background interference for the Method 8330
chromatography.
435
-------
Table 6. Blank, Control, and Matrix Spike Results
(Percent Recovery)3 for Plant Biota
Biota controls
Method blank
Control 1
Control 2
Matrix spike
terrestrial
Plant No. 1
Matrix spike
terrestrial
Plant No. 2
Root plant No . 3
Root plant No . 4
RDX
< QL
52%
60%
57%
45%
49%
54%
TNB
< QL
51%
59%
Interference13
21%
37%
43%
Tetryl
< QL
78%
79%
~ 37%°
~ 53%°
63%
42%
TNT
< QL
43%
45%
~ 35%d
39%
28%
29%
2,4-DNT
< QL
40%
44%
37%
43%
35%
38%
a Spike levels in control and matrix spikes were 20 /tg/g f°r
indicated analytes.
b TNB recovery indeterminate due to chromatographic interference.
0 Tetryl recoveries are estimated because the extract concentrations
were near the quantitation limit.
d Estimated recovery for TNT due to high native level.
METHOD ENHANCEMENTS
The most significant technical improvement MRI has made to the SW-846
method has been the use of an internal standard (a chemically-similar
compound which is spiked into sample extracts and standards just prior
to analysis). This addition serves several important functions. First
and foremost, it is used to establish relative retention time windows
for qualitative identification for the Method 8330 analytes; it is
relative in the sense that chromatographic retention times are
normalized to the internal standard marker within each chromatogram.
This technique provides a much more precise determination of close
eluting compounds. For example, because the retention times for the
2,4-DNT and 2,6-DNT isomers vary by only a few seconds, normal time
variances between one injection to the next can make it difficult to
absolutely distinguish between the two compounds. However, when an
internal standard marker is added to set relative retention time windows
for these two analytes, there is an easily discernable difference. The
table below (Table 7) presents a summary of how precision is
dramatically improved by the use of relative retention time over normal
Method 8330 retention time identification criteria. Relative percent
differences decrease for all analytes when using RRT values which
reduces the chance of reporting false positive results.
436
-------
Table 7. Comparison of Absolute Retention Time vs.
Relative Retention Time (Internal Standard)
for Explosive Standards
HMX
RDX
TNB
DNB
NB
TNT
2,4-DNT
Tetryl
2,6-DNT
2 -NT
4 -NT
3 -NT
Average RTa
2.8
4.1
5.4
6.5
7.6
8.8
10.5
7.3
10.2
12.8
13.8
14.8
RPDC
1.4
1.7
1.3
2.0
1.6
1.6
1.9
2.1
2.8
2.8
3.1
2.9
Average RRTb
0.137
0.199
0.264
0.317
0.370
0.428
0.509
0.359
0.495
0.618
0.667
0.715
RPDC
0.73
0.50
0.38
0.95
0.27
0.42
0.39
0.56
0.61
0.32
0.45
0.28
a Average Retention Time (RT) in minutes.
b Average Relative Retention Time (RRT) - analyte RT over the internal
standard RT.
c Precision expressed as Range Percent Difference over 20 injections
and 7 different concentrations.
The internal standard also provides the advantage of improved
quantitation by establishing the linear relationship between an analyte
concentration and its corresponding relative response, i.e., analyte
response vs. internal standard response. This technique corrects for
minor fluctuations in injection volume and sample preparation volume
after the internal standard is added. And finally, the internal stan-
dard can be used to monitor injection precision during automatic (auto-
injector) analysis runs.
SUMMARY
Method 8330 can be adequately adapted to a wide variety of matrices,
including surface wipes, solid materials, plants, and TCLP extracts.
Method performance, especially precision, can be improved through more
rigorous preparation of soil/sediment samples prior to extraction and by
double-extracting aqueous samples. Addition of an internal standard to
sample extracts just prior to analysis provides more accurate qualita-
tive identification of target analytes and improves quantitative analy-
sis. Method 8330 has proven itself to be a rugged and reliable test for
determining trace-level explosives in environmental investigations.
437
-------
LIQUID CHROMATOGRAPHY/MASS SPECTROMETRY OF HIGH-MOLECULAR-WEIGHT
POLYCYCLIC AROMATIC HYDROCARBONS USING THE PARTICLE BEAM INTERFACE
L. Don Betowski, U.S. EPA, EMSL-LV, Las Vegas, NV, 89193
Chris Pace, Lockheed Environ. Sys. & Tech. Co., Las Vegas, NV, 89119
Mark Robv. SAIC, San Diego, CA, 92121
Most studies of polycyclic aromatic hydrocarbons (PAH) in
environmental samples have focused on PAH containing five fused rings
or less (MW < 300). Because of analytical difficulties, including low
vapor pressure, lack of analytical reference standards, lower sample
concentration, and the large number of isomeric possibilities, the
higher molecular weight PAH (MW > 300) are seldom examined. There are
no standardized methods for the measurement of high-molecular-weight
PAH. Despite analytical difficulties, the presence of high-molecular-
weight PAH has been demonstrated in many environmental samples
including carbon black, coal tar, petroleum, diesel and other air
particulates, and in soils. Although the number of toxicity studies
on high-molecular weight PAH are limited, several dibenzopyrenes (MW
302) are known to be highly carcinogenic.
In this work, particle beam liquid chromatography/mass spectrometry
(PB LC/MS) was investigated as a means to measure high-molecular-
weight PAH. Instrument performance was evaluated with 17 PAH covering
the molecular weight range 300 to 450 amu. The PAH were separated by
conventional high performance liquid chromatography (HPLC) using a
polymeric octadecylsilica (C-18) packing (Vydac 201-TP) and gradient
elution with methanol/tetrahydrofuran (THF). Instrument detection
limits as measured by selected ion monitoring (SIM) on the singly
charged molecular ion of each PAH were found to be 0.15 to 0.6 ng on
column for PAH up to 352 amu and 2 to 4 ng on column for PAH greater
than 352 amu. Linear response was observed for those PAH with
molecular weight 300 to 352 amu over the concentration range 0.05 to
5.0 ug/mL. Non-linear response was observed for PAH with molecular
weight greater than 352 amu over the concentration range 0.25 to 25
ug/ml. The PB electron ionization (El) mass spectra of the PAH were
found to be variable with the ion distribution ratio of the singly
charged molecular to the doubly charged molecular ion being dependent
on molecular weight, ion source temperature, and on concentration.
Particle beam negative chemical ionization (NCI) of the PAH was also
examined. Substantial signal enhancement (10 to 600 fold) was
observed but found to be highly dependent on the amount of THF in the
mobile phase.
Notice: Although the research described in this abstract has been
funded wholly by the U.S. EPA through contract #68-CO-0049 to
Lockheed, it has not been subjected to agency review. Therefore, it
does not necessarily reflect the views of the agency.
438
-------
58 Quantitation of Pentachlorophenol (PCP) by a Rapid Magnetic
Particle-Based Solid-Phase ELISA in Water and SoiL
Fernando M. Rubio. Timothy S. Lawruk, Scott W. Jourdan, Charles S.
Hottenstein, Van Zha, and David P. Herzog, Ohmicron Corporation,
Newtown, Pennsylvania 18940.
ABSTRACT
Traditional testing methods for pentachlorophenol (PCP) are time
consuming, expensive and require specialized instrumentation.
Immunochemical assays provide methods that are sensitive, rapid, reliable,
which are suitable for lab and field analysis. A magnetic particle-based
ELISA for the determination of PCP in water and soil samples has been
developed. Paramagnetic particles, used as the solid support, allow for
precise antibody addition and superior assay kinetics. The assay procedure
and detailed performance characteristics, including correlation with GC/MS
and HPLC, are discussed. The method has a test range of 0.06 to 10 ppb
in water, and typical within assay %CV of less than 10%. Recovery studies
averaged 105%. The application of this method permits cost-effective
evaluation of samples without solvent disposal and can result in savings of
time and money.
INTRODUCTION
Pentachlorophenol (PCP) is used worldwide in commercial wood treatment,
paper production, the leather industry (as a preservative and fungicide), and
agriculture (as an insecticide and herbicide). Its extensive use has made
PCP an ubiquitous environmental pollutant; the chemical has been found in
water, soil, food, and air, often in high concentrations. Like many
organochlorine compounds, PCP has toxic effects in humans. It is rapidly
absorbed through the skin, respiratory and gastro-intestinal tract, inducing
skin rashes, respiratory diseases, increased hepatic enzyme activity and
renal failure.
Pentachlorophenol is one of the compounds regulated by USEPA under the
NPDWR with a Maximum Contaminant Level of 1 ppb in drinking water.
Under the RCRA Program (Toxicity Characteristic), the maximum
concentration of PCP allowable in solid waste is 100 ppm.
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 ELISAs 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 PCP magnetic-based ELISA described
439
-------
in this paper combines an antibody specific for PCP with enzyme labeled
PCP. The presence of PCP 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-PCP serum
and PCP-HRP conjugate (Ohmicron, Newtown, PA). Hydrogen peroxide and
TMB (Kirkegaard & Perry Labs, Gaithersburg, MD). PCP and related
compounds, as well as non-related cross-reactants (Chem Service, West
Chester, PA).
The anti-PCP 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 PCP
antiserum and the activated particles were incubated overnight at room
temperature with agitation. The unreacted glutaraldehyde was quenched
with glycine buffer and the covalently coupled anti-PCP particles were
washed and diluted with a Tris-saline/BSA preserved buffer.
Water samples (200 uL) and horseradish peroxidase (HRP) labeled PCP (250
uL) are incubated for 30 minutes with the antibody coupled solid-phase (500
uL) (step 1). A magnetic field is applied to the magnetic solid-phase to
facilate washing and removal of unbound PCP-HRP and eliminate any
potential interfering substances (step 2). The enzyme substrate (hydrogen
peroxide) and chromogen (3,3',5,5'-tetramethyl benzidine [TMB]) are then
added and incubated for 20 minutes (step 3). The reaction is stopped with
the addition of acid and the final colored product is 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, 0.1, 2.0, and
10.0 ppb of PCP. If the assay is performed in the field (on-site), a battery
powered photometer such as the RPA-III™ is used.
When analyzing soil samples, a simple extraction is performed prior to
analysis: 10 g of soil is shaken for 30 minutes (extractions as short as 1
minute could be performed) with 20 mL of a solution of 0.5% sodium
hydroxide/75% methanol/25% water (w/v/v). After settling, the sample
supernatant is diluted at least 1:500 in pentachlorophenol zero standard and
assayed as in the case of water samples.
RESULTS AND DISCUSSION
Figure 1 illustrates the mean standard curve for the PCP calibrators collected
over 79 runs, error bars represent one standard deviation (SD). The
displacement at the 0.1 ppb level is significant (86.9% B/Bo, where B/Bo is
the absorbance at 450 nm observed for a sample or standard divided by the
440
-------
absorbance at the zero standard). The assay sensitivity based on 90% B/Bo
(Midgley et al, 1969} is 0.06 ppb.
A precision study in which surface and groundwater samples were fortified
with PCP at 4 concentrations, and assayed 5 times in singlicate per assay,
on five different days is shown in Table 1. Coefficients of variation (%CV)
within and between day (Bookbinder and Panosian, 1986) were less than
13% and 12% respectively.
Correlation of twenty groundwater samples using values obtained by the
ELISA method (y) and GC/MS EPA Method 625 (x) method is illustrated in
Figure 2. The regression analysis yields a correlation of 0.980 and a slope
of 1.08 between methods.
Table 2 summarizes the accuracy of the PCP ELISA. Added amounts of PCP
(0.50, 1.50, 3.0, an 8.0 ppb) were recovered correctly in all cases with an
average assay recovery of 105%.
Table 3 summarizes the cross-reactivity data using a variety of chlorophenol
compounds. The percent cross-reactivity was determined as the amount of
analogue required to achieve 90% B/Bo (Least Detectable Dose). Many
non-structurally related agricultural compounds demonstrated no reactivity
at concentrations up to 10,000 ppb (data not shown).
Table 4 indicates that no interferences are present from compounds
commonly found in groundwater samples at concentrations much higher
than usually found in those waters (American Public Health Association,
1989).
Recovery data from two different soil types (Piano loam and Sassafras
sandy loam) fortified with pentachlorophenol at 5, 10, and 50 ppm is
summarized in Table 5. To compare extraction efficiencies, soils were also
extracted with methanol using a procedure similar to the procedure listed in
materials and methods. Average recoveries of PCP for the spiked soils were
100% with NaOH/MeOH/H20 and 78% with methanol.
To compare immunoassays results with HPLC, the same extracts prepared
in the above soil recovery procedure were acidified with HCL and filtered
thru 0.7 um glassfiber microfilters. Samples were analyzed by HPLC at the
Air Pollution Lab of New Jersey Institute of Technology. Figure 3 shows
good agreement (correlation = 0.977, slope = 0.975) between the
immunoassay and HPLC methods for the twelve (12) spiked soil samples
listed in Table 5.
SUMMARY
This work describes a particle-based ELISA for the detection of PCP and its
performance characteristics using water and soil samples. The assay
compares favorably to GC/MS and HPLC determinations, 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
441
-------
can be obtained in less than one hour without the problems of variability
encountered with antibody coated tubes and microtiter plates (e.g. coating
variability, antibody leaching, etc.). This system is ideally suited for the
adaptation to on-site monitoring of PCP in water, soil, and solid waste
samples.
REFERENCES
American Public Health Association, Standard Methods for Examination of
Water and Wastewater. American Public Health Association, Washington
DC, 1989.
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, DJ.; 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, 1, 37-51.
Itak, J.A.; Olson, E.G.; Flecker, 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, 51, 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.; Flecker, J.R.; Herzog, D.P.;
Rubio, F.M. Quantification of Cyanazine in Water and Soil by a Magnetic
Particle-Based ELISA. J. Agric. Food Chem. 1993, in press.
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, F.M.; Itak, J.A.; Scutellaro, A.M.; Selisker, M.Y.; Herzog, D.P.;
Performance Characteristics of a Novel Magnetic Particle-Based Enzyme-
442
-------
Linked Immunosorbent Assay for the Quantitative Analysis of Atrazine and
Related Triazines in Water Samples. Food & Agricultural Immunology 1991,
3, 113-125.
Weston, PD, Avrameas, S.; Proteins coupled to polyacrylamide beads using
glutaraldehyde. Biochem. Biophys. Res. Commun. 1971, 45, 1574-1580.
443
-------
Table 1
Pool Number Pool 1 Pool 2
Replicates 5 5
Days 5 5
N 25 25
Mean(ppb) 0.51 1.67
% CV (within) 12.5 8.8
% CV (between) 11.4 8.6
Pool 3 Pool 4
5
5
25
3.16
7.7
1.8
5
5
25
8.63
6.7
3.2
Table 2
PCP added
(DDb)
0.50
1.50
3.00
8.00
Average
PCP observed
(ppb)
0.49
1.63
3.34
8.43
SD
(ppb)
0.09
0.17
0.29
0.71
Recovery
98
108
111
105
105
444
-------
Table 3
Compound
Pentachlorophenol
2,3,5,6-Tetrachlorophenol
2,3,3,6-Tetrachlorophenol
2,3,6-Trichloropheno!
2,3,5-Trichlorophenol
Tetrachloro-hyoroquinone
2,4,6-Trichlorophenol
2,4,5-Trichlorophenol
2,3,4-Trichlorophenol
2,5-Dichlorophenol
2,6-Djchlorophenol
2,3-Dichlorophenol
2,4-Djchlorophenol
3,5 -Dichlorophenol
Hexachlorobenzene
Hexachlorocyclo-hexane
3,4-Dichlorobenzene
4-Chlorophenol
Pentachlorobenzene
Pentachloronitro-benzene
90% B/Bo
LDD (ppb)
0.06
0.21
0.91
2.44
1.52
8.7
15.1
21.5
53.2
62.9
266
611
887
1670
1560
5790
NR
NR
NR
NR
50% B/Bo
(ppb)
2.20
4.06
14.6
62.9
119
148
463
574
1730
7830
5990
> 10000
> 10000
> 10000
> 10000
> 10000
NR
NR
NR
NR
% Cross
Reactivity
100
54.1
15.1
3.4
1.8
1.4
0.5
0.4
0.1
0.03
0.04
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
-------
Table 4
Compound Max. Cone. Tested Interference
Nitrate 250 ppm No
Copper 250 ppm No
Nickel 250 ppm No
Thiosulfate 250 ppm No
Sulfite 250 ppm No
Sulfate 10,000 ppm No
Iron 250 ppm No up to 50 ppm
Magnesium 250 ppm No
Calcium 500 ppm No
NaCI 1.0 M No up to 0.65 M
Humic acid 250 ppm No up to 10 ppm
Silicates 2500 ppm No up to 1000 ppm
Zinc 250 ppm No
Mercury (+ 2) 250 ppm No
Phosphate 250 ppm No
Manganese 250 ppm No
446
-------
Table 5
Extractant
Methanol
Methanol
Soil Type
Piano loam
Sassafras
sandy loam
PCP Spiked
(ppm)
0
5
10
50
0
5
10
50
PCP Observed
(ppm)
nd
2.6
7.7
40.1
nd
5.0
8.6
36.8
Average
Recovery
52
77
80
100
86
74
78
NaOH/MeOH/H20'
NaOH/MeOH/H20
Piano loam
Sassafras
sandy loam
0
5
10
50
0
5
10
50
nd
5.5
8.8
47.2
nd
5.5
10.5
47.6
Average
110
88
94
110
105
95
100
nd = below LDD of assay
1 Methanol =100% methanol
2 NaOH/MeOH/H2O =0.5% NaOH/75% MeOH/25% H2O (w/v/v)
-------
o—<
«—*
A
D
Magnetic Particle Immunoassay
Magnetic Particle with
Antibody Attached
Pesticide Conjugated
with Enzyme
Pesticide
Chromogen/Substrate
Colored Product
1. Immunological Reaction
00
2. Separation
3. Color Development
-------
90.0
o
CD
80.0
70.0
60.0
m 50.0
40.0
30.0
20.0
0.1
10
PCP Concentration (ppb)
Rgure 1. Dose response curve for PCP. Each point represents the
mean of 79 determinations. Vertical bars indicate +/- 1 SD about
the mean.
449
-------
_Q
Q.
Q_
D
W
9
Q.
o
on
140
120 -
100 -
20 -
0
20 40 60 80 100 120
GC/MS Method 625 (ppb)
Figure 2. Correlation between PCP concentrations as determined
by ELISA and GC/MS method 625, in water samples, n = 20, r
= 0.980, y = 1.08 + 2.30
450
-------
50
40
E
D
CO
CO
9 20
Q_
O
Q:
10
0
0
10
20 30
HPLC (ppm)
40
50
Figure 3. Correlation between PCP concentrations as determined
by ELISA and HPLC in soil samples, n = 12, r = 0.977, y =
0.98 + 1.39.
451
-------
59 CHROMATOGRAPHIC OPTIMIZATION FOR THE ANALYSIS OF
AN EXPANDED LIST OF VOLATILE ORGANIC POLLUTANTS
DAVID M. SHELQW AND MICHAEL J. FEENEY
Restek Corporation 110 Benner Circle, Bellefonte, PA 16823
Volatile organic analysis (VOA) is one of the most common environmental test
procedures for analysis of pollutants in air, water, and spil. Current methodology specified
in EPA methods involves concentrating the trace volatile organics onto an adsorbent
material and then thermally desorbing them onto a chromatographic column. The list of
target compounds to be analyzed are similar for air, water and soil matrices and there are
indications that additional compounds will be added to the target compound lists in the
future. When the target lists of EPA methods TO-14, 524.2, 624, 8240 and 8260 are
combined, including various surrogates and internal standards, the list contains over 100
compounds. Resolution of all of these compounds in a single chromatographic analysis can
be difficult if not impossible, however, through proper column selection and
chromatographic optimization most compounds can be resolved.
This paper will describe a new technique for GC method development using
thermodynamic retention indices and computer modeling of the chromatographic process.
From the compound retention times using two different chromatographic analyses, it is
possible to predict the best oven temperature and flow conditions for the target volatile
compounds. This greatly simplifies the task of optimizing the GC method development,
especially for complex mixtures. The application of thermodynamic retention indices will
be demonstrated for the analysis of an expanded list of volatile organic compounds.
Examples will be shown using the 502.2 and 624 stationary phases which are the most
commonly used phases for these analyses.
452
-------
APPLICATION OF SW-846 METHODS TO THE IDENTIFICATION OF UNUSUAL
BROMINATED COMPOUNDS IN HIGH-CONCENTRATION PROCESS STREAMS.
D.V. Smith and L.P. Pollack, Science Applications International
Corporation, 7600A Leesburg Pike, Falls Church, Virginia 22043, and F.T.
Varcoe, Triangle Laboratories Incorporated, P.O. Box 13485, Research
Triangle Park, North Carolina, 27709.
Abstract
This presentation reports work performed to characterize
manufacturing residuals containing brominated and other organic compounds.
SW-846 methods were used for analysis of volatile and semi-volatile
organics and selected metals. Methods developed for the TSCA testing
program were used for analysis of brominated dioxins and furans. Metals
analysis was performed using ICP/MS instrumentation to utilize its
superior dynamic range. However, it was found that selenium could not be
reliably determined by this method because of interference from bromine
isotopes of similar atomic mass. GFAA was evaluated as a specific
confirmatory method for selenium in this environment. It was also noted
that the use of SW-846 methods 8260 and 8270 as written may not provide
the level of QC needed for complete evaluation of the volatile and
semivolatile organic data. To quantify the predominant organic compounds
in the samples collected, significant dilutions were required resulting in
loss of some spike recovery data. The paper presents options for sample
extraction and analysis which may assist in overcoming these difficulties.
Introduction
The purpose of this presentation is to describe some special challenges
that were encountered in analyzing a series of high-concentration
industrial wastes containing brominated compounds.
Background
A study of flame retardant manufacturing resulted in a need to sample
several process waste streams containing brominated compounds. These
samples were analyzed for volatile and semi-volatile organics, metals, and
brominated dioxins and furans. Several aspects of this work were unusual
and demonstrated issues not normally encountered in analysis of other
industrial wastes or environmental media:
o Brominated dioxins and furans were suspected to be present in these
samples and newly developed methods were used to identify and
quantitate these species.
o It was found that bromine present in the samples interfered with the
determination of selenium by ICP-MS.
o Many of the samples contained elevated concentrations of one or two
semivolatile organics which made determination of other species more
difficult.
o Some samples consisted of 20% NaOH, making field adjustment of pH
impossible and complicating laboratory handling.
453
-------
Dioxins/Furans
SW-846 Method 8290 addresses the determination of chlorinated dioxins and
furans in solid wastes, but does not include their brominated analogs.
Methods have been developed for these compounds [1-3] to meet the
requirements of EPA's dioxin product test rule (40 CFR §766). Methods
development was sponsored by the Brominated Flame Retardant Industry Panel
(BFRIP). In this work, we applied these methods to industrial wastes.
Some samples contained brominated diphenyl ether flame retardants, which
could interfere with the determination of the furans. The methods include
HPLC cleanup procedures to remove \these compounds.
A typical mass of spectrum of 1,2,3,7,8-pentabromodibenzofuran is shown in
Figure 1. Criteria ions used for identification of this compound are as
follows:
Mass
Composition
Abundance'
559.608
561.606
563.604
C12H379Br481BrO
C12H379Br3Br20
C12H379Br281Br30
0.51
1.00
0.98
(1)
Theoretical abundance relative to peak at 561.606
Brominated compounds are immediately recognizable in a mass spectrum
because the element consists of two isotopes of almost equal abundance
(50.54% 79Br and 49.46% 81Br) giving rise to series of doublets, triplets,
etc. Note that the major ions present have almost twice the mass of those
which would be seen in the spectrum of the corresponding chlorinated
dibenzofuran.
07
135 Er'^^^^Cr^^^'^Br
163 214 2f2
1 • i1, -r 1- ' "
A
1
295
1
323
t
125 150 175 200 225 250 275 300 325 350
323
1] 374
404
,1 «5 «2
56]
It
325 350 375
400
425 450 475
500
525 550
Figure 1. Mass Spectrum of 1, 2, 3,7,8-Pentabromodibenzofuran (Courtesy
of Mike Re, Radian Corporation)
454
-------
A unique problem encountered in the application of these methods is the
fact that methylene bromide, the extraction solvent of choice, is not
available commercially in the same high purity as methylene chloride. The
analyst may have to purify the methylene bromide and verify that it is
free of contaminants before beginning sample extractions.
ICP/MS
All of the metals included in the definition of the Toxicity
Characteristic, except mercury, were quantitated using Method 6020 (ICP-
MS) . This method generally gives superior detection limits together with
broad dynamic range. However, we found that this method cannot be applied
to the determination of selenium in samples containing bromine because of
mass-number interference. Selenium has six stable isotopes which span the
same range as the isotopes of bromine:
Isotope
7
-------
Sample
17903
100
I I
125
As
Se
Ag
Calibration
Blank
,Cd
i I i
100
125
Figure 2. ICP-MS Spectrum of (top) a brominated -sample containing no
selenium, and (bottom) the calibration blank.
Acknowledgement s
The authors wish to thank Edwin Rissmann, U.S. Environmental Protection
Agency, Bruce Colby and the staff of Pacific Analytical Incorporated, Mike
Re, Radian Corporation, and Siu-Fai Tsang, SAIC, for their assistance and
guidance during this project.
References
[1] Y. Tondeur, "Analytical Protocol for the Determination of
Polybrominated Dibenzo-p-dioxins and Dibenzofurans by High-
Resolution Gas Chromatography/Medium High-Resolution Mass
Spectrometry in Pentabromodiphenyl Oxide", Prepared by Triangle
Laboratories for the Brominated Flame Retardant Industry Panel,
January 1991. Similar protocols have been prepared for
decabromodiphenyl oxide, octabromodiphenyl oxide, tribromophenol,
and tetrabromobisphenol A.
[2] J.R. Donnelly, W.D. Munslow, T.L. Vonnahme, N.J. Nunn, and C.M
Hedin, "The Chemistry and Mass Spectrometry for Brominated Dibenzo-
p-dioxins and Dibenzofurans", Biom. and Environ. Mass Spect.,
14:465-472 (1987).
[3] J.R. Donnelly, W.D. Munslow, A.H. Grange, T.L Petitt, R.D. Simmons,
K.S. Kumar, and G.W. Sovocool, "A Gas Chromatographic/Mass
Spectrometric Approach for Isomer-Specific Environmental Monitoring
of the 1700 Bromo-, Chloro-, and Bromochloro- Dibenzo-p-dioxins",
Biol. Mass Spec., 20:329-337 (1991).
Notice: This work was funded by the U.S. Environmental Protection Agency under contract, no.
68-WO-0027 It has not been reviewed by the Agency and the contents do not necessarily
reflect the views and policies of EPA. Mention of trade names, commercial products, or
publications does not constitute1 endorsement or recommendation for use.
456
-------
R1 RESULTS OF LABORATORY TESTS OF THE ACCELERATED ONE-STEP™ LIQUID-
w ' LIQUID EXTRACTOR
Roy-Keith Smith. Ph. P.. Robert G. Owens, Jr.- Denise S. Geier, Timothy Allen Shaw and
Donna Phinney
Analytical Services, Inc, 390 Trabert Ave, Atlanta, GA 30309
James B. Carl
Coming Incorporated, Corning, NY 14831
ABSTRACT
A series of experiments were conducted using the Corning Accelerated One-Step™ liquid-liquid
extractor/concentrator for BNA and TPH analyte isolation on real samples. The results indicate
the device gives comparable results to the traditional designs of continuous liquid-liquid extractors
in a fraction of the time and solvent usage. The device was evaluated as a replacement for the
separatory funnel liquid-liquid technique, and for the most part offers distinct advantages.
INTRODUCTION
Continuous liquid-liquid extractors play a key part in the laboratory analysis of extractable
semivolatile organics such as PAHs, PCBs, chlorinated insecticides, acidic herbicides, phthalates,
and a host of other pollutants from industrial wastewaters. Use of the extractors avoids formation
of intractable emulsions, a common problem in the separatory funnel liquid-liquid extraction
methods. Further the extractors are not prone to unexpected selective retention of target analytes
or plugging due to particulates, as are the proposed solid phase extraction devices. However, the
major drawback to the use of continuous liquid-liquid extractors is the 18-24 hours required for an
extraction. This is particularly a problem when the sample must be made acidic, extracted, then
made basic and re-extracted. A second problem is the decomposition of sensitive target analytes
in the strongly basic or acidic solution during the lengthy extraction. A third, and potentially the
worst problem, is the relatively large volumes of chlorinated solvents which are used and which
must be either disposed or recycled at the end of the extraction.
In 1991 Corning Glass Works invented a fundamental modification of the continuous liquid-liquid
extractor, by addition of a hydrophobic membrane to the bottom of the extraction chamber. The
membrane allows passage of non-polar organic solvents and dissolved materials, while retaining
polar and aqueous liquids. Use of the membrane eliminates the solvent pool at the bottom of the
extraction chamber and renders obsolete the solvent siphon tube. The membrane serves a second
function in drying the solvent as it passes through, eliminating the need for sodium sulfate drying
columns after the extraction process.
