The 14th Annual Waste Testing & Quality Assurance Symposium
PROCEEDINGS
Sponsored by the
American Chemical Society and the
U.S. Environmental Protection Agency
July 13-15, 1998 Crystal Gateway Marriott Arlington, VA
-------
Proceedings
The Fourteenth Annual
Waste Testing & Quality Assurance
Symposium
(WTQA '98)
July 13-15, 1998
Crystal Gateway Marriott
Arlington, VA
-------
HIGHLIGHTS
14th Annual Waste Testing & Quality Assurance Symposium
(WTQA '98)
Using A Performance-Based Measurement System (PBMS)
WTQA '98, cosponsored by the American Chemical Society (ACS) Division of Environmental Chemistry and the
U.S. Environmental Protection Agency (EPA) Office of Solid Waste and Emergency Response (OSWER), will be
held at the Crystal Gateway Marriott in Arlington, VA, July 13-15, 1998. This year's symposium continues the
partnership between the regulated community and their supporting laboratories and state and federal regulators
from the Research Conservation and Recovery Act (RCRA) and the Comprehensive Environmental Response
Compensation and Liability Act (CERCLA) programs.
Highlights
EPA is actively working to implement the President's program for "reinventing" government and reforming
regulatory policy. As part of this endeavor, EPA has been trying to break down barriers to using new monitoring
techniques. One barrier that OSWER is tackling is the requirement to use specific measurement methods or
technologies in complying with Agency regulations. EPA's Environmental Monitoring Management Council
(EMMC), members of the regulatory community, and Congress all agree that EPA needs to change the way it
specifies monitoring requirements in regulations and permits. There is broad acceptance for use of a
performance-based measurement system (PBMS). The EPA Office of Solid Waste and Emergency Response
(OSWER) strongly supports this position and is committed to using this approach in both the RCRA and
CERCLA monitoring programs. This year's plenary session speakers, including Fred Hanse, Deputy
Administrator of EPA; Brad Campbell, Associate Director for Toxics & Environmental Protection; Steve Koorse,
Hunton and Williams; Elizabeth Cotsworth, Acting Director, Office of Solid Waste; and Larry Keith, Waste Policy
Institute, will focus their remarks on various aspects of PBMS.
PBMS Implementation Session
This session, organized by the International Association of Environmental Testing Laboratories Section of the
American Council of Independent Laboratories (ACIL-IAETL), will include a presentation of EPA Program Office
PBMS Implementation Plans, followed by speakers presenting the laboratory community, regulated community,
and state regulatory agency perspectives on PBMS. The focus of the talks from the regulated community will be
on the barriers they expect to face when trying to obtain the benefits of PBMS and what we all need to do to
overcome the problems.
PBMS Methods Validation Session
Now that the EPA has moved towards implementing a PBMS approach to environmental monitoring, the
regulated community needs to know how to validate methods under this new system. This session, also
organized by ACIL-IAETL, will focus on how to tailor measurement system validation to the required data quality;
how validation criteria should be specified in order that it not serve as a barrier to using alternative measurement
technologies; and what is the minimum methods validation data set that one needs for the data obtained from
the analysis to be of known and documented quality.
EPA's Environmental Monitoring Research Program
This session will focus on environmental monitoring methodology research supported under EPA's Science to
Achieve Results (STAR) competitive, extramural grant program. The program has funded a number of research
projects whose goal it is to develop unique or novel techniques for monitoring pollutants in the environment.
Methodology to monitor air, water, soil and other media will be presented. The session will review the results
from projects funded in prior years, and discuss the objectives and approaches to be undertaken in research
studies that received funding this past year.
Superfund Session
The keynote for this Superfund session will be "Times Are Changing." Planned highlights include (1) a description
of the Superfund pipeline over time, emphasizing the near and longer term future and how that relates to
analytical service need; (2) information on how Superfund is working with other EPA program offices to enhance
our operation; (3) the trend to encouraging economic redevelopment at Superfund sites; (4) how Superfund is
implementing a PBMS approach; (5) initiatives related to data quality and minimal Quality Systems; (6)
iii
-------
electronic data delivery and validation; (7) usefulness of the National Environmental Laboratory Accreditation
Conference (NELAC) accreditation for Superfund analytical work; (8) future direction for the Contract Laboratory
Program (CLP); (9) Contracts 2000, Performance-Based Contracting, and what this means to the laboratory
community, and (10) laboratory perspective on working for Superfund (a panel discussion).
SYMPOSIUM-AT-A-G LANCE
Sunday, July 12
8:00 a.m. - 9:30 a.m.
9:00 a.m. -12:00 noon
9:00 a.m. -4:00 p.m.
1:00 p.m. - 5:00 p.m.
Registration for Short Courses
Short Course: Analytical Strategy for the RCRA Program
Short Course: An Introduction to Practical Ethics for Environmental Laboratories
Short Course: Clean Chemistry for Trace Elemental Analysis
Monday, July 13
7:00 a.m. - 4:00 p.m. Registration Open
8:00 a.m. -12:00 noon SW-846 Workgroups (closed)
8:00 a.m. 12:00 noon Short Course: Quality Systems, PBMS and NELAC: Putting It All Together
11:00 a.m. -12:00 noon Trial Exam for Environmental Analytical Chemists
2:00 p.m. - 4:30 p.m. Opening Plenary Session
David Friedman, Office of Research and Development
Gail Hansen, Office of Solid Waste
Fred Hansen, Deputy Administrator of EPA
Brad Campbell, Associate Director for Toxics & Environmental Protection
Steve Koorse, Hunton & Williams
Elizabeth Cotsworth, Acting Director, Office of Solid Waste
Larry Keith, Waste Policy Institute
5:00 p.m. - 7:00 p.m. Opening Reception
Tuesday, July 14
7:00 a.m. - 5:00 p.m.
8:15 a.m. -12:00 noon
8:15 a.m. -12:00 noon
10:00 a.m. -10:30 a.m.
12:00 noon -1:00 p.m.
1:15 p.m. -5:00 p.m.
1:15 p.m. - 5:00 p.m.
3:00 p.m. -3:30 p.m.
7:00 p.m. 8:00 p.m.
Wednesday, July 15
7:00 a.m. -5:00 p.m.
8:15 a.m. -12:00 noon
8:15 a.m. -12:00 noon
10:00 a.m. 10:30 a.m.
12:00 noon -1:00 p.m.
1:15 p.m. -5:00 p.m.
1:15 p.m. -5:00 p.m.
3:00 p.m. - 3:30 p.m.
7:00 p.m. - 8:00 p.m.
Thursday, July 16
9:00 a.m. -12:00 noon
9:00 a.m. -4:00 p.m.
Registration Open
Organic I Session
Inorganic Session
Break
Lunch Break
Organic II Session
Environmental Monitoring Research Session
Break
Trial Exam for Environmental Analytical Chemists
Registration Open
PBMS Implementation Session
QA Session
Break
Lunch Break
PBMS Validation Session
General Session
Break
Trial Exam for Environmental Analytical Chemists
Short Course: Analytical Strategy for the RCRA Program
Superfund Session
IV
-------
CONTENTS
ORGANIC
Paper page
Number Number
1 Overview of Current Status of the RCRA Organic Methods Program. B. Lesnik 3
2 Appropriate Ways to Modify Existing Methods for New Applications. B. Lesnik 3
3 Questionable Practices in Organic Laboratories. J.F. Solsky 3
4 Modifications to SW-846 HPLC Methods 8330 and 8310. S. Weisberg, M.L. Ellickson 4
5 PCB Separations Using 2 Dimensional GC for Confirmation: Use of a Heart-Cutting 4
Switching Valve. D.R. Gere, A. Broske, R. Kinghorn
6 Estimating the Total Concentration of Volatile Organic Compounds in Soil at Sampling 5
Locations: Field Trials. A. Hewitt, M.H. Stutz
7 Final Evaluation of Method 3546: A Microwave-Assisted Process (MAP) Method for 11
the Extraction of Contaminants Under Closed-Vessel Conditions. J.R.J. Pare, J.M.R.
Belanger, C. Chin, R. Turle, B. Lesnik, R. Turpin, R. Singhvi
8 Phenoxyacid Herbicide Screening. M. Bruce, K.L. Richards, R.M. Risden 17
9 Field Demonstration of a Portable Immunosensor for Explosives Detection. A.W. 26
Kustarbeck, P.T. Charles, P.R. Gauger, C. Patterson
10 Environmental Applications of a Fiber Optic Biosensor. L.C. Shriver-Lake, I.B. 26
Bakaltcheva, S. van Bergen
11 Development and Validation of an Improved Immunoassay for Screening Soil for 27
Polynuclear Aromatic Hydrocarbons. T.S. Fan, B.A. Skoczenski
12 A New Dioxin/Furan Immunoassay with Low Picogram Sensitivity and Specificity 27
Appropriate for TEQ Measurement: Applications Development. R.O. Harrison, R.E.
Carlson
13 Development and Validation of an Immunoassay for Screening Soil for Polychlorinated 28
Biphenyls. T.S. Fan, B.A. Skoczenski
14 Gasoline Range Aromatic/Aliphatic Analysis Using Pattern Recognition. S.E. Bonde 28
15 Bioremediation Assessment Using Conserved Internal Biomarkers. P. Calcavecchio, 29
E.N. Drake, G.C. VanGaalen, A. Felix
16 Method 8270 for Multicomponent Analyte Analysis. E.A. LeMoine, H. Hoberecht 35
17 Benzidine? Really? R-K. Smith 40
18 Comparison of Volatile Organic Compound Results Between Method 5030 and Method 43
5035 on a Large Multi-state Hydrocarbon Investigation. R.J. Vitale, R. Forman, L.
Dupes
INORGANIC
Paper
Number
19
20
21
SW-846 Inorganic Methods Update. O. Fordham
Direct Mercury Analysis: Field and Laboratory Applications, H.M. Boylan, H.M.
Kingston, R.C. Richter
Mercury in Soil Screening by Immunoassay. M.L. Bruce, K.L. Richards, L.M. Miller
Page
Number
53
53
55
-------
INORGANIC (continued)
Paper
Number
22
23
24
25
26
27
Utilization of a Field Method for the Semiquantitative Detection of Silver in
Environmental Samples in the 0-50 ppb Range. D. Kroll
Diagnosing Errors in Species Analysis Procedures Using SIDMS Method 6800. H.M.
Kingston, D. Huo, Y. Lu
The Use of 210Pb Dating Stratigraphy to Determine the Significance and Fate of
Chromium in Sediments near a Hazardous Waste Site. R. Rediske, G. Fahnenstiel, C.
Schelske, M. Tuchman
Air-Force Wide Background Concentrations of Inorganics Occurring in Ground Water
and Soil. P.M. Hunter
New Tools for Liquid Sampling. J.D. Hoover, S.R. Somers
Analysis of Chemical Warfare Agent Decontamination Brines for Lewisite Degradation
Products Using Gas Chromatography with Atomic Emission Detection. K.M.
Morrissey, T.R. Connell, J. Mays, H.D. Durst
Page
Number
60
64
65
73
77
88
EPA'S ENVIRONMENTAL MONITORING RESEARCH PROGRAM SESSION
Paper
Number
28
29
30
31
32
33
34
35
36
37
38
39
40
Introduction, Session Scope and Purpose. W. Stelz
Bioavailabilty and Risk Assessment of Complex Mixtures. K.C. Donnelly, W.R.
Reeves, T.J. McDonald, L.-Y. He, R.L. Autenrieth
Field Determination of Organics from Soil and Sludge Using Sub-critical Water
Extraction Coupled with SPME and SPE. S.B. Hawthorne, C.B. Grabanski, A.J.M.
Lagadec, M. Krappe, C.L. Moniot, D.J. Miller
A Field Portable Capillary Liquid/Ion Chromatograph. T.S. Kaphart, C.B. Boring, P.K.
Dasgupta, J.N. Alexander IV
Rapid Determination of Organic Contaminants in Water by Solid Phase
Microextraction and Infrared Spectroscopy. D.C. Tillotta, D.C. Stahl, S.A.
Merschman, D.L. Heglund, S.H. Lubbad
Intrinsic Stable Isotopic Tracers of Environmental Contaminants. S.A. Macko, P.J.
Yanik, N.A. Cortese, M.C. Kennicutt II, Y. Quin
Recent Developments in Immunobiosensors and Related Techniques for the
Detection of Environmental Pollutants. M. Masila, H. Xu, E. Lee, O.A. Sadik
Multiplexed Diode Laser Gas Sensor System for In-situ Multi-species Emissions
Measurements. R. Hanson
Overview/Future of NCERQA Research Program. B. Krishnan
Advanced Analytical Methods for the Direct Quantification and Characterization of
Ambient Metal Species in Natural Waters. J.G. Hering, J.H. Min
Radical Balance in Urban Air. R.J. O'Brien, L.A. George, T.M. Hard
Environmental Applications of Novel Instrumentation for Measurement of Lead
Isotope Ratios in Atmospheric Pollution Source Apportionment Results. Keeler
Remote Sampling Probe with Fast GC/MS Analysis: Subsurface Detection of
Environmental Contaminants. A. Robbat, Jr.
Page
Number
93
93
93
94
101
102
110
117
117
117
118
119
119
VI
-------
ERA'S ENVIRONMENTAL MONITORING RESEARCH PROGRAM SESSION (Continued)
Paper
Number
41
42
43
44
45
46
47
48
49
50
QUALITY
Paper
Number
51
52
53
54
55
56
57
58
59
60
61
An Integrated Near Infrared/Spectroscopy Sensor for In-situ Envrionmental
Monitoring. R.A. Levy, J.F. Federici
A Nitric Oxide and Ammonia Sensor Array for Fossil Fuel Combustion Control. B.T.
Marquis, J.F. Vetelino, T.D. Kenny
Relevance of Enantiomeric Separations in Environmental Science. D.W. Armstrong
Development of Aerosol Mass Spectrometer for Real Time Analysis of PAH Bound to
Submicron Particles. J.T. Jayne, D.R. Worsnop, C.E. Kolb, X. Zhang, K.A. Smith
Real-Time Trace Detection of Elemental Mercury and Its Compounds. R.B. Barat
Orthogonal Background Suppression Technique for EPA's Field Infrared Data
Processing. Blaterwick
Development of a Continous Monitoring System for PMio and Compounents of PM2s,
M. Lippmann, J.Q. Ziong, W. Li
A Real-Time Sampler, Rams, for the Determination of PM25, Including Semi-Volatile
Species. F Obeidi, E. Patterson, DJ. Eatough
Dynamic Nuclear Polarization (DNP): A New Detector for Analysis of Environmental
Toxicants. H.C. Dorn
Partitioning Tracers for In-situ Detection and Measurement of Nonaqueous Liguids in
Porous Media. Brusseau
ASSURANCE
CD-R Archive and Catalog of Hewlett-Packard Low Resoltuion Format Data from
Networked GC/MS Systems. R.G. Briggs, H. Valente
The Use of Acceptable Knowledge for the Characterization of Transuranic Waste in the
Department of Energy Complex. R.V. Bynum, R.B. Stroud, R.D. Brown, L.D. Sparks
Performance Evaluation Soil Samples for Volatile Organic Compounds Utilizing
Solvent Encapsulation Technology. J. Dahlgran, C. Thies
U.S. EPA and U.S. A.F Interagency Agreement for Field Analytical Services. R.A.
Flores, G.H. Lee
Several Organic Parameters on Underlying Hazardous Constituents List Can Not Be
Measured at the Universal Treatment Standards. H.C. Johnson, C.S. Watkins
Ignitability Performance Evaluation Study - Are Your Waste Streams Being Correctly
Characterized? L.J. Dupes, R.J. Vitale, D.J. Caillouet
Techniques for Improving the Accuracy of Calibration in the Environmental Laboratory.
D.A. Edgerley
Quality Control Protocol for Analysis of Drugs and Explosives Using Ion Mobility
Spectrometry. J. Homstead, E.J. Poziomek
Reference Materials for the Analysis of Metals in Sludge. S.J. Nagourney, N.J.
Tummillo, Jr., J. Birri, K. Peist, B. MacDonald, J.S. Kane
Perspectives on Dioxin Analysis. S. Wilding
Interpretation of Ground Water Chemical Quality Data. G.M. Zemansky
Page
Number
124
125
131
131
132
137
137
146
147
148
Page
Number
151
152
161
165
165
173
181
187
188
191
192
vii
-------
GENERAL
Paper Page
Number Number
62 Sample Introduction Techniques for Fast GC Analysis of Organochlorine Pesticides. 205
C.E. Boswell, M.S. Clark, R.M. Woodard
63 lonspray LC/MS Method for Quantitation of Major Degradation Products of Benomyl. 209
C. Eilcone, T. Xuan, S.M. Reddy
64 Methods Optimization of Microwave Assisted Solvent Extraction Techniques for 214
Various Regulatory Compounds. G. LeBlanc, M. Miller
65 New Developments in Closed Vessel Microwave Digestion Technology for Preparation 215
of Difficult Organic Samples for AA/ICP Analysis. G. LeBlanc, B. Haire, B. Fidler
66 Evaluation of ICP-OES and ICP-MS for Analysis fo the Full TCLP Inorganic Target 215
Analyte List. M. Paustian, Z. Grosser
67 Clean Metals Sampling. R.E. Stewart II 219
Author Index 227
VIII
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
ORGANIC
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
OVERVIEW OF CURRENT STATUS OF RCRA ORGANIC METHODS PROGRAM
Barry Lesnik
US EPA, Office of Solid Waste, 401 M ST., SW, Washington, DC
A/0 ABSTRACT AVAILABLE
APPROPRIATE WAYS TO MODIFY EXISTING METHODS FOR NEW APPLICATIONS
Barry Lesnik
US EPA, Office of Solid Waste, 401 M ST., SW, Washington, DC
NO ABSTRACT A VAILABLE
QUESTIONABLE PRACTICES IN THE ORGANIC LABORATORY
Joseph F. Solsky
US Army Corps of Engineers (CENWO-HX-C), 12565 W Center Road, Omaha, NE 68144
Phone: (402) 592-9542, Fax: (402) 592-9595
e-mail: joseph.f.solsky@usace.army.mil
SW-846 is a collection of performance-based methods used by the US Army Corps of Engineers (USAGE) for
execution of environmental restoration projects. These methods provide guidance for the running of various
organic and inorganic protocols that can be used for the analysis of samples from a variety of sample matrices.
During recent laboratory audits conducted by the USAGE, certain 'questionable practices' have been observed,
especially in the organic analysis areas.
Most people have a relatively good idea of what constitutes a fraudulent activity today. The concepts of
'dry-labing,' 'peak shaving,' 'peak enhancing,' or 'time-traveling1 are well understood. These practices clearly
involve the deliberate direct manipulation and/or alteration of data, often to achieve or meet method QC criteria.
However, there are a new group of 'questionable practices' now being observed that often involve the selective
exclusion of data to achieve or meet method QC criteria.
This presentation will focus on several of these 'questionable practices'. Examples of some of these practices
include the following: (1) One such practice involves the determination of initial calibration curves. Laboratories
have been observed running eight or more standards for methods that state 'a minimum of five points should be
used to establish the initial calibration curve.' Up to three points are then discarded, even from the middle of the
curve until the appropriate QC criteria can be met. No technical justification existed for the deletion of these
points other than to meet the method QC criteria. (2) Another such practice involves the evaluation of the
continuing calibration verification solution. Laboratories have been observed averaging the % difference or %
drift across all target analytes even when several of the target analytes exceed the criteria by a significant
amount such that the average still meets the criteria as stated in the method. (3) Another such practice involves
the reporting of acceptance ranges for surrogates or laboratory control samples. Laboratories have been
observed reporting a very tight range indicating that they have good method control. However, an examination of
the actual control charts maintained by the laboratory shows a significantly wider range. (4) Another such
practice involves the determination of the method detection limit (MDL). Laboratories have been observed
running ten or more standards and then discarding points to achieve a lower MDL. No technical justification
existed for the deletion of these points other than to achieve a lower MDL.
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Should the above 'questionable practices' be considered as examples of fraudulent activities? Some of the
laboratories have described these practices as 'the common approach used by everyone.' Yet when described to
people within EPA, the clear response is that these approaches were never intended. The background history will
be discussed as to how and why these practices have developed and what can be done to overcome them.
MODIFICATIONS TO SW-846 HPLC METHODS 8330 AND 8310
C.A. Weisberg. M.L. Ellickson
U.S. EPA, Region III, Office of Analytical Services and Quality Assurance,
839 Bestgate Road, Annapolis, MD 21401
Method 8000 of SW-846 allows for modification of the existing methodology, provided that the analyst has
documented the ability to generate acceptable results by successfully performing an initial demonstration of
proficiency with the altered conditions prior to sample analysis. Analysts may vary the extraction procedures
and/or HPLC parameters (mobile phase composition, elution program, injection volume, etc) in order to improve
sensitivity and chromatographic resolution, or to reduce interferences resulting in coelution with analytes of
interest. The use of analytical columns different from those specified in the methods may require that different
HPLC conditions and flow rates be used. Generally, the HPLC methods 8330 and 8310 have been found to
perform acceptably as written. However, some method modifications need to be employed in order to improve
analyte resolution, achieve desired quantitation limits and obtain definitive confirmation of the primary column
results. Some of the modifications to Method 8330 and Method 8310 reported in this paper include: nitrogen
blow-down to adequately concentrate the final acetonitrile extracts after the salting-out liquid-liquid extraction
procedure or the solvent exchange procedures; reduced flow rates and gradient elution schemes for the C-18
primary columns; and, the use of an acetonitrile/water mobile phase for the cyano secondary column
confirmations.
When reporting quantitated results for these analyses, it is more important to positively confirm the primary
column results than it is to completely resolve all of the target compounds. The use of analytical columns
different from those specified can produce coeluting pairs which may differ from those mentioned in the
methods. This paper will also discuss the reporting of coeluting compounds as mixtures, and the use of
confirmatory techniques (e.g. dissimilar secondary column analyses, spectral library matching, LC/MS, GC/MS
and GC-ECD).
PCB SEPARATIONS USING 2 DIMENSIONAL GC FOR CONFIRMATION:
USE OF A HEART-CUTTING SWITCHING VALVE
Dennis R. Gere and Allen Broske
Hewlett-Packard, 2850 Centerville Rd., Wilmington, DE 19808
Phone (302) 633-8162
email dennis_gere@hp.com
Russell Kinghorn
SGE International Pty. Ltd., 7 Argent Place Ringwood, Victoria 3134 Australia
Environmental gas chromatography analysis continues to be a difficult but necessary effort for the detection and
identification of PCBs. One approach is the use of a very good separation column, which would resolve most if
not all of the 140-150 key congeners from one another. Another approach uses a more user-friendly approach to
the congener specific analysis of PCBs. This approach is not a new or novel approach, but a revisit of a variation
of a general column-confirmation column method. The use of two different stationary phase columns
(orthogonal) allows a geometric gain in separation power compared to the algebraic gain by making the column
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
longer or more efficient.
The approach we used involves a heart cutting switching valve, one injector and two detectors. Some discussion
will be given concerning the choice of the stationary phases, the choice of carrier gas and especially the judicious
combination of temperature programming rate and carrier linear velocity. Focus will be upon key critical PCB
congeners as described by the European Union protocols, that is, congener specific detection and confirmation.
The configuration uses two columns and two detectors, with a valve between the two detectors and a second
selective column before the second detector. In the initial analysis, the sample is nominally separated on a
general-purpose column. As sets of unresolved PCBs elute, they are cut out of the first separation scheme and
sent to the second more selective column. The front of the second column may be cooled cryogenically (we will
also demonstrate what happens when the cooling of the second column is not used) to hold the analytes at the
front of the second column until the analysis of the first column is complete. Then a second temperature program
is begun to resolve previously unresolved pairs.
The first column is a 5 % phenyl phase and the second column we used was either a 35 % phenyl or a 50 %
phenyl phase of intermediate polarity. Either column is useful, and a specific choice may be made depending
upon the sensitivity level (the 35 % phenyl is a much lower bleed column) or whether additional selectivity is
needed. There are a great many permutations of this approach possible, and we will outline what we did and
what can be done if even greater resolution is required.
ESTIMATING THE TOTAL CONCENTRATION OF VOLATILE ORGANIC COMPOUNDS
IN SOIL AT SAMPLING LOCATIONS: FIELD TRIALS
Alan D. Hewitt
U.S. Army Cold Regions Research and Engineering Laboratory, 72 Lyme Road, Hanover, NH 03755-1290
(603) 646-4388
Martin H. Stutz
U.S. Army Environmental Center, Aberdeen Proving Ground, MD 21010-5401
(410)612-6856
ABSTRACT
This report describes a method for estimating the total concentration of volatile organic compounds (VOCs) in
soil relative to a site-specific 0.2-mg/kg working standard and presents the results from three separate field trials.
This method was developed to provide a decision tool for field or laboratory personnel so they can implement the
appropriate soil sample preparation procedure for the selected method of instrumental analysis. Coupling a rapid
method for estimating the total VOC concentration with sample collection, handling, and preparation procedures
that limit substrate dissaggregation and exposure complements efforts to achieve site-representative estimates
for vadose zone contamination.
INTRODUCTION
Since the beginning of the Superfund and the Resource Conservation and Recovery Act (RCRA) programs, gas
chromatography/mass spectrometry (GC/MS) (via Methods 8260 and 8240) has served as the major laboratory
instrument for identifying and quantifying VOCs in soils1 The principal reason for the selection of this analytical
detection system is that it provides an unambiguous identification of analytes present. Unfortunately, this very
desirable quality comes with the limitation that for quantification purposes the individual analytes must fall within
a concentration range of 2 to 3 orders of magnitude. High analyte concentrations can degrade the performance
of the MS detection system, which interrupts scheduled runs and may lead to expensive instrument repairs.
Therefore, one of the challenges when using an MS is how to couple it with a sample collection, handling, and
preparation protocol when analyte concentration can range over 7 orders of magnitude (percent levels to the
current levels of instrumental detection, approximately 0.005 mg/kg). To cope with this concern, samples thought
to be contaminated with VOCs at levels greater than 0.2 mg/kg are prepared by extraction (and perhaps further
dilution) with methanol (MeOH), i.e., the high-level method. In contrast, samples thought to have concentrations
less than 0.2 mg VOC/kg are analyzed directly, which is referred to as the low-level method. Many other
commonly used laboratory instruments and their respective methods for VOC detection (e.g., Methods 8015 and
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
8021B) also benefit from using these two approaches to sample preparation.
A second challenge is that VOCs in soils fail to maintain their concentration integrity if they are not collected and
handled with limited disruption and exposure and if preventive measures are not taken to limit biological
degradation of aromatic compounds. Today it is generally recognized that the sample collection and handling
guidance, provided in the past by Method 5030, often resulted in a greater than 90% loss of the VOCs from soil
samples prior to laboratory analysis2-6. To minimize losses due to volatilization and biodegradation, new sample
collection and analysis protocols were included in the third update of SW-8461: Method 5035, "Modified
purge-and-trap and extraction for volatile organics in soil and waste samples," and Method 5021, "Volatile
organic compounds in soils and other solid matrices using equilibrium headspace analysis."
The two most effective collection and handling protocols that can be used with these new methods for preventing
the loss of VOCs are 1) the on-site, rapid transfer of discrete samples with a small coring tool to a vessel that
hermetically seals and already contains the appropriate dispersion/extractant solution for the chosen method of
analysis7, or 2) obtaining and temporarily storing (two days at 4ฐC) a sarqple in an En Core (En Novative
Technologies, Inc., 1241 Bellevue St., Green Bay, Wise. 54302) sampler before transferring it into an
appropriately prepared vessel8. In addition, it should be recognized that, if the sample is to be held for more than
two days before analysis, then some form of chemical preservation may be necessary in addition to storage at
4ฐC. For example, acidification can be used for low-level sample preparation procedures when carbonates are
not present6
Because there is often no a priori knowledge of the VOC concentrations at a given location, the data quality
objectives for site characterization activities often require that samples be collected and prepared for both the
lowand high-level analysis procedures. To avoid collecting and processing samples through both of these
preparation procedures for every location, it has been suggested that a rapid screening analysis be performed to
establish an estimate for the total VOC concentration9 This screening indicates the levels of VOCs to expect,
before the sample is prepared for analysis, and thus whether collocated sample(s) taken for laboratory analysis
should be prepared using the low- or high- level procedures, or both. The method developed is based on the
comparison of responses of a hand-held photoionization detector (PID) to a sample relative to a 0.2-mg VOC/kg
site-specific working standard. Recognition of the potential effort and cost savings by using screening as a
decision tool are two reasons why this method is being considered for inclusion in the fourth update of SW-846
(proposed Method 3815). This paper briefly outlines this screening method and presents the results from three
case studies. Additional information concerning the development of this method for screening is available
elsewhere9
SCREENING METHOD
Materials
The necessary equipment and reagents are as follows:
1) Modified VOA vials (40 or 44 ml), Teflon-lined septa with 5- to 6-mm hole
punched through the middle and 3- x 3-cm squares of light-gauge
aluminum foil for temporary covers (see Fig. 1).
2) Coring tool for the collection and transfer of discrete soil samples, e.g.,
disposable 10-mL plastic syringes with the Luertip and rubber plunger cap
removed, or an equivalent metal tube and plunger.
3) A portable photoionization detector (PID) analyzer with a 10.6-eV or greater
electrode discharge tube, digital display, inlet flow rate greater than 200
mL/min, and sample inlet tube of 3 to 4 mm o.d. and at least 3 cm in
length.
4) A 10-mL liquid syringe.
5) Reagent-grade water (i.e., water with no detectable VOCs), polypropylene
glycol (PPG, or similarly low-vapor-pressure organic solvent), and principal
VOC(s) of site interest.
6) A cylinder of calibration gas for the PID, e.g., 100 ppm isobutylene.
Figure 1. Modified VOA vials for rapid total VOC screening of soil samples.
Screw Cap
Teflon-lined septum
with access hole
Standard or soil sample
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Standards
A stock standard is prepared by transferring the VOC of interest into PPG. The stock standard concentration
should be based on the density of the analyte of interest, so that a 1- to 3-uL volume transferred to a 40-mL VOA
vial containing 10 mL of reagent water and 10 g of the site-specific soil matrix results in a 0.2-mg-VOC/kg
working standard. For example,
Stock standard: 1.34 g/mL* x 2.0 ul_/2.5 mL -1.1 mg TCE/mL
Working standard: 1.1 mg TCE/mL x 1.8 uL/10 g soil = 0.2 mg TCE/kg
*nfinsit\/ nf TCP
*Density of TCE
Immediately after spiking, these working standard vials are covered with a single sheet of aluminum foil that is
held tightly in position with a septum with a hole punched in the middle and a screw cap (Fig. 1). The vial
contents of the working standards should be thoroughly mixed by handshaking, then transported to the location
of the sampling activity, stored out of direct sunlight, and allowed to equilibrate for 1 hr prior to use. Working
standards should be prepared daily.
The PID response to the working standard should be at least 10x greater than its response to a blank (reagent
water, contamination-free site-specific matrix, and appropriate volume of PPG). For analytes with high vapor
pressures or low octanol water partition coefficients, or both, and soil matrices with low organic carbon contents,
it may not be necessary to include the site-specific soil matrix in the working standards. This should be
established on a site-by-site basis by comparing the means of triplicate working standards with and without the
soil matrix. As a general rule, if the means differ by more than 20%, it is recommended that the soil matrix be
included in the working standards.
Sample Collection and Analysis
Before field sampling, 10 mL of reagent water is added to the modified VOA vials. Once prepared, the VOA vials
for screening samples should be transported to the sampling location and stored with the working standards until
they are used. The native structure of the material being sampled for screening should be kept intact, thus
experiencing as little disaggregration as possible during the collection and transfer process. This can often be
accomplished with a coring tool designed to obtain a discrete sample. For example, a modified 10-mL syringe is
a practical tool for obtaining up to a 10-g soil sample. If 10 g cannot be easily obtained in a single transfer, more
than one corer can be used, or a couple of transfers with a single corer can be made. This coring device is
transparent and comes with gradient markings so the volume/weight relationship for a given material can easily
be established with a portable balance. The location of samples taken both for screening purposes and for
laboratory analysis should be as close as possible to each other (generally within a 10-cm radius) and from the
same stratum. Before preparing (or exposing) a fresh sampling surface, for instance, opening a split spoon or
scraping away the top layer of a material, the cap and aluminum foil should be removed from the screening VOA
vial. After retrieving a discrete sample, the core barrel should be inserted into the mouth of the screening VOA
vial and the sample extruded. Once the sample has been extruded, the aluminum foil and cap should
immediately be replaced on the vial. This collection and transfer process should take less than 10 seconds, and
the sample weight only has to approximate 10 g (plus or minus 2 g).
Before a working standard or sample is analyzed, the VOA vial should be shaken by hand for 10 to 15 seconds.
Cohesive materials, such as silts and clays, do not break apart rapidly upon shaking and may require more than
15 seconds for complete dispersion. The vial is then visually checked both for the complete dispersion of the
sample matrix and for particles adhering to the aluminum foil cap liner (knock large particles off the aluminum
foil if present). Then the inlet tube of the PID is pushed through the foil liner to a set position about 3 cm below
the rim. A maximum response will be achieved within 2 to 3 seconds of punching through the foil liner. The
maximum response for each sample screened and for the analysis of each working standard should be recorded.
Daily Operating Procedure for VOC Screening
The PID should be initially calibrated with a cylinder of standard gas (e.g., 100 ppm isobutylene) at the beginning
of each day. This task can be performed before going to the sampling location. However, both the analysis of
site-specific working standards and the screening of a sampling location should be performed under the same
conditions, thereby normalizing meteorological influences on the performance of the PID. Site-specific working
standards should be prepared daily and in sufficient quantity to satisfy the study's objectives. At a minimum, one
working standard should be analyzed for every hour of site activity.
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Collection of samples for VOC analysis should always be the first operation performed after a surface to be
sampled has been exposed to the atmosphere. This includes samples both for screening and for laboratory
analysis. To establish how to handle and prepare the discrete sample for laboratory analysis (low, high, or both
procedures), a total VOC screening analysis should be performed at each sampling location. Therefore, before
opening a split spoon, scraping a fresh surface on a pit wall, removing surface vegetation and the appropriate
amount of top soil for a surface grid location, or removing the first several inches of some other type of waste
material, the PID of choice should be operating. Furthermore, if a working standard is being utilized to verify
performance of the PID for the sampling location, the analysis of a working standard should be completed before
exposing a fresh sampling surface.
Once a fresh surface has been exposed, a sample
should be quickly obtained, transferred to a screening
VOA vial, dispersed, and analyzed. If the maximum
response is greater than the working standard (or the
running average), the sample or samples taken for
laboratory analysis should be prepared using the
high-level procedure (i.e., MeOH extraction). If the
maximum response is below the working standard, the
laboratory sample(s) should be prepared using a
low-level procedure. The total elapsed time between
exposing a fresh surface, screening a sample, and
obtaining samples for laboratory analysis should be
less than 2 minutes. As a precaution against false
positive and false negative screening estimates
relative to the decision point, locations where
screening results are between 0.5 and 2x the working
standard response should have samples prepared by
both high- and low-level procedures.
Method Limitations
For this method of sample location screening to work,
the VOC(s) of interest must be detectable by
photoionization. If more than one analyte is of interest,
and there are large discrepancies (greater than a
factor of 2) in photoionization potentials, then the
range around the decision point where samples are
prepared by both high- and low-level procedures
should be increased proportionally. That is, if the
responses for the VOCs of interest differ by a factor of
3x, and the analyte with the highest response is used
to make the working standard, then laboratory samples
from locations where screening results are only 0.3x
the working standard should be prepared by both
procedures. However, this often will not be a problem
for sites contaminated with common chlorinated and
aromatic compounds because they have similar
photoionization potentials. This approach may not be
effective for sample matrices that are not readily
dispersed in water (e.g., some clays and cementitious
materials).
Table 1. Field screening measurements and sample
preparation procedures.
Screening (response)*
No.
Site 1
S1-1
S1-2
S1-3
S1-4
S1-5
S1-6
S1-7
S1-8
S1-9
S1-10
Site 2
S2-1
S2-2
S2-3
S2-4
S2-5
S2-6
S2-7
S2-8
S2-9
S2-10
SiteS
S3-1
S3-2
S3-3
Working Std
6.5
6.5
7.0
8.0
6.3
2.3,2.5
3.2
4.2
3.9
3.9
3.8
4.8
5.6
4.8
5.3, 5.5, 5.2, 4.9
Sample
0.0
0.0
0.4
790
480
1400
26
2.2
12
4.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Sample
Preparation
Method
5021 , LL**
5021, LL
5021, L
5035, HLf
5035, HL
5035, HL
5035, HL
5021, LL
5035, HL
5021 , LL
5035, LL"
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5021, LL
5021 , LL
5021, LL
FIELD TRIALS t
This method for rapidly estimating the total
concentration of VOCs was tested during three tt
different sampling activities performed under the
supervision of personnel from EPA Region 1. At the
sites visited, samples were obtained from near the surface with the aid of hand tools and from split-spoon core
barrels. All samples, whether collected for on-site screening or for off-site analysis, were transferred using a
PID field screening: Working Std-Results of
analyzing a site-specific working standard 0.2 mg
TCE/kg for Site 1 and 0.2 mg PCE/kg for Sites 2
and 3. Sample --Results from rapidly (<30 s)
screening 10ฑ2 g soil.
5021, LL-Sample placed in 22-mL VOA vial
containing 10 mL water.
5035, HL--Sample placed in 40-mL VOA vial
containing 5 mL MeOH.
5035,LL--Sample placed in 40-mL VOA vial
containing 5 mL water.
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
modified syringe. Samples collected for off-site analysis were placed into vials containing methanol or
organic-free water, as appropriate for the intended method of sample preparation (Method 5035 or 5021) and
analysis (Methods 8260, 8015, 8021), and analyzed within 48 hours.
The results of the screening analysis for both the working standards and samples are shown in Table 1, and the
results of the laboratory analysis are shown in Table 2. During these field trials, the screening results were only
used to decide whether to prepare samples by a low- or high-level procedure. With the possible exceptions of
sampling locations S1-3 and S1-10, the samples were prepared appropriately for the intended method of
analysis. That is, the analysis system was not exposed to an unexpectedly high analyte concentration. Indeed,
this statement applies to all of the samples, since the concentrations were not much greater than 0.2 mg/kg for
the individual analytes found in these two samples. Therefore, a scheduled run would not have been delayed nor
would the detector have been damaged; however, there may have been individual analyte responses greater
than the highest calibration standard.
Table 2. Laboratory results.
Sitel.
HS/GC-PID/FID (mg/kg)
No.
S1-1
S1-2
S1-3**
S1-4**
S1-5
S1-6
S1-7
S1-8
S1-9
S1-10
Site 2.
No.
S2-1
S2-2
S2-3
S2-4
5
6
7
8
9
10
CDCE TCE Tol
0.021 0.15
5.2 29
250 19
0.040 0.020
PT/GC/MSf
Trichlorofluoromethane
0.004
0.001
0.002
PCE
<0.003*
<0.003
0.008
140
240
7.4
0.17
1.0
0.24
(mg/kg)
PCE
0.001
0.001
...
EBen
<0.003
0.074
15
33
...
...
...
p/m-Xyl
<0.003
0.076
15
76
...
Site 3.
o-Xyl
<0.009
0.082
3.5
68
...
Total
<0.003
0.41
68
140
690
7.4
0.23
1.0
0.24
HS/GC-PID/FID
Total
ND
0.001
0.001
0.004
ND
ND
0.001
0.002
ND
ND
(mg/kg)
No.
1.
2.
3.
PCE
<0.003
Total
<0.003
ND
ND
<0.003-Peak identified but below quantation.
** Unidentified peaks present in chromatogram.
T Samples analyzed at EPA Region 1 Laboratory in Lexington, Mass.
In the case of sample S1-10, the recommendation that samples be processed through both procedures when
sample screening results are within a factor of 2 to 0.5 of the working standards would have provided the
necessary precaution However, in the case of S1-3, where the screening results were well below 0.5x the
working standard, this same logic would not have succeeded. Samples S1-3 and S1-4 were taken within 5 cm
(vertical) of one another, and both the screening and laboratory results (Table 2) showed that this area had a
large vertical gradient in VOC concentrations. In this case, a review of the data and perhaps the site history
would alert the sample collectors to a potential problem and therefore the need to implement an additional
precaution so as not to overload the analytical system. Two potential solutions that would have worked at the
site where'samples S1-3 and S1-4 were taken are 1) to have taken screening samples on either side of Oust
above and below) the sample taken for laboratory analysis, 2) or alternatively, to have automatically prepared
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
samples by both low- and high-level procedures based on knowledge of where the source regions were on this
site. Because of this experience, an additional recommendation is to use one of these two procedures (e.g.,
bracketing laboratory samples with screening samples or taking laboratory samples for both low- and high-level
preparation procedures) when sampling near known or suspected source regions. Whenever samples are
prepared by both a low- and high-level procedure, the sample prepared by the high-level method should be
analyzed first. Furthermore, although not reported here, it has become evident that when a screening exceeds
the scale of the PID, which is typically greater than 2000 ppm, further dilution of a sample processed by the
high-level procedure is most likely warranted prior to analysis.
SUMMARY
The problems of underestimating the concentration of VOCs in samples taken from the vadose zone has
facilitated the acceptance of new sample collection, handling, and preparation protocols (e.g., Methods 5035 and
5021). These changes not only present challenges to field sampling teams but to the laboratories responsible for
sample analysis as well. For this reason both parties must be involved in the initial design of the sample
collection plan and remain in contact throughout the project. To assist in deciding how samples should be
prepared for instrumental analysis, a simple total VOC screening procedure has been developed. The main
purpose of this screening method is to provide a decision tool during the sampling activity to help establish
whether samples taken for laboratory analysis should be prepared by a low-level or a high-level procedure, or by
both. This screening process is, however, not foolproof. Likewise, neither are any others that must contend with
the possibility of a heterogeneous analyte distribution. For this reason, there are additional precautions that
should be taken when using this method. One is for the case when screening results for samples are within a
factor of 2 to 0.5 of the working standard results, another is for the case when sampling near known or suspected
source regions, and the third is for when the PID's response to a screening sample is over range (greater than
2000 ppm). As demonstrated here, this screening procedure has the potential to greatly reduce the number of
samples that would have to be collected and processed during a site investigation.
ACKNOWLEDGMENTS
Funding for this work was provided by the U.S. Army Environmental Center, Martin H. Stutz, Project Monitor,
and from the Advanced Site Monitoring System AF-25 program, Dr. James Brannon, Project Monitor. The
authors thank Louise Parker and Thomas Ranney for critical review of the text.
This publication reflects the view of the author and does not suggest or reflect policy, practices, programs, or
doctrine of the U.S. Army or of the government of the United States.
REFERENCES
1 U.S. Environmental Protection Agency (1986) Test Methods for Evaluating Solid Waste. Vol. 1B. SW-846.
2. Urban, M.J., J.S. Smith, E.K. Schultz, R.K. Dickinson (1989) "Volatile organic analysis for a soil, sediment or
waste sample," in Proceedings of the 5th Annual Waste Testing & Quality Assurance Symp., U.S.
Environmental Protection Agency, Washington, DC, pp. 11-87-11-101.
3. Illias, A.M., C. Jaeger (1993) "Evaluation of sampling techniques for the analysis of volatile and total
recoverable petroleum hydrocarbons (TRPH) by IR, GC, and GC/MS methods," in Hydrocarbon
Contaminated Soils, Chelsea, Michigan: Lewis Publishers. 3: 147-165.
4. Lewis, T.E., A.B. Crockett, R.L. Siegrist, K. Zarrabi (1994) Soil sampling and analysis for volatile organic
compounds. Environmental Monitoring and Assessment, 30: 213-246.
5. Hewitt, A.D., T.F Jenkins, C.L. Grant (1995) Collection, handling and storage: Keys to improved data quality
for volatile organic compounds in soil. American Environmental Laboratory, 7(1): 25-28.
6. Liikala, T.L., K.B. Olsen, S.S. Teel, D.C. Lanigan (1996) Volatile organic compounds: Comparison of two
sample collection and preservation methods. Environmental Science and Technology, 30:3441-3447.
7. Hewitt, A.D., N.J.E. Lukash (1996) Sampling for in-vial analysis of volatile organic compounds in soil.
American Environmental Laboratory, 7(8): 15-19.
8. Hewitt, A.D. (1998) A tool for the collection and storage of soil samples for volatile organic compound
analysis. American Environmental Laboratory, 9(10): 14-16.
9. Hewitt, A.D., N.J.E. Lukash (1997) Estimating the total concentration of volatile organic compounds in soil: A
decision marker for sample handling. Special Report 97-12. U.S. Army Cold Regions Research and
Engineering Laboratory, Hanover, New Hamsphire.
10
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
FINAL EVALUATION OF METHOD 3546: A MICROWAVE-ASSISTED PROCESS (MAP)* METHOD FOR
THE EXTRACTION OF CONTAMINANTS UNDER CLOSED-VESSEL CONDITIONS
J.R. Jocelyn Pare and Jacqueline M. R. Belanger
Microwave-Assisted Processes Division
Chung Chin and Richard Turle
Air Analysis and Quality Division, Environment Canada,
Environmental Technology Centre, Ottawa, ON, Canada K1A OH3
Barry Lesnik
US EPA, Office of Solid Waste, Method Section, 401 M. St., SW, Washington, District of Columbia 20460, USA
Rodney Turpin and Raj Singhvi
US EPA, Emergencies Response Branch, 2890 Woodbridge Ave., Edison, New Jersey 08837-3679, USA
ABSTRACT
Microwave-assisted extraction (MAP) has been the subject of enhanced interest from the environmental sector
in the past year as a result of the need for methodologies that will improve sample preparation without
compromising the quality of the data while being sustainable environmentally. Liquid-phase microwave-assisted
extraction (MAP) offers such advantages: it is a very fast extraction technique, it consumes less solvent and
energy, and it is cost effective. A preliminary validation study involving closed-vessel apparatus and
contaminants such as PAHs, PCDDs/PCDFs, chlorinated pesticides, and PCBs was performed. Excellent
performance and precision were achieved for these analytes. In order to fully evaluate the method for the range
of analytes an inter-laboratory study was performed. A round-robin study was performed with five laboratories
and involved thermally labile RCRA target analytes such as phenols, phenoxyacids herbicides and
organophosphorus pesticides. Three split samples were used along with a single standard operational procedure
(SOP). All analyses were performed by a single laboratory in order to minimise the variability of the results due
to the determinative procedure. Clean up was performed using standard procedures and analysis was done
according to the appropriate SW-846 methods. The broad range of applicability, the reduced sample preparation
time and the reduced amount of solvent used all contribute to reach sustainable environmental protection goals.
Furthermore, the reduced operational costs associated with the protocol compared to conventional Soxhlet for
example - are significant and will prove valuable in these times where the "greening" of the laboratory usually
gives rise to higher operating costs. Further work involving open-vessels apparatus is under way.
INTRODUCTION
The microwave-assisted process (MAP) is a technology patented by Environment Canada1'3 The most widely
used applications to date make use of microwave energy to extract soluble materials from different matrices,
mostly using organic solvents45. Microwave energy has been used in various ways to extract organic compounds
from a variety of matrices6. For example, the technology has been applied to organochlorinated pesticides from
sediments and PCBs from water7, petroleum hydrocarbons from soil8, and to herbicides from soils910.
Lopez-Avila et a/, used a MAP-approach to extract several groups of pollutants such as PAHs, PCBs, pesticides,
phenols and base/neutral compounds in soils and sediments1113. In all these studies, microwave-assisted
extraction proved to be similar or more efficient than methods based upon the use of Soxhlet or ultrasound.
More recently, we reported on a preliminary validation of a draft method for inclusion into US EPA Test Methods
for Evaluating Solid Waste Physical/Chemical Methods (SW-846)14 Table 1 presents the types of compounds
that have been subjected to that preliminary work.
The method reported herein makes use of partially microwave-transparent solvents (or a mixture of such
solvents) contained into a closed vessel. Although not as elegant and efficient as methods using open-vessel
microwave-transparent solvents, it provides the possibility of combining the benefits of heat (enhanced solubility
and diffusivity) to the action of the microwaves on the matrix. Work is currently underway to validate open-vessel
methods and the results will be reported elsewhere in due time along with the parameters to be controlled to
effect even more efficient extraction procedures using MAP. This paper reports on the final evaluation of a
closed-vessel microwave-assisted extraction procedure for environmental pollutants from soils and sediments
that recently met with the US EPA approval and will be referred to as Method 3546 under SW-846. It is a
procedure for extracting water insoluble or slightly water soluble organic compounds from soils, clays, sediments,
* MAP is a Trade-mark of Her majesty The Queen in Right of Canada as represented by the Minister of the
Environment.
11
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
sludges, and other solid wastes. The method is applicable to the extraction of semi-volatile organic compounds,
organo-phosphorus pesticides, organo-chlorine pesticides, chlorinated herbicides, phenoxy acid herbicides,
substituted phenols, PCBs, and PCDDs/PCDFs which may then be analyzed by a variety of chromatographic
procedures.
TABLE 1. List of target analytes used in preliminary validation package
Base Neutral and Acid PAH Organo-chlorines
Hexachloropentadiene Acenapthylene PCB-I
Dimethylphthalate d10-Acenaphtene (surrogate) PCB-2
Diethylphthalate Fluorene Hexachlorobenzene
Di-n-butylphthalate dio-Phenantrene (surrogate) Simazine
Butylbenzylphthalate Phenanthrene Atrazine
ฃ>/'s(2-Ethylhexyl)adipate Anthracene Lindane
b/s(2-ethylhexyl)phthalate Pyrene PCB-3
Benzo(a)anthracene Alachlor
Pentachlorophenol dn-Chrysene (surrogate) Heptachlor
Chrysene PCB-4
Benzo(b)fluoranthene Heptachlor epoxide
Benzo(k)fluoranthene PCB-5
Benzo(a)pyrene gam/na-Chlordane
d12-Perylene (surrogate) a/prta-Chlordane
lndeno(123-cd)pyrene frans-Nonachlor
Dibenzo(ah)anthracene PCB-6
Benzo(ghi)perylene Endrin
PCB-7
di4-Terphenyl (int. std) Methoxychlor
PCB-8
This method has been validated for solid matrices containing 50 to 10,000 ug/kg of semi-volatile organic
compounds, 250 to 2,500 ug/kg of organo-phosphorus pesticides, 10 to 5,000 ug/kg of organo-chlorine pesticides
and chlorinated herbicides, 50 to 2,500 ug/kg of substituted phenols, 100 to 5,000 ug/kg of phenoxy acid
herbicides, 1 to 5,000 ug/kg of PCBs, and 10 to 6000 ng/kg of PCDDs/PCDFs.
EXPERIMENTAL
The experimental procedures for the preliminary validation work has been presented elsewhere, hence all the
text presented herein refers exclusively to the inter-laboratory work and is relevant to Method 3546 as approved.
This method is applicable to solid samples only with small particle sizes. If practical, soil/sediment samples may
be air-dried and ground to a fine powder prior to extraction. Alternatively, if worker safety or the loss of analytes
during drying is a concern, soil/sediment samples may be mixed with anhydrous sodium sulfate or pelletised
diatomaceous earth. The total mass of material to be prepared depends on the specifications of the
determinative method and the sensitivity required for the analysis, but 2 - 20 g of material are usually necessary
and can be accommodated by this extraction procedure.
Safety
The use of solvents combined with the operational parameters associated with this method will raise
temperatures and pressures in the extraction vessels to values that can be a safety concern in the laboratory.
Only equipment designed for laboratory use and manufactured under legitimate rights should be used to ensure
that proper safety devices are built into the apparatus. Common sense laboratory practices can be employed to
minimize this concern. For example, the following sections describe some additional steps that should be taken.
The extraction vessels are at elevated temperatures and pressure after the extraction stage. Allow the vessels to
cool before opening (the use of a water bath is recommended for this purpose) and always monitor the
temperature and pressure by re-connecting the control vessel to the apparatus prior to opening the vessels.
During the heating step, some solvent vapors may escape through the vessel liner/seal cover. Follow the
12
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
manufacturer's directions regarding the vessel assembly and instrument set up to prevent release of solvent
vapors to the laboratory atmosphere. The instrument may contain flammable vapor sensors and should be
operated with all covers in place and doors closed to ensure proper operation of the sensors. If so equipped,
follow the manufacturer's directions regarding replacement of extraction vessel seals when frequent vapor leaks
are detected.
Extraction
Decant and discard any water layer on a sediment sample. Mix the sample thoroughly, especially composite
samples. Discard any foreign objects (sticks, leaves, rocks, etc.). Air dry the sample at room temperature for 48
hours in a glass tray or on hexane-rinsed aluminum foil. Alternatively, mix the sample with an equal volume of
anhydrous sodium sulfate or pelletised diatomaceous earth until a free-flowing powder is obtained.
If multiphase waste samples are used, then they must be prepared by the phase separation method in Chapter
Two of SW-846 before extraction. Dry sediment/soil and dry waste samples amenable to grinding. Grind or
otherwise reduce the particle size of the waste so that it either passes through a 1-mm sieve or can be extruded
through a 1-mm hole. Disassemble grinder between samples, according to manufacturer's instructions, and
decontaminate with soap and water, followed by acetone and hexane rinses.
Gummy, fibrous, or oily materials not amenable to grinding should be cut, shredded, or otherwise reduced in size
to allow mixing and maximum exposure of the sample surfaces for the extraction. The analyst may add
anhydrous sodium sulfate, pelletised diatomaceous earth, sand, or other clean, dry reagents to the sample to
make it more amenable to grinding.
Grind a sufficient weight of the dried sample to yield the sample weight needed for the determinative method
(usually 10-30 g). Grind the sample until it passes through a 10-mesh sieve. Prepare a method blank using an
aliquot of a clean solid matrix such as quartz sand of the approximate weight of the samples. Add the surrogates
listed in the determinative method to each sample and method blank. Add the surrogates and the matrix spike
compounds appropriate for the project to the two additional aliquots of the sample selected for spiking.
A volume of about 30 ml_ of the appropriate solvent system is added to the vessel and sealed. The extraction
vessel containing the sample and solvent system is heated to the extraction temperature and extracted for 10
minutes. The solvent systems used for this procedure vary with the analytes of interest and are listed below. The
mixture is allowed to cool. The vessel is opened and the contents are filtered. The solid material is rinsed and the
various solvent fractions are combined. The extract may be concentrated, if necessary, and, as needed,
exchanged into a solvent compatible with the cleanup or determinative step being employed.
Six vessels were always placed in the microwave oven at any one time to standardise conditions. After
extraction, the sample carousel was removed from the microwave and cooled in a water bath. To ensure that it
was safe to proceed with the filtration step the control vessel was returned to the microwave oven and the
temperature was monitored before opening. Solvent loss was checked randomly in some instances and found to
be below 1%.
Interferences
Refer to Method 3500 of SW-846. If necessary, Florisil and/or sulfur cleanup procedures may be employed. In
such cases, proceed with Method 3620 and/or Method 3660 of SW846.
Apparatus and Supplies
CEM Corporation (Matthews, NC) MAP solvent extraction systems equipped with appropriate
microwave-transparent extraction vessels should be transparent to microwave energy and capable of
withstanding the temperature and pressure requirements (200ฐC and 200 psi) for this procedure. Models
MES-1000 or MSP-1000 have been used for the present work.
Solvents Systems and Reagents
All solvents should be pesticide quality or equivalent. Solvents may be degassed prior to use.
Organo-chlorine pesticides, organo-phosphorus pesticides, semi-volatile organics may be extracted with
acetone/hexane (1:1, v/v) or acetone/methylene chloride (1:1, v/v).
13
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
e extracted with acetone/hexane (1:1, v/v), or acetone/methylene
<
chloride (1:1, v/v), orhexane.
Phenoxy acid herbicides and phenols may be extracted with acetone/hexane (1-1 v/v) and the phosphate buffer
solution.
Reagent grade chemicals shall be used in all tests. Organic-free reagent water should be used. Sodium sulfate
(granular anhydrous), Na2SO4 and pelletised diatomaceous earth can be used as desiccant. They should be
purified by heating at 400ฐC for 4 hours in a shallow tray, or by extraction with methylene chloride. If the latter
approach is used, then a reagent blank should be prepared to demonstrate that the drying agent is free of
interferences.
Phosphate buffer solution - for use in extraction of phenols and phenoxyacid herbicides. Prepare a 0.1 M
phosphate buffer solution by adding 1.2 g reagent grade sodium phosphate into a 250-mL beaker, add 100 ml of
reagent water and thoroughly mix. Adjust the solution pH to 2 with the addition of reagent grade phosphoric acid
Quality Control
Chapter One and Method 8000 of SW-846 should be followed for specific Quality Control procedures and
Method 3500 should be followed for sample preparation quality control procedures. Surrogate standards should
be added to samples when listed in the appropriate determinative method.
RESULTS AND DISCUSSION
Reference 14 presents a large body of information and specific data on a number of analytes. It provides the
basis for a major portion of the performance work associated with this procedure. References 12 and 15 are
reports of similar, more specific studies. References 16 to 18 deal specifically with phenols. Representative data
sets are presented in Tables 2 to 6. They are not exhaustive and are reported here as they are new data. Other
data can be found in the references cited herein including references 19 and 20.
Chlorinated pesticides: Single-laboratory accuracy data were obtained for chlorinated pesticides using natural
soils, glass-fiber, and sand matrices. Concentrations of each target analyte ranged between 0.5 to 10 ug/g. Four
real-world split samples contaminated with pesticides and creosotes were also used (obtained from US EPA
ERT, Edison, NJ). The latter were extracted by an independent laboratory using standard Soxhlet procedures
and results compared to those obtained with this procedure. Extracts were analyzed by the appropriate method.
Method blanks and five spiked replicates were included. Work was also carried out to assess the level of
degradation of thermally labile pesticides and it was found that no significant degradation takes place under the
procedure described herein. The data are reported in detail in Reference 4. Data summary tables are included in
Method 8081 .
TABLE 2. Single-laboratory organochlorine pesticides analysis data from a real contaminated soil _
Compounds Average Std. Dev. RSD n REAC value
(ppb) _ (%) _ (ppb)
DDE+Dieldrin
Endrin
*DDD
*DDT
*Methoxychlor
a-Chlordane
Y-Chlordane
3380
21500
40000
62670
16500
730
720
340
2290
5750
8430
1980
100
90
10.06
10.66
14.38
13.45
12.03
13.37
12.47
3
3
3
3
3
3
3
7100
22000
45000
62000
16000
750
910
"(dilution 1:5); Soil samples obtained from US EPA Emergency Response Center archive bank through their
contract laboratory REAC (Edison, NJ). The standard Soxhlet extraction procedures were performed by REAC
three years earlier; this long storage period is believed to account for the low DDE+Dieldrin recovery data in the
present study. DDE+Dieldrin is the sum of the compounds since they were not resolved by chromatography.
Semivolatile orqanics: Single-laboratory accuracy data were obtained for semivolatile organics using natural
soils, glass-fiber, and sand samples. Concentrations of each target analyte was about 0.5 ug/g. Extracts were
analyzed by the appropriate method. Method blanks and five spike replicates were included. The data are
reported in detail in Reference 14. Data summary tables are included in Method 8270.
14
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
PAHs: Single-laboratory accuracy data were obtained for PAHs using five reference materials comprising marine
sediments (HS-3, HS-4, and HS-5, all from the National Research Council of Canada), lake sediments
(SRM-1491, from the National Institute of Science and Technology), and a natural soil contaminated with
creosote (SRS103-100, from Fisher Scientific, Fairlawn, NJ). Natural soils, glass-fiber, and sand samples were
also used in spiked matrices work. Concentrations varied between 0.1 and 2000 ug/g. One real-world split
sample contaminated with creosote and pesticides was also used (obtained from US EPA ERT, Edison, NJ). The
latter was extracted by one laboratory using standard Soxhlet procedures and results compared to those obtained
with this procedure. Extracts were analyzed by the appropriate method. Method blanks, spikes and five spiked
replicates were included. Surrogates were used in real-world split sample. The data are reported in detail in
Reference 14. Data summary tables are included in Method 8270.
PCBs: Single-laboratory accuracy data were obtained for PCBs using three reference materials EC-1, EC-2,
EC-3 (from Environment Canada). Natural soils, glass-fibre, and sand samples were also used in spiked matrices
work. Concentrations varied between 0.2 and 10 ug/g (total PCBs). Extracts were analyzed by the appropriate
method. Method blanks, spikes and spike duplicates were included for the low concentration spikes; matrix
spikes were included for all other concentrations. The data are reported in detail in Reference 14. Data summary
tables are included in Method 8082.
TABLE 3. Single-laboratory PCB recoveries data from certified Great Lake sediment materials
Sediment Aroclor Std Dev. RSD nCertified value
(ppb) (%) (ppb)
EC-1
EC-2
EC-3
1850
1430
670
0.07
0.09
0.02
3.78
6.60
3.12
3
4
3
2000 ฑ 54
1160 ฑ70
660 ฑ 54
Sample size = 2 g extracted into a final volume of 4 ml; EC-2 and EC-3 certified values were provisional values
only, at the time the work was conducted. The data presented herein were part of the validation data package
used to confirm the certified values. Real samples were also tested when fortified with mixtures of native Aroclor
(1242, 1254, and 1260) to a 600 ppb level. Recoveries were in the 88% range with a reproducibility of 2% RSD.
Chlorinated herbicides (phenoxyacid herbicides): Multi-laboratory accuracy data were obtained for chlorinated
herbicides spiked at 100 ng/g in one soil type. A certified spiked material was used (obtained from ERA, Arvada,
CO). Extracts were analyzed by Method 8151. Method blanks and three replicates from five laboratories were
included. Data summary tables are included in Method 815 1.
Phenols: Single-laboratory accuracy data were obtained for phenols using a number of spiked natural soils and a
number of real-world split soils. Concentrations varied between 0.2 and 10 ug/g. Extracts were analyzed by the
appropriate method. The data are reported in detail in References 14 to 18. Data summary tables are included in
Method 8041. Multi-laboratory accuracy data were obtained for phenols spiked at 250 ug/kg in one soil type. A
certified spiked material was used (obtained from ERA, Arvada, CO). Extracts were analyzed by Method 8041.
Method blanks and three replicates from five laboratories were included. Data summary tables are included in
Method 8041.
TABLE 4. Multiple-laboratory phenoxyacid herbicides recoveries from certified spiked material
Compounds Average Recovery RSD
(%)
2,4-D
2,4-DB
2,4,5-T
2,4,5-TP (Silvex)
Dicamba
Dichlorprop
Dinoseb
81
122
74
68
50
87
118
81
122
74
68
50
87
118
13.0
15.1
11.8
17.9
17.6
20.7
29.4
Material spiked at 100 ug/kg. Number of participating laboratories = 4. N = 3
15
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
TABLE 5. Multiple-laboratory phenols recoveries from certified spiked material
Compounds
2,4-D
2,4-DB
2 4 5-T
2',4',5-TP (Silvex)
Dicamba
Dichlorprop
Dinoseb
Average
(M9/kg)
81
122
74
68
50
87
118
Recovery
(%)
81
122
74
68
50
87
118
RSD
13.0
15.1
11.8
17.9
17.6
20.7
29.4
Material spiked at 250 ug/kg. Number of participating laboratories = 4. N = 3
Omanoohosphorus pesticides and chlorinated herbicides: Multi-laboratory accuracy data were obtained for
organophosphorus pesticides spiked at 250 ug/kg in one soil type. A certified spiked material was used (obtained
from ERA, Arvada, CO). Extracts were analyzed by Method 8141. Method blanks and three replicates from five
laboratories were included. Data summary tables are included in Method 8151.
TABLE 6. Multiple-laboratory organophosphorus pesticides recoveries from certified spiked material
Compounds Average Recovery RSD
(M9"
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
ensure even shorter extraction times as well as higher extraction efficiencies.
ACKNOWLEDGEMENTS
The split samples were provided by Mr. Yi-Hua Lin and Vinod Kansal by Roy Weston Laboratory, REAC, Edison
through arrangements of Mr. Phil Campagna of the US EPA. Some supplementary funding from the
Environmental Protection Service of the Department of Environment of Canada is acknowledged. We are
especially indebted to the laboratories, other than our own, that contributed to the inter-laboratory work, namely
APPL (Fresno, CA), US EPA Region VI (Houston, TX), Emergencies Science Division of Environment Canada
(Ottawa, ON, Canada), and the US EPA/AED (Narragansett, Rl). The support provided by SAIC (Reston, VA) is
also acknowledged. The use of trade names and Trade-marks throughout this report does not imply an
endorsement or recommendation of use. The processes described herein may be subject to certain patents rights
and this paper does not in any way suggest or condone their use without proper rights to do so.
REFERENCES
1. J.R.J. Pare, M. Sigouin, and J. Lapointe, U.S. Patent 5,002,784, March 1991 (various international
counterparts).
2. J.R.J. Pare, U.S. Patent 5,338,557, August 1994 (various international counterparts); U.S. Patent 5,458,897,
October 1995 (various international counterparts).
3. J.R.J. Pare, U.S. Patent 5,377,426, January 1995 (various international counterparts); U.S. Patent
5,519,947, May 1996 (various international counterparts).
4. J.R.J. Pare, J.M.R. Belanger, and S.S. Stafford, Trends Anal. Chem. 13 (1994) 176.
5. L.G. Croteau, M.H. Akhtar, J.M.R. Belanger, and J.R.J. Pare, J. Liquid Chromatogr.
17(1994)2971.
6. K. Ganzler, A. Salgo, and K. Valko, J. Chromatogr. 371 (1986) 299.
7. F.E. Onuska and K. A. Terry, Chromatographia 36 (1993) 191; J. High Resol. Chromatogr. 18 (1995) 417.
8. E. Hasty and R. Revesz, American Laboratory (2) (1995) 66.
9. R. Steinheimer, J. Agric. Food Chem. 41, 588 (1993).
10. S. J. Stout, A.R. daCunha, and D.G. Allardice, Anal. Chem. 68 (1996) 653.
11. V. Lopez-Avila, R. Young, J. Benedicto, P Ho, R. Kim, and W.F. Beckert, Anal. Chem. 67 (1995) 2096.
12. V. Lopez-Avila, R. Young, and W.F. Beckett, Anal. Chem. 66 (1994) 1097.
13. V. Lopez-Avila, J. Benedicto, C. Charan, R. Young, and W.F Beckert, Environ. Sci. Technol. 29 (1995) 2709.
14. K. Li, J.M.R. Belanger, M.P Llompart, R.D. Turpin, R. Singhvi, and J.R.J. Pare, Spectres. Int. J. 13 (1996) 1.
15. R. McMillin, L.C. Miner, and L. Hurst. Abbreviated microwave extraction of pesticides and PCBs in soil.
Spectros. Int. J. 13 (1), 41-50 (1997).
16. M.P.M.P Llompart, R.A. Lorenzo, R. Cela, and J.R.J.Pare. Optimization of a microwave-assisted extraction
method for phenol and methylphenol isomers in soil samples using a central composite design. Analyst, 122,
133-137(1997).
17. M.P Llompart, R.A. Lorenzo, R. Cela, J.R.J. Pare, J.M.R. Belanger, and K. Li. Phenol and methylphenol
isomers determination in soils by in-situ microwave-assisted extraction and derivatisation. J. Chromatogr. A
757, 153-164(1997).
18. M. P. Llompart, R. A. Lorenzo, R. Cela, K. Li, J.M.R. Belanger, and J. R. J. Pare. Evaluation of supercritical
fluid extraction, microwave-assisted extraction and sonication in the determination of some phenolic
compounds from various soil matrices. J. Chromatogr. A, 774, 243-251 (1997).
19. C. Chiu, G. Poole, Y. Shu, R. Thomas, and R. Turle. Microwave vs. Soxhlet for the extraction of dioxins and
furans from solid samples. Organohalogen Compounds 27, 333-338 (1996).
20. C. Chiu, G. Poole, M. Tardif, W. Miles, and R. Turle. Determination of PCDDs/PCDFs Using
Microwave-Assisted Solvent Extraction. Organohalogen Compounds 31, 175-180 (1997).
PHENOXYACID HERBICIDE SCREENING
Mark L. Bruce. Kathleen L. Richards and Raymond M. Risden
Quanterra Inc., 4101 Shuffel Dr. NW, North Canton, OH 44720
brucem@quanterra.com
17
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Abstract
EPA SW-846 Methods 8150 and 8151 have been traditionally used for phenoxy acid herbicide analyses. These
methods require the use of hazardous derivatization reagents and a highly flammable solvent. The sample
preparation process is very long, complex, labor intensive and time consuming. Immunoassay, as used in EPA
SW-846 Method 4015, uses aqueous based chemistry with minimal solvent volume and much simpler and faster
sample preparation.
Quanterra has demonstrated that EPA SW-846 Method 4015 (2,4-D) can be extended to include 2,4,5-T and
silvex by using both the 2,4-D and silvex kits from Strategic Diagnostics Inc. It is useful for screening both water
and soil samples for phenoxyacid herbicides. Non-detect samples from the immunoassay (IA) screen are
reported as nondetect at the applicable reporting limit. Samples which have responses greater than the threshold
(i.e. positive response) are confirmed by the traditional methods. The overall false negative rates of 0.5% for
waters and 1.0% for soils were well below the EPA Office of Solid Waste criteria of 5% at the reporting limits
listed below. The false positive rates were 12.5% and 11.5% for waters and soils respectively. Water reporting
limits were 2 ug/L, 10 ug/L and 10 ug/L for 2,4-D, silvex and 2,4,5-T respectively. Soil reporting limits were 1
mg/kg, 1.5 mg/kg and 1.5 mg/kg respectively.
Switching to IA improves laboratory safety, reduces organic solvent usage and disposal, improves
turn-around-time and reduces analytical costs.
Introduction
Many herbicide analysis requests target three main analytes of interest: 2,4-D, 2,4,5-T and silvex (2,4,5-TP).
Also, >95% of these samples are reported as "non-detect" for these analytes. Thus, it would be very useful to
screen samples submitted for the. analysis of these herbicides. Only those samples with herbicides above the
reporting limit would be subjected to quantitation by the more exhaustive traditional analysis methods.
Immunoassay is capable of providing the screening information with an aqueous chemistry based method that is
simpler, faster, safer and less expensive.
Currently Immunoassay Method 4015 is only applicable to 2,4-D. However, this validation study demonstrates
that it is possible to extend the method by including a Strategic Diagnostics Incorporated (SDI) analysis kit
developed specifically for silvex by Ohmicron. Both the 2,4-D and silvex immunoassay kits have cross reactivity
for 2,4,5-T because of structural similarities. This cross reactivity enables the combined use of the two kits to
effectively screen for all three compounds, The cross reactivity is shown below expressed as least detectable
dose (LDD).
Table 1. Cross Reactivity
Compound
2,4-D
2,4,5-T
Silvex
2,4-D Assay
LDD (ug/L)
0.70
3.0
170
Silvex Assay
LDD (ug/L)
100
1.0
1.4
The Ohmicron (now part of SDI) 2,4-D kit has been validated following Office of Solid Waste (OSW) guidelines.
It was approved by the OSW Organics workgroup and incorporated into Method 4015. Method 4015 was part of
the Update III package for SW-846 promulgated in mid 1997. Ohmicron had established threshold test levels of
10 ug/L in water and 150 ug/kg in soil with their magnetic particle based assay kit. The sensitivity difference
between the two matrices is mostly due the large dilution of the methanol soil extract needed to reduce antibody
exposure to methanol.
The Ohmicron Silvex kit was validated similarly to the 2,4-D kit but the results were not submitted to the OSW
workgroup. The silvex kit has comparable sensitivity to the 2,4-D kit. Ohmicron has documented that the silvex
assay responds similarly to 2,4,5-T (i.e. cross reactivity).
Immunoassay Method Summary
The immunoassay form applied in these two kits is enzyme linked immunosorbent assay (ELISA). Information is
available from SDI product literature and the EPA - Las Vegas (web site http://www.epa.gov/crdlvweb/asb/
immuchem/forum.htm) describing IA basics.
18
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
For these SDI kits a subaliquot (250 uL) of aqueous sample is combined with 250 uL of enzyme conjugate
solution. One portion of the conjugate molecule resembles the analyte of interest. Antibody which is bound to
small magnetic particles is added to this analyte/conjugate mixture. The analyte and conjugate compete for
antibody binding sites. The higher the analyte concentration the less conjugate is bound to the antibody. The
lower the analyte concentration the more conjugate is bound to the antibody.
A magnet is used to retain the antibody particles and the analytes and conjugates that have bound to the
antibodies. All other analytes, conjugates and matrix components are washed away. A color development
reagent is added to the antibodies. This reagent reacts with another portion of the conjugate molecule and
develops color proportional to the amount of bound conjugate present. Thus, the color (i.e. absorbance
measured on the filter photometer) is highest (~1) for blanks and clean samples and the least absorbance is
produced for high concentration standards and samples. This inverse relationship between analyte concentration
and absorbance response has caused some confusion.
Thus, when using this IA format for threshold testing the results scenarios are:
a) if sample absorbance > threshold standard absorbance then analyte concentration is < threshold (e.g. non
detect at 10 ug/L), b) if sample absorbance < threshold standard absorbance then analyte concentration is >
threshold (e.g. positive analyte > 10 ug/L)
Since the color response will always have some variability associated with both the threshold standard and the
sample, it is customary to prepare the threshold standard at a concentration slightly below the reported test
threshold. This "low bias" on the standard favors the assay toward producing false positives in order to reduce
the false negative rate. The following 10 ug/L report threshold example illustrates:
Table 2. Example Assay Batch
Assay
threshold standard #1 (7 ug/L)
threshold standard #2 (7 ug/L)
LCS(10ug/L)
sample #1
sample #2
sample #3
sample #4
sample #5
sample #6
sample #7
NIS of ND sample #1(>10 ug/L)
MS of ND sample #2(>10 ug/L)
Absorbance
0.8
0.88
0.6
0.97
0.92
1.02
0.86
0.4
0.55
0.7
0.57
0.86
Comment
use the avg std response 0.84 for cutoff
pass
ND<10
ND<10
ND<10
ND<10
positive >10
positive >10
positive >10*
pass
fail - false negative**
sample #7 may be a false positive at the test threshold since it falls between the. threshold standard and the
LCS. In other words it may have herbicide present at 8 ug/L. It would be sent for 8150 confirmation.
** This MS illustrates a documented false negative which would trigger corrective action.
The threshold standard (7 ug/L) is intentionally lower than the reported test threshold (10 upg/L) to reduce the
false negative rate due to normal statistical variability and minor preparation losses and matrix interferences.
Quanterra intended to improve the sensitivity of the soil assay by reducing or eliminating the need for the large
dilution used to reduce the impact of methanol on the antibodies. Evaporating the methanol from a small aliquot
of sample extract, followed by reconstitution in water was effective at removing the methanol. However, this
process also concentrated the interferences in the soil matrices and raised the level of non-specific antibody
binding. Unfortunately, this raised the false positive rate to an unacceptable level. Thus, this preparation
modification was not used in the validation study.
The silvex kit has not been "validated" for 2,4,5-T but the SDI cross reactivity data suggest that the silvex
antibodies work similarly for 2,4,5-T threshold testing. 2,4,5-T was investigated along with 2,4-D and silvex for
which the IA kits were designed. In fact, 2,4,5-T was detected by both the 2,4-D and silvex IA kits during the
validation study. At least one kit produced an acceptable 2,4,5-T response for each sample included in the study.
19
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Sympo:
mum
This validation study builds on top of the existing studies and knowledge discussed above. Quanterra has
demonstrated that the two IA herbicide kits are appropriate to screen out water and soil samples that have no
detectable levels of 2,4-D, 2,4,5-T or silvex at the applicable reporting limits. This is not intended to replace
Method 8151 for quantitation but to focus its use where it is most applicable. Since the extended IA method is not
intended for quantitation and much validation work has already been completed for the IA kits very little
additional direct comparison between IA and GC data is necessary to establish that the IA method is capable of
screening out non-detect samples. We intended to utilize matrix spikes of well characterized matrices, matrix
spikes of real samples (ND for herbicides by 8150/51) and a few real samples with herbicides that have been
previously analyzed by 8150/51. Hereafter, herbicides that entered the sample by environmental transport
mechanisms and have undergone aging and weathering are referred to as native herbicides, Unfortunately, the
native herbicide samples that were available at the time of the validation study could not be used as intended
because the analyte concentrations were much lower than the final test thresholds. To compensate, many of
these samples were fortified with known amounts of each herbicide and included in the study as matrix spiked
samples.
Validation Goal
The main goal was to demonstrate that the IA method responds to herbicide levels known to be at or above the
IA reporting limit (threshold) for each analyte. Since samples with native herbicides above the final test
thresholds were not available, the number of matrix spiked samples was significantly increased over the original
validation plan. Water matrix spiked samples were increased from 6 to 18. Soil matrix spiked samples were
increased from 7 to 12. Also, using matrix spikes allowed us to more accurately document performance at the
critical area near the reporting limit. It was not necessary to analyze the matrix spiked samples with 8150/51
since the herbicide concentrations were already known. The matrix spiked soil samples were stored at 4ฐC at
least overnight and usually for several days between spiking and extraction in order to age them and more
closely mimic native analytes.
Validation Protocol
Matrix spiked samples were prepared using techniques previously employed for SW-846 methods development
work. Separate spike solutions containing known amounts of each of the three target analytes were prepared at
appropriate concentrations in acetone. These high concentration standards were further diluted in reagent water.
Water samples were spiked with these aqueous solutions, homogenized by shaking and immediately assayed.
Soil samples were spiked, acetone solvent evaporated at room temperature and tumbled overnight in a rotary
mixer. The spiked soil samples were stored at least overnight at 4ฐC and usually for several days prior to
extraction in order to "age" the samples and thus more closely mimic native soil samples. Obviously this process
does not duplicate the extensive weathering which can occur in real environmental samples, but the combined
use of short term "aging" and fuller's earth that is known to be difficult to extract should simulate many difficult
samples. Each of the following sample-analyte combinations in Tables 3 & 4 was assayed at least once.
Replicate analyses indicated in Q.
Analytical Procedure Summaries
Water preparation
Allow particulates in water sample to settle, filter (0.45 urn PTFE) if sample is cloudy with suspended solids,
aliquot 250 uL of sample into plastic assay tube. Aqueous matrix spike solution added as appropriate. TCLP
buffers and samples were spiked as appropriate in a small vial. A 25 uL aliquot was transferred to the assay tube
and 225 uL of diluent was added.
So/7 preparation
Weigh 10 g of soil sample into 50 ml plastic centrifuge tube. Add acetone matrix spike solution as appropriate.
Allow solvent to evaporate. Add 2-3 ball bearings. Tumble overnight (note: some wet clay samples formed large
clumps during tumbling and were not tumbled when matrix spiking was required for additional analyses). After
"aging", add 30 mL of extraction solvent (75% methanol, 23% reagent water, 2% acetic acid). Recap centrifuge
tube, rotate onto side and mechanically shake at about 200 cycles/minute for 30 minutes. Stand centrifuge tubes
up and allow soil to settle for 1 hour and/or centrifuge for 3-5 minutes. Filter (0.45 urn PTFE) a few milliliters of
extract. Transfer 5 uL of extract into plastic assay tube and add 245 uL of diluent.
20
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 3. Water Sample List and Matrix Soike Levels
Sample
ground water
industrial waste water
TCLP buffer #1
TCLP buffer #2
TCLP Buffer #1 samples t
X3
02
43
01
ZX
Analyte
No spike
(10)
(4)
(13)
Water samples t
HN
HV
QK
5G
DA
W2
4JL
(6)
(7)
(5)
(4)
(5)
(6)
(5)
Performance Evaluations
Wise. PE
WS038 PE
(3)
(3)
2,4-D
high
10
10
100
100(7)
100
100
100
100
100
10
10
10
10
10
10
10
low
2(4)
2
2
2
2
2
2
2(8)
2
si I vex
high low
10 5
10 5
100(2)
100(7)
100
100
100
100
100
10 5
10 5
10 5
10 5
10 5
10 5
10 5
2,4,5-T
high low
10(27) 5
10 5
100(27)
10 5
10 5
10 5
10 5
10 5
10 5
10 5
Additional water samples screened for false positives t
8H
AH
HK
we
5V
65
6
K1
K2
W3
W6
W7
WA
JW
KO
(4)
(4)
t Samples previously analyzed by 8150 but no herbicides detected.
Immunoassay
Allow all IA reagents (particularly the enzyme conjugate and antibodies) to warm to room temperature. Add
sample or standard aliquot (250 uL final volume). Add 250 uL of enzyme conjugate. Add 500 uL of suspended
antibody coupled magnetic particles and vortex mix for 1-2 seconds. Incubate at room temperature for 30
minutes. Apply magnetic rack base for 2 minutes to separate magnetic particles from bulk fluid in tubes. Pour out
tube contents while the magnetic particles are retained at the bottom of the tubes. Rinse the antibodies twice with
wash solution. Remove magnetic rack base. Add 500 uL of color development reagent and incubate for 20
minutes. Add 500 uL of sulfuric acid stop solution and read absorbance at 450 nm.
The assay tubes were arranged in a 3 X 10 layout in the SDI magnetic rack. Often two batches of 30 assays were
performed simultaneously in the rack which holds a maximum of 60 tubes. A typical 30 tube layout is shown
below. It was quite common in the early batches of the validation study to assay standards prepared by both SDI
21
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
and Quanterraฎ (QES) as a means of verifying the standards prepared in-house.
Table 4. Soil Sample List and Matrix Spike Levels
Sample
sandy soil
loam soil
fuller's earth* (dry)
fuller's earth* (50%
moisture)
Soil -samples t
03
04
2N
36
20
20
22
25
1V
NO
spike
(5)
(4)
(4)
(3)
(4)
(4)
high
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
2,4-D
ug/kg
low
0.3 (2)
0.3 (2)
0.3 (2)
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
Analyte
silvex
ug/kg
high low
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5(8)
1.5
1.5
1.5
2,4,5-T
ug/kg
high low
1.5(2)
1.5(2)
1.5(2)
1.5
1.5
1.5
1.5
1.5
1.5
1.5(8)
1.5(8)
1.5
1.5
* Fuller's earth is an absorptive clay known to challenge the efficiency of solid extraction procedures.
t Samples previously analyzed by 8150 but no herbicides detected above (or near) spike concentrations.
Raw Response Data
The average difference in response between duplicate SDI and QES standards was 4%. The response difference
between silvex and 2,4,5-T was small (average 12.5%) and 2,4,5-T consistently produced less response. Thus,
2,4,5-T was used as the threshold compound for all silvex kit assays in order to reduce the false negative rate,
particularly for 2,4,5-T containing samples. The test threshold standards were prepared at 70% of threshold
concentration (0.7 x expected reporting level) in order to reduce false negatives for water matrices. Threshold
standards were prepared at 50% of the threshold concentration for the soil assays.
Table 5. Example IA Batch Layout for 2,4-D kit assays - 10 ug/L Reporting Level and matrix spikes
1
2
3
ABODE FGH I J
QES
"10"std
W2
SDI
"10"std
GW
W2+
D
W6
GW+
D
W3
W6+
D
TCL
P
W3+
D
W7
TCLP j WW
_jUL^A
QES LCS
atlORL
W7+
D
SDI LCS
atlORL
Wise
PE
WW
+ D
W4
Wise
PE
W1
W4+
D
WS038
PE
W1 +
D
W5
WS038
PE
SDI
"10"std
W5+
D
QES
"10"std
Key:
QES "10" std = 7 ug/L 2,4-D standard prepared by Quanterra
SDI "10" std = 7 ug/L 2,4-D stock standard prepared by SDI
QES LCS at 10 RL = 10 ug/L 2,4-D standard prepared by Quanterra
SDI LCS at 10 RL = 10 ug/L 2,4-D stock standard prepared by SDI
D = 2,4-D spiked sample (10 ug/L)
GW = Ground water
TCLP = TCLP buffer #1 [10 X dilution]
WW = Industrial waste water
W1, W2, W3, W4, W5, W6, W7 = real water samples without native herbicides
PE = Performance evaluation samples from Wisconsin and EPA WS038 programs.
Results and Discussion
The key questions when evaluating the reliability of this screening method are:
1) If an analyte is present in a sample at a concentration greater than or equal to the threshold (i.e. reporting
22
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
limit), what are the chances of the assay generating a false negative response?
2) If no analyte is present in a sample at or above the threshold concentration, what are the chances of the assay
generating a false positive response?
Table 6. Threshold Standards
Threshold Level
(reporting limit)
2.0 ug/L
10. ug/L
1 .0 mg/kg
1.5 mg/kg
Compound
water samples
2,4-D
silvex or 2,4, 5-T
soil samples
2,4-D
silvex or 2, 4, 5-T
actual
standard concentration
1 .4 ug/L 2,4-D
7.0 ug/L 2,4,5-T
3.3 ug/L 2,4-D
5.0 ug/L. 2,4,5-T
False negatives are undesirable since they would report a sample as "clean" with regard to the target herbicides
when it was not and potentially increase environmental health risks. False positives are undesirable since they
would needlessly "trigger" a batch of traditional 8151 analyses which would increase organic solvent usage,
analyst exposure to hazardous reagents, turn-around time and cost.
Since the same number of replicates were not run for each sample it is not appropriate to simply divide the
number of false negatives or positives by the total number of assays. The false negative or positive rate was
determined for each sample - analyte combination or unspiked sample. These individual rates were then
averaged to determine the overall false negative and positive rates. LCS results were not included in these
calculations because they do not contain real sample matrix. The false negative rate must meet the normal
Office of Solid Waste criteria, <5%. The false positive rate was expected to be <10%.
The matrix spike results from samples producing consistent false positive responses were not included when
calculating the false negative rates for either water or soil samples.
Water Samples
Production threshold levels of 2 ug/L for 2,4-D and 10 ug/L for silvex and 2,4,5-T were selected after evaluating
the initial water sample data. Table 7 summarizes the percentage of false positives and false negatives for each
sample - analyte combination.
The overall false negative rate of 0.5% for water samples was excellent. Sample W2 had one false negative
among 8 replicates at the 2 ug/L threshold for 2,4-D for a false negative rate of 12.5% for this sample - analyte
combination. The ground water (GW) and TCLP buffer #1 produced many false negatives for 2,4,5-T when
assayed with the silvex kit at the 10 ug/L (GW) and 100 ug/L (TCLP) levels. Keeping the spike levels and analyte
the same but switching to the 2,4-D kit significantly improved the assay reliability. The false negative rates
dropped to 16.7% and 5.9% for these two matrices respectively. Since 2,4,5-T is not a normal TCLP analyte
these false negatives have little practical impact on assays of TCLP samples, although this does indicate that
2 4 5-T may be more susceptible to false negatives than the other two analytes. The ground water false negative
rate was reduced by assaying for 2,4,5-T with the 2,4-D kit as noted above. This indicates that either
immunoassay kit will respond to 2,4,5-T and that at least one kit is likely to produce a positive response at the 10
ug/L threshold when 2,4,5-T is present.
The false positive rate was determined by dividing the total number of "non-detect" samples included in the study
into the number of samples that produced false positives for a false positive rate of 12.5%. Replicate assays
were performed on two of the samples that produced false positives. This confirmed that a matrix interference
existed which produced the false positive. One of the false positive sample responses was near the response of
the threshold standard and probably would not produce a false positive in all instances if replicate assays were
performed It appears that false positives are primarily generated by matrix interferences. Thus, positive matrix
interferences are likely to be site specific and the false positives that do occur in actual production use of the IA
kits should be clustered together in a limited number of sample lots. This means that most of the sample lots
assayed are expected to be free of false positives.
23
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 7. False Negative and False Positive Rate Calculations for Water Samples
Matrix or sample
ground water
waste water
TCLP buffer #1
TCLP buffer #2
X3
02
43
01
zx
HN
HV
QK
5G
DA
W2
JL
Wise PE
WS038 PE
Matrix or sample
8H
AH
HK
we
5V
63
65
6
K1
K2
W3
W6
W7
WA
JW
KO
Average rates
No spike
% false positives
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
No spike
% false positives
0
100
100
0
0
0
0
0
0
0
0
0
100
0
0
100
12.5%
2,4-D
% false negatives
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1/8 = 12.5
0
0
0
2,4-D
% false negatives
0.7%
overall false
Silvex
% false negatives
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Silvex
% false negatives
0%
negative rate 0.5%
2,4,5-T
% false negatives
3/18=16.7
0
1/17 = 5.9
0
0
0
0
0
0
0
2,4,5-T
% false negatives
2.3% |
Soil samples
Attempts to improve the sensitivity of the soil method were unsuccessful. Initial information from SDI indicated
that the 50X extract dilution included in the SDI soil preparation method was designed to remove the deleterious
effect that methanol has on the immunoassay. We intended to remove the methanol interference by evaporating
a small aliquot of the extract to dryness then redisolving the analytes in diluent solution or water. Analyte
recovery appeared acceptable and there was no visible methanol residue, but there was still some small positive
interference which probably would have prevented reliable assays at the 30 pg/kg target threshold. Real soil
extracts showed very large positive interferences when 250 uL of extract were concentrated for the assay. Thus,
it was necessary to restrict to soil preparation to the original SDI/Ohmicron soil prep method. This was
accomplished by diluting 5 pL of extract with 245 \iL of diluent. When this SOX dilution was combined with higher
analyte spike levels the interference problems were reduced to an acceptable level.
The following preparation method was used for most soils: 10 g soil + 30 ml extract solvent, shake for 30
minutes, settle or centrifuge, filter, 5 pL of extract + 245 |jL diluent. Thus, the final concentration of analytes
24
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
presented to the assay were 150X more dilute than in the original soil sample. Quantities of some samples were
limited so the preparation amounts were scaled back to 5 g soil and 15 mL of extraction solvent. The assay
threshold standard concentrations reflected the dilution built into the sample prep. Most spike levels were 1.0
mg/kg for 2,4-D and 1.5 mg/kg for silvex and 2,4,5-T. Some low level (0.03 and 0.3 mg/kg) work was attempted,
but the false positive rate was unacceptably high.
The concentration of the test threshold standard concentration was also reduced from 70% to 50% of threshold
for soil assays. This decreased the false negative rate without unacceptably raising the false positive rate. Table
8 below shows 3 soil - analyte combinations had false negative results for silvex or 2,4,5-T when spiked initially
at 1.5 mg/kg. Replicate sample aliquots were spiked, extracted and assayed. No additional false negative results
were produced. Thus, the overall false negative rate was 1.0%, which is excellent. False positive results were
reported for two samples at the 1.0 mg/kg 2,4-D or 1.5 mg/kg silvex/2,4,5-T assay levels. The overall false
positive rate for soil samples was 11.5%. It is expected that false positives will be primarily caused by specific
constituents in the soil samples and are thus more likely to be related to specific sites rather than be evenly
distributed through all sample lots.
Table 8. False Negative and False Positive Rate Calculations for Soil Samples
Matrix or sample
ground water
loam soil
fullers earth dry
fullers earth wet
03
04
2N
36
2E
20
22
25
1V
Average rates
No spike
% false positives
0
0
0
0
50
0
0
0
100
0
0
0
0
1 1 .5%
2,4-D Silvex
% false negatives % false negatives
0
0
0
0
0
0
0
0
0
0
0
0
0%
average false negative
0
0
0
0
0
0
0
0
1/8= 12.5
0
0
0
1 .0%
rate = 1 .0%
2, 4, 5-T
% false negatives
0
0
0
0
0
0
0
0
1/8= 12.5
1/8 = 12.5
0
0
2.1%
The soil extraction procedure did not appear to suffer any serious extraction efficiency problems despite its
simplicity and short time frame. Even though the IA results were not quantitative, the low false negative rate
indicates that analyte recovery was > 50% and many recoveries were > 70%. Previous Quanterra work with wet
fullers earth with nonpolar analytes and solvents, showed very low (<20%) analyte recovery for hydrocarbons.
Good extraction efficiency of the phenoxy acid herbicides with the simple shake extraction used with these IA
kits was probably due to the following reasons: 1) The polar extraction solvent (methanol / water / acetic acid)
readily permeated the wet clay matrix. 2) The polar solvent molecules could effectively displace polar analyte
molecules from the polar sorption sites on the matrix. 3) The polar analytes were readily soluble in the polar
extraction solvent.
Conclusion
The validation study results are summarized in Table 9 below. The false negative rates easily meet the normal
EPA Office of Solid Waste criteria of <5%. The. false positive rates although slightly higher than the target 10%,
are still acceptable. The water reporting limits are in the low part-per-billion range and meet the validation plan
minimum objectives. In particular, analyses from TCLP buffers and samples demonstrated excellent
performance for 2,4-D and silvex at levels well below the regulatory limits. The soil reporting limits although
higher than originally expected should still be useful for many types of herbicide samples.
Table 9. Summary of Performance Results and Reporting Limits
Matrix
water
2,4-D
2ug/L
1 mq/kg
Silvex
10ug/L
1 .5 mg/kg
2, 4, 5-T
10ug/L
1 .5 mg/kg
False Negative
0.5%
1 .0%
False Positive
12.5%
1 1 .5%
25
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Acknowledgments
The authors would like to thank Allen Debus from US EPA, Region 5, Craig Kostyshyn and Timothy Lawruk from
Strategic Diagnostics and Susie Dempster, Carolynne Roach, Chris Lee and Richard Burrows from Quanterra
Inc.
References
1. Strategic Diagnostics Incorporated Product Literature - 2,4-D Rapid Assay
2. Strategic Diagnostics Incorporated Product Literature - Silvex Rapid Assay
3. SW-846 Method 4015 - Screening for 2,4-Dichlorophenoxyacetic Acid by Immunoassay.
FIELD DEMONSTRATION OF A PORTABLE IMMUNOSENSOR FOR EXPLOSIVES DETECTION
Anne W. Kusterbeck1, Paul T. Charles1, Paul R. Gauger2 and Charles Patterson1
1Centerfor Bio/Molecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375
phone: 202 404-6042, fax: 202 404-8897
akusterbeck@cbmse.nrl.navy.mil
2GeoCenters, Inc., 1801 Rockville Pike, Suite 405, Rockville, MD 20852-1633
Environmental biosensors are being developed at the Naval Research Laboratory for detection of the explosives
TNT and RDX in groundwater and monitoring of cleanup progress for these compounds at remediation sites.
Based on a displacement immunoassay, the portable sensor, known as the FAST 200, has been engineered by
Research International (Woodinville, WA) to quantitate water samples with no sample preparation or reagent
addition. Analysis is complete within five minutes. The sensor, along with a fiber optic biosensor, recently was
extensively tested in field trials at several U.S. EPA Superfund sites to validate sensor performance. Results of
these studies and application of the technology will be described.
ENVIRONMENTAL APPLICATIONS OF A FIBER OPTIC BIOSENSOR
Lisa C. Shriver-Lake1, Irina B. Bakaltcheva2, and Saskia van Bergen3
1Center for Bio/Molecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375
phone: (202) 404-6045, fax: (202) 404-8897
email: lcs@cbmse.nrl.navy.mil
2GeoCenters, Inc., 1801 Rockville Pike, Suite 405, Rockville, MD 20852-1633
3George Mason University, 4400 University Drive, Fairfax, VA 22030-4444
Detection and remediation monitoring of the explosives TNT and RDX on-site requires a sensitive and preferably
portable method. The fiber optic biosensor is based on a competitive fluoroimmunoassay being performed on the
core of an optical fiber probe. A portable version of the sensor was engineered by Research International
(Woodinville, WA) and is known as the Analyte 2000. With this sensor, four optical probes can be monitored
simultaneously and relatively Adirty@ samples can be employed. Analysis takes 16 minutes for the four probes.
This sensor, along with the FAST 2000, was extensively evaluated at three on-site trials to validate sensor
performance. Results of these studies and other applications for the fiber optic biosensor will be described.
26
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
DEVELOPMENT AND VALIDATION OF AN IMPROVED IMMUNOASSAY FOR SCREENING SOIL FOR
POLYNUCLEAR AROMATIC HYDROCARBONS
Titan S. Fan and Brian A. Skoczenski
Beacon Analytical Systems, Inc., 4 Washington Ave., Scarborough, ME 04074
Polynuclear aromatic hydrocarbons (PAHs) are a group of fused ring compounds most typically found as
combustion by-products. Over the past 4 - 5 years a number of immunoassay kit manufacturers have developed
and commercialized soil screening method for PAHs. Briefly, there methods involve a rapid extraction of soils by
shaking with methanol followed by analysis of the filtered sample extract using competitive enzyme
immunoassay. We have developed an improved method for screening of PAHs in soil samples. The new method
utilizes a modified sample extraction step that results in improved extraction efficiency compared to earlier
methods. In addition, the specificity of the antibody used in the method allows for a better estimate of the total
PAHs present. Immunogen design and antibody specificity will be described in detail. Results of concordance
study with a gas chromatographic method will be presented.
A NEW DIOXIN/FURAN IMMUNOASSAY WITH LOW PICOGRAM SENSITIVITY AND SPECIFICITY
APPROPRIATE FOR TEQ MEASUREMENT: APPLICATIONS DEVELOPMENT
Robert O. Harrison
CAPE Technologies, L.LC, 3 Adams St., South Portland, ME 04106
Robert E. Carlson
ECOCHEM Research, Inc., 1107 Hazeltine Blvd, Chaska, MN 55318
Since 1990 the commercial development of immunoassay kits his opened a new market in environmental
analysis. The US EPA has approved more than 10 immunoassay screening methods under the 4000 series of
Field Screening Methods within SW-846. These tests are now widely used in assessment and remediation of
hazardous waste sites.
This recent success has not included the development of a useful immunoassay method for polychlorinated
dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs). The development and application of
immunoassays for PCDD/Fs pose unique challenges not found in immunoassays for other analytes. First, the
sensitivity requirements of PCDD/F analysis are typically in the ppt range rather than the high ppb to mid ppm
range of the existing 4000 series screening methods. Second, if the test is to provide useful data, the EIA
response must correlate to the relative toxicity of the 17 most toxic PCDD/F congeners. No previous
immunoassay has demonstrated the necessary combination of sensitivity and specificity required for
measurement of toxic equivalency (TEQ) at ppt levels. No immunoassay specific sample preparation methods
have been developed because of the obvious lack of commercial potential demonstrated by all previous PCDD/F
immunoassays
A new enzyme immunoassay (EIA) for PCDD/Fs has been developed using novel chemistry. The sensitivity of
this test is approximately 4 pg of 2378-TCDD, which is more than an order of magnitude better than previous
PCDD/F immunoassays. Based on typical sample size, this sensitivity is sufficient to measure low ppt TEQ
levels in solid samples or 0.1 ng/m3 TEQ in stack gases using only a small fraction of the prepared sample
extract. This sensitivity allows detection of 2378-TCDD in a 10 uL sample aliquot at the same concentration as
the lowest calibration solutions typically used for HRGC/HRMS based PCDD/F methods such as EPA Methods
1613, 8290, and 23. These sensitivity comparisons indicate that the EIA is capable of screening samples prior to
HRGC/HRMS analysis without consuming an unacceptably large proportion of the sample.
The dioxin/furan congener cross-reaction profile of this EIA is suitable for TEQ measurement. The test is most
sensitive to the three most toxic congeners, 2378-TCDD, 12378-PnCDD, and 23478-PnCDF EIA specificity data
plus HRGC/HRMS data from previously analyzed samples have been utilized in a simple, additive response
model to predict the EIA response for each sample. The resulting correlation between predicted EIA response
and TEQ validates the concept of TEQ screening by EIA for a variety of samples.
27
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
A muliti-laboratory collaborative effort is now in progress for evaluation of kit performance and development of
sample preparation methods. Results for fly ash and soil validate the additive response model for TEQ
measurement by EIA. These results also validate the use of the kit for screening fully cleaned samples.