Analytical Services, Inc. began testing the Accelerated One-Step™ Extractor in January, 1993.
Objectives of the test experiments were to determine the amount of time necessary for extraction
of analytes, determine the range of extractions which could be performed with the device, examine
which solvents were compatible with the device and produce date comparing the effectiveness of
the device with the liquid-liquid separatory funnel and more tradition continuous liquid-liquid
extractors. The results of these test objectives and some observations of the device in use are the
subject of this paper.
METHODS AND MATERIALS
BNA surrogates, matrix spike solutions and internal standards were purchased as concentrates
from Supelco (Belafonte, PA) or Ultra Scientific (Kingston, RI) and were those specified in SW-
846, 3rd Edition. Final Update. Methylene chloride, 1,1,2-trichlorotrifluoroethane (Freon 113™),
carbon disulfide, acetone and hexane were of the highest quality available (Pesticide Residue
grade or Nano-grade™) from Fisher Scientific. All volumetric measurements were performed in
Class A volumetric glassware. BNA analyses were performed by EPA method 8270A, 1990 on a
Hewlett-Packard 5890A, Series II GC with 5971 Mass Selective detector. The capillary column
used was a Hewlett-Packard HP5-MS column. Daily QC acceptance criteria for DFTPP tuning,
457
-------
SPCC, and CCC were met prior to analysis, as specified in the method. TPH analyses were
performed by modified EPA method 418.1.
The extractor was cleaned prior to each use with detergent, brush and hot water, rinsed with hot
water, rinsed with chromic acid solution, rinsed with hot water, rinsed twice with DI (organic free)
water and allowed to air dry. Immediately before use it was rinsed with the extraction solvent.
The membrane was rinsed immediately prior to use with the extraction solvent. After assembly (3
to 5 Teflon™ boiling chips were added to the concentrator flask), 50 mL of solvent was added to
the extraction chamber with the stopcock closed. The sample was added to the extraction
chamber, then the container rinsed with 50 mL of the extraction solvent which was added to the
chamber. The surrogates and/or matrix spikes were added to the sample with a volumetric pipet
followed by pH adjustment with 1:1 sulfuric acid or 10 M sodium hydroxide. pH was read with
wide range paper. The stopcock was opened and about half of the solvent was allowed to drain
into the concentrator flask. Heat was applied to the concentrator flask. Extraction times were
established from the first drops of solvent to fall from the condensor into the sample. At that time
the stopcock was opened and all the solvent allowed to drain into the concentrator. A distillation
rate of 15 to 25 mL per minute was maintained. At the end of the extraction the stopcock was
closed and the solvent concentrated to about 10 mL, at which time the heat source was removed.
The apparatus was allowed to cool, then the concentrator was removed. The solvent was removed
to 2 mL under nitrogen blowdown. The concentrate was quantitatively transfered to a Class A 2
mL volumetric ground glass stoppered tube which had previously been calibrated at the 1.00 mL
level. The sample was reduced to 1.00 mL with nitrogen blowdown, internal standards were
added and the sample sealed and shaken followed by transfer to an autosampler vial for analysis.
TPH extractions were similar except there was no solvent concentration.
RESULTS AND DISCUSSION
The first set of experiments were performed by adding acid matrix spike (200 ng in methanol per
method 3500) and BN surrogate (100 ng in methanol per method 3500) compounds to DI water at
neutral pH and extracting them with methylene chloride for 1, 2,4 and 16 hours. The results are
in Table 1 with example plots shown in Figures 1 and 2 . The point at time 30 hours (PES)
represents the results of a post-extraction spike and analysis of the surrogate mixture. Poor results
were obtained for the acidic spikes (phenol, 2-chlorophenol, 4-chloro-3-methylphenol and 4-
nitrophenol) as expected for pH 7. The BN surrogates (nitrobenzene -ds, 2-fluorobiphenyl and
terphenyl-di4) were recovered with above 90% efficiency within 4 hours. Surprisingly
pentachlorophenol was recovered in 90% efficiency in 4 hours at pH 7 (Figure 3).
Time 1 2 4 18 PES
Compound
Phenol 35.1 61.3 58.8 69.1 194
2-Chlorophenol 18.8 43.2 37.6 40.1 172
Nitrobenzene d5 45.3 72.2 91.5 93.0 95.4
4-Cl-3-methylphenol 8.85 23.8 14.5 168
2-Fluorobiphenyl 56.9 75.8 90.9 90.6 99.8
4-Nitrophenol 23.2 45.2 66.7 86.3 174
Pentachlorophenol 52.9 123 175 192 195
Terphenyldl4 82.2 74.9 86.4 86.6 82.1
Table 1. Acid matrix spike compound and BN surrogate recoveries (ng) for extractions at various
times (hours) at pH 7.
458
-------
PES Result
ng Phenol spike recovery at pH 7
15 20
Time Hours
Figure 1. Phenol recovery (in ng) vs. time (hours) extracted at pH 7. The PES represents the
amount of spike placed in the extractor as determined by neat analysis.
-o
-O PES Result
ng Nitrobenzene d5 recovery at pH 7
10
1 5 20
Time Hours
25
30
35
Figure 2. Nitrobenzene ds recovery (ng) vs. time (hours) extracted at pH 7. The PES represents
the amount of spike placed in the sample as determined by neat analysis.
459
-------
PES Result
ng Pentachlorophenol spike recovery at pH 7
25
10
25
30
35
15 20
Time Hours
Figure 3. Pentachlorophenol recovery (ng) vs. time (hours) extracted at pH 7. The PES represents
the amount of spike placed in the extractor as determined by neat analysis.
The experiment was repeated with initial acidification to pH < 2 with extraction times of 1,
2, and 4 hours. Over 80% recoveries of all phenolic compounds (Table 2.), except 4-nitrophenol,
were obtained, as illustrated for phenol (Figure 4). The 4-nitrophenol results are illustrated in
Figure 5. These results indicated that 4 hours extraction for each pH change would give suitable
results.
Tune 1
Compound
Phenol 105.7
2-Chlorophenol 127.2
Nitrobenzene d5 69.2
4-Cl-3-methylphenol 128.3
2-Fluorobiphenyl 69.9
4-Nitrophenol 94.9
Pentachlorophenol 129.1
Terphenyl d!4 72.5
131.9
144.8
85.7
165.8
82.0
146.3
187.1
80.2
165.2
165.2
92.3
190.9
90.1
162.4
205.7
92.6
PES
188.4
185.1
98.4
191.4
100.7
210.8
197.0
89.1
Table 2. Acid matrix spike compound and BN surrogate recoveries (ng) for extractions at various
times (hours) at pH <2.
460
-------
PES Result
ng Phenol spike recovery at pH <2
0 5 10 15 20
Time Hours
Figure 4. Phenol recovery (ng) vs. time (hours) extracted at pH <2. The PES respresents the
amount of spike placed in the extractor as determined by neat analysis.
PES Result
ng 4-Nilrophenol spike recovery at pH <2
10
25
30
35
15 20
Time Hours
Figure 5. 4-Nitrophenol recovery (ng) vs. time (hours) extracted at pH <2. The PES respresents
the amount of spike placed in the extractor as determined by neat analysis.
The extraction of TPH samples for IR analysis with Freon 113™ (1,1,2-trichlorotrifluoroethane)
was examined. Spikes of the calibration standard mix (isooctane, hexadecane and chlorobenzene),
were made into Dl water and extracted for various times. The results are presented in Figure 6.
The first three data points were collected and analyzed on one day and the 8 and 22 hour
461
-------
extractions performed on another day. The results are equivalent given the ±10% acceptance
criteria of the calibration curve and indicate that a 1 to 2 hour extraction time is suficient.
120 -i
100 -
-Q
2.5
7.5
10 12.5 15 17.5 20 22.5
Hours
Figure 6. THP standard recovery (percentage) from DI water vs. duration (hours) of extraction
using Freon 113 as the solvent.
The ability of the hydrophobic membrane to pass organic compounds was evaluated by placing 50
mL of a solvent in the extraction chamber and recording the time required to drain. The data are
presented in Table 3. Most of the solvents tested passed through the membrane within the same
time window, Freon 113 being marginally faster. As expected the hydrogen bonded solvent, n-
butanol was significantly retarded. Carbon disulfide has been proposed as a possible Freon 113
replacement for the IR-TPH test, however it is not suitable for use with the Accelerated One-
Step™ membrane, as shown in Table 3. Carbon disulfide is an opposite polarity compound
(carbon is polarized negative and sulfur positive) as compared to the other halogenated and
hydrocarbon solvents tested. These data support the observations of the lower extraction
efficiency of the hydrogen bonding carbamates, amines and highly acidic phenols, such as the
mono- and di-nitrophenols, as reported by Bruce (1993).
Solvent
Freon 113
Hexane
Benzene
Chloroform
Methylene chloride
tert- Butyl methyl ether
Acetone
n-Butanol
Carbon disulfide
Water
Minutes. Seconds
8.45
10.40
16.07
11.43
9.55
12.42
10.17
55.0
< 10 mL after 55.0
none after 120
Table 3. Time (minutes) for 50 mL of solvent to pass through the hydrophobic membrane of the
Accelerated One-Step™.
462
-------
A number of real samples were then extracted using 4 hours extraction times each for the acid and
basic extractions. Two different samples, a TCLP extraction and a wastewater sample were
performed as BNA matrix spike (MS) and matrix spike duplicate (MSD), as part of a normal batch
of samples. The MS was extracted using the separatory funnel technique of Method 3510 with
sodium sulfate drying column and Kuderna-Danish concentration, while the MSD was extracted
using the One-Step™ extractor and direct concentration, followed by nitrogen blowdown. The
results are presented in Table 4 and demonstrate that the One-Step™ extraction is at least as good
as the separatory funnel technique of Method 3510 for normally encountered samples.
TCLP Sample Wastewater Sample
Surrogate Compounds Sep Funnel One-Step™ Sep. Funnel One-Step™
2-Fluorophenol 35 62 33 49
Phenol d5 28 79 25 65
Nitrobenzene d5 88 86 89 85
2-Fluorobiphenyl 75 83 82 77
2,4,6-Tribromophenol 84 89 75 81
Terphenyldl4 90 88 83 73
Matrix Spike Compounds
Phenol 15 71 11 60
2-Chlorophenol 31 73 29 65
1,4-Dichlorobenzene 39 77 15 65
N-Nitrosodipropylamine 57 103 45 81
1,2,4-Trichlorobenzene 43 80 20 66
4-Chloro-3-methylphenol 35 86 26 85
Acenaphthene 49 88 45 79
2,4-Dinitrotoluene 42 74 42 80
4-Nitrophenol 14 74 15 75
Pentachlorophenol 48 97 39 78
Pyrene 55 83 44 77
Table 4. Comparison of percent recoveries of BNA matrix spike and surrogate compounds on the
same samples using Method 3510 (separatory funnel) and the modified 3520 using the
Accelerated One-Step™
An industrial wastewater sample from a industrial cleaner formulator was analysed for acid
extractables as a sample and as a matrix spike with separatory funnel extraction. Past history of
the sample indicated it needed to be diluted 1:20 (50 mL sample to 950 mL Dl water) to avoid
swamping the GC/MS detector. The same sample at 1:20 dilution was extracted on 4 consecutive
days as a BNA matrix spike with the Accelerated One-Step™. The results are presented in Table
5. The One-Step™ consistently produces higher recoveries of the more volatile compounds such
as phenol, 1,2-dichlorobenzene, 2-chlorophenol and benzoic acid, while prerforming about as well
as Method 3510 on the rest of the compounds.
463
-------
Compounds
Analytes
1,2-Dichlorobenzene
2-Methylphenol
4-Methyl phenol
Isophorone
Benzole acid
Naphthalene
Diethylphthalate
Di-n-butylphthalate
Bis(2-ethylhexyl)phthalate
Surrogates
2-Fluorophenol
Phenol d5
Nitrobenzene d5
2-Fluorobiphenyl
2,4,6-Tribromophenol
Terphenyl d!4
Matrix Spike Compounds
Phenol
2-Chlorophenol
1,4-Dichlorobenzene
1,2,4-Trichlorobenzene
4-Chloro-3-methylphenol
Acenaphthene
2,4-Dinitrotoluene
4-Nitrophenol
Pentachlorophenol
Pyrene
Sep. Funnel
Sample MS
32 32
Accelerated One-Step
MSI MS2 MS3
,TM
3
6
103
4
5
28
4
60
80
75
58
177
85
2
1
72
3
1
12
3
54
74
92
47
153
63
67
119
65
62
162
53
63
48
130
47
97
15
9
5
173
7
12
26
6
90
201
95
59
157
71
205
197
102
74
182
68
77
116
88
71
97
139
2
249
8
15
12
7
63
176
86
49
148
68
155
175
95
67
177
57
79
119
85
63
91
128
14
3
189
7
11
18
5
72
182
88
55
146
72
160
172
82
63
173
67
82
132
94
73
MS4
65
10
1
120
5
10
11
1
87
168
80
35
128
57
171
151
75
52
165
44
73
131
77
47
Table 5. Comparison of recoveries of target analytes and matrix spike compounds for a strong
industrial effluent processed with a separatory funnel by method 3510 and processed on 4
consecutive days with the Accelerated One-Step™ extractor. Acid compounds are spiked at the
200 ng level while BN compounds are at the 100 ng level.
A wastewater treatment plant sludge (6% solids) was extracted for BNA analysis using a
continuous liquid-liquid extractor and the Accelerated One-Step™ extractor. Another portion of
the same sample was diluted 1:30 with DI water and 1 L extracted using Method 3510 with a
separatory funnel. The membrane of the Accelerated One-Step™ did not plug, however
channeling of the solvent through the sludge was noticed. The surrogate recoveries for the three
extraction techniques are presented in Table 6. Although none of the recoveries were quantitative,
the Accelerated One-Step performed significantly better than the other two techniques.
464
-------
Sludge sample with 6% solids
Surrogate Compounds Sep Funnel Cont. Liq-Liq One-Step™
2-Fluorophenol 006
Phenol d5 5 14 26
Nitrobenzene d5 47 33 63
2-Fluorobiphenyl 27 31 60
2,4,6-Tribromophenol 2 3 49
TerphenyldH 31 45 65
Table 6. Comparison of percentage suggogate recoveries of BNA separately funnel, continuous
liquid-liquid and Accelerated One-Step™ extraction of a sludge sample containing 6% solids.
The separatory funnel extraction was performed on a 1:30 dilution of the sludge.
Three other wastewater samples have been extracted for BN or BNA analysis. The surrogate
recoveries are presented in Table 7. In general, use of the Accelerated One-Step™ extractor gives
recoveries which are comparable to those obtained using separatory funnel techniques in each
respective batch.
Surrogate Compounds 41082-1 40619 40598-1
2-Fluorophenol - 65 (56)
Phenol d5 93 (62)
Nitrobenzene d5 115(121) 101(108) 69(73)
2-Fluorobiphenyl 105(88) 74(96) 69(74)
2,4,6-Tribromophenol 105 (103)
Terphenyl d!4 87(98) 74(79) 50(61)
Table 7. Percentage surrogate compound recoveries for wastewater samples extracted and
concentrated using the Accelerated One-Step™ Samples 41082-1 and 40598-1 were extracted
for base/neutrals only. The values in parentheses are the average recoveries for the batch.
Two instances of membrane plugging were encountered. The first was a sample which contained
a fine dense grit which coated the membrane and was not displaced by the methylene chloride
solvent. The sample was successfully prepared by separating the solid from the liquid layer,
separately extracting the solid with sonication and the liquid with the One-Step™, then combining
the extracts. The second was a sample which formed a tenacious emulsion as the solvent fell from
the condenser. The surface of the bubbles coated the membrane and stopped the flow of solvent
through it. The sample had to be processed by continuous liquid-liquid extraction followed by
removal of the large amounts of water in the extract by slurrying with sodium sulfate.
SUMMARY
The Corning Accelerated One-Step™ liquid-liquid extractor/concentrator has been shown to offer
rapid extraction of target analytes with acceptable recoveries and minimal solvent usage as
compared to more traditional liquid-liquid extractors. When compared to separatory funnel
liquid-liquid extraction, the Accelerated One-Step™ is shown to be a viable alternate with
demonstrated savings through reduced solvent usage and elimination of the drying column and
Kuderna-Danish concentrator with no decrease in the recovery of most analytes.
REFERENCES
Bruce, M.L., 1993. Extraction of Appendix IX BNAs, Pesticides and PCBs by Accelerated One-
Step™ Liquid-Liquid Extractor/Concentrator with Analysis by GC/MS & GC/ECD. A Single
Method Validation Report. Enseco-Wadsworth/ALERT Laboratories, 4101 Shuffel Dr NW
North Canton OH 44720
465
-------
A MATCHED DUAL CAPILLARY COLUMN SYSTEM FOR PESTICIDES
Mitchell R. Hastings, Alan J. Spilkin. J&W Scientific, 91
Blue Ravine Road, Folsom, California 95630. Phone (916) 985-
7888
ABSTRACT
In early and existing GC pesticide analyses methods, packed
columns were and are employed. The use of packed columns
has, in many cases, given way to the use of Megabore (0.53
mm I.D.) columns in these methods. Many laboratories and
regulatory agencies have realized that, relative to packed
columns, Megabore columns offer greater reproduceability,
longer lifetimes, increased resolution, and decreased
overall analysis times.
Many laboratories, among them those doing pesticide analyses
for the EPA and FDA, have both a high sample throughput and
require quick sample turnaround times. It follows that any
enhancement to the GC system that increases resolution and
decreases run times will enhance the performance and profits
of these laboratories. This decrease in run time becomes
even more attractive if minimal to no hardware changes need
be made. What follows is an experiment converting the
chromatography of an EPA Contract Laboratory Program (CLP)
Chlorinated Pesticide analysis currently done by two
Megabore columns to an analysis utilizing two 0.32 mm I.D.
columns. A reduction in run time of over 25% is realized
while still maintaining the ability to use a Splitless or
Megabore direct injection inlet.
INTRODUCTION
BACKGROUND
Quantitative and confirmatory qualitative information can be
obtained for a list of analytes by running the analytes on
two columns with significantly dissimilar stationary phases.
Many state and federal regulatory agencies accept this data
as conclusive. The benefit of single injection, dual column
analysis is a time savings by having both quantitative and
confirmatory qualitative information available from one run.
For example, from analysis of a sample using one column,
tentative identification of the analyte of interest is made
if its retention time matches the retention time of the
standard run on that column. Quantitative confirmation the
presence of that component on the column with the different
stationary phase if the component's retention time also
matches the standard's retention time on that column. If
466
-------
components are resolved on both columns, quantitation data
can be compared to insure quantitative accuracy. This
technique becomes even more powerful if a selective detector
is employed.
In the latest available version (OLMO1.3, 2/91) of the CLP
Chlorinated Pesticide Method, two 30 m x 0.53 mm I.D.
columns with dissimilar stationary phases are used (DB-608,
50% phenyl 50% methyl polysiloxane; DB-1701, 14%
cyanopropylphenyl, 86% methyl polysiloxane) and a selective
detector (ECD) is employed in the analysis of 20 chlorinated
pesticides. It was the purpose of this experiment to
determine if two 0.32 mm I.D. columns of the same stationary
phase type could be utilized in order to decrease the run
time, while still meeting the required resolution criteria,
and without changing hardware or increasing hardware costs.
INLET AND COLUMN CONNECTION CONSIDERATIONS
A number of different options were considered when designing
this single injection, dual column system. Five criteria
were used to evaluate certain inlet and column connection
options. The inlet and column connection design had to be:
1) equal or lower in cost compared to the current low cost
of a Megabore configuration--the least expensive mode,
2) adaptable to splitless or Megabore direct injectors since
both injection techniques are currently in use,
3) highly inert (some analytes degrade easily when exposed
to active sites) by eliminating any metal or other active
sites in the sample flow path,
4) easily, reliabilly, and reproducibly installed,
5) amenable to the use of two 0.32 mm I.D. columns.
Table I is a report card of how each option scored in the
above categories.
Table I
Option
Qualified Unqualified
Graphpak Divider for Simul-
taneous Sampling to Two
Capillary Columns (Hewlitt
-Packard)
Injection Tee Kit for
Simultaneous Analyses
on Two Wide Bore Columns
(Supelco)
4,5
1,2,3
1,3
2,4,5
467
-------
Two Hole Ferrule 1,2,3,5 4
Deactivated Megabore Guard
Column with a Deactivated
0.53/0.32/0.32 mm 1,2,3,4,5
3-Way "Y" Union
To reduce activity, a deactivated, splitless liner would
have been used if splitless injection mode was used. This
particular experiment employed the Megabore direct injection
mode as it was thought that this technique could potentially
pose sample transfer and solvent front problems1 that would
not result from the use of the splitless mode (see "Inlet
and Column Flow Considerations" below). A deactivated,
Direct Flash Vaporization liner (J&W Scientific)—the type
of liner that allows the Megabore guard column to seal into
a tapered region below an expansion chamber—was used with
the Megabore direct injection mode. The configuration
decided upon is shown in Figure I.
COLUMN DIMENSION CONSIDERATIONS
Provided that temperature, phase ratio, stationary phase
type, linear velocity, test compound, and length are kept
constant—smaller diameter columns yield an increased number
of theoretical plates and sharper peaks than larger diameter
columns.
As a demonstration of this, two 30 meter, DB-608 columns of
the same phase ratio but with differing diameters (0.53 and
0.32 mm I.D.'s) were tested under identical temperature,
carrier gas, and linear velocity conditions. To simplify
the mathematics, the equation used to describe the area of
the Gaussian peak ideally obtained from a chromatogram was
that of a triangle: Area = 1/2 Base x Height. Theoretical
plates were determined using the same compound, and width
was calculated using the following formula2:
w = 4t / n1/2
where
w = width at the base of the peak
t = retention time of the peak
n = calculated theoretical plates
If the same amount is injected on each column, we will
assume this amount generates the same number of area counts
whether chromatographed using the 0.53 or 0.32 mm I.D.
column. So if the area remains the same and the width of
the peak decreases, the height of the triangle must
increase.
468
-------
So we can say that:
0.5 x Base0_53 x Height0>53 = 0.5 x Base0>32 x Height0-32
or
Base0_53/Base0_32 = Height0i32/Height0>53
The 0.53 mm I.D. column generated 40400 theoretical plates
and the retention time of the test compound was 11.523 min.,
therefore the width of the base of this peak is 0.229 min.
The height was set at 1 unit since we are trying to see the
resulting ratio difference in height.
The 0.32 mm I.D. column generated 85100 theoretical plates
and the retention time of the test compound was 11.460 min.
which makes the base of the peak 0.157 min wide. Therefore,
all of the above parameters remaining the same, the increase
in the peak height resulting from the use of the 0.32 mm
column as opposed to the 0.53 mm I.D. is approximately 1.5
for the same compound when using the triangle approximation.
This figure cannot be exactly correlated to peaks of
differing retention, or peaks eluting on the ramp portion of
a temperature programmed run, but it does demonstrate a
benificial peak height, or signal to noise increase which
implies a resulting increase in system sensitivity when
using a smaller, rather than a larger diameter column.
INLET AND COLUMN FLOW CONSIDERATIONS
With the above column considerations in mind, why not use
0.25 mm I.D. or smaller diameter columns in this single
injection, dual column setup and really minimize run times
and system sensitivity?
In order for sample transfer to occur efficiently (in a
short period of time and with minimal band broadening) from
a Megabore direct inlet to a column(s), their generally
needs to be a minimum of 4-5 mL/min. of carrier flow through
the inlet3. Therefore, the summed flow through the two
columns must be at least 4-5 mL/min.
As demonstrated by Van Deemter curves4, carrier gases have
flow ranges for particular length, diameter, and film
thickness columns in which they will yield the most
efficient (greatest number of theoretical plates)
chromatography. Table II shows what are considered general,
effecient flow ranges for different diameter, length, and
film thickness columns for components of varying retention
using Helium or Hydrogen as a carrier gas5.
469
-------
Table II
He Efficient H2 Efficient
Inner Diameter (mm^ Flow Ranae (mL/min) Flow ranae (mL/min)
0.25
0.32
0.53
0.9-1.3
1.4-2.2
4.0-6.0
1.3-1.8
2.2-2.9
6.0-7.9
Since most analysts doing the CLP Chlorinated Pesticide
analysis are using Helium as a carrier gas, and since
Helium's efficient flow range is lower than that of
Hydrogen, Helium was considered the limiting parameter. It
was selected as the carrier gas in this experiment as its
lower efficient flow range could potentially cause flow
problems in the Megabore inlet, or chromatography problems
if it was necessary to operate above the efficient flow
range to avoid inlet problems.
Allowing for the flow limitations of the Megabore inlet and
the Helium carrier gas, it becomes apparent that the
smallest diameter column that could be most efficiently
employed in a single injection, dual column analysis is the
0.32 mm I.D. column, the efficient flow range for two
columns summed being 2.8-4.4 mL/min.
If sample transfer problems from the inlet to the column
exist, they would be manifested in two ways. First, if the
solvent front interfered with the chromatography of the
first eluting peak, it would tail, be excessively broad,
and/or possibly rounded at the top, especially on the
thinner film (0.25 jum) DB-1701 column. Second, if the flow
through the inlet was inadequate, all peaks might tail.
ADDITIONAL COLUMN CONSIDERATIONS
There are a number of advantages, relavent to this
experiment, that smaller diameter columns have over larger
diameter columns of the same length and phase ratio. They
produce less bleed because there is less stationary phase
available to degrade. Under proper flow conditions, as
demonstrated above, smaller diameter columns generally yield
sharper peaks. Additionally, because there are more
theoretical plates generated by smaller diameter columns,
the same resolution can be achieved in a shorter period of
time. This means compounds spend less time in the column
and, since the width of a peak is directly related to the
amount of time the compound spends in the column (the amount
of time it has to diffuse longitudinally in the mobile
phase), the compound suffers less from band broadening when
it elutes earlier. For the same amount of analyte on
column, this results in narrower, taller peaks. All of
470
-------
these factors translate directly into a greater signal to
noise ratio, and, more importantly, lower, more easily
reached detection limits.
There are, however, relavent disadvantages of using smaller
diameter columns as opposed to Megabore columns. Smaller
diameter columns of the same phase ratio are generally
fouled more quickly by non-volatile sample residues. This
fouling can be minimized by the use of an adequate length of
guard column and/or by the use of a plug of silanized glass
wool in the inlet. Smaller diameters of the same phase
ratio are also thought to be slightly more active than
larger diameter columns because the thinner stationary phase
allows the analytes more interactions with the surface of
the fused silica tubing than with the thicker film Megabore
columns. Also, assuming analytes are well separated from
the solvent front on a 1 or 2/iL injection, up to 8p.L can be
injected on a megabore column with a standard film thickness
without significant negative chromatogrphic consequences
(retention time shifts from run to run and noticeable band
broadening of components) . At most, only 3-4/iL should be
injected on the 0.32 mm I.D. dual column setup. So, if
injection volumes greater than 3-4juL are normally used, the
greater signal to noise generated by using a 0.32 as opposed
to a 0.53 mm I.D. column would be somewhat negated.
MEETING THE RESOLUTION REQUIREMENTS
In the latest available version (OLMO1.3, 2/91) of the CLP
Chlorinated Pesticide Method, certain resolution
requirements must be met. All components must be adequately
resolved, particularly the targeted pesticides in Table III
which are ordinarily difficult to resolve on the Megabore
system. Acceptable resolution is defined as the depth of
the valley between two adjacent peaks as being greater than
or equal to 60% of the height of the shorter peak.
Table III Pesticides in CLP Resolution Check Mixture
gamma-Chlordane Endosulfan Sulfate
Endosulfan I Endrin Ketone
p,p'-DDE Methoxychlor
Dieldrin
SUMMARY
As shown in Figure 2, peak shapes and widths with a 2 /xL
injection for tetrachloro-m-xylene, the earliest eluting
peak on both columns, and decachlorobiphenyl, the latest
eluting peak on both columns, do not display symptoms that
471
-------
would indicate sample transfer problems from the inlet to
the column.
Figure 3 contains the compound list with the conditions
under which the chromatograms in Figures 2,4-6 were
generated.
Figure 4 demonstrates the chromatography of the list in
Figure 3 on the DB-608, and Figure 5 shows the measured and
calculated resolution of the only partially co-eluting pair
on this column: alpha-Chlordane and Endosulfan I. Figure 6
demonstrates the chromatography for the compounds listed in
Figure 3 on the DB-1701, and Figure 7 shows the measured and
calculated resolution between the only partially coeluting
pair on this column: Endosulfan Sulfate and Methoxychlor.
The GC/FID instrumentation was used for two reasons: 1) it
was considered adequate to determine if the chromatography
of the dual column capillary system would or would not work,
and 2) a GC/dual BCD instrument was not available.
The above example chromatograms demonstrate acceptable
resolution with on-column amounts in the 2-5 nanogram range
by GC/FID. CLP Chlorinated Pesticide analysis, however,
uses GC/ECD and routinely analyzes for compounds in the
picogram range. Knowing that peak width is dependent
partially on the amount of compound present, it is expected
that the peaks should be narrower as there is at least 2
orders of magnitude difference in the amount on column.
This difference should result in analyte resolution better
than that demonstrated, especially for critical pairs.