Extension of this validation to partially cleaned samples is in progress, with positive initial results for fly ash and
soil. Work on immunoassay specific sample preparation methods, including rapid extraction of soil and fly ash, is
also in progress. The ultimate goal of this coli program is to develop sample preparation protocols which will
maximize throughput and cost-effectiveness of immunoassay based PCDD/F screening.
DEVELOPMENT AND VALIDATION OF AN IMMUNOASSAY FOR SCREENING SOIL
FOR POLYCHLORINATED BIPHENYLS
Titan S. Fan and Brian A. Skoczenski
Beacon Analytical Systems, Inc., 4 Washington Ave., Scarborough, ME 04074
Polychlorinated Biphenyls (PCBs) were commonly used in electrical applications due to their properties of high
thermal transfer and low conductivity. Since being identified as toxic substances their use has been banned in
the US. Due to the once widespread use and their stability in the environment a large number of sites are
contaminated with PCB residues. A number of immunoassay kit manufacturers have developed and
commercialized soil screening methods for PCBs. Briefly, these methods involve a rapid extraction of soils by
shaking with methanol followed by analysis of the filtered sample extract using competitive enzyme
immunoassay. We have developed an improved method for screening of PCBs in soil samples. The new method
utilizes a modified sample extraction step that results in improved extraction efficiency compared to earlier
methods. The immunoassays performed on the sample extract and yields qualitative results at 1, 5, 10 or 50
ppm. The test can be used for measuring Aroclors 1016, 1242, 1248, 1254 and 1260. Results of a concordance
study with a gas chromatographic method will be presented.
GASOLINE RANGE AROMATIC/ALIPHATIC ANALYSIS USING PATTERN RECOGNITION
Steven E. Bonde
Petro Star Inc., 201 Arctic Slope Ave., Suite 200, Anchorage, Alaska 99518-3030
Finding an analytical technique for the analysis of aromatic and aliphatic compounds in the gasoline range of
hydrocarbons is of great interest to not only the laboratory but also the regulated community. In many states,
including Alaska, great pressure to apply risk based cleanup standards is driving the need to separate "high risk
compounds" from a given matrix in an economical manner with high confidence. False positives can mean an
unneeded costly cleanup, while false negatives can mean non-compliance and possible fines if found. Analysis
of aromatic and aliphatic compounds in the gasoline range has been accomplished using pattern recognition
algorithms based on PCB matching criteria. Using existing methodologies for the analysis of gasoline range
hydrocarbons in multimedia samples by GC-FID-PID this algorithm has been developed using widely available
software to recognize and quantify aromatic hydrocarbons in a given sample. Soil, sediment, and water samples
were analyzed using standard Alaska Department of Environmental Conservation methodology
(AK101/EPA8021) and the results of pattern matching showed a significantly reduced number of false positives
for the aromatic portion of the analysis.
28
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
BIOREMEDIATION ASSESSMENT USING CONSERVED INTERNAL BIOMARKERS
P. Caleavecchio. E.N. Drake, G.C. VanGaalen and A. Felix
Corporate Research Laboratories, Exxon Research and Engineering Company,
Rt. 22 East, Annandale, New Jersey 08801
(908) 730-2308
ABSTRACT
The lack of homogeneity of field samples is often a primary concern in the analytical assessment of hydrocarbon
bioremediation treatment efficacy. One approach to handling the sampling variability is to take a statistically
significant number of samples to adequately represent the distribution of the contaminant. This approach is
limited in that small biodegradation changes often cannot be seen within the inherent sampling variability. An
additional drawback of this approach is the high cost of analyzing sufficiently large numbers of samples to draw
conclusions.
An alternative approach, which can minimize the effects of sampling variability, is to use naturally occurring
molecules, which resist biodegradation as conserved internal "biomarkers". There are a number of marker
classes commonly found in petroleum products, which can be analyzed by GC/MS, these include hopanes,
steranes and isoprenoids. These markers can be used as the basis for relative comparisons of target analytes
before and after treatment. The ratios of two or more markers can also be used as a fingerprint to help identify
the petroleum product source.
This poster demonstrates the application of biomarker normalized GC/MS data to evaluate the biodegradation of
petroleum products from well-mixed refinery sludge waste. A comparison is shown tracking the treatment
progress using absolute target analyte quantities and relative amounts normalized to biomarkers. The results
show that the biomarker normalized data provided a good basis for determining the percent biodegradation of
total oil as well as the molecular components, similar to the results obtained using absolute quantities.
The conclusions of the above study indicate that although significant amounts of 2, 3, 4 and 5 ring polynuclear
aromatic hydrocarbons (PAHs) were removed during treatment the biomarkers used to normalize the data were
conserved, resisting significant biodegradation. The implication of this work is that the use of biomarkers can be
effective in situations that are less homogeneous to assess biodegradation of petroleum constituents from
products with similar sources.
INTRODUCTION
The objective of the following study was to remediate a refinery process sewer sludge, through aerobic
biodegradation, using composting to accelerate this process. Composting is a modification of the "biopile"
process, which includes the addition of biodegradable ammendments (straw, wood chips, manure, etc.) to help
aerate and supply energy for microbial growth. The laboratory scale evaluation was conducted in three-liter
insulated glass beakers designed to minimize the loss of heat generated by the compost.
To minimize the effects of sludge heterogeneity on analytical results, a number of steps were taken which
included mixing the sample well. An additional approach was to use naturally occurring "biomarker" molecules,
which resist biodegradation, as internal standards. The latter approach was taken considering that future field
applications would probably have much higher sampling variability due to the difficulty in homogenizing large
volumes of compost.
The sludge used in these tests had significant amounts of two biomarkers typically found in petroleum products,
C30 17a, 21B(H)-Hopane ("hopane") and C29 25-Nor-17a(H)-Hopane ("norhopane"). The above molecules are
related by a mechanism proposed by Peters and Moldowan that suggests the origin of norhopane through the
biological demethylation of hopane (see figure 1). Both biomarkers were found by selective ion (SIM) GC/MS at
m/z 191, norhopane also has a characteristic peak at m/z 177.
The progress of the composting experiments were tracked by periodic sampling during the period of peak
biological activity as indicated by naturally occurring elevated temperatures within the cells. The analytical tests
performed on the extracted hydrocarbon included Total Petroleum Hydrocarbon (TPH, modified New Jersey
method OQA-QAM-025-10/91) and GC/MS analysis for priority pollutant (PP) polynuclear aromatic hydrocarbons
(PAHs). The TPH data was quantified by total ion GC/MS response using an external calibration of free oil
29
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Aerobic
Blodegradatton?
2ป 24
C3017a(21/J(H)-HoPnne
C,g 2S-Nor-17a(H)-Hopanซ
O0-Deปmeซhylhopane")
Proposed origin of 25-norhopanes by
bacterial demethylation of 17a (H)-hopanes.
The methyl group attached to the C-10
position in the C3o 17a (H)-hopane (left) is
removed to produce the C2g 25-norhopane
(right).
Figure 1. Proposed origin of norphopane
from hopane
Figure 2. Biomarker ratio of hopane/norhopane
collected from the sludge. The PAHs were quantified using internal standards spiked into the extracts as
prescribed in EPA SW846 method 8270.
The "biomarker normalized" data was obtained by dividing the GC/MS response of the analyte by that of hopane.
The degradation measured relative to hopane was then compared to the "absolute" values obtained using the
traditional approach described in the above paragraph. The conservation of hopane throughout the test, was
monitored by comparing its GC/MS response relative to that of norhopane.
SUMMARY
The results show that rapid degradation of both TPH and PAHs occurred over the six-week period of the
composting experiments, using both the absolute measurement and the hopane normalized approach. During
this period there was no indication to suggest that significant hopane degradation occurred as shown by little
change in the ratio of hopane to norhopane (see figure 2). Also qualitative examination of the biomarker's ion
chromatograms at m/z 191 and 177 show little change before and after treatment (see figures 3 and 4). In
contrast, the total ion chromatogram of the extracted hydrocarbon shows significant removal of the oil during
treatment (see figure 5).
The TPH removal during treatment is shown in figure 6, which compares the absolute measurement to the
hopane normalized results. These results show somewhat higher degradation estimates relative to hopane,
ranging from 4% higher after one week to 30% higher after 6 weeks of treatment. The comparison of a selection
of the more abundant 3, 4 and 5 ring PP PAHs using the absolute and hopane-normalized approaches is shown
in figures 7 and 8 respectively. This comparison shows good agreement for phenanthrene and pyrene
measurements using both approaches, with a somewhat lower estimate for the hopane normalized
benzo(a)pyrene degradation (~ 30% lower). The lower concentration of benzo(a)pyrene (1 ppm) compared to
30
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
phenanthrene (27 ppm) and pyrene (15 ppm) may have attributed to the variability seen in its degradation
measurements.
Abundance 1911 191.00 (150.71) to lt>1.70J : 2Ptt-H-I.l3
2000J
Norhopane
1500 J
1000 J
500 J
Time-->
Hopane
37.32
m/z!91
37! do' ialdo' 'aslo'o' 40! do' '411 do' ซloo' 43!do' 4* loo'
700 -1
600
500
400
300 J
200
100
0
36
Norhop.
Iqn 177.00 (176.70 to 177.70): ZDB-14-1.D
46
m/z!77
ritne-->
so 3700 3800 39o 40o 4lo
4300 4*00
Figure 3. Before treatment ion chromatograms with hopane and norhopane
The degradation of the C1, C2 and C3 alkylated homologs of phenanthrene and anthracene are shown in figures
9, 10 and 11 respectively. This comparison shows very close agreement between the absolute and hopane
normalized approaches (less than 10% difference) over the duration of the composting experiments.
CONCLUSIONS
This work demonstrates a practical application using selective ion GC/MS to characterize the biodegradation of
petroleum contaminants relative to naturally occurring biomarkers. The hopane and norhopane biomarkers used
in the above composting experiments showed no sign of significant biodegradation over the tests duration, while
234 and 5 ring PAHs as well as TPH showed rapid biodegradation. A comparison of biodegradation
measurements using conventional quantification and biomarker normalized approaches showed good agreement
for the most abundant analytes with somewhat higher variability for those in low concentrations.
The biomarker approach was easy to use and is independent of many sources of analytical and sampling
variability associated with absolute measurements of analytes. Although this was a controlled experiment using a
relatively homogeneous sludge, many pilot and field applications often deal with a wide range of contaminant
distribution resulting in sampling variability which is difficult to handle. The large numbers of samples needed to
measure statistically significant changes can be avoided using biomarkers since analyte responses are
normalized to stable (naturally occurring) internal standards.
31
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposiun
An additional benefit of biomarkers is in site characterization of petroleum spills where ratios of two or more
biomarkers can be used as a fingerprint to help identify the source of the contaminant. The degree of weathering
and natural attenuation in various areas of a spill can also be assessed using this approach.
Abundance
1800-
1600 .
1400
1200 -
1000 -
BOO -
600 -
400 .
200
lime--*
HOT
-^WjvV^-V\ftW
Ion 191. dO (190.70 b'6 m. 70) .- 30B-27-J.il
'
hopane
37.32
A-JLu\
opane
m/z!91
I
j
1
II
.UuJ 1 UJuLxvJ^^^^ ^ w .
3e'.00 37100 38.00 39.00 4oldo 4l!6o 42lo'o 43ldo 44^00
Abundance
800 ,
Time-
700 .
500 .
500
400 -
300 -
200 -
100 -
0 -
- >
36
He
'^Wu
Ion. 177.00 U7S.70 to 177.70): 20ซ-av-3.B
45
rhopane
m/zl77
ft 1.
36^00 37^00 38.00 39^00 4o'.00 4l'.00 iizloO 43.00 44.00
Figure 4. After treatment ion chrmatograms with hopane and norhopane
ACKNOWLEDGMENTS
Dr. Roger C. Prince for his insights into the application of biomarkers to oil spills.
REFERENCES
The Biomarker Guide - Interpreting Molecular Fossils in Petroleum and Ancient Sediments. K. E. Peters and J.
M. Moldowan, Prentice Hall, Englewood Cliffs, New Jersey 07632
32
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Abundance'
200000.
1SOOOO-
100000-
50000
TIC: 20B-I4-1.D
Before
Time--> ' 10 lob' is lob 'a'oio'o' as loo '30 lob' 35 lob 4o!oo
Abundance
14000O -
120000-
100000-
eoooo .
60000 -
40000 -
20000 -
0 -
rime-->
J
'
10
. o'o '
. ,11, I-l"-
TIC: 2U8-27-3.D
UU^L.
After
^u^-
^--
1 1 I ' 1 1 1 1 1 1 r-^-1 1 1 1 T^-T Y ] 1" I 1 T J 1 1 1 1 "T
15.00 20.00 25.00 30 00 35.00 40.00
Figure 5. Before and after
treament total ion
chromatograms
Figure 6. Comparison of
absolute and relative TPH (GC)
degradation
Tims (woks)
33
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
100
90
SO
70
80
* 40
30
20
e j
Pyrana
Figure 7. Degradation of 3, 4
and 5 ring PAHs absolute
values
Time (wuki)
Figure 8. Degradation of 3, 4
and 5 ring PAHs relative to
hopane
i
a Absolute
Relative to HopollB
Figure 9. Comparison of
absolute and relative
C1-phenanthrene / anthracene
degradation
Tunซ [WMkป
34
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Figure 10. Comparison of
absolute and relative
C2-phenanthrenes /
anthracenes degradation
Figure 11. Comparison of
asolute and relative
C3-phenanthrenes /
anthracenes degradation
a Absolute {
Relative to Hopane I
Tim* (wMk>)
METHOD 8270 FOR MULTICOMPONENT ANALYTE ANALYSIS
Elaine A. LeMoine and Herman Hoberecht
The Perkin Elmer Corporation, 761 Main Avenue, Norwalk, Connecticut 06859
ABSTRACT
The identification and quantitation of multicomponent analytes (yielding more than one chromatographic peak)
can be an analytical and productivity challenge. Multicomponent analytes such as Aroclors, Toxaphene, and
technical Chlordane tentatively identified by another method may be confirmed using SW846 method 8270.
Alternate confirmation of a tentative identification may be made using an electron capture detector (ECD)
method such as 8081 or 8082 with a second column. For instruments with sufficient sensitivity, the mass
spectrometer and ECD can be used in parallel for a simultaneous tentative identification and quantification. This
paper will investigate the utility of a new mass spectrometer system for the quantitative identification of a mixture
of multicomponent analytes. The method will be evaluated for detection limits, linearity, accuracy, and precision.
The GC-MS method will be compared with the dual column method for analytical capability, productivity, and
compliance.
35
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
INTRODUCTION
While the ability to positively identify sample analytes can be accomplished with the use of two columns, it is not
necessarily the most desirable of options. In many cases the confirmational column alone is not sufficient and
additional clean-up procedures need to be performed to eliminate co-eluting analytes. The additional equipment
and analysis time required places productivity burdens on a laboratory. Gas Chromatography/Mass Spectrometry
is widely used because its selectivity enables positive identification without additional sample processing. Along
with the ability to make qualitative determinations, GC/MS is an invaluable tool for providing quantitative results.
Mass Spec methods however, are generally considered less sensitive than conventional detector methods,
although sensitive enough for most applications. The analysis of multicomponent analytes, such as Toxaphene
and the Aroclors is more of a challange. EPA Method 8270C states, "In most cases, Method 8270 is not
appropriate for the quantitation of multicomponent analytes, e.g., Aroclors, Toxaphene, Chlordane, etc., because
of limited sensitivity for those analytes. When these analytes have been identified by another technique, Method
8270 is appropriate for confirmation of the presence of these analytes when concentration in the extract
permits."1 The development of more sensitive quadrupole mass spectrometry technology along with innovative
sample introduction techniques, allow for the quantitation of many of these analytes at levels previously not
achievable. The data to follow illustrates the ability of quadrupole mass spectrometry to quantitate
multicomponent analytes at these lower levels. The ability to accurately identify and quantitate using GC/MS can
eliminate the need for additional confirmatory analyses and reduce the amount of sample preparation required.
EXPERIMENT DESCRIPTION
Identical standards were analyzed using two sets of experimental conditions. A 50 uL Large Volume Injection
was used in both cases. One set of standards was analyzed using the GC/MS Full Scan mode (FS-50) and the
second using the Selected Ion Recording mode (SIR-50). Table 1 lists the chromatographic conditions used for
both experiments, while Tables 2 and 3 list the Mass Spec conditions used for each set. The results are
evaluated with respect to the accepted standard analytical techniques.
Perkin-Elmer AutoSystem XL
Column:
Pre-Column:
Oven Temperature Program:
Programable Pneumatic Control (PPC):
Programable Split/Splitless (PSS) Injector:
Injection Volume:
PE-5MS 30 to x 0.25 mm; 0.25 urn film thickness
1 m x 0.32 mm deactivated fused silica
55ฐC for 5 min., 45ฐC/min. to 160ฐC; 6ฐC/min to 320ฐC
Helium 1.0 mL/min.
55ฐC for 4 min.; ballistic to 250ฐC; Solvent Purge Mode
50 uL
Table 1. Chromatographic Conditions
FS-50
Perkin-Elmer TurboMass Mass Spectrometer
Mass Scan Range:
Scan Speed:
Filament Delay:
Ion Source Temperature:
Transfer Line Temperature:
lonization Mode:
50 - 350 m/z
2.0 scans/sec
5 min.
150ฐC
250 ฐC
El
Table 2. Full Scan Mass Spectrometer Conditions
SIR-50
Perkin-Elmer TurboMass Mass Spectrometer
Selected Scan Masses:
Scan Speed:
Filament Delay:
Ion Source Temperature: ,
Transfer Line Temperature:
lonization Mode:
159,231,233
m/z
2.0 scans/sec.
5 min.
150ฐC
250ฐC
El
Table 3. Selected Ion Recording (SIR) Mass Spectrometer Conditions
RESULTS
Toxaphene standards at 0.10 ng/uL, 0.20 ng/uL, 0.50 ng/uL, 1.00 ng/uL, and 5.00 ng/uL concentrations were
36
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
analyzed using both methods. The chromatograms shown in Figure 1 were obtained using the SIR mode. All the
calibration standards clearly exhibit the characteristic Toxaphene pattern.
SO uL SUM
STD126....i3
1Oฐ1 5. Ong Toxaphene
ป %-j
STD12& 5
1 1 Q ng Toxaphene
H
STD128 4 ' ' ' ' ' "
rt
1 05 ng Toxaphene
STD12B 3
1 D.2 ng Toxaphene
% J
nl **
STDI28 2
1 0.1 ng Toxaphene
^f-i 23C)3
ill I 231 1
_-A ^^-v^AuMt>A^JSt'J\k^vVltjlj(>lji,_^,
20.8B
231 1 23,02
,11 i 231 1
^ _Aw- oA.^A.H^X/MA<^wU^.^mi,, .
i i ... i .,.,..,,,..,..,.,.
2O H7 21 76 23.02
231 .1 231 231
j^^^^iJ^^
' ' ' ,....,..,,, ....,.,, , .
20.87;231
Jt_j^_u^JU^^^
''" 1 .... I ....,...., , .
20.4S_ 21.76 2301
231 K I 231 231
KIR of 3 Cliannels EH
1SS.BO
1 .4667;
-
SIR of 3 Channels El-t-
159 OD
SIR of 3 Channels El-i-
159 DO.
1 S1e6
SIR of 3 Channels ฃ!+
1SS.OD
6.7165
SIR of 3 Channels El-t-
1 S9 l)D
10.00 12.0O 14.00 16.00 18.00 2000 22.00 24.00 26.00 28.00
Figure 1. Calibration Standards show recognizable pattern for all levels.
Full Scan Mode
Compound 3 name: Toxaphene-3
Correlation coefficient: r ^-- 0.99931 2, rA2 = 0.999624
Calibration curve. 674.670 * x H- 79.0644
Response type. External std. Area
Curve t/pe: unear. Origin: txciude, Welgrtting: Null, Axis tran
6.83e3-|
Response-
0.0 2.0 4.0
6.0
8.0
rrn ng
10.0
Four chromatographic peaks were selected and
determined as representative of the
multicomponent analyte Toxaphene.
Calibration Factors (CF) were calculated based
on the integrated peak areas and the known
standard concentrations. From these results,
the Relative Standard Deviation (RSD) for each
multilevel concentration range was determined.
These results were averaged providing a final
Toxaphene RSD. Correlation coefficients were
calculated in a similar fashion and are
illustrated in Figures 2 and 3.
The results of all the calibration data and
acceptance criteria are listed in Tables 4 and 5.
Both experimental results easily comply with
method performance specifications.
Figure 2. Toxaphene peak #3 Calibration Curve.
37
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
SIR Mode
Compound 2 name: Toxaphene- 2 '.
CoeiTicient or Determination. 0.993732 Tne Method Dptprtinn Limits (MDLsl Iktprl in
caiioration curve 13SSS3 * x + -2213 52 i ne lvieinoa ueiecuon umiis JMUUS; usiea in
Response type: external std. Area Table 6 are the result of seven (7) replicate
Curve type:
6 7365-1
.
Response-
-2 21 e3-
O
Calibration
Peak #2
Peak #3
Peak #4
unear. origin: include, weighting: NUM. AXIS trans injections of a 0.10 ng/|jL standard using the
^ standard deviation and the t-statistic.
/' Integrated peaks representative of the entire
/ calibration range can be seen in Figure 4. The
S bottom chromatogram was obtained from a
/" 0.05 ng/^L standard which is below the lowest
s^ calibration standard of 0.10 ng/uL. The peaks
y are readily discernible above the noise and
/ can be easily integrated.
^ Figure 3. Toxaphene peak #2 Calibration
sX Curve
.O 1 .Q 2.O 3.Q A.O 5.0
Full Scan-50
Peaks
Toxaphene (Average of 4 peaks)
RSD
Actual
11.0
13.6
12.6
8.4
11.4
Acceptance
Limit
f\ " i^fi^lf^fef^^f&Mait^^^^^Ss
15.0
Correlation Coefficient
Actual
0.99934
0.99949
.99962
0.99948
0.9995
Acceptance
Limit
:'-M?iyi
0.99
Table 4. Comparison with Calibration Acceptance Criteria using Full Scan mode.
Calibration
Peak#1
Peak #2
Peak #3
Peak #4
SIR-50
Peaks
Toxaphene (Average of 4 peaks)
RSD
Actual
10.4
8.2
10.8
7.5
9.2
Acceptance
Limit
iiliSip|u||ilpi^^p
15.0
Correlation Coefficient
Actual
0.99936
0.99973
0.99967
0.99934
0.9995
Acceptance
Limit
0.99
Table 5. Comparison with Calibration Acceptance Criteria using Selected Ion mode.
Calibration Peaks
Peak #1
Peak #2
Peak #3
Peak #4
Toxaphene (Average)
Calculated Analytical Detection Limits
FS-50
(ng/ML)
0.073
0.089
0.105
0.035
0.07
SIR-50
(ng/ML)
0.065
0.009
0.021
0.014
0.02
Table 6. Calculated Detection Limits
38
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Toxaphene
STD 1 28_S
100-
Oj
19.47 1 00 ng/uL
231
2G2BE
^A^,.J\^^^Vyv\J
STO12S 4 ' ' '
100-
Area
Ul J
o-
12s17 ฐ 5ฐ n9/UL
14546
--"^- X_^-
STD12R_L>
100-
Area
%
STD1
100-
0-
I1947 0.10ng/ul_
, 5642 p
2B 1
1947 0 Q5 ng/uL
231
3612 ^ %_/"
19.00 1950 20. QQ
20.87
A ii^a ai-78=2J1-aM=4
\yV*xJk
ปA-^/v^xJLWV
_20.87
A ifii^ 2175.231:49710
rA\ \ k A
" ~"^*
"N,-/1^^7 ^^t_X^ V
ft 20.B6;231;)9258
/ ^ A i. 21. 75.231. 119B6
U ^-ซk^,-\. - - ^ '"^^
I i.i i J.
20.87
JWi
20. 5D 21
21.75
231
8311
OD 21.50 22.00
SIR Of 3
^23.02
IT 231
1175806
v,. -^^w^A-JLys.^
BIR Of 3
23.02:231 :3B978
k, *-*' \.^-j"^\jf^^A^\__
3IR of 3
23.01:231:8966
Channels EI+
159.00,
2,5766
~^_
''M'"
Channels EI+ ,
159.00'
1 43e6
-V . -
Channels tl*
159 00
5.2165'
SIR nf 3 Channels EI+
23.02
231
5857
22.50 23.00
159 OO
6. OSes
23. 5O :
Figure 4. Integrated Toxaphene Peaks
STD209_18
1OO-I
scan EM- ,
66
di u^2ny_ i fct
1QO-,
Toxapfnene
A - ^A.
^-^- . . ^-- '^.,s^-r^-^- ^- '*
A--,
i
A
>_^A/W
4
\/K,...^
1 OO ng Toxapfiene and O.1 ng Pesticide Mix
1 cap 1 oso i 1 op i.i_s.Q _^2oo 12so 1300
Figure 5. Aldrin and Toxaphene Extracted Ions
l^Q9_1 8
1 DO-,
Scan ei-t- .
SSi
a i Li:?uy _i a
oxaphene
1OO ng Toxapriene arid O.1 r>8 Pesticide Mix
366 ' ib'o'o ' " ib's'p ' II'DO .t.i.sp. iapo .1.25O 1?QQ ISSD
Figure 6. Library Searchable Spectra
39
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
What sets the Mass Spectrometer apart from other forms of detectors is the ability to selectively-identify
individual masses. Figure 5 shows the Total Ion Chromatogram (TIC) of a mixture of 100 ng/uL Toxaphene and
0.10 ng/uL Pesticide Mix. The Extracted Ion (El) mass 159 is Toxaphene, and the Extracted Ion (El) mass 66 is
Aldrin which was confirmed by a NIST library search as seen in Figure 6. Aldrin is easily identified and integrated
without additional preparatory procedures.
The Contract Laboratory Program (CLP) lists 0.2 ng/uL as the quantitation limit for Aroclor 1221 using an
Electron Capture Detector. Figure 7 shows Aroclor 1221 well above the noise level at the 0.20 ng/uL quantitation
level, using GC/MS in the SIR mode and large volume injection.
XM-Qcldr 1 221
O.2O nti/uL_
SIF? MortR
L.V1-5O
31.2231
326 3i
21.IJU '^2-ua 23!oU :iM LJU 2S QO i>6.OU 'i7.OO 2Q.OO 29 OO 3Q-Otl 31.PO S
Figure 7. Quantitation Limit Pattern Recognition
SUMMARY
The ability of GC/MS to selectively identify a component based on an extracted ion chromatogram from a
mixture of compounds not only assures a positive identification, but also saves time by eliminating additional
cleanup and analyses. Recent technological advances in quadrupole Mass Spectrometry have increased the
instrument's sensitivity. The use of Selected Ion Recording provides further sensitivity enhancements. In addition
to the detector and it's mode of operation, the use of large volume injection with a programmable inlet system
allow for introduction of larger sample volumes. The combination of these elements enhances the sensitivity of a
GC/MS system so multicomponent analytes can be identified and quantified in an efficient and productive
manner.
References
1. "Method 8270C, Semivolatile Organic Compounds by Gas Chromatography/Mass Spectrometry (GC/MS)," in
Test Methods for Evaluating Solid Waste Physical/Chemical Methods, SW-846, Third Edition, U.S.
Government Printing Office, Springfield, VA (1996).
BENZIDINE? REALLY?
Roy-Keith Smith
Analytical Services, Inc., 110 Technology Parkway, Norcross, GA 30092
(770) 734-4200, Fax (770) 734-4201
ASI has been performing sample analysis using GC-MS Method 8270, with a variety of sample preparations from
SW-846 for many years. In 1995 soil samples were received from a manufacturing site for analysis. The samples
were part of a general site survey of the manufacturers facility to determine what may need remediation efforts in
the future. ASI performed the requested analyses, and found and reported benzidine in several of the samples at
levels approaching 1000 mg/kg. Unfortunately the data reports were simply filed by the manufacturer.
In late 1997, the manufacturing site became under consideration for sale to another company. As part of the
40
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
pre-sale investigation of the site, past analytical records were examined and the benzidine results came to light.
As there had never been any benzidine used or stored on the site and the manufacturing processes involved no
chemical syntheses, there were some questions about the validity of the reported results.
The raw data generated during the sample analyses was examined in detail. The initial calibration was
acceptable, as were the daily calibration verifications and tunes. The system performance criteria were being
met. The initial calibration was used for quantitation, with retention times and user generated library spectra
being updated on a daily basis. Examination of the raw chromatograms (Figure 1) from the samples revealed a
hump-o-gram. Random MS scans from the hump suggested a petroleum-based background interference.
Although the benzidine hits were buried in the hump-o-gram, rather than being isolated well defined peaks, the
candidates matched up perfectly with the retention time and daily generated mass spectrum of the benzidine
standard.
The client had not requested a library search for TIC with the original analysis. By happenstance during the data
review, a library search of the questioned peak was performed. Quite surprisingly the search generated a match
for dibenzothiophene, rather than the expected benzidine. Spectra of both compounds were pulled from the
database for examination.
It is rather startling how similar the mass spectra of benzidine and dibenzothiophene appear in a fast visual
comparison (Figures 2 and 3). Both have a dominant peak at m/z 184 and an assortment of low intensity smaller
m/z signals. The two compounds have the same unit mass molecular weight (184), the only difference in the
molecular formulas being the two amino groups in benzidine and the sulfur in dibenzothiophene, Ci2H8(NH2)2 vs.
Ci2H8S. By coincidence the mass of the two amino groups (32) is the same as the sulfur (32).
Detailed peak mass matching and relative abundance comparisons of select peaks reveal definitive mass
spectral variances between the compounds that allows conclusive identification of either analyte. First off, both
compounds have a M + I isotopic peak, however only dibenzothiophene has a very prominent M + 2 signal due
to the sulfur. The fragmentation patterns of the molecular ions of the two compounds are nowhere near alike.
Both compounds can loose a hydrogen to give M-1 signals, however only benzidine will continue to lose
hydrogens giving the M-2 and M-3 peaks. Benzidine can also lose a -NH2 group to generate the M-16 peak at
m/z 168 or a NH3 group, forming a benzyne at m/z 167 (M-17). Dibenzothiophene has no fragmentation pathway
to generate either m/z 167 or 168. Extrusion of sulfur from dibenzothiophene gives rise to a prominent peak at
m/z152(M-32).
The stability of the molecular ion of benzidine is probably enhanced through ring-expansion of one of the
aromatic rings to include a nitrogen in a seven-membered aromatic ring (aza-tropyllium ion). Concerted ejection
of a CNH4 unit (M-30) from this ring generates m/z 154. A similar ring expansion, followed by concerted ejection
of SCH from dibenzothiophene forms m/z 139 (M-45) as a significant signal that is quite undistinguished in the
benzidine spectrum. Other important differences are indicated in Figures 2 and 3, and include m/z 65 and 77 in
the benzidine spectrum, while Dibenzothiophene exhibits 69 and 79.
Although the distinguishing features of the two spectra are easy to overlook by eye, it was obvious that the
computer spectral matching algorithm was having no such problem, and further investigation focused upon the
user generated spectra. This is displayed in Figure 4, along with the spectra from one of the challenged
identifications. What was in the sample matched up almost perfectly with what was stored as a spectrum of
benzidine from the standard. Using the identification criteria listed above to examine the library spectra led to the
inescapable conclusion that the standard used for initial and continuing calibration was dibenzothiophene rather
than benzidine.
The retention time of dibenzothiophene is slightly less than that of benzidine, however not so much as to be
really startling. Benzidine itself exhibits shifts in absolute and relative retention times as columns are changed in
the GC-MS. As it is our habit to replace columns with recalibration, the slight shift in retention time from one
initial calibration to the next was unexceptional. The quick visual examination of the mass spectra that
accompanied the recalibration failed to detect any differences.
Dibenzothiophene is a naturally occurring substance commonly found in high sulfur crude oils. Discussions with
the client revealed that the samples with "benzidine" all came from the soil underneath a storage area where
several barrels of high sulfur Venezuelan fuel oil #6 had been placed. The high sulfur levels had made the oil
41
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Abundance"
TIC: BU27L.D
85OOOOB j
8000000 :
7500000^
1
1
7000000 1
6500000 \
5000000
S5000OO.
50000OO .
4SOOOOO .
40000OO-.
3500000 .
30OOOOO
2500000
200DOOO _
1500OOO 1
1000OOO .
soooao-
0
!
i
5.00
10.00
15.00
20.00
25.00
30.00
unacceptable for use in the boilers at
the facility. The GC-MS chromato-
grams contained hydrocarbon hump-
o-grams along with the "benzidine"
and other sulfur-containing aromatics.
As all the evidence was now
consistent, the reports were re-issued
deleting the benzidine hit.
The investigation was expanded to
include examination of the calibration
spectra both prior to and after these
particular samples were analyzed. It
was found that the problem began
several months prior to October, 1995
surprisingly in the middle of a
supplier's lot number. The problem
continued after October, 1995, through
the next lot number of benzidine
standard purchased from the supplier.
It was not until Summer, 1996 and a
further lot number change that the
spectra reverted back to the correct
benzidine mass spectrum.
Figure 1. Chromatogram of sample
ASI contacted the supplier of the benzidine standard and presented them with the above evidence. The supplier
was not making any spectral checks upon purchased stock standards as part of their QA program. Only the
technically outdated melting point determination was being verified, and it was not being performed as a mixed
melting point. There were no ampules of the particular lot numbers available for examination, however the
supplier offered to reimburse the cost of the standards.
We went back through every sample analyzed during the year that the incorrect standard had been used.
Fortunately we found that no other samples had been reported with hits for benzidine, thus there were no false
positives. We are still in the process of doing a manual search of the tape archives for any false negative
benzidine hits in these samples. None have been found to date.
How could we have caught the error and
prevent it's happening again in the future?
The corrective actions we have instituted
are to: 1) inject new GC-MS standards
under old calibrations prior to changing out
the column and re-calibrating; 2) use second
source standards to check each new
calibration; and 3) closely examine the mass
spectrum of each new benzidine standard
that is purchased.
The lesson learned? Don't trust anyone's
claims or documentation of purity or
authenticity. Their acceptance standards
may well be different than your own.
Figure 2. Mass spectrum of benzidine from
Wiley-NIST library
1SB
111
ill
i !'!'> ii i I'l'ii 111 iTI'i
US 168 17B IBB
Wafc1"*'
unr
Utn
lll'l'lll. I'I'll I I l'1'l'l I
11B its l^B 1ซ
C12H12N2
42
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
in
iti ii i^11!^"^1^11^"^"^"^1 ^"^"^"y,;1^1 ^.
Figure 3. Mass spectrum of dibenzothiphene
from Wiley-NIST library
zfc1 ak
oB yB OB ?B
D Lie troth lophene
tft 119 \St
C1ZHBS
Abu'ndance
5000 .
Figure 4. Mass spectra
from sample and lab
generated library
Abundance
41 55 69 92 11S9 I "'
~rVJ,' ,*^ ,-VlA"l'VV"l'Jl'Jll-| .' W\- \ I'rvf'i' 'i ''l''i'
40 60 80 100 120 140 II
196
' ' '
218232246 264278
5000
I J , 1U11. , -. ซ-ซ.ปw ซu-Kซ.fU
160 180 '266 zio''aJo^^o^aS"OT
Scan" TffTB~T21.333 minT: BOX68TD" ~{- ",~*T
184
3SIS
79
92
139
113 13 2
" -LT-T-T'IT i1 r 1 i rV ^I'V-l-'t"!''!1'i'[T-t~flVrT"r-rji'-'[ i'i "I j'i-i1 i "n i'V''i i I-T-T i i i i--i ' i i i i-r-T-r-T-, , , r-pi-
Ti/Z--> 40 60 80 100 120 140 160 180 200 220 240 260 280
204
250 268 286
COMPARISON OF VOLATILE ORGANIC COMPOUND RESULTS BETWEEN METHOD 5030
AND METHOD 5035 ON A LARGE MULTI-STATE HYDROCARBON INVESTIGATION
Rock J. Vitale, Ruth Forman, and Lester Dupes
Environmental Standards, Inc., 1140 Valley Forge Road, Valley Forge, Pennsylvania 19482
ABSTRACT
With the promulgation of SW-846 Update III during June of 1997, elimination of Method 5030 and
implementation of Method 5035 have created significant challenges for the regulatory and laboratory
communities (US EPA, 1997). Based on historical data, the results for volatile organics in certain sample types
using the previously approved direct heated purge technologies were observed to be biased low (Hewitt, 1994).
The loss of volatile organics was not observed to be the determinative process of Method 5030 but of the
sampling, preservation, and preparatory aspects of the methodology (Hewitt, 1997, Siegrist, 1992). The
promulgation of Method 5035 requires training of field samplers and a decision-tree approach to collecting and
analyzing samples.
43
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
As recent as the first quarter 1998, for previously initiated on-going projects, various regulatory agencies have
been issuing directives to either implement Method 5035 or continue with Method 5030 (US EPA, 1997). In order
to understand the implications of implementing Method 5035 after several years of using "traditional" soil
sampling methods and analysis by Method 5030 on a large multi-state hydrocarbon investigation, a 4-week study
was performed to ascertain the differences in field activities, documentation issues, and analytical results.
A description of the comparative study, inclusive of observations between the traditional Method 5030 and the
three Method 5035 options (e.g., Encoreฎ, sodium bisulfate, and methanol), will be presented. In addition, some
of the lessons learned from the study will be discussed.
INTRODUCTION
A two-phased study was performed to determine the comparability between investigative samples collected and
analyzed by the "traditional" sampling/analytical method (SW-846 Method 5030) and investigative samples
collected and analyzed by the recently promulgated sampling/analytical method (SW-846 Method 5035). Soil
samples included in this study were collected from four states with different soil types and contaminant
concentrations. The soil samples were analyzed for a list of 18 volatile compounds by gas chromatography/mass
spectroscopy (SW-846 Method 8260A).
STUDY DESIGN
The study included the collection of soil samples in two phases. The first phase of sampling was conducted at
select locations in Ohio, West Virginia, Pennsylvania, and Maryland from January 27 through February 11,1998.
Soil samples were collected at 56 sample locations. Nine field duplicates were collected during this phase. A trip
blank (one sodium bisulfate and one deionized water) was collected for each day of sample collection. Based on
previous analyses and remedial activities, samples collected during phase one were expected to either be
"clean" or require low-level analysis. Accordingly, methanol samples were not collected for this phase.
Specifically, for the first phase of sampling, samples were collected utilizing three techniques as follows:
The traditional method of sample collection (in a 125 ml wide-mouthed glass jar);
Utilizing a plastic syringe and placing five grams of soil into a 40 ml glass vial pre-preserved (by the
laboratory) with sodium bisulfate; and
Utilizing an Encoreฎ sampler (the Encoreฎ analyses for both phases were performed with the modifications
recommended by the International Association of Environmental Testing Laboratories [IAETL]).
The second phase of sampling was performed the week of March 2, 1998. Samples were collected at 33 select
locations (including two field duplicates) in West Virginia. The selection criteria for sample locations collected for
the second phase of the project were based upon available historical sample concentration data. The sample
locations were selected to include samples that contained low, medium, and high concentrations of volatile
compounds (based on the traditional sample collection historical data).
Trip blanks (one sodium bisulfate, one deionized water, and one methanol) were collected for each day of
sample collection. For the second phase of sampling, samples were collected utilizing four techniques as follows:
The traditional method of sample collection (in a 125 ml wide-mouthed glass jar;
Utilizing a plastic syringe and placing five grams of soil into a 40 ml glass vial pre-preserved (by the
laboratory) with sodium bisulfate;
Utilizing an Encoreฎ sampler (the Encoreฎ analyses for both phases were performed with the modifications
recommended by the IAETL); and
Placing five grams of soil into a vial containing methanol.
All samples were packed in coolers at 4ฐC under Chain-of-Custody and shipped via overnight courier to a
reputable commercial environmental laboratory. All samples were analyzed within a holding time of 14 days of
samples collection. In addition, each sample type (i.e., traditional, Encoreฎ sodium bisulfate and methanol)
44
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
collected at a given sample location was analyzed within 24 hours of the other sample types. In order to
minimize possible confounding effects of analytical holding times, all sample types at a given location were
analyzed within the sample 24-hour time period.
DATA REVIEW
Reduced data package deliverables were prepared by the laboratory for all samples. The reduced data package
deliverables included a summary of the reported analytical results for all field samples (including samples, field
duplicate samples and trip blanks), the associated method blank results, the associated laboratory control sample
recoveries, and the surrogate recoveries. The analytical data for both phases were reviewed for completeness of
the data package deliverables, compliance with the SW-846 Methods 5035 and 8260A, and usability of the
reported analytical results (Clark and Vitale, 1996). Although important for comparability, compliance with
methodologies often provides little information on data quality (Blye and Vitale, 1995).
The initial and continuing calibration criteria for Method 8260A were met for all study samples. With the
exception of one compound in one laboratory control sample (LCS), the LCS recoveries were within
study-specified limits (75-125%). The recoveries for one or more of the three volatile surrogate compounds were
outside the limits specified (varying limits for each surrogate with limits between 70-121%) for many (38) of the
study samples. Because the "true value" accuracy was not important for this study, the surrogate recoveries are
not expected to affect the operational definition for comparing techniques on respective sample aliquots, which
was most relevant in a comparative study. Analysis of the study trip blanks and laboratory method blanks did not
reveal the presence of target analytes, with the exception of methylene chloride and acetone. The positive
sample results for these two compounds were rejected from further consideration.
Eleven field duplicate samples were collected for the study; approximately one field duplicate sample was
collected per day of sample collection. Field duplicates provide valuable information on precision and sample
representativeness when evaluated properly (Zeiner, 1994). Acceptable precision (<50% RPD, as defined for this
project) was noted between six field duplicate pairs. High RPDs (>50% RPD) were noted for five of the duplicate
pairs. Such sample variability appears to have complicated a meaningful comparison of techniques as separate
sample aliquots were collected and analyzed for each of the study sampling techniques.
SUMMARY OF RESULTS
A summary of the reported analytical results is provided on Table 1. All results are reported on a dry-weight
basis. Variations between sample-specific quantitation limits were evident due to the sample collection volume,
the percent moisture of the sample, and the sample-specific dilutions performed. The sample collection volume
at a given sample location is different for each sample type due to the manner in which the sample was
collected. The quantitation limits for samples preserved with methanol were raised by the laboratory due to the
medium-level sample preparation.
Fourteen of the sodium bisulfate samples were not analyzed due to the observed concentrations of non-target
compounds in these samples (samples containing sodium bisulfate cannot be analyzed using medium-level
protocol). The corresponding traditional and Encoreฎ samples at these sample locations were analyzed at a
medium-level due to the same reason. In addition, six of the sodium bisulfate samples could not be analyzed due
to observed sodium bisulfate effervescence.
In total, 79 Encoreฎ samples, 79 traditional samples, 62 sodium bisulfate preserved samples, and 23
methanol-preserved samples were collected for the volatile pilot study. Of the 79 sample locations examined,
positive results were reported in 33 of the sample locations. Positive results were reported for four aromatic
compounds (benzene, toluene, ethylbenzene, and total xylenes); two ketones (4-methyl-2-pentanone and
2-butanone); and four chlorinated aliphatic hydrocarbons (1,1-dichloroethene, 1,1,2,2-tetrachloroethane,
1,1,1-trichloroethane, and tetrachloroethene).
45
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
TABLE 1. SUMMARY OF ANALYTICAL RESULTS
Sample Number
SAMPLE 1
total xylenes
SAMPLE 2
benzene
ethylbenzene
total xylenes
SAMPLE 3
toluene
2-butanone
SAMPLE 4
toluene
SAMPLE 5
toluene
SAMPLE 6
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 7 (DUPLICATE OF SAMPLE 6)
tetrachloroethene
toluene
total xylenes
SAMPLE 8
total xylenes
SAMPLE 9 (DUPLICATE OF SAMPLE 8)
total xylenes
SAMPLE 10
total xylenes
SAMPLE 11
1,1-dichloroethene
1,1,1-trichloroethane
toluene
ethylbenzene
total xylenes
SAMPLE 12
1 ,1 ,2,2-tetrachloroethane
ethylbenzene
total xylenes
SAMPLE 13
toluene
SAMPLE 14 (DUPLICATE OF SAMPLE 13)
1 ,1 ,2,2-tetrachlorethane
SAMPLE 15
4-methyl-2-pentanone
SAMPLE 16
4-methyl-2-pentanone
SAMPLE 17
toluene
ethylbenzene
total xylenes
SAMPLE 18
4-methyl-2-pentanone
ethylbenzene
total xylenes
Encore
6U
6U
6U
6U
6U
95U
6U
6U
6U
6U
6U
6U
5U
5U
5U
36
610
760
200J
2200
1200
21 OJ
930
310
200J
1200
5U
6U
31,000
13,000
1000
6200
30,000
2700U
2000
20,000
Traditional
6U
6U
6U
15
5U
110
6U
7U
6U
6U
6U
6U
6U
6U
6U
140J
0
440U
0
130U
320U
570
110J
320U
160J
330U
280J
2100
6U
6U
9700
8400
1600
12,000
63,000
6500U
1400
18,000
Sodium Bisulfate
6
6
12
14
8
110U
7
12
23
23
15
260
13
9
7
45
0
N/A
0
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
6
16
N/A
N/A
15
16
53
1,300
790
240
Methanol
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
46
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
TABLE 1. (ConU
Sample Number
SAMPLE 19
toluene
ethylbenzene
total xylenes
SAMPLE 20
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 21
toluene
2-butanone
SAMPLE 22
1,1-dichloroethene
toluene
SAMPLE 23
benzene
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 24 (DUPLICATE OF SAMPLE 23)
benzene
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 25
benzene
toluene
ethylbenzene
total xylenes
SAMPLE 26
ethylbenzene
total xylenes
SAMPLE 27
benzene
ethylbenzene
SAMPLE 28
benzene
toluene
ethylbenzene
total xylenes
SAMPLE 29 (DUPLICATE OF SAMPLE 28)
benzene
toluene
total xylenes
SAMPLE 30
ethylbenzene
total xylenes
SAMPLE 31
benzene
ethylbenzene
total xylenes
Encore
63
20
160
6300
5U
5U
11
7U
130U
5U
5U
250J
901
1300
470
4600
250J
80J
1200
570
5600
5U
5U
5U
5U
600
370
6U
6U
2000
7400
1500
14,000
6U
6U
6U
120J
60J
5U
5U
5U
Traditional
34
11
59
5000
6U
6U
6U
6U
120U
6U
6U
220J
300U
1100
550
5500
220J
300U
1500
430
4500
6U
6U
6U
6U
230J
110J
6U
15
6U
6U
6U
6U
7J
30
4600
120U
320U
13J
23J
33
Sodium Bisulfate
61
31
280
2900
14
9
24
7
120
16
6U
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
45
90
26
130
N/A
N/A
12
7U
N/A
N/A
N/A
N/A
6U
6U
23
N/A
N/A
6U
6U
7
Methanol
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
530U
100J
670
200J
3300
870
8400
400J
200J
1900
740
7200
570U
570U
570U
300J
1200
930
61 OU
610U
650U
200J
650U
600J
540U
100J
780
0
340U
80J
260U
60J
670
47
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
TABLE 1.(Cont.)
Sample Number
SAMPLE 32
toluene
ethylbenzene
total xylenes
SAMPLE 33
benzene
toluene
total xylenes
Encore Traditional
21 OJ 280U
540 280U
3300 280U
6U 6U
6U 6U
6U 6U
Notes: U - Not detected at or above the associated numerical value
N/A Not analyzed
J - Estimated value; result less than the quantitation limit
E Estimated value; result exceeded the calibration range and
Sodium Bisulfate
N/A
N/A
N/A
8
14
8
not reanalyzed
Methanol
470U
470U
550
61 OU
61 OU
610U
COST ANALYSIS
For the reputable commercial laboratory that was utilized for this study, the cost of analysis of all four sample
collection types is identical, and there are not any cost implications for analyzing a volatile soil sample by any of
the four sample collection methods. The cost of sample bottleware preparation is different between the four
sample collection types. The cost for traditional sample collection bottleware is built into the contract with the
laboratory. The cost for the Encoreฎ sampler is roughly $10.00 per sampler. Three Encoreฎ samplers are filled at
each sample location at a total of <$30.00 per sample location. The cost for a methanol-preserved vial (two vials
are sampled at a given location when the methanol preserved sample is not collected in tandem with a sodium
bisulfate preserved sample; otherwise, only one vial is submitted) is <$35.00 per sample or <$60.00 per sample
location. The cost for a sodium bisulfate preserved vial is approximately $50.00 per sample. Two sodium
bisulfate preserved vials are collected per sample location at a cost of approximately $100.00 per sample
location.
The labor costs for the collection of the various sample types can vary and exact costs could not be precisely
calculated based on information available. The Encoreฎ sampler and the traditional method of sample collection
were found to be the quickest and easiest sample types to collect. The sodium bisulfate samples were found to
take the longest time to collect and provided the most difficulty in the field.
FIELD OBSERVATIONS
Sodium Bisulfate and Methanol Sampling Techniques
The sodium bisulfate and methanol techniques are very similar and are reviewed here jointly. These methods
were by far the most cumbersome of the selected sampling techniques. The necessity of using an analytical
balance under imperfect field conditions proved frustrating. The balance readings would fluctuate wildly with the
slightest amount of wind or movement of the sample preparation surface. In addition, unless the sample
preparation surface was perfectly horizontal, the scale could not calibrate nor zero out. Furthermore, calibration
of syringes with soil matrix was time consuming in comparison to the traditional and Encoreฎ methods. Using
razor knives to cut syringe tips evenly required practice and care. Finally, preservative solution was prone to
spillage unless great care was taken while placing the soil in the vials.
The methanol technique has an additional drawback. When shipping samples via popular air couriers, samples
must be packaged according to IATA regulations. The regulations regarding shipping of flammable compounds
are very strict and may result in lost or delayed delivery if the sample containers are not properly packaged. In
addition, the appropriate paperwork and package labeling must be completed in a precise manner, or the
shipment will be delayed or returned to the sender. From a field sampling perspective of the methods used
during the pilot study, these were by far the most time consuming and frustrating.
Encoreฎ Sampling Technique
The Encore sampling system was very straightforward in its approach and implementation, although some
minor problems were encountered during the pilot study. Problems were encountered when trying to place loose
and/or wet soils into the Encoreฎ sampler. Soils of this type had to be manipulated into the Encoreฎ with another
sampling device (such as a spatula). The only other sample collection issue involved improper seating of the cap
on the plunger. However, by pushing down on a hard surface with the T-handle, the cap could be seated
properly. Overall, the Encoreฎ system appeared to be easy to use even under adverse field conditions.
48
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Traditional Sampling Technique
This sampling method is well-known to most field technicians and has been employed under a wide variety of
field conditions. It was apparent during the field study that this technique was easily implemented and rivaled the
Encoreฎ sampling technique for ease of use. However, from a field perspective, some clays, silts and other
tightly compacted soils are difficult to place into a sample container so that no head space is allowed. Breaking
up soils to place into sample containers may result in loss of volatiles, thereby lowering detectable
concentrations. However, from a field perspective, this technique has been historically easily implemented.
CONCLUSIONS
The following observations can be made from the overall study. With the exception of the first bullet item, the
conclusions presented below should not be interpreted to be applied to other sites, soil types, and concentrations
of analytes. These are general conclusions regarding this particular data set; there may not be an equivalent
trend noted in other data sets. Our findings suggest that inherent difficulties associated with analyzing soil
samples makes definitive states regarding data comparability difficult. Furthermore, the number of positive data
points and the disparity observed for half of the collected field duplicates makes statistical trend analysis
problematic at best.
At sample locations where methanol-preserved samples were collected and where the concentration of target
analytes was within range of the medium-level analysis, the concentration of target analytes of the
methanol-preserved sample type was greater than the concentration of the analytes in the other sample
types at the same sample location. This is consistent with other studies appearing in peer-reviewed literature.
At sample locations where methanol-preserved samples were not collected and where the concentration of
target analytes was within range of the medium-level analysis, the concentration of the aromatic analytes
appeared to be greater in the traditional sample collection type than the other sample collection types at the
same location.
At sample locations where methanol-preserved samples were not collected and where the concentration of
analytes was within range of the medium-level analysis, the concentration of the non-aromatic target
compounds appeared to be greater in the Encore sample collection type than the other sample collection
types at the same location.
At sample locations where methanol-preserved samples were not collected and where the concentration of
target analytes was not within range of the medium-level analysis, the concentration of the analytes
appeared to be greater in the sodium bisulfate sample collection type than the other sample collection types
at the same location.
REFERENCES
Blye, D.R. and R.J. Vitale. "Data Quality - Assessment of Data Usability Versus Analytical Method Compliance."
Proceedings from the Eleventh Annual Waste Testing and Quality Assurance Symposium. Washington, DC,
23-28 July 1995.
Clark, M.A. and R.J. Vitale. "How to Assess Data Quality for Better Decisions." Clearwater, New York Water
Environmental Association (NYWEA), Vol. 26, No. 2 (1996).
Hewitt, A.D. "Losses of Trichloroethylene From Soil During Sample Collection, Storage and Laboratory
Handling." SR 94-8, US Army Corps of Engineers, Cold Regions Research & Engineering Laboratory. 1994.
Hewitt, A.D. "Preparing Soil Samples for Volatile Organic Compound Analysis." SR 97-11. US Army Corps of
Engineers, Cold Regions Research & Engineering Laboratory. 1997.
Siegrist, R.L. "Volatile Organic Compounds in Contaminated Soils: The Nature and Validity of the Measurement
Process." J. Hazard Mater Vol. 29 (1992):3-15.
US EPA. "Determination of Volatiles in Soil - Directive for Change." Memorandum from Mr. Norman R.
Niedergang, Director of Waste, Pesticides and Toxics, US EPA Region 5, to Corrective Action Project
Managers and QA Staff. 1997.
US EPA. "Test Methods for Evaluating Solid Waste." SW-846, Update III, United States Environmental
Protection Agency. 1997.
Vitale R.J. "Balancing Regulatory Compliance with Technical Validity." Proceedings from Soil Sampling for
Volatile Organics Workshop. O'Hare International Holiday Inn, Chicago, IL, 10 December 1997.
Zeiner, ST. "Realistic Criteria for the Evaluation of Field Duplicate Sample Results." SUPERFUND XV.
Washington, DC, 29 November-1 December 1994.
49
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
50
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
INORGANIC
51
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
52
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
SW-846 INORGANIC METHODS UPDATE
Oliver Fordham
U.S. EPA, Office of Solid Waste, 401 M. St., SW, Washington, DC
NO ABSTRACT AVAILABLE
DIRECT MERCURY ANALYSIS: FIELD AND LABORATORY APPLICATIONS
Helen M. Boylan. H.M. "Skip" Kingston, and Robert C. Richter
Duquesne University, Department of Chemistry & Biochemistry, Pittsburgh, PA 15282-1503
ABSTRACT
EPA Method 7473 is designed for the determination of total mercury in solid and aqueous samples. This method
is based on the instrumental methodology of the Direct Mercury Analyzer-80 (DMA-80) (Milestone, Inc.) in which
sample preparation and analysis are essentially integrated into a single analytical step. The method's unique
capability for direct analysis allows for application in either laboratory or field settings. Method 7473 has been
validated by analysis of various Standard Reference Materials (SRMs) in both the laboratory and in the field.
This validation data has been presented1 Results from Method 7473 have also been confirmed by independent
analysis using traditional methods. Method 7473 has been used on-site in conjunction with mercury remediation.
Real-time analysis using this technique has provided an accurate and cost-effective risk assessment of mercury
contaminated sites.
INTRODUCTION
There are several analytical techniques that may be applied for the determination of mercury in solid waste.
Existing EPA methods for the analysis of mercury include inductively coupled plasma atomic emission
spectroscopy (ICP-AES) (Method 601 OB), cold vapor atomic absorption spectroscopy (CV-AAS) (Method
7471A), and anodic stripping voltammetry (ASV). Regardless of the method used, sample preparation is
required. "Soils, sludges, sediments, and other solid wastes require digestion prior to analysis"2. Method 7473 has
an advantage over traditional mercury analysis because it eliminates the need for a discrete sample preparation
step. Direct analysis is performed by integration of thermal decomposition, amalgamation, and atomic absorption
spectroscopy. While the fundamental theory for this type of analysis has been available in the literature, the
DMA-80 is the firs instrumental implementation of these integrated concepts.
A schematic of the DMA-80 is shown in Figure 1. The sample is automatically inserted into the quartz
deomposition tube, where it is first dried and then thermally decomposed. The gaseous decomposition products
are carried by a flow of oxygen to the catalytic core, which is maintained at a temperature of 750 ฐC to ensure
complete thermal decomposition. The oxygen flow continues to carry the gases to the gold amalgamator, where
mercury is selectively trapped. Continuous oxygen flow removes any remaining decomposition products. The
amalgamator is subsequently heated, releasing the mercury vapor to the absorbance cuvettes where peak height
is measured at 253.7 nm as a function of ng of mercury.
Calibration for Method 7473 can be performed in two ways. One method is by the traditional analysis of aqueous
standards. The ability for direct analysis also allows for unique calibration using solid standards with a certified
mercury content. Method 7473 provides the option to perform calibration using solid samples, "An alternative
calibration using standard reference materials may be used..."3 This option is beneficial, especially for on-site
analysis, when transport and storage of aqueous standards may be problematic.
Subsequent to validation of Method 7473 in both laboratory and field settings, the method was applied to the
laboratory analysis of a series of contaminated soils. Duplicate soil samples were sent to a commercial
laboratory for independent analysis. Only those soils with a mercury content less than 10 mg/kg were analyzed
directly due to the extreme sensitivity of the instrument. Soils above 10 mg/kg were leached using EPA Method
3051A prior to analysis. Mercury content in these soils ranged from 1-700 mg/kg. Regardless of the mercury
content, results using Method 7473 agree with results using the traditional cold vapor Method 7471A and show
53
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
greater precision. Average RSDs for Methods 7473 and 7471A were less than 10% and 15%, respectively.
In collaboration with a local engineering firm, Method 7473 has been used on-site to evaluate the remediation
efforts of a large national utility company. A site near Gettysburg, Pennsylvania, was first evaluated. A sample
was taken from each wall and the floor of the 124 ft3 excavated area. It was determined that remediation efforts
were successful, as the mercury content in all samples was below the action level of 20 mg/kg. These results
have been confirmed by independent commercial laboratory analysis.
A second on-site evaluation was performed in Hocking County, Ohio. Real-time results were used to direct the
extent of excavation. The original scope of work based exclusively on the site characterization was an
excavation area of 750 ft3 Use of Method 7473 on-site produced real-time information as to the level of
remediation required and allowed remediatiors to reduce the planned excavation area by more than half to 250
ft3. Real-time Method 7473 results indicated that the reduced excavation was adequate in all areas except one.
Further excavation in that location was thus performed, providing a more accurate remediation and eliminating
the need for a return trip to the site. Approximately $9,000 in savings resulted from the reduced amount of soil
remediated and elimination of a return trip.
SUMMARY
A method for the direct determination of total mercury in both laboratory and field environments has been
established. Method 7473 has been validated by analysis of SRMs as well as by independent analysis. Data
indicates that Method 7473 can achieve lab-quality results in a field setting, use of Method 7473 in the field can
lead to more accurate and cost-effective risk assessment of mercury contaminated sites. The next step will be to
investigate compatibility of Method 7473 with speciation technology.
CtMctor
I ami
I
rO
Hrix*
mnot
CVKtaf
1
X
t~-
^
"^
ป
)
Figure 1. Schematic
of the DMA-80
Table 1. Comparison of techniques use for method validation (n > 5).
EPA Method 7473 7471 A
Technique direct mercury analysis CV-AAS
Average RSD <10% <15<>/0
Agreement? V y
6020
ICP-MS
-15%
54
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
REFERENCES
1. Walter, P.J., Kingston, H.M.; Boylan, H.M.; Han, Y. Presented at the Waste Testing and Quality Assurance
Symposium, Arlington, VA 1997.
2. US EPA Test Methods for Evaluating Solid Waste, Physical/Chemical Methods; SW-846, 2nd ed.; US
Government Printing Office: Washington, DC, 1997.
3. US EPA Test Methods for Evaluating Solid Waste, Physical/Chemical Methods; SW-846, Update IV ed.; US
Government Printing Office: Washington, DC, 1998.
MERCURY IN SOIL SCREENING BY IMMUNOASSAY
Mark L. Bruce and Kathleen L. Richards
Quanterra Inc., 4101 Shuffel Dr. NW, North Canton, OH 44720
brucem@quanterra.com
Lorraine M. Miller
Olin Corporation, 1186 Lower River Road NW, Charleston, TN 37310
lmmiller@corp.olin.com
Abstract
EPA SW-846 Method 7471 (Cold Vapor Atomic Absorption, CVAA) has been traditionally used for mercury
analysis. This method requires several hours of labor and total processing time to prepare soil samples and
analyze them. The sequential nature of the analysis step limits sample throughput and capacity. Also, the
method is difficult to perform at a remediation site. In contrast, immunoassay (IA), as used in proposed Method
4500, prepares and analyzes the samples in parallel and can provide very large sample capacity in a 2-3 hour
processing time. Also, the IA kit is much more field portable.
Quanterra performed an evaluation of the BiMelyze Soil Extraction and Mercury Immunoassay kit from
BioNebraska for the Olin Corporation. Four different site soil samples were analyzed in quadruplicate by CVAA
and IA following the New York State Department of Health (NYSDOH) Alternate Testing Procedure guidelines.
Threshold standards were setup at 1 mg/kg and 15 mg/kg. No false negatives were observed. Thefalse positive
rate was 6%. The overall accuracy rate was 94%.
Switching to immunoassay improves turn-around-time, sample capacity and field portability.
Introduction
Community, environmental and economic concerns are exerting pressure on the environmental market to reduce
analysis turn-around time and cost. At the same time, some data end users have realized that traditional test
methodologies and their QA/QC requirements are not necessary for all environmental decisions. There is a
growing need to provide reliable, quick turn-around field testing in order to expedite site remediation. Accelerated
testing can be instrumental in reducing the impact that remediation activities may have on the local community.
It also facilitates faster site closure, thus reducing the time the excavation site is exposed to the effects of
weathering.
Immunoassay has been widely used in biochemical testing for health services for decades. Several companies
have adapted this technology to test environmental pollutants including organic compounds and metals.
Immunoassay has proven to be a low cost, fast turn-around, high capacity analysis for the health sciences.
These same characteristics make it very attractive for the environmental market. Immunoassay (IA) for metals
does not provide data that is identical to the traditional inductively coupled plasma or atomic absorption tests.
The specificity of IA should make it less susceptible to the interferences that limit spectroscopic analyses.
However, the biochemical nature of IA may make it sensitive to new interferences. The QA/QC data normally
available from IA can include replicates and matrix spikes.
This study compared the results from proposed US EPA SW-846 Method 4500 (Mercury in Soil Sample by
Immunoassay) with EPA SW-846 Method 7471 (Cold Vapor Atomic Absorption, CVAA). Method 4500 was
performed using the BioNebraska BiMelyze Soil Extraction and Mercury Immunoassay kit. Actual field samples
55
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
from a remediation site containing the analyte of interest (mercury) were used for this study.
Immunoassay Method Summary (excerpted from Section 2.2 of Method 4500)
"Solid samples are prepared by extraction with a mixture of hydrochloric and nitric acids for ten minutes and then
buffered prior to analysis. The sample is added to a tube (treated with BSA-glutathione) and incubated at
ambient temperatures for five minutes. The mercuric ions bound to the sulfhydryl groups of the BSA-glutathione
are now reacted with a reconstituted antibody specific for mercury and incubated for five more minutes. A
peroxidase conjugate is added to the sample, reacting with any mercury specific antibody. The substrate is then
added forming a color that is in proportion to the amount of mercury originally present in the sample. The color
produced is then spectrophotometrically compared with the control standards."
Experimental
This project evaluated the BioNebraska BiMelyze Soil Extraction and Mercury Immunoassay kits (BN-IA-Hg),
proposed Method 4500 by comparison to the traditional mercury analysis method (cold vapor atomic absorption,
SW-846 7471). NYSDOH Alternative Testing Procedure guidelines (4/1/86) were followed.
The NYSDOH Alternative Testing Procedure specifies that limited approval for a "new" test method requires the
following analytical work:
1) Three samples from the soil source.
2) Analyze four aliquots of each sample with the approved method (7471).
3) Analyze four aliquots of each sample with the alternative method (4500).
4) Analyte concentrations should range from the detection limit to 20% greater than the regulatory limit.
5) The result from the alternative method must fall within the confidence interval of the approved method (based
on two times the published standard deviation).
An additional sample from the soil source which had a mercury concentration previously determined to be well
below the lowest action threshold was evaluated to test the likelihood of the IA kit producing false positives.
False positives occur when interferences or analytical error produces a test result that is greater than the control
threshold when the analyte concentration is actually below the threshold. A false positive can result in
unnecessary remediation work.
Also, it is common to intentionally bias the results of semi-quantitative field tests slightly high. This reduces the
false negative rate. Many organic immunoassays use a bias of 30%. A 20% bias had been used previously by
others working with the BioNebraska mercury test kit. A 30% bias on the threshold standards was selected for
this study.
The NYSDOH requirements described above were met as follows:
1) Four samples from the soil sample source were selected based on previous CVAA analysis to cover the
appropriate concentration ranges. Approximately 150 g of each sample was frozen and then homogenized
cold to reduce analyte losses. The samples were stored at 4ฐC until appropriately sized aliquots were
removed from the sample containers for each test method. No elemental mercury was visible and the
samples appeared visually to be homogeneous before sub-aliquots were removed.
2) Four aliquots of each of the 4 samples were analyzed by Method 7471.
3) Four aliquots of each of the 4 samples were analyzed by Method 4500 following the procedure described in
the BioNebraska test kit product literature. The two control thresholds were 1 and 15 mg/kg since those
correspond to the two action limits fortheapplicable site.
4) Ideally one sample would have been available for the middle of the following ranges: x < 0.5 ppm, 1 ppm < x
< 5 ppm, 10 ppm < x < 15 ppm, 15 ppm < x < 20 ppm
The quadruplicate CVAA results show that 3 of the 4 ranges were covered by the samples selected.
5) Since the alternative method is semi-quantitative, the range determined for each individual sample was
compared to the corresponding average quantitative result from Method 7471
Results and Discussion
Accuracy
The mercury field test kit is semi-quantitative for this site (e.g. x < 1 mg/kg or 1 mg/kg < x <15 mg/kg or 15
mg/kg < x). The concentration range determined by method 4500 (IA) was compared to the concentration
56
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
measured by 7471 (CVAA). IA results which placed the sample in the correct range were considered "accurate".
Results that were too low were labeled "false negative". Results that were too high were labeled "false positive".
Normal US EPA criteria for field test kits specify the false negative rate should not exceed 5%. NYSDOH criteria
were not available.
Table 1. Comparison of Average Results, CVAA vs IA
Sample ID
B41 -24-1 02297
B1 1-02-1 021 97
B44-02-102197
B1 -24-1 021 97
CVAA
avg
mg/kg
0.12ฑ0.02
9.4 ฑ0.88
35.2 ฑ9.2
16.9ฑ14
IA
avg
abs
0.21
0.43
1.20
0.77
IA interpretation (mg/kg)
x<1 1
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The average IA results agreed with the average CVAA results for each of the four samples.
As expected the individual results showed greater variability and slightly less agreement between the two
analysis types One immunoassay result from B41-24-102297 (x < 1 ppm) was a false positive since its
absorbance equaled that of the 1 ppm threshold standard. Occasional results of this type are to be expected
when using low biased threshold standards to reduce the false negative rate. The IA and CVAA results for
samples B11-02-102197 (1 < x < 15 ppm) and B44-02-102197 (15 ppm < x) showed excellent agreement.
Sample B1-24-102197 had a concentration near the 15 ppm threshold and was apparently more heterogeneous
than the other samples (despite the extensive homogenization that had been performed). Both the IA and CVAA
results were more variable and were evenly split on either side of the 15 ppm threshold. As a general data set
the results from the two methods are in agreement. Please note that since the individual IA and CVAA analyses
were performed on separate aliquots of the same sample, it is not appropriate to directly compare the individual
test results in order. Rather the quadruplicate results from each method must be considered as a set.
Given the results discussed above and shown in the table below, no false negatives (0%) and only one false
positive (6%) were observed in the 16 sample assays. The agreement rate between the two tests was 94%.
The average IA kit absorbances correlated well with the CVAA average results. This indicates it may be possible
to use the IA kit for quantitative analysis, particularly if standards were prepared in or made from soil samples
on-site. The chart below shows IA absorbances plotted
relative to the CVAA results. The IA standard soils
(from BioNebraska) are also shown relative to their
actual prepared concentrations. The least squares
linear equation for the 4 soil samples was :
IA abs = 0.02879 (Hg cone mg/kg) + 0.2089 with an R2
value of 0.9848.
Figure 1. Comparison of IA Absorbance vs CVAA
Concentration
in
5
1 .
0.6
0.4
0.2
0
(
X
X
*
X soil samples
o soil standards
) 10 20 30 40
Hg mg/kg (CVAA)
Precision
The standard deviation of the absorbances produced by the IA test were useful when interpreting results near a
control threshold. It is not practical to calculate the standard deviation of the concentration since this is a
semi-quantitative test in its current configuration.
The percent relative standard deviation (%RSD) of the result (cone for CVAA and absorbance for IA) for each
method was comparable. This indicates the two methods have the same precision for this group of samples.
Most likely sample homogeneity was the factor which limited the precision of both methods.
Table 3. Comparison of Precision, CVAA vs IA
Sample ID
B41 -24-1 02297
B1 1-02-1 021 97
B44-02-102197
B1 -24-1 021 97
CVAA
Std Dev
0.01
0.44
4.6
7.1
%RSD
8.6
4.7
13
42
IA
Std Dev
0.02
0.02
0.23
0.18
%RSD
10
4.9
19
24
average %RSD
17
Information from a BioNebraska representative indicates that the use of volumetric pipettes instead of the "eye
dropper" volume measurements described in the product literature may have improved precision by a few
percent.
Sample Turn-Around Time
Complete kit preparation, extraction and analysis of a batch of soil samples took approximately 3 hours for a
58
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
batch of 19 assays (16 samples). Kit preparation (including reagent prep and sample container labeling) took
about 0.5 hours. Extraction and filtration of the soil samples took 1.75 hours. Performing the immunoassay took
0.75 hours. Two analysts were working together performing the soil filtrations. If a single analyst was performing
all the filtrations with the filters supplied in the kit, add about 0.5 hours to the total time Obviously smaller
batches of samples would reduce the time required for kit preparation, extraction and filtration but the assay step
would be only slightly shorter. Also, autopipettes and repeating pipettes were used (at client request) to improve
accuracy and precision. This equipment (not included in the field kit) also shortened the time necessary for some
of the liquid handling steps in the immunoassay.
Sample Capacity
The field portable version of the test kit is designed to process 13 soil samples in each batch. Smaller batches
can be processed but the cost per sample increases because the test kits are used less efficiently and additional
control soil samples must be purchased. The laboratory version of the test can be configured for much larger
batch sizes. Using laboratory pipettes and previous experience with immunoassay, two field kits were combined
to process all 16 tests (4 samples in quadruplicate) in the same batch. Laboratory kits are available to process up
to 96 assays (-80 samples) in a batch.
Analyst Skill
The field test kit was designed for use by field technicians with average manual dexterity and attention to detail
but limited experience with scientific instrumentation. Based on the experience during this study, this expectation
is true. Appropriate hands-on training with the IA kit on known samples is essential before beginning work on
unknown samples. Also, two problems occurred during sample preparation which might happen for other
samples on this site.
1) B44-02-102197 produced a large amount of foam when the acid was applied to the sample at the being of the
extraction. During the study, the acid was added very slowly and the bottle continuously tapped on the bench top
to break the foam bubbles. Even so, 2 of 4 aliquots lost about 1% of the sample due to foaming out of the
extraction bottle. This did not appear to affect the assay results but required considerable attention, persistence
and time to avoid significant sample loss. An experiment conducted after the assays were complete showed that
adding the recommended amount of acid solution to a 5g sample aliquot resulted in foam overflowing the normal
32 ml_ sample bottle and a 67 mL bottle. When a 140 ml_ bottle was used the foam filled 80% of the bottle before
bursting and settling into the bottom of the bottle. Thus, for samples similar to this one it would be advisable to
use a 140 mL bottle to avoid accidental sample loss, slow processing and increasing the skill requirements.
2) The extract filtration step was much slower for the samples collected at the site than for the standard soils
supplied by BioNebraska. Sample B41-24-102297 in particular was very difficult to filter. Several samples
required squeezing the sample bottle with pliers in order to force sufficient extract filter. This would make
reproducible results difficult using "eye dropper" volume measurements at this point in the process. Other types
of IA kits have more "user-friendly" filtration processes. It may also be possible to adapt the current filtration
process to improve its performance.
Conclusion
Method 4500 (IA) produced semi quantitative results which matched the Method 7471 (CVAA) results for 15 of
the 16 tests (94%) performed. There was one false positive where the immunoassay overestimated the mercury
concentration. The false negative rate in this four sample evaluation study was 0%, while the false positive rate
was 6%. This meets the requirements of the NY DEC Alternative Testing Procedure. In addition, the average IA
absorbances from each quadruplicate set of data correlated very well with the CVAA results. This indicates that
the IA kit may also be useful for quantitative analyses in the future.
Summary of Reporting Limits and Performance Results
Matrix
soil
Mercury
1 mg/kg
Accuracy
94%
False Negative
0%
False Positive
6%
Acknowledgments
The authors would like to thank Veronica Bortot from Quanterra - Pittsburgh for overall project coordination and
Craig Schweitzer from BioNebraska for IA product and technical support.
59
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
References
1. BioNebraska - BiMelyzeฎ Soil Extraction and Mercury Assay Product Literature
2. US EPA SW-846 Method 4500 (proposed Update IVA)
3. DOE Method MB 100 Rev 1 (draft)
4. California EPA Evaluation Report Certificate No. 95-01-014
UTILIZATION OF A FIELD METHOD FOR THE SEMIQUANTITATIVE DETECTION OF SILVER IN
ENVIRONMENTAL SAMPLES IN THE 0 - 50 ppb RANGE
Dan Kroll
Hach, 100 Dayton Road, Ames, Iowa 50010
Abstract
Silver is commonly used in industry, and its bactericidal properties have also lead to its use for water disinfection
purposes. Excess concentrations of silver may damage human health and be toxic to aquatic life. Current
methods of silver analysis, in the ppb range, require expensive equipment and careful technique. There is a need
for a quick, easy screening method for silver at these levels. The procedure described employs the
bromopyrogallol red/1,10-phenanthroline method, described by Dagnall and West (1964), combined with a novel
concentration/detection method. At a pH of 7, a ternary complex is formed with two 1,10-phenanthroline
molecules binding to each silver ion, and then two of these complexes bind to a bromopyrogallol red molecule.
This results in a blue precipitate. The colored precipitate is caught on a 13 urn pore size filter, and the filter is
compared to a precalibrated (0, 5, 10, 25, and 50 ppb) printed color chart for quantification. All the reagents are
combined in a single powder that contains both dyes, a buffer, and a masking reagent. The system is easy to
use, fast, portable, and all reagents are stable for at least one year making the system ideal for field testing. This
method has been evaluated on a variety of tap water, pool & spa water, river water, and sewage effluent samples
that have been spiked with known amounts of silver. Some of the river water samples and the sewage effluent
required a sulfuric acid digestion, but all samples resulted in good recovery of the spikes and correlated well with
numbers generated using AA techniques. Various soil samples that were spiked with 25 ppb Ag resulted in good
recoveries that corresponded to the appropriate color spot on the chart. Use of this method as a screening
process may help to save time and money by cutting down on the need to do more accurate analysis of all
samples.
Introduction
Silver is a common contaminant of industrial process and wastewater. In private industry, silver is used in
applications such as jewelry, coins, dentalware, silverware, solder, electroplating, photography, and battery
production. In low concentrations, silver's antibiotic properties make it desirable for use as a fungicide and for
drinking water disinfection purposes, and it has been gaining in popularity as a pool and spa biocide. However,
according to the World Health Organization (WHO), continuous exposure to silver in drinking water (0.4 mg or
more) in humans causes arygaria, an irreversible condition which produces a bluish-gray discoloration of the
skin, hair, nails and eyes.4 Long term continuous exposure to silver has also been implicated in liver damage and
enzyme inactivation in humans."
Unpolluted surface water levels of silver usually range between 0.1 4 ug/L. Drinking water levels range between
0 - 2 ug/L; average = 0.13 ug/L.3 The WHO has not as yet set limits for safe silver concentrations in drinking
water.5 The USEPA has adopted the Public Health Service (PHS) standard that silver in domestic water not
exceed 50 ug/L.3 The USEPA-adopted PHS standard was set to protect aquatic life and human health. Canada
has adopted a similar 50 ug/L standard while the EEC standard is 10 ug/L.5 Silver is also on the list of seven
priority pollutant metals that must be monitored in landfill leachate.7
There are many current methods to measure silver in the ppb ranges. These include atomic adsorption by flame
or electrothermal techniques, inductively coupled plasma, or colorimetry. Each of these methods requires
complicated and expensive apparatus, hazardous chemicals and/or a large investment in time and equipment.
Each also has its drawbacks. AA is accurate at moderate concentrations, but displays sensitivity to ion
interference. ICP techniques have higher minimum detection limits and are sensitive to refractory elements
60
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Colorimetry loses sensitivity at these ranges and uses hazardous chemicals.6
There is a need for a portable, rapid, easy, and safe screening method for low levels of silver. The method
described in this article meets all of these requirements and requires less time than conventional methods. The
user obtains an on-site reading that only takes minutes to obtain. The procedure features a simple visual
comparison for obtaining semiquanitative measurements for silver between 0-50 (jg/L. Single-reagent addition
and quick results make this method ideal for field testing.
Method and Chemistry
This test makes use of the fact that under certain specified conditions, metals form brightly colored precipitates
with certain dyes or mixtures of dyes. In the case of saver, the complex formation is with 1-10 Phenanthroline
and Bromopyrogallol Red. This chemistry is an adaptation of a reaction described by Dagnall and West.1 In this
reaction each Ag ion first reacts with two 1-10 Phenanthroline molecules to form a colorless complex. Two
molecules of this complex then react with a molecule of the Bromopyrogallol Red to form a blue precipitate.
Dagnall and West made use of this system for the spectrophotometric detection of silver with a lower detection
limit of 20 ppb. The system that I am currently using follows this chemistry until the method of detection. I make
use of the fact that the blue complex is an insoluble precipitate in an aqueous sample.
A reagent powder is prepared that contains a Sodium Citrate - Citric Acid buffer system with a pH of 7. The
powder also contains Tetrasodium EDTA as a masking reagent to remove interference from other metals. An
excess of 1-10 Phenanthroline is added to make sure that there is more than enough present to complex the
silver along with any iron that may be present. If excess 1-10 Phenanthroline is not added, it may all bind to
contaminant iron leaving none available for the silver. Finally, the reagent powder contains the Bromopyrogallol
Red. The final reagent powder is then packaged at a weight fill of 2.0 g in unit dose form. This is the amount
needed to react with a 100 ml sample.
After the initial chemical reaction is carried out, 100 ml_ of the reacted sample is filtered by being forced with a
syringe through a nitrocellulose microporous filter. (Schleicher & Schuell) The blue precipitate is trapped on the
filter. The intensity and hue of the filter is dependent in a quantitative manner upon the original concentration of
silver present in the sample. The colored filter is then compared to a color matching chart with different shades
of blue corresponding to different concentrations of silver. By manipulation of the dye concentrations, filter size,
pore size, and sample volume, the levels to which the test can be made effective can be altered. Using a 100 mL
sample, 5 mm dia. filter size (a Gelman #4317 13mm plastic filter holder is modified with washers to expose a 5
nun surface area on the filter), and 12 urn pore size, visual levels of detection down to 5 ppb Ag can be
achieved. The final color coding chart for silver has gradations
of 0, 5, 10, 25, and 50 ppb Ag (See Fig. 1). This method has
been validated on a number of different water matrices using
NIST standard spikes (SRM3151) and compared to AA (Varian
SPECTRAA 20-plus) for verification. This method allows visual
detection of silver in ranges that are comparable to, and in
some cases lower than, the levels of detection that are possible
with more costly and time consuming methods, such as AA.
Figure 1. Color Comparator Chart pp
Hach Rapid Silver Test Chart
0 - 50 ppb Ag
Summary of Method
Silver ions at a pH of ~7 combine with two molecules of 1,10-Phenanthroline to form a colorless water soluble
complex. Two molecules of this complex then combine with a molecule of Bromopyrogallol Red to form a water
insoluble blue precipitate. (See Fig. 2) This blue precipitate is then captured on a nitrocellulose filter. The filter is
then compared to a color chart and matched to the appropriate spot for a semi-quantitative reading. The reagent
pillow contains appropriate buffer, indicator, and EDTA as a masking reagent. A pretreatment of ascorbic acid is
necessary to remove CI2 in excess of 2 mg/L. Digestion is needed for some samples. Other samples such as
very concentrated soil digests that contain large quantities of interfering metals may require the addition of extra
EDTA and or sodium citrate as chelating reagents.
Sampling and Storage
Collect samples in an acid-cleaned glass or plastic container. Adjust the pH to 2 or less with Nitric Acid (about 2
61
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
rx^ . ,ซ,*ซ,, .AA.AA-
OT sllvar' Jc>
^^ no, Ehenanthroline HlT
Brcmopyjcogallol ^5^
Bluป Precipitate
mL per liter). Store preserved samples at
room temperature for up to six months.
Adjust the pH to ~7 with Ammonium
Hydroxide before analysis. Do not use a
pH meter as silver contamination from
the electrode may occur. Phenol Red
may be used as an indicator. Correct the
test results for volume additions.
Figure 2. Reaction Mechanism
Interferences
Interference studies were conducted by preparing a known silver solution (approximately 0.010 mg/L) and the
potential interfering ion. Positive interference was tested by running blanks of deionized water that contained the
potential interfering ion. The ion was said to interfere when the resulting change threw off the color match by
more than 1/2 step on the chart. The following substances, at the stated levels, show no interference on a 10 ppb
Ag standard test. There should be no problem with color matching or bad blanks at these levels. These tests
were run at the levels indicated, although greater concentrations may be tolerated.
Table of Interference Study Results
Aluminum Al
Antimony Sb
Bismuth Bi
Borate
Na2B4O7
Barium Ba
Chlorine CI2
Chloride Cl"
Chlorite
HOCI
Calcium Ca+2
Chromium Cr+3
Chromium Cr+6
Cadmium Cd+2
Cobalt Co
Copper Cu+2
Flouride F~
Gold Au
Iron Fe+2
Iron Fe+3
Lead Pb
Magnesium
10 ppm
10 ppm
10 ppm
1 g/Lor215
ppm as B
10 ppm
5 ppm
400 ppm
1 Ppm
1000 ppm
10 ppm
10 ppm
10 ppm
10 ppm
10 ppm
10 ppm
10 ppb
10 ppm
10 ppm
10 ppm
1000 ppm
none
none
none
none
(fades with time)
none
Dingy but
readable > OK
with ascorbic
acid
none > amounts
interfere
none
none
none
none
none
none
none
none
none > negative
interference
none
none (lite pink
tint) >with
addition extra
none
none
Manganese Mn
Mercury Hg
Molybdenum Mo
Nickel Ni
Nitrate NO3 as N
Nitrite NO2 as N
Ammonia
NH3+ as N
Palladium Pd
Phosphate PO4
asP
Potassium K
Platinum Pt
Sulfate SO4
Selenium Se
Silica SiO2
Thallium Tl
Tin Sn
Titanium Ti
Phenanthroline
Zinc Zn
8lSteftfeB*T\"jซ^llfeB
10 ppm
10 ppm
10 ppm
10 ppm
250 ppm
100 ppm
1 000 ppm
10 ppm
1000 ppm
1 000 ppm
50 ppb
1 000 ppm
10 ppm
1000 ppm
1 ppm
10 ppm.
10 ppm
1 ppm
10 ppm
none
none
none
none
none > amounts
appear to fade color
none > amounts
appear to fade color
none
none
none
none
none > negative
interference
none
none
none
none
none
none
none
none
Turbid samples should be pre-filtered through a glass fiber filter. Oils and surfactants may interfere by preventing
62
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
formation of the insoluble complex; however, appropriate digestion may eradicate some of these interferences.
Materials that were not tested, that according to the literature do interfere, are Uranium (VI), Thorium (IV), and
Niobium (V). They all form blue colored complexes with Bromopyrogallol Red. They can be masked at a 10-fold
excess over silver by the addition of excess fluoride for Uranium and Thorium, and of hydrogen peroxide for
Niobium.1
Precision
In a single laboratory, using a standard solution of 0.010 mg/L silver and three representative lots of reagents, a
single operator performing 50 tests per lot obtained no results that were not properly matched to the 10 ppb color
dot within 1/2 step. In other words all results were within a range of 7.5 to 17.5 ppb. These results were compared
to results obtained on a AA (Varian SPECTRAA 20-plus). The results for the AA on 30 repetitions using the same
standard gave an average of 6 ppb with a standard deviation of 4.98 ppb. The recommended lower level of
detection for the AA was 20 ppb.
Assuming the worst case scenario where the standard deviation for the visual test method is 7.5 ppb a z test was
performed to compare the mean results from the two methods. This resulted in a calculated z value of 2.86. This
value for the z statistic indicates that the means are not the same with over 99% confidence. In this case, where
the amounts of silver to be detected are below the recommended level of detection for the AA, the visual method
out performs the AA.
Performance on Environmental Samples
Samples from a hot tub utilizing bromine as a disinfectant, Ames, Iowa tap water and pool water from Carr pool,
also in Ames, were spiked at a level of 10 ppb silver. Tests run in the field using the visual method resulted in
100% recovery on these spiked samples. All samples matched the 10 ppb spot.
Samples run on sewage effluent from the Ames, Iowa sewer plant and surface water from the Skunk River near
Ames resulted in poor or no recovery. This is probably due to the binding of the silver ions by humic or fulvic
acids present in the sample, or possibly because of reduction of the silver ions to silver metal upon addition.
100% recovery on spiked 25 ppb silver samples was found after treatment of the samples with a simple sulfuric
acid - hydrogen peroxide digestion procedure (Hach Didesdahl Procedure)2. Strongly reducing samples,
samples with high organic content, and samples which contain thiosulfate or cyanide should be digested before
testing.
Summary of Results
\ *f ?" งpyFce of Sample
i f' t
Tap Water
Pool Water
Spa Water
Sewage Efluent
River Water
Crowley Silt Loam
Coland Clay Loam
Clarion Loam
NIST SRM 271 1
Aeitow;1.'
10
10
10
25
25
30-32
25
25
25
1 tJbserve^ Cone, "^ IJ
A9.(fl8*)
10
10
10
25
25
35
25
25
25
Tests were run on digested samples of a Crowley Silt Loam soil obtained from the Louisiana State University
Extension Service. 0.5g of the sample was digested using the Hach Digesdahl procedure. The digests were
then filtered through a glass fiber filter and diluted to a final volume of 1 liter. These tests resulted in no recovery.
After the addition of 1.5 g of extra disodium citrate as a complexing reagent to remove interfering ions, the
samples were found to give a blank of *8 ppb for the visual test and 4.66 with a standard deviation of 5.07 ppb
on the AA. 0.5g samples of the soil were then spiked with 25 ug of silver and the procedure was repeated. All of
30 samples measured by the visual test read between the 25 and 50 ppb dots for an extrapolated average of ซ35
ppb. The same samples ran on the AA gave an average result of 36.66 ppb with a standard deviation of 4.34
ppb. Assuming the worst case scenario where the standard deviation for the visual test method is 12.5 ppb a z
test was performed to compare the mean results from the two methods. This resulted in a small z statistic of
-0.69. This value for the z statistic indicates that the means for this sample using the visual and AA methods are
63
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
statistically the same. The initial silver content of the soil explains the differences in the spiked and recovered
amounts. Tests performed on a Coland Clay Loam and a Clarion Loam obtained in the Ames area gave similar
results, but with no detectable blanks.
Finally a 0.5407 g sample of Montana soil NIST SRM 2711 was digested using the Digesdahl procedure. This
amount of test soil results in a final solution, when diluted to 100 mL, with a reported assay value of 25 ppb
silver. All of the samples tested correlated well with the reported values. All samples tested matched the 25 ppb
dot using the visual method.
Stability
Reagents kept at 35ฐC for 1 year still function properly.
Summary
The method described was utilized to test various soil and water samples for the presence of Ag. All of the
matrices that were examined gave acceptable results. The results were comparable to those obtained with an
AA, and in the case of levels of 10 ppb and below, the visual method was found to be superior to the AA. This
method should be useful as a field method for determining trace amounts of silver in a variety of environmental
samples.
The use of this method should save time and money without compromising the accuracy of analysis.
References
1. Dagnall, R.M, and West, T.S. 1964. A selective and sensitive colour reaction for silver. Talanta, 1964, vol 11.
pp. 1533-1541.
2. Digestion and Analysis of Wastewater, Solids, and Sludges. 1987, Hach Company. Loveland, Colorado.
3. Guidelines for Drinking Water Quality, vol. 1: Recommendations, World Health Organization, 1984.
4. Guidefines for Drinking Water Quality, vol. 2: Health Criteria and Supporting Info, 1984.
5. Hach Water Analysis Handbook, 2nd edition, 1992, Hach Company. Loveland, Colorado.
6. Standard Methods for Examination of Water and Wastewater, 18th edition, 1992.
7. 40 CM 7-1-96 edition. Section 261.24.
DIAGNOSING ERRORS IN SPECIES ANALYSIS PROCEDURES USING SIDMS METHOD 6800
H.M. "Skip" Kingston, Dengwei Huo, Yusheng Lu
Duquesne University, Department of Chemistry and Biochemistry Pittsburgh, PA 15282-1503
412-396-5564
The determination of chemical species in environmental samples is difficult when the species are easily altered
during the analysis process. A method has been developed to permit the species to undergo chemical reactions
during analysis and correct for these changes. Speciated Isotope Dilution Mass Spectrometry (SIDMS) is
especially suited for species that equilibrate in solution quickly and also degrade during extraction, oxidize or
reduce during analysis or are difficult to separate quantitatively using conventional methods1 This method is a
new draft EPA RCRA method (Method 6800) that utilizes isotopically enriched speciated spikes combined with
isotope dilution to accurately determine and correct for species transformations that occur in sample processing.
The errors in the measurement are those that are limited by the ability of the ratio measurement and the
equilibrium of the species during spiking. This method was specifically developed to address the problems of
accurately quantifying different species in complicated matrices that also play a role in the stability of the species
during extraction, separation or manipulation. Additionally, it is a diagnostic tool for identifying both the error and
bias inherent in specific method steps prior to and during sample analysis such as sampling process, storage,
sample preparation, and chemical modifications prior and during measurement. The basic SIDMS method is
applicable to many species of elements with multiple isotopes and extends to various oxidation states,
organometallics, and molecular forms of species. The method is used as a diagnostic tool and reference method
assisting with error identification in other methods permitting their development as more accurate and precise
species analysis tools.
64
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Validation data for Cr(VI) and Cr(lll) demonstrate the ability of SIDMS to examine other methods in a diagnostic
manner. The extraction procedure Method 3060A and analysis Method 7196A are used to demonstrate the
identification of specific chemical changes that take place in these methods. These changes can be corrected by
use of this very sensitive internal speciated tracer2.
The objective of EPA Method 6800 is to provide a new reference method that is also legally defensible as a
reference method for measurements that have high degrees of uncertainty and error due to highly reactive
species.
1. Kingston, H.M. Skip, Dengwei Huo, Yusheng Lu, Stuart Chalk, "Accuracy in species analysis: Speciated
Isotope Dilution Mass Spectrometry (SIDMS) exemplified by the evaluation of chromium species"
Spectrachemica Acta B, (accepted) 1998.
2. Yusheng Lu, Dengwei Huo and H.M. Skip Kingston. "Determination of Analytical Biases and Chemical
Mechanisms in the Analysis of Cr(VI) Using EPA Protocols" ES&T (submitted) 1998.
THE USE OF 210Pb DATING AND DETAILED STRATIGRAPHY TO DETERMINE THE SIGNIFICANCE
AND FATE OF CHROMIUM IN SEDIMENTS NEAR A HAZARDOUS WASTE SITE
Richard Rediske
The Robert B. Annis Water Resources Institute, Grand Valley State University, Allendale, Ml 49401
616-895-3047
Gary Fahnenstiel
National Oceanic and Atmospheric Administration, Great Lakes Environmental Research Laboratory,
Lake Michigan Field Station, 1431 Beach St., Muskegon, Ml 49441
Claire Schelske
Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, FL
Marc Tuchman
USEPA Great Lakes National Program Office, 77 West Jackson Blvd., Chicago, IL 60604
ABSTRACT
An innovative investigation using 210Pb dating and detailed stratigraphy was conducted to determine the
significance and fate of chromium in the sediments of White Lake (Michigan). Elevated sediment concentrations
of chromium, arsenic, and mercury were found in the vicinity of the historical effluent discharge point of a
tannery. The chromium levels found in the sediments were among the highest concentrations reported in the
Great Lakes basin (20,000 mg/kg). Since the direct discharge of effluent from the tannery was discontinued in
1976, vertical depositional patterns may reflect changes in the flux of chromium into the system. Historical levels
of metals may be covered by less contaminated material or resuspended by physical events. Information on
sediment stability and deposition rates was critical to the development of remediation options for the site.
Traditional sampling and analytical methods would not provide information on sediment stability and
accumulation patterns. Radiodating using 210Pb, a technique commonly used in linmology, was employed to
determine the history of sediment deposition. This technique was augmented with detailed stratigraphy analysis
to provide a current and historical record of chromium deposition in the sediments. Two piston core samples
were collected in the tannery discharge area and sectioned in 2 cm intervals. Total chromium was analyzed by
ICP. Radiometric measurements were made using a low-background gamma counting system with a well-type
intrinsic germanium detector. Total 210Pb activity was obtained from the 46.5 kev photon peak, and 226Ra activity
was obtained from the 609.2 kev peak of 214Bi. The 661.7 kev photon peak was used to measure 137Cs activity.
The peak in 137Cs activity was measured to evaluate its usefulness as an independent time marker for the peak
period of fallout from nuclear weapons testing in 1962-63.
Chromium stratigraphy in the tannery discharge area indicated that the top 15-20 cut of sediment was less
contaminated (2,000-4,000 mg/kg) than sediment located at >30 cm (>5,000 mg/kg). Radionuclide results
suggested that this surface sediment layer was well mixed, however, distinct from the deeper more highly
65
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
contaminated sediments. Presently this surface sediment layer (15-20 cm) does not physically mix with the
deeper, more contaminated sediment. The surface layer was followed by a region (30-80 cm) that contained
chromium levels in excess of 20,000 mg/kg. Since the direct discharge of tannery effluent to this area ceased in
1976, evidence of the deposition of sediment with less chromium contamination should have been apparent. The
lack of a decreasing gradient of chromium concentration in the near surface zone sediments suggested that the
processes of mixing and resuspension continue to be active. In addition, chromium transport to the 0-20 cm
sediment zone may also have occurred by other mechanisms including surface runoff of contaminated soils and
possibly groundwater advection. The lack of a significant 137Cs horizon in the sediments indicated that
groundwater was discharging in this region; however, the linkage with chromium mobility required further
investigation.
INTRODUCTION
White Lake is a 2,571 acre, drowned-rivermouth lake located on the eastern shore of Lake Michigan in
Muskegon County. The Lake is part of the White River Watershed and discharges directly to Lake Michigan
through a channel located on the western end. White Lake was designated an Area of Concern (AOC) in 1985 by
the International Joint Commission because of historical discharges of heavy metals and organic chemicals.
Chromium, mercury, arsenic, and animal hides have been discharged into White Lake by Whitehall Leather. The
tannery began operating in Whitehall near the turn of the century and used wood bark as the original tanning
agent. In 1940, the tanning agent was changed to chromic sulfate, and a series of six waste treatment lagoons
were constructed near an area of the shoreline called Tannery Bay. Effluent from these lagoons containing heavy
metals and leather byproducts was discharged directly into the bay. In addition, dredged materials from the
lagoons and other process wastes were disposed of in landfill areas adjacent to the shore. The direct discharge
of wastewater effluent from the tannery ceased in 1976. Previous investigations have indicated extensive
contamination of sediments in White Lake. Elevated levels of chromium, lead, arsenic, and mercury were
detected in the northeastern section of the lake in 1982 (WMSRDC 1982) and in 1994 (Bolattino and Fox 1995).
This area was the historical discharge point for tannery effluent from Whitehall Leather. The chromium levels
found in the sediments of this area were some of the highest reported from any site in the Great Lakes. Since the
direct discharge of effluent to Tannery Bay was discontinued in 1976, vertical depositional patterns may reflect
changes in the flux of chromium into the system. The stability of the sediments in this region was also unknown.
Without more information on sediment stability and accumulation rates, it would difficult to determine the
residence time of contaminants within any specific region of the sediments. Whether historical levels of metals
are being covered by less contaminated material or being resuspended by physical events are critical questions
that need to be answered before evaluating remediation options.
An innovative investigation using 210Pb dating and detailed stratigraphy was conducted to determine the
significance and fate of chromium in the sediments of White Lake. Radiodating using 210Pb provides a
continuous sequence of dates from a single core utilizing the natural decay of 210Pb. This technique has been
widely used in limnology and has been independently verified by comparisons with other techniques (e.g.,
Robbins et al. 1978; Appleby et al. 1979; and Wolfe et al. 1994). 210Pb is a naturally occurring radioisotope that
enters lakes through wet and dry deposition following the decay of atmospheric 222Rn. Once in the take, 210Pb is
rapidly scavenged by particles and settles to the bottom. The concentration of 210Pb can then be analyzed at a
series of depths in the cores from the surface to the depth where excess 210Pb is no longer measurable,
approximately 5-8 half-lives or 150 years. From this 210Pb profile, dates and sediment accumulation rates are
calculated using one of several mathematical models, such as the constant rate of supply method. Using a
combination of 210Pb dating and detailed metal stratigraphy, critical information related to contaminant profiles
and sediment stability was obtained. The technique described can be used at hazardous waste sites where the
evaluation of contaminated sediments is required for remediation.
EXPERIMENTAL
The sediment cores were collected with a VibraCore device with core lengths ranging from 6 8 ft. The core
samples were then sectioned in three equal lengths for chemical analysis. Ponar samples were also collected at
these locations to provide an assessment of the near surface zone sediments. A piston corer (Fisher et al. 1992)
was used to obtain the samples for stratigraphy and radiodating since the VibraCore causes some degree of
internal mixing in the core tube. VibraCore samples were collected during 1994. The ponar and piston core
samples were collected during 1996. Sampling locations are shown in Figures I and 2. All samples were collected
using the USEPA R.V. Mudpuppy.
Piston core samples were extruded and cut into 2 cm intervals. Each interval was weighed and an aliquot was
66
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
removed for metals analysis. Sample preparation and analysis methods (EPA 1994) for metals are listed below:
Arsenic
Chromium
Mercury
Graphite Furnace
Inductively Coupled Plasma
Cold vapor
7060/3050
6010/3050
7471
All metals results were reported on a dry weight basis.
The preparation for radiometric analysis consisted of freeze drying the sediment and the grinding the sample to a
homogenous mixture. Sub samples were then packed and sealed with an epoxy resin in polypropylene tubes in
preparation for radiometric analysis.