If the system described above were to be used for routine
sample analysis, it would be highly recommend that a guard
column of adequate length (2-5 meters) be used in order to
protect the analytical columns from non-volatile sample
residues.
The dual column system using 0.32 mm I.D. columns in the CLP
Chlorinated Pesticide analysis, on the surface, appears to
be a viable alternative setup which would result in
increased sample throughput and greater profits. Also, more
samples could be run between quality control checks as their
analysis is required every 12 hours as opposed to every so
many samples. Detection limits should be more easily
achieved by using the smaller diameter columns as they
result in an increased signal to noise ratio of analytes.
The overall cost of the dual column setup would be less as
column price is directly proportional to column diameter.
The dual 0.32 mm I.D. setup could be used in either the
Megabore or splitless injection mode. In this particular
472
-------
CLP protocol, as with most EPA methodology, substitution of
columns is explicitly allowed. Section 2.2.1 says
"Equivalent columns may be employed if they meet the
requirements for resolution, initial calibration, and
calibration verification..-" If a capillary system such as
this is proven effective for pesticide analysis, PCB
analysis should follow easily as identification is
accomplished on a pattern recognition basis, and
quantitation should not change from whatever mode
laboratories are currently using.
Future experiments might include: single injection sample
splitting into three dissimiliar stationary phase 0.25 mm
I.D. columns; use of hydrogen as a carrier as opposed to
helium on the system described above; an investigation into
the significance of using the thinner film capillaries on
sensitive compound degredation at the 10-100 picogram level
using GC/ECD; and the investigation of other Megabore and
capillary dual column methods that could be adapted to this
dual column setup, or dual column setups with other types of
stationary phases. The possibility of using shorter
capillaries that would still generate the same number of
theoretical plates as a Megabore column could also be
investigated.
1. Jennings, W. Analytical Gas Chromatography; Academic
Press, San Diego, California, 1987; p 10.
2. Rood, D. A Practical Guide to the Care, Maintenance, and
Troubleshooting of Capillary Gas Chromatographic Systems;
Huthig Buch Verlag GmbH, Heidelberg, 1991; p 109-113.
3. Rood, D., Ibid.; p 108.
4. Jennings, W. Ibid.; p 111-151.
5. Rood, D., Ibid.; p 32.
473
-------
Figure 1
INSTRUMENT CONFIGURATION
INJECTION PORT
ECD
ECD
ANALYTICAL
COLUMNS
GUARD
COLUMN
0.5 - t METER
GLASS"Y"
CONNECTION
GAS CHROMATOGRAPH
-------
900C
Lie*.
5.60
1.0e
980Q
960Q
940Q
920Q
Figure 2
5.80 6.90
Time (min.)
22
.0
24.5 25.0
Time (min.)
6.20 6.40
25.5
26.0
475
-------
Figure 3
Column: DB-608
30m x 0.32 mm I.D., O.Sum
J&W Scientific P/N 123-1730
Column: DB-1701
30m x 0.32 mm I.D., 0.25 Mm
J&W Scientific P/N 123-0732
Carrier: Helium at 48 cm/sec
(measured at 150" C)
Oven: 110*Cfor0.5min
110-140*Cat20*/min
140-235*C at 9*/min
235* C for 4.5m in
235-280*C at 15*C/min
280* C forSmin
Injector: Megabore direct, 250*C, 2uL
Direct Flash Vaporization Liner
2-5 ng each component
Conector: 2 m x 0.53 mm I.D. deactivated guard column
0.53/0.32/0.32, 3-way "Y" union
Detector: FID, 280*C
Nitrogen make-up gas at 30 mL/min
1. Tetrachloro-m-Xylene
2 alpha-BHC
3. gamma-BHC
4. beta-BHC
5. Heptachlor
6. delta-BHC
7. Aldrin
8. Heptachlor Epoxide
9. gamma-Chlordane
10. alpha-Chlordane
11. Endosulfan I
12. p, p'-DDE
13. Dieldrin
14. Endrin
15. p,p'-DDD
16. Endosulfan II
17. p,p'-DDT
18. Endrin Aldehyde
19. Endosulfan Sulfate
20. Methoxychlor
21. Endrin Ketone
22. Decachlorobiphenyl
23. Unknown
476
-------
1.3e
.2e
a
1.0e
900(
Figure 4
DB-608: Chlorinated
Pesticides
20
I f
Jij
23
15 17
10
11
12
13 14
16
18
19
10 15
Time (min.)
20
25
-------
CO
9500
9400
9300
920O
9100
9000
890G
8800
87001
860
:
85004
14.20
Figure 5
77.3% RESOLUTION
14.40 14.60
Time (min.)
14.80
-------
1.4e
1.3e4
(O
1.0e
good
5
Figure 6
DB-1701: Chlorinated
Pesticides
22
UJLJW
15
I
11
10
2
14
20
16
17
21
23
10 15
Time (min.)
20
25
-------
9600
950
940
9300
920O
9100
9000
16.00
Figure 7
77.3% RESOLUTION
16.20
Time (min.)
16.40
16.60
-------
53 PERFORMANCE DATA FOR THE ANALYSIS OF PHENOLS, NITROAROMATICS,
CYCLIC KETONES, HALOETHERS AND CHLORINATED HYDROCARBONS
Siu-Fai Tsancr. Nellie Chau, and Paul Marsden, SAIC, San Diego, CA
92121 and Barry Lesnik, Organics Methods Program Manager, OSW/EPA
22403
The U.S. EPA, Office of Solid Waste (OSW) is responsible for
providing reliable, robust analytical methods with documented
performance. This mission requires that the OSW regularly evaluate
developments in analytical technologies that could improve the
measurement of chemicals regulated under the Resource Conservation
and Recovery Act (RCRA). Whenever modifications to existing
methods are proposed by research scientists, government
laboratories, commercial laboratories, or instrument manufacturers,
the Methods Section of the OSW compares them with existing SW-846
methods. If a technique is promising, single- or multi-laboratory
method studies are conducted to document method performance.
This report provides performance data for the analysis of
phenols [Method 8041], nitroaromatics and cyclic ketones [Method
8091], haloethers [Method 8111], and chlorinated hydrocarbons
[Method 8121A] using capillary gas chromatography (GC). Target
analytes were extracted from spiked soil using Method 3540
(Soxhlet) and analyzed with 0.53 mm id a DB-5 capillary column.
Phenols were analyzed with and without derivatization using
diazomethane. Initial conditions for some chromatographic
separations were derived from reports presented by the
Environmental Monitoring Systems Laboratory at Las Vegas (EMSL-LV)
during the 1990 and 1991 Symposia.
481
-------
RAPID CONFIRMATION OF NITROAROMATICS AND NITRAMINES
USING UV DIODE ARRAY SPECTRAL COMPARISON
Bradley A. Weichert, Manager, GC/HPLC Department, Robert D.
Baker, Supervisor, HPLC Group, Dr. Curtis R. Campbell, Sr.
Staff Scientist, HPLC Group.
ABSTRACT
The present proposed SW846 method for the determination
of nitroaromatics and nitramines, designated EPA 8330,
requires a second column confirmation for any positive
identification from the primary analysis. Since this confir-
mation is sequential to the primary analysis it results in
additional turnaround time for an already complex determina-
tion. By comparing spectral data collected using a diode
array detector the confirmation can be accomplished with a
single analytical run.
Spectra from the UV range of 200-400 nm are collected
for each peak within the retention windows determined by a
standard curve run prior to sample analysis. These spectra
are stored electronically for comparison to the standard
curve spectra in the method library. Peaks of interest may
also be checked for peak purity by comparing spectra from
the leading edge, apex, and downslope of each peak. Criteria
for confirmation are established by comparing spectral
integrity over the range of the standards run for calibra-
tion.
This technique presents a three-dimensional approach to
analyte determination and reduces the total time of analy-
sis. Some of the remaining problems associated with a second
column confirmation, such as nonresolution of all the
analytes and interferences which do not appear in the
primary analysis, can also be eliminated or reduced.
482
-------
65
Tom Williq, Chemist and Jon S. Kauffman, Ph.D., Group Leader GC/MS
Semivolatiles, Lancaster Laboratories, Inc., 2425 New Holland Pike,
Lancaster, PA 17601-5994
ABSTRACT
Gel Permeation Chromatography (GPC) is a size exclusion type of
chromatography which efficiently removes high boiling materials that may
interfere with the analysis of target compounds in semivolatile extracts.
Currently, the use of GPC to clean up environmental soil samples is
required under the USEPA Contract Laboratory Program (CLP) Statement of
Work. The purpose of this study was to evaluate the effects of GPC clean
up on GC/MS performance in order to justify the cleanup of all samples with
difficult matrices by GPC prior to analysis. Potting soil was chosen as
the subject matrix due to the high level of organic interferences found in
this matrix. A sample was split into two portions both of which were
extracted by USEPA SW-846 Method 3550. One extract was then cleaned up
using GPC. Each extract was then injected four times on a Hewlett-Packard
GC/MS. In order to monitor the performance of the GC/MS system, a standard
containing all of the semivolatile Priority Pollutant List compounds at a
concentration of 50 ppm was injected before and after each sample. The
degradation of the system was monitored by measuring the decrease in
response factors against the number of injections of the potting soil
extracts. Results were compared for the GPC extract versus the non-GPC
extract. The response factors dropped drastically for five acid compounds:
2,4-dinitrophenol, 4-nitrophenol, pentachlorophenol, 2,4,6-trichlorophenol
and 4,6-dinitro-2-methylphenol after only one or two injections of the non-
GPC extract. However, the response factors for these compounds were much
more stable even after four injections of the GPC extract. A similar trend
was observed for the internal standards, which are used to monitor system
performance. This loss of system sensitivity after injecting the non-GPC
extract can be explained by an increase of active sites in the injector
port caused by the buildup of high molecular weight compounds in the
injector liner and on the stainless steel seal. Although the experiment is
very simple, the results make the utility of GPC cleanup evident and
therefore justify the extra time and effort spent in adding this step to
semivolatile sample preparation.
483
-------
66
EVALUATION OF THE ENSYS PAH-RISc TEST KIT
R. Paul Swift. Ph.D.. Senior Chemist, John R. Leavell, Chemist, Chris W.
Brandenburg, Chemist, ESAT Region 10, ICF Kaiser Engineers, Inc., 7411
Beach Drive East, Port Orchard, Washington, 98366
ABSTRACT
A polynuclear aromatic hydrocarbon field screening method, utilizing
enzyme-linked immunosorbent assay techniques, was recently evaluated. The
validation study was conducted to evaluate the performance and assess the
utility of this method for use in field screening activities at hazardous
waste sites. Design criteria included analysis of field samples, analysis
of well-characterized reference materials, test kit response to various
soil types (including analyte-spiked soils), and a general performance
evaluation using selected soil types with varying PAH concentrations.
INTRODUCTION
The authors recently conducted a validation study for an Enzyme-Linked
Immunosorbent Assay (ELISA) field-screening test method, PAH-RISc*. This
semiquantitative chromogenic analytical method, developed by Ensys, Inc.,
Research Triangle Park, NC, was evaluated for use in the determination of
Polynuclear Aromatic Hydrocarbons (PAHs) in environmental soil and
sediment samples. The experimental design and interpretation of results
obtained are described below.
BACKGROUND
Field Screening
Field screening for contaminants has applications in a wide variety of
situations (1), from collecting real-time data relating to worker safety
to monitoring plume boundaries resulting from materials spills. Low cost,
rapid-turnaround sample analysis is beneficial to site characterization
and assessment, as well. Sampling locations of possible interest and
local "hot-spots" may be identified quickly, aiding in the selection of
samples to be collected for subsequent CLP analysis. Other considerations
such as remediation efforts and effluent compliance may also be monitored.
The need for rapid, reliable, semiquantitative methods for measuring
environmental contaminants has resulted in the introduction and develop-
ment of new technologies. In addition, laboratory-based analytical
methods such as GC/MS and HPLC, among others, are finding increased use in
field applications (2).
One novel field screening method gaining regulatory acceptance utilizes
immunoassay-based testing (3). Immunoassay techniques have been used for
more than 15 years in medical and clinical settings. This technology has
received Agency approval for a few classes of analytes (4), and is
currently being applied to the detection of PCBs, petroleum hydrocarbons,
and pentachlorophenol. One of the earlier environmental applications in
which this technology was used was in the detection of pesticides and
pesticide residues (5).
Enzyme-Linked Immunosorbent Assays
Immunochemical analysis is a well-established clinical diagnostic
technique which is gaining wider acceptance in other areas. In the field
of environmental analysis, antibodies have been or are being developed for
484
-------
a number of contaminants. All these antibodies are of animal origin, but
advances in monoclonal preparation have yielded in-vitro propagation
methods.
In order to obtain antibodies of the desired specificity, an immunological
response must first be generated in the host. Typically, the target
analyte, or hapten, is derivatized to produce a "handle" for use in
attachment to a macromolecule. Historically, proteins such as bovine
albumin have been used as the carrier. This step is essential because
direct injection of the analyte, due to its small mass (on a biological
scale), would result in detoxification by the animal's liver. In order to
be recognized as an antigen, or immunogen, by the immune system, a foreign
substance must have a mass which is large on a molecular scale (>10,000
Da). When injected into the host animal these immunogens cause the
animal's immune system to generate antibodies in response to the foreign
substance. The hapten is then bound to an enzyme, such as horseradish
peroxidase, to form an enzyme-conjugate. The conjugated compound is used
as the chromogenic reagent in this test.
After the antibodies are extracted from the host, they are sorted based on
sensitivity and specificity to the hapten. In the polyclonal method, the
above steps are repeated until a desired quantity of antibody is obtained.
The test kit used in this study utilizes a monoclonal preparation of the
desired antibodies using hybridoma technology, in which antibody-producing
(spleen) cells are fused with myeloma cells. The resulting progeny cells
are able to produce relatively large amounts of the specified antibody.
In a typical test kit, antibodies of the desired specificity are then
immobilized on a solid substrate such as a small test tube. A predeter-
mined amount of sample, possibly containing analyte, is extracted with a
suitable solvent. An aliquot of extract and a fixed amount of enzyme-
conjugate are added to the antibody tube and allowed to incubate for a
given period of time. Competition between the analyte and enzyme-
conjugate for a limited number of antibody binding sites results in
binding of the haptens in proportion to their relative concentrations.
Unreacted haptens are washed from the tube and two color development
reagents added. These reagents react with the bound enzyme-conjugate,
producing a depth of color proportional to the amount of enzyme-conjugate
bound. The "intensity" of color, compared photometrically to a calibra-
tion solution containing analyte, is inversely proportional to the amount
of analyte present in the original sample.
This test method utilizes ten-fold serial dilutions of the solvent extract
for comparison with a 1 ppm calibration standard prepared in the same
manner as the samples. A comparative photometer measures the difference
in absorbance between the sample and the standard, and uses the difference
measurement rather than the parametric value for quantitation.
EXPERIMENTAL DESIGN
Four main design parameters were selected for this validation study,
including analyses of 30 samples for comparison with CLP-generated data
analysis of four "worst-case" Superfund-class reference samples, false-
positive and then false-negative (1 ppm spike level) reaction to soil
types, and a general reliability/performance test using a given soil type
spiked with varying amounts of analyte. Given a fixed number of test
kits, the key design criterium was to conduct a comprehensive yet
conclusive set of experiments.
485
-------
Polynuclear aromatic hydrocarbons are a class of compounds which have one
or more aromatic rings, generally in conjugation. The majority of common
PAHs contain three, four or five rings, may have alkyl- or aryl- groups
bound to them, and the rings themselves may be heterocyclic. One
important three-ringed PAH is phenanthrene, because it has a physical
structure which resembles the skeleton of many other PAH compounds.
Table I PAH RISc" Soil Test Sensitivity to PAH Compounds
Number of Rings
2 rings
3 rings
4 rings
5 rings
6 rings
PAH Compound
Naphthalene
Acenaphthene
Acenaphthylene
Phenanthrene
Anthracene
Fluorene
Benzo[a]anthracene
Chrysene
Fluoranthene
Pyrene
Benzo[jb]fluoranthene
Benzo[£]fluoranthene
Benzo|a]pyrene
Dibenzo[a,/)]anthracene
lndeno[/,2,3-c]pyrene
Benzo[flr,/7,/]perylene
Concentration Necessary to Result in
Positive Test (ppm)*
200
8.1
7.5
1.0
0.81
1.5
1.6
1.2
1.4
3.5
4.6
9.4
8.3
>200
11
>200
* Samples with stated concentration will give positive result greater than 95% of the time when
tested at stated concentration level.
The antibody used by this test kit was developed to target, or key,
phenanthrene. According to the manufacturer, the kit is designed such
that concentration of phenanthrene necessary to result in a positive test
is 1 ppm (6). With the exception of anthracene, which is detectable at
0.81 ppm, the test is less sensitive to other PAHs. The concentrations
necessary to produce a positive result range from 1.2 ppm for chrysene to
200 ppm for naphthalene. A list of several PAH compounds (SW-846 Method
8310 Target List) and their respective detection limits has been provided
by the vendor (Table I). The reported sensitivities have been used as
486
-------
scaling factors in some sections of this report. For example, to produce
an equivalent response using naphthalene rather than phenanthrene as the
target analyte, 200-times more naphthalene (on a mass basis) is required.
This relationship will be further investigated in a later section.
Phase I: Analysis of Field Samples
Several issues were raised during the initial design of the validation
study, the most obvious of these being the evaluation of the test kit with
respect to the manufacturer's claim:
This method correctly identifies 95% of samples that are PAH-
free and those containing I ppm or 10 ppm of PAHs. A sample
that develops less color than the standard is interpreted as
positive. It contains PAHs. A sample that develops more
color than the standard is interpreted as negative. It
contains less than 1 ppm or 10 ppm PAHs.
To substantiate this claim, 30 samples were selected from the Manchester
Environmental Laboratory (MEL) soils storage facility. These samples were
analyzed previously using either GC/MS (tentatively identified compounds
excluded but noted) or HPLC methodology; total PAH values (16 analytes)
ranged from zero (non-detect) to >182 ppm. Soil types ranged from
weathered sand to dark loamy humus.
Phase II: Analysis of Reference Samples
Four well-characterized reference samples were selected for use in this
study. These samples have been independently analyzed by a number of
laboratories and are representative of materials present at various
preremediated Superfund sites. The matrices range from marine sediment to
composited hazardous waste. All samples contain a variety of contami-
nants, including pesticides, PCBs, and in one case, tributyl tin.
Clearly, potential interferents are present in these samples. A scheme in
which three analysts conduct the analyses, with one analyst performing the
tests in duplicate, provides an empirical estimate of both precision and
accuracy.
Phase III: Reaction to Soil Types
Soil compositions vary greatly on a regional basis. Extractable humic
components, among many others, have the potential to negatively affect the
performance of the test kit. Analysis of soils of various compositions
would provide information necessary to estimate the effect of soil type on
test kit reliability. To that end, 11 samples collected from different
soil horizons throughout Washington state by the US Geological Service
were analyzed. The PAH content of each of these library soils has been
previously determined to be below the detection limit of the test kit.
Therefore, any positive test result(s) could be attributed to what have
been termed "relatively unremarkable" contaminants, which possess
structural features similar to those of phenanthrene (7).
PAH compounds adsorbed onto soil surfaces may resist methanol extraction
and the degree of adsorption may be dependent on the nature of the soil
In order to investigate this effect, a subset of the library soils was
selected and the soils fortified, or spiked, with 1 ppm phenanthrene
This quality control measure was used to qualitatively estimate the
distribution of phenanthrene between the soil matrix and the methanol
extract.
487
-------
Phase IV: Reliability/Performance Evaluation
A set of samples consisting of weathered, sandy soil was composited for
use in this phase of the validation study. Each of the samples were
previously confirmed to be PAH-free. Eight spiking levels were used,
ranging from 0.1 to 10.0 ppm normanlized to the phenanthrene response.
For the spike, a three-component mixture containing a three, four and five
ringed PAH - phenanthrene, pyrene and benzo[.fc]fluoranthene - was
prepared. The relative amount of each component was scaled based on the
sensitivity of the test kit to the individual components.
The objective of this set of measurements was to provide an estimate of
error with respect to false positive results below the detection limit as
well as false negative results for samples with PAH concentrations above
the detection limit.
It was also hoped that these measurements would provide insight relating
to test kit response for mixtures of PAHs. Because the assay utilizes
competitive binding of the antibody-antigen for detection of the analyte,
we were interested in determining whether the presence of one PAH would
affect the sensitivity of the test to another PAH. Rather than having
only one PAH competing with the enzyme conjugate for antibody binding
sites, there could be two or more competitors. There is no a priori
reason to believe that the test response may be calculated using a simple
weighted average concentration of the individual PAHs and the stated test
sensitivities to each.
RESULTS AND DISCUSSION
This experiment was conducted under relatively ideal conditions. Three
experienced chemists performed several trial analyses in order to gain
familiarity with the kit. Work was performed at the laboratory benchtop
in a controlled environment (temperature, humidity, etc.). Quality
control/quality assurance, as specified through personal communications
and training by the manufacturer, included analysis of method blanks,
regular pipet delivery calibration verification, and rejection of
analytical sequences in which the relative absorbance of replicate
calibration standards varied by more than 0.2 absorbance units. The
analytical instructions supplied with the kit and QA/QC guidelines
provided by the manufacturer were followed without exception.
The laboratory results reported for the GC/MS and HPLC data were generated
in accordance with full QA/QC requirements following USEPA CLP and SW-846
protocols. All data reported have been validated. To maintain client
confidentiality, any unnecessary site-specific references have been
omitted.
One set of measurements was made to evaluate the response of the
comparative photometer. A series of five spiked-blank solutions were
prepared in duplicate, with concentrations of 0.5, 0.75, 1.0, 1.25 and 1.5
ppm phenanthrene. Within each set, the 1.0 ppm was treated as the
calibration standard. Aliquots of these solutions were withdrawn and
processed as if they were actual soil extract. The absorbance of each
solution was measured and recorded. The resulting "calibration" curve was
determined through a linear least-squares fit of absorbance vs. concentra-
tion. The results are illustrated in Figure I.
From the data, it appears as though the instrument response is linear
around the zero (difference) absorbance. However, the slope is relatively
small - therefore variations in absorbance affect the apparent sample
488
-------
Figure 1 Instrument Response
concentrations.
cautiously.
Instrument readings near zero must be interpreted
Phase I: Thirty soil samples were selected for ELISA-PAH analysis. No
attempt was made to target the one or ten ppm concentration levels; this
portion of the study was designed to emulate characterization of samples
in the field. These samples were previously extracted (EPA Method 3540)
and characterized for base/neutral/acid extractable compounds (BNAs,
including PAHs; EPA Method 625), volatile organic analytes (VOAs; EPA
Method 624), metals (ICP-AES, EPA-CLP Method 200.8), and pest-
icides/polychlorinated biphenyls (pest/PCBs; GC/ECD, EPA Method 8080). In
addition to quantifying total PAH content and relative amounts of the
individual compounds, the supporting analyses were used to characterize
the soil matrices and identify potential interferences. The results are
shown in Tables II and III.
In Table II, the total PAH concentration for each sample is compared with
the ELISA results. False-positive and false-negative results are
indicated by "+" or "-", respectively. Results from GC/MS analysis of
sample 2361 indicated the presence methyl-phenanthrenes at 3 ppm.
Although the associated data were qualified "NJ" to indicate "there is
evidence that the analyte is present, the associated numerical result is
an estimate", the levels apparently present would be sufficient to yield
positive identification at 1 ppm. This result slightly increases the
accuracy rates (indicated in parenthesis). The estimated accuracy for the
ELISA test, based analysis of the thirty samples, was 86.7% (90%) at 1 ppm
and 73.3% at 10 ppm. The frequency of false-positive results at 1 ppm was
13.3% (10%), with no false-negatives. The frequency of false- positive
results at 10 ppm was 20%, and false-negatives, 6.7%. One sample,
selected randomly, was analyzed in duplicate.
489
-------
Table II Analysis of Field Samples / Total PAH Content
Sample ID
4005
4006
4007
4010
4659
4300
4301
4302
4303
4640
2107
2358
2358D
2359
2360
2361
2362
2363
2364
2365
2366
2368
2369
2370
2371
2372
2373
2374
1 ppm Test
<1
ft
*
ft
*
*
*
ft
•ft
>1
*
ft
ft
ft
ft
ft
ft
*
ft
ft
10 ppm test
<10
*
*
#
»
*
«
*
*
«
>10
ft
*
ft
ft
ft
ft
ft
ft
ft
-
«
»
#
*
*
Lab Result (ppm)
0.2
12.2
16.0
0.0
0.5
8.7
147.7
182.3
4.4
0.2
0.0
85.4
85.4
28.5
0.3
0.6
0.0
1.8
3.4
6.7
0.9
43.2
72.8
1.3
0.3
0.4
27.9
0.0
False +/-
Evaluation
@ 1 ppm
+
+
+ '
+
Evaluation
@ 1 0 ppm
+
+
+
+
+
+
490
-------
Table II Analysis of Field Samples / Total PAH Content
Sample ID
2375
2376
2377
1 ppm Test
<1
*
>1
•
10 ppm test
<10
•
•
•
>10
Lab Result (ppm)
16.4
0.4
9.5
False
Evaluation
@ 1 ppm
+ /-
Evaluation
@ 1 0 ppm
' Tentatively identified compounds list indicates methyl-phenanthrenes present at 3 ppm
Table III Analysis of Field Samples / Normalized PAH Content
Sample ID
4005
4006
4O07
4010
4659
4300
4301
4302
4303
4640
2107
2358
2358D
2359
2360
2361
2362
2363
2364
2365
1 ppm Test
<1
•
•
•
•
•
>1
•
•
•
•
•
•
•
•
10 ppm test
<10
•
•
•
*
>10
•
•
•
•
•
•
•
•
•
•
•
•
Normalized Lab
Result (ppm)
0.1
8.1
9.0
0.0
0.2
5.2
56.9
73.2
0.1
0.0
0.0
47.3
47.3
11.5
0.2
0.5
0.0
1.2
1.7
3.6
False +/-
Evaluation
@ 1 ppm
+
+
+
+ '
Evaluation
@ 10 ppm
+
+
+
+
+
+
+
491
-------
Table III Analysis of Field Samples / Normalized PAH Content
Sample ID
2366
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
1 ppm Test
<1
*
*
*
*
>1
*
*
*
1 0 ppm test
<10
«
*
*
«
«
*
«
»
>10
*
•
tt
Normalized Lab
Result (ppm)
0.6
27.5
49.2
0.8
0.1
0.2
13.5
0.0
6.4
0.2
2.8
False +/-
Evaluation
@ 1 ppm
+
+
Evaluation
@ 10 ppm
+
-
' Tentatively identified compounds list indicates methyl-phenanthrenes present at 3 ppm
Table III shows the comparison of normalized PAH concentrations with the
ELISA results. As mentioned previously, the sensitivity of the test
varies with different PAH compounds. In order to account for this effect,
each PAH concentration was normalized by dividing its actual value by that
concentration required to produce a positive result at 1 ppm (phenanthrene
= 1). This was done to minimize any bias introduced into the test by
anti-PAH specificity. A critical assumption is that the manufacturer-
provided list of PAH sensitivities are valid in the presence of two or
more PAH compounds. Sensitivity to the methyl-phenanthrenes tentatively
identified in sample 2361 is not known, but was assumed to be similar to
that of phenanthrene. As in the previous data set, estimated accuracies
including any contribution from the methyl-phenanthrenes are enclosed in
parenthesis. The estimated accuracy was 80% (83.3%) at 1 ppm and 70% at
10 ppm. The frequency of false-positive results at 1 ppm was 20% (16.7%),
with no false-negatives. At 10 ppm, the frequency of false-positive
results was 26.6%, and false-negatives, 3.3%.
The results using total and normalized PAH concentration values are
similar. If samples with PAH concentrations exclusively near 1 and 10 ppm
had been used, a greater discrepancy would have been observed. The
majority of these samples contained levels of PAHs separated from either
test kit detection level. This demonstrates the need to apply the
screening test judiciously, with operator intervention/interpretation
applied to those samples near either limit.
Phase II: Four soil reference samples were analyzed by three different
analysts, with one analysis performed in duplicate ("D" in Table IV), to
illustrate the variation introduced by different analysts and also the
duplicate precision. Comparison of the data is shown in Table IV. There
was considerable variation in the results, apparently due to the presence
492
-------
of interferents, including compounds to which the ELISA test exhibits
cross-reactivity. The summary in Table IV lists the total and normalized
PAH concentrations.
These samples were selected because they present a real challenge to the
ELISA test. They are each representative of the types of materials found
at Superfund sites. Prior to analysis, the samples were sieved and
homogenized. Non-representative material was removed, but an effort was
made to preserve the physical character of the original sample. A brief
summary of each soil follows.
Sequim Bay: This matrix is typical of marine or lower-wetlands sediment,
with corresponding levels of marine salts. The consistency was
similar to water-saturated silt. Although the concentration of
various contaminants were each less than 1 ppm, a large number of
different compounds were present. This particular reference sample
has been independently characterized as many as 50 different times.