Radiometric measurements were made using low-background gamma counting systems with well-type intrinsic
germanium detectors (Schelske et al. 1994). To prepare samples for radiometric analysis, dry sediment from
each section was packed to a nominal height of 30 min in a tared polypropylene tube (84 mm high x 14.5 mm
outside diameter, 12 mm. inside diameter). Sample height was recorded and tubes were weighed to obtain
sample mass. Samples in the tubes were sealed with a layer of epoxy resin and polyamine hardener, capped,
and stored before counting to ensure equilibrium between 226Ra and 214Bi. Activities for each radionuclide were
calculated using empirically derived factors of variation in counting efficiency with sample mass and height
(Schelske et al. 1994). Total 210Pb activity was obtained from the 46.5 kev photon peak, and 226Ra activity was
obtained from the 609.2 kev peak of 214Bi . 226Ra activity was assumed to represent supported 210Pb activity.
Excess 210Pb activity was determined from the difference between total and supported 210Pb activity and then
corrected for decay from the coring date. The 661.7 kev photon peak was used to measure 137Cs activity. The
peak in 137Cs activity was measured to evaluate its usefulness as an independent time marker for the peak period
of fallout from nuclear weapons testing in 1962-63.
Sediments were aged using measurements of the activity of naturally occurring radioisotopes in sediment
samples. The method is based on determining the activity of total 210Pb (22.3 yr half-life), a decay product of
226Ra (half-life 1,622 yr) in the 238U decay series. Total 210Pb represents the sum of excess 210Pb and supported 2
10Pb activity in sediments. The ultimate source of excess 210Pb is the outgassing of chemically inert 222Rn (3.83
d half-life) from continents as 226Ra incorporated in soils and rocks decays. In the atmosphere, 222Rn decays to
210Pb, which is deposited at the earth's surface with atmospheric washout as unsupported or excess 210Pb.
Supported 210Pb in lake sediments is produced by the decay of 226Ra that is deposited as one fraction of erosional
inputs. In the sediments, gaseous 222Rn produced from 226Ra is trapped and decays to 210Pb. By definition,
supported 210Pb is in secular equilibrium with sedimentary 226Ra and is equal to total 210Pb activity at depths
where excess 210Pb activity is not measurable due to decay. Because the decay of excess 210Pb activity in
sediments provides the basis for estimating sediment ages, it is necessary to make estimates of total and
supported 210Pb, activities so
excess 210Pb activity can be deter-
mined by difference. Excess 210Pb
activity was calculated either by
subtracting 226Ra activity from total
210Pb activity at each depth or by
subtracting an estimate of sup-
ported 210Pb activity based on
measurements of total 210Pb
activity at depths where excess
210Pb activity is negligible.
Cr ISM m(/kg
Aft 9.0 mg/kg
He O.TIm{/kg
Cr SIS n|/kf
AJ 8,6
Hi 0.39 mg/k>
Cr 1650 mi/kg
Al 1 Dug/lie
Hg 1.04mj/ig
Figure 1. Conentration of
chromium, arsenic, and mercury in
ponar samples collected from
Tannery Bay, Whitehall, Michigan
(1996)
Sediment ages were calculated using a CRS model (Appleby and Oldfield 1983). This model calculates ages
based on the assumption that the flux of excess 210Pb, to the lake was constant and therefore that variation in
210Pb activity from a pattern of exponential decrease with depth depends on variation in rate of sedimentation.
67
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Errors in age and mass sedimentation rate were propagated using first-order approximations and calculated
according to Binford (1990).
RESULTS AND DISCUSSION
The results of ponar and VibraCore samples are shown in Figures 1 and
Depth Cr
0--24" <360mg/kc
1610 mg/kg
W-56" 88 me/kg
WL9417. WL9411
Depth Cr
0"-15" 6700 mg/kg
150-35" 2420 mg/kg
35"-4T 30 rag/kg
Cr
8100 mg/kg
7SIOmg/kg
390 mg/kg
Depth Cr
fl"-lซ" 9520 mg/kg
W-35" 2470 mg/kg
35"-49" 179 mg/kg
Depth Cr
>. (T-15- 5540 m(/kg
JS"-23" 14300 mg/kg
23"-40" 5620mg/ki
2 respectively. Ponar samples provide
an indication of the conditions present
in the near surface zone of the
sediments. The penetration of the
ponar is variable and can range from
0-15 cm. depending on the condition
of the sediment. A comparison of the
results from both collection methods
suggests that the highest degree of
the chromium contamination is
present at depths below the pene-
tration of the ponar. In addition, the
ponar results suggest that chromium
continues to enter the sediments of
Tannery Bay since concentrations
range from 1,000 mg/kg to 4,600
mg/kg near the surface.
Figure 2. Concentration of chromium
in core samples collected from
Tannery Bay, Whitehall, Michigan
(1994)
The results of the stratigraphy analyses for total chromium are given in Figures 3 and 4 for I-5M and I-7M
respectively. The I-5M core shows a relatively uniform region of chromium concentrations ranging from 2,500
mg/kg to 3,600 mg/kg between 0 and 26 cm. This region is followed by more concentrated strata that vary from
approximately 5,000 mg/kg to 23,000 mg/kg in the interval from 26-84 cm. Chromium in the remainder of the
core decreases after 84 cm. Since this station was located in the discharge area of the waste treatment lagoons,
Chromium (mg/kg)
5000
10000
15000
20000
25000
Figure 3. Results of chromium stratigraphy analysis on the core sample from I-5M
the variations in chromium concentrations observed reflect differences in effluent composition over time. Sudden
68
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
reductions in chromium levels also correspond to strata that contain animal hair and hide fragments. The I-7M
core follows a different depositional pattern. Concentrations of chromium gradually increase from approximately
2,000 mg/kg to 5,000 mg/kg over the interval from 0-36 cm. Concentrations then rapidly rise and remain
elevated in the region from 38-128 cm. Chromium concentrations began to decrease after 128 cm. Higher levels
of chromium were found in the I-7M core than at I-5M. The highest level found at I-5M was 22,800 mg/kg while
several 2 cm strata at I-7M ranged from 34,000 mg/kg to 61,100 mg/kg.
Chromium (mg/kg)
5000 10000 11000 JOOOO 35000 30000 31090 40000 46600
Figure 4. Results of chromium stratigraphy analysis on the core sample from I-7M
The radiochemistry data is summarized in Tables 1 and 2. The profiles of 210Pb activity for Station I-5M and
Station I-7M (Figure 5) provide other information about historical sedimentation at the Tannery Bay sites. First,
210Pb activity generally decreases with depth. Second, several stratigraphic layers can be identified based on
210Pb activity. Four layers are present in core I-5M: 0-15 cm, 15-30 cm, 30-50 cm, and 50-65 cm; and four layers
can also be identified in I-7M: 0-20 cm, 20-35 cm, 35-45 cm, and 45-70 cm. The total 210Pb activity in the top
layer in both cores was similar, ranging from 10-12 dpm/g, and the activity in the lowest layer was also similar in
both cores. Supported 210Pb is by far the largest fraction in the lowest layer. Finally, because excess 210Pb is
generally not measurable in sediments with ages older than five or six half lives, we can conclude that the ages
in sediments above the bottom layer with measurable levels of excess 210Pb activity are probably not older than
110 to 130 years.
Table 1. Results of radiochemistry analysis of the core sample from I-5M
Depth
(cm)
5
10
15
20
25
30
35
40
45
50
55
60
Total
Pb-210
Activity
(dpm/g)
11.951
11.552
10.70
8.55
8.81
7.455
4.931
2.641
2.696
2.889
1.243
0.44
Ra-226
Activity
(dpm/g)
1.735
1.592
1.73
1.276
1.015
1.423
2.967
1.753
1.398
1.224
0.703
0.757
Cs-137
Activity
(dpm/g)
1.565
1.437
1.614
1.449
1.598
2.414
3.769
3.142
2.468
1.139
0.437
0.024
Excess
Pb-210
Activity
(dpm/g)
10.358
10.103
9.1
7.38
7.911
6.123
1.996
0.904
1.32
1.689
0.549
-0.322
Age
(years)
3.965
10.534
17.992
29.222
46.450
68.841
81.028
90.010
104.612
149.286
Age
Error
(1s)
1.222
1.339
1.490
1.848
2.604
4.523
5.603
5.747
6.991
21.588
Date
1992.7
1986.2
1978.7
1967.5
1950.2
1927.9
1915.7
1906.7
1892.1
1847.4
Mass
Sedimentation
Rate
(mg/cm2/yr)
167.23
145.68
130.06
120.21
72.50
50.96
89.93
142.5
67.94
22.63
MSR
Error
ds)
8.36
7.68
8.22
7.18
5.86
5.95
22.61
61.92
22.25
9.26
69
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Combined plots of chromium stratigraphy and radiochemistry data are shown in Figure 6 for I-5M and I-7M. The
four regions in each core described earlier suggest distinct layers. Sediments that are well mixed would have
relatively uniform 210Pb activity as illustrated from the surface to the 10-20 cm zone. These results are significant
as the 210Pb profile demonstrates a mixed zone near the surface that is isolated from the sediments below
approximately 20 cm. Levels of chromium in excess of 20,000 mg/kg begin at 40 cm at I-7M and at 30 cm at
I-5M. Based on the 210Pb profile, a region of unmixed sediment lies between the heavily contaminated strata and
the mixed sediment zone. The zones of greatest chromium contamination therefore appear to be isolated from
the surface sediments that are subject to mixing. Sedimentation rate data for both Table 1. Results of
radiochemistry analysis of the core sample from I-5M stations suggest that I-7M has a greater rate (225
mg/cm2/yr) than I-5M (167 mg/cm2/yr). This observation is supported by the chromium profile discussed above.
Total 2. Results of radiochemistry analysis of the core sample from I-7M.
Depth
(cm)
5
10
15
20
25
30
35
40
45
50
55
Total
Pb-210
Activity
(dpm/g)
12.510
12.700
1 1 .090
12.150
8.770
8.320
7.980
7.000
5.780
5.440
2.300
Ra-226
Activity
(dpm/g)
1.550
1.990
2.030
1.730
1.740
1.840
1.800
2.350
3.380
3.640
3.360
Cs-137
Activity
(dpm/g)
1.280
1.440
1.350
1.530
1.420
1.660
1.640
1.950
2.820
1.220
0.850
Excess
Pb-210
Activity
(dpm/g)
11.110
10.860
9.190
10.570
7.130
6.580
6.270
4.710
2.440
-1 .220
-1.080
Age
(years)
2.410
5.460
1 1 .620
21.900
31.670
44.500
65.110
101.820
Age
Error
(1s)
1.710
1.790
1.940
2.380
2.870
3.740
6.130
16.450
Date
1994.3
1991.2
1985.1
1974.8
1965.0
1952.2
1931.6
1894.9
Mass
Sedimentation
Rate
(mg/cm2/yr)
225.77
212.07
217.52
146.84
159.07
121.58
76.58
43.32
MSR
Error
ds)
14.85
16.31
19.10
9.77
16.91
15.67
12.89
13.97
-2.5
Activity (dpm/g)
2 0 2 4 6 8 10 12
u
**
v
a
fl-
I-5M
tCB-
132-
H4-I
Excess "Pb
I-7M
The peak input of fallout 137Cs in the late 1950s and early
1960s has been used to provide a time-dependent horizon
in cores. This approach was used to verify CRS dates in
Lake Erie cores (Schelske and Hodell 1995). Neither a
sharp peak nor a large peak in 137Cs activity was found in
the Tannery Bay cores. Therefore, this measurement was
not useful in establishing the 137Cs horizon. The low
inventory of 137Cs activity in both cores is in sharp contrast
to the high inventory of 210Pb activity. These results
indicate that 137CS was deposited and not retained at these
sites for the following reasons:
sediment resuspension focused the 137Cs to other
locations
the 137Cs was diluted by the introduction of large
quantities of tannery wastes
ionic 137Cs was advected with pore waters from the core
site
Figure 5. Activity versus depth of excess 210Pb and 137Cs
at stations I-5M and I-7M
The latter mechanism is prevalent at locations where groundwater is moving through deposited sediments. It
seems unlikely that resuspension or dilution was a primary mechanism because of the large inventories of 210Pb
activity at both sites. The most plausible explanation for the absence of the 137Cs horizon is therefore,
groundwater advection.
70
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Cr (rag/kg)
5000 10000 15000 20000 25000
Cr(mg/kB)
1200024000360004800060000
The influence of groundwater advection on chromium may also be a factor in its fate and transport. As discussed
previously, the absence of the 137Cs horizon suggested that the movement of local groundwater through the
sediments was responsible for advective losses. Since the local groundwater is known to discharge in the near
shore area of Tannery Bay, chromium may also be mobilized from the deeper layers and transported to the
surface. While the solubility of trivalent chromium is
generally limited due to the precipitation of insoluble
hydroxides, the formation of organic complexes has
been shown to increase its solubility. Kaczynski and
Kleber (1994), James and Bartlett (1983), and Hassan
and Garrison (1996) noticed that the solubility of
trivalent chromium was increased in the presence of
organic complexing agents. The latter authors noticed
an increase in solubility in the presence of cysteine
under low Eh conditions. The low Eh environment
present in the sediments of Tannery Bay, in addition to
the organosulfur compounds produced during the
decomposition of animal hides and hair, may produce
conditions that promote chromium complexation. It was
also noted that a large amount of humic material was
released from the Tannery Bay sediments during
alkaline digestion. These materials may also serve as
complexing agents to increase chromium complexation.
The presence of a complexed chromium fraction in the
sediment pore water and its potential role in the
advection of chromium needs to be evaluated as long
as groundwater continues to enter Tannery Bay.
Figure 6. Chromium concentrations and excess 210Pb
versus depth at stations I-5M and I-7M
2.5 5.0 7,5 10,0 12.5
Item "Pbfdpm/g)
-c Chromium
Excess "Pb
10.0
ExcMcMPb(dpm/g)
Chromium
Excess Pb
Since the direct discharge of tannery effluent to Tannery Bay ceased in 1976, evidence of the deposition of
sediments with less chromium contamination should be evident. The lack of a decreasing gradient of chromium
concentrations in the surface zone sediments (0-20 cm) may be explained by several mechanisms:
continued surface runoff
groundwater advection
continual sediment mixing and resuspension in the 0-20 cm zone
Since the levels of excess 210Pb in the surface zone sediments are normal and do not reflect excessive dilution
with terrestrial soil, surface runoff would only be significant if small amounts of highly contaminated material
were continuously eroding into Tannery Bay. As discussed previously, groundwater advection may be
responsible for some migration of chromium from deeper sediment layers to the surface. It is, however, doubtful
that this mechanism would be responsible for chromium levels in excess of 2,000 mg/kg. The most likely process
that would produce the observed chromium levels is that of sediment mixing and resuspension. The flocculent,
fine-grained sediments in Tannery Bay may be mixed to a degree that prohibits the formation of concentration
gradients. The continued mixing of flocculent materials would result in unstable, resuspended sediment that
could readily be exported into White Lake by currents and wave action. The prevalence of the high levels of
chromium in the White Lake samples collected down gradient from Tannery Bay would support the continued
export of resuspended sediments (Rediske et al. 1998)
The CRS model (Appleby and Oldfield 1983) was selected to calculate ages because of the absence of an
exponential 210Pb gradient (Table 1). Ages calculated from the model placed 1892 at 45 cm and 1847 at 50 cm
for core I-5M, and 1895 at 40 cm for core I-7M. These ages, however, were much younger than expected based
on other information available from the core. For example, the chromium concentration exceeded 10,000 mg/kg
from 62-82 cm in I-5M or in sediments older than 1847 according to the calculated 210Pb ages. The chromium
concentration exceeded 10,000 mg/kg at depths down to 84 cm in I-7M. By contrast only the upper 40 cm of this
core contained sufficient levels of excess 210Pb for dating. This lack of conformity shows that the calculated ages
71
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
are not credible. Results from the CRS 210Pb age model that are not credible can be a product of the point
transformations that are used in the CRS model (Robbins and Herche 1993). Independent assessment of dating
is therefore required for the CRS model. For the piston cores, data for chromium and tannery waste by-products
provide independent time markers that are at variance with the calculated 210Pb ages. Since high levels of
chromium and tannery waste byproducts (hair and dye coloration) persist well below the calculated 1847 date,
the dating chronology must be rejected. Large inputs of waste materials could confound the chronological record
by diluting the natural sediments and by altering the physical-chemical environment.
Given the insolubility of trivalent chromium in natural water (Palmer and Puls 1994), the predominant
mechanism driving the flux of this metal in White Lake appears to be sediment export. The hydrodynamics of
White Lake support the progressive transport of sediments in a westerly direction following the natural water
currents. Prevailing winds function to mix the near shore sediments and move the resuspended material out into
the main lake. The 0-20 cm zone of sediment mixing determined by the 210Pb data reflects the action of the
prevailing winds and the wave induced resuspension. The well mixed nature of the top 20 cm zone also suggests
that these sediments are unstable and easily exported. In addition, the differences in stratigraphy between I-5M
and I-7M are consistent with wind induced wave action. Station I-5M has a greater exposure to the westerly wind
and has a lower calculated sedimentation rate (167.23 mg/cm2/yr) and a shallower interval of sediment above the
highly contaminated zone. In contrast, station I-7M is more protected from wave action and exhibits a greater
calculated sedimentation rate (225.77 mg/cm2/yr) and a stratigraphy profile reflecting a greater depth of less
contaminated material.
The discharge of tannery waste was located near the shore of Tannery Bay in the northeast corner. The
EPA/MDEQ core samples indicate heavy sediment contamination with chromium in the near shore and middle
areas of Tannery Bay. Stations near the confluence with White Lake have considerably less chromium in the
sediments. This pattern reflects a discharge of insoluble chromium that was rapidly incorporated into the
sediments. Based on this information, the historical and current mechanism for chromium transport in White
Lake is sediment export from Tannery Bay by the prevailing circulation pattern and wave action. Chromium
export from Tannery Bay into White Lake proper will continue as long as the contaminated sediments are
influenced by hydrodynamic circulation patterns.
CONCLUSIONS
By using a combination of combination of traditional chemical analyses, radiometric determinations, and
stratigraphy, important information concerning the nature and fate sediment contamination in the Tannery Bay
area of eastern White Lake was obtained. Chromium stratigraphy in indicated that the top 15-20 cm of sediment
were less contaminated (2,000-4,000 mg/kg) than sediment located at >30 cm (>5,000 mg/kg). Radionuclide
results suggested that this surface sediment layer was well mixed, however, distinct from the deeper more highly
contaminated sediments. Presently this sediment layer (15-20 cm) does not physically mix with the deeper, more
contaminated sediment. The surface layer was followed by a region (30-80 cm) that contains chromium levels in
excess of 20,000 mg/kg. Since the direct discharge of tannery effluent to this area ceased in 1976, evidence of
the deposition of sediment with less chromium contamination should have been apparent. The lack of a
decreasing gradient of chromium concentration in the near surface zone sediments suggests that the processes
of mixing and resuspension continue to be active in Tannery Bay. In addition, the high inventories of 210Pb in the
0-20 cm zone show that surface runoff from the waste piles has not contributed significantly to the recent
sediment record. The lack of a significant 137Cs horizon in the sediments indicates that groundwater is
discharging in this region; however, the linkage with chromium mobility requires further investigation. While
traditional chemical analyses provide important information for determining the spatial extent of contamination,
additional techniques are required to describe chemical fate and transport. Stratigraphy and radiometric analysis
using 210Pb can provide critical information related to sediment stability, depositional patterns, and chemical flux
that is essential for the analysis of remediation alternatives.
REFERENCES
Appleby, P.G., G.F Oldfield, R. Thompson, P Huttunen, and K. Tolonen. 1979. Pb-210 dating of annually
laminated lake sediments from Finland. Nature 280:53-55.
Appleby, P.G. and F Oldfield. 1983. The assessment of 210Pb data from sites with varying sediment
accumulation rates. Hydrobiologia 103: 29-35.
Binford, M.W. 1990. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake sediment cores. J.
Paleolim. 3:253-267.
Bolattino, C. and R. Fox. 1995. White Lake Area of Concern: 1994 sediment assessment. EPA Technical Report.
72
-------
WTO A '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Great Lakes National Program Office, Chicago.
EPA, 1994. Test Methods for Evaluating Solid Waste Physical/Chemical Methods. US. Environmental Protection
Agency. SW-846, 3rd Edition.
Fisher, M.M., M. Brenner, and K.R. Reddy, 1992. A simple, inexpensive piston corer for collecting undisturbed
sediment/water interface profiles. Journal of Paleominology 7:157-161.
Hassan, S.M. and A.W. Garrison. 1996. Distribution of chromium species between soil and porewater. Chemical.
Speciation and Bioavailability. 8:(3/4)85-103.
James, B.R. and R.J. Bartlett. 1983. Behavior of chromium in soils. V. Fate of organically complexed Cr(lll)
added to soils. J. Environ. Qual. 12:169-172.
Kaczynski, S. E, and R.J. Kiebler. 1994. Hydrophobic Ci8 bound organic complexes of chromium and their
impact on the geochemistry of chromium in natural waters. Environ. Sci. Technol. 28:799-804.
Palmer, C.D. and R.W. Puts. 1994. Natural Attenuation of Hexavalent Chromium in Groundwater and Soils. U.S.
Environmental Protection Agency. EPA/540/5-94/505.
Robbins, J.A., D.N. Edgington, and A.L.W. Kemp. 1978. Comparative 210Pb, 137Cs, and pollen geochronologies of
sediments from Lakes Ontario and Erie. Quat. Res. 10:256-278.
Robbins, J.A., and L.R. Herche. 1993. Models and uncertainty in 210Pb dating of sediments. Int. Ver. Theor.
Angew. Limnol. Verh 25:217-222.
Rediske, R.R., G. Fahnensteil, C. Schelske, P Meier. T. Nalepa, and M. Tuchman, 1998. Preliminary
Investigation of the Extent and Effects of Sediment Contamination in White Lake near the Whitehall Leather
Tannery. Final Report to U. S. EPA. Great Lakes National Program Office. Chicago, II.
Schelske, C.L., A. Peplow, M. Brenner, and C.N. Spencer. 1994. Low-background gamma counting: Applications
for210Pb, dating of sediments. J. Paleolim. 10:115-128.
Schelske, C.L. and D. Hodell. 1995. Using carbon isotopes of bulk sedimentary organic matter to reconstruct the
history of nutrient loading and eutrophication in Lake Eric. Limnol. Oceanogr. 40:918-929.
Wolfe, B., H.J. Kling, G.J. Brunskill, and P Wilkinson. 1994. Multiple dating of a freeze core from Lake 227, and
experimental fertilized lake with varied sediments. Can. J. Fish. Aquat. Sci. 51:2274-2285.
ACKNOWLEDGEMENTS
This work was supported by an Interagency Agreement (IAG) #DW13947766-01 between the Environmental
Protection Agency Great Lakes National Program Office (GLNPO) and the National Oceanic and Atmospheric
Administration (NOAA) Great Lakes Environmental Research Laboratory (GLERL). Other funding was provided
by NOAA/GLERL and Grand Valley State University (GVSU) Water Resources Institute (WRI). The authors
would also like to acknowledge the assistance of the following individuals:
NOAA/GLERL C. Beckman, S. Beattie, R. Stone, and G. Carter
GVSU A. Stiop, F Winkler, J. Oxford, D. Graeber, T. Hudson, and K. Onwuzulike
University of Florida Dr. Jaye Cable
R/V Mudpuppy (USEPA) J. Bonam
AIR-FORCE WIDE BACKGROUND CONCENTRATIONS OF INORGANICS
OCCURRING IN GROUND WATER AND SOIL
Philip M. Hunter
Air Force Center for Environmental Excellence, Consultant Operations Division,
3207 North Road, Brooks AFB, Texas 78235
ABSTRACT
Background concentrations of naturally occurring inorganics are important to site characterization, establishing
cleanup levels, conducting risk assessments, designing and operating long-term monitoring programs, and the
like. The Air Force Center for Environmental Excellence (AFCEE) has streamlined the process for determining
background concentrations by using computer algorithms that interrogate the Air Force's Environmental
Resources Program Information Management System (ERPIMS). Analysis of this database reveals that there are
a wealth of existing sampling locations that are known to be uncontaminated and available for use in these
important calculations. Air-Force wide background concentrations for ground water and soil have been calculated
73
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
for inorganics using all available sampling data from the ERPIMS Database. Insight can be gained by using
these concentrations as a representative baseline of numbers for ongoing and future investigations concerned
with monitoring and remediation of inorganic contamination.
INTRODUCTION
Analysis of the ERPIMS database reveals that most contamination across the Air Force is organic in nature and
is typically associated with chlorinated solvents and fuels (i.e. BTEX and related compounds). The presence of
key organics detected in groundwater and soil samples is a good indicator of both inorganic and organic
contamination. Often Air Force resources are expended needlessly to perform a separate background
investigation at individual installations using newly-captured data rather than relying on existing data.
AFCEE-developed computer algorithms were used to automate the process of identifying background locations
using ERPIMS data. Over 10 years worth of project data is available as an existing resource for background
determinations at installations across the Air Force. This methodology reduces and some cases eliminates the
need to perform a separate background investigation which can cost hundreds of thousands of dollars to
accomplish. This paper will discuss AFCEE's automated approach for identifying background locations, the
statistical methodology used to calculate Air-Force wide background concentrations, and the nature of these
background concentrations for both ground water and soil.
ERPIMS DATABASE
ERPIMS (previously known as IRPIMS) stores some 12.5 million analytical sampling results from 196 Air Force
installations. Data from 40,000 distinct sampling locations (wells, borings, etc.) is captured by the system. The
ERPIMS hardware consists of a Digital Equipment Corporation (DEC) Alphaฎ 4100 computer that runs Oracle*
7.3 on a VMS operating system. The system has been operational since 1987 and is managed by AFCEE/MSC.
DETERMINATION OF BACKGROUND LOCATIONS
Even when investigations specifically target contamination by drilling wells and borings into areas known to be
hazardous waste sites, the non-detect (ND) rates for organics are surprisingly high. For example, the ND rates
for TCE, which is highly mobile and known to be the most ubiquitous constituent found on Air Force installations,
are on the order of 65% in ground water. For most other organic constituents, the ND rates are approximately
90% for ground water. For soil, the ND rates for organics tend to be even higher. As a result, a wealth of existing
sampling locations are known to be uncontaminated and available for use as locations for background
calculations. This knowledge was used to construct a computer algorithm that identifies background locations
across the entire Air Force. The algorithm, which was written in Structured Query Language (SQL), searches out
all locations that have been sampled for both inorganics and organics. Sampling locations showing evidence (i.e.
detects) of organic contamination are then eliminated from the search and those remaining are retained for
further consideration as background locations. Both upgradient, downgradient, and sidegradient locations could
potentially be identified as background sampling locations. There were substantially more background locations
identified for soil as opposed to ground water. On average, at least 25 background well locations and 50
background borehole locations per Air Force installation have been identified using these procedures. As
indicated in the next section that follows, the magnitude of distinct sample locations and the sample sizes
generated from these locations will more than adequately meet the requirements for the statistical calculations
used to determine background levels.
DATA ANALYSIS AND CALCULATION OF BACKGROUND LEVELS
For calculation of background levels at individual installations and sites, AFCEE's approach and statistical
methodology are similar to guidance published by EPA (EPA 1989, 1992), ASTM (1996), and Gibbons (1994).
Using this guidance, normal 95% confidence 95% coverage upper tolerance limits are used. Depending on the
distribution of individual data sets and the percentage of detects, nonparametric tolerance limits are also typically
used. AFCEE, like EPA and others, requires at least 2 sampling locations to minimally account for spatial
variability and a total sample size of at least 8 (n = 8 for each constituent) to provide ample statistical power for
the background calculations. However, Air-Force wide inorganic data is complicated by multiple detection limits,
diverse hydrogeologic terranes, variability over 3-dimensional space, a variety of types of hazardous waste sites,
multiple Air Force bases, different waste handling practices, and the like. All of these issues force one ultimately
to discriminate background levels across more than one hydrostratigraphic unit or more than one soil horizon. As
a result and for the purposes of this investigation, the 95th percentile (Prc95) of the data associated with each
analyte was used as the statistic of choice to best represent background. This approach parallels AFCEE's
guidance for individual installations in that the 95% upper tolerance limit focuses on the 95th percentile of the
data (i.e. a tolerance limit is similar to an upper confidence limit on a specified percentile or coverage of the
74
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
data, in this case the 95th percentile). Calculation of the median and the 99th percentile of background levels for
analytes detected in both ground water and soil were also calculated and can be found in Tables 1 and 2. These
tables also provide information that qualifies the results of the background calculations including: sample size,
the number of distinct background well locations, the number of Air Force bases having background locations,
and the detection frequency. All statistical analysis was performed using SASฎ statistical software.
Table 1. Air-Force Wide
Background Levels of Inorganics in Ground Water
As of May 1998
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Cyanide
Fluoride
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Nitrate
Nitrite
Potassium
Selenium
Silver
Sodium
Strontium
Sulfate
Sulfide
Thallium
Vanadium
Zinc
Sample
Size
5656
5839
7259
6828
5891
812
7153
7892
5121
6412
1796
2132
6844
8916
6465
6017
3481
6471
1788
662
6561
6794
6812
6561
31
3175
176
5698
5378
6820
Wells
Sampled
2493
2843
2996
2750
2843
461
3202
3169
2538
2873
987
1351
2838
3552
2718
2808
1779
2915
1074
468
2750
2934
3165
2750
22
1794
122
2780
2443
2863
AF
Bases
86
97
107
98
98
26
110
112
83
102
49
60
93
118
92
105
58
104
61
33
91
105
108
91
2
77
14
97
84
104
Detection
Freq (%)
54
7
32
83
9
65
9
35
13
29
2
61
78
33
83
9
17
25
67
6
98
12
4
98
100
92
11
4
37
67
Median
(mg/L)
0.0735
ND
ND
0.07715
ND
0.042
ND
ND
ND
ND
ND
0.24
0.56
ND
0.0708
ND
ND
ND
0.8
ND
27.4
ND
ND
27.4
0.17
36.09
ND
ND
ND
0.02
Prc95
(mg/L)
44
0.014
0.044
0.6
0.002
1.46
0.0049
0.195
0.031
0.086
ND
2.263
54.4
0.047
2.840
0.00036
0.021
0.2
24.600
0.02
452
0.0081
ND
452
3670
430
0.14
ND
0.11
0.33
Prc99
(mg/L)
201
0.2
0.171
2.03
0.009
12
0.017
1.52
0.12
0.371
0.015
4.9
240
0.23
9.2
0.0017
0.147
0.83
67
1.1
4050
0.1
0.0155
4050
9070
2420
9.3
0.16
0.464
1.67
BACKGROUND LEVELS FOR GROUND WATER
The universe of distinct monitoring wells that were sampled simultaneously for both inorganics and organics
across the Air Force is approximately 4000 wells as of this writing. The query used to identify the background
data set resulted in the analysis of over 145,000 analytical records. Depending on the analyte, the number of
background wells used in the analysis varied from 22 (strontium) to 3552 (lead) and sample sizes varied from 31
(strontium) to 8916 (lead). Background data was captured from as many as 118 Air Force installations for lead
and as little as 2 installations for strontium. Potassium and sodium were detected 98% of the time while cyanide
was detected 2% of the time. Other analytes such as strontium and sulfate were detected over 90% of the time;
however, the detection frequency for strontium was represented by only 2 Air Force bases and 22 monitoring
wells. Some constituents were not typically detected at background locations. The following analytes had median
concentrations that were below the method detection limit (MDL): antimony, arsenic, beryllium, cadmium,
chromium, cobalt, copper, cyanide, lead, mercury, molybdenum, nickel, nitrite, selenium, silver, sulfide, thallium,
and vanadium. The 95th percentile of the data sets for cyanide, silver, and thallium were also below MDL. This
indicates that they are rarely detected in ground water and was substantiated by detection frequencies that were
found to be in the neighborhood of 2% - 4%. Conversely, some inorganic constituents were detected frequently
and at levels that exceeded important environmental thresholds such as Maximum Contaminant Levels (MCLs)
75
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
or Action Levels for drinking water. The following analytes had background levels (95th percentile) that exceeded
MCLs: antimony, chromium, and nitrate. The background level for lead exceeded the Action Level of 0.015 mg/L
set for drinking water measured at the tap. This may suggest that some regulatory limits are placed artificially
close to observed background levels.
Table 2. Air-Force Wide Background Levels of Inorganics in Soil
As of May 1998
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Cyanide
Fluoride
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Nitrate
Nitrite
Selenium
Silver
Sodium
Strontium
Sulfate
Sulfide
Thallium
Vanadium
Zinc
Sample
Size
13077
15051
17212
15290
14724
790
17464
17549
11815
15396
3220
1270
13719
20784
13495
15465
10584
15167
1400
107
16966
17600
12161
92
1416
204
15186
12342
16017
Wells
Sampled
4840
5683
6165
5765
5513
396
6738
6689
4359
5764
1299
224
4939
7523
4837
5492
3581
5677
273
30
6019
6598
4466
24
273
162
5580
4645
5996
AF
Bases
80
90
101
98
89
16
103
103
81
89
47
8
82
113
80
94
56
92
12
4
99
103
81
2
7
10
89
80
90
Detection
Freq (%)
99
8
66
98
64
63
20
93
60
83
5
79
99
76
99
8
8
68
47
50
8
7
64
100
96
12
6
97
98
Median
(mg/L)
6,510
ND
1.6
56.25
0.3
24.7
ND
9.1
3
7.9
ND
3.4
9180
5
187
ND
ND
6.1
ND
0.008
ND
ND
120
25.1
13
ND
ND
18.6
25.2
Prc95
(mg/L)
23700
5.5
13.8
332.64
1.1
108
2.56
51.8
15.2
53
0.155
9.9
33600
54
856
0.11
1.8
38.3
7.95
0.499
0.87
0.93
1300
111
200
24
0.352
66.6
111
Prc99
(mg/L)
84600
29.3
43.3
995
2.4
201
10
388
28.4
230
2.3
17
82900
340
2380
0.58
7.99
160
42.75
1.1
23.3
7.35
3260
8020
1340
99.7
19
142
540
BACKGROUND LEVELS FOR SOIL
The universe of distinct boreholes sampled for both inorganics and organics across the Air Force was
approximately 8100 boreholes. The query used to identify the background data set resulted in the analysis of
over 325,000 analytical records. Depending on the constituent, the number of background boreholes used in the
analysis varied from 24 (strontium) to 7523 (lead) and sample sizes varied from 92 (strontium) to 20784 (lead).
Background data was captured from as many as 113 Air Force installations for lead and as little as 2 installations
for strontium. Since inorganics tend not to be particularly mobile in ground water, it is not surprising that they are
detected at higher frequencies in soil vis a vis ground water. The following constituents had detection frequencies
exceeding 95%: aluminum, barium, chromium, iron, manganese, strontium, sulfate, vanadium, and zinc. The
high detection frequencies for strontium and sulfate, however, are misleading since the number of Air Force
bases represented is 2 and 7, respectively. Strontium, in particular, had only 24 boreholes that were sampled and
identified as background locations across the Air Force. Some analytes were not commonly detected at
background locations. The following analytes had median concentrations that were below MDLs- antimony,
arsenic, cadmium, cyanide, mercury, molybdenum, nitrate, selenium, silver, sulfide, and thallium. These same
constituents also had median concentrations below MDLs for ground water. None of the 95th percentiles of any of
the data sets for soil fell below MDLs, unlike the situation found for ground water The following analytes had
detection frequencies that were below 10%: antimony, cyanide, mercury, molybdenum, nitrate selenium silver,
sulfide, and thallium. On rare occasion, inorganic constituents were detected at levels that exceeded important
76
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
environmental thresholds (using residential criteria) such as Preliminary Remediation Goals (PRGs), Human
Health Screening Levels (HHSLs), and Risk-Based Concentrations promulgated by various EPA regions. The
background level for both arsenic and beryllium exceeded the PRGs and HHSLs for EPA Region 9 and Region
6, respectively. Iron exceeded both the HHSLs and the PRGs for EPA Regions 6 and 9, respectively. As in the
case for ground water, these results also suggest that regulatory limits may be artificially placed too close to
observed background levels.
SUMMARY
Computer algorithms developed by the Air Force were used to automate the process of identifying background
locations for inorganics occurring in ground water and soil. These procedures identified large numbers of
background locations and a more than adequate sample size which was used to determine Air-Force wide
background levels for some 30 inorganic constituents. This baseline of numbers which was calculated in this
study provides insight on the nature of background variability across the Air Force and gives decision makers a
"feel" for representative background levels. The 95th percentile statistic calculated from individual constituent
data sets is believed to best represent background levels given the inherent complexities associated with
analyzing these large and diverse data sets. Potassium and sodium were highly detected in ground water; while
aluminum, barium, chromium, iron, manganese, strontium, sulfate, vanadium, and zinc were frequently detected
in background soil. Some constituents were not commonly detected at background locations across the Air
Force. The following analytes were not typically found in ground water: antimony, arsenic, beryllium, cadmium,
chromium, cobalt, copper, cyanide, lead, mercury, molybdenum, nickel, nitrite, selenium, silver, sulfide, thallium,
and vanadium. For background soil, the following analytes were not typically detected in soil: antimony, cyanide,
mercury, molybdenum, nitrate, selenium, silver, sulfide, and thallium. The results of this investigation suggest
that some regulatory limits may be placed too close to observed background levels for selected analytes.
Analytes that may fall into this category include antimony, chromium, and nitrate for ground water; and arsenic,
beryllium, and iron for soil. Background levels of these constituents were found to exceed important
environmental thresholds. This automated approach of performing background investigations using existing data
which has already been paid for, affords the Air Force many cost benefits. Use of this methodology can eliminate
the need to conduct a separate background investigation for individual sites or at the installation level and can
save hundreds of thousands of dollars that would otherwise be needlessly spent.
REFERENCES
American Society for Testing Materials (ASTM), 1996, Provisional Standard Guide for Developing Appropriate
Statistical Approaches for Ground-Water Detection Monitoring Programs, Designation: PS 64 - 96.
EPA, Office of Solid Waste, 1989, Statistical Analysis of Ground-Water Monitoring Data at RCRA Facilities,
Interim Final Guidance.
EPA, Office of Solid Waste, 1992, Statistical Analysis of Ground-Water Monitoring Data at RCRA Facilities,
Addendum to Interim Final Guidance.
EPA Region 6, 1996, Human Health Media-Specific Screening Levels.
EPA Region 9, 1996, Preliminary Remediation Goals (PRGs).
EPA Region 3, 1997, Risk-Based Concentrations.
Gibbons, Robert D., 1994, Statistical Methods for Groundwater Monitoring; John Wiley & Sons.
NEW TOOLS FOR LIQUID SAMPLING
'Evaluation and Comparison of the Performance of Liquid Sampling Devices In Stratified Liquids"
James D. Hoover
Department of Civil and Environmental Engineering,
Washington State University Tri-Cities, 100 Sprout Road, Richland, WA 99352
509-372-6972
Scott R. Somers
Advanced Concepts & Design, Inc., 77 Symons Street, Richland, WA 99352
509-943-1431
77
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
EXECUTIVE SUMMARY
Systematic differences in sampling performance were found to exist between the Drum Thief, COLIWASA, and
ACD liquid sampling devices in laboratory tests simulating the containerized sampling of stratified liquids. The
study involved two phases of testing: (1) the collection of 250 samples with these devices using water and corn
oil at a ratio of 1:1 by 17 professional and 35 inexperienced sampling personnel, and (2) the collection of 216 test
samples at wateroil ratios ranging from 95:5 to 5:95 by a single user. Both small volume (<250 ml) and large
volume (wlOOO ml) device models were evaluated. Statistically significant differences in the accuracy, precision,
spillage, and time of sampling were found to exist among these devices. The relative performance of the devices
were also found to coincide with the rating of the devices by 52 volunteer users.
Nearly all differences in sampling performance were found to vary systematically with device type. Differences in
sampling accuracy between devices range from 18% to 175%, and sampling accuracy was found to be highly
dependent on the liquid ratio for some devices. Sampling accuracy with tile COLIWASA and Drum Thief
decreased sharply when the proportion of either liquid was <30%, indicating that there are limits on the liquid
proportions that can be sampled with these devices. Sampling reproducibility among users also varied
systematically with device type, but when used by a single individual, precision was >97% for all devices. The
amount of spillage associated with sampling ranged from zero with some devices to as much as 10% of the
sample volume with others. Sampling time among devices differed by as much as 300%.
The performance of the small volume and large volume models of the ACD devices exceeded that of the other
devices in essentially all categories of evaluation. Nearly all aspects of sampling performance appear to be
dominated by differences in the inherent design and function of the devices, which cause some devices to also
be more user intensive than others. Among the devices tested, sampling performance with the ACD devices was
found to be the least dependent on user factors, including experience. The ACD devices also received the
highest ratings in the user survey.
These results have important implications for the sampling of liquids that is routinely performed for the purpose
of characterizing their composition and potential hazard. These implications include health, safety, economic,
legal, and practical considerations for the handling and disposition of the sampled liquids. These results also
provide a quantitatively basis for comparing the performance of liquid sampling devices used in the
environmental industry and to provide a basis for improving sampling performance and sampler design.
PURPOSE AND BACKGROUND
The sampling of liquids is an activity that is performed thousands of times daily by numerous federal, state, and
local and agencies for the purpose of characterizing type, nature, and/or hazard of unknown liquid substances.
Environmental regulations also require that essentially "all liquid waste materials be sampled and appropriately
characterized for the purposes of appropriate handling, transportation, treatment and/or disposal. Although a
relatively small number of sampling devices are used to collect these samples, there is little documented
information regarding the performance of these sampling devices or a quantitative comparison to serve as a
benchmark for assessing quality control or product improvement
These studies were undertaken to assess the performance of five liquid sampling devices, The purpose of these
studies was to establish a quantitative baseline for the evaluation and comparison of liquid samplers used in the
acquisition of representative samples of stratified liquids. These studies were motivated by the need to establish
a quantitative basis for evaluating the quality of samples that are obtained with various sampling devices.
The sampling and analysis of waste materials, contaminated media, and other materials of unknown composition
are integral components of many environmental activities including regulatory compliance in accordance with the
Resource Conservation and Recovery Act (RCRA; 1976) and the Comprehensive Environmental Response
Compensation and Liability Act (CERCLA; 1980). Sampling protocols have been established to ensure uniformity
in the generation on of sampling data (e.g., Environmental Protection Agency's (EPA's) Office of Solid Waste
and Emergency Response (OSWER (e.g., EPA, 1986). Although general guidelines and procedures for liquid
sampling exist (e.g., ASTM 1994; EPA 1991), there are no standards pertaining to sample quality (e.g.,
accuracy), and essentially no information on the quality of the samples that call be obtained with various types of
sampling devices.
EXPERIMENTAL DESIGN AND TEST METHODS
This Study focused on the evaluation of three types of liquid sampling devices for the sampling of stratified
78
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
liquids from containers such as drums. The three types of liquid sampling devices evaluated were the Drum
Thief, the Composite Liquid WAste SAmpler (COLIWASA), and a new product, the Advanced Concepts &
Design (ACD) liquid samplers. Further description of the sampling devices is provided in Attachment 1. These
liquid sampling devices were chosen because they have all been recognized by the standards organizations as
devices appropriate for liquid sampling (e.g., ASTM 1994, 1995, 1996, 1998; EPA 1986). The Drum Thief and
the COLIWASA have also been the most commonly used liquid sampling devices in the collection of composite
liquid samples. The Drum Thief and small volume models of the other two devices (COLIWASA-S and ACD-S)
were evaluated for the collection of small volume liquid samples <250-ml. The COLIWASA-L and ACD-L models
were evaluated for the collection of larger volume samples (i.e., 1000 ml).
Conditions of stratified liquid sampling were chosen because most data quality issues in containerized liquid
sampling involve representative sampling of segregated liquids with different physical and chemical properties.
Two essentially immiscible liquids, distilled water (p=1.0) and more viscous corn oil (p=0.9), were used in these
tests to simulate simplified test conditions for stratified liquids. The evaluation involved two phases of study. In
the first phase of testing, accuracy, precision, spillage, and sampling time were evaluated for samples collected ^<
from simulated waste drums containing equal proportions of water and oil (ratio -1:1). Phase II tests involved the j]j
assessment of the accuracy and precision of the sampling devices for seven different watenoil ratios ranging
from 95:5 to 5:95. (/)
D
Test Conditions ^
All samples were collected under controlled laboratory conditions simulating the sampling of stratified liquids CC
from 55-gallon waste drums. In the Phase I tests, 17 professional and 35 inexperienced volunteers collected ^
samples with each of the five sampling devices from simulated waste drums under the supervision of laboratory QJ
personnel. The simulated waste drums consisted of cut-away 55-gallon barrels fitted with 4" ID acrylic cylinders j
34" in length (5.6 liter capacity per cylinder) sealed on the bottom and mounted below the bung hole. In the
Phase II tests, samples were collected from freestanding acrylic cylinders by a single individual. Measured ^
volumes of water and corn oil were placed into the cylinders to simulate stratified liquid conditions. A user survey UJ
was also conducted in conjunction with the Phase I tests to obtain all independent assessment of the overall JB
performance of each device from the perspective of user personnel. Details regarding, the experimental ?>
procedures of the Phase I and II tests are described in Attachment 2. 96% accuracy across most of the range
of watenoil ratios. For tests performed with a watenoil ratio of 1:1, under-sampling of one liquid resulted in
complementary oversampling, of the other liquid by an equivalent amount. Sampling biases for tests performed
at other ratios were not complementary (Figure 1).
Precision
Measurement reproducibility, i.e., precision, varied systematically with device type among multiple users, but
varied little when used by a single individual. For the tests performed with a watenoil ratio of 1:1, the ACD
devices produced the greatest precision among both experienced and inexperienced users. For the tests with a
watenoil ratio of 1:1, measurement uncertainty (at the 95% confidence interval) with the Drum Thief and
79
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
COLIWASAs were about twice as large as that with the ACD devices (Figure 2). However, when all samples
were collected by a single user, the sampling precision (at the 95% confidence interval) for all devices was >97%
for all wateroil ratios. Thus, the level of precision produced by an individual user was significantly better that that
produced among multiple users. Sampling precision appears to be only slightly influenced by user experience
and to be generally insensitive to differences in liquid ratios.
Spillage
The amount of spillage from sampling and the variability in the spillage measurements differed systematically
with device type. The average spillage with the small volume devices ranged from essentially zero with the ACD
device (0.006 ml, 1a = 0.02 ml), to as much as 10% of sample volume (24 ml, 1c =13 ml) with the Drum Thief
(Figure 3). For the tests with large volume devices, the COLIWASA-L yielded the greatest spillage and variability
(up to 11 ml, 1a =6.7ml), and the ACD-L yielded the least spillage and variability (0.02 ml, 1a =0.07 ml) (Figure
3). The variability in spillage was generally observed to increase with the volume spilled. User experience
resulted in reduced spillage only with the Drum Thief (by about 30%).
Sampling Time
The time required to collect samples with each of the devices varied systematically with the type of device used,
and was lower among experienced users. Average sampling times for the devices ranged from 40 seconds to
120 seconds. The small volume samples obtained by inexperienced users were collected fastest with the ACD
device. The COLIWASA-S required about 30% longer, and the Drum Thief took nearly three times as long
(Figures 4). The same pattern of device performance was found for samples collected by experienced users, but
with a 20%-30% reduction in sampling times with all devices. The sampling times with the larger volume devices
required only about 10-15 seconds longer. Sampling with the ACD-L was consistently about 10 seconds (10-12%)
faster than with the COLIWASA-L.
User Ratings
The ratings of specific devices by experienced and inexperienced users were nearly identical. The small and
large volume ACD devices received the highest ratings (4.2 and 4.3 out of a maximum of 5.0: Figure 5). The two
COLIWASA devices received the second highest ratings; about 3.0 for the smaller device, and 2.7-3.2 for the
larger device. The Drum Thief was ranked lowest by both groups of users, receiving average scores of 1.3-1.5.
The average ratings and the 95% confidence intervals for these scores are shown in Figures 5.
Other Comparisons
It was hypothesized that some aspects of sampling performance may be related to factors such as the height of
the user and/or arm length for devices requiring 42" tubes to be lifted clear of the sampling vessel for the
collection of the sample. However, no obvious correlation was found between any of the sampling performance
parameters and physical characteristics of the user including height, user stature, or sex.
SUMMARY AND INTERPRETATIONS
The relative performance of the five liquid sampling devices evaluated in these tests is summarized in Table 1.
As indicated in Table 1, the performance of the ACD-S device exceeded that of the Drum Thief and COLIWASA
in essentially all categories of evaluation for small volume sampling devices, and the performance of the ACD-L
exceeded that of the COLIWASA-L for large volume devices. The performance of the COLIWASA-S was
somewhat better than that of the Drum Thief for most test parameters, but the accuracy obtainable with these
two devices appears to vary from user to user. The relative performance of all five devices is also consistent with
the ranking of device performance based on the survey of 17 professional and 35 inexperienced users. The two
ACD devices received the highest ratings, followed by the two COLIWASA devices, with the Drum Thief
receiving the lowest user ratings.
The systematic differences in sampling performance among the devices can largely be attributed to their design
and function. Design factors that affect performance include the relative size of the opening through which liquids
enter the sampler, closing mechanisms, and the tendency for sample leakage. These factors also appear to
control the extent to which device performance is affected by the user.
Much of the sampling error with the COLIWASA and Drum Thief is related to the relatively small diameter
opening on the sampling tube, which causes disproportionate amounts of the liquids to enter the tube. The
magnitude of this error appears to depend on the opening diameter, the rate of insertion of the sampling tube into
the liquids, and on the proportions and viscosities of the liquids sampled. Although the guidelines for use of the
80
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
COLIWASA and Drum Thief recommend a slow insertion, the rate of insertion necessary to minimize this error is
never known in blind sampling. In practice, disproportionate amounts of water and oil were obtained with these
devices by all users. These relationships are discussed further in manuscripts being prepared for publication.
Another important manifestation of this source of bias is the dependency of sampling accuracy on the wateroil
ratio with the Drum Thief and COLIWASA because this error is magnified as the proportion of a liquid becomes
smaller. Although sampling accuracy for all devices was >80% when the proportions of both liquids were greater
than 10%, accuracy with the COLIWASA and Drum Thief decreased substantially when the fraction of the liquid
was less than 30% (Figure 1a,b). At extreme wateroil ratios (e.g., 95:5) the accuracy with these two devices was
as low as 10%. These results indicate that there are limits on the proportions of a liquid that can be sampled with
these devices. Sampling accuracy with the ACD devices, however, was >96% for most liquid ratios and >85% for
extreme liquid proportions. Although there may also be sampling limits with the ACD devices for other testing
conditions, these limits appeal to be much smaller than those for Drum Thief and COLIWASA. Liquid viscosity is
expected to affect the accuracy and limits of sampling with all these devices at some point. However, it is
indicated from the results of this study that sampling accuracy with the ACD devices should be significantly
greater than with the Drum Thief or COLIWASA for most sampling conditions.
These design and function factors also affect sampling accuracy with the Drum Thief and COLIWASA because
they cause performance of these devices to vary from user to user. It is possible for individual users to obtain
highly reproducible results with the COLIWASA and Drum Thief, whether or not they are accurate. But it is
indicated from the results of this study that the uncertainty in sampling accuracy among users is two to three
times higher with the COLIWASA and Drum Thief than with the ACD devices. The Drum Thief appears to be the
most prone to user errors. The high sampling accuracy and precision obtained with the ACD devices among all
users indicates that sampling quality with these devices is essentially independent of user factors, including user
experience.
Design-related performance factors unique to the Drum Thief include the leakage of liquid from tire Drum Thief
and the small tube volume necessitating multiple strokes for the collection of samples. Spillage and the potential
for sampling error related to spillage is greatest with the Drum Thief because significant leakage of liquid from
this device during transfer is inherent to its' use. Although leakage is negligible with the COLIWASA and ACD
devices, spillage resulting from liquid on the outside of the sampling tube occurs with both the COLIWASA and
Drum Thief. This Source of spillage occurs because the 42-inch tubes must be removed from the vessel being
sampled, and lifted above eye-level to transfer the liquid to a sample container. With the Drum Thief, both types
of spillage are compounded because multiple strokes are required to obtain a sufficient sample volume all forms
of spillage are essentially eliminated with the ACID devices because they are designed for transfer of the liquid
directly to a sampling container without removal of the tube from the vessel (e.g., drum). These factors also
affect the time required to collect a sample. Samples were systematically collected more rapidly with the ACD
devices than with the COLIWASA or Drum Thief because direct transfer of samples requires less time than
removal of the tube for transfer. Sampling times with the Drum Thief are also up to three times longer because
multiple strokes are required to obtain samples.
Although statistically significant differences were found in the performance of liquid samplers, these differences
are specific to these test conditions. Based on the evaluation of these results, similar patterns of performance are
expected for most other sampling conditions. The magnitude of the differences in device performance is
expected to vary systematically with changes in physical properties of liquids such as viscosity and density.
However, these relationships, and the effects of other factors such as suspended solids oil device performance
are not presently known, and are the subject of further investigations.
There are many potentially important implications for the results obtained in this study. Although the practical
significance of the differences in device performance must be assessed oil a case-by-case basis, these results
have a number of broad implications for liquid sampling. The implications for significant differences in sampling
accuracy are important because the objective of sampling is to determine the types and amounts of liquids
present so they can be properly (and legally) handled and dispositioned. Representative sampling has safety,
economic, legal, and practical implications because the accuracy of sampling determines, correctly or
incorrectly, the type and level of hazards associated with the liquids. Errors in sampling accuracy have direct and
indirect economic implications involving billings and associated costs that are based on the amounts of specific
chemicals that must handled or otherwise dispositioned. The level of sampling accuracy required also depends
on the inherent hazard of the materials which is usually not known prior to sampling. Thus, sampling should be
81
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
carried out in a manner capable of providing the highest practicable accuracy. The indication that the proportion
limits on sampling with the Drum Thief and COLIWASA are relatively high (5%-30%) is especially notable
because this implies that liquid fractions smaller than 30% are subject to significant sampling errors with these
devices. Sampling precision requirements may also vary from case to case, but it is important that the quality of
sampling be comparable and reproducible from user to user. Spillage has direct implications for worker safety
and for the costs of handling spillage and any subsequent contamination in accordance with regulatory
requirements Sampling time differences of up to 300% can have cost-benefit implications, but only if the quality
of sampling is not compromised. Sampling time and ease of use can also be important when sampling must be
performed in personal protection equipment (PPE) or under adverse (e.g., weather) conditions. Long-term
implications of sampling performance also include the integrated cost savings associated with improved
sampling efficiency and minimizing the negative implications of poor quality sampling.
CONCLUSIONS & RECOMMENDATIONS
Systematic differences in sampling performance were found between the Drum Thief, COLIWASA, and ACD
liquid sampling devices in tests simulating the sampling of stratified liquids. Accuracy and precision were found
to be highest, and spillage and sampling time least, with the small volume (< 250 ml) and large volume (1000 ml)
ACD devices. Sampling accuracy with the COLIWASA and Drum Thief appear to depend on the relative
proportions of the liquids, viscosities, and on the users themselves. The ACD devices were found to be largely
independent of user factors, including user experience for the test conditions evaluated, and to exhibit high
sampling accuracy for proportions ranging from 5% to 95% of each liquid. The Drum Thief and COLIWASA
appear to have limitations on proportion of a liquid that can be reliably sampled between 5%-30%. The overall
performance of the devices based on a survey of 17 professional samplers and 35 inexperienced nolunteer
samplers indicate that both groups of users rated the ACD devices highest followed by the COLIWASA, and the
Drum Thief.
These results have potentially important implications for the handling and disposition of potentially hazardous
liquids, and immediate implications for sampling efficiency and the reduction of risk to sampling personnel from
spillage. The results provide a baseline for quantitatively comparing the performance of liquid sampling devices,
for matching sampling needs with product performance, and also provides a basis for the improvement of
sampling performance and sampler design.
Although these tests are specific to the test conditions used, these results provide insight regarding the factors
that affect liquid sampling under these and other conditions. Tests with liquids over wider range of densities,
viscosities, and liquid conditions including suspended solids, and numbers of liquids, will be required to establish
functional relationships for the performance of these devices over the range of sampling conditions encountered
in the field. Based on the results of this study, the ACD devices appear to provide significantly better sampling
performance and data quality than the COLIWASA and Drum Thief for most conditions of stratified liquid
sampling.
REFERENCES
American Society for Testing and Materials, 1994, Standard Practice for Sampling With a Composite Liquid
Waste Sampler (COLIWASA), ASTM D5495-94, 2pp.
American Society for Testing and Materials, 1995, Standard Practice for Sampling Single or Multilayered Liquids,
With or Without Solids, in Drums or Similar Containers, ASTM D5743-95, 5pp.
American Society for Testing and Materials, 1996, Standard Guide for Sampling of Drums and Similar
Containers by Field Personnel, ASTM D6063-96,18pp.
American Society for Testing and Materials, 1998, Standard Practice for Sampling Single or Multilayered Liquids,
With or Without Solids, in Drums or Similar Containers, ASTM D5743-98, 7pp.
Coin prehensive Environmental Response, Compensation, and Liability Act (Superfund), 42 DSC 9601 et seq.,
1980.
EPA, 1986, Compendium of ERT Waste Sampling Procedures, EPA/540/P-91/008, U.S. Environmental
Protection Agency, Washington, D.C.
EPA, 1991, Test Methods for the Evaluation of Solid Waste Physical/Chemical Methods, SW-846 3rd ed., U.S.
Environmental Protection Agency, Washington, D.C.
The Resource Conservation and Recovery Act of 1976, 42 USC 6901 et seq., 1976.
82
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
-
e> lit
DJ>
"5. ซ
ฃ
-211
w
ซ*
O -40
S?
-61)
-NO
-1110
lllll
XI)
61)
01) 20
--Drum Thief
-*-t'OLIW*S.\-S
-ป-ACD-S
-"-COLIVVASJl-L
--ACD-L
IB 20 3D 40 50 69 70
Percent Water In Stratified Liquid
II
-20
-41)
-lit)
-Nil
-101)
-Drum Tliicf
-COMWASA-S
-ACD-S
-C'OLIWASA-L
-ACD-
Ib
10 2fl 30 40 58 60 70
Percent Oil in Stratified Liquid
80
Figure 1a,b. Accuracy for test results at various watenoil ratios.
Accuracy of sampling for five liquid sampling devices at wateroil ratios ranging from 5:95 to 95:5. Sampling
accuracy is indicated by the percent of over-sampling or under-sampling (bias); e.g., zero bias corresponds to
100% accuracy. Figure 1a shows the amount of over-sampling and/or under-sampling of water obtained with
each device as a function of the amount of water present in the stratified liquids that were sampled. Figure 1b
shows the amount of over-sampling and/or under-sampling of oil obtained with each device as a function of the
amount of oil present in the stratified liquids that were sampled.
83
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
GHucufmrienceil llsorn
Experienced Uierป
OSiiigte Uner
COUWASA-S
Figure 2. Accuracy and precision test results.
Accuracy and precision for liquid samples collected with five sampling devices by experienced inexperienced,
and individual users. Sampling accuracy is indicated in terms of the average percent of over-sampling and/or
under-sampling (bias) for water at a wateroil ratio of 1:1. Zero over-sampling corresponds to 100% accuracy.
Bias levels for water mirror bias levels for oil (i.e., 10% over-sampling of water corresponds to 10%
under-sampling of oil). The vertical lines at the top of each bar represent tile precision of the measurements at
the 95% confidence interval (i.e., 95% confidence that the mean values lies within the range of these error bars).
0 Inexperienced I'serx
Experienced ( ncrs
M'O-S
Drum I Kief
ACU-I.
OOLIH ASA-!.
Figure 3. Spillage Test Results.
Amount of spillage resulting from the collection of liquid samples with five sampling devices The height of the
columns denote the average volume of liquid spilled (ml) by experienced and inexperienced in the process of
KSLTP < ffromhsimulated waste drums. Vertical lines at the top of each bar represent the 95%
confidence intervals for the measurements.
84
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
S
a
E3 Inexperienced Users
Experienced Users
,U:D-S
COLIWASA-S
Drum Thtef
ACD-I.
Figure 4. Sampling Time Test Results.
Time required for the collection of a sample with each of five sampling devices. The height of the columns
denote the average time (in seconds) required for experienced and inexperienced users to collect a liquid sample
from a simulated waste drum and transfer the liquid to a sampling container. Vertical lines at the top of each bar
represent the 95% confidence intervals for these measurements.
ฃ3 Inexperienced Users
Experienced Users
A( TKS
COUWASA-S
Drum 1 hief
ACD-L
fOI.IVVASA-l.
Figure 5. User Rating Results
Results of a survey of 17 professional and 35 inexperienced users regarding the performance of five liquid
sampling devices. After collecting samples with each device, users scored each device on a scale of 1 to 5, with
5 = best, and 1 = worst. The heights of the columns denote the average user score. Vertical lines at the top of
each bar represent 95% confidence intervals for the mean values.
85
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
EXPERIENCED
USERS (n= 17)
Accuracy
Precision
Spillage
Sampling Time
User Ratings
INEXPERIENCED
USERS (n=35)
Accuracy
Precision
Spillage
Sampling Time
User Ratings
SINGLE USER
(n=80)
Accuracy
Fraction = 1 1
FiMion IOซ
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
jar.
ACD Samplers
The third type of sampling device tested was a product developed and manufactured by Advanced Concepts &
Design (AC&D), Inc. The ACD devices are essentially have an open tube design (0.87" and 1.5" ID) in which fluid
in the sample tube is manually drawn by pulling a small cone-shaped plastic plunger upward from the bottom of
the tube using an attached rod or cord extending the length of the collection tube. A sample container Gar)
adapter is attached at the upper end of the tube to receive the sample. Samples are obtained with this device by
lowering the collection tube into the liquid and transferring the liquid from the tube to the sampling container Gar)
by pulling the plunger up through the tube with the rod or cord, essentially pumping the liquid from the tube into
the aligned sample jar without removing the collection tube from the sampled vessel (e.g., drum).
The COLIWASA and the ACD devices both have small volume (-200-250 ml) and large volume (-1000 ml)
models. All three devices are available in glass and high density polyethylene (HOPE) models.
ATTACHMENT 2
Laboratory Procedures
Phase / Testing
Samples were collected with each of the five sampling devices from simulated waste drums under the
supervision of laboratory personnel. The simulated waste drums consisted of cut-away 55-gallon barrels fitted
with 4" ID acrylic cylinders 34" in length (5.6 liter capacity per cylinder) sealed on the bottom and mounted below
the bung hole. Sampling was performed by two groups of volunteer. A total of 250 samples were collected by 17
experienced (professional) samplers, and 35 inexperienced participants with no prior sampling experience. A
description of the study and procedures for the use of each device was provided to each participant (herein
referred to as users) prior to testing. Each user was tasked with drawing a sample from each of five simulated
waste drums using one of the five liquid sampling devices at each drum station, and transferring the samples to
clean sampling containers Gars) placed on top of the drum.
Data on the sample volume, liquid ratios, sampling time, and spillage were obtained for each sample collected.
Pre-weighed drip trays and absorbent material (e.g., paper towels) were placed on the top of the barrels to
capture spillage associated with the transfer of liquid from the drum to the sampling jars. At the conclusion of
each sampling series, laboratory personnel measured the wateroil volume ratio and total volume in each
sampling jar in graduated cylinders. The drip trays were then re-weighed to quantitatively assess the spillage
associated with each device. The sequence in which the devices were used was rotated randomly between users
to minimize bias due to the sequence in which the devices were used. All devices and testing materials were
thoroughly cleaned between sampling series, liquid columns refilled and recalibrated, and pre-measured spill
trays replaced.
All participants completed a user survey at the conclusion of their sampling series. In the survey, users rated
each device in terms of overall performance, and commented on device features, attributes, and shortcomings.
Users also provided personal information on experience and physical characteristics such as height, sex, and
age. Measurements were also made of wrist girth and forearm length to assess correlation between sampling
performance and physical characteristics of the users.
Phase II Tests
Sampling procedures similar to those used in the Phase I tests were employed in the Phase II tests. However, all
samples were collected from freestanding acrylic cylinders by a single individual. Multiple samples were
collected with each sampling device for wateroil ratios ranging from 95:5 to 5:95. A total of 216 samples were
collected in Phase II testing. The following are the wateroil ratios, and the number of measurements made at
each ratio.
Numbers of samples collected in Phase II testing at each wateroil ratio
Preliminary test, were performed with glass and plastic models of the COLIWASA and Drum Thief to assess
performance among these models of the same brand. The samples were collected with each of the seven device
models at a wateroil ratio of 50:50, including ten single-stroke and double-stroke samples with both of the
COLIWASA-S models. The best performing models of the Drum Thief and COLIWASA-S were then used in tests
at other wateroil ratios. The glass model of the Drum Thief was selected based on slightly better sampling
accuracy than the high-density polyethylene (HOPE) model. The COLIWASA-S with the (HOPE) plunger head
87
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
was selected for use in the remaining tests because its' sampling accuracy was slightly better than that with the
model with a borosilicate plunger head. The COLIWASA-L and ACD devices tested were all composed of HOPE.
Single stroke measurements were also made for the purpose of independently evaluating the accuracy and
uncertainty of samples obtained from multiple stroke composites.
RATIO
water/oil
95/5
90/10
70/30
50/150
30/70
10/90
5/95
COLI WASA-Small
- Glass head
- Plastic head
Single
stroke
3
3
3
10
3
3
3
Double
stroke
3
3
3
10
3
3
3
Drum Thief
- Glass model
- Plastic model
Single
stroke
3
3
3
10
3
3
3
Five
stroke
3
3
3
10
3
3
3
COLIWASA
Large
Single
stroke
3
3
3
10
3
3
3
ACD
Small
Single
stroke
3
3
3
10
3
3
3
ACD
Large
Single
stroke
3
3
3
10
3
3
3
The following steps were routinely followed in the conduction of Phase I and Phase II testing:
Each sampler was cleaned before use.
The plunger head was consistently lifted to a height of 4-inches prior to inserting tile sampling tube into the
liquid-filled acrylic cylinder.
A constant insertion rate of approximately 0.5 inches per second was used with all devices.
Sampling devices were closed and sample transfer was initiated immediately after the head of the sampling
tube reached the bottom of the acrylic cylinder.
The Drum Thief and COLIWASA samplers were removed from the acrylic cylinder by pulling them up
through a rag to wipe out excess fluid on the outside surface.
Samples obtained with the Drum Thief and COLIWASA devices were transferred directly to graduated
cylinders.
Samples obtained with the ACD-S and ACD-L devices were first collected in 250ml and 1000ml jars
respectively, and then transferred to graduated cylinders. The jars were allowed to drain into the graduated
cylinders for up to 60 seconds so that transfer loss was negligible.
Sufficient amount of time (up to 20 minutes) was allowed for the sample in the graduated cylinder to be
segregate into layers of water and oil prior to recording the total volume and water and oil volumes of.
Ambient room temperature was recorded at the beginning of each series of tests. An ambient room
temperature from 19-21ฐC was maintained for all tests.
The density of oil in each simulated waste cylinder was determined on daily basis at the beginning of each
experiment. Distilled water was used in all tests.
ANALYSIS OF CHEMICAL WARFARE AGENT DECONTAMINATION BRINES FOR
LEWISITE DEGRADATION PRODUCTS USING GAS CHROMATOGRAPHY
WITH ATOMIC EMISSION DETECTION
Kevin M. Morrissey. Theresa R. Connell, and Jeffrey Mays
EAI Corporation, 1308 Continental Drive, Suite J, Abingdon, Maryland 21009
Tel: (410) 612-7332/Fax: (410) 612-7317
E-mail: kmmorris@eaicorp.com
H. DuPont Durst
US Army, CBDCOM, Edgewcod Research, Development, and Engineering Center, Edgewood, Maryland 21009
Tel: (410) 671-5270/Fax: (410) 671-2081
E-mail: hddurst@cbdcom.apgea.army.mil
88
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Bulk chemical warfare agent (CWA) storage containers which contain decontaminated CWA waste are stored at
several locations throughout the United States. Most of these containers have been in storage since the mid
1970's, and little analytical work has been performed to characterize the contents. This work reports on the
efforts to determine whether these ton containers may have contained Lewisite (L, 2-chlorovinylarsonous
dichloride, CAS No. 541-25-3), an arsenic containing CWA. This work summarizes the efforts at analyzing 250
ton containers stored at a single location.
In the presence of water, Lewisite quickly hydrolyzes to 2-chlorovinylarsenous acid (CVAA, CAS No.
159939-86-3), and this hydrolysis product is not directly amenable to gas chromatographic analysis. The CVAA
was derivatized with 1,3-propanedithiol (PDT, CAS No. 109-80-8), and analyzed by gas chromatography/atomic
emission detection (GC/AED). Quantitation was accomplished using response in the arsenic channel, with
supporting data collected in the sulfur and carbon channels. Spike recovery experiments were performed at 7
different levels, and data will be reported. In addition, total arsenic was determined by ICP/MS, and supporting
techniques of GC/MSD, LC/MS and CE were used to confirm the presence of CVAA and other L hydrolysis
products.
Detectable levels of total As were observed in 25 of the ton containers by ICP/MS. The total As values ranged
from just over the detection limit of 1 ppm, to well over 7800 ppm. Detectable levels of CVAA were observed in
17 of the ton containers by GC/AED. The CVAA levels ranged from just over the detection limit of 0.008 ppm to
2.4 ppm. In addition to the CVAA, additional organo-arsenic compounds were detected in several of the ton
container samples. These additional organo-arsenicals may be indicative of the presence of other As containing
CWA, interaction of CVAA with sulfur mustard (HD, CAS No. 505-60-2) hydrolysis products, or As containing
industrial waste. Correlations will be made between the presence of CVAA and other CWA hydrolysis products.
89
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
90
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
ERA'S
ENVIRONMENTAL
MONITORING
RESEARCH
PROGRAM
91
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
92
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
INTRODUCTION, SESSION SCOPE AND PURPOSE
William Stelz
US EPA, Office of Research and Development, 401 M St., SW, Washington, DC
NO ABSTRACT AVAILABLE
BIOAVAILABILITY AND RISK ASSESSMENT OF COMPLEX MIXTURES
K.C. Donnelly. W.R. Reeves, T.J. McDonald, L.-Y. He and R.L. Autenrieth
Departments of Veterinary Anatomy & Public Health and Civil Engineering,
Texas A&M University, College Station, TX
There is an urgent need to develop an accurate method to assess the risk associated with contaminated soils
and complex mixtures. Perhaps more importantly, this method should provide a means of defining acceptable
residue levels to allow a more cost effective approach to site remediation. This research program is developing a
methodology which can be used to estimate bioavailability. Two soils have been prepared for evaluating the
bioavailability extraction method. One soil is a Weswood silt loam amended with model chemicals, including
chrysene, pyrene, phenanthrene and anthracene; while the second soil is a Weswood silt loam amended with
10% (wt/wt) wood preserving waste (WPW). The soil was spiked with either the model chemicals or the complex
mixture and samples collected immediately after spiking, as well as after 60 and 360 days of incubation (note
day 360 samples will be collected in the fall of 1998). Soil was extracted with pH 7 water or a 1:1 methanohwater
mixture. Other solutions to be tested will include a gastric solution (pH 2) and an intestinal solution (pH 7) and a
3:1 methanol:water mixture. Extractions are performed by shaking 40 g of soil with a 200 mL volume of
extracting solution for 2, 3, or 5 hours (depending on the extractant) at 37ฐC. Recoveries were determined using
GC-FID. In addition, these extractions will be compared to results from desorption kinetics studies. Results from
the digestive fluid extractions indicate that the stomach to intestinal fluid conversion (Gl) extracted only 5.4%
and 0.11% of that recovered by standard methods for chrysene and pyrene, respectively. Using these numbers
as an estimate, the hypothetical excess lifetime cancer risk for the hexane:acetone extraction would be 7.1E-5,
while the estimate for the Gl fluid was 1.3E-7. The desorption study reveals two compartments: one slowly
desorbing and solubility limited, and one limited by desorption/diffusion, which increases in size as the soil ages.
An animal study is planned for this summer using the soil from this study as a means of evaluating these
methods in a rodent model. This research is supported by USEPA Grant No. R825408.
FIELD DETERMINATION OF ORGANICS FROM SOIL AND SLUDGE USING
SUB-CRITICAL WATER EXTRACTION COUPLED WITH SPME AND SPE
Steven B. Hawthorne. Carol B. Grabanski, Arnaud J.M. Lagadec, Martin Krappe,
Cedric L. Moniot, and David J. Miller
Energy and Environmental Research Center, University of North Dakota,
Grand Forks, North Dakota, USA 58202
We have demonstrated that subcritical water (hot water maintained as a liquid by a few bar pressure) is an
excellent solvent to quantitatively extract polar and non-polar organics from soils and sludges. Subcritical water
extractions can be highly selective; polar organics extract at lower temperatures (e.g., phenols and amines
extract at 50 to 100 ฐC), and non-polar organics extract at high temperatures (e.g., 200 to 250 ฐC). By heating
water under low pressure, solubilities of polar organics increase dramatically, and even non-polar organics such
as PAHs can increase solubilities by > 106-fold.
93
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
For water samples, both SPE (solid phase extraction) and SPME (solid phase microextraction) can be used to
extract and concentrate organics in the field for subsequent analysis (e.g., field-portable GC), but are not
applicable to extracting organic pollutants from solid samples. If organic pollutants on soils and sludges could be
efficiently transferred to water, both SPE and SPME could be very useful for field determinations of organic
pollutants from solids. The primary purpose of the proposed investigations is to couple SPE and SPME with
subcritical water extraction of soils and sludges to allow field-portable water methods to be applied to
contaminated solids.
We have coupled subcritical water extraction with SPME using very simple, inexpensive, and field-portable
equipment. The method uses a static extraction (no pump), no flow restrictor, and no organic solvent. Soil, water,
and internal standards are placed in an extraction cell and heated for 15-60 minutes. The cell is then cooled and
the water extracted using a SPME fiber followed by direct desorption in a GC injection port. Although the method
involves multiple partitioning steps (water/soil, and water/SPME), quantitative results can be obtained using
proper internal standards, e.g., deuterated PAHs are added to calibrate for PAH determinations. Methods have
been developed for PCBs, PAHs, and aromatic amines which give good quantitative comparisons to
conventional (Soxhlet) extraction. Typical sample preparation time is < 1 hour, and detection limits of < ppb are
obtained.
In contrast to the multiple partitioning steps involved in the coupled subcritical water/SPME method, coupling
subcritical water with SPE discs (e.g., "Empore" discs) should allow quantitative extraction and collection of
organic analytes. For example, when a static extraction cell contains the soil, water, and an SDB disc, PAHs
extract from the soil into the water during the 250 ฐC heating step, but then are efficiently collected (ca. 90 %) on
the sorbent disc as the extraction cell is cooled to room temperature. The PAHs are then eluted from the disc in a
few ml of solvent, and the extracts analyzed by conventional GC methods. Similar approaches are being
developed for PCBs. In addition, the use of subcritical water to aid in derivatization reactions for the SPME or
SPE collection and analysis of more polar solutes (e.g., acid herbicides, natural pyrethrins) will be presented.
A FIELD PORTABLE CAPILLARY LIQUID/ION CHROMATOGRAPH
T. Scott Kephart, C. Bradley Boring and Purnendu K. Dasgupta
Department of Chemistry and Biochemistry, Texas Tech University, Lubbock, TX 79409-1061
James N. Alexander, IV
Rohm and Haas Company, 727 Norristown Road, P.O. Box 904, Spring House, Pennsylvania 19477-0904
INTRODUCTION
The need for on-site, real time characterization of environmentally hazardous sites has led to a considerable
interest in the development of self-contained field-portable instrumentation. Presently, two factors limit the use of
field portable instruments for environmental analysis. First, most portable instruments do not compare favorably
to laboratory-based instruments with respect to reliability and performance. Second, the availability of
stand-alone field portable equipment is limited primarily to chemical analyzers or sensors which measure a single
physicochemical property such as pH, temperature, or UV/VIS absorbance, although more sophisticated
instruments such as X-ray fluorescence analyzers, mass spectrometers, and Fourier transform IR systems have
recently been developed1. Bringing samples collected in the fields back to the laboratory for analysis results in a
time lag that can compromise sample integrity as well as delay any needed response prompted by the analytical
result.
Analysis in the field often requires the separation of multiple analyte species before detection and quantitation.
For the environmental analyst, liquid chromatography (LC), including ion chromatography (1C), and gas
chromatography (GC) remain the primary techniques of choice. Although field portable GC systems have been
commercially available for some time, field portable LC systems are virtually non-existent. If used in the field, LC
systems typically have to be located in a mobile laboratory, making them at best only moderately portable.
The practice of capillary LC has undergone extensive development since its introduction twenty years ago2. The
primary advantages of moving from conventional size columns (> 4 mm i.d.) into the capillary domain include
94
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
higher efficiencies, higher mass sensitivities, low eluent consumption, and a very small sample requirement. LC
hardware components other than the absorbance detector have been miniaturized with the development of
capillary chromatography. Fully automatable injection valves are available down to 20 nl_; small, compact, and
inexpensive syringe pumps with very low power requirements have been developed.
Recently, we described a capillary 1C with suppressed conductometric detection for anion analysis3. The
miniaturization of LC hardware components in combination with the excellent performance of this bench-top
capillary 1C, led us to investigate the feasibility of a field portable capillary IC/LC system.
In this paper, we present a fully computer controlled, stand-alone field portable capillary 1C. Further, we describe
a dual syringe pump capillary LC system that is in the process of being prepared for field use. The construction
and performance of each of these capillary systems is reported.
EXPERIMENTAL
The layout of the capillary 1C and capillary HPLC systems are shown in Figure 1a and 1b, respectively. The
pumps used were fully computer controlled 48 000 step, motor driven syringe-type dispensers (Model 50300,
Kloehn Inc., Reno, NV) equipped with syringes of appropriate size. A 500 uL glass syringe was used for capillary
1C; stainless steel syringes (constructed in-house) were used for capillary HPLC because operating pressures are
well above 1000 psi. A stainless steel block was machined in-house with appropriately sized ports to
accommodate the appropriate syringe head, a low leak dual ball and seat inlet check valve (P/N 44541, Dionex
Corp., Sunnyvale, CA), and a liquid output port, A small volume eluent reservoir was affixed next to the syringe
dispenser and was connected to the check valve with PEEK tubing to avoid CO2 intrusion. A pressure sensor
(Model SP70-A3000, Senso-Metrics, Simi Valley, CA) was connected to the liquid output port of the stainless
steel block using 0.25 min i.d. PEEK tubing. The column backpressure was continuously monitored to insure
proper system performance.
Capillary 1C System
Water pumped by the syringe was passed through a mixed bed ion exchange resin to remove any impurities
leached from glass and metal parts in the upstream components. A previously described microscale
electrodialytic sodium hydroxide generator3 (EDG) was used for eluent production. A mildly pressurized reservoir
of 25 mM sodium hydroxide was used as the donor solution for the EDG. A 10 cm long polystyrene capillary, -80
urn i.d., 250 urn o.d., was placed at the exit of the EDG to remove the H2 gas in the eluent stream by permeation
through this tube. The polystyrene capillary was able to perform gas removal at pressures > 900 psi at NaOH
concentrations > 40 mM.
A hollow fiber suppressor, which has been previously described,3 was deployed prior to the detector. A H2SO4
regenerant reservoir, mildly pressurized (< 1 psi), was connected to the suppressor using Tygon tubing. The
suppressor was able to suppress NaOH concentrations ranging from 0.5 to 40 mM to a background of < 2 uS/cm
with eluent flow rates of 1.5-2.2 uL/min. A conductivity cell was connected at the suppressor exit. A bipolar pulse
conductivity detection system4 was used for the 1C system. A laptop computer in combination with an executable
program, written in C, provided a user interface for the data acquisition system.5
Capillary HPLC System
The dual syringe pump capillary HPLC system is shown in Figure 1b. The syringe pumps were coupled to a
mixing chamber, having an internal volume of 2 uL, for isocratic or gradient eluent production in the capillary
HPLC system. To produce a constant flow rate while operating in the gradient mode, this setup required a
custom control program to be written. An executable program written in Microsoft Visual Basicฎ provided a user
interface for instrument control. A typical gradient program utilizes a six step gradient, although more steps can
easily be added. The program calculates the appropriate delay between steps for each pump, given the total flow
rate, in terms of the percentage of flow necessary from each pump. The program then creates a command string
with the appropriate delays for each pump and downloads this into the resident memory of the pump hardware
via an RS-232 serial communication port. A Linear model UVIS 200 absorbance detector
(Spectra-Physics/Thermoseparation systems), designed for on-column detection with capillaries, was used for
HPLC detection. A smaller detector will be developed in the future.
An electrically actuated injection valve equipped with 20-100 nL internal sample loops (Valco Instruments,
Houston, TX) was used for sample introduction.
95
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Analytical columns, -50 cm long, 180 um i.d. fused silica capillaries (Polymicro Technologies), were packed
in-house. The 1C columns were packed with the same packing as commercially available Dionex AS-11 columns.
The reverse phase HPLC columns were packed with 5 urn PRP-1 and 5 urn HSC-18 particles, respectively. A frit
was made at the exit end of the column by placing several short pieces of glass wool into 0.3 mm i.d. PTFE
tubing and push fitting the end of the column and a 75 um i.d., 365 um o.d. fused silica capillary on either side of
the glass wool. No frit was needed at the front of the column. This configuration allowed the front of the column
to be easily trimmed when compression of packing material at the head of a column led to a void over a period
of time.
For 1C experiments using sample preconcentration, the preconcentration column consisted of an -1.5 cm long
piece of 250 um i.d. fused silica capillary packed with AS-11 packing. The exit frit was constructed by first
pushing a small piece of glass fiber filter (Whatman type GF/A, Maidstone, England) into this capillary -1 cm
from the end of the packing. A 50 um i.d., 150 um o.d. fused silica capillary was then pushed inside the larger
capillary against the glass fiber filter and epoxied into place. The total length of the preconcentrator column was
-8 cm. The electrically actuated sample injector was equipped with a six port valve (Cheminert Model
C3-1006-EH, Valco Instruments Co. Inc.), having internal dead volumes of 200 nl_ between each port, to
accommodate the preconcentrator column.
A 24 Vdc power supply (Lambda Electronics, Melville, NY) was used to power the pumps and pressure sensor. A
24Vdc-10Vdc converter was built in-house to provide the pressure sensor with 10 Vdc operating voltage. The
bipolar pulse conductivity detector (capillary 1C system only) used a 5 Vdc power supply for operation.
Atmospheric Sampling
The capillary ion chromatograph was interfaced to a miniaturized parallel plate diffusion denuder (PPDD) to
monitor ambient levels of sulfur dioxide. The PPDD construction is shown in Figure 2. The PPDD was
constructed from two Plexiglas plates, each measuring 2x17 cm. The active area of the PPDD (0.6 x 10 cm)
was prepared by thermally pressing silica gel particles (120 mesh or smaller) onto the Plexiglas plates. The two
plates were separated by 1.5 mm thick Teflon coated Plexiglas spacers, 0.7 x 17 cm, which completely cover the
untreated edges of the plates. Holes were machined in the top and bottom of the silica coating to provide a liquid
input and output, respectively. Stainless steel tubing (23 gauge) was push fit into these holes and epoxied to the
plates to provide rigid liquid input/output ports. The two plates were clamped together along their edges. Tubing
for the air inlet and air outlet was fixed by epoxy adhesive at the bottom and top of the PPDD, respectively.
Hydrogen peroxide (0.5 mM, 30 uL/min flow rate) was used as the denuder liquid. The denuder effluent was
loaded onto a preconcentration column at a flow rate of 18 uL/min for 10 minutes for analysis. The PPDD
displayed -100 % collection efficiency up to an air sampling rate of 0.5 standard liters per minute (SLPM). Data
presented here was obtained using this sampling rate.
RESULTS AND DISCUSSION
Portable Capillary 1C
System Performance
The day to day reproducibility of the portable 1C is shown in Figure 3 for repeated injections of fluoride, chloride,
sulfate, and phthalate. The chromatograms were obtained under isocratic conditions using an -20 mM NaOH
eluent at a flow rate of 1.5 uL/min. The relative standard deviation of retention times ranged from 0.1% to 0.7%
within one day and 0.3% to 0.8% day-to-day. Peak efficiencies for chloride, sulfate, and phthalate were 27 133,
21 018, and 15 422 plates/m, respectively.
Response linearity was studied under the same chromatographic conditions as above. A sample solution
containing chloride, sulfate, and phthalate over a concentration range of 10-200 uM was used; fluoride eluted
near an impurity and therefore was not used for evaluating response linearity. Linear r2 values for peak area
response vs. injected concentration for chloride, sulfate, and phthalate were 0.9959, 0.9988, and 0.9974,
respectively. Above a concentration of 200 uM, peak broadening resulted from column overloading.
The three constituent mixture was also evaluated in terms of attainable limits of detection (LOD) under the same
isocratic conditions. The intrinsic electronic noise of the bipolar pulse detector electronics was 0.3-0.4 nS. The
noise increased to 2-3 nS/cm during chromatography, regardless of NaOH concentration or flow rate employed.
Based on the performance at an injection concentration near the baseline and the peak-to-peak noise level, the
S/N = 3 LOD for the three anions are as follows (LOD in uM indicated in parenthesis): chloride (0.03), sulfate
(0.12), and phthalate (0.25). This performance is comparable to conventional size 1C systems. However, an
96
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
increase of >2 orders of magnitude in terms of mass sensitivity is realized in the present system compared to the
performance of a bench-top 1C using conventional size columns.
Gradient Chromatography
Incorporation of the EDG on the high pressure side of the pump also allows gradient Chromatography to be
performed easily. The lag time, or time required for the produced NaOH to reach the head of the column from
the EDG, was measured to be 1.5 minutes using an eluent flow rate of 1.5 uL/min. Therefore, only a short time is
needed for a specific programmed NaOH concentration to reach the head of the column. A gradient
chromatogram of a sample containing 15 anions is shown in Figure 4. A linear gradient of ~2 mM to 38 mM
NaOH from 5 min to 17 min was used for the separation. This corresponded to a current requirement of 5-95 uA
using a water flow rate of 1.5 uL/min. The resulting separation was excellent. Peak efficiencies ranged from 10
648 plates/m for acetate to 240 152 plates/m for chromate, with an average of 80 000 plates/m being observed
for the separation.
Miniaturized PPDD Coupled Capillary Ion Chromatography System
An air sampling rate of 0.5 SLPM was chosen for evaluation of the PPDD-capillary 1C system due to the
collection efficiency of the PPDD being -100% at this sampling rate. The response linearity was studied over an
SO2 concentration range of 23 pptv to 1944 pptv (at an SO2 concentration > 2000 pptv, sample peak height
reached the maximum value permitted with the conductivity detection system). The response linearity over this
concentration range was excellent. A log-log plot of peak height vs. SO2 concentration resulted in a linear r2
value of 0.998. The reproducibility of the data over this concentration range was < 3.2% RSD for each point
sampled. Figure 5 shows a chromatogram resulting from the sampling of clean air and 80 pptv SO2. These data
lead to a computed limit of detection of 1.6 pptv SO2.
Ambient Air Studies
The ambient concentration of SO2 was studied in Lubbock, TX over a period of 48 h. The system operated over
this time period without any user intervention. The results are shown in Figure 6. These results correlate well with
the ambient SO2 levels at this location.
Capillary HPLC System
Isocratic Elution
Experiments evaluating retention time reproducibility were performed on the PRP-1 column injecting samples of
biological interest. A solution of 100 mM ammonium formate (pH 4.25) was contained in pump A and the same
solution containing 10% acetonitrile was contained in pump B. Figure 7 shows system reproducibility for isocratic
elution of 8 sample components with a 50:50 A and B mix.
The average RSD in retention times for the 8 component mixture was 0.825%. Using only a single syringe pump
and employing the same eluent conditions, the average RSD in retention times was 0.921%. The fact that the
dual pump system actually has a lower average RSD indicates that the main source of retention irreproducibility
is not from the pumping system but comes from other components.
Gradient Elution
The gradient capabilities of the system were examined by separating a series of benzene derivatives on the
HSC-18 column. Pump 1 contained a mixture of acetonitrile and water (50:50); pump 2 contained only
acetonitrile. Figure 8 shows a sample chromatogram that also indicates the gradient profile. Figure 9 shows the
dual pump gradient reproducibility. The average RSD in retention times under gradient conditions was 0.545%.
This corresponds to a variation of 2.05 (ฑ.88) seconds for 10 peaks eluting in under 8.5 minutes.
System performance in terms of peak efficiency for the gradient HPLC system was also evaluated. The
maximum peak efficiency was observed for ethylbenzene, which had 320 000 theoretical plates per meter. The
average peak efficiency for the 10 components was 220 000 theoretical plates per meter. This correlates to an
average of 17 000 plates per minute.
ACKNOWLEDGMENT
This research was supported by the U. S. Environmental Protection Agency through Grant R82-5344-01-0. The
manuscript has not, however, been reviewed by the agency and no endorsements should be inferred.
97
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
REFERENCES
1. Newman, A. R. Anal. Chem 1991, 63, 641A-644A.
2. Ishii, D.; Asai, K.; Hibi, K., Jonokuchi, T.; Nagaya, M. J.
Chromatogr. 1977,144, 157-168.
3. Sjogren, A.; Boring, C.B.; Dasgupta, P.K.; Alexander, J.N. Anal.
Chem 1997, 69, 1385-1391.
4. Kar, S.; Dasgupta, P. K.; Liu, H.; Hwang, H. Anal. Chem 1994, 66,
2537-2543.
5. Boring, C.B.; Dasgupta, P.K. and Sjogren, A. J. Chromatogr. 804,
45-54(1998).
Figure 1a. Schematic layout of portable 1C system. Figure
designations: SP, syringe pump; AP, air pressure pump; PS, pressure
sensor; ITC, ion trap column; EDG, electrodialytic sodium hydroxide
generator; PC, polystyrene capillary; I, motorized injector; C, capillary
column; SU, chemical suppressor; D, detector; EB, electronics box.
Top
hlnga
Figure 1b. Schematic layout of capillary HPLC system.
Pump A Pump B
2 3
Computer
12.0
8.0
Day 2 Run 3+10 uS/cm
Day 2 Run 2 + 8 uS/cm
PP-
-TF
VPS
Side View
Figure 2. Wet parallel plate diffusion denuder. Figure
Day 2 Run 1+5 us/cm designations: AO, air outlet; LI, liquid inlets; PS, Teflon
coated Plexiglas spacer; SC, silica coating; LO, liquid
outlet; PP, Plexiglas plates; Al, air inlet; TF, Teflon film
Day 1 Run 3 + 4pS/cm
Day 1 Run 2 + 2 uS/cm
Day 1 Run 1
Figure 3. Day-to-day system reproducibility; repeated
100nL injections with ~20 mM NaOH eluent. Injected
concentration was 20 uM for each ion. Peak identities: 1,
fluoride; 2, chloride; 3, sulfate; 4, phthalate.
98
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
3.0 -
13
14
15
Figure 4. Background subtracted gradient
chromatogram. A linear gradient from 2 mM NaOH to
38 mM NaOH from 5 to 17 minutes was used. Peak
identities: 1, acetate; 2, formate; 3, methanesulfonate;
4, monochloroacetate; 5, bromate; 6, chloride; 7,
nitrite; 8, trifluoroacetate; 9, dichloroacetate; 10,
bromide; 11, nitrate; 12, chlorate; 13, sulfate; 14,
phthalate; 15, chromate. All ions were 50 uM except
dichloroacetate which was 60 uM.
10 20
Time (min)
Figure 5. Chromatograms resulting from
sampling blank air (lower trace) and 80
pptv SO2 (upper trace). Peak identities: 1,
chloride; 2, carbonate; 3, sulfate.
o
o
80 pptv sulfur
dioxide -f 1 uS/cm
Blank
4.0 B.O
Time (min)
12.0
99
-------
1000 i
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
eoo
Figure 6. Ambient air levels of SO2 in Lubbock, TX
for a 48-hr period beginning in the afternoon of April
28, 1998.
12
1 I I
24 12 24
Diurnal Time (hrs)
zsoo
Figure 7. Sample of isocratic system reproducibility.
RSD in retention time is < 1%. All samples are 500
\M. Peak identities from left to right: cytosine, uracil,
adenine, uridine, thymidine, adenosine, xanthosine,
and inosine.
-500
e
o
0.20 ,
0.15
0.10
0.05
0.00
u
, 100.00
80.00
80.00
40.00
6789
Time (min)
12 13
20.00
Figure 8. Sample chromatogram with gradient
profile using an eluent flow rate of 5.0 uL/min.
All samples are 0.5 ml / 100 ml solution in
acetonitrile. Peak identities from left to right:
phenol, benzaldehyde, benzonitrile, nitro-
benzene, benzene, bromobenzene, toluene,
ethylbenzene, propylbenzene, and t-butyl-
benzene.
0.00
2.00
4.00 6.00
Timp fminl
8.00
100
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
0.1 AU
Figure 9. Dual pump gradient reproducibility.
Average RSD in peak retention time is 0.545%.
Peak identities from left to right: phenol,
benzaldehyde, benzonitrile, nitrobenzene,
benzene, bromobenzene, toluene, ethylbenzene,
propylbenzene, and t-butylbenzene.
Time (min)
RAPID DETERMINATION OF ORGANIC CONTAMINANTS IN WATER BY SOLID PHASE
MICROEXTRACTION AND INFRARED SPECTROSCOPY
David C. Tilotta. Danese C. Stahl, Sheila A. Merschman, Daniel L Heglund, and Said H. Lubbad
Department of Chemistry, P.O. Box 9024, The University of North Dakota, Grand Forks, ND 59202
The objectives of this research project are to: identify suitable solid phase films for determining organic
contaminants in water by SPME/IR, determine which organic contaminants are amenable to the SPME/IR
method, and adapt the basic methodology to field use.
Solid Phases
Of the 15 solid phases examined to date, three polymers have been found to be useful, for SPME/IR: Parafilm
M (a wax-impregnated polymer/rubber composite), poly(dimethylsiloxane) (PDMS, an important solid phase
material of the SPME syringe technology) and Teflon PFA (a perfluoroalkoxy teflon polymer).
Analyte Classes
To date, three classes of compounds have been examined for their suitability as analytes for SPME/IR using the
three aforementioned films. Table 1 shows formal equilibration times, linear dynamic ranges, detection limits,
and precision data (expressed as percent relative standard deviation) for these classes for the appropriate solid
phase film(s). Multiple entries in this table for a given analyte imply that more than one film is useful Conversely,
the absence of an entry for a given analyte/film combination indicates that that film is not suitable for the
analysis.
Volatile organic compounds (VOCs) examined include the BTEX compounds (benzene, toluene, ethylbenzene,
xylenes), and halocarbons such as carbon tetrachloride, chlorobenzene, chloroform, and p-chlorotoluene.
Parafilm M and PDMS both have been useful for the analytical determination of these compounds. SPME/IR
analyses using Parafilm, have demonstrated the ability of SPME/IR in distinguishing four of the six
alkylbenzenes (benzene, o-xylene, m-mylene, p-xylene) in petroleum industry wastewater samples. Quantitation
by simple univariate calibration based on absorbance band heights have provided good agreement with purge
and trap GC/MS standard methods. Analytical determinations of ethylbenzene and toluene are, however,
complicated by the spectral overlap of other components in gasoline.
101
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Gasoline fuels include the more volatile organics such as the short chain hydrocarbons (e.g.,
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
produced or processed for human use, are released into the environment. Upon reaching the environment, the
contaminants can be transported in a variety of ways until they reach a depositional end point, possibly in soils,
waters or in biological tissues. Depending on the reactivity and toxicity of the contaminants, there is a multitude
of possible implications for both the environment and human populations. It is necessary, therefore, to
understand the nature of the contaminants, where they came from, and how they reached their depositional
location. With this understanding, cleaning contaminated sites and preventing further contamination become
more manageable endeavors.
There are many compound classes that have been contributors of contaminants into the environment. Of
growing interest in recent decades are polycyclic aromatic hydrocarbons (PAH), organochlorine (OC) pesticides
and polychlorinated biphenyls (PCBs). PAH are of interest because they can be indicative of a variety of
contaminant sources, such as petroleum spills or combustion processes and are known to be carcinogenic
(LaFlemme and Hiles, 1978 and Hites et a/., 1980). Although some organisms, such as fish, are able to
metabolize PAH, other organisms such as mollusks and crustaceans are unable to do so (Law and Biscaya,
1994; Hickey et a/., 1995; Roper et a/., 1997). Thus, PAH tend to accumulate in the tissue of these organisms.
Pesticides, such as DDT, and PCBs are of interest because they not only can reside in soil and water reservoirs,
but due to their lipophilic nature, they can accumulate in tissues and are known to be toxic (Killops and Killops,
1993 and Hickey et a/., 1995). The storage of such contaminants in tissue results in heightened concentrations in
organisms residing at higher trophic levels, thus enhancing the risk to many predatory species.
In many earlier attempts at deciphering the sources of contaminants such as PAH and PCBs, the approach has
been to look at absolute concentration levels and relate a concentration gradient to a point source. Another
approach was to determine the concentrations of compound classes relative to one another and compare these
relationships to possible source relationships. One difficulty with the first approach is that although measurable
amounts of a contaminant may be located near a source known to produce such compounds, there is no
absolute proof that the contaminant came from that source. Furthermore, owing to the off-site acquisition of
pesticides, there is rarely a point source nearby which can serve as a possible answer to contaminant
apportioning scenarios. A problem with the second approach is that chemical or biological activities such as
evaporation, water washing or biodegradation could alter the concentration of one compound relative to the
others. This chemical or biological alteration could change the original "signature" of the compound class that
could lead to an incorrect assignment of a source (O'Malley et a/., 1994). In the past, traditional analytical
methods using gas chromatography (GC) and gas chromatography/mass spectrometry (GC/MS) have been used
to characterize contamination sites. As the above arguments show however, these techniques may yield
ambiguous results (Mansuy et a/., 1997).
A Different Technique
In recent years another approach has developed which is based on the stable isotopic composition of the
compounds of interest. Carbon, for instance, has two naturally occurring stable isotopes, 13C and 12C. It is
reasonable to assume that the ratio of the amounts of these two isotopes is unique for each compound derived
from a different source. If two chemical companies were to produce the chemically identical PCBs, for instance,
it is probable that the identical compounds will have isotope ratios characteristic of different feedstocks or
different manufacturing processes. Upon entering the environment, therefore, a contaminant should be able to
be linked to a source by its isotopic composition.
The technique for compound specific isotopic analysis (CSIA) involves coupling a gas chromatograph (GC) to a
combustion furnace, which is then attached to an isotope ratio mass spectrometer (GC/C/IRMS). The effluent of
the gas chromatograph is introduced into a microcombustion/CO2 purification interface. Within the GC, fused
silica columns have been found to be most effective because carrier gas flows are low (a few ml/min. of He),
resolving power is excellent (30 to 60 meter column lengths are routine), and column bleed is minimal (bonded
phase columns). The column effluent is combusted to CO2 in the interface in the presence of CuO and water is
removed by a cryogenic trap. Ion current intensities of masses 44, 45 and 46, which represent the major isotopic
forms of CO2, are recorded simultaneously using a high speed on-line acquisition system. A schematic of the
system is shown below (Figure 1).
The ratio of the ion current intensities of mass 45 to mass 44 is a measure of the ratio of 13C/12C and is compared
to the reference value of 13CI/12C. The reference and sample ion current intensities are measured in an
alternating manner, which ensures a reliable comparison between the reference and sample isotope ratios. A
computer interfaced to the instrument is used to calculate in the "per mil" (%o) notation. Isotope compositions are
103
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
reported as 5 (delta) values, using the following formula:
R = 13C/12C
5 13C = (Rsample/Rreference - 1) X 1000
Gas Chromatograph
Injector Detector
Heart split
valve
Reference Gas-
Injector
Figure 1. Schematic
showing the basic structure
of a GC/C/IRMS system
(adapted from Freedman et
a/., 1988).