Some of the compounds present were:
Phenol - 0.24 ppm; 4-methyl phenol - 0.24 ppm; pentachlorophenol - 0.42
ppm; biphenyl - 0.79 ppm; methylnaphthalenes - 0.25 ppm; tributyl tin
chloride - 0.11 ppm; halophenyl phenyl ethers - 0.4 ppm; tetrachloroguicol
- 0.47 ppm
Table IV Analysis of Reference Samples
Sample ID
Sequim Bay
- marine sediment
- low levels of many
pesticides, chloro-
benzenes, phthalates
H.I. BSRM #XX
Composite soil
PCBs: 130 ppm
Lead: 5000 ppm
- contains slag/ash
RTC Sample #XXX
- PCP: 2190 ppm
- substituted
naphthalenes:
2300 ppm
Soils Bldg 0003
Isophorone: 4.4ppm
(dimethyl hexanone)
Trial
1
2
3
3D
1
2
3
3D
1
2
3
3D
1
2
3
3D
1 ppm test
<1
>1
•
•
•
I
•
•
•
•
•
•
•
•
10 ppm test
<10
•
•
•
I
•
•
>10
*
•
•
•
•
"
*
•
•
•
100 ppm test
<100
•
•
•
*
•
«
•
>100
•
*
«
•
•
•
tt
•
•
GC/MS Results
Lab Result
(ppm)
1.73
1.73
1.73
1.73
18.3
18.3
18.3
18.3
8611
8611
8611
8611
24.8
24.8
24.8
24.8
Normalized Lab
Result (ppm)
<0.1
<0.1
<0.1
<0.1
12.1
12.1
12.1
12.1
5764
5764
5764
5764
12.4
12.4
12.4
12.4
493
-------
As indicated in Table IV, the actual PAH concentration was less than
2 ppm. When normalized based on specificity to individual PAH
compounds, the apparent PAH concentration was less than 0.1 ppm.
Three of four ELISA analyses indicated that the PAH concentration
was greater than 10 ppm, and two tests yielded results greater than
100 ppm. Clearly, materials were present which respond to the test
in a similar manner as do PAHs, and exhibit a strong interference.
Unfortunately, no clear pattern was apparent in the spurious
results. The test did appear to yield either correct or false-
positive results at the 1 ppm level, depending on which set of lab
results were considered.
HI BSRM #XX: This bulk site reference material composite sample was
manufactured from soils collected at a Superfund site by the USEPA
in Region 10 and known to contain PAHs and PCBs. The matrix
consisted largely of marine sediments and silt, although there is a
significant amount of ash, slag and related fallout originating from
the operation of a secondary smelter for many years. Other man-made
debris and residue include cinders, brick material and sandblasting
waste from ship refinishing. Local "hot-spots" resulting from wood
and pole treating and indiscriminate dumping of waste oils and
electrical transformer/capacitor dielectric contributed PCBs and
PAHs to the BSRM.
The reference sample exhibited a fairly high degree of matrix
inhomogeneity on both the macroscopic and microscopic level. A
concerted effort was made to homogenize the sample prior to
analysis. The variability is reflected, in part, by the standard
deviations associated with the analytical results:
Lead - 5480 ± 2620 ppm; Arsenic - 37.8 ± 7.3 ppm; Total PAHs (7 carcino-
genic PAHs identified in MTCA) - 18.3 ± 4.6 ppm; Total PCBs - 127 ± 39
ppm.
A comparison of the sample results is shown in Table IV. All ELISA
analyses at the 1 ppm level correlate with the laboratory data. The
first two trials produced identical results at 10 and 100 ppm (true-
negatives), but disagree with the duplicate analyses performed by
the third analyst. The duplicate analyses were self-consistent.
Because the duplicate analyses were performed using the same soil-
extract, sample inhomogeneity rather than ELISA precision may have
contributed to the disparity.
RTC Sample #XXX: This sample was prepared under contract to the EPA as
part of a RCRA study and made available for use in this study.
Laboratory analyses were performed for both total metals and BNA
compounds. The levels of sodium, potassium, nickel and total PAHs
were relatively high. Some of the analytes present at significant
levels were:
Chromium - 16100 ppm; Phenanthrene - 2430 ppm; fluoranthene - 1840 ppm;
pentachlorophenol - 2190 ppm; methyl-naphthalenes - 200 ppm; carbazole -
81 ppm;
Of the reference samples, this was the most highly contaminated.
All ELISA results indicated PAHs present at 1 ppm. One analysis
yielded false-negative results at 10 and 100 ppm. This analysis is
clearly in error; with the levels of contaminants known to be
present, it seems unlikely that normal experimental or method
uncertainties were the cause. Since the 10 and 100 ppm results were
494
-------
generated using serial dilution of the 1 ppm extract, we believe
their inaccuracy may be attributed to dilution error.
Soils Building 003: This sample was also a composite soil manufactured
from materials which had collected in a soils laboratory over a
period of time. A large number of assorted samples of varying
origin were blended, with aliquots withdrawn for waste disposal
characterization. Some of the material was dated and all had been
stored at room temperature. Microbial and oxidative decomposition
were expected to have yielded a variety of degradation products of
the PAHs and other contaminants. Subsequent analysis showed that
with the exception of 4.4 ppm isophorone (dimethyl hexanone), the
composition of contaminants was due predominately to PAHs, with
total and normalized concentrations of 24.8 and 12.4 ppm, respec-
tively. The matrix of the composite soil sample included clay,
silt, sand and loam.
All ELISA results indicated PAHs present at the 1 ppm level. Two
replicates yielded false-negative results at 10 ppm. Duplicate
analysis yielded correct results at 10 ppm, but false positive
results at 100 ppm. No simple explanation can be offered for this
disparity.
Phase III: Eleven soil samples collected from various soil horizons across
Washington state were analyzed using the ELISA method. The soils were
previously determined to contain undetectable amounts of PAH compounds.
The purpose of this phase of the experiment was to determine the effect of
soil type with respect to false-positive ELISA results.
Analysis of these soils yielded negative results in each case. On this
basis, soil type did not appear to negatively impact the accuracy of the
test with respect to false-positive results.
Four soils from this set were spiked with PAH to provide an estimate of
false-negative results with respect to soil type. These soils were
primarily clay/silt with a fair amount of humic or loamy material. A 10-
gram aliquot of each soil was spiked directly with phenanthrene to give a
final concentration of 1 ppm, allowed to weather at room temperature for
24 hours, extracted, and analyzed in duplicate. This method was preferred
to that in which the extract is spiked because adsorption of phenanthrene
on the soil surface is a potential physical interference. Resistance to
methanol-extraction would result in a negative bias and, therefore, false-
negative results. Phenanthrene was selected for use as the spike on the
basis of ELISA sensitivity, and to reduce any possible cross-reactivity
questions which may arise from the use of a mixed-PAH spike. The results
are shown in Table V.
Results from the spiking study show a high frequency (75%) of false-
negative results, indicative of the inability of the methanol extraction
fluid to completely liberate the phenanthrene from the matrix. Although
it is reasonable to expect that the methanol-based extraction is not 100%
efficient, these results yield an unacceptable percentage of false-
negative results. The bias built into the ELISA test did not appear to be
able to provide a large enough margin of uncertainty in this case.
Because a CLP-quality extraction and analysis was not performed, it was
not possible to estimate the actual efficiency of the extraction usinq the
ELISA method.
These limitations should be compared with those generally encountered
during laboratory-based analysis. Standard CLP soxhlet extraction
495
-------
involves refluxing the solid sample for a minimum of 18 hours in boiling
solvent. Even then, extraction efficiency may only approach 80% or so.
Matrix spike analysis generally involves adding the spike compound to the
extraction solvent rather than to the soil itself. Obviously, no
weathering of the PAH/soil occurs. However, this does not change the fact
that the kit was unable to detect PAH at 1 ppm in soil.
Table V Reaction To Soil Types
Sample ID
8114
8123
8129
8136
8142
8500
8500 S
8501
8501 S
8506
8512
8512 S
8513
8513 S
8107
1 ppm test
<1 ppm
*
*
*
*
*
*
*
*
«
*
*
«
*
*
>1 ppm
#
10 ppm test
<10 ppm
•
•
*
*
«
*
*
«
*
*
*
«
*
«
w
>10 ppm
Spike result
phenanthrene
1 ppm
n/a
n/a
n/a
n/a
n/a
n/a
-
n/a
-
n/a
n/a
n/a
-
n/a
Soil description
yellow brn sandy loam
pale brn silty loam
brn silty loam
pale brn sand
pale brn sand
drk brn silt loam
-
drk brn gry sandy loam
»
lake sediments
It brn coarse loam sand
••
drk brn stony silt loam
»
It grey silty clay
Two papers previously published by the manufacturer described spiking
studies performed during validation of their PCB and pentachlorophenol
immunoassay methods (8-9). In each case the extracts, rather than the
native soils, were spiked. A similar study reported poor extraction
efficiency for PCBs spiked directly onto the soil (10). We felt it
important and more procedurally valid to spike and then weather the soil.
Phase IV: Another spiking study was performed to estimate reliability of
the ELISA method as a function of analyte concentration. Sandy, weathered
soil previously determined to be PAH-free was divided into 57 - 10 gram
aliquots, spiked directly, and allowed to interact for one hour.
A spiking solution was prepared using a three, four and five ring PAH
compound - phenanthrene (3), pyrene (4) and benzofk] fluoranthene (5) -
such that their relative, normalized contributions on a mass basis to the
spike cocktail were equivalent. Using the 1 ppm spike as an example, a
solution was prepared such that 0.33 ppm of phenanthrene would be added to
the soil. The same contributions were desired from the four and five-
496
-------
ringed compounds, but the test is less sensitive to them. Since 3.5
of pyrene is required to produce the same response as 1 ppm phenanthrene,
3.5 times more pyrene was used (3.5x0.33). Similarly, the test is 9.4
times less sensitive to the benzo-compound, so 9.4x0.33 ppm benzo[k]fluor-
anthene was used. Calculation shows that the total PAH concentration in
the spiked soil was actually 4.6 ppm. A similar scheme was used in the
preparation of the soils at the other spiking levels. With the exception
of the blank sample, which was analyzed once in duplicate, seven
replicates were analyzed in duplicate at each of 0.1, 0.5, 0.8, 1.0, 1.5,
2.0, 5.0, and 10.0 ppm (normalized) concentrations. Seven extractions
were performed at each level, and each extract analyzed in duplicate. The
results for normalized and total PAH concentrations are shown in Tables
VI(a) and VI(b), respectively; agreement of estimated error rates for the
different spike levels is coincidental.
Table Vl(a) Reliability Test Normalized Concentrations
Normalized Value
(ppm)
Estimated Rate of
False Positives (%)
Estimated Rate of
False Negatives (%)
0.0
0,0
0.1
0.0
0.5
78.5
0.8
78.5
1.0
21.5
1.5
0.0
2.0
0.0
5.0
0.0
10.0
0.0
Table Vl(b) Reliability Test Actual Concentrations
True Value (ppm)
Estimated Rate of
False Positives (%)
Estimated Rate of
False Negatives (%)
0.0
0.0
0.46
0.0
2.32
21.5
3.71
21.5
4.63
21.5
6.9
0.0
9.3
0.0
23.2
0.0
46.3
0.0
Three false-positive results were observed at both the 0.5 and 0.8 ppm
levels. The 0.5 ppm trial was repeated (extraction and analyses), and the
results were identical to those obtained in the first trial. Three false-
negative results were observed at the 1 ppm level. No false-negative
results were seen above 1 ppm.
The total PAH results showed a high rate of false-negatives. This data
would seem to indicate that normalizing the PAH concentration based on
sensitivity of the anti-PAH to the various PAH compounds is somewhat
valid. Also, the discrepancy between the spike recoveries in Phase V and
Phase VI is attributed to the difference in soil matrices. More PAH was
extracted from the sandy soil than from the loamy soil. This further
supports the conclusion that adsorption may introduce a low bias into the
test results.
CONCLUSION
It was the intent of this validation study to investigate and evaluate the
performance of the Ensys PAH-RISc* test kit. The goal was to conduct a
comprehensive set of experiments while working under the constraint of a
predetermined number of available test kits.
497
-------
This immunoassay test method performed favorably, although not quite as
well, as claimed by the manufacturer. It does appear to have the
potential to be used as a field screening tool, provided that the
limitations of its use be borne in mind.
The primary caveat is that this kit should be used by personnel capable of
proper interpretion of the results from a scientific standpoint. A
negative result at 1 ppm does not necessarily indicate that less than 1
ppm of PAH contamination exists. Relatively high concentrations of
particularly carcinogenic PAHs, such as benzo[a]pyrene, may be present on-
site at levels which are well above 1 ppm but below the levels necessary
to generate a positive test response. The sensitivity of the test kit to
the various compounds must be borne in mind when interpreting the results.
PAH-extractability may introduce low-bias at 1 ppm; the implications of
this must be considered when concentrations of PAHs are near the 1 ppm
level.
Also, proper soil sampling is a scientific technique. In order to obtain
accurate, representative results using this test kit, the operator should
follow appropriate soil collection methods.
Any samples which generate positive results, as well as a percentage of
those which test negative, should be submitted for subsequent confirma-
tional analysis by classical analytical methodology.
ACKNOWLEDGEMENTS
The authors would like to thank the U.S. EPA for supporting the study
under the Environmental Service Assistance Teams Zone 2 Contract 68D10135.
We would also like to thank Ensys Inc for providing the test kits used in
this study, and for the training rendered in their proper use. The USGS
soils and their characterization data were provided by Mr. Kenneth Ames,
USGS, and Mr. Dickey Huntamer, Washington State Department of Ecology.
Mr. Gerald Muth, USEPA, provided initial technical direction on the
experimental design.
DISCLAIMER
Although the research described in this manuscript has been funded wholly
or in part by the EPA Contract 68D10135 to ICF Technology Inc., it has not
been subjected to the Agency's review and therefore does not necessarily
reflect the views of the Agency; no official endorsement should be
inferred.
REFERENCES
(1) H.M. Fribush, J.F. Fisk, "Field Analytical Methods for Superfund",
Field Screening Methods for Hazardous Wastes and Toxic Chemicals,
Proceedings of the Second International Symposium, ICAIR Life
Systems, Inc., Las Vegas, NV, pp 25-29, February 12-14, 1991.
(2) Office of Emergency and Remedial Response, Hazardous Site Evaluation
Division, Field Screening Methods Catalog, Users Guide, EPA/540/2-
88/005, U.S. EPA, Washington, D.C., July, 1990.
(3) J.M. Van Emon and R.O. Mumma, Immunochemical Methods for Environmen-
tal Analysis, ACS Symposium Series 442, American Chemical Society,
Washington, D.C., 1990.
498
-------
(4) U.S. EPA Office of Solid Waste, SW-846 Draft Methods 4010, 4020,
4030, October, 1992.
(5) H.W. Newsome and P.G. Collins, Assoc. Off. Anal. Chem, 70, 1025,
1987.
(6) Personal communication, K.R. Carter, Ensys, Inc., 1993.
(7) S.B. Friedman, CHEMTECH, 732, 1992.
(8) J.p. Mapes, et.a.Z., "Rapid, On-Site Screening est for Pentachloro-
phenol in Soil and Water-PENTA RISc*, Superfund *91 Proceedings of
the 12th National Conference, HMCRI, Washington, D.C., December
1991.
(9) J.p. Mapes, et. al., "PCB-RISc* - An On-Site Immunoassay for
Detecting PCBs in Soil", Superfund *91 Proceedings of the 12th
National Conference, HMCRI, Washington, D.C., December 1991.
(10) M. Chamberlik-Cooper, R.E. Carlson, R.O. Harrison, "Determination of
PCBs by Enzyme Immunoassay", Field Screening Methods for Hazardous
Wastes and Toxic Chemicals, Proceedings of the Second International
Symposium, ICAIR Life Systems, Inc., Las Vegas, NV, pp 625-628,
February 12-14, 1991.
499
-------
67 RESULTS OF ANALYTICAL FIELD TRIALS FOR PCBs USING AN
IMMUNOASSAY TECHNIQUE
Andrew Parsons and Alan Weiss, Millipore Corporation, Bedford,
Massachusetts 01730.
The use of immunoassays (IA) for field screening of environmental
samples has increased dramatically over the past few years. Their
utilization in remedial investigation and monitoring has gained widespread
acceptance due in large part to the efforts of the US EPA to demonstrate
the technology and begin the process of promulgating methods in which IA
methodology is incorporated.
As part of the Superfund Innovative Technology Evaluation (SITE)
program, the EnviroGard™ PCB Test Kit (Millipore Corp., Bedford, MA)
was evaluated as a field screening method for identifying soil samples
containing less than 10 mg/kg PCB. As part of the field screening activities,
the SITE demonstration contractor (PRC Environmental Management,
Inc., Kansas City, Kansas) standardized the IA with Aroclor 1242 (the
major PCB contaminant at the site) and assessed the lA's utility in
quantitating the level of PCB contamination. The screening results, which
have been published in the SITE Technical Evaluation Report (TER), were
consistent with the assay's performance claim of less than 5% false
negatives (0/96 in the actual demonstration). To assess the accuracy of the
EnviroGard PCB test kit, 89 samples that were quantitated using the IA
with Aroclor 1242 standards, were compared to values obtained by
GC/ECD analysis (SW846-8081). The correlation coefficient (r2) for this
data set was 0.45. Six of the 89 samples, all with high (> 300 mg/kg) PCB
levels, skewed the data so that when they were removed (i.e., N = 83), r2
increased to 0.87. However, when only samples with GC/MS values of <
100 mg/kg were considered, r2 dropped to 0.33. Zero to 100 mg/kg, which
encompasses most regulatory contamination limits, is the approximate
dynamic range of the IA: Quantitation of levels higher than 100 mg/kg
normally requires sample extract dilution. The high degree of scatter that
was reported in the TER might be attributed to the imprecision associated
with the field screening procedure: Analyses were performed in a non-
laboratory environment, no reference (e.g., performance evaluation)
samples were included, all results were based on single determinations, and
the analyst, although competent, was only marginally trained in IA
methodology. In order to determine if improving the method precision will
improve its accuracy, the SITE demonstration samples will be reanalyzed
under more rigorous, laboratory conditions.
If no improvement in accuracy is seen, it would suggest that IA
technology, at least in terms of PCB analysis, might be limited to screening
applications. On the other hand, if upgrading the analytical data quality
and running the method under controlled laboratory conditions improve
the lA's accuracy (i.e., r2 > 0.8) in the 0 - 100 mg/kg range, it may be
possible to bring many of the advantages of this innovative technology to
quantitative analyses.
Data from the reanalysis will be presented and compared to the
results obtained during the field demonstration. Based on the reanalysis,
recommendations with regard to the scope and limitations of PCB
immunoassay will be made.
500
-------
58 IMPORTANT FACTORS IN ENHANCING SUPERCRITICAL FLUID EXTRACTION
EFFICIENCIES FOR ENVIRONMENTAL APPLICATIONS
J.M. Levy. L. A. Dolata, R. M. Ravey, Suprex Corporation, 125
William Pitt Way, Pittsburgh, PA 15238 (412) 826-5200
ABSTRACT - Poster Presentation
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. On-line SFE/GC-MS is a powerful
technique to accurately analyze and quantitate environmental
analytes. In addition, 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 (i.e., GC, LC, IR, NMR and UV) .
Moreover, the advantages of SFE can be further augmented by the
use of automation for greater sample throughput which can be
especially important for environmental applications.
Examples will be presented showing the use of SFE/GC-MS and FID
methodologies for the determination of different target analytes
in environmental matrices, such as polynuclear aromatic
hydrocarbons (PAHs), total petroleum hydrocarbons and pesticides
in the soil. The discussion will also focus on the experimental
verification of optimized SFE variables to achieve efficient and
quantitative extractions of the target analytes in environmental
solids. An example is shown in Table 1 where an off-line SFE/GC
comparison was made between the extraction of PAHs from soil at
different pressures, indicating that higher pressures were
necessary for the complete recovery of the PAHs, especially the
four and five ring PAHs.
501
-------
Table I
Compound
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)Anthracene
Chrysene
Benzo(b,k)-
Fluoranthene
Benzo(a)Pyrene
EPA Method
8270
Acceptance
Range fppm)
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 350 atm 450 atm
23
20
566
445
1682
357
1028
703
74
74
23
*
601
471
1978
439
1459
1153
235
251
107
64
25
22
614
458
1911
400
1571
1269
284
314
155
89
502
-------
INORGANICS
-------
69 EVALUATION OF A RAPID STEAM-DISTILLATION PROCEDURE FOR THE
EXTRACTION OF CYANIDE FROM LIQUID AND SOLID WASTES
Edward M. Heithmar. U. S. Environmental Protection Agency, P. O. Box 93478, Las
Vegas, NV 89193-3478; Richard Herman and Michael R. Straka, Perstorp Analytical
Environmental, 9445 SW Ridder Road, Suite 310, Wilsonville, OR 97070
Cyanide in waste occurs as free cyanide (CN' or HCN) or in complexes with metal ions.
Depending on the thermodynamic stability and kinetic lability of the complexed forms,
they may or may not be toxic to aquatic organisms if leached into surface water. For
the estimation of potential water pollution, two operationally defined cyanide
determinations are often performed on wastes. Total cyanide is usually determined by
air-purged refluxing of the waste acidified with dilute sulfuric acid in the presence of
magnesium chloride catalyst (1). The volatilized HCN is trapped in an alkaline scrubber
solution and determined by one of a variety of methods. Strongly complexed cyanide
(e.g., ferricyanide and ferrocyanide) is often determined by a second air-purged reflux
after oxidation of other cyanide forms with alkaline hypochlorite. The presumed-toxic
"cyanide amenable to chlorination" is determined by difference (1). An alternate
estimation of the more toxic cyanide species can be made by the "weak-acid dissociable
cyanide" procedure, which utilizes air-purged refluxing to liberate HCN from a sample
containing zinc and buffered at pH 4.5. Under these conditions, iron cyanide complexes
do not generate HCN (2).
All of the above procedures for the measurement of total cyanide and the estimation of
the more toxic cyanide fraction are problematic. In addition to being prone to numerous
interferences, they all rely on the air-purged reflux for the liberation of HCN. The
reflux requires a minimum of about 90 minutes per sample. The procedures prescribed
in SW-846 require a large reflux apparatus that precludes the simultaneous refluxing of
several samples. None of the procedures have been validated for the determination of
cyanide in solid waste. An alkaline extraction procedure for such samples is provided
in SW-846, but recoveries are poor (3).
A steam-distillation procedure for the liberation of HCN from samples containing free
and complexed cyanide will be described in this presentation. The method is rapid (< 15
minutes per sample) and provides good cyanide recovery from synthetic samples.
Statistical analysis of the effect of various experimental parameters will be presented.
Results for liquid and solid wastes using steam distillation will be compared to standard
methods of analysis.
REFERENCES
(7) "Total and Amenable Cyanide," Method 9010A, Test Methods for Evaluating
Solid Wastes (SW-846), U.S. Environmental Protection Agency, Washington,
503
-------
D.C., November 1990.
(2) "Standard Test Methods for Cyanide in Water: Test Method C - Weak Acid
Dissociable Cyanides," Standard Test Method D 2036-91, Annual Book ofASTM
Standards, Volume 11.02, pp. 83-99, American Society for Testing and
Materials, Philadelphia, PA, 1992.
(J) "Cyanide Extraction Procedure for Solids and Oils," Method 9011, Test Methods
for Evaluating Solid Wastes (SW-846), U.S. Environmental Protection Agency,
Washington, D.C., November 1990.
Notice: Although the research described in this presentation was conducted in part by
the U.S. Environmental Protection Agency, it has not been subjected to Agency review.
Therefore, it does not reflect the views of the Agency. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
504
-------
70 ABSTRACT
A METHOD EVALUATION STUDY FOR THE ANALYSIS FOR HEXAVALENT
CHROMIUM IN SOLID SAMPLES USING A MODIFIED ALKALINE DIGESTION
PROCEDURE AND COLORIMETRIC DETERMINATION
R.J. Vitale. Quality Assurance Specialist/Principal, Environmental Standards, Inc., 1220 Valley
Forge Road, P.O. Box 911, Valley Forge, PA 19482.
ABSTRACT
The difference between the levels of hexavalent chromium (Cr(VI)) and trivalent chromium
(Cr(m)) in environmental samples is important from a human health concern due to the
significantly greater toxicity of Cr(VI) compared to Cr(m). Consequently, knowledge of the
chemical speciation and distribution of these two valence states of chromium in solid matrices
is essential for adequate characterization and risk assessment of sites containing elevated levels
of chromium. While current approved analytical methodologies exist to successfully differentiate
between the 3+ and 6+ valence states in the liquid (aqueous) medium, there is currently no
approved method for the determination of total Cr(VT) in the solid medium. An alkaline
digestion method (Method 3060) appeared in the SW846 2nd Edition as a preparatory method
for Cr(VI) analysis in soils. However, subsequent research and evaluation of the method in
1986 resulted in Method 3060 being omitted from SW846 3rd Edition due to a lack of
predictability in environmental samples relating to oxidation of Cr(in) and/or reduction of
Cr(VI). More than a thousand field samples and the results of a recently completed method
evaluation study (MES) have demonstrated significant improvement over Method 3060 (SW846
2nd Edition). The MES which is the focus of this presentation, was divided into four distinct
portions with the objective of showing that the modified Method 3060 is predictable and reliable
within the constraints of the sample types evaluated. These include: (1) sample
homogenization, (2) spiking studies using Cr(HI) and Cr(VI), (3) a comparison of the modified
Method 3060 to Method 3060 and (4) a mass balance study to determine the ultimate fate of
Cr(VI) spikes with respect to chromium oxidation and/or reduction. Nine different types of solid
matrices were selected for testing, ranging from quartz sand to chromite ore processing residues.
All samples were characterized for auxiliary parameters including total organic carbon (TOC),
total sulfides, pH, oxidation/reduction potential (ORP), and percent solids. The results of the
study indicate that good matrix spike recoveries (76-115%) and duplicate precision were
achieved in non-reducing samples using the modified alkaline digestion (Method
3060)/colorimetric technique (Method 7196A). In contrast, for highly reducing samples (e.g.,
anoxic sediments), zero percent Cr(VI) matrix spikes were obtained. For these samples,
auxiliary characterization parameters and mass balance results demonstrated that such samples
are not capable of maintaining a Cr(VI) matrix spike in the 6+ valence state either in the
laboratory or in the environmental settings from which they were collected. Spiking studies
using soluble forms of Cr(IH) indicate that less than 1% oxidation of Cr(DI) to Cr(VI) is
possible on the freshly precipitated Cr(m) that forms during the alkaline digestion, which
should not hinder an accurate characterization of Cr(VE) in non-reducing solid matrices. A more
difficult challenge for analytical chemists who use this method will be to avoid the interpretation
that poor spike recoveries automatically mean that the method does not yield acceptable results.
505
-------
71 THE ANALYSIS OF REDUCING SOILS FOR HEXAVALENT CHROMIUM
Don P. Miller. Chemist, U.S. EPA Region VH Laboratory, 25 Funston Road, Kansas
City, Kansas 66115.
ABSTRACT
The Second Edition of SW-846 contained a method (3060, 7196) for the
determination of hexavalent chromium in soil. The method did not perform
adequately for reducing soils and was dropped from later editions of SW-846. A
need to analyze soil samples for hexavalent chromium exists; this paper presents a
method suitable for reducing soils. The Second Edition method worked under two
conditions: 1) the soil contained hexavalent chromium, or 2) the soil did not contain
hexavalent chromium and was non-reducing. The Second Edition method failed the
matrix spike recovery test if the soil did not contain hexavalent chromium and was
reducing. The method given in this presentation uses capillary electrophoresis (CE),
passes the matrix spike recovery test for reducing and non-reducing soils, and has a
method detection limit of 4 mg Cr(VI) per kg soil (dry weight basis). This CE
method is presently used in our laboratory for the determination of hexavalent
chromium in soil.
INTRODUCTION
Previously presented methods and subsequent variations for the determination of
hexavalent chromium, Cr(VI), in soil (for example, SW-846, Second Edition, Method
3060, 7196) have given acceptable results if Cr(VI) is present in the soil or the soil
is non-reducing. Many of the methods fail a spike recovery quality assurance, QA,
test if Cr(VI) is not present in the soil and the soil is reducing. Experience with
actual samples indicates that the previous methods failed often and that a need
existed for a new method to determine Cr(VI) in reducing soils.
This method determines the concentration of Cr(VI) in soil samples. The range is
from 0.5 to 500 ppm (mg of Cr(VI) per kg of soil). This method determines easily
extracted Cr(VI) from friable soil and other similar solid or semi-solid matrices, but
is not acceptable for oily sludge, bricks, and other similar matrices. This method
tolerates colored interferences in the soil and passes QA tests for soils which contain
reducing agents.
The data quality objectives are: 1) The result of the measurement is the
concentration of Cr(Vl) in the soil sample expressed in units of mg/kg on a dry
weight basis. 2) The desired detection limit is 4 mg/kg (dry weight) or better. The
actual detection limit based on preliminary measurements for this method appears
506
-------
in the Method Performance section. 3) The desired matrix spike recovery is 75%
or better. The control limits for the matrix spike based on preliminary measurements
appear in the Control Limits section. 4) A frequent use of the result of this method
is to decide whether the concentration of Cr(VI) in a soil sample is equal or greater
than an action level.