Isotopic compositions are measured as ratios owing to the normally low absolute abundances of each chemical
species of interest. Samples with more 13C than the reference compound will have positive 5 13C values, and are
said to be enriched in 13C. Samples with less 13C than the reference compound are said to be depleted. It should
be noted that currently, while there are many elements with multiple stable isotopes, the most commonly used in
analyses are carbon, nitrogen, oxygen, sulfur and hydrogen. Like carbon, each of the other elements are
analyzed in a similar way, with deviations occurring in the combustion and subsequent gas purification processes
occurring just before mass abundances are measured. In every case, a sample with more of the heavy stable
isotope is said to be enriched for that element under investigation. For instance, with respect to nitrogen, using
new modifications of the combustion system, it is now possible to perform CSIA on nitrogenous compounds
(Macko et a/., 1997). In this case, the heavy isotope is 15N and the mass spectrometer measures the ion current
intensities of masses 29 and 28, representingthe two most common forms of molecular nitrogen.
With a state-of-the-art instrument, reproducibility is
measurements typically better than 0.5 %o (Tables 1-3).
quite good, with standard deviations on replicate
Table 1. Average 8 13C values (relative to PDB reference) with associated standard deviations for five PAH
standards.
Compound
Naphthalene
Phenanthrene
Anthracene
Chrysene
2,3,6 - Trimethylnapthalene
Averaae 8 13C value
-26.2
-24.4
-23.7
-23.9
-22.6
Standard Deviation (+/-)
0.4
0.45
0.44
0.58
0.20
Table 2. Average 8 13C values (relative to PDB reference) with associated standard deviations for four alkane
standards.
Compound
Decane
Undecare
Dodecane
Tridecane
Averaqe 8 13C value
-29.0
-27.1
-33.3
-32.6
Standard Deviation
0.30
0.30
0.30
0.40
Also, with a state-of-the-art instrument, performing CSIA with a GC/C/IRMS system has many advantages over
other techniques for determining contaminant identities and sources. Because a GC is coupled to the mass
spectrometer, there is the capability of resolving complex mixtures. Also, with such a system, there is very high
104
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
sensitivity, which can be a very important consideration when dealing with trace contaminants. With a
GC/C/IRMS system, detection is possible in the sub nM region.
Table 3. Average 6 15N values (relative to atmospheric N2 reference) with associated standard deviations for four
nitrogen-containing compounds.
Compound
Pyrazine
Tripropylamine
Quinoxaline
Nicotine
Averaae 8 15N value
0.7
9.5
-0.6
-1.7
Standard Deviation
0.12
0.55
0.16
0.41
The Approach
In applying CSIA to the problem of tracing environmental contaminants, measures must be taken to determine
whether the techniques are viable under natural conditions. As was mentioned above, using GC/IRMS to
measure isotopic compositions for the purpose of tracing contaminants seems to be an exciting alternative to
traditional methods. The first steps taken toward application of the technique must be field tests, however. This
can be stated in a series of objectives:
1) Establish extraction, isolation, and purification techniques for the contaminants of interest. Of particular
importance is obtaining high purity samples that have not been altered in isotopic composition.
2) Determine the stable isotopic composition of the compounds of interest obtained from primary sources, such
as manufacturers, petroleum suppliers and chemical storage locations.
3) Determine the stable isotopic composition of the contaminants at locations where they are introduced into
the environment. Possible locations include combustion areas (automobile exhaust and industrial exhaust),
industrial emission, agricultural runoff, urban runoff and sewage effluent.
Once the first three objectives are fulfilled, the next step is to apply them to well-characterized sites where
pollution sources and times of occurrence are well known. In this initial phase, carbon isotopes have been the
focus, and within this framework, early efforts have concentrated on the most common PAH, OC pesticides and
PCBs (Table 4). The reasons for first analyzing the most common contaminants are two-fold. First, in trying to
determine the validity and applicability of this technique, it is undesirable to be sample-limited. From a sample
extraction and purification standpoint, it is necessary to deal with sufficient sample quantities. Once the method
proves successful, samples of lower of concentration can be addressed. The second reason for focusing on the
most common contaminants is time constraints. This approach is fairly time consuming, and thus only a limited
number of samples can be analyzed during the study. Choosing the most abundant contaminants for analysis
should allow for the greatest chance of success.
Methods
In this study, the matrices of interest are soil and biological tissue. Typically, the first step is the extraction
process. Soil samples are Soxhlet extracted with methylene chloride and then separated based on compound
class with an alumina/silica column. Aliphatic hydrocarbons are eluted from the column with 50 ml of pentane.
Aromatic hydrocarbons, PCBs and OC pesticides are eluted with 150 ml of pentane/methylene chloride (1:1).
The second fraction is further separated on a silica column. PCBs are eluted with an additional 90 ml of pentane.
High performance liquid chromatography (HPLC) using a cyano/amino bonded phase column as per Killops and
Readman (1985) is used to purify the PCBs from fraction 2. Pentane is used as the eluting solvent. Purity is then
checked with gas chromatography/electron capture detector (GC/ECD) and GC/MS systems. Fraction two from
the silica column is separated into several subfractions, with the aromatic compounds being separated based on
the number of double bonds they contain (Killops and Readman, 1985). A pentane and methylene chloride
gradient is used for the elution. It should be noted that others have shown that aromatic species can also be
separated with molecular sieves based on the arrangement of any alkyl substitutions present (Ellis et a/., 1992;
1994). If need be, this is another viable alternative.
For the tissue samples, 20-30 grams of wet tissue are mixed with approximately 50 g of anhydrous sodium
sulfate. The extraction is performed with three 100 ml methylene chloride aliquots while macerating with a
homogenizer. The combined extract is then separated as described above except there is not an aliphatic
hydrocarbon component to contend with. Before analyzing by GC/C/IRMS, if the samples still are not pure
105
-------
MTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
enough, thin layer chromatography (TLC) can be utilized for further purification.
Table 4. List of contaminants that are initially being
investigated.
Contaminant
PAH
Pesticides
PCBs
Upon final separation and purification, the analytes of
interest are taken up in an appropriate solvent (pentane
and methylene chloride, for example) to a concentration
in the 5-100 ng/ul range, depending on the sensitivity of
the instrument for a particular compound class. Samples
are then analyzed for isotopic composition by the
GC/C/IRMS system as described above.
Other Considerations
The stable isotopic composition of a contaminant in the
environment is the end-result of a complex chain of
events. Naturally occurring pollutants such as spilled
crude oil are most easily traced in the environment
because the starting material is usually readily available
for analysis. Furthermore, intrinsic tracers in spilled crude
oil would also be directly reflected in environmental
samples since few complicating processes would
intervene. Many synthetic materials are produced from
petrochemical precursors through manufacturing
processes such as distillation, catalytic cracking,
chlorination and polymerization to name a few. A change
in isotopic composition might occur during the
manufacturing process if catalysis or high temperature is
involved. Next the product is applied for its intended
purpose which could include combustion (gasoline),
lubrication (lube oils), pesticide application (DDT), or use
as a transformer oil (PCBs) for example. These
applications may cause an additional shift in isotopic and
molecular composition. During the application, either
intentionally or indirectly, some portion of the
contaminant is released to the environment (such as soot
from combustion or the disposal of waste materials). Once released to the environment, the contaminant is then
subjected to redistribution throughout various matrices including air, water, sediments, and biological tissue
depending on its chemical properties and stability. The partitioning of the chemical among various phases might
be accompanied by a shift in isotopic composition as well as chemical transformation. Environmental
transformations are brought about by physical, chemical and microbiological processes. Each process defines an
independent set of isotopic and compositional changes. The complex history of a pollutant suggests that a
combination of compositional and stable isotopes can be linked to a specific series of events and processes
(Figure 2). Chemicals of identical structure may have different isotopic composition if they have witnessed
different histories from manufacture to environmental deposition. By analyzing contaminant samples at different
stages of transport it should be possible to understand the dominant processes acting on them and to elucidate
the complex chain of events which led to the isotopic composition of the contaminants at their environmental end
point.
The issue of complex mixtures
As was mentioned above, compound specific isotopic analysis using GC/C/IRMS has the potential to handle
complex mixtures. It is inevitable, however, that during some analyses, there are two or more chemical species
present in a mixture that are similar chemically and therefore coelute. Perhaps in some of the separated fractions
there could be additional compounds present other than the desired analytes of interest that are similar to the
desired compounds. If there is only one other compound present, temperature programs for the GC could be
altered, columns could be changed, either in length or composition, or additional separation steps could be
added to alleviate this problem. When there is a high amount of background, however, owing to more than one
additional chemical species being present, the problem is more difficult to handle. The coelution will lead to
erroneous isotopic results because what was thought to be the composition of one compound is actually the
combined isotopic composition of two or more compounds. This issue is complicated further by the fact that
isotope fractionation occurs during elution, with portions of the peak varying dramatically in isotopic composition
Naphthalenes
Fluorenes
Phenanthrenes/Anthracenes
Dibenzothiophenes
Fluoranthenes/Pyrenes
Benzanthracenes
Benzofluoroanthenes
Benzopyrenes, Dibenzanthracenes
Perylene
Aldrin, Heptachlor, Endrin, Mirex
DDT, Dieldrin, Transnonachlor
Dichlorobiphenyls
Trichlorobiphenyls
Tetrachlorobiphenyls
Pentachlorobiphenyls
Hexachlorobiphenyls
Heptachlorobiphenyls
Octachlorobiphenyls
Nenachlorobiphenyls
106
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
(Figure 3). Although the isotope values can be averaged
across a peak, with coelution, the modifying effect of the
additional species can also vary across the peak if the
analyte of interest and the additional species do not
elute at exactly the same time. Some researchers have
used internal standards in their analyses to monitor the
degree to which unresolved complex mixtures (UCM) or
other coelution problems affect the results of the isotopic
analyses. By knowing how much the isotopic values of
compounds of known composition have shifted,
corrections can be made for the unknowns (Mansuy et
a/., 1997). Other researchers have relied on certain
separation techniques to try to keep the problem to a
minimum (Ellis et a/., 1994). Also, within some software
packages, monitoring of the background is possible,
along with the subsequent subtraction of unwanted
contributions to a peak. Based on the above discussion,
it is obvious that this issue will require careful monitoring
in order to insure results are as accurate as possible.
Initial Applications
Extraction, purification, and analysis on some of the
analytes of interest has already been performed (Table
5). PAH mixtures from two sediment sites as well as a
sample of creosote were analyzed by GC/C/IRMS
yielding variable carbon isotope values. This variability
could indicate different sources of the PAH influence
into the various reservoirs.
Figure 2. Schematic representing possible pathways for
contaminants that enter the environment.
Contaminant
Source
(Original Compounds)
-manufซclurins -
A S13C.
AS'3C,T
Possible Transformation
(Biodegradation, other chemical
reactions?)
Continued Transport
Deposition in Environment
A S
Soil
Water
Tissue
Atmosphere
Table 5. 8 13C values (relative to PDB reference) for PAH mixtures derived from three different sources.
Compound
Naphthalene
2-Methylnaphthalene
1-Methylnaphthalene
Biphenyl
2,6-Dimethylnaphthalne
Acenaphthalene
Fluorene
Phenanthrene
1 -Methylphenanthrene
Fluoranthene
Pyrene
Chrysene
Oreaon Inlet Sediment
-27.1
-27.8
-28.2
-28
-28.8
-27.4
-27.7
-25.6
-25.8
-24.6
-23.0
-23.0
Creosote
-23.5
-23.1
-21.2
-21.4
-18.4
-24.3
-25.2
-25.2
Casco Bay Sediment
-27.9
-29.5
-26.4
107
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
nf XC. / TRMS
t.UOB-?
o 1.120B-2-
1.100B-2:
1.080B-2:
> I
0 200 400 ฃ00
l.OOE-31
8. COB-9-
6.00E-9-
4.00E -9-
2.00E-*
O.OOE+I)
Figure 3. Plot of the carbon isotope signal of three hydrocarbon species. The upper trace is the ratio of ion
current intensities of masses 45 and 44. The lower trace is the ion current intensity of mass 44. Note the
changing isotope ratio across the peak.
Stable Isotopic Analysis as a Versatile Tool
In recent years, more attention has been focused on stable isotopic analysis, and in recent years, on compound
specific isotopic analysis. While traditional molecular approaches are still in active use today, stable isotope
analysis has gained popularity as a complement to other techniques or a stand-alone technique when other
approaches are ineffective or too time consuming. Furthermore, as more attention is paid to stable isotopic
analysis, the versatility of the technique has become more evident. As was mentioned in the above discussion,
stable isotope analysis can be used as an effective tracer or method to apportion sources of compounds to a site.
GC/C/IRMS has been used to distinguish oils based on the isotopic composition of alkane and isoprenoid
constituents present in the oil (Bjoroy et a/., 1991). This technique has been applied to oil spills as well; even
under conditions where some of the oils have been weathered (Mansuy et a/., 1997). With respect to the analytes
108
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
of interest in this study, CSIA has been used to not only trace the transport of PAH in aerosols (Ballentine et a/.,
1995) but to distinguish between biomass burning and fossil fuel burning sources of PAH (Ballentine et a/., 1996).
Also, sources of PAH have been identified along with the relative amount of influence of each in marine soils.
(O'Malley el a/., 1994). With respect to PCBs, it was discovered that individual compounds tend to be more
depleted in 13C the more chlorine atoms there are present in the structure (Jarman et a/., 1998). Stable isotopic
analyses have also been used in groundwater studies, both as a source apportioning tool (Kelley et a/., 1997) and
also as a means of tracing compounds during groundwater flow (Dempster et a/., 1997). Perhaps one of the most
unique characteristics of stable isotope analysis is that is has been applied to many different scientific areas. Two
areas that are utilizing this approach more and more are biology and ecology. Characteristics of food webs, from
trophic structure to influences on the diet on individual organisms have found application in stable isotope
analysis (Fang et a/., 1993 and Jarman et a/., 1996). As has been shown, beyond the realm of just contaminant
studies, compound specific stable isotope analyses has a wide range of applicability. This should prove very
beneficial as the call for more interdisciplinary science continues.
Conclusions
With the current heightened awareness of the need to understand and control contamination sites and sources,
the demand for powerful and accurate analytical tools to answer these questions is very high. Compound specific
isotopic analysis has been suggested as a new tool to be used to answer these questions. Upon purifying
contaminant fractions, GC/C/IRMS can be used to determine the isotopic composition of the individual
components in these fractions. These stable isotope compositions can compared to the isotopic compositions of
different source materials and contaminants entering the environment through various pathways. With this
information, the analyst can begin decipher together the history of the contaminants, linking the source and the
process of introduction to the contaminants of interest. Whereas first only being applied to well characterized
sites, upon finding that the techniques are valid, the methodology can be applied to sites of unknown composition
and influence. Compound specific isotopic analysis could then be used to complement other analytical methods
or as a stand-alone technique where the other methods prove to be ineffective.
Works Cited
Ballentine, D.C., Macko, S.A., Turekian, V.C., Gilhooly, W.P. and B. Martincigh. 1995. Transport of biomass
burning products through compound specific isotope analysis, Selected Papers from the 17th International
Meeting on Organic Geochemistry, 644-646.
Ballentine, D.C., Macko, S.A., Turekian, V.C., Gilhooly, W.P and B. Martincigh. 1996. Compound specific
isotope analysis of fatty acids and polycyclic aromatic hydrocarbons in aerosols: implications for biomass
burning, Organic Geochemistry, 25, 97-104.
Bjoroy, M., Hall, K., Gillyon, P and J. Jumeau. 1991. Carbon isotope variations in n-alkanes and isoprenoids of
whole oils, Chemical Geology, 93, 13-20.
Dempster, H.S., Sherwood Lollar, B. and S. Feenstra. 1997. Tracing organic contaminants in groundwater: a new
methodology using compound specific isotopic analysis, EnvironmentalScience and Technology, 31,
3193-3197.
Ellis, L, Alexander, R., and R.I. Kagi. 1994. Separation of petroleum hydrocarbons using dealuminated
mordenite molecular sieve-ll. Alkylnapthalenes and alkylyphenanthrenes, Organic Geochemistry, 21,
849-855.
Ellis, L., Kagi, R.I., and R. Alexander. 1992. Separation of petroleum hydrocarbons using dealuminated
mordenite molecular sieve. I. Monoaromatic hydrocarbons, Organic Geochemistry, 18, 587-593.
Fang, J., Abrajano, T.A., Comet, P.A., Brooks, J.M., Sassen, R. and I.R. MacDonald. 1993. Gulf of Mexico
hydrocarbon seep communities XI. Carbon isotopic fractionation during fatty acid biosynthesis of seep
organisms and its implications for chemosynthetic processes, Chemical Geology, 109, 271-279.
Freedman, P.A., Gillyon, E.G.P. and E.J. Jumeau. 1988. Design and application of a new instrument for
GC-isotope ratio MS, American Laboratory, June, 114-119.
Mickey, C.W., Roper, D.S., Holland, P.T. and T.M. Trower. 1995. Accumulation of organic contaminants in two
sediment-dwelling shellfish with contrasting feeding modes: deposit- (Macomona liliana) and filter-feeding-
(Austrovenus stutchburyi), Archives of Environmental Contamination and Toxicology, 29, 221-231.
Hites, R.A., LaFlamme, R.E., Windsor, Jr., J.G., Farrington, J.W., and W.G. Denser. 1980. Polycyclic aromatic
hydrocarbons in an anoxic sediment core from the Pettaquamscutt River (Rhode Island, USA), Geochim.
Cosmochim. Acta, 44, 873-878.
Jarman, W.A., Hobson, K.A., Sydeman, W.J., Bacon, C.E., and E.B. Mclaren. 1996. Influence of trophic position
and feeding location on contaminant levels in the Gulf of Farallones food web revealed by stable isotope
109
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
analysis, Environmental Science and Technology, 30, 654-660.
Jarman, W.M, Hilkert, A., Bacon, J.W., Ballschmoter, K. and R.W. Risebrough. 1998. Compound specific carbon
isotopic analysis of Aroclors, Clophens, Kaneclors, and Phenoclors, Environmental Science and Technology,
32, 833-836.
Kelley, C.A., Hammer, B.T. and R.B. Coffin. 1997. Concentrations and stable isotope values of BTEX in
gasoline-contaminated groundwater, Environmental Science and Technology, 31, 2469-2472.
Killops, S.D. and V.J. Killops. 1993. An introduction to organic geochemistry, Longman Scientific and Technical,
265 pp.
Killops, S.D. and J.W. Readman. 1985. HPLC fractionation and GC-MS determination of aromatic hydrocarbons
from oils and sediments, Organic Geochemistry, 8, 247-257.
LaFlamme, R.E. and R.A. Hites. 1978. The global distribution of polycyclic aromatic hydrocarbons in recent
sediments, Geochim. Cosmochim. Acta, 42, 289-303.
Law, R.J. and J.L. Biscaya. 1994. Polycyclic aromatic hydrocarbons (PAH)-problems and processes in sampling,
analysis and interpretation, Marine Pollution Bulletin, 29, 235-241.
Macko, S.A., Uhle, M.E., Engel, M.H. and V. Andrusevich. 1997. Stable nitrogen isotope analysis of amino acid
enantiomers by gas chromatography/combustion/isotope ratio mass spectrometry, Analytical Chemistry, 69,
926-929.
Mansuy, L, Philp, R.P. and J. Allen. 1997. Source identification of oil spills based on the isotopic composition of
individual components in weathered oil samples, Environmental Science and Technology, 31, 3417-3425.
O'Malley, V.P., Abrajano, Jr., T.A., and J. Hellou. 1994. Determination of the 13C/12C ratios of individual PAH
from environmental samples: can PAH sources be apportioned?, Organic Geochemistry, 21, 809-822.
Roper, J.M., Cherry, D.S., Simmers, J.W. and H.E. Tatem. 1997. Bioaccumulation of PAHs in the Zebra mussel
at Times Beach, Buffalo, New York, Environmental Monitoring and Assessment, 46, 267-277.
RECENT DEVELOPMENTS IN IMMUNOBIOSENSORS & RELATED TECHNIQUES
FOR THE DETECTION OF ENVIRONMENTAL POLLUTANTS
M. Masila, H. Xu, E. Lee, O.A. Sadik*
Department of Chemistry, State University of New York at Binghamton,
P.O. Box 6016, Binghamton, New York 13902
Fax: (607)777-4478
E-mail: osadik@binghamton.edu
ABSTRACT
This paper describes inimunobiosensors and other related multianalyte detection methods of identification and
quantitation of various environmental compounds and metabolites. It presents the synthesis and characterization
of polymers, biological conjugates and metabolites for the detection of a range of toxic chemicals, including
chlorinated phenols, s-triazine herbicides, polychlorinated biphenyls, and heavy metals. A new multi-analyle
detection technique utilizing polypyrrole derivatives was developed for the detection of chlorinated phenols and
other organics. The detection of polychlorinated biphenyls (PCBs), volatile and semivolatile, halogenated organic
compounds of environmental interest was conducted using the new polymer sensors. A new promising approach
for heavy metal detection is described that utilizes o-hydroxypyridylazo metal-protein conjugates which may also
be used in the development of nonradioactive immunosensor labels for environmental compliance monitoring
and clinical applications.
INTRODUCTION
Very few analytical methods for environmental monitoring that are fast, low-cost and continuous are currently
available. The monitoring of residue or contamination in soil, water and air can be classified into two main
categories. These are: (i) screening or diagnostic techniques in which only a yes-or-no (qualitative) answer is
required, and (ii) semi-quantitative or quantitative techniques in which the detection of unwanted chemicals, and
the testing of whether or not the residues of the contaminants are within permissible levels are required. It is
possible for the former methods to generate false positive or negative results if the sensitivities are insufficient
for the detection of the threshold levels.
110
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The successful coupling of suitable transducers (e.g. electrochemical, optical or mass) with biomolecular
systems is of great importance in the search for novel sensing technologies that are inexpensive, highly selective
and capable of generating early warning signals in the presence of toxic environmental chemicals. The
emergence of on-site chemical and immunochemical biosensors for environmental pollutant monitoring has
tremendous potentials due to their small size, low costs and the case of analytical signal generation in real time.
These sensors represent a step forward over the conventional laboratory analytical methods.
In recent years, biosensors; that can detect a range of analytes in environmental samples have been
developed12 Basically, a sensor consists of a chemically selective layer, a transducer, and a signal processor. If
the selective layer utilizes a biological or biochemical species, then it can be classified as a biosensor. Thus, an
immunosensor is a subset of biosensor since it comprises either an antibody or an antigen. Each sensor has a
number of desirable characteristics depending on its applications. Essentially, a practical biosensor for the
monitoring of environmental pollutants must be specific, reversible, able to provide fast response time, and
capable of direct detection of an immunoreaction with minimal sequential addition of immunoreagents. Also, the
sensor should be capable of continuous flow measurements and capable of determining multiple analytes in
complex samples with little or no need for sample preparation steps. Finally, the sensor must be able to process
signals, or capable of being integrated into other devices that can exercise real-time feedback as required for
pollution monitoring or surveillance studies. Although, a number of pollutant measurement techniques have been
reported, only few possess these specific requirements.
One of the major objectives of our research is to develop field-portable sensors that meet or exceed the above
sensor requirements for use in the assessment of toxic chemical residues in various environmental media. This
paper discusses sensors developed in our laboratory for the identification and quantitation of environmental
contaminants.
MATERIALS AND METHODS
Instrumentation
The following instruments were used to conduct the experiments described in this paper: A Hewlett-Packard
Diode-array UVA/is spectrophotometer was used for the characterization of all protein conjugates. ELX 800 UV
Plate Reader (from Bio-Tek Instruments) was used for all of the enzyme-linked immunosorbent assay (ELISA)
experiments. EG&G PAR potentiostat/galvanostat Model 263A and EG&G 270 software were employed for the
electrochemical experiments with silver/silver chloride reference electrode, platinum wire counter electrode and
gold (A = 0.2 cm2) as working electrode. Quartz crystal microbalance (QCM) measurements were carried out
using EG&G quartz crystal analyzer (Model QCA917). A 9MHz EG&G At-cut quartz crystals was sandwiched
between two gold electrodes (A = 0.186 cm2). AromaScanner Model A32S (from AromaScan, Inc., NH) was used
for the multiarray electronic nose experiments.
Sensor Preparation and Characterization
Immobilization on Quartz: The Au-coated quartz crystal was initially pretreated by cycling the potential between
1.4 and 0.0V for a minimum of 15 minutes in 0.2M perchloric acid. The cell was then rinsed with copious amount
of water, and one surface of the crystal was soaked in a 0.02 M cystamine solution. The Au surface was
thoroughly rinsed with water to remove any physically adsorbed cystamine before being soaked in 3mM
cyanazine hapten solution containing 0.01 M HEPES (N-[2-Hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid)
buffer solution (pH = 7.3) using 10mM 1-ethyl-3-(3-dimethylamino-propyl) cabodimide EDO coupling reagent.
Electrochemical Immobilization: The Au electrode was pretreated as described above before being modified with
the cyanazine hapten using EDC as the coupling reagent. The modified electrode was used in the
electrochemical analysis, first without soaking in an antibody solution. Later the electrode was incubated in an
anti-cyanazine antibody solution at 35ฐC using a thermostated water-bath. All cyclic voltammetry experiments
were conducted at the same temperature. The other electrochemical immobilization procedures were as recently
reported34
Polymer Synthesis: Various pyrrole derivatives were polymerized by electrochemical oxidation to enable the
conducting polymer films to be used for conductivity, electrochemical, and mass measurements. Some selective
electrodes for phenols, PCBs and s-triazines were prepared by the electropolymerization of pyrrole onto platinum
electrodes in the presence of tetrabutyl ammonium perchlorate. The selectivities were comparable to a range of
structurally similar organic compounds, including 2,3,5,6-tetra chloroanisole, 2,3,4-trichloroanisole,
2-chloroanisole, 2,4,5-trichlorophenol, simazine, cyanazine, and substituted benzenes.
111
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Antibody Production: The design and preparation of analyte analogs and immunogens are essential steps in the
development of low molecular weight immunosensors. Triazine analogs were obtained from Dr. J.R. Fleeker's
laboratory. These were prepared from active esters of the carboxylic acid analog of the triazine haptens using
N-hydroxysuccinamide5 The triazines were coupled to a high molecular weight carrier, bovine serum albumin
(BSA), or keyhole limpet hemocyanin (KLH) which were used for the production of the antibodies. The antibodies
were purified by gel filtration and protein-A immunoaffinity columns, and were subsequently characterized using
ELISA and nuclear magnetic resonance (NMR) techniques. By using these antibodies, sensors for s-triazine were
developed based on the antibody inhibition of the current generated by the ferricyanide mediator on
antigen-immobilized gold electrodes.
Pesticide Immunosensors
Several pesticides and herbicides are routinely used to improve crop harvesting and pest-control. Due to the
growing concern about health effects, several investigations have been conducted in order to understand how
pesticides and herbicides degrade in the environment67 Current methods of monitoring pesticides include liquid
chromatography and gas chromatography with mass spectrometry. The high costs and labor involved in the
chromatographic methods have led to the search for low-cost alternatives capable of providing rapid analysis. In
this paper, we report on the development of immunosensors for atrazine, cyanazine, simazine and their
metabolites. The sensor chemistries are shown in Scheme 1 below:
NH-C(CH3)3
x-^NH2 HEPES buffer, ฐ-ฐ' M. pH=7.4 * EDC
Au Au
s-VN-C
!TpI
Scheme 1. The assembly of a cyanazine hapten monolayer on Au electrode.
The hapten monolayer electrode sensor assembly was used in the detection of cyanazine in a flow injection
analysis mode. The interaction of the electrode with different antibody concentrations resulted in the formation of
an antibody-antigen (Ab-Ag) complex which insulated the electrode towards the [Fe(CN)6]47Fe(CN)6]3" redox
probe, hence resulting in no charge transfer. The extent of the insulation depended on the antibody concentration
and the time of exposure to the antibody solution. The decrease in the amperometric response of the antigenic
monolayer to corresponding antibody solution for a fixed time produced a quantitative measurement of the
antibody concentration (Figure 1). Typical voltammetric responses obtained for the cyanazine hapten monolayer
electrode to different antibody concentrations are shown in Figure 2. The lowest detection limit achieved for the
cyanazine sensor was 4.0 ug/ml with a response time of a few minutes and a less-than 2% cross-reactivity to
atrazine, simazine and other metabolites.
Multianalyte Sensors
The presence of halogenated organic compounds in the environment has posed a great concern due to their
persistent toxicity and the ability to bioaccumulate. Of all the 19 known chlorinated phenols, the most important
congeners include the 2,4-Dinitrophenol (2,4-D), 2,4,5-trichlorophenol, (2,4,5-TCP) and pentachlorophenol. While
these compounds can be determined using mass spectrometry and gas chromatographic techniques, the
structural similarities of substituted phenols and their derivatives enable the development of a rapid, multianalyte
method rather than for one or two analytes. The combination of gas sensor arrays and pattern recognition
techniques has resulted in a fast and objective method for the simultaneous measurement of a wide range of
volative and semi-volative organics.
A 32-array conducting polymer sensor was used for the rapid measurement of volatile and semivolatile
halogenated organic compounds of environmental interest. The mathematical expressions for the microscopic
polymer network model was described in a recent article4. A classical, nonparametric, and unsupervised
technique of cluster analysis was used to discriminate between the polychlorinated organic phenol vapor
112
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
response vectors in a 2-dimensional space, and to identify clusters, or groups, to which unknown vectors were
likely to belong. Consequently, the characteristic pattern for each sample was generated. The pattern was used
to generate the database employed in the determination of the Euclidean distances between two given patterns
and the normalized sensor response. Also, this was used to develop the 2-dimensional mapping from a
multi-dimensional space to quantify the distinctions of the samples.
The resulting sensor arrays were found to recognize small molecules on the basis of their chemical structures
which were related to the nature of the chemical class, the type and the position of the functional groups. Each
sensor responded in varying degrees to chlorinated organic molecules with standard deviation of less than 0.05.
The time averages for the sensor response databases, datamaps, response patterns, and the intensity profiles
were obtained for different phenols. Tables 2 and 3 showed the representative databases obtained for
2,4,6-trichlorophenol and 2-CP created from the raw sample data files by selecting sample data between 60 and
120 sec. The limit of detection obtained for 2,4,6-trichlorophenol and 2-chlorophenols using the conducting
polymer sensor array were 0.1 and 0.25 ng/mL respectively. This results demonstrated the viability of conducting
polymer sensor arrays for the identification and quantitation of chlorinated organic phenols based on the
differences in their Euclidean distances. The qualitative differences as defined by the Euclidean distance
measurements were most clearly visible when the nature and the type of the functional groups were considered.
Direct Electrochemical Sensors for Polychlorinated Biphenyls (PCBs)
A direct electrochemical immunosensor has been developed for the determination of PCBs in water. The assay
was based on the measurement of the current due to the specific binding between PCB and anti-PCB
antibody-immobilized conducting polymer matrix. The linear dynamic range of the immunosensor was between
0.3-100 ng/mL with a correlation coefficient of 0.997 for Aroclor 1242. A typical flow injection analysis signal
obtained for Aroclor 1254 is shown in Figure 3. Well defined responses were recorded for all aroclors. The
method detection limits for Aroclors 1242, 1248, 1254 and 1016 were 3.3, 1.56, 0.39, and 1.66 ng/mL
respectively, and a signal-to-noise (S/N) ratio of 3. The immunosensor exhibited high selectivity for PCBs in the
presence of potential interference such as chlorinated anisoles, benzenes and phenols. The highest
cross-reactivity measured for chlorinated phenolic compounds relative to Aroclor 1248 was less than 3%. The
recoveries of spiked Aroclors 1242 and 1254 from industrial effluent water, rolling mill and seafood plant
pretreated water at 0.5 and 50 ng/mL ranged from 103-106%. The effect of ionic compounds on the detection
indicated that no significant change in immunosensor signal was observed within the uncertainty of the assay
procedure. The detection method can be used for continuous monitoring of effluent such as waste streams and
ground water.
Rational Design of Immunosensors: Sensors for Heavy Metals
In order to increase the sensitivity of immunosensing methods, a rational design of sensors using
o-hydroxypyridylazo compounds was explored. The two most important of these compounds employed were
1-(2-pyridylazo)-2-naphthol (PAN) and 4-2-pyridylazo resorcinol (PAR). Both PAR and PAN have been used
extensively for the analysis of metals, and they posses lots of useful spectroscopic and luminescence properties.
The use of 2-pyridylazo compounds as precursors for the preparation of protein conjugate by coupling the ligand
to BSA, KLH, and ovalbumin was considered. It was anticipated that using these conjugates would lead to the
development of new biosensing chemistries and transduction principles. Ultimately, any protein conjugate
developed may become useful in developing novel non-radioactive molecular labels for immunoassay,
molecular labeling and environmental compliance monitoring applications. The metal-chelate conjugates were
tested to determine if the system was simpler and rapid for the identification and quantitation of lead and other
heavy metals.
Finally, the PAR-lead-BSA, PAR-lead-Ovalbumin, and PAR-lead-alkaline phosphatase enzymes were
successfully designed and synthesized8 These conjugates were characterized using UV/Vis, intra-red
spectroscopy, NMR, and electrochemical techniques. Figure 3 shows the absorption spectrum obtained for the
coupling of PAR (510 nm) and BSA (280 mn) conjugate. A preliminary test of the PAR-conjugates and the
detection of lead and mercury were conducted using optical, differential pulse voltammetry and anodic stripping
voltammetry techniques. The binding strategies employed include a sandwich configuration using the
synthesized PAR-protein conjugates.
Pb2+ binding was monitored by recording the change in the cathodic reduction of the ion and the absorbance of
the lead-PAR chelate. The binding affinity was controlled by an electro-optical technique which influenced the
113
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
PAR-chelate's geometry. However, it is our understanding that no published literature exists on protein that
utilizes PAR metal-ion binding and the chromophoric PAR-protein conjugates. Using the electro-optical
techniques it became possible to quantitatively determine mercury and lead at approximately 3x1 Q-7 and
1x10*M respectively. Thus, a good knowledge of the selective interaction of the conjugates with biological
macromolecules may result in new applications of metal-ion immunoselective adsorbents.
CONCLUSIONS . , ,
We have developed various sensors for the detection of pesticides, PCBs and heavy metals for environmental
monitoring Other sensors developed and their analytical characteristics are summanzed in Table 3. The new
metal-chelate protein conjugates reported in this manuscript could lead to the development of new immunoassay
formats and instrumentation capable of providing rapid and selective environmental monitoring. Details of the
design synthesis and characterization of the conjugates and their analytical applications for metal detection are
being compiled for journal publication. This work demonstrates that new and promising applications of the
chemical and immunobiosensors and the emerging immunoassay labels will continue to make immunochemical
methods more valuable to environmental monitoring.
ACKNOWLEDGMENTS ^^ AJ
The authors would like to acknowledge the following: Office of Research & Development, NCERQ Advance
Monitoring Program of the United States Environmental Protection Agency for funding. AromaScan Inc., NH, for
loaning their AS 32A Electronic Nose instrument used in the evaluation of the chlorinated organic sensors
developed in this project, and Dr. James R. Fleeker of North Dakota State University for the s-triazine analogs
used in the pesticide sensors.
REFERENCES
1. Sadik O.A., Van Emon J.M., Designing Immunosensors for Environmental Monitoring, ChemTech, Vol. 27
No. 6, pp. 39 - 46, June 1997.
2. Sadik, O.A., Van Emon J.M., Biosensors & Bioelectronics, Vol 11 (8), AR1, 1996.
3. Bender Sharin, Omowunmi Sadik, "Direct Electrochemical Immunosensor for Polychlorinated Biphenyls
(PCBs)," Environmental Science & Technology, Vol. 32, No.6, pp. 788-797, 1998.
4. M. Masila, A. Sargent, F Van, O.A. Sadik, "Pattern Recognition Studies of Chlorinated Organic Compounds
Using Polymer Sensor Arrays," Electroanalysis, Vol 10, No.4, pp. 1-9, 1998.
5. Lawruk T.S., Hottenstein C.S., Fleeker J.R., Rubio F.M., Herzog D.P., ACS Symposium Series No. 630,
Herbicide Metabolites in Surface Water and Groundwater, (M.T., Meyer and E.M., Thurman - Editors) pp
43-52,1996.
6. Roy-Keith Smith, Handbook of Environmental Analysis, 2nd Edition, Genium Publishing Corp., USA, 1995.
7. Barnett D., Laing D.G., Skopec, S., Sadik O.A., Wallace G.G., Analytical Letters, 27(13), 2417, 1994.
8. Hongwu Xu, E. Lee, S. Bender, O.A. Sadik, "Immunosensors Based on Metal-Chelates for Monitoring Heavy
Metals," PittCon 98, New Orleans, Louisiana, March 1-5, 1998, paper #1244.
9. Sadik O.A., Wallace G.G., Anal. Chim. Acta, 279, 209, 1993
10. Sadik O.A., John M. J., Wallace G.G., Barnett D., Clarke C., Laing D., Analyst, 119, 1997, 1994.
11. Anita Sargent, Omowunmi Sadik, "Pulsed Electrochemical Technique for Monitoring Antibody-Antigen
Reactions at Interfaces," Trends in Analytical Chemistry, 1998 (In Press).
12. Sadik O.A., Wallace G.G., Electroanalysis, 5, 555, 1993.
Predicted Y
10 20 30
Ab cone. (uL/mL)
40
Figure 1. Cathodic peak current versus
antibody concentration at constant
incubation time of 5 minutes.
114
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1. Representative database obtained for 2,4,6-Trichlorophenol at conducting polymer sensor arrays
Sensor
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Intensity
-2.560
-2.450
-2.710
-2.700
-2.540
-2.460
-2.730
-2.600
-2.490
-2.700
-2.320
-2.450
-2.310
-2.720
-2.310
-2.260
-2.150
-2.500
-2.720
-2.790
-2.570
-2.170
-2.300
-2.120
-2.070
-2.440
-2.520
-2.350
-2.430
-2.700
-2.810
-2.570
Pattern
-3.219
-3.079
-3.407
-3.392
-3.198
-3.096
-3.428
-3.269
-3.132
-3.393
-2.920
-3.080
-2.905
-3.426
-2.902
-2.839
-2.709
-3.145
-3.424
-3.508
-3.238
-2.728
-2.892
-2.659
-2.602
-3.068
-8.175
-2.957
-3.053
-3.398
-3.530
-3.229
SD
0.038
0.003
0.021
0.024
0.037
0.033
0.009
0.013
0.005
0.004
0.014
0.008
0.017
0.007
0.041
0.036
0.070
0.013
0.030
0.006
0.056
0.110
0.109
0.194
0.010
0.012
0.028
0.009
0.039
0.037
0.029
0.054
SD = Standard deviation, experimental conditions.
30 min.
relative humidity 50%, temperature 25ฐC, equilibration time
Figure 2. Voltammetric responses obtained
for 1.1 mM K4Fe(CN)6 with cyanazine hapten
monolayer electrode using different
concentrations of anti-cyanazine antibody at
5-minute incubation time interval, (a) Blank
(phosphate buffer), (b) 5ML/ml; (c) 15 |jl_/ml;
(d) 30gL/ml; (e) 40ul_/ml.
-400 -200 0 200
E (mV) vs. Ag/AgCI
400
600
115
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 2. Representative database obtained for 2-Chlorophenol at conducting polymer sensor arrays
Sensor
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Intensity
-0.21
-0.08
-0.15
-0.15
-0.17
-0.18
-0.08
-0.12
0.11
-0.05
-0.08
-0.09
0
-0.09
-0.17
-0.18
0.32
0.46
0.02
0.36
-0.24
0.6
0.73
0.81
-0.04
-0.03
-0.11
-0.01
-0.17
-0.11
-0.09
-0.24
Pattern
-3.63
1.26
-2.58
-2.59
-2.94
-3.1
-1.56
-2.06
-1.66
-0.97
-1.36
-1.56
0.1
-1.62
-3.09
-3.16
4.77
6.05
0.43
5.44
-4.31
8.45
10.09
11.67
-0.59
-0.64
-1.96
-0.15
-3
-2.07
-1.65
-4.22
SD
1.11
0.51
0.85
0.8
0.86
0.99
0.58
0.59
0.66
0.62
0.37
0.5
0:090
0.55
1.03
1.04
1.03
4.58
0.1
1.51
1.49
3.39
4.38
3.26
0.16
0.42
0.61
0.14
0.9
0.72
0.47
1.45
SD = Standard deviation, experimental conditions: relative humidity 50%, temperature 25ฐC, equilibration time
30 min.
Table 3. Immunobiosensors & Chemical Sensors Developed
Compound
Chlorinated
Phenols
PCBs
Cyanazine
Heavy metals
Cyanide
Anions
Atrazine
HSA
P-Cresol
P-Athau
* MDL = Method
MDL
0.25 ug/ml
0.05 ng/ml
4.0 ug/ml
low ppt
3X10-8M
4.0 ng/ml
3X10'7M
0.5 mg/l
0.01 mg/ml
Detection Limit. The MDL
Detection TechniquelRemarks
AR
FIA mode
A[Ab]
absorbance measurement
Fluorescence
Current
A[Ab]
Regenerate
Dynamic range 3 orgers of magnitude,
reusable
Regenerable
was computed using MDL = t(n-ri^=o99i *
value, S = standard deviation of the replicate analyses.
b Currently under
investigation.
Ref.
2
1-3
b
b
b
12
b
9
7
10
S, where t = the students t
116
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Figure 3. UVA/is Spectrum recorded for the coupling of
BSA (labeled A at 280 nm) and PAR (labeled B at 510 nm)
conjugate.
MULTIPLEXED DIODE LASER GAS SENSOR SYSTEM
FOR IN-SITU MULTI-SPECIES EMISIONS MEASUREMENTS
R. Hanson
NO ABSTRACT AVAILABLE
OVERVIEW/FUTURE OF NCERQA RESEARCH PROGRAM
Bala Krishnan
US EPA, Office of Research and Development, 401 M St, SW, Washington, DC
WO ABSTRACT AVAILABLE
ADVANCED ANALYTICAL METHODS FOR THE DIRECT QUANTIFICATION
AND CHARACTERIZATION OF AMBIENT METAL SPECIES IN NATURAL WATERS
Janet G. Hering and Joon H. Min
California Institute of Technology, Environmental Engineering Science,
1200 E. California Blvd. (138-78), Pasadena, CA 91125
Neither the biogeochemical cycling of trace metals nor their ecotoxicological effects can be fully understood
without careful consideration of metal speciation. Investigations of metal speciation in natural waters have
demonstrated that, for many metals (particularly Cu, Zn, and Fe), the predominant ambient metal species are
nonreactive in chemical and/or biological assays. However, the methods commonly used to monitor metal
speciation cannot provide definitive information on the nature of ambient metal species.
Electrospray-mass spectrometry (ESMS) offers a powerful tool for the investigation of ambient metal species.
Unlike more common GC-MS techniques, ESMS can be applied to non-volatilizable species and the
comparatively gentle ionization in ESMS allows compounds to be analyzed with minimal fragmentation thus
117
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
preserving the molecular ion signature.
Preliminary studies with model compounds have been performed to investigate the application of ESMS for
analysis of metal-organic complexes. Work to date with the strong organic complexing agent EDTA
(ethylenediaminetetraacetic acid) has demonstrated that both uncomplexed EDTA and metal-EDTA complexes
can be detected in the positive ion mode as protonated species with a single positive charge. Despite the
extreme conditions imposed by the electrospray interface, protonation (and, in the case of uncomplexed EDTA,
formation of the Na+ adduct) appears to be the only perturbation of the initial distribution of EDTA species.
Molecular ions were detected for EDTA and its complexes with Cu, Pb, Cd, Al, and Fe(lll). In the case of the
Pb-EDTA complex, the isotopic signature of Pb was also observed. Based on this work, the application of ESMS
for the detection of ambient metal species in natural waters appears promising but is limited by the sensitivity of
the technique and by variable response to different compounds. Detection limits for EDTA and its metal
complexes are approximately micromolar though better sensitivity has been reported for other metal-organic
complexes.
Work is in progress to characterize the ESMS response to organic ligands of varying structures (and their metal
complexes) and to improve the sensitivity of the technique. The instrument currently in use is a Hewlett Packard
Series 1100 LC/MSD; adjustable instrumental parameters include drying gas flow rate and temperature, capillary
and fragmentor voltages, gain and nebulizer pressure. The effects of the sample matrix (e.g., pH and methanol
concentration) on sensitivity will also be tested for selected model compounds. Model compounds have been
selected to include a range structural characteristics including: type of heteroatom(s) and complexing
functionalities, metal-ligand stoichiometry, and ligand charge and hydrophobicity. Preliminary, screening studies
will be performed to examine ESMS response to reference humic and fulvic acids.
RADICAL BALANCE IN URBAN AIR
Robert J. O'Brien
Chemistry Department
Linda A. George
Center for Science Education
Thomas M. Hard
Chemistry Department, Portland State University, PO Box 751, Portland, Oregon 97207
Atmospheric free radicals hydroxyl and hydroperoxyl (OH and HO2, collectively HOX) are the catalysts which
cause secondary or photochemical air pollution. Chemical mechanisms for oxidant and acid formation, on which
expensive air pollution control strategies are based, must accurately predict these radical concentrations. We
used the PAGE technique to carry out the first simultaneous, in-situ, measurements of these two radicals in
highly polluted air at downwind sites in the Los Angeles airshed.
To compare the measured OH and HO2 concentrations with photochemical models, a complete suite of
simultaneous ancillary measurements was necessary, and was obtained during each measurement campaign.
The suite included speciated hydrocarbons, carbonyl compounds, carbon monoxide, nitric oxide, nitrogen
dioxide, ozone, and meteorological parameters. With this suite as input, we tested the ability of a lumped
chemical mechanism to accurately predict the measured OH and HO2 radical concentrations.
Due to the short photochemical lifetime of HO* (less than 1 minute), this test of radical balance in urban air
depends directly and quantitatively on the measured parents and reaction partners of the radicals, and only
indirectly on the upstream history of the sample.
Results of the measurements, and of the radical balance tests, will be presented, with acknowledgments to the
organizations and scientists who provided assistance.
118
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
ENVIRONMENTAL APPLICATIONS OF NOVEL INSTRUMENTATION FOR MEASUREMENT OF LEAD
ISOTOPE RATIOS IN ATMOSPHERIC POLLUTION SOURCE APPORTIONMENT RESULTS
Keeler
NO ABSTRACT AVAILABLE
REMOTE SAMPLING PROBE WITH FAST GC/MS ANALYSIS:
SUBSURFACE DETECTION OF ENVIRONMENTAL CONTAMINANTS
Albert Robbat. Jr.
Center for Field Analytical Studies & Technology, Chemistry Department,
Tufts University, Medford, Massachusetts 02155
ABSTRACT
This paper describes the results of an in situ sampling probe that is capable of thermally extracting volatile and
semivolatile organics bound to soil from depths of 30 ft. The organic vapor is swept to the surface by an inert
carrier gas and trapped (volatiles) or condensed (semivolatiles) in appropriate sampling tubes. The organics are
subsequently thermally desorbed into a gas chromatograph/mass spectrometer and analyzed in under 5 minutes.
INTRODUCTION
The EPA estimates that the cost for hazardous waste site cleanups will exceed $300 billion over the next 10
years,1 with the cost for Superfund alone exceeding $26 billion since 1980, The following questions can be
posed: Do inadequate site investigations and, therefore, a lack of understanding with respect to the chemical and
physical dynamics affecting the cleanup contribute to these costs? Can field-based analytical instrumentation
and methods give on-site project engineers the kind of data needed that will lead to faster, better, and cheaper
cleanups?
Toward this end, research is leading to the development technology and methods that can produce quantitative
analysis of environmental contaminants in minutes by thermal desorption gas cnromatography/mass
spectrometry (TDGC/MS). The analysis is based on a ballistically heated thermal desorber to achieve large
volume sample introduction and mass spectral data analysis algorithms that can "look through" complex matrix
signals to identify and quantify target compounds.2 The TDGC/MS when used in a dynamic workplan framework
can provide data fast enough to influence the on-site decision making process.3 We have shown that on-site
chemical analyses employing dynamic workplans can reduce the time and cost of hazardous waste site
investigations.4
Cone Penetrometer (CP) systems can collect samples at much faster rates than can traditional drilling rigs.
Figure 1 depicts the sampling probe used to collect subsurface soil and water samples. Typically, 5 cm o.d. pipes
are threaded together and pushed underground by truck weights of up to 40 tons. The challenge therefore, is to
design 1) a flexible heated, 300 ฐC, transfer line that can be woven through each pipe section and 2) a
programmable thermal extraction sample collection probe that can heat the soil to at least 350 ฐC. These target
temperatures are based an past studies aimed at developing direct measuring thermal desorption gas
chromatography (TDGC) sample introduction system.56789 The design of a thermal extraction cone
penetrometer (TECP) system for subsurface sampling of soil-bound organics from depths of up to 25 m is
described in this paper as well as a new data analysis system that provides unique compound identification and
quantification capabilities under fast TDGC/MS conditions.
MATERIALS
Heated Transfer Line
The materials used to fabricate the heated transfer line include: deactivated fused silica lined stainless steel
tubing 30 in x 1 mm, i.d. 0.76 mm Silcosteelฎ (Restec Corp., Bellefonte, PA); Nextel 312 thermal insulation
sleeving and Viton shrinkable tubing (Omega, Stamford, CT), heat shrinkable Teflon tubing (Patriot Plastics,
119
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Woburn, MA); aluminum foil with silicon adhesive backing
(COMCO), polyimid moisture insulation tape (Newark Electronics,
Chicago, IL); thermal insulated fiber glass cloth tape (Fisher
Scientific, Pittsburgh, PA). The heated transfer line is heated by
connecting high temperature power lead wires (Newark
Electronics) to both ends of the Silcosteer* tube. Temperature was
measured using Thermocouples C01-K and C02-K (Omega).