Summarizing the method: extract soluble Cr(VI) from the soil by using an alkaline
solution, separate Cr(VI) from interferences in the extract by capillary
electrophoresis, and measure Cr(VI) directly by absorbance at 270 nm. The
extracting solution is alkaline to inhibit reduction of Cr(VI) by any reducing agents
present in the soil.
Interferences are any soluble substances in the soil absorbing at 270 nm and co-
migrating with Cr(VI). Any reducing agent in the soil capable of reducing Cr(VI)
in alkaline solution, assuming such an agent and Cr(VI) co-exist in the soil when the
soil was sampled, will be a negative interference. Any unfilled binding sites (mineral
or organic) in the soil, assuming such sites co-exist with available Cr(VI) in the soil
at the time of sampling, will be a negative interference. The usual reducing agents
found in natural soils do not reduce Cr(VI) in alkaline solutions in the time frame
of this method.
APPARATUS AND MATERIALS
1. Capillary electrophoresis (CE) with electromigration injection, absorbance
detector at 270 nm, and a device to flush the capillary between runs. A number of
instruments are commercially available. The instrument used to gather the data in
the Method Performance section was: ISCO Model 3850 Electropherograph;
uncoated silica, 60 cm, 50 pm ID capillary; detector at 270 nm, 0.002 absorbance
range, 0.10 seconds rise time; voltage at 15 kilovolts, negative polarity (the injector
end is negative); injection by electromigration, 5 seconds at 2 kV.
2. Integrator or Strip Chart Recorder.
3. 40 mL sample vials with lids.
4. Disposable Pasteur pipettes with cotton or tissue wrapped tips.
5. pH paper.
6. Micropipettes, adjustable, 10- to 100-^L and 100- to 1000-//L.
7. Volumetric flasks, 100 mL.
507
-------
1. Extraction Solution: 0.1 gram sodium acetate, 0.04 gram alkyl
trimethylammonium bromide (Sigma M-7635 or equivalent), 0.1 mL of 10 N sodium
hydroxide, all diluted to 100 mL with deionized water.
2. 10 N NaOH: Dissolve 40 grams sodium hydroxide pellets in 50 mL deionized
water. Dilute to 100 mL with deionized water.
3. 1 N NaOH: Dilute 10 mL of above 10 N NaOH to 100 mL with deionized
water.
4. Dry reagent grade potassium dichromate (K2Cr2O7).
5. Cr(VI) Calibration Stock Solution, 2000 mg Cr(VI) per liter of solution:
Dissolve 0.566 gram potassium dichromate and 0.1 gram sodium acetate in 50 mL
deionized water. Dilute to 100 mL with deionized water.
6. Cr(VI) Calibration Standards: Dilute the following volumes of Cr(VI)
calibration stock solution to 100 mL with Extraction Solution:
Cr(VT).(m CrVIL soln)
0 0
1 50
3 150
5 250
10 500
20 1000
30 1500
50 2500
7. Dry sand known to be free of Cr(VI).
8. Method Standard Stock Solution: Dissolve 0.283 grams of potassium dichromate
and 0.1 gram of sodium acetate in 50 mL of deionized water, dilute to 100 mL with
deionized water to make stock solution with a concentration of 1000 mg Cr(VI) per
liter of solution. Prepare this solution using a separate source of potassium
dichromate than the Calibration Stock Solution in Step 5 above, and keep it as
independent as possible from the Calibration Standards.
508
-------
PC SAMPLES
The definitions of recommended QC samples are:
1. Method Blank: a sample of sand known to be free of Cr(VI) put through the
same steps as a regular soil sample.
2. Matrix Spike: an amount of the Cr(VI) Calibration Stock Solution added to
a replicate of a soil sample at the appropriate step (see Analytical Procedure section:
Sample Preparation) and analyzed as a regular soil sample.
3. Matrix Spike Duplicate: an amount (as above) of the Cr(VI) solution added
to a second replicate of a soil sample to give the same increase in concentration as
the above matrix spike, and analyzed as above.
4. Method Standard: an independent check sample of sand known to be free of
Cr(VI) spiked (as above) with Method Standard Stock Solution to give 40 mg
Cr(VI)/kg sand (or other amount as specified in a QA project plan) and put through
the same steps as a regular soil sample.
SAMPLE COLLECTION. PRESERVATION. AND HOLDING TIMES
Collect soil samples using standard field procedures. Since the Cr(VI) and other
interfering substances may be unstable in the disturbed soil, analyze the samples as
soon as possible after sample collection. Do not add preservatives to the sample.
Holding time studies were not conducted.
ANALYTICAL PROCEDURE
1. SAMPLE PREPARATION
Tare a 40 mL wide-mouth glass vial (as commonly used for VGA analysis) and
transfer approximately 20 grams of the sample into the vial. Measure and record the
actual sample weight. Add the extraction fluid to the sample in the vial, 15 mL or
enough to make a free-flowing slurry, and mix the solid and liquid phases by shaking.
Calculate and record the actual volume of extraction fluid using the measured weight
and the density of the fluid. Test the pH of the slurry immediately and, if necessary,
adjust the pH with 10 N NaOH to greater than 8.5. For matrix spike and matrix
spike duplicate samples, add the spike to the sample after the pH is adjusted.
Experience showed that analysis should be completed within two hours after adding
the extraction solution to prevent loss of Cr(VI).
509
-------
2. INSTRUMENT
Set up the CE instrument with electromigration injection according to the
manufacturer's instructions. Use extraction fluid to fill both reservoir cups.
3. CALIBRATION
Analyze each of the calibration standards. Flush the capillary between standards
with three drops of extraction fluid.
Construct a calibration curve. Initial trials indicated the correlation coefficient was
0.995 or greater over a range of zero to 50 mg/L.
4. SPIKING PROCEDURE
Calculate the amount of spike (S) to add to the sample as follows:
Let "D" equal the desired increase in concentration of Cr(VI) due to the added spike
to the soil sample, in units of mg Cr(VI)/kg soil, dry weight basis.
Then,
S = V * 1.000.000
V'2000 - 1
D * W* F
where:
S = the volume of calibration stock solution (which is
2000 mg Cr(VI) per liter) added as a spike to the soil
sample, in units of //L,
V = the volume of extraction fluid added to the soil
sample, in units of L,
1,000,000 = the units conversion factor to convert the
units of S from L to
2000 = the concentration of the calibration stock solution
in units of mg Cr(VI)/L,
D = the desired concentration increase, in units of
mg Cr(VI)/kg, dry weight basis,
510
-------
W = the weight of the soil sample in the vial, "as is", in
units of kg, wet weight basis, and
F = the unitless fraction of dry material in the soil
sample.
Add the spike, S (jiL), to the sample at the appropriate step given above, and mix
thoroughly by shaking.
5. ANALYSIS
Draw about 100 /*L of the liquid from the slurry through a tissue or cotton-tip into
a Pasteur pipet and transfer the clear liquid into the sample cup in the CE
instrument. Inject the sample by electromigration into the instrument using a 5-
second injection time. Apply the CE voltage and record the absorbance measured
at 270 nm with a chromatographic integrator or stripchart recorder. For most
samples, the Cr(VI) absorbance peak occurs at about two minutes. The retention
time depends on the concentration of the buffer, the flush history of the capillary, the
temperature, and the voltage.
Since other electrophoretic peaks appear in the electrophoreogram and the capillary
becomes coated with contamination when running regular soil samples, flush the
capillary before every injection using three drops of 1.0 N NaOH, followed by six
drops DI water, and finally six drops of extraction fluid. Indications of inadequate
flushing are shifts in retention time, ghost peaks, and broad peaks. A small peak
observed at the correct time window should be confirmed with Another injection after
flushing the capillary.
CALCULATIONS
1. Calculate the result, H, as follows:
H (mg Cr(VI) /kg soil, dry) =
C (mg Cr(VD / L soln) * V (L sohri
W (kg soil, as is) * F (unitless, solids in soil)
where:
H = the measured concentration of Cr(VI) in the soil, dry
weight basis, in units of mg/kg,
511
-------
C = the concentration of Cr(VI) measured in the extraction
fluid, taken from the detector response and the
calibration graph, in units of mg/L,
V = the volume of extraction fluid added to the soil sample
in units of L,
W = the weight of the soil sample in the vial, "as is" or
"wet weight", in units of kg,
F = the unitless fraction of dry material in the soil
sample.
2. Calculate the spike recovery for the matrix spike and the matrix spike duplicate
as follows:
% Recovery (spike) = 100 * R/(H + D)
where:
% Recovery (spike) = the matrix spike recovery in units of
percent, dry weight basis,
H = the measured unspiked sample result given above, in
units of mg/kg, dry weight basis,
D = the increase in concentration of Cr(VI) in the sample due
to the spike added, in units of mg/kg, dry weight basis,
R = the measured spiked sample result, obtained using the
same expression as H above, in units of mg/kg, dry
weight basis.
3. Calculate the method standard recovery as follows:
% Recovery (standard) = ( H / STD ) * 100
where:
% Recovery (standard) = percent recovery of the method
standard in units of percent, dry
weight basis,
512
-------
H = the measured sample result given above, in units of
m8/kg, dry weight basis, and
STD = the actual or true value of the concentration of
Cr(VI) in the standard sample, ip units of mg/kg dry
weight basis. The default value of STD was set
to 40, but may be changed to meet project objectives
specified in a QA project plan.
4. Calculate the percent relative standard deviation of the matrix spike and the
matrix spike duplicate (the two replicate samples may have different weights, but
each must have the same increase in concentration due to the spike, see Analytical
Procedures section: Spiking Procedures) as follows:
%RSD = sqrt(2) * abs( MS - MSD ) / (MS + MSD)
where:
%RSD = percent relative standard deviation in units of
percent,
sqrt(2) = the square root of 2,
abs(X) = the absolute value of X,
MS = the measured result for the matrix spike sample,
obtained as H above, in units of mg/kg dry weight
basis, and
MSD = the measured result for the matrix spike
duplicate sample, as MS above.
CONTROL LIMITS
Using the data given in the Results Section, the 95% confidence interval for the
control limit for the percent recovery of the method standard and the matrix spike
was from 60% to 140%. This assumes the method standard and the matrix spike
behave the same.
The 95% control limits of the percent relative standard deviation of the lab matrix
spike and the lab matrix spike duplicate were: lower limit = 5%, upper limit =
10%.
513
-------
When conducting field spiking, note that field spiked samples are potentially
unstable, since Cr(VI) and reducing agents could co-exist indefinitely only in a
completely dry soil spiked with a dry substance.
METHOD PERFORMANCE
The following is a summary of results for a number of different matrices using this
method. All following values were the result of seven separate replicate samples
analyzed in one batch. All values are in units of mg Cr(VI) per kg soil, dry weight
basis. All entries in the following tables were calculated using the expressions
specified above, except the MDL which followed 40 CFR Part 136 Appendix B. The
naming and data reporting conventions are:
Av.Obs. the average observed value,
n = 7 indicates seven replicate samples were used
to determine the reported value,
A%R the average percent recovery,
MDL the detection limit,
df the degrees of freedom of the reported value,
AL%RSD the average laboratory percent relative
standard deviation.
1. Fertilite (tm) 'Top Soil", UPC# 7270140030, Hyponex Corporation, 14111
Scottslawn Road, Marysville, Ohio 43041. (This soil failed QA using the previous
methods.)
Spike Added Av.Obs.(n = 7>> A%Rfn = 7^ MDL(df=6^ AL%RSDrdf=6^
0 non-detect -
10.3 8.4 84.1 3.2 119
102.8 96.9 94.3 - 4.5
514
-------
2. U.S. EPA Synthetic Soil Matrix, Risk Reduction Engineering Laboratory, Release
Control Branch, Edison, New Jersey 08837-3079. (This soil passed QA using
previous methods.)
Spike Added Av.Obs.(n=7) A%R(n=7^ MDUdf=6) AL%RSD(df=6^
0 non-detect ...
4.0 4.3 106.5 4.3 32.5
40.0 53.4 133.5 - 9.2
3. The above Synthetic Soil Matrix fortified with 2% glucose by weight. (This
soil failed QA using previous methods.)
Spike Added Av.Obs.(n=7^ A%R(n=7^ MDL(df=6^ AL%RSD(df=6)
0 non-detect - -
4.0 6.4 161. 4.3 21.9
40.0 27.0 67.5 - 8.9
4. Loess, underlying material obtained from road cut on Interstate 70 between the
57th Street exit and the 61st Street overpass, Section 11, T11S, R24E, Wyandotte
County, Kansas. (This soil passed QA using previous methods.)
Spike Added Av.Obs.(n=7^ A%R(n=7) MDL(df=6) AL%RSD(df=6)
0 non-detect ...
5.52 4.2 75.9 0.55 4.8
55.2 52.7 95.5 - 2.8
SUMMARY
The above text is an abridged version of the SOP used in our laboratory for the
analysis of soil samples for hexavalent chromium. The throughput for one person
using this method is 20 to 30 samples per day. This compares with a throughput of
2 to 3 samples per day using the older method. Analysts familiar with both methods
prefer the CE method. Background reading for this method includes: 1) P. Jandik
et al., "Electrophoretic Capillary Ion Analysis: Origins, Principles, and Applications."
LC-GC 9 634-645, 1991, and 2) J. E. McLean & B. E. Bledsoe, "Behavior of Metals
in Soils." EPA/540/S-92/018.
515
-------
79 FACTORS AFFECTING SAMPLE THROUGHPUT FOR AN ICP-OES SYSTEM
' WITH A SEGMENTED-ARRAY CCD DETECTOR
Z.A. Grosser. K.J. Fredeen, C. Anderau, D.A. Yates, K.W. Barnes, T.J. Gluodenis, Jr.
The Perkin-Elmer Corporation, 761 Main Ave., Norwalk, CT 06859-0215
Introduction
Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) is a fast,
economical technique commonly employed for the determination of metals in a
variety of environmental matrices. Advances in optics and detection systems for
ICP-OES have led to instruments that have new capabilities and better
performance (1, 2). In this work we examine the potential to increase sample
throughput by evaluating the factors that may be varied versus the data quality
objectives for the analysis. We will characterize several of the ICP performance
characteristics and evaluate several real applications for potential time savings.
Data quality objectives (DQOs) are critical in evaluating the performance that
must be achieved with an ICP-OES method. DQOs that will be particularly
important in evaluating sample throughput are detection limits, accuracy
requirements, precision, and dynamic range. If these are well-characterized for a
particular analytical situation then potential sacrifices can be identified and
judicious choices made in pursuing sample throughput increases.
Several ICP-OES instrumental factors will affect the DQOs and analysis time.
These factors were investigated to quantitate the magnitude of each effect.
Integration time will affect both the detection limits and precision of an analysis.
Fifteen elements representing the full wavelength range were studied and showed
that increasing the integration for each replicate will improve both detection
limits and precision. The greatest improvement is for elements with transitions
below 350 nm where photon shot noise is the predominant source of noise
affecting the signal and background.
Background correction can remove spurious radiation from the desired signal by
comparison with an off-line measurement. If background correction is done
simultaneously, rather than sequentially, detection limits can be improved for
elements with transitions above 250 nm and exceed detection limits without
background correction above 350 nm. These transitions take place in a region of
the spectrum dominated by flicker noise, which can be partially removed by the
correlation in simultaneous background correction.
In a complex sample, background correction may be even further improved by the
use of a multivariate correction procedure such as Multicomponent Spectral
Fitting (MSF) (3). MSF uses the full spectrum to create a model of peak shapes
used in correcting the analytical signal. Figure 1 demonstrates the magnitude of
improvement that can be expected with different types of background correction
procedures. MSF improves the detection limit beyond that which can be achieved
without background correction.
516
-------
No BGC, 8 sec
Simult BGC, 8 sec
Seq BGC, 24 sec
MSF, 8 sec
Zn Pb Cd Ni Mn Fe Cr V Cu Ti La Al Sc K
Element
Figure 1. Effect of Background Correction on Detection Limits
Three applications were studied to demonstrate the practical application of the
decision-making process. The first application is to increase the sample
throughput for a wear metals trend analysis (WMTA) requiring moderate
detection levels for 16 elements. An oil additive and trace analysis (OATA)
requires lower detection limits and precision better than 1%. Finally a TCLP
analysis for 7 elements extracted into an acetic acid buffered matrix was
examined. Table I lists the data quality objectives for the three analytical
situations.
Table I
Data Quality Objectives of
Three ICP-OES Applications
WMTA
No. Elements 16
No. Reads/Sample 1
Detection Limits (mg/L) 0.01-0.5
Accuracy ± 10%
Precision (RSD) 3%
LDR (orders/elem.) 3
Rinse Out (orders) 3
OATA
16
2
0.001-0.1
4-5
TCLP
7 (14 lines)
3
<0.1xMCL
±10%
3%
4-5
LDR to 3x DL LDR to 3x DL
Results
After careful evaluation of the various instrumental factors the WMTA was
speeded up to allow the analysis of a standard, blank, and 20 samples in 13.5
minutes. Including a QC check sample every fifth sample, 98 samples can be
analyzed in 1 hour. The average precision obtained at concentrations above 50x
the detection limit were 1.4% RSD.
517
-------
The OATA speed was increased to analyze 35 samples per hour. For low
concentration samples an analysis time of 2 minutes was necessary to meet the
DQOs, but when concentrations approach 20 mg/kg the analysis time can be
reduced to 1.2 minutes/sample.
Two emission lines were monitored for each TCLP element to add confirmation
information to the analysis. The analysis was performed in less than one minute
with 65 samples per hour analysis rate, including the QC checks. If the ICP
periodic table is monitored to evaluate the sample for other possible
contaminants the analysis is still performed in less than 2 minutes per sample.
Summary
Advances in ICP-OES technology have provided the base for improvements in
sample throughput. ICP-OES factors have been evaluated to demonstrate the
relationship with sample throughput and the concepts applied to three real-world
applications. In all cases the analyses were speeded and the sample throughput
increased significantly. The principles examined can be applied to other
situations to evaluate the potential for time savings.
References
1. T.W. Barnard, M.I. Crockett, J.C. Ivaldi, and P.L. Lundberg, Design and
Evaluation of an Echelle Grating Optical System for ICP-OES, Anal. Chem. 65,
pp. 1225-1230, 1993.
2. T.W. Barnard, M.I. Crockett, J.C. Ivaldi, P.L. Lundberg, D.A. Yates, P.A.
Levine, and D.J. Sauer, Solid State Detector for ICP-OES, Anal. Chem. 65, pp.
1231-1239, 1993.
3. J.C. Ivaldi, D. Tracy, T.W. Barnard, and W. Slavin, Multivariate Methods for
Interpretation of Emission Spectra from the Inductively Coupled Plasma
Spectrochim. Acta 478(12) pp. 1361-1371, 1992.
518
-------
73 DESIGN, PERFORMANCE, AND ENVIRONMENTAL APPLICATIONS
OF AN ICP ARRAY DETECTOR SPECTROMETER
D. Nvgaard. ICP Product Manager, Frank Bulman, and Manny Almeida, Baird Corporation,
125 Middlesex Turnpike, Bedford, MA 01730
This paper will describe and evaluate a new kind of ICP-optical emission spectrometer — one
utilizing a mega-pixel, two dimensional, solid state array detector. With more than one million
individual detector elements, the array captures a high resolution electronic photograph of the
complete UV-visible emission spectrum. All elements in the periodic table that emit UV-visible
light in an ICP discharge can be detected and quantified at any number and any combination of
wavelengths. This capability enables rapid qualitative and semiquantitative screening of
unknown samples, an application that will be discussed in detail especially as it pertains to waste
analysis. In addition, rigorous quantitative analysis of water and waste will be demonstrated to
the quality control standards established by the Contract Laboratory Program.
519
-------
74 VARIABILITY IN TCLP METALS RESULTS FROM
STABILIZED WASTES
Gene Merewether. Analytical Chemist, Nilesh Shah, Senior Analytical Chemist;
Quality/Analytical Programs, Chemical Waste Management, Inc., 7250 West
College Drive, Palos Heights, Dlinois 60643.
ABSTRACT
The purpose of this multi-laboratory study was to identify variables in Toxicity
Characteristics Leaching Procedure (TCLP) and subsequent analysis and to recommend
standard protocols to improve inter-laboratory precision. The TCLP, together with
subsequent digestion and instrumentation methods, allow choices of laboratory practices.
The choices made by individual laboratories contribute to inter-laboratory variability. For
instance, TCLP allows extraction of any particle size less than 3/8 inch, and any extraction
time from 16 to 20 hours. Additionally, choices made about specific digestion methods
and instrument methods may also allow further inter-laboratory variation.
For analyzing TCLP metals, a laboratory can choose from any appropriate digestion or
instrument procedures. For example, lead can be analyzed either by graphite furnace
atomic absorption (GFAA) spectroscopy (Method 7421) or inductively coupled plasma
(ICP) atomic emission spectroscopy (Method 6010). The digestion procedure associated
with each analysis is Method 3020 for GFAA or Method 3010 for ICP. Inevitably, there
are differences between the methods, and there is variability within each method.
Individual waste matrices can also add variability to the results.
The goal of improving inter-laboratory precision therefore required identifying and
quantitating variables that lead to differences in TCLP results. The approach taken in this
study was to have each participating laboratory analyze a "tailor made" proficiency sample
so that inter-laboratory and intra-laboratory variability could be measured. Existing
QA/QC data from proficiency samples and standard reference materials were also
examined. The results from round robin analyses and the recommendations for analytical
protocols are discussed.
520
-------
75 AMALGAMATION CVAA TO IMPROVE THE DETECTION LIMIT OF MERCURY
IN ENVIRONMENTAL SAMPLES
Steven Sauerhoff. Susan Mclntosh, and Zoe Grosser
The Perkin-Elmer Corporation, 761 Main Ave, Norwalk, CT 06859-0219
Introduction
Mercury is of continuing interest in the environment due to its persistence and
toxicity. Mercury determination has traditionally been performed using cold vapor
generation coupled with atomic absorption (CVAA) to achieve a detection limit of
200 ng/L (ppt) in a prepared sample. Each sample is handled individually and the
technique is very labor and time intensive. Manual sample handling may also
expose the sample to additional sources of contamination, increasing the variability
in the results obtained with the technique.
Flow injection techniques (FI), where reagents are mixed with the digested sample
automatically, provide several advantages. The samples are isolated from
atmospheric and handling contamination. The volume of sample necessary to
generate the signal is greatly reduced, which also reduces the consumption of
reagents and waste generated from the process. Interferences are reduced because
the mercury vapor is separated from the mixture very quickly, limiting the time for
adverse reactions to occur. The sample throughput is increased and the procedure
can run unattended (1).
Although FI can significantly improve the quality and ease of a classical mercury
determination, lower detection limits would also be useful. Coupling FI with
newer light sources, such as the EDL System 2 can lower the detection limit to 60
ng/L for routine analyses. However, research into bioaccumulation of mercury and
naturally occurring marine levels require the establishment of still lower detection
levels. This work describes the development of a FI system coupled with
preconcentration of mercury by amalgamation to reduce mercury detection limits
to 2 ng/L for an 8.5-mL sample size (2).
Figure 1 shows the FI system coupled to the amalgamation accessory for mercury
preconcentration. Mercury vapor is generated in the mixing manifold and
separated in the gas-liquid separator. The mercury vapor is trapped on a
gold/platinum gauze and held there until a sufficient amount is deposited. The
deposition is halted and the gauze heated to release the mercury as a concentrated
plug into the absorption cell of the atomic absorption spectrometer, where the
absorbance is measured and quantitated.
Results
Figure 2 compares the signals obtained for 1000 ng/L mercury with and without
amalgamation preconcentration. The method detection limit, measured using the
procedure in CFR, Part 136 Appendix B, was found to be 2 ng/L for an 8.5-mL
sample and 60 seconds of deposition time. This represents the detection of 17 pg
of mercury on an absolute basis. Larger samples and longer deposition times might
decrease the detection limit, but eventually reagent and handling contamination
are going to limit further improvement.
521
-------
KSnCI,
Figure 1. FI Mercury Amalgamation System
The mercury vapor is channeled through the FI valve to the amalgamation system.
This controlled vapor deposition technique reduces sample carryover and improves
precision. The precision for replicate samples was 2.5% RSD at 20 ng/L of
mercury.
8.244
1.88 ppb
1.86 ppb Hg CU
i<— 1.8 ppb Hg 68 sec. (Wilgo.
<—1.8 ppb Hg Cold Va
h^^^Mrib
(nc)
2fl.ee
Figure 2. Mercury Signals With and Without Amalgamation
522
-------
Summary
The system described in this work provides several advantages for the
determination of mercury compared to conventional CVAA batch analysis. The
detection limit using amalgamation is reduced to levels that allow the study of
marine and estuary ecosystems. The addition of reagents is automatic and the
sample throughput is improved to allow the determination of 30 samples per hour
in unattended operation.
References
1. S. Mclntosh and B. Welz, Perkin-Elmer Method 245.1A; Determination of
Mercury in Drinking Water and Wastewater by Flow Injection Atomic Absorption
Spectrometry (Cold Vapor Technique), ENV-12A, available from The
Perkin-Elmer Corporation, 761 Main Ave., Norwalk, CT, 06859-0012.
2. S. Mclntosh, The Determination of Mercury at Ultra-Trace Levels Using an
Automated Amalgamation Technique, Atom. Spectrosc., 14 (2), pp.47-49 (1993).
523
-------
7fi THE DEVELOPMENT OF EPA TOTAL MERCURY METHOD USING COLD VAPOR FLUORESCENCE
' MERCURY DETECTION SYSTEMS. THE MERCURY METHOD WILL BE USED FOR THE ANALYSE
OF MERCURY IN WATER, WASTEWATER, SEA WATER AND RELATED MATRICES.
Authors: Billy B. Potter. Research Chemist, Inorganic Chemistry Branch,
Chemistry Research Division, Environmental Monitoring Systems Laboratory, U.S.
Environmental Protection Agency, 26 W. Martin Luther King Dr., Cincinnati Ohio
45268.
Stephen E. Long and Jane A. Doster, Technology Applications, Inc., Cincinnati,
Ohio.
Robert Davis, Computer Services Corporation, Cincinnati, Ohio.
ABSTRACT:
The U.S. Environmental Protection Agency (USEPA) is developing a method for
the determination of total mercury found in water and sediment at the part per
trillion (ppt) level. The total mercury method has an estimated method
detection limit (MDL) of 5 ppt to 20 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:
Methylmercury poisoning of humans and cats in Japan known as "Minamata" disease
(1950-1970) was caused by the consumption of mercury (10 to 24 parts per million
[ppm]) contaminated fish. This environmental catastrophe marked the beginning
of a world-wide concern that mercury may be a global pollution problem. The
history, hazards, and concerns about mercury pollution were documented by 19721.
The Minamata Bay tragedy led to independent investigations of the Great Lakes
region by the United States and Canada. These studies found high levels of
mercury in fish. From these early studies, it became apparent that a
standardized mercury method was needed for the exchange of mercury data between
the two countries. Jacobs et al2 in 1960 introduced the oxidation/digestion of
biological samples by wet digestion. The wet digestion technique of mercury
samples using potassium permanganate and sulfuric acid gained acceptance in
Canada and United States. The Canadian government made strides to standardize
this approach. The credit for developing wet digestion and the cold vapor
(flameless) atomic absorption technique is credited to the Canadian team Hatch
and Ott3 However, their approach to digestion of mercury samples varied from
the Jacobs' approach. This variation of sample digestion and a lack of
standardization of methodology created data compatibility problems between
countries. The Canadian government elected to standardize on one of these sample
digestion techniques. By 1970 the cold vapor technique and the permanganate
sulfuric acid oxidation/digestion was accepted by the Fisheries Research Board
of Canada and Federal Water Quality Administration (FWQA) as provisional methods
with minor differences in sample digestion. The United States reviewed the
524
-------
results of the Canadian methods, then provided the Water Pollution Control Agency
(WPCA, 1970) a provisional method for mercury analysis.
The USEPA, established in 1971, took over the responsibilities of the WPCA. The
USEPA, Environmental Monitoring and Support Laboratory Cincinnati (EMSL-
Cincinnati) continued the research on the WPCA provisional cold vapor method.
The USEPA, EMSL-Cincinnati proceeded to standardize the method. The EPA's method
development protocol required a feasibility study and single laboratory
validation study. The method feasibility study was conducted by J.F. Kopp, M.C.
Longbottom, and L.B. Lobring.A These researchers used an experimental design
that included the analysis of known interferences (chloride, sulfide, organics)
and other variables (digestion, heat, time). The result of this research in 1972
lead to the development of the optimized "Cold Vapor" USEPA Mercury Method 245.1
for mercury-total5 that was completed in 1976 and revised in 1983. The USEPA
adopted an extensive method validation policy and then proceeded to confirm the
single operator precision by expanding to multi-laboratory confirmation. The
USEPA, EMSL-Cincinnati, Quality Assurance Branch conducted a nation-wide
validation study. This study validated the USEPA Mercury Method 245.1 that
produces data of known quality at the part per billion (ppb) level. However, new
regulatory demands are now being made to detect mercury at the parts per trillion
(ppt) level. These regulatory demands are a result of new consensus among
researchers6-7-8'9 over growing mercury pollution of the aquatic food chain.
There are at least 23 states with existing fish consumption advisories for
mercury and numerous USEPA Superfund sites with significant mercury contamination
of soils, ground and surface waters. The fish advisories and the need to monitor
trace level mercury migration from Superfund sites has prompted many requests for
a mercury method able to detect mercury at the ppt level. To protect the aquatic
environment, many states are now implementing monitoring and enforcement of
National Pollutant Discharge Elimination System (NPDES) permits established by
water quality-based effluent limitations. The effluent limitation for mercury
uses a bioaccumulation factor in the equation that has resulted in permit values
being set below the USEPA Method 245.1 detection limit for total mercury. The
current USEPA Method 245.1 has a detection limit for mercury of .2 /ig-Hg/L (.2
ppb) in water. This current MDL is too high to satisfy monitoring requirements
as established by the USEPA and state water quality-based effluent limitations
requirements. A mercury method with a low detection limit (20 to 2 ppt) is
desired for the determination of mercury in natural waters. This would support
studies that are attempting to define probable sources of mercury contamination
in the aquatic food chain. There is a study being proposed for the Florida
Everglades that will require ppt detection limits.
The State of Florida has determined that many sport fish caught in the South
Florida Everglades are contaminated with high levels (0.5 to 4 ppm) of mercury.
This has caused a fish advisory to be issued for the Florida Everglades. The
search for the cause of mercury contamination is hampered by a mercury method
having a detection limit of 0.2 ppb. This detection limit is too high for
ambient water analysis where mercury levels are expected to be in the low ppt
level. The State of Florida is located in USEPA Region 4. The USEPA,
525
-------
Environmental Monitoring Systems Laboratory Cincinnati (EMSL-Cincinnati)
received a request to support USEPA, Region 4, by providing methods development
and quality assurance resources for the Everglades study. Region 4 will provide
a plan of study for the Everglades ecosystem10. The study of the Everglades will
require methods that can measure mercury below the ambient levels found in
nature.
The current EPA Mercury Method 245.1 may be applied to many sample types
(industrial and domestic). Although it is a rugged method, it suffers from
interferences that absorb UV radiation that can cause a positive bias. ^There are
two ways to overcome these interferences and lower the detection limits:
(a) The atomic fluorescence of mercury is independent of UV absorption
interferences. Atomic fluorescence detects mercury by emission rather than
by absorption. Since emission is inherently more sensitive than atomic
absorption, it is expected that detection limits may be lowered to parts
per trillion (ppt, ng/L) range.
(b) Interferences may be eliminated by pre-concentration of mercury on gold
amalgam. Gas vapors containing interferences pass through the instrument
before the mercury analysis. The gold/mercury amalgam trap is heated,
releasing the mercury vapor for analysis by atomic fluorescence detector.
This increases mercury sensitivity by producing a highly concentrated peak.
EXPERIMENTAL:
A statistically-based experimental design or chemometric approach as described
by Deming and Morgan (1987)11 was selected for the evaluation of the mercury
method. 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 Room Protocol.
526
-------
The instrument selected for the study is a QuesTron "PSA" Automated Mercury
Analyzer (Merc 22) . This instrument offers a complete system containing
Autosampler, Gold amalgam pre-concentrator, fluorescence detector, an IBM
compatible USEPA Contract Laboratory Program (CLP) quality control software
program and data acquisition system.
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
portion of the method development and is not part of the scope of this
experimentation.
FLUORESCENCE
DETECTOR
PERISTALTIC
PIMP
CARBON
WASTE FILTER
CVent to HcxxO
AUTOSAIiPLER
GABBIER y
S
«
I
• /
f
•/c
I
\LIQUID
WASTE
COLLECTION
vant to Hood}
GASM-IQUID
SEPARATOR
' If*
d u
ARSON GAS
FIGURE 1: PSA AUTOMATED MERCURY FLUORESCENCE SYSTEM
527
-------
RESULTS:
Instrument Optimization:
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 Simplex program written in BASIC was used
for algorithm computations. The optimized flow rates calculated by the program
were 156 mL/min for the sheath gas and 329 mL/min for the carrier gas, although
there was some evidence from the Simplex convergence behavior that the two
variables were not interactive. The timing sequences on the hydride generator,
which define the analytical cycle, were also optimized manually to obtain good
peak shapes. Work continued on the evaluation and optimization of instrument
performance and assessment of sample preparation procedures.
Improvements were made on the system to increase precision and sensitivity. The
rotameters for the carrier, sheath, and dryer tube gas were replaced by new
rotameters from PS Analytical. The new rotameters for the carrier and sheath gas
flows were modified to a lower and more practical range. The dryer tube
rotameter was modified to allow much higher flows to remove moisture from the
sample more effectively. The sensitivity of the instrument was also increased
over 30% upon replacement of the worn dryer tube element. After these changes
were made, a check on the precision of the instrument was performed. The system
was calibrated once an hour for seven hours and the slope of the curve was
monitored as an indication of the instrument drift. The RSD of the slope over
the day was 3.6%, a noticeable improvement over earlier drift that ranged from
about 10-20%.
Before conducting any detailed analytical benchmark studies, it was decided to
optimize the instrumentation with respect to maximizing the slope of the
sensitivity curve, while maintaining acceptable precision, accuracy and peak
shape dynamics. As only two user-adjustable variables were considered to have
a critical effect on the analytical performance of the instrument, namely the
carrier and sheath gas flow rates, it was not considered necessary to use a
simplex optimization process. The experiment consisted of running two
repetitions of a 32-factorial design. Each repetition was conducted on a
separate day to find out the day-to-day reproducibility of the system. The order
of the experiments within each repetition was randomized. The levels of carrier
gas flow chosen for the experiment were 200, 350 and 500 mL/min, while the sheath
gas flow rates chosen were 0, 350 and 700 mL/min. For each experiment, the
instrument was calibrated using a calibration blank and three calibration
standards (50, 100 and 150 ng/L). Using this calibration curve, the instrument
was then used to determine three test concentrations (25, 75 and 125 ng/L). The
output variables recorded for each experiment were the maximum peak height for
the 50 ng/L calibration concentration, the slope of the calibration line, the
correlation of the calibration line and the estimates provided by the system for
each of the three test concentrations. The results of the experiments are
summarized in Table 1 below.
528
-------
TABLE 1: DATA FROM OPTIMIZATION EXPERIMENTS
Carrier
mL/min
200
200
200
200
200
200
350
350
350
350
350
350
500
500
500
500
500
500
Sheath
mL/min
0
0
350
350
700
700
0
0
350
350
700
700
0
0
350
350
700
700
Day
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
Peak
Height
146.2
164.6
101.3
108.4
63.1
71.6
106.8
118.6
84.5
93.7
61.3
68.2
76.4
93.6
68.0
68.5
48.8
54.3
Slope
2.24
2.52
1.52
1.68
0.95
1.04
1.60
1.86
1.27
1.38
0.91
1.01
1.13
1.39
0.97
1.11
0.71
0.82
r
1.0000
1.0000
0.9999
1.0000
0.9998
0.9997
1.0000
1.0000
0.9998
0.9999
0.9999
0.9996
0.9999
1.0000
0.9998
0.9999
0.9999
1.0000
25
ng/L
24.4
24.1
26.4
24.6
25.1
23.6
25.1
24.3
24.8
23.2
22.9
25.8
24.2
23.5
21.5
24.4
24.5
25.8
75
ng/L
75.5
****
79.4
77.1
75.7
74.2
75.3
71.4
73.5
74.2
76.5
76.0
73.5
71.7
71.4
74.0
76.2
75.4
125
ng/L
****
****
118.7
****
124.8
125.6
****
****
122.8
120.9
127.5
127.1
126.5
119.8
118.6
125.1
126.4
125.0
****Test concentration signal off-scale r correlation coefficient
For the (500,0) run on day 1, two duplicates of each calibration and test
concentration were run through the system. However, it was decided that running
duplicates for all factor setting combinations would be too time consuming. All
of the remaining runs were based on just one sample of each of the calibration
and test concentrations. To equalize the underlying variance, only the numbers
for the first of the duplicates on the first run were used. The calibration
curve,' correlation coefficient and estimates for the test" concentrations were
computed manually from the peak heights produced by the instrument.
529
-------
To characterize the accuracy of the readings of the three test concentrations
the absolute percent deviation from the true concentration was used. Originally
a multivariate analysis of variance (MANOVA) was proposed for the analysis or tne
data. However, the missing values in Table 1 meant that only thirteen
observations could be used in the MANOVA. Individual statistical analyses were
therefore, run using the response variables, calibration slope and percent
deviation for each of the three sample concentrations. Peak height was not
included because it provides similar information as the slope of the calibration
line. The correlation coefficient was omitted because the calibration line was
almost always a perfect fit (r values ranging from 0.9996 to 1.0000). Type III
sums of squares were used because it could not be assumed that interaction
effects were insignificant. The day was included in the analysis because it was
of interest to know whether the performance of the system changed from day to
day. Table 2 contains the F-statistics and p-values for each of the four
response variables for the factors carrier gas flow, sheath gas flow, the
interaction term between the two and the day. This information shows that the
accuracy of the system does not change significantly when the flow levels are
varied. However, the slope of the calibration line differs significantly based
on all of the independent variables in the model.
TABLE 2: OPTIMIZATION TEST STATISTICS
Dependent
Variable
Slope of
Calibration Line
Absolute
Percentage Error
(25 ng/L)
Absolute
Percentage Error
(50 ng/L)
Absolute
Percentage Error
(125 ng/L)
Source of Error
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
F-statistic
153.02
203.71
392.61
35.53
42.44
0.43
0.39
0.46
0.54
0.00
0.73
0.16
1.28
0.91
0.11
0.86
0.22
2.33
0.82
0.05
Degrees of
Freedom
9,8
2,8
2,8
4,8
1,8
9,8
2,8
2,8
4,8
1,8
9,7
2,7
2,7
4,7
1,7
7,5
2,5
2,5
2,5
1,5
p- value
0.0001
0.0001
0.0001
0.0001
0.0002
0.8867
0.6924
0.6471
0.7129
0.9714
0.6765
0.8566
0.3350
0.5083
0.7495
0.5876
0.8067
0.1928
0.4903
0.8262
530
-------
Inspection of the data in Table 1 shows that the slope of the line increases as
either the carrier flow or sheath flow is lowered. Unfortunately, the peak shape
produced by the system at the lower gas flows is not of an acceptable shape for
routine measurements and although higher slopes would maximize the detection
power of the system it also would compromise the robustness of the measurement
process. The experimental results clearly show that the calibration line slope
is inversely related to the carrier and sheath gas flow rates. The flow rates
selected for continuing work were, therefore, based on the lowest values of the
gas flows that were consistent with the provision of reasonable peak shapes.
These experiments have indicated that adjustment of the flow rates within the
limits employed here has no effect on the overall accuracy of the system. The
flow settings for the system were selected based on a compromise between these
two parameters. The final results of the optimized instrument setting are in
Table 3 below. These optimized settings and procedures were held constant for
the remaining experiments.
TABLE 3: EXPERIMENTAL CONDITIONS FOR SINGLE LABORATORY VALIDATION
Fluorescence Instrument Condition Instrument
PSA Merlin Series AFS
Flow rate, blank
Flow rate, SnCl2
Flow rate, sample
Carrier gas rate
Sheath gas rate
Drier tube gas rate
8.00 mL/min
3.00 mL/min
8.00 mL/min
0.35 L/min
0.70 L/min
3.00 L/min
Delay time
Rise time
Analysis time
Memory time
10 s
25 s
30 s
60 s
INSTRUMENT PERFORMANCE BENCHMARKS
As part of the familiarization phase of the project, initial assessments of the
instrument detection limit (IDL) and analytical precision were made. Precision
evaluated as relative standard deviation (n=10) was estimated at 100, 150, and
200 ng/L concentration levels to be 4.2, 1.6 and 0.9%, respectively. The IDL was
evaluated using three different variations in sample preparation. All three
initial assessments yielded an IDL of around 3 ng/L, where the IDL was defined
as that concentration giving an analytical signal equal to three times the
standard deviation of the blank signal.
Once the instrument settings were optimized and improvements to the system had
been made the instrument performance was assessed in more detail. The IDL was
determined to be 0.8 ng/L. Short term precision was measured by running ten
replicates of an inorganic mercury standard at 10, 50, and 100X the IDL and
digested using the bromate/bromide technique. The RSD of the instrument response
at each concentration was 2.3, 1.3, and 0.79% respectively. The short term
precision experiment was repeated using an organic mercury standard at the same
concentrations. The RSD at 10, 50, and 100X the IDL were 2.1, 1.1, and 1.0%
respectively. Instrument stability was determined by measuring the long term
531
-------
precision of the analysis of a mid-level mercury standard over the course of one
work day. The RSD of seven separate analyses over six hours was 1.2%. A control
chart that represents the day to day long term stability of the instrument was
constructed by plotting the sensitivity slopes for twenty-five calibration
curves produced between 8/13/92 and 10/14/92. The average slope over the two
month period was 0.914 with an RSD of 6.9%. All slopes were within three
standard deviations of the average, indicating no excessive drift in the system.
COMPARISON OF SAMPLE DIGESTION METHODS
Two methods for sample digestion were assessed to determine which should be
recommended for the proposed EPA Method. The standard EPA type digestion defined
in Method 245.1 using KMn04/K2S208, and a semi-automated method for total mercury
is being evaluated for acceptance as an official USEPA method. This method is
used routinely by the Yorkshire Water Authority (YWA) in the United Kingdom12.
In the YWA method, a sample aliquot is digested using free bromine reagent,
resulting in the break down of the commonly occurring organomercury species to
mercury (II )13'1/(. Elemental mercury vapor is generated from the digested
sample by reduction with stannous chloride and is purged from solution by a
carrier stream of argon. The mercury vapor is determined by atomic fluorescence
spectrometry at 253.7 run15. The proposed method procedure is simplified and
summarized as follows:
(1) Transfer 35-40 mL of sample to a 50-raL tared container.
(2) Add 5 mL (1+1) hydrochloric acid and 1 mL 0.1N potassium bromate/potassium
bromide.
(3) Allow samples to stand for at least fifteen minutes before analysis.
(4) If the yellow color does not persist after fifteen minutes, add more
KBr03/KBr solution.
(5) Add 12% (w/v) hydroxylamine hydrochloride (NH2OH-HC1) at a cone, of 6.0 ftL
per 10 mL of the sample, to eliminate the excess bromine until yellow color
disappears.
(6) Turn on the automated instrument/detector and allow to stabilize.
(7) The sample enters gas/liquid separator with SnCl2 to form mercury vapor.
(8) The vapor is analyzed by cold vapor atomic fluorescence spectrometry.
Each digestion procedure was optimized for reagent volume and final dilution.
Sample containers for digestion and analysis were also evaluated. The original
EPA Method 245.1 utilizes glass BOD bottles. Some initial studies for this work
were done in glass volumetric flasks. The difficulty in using glass containers
for sample digestion is that their capital cost can be prohibitive for disposable
use and they must be cleaned using a rigorous and time consuming wash regimen if
532
-------
they are to be reused. At ng/L detection levels any contamination from the
sample container becomes a crucial issue. To reduce the risk of contamination,
disposable polypropylene sample containers for digestion and immediate analysis
were introduced. Initially, the digestion procedures were experimented with
using inorganic standards in deionized water to determine the stability of
mercury in each digestion matrix. Early studies showed that both procedures
processed the mercury standards equally well. The effects of heat on the
performance of the permanganate method were also initially tested with inorganic
and organic standard solutions in deionized water. No significant difference in
recovery was observed between heated and non-heated samples (Table 3) . The
importance of the heating step for the permanganate method became apparent when
more complex matrices were studied. The current methods using permanganate
require that additional reagents be added to digest adequately samples high in
chloride concentrations. This complicates the analysis procedure since it
requires the preparation of a separate blank for samples treated with additional
reagents.
To compare the permanganate method with the bromate method in high concomitant
chloride conditions, an artificial seawater with an estimated chloride content
of 18,000 ppm was used as a sample matrix. Three aliquots of artificial seawater
were spiked to a concentration of 100 ng/L with either 100% inorganic mercury,
50:50 inorganic/organic mercury, or 100% organic mercury. Two samples of each
spike were digested and analyzed for each digestion procedure. The permanganate
method was tested both with and without heating. The average recoveries for the
inorganic, 50:50, and organic mercury spiked samples using the bromate digestion
were 93.8, 92.1, and 90.4% respectively with close agreement between replicates.
For the permanganate digestion using heating, average recoveries were 95.3,
85.6, and 81.4% respectively with a high relative percent difference between
replicates. Repeating the permanganate digestion without heating yielded
recoveries of 108.6, 83.3, and 62.1% for the inorganic, 50:50, and organic spiked
samples also with high RPDs between replicates.
TABLE 3: COMPARISON OF BROMATE AND PERMANGANATE DIGESTION RECOVERIES
(100 NG/L INORGANIC, 50:50, AND ORGANIC MERCURY) IN ARTIFICIAL SEAWATER
% Recovery/(RPD)
Bromate Permanganate with Heat Permanganate no Heat
Inorganic 93.8 (0.43) 95.3 (6.7) 108.6 (4.2)
50:50 92.1 (2.3) 85.6 (5.1) 83.3 (12.9)
Organic 90.4 (0.11) 81.4 (18.7) 62.1 (5.6)
Several advantages were found using the bromate digestion. It provided better
accuracy and precision and increased sensitivity about the permanganate method.
The bromate method also is preferable because it requires less reagents and can
be performed at ambient temperature in a shorter amount of time. Specifically
for saline matrices, the bromate digestion does not require further additions of
533
-------
reagents or supplemental standard preparation to complete the digestion, thus
eliminating this time consuming procedure that is necessary for the permanganate
digestion. From the evaluation of these data it was decided to incorporate the
bromate digestion into the experimental development of the method.
INTERFERENCE STUDIES:
Studies of the affects of chloride on the recovery of mercury using the bromate
digestion and the permanganate digestion techniques were performed. It was
noticed upon heating that the samples lost the purple permanganate color and
formed a brown precipitate. This was not observed in the standards that were
heated. A noticeable difference in sensitivity was observed between the
digestion procedures. Slopes of calibration standard curves prepared with
bromate digestion ranged from 0.9 to 1.0 and a 100 ppt sample had a peak height
of about 100, whereas slopes for the permanganate digestion ranged from 0.5 to
0.6 and the peak height of the 100 ppt sample was about 60. The bromate
digestion appears to be superior to the permanganate digestion in sensitivity,
accuracy and precision and is faster and easier to perform.
Further interference studies (Table 4) were performed using the bromate digestion
only. Potential interferences from sample matrices containing fluoride, sulfate,
nitrate, phosphate, gold, silver, or a mixture of calcium, copper, lead,
manganese, barium and iron were evaluated. Recoveries for matrices of fluoride
up to 100 ppm and phosphate up to 1000 ppm were over 93%. The composite sample
containing 1 ppm of Cu, Pb, Mn, Ba, and 10 ppm of Ca and Fe had 100% recovery.
The samples used for preparing the sulfate and nitrate matrices were too high in
residual background mercury to analyze. The determination of mercury in a 1 ppm
gold matrix had a pronounced interference that was thought to be from a reaction
of mercury vapor combining with gold from the sample matrix in the gas/liquid
separator during analysis. A sample matrix containing silver at 1 ppm also
exhibited this effect, but to a lesser extreme. Potential interferences from
volatile organic compounds were also evaluated using the bromate digestion.
Recoveries were 100% for sample matrices containing up to 50 ppb chloroform or
500 ppb toluene.
534
-------
Table 4: Interference Study
Interferent
Fluoride
Sulfate1
Nitrate1
Phosphate
Gold2
Silver2
Composite:
Ca
Cu
Pb
Mn
Ba
Fe
Chloroform
Toluene
Level
SOppm
lOOppm
lOOppm
SOOppm
lOOppm
lOOppm
SOOppm
lOOOppm
Ippm
Ippm
lOppm
Ippm
Ippm
Ippm
Ippm
lOppm
5ppb
25ppb
SOppb
SOppb
2 SOppb
SOOppb
Recovery
98.2%
101.4%
115.4%
197.6%
151.5%
93.0%
97.0%
100 . 3%
76.1%
182.2%
101.6%
98.9%
102.1%
102 . 6%
100 . 9%
101.9%
99.1%
1 Sulfate and nitrate solutions obtained from an outside source contained
high levels of mercury.
2 Chemical Interference
ppm = mg/L ppb
535
-------
SINGLE LABORATORY VALIDATION:
The single lab validation study was completed. Precision and accuracy data were
generated and the linear range of the method was determined to be 25 Mg/L-
Recoveries for duplicate samples fortified at 25, 50 and 100 ng/L and precision
for six replicate samples fortified at 50 ng/L were determined for each of the
six aqueous matrices selected for the study. The Trade Effluents (T&E) primary
effluent matrix yielded lower recoveries than the other matrices and the
recoveries also had a higher standard deviation. Matrix interference was
suspected in this case, so a standard additions curve was constructed using this
matrix. The curve generated showed no interference from the matrix. Some
laboratories have claimed difficulty when using the bromate-bromide digestion
procedure for samples high in organics. Further studies with the T&E matrix are
planned using a modification of the digestion procedure similar to one for trade
effluents described by Yorkshire Water Authority in the United Kingdom. This
procedure uses more vigorous digestion conditions.
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:
METHOD DETECTION LIMITS FOR MERCURY
MATRIX ng/L
Reagent water 1.8
Florida marsh water 3.3
Synthetic seawater 2.6
The digestion procedures evaluated seem equivalent except sea water where the
bromide/bromate reagent tends to perform better than the permanganate procedure.
The statistical approach requires that the interferant studies should be resumed.
This will include field testing of many kinds of sample matrix and ruggedness
testing.
The gold amalgam accessory may be useful when dealing with certain kinds of
interferences (organics and inorganics) not yet identified. Interferences may
be eliminated by preconcentration of mercury on gold. Gas vapors containing
interferences are passed through the instrument before analyzing mercury. The
gold/mercury amalgam trap is heated to release the mercury for cold vapor atomic
detection analysis. The gold amalgam accessory would be inserted between the
mercury vapor generator and the fluorescence detector. The future of this
accessory will be evaluated during short-term and long-term stability studies.
These studies are needed for this accessory before its inclusion in an official
USEPA method for Total Mercury.
536
-------
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.
REFERENCES
1. Selikoff, I.J. (Editor-in-Chief); "Hazards of Mercury", Environmental
Research. An International Journal of Environmental Medicine and Environmental
Sciences, Vol. 4, No. 1, Mar. 1971.
2. Jacobs, M.B.; Yamaguchi, S.; Goldwater, L.J.; Gilbert, H.; Amer. Ind. Hyg.
Assoc. J., Vol. 21, 1960.
3. Hatch, W.R.; Ott, W.L.; "Determination of sub-microgram quantities or
mercury by atomic absorption spectrophotometry" Anal. Chem. Vol.40, pp.2085-
2087, 1968.
4. Kopp, J.F.; Longbottom, M.C.; Lobring, L.B.; " 'Cold Vapor' Method for
Determining Mercury"; Jo. AWWA, Vol. 64. No.l, Jan. 1972.
5. U.S. EPA Method for Chemical Analysis of Water and Wastes. EPA-600/4-79-
020, Revised March 1983, Method 245 (Total Mercury).
6. Fitzgerald, W.F.; Clarkson, T.W.; "Mercury and Monomethylmercury: Present
and Future Concerns", Environmental Health Perspectives, Vol. 96, pp. 159-166,
1991.
7. Slemr, F.; Langer E.; "Increase in global atmospheric concentrations of
mercury inferred from measurements over the Atlantic Ocean", Nature, vol. 355,
Jan. 30, 1992.
8. Swain, E.B., Engstrom, D.R.; Brigham, M.E.; Henning, T.A.; Breonik, P.L.;
"Increasing Rates of Atmospheric Mercury Deposition in Midcontinental North
America", Science, Vol 257, pp.784-787, Aug. 1992.
9. Sorensen, J.A.; Glass, G.E.; Schmidt, K.W.; Huber, J.K.; Rapp Jr., G.R.;
"Airborne Mercury Deposition and Watershed Characteristics in Relation to
Mercury Concentrations in Water, Sediments, Plankton, and Fish of Eighty
Northern Minnesota Lakes"; Enviro. Sci. Technol. Vol. 24, No. 11, 1990.
10. Stober, J.; Hicks, D.; "MERCURY CONTAMINATION IN THE EVERGLADES
ECOSYSTEM", A Plan of Study for USEPA Region IV, Atlanta, Ga.; Prepared by
Environmental Services Division, Ecological Support Branch, Athens, GA.,
Draft-Copy 1992.
537
-------
11. S.N. Deming, S.L. Morgan; "Experimental Design: A Chemometric Approach",
Elsevier, Amsterdam, 1987.
12. Yorkshire Water Methods of Analysis, 5th Ed. 1988.
(ISBN 0 905057 23 6)
13. Farey, B.J.; Nelson, L.A.; Rolfe M.G.; "A rapid technique for the
breakdown of organic mercury compounds in natural waters and effluents."
Analyst, 1978, 103. 656.
14. Nelson, L.A.; "Brominating solution for the preconcentration of mercury
from natural waters.", Anal. Chem., 1979, 51, 2289.
15. Stockwell P.B. and R.G. Godden, J. Anal. At. Spec, 1989, 4, 301.
538
-------
77 DETERMINATION OF TETRAETHYLLEAD
IN GROUND WATER
Dante J. Bencivengo, Bruce N. Colby and C. Lee Helms, Pacific Analytical, 6349 Paseo
Del Lago, Carlsbad, California 92009
Abstract: A chemical analysis method for determining tetraethyllead (TEL) in ground
water has been devised based on purge and trap GC/MS. It incorporates TEL-d^ as an
internal standard and naphthalene-d8 as a surrogate. The method detection limit (MDL)
has been calculated as 2.1 |ig/L and recoveries averaged 102 percent.
INTRODUCTION
Organic lead compounds constitute the largest single application of industrial
organometallic chemistry. Of those generated commercially, tetraethyllead (TEL), an
octane-boosting agent, has been produced in the largest quantity. DOT classifies TEL as
a Class B poison. TEL has a strong affinity for lipids and is known to affect the central
nervous system causing symptoms such as fatigue, ataxia, psychosis and convulsions.
Due to the quantities produced and its highly toxic nature, TEL is a potentially
significant threat to the environment.
At present there are no "official" TEL analysis methods. The most common approach to
inferring TEL levels, especially those presumably associated with leaking underground
gasoline tanks, is to determine "organic lead" via solvent extraction using xylene
followed by reaction iodine, then tri-capryl methyl ammonium chloride in methyl
isobutyl ketone (MIBK); finally, elemental lead is determined using flame AA. The
detection limit using this method is only about 100 (Xg/L and it is nonspecific.
Because tetraethyllead has a vapor pressure exceeding that of naphthalene, it is amenable
to analysis using a purge and trap gas chromatography based analysis system.
METHOD SUMMARY
A 25 mL aliquot of ground water is spiked to 50 (ig/L each with tetraethyllead-d20 and
naphthalene-dg. It is then subjected to GC/MS analysis using conditions equivalent to
those identified in Method 8260. The GC/MS is calibrated with standard solutions
containing 10, 20, 50, 100 and 200 jig/L of TEL and naphthalene-d8, plus 50 jig/L of
TEL-d2o- All quantitation and quality control are performed per Method 8260 except that
ao is used as the internal standard and naphthalene-dg is the surrogate.
EXPERIMENTAL
Data were acquired using an OI purge and trap interfaced to a Fisons MD800 GC/MS
using a 20-to-l split where the transfer line connects to the head of the GC column. A
60 m x 0.32 mm ID, 1.8 micron film DB-624 column was employed using a temperature
539
-------
o
program which included a 4 min initial hold at 40 °C, then a program of 8 °C/min to 200
°C and a final hold of 10 min. This arrangement provided good performance without the
added complexity of cryogenic trapping. All data reduction and report generation was
performed using the Lab-Base data system.
Neat TEL was obtained from Ethyl Corporation, TEL-d20 was obtained from Merck
Isotopes (Merck stock is now available from Cambridge Isotopes), and naphthalene-d8
was obtained from Cambridge Isotopes. An MDL study was carried out using seven
laboratory water samples spiked at 5 jig/L. Approximately 60 field samples from a site
suspected to involve TEL contamination were analyzed. Matrix and matrix spike
duplicates were all spiked to 50 }ig/L.
RESULTS AND DISCUSSION
Mass chromatograms showing the elution of TEL, TEL-d20 and naphthalene-d8 are given
in Figure 1 for a 50 ng/L standard solution. Peaks shapes are good and retention times
are reproducible. Mass spectra for TEL and TEL-d20 are shown in Figure 2. Masses
chosen of quantitation were 295 for TEL and 310 for TEL-d20.
In performing the initial analyses it was noted that both TEL and TEL-d20 were
apparently being lost over time as they sat in the autosampler. Initial suspicions were that
the TEL was decomposing. This was difficult to justify, however, because TEL could be
detected in ground water samples which most likely were many years old. Subsequent
review of the data suggested that ethyl groups were exchanging between labeled and
natural TEL. If a TEL/TEL-d20 spiked sample was allowed to stand for several hours,
spectra for TEL-dJ5i TEL-dio and TEL-d5 would start to appear. This exchange process
was not significant when samples were allowed to stand less than an hour prior to
initiating the purge cycle. Consequently, all samples were purged immediately after
spiking.
A typical calibration curve is shown in Figure 3. Response factors are constant to within
25% RSD, but a marked tendency for the response factor to increase with increasing
concentrations of TEL was noted. Consequently, calibration via linear regression is
recommended if accurate results are more important than strict adherence to SW-846
protocol.
Matrix spikes and matrix spike duplicates using field samples yielded an average
recovery of 102% and an average difference of 1.5% based on three sets of data. The
method detection limit study provided a calculated MDL of 2.1 (ig/L.
TEL was detected in field samples even when they were very highly contaminated. The
example shown in Figure 4 is for a well water sample with 32 jig/L of TEL. The primary
reason TEL can be measured interference-free in samples like these is due to the high
masses used for quantitation.
540
-------
Figure 3 - TETRAETHYLLEAD CALIBRATION CURVE
250
200 -
150 -
on
-------
SUMMARY
X
Purge and Trap GC/MS is a viable technique for determining tetraethyllead in ground
water. It can be both accurate and precise if care is taken to avoid ethyl group exchange
between labeled internal standard and natural material.
Figure 1 - TETRAETHYLLEAD STANDARD
Pacific Analytical
Sample: USTD885
Instrument: HP8BB
7K05A8?
X.FS-
a
100
0
100
X.FS
Son
26,06
TEL-d2B
26.63
TEL
30,53
NaphthaIene-d8
2608
2800
38O0
3200
33792
24!
8888
832
X
3400
Figure 2 - TETRAETHYLLEAD MASS SPECTRA
Pacific Analytical
Samp
7K8£
ioa-
X.FS
•«xx:
iaa-
XFS-
0-
lo: VSTDBBS
;A87'2&46 (26.852) REFINE
TEL-d28
21
206
« 153 |
i« 188 130 ' 200
TEL
21
206
a 100 i ^a **aa
J* *y^ 13S 2QQ
,8 24fl
^18
2:
233
18
209
Instrument : HD8B8
12 35328
3 O
308
243 276 306 ?"•
i' |L -281 "I (
258 naa ' T;Q
7 22016
295
293
266 396
264-1-267 /
— 239 300 358
542
-------
78 USE OF A TELEPHONE SERVICE AND DATABASE TO PROVIDE GUIDANCE AND
INCREASE PUBLIC INVOLVEMENT IN OSW METHODS
D. Anderson. R. Carlston, and S. Hartwell, Science Applications International Corporation,
7600A Leesburg Pike, Falls Church, Virginia 22043, and G. Hansen and K. Kirkland, U.S.
Environmental Protection Agency, OSW Methods Section, 401 M Street S.W., Washington
DC 20460.
ABSTRACT
The Methods Section of EPA's Office of Solid Waste (OSW) is responsible for
approving and issuing methods used to evaluate solid waste and determine whether the
waste is hazardous under RCRA. These methods are contained in EPA's methods manual
'Test Methods for Evaluating Solid Waste, Physical/Chemical Methods", also known as SW-
846. Individuals in the Methods Section have specific areas of responsibility for these
methods, including organic, inorganic, miscellaneous, and characteristic test methods. Each
member receives many inquiries and comments regarding method status, selection,
performance, and detection capabilities. The number of inquiries (particularly phone calls)
is often overwhelming and, in order to fully answer a question, more than one member of
the Methods Section is frequently involved.
The Methods Section, therefore, identified a need to more efficiently respond to
inquiries and to document the transmission of the information for future reference. In order
to accomplish this task, the OSW Methods Section established a telephone service known
as the Methods Information Communication Exchange (MICE) Service. Using a voice-mail
system, MICE is set up to contain informational messages of the most commonly asked
methods questions and record questions and comments from callers regarding technical
difficulties or issues regarding the use of the SW-846 manual. After compiling the incoming
messages for one day, the calls are sorted and distributed to appropriate technical staff
members for a telephone response. Each response is summarized, categorized, and entered
into a computer database on a periodic basis so that internal reports can be sent to the
OSW Methods Section.
Once categorized and entered into the database, the resulting call responses and any
issues raised are used internally at the Agency to document typographical errors in the
manual and recurring technical problems with a particular method. The data are also used
to record ideas and comments for possible directions of future analytical research and
manual expansion. The MICE Service voice-mail system has received up to 821 calls and
responded to 349 individuals requesting information in one 4 week period.
543
-------
EXTRACTION OF METAL IONS FROM SOLID MATRICES BY COMPLEXATION SFE
W. F. Beckert. U.S. Environmental Protection Agency, EMSL-LV, Las Vegas, Nevada 89119 and
Y. Liu, V. Lopez-Avila, and M. Alcaraz, Midwest Research Institute, California Operations,
Mountain View, California 94043.
ABSTRACT
Complexation combined with solvent extraction has been widely used to extract metal ions from
various matrices. Despite its good performance, this technique is usually time and labor intensive.
Since we have been working on SFE techniques for other classes of compounds, we undertook s
feasibility study to determine whether SFE could be used to extract metal ions from solid matrices.
We knew from work published by Wai and coworkers that metal complexes formed with fluorinated
complexing agents such as bis(trifluoroethyl)dithiocarbamate (FDDC) are soluble in supercritical
carbon dioxide; thus, we were expecting that these complexes could be extracted by SFE.
Experiments were performed to derivaiize the metal ions with FDDC, extract the complexes by SFE,
and then analyze the extracts by gas chromatography with atomic emission detection. Data will be
presented on the extraction of C\? +, Cd2+, Zn2+, and Co2"1" from spiked filter paper and spiked
soil by complexation SFE and then on the analysis of the complexes by gas chromatography with
atomic emission detection.
544
-------
80 MICROWAVE SAMPLE PREPARATION FOR MERCURY ANALYSIS VIA
COLD VAPOR ATOMIC ABSORPTION SPECTROMETRY
W.G. Engelhart. S.E. Littau and E.T. Hasty,
GEM Corporation, Matthews, NC 28105
ABSTRACT
Mercury contamination in the environment has become an increasing concern due
to the toxicity of this element. The volatility of elemental mercury and
organomercury compounds requires precautions be taken to avoid losses during
sample preparation.
Microwave acid digestion of environmental samples inside sealed vessels is an
EPA approved method for metals analysis by AA and ICP spectroscopy. This study
will describe microwave digestion methodology required for mercury analysis via
cold vapor AAS.
Analytical results obtained following microwave digestion of standard reference
materials and real world samples spiked with organomercury standards will be
presented. Microwave sample preparation methods for SRM 1575 Pine Needles,
US EPA Fish, US EPA Dried Sludge, effluent municipal wastewater, waste motor
oil and mixed solvent fuel will be described.
545
-------
STUDY OF EPA METHOD 300.0:
APPLICATION TO HAZARDOUS WASTE ANALYSIS
Harish Mehra, Chemical Waste Management, Inc., P.O. Box 4249,
Modesto, CA 95352 Tele: 209 491 7511
Douglas T. Gjerde, Sarasep, Inc., 1600 Wyatt Dr., Suite 10, Santa
Clara, CA 95054, Tele: 408 492 1029
ABSTRACT:
One of the first rules of using chemical analysis in the chemical
waste industry is that standard analytical methodology must be
employed. The second rule is that even if a standard method is
used, it must be verified. Analytical data are used to determine
proper disposal, secure information that may then be used in
litigation, etc. So it is important to get the right answer and
be able to prove it.
This work is a study of EPA Method 300.0 ion chromatography
method for anions. The method has been certified under ASTM
method D4327. Two commercial columns were studied for
selectivity accuracy and precision. The effectiveness of EPA
Method 300.0 was studied using standard reference materials.
Several samples of hazardous waste were examined. Of particular
interest are halogenated organic solvent sample containing high
amounts of chloride, fluoride, and bromide. Prior to anion
determination, these sample were digested using a combustion
bomb. Residues were taken up with an aqueous bicarbonate medium
and the samples analyzed. Halide values are used in determining
the feasibility of hazardous waste disposal and recycling.
Furthermore, the cost of disposal is directly dependent upon the
halogen concentration of the waste. Bromide containing waste are
more expensive to incinerate than chloride containing waste.
546
-------
INTRODUCTION:
Characterization of a waste as hazardous is performed by
regulatory agencies, including EPA, OSHA, and DOT. Generally
hazardous waste is either corrosive, reactive, toxic, or
flammable. Once a hazardous waste is generated, it must be
properly managed. Proper management includes the cost effective,
safe, and legal disposal of the waste. However, to accomplish
this, the content of the waste must be known accurately- For
example, whether a waste is incinerated and how it is incinerated
depends on the individual concentrations and total concentration
of the halogens in the waste.
Because of this and of possible litigation considerations, it is
important to use standard analytical methodology. But even if
standard analytical methodology is used, it is still important to
be able to show that accurate results are being achieved.
Accurate results are gnerally obtained with standard methods such
as EPA method 300.
Ion chromatography is widely used for the analysis of
environmental samples (Ref. 1-3). The most common method for
anion determinations in ion chromatography is EPA Method 300.0.
The method has been certified as ASTM Method D4327. The method
is based on ion exchange chromatography and therefore, is largely
sample independent and rugged. The ion exchange column acts as a
buffer to the sample as it is injected. Thus, retention times of
sample peaks are much more constant (matrix independent) than
other forms of chromatography when a variety of sample are
injected.
This work examines the accuracy and precision of the method for
organic solvent digests. Two different analytical columns are
compared.
EXPERIMENTAL:
The ion chromatograph used in this work was a Dionex 4000i
equipped with a AMMS™ suppressor and conductivity detection.
The regenerant was 25 mM sulfuric acid at 2.8 mL/min flow. A 25
uL loop was used. Except for a standard digest material used for
column comparison, the normal Method 300.0 eluant concentration
and flow rate was decreased to increase resolution of the anions:
The eluant concentration was 0.9 mM Na2C03 / 0.85 mM NaHC03 and
the flow rate was 1.5 mL/min. All reagents were analytical or
reagent grade.
Two different analytical columns were used in this study: Dionex
HPIC AS4A™, 250 x 4 mm and Sarasep AN300™ 100 x 7.5 mm. The two
columns are designed to operate under identical eluant
concentrations and flow rates. No changes were made in the
instrumentation. Each column was equipped with a 5 cm guard
column available from their respective companies.
547
-------
All samples and standards were prepared with bomb calorimeter
(Ref. 4) digestion (PARR Model 1261 calorimeter). Approximately
0.5 g of a representative (homogenized) sample was accurately
weighed into a capsule. Some samples are not highly combustible.
If necessary for combustion, 0.2 - 0.5 g of halogen-free kerosene
was added to the capsule. All weights were recorded to 3 decimal
places. To a clean bomb 10 mL of 5% aqueous NaHC03 was added,
coating the internal surface of the bomb. The capsule was placed
in the bomb seat. A Ni alloy wire was tied to the ignition
electrodes. If the sample was liquid, the wire was placed close
the sample, and if solid, the wire touched the sample. The PARR
cover was screwed tight and about 30 psi of oxygen was charged to
the bomb. After combustion, the bomb was shaken. The solution
and 3 rinses were made up to 50 mL. The sample was filtered 0.45
micron filter prior to injection into the ion chromatograph.
RESULTS AND DISCUSSION:
Digest of sodium salt standards. The selectivity of two
commercial columns were compared under identical separation
conditions specified by method 300.0. A liquid standard
containing 10 - 50 ppm of sodium salts of fluoride, chloride,
bromide, nitrate, and sulfate were digested and injected into the
chromatograph. The results are shown in Figures 1 and 2.
Figure 1 shows that for the AS4A column, fluoride is non-retained
and elutes in the water dip. Other materials such as non-ionic
organics and cation are also non-retained and elutes with
fluoride. Counter ions are suppressed by the AMMS and will not
affect quantification of fluoride provided the capacity is not
exceeded. No study has been published on the effect of non-ionic
organics on quantification.
Figure 2 shows exactly the same separation on an AN300 column.
The selectivity is identical expect that the peak window is
retained. Non-retained material is elutes well before the
fluoride peak. Normally a water dip would elute first, but in
this case the digest contains material that gives a positive
peak. A unknown positive peak eluting with fluoride will give a
false positive result. It is unknown what causes the positive
peak or whether it is present in all digests.
Digests of Standard Reference Materials. Several organic
standard reference materials (SRM) were obtained from the
National Institute of Standards and Technology (NIST) and
Environments Resource Associates (ERA) (see Table 1) . The
samples contains elemental bromine, fluorine, chlorine, and
sulfur; however ions generated in the digest were bromide
fluoride, chloride, and sulfate.
The results of the ion chromatography analysis are shown in Table
2. The data are a compilation of the injection of individual
SRMs listed in Table 1. The data show that the two commercial
column gave virtually identical results. The concentrations
obtained were all within the accepted range of elemental
548
-------
concentrations reported for the SRMs.
Note the precision of the method, measured over a eight day
period, was < 5%.
Digests of Hazardous Waste Samples. Three actual hazardous waste
samples were obtained from generator companies. When the sample
is received, the generator also provides a profile of the
elemental composition. The types sample received can range from
pesticides, oil, organic flux, The samples were digested and
analyzed by ion chromatography using the two columns. The
results, reported as elemental wt/wt percent, are shown in Table
3. All of the samples were also spiked with a known
concentration of each ion. The recovery of the standard spikes
ranged from 90 - 113%. While the precision of the analyses were
< 5%.
The peak was sharper and higher for bromide with the AS4A column.
Therefore, under identical conditions, limit of quantification
(LOQ), was 0.2% for the AS4A column and was 0.5% for the AN300
column. On the other hand, the AN300 column produced sharper
peaks for fluoride (LOQ 0.2%) and chloride (LOQ 0.2%).
All other results were comparable for the two columns.
Table 1. SRM for Ion Chromatography.
Digest Ion Generated SRM
Fluoride Spex Lot tt W5-378
Bromide NIST #3184
Sulfate NIST #2684A
Chloride ERA #03042
549
-------
Table 2.
ANALYTICAL RESULTS OF STANDARD REFERENCE MATERIALS
ON TWO DIFFERENT COLUMNS
Sample ANIONS IN PPM
Standard
Reference
Materials
(NIST) from
Environment
Resource
Associates
cr
AS4A
255
246
AN300
236
246
Br~
AS4A
1012
—
AN300
1012
—
S042'
AS4A
8.41
8.20
AN300
8.32
8.32
F
AS4A
1.57
1.70
-
AN300
1.35
1.51
Table 3. ANALYTICAL 1C RESULTS ON TWO DIFFERENT COLUMNS
Sample ANIONS IN PERCENTAGE
Rinse water
containing
oil and DDT
% Error
DY-90
% Error
Organic Flux
% Error
Cl
AS4A
<0.5
*94%
—
<0.5
*92%
—
2.03
•"106%
2.3
AN300
0.38
*96%
3.8
0.437
*94%
—
2. 18
*112%
2.6
Br'
AS4A
<0.2
*107%
—
<0.2
*108%
—
<0.2
*108%
—
AN300
<0.4
*108%
—
<0.4
*109%
—
<0.4
*111%
—
F-
AS4A
<0.6
*97%
—
<0.6
*99%
—
<0.6
*102%
—
AN300
<0.2
*99%
—
<0.2
*105%
—
<0.2
*109%
—
SO <2~
AS4A
2.89
*95%
3.2
<0.6
*92%
—
<0.6
*113%
—
AN300
2.79
*94%
3.9
<0.6
*95%
—
<0.6
*108%
—
% Accuarcy
550
-------
ACKNOWLEDGEMENTS:
The authors would like to express their appreciation to Dr.
Nilesh Shah, Ghee Lee, and Gayle King for their input in this
paper.
REFERENCES:
1. D. T. Gjerde and J. S. Fritz, "Ion Chromatography" 2nd ed.,
Huethig: Heidelberg and New York (1987).
2. R. J- Joyce and A. Schein, "1C: A powerful analytical
Technique for environmental laboratories" Am. Lab., 11 (1989) 46,
3. W. T. Frankenberger, Jr., H. C. Mehra, and D. T. Gjerde, J.
Chromatogr. 504 (1990) 211.
4. ASTM D240-87.
551
-------
Cl
in
01
DIGEST STANDARD
Dionex AS4A
Br
SO,
Fig. 1
8 minutes
-------
Ul
U1
w
DIGEST STANDARD
Sarasep AN300 ~* n
Cl
SO4
10 minutes
Fig. 2
-------
The Effects of Size Reduction Techniques on TCLP analysis of
Solidified Mixed Wastes
R. D. Thiel, Research Engineer, EG&G Rocky Flats, Inc. P. 0. Box 464, Golden, CO
80401.
The Rocky Flats Plant (RFP) is presently engaged in research and
development activities for the stabilization of mixed wastes. Two
stabilization technologies (microwave melting and polymer
encapsulation) are being developed to process various waste forms
produced at RFP. At the Nevada Test Site storage facility one of
the acceptance criteria for low level mixed waste is passing the
Toxicity Characteristic Leaching Procedure (TCLP).
The microwave melting and polymer encapsulation processes rely, at
least partially, on a physical separation of toxic metals from the
environment via formation of a monolith containing the mixed waste.
The most important step, but least specified, in the TCLP analysis
of these types of waste forms is size reduction.
This paper is a study of the effects of several different
approaches for size reduction of the products of the above
mentioned waste stabilization processes. Inductively Coupled
Plasma Atomic Emission analysis results for several TCLP metals are
presented to prove how the size reduction step drastically affects
the TCLP analysis results. Also presented are results that show
how the three stabilization techniques respond differently to the
size reduction step of the TCLP analysis.
554
-------
AIR AND GROUNDWATER
-------
CANISTER ANALYSIS BY USING
GAS CHROMATOGRAPHY/fflGH RESOLUTION MASS SPECTROMETER
Jong-Pvng Hsu. Director, Greg Miller and Joseph C. Pan,
Department of Environmental Chemistry
Southwest Research Institute
6220 Culebra Road
San Antonio, Texas 78228-0510
Summa*-passivated canisters used for the collection of air samples have distinct advantages
over tenex and charcoal tubes since the whole air sample is collected with long storage stabilities and
allows for multiple and flexible analysis schemes to be used.
When very low detection limits are required, pressurization of the canister is generally
required which must use specialized and expensive sampling devices which will pressurize the
canister above atmospheric pressure. A simpler approach and the preferred one is to do the sampling
in evacuated canisters using inexpensive flow controllers and sample up to atmospheric pressure.
Clean humidified air is when added in the laboratory to bring the pressure above atmosphere and the
sample is when concentrated cryogenically before analysis. Addition of humidified air dilutes the
sample and raises the detection limits and therefore places sensitivity demands on the instrumentation
when low detection limits are required.
A gas chromatography/high resolution mass spectrometer has always offered better sensitivity
and selectivity when low resolution quadrople instruments. Operating at resolution of 3000 in full
scan-mode has quite a number of distinct advantages in highly contaminated air samples since due
to the high resolutions used, it is possible to detect target analysis even in the presence of high
background contaminants. We have found that several modifications are required in sample
introduction to prevent high voltage flashovers to occur. The sensitivity and selectivity of the
approach is indeed improved and the results will be presented.
555
-------
AN EFFECTIVE MOISTURE CONTROL SOLUTION FOR THE GC/MS
DETERMINATION OF VOCS IN CANISTER SAMPLES COLLECTED
AT SUPERFUND SITES
Michael G. Winslow. Manager, Organic Analytical Division, Dwight F. Roberts, Manager,
GC/MS Department, and Michael E. Keller, Supervisor, Air Toxics Group,
Environmental Science & Engineering, Inc. (ESE), P.O. Box 1703, Gainesville, FL
32602.
ABSTRACT
EPA Compendium Method TO-141 has become the preferred guidance method for the
sampling and analysis of volatile organic compounds (VOCs) in ambient air samples . A
Contract Laboratory Program (CLP) protocol2 for the analysis of air samples at
Superfund sites, which is based on TO-14, has recently been drafted by the USEPA .
Like TO-14, the CLP draft protocol recommends cryogenic preconcentration of the air
sample followed by thermal desorption into a chromatographic system. Because
moisture in the air sample is cryogenically preconcentrated along with the VOCs, it will
also delivered into the chromatographic system during themal desorption. The moisture
can adversely affect chromatography and/or detector sensitivity and reliability.
Therefore, unless the moisture in the sample is removed or significantly reduced,
sample size must be limited with consequent limitations to method sensitivity.
Compendium Method TO-14 and the proposed CLP analytical protocol recommend a
Nafion dryer for moisture removal. However, this technique limits the method to non-
polar analytes and introduces potential carryover and reproducibility problems.
Because many polar VOCs are of great interest to the EPA, a canister analysis procedure
that allows these analytes to be determined with acceptable precision and accuracy
should be preferred over an approach that eliminates that possibility. This paper
decscribes a commercially available system that effectively controls moisture in large
volume air samples, thereby allowing both non-polar and polar VOCs to be measured at
very low concentration levels with acceptable precision and accuracy.
INTRODUCTION
In response to CERCLA and SARA mandates for assessment of potential air emissions and
air quality impacts of toxic contaminants before and during Superfund site remediations,
the USEPA's Analytical Operations Branch (AOB) within the Hazardous Site Evaluation
Division (HSED) of athe Office of Emergency and Remedial Response (OERR) has drafted
a CLP SOW for the analysis of air samples collected at Superfund sites. One of the four
analytical protocols which has been drafted for the proposed CLP air program involves
the determination of VOCs in whole air samples collected in passivated stainless-steel
canisters. The CLP draft protocol is based on Compendium Method TO-14 and expands
the target compound list from 41 to 61 analytes, including several polar VOCs. Air
samples are collected in 6-L canisters, 500 mL of which is preconcentrated by
collection on a cryogenically cooled trap, revolatilized and backflushed onto a capillary
GC column for separation, and quantitated with a mass spectrometer operated in the
full-scan mode. The full-scan mode is required in order to tentatively identify non-
target or unknown responses.
556
-------
The CLP draft protocol recommends removal of excessive moisture from the sample gas
stream with a Nafion® dryer, if applicable. Excessive moisture can cause the cryotrap
or GC capillary column head to plug with ice, degrade chromatographic resolution,
especially of the early eluting gases, and cause MS source overpressure. However, use
of a Nafion® dryer as a moisture reduction technique limits quantitative determinations
to non-polar VOCs only, because polar VOC compounds are coincidentally removed, or
partially removed, from the sample stream. Because the protocol requires a 500 mL
sample to be analyzed, a Nafion® dryer will have to be employed for most environmental
air samples. Analysis of sample volumes greater than 100 ml generally requires
moisture reduction. Unfortunately the proposed target compound list does contain
several polar VOCs. This apparent conflict is not clearly addressed by the protocol.
This paper presents results of a preliminary evaluation of an Entech Model 2000
preconcentrator interfaced to a Hewlett-Packard 5890 gas chromatograph
(GC)/Finnigan Incos 50 mass spectrometer (MS). The analytical system can reliably
measure VOC concentration levels as low as 0.1 ppbV while avoiding the typical
problems caused by the introduction of excessive moisture into the analytical system. At
the same time, the preconcentrator system allows both non-polar and polar VOCs to be
determined for target and unknown compounds alike. The components and operating
configurations of the system are described and the results of system performance and
method detection limit (MDL) studies are also presented.
EXPERIMENTAL
Canister Cleaning and Leak Checking
To determine ambient air VOC concentrations with accepable precision and accuracy at
ppbV and sub-ppbV concentration levels, it is imperative that the laboratory establish
the cleanliness and integrity of each sampling canister. ESE's laboratory in Gainesville,
Florida, utilizes an Entech Model 3000 Canister Conditioning System for cleaning and
leak checking canisters prior to sampling and analysis.
Canister leak checking is performed prior to canister cleaning. The canisters are
connected to the system manifold via 1/4-inch stainless-steel tubing and the canister
valves are left closed. A high vacuum pump is engaged and the vacuum pressure is
monitored using an in-line pirani gauge (0-2000 mtorr sensor). If a significant leak
exists, it will manifest itself by not allowing the system to quickly evacuate to
approximately 0.3 torr, and/or the system pressure will quickly rise once the high
vacuum pump is disengaged. Quick disconnects are situated between each canister and the
manifold to allow easy isolation of individual canisters which may have leaking valves.
After the repetitive fill/evacuate cycling is completed, the high vacuum pump is used for
final canister evacuation. Leaks will not allow the canisters to achieve a vacuum reading
of approximately 0.5 torr.
Four to eight 6-liter canisters are cleaned simultaneously over a 4-hour period through
automatic, unattended cycling between canister filling and evacuation modes. Each of the
canisters is heated to 100°C for the duration of the cleaning process. A rough-vacuum
oilless diaphragm pump cycles between eight 8-minute fillings and eight 8-minute
evacuations. Nitrogen vent gas from a liquid nitrogen tank is used for the fill gas and a
humidification chamber containing ASTM Type II HPLC grade water is used to add
557
-------
moisture to the fill gas in order to assist in the displacement of VOCs from the interior
surface of the canisters. In the fill mode the canisters are pressurized to about 25 psig.
After completion of eight cycles with the low-vacuum pump, the canisters are evacuted
with a high-vacuum oil-based pump for about 75 minutes to an absolute pressure of 0.5
torr, which is measured by an in-line pirani gauge. A cryogenic trap containing liquid
nitrogen is placed between the pump and the system controller to keep oil vapor out of
the system.
At the completion of the cleaning cycle, the canisters are pressurized to 30 psig with
humidified nitrogen and analyzed by GC/MS. No target analyte should be detectable at or
above the lowest calibration standard (0.1 ppbV for this study).. After analysis, each
canister is re-evacuated to 0.5 torr.
Standards Preparation
Working calibration standards were prepared in canisters at a minimum of five
concentration levels in the range of 0.1 - 15 ppbV. These working standards were
prepared from stock calibration mixtures containing approximately 100 ppbV of each of
the target analytes in dry nitrogen in high pressure (2000 psi) cylinders. These
calibration mixtures were purchased from Scott Specialty Gases and are certified to ± 5
% for each of the VOCs.
The working calibration standards were generated by dynamic flow dilutions of the
purchased mixtures with cleaned, zero grade nitrogen humidified with ASTM Type II
HPLC-grade water. ESE utilizes an Entech Model 4560 Dynamic Dilution System for
this process. This system utilizes up to six mass flow controllers for simultaneous
blending from multiple cylinders. Canisters are filled to 15-25 psig. Newly prepared
calibration standards are allowed to equilibrate at least 24 hours.
The 40 compounds listed in Compendium Method TO-14 (ethyl toluene was not included)
were evaluated for this paper.
Instrumental Analysis
Canister samples are analyzed with an Entech Model 2000 automated VOC cryogenic
preconcentrator equipped with an Entech 2016 16-position autosampler manifold and
interfaced to a HP 5890 GC/ Finnigan INCOS 50 MS operated in the full scan mode (m/e
35 to 270 amu). Since the preconcentrator system is under software control (IBM
compatible), a QA/QC report is a standard feature for documentation of actual run
conditions.
Canisters are attached to the manifold with 1/8 inch stainless-steel tubing. A leak check
of the system is then performed. Sample flow is set under mass flow control to 150
mL/min. to give an integrated total volume of 1000 ml_. Both subambient and
pressurized samples can be analyzed because the system employs a mechanical pump to
draw samples across the preconcentor. The internal standards, bromochloromethane,
1,4-difluorobenzene, and d5-chlorobenzene, are also added to each analysis at 150
mL/min. to a total volume of 100 ml. All rotary valves and transfer lines exposed to the
sample matrix are heated at 70-100°C.
558
-------
The Entech preconcentrator is configured with a 3-stage trapping approach which allows
a large sample volume to be concentrated and the moisture and carbon dioxide removed
without significant loss of non-polar or polar VOCs. Entech refers to the technique as
microscale purge and trap because it is based on classical purge and trap principles.
(15-20 u.L of water is purged with 30-40 ml of nitrogen instead of 5000 jiL being
purged with 400 mL of nitrogen.) The first stage is a high-volume cryogenic trap
(1/8-inch nickel tube containing glass beads) cooled during sampling to -150 °C with
liquid nitrogen. After cryotraping 1000 mL of air sample, the first stage trap is heated
rapidly to room temperature and slowly purged with about mL of nitrogen to transfer
the trapped VOCs to a second stage trap (1/8-inch nickel tube) containing hydrophobic
Tenax TA held at 0°C. Only the amount of water vapor that can saturate 30-40 mL of
nitrogen (.7 - .9 jiL) will be delivered to the second stage Tenax trap. Carbon dioxide
and some the water will not be trapped by the Tenax. The second stage trap is then heated
to 170°C and backflushed with helium to a third stage.megabore focusing trap cooled
with liquid nitrogen to -150°C. After heating and backflushing of the second stage trap
is complete, the third stage trap is then very rapidly heated to above 100°C to allow a
rapid injection of the VOCs onto the GC analytical column.
The target analytes are separated on a DB-1 fused silica capillary column, 75 m x 0.32
mm I.D., with a 1 micron film thickness. The chromatographic run is started at 35°C
and held for 5 min. A subambient starting temperature is not required because of the
rapid transfer of VOCs from the preconcentrator's third stage megabore focusing trap of
the Entech preconcentrator to the head of the GC column. Sharp peaks with 2-5 sec.
widths are maintained, even for early eluting compounds. The column is then
temperature programmed at 6°C/min. to 180°C and then at 7.5°C/min. to 225°C. The
scan rate is approximately 3 scans/sec. The total run time for preconcentration of the
sample and chromatographic separation is approximately 60 min. Figure 1 shows a
mass chromaogram of a 1 ppbV standard of target VOCs.
The MS scan rate is approximately 3 scans/sec. The Incos 50 turbomolecular source
pump draws about 170 L/min. This pumping rate effectively prevents source pressure
increases from any residual sample moisture.
System Performance and Method Detection Limit (MDL) Studies
The reproducibility and linerarity of the analytical system were evaluated over a two-
week period. Eight calibration standards and a blank were prepared as described above
in separate 6-liter canisters. The blank canister was prepared with humidified zero
grade nitrogen. The calibration standards contained the 40 compounds listed in Method
TO-14 and were prepared at concentration levels of 0.1, 0.25, 0.5, 1, 2, 5, 10, and 15
ppbV. Another canister containing the three internal standards was also prepared. All
canisters were attached to the autosampler manifold and analyzed, after instrument
tuning, as described above. Internal standards were added to each analysis at 5 ppbV.
The canisters were analyzed three times, with a week between analytical runs. Each
analytical sequence lasted about 9 hours after tuning. Initial canister pressures ranged
from 10 to15 psig for the first day's run, from 1 to 2 psig for the second day's run, and
from -1 to -2 psig for the last day's run.
In addition to the reproducibility study described above, an MDL study was performed.
Lower limits of detection for the target compounds were estimated for each instrument
559
-------
by determining the method detection limits (MDLs) as specified by the U.S.
Environmental Protection Agency (U.S. EPA).3 Nine canisters samples were prepared at
about 0.5 ppbV for each of the target compounds used the standards preparation system
described above. The nine replicate samples were then analyzed and quantitated against
a calibration curve containg eight standards ranging from 0.1 to 15 ppbV. The compound
responses divided by the appropriate internal standard response were plotted against
concentration. The MDL for each target compound is calculated by multiplying the
standard deviation of the seven replicate concentration measurements by the appropriate
one-sided t-value corresponding to n - 1 (8) degress of freedom. The corresponding t-
value for seven measurements is 2.897.
RESULTS AND DISCUSSION
The reproducibility, or precision, of the analytical system was evaluated by comparing
the responses of the target compounds and their corresponding internal standards over
the three analysis days at different calibration standard concentrations. In addition
compound retention time variability was evaluated. Table 1 summarizes response and
retention data for vinyl chloride, benzene, and hexachlorobutadiene (HCBD) at 0.1, 1
and 10 ppbV. These compounds represent low boiling (vinyl chloride = -13.4°C),
middle boiling (benzene = 80.1°C, and high boiling (HCBD = 215°C) VOCs, respectively.
For each of the three compounds listed in Table 1, the relative stardard deviations
(RSDs) of the absolute responses for the three analysis days over the three
concentration levels averaged less than 6 percent . The relative response factors
(RRFs) averaged less than 5 percent. The RSDs of the absolute responses of each of the
internal standards for the three analysis days and within each analysis day over the
three concentration levels averaged less than 5 percent. Retention times had ranges no
greater than .03 minutes for the three compounds. Response and retention time
reproducibility for the other target VOCs were also very good.
Table 2 summarizes calibration data for vinyl chloride, benzene, and HCBD on the second
analysis day. The compound responses were very linear over the 0.1 to 2.0 ppbV range,
as measured by either a linear regression curve or the %RSD of the RRFs.
Table 3 lists the calculated MDLs for this study. The calculated MDLs averaged 0.11
ppbV for all the TO-14 compounds. Seven of the calculated MDLs were less than 10% of
the 0.5 ppbV nominal concentration analyzed. This indicates that the replicate analysis
for those compounds should have been targeted at a lower concentration (probably 0.25
ppbV). The 0.1 ppbV responses for all the TO-14 compounds were well above
background noise, in the range of 10 - 80,000 area counts. Compound responses were
very linear over the 0.1-2 ppbV range and generally approached a quadratic fit across
the full calibration range (0.1 - 15 ppbV).
Polar VOCs
Although this paper does not discuss experimental results for polar VOC determinations
using the analytical system described above, preliminary analysis of eight polar VOCs of
interest have been good. The compounds that are currently being investigated are those
listed in the CLP draft protocol target compound list. They include acetone, acetonitrile,
acrolein, acrylonitrile, methyl ethyl ketone, methyl isobutyl ketone, methyl
methacrylate, and vinyl acetate. Other polar analytes will also be examined.
560
-------
Several limited studies have been performed elsewhere that have examined the stability
of selected polar compounds in canisters. There is sufficient evidence to indicate that
some polar VOCs are unstable in canisters if held over extended periods due to canister
and/or matrix effects. However, considerable experimentation is still needed,
especially at low concentration levels.
SUMMARY
Routine measurement of non-polar and polar ambient VOCs at ppbV and sub-ppbV levels
by GC/MS-SCAN can be acheived with acceptable precision and accuracy if sample
moisture is effectively reduced. This has been demonstrated with a analytical system
comprised of an Entech preconcentrator, a Hewlett-Packard GC, and a Finnigan MS.
ACKNOWLEDGEMENTS
The authors thank Norm Staubly for his laboratory assistance.
REFERENCES
1. W.T. Winberry, Jr., NT. Murphy and R. M. Riggin, "Method TO-14,"
Compendium of Methods for The Determination of Toxic Organic Compounds in
Ambient Air. EPA-600/4-89-017. U.S. Environmental Protection Agency,
Research Triangle Park, NC, 1988.
2. Statement-of-Work for the Analysis of Air Toxics at Superfund Sites (Draft),
"Volatile Organics Analysis of Ambient Air in Canisters," U.S. Environmental
Protection Agency, Washington, D. C., December, 1991.
3. Federal Register. "Definition and procedure for the detemination of the method
detection limit," Code of Federal Regulations. Part 136, Appendix B, Oct. 26
(1984).
4. M. G. Winslow, D.F. Roberts, and M.E. Keller, "Determination of VOCs in Ambient
Air at 0.1 ppbV for the Clean Air Status and Trends Network (CASTNet)",
presented in May 1993 at the EPA/AWMA Internation Symposium on the
Measurement of Toxic and Related Air Pollutants held in Durham, N.C.
561
-------
Table 1: Area Response and Retention Time Reproducibility of
Selected VOCS
10 ppbV
Resp. (xlO3)
IS Resp. (x103)
RRF
Ret. Time
t ppbV
Resp. (x1Q3)
IS Resp. (x103)
RRF
Ret. Time
0.1 opbV
Resp. (xlO3)
IS Resp. (x1Q3)
RRF
Ret. Time
P
vc
2058
764
1.35
4.03
278
808
1.72
4.03
29.7
739
2.01
4.02
ressuriz
2/17/93
Bzn
2379
2078
0.57
10.16
302
2143
0.70
10.15
27.1
1968
0.69
10.15
ed
I
HC8D
883
1665
0.27
28.11
124
1652
0.38
28.11
13.5
1479
0.46
28.10
Atmospheric
2/23/93
vc
2067
702
1.47
4.02
271
762
1.77
4.04
27.9
724
1.93
4.02
Bzn
2256
1947
0.58
10.15
279
2000
0.70
10.16
26.1
1894
0.63
10.14
HC8D
910
1610
0.28
28.11
111
1577
0.35
28.11
13.8
1440
0.48
28.09
Sub-Atmospheric
3/04/93
vc
1996
732
1.36
4.03
290
788
1.84
4.03
31.3
748
2.09
4.02
Bzn
2195
1978
0.56
10.16
300
2094
0.72
10.16
29.8
1985
0.75
10.15
HCBD
843
1582
0.27
28.12
106
1614
0.33
28.11
14.8
1551
0.48
28.11
562
-------
Table 2: Calibration and MDL Results for Selected VOCs
Calibration:
Standard VINYL CHLORIDE
(ppbV) Response RRF
2 515511 1.67
1 270598 1.77
0.5 141763 1.93
0.25 74681 2.03
0.1 27932 1.89
0 0
LR. Corr. Coeff. = 0.9995
y-lntercept = 6895
Average RRF = 1.86
%RSD Of RRFs = 7.5
BENZENE HCBD
Response RRF Response RRF
575498 0.700 219910 0.340
279430 0.699 111351 0.353
132341 0.672 51646 0.355
68616 0.733 27152 0.371
26112 0.652 13810 0.461
3248 254
0.9993 0.9997
-5397 264
0.691 0.376
4.5 13
563
-------
Table 3: Calculated MDLs of TO-14 Compounds
Compound . MDL (ppbV)
Dichlorodifluoromethane 0.10
Chloromethane 0.12
1,2-Dichloro-1,1,2,2-tetrafluoroethane 0.15
Vinyl chloride °-07
Bromomethane 0.09
Chloroethane 0-04
Trichlorofluoromethane 0.12
1,1-Dichloroethene 0.11
Methylene chloride 0.40
1,1,2-Trichloro-1,1,1-trifluoroethane 0.20
cis-1,2-Dichloroethene 0.07
1,1-Dichloroethane 0.07
Chloroform 0.06
1,2-Dichloroethane 0.06
1,1,1-Trichloroethane 0.29
Benzene 0.03
Carbon tetrachloride 0.03
1,2-Dichloropropane 0.09
Trichloroethene 0.04
cis-1,3-Dichloropropene 0.09
trans-1,3-Dichloropropene 0.12
1,1,2-Trichloroethane 0.10
Toluene 0.07
1,2-Dibromoethane 0.11
Tetrachloroethene 0.06
Chtorobenzene 0.04
Ethylbenzene 0.06
m,p-Xylenes 0.14
Styrene 0.03
1,1,2,2-Tetrachloroethane 0.30
o-Xylene 0.04
1,3,5-Trimethylbenzene 0.06
1,2,4-Trimethylbenzene 0.13
Benzyl chloride 0.30
1,3-Dichlorobenzene 0.05
1,4-Dichlorobenzene 0.06
1,2-Dichlorobenzene 0.04
1,2,4-Trichlorobenzene 0.08
Hexachlorobutadiene 0.09
Average MDL = 0.11
564
-------
Figure
Mass Chromatogram of a Canister VOC Standard
RIC DATA: 02179387 #1 SCANS 556 TO 5580
92/17/93 15:39:09 CALI: 02179307 #3
SAMPLE: 1.0PPBU
CONDS.:
RANGE: G 1,6881 LABEL: N 5, 5.0 QUAH: A 5, 5.8 J Q BASE: U 29.. 3
RIC
1000
2800
3000
17:13
4000
23:05
5000
23:52
SCAN
TIME
-------
SELECTION CRITERIA FOR GROUND-WATER MONITORING WELL CONSTRUCTION
MATERIALS
J.R. Brown. US Environmental Protection Agency, Washington, DC
20460, L. Parker, US Army Cold Regions Research and Engineering
Laboratory, Hanover, NH 03755-1290, A.E. Johnson, Science
Applications International Corporation, McLean, VA 22102
ABSTRACT
This presentation discusses a regulatory perspective on the
selection of ground-water monitoring well casing materials for
hazardous and solid waste disposal facilities that must comply
with the Resource Conservation and Recovery Act requirements.
The primary focus will center on EPA's current recommendations
for selecting well casing materials, with special consideration
given to regulatory requirements, data quality objectives, common
types of well casing materials and their physical and chemical
characteristics, and geochemical considerations. A table
summarizing the recommended selection criteria and a series of
case studies will also be presented.
566
-------
AUTO-GC DESIGN AND OPERATION FOR
REMOTE UNATTENDED VOC DETERMINATIONS
James F. Ryan. Environmental Marketing Manager
and
Ian Seeley, Senior Technical Specialist
The Perkin-Elmer Corporation
50 Danbury Road, Wilton, Connecticut 06897
ABSTRACT
Recent emphasis on the design considerations for ozone precursor analytical systems using gas
chromatography has focused on the practical and logistical difficulties of using liquCid cryogen
for trapping volatile organics. Conventional approaches to the problems of trapping species in
the range of C2 to C10 require liquid nitrogen for cold-trapping and for sub-ambient operation of
the chromatographic oven. Liquid nitrogen is not only expensive, but difficult to supply in a
reliable fashion for long-term unattended operation at remote locations.
A system has been developed and evaluated for the on-line ambient detection of hydrocarbon
ozone precursors that does not use liquid cryogen. This novel chromatographic approach has
been used to optimize the separation of C2 to C10 compounds using a pressure switching facility.
Such a chromatography system has the added benefit of stabilizing the column system resulting
in improved retention time reproducibility. This is consistent with the requirements of a robust,
stable system capable of long-term unattended operation.
INTRODUCTION
U.S. Environmental Protection Agency, namely carbon monoxide, sulfur oxides, nitrogen oxides,
ozone, particulate matter less than 10 microns in size, and airborne lead. Though EPA has
promulgated these levels for some time, a number of urban areas in the United States continue
to be unable to bring their normal levels into compliance with EPA regulatory levels. This "non-
attainment" problem is especially acute with respect to ozone. There are 23 cities in the U.S.
that are considered non-compliant with ozone regulations.
In order to address this problem, the 1990 Clean Air Act Amendments (CAAA) required that EPA
develop an ozone abatement strategy based not only on the chemical species itself (i.e., ozone),
but based as well on hydrocarbon precursors of ozone that react with sunlight and nitrogen
oxides to form the unhealthful, ozone-rich, summer's haze so prevalent in urban areas.
In order to limit the release of low molecular weight hydrocarbon precursors into ambient air, it is
necessary to determine the hydrocarbon source. Therefore, an instrumental approach has been
developed permitting the automatic collection and chromatography of C2 to C10 hydrocarbons
from ambient air.
INSTRUMENTAL DESIGN
This system is based on the Perkin-Elmer ATD-400 Automatic Thermal Desorption instrument
that has a Peltier-cooled adsorbent trap. After being electrically cooled to -30 oC, ambient air is
567
-------
drawn through the trap. Following a predetermined sampling time, the trap is automatically
connected to the Perkin-Elmer 8700 gas chromatograph via a 6-port valve, and the trap is
programmably heated at 40 oC/sec; hydrocarbons are flashed off the adsorbent and onto a
capillary GC column.
Chromatographic separation is mutti-dimensional, i.e., a combination of a PLOT column for the
extremely volatile low molecular weight compounds and a BP1 boiling point column for the
higher molecular weight species. All the organic analytes from the trap are transferred to the
BP1 column. For the first 13 to 14 minutes, the effluent from the BP1 column is directed onto
the AI2O3.Na2SO4 PLOT column. Thus early eluting poorly resolved components from the first
column are transferred for further Chromatographic separation on the PLOT column. After 13 to
14 minutes, the effluent from the BP1 column is switched (using the Deans system) directly to a
2nd FID. Higher boiling, well-resolved components are analyzed on this second detector and are
prevented from contaminating the PLOT column. For the remainder of the analysis, components
elute simultaneously on both columns. However, as each peak only elutes on one or the other
column, and not on both columns, data handling is not complex. This system is illustrated in
Figure 1.
I Vacuum |_| M
Rump | | C
Air
Cr»rrl«»r
•»O_ Column
Hpnoumotlc I
Swttcti I
Figure 1: Schematic diagram of ambient air monitoring system.
The resulting Chromatographic output comes from two GC detectors, one for the PLOT column,
and a second for the BP1 column. Figure 2 shows a typical GC output for a 55-component
standard.
568
-------
Perkin Elmer Ozone Precursor
12-
10-
8-
6-j
14-
12-
1 OJ
8-
61
C
joo_O3O
C»bn 1 PLOT
< t
| iLl_JJJj
i
. . . . , , . . , , , r i . | . . . . | i . . . , .
job_O3O
C.bnut2BP-l *^* **:
"\\ "\ »
XV *^^**JI \ ll i ||
27 | 2* | 1 I k 1 ll
, .IklJUJ^
5 5 1O 15 2O 25
*
LI-
«a 44
JL
A
3O
!•
13
,
System
20
-1« '
X^\
1
• i J
» 4
•
93
1
I
\
JJ
*»
2<
JL ^
i
i
iu^__
35 AQ 45
Figure 2: Multidimensional chromatographic output from ambient air monitoring system.
Peak identities are shown in Table 1.
VOCS Chromatographed on PLOT Column
01
02
03
04
OS
06
07
08
09
10
11
12
13
14
Ethane
Ethytene
Propane
Propene
teobutane
n-Butane
Acetylene
trans-2-Butene
1-Butene
cis-2-Butene
Cyclopentane
teopentane
n-Pentane
2-Methyl-2-butene
15
18
17
18
19
20
21
22
23
24
25
26
Cyclopentene
Trans-2-pentene
3-MethyM-butene
1-Pentene
cis-2-Pentene
2,2-Dimethylbulane
3-Methytpentane
2-Metnylpentane
2,3-Dimethylbutane
Isoprene
4-Methyl-1-pentene
2-Melhyl-1-pentene
VOCs Chromatographed on BP1 Column
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
n-Hexane
trans-2-Hexene
cis-2-Hexene
Methylcyclopentane
2,4-Dimethylpentane
Benzene
Cyclohexane
2-Methylhexane
2,3-Dimethylpentane
3-Methylhexane
2,2,4-Trimelhylpenlane
n-Heptane
Methylcyclohexane
2 ,3,4-Trimethylpenlane
Toluene
42
43
44
45
46
47
48
49
50
51
52
53
54
55
2-Methylheptane
3-Methylheptane
n-Ociane
Ethyl benzene
p-Xylene
2-Methylheptane
3-Methylheptane
n-Nonane
I sopropyl benzene
n-Propyl benzene
a-Pinene
1 ,3,5-Trimethylbenzene
b-Pinene
1 ,2,4-Trimethylbenzene
Table 1: Identification of chromatographic peaks in Figure 2.
The advantages of this instrumental system are that the 55 compounds of interest are separated
cleanly, and without the use of liquid nitrogen. Neither the VOC collection (using the Peltier-
cooled trap) nor the chromatography (which starts at 40 oC) require liquid nitrogen.
In defining the limits of this system, questions naturally arise concerning trap breakthrough
volume and overall detection limits. Figure 3 below shows three of the key C2 compounds,
introduced to the system at 16 to 19 ng in absolute amounts. After introduction of the three
hydrocarbons, zero grade bottled air was drawn through the Peltier-cooled trap and the organic-
laden adsorbent. The trap was then heated to release the three VOCs for chromatography. As
shown, acetylene response began to fall off only above 600 ml of air volume. Therefore, 600
mL of air was chosen as the sample size for analysis.
569
-------
400000 -i
C2 HYDROCARBONS
COMPONENT RECOVERY v VOLUME SAMPLED
SAMPUNG TIME: 40min
16 to 19ng each component
300000^
200000-
100000-
Ethan* (75«RH)
EthyUn* (75«RH)
AMfylww (75«RH)
200
FLOW RATE (mL/mln)
15 25
1 ' I ' '.' ' '.' M'
35
600 1000
SAMPLE VOLUME (ml)
Figure 3: Determination ot air volume breakthrough values tor acetylene, ethylene and ethane.
That modest sample size notwithstanding, Figure 4 shows that this system achieves low level
quantrtation levels well within the part-per-trillion range. Methyl cyclohexane, the second labeled
peak, has been quantified at 0.06 ppb.
TYPICAL LOW LEVEL DETERMINATION
|
AlLJL
Figure 4: Illustration of low quantrtation levels achievable with air monitoring system.
As an example of how this system can be used, consider Figure 5. In this chart, concentrations
of three VOCs, namely isobutane, n-butane and ethane (based on ppb of carbon), in laboratory
ambient air have been determined over a 24 hour period. Note that the X axis is based on a
starting time of 9:00 AM, i.e., the first series of bars is at 9:00 AM, the second at 10:00 AM, etc.
570
-------
At 6:00 PM, there is a dramatic increase in isobutane. This is when the evening laboratory
cleaning crew begins their work, and one can speculate that the isobutane and n-butane rise
might be due to the cleaning reagents they use, specifically if isobutane is a propellant in one of
their spray cans. The butane data, along with charts derived from other hydrocarbon data, is
presented in Figures 5 to 8.
Diurnal Variations of Ozone Precursors
HOUR from fcOOun 12/1/82
Dumal Variations of Cfeone Precursors
Ham torn Ittam 12/1/n
Fig. 5: Concentration of butanes and ethane
over 24 hour period.
Fig. 6: Concentration
and toluene
of
benzene
Diurnal Variation* of Ozona Praeuraore
J J J iiJL
Haora Iran IMim 11/1/tl
Diurnal Variationa of Ozona Pracuraon
B^jalii Jj. ..•.
Ham fram Sfivdm llri/»
Fig. 7: Concentration of nonane and decane Figure 8:
over 24 hours.
Concentration of isoprene and two
pinenes over 24 hours
It is interesting to compare the four charts, even if explanations for the data are not readily
apparent. As noted above, the butanes might be explained as a propellant used in a spray can.
However, no similar explanation for the presence of nonane, decane, isoprene and the pinenes
readily comes to mind, other than that the engineering laboratory in which this instrument is
located is adjacent to a major roadway. Investigations to determine the source of these
chemicals continue.
It should also be noted that an intrinsic part of this ambient air monitoring system is the ability to
not only collect data at a remote site, but also to transmit information on the status of the
instrument operation and reports on the collected data to a central operation. Using personal
computers, modem telecommunications software such as PC-Anywhere, and Perkin-Elmer
Turbochrom 3 chromatographic data handling software, the status of the ambient air monitoring
system can be evaluated, raw data can be examined, and summary reports generated for
transmission to a central operation where they can be combined with reports from other stations.
Raw data files, because of their size, would likely be collected once per week during routine
visits to the monitoring site. Chromatographic run and summary reports could be collected
several times a day as needed.
571
-------
SUMMARY
There are unique advantages to an ambient air monitoring system such as that described above.
First and foremost, the system routinely operates unattended. The instrument used to collect
data for this paper has been in continuous operation for over 1500 hours, and has consumed
only 1 tank of hydrogen. No liquid nitrogen was required and zero grade air is generated using a
laboratory compressor. Secondly, even though the instrument operates unattended, through the
use of appropriate data handling and remote operation software, reports on instrument operation
and chromatographic results can be collected as often as necessary.
572
-------
AUTHOR INDEX
-------
Author Index
Author
Paper Number Author
Paper Number Author
Paper Number
Adolfo, N 7
Alcaraz, M 79
Allen, R.L 36
Almeida, M 73
Almond, R.E 36,38
Anderau, C 72
Anderson, D 46,78
Arrowood, S.P 36
Baker, R.D 64
Barnes, K.M 72
Barton, J 1
Bath, R.J 10
Beckett, W.F 79
Bencivengo, D.J 77
Berman.R 69
Betowski, L.D 57
Bollman.M 31
Brandenberg, C.W 66
Brown, J.R 85
Brown, W.T 35
Bruce, M 39,43
Budde, W.L 44
Bulman, F 73
Burkitt, D 39
Campbell, C.R 64
Carl, J 39, 61
Carlson, R.E 55
Carlston, R 78
Carpenter, S.C 14
Caruso, M 2
Caton, J.E 48
Chapnick, S 4
Chau,N 63
Chen, P.H 51
Christopherson, M 56
Cline-Thomas, T 28
Colby, B.N 50, 77
Converse, J.C 16
Cross-Smiecinski, A 18
Curl, P 30
DalSoglio, J 31
Davis, A 12
Davis, R 76
Deloatch, 1 6
DiBussolo, J.M 52
Diebold, T 26
Dobb, D.E 16
Doherty, L.C 49,53
Dolata, L.A 68
Dong, M.W 52
Doskey, P.V 14
Doster, J.A 76
Douglass, O.B 9
Dutta, S 29
Dux, T.P 13'56
Eichelberger, J.W 44
Engelhart, W.G 80
Erickson, M.D 14
Fadeff, S.K 20
Farrell, J 3
Feeney, M.J 59
Forsberg, D 17
Fowler, R 32
Fredeen, K.J 72
Friedman, S.B 36,38
Garner, F 5
Geier, D.S 61
Gere, D.R 42, 53, 54
Gillet, C 31
Gjerde, D 81
Gladwell, A 8
Gluodenis, T.J 72
Godfrey, C.B 8
Goheen, S.C 20
Gore, J.C 21
Greenlaw, P 10
Griest, W.H 48
Grosser, Z.A 72, 75
Haley, A.E 11
Hall, J.R 3
Hall, K 13
Hansen, G 78
Harrison, R.0 55
Hartwell, S 78
Hastings, M.R 62
Hasty, E.T 80
Hayes, M.C 37
Heithmar, E 69
Helms, C.L 77
Henderson, G 31
Herzog, D.P 37, 58
Hill, B.R 3
Ho, J.S 44
Hooton, D 56
Hottenstein, C.S 58
Hsu, J.P 83
Jackson, K.W 2
Jackson, T 46
Janiec, G 9
Jenkins, T.F 45
Johnson, A.E 85
Johnson, D.P 36
Jourdan, S.W 37,58
Jungclaus, G 13, 56
Kauffman, J.S 65
Keeran, W.S 51
Keller, M.E 84
Kennedy, M 47
Kibler, L 19
Killough, B 39
King, R 31
Kirkland, K 78
Kleiser, H.R 30
Knipe, C 42, 53, 54
Kuehl, M.A 15
Larson, A 30
Lawruk, T.S 37,58
Leavell, J.R 66
Lee, H.B 42
Lesnik, B 33,47, 63
Levy, J.M 68
Lillian, D 10
Lindahl, P.C 14
Linder, G 31
Littau, S.E 80
Liu, Y 79
Lohner, W 21
Long, S.E 76
Lopez-Avila, V 79
Loring, D 4
Low, N 49
Macdonald, D 17
Mapes, J.P 36, 38
Marsden, P 34,46,47, 63
Mauro, D 40
McCallister, R 16
McCormick, E.F 45
McCulloch, M 20
McDonald, P.P 36, 38
Mclntosh, S 75
Mehra, H 81
Merewether, G 74
Meszaros, T.J 16
Miller, D.P 71
Miller, G 83
Mills, W 22, 28
Miyares, P.H 45
Mong, G.M 20
Moore, M 3
Murarka, 1 40
Myers, K.F 45
Nwosu, J 31
Nygaard, D.D 73
O'Brien, W 22
O'Brien, K 8
O'Neil, G 41
Organ, B 16
Ott, S 31
Owens, J-17
Owens, R.G 61
Pace, C 57
Pan, J.C 83
Parker, L 85
Parr, J.L 3
Parsons, A 67
Parsons, C.S 50
Pfleeger, T 31
Pflug, A.D 14
-------
Author Index
Author
Paper Number Author
Paper Number Author
Paper Number
Phinney, D 61
Pipkin, W. 42, 53, 54
Ploscyca, J 3
Pollack, L.P 60
Poppiti.J 20
Potter, B.B 76
Poziomek, E.J 18
Purdy.C 14
Randa, H 56
Randall, L.G 42, 53, 54
Ravey, R.M 68
Ray, E. 21
Richardson, L.A 11
Riga, T.J. .. 41
Riley, R.G 20
Roberts, D.F 51, 84
Robertson, G 5
Roby, M '. 34, 57
Rosecrance, A. 7, 19
Rubio, F.M 58
Ryan, J.F. 86
Salesky, J 21
Sauerhoff, S 75
Sauter, R 56
Schenley, R.L 48
Schindler, D.R 51
Schmick, M 8
Schneider, J 40
Schutte, W.E 14
Seeley. 1 86
Shah, N 74
Shaw, T.A 61
Shelow, D.M 59
Sklarew, D.S 20
Sleevi, P 3
Smith, D.V 60
Smith, J.S 50
Smith, R.K 61
Spear, R.D 10
Speis, D 3
Spilkin, A 62
Spitz, H ., 21
Stapanian, M 5
Stewart, T.N 36
Straka, M 69
Strong, A.B.... 45
Studebaker, W.B 36
Swift, R.P 66
Tang, P.H 44
Taylor, B 40
Thiel, R.D , 82
Thomas, B 20
Tomczak, L 21
Tracy, A 22,28
Tsang, S.F '. 34,46,47,63
Tucker, S ; 3
Valkenburg, C.A: 35
Van Ausdale, W.A 51
Van Stiver, C 27,32
Vance, L.W 23
Varcoe, F.T 60
Vernon, J 12
Vinson, C 3
Vitale.RJ 70
Wallace, E 27
Weichert, B.A 64
Weiss, A.J 55,67
Wilborn, D 31
Williams, A 13
Willig, T 65
Winslow, M.G 84
Winward, R.T 24
Withers, T.A 36
Yates, D.A 72
Young, M 40
Zha, Y 58
Zine, T 39
-------
|