Heated Probe
The probe was made from a 1 m x 4.5 cm, 2.5 cm i.d., threaded
steel pipe. A 20 mm i.d. hole was cut in the pipe at one-third the
distance from the bottom. The heat was supplied by inserting an
aluminum casing into the pipe, which contained a 10 cm x 1.5 cm
heating cartridge L4A712, 240V/1000W. The same model
Thermocouples used to measure the transfer line temperature
were used in the Probe.
Figure 1. Cone penetrometer and thermal extraction sampling
probe.
Equipment
The Silcosteelฎ was heated by passing current through the tube using an electrical isolation step-up transformer
(Grainger, Haverhill, MA) with power and temperature controllers model DCIP-50245-FOO and model
988A-10FD-AARG, respectively (Watlow, St. Loius, MO). A Hewlett Packard (Palo Alto, CA) model 5972 mass
spectrometer was ruggedized for the field and used in combination with a Tufts University (Medford, MA)
designed thermal desorption gas chromatograph. All GC/MS total ion current chromatograms (TIC) were
acquired by HP's data acquisition system. A new mass spectrometry data analysis software developed at Tufts
was used to identify and quantify polycyclic aromatic hydrocarbons (PAHs). The Ion Fingerprint Detection
software is available from Ion Signature Technology (Cambridge, MA).
UEfr
l.OEfr
11.8
Figure 2. GC/MS analysis of a soil
sample collected from Hanscom Air
Force Base (Bedford, MA);
compounds found: 1) 1,1,1-trichloro-
ethene, 2) methylene chloride, 3)
1,1 -dichloroethene, 4) 1,4-difluoro-
benzene (surrogate), 5) toluene-da
(internal standard), 6) ethylbenzene,
7) m/p-xylene, 8) o-xylene, 9) styrene,
10) 4-bromofluorobenzene (surro-
gate).
Total Ion Chromatograra
AAJ
Reconstructed Ion Chromatogram
RESULTS
Two key breakthrough technologies have been developed that meet the EPA data quality measurement
objectives and EPA Soil Screening Level quantitation levels under fast GC/MS conditions. First is the mass
spectral data analysis software. The software extracts between three and ten characteristic fragment ions for
each targeted organic and then, based on a patented set of algorithms, compound identity and concentration are
determined. Algorithmic details can be found elsewhere.2 Although all MS systems can extract ions, they cannot
handle the amount of extracted ion information and determine compound presence using current statistical or
120
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
library matching routines when high levels of interferents are present in the sample. For example, the software
can provide compound identification in complex environmental samples without the need for extensive sample
cleanup. The second technology breakthrough is the thermal desorber. Unlike other commercially available units,
the TD can be ballistically heated from subambient temperatures to 320 ฐC in 8-sec. The TD uses a standard
Tenax tube for trapping VOCs and an empty glass sleeve for semivolatiles that have been swept from depth to
the surface by the carrier gas. Direct desorption of organics from solid materials or an organic extract into the GC
is made by placing known quantities into an empty glass sleeve.
Figure 2 illustrates a representative total ion current chromatogram (TIC) and a reconstructed ion current (RIG)
chromatogram where the data analysis software was able to "see through" the matrix. Note that the internal
standard (peak #5) and target compounds 1,1,1-trichloroethene and 1,1-dichloroethene (peaks 1 and 3) are
buried within the matrix and that their corresponding signals are 102 to 104 times smaller than the matrix signal.
Typically, analysts would dilute this sample prior to analysis based on visual inspection of the petroleum present
in the sample. This practice will result in the loss of low level target compounds such as the chlorinated solvents
found in this sample.
Figure 3 shows the TIC and RIC chromatograms produced from a 10-min TDGC/MS analysis
of a standard mixture of polychlorinated biphenyls (PCBs, Aroclor 1248), polycyclic aromatic
hydrocarbons (PAHs), chlorinated pesticides, and engine oil (25% v/v). A total of 1,000-ng
PCBs was thermally desorbed into a 15-m GC column along with 19 chlorinated pesticides
(20-ng/compound), 16 PAHs (40-ng/compound), and pyrene-dio (50-ng) added as an internal
standard. An expanded view of the RIC chromatogram between 6.7 min and 6.9 min is shown
here. Note that there are six compounds that elute within this time domain. Compound
identification was made based on a set of algorithms that extracted 3-6 fragment ions per
compound from the TIC and computed their match against standard reference spectra in
seconds. Each compound's RIC signal, based upon preselected quantitation ions, are then
used to produce the RIC chromatogram and to quantify compound concentration. The
algorithms and results will be presented documenting measurement accuracy, precision, and
sensitivity.
Figure 4 depicts the schematic of the heated transfer line and the electronic circuitry used to control the power
and temperature. The figure shows the various layers including moisture, electrical, and thermal insulation
sleeves as well as the fused silica coated stainless steel tube Silcosteelฎ The goal was to heat the transfer line to
300 ฐC and achieve a 15 cm bend radius so that the pipes in the truck could be stacked efficiently. This feature is
important since cone penetrometer systems can reach subsurface depths of up to 60 m when the geology is
amenable. Based on the design shown in the figure
a 10 cm bend radius was obtained, with the
Silcosteelฎ temperature programmable from
ambient soil temperatures up to 300 ฐC. To date, a
30 m transfer line has been made, which can be
woven through ten 1 m pipe sections to achieve
subsurface depths of 10 m. The transfer line is
heated by passing direct current through the
stainless steel tube. Imbedded in the transfer line
are the electrical and thermal couple wires needed
to carry current to the probe head and to monitor
both the transfer line and probe temperatures.
Figure 3. 10-min TDGC/MS analysis of a soil
sample fortified with a standard mixture of PCBs,
PAHs, pesticides and gasoline/engine oil (1:3 by
vol).
Reconstructed Ion Current
Total Ion Current
121
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
collection gas
Figure 4. TECP system, which includes 1)
heat shrinkable sleeve, 2) fiber glass
insulation, 3) high temperature electrical
wires, 4) Viton shrinkable tubing, 5) heat
shrinkable Teflon tubing, 6) Nextel 312
thermal insulation sleeving, 7) deactivated
fused silica lined stainless steel tubing 30
m x 1 mm, i.d. 0.76 mm Silcosteelฎ 8)
aluminum foil with silicon adhesive
backing, 9) polyimid moisture insulation
tape, 10) thermal insulated fiber glass
cloth tape, 11) polyimid moisture
insulation tape, 12) thermocouple wire,
13) isolation step-up transformer, 14)
temperature and power controllers, 15)
transfer line thermocouple, and 16) high
temperature electrical wires.
Table 1. Material Balance of Closed Chamber
(Rem = 5700; V0 = 32; wg = 0.6 m/s; dry sand)
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluorarthene
Benzo(a)pyrene
Dibenz(a,h)anthracene
lndeno(1 ,2,3-ed)pyrene
Benzo (g,h,i)perylene
250 ฐC
% Rec
5
61
74
72
76
75
79
86
56
50
41
41
65
NA
NA
NA
% Soil
ND
ND
ND
ND
5
6
5
9
42
40
53
53
31
NA
NA
NA
% Ads
ND
ND
3
2
2
3
3
2
2
2
6
6
4
NA
NA
NA
300 ฐC
% Rec
15
63
78
75
83
78
80
83
67
61
55
55
65
37
37
23
% Soil
ND
ND
ND
3
5
6
4
6
30
31
40
40
23
59
59
67
% Ads
ND
ND
ND
0.6
2
2
3
1
1
1
4
4
3
2
2
4
ND - not detected; NA - not analyzed
Little degradation of the silica lining is observed as long as air is purged from the system. Air is flushed from the
transfer line by nitrogen prior to transfer line heating. After the Silcosteelฎ has been conditioned the gas valve is
switched to re-direct nitrogen into the carrier gas line. At this point, a vacuum pump is turned on to collect the
soil/organic vapor at the collection window and to transport the organics to the surface through the Silcosteel*
tube. The valve can then be re-positioned to cleanse the transfer line tube when high levels of sample are
collected. This step is intended to eliminate sample carry over from one sample location to the next. Work is in
progress to automate the TECP system to control the probe and transfer line temperatures as well as the carrier,
flush, and collection gas lines.
Material balance experiments were conducted to determine the thermal extraction efficiency for PAHs at 250ฐC
and 300ฐC, see Table 1. At optimum conditions, i.e., soil temperature, carrier gas flow rate, collection volume,
and under closed cell conditions, greater than 55% recovery was obtained. Closed cell conditions represent the
maximum amount that can be extracted, as opposed to the TECP, since none of the organic vapor produced is
122
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
lost to the environment and all soil is uniformly heated. The organic vapor is efficiently flushed from the cell into
Tenax for VOCs or the cold trap for semi-VOCs. Note that less than 4% of the collected organics remained in the
transfer line after collection. Research is in progress to chemically modify the surface to minimize (eliminate)
organic absorption. Nonetheless, back-flushing for 5-min reduced the percent adsorbed to non detectable levels.
Table 2. Comparison of TECP vs. Closed System
T^pe = 450 ฐC; Tsoi, = 280 ฐC; Rem = 6000; wg = 1.2 m/s; Ud = 15
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Chrysene
Benzo(a)anthracene
Benzo(b&k)fluoranthene
Benzo(a)pyrene
lndeno(1 ,2,3-ed)pyrene
Dibenz(a,b) anthracene
Benzo(g,h,i)perylene
TECP % Recovery
50-ppm
20
34
41
49
37
37
65
64
27
49
29
25
26
26
17
25-ppm
37
36
39
36
32
32
53
53
37
37
27
23
19
19
16
1 5-ppm
26
36
33
38
35
34
48
45
31
36
29
36
20
19
16
5-ppm
37
33
30
37
34
34
36
68
41
41
23
24
23
23
16
Ave Rec
30
35
36
40
35
34
50
47
34
40
27
27
22
22
10
% RSD
23
5
18
17
7
7
34
15
16
12
12
26
14
14
4
Closed
Chamber
% Recovery
18
62
77
73
80
77
80
85
57
62
48
65
65
62
20
%
Diff
68
44
53
46
57
56
37
32
40
36
44
58
31
32
19
Note: % Difference is between Average TECP Recovery and Closed Chamber Recovery
The TECP and closed cell data comparison study is shown in Table 2. Recall that the closed system represents
maximum recovery of analyte while the TECP is what one should expect to achieve in the field. The TECP
measurement precision determined over an order of magnitude for PAHs is excellent. The results are
remarkable and are as good or better than what is achievable through soil/solvent extraction. The accuracy for
the TECP approximates one-half that of the closed (ideal) system. Results will be presented illustrating collection
efficiency as a function of soil type, probe depth and moisture content.
Table 3. TECP Field Study, Berlin Vermont
TDiDe = 450 ฐC; Tsoji = 280 ฐC; Ttmne = 280 ฐC; Rem = 6000; wfl = 1.6 m/s
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Dibenz(a,h)anthracene
lndeno(1 ,2,3-cd)pyrene
Benzo(g,h,i)perylene
TECP Measured PAH/Soil Concentration
Boring 1
97-cm
2
9
2
7
11
11
44
29
45
45
5
6
12
2
3
6
142-cm
2
6
2
3
5
5
3
4
11
11
0.4
0.3
0.6
ND
ND
ND
Boring 2
1 30-cm
4
14
3
10
5
7
6
6
6
6
0.5
0.5
ND
ND
ND
ND
155-cm
ND
ND
2
4
2
2
2
5
3
3
ND
ND
ND
ND
ND
ND
1 90-cm
ND
5
3
3
4
4
3
4
2
2
ND
ND
ND
ND
ND
ND
The TECP and heated transfer line was tested in the field employing ARA's CPT in Berlin, Vermont. The location
was a Vermont state central maintenance facility known to contain petroleum hydrocarbon contamination. The
123
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
TECP was tested for mechanical ruggedness and its ability to collect subsurface bound organics. The vertical
profile of two borings are shown in Table 3. PAHs were found at depths of 97-cm and 142-cm at boring one and
130-cm. The hot organic vapor was cooled by dry ice and collected in an empty glass sleeve attached to the
transfer line. Total collection and analysis time required at each depth location was approximately 15-min per
sample inclusive of sample collection, transfer line back flushing, and analysis time. Unfortunately, comparison
measurements between the TECP measured concentrations and actual soils collected at depth and analyzed by
standard GC/MS methods were not possible since ARA was unable to re-enter the hole with a soil sample
collection probe without breaking the CP rod. The results produced to date are promising suggesting that direct
on-line chemical measurements of subsurface contaminants may be possible within a couple years.
REFERENCES
1. Cleaning Up the Nation's Waste Sites: Markets and Technology Trends, Office of Solid Waste and Emergency
Response, U.S. EPA, Washington, DC, 1993, EPA 542-R-92-012, 164 pp.
2. Y. Gankin, A. Gorshteyn, S. Smarason, and A. Robbat, Jr., Anal. Chem, 70:_1655-1663 (1998).
3. A. Robbat, Jr., "Field Analytics, Dynamic Workplans" The Encyclopedia of Environmental Analysis and
Remediation, ed. by R.A. Meyers, John Wiley & Sons, Inc. New York, NY., July 1998.
4. A. Robbat, Jr., "A Dynamic Site Investigation Adaptive Sampling and Analysis Program for Operable Unit 1 at
Hanscom Air Force Base, Bedford, Massachusetts", U.S. Environmental Protection Agency, Region I, October
1997; see http://clu-in.com/char1.htmtfregional.
5. A. Robbat, Jr., T-Y Liu, and B. Abraham, Anal. Chem., 64:358-364 (1992).
6. A. Robbat, Jr., C. Liu, and T.-Y. Liu, J. Chromatography, 625: 277-288 (1992).
7. A. Robbat, Jr., C. Liu, and T.-Y. Liu, J., Anal. Chem., 64: 1477-1483 (1992).
8. K M. Abraham, T-Y Liu, and A. Robbat, Jr., Hazardous Waste & Management, 10: 461-473 (1993).
9. K. Jiao and A. Robbat, Jr., J. of AOAC International, 79: 131-142 (1996).
AN INTEGRATED NEAR INFRARED SPECTROSCOPY SENSOR
FOR IN-SITU ENVIRONMENTAL MONITORING
Roland A. Levy and John F. Federici
New Jersey Institute of Technology, University Heights, Newark, NJ 07102
Tel. (973)-596-3561
The monitoring of environmental organic contaminants currently involves off-site methods which prohibit optimal
usage. This study explores the possibility of combining the principles of interferometry with that of near infrared
evanescent wave absorption spectroscopy to produce a novel integrated sensor technology capable of
monitoring and determining in-situ the concentration of numerous organic analyte species simultaneously. This
novel sensor promises to be non-intrusive and to provide accurate, rapid, and cost effective data. The overall
instrument is envisioned to be compact, portable, rugged, and suitable for real time monitoring of organics. The
sensor consists of a symmetric, single mode Mach-Zehnder interferometer with one arm (sampling) that is either
exposed directly to the analyte or coated with a thin hydrophobic layer that enhances the binding of pollutant
molecules onto its surface. A glass buffer layer protects the second arm (reference) from the influence of
pollutants. Light is coupled into the waveguide and split between the sampling and reference arms using a
Y-splitter configuration. Changes in the refractive index caused by the presence of organic contaminants result in
a measurable phase difference between the sampling and reference arm. Selectivity of the sensor is achievable
by utilizing evanescent wave absorption spectroscopy in the near infrared, a technique which measures
wavelength dependent refractive index changes. The waveguide structures used in this study are fabricated on
10 cm silicon wafers. V-grooves are first formed in the silicon substrate to hold the fibers which couple to the
ends of the Mach-Zehnder interferometer. A 10 urn thick SiO2 film is synthesized by low pressure chemical vapor
deposition (LPCVD) to act as cladding material for the waveguide and prevent light from coupling with the
underlying silicon. A 4 urn thick phosphorus-doped (7.5 wt% P) LPCVD SiO2 film is then deposited to act as core
material for the waveguide. This layer is patterned using standard lithographic exposure and plasma etching
techniques and subjected to a 1050 ฐC anneal to cause viscous flow and round off the edges. This rounding-off
procedure is necessary to minimize coupling losses between fiber and waveguide. The refractive index of the
doped glass is near 1.48, thus, producing with the underlying SiO2 (n=1.46) substrate a single mode waveguide
124
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
device. Deposition of a 0.5 |jm thick LPCVD undoped SiO2 buffer layer over the entire wafer and a subsequent
lithographic step results in selective removal of that layer over the sampling arm of the interferometer. This
configuration allows for exposure of the sampling arm (uncoated or coated) to various contaminants in the
environment which cause a change in the effective refractive index of that arm. The arm coated with the SiO2
buffer layer sees a constant refractive index of n=1.46. A lithographic mask is used to produce the required
patterns. Five micron wide waveguides form the two interferometer paths using a splitting angle of 2ฐ. The
sampling and reference arms are at a fixed separation of 50 urn and variable lengths (4, 6, 8, and 10 mm). A
detailed analysis of these processing steps together with the principles of operation of the sensor will be
discussed.
A NITRIC OXIDE AND AMMONIA SENSOR ARRAY FOR FOSSIL FUEL COMBUSTION CONTROL
B.T. Marquis, J.F Vetelino and T.D. Kenny
Laboratory for Surface Science & Tech. and Department of Electrical and Computer Engineering,
University of Maine, Orono, Maine 04469
ABSTRACT
In fossil fuel combustion processes, nitric oxide (NO) emissions are minimized by a selective catalytic reduction
(SCR) technique where ammonia (NH3) is injected into the flue gas stream to react with NO to form
environmentally safe gases such as nitrogen and water vapor. Unfortunately, this process is usually incomplete,
resulting in either NO emissions or excess NH3 (NH3 slip). Therefore, a critical need exists for an in situ sensor
array to continuously monitor the NO and NH3 levels at the output of the SCR system near the stack to provide
real time control of the NH3 injection and hence minimize the NO emissions to the environment. Chemiresistive
sensor technology is being used to develop a small, portable, sensitive and selective sensor array that has
potential to continuously measure NO and NH3 emissions. The sensor array utilizes a tungsten trioxide (WO3)
film as the sensing element to simultaneously identify NO and NH3 concentrations present in the fossil fuel gas
exhaust. Several film parameters such as thickness, dopant (gold vs. ruthenium), doping method (post-sputter
vs. co-sputter), doping amount, deposition temperature, annealment procedure and operating temperature were
varied to determine their effect on the film's sensing properties. As a result, a 500A WO3:16A Au film
post-sputtered at 200ฐC demonstrated high sensitivity and selectivity to NH3 whereas a 1000A WO3 undoped film
sputtered at 200ฐC exhibited greater sensitivity to NO.
INTRODUCTION
The detection and measurement of flue gases are critical not only for achieving real time process control of new
clean combustion systems, but also to minimize their emissions of dangerous air pollutants. Among the most
dangerous of these air pollutants are nitric oxide (NO) and nitrogen dioxide (NO2), collectively referred to as NOX.
Currently about one half of all NOX emissions into the environment are due to power plants and industrial boilers.
NOX gas which is the precursor to nitric and nitrous acid, causes acid rain and photochemical smog and is a
critical factor in the formation of ozone in the troposphere. Ground level ozone is a severe irritant, responsible for
the choking, coughing and burning eyes associated with smog. Ozone often damages lungs, aggravates
respiratory disorders, increases susceptibility to respiratory infections and is particularly harmful to children.
Elevated ozone levels can also inhibit plant growth and cause widespread damage to trees and crops. Therefore,
exceeding critical NOX levels poses immediate health and environmental problems.
In fossil fuel combustion NOX is formed by high temperature chemical processes from both nitrogen present in
the fuel and oxidation of nitrogen in air. Typically, the NOX emissions consist of 90-95% NO with the remainder
being N2O and NO2.1
Several methods have been examined as potential control systems for the reduction of NO emissions in
combustion processes including combustion control and flue gas treatment techniques such as selective
noncatalytic reduction (SNCR) and selective catalytic reduction (SCR)23. SCR technology achieves the highest
overall control efficiency of 60-80% NO reduction. In this process, NH3 is uniformly injected into the flue gas
stream that passes through a catalyst bed to enhance the kinetics of the NO/NH3 reaction. This can be further
improved with air staging to provide a longer residence time allowing more NO to react with the NH3 , thus
125
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
increasing NO reduction and decreasing NH3 emissions (commonly called NH3 slip). Although there are several
chemical reactions that may take place in the catalyst bed, the most common reactions are as follows,
4 NH3 + 4 NO + O2 -> 4 N2 + 6 H2O, (1)
4 NH3 + 6 NO -> 5 N2 + 6 H2O, (2)
4 NH3 + 3 02 -ป 2 N2 + 6 H2O (3)
and
4 NH3 + 5O2 -- NO + 6 H2O. (4)
Reaction (1) is the predominant reaction1 and reduces NO into harmless nitrogen and water vapor. The reaction
given in equation (3) neutralizes some of the excess NH3 that did not fully react with the NO in equations (1) and
(2) and hence decreases the NH3 slip. An unwanted side reaction given by equation (4) oxidizes NH3 to form
water vapor and NO. This harmful oxidation process however usually takes place at elevated temperatures, and
can be minimized by temperature control. If a judicious choice of NH3 injection levels is made, NO emissions in
SCR systems, as well as NH3 slip into the atmosphere can be reduced significantly.
In order to control the precise levels of NH3 injection, a critical need exists for an NO/NH3 sensor system in the
flue gas exhaust prior to the stack emission. The output from the sensor would be fed back to the SCR system
via a real time control system to adjust NH3 injection levels thereby maximizing NO reduction with minimal NH3
slip into the atmosphere. Current NO and NH3 measurement techniques, such as chromatography,
chemiluminescence and infrared adsorption are very expensive and too bulky for in situ operation.
Chemiresistive semiconducting metal oxide (SMO) films offer the technology to develop a small, inexpensive
and reliable in situ sensor array for the simultaneous identification of NO and NH3 present in a flue gas exhaust.
The first report on Chemiresistive SMO films for gas sensing appeared in 19624 Since that time, considerable
efforts have been made5 to study SMO films for the detection of a variety of gases. Investigators have examined
SMO films such as SnO2 615, TiO215'16, indium tin oxide (ITO) 17, ZnO 1S1718 and WO319'22 for detecting NO, as
well as ZnO 2328, MoOs/TiO2 29 and WO3 3034 for detecting NH3. WO3 films can operate at elevated temperatures
for long periods of time and selectively detect NO and NH3 in the presence of interferent gases such as H2, CO,
CO2, CH4 and various other hydrocarbons. These films are also electrically and structurally stable at elevated
temperatures.
The electrical conductivity of thin WO3 films doped with metals such as gold (Au) and ruthenium (Ru), change
upon exposure to NO and NH3. The sensitivity of WO3 to NO and NH3 depends significantly upon film
parameters such as thickness, dopant type, doping method, deposition temperature, annealing procedure and
operating temperature. Furthermore, the functional relationship between sensitivity and each of these parameters
is different for both gases.
THEORY
The WO3 film conductivity changes as a function of NO and NH3 gas concentrations. These films exhibit very
fast response and recovery times and, after initial film conditioning, show no appreciable aging effects after
repeated gas exposures. The basic chemical sensing mechanism involves the dissociative chemisorption of the
target molecules and the formation of transitory concentrations of chemisorbed atoms on the WO3 film surface.
The rate of dissociation of the target molecules on the surface can be greatly enhanced by the addition of
catalytic metals such as gold or ruthenium. Although the overall chemistry of the possible interactions of NO and
NH3 with WO3 is complex and not well defined, certain primary interactions dominate.
In the case of NH3, which acts as a reducing agent (an oxygen scavenger), the carrier concentration in the film
rises as a result of a decrease in adsorbed surface oxygen, as follows,
2 NH3 + 3O'(ad)^ N2 + 3 H2O + 3 e- (5)
This rise in carrier concentration within the film is then manifested as a decrease in resistivity.
Unlike NH3, NO acts as an oxidizing agent (oxygen donor) at the temperatures of interest. Thus, reactions with
NO result in an increase in chemisorbed oxygen in the film, decreasing the free carrier concentration, as follows,
2 NO + 2 e- -* N2 + 2 O (ad). (6)
126
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
This decrease in carrier concentration causes the film resistivity to rise. This behavior is completely opposite to
that observed with NH3.
EXPERIMENTAL SETUP
The WO3 sensing films were deposited on alumina and sapphire substrates by reactively sputtering a pure
tungsten target in an 80/20 argon-oxygen atmosphere using an RF magnetron sputtering system. Alumina and
sapphire substrates were used because they are good electrical insulators and thermodynamically stable. A DC
magnetron gun was used to precisely dope the WO3 film with gold or ruthenium to selectively catalyze the
reaction with NO or NH3. After the films were sputtered, they were subjected to an annealing process which
transformed the as-deposited amorphous film into a polycrystalline film. This results in a more stable film that is
chemically and conductiometrically inert to moisture and many potential interferent gases. Microheaters which
controlled the temperature of the sensing film were fabricated by depositing a thin chrome serpentine structure
on alumina.
The chemiresistive sensing element is placed on the surface of the serpentine microheater with silicone heat sink
grease. The heater and a thermocouple are used to control
the film's operating temperature. The entire apparatus is
suspended on two wires attached to the sensor package to
provide thermal isolation from other nearby sensors and
electronics. Electrical connections to both the sensing film
and microheater are made with 100um aluminum bond wire
to the sensor package. The chemiresistive sensing element
and microbeater is shown in Figure 1.
Thermocouple
Chemiresistive
Sensing Element
Serpentine
Microheater
Figure 1. Chemiresistive sensing element and microheater.
In order to determine the electrical conductivity of a large number of WO3 films exposed to a wide range of gas
concentrations, a system capable of simultaneously testing and controlling up to eight sensors was designed and
built. This system improved testing efficiency and insured that all films were subjected to the same gas
environment during each test. A block diagram of this system is shown in Figure 2.
The sensing elements shown in Figure 1 reside inside the sealed teflon gas chamber. Two-point conductivity
measurements of each sensing film are performed with an electrometer and read into the computer via a HPIB
interface. The computer outputs an analog
voltage to the microheater that is
determined by feeding the measured film
proportional
temperature
Gas Exhaust
Analog Interface
Digital Interface
i AV >
f
Tefloii Gaa Chamber
with Eight Separately
Addressable Sensors
and Heaters
Electrometer ^ ^
Multiplexing; ^J Electrometer' HPFB Interface
' r\
Gas Delivery
i System
Interface Electronics
Computer Control/
Data Acquisition
temperature through a
integral differential (PID)
control algorithm. The support electronics
include power stages to supply the
necessary current to all the microheaters
and amplifiers that linearize the ther-
mocouples' temperature measurements
into a 10mV/ฐC analog signal for the
computer's temperature control algorithm.
Figure 2. Block diagram of the
experimental testing system.
The delivery of the NO and NH3 gases to the sensor is achieved with the system shown in Figure 3. The
apparatus consists of a bottle of the dilution gas (simulated flue gas, in this case), bottles of each test gas (NO
and NH3), an electronic three-way solenoid valve (Y-valve) and a system of mass flow controllers and plumbing.
The entire system is computer controlled and is capable of delivering precise concentrations of NO and NH3 in
an atmosphere of controlled humidity for precise lengths of time. It is fully programmable and capable of running
127
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
multiple tests and collecting data without user input or intervention.
RESULTS
In order to engineer the film to respond to NH3 and NO in a fashion unique to each gas, a large number of
experiments were performed to determine the film parameters that would result in responses to NH3 and NO that
were clearly different. In particular, relationships between film thickness, gold doping, annealing environment,
operating temperature and electrical
ill
conductivity when the film is exposed to
NH3 and NO gas were determined. A few
of these experiments are discussed below
in order to demonstrate the feasibility of
using WO3 films in a sensor array to
selectively detect NH3 and NO.
Figure 3. Gas Delivery System
Simulated
Flue Gas
(Diluent)
Exhaust
iP'
-.*:.--\-
r
[*.. J- -IX)
Gas
Chamber
Bubble
~l
Exhaust
15
SOppm NO
40ppmNO
SOppmNO
20ppm NO
10ppm NO
30
60 90
Time (minutes)
120
150
A 100Q.A WO3 undoped film was
sputtered at 500ฐC on a sapphire
substrate and annealed in dry air at
300ฐC for 10 hours and exposed to
concentrations of NO ranging from
0-50ppm at a temperature of 300ฐC in
an environment of dry CO2. As can be
seen in Figure 4, the film demon-
strated a very fast linear response to
0-50ppm NO. It also recovered to the
baseline resistance when the gas was
taken away which allows the sensor to
measure absolute NO concentrations
in real time.
Figure 4. Response of a 1000A W03
film to 0-50 ppm NO at 300ฐC in dry
CO2
Likewise, Figure 5 shows the same 1000A WO3 film responding to a SOppm pulse of NH3 in 50% humid air. It is
important to note that the response to NO is manifested as an increase in resistivity, while NH3 results in an
increase in conductivity (i.e. a decrease in resistivity), as predicted by equations (5) and (6) above. The film's
change in resistivity is 6 times greater to 50ppm NO than to SOppm NH3. In fact, the film's response to only
10ppm NO was still greater than the response to SOppm NH3. The typical NH3 and NO concentration ranges
found in the fossil fuel combustion exhaust are 0-5ppm and 5-75ppm, respectively. Therefore, this film could
selectively measure up to 10-50ppm NO in the presence of up to SOppm NH3.
A 500A WO3:16A Au film post-sputtered at 200ฐC on alumina and annealed in dry air at 350ฐC for 5 hours was
exposed to 3 pulses of 10ppm NH3 and 3 pulses of SOppm NO in dry air at 350 ฐC. Figure 6 shows the film's high
sensitivity and reproducibility to 10ppm NH3 and relatively low sensitivity to SOppm NO. Again, the response to
NH3 is manifested as an decrease in resistivity, while NO results in an increase in resistivity. The film
demonstrates adequate resolution to potentially measure NH3 concentrations between 0-10ppm in the presence
of up to 5-75ppm of NO.
SUMMARY
As a result of the tests described above, it can be deduced that the film parameters, such as thickness, doping,
annealing procedure, and operating temperature significantly effect the film's sensitivity and selectivity to NH3
and NO gas concentrations. Both, the 1000A WO3 undoped film operated at 300ฐC and the 500A WO3:16AAn
post-sputtered film operated at 350ฐC show potential as a two sensor array capable of simultaneously identifying
concentrations of NO and NH3. Eventually, a neural network could be trained and integrated with the
128
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
chemiresistive sensor array to improve
real time process control and correlate
the sensors' responses to the NO and
NH3 concentrations present in a flue
gas exhaust. In future work, this
technology could be extended to
include other combustion gases such
as SOX and H2S
Figure 5. Response of 1000A WO3
film to SOppm NH3 at 300ฐC in humid
air.
16
? 13
0 10
S 7
30
60
90
Time (minutes)
15
12
9
6
3
0
60
50 -
40 |
30 |
20 o
ID
O
10
60
120
180 240
Time (minutes)
300
360
420
ACKNOWLEDGMENTS
The authors wish to thank the
Environmental Protection Agency for
their financial support (Grant No.
GAD#R826164). Also, T.D. Kenny
was funded through the National
Science Foundation Research Experi-
ence for Undergraduates program
(Grant No. EEC9531378).
Figure 6. Response of a 500A
WO3:16A Au film to (3) 10ppm pulses
of NH3 and (3) SOppm pulses of NO at
350ฐC in humid air, respectively.
REFERENCES
1. J.H.A. Kiel, A. Edelaar, W. Prims and W. Van Swaajj, "Selective Catalytic Reduction of Nitric Oxide by
Ammonia", Applied Catalysis B: Environmental 1, pp. 41-60 (1992).
2. J.A. Eddinger, "Status of EPA Regulatory Development Program for Revised NO* New Source Performance
Standards for Utility and Nonutility Units-Performance and Costs of Control Options", Joint Symposium on
Stationary Combustion NOX Control 6 (1995).
3. K.J. Fewel and J.H. Conroy, "Design Guidelines for NH3 Injection Grids Optimize SCR NOX Removal", pp.
56-64(1993).
4. T. Seiyama, A. Kato, K. Fujiishi and M. Nagatani, "A New Detector for Gaseous Components Using
Semiconductive Thin Films", Anal. Chem. 34, 1502 (1962).
5. See, for example, Proceedings of the First, Second, Third, Fourth and Fifth International Meetings on
Chemical Sensors, 1983, 1986, 1990, 1992 and 1994 and references contained therein. Editors Elsevier
(1983, 1992). Unpublished (1986, 1990, 1994).
6. K. Tanaka, S. Morimoto, S. Sonoda, S. Matsuura, K. Moriya and M. Egashira, "Combustion Monitoring
Sensor Using Tin Dioxide Semiconductor", Sensor and Actuators B 3, 247 (1991).
7. S. Sberveglieri, G. Faglia, S. Groppelli and P. Nelli, "Methods for the Preparation of NO, NO2 and H2 Sensors
Based on Tine Oxide Thin Films Grown by Means of RF Magnetron Sputtering Techniques", Sensors and
Actuators 68,79(1992).
8. G. Williams and G.S.V. Coles, "NOX Response of Tin Dioxide Based Gas Sensors", Sensors and Actuators B
15-16.349(1993).
9. F.J. Gutierrez, L. Ares, J.I. Robla, M.C. Horillo, I. Sayago, J.M. Getino and J.A. de Agapito, NOX Tin Dioxide
Sensor Activities as a Function of Doped Materials and Temperature", Sensors and Actuators B 15-16, 354
129
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
(1993).
10. G.B. Barbi and J.S. Blanco, "Structure of Tin Oxide Layers and Operating Temperature as Factors
Determining the Sensitivity to NOX", Sensors and Actuators B 15-16, 372 (1993).
11. C. DiNatale, A. D'Amico, F.A.M. Davide, G. Faglia, P. Nelli and G. Sberveglieri, "Performance Evalation of
an SnO2-based Sensor Array for the Quantitative Measurement of Mixtures of H2S and NO2", Sensors and
Actuators 820,217(1994).
12. G. Wiegleband J. Heitbaum, "Semiconductor Gas Sensor for Detecting NO and CO Traces in Ambient Air of
Road Traffic", Sensors and Actuators B17, 93 (1994).
13. I. Sayago, J. Gutierrez, L. Ares, J.I. Robla, M.C. Horillo, J. Getino, J. Rino and J.A. Agapito, "The Effect of
Additives in Tin Oxide on the Sensitivity and Selectivity to NOX and CO", Sensors and Actuators B 26-27.19
(1995).
14. Ibid., "Long-term Reliability of Sensors for Detection of Nitrogen Dioxides", Sensors and Actuators B 26-27,
56(1995).
15. K. Satake, A. Kobayashi, T. Inoue, T. Nakahara and T. Takeuchi, "NOX Sensors for Exhaust Monitoring",
Proc. Third Int. Mtg. on Chem. Sensors, Sept. 24-26, 1990, Cleveland Ohio, pp. 334-337.
16. J. Huusko, V. Lantto and H. Torvela, "TiO2 Thick Film Gas Sensors and Their Suitability for NOX Monitoring",
Sensors and Actuators B 15-16. 245 (1993).
17. G. Sberveglieri, P. Benussi, G. Coccoli, S. Groppelli and P Nelli, "Reactively Sputtered Indium Tin Oxide
Polycrystalline Thin Films as NO and NO2 Gas Sensors", Thin Solid Films 186, 349 (1990).
18. S. Matsushima, D. Ikeda, K. Kobayashi and G. Okada "NO2 Gas Sensing Properties of Ga-Doped ZnO Thin
Films", Proc. Fourth Int. Mtg. on Chem. Sensors, Sept. 13-17, 1992, Tokoyo, Japan, pp. 704-705.
19. M. Akiyama, J. Tamaki, N. Miura and N. Yamazoe, "Tungsten Oxide-Based Semiconductor Sensor Highly
Sensitive to NO and NO2", Chem. Letters, 1611 (1991).
20. M. Akiyarna, Z. Zhang, J. Tamaki, T. Harada, N. Miura and N. Yamazoe, "Tungsten Oxide-Based Sensor for
Detection of Nitrogen Oxides in Combustion Exhaust", Sensors and Actuators B 13-14, 619 (1993).
21. J. Tamaki, Z. Zhang, K. Fujimori, M. Akiyama, T. Harada, N. Miura and N. Yamazoe, "Grain-Size Effects in
Tungsten Oxide-Based Sensor for Nitrogen Oxides", J. Electrochem. Soc. 141. 2207 (1994).
22. G. Sberveglieri, L. Depero, S. Gropelli and P Nelli, "W03 Sputtered Tin Films for NOx Monitoring", Sensors
and Actuators B 26-27. 89 (1995).
23. H. Nanto, H. Sokooshi, and T. Usuda, "Smell Sensor Using Aluminum-Doped Zinc Oxide Thin Film Prepared
by Sputtering Technique", Sensors and Actuators B10_,_pp. 79-83 (1993).
24. Hidehito Nanto, Tadatugu Minami, and Shinso Takata, "Ammonia Gas Sensor Using Sputtered Zinc Oxide
Thin Film", Proceedings of the 5th Sensor Symposium, pp. 191-194 (1985).
25. Hidehito Nanto, Tadatugu Minami, and Shinso Takata, "Zinc-Oxide Thin-Film Ammonia Gas Sensors with
High Sensititivity and Excellent Selectivity", Journal of Applied Physics 60 (2). pp. 482-484 (1986).
26. Arya, D'Amico, and Verona, "Study of Sputtered ZnO-Pd Thin Films as Solid State H2 and NH3 Gas
Sensors", Thin Solid Films 157, pp. 169-174 (1988).
27. H. Nanto, H. Sokooshi, T. Kawai, and T. Usuda, "Zinc-Oxide Thin-Film Trimethylamine Sensor With High
Sensitivity and Excellent Selectivity", Journal of Materials Science Letters H, pp. 235-237 (1992).
28. G. Sberveglieri, S. Groppelli, and P. Nelli, "A Novel Method for the Preparation of ZnO-ln Thin Films for
Selective NH3 Detection", 5th International Meeting on Chemical Sensors, pp. 748-751 (1994).
29. A.R. Raju, C.N.R., (MoO3/TiO2 and Bi2MoO6 as Ammonia Sensors", Sensors and Actuators B, 21, pp. 23-26
(1994).
30. A. Bryant, K.-Poirer, D. Lee, and J.F. Vetelino, "Gas Detection Using Surface Acoustic Wave Delay Lines",
Sensors and Actuators 4_, pp. 105-111 (1983).
31. Tomoki Maekawa, Jun Tamaki, Norio Miura, and Noboru Yamazoe, "Gold-Loaded Tungsten Oxide Sensor
for Detection of Ammonia in Air", Chemistry Letters, pp. 639-642 (1992).
32. Tomoki Maekawa, Jun Tamaki, Norio Miura, and Noboru Yamazoe, "Promoting Effects of Noble Metals on
the Detection of Ammonia By Semiconducting Gas Sensor", New Aspects of Spillover Effects in Catalysis,
pp. 421-424(1993).
33. G. Sberveglieri, L. Depero, S. Groppelli, and P Nelli, "WO3 Sputtered Thin Films for NOX Monitoring",
Sensors and Actuators B 26-27. pp.89-92 (1995).
34. H. Meixner, J. Gerblinger, U. Lampe, and M. Fleischer, "Thin-Film Gas Sensors Based on Semiconducting
Metal Oxides", Sensors and Actuators B 23, pp. 119-125
(1995).
130
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
RELEVANCE OF ENANTIOMERIC SEPARATIONS IN ENVIRONMENTAL SCIENCE
Daniel W. Armstrong
Department of Chemistry, University of Missouri-Rolla, Rolla, MO
A significant number of all organic chemicals that are released into the environment are racemic mixtures. Most
environmental regulations and scientific environmental studies treat racemic mixtures as though they were
single, pure compounds. This can lead to incorrect lexicological, distribution, degradation and other data. A
series of new enantioselective analytical techniques have been developed that allow the facile separation and
quantitation of chiral compounds of environmental importance.1 Table 1 shows a typical example of the disparate
biological activities of paclobutrazol stereoisomers.2
Table 1. Biological activity of paclobutrazol diastereoisomers and enantiomers.
Compound Fungicidal activity Plant growth regulatory
cc-cwCMCK-<' *>a (cereal mildews and rust) activity (apple seedlings)
c i ~ 2RS.3RS High High
2R.3R (+) High Low
2S,3S(-) Low High
Note: The 2R,3R(+) enantiomer has a high fungicidal but a low plant growth regulatory activity. For the 2S.3S (-)
enantiomer the reverse situation holds true. Separation of the enantiomers implies separation of the desired and
the undesired action
We have examined the enantioselective biodegradation of chlorinated pesticides and the herbicides dichlorprop
and mecoprop. While it is known that the herbicidally active enantiomer is the (+) enantiomer for both herbicides,
and fate of each enantiomer in broadleaf weeds and grass has not previously been reported. Both dichlorprop
and mecoprop are sold as racemic mixtures and are among the most commonly used herbicides for the control
of broadleaf weeds in grass. This presentation compares the biodegradation of each enantiomer of dichlorprop
and mecoprop in several types of broadleaf weeds and common grasses. The results indicate that one
enantiomer is degraded faster in weeds, while both enantiomers degrade at equal rates in grass.
References
1. D.A. Armstrong, G.L. Reid, M.L. Hilton, C.-D. Chang, Environ. Pollution 79 (1993) 51-58.
2. E. J. Ariens, in Chiral Separations by HPLC, Ed. A. M. Kistulovic, John Wiley & Sons, New York, 1989, Ch. 2,
pp. 69-80.
DEVELOPMENT OF AEROSOL MASS SPECTROMETER FOR
REAL TIME ANALYSIS OF PAH BOUND TO SUBMICRON PARTICLES
J.T. Jayne, D.R. Worsnop and C.E. Kolb
Aerodyne Research, Inc. 45 Manning Road Billerica, MA 01821
Tel: 978-663-9500, Fax: 978-663-4918
X. Zhang and K.A. Smith
Massachusetts Institute of Technology, Department of Chemical Engineering, Cambridge, MA 01239
Tel: 617-253-1973, Fax: 617-253-2701
Polycyclic aromatic hydrocarbon (PAH) are mutagenic pollutants formed as by-products of combustion.
Measurements of the distribution of PAH species with different aerosol particles of different sizes are critical for a
complete understanding of the environmental fate of the human exposure to PAH. We present here an aerosol
mass spectrometer (AMS) designed to simultaneously measure particle size, particle number density and
size-resolved particle composition for volatile and semi-volatile compounds. The instrument combines a unique
aerodynamic sampling inlet which focuses the particles into a narrow beam and efficiently transports them from
131
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
atmospheric pressure into a vacuum chamber where particle mass/size is determined by time-of-flight (TOP)
measurement. Size-resolved single particles are flash vaporized on a heated filament and ionized by resonance
enhanced multi-photon ionization (REMPI) using an excimer laser (248 nm). The PAH ions are detected by a
molecular TOP mass spectrometer. The ionization process is highly selective for aromatic PAH.
The AMS described here provides size-resolved single particle composition detection down to 50 nm. Molecular
mass spectra for a number of PAH aerosols have been measured. The sampling efficiency and the size
resolution of the instrument have been investigated using a differential mobility analyzer (DMA) and
condensation nucleus counter. Size resolution defined as ADp/Dp = 4 (where Dp = particle diameter) has been
measured for tenth micron size particles with sampling efficiencies approaching unity, and found to be about
25%.
REAL-TIME TRACE DETECTION OF ELEMENTAL MERCURY AND ITS COMPOUNDS
Robert B. Barat
New Jersey Institute of Technology, Dep't. of Chemical Engineering,
Chemistry, and Environmental Science Newark, NJ 07102
Introduction
Emission of elemental mercury [Hg] vapor and volatile mercury compounds [e.g. HgCI2, Hg(CH3)2] from
combustion and other processes is an important environmental issue [Von Burg and Greenwood, 1991]. The
present research addresses the need to develop real time stack monitoring of emissions of Hg and its
compounds. Such continuous emission monitor (CEM) technology would enable identification of peak emission
events and the possibility of corrective action. Realistic emission inventory will enable regulatory agencies to
better assess health risks associated with future siting of emission sources, such as waste incinerators.
To be useful to a wide range of applications, the CEM should be capable of detection in the range of 1-5000
ug/m3, with an ultimate sensitivity limit on the order of 0.1 ug/m3 (ca. 10 pptv). Current best technology for Hg
detection uses cold vapor trap resonant atomic fluorescence [Tekran, 1998]. However, this technology provides
no information on mercury compounds.
In this work, detection of mercury compounds is based on photo-fragment fluorescence (PFF) excited by deep
ultraviolet (UV) light. The fluorescence spectra can facilitate identification of the original mercury compounds.
Photo-fragment fluorescence has been successfully applied for the gas-phase analysis of HgCI2, Hg(CH3)CI, and
Hgl2 (Barat and Poulos, 1998).
The detection for Hg uses Doppler-shifted resonant atomic fluorescence excited by a UV laser or a low pressure
mercury lamp. The Doppler-shifted fluorescence will be separated from the unshifted background signal by use
of an optically dense Hg vapor filter precisely matched to the spectral linewidth of the source. An alternative to
the vapor filter is precise use of time gating to distinguish the nearly instantaneous background scattering from
the relatively long fluorescence signal.
The experimental program involves three stages. In the first, a static cell (no flow) containing mercury (elemental
or compound) vapor is probed with a deep UV laser to generate spectra. An atmospheric pressure flow cell is
used in the second stage. The third stage utilizes the expanding jet.
Compounded Mercury
In general, mercury compounds absorb light strongly below 250 nm [Gowenlock and Trotman, 1955], and these
absorption bands are generally dissociative. In PFF, a photolyzing UV photon dissociates the target molecule
into fragments, some of which are imparted with excess energy. The energy might then be lost by fluorescence:
hv
A-B-->A + B* (1a)
B* > B + hv' (1b)
132
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
where A and B can be atoms or polyatomic species. The fragment identities and distributions, as revealed in the
fluorescence spectrum, can, in principle, provide information on the parent species, in a manner analogous to
mass spectrometry and other fragmentation spectroscopies.
For example, Figure 2 shows the PFF spectrum from 193 nm excimer laser excitation of Hg(CH3)CI vapor in a
static cell (Figure 1). Two features are evident: atomic Hg emission lines at 546 and 579 nm; and a broad
continuum assigned to the B->X system of HgCI** excited state (Mandl and Parks, 1978; Whitehurst and King,
1987). Likely photochemical processes resulting in these observations are:
hv
Hg(CH3)CI > HgCI** + CH3 (2)
HgCI** > HgCI + hv1 (2a)
> Hg + Cl (2b)
Steps (2a) and (2b) are clearly implicated, but the origin of the Hg atomic emission lines is open to question.
Focussing the laser beam increases the emission from Hg, and generates the blue emission at 431 nm from
CH*.
The lowest concentration of mercury compound [Hg(CH3)CI] measured in the static cell was approximately 50
pg/m3. Using reasonable improvements in optics and electronics, it is estimated that the limit-of-detection can be
lowered by at least a factor of 500.
The concentration of the target compound is related to the fluorescence intensity from a hot fragment. Excitation
laser energy of about 2 mj/pulse at a repetition rate of 10 Hz at 222 nm was applied to HgBr2 vapor (in Argon) in
an atmospheric pressure flow cell (Figure 3). The PFF was monitored using a photomultiplier tube and narrow
interference filters centered at 254 nm - a strong Hg* emission line - (see Figure 4). Excellent linearity was
obtained over a wide concentration range. Similar results were obtained using a compact monochromator + CCD
system.
The supersonic jet spectroscopy (to be discussed in the next section) is expected to further improve sensitivity of
PFF detection. Spectra will be sharpened, leading to better discrimination of fragment vibrations! structure.
Quenching by O2 will be reduced due to the low pressure.
Elemental Mercury
Atomic Fluorescence Spectroscopy (AFS) is a highly sensitive spectroscopic marker for elemental Hg detection.
Current AFS instruments (e.g. Tekran, 1998) use a cold vapor trap for collection + concentration of the air
sample, purging (to remove O2), desorption, excitation with an Hg vapor lamp (253.7 mn), and then
measurement of the resonant fluorescence (at the same wavelength). Sensitivity is limited by the elastically
scattered light from the exciting source.
The technique under study, shown in Figure 5, will expand the Hg-contaminated air stream across a supersonic
nozzle into a high vacuum chamber. Light at 253.4 rim will be directed across the jet. Atomic Hg fluorescence
will be Doppler-shifted by between 1 and 3 GHz due to the jet motion. Total collected light, comprised of the
shifted fluorescence and stray elastic scattering, will be passed out of the vacuum chamber and through (Figure
6) an optically dense, sharp cut-off Hg vapor filter (centered at the excitation wavelength) to reduce elastic
scattering while transmitting the fluorescence signal (Miles, 1991; Finkelstein, et al., 1994).
The expansion reduces the collision rate of Hg* with O2, so collisional quenching and collisional broadening are
both substantially reduced. The low temperature associated with this expansion further reduces the fluorescence
linewidth, enhancing the performance of the Hg vapor filter.
In the absence (or in conjunction with) the atomic filter, time gating of the collected signal offers an alternative
means to extract the fluorescence signal from the scattering background. It has been observed in the
atmospheric pressure flow cell experiments with HgBr2 that scattering is essentially instantaneous, lasting for the
duration of the laser pulse (i.e. 10 nanoseconds). The relatively long lifetime for Hg* (117 nanoseconds at 253.7
nm - Dodd et. al., 1970) results in the fluorescence signal appearing as a long-tailed shoulder on the scattering
signal. Prudent placement of the signal detection gate, as well as subtraction of the background signal measured
in the absence of Hg vapor, should allow for extraction of the desired fluorescence signal.
133
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Quality Assurance / Performance Assessment
For comparison of time-averaged test concentrations, the mercurycontaining sample stream will be diverted to a
reference method, such as optical absorption. Testing of the research technology will consist of obtaining data on
five performance measures:
Relative Accuracy: the absolute mean difference between the metals concentration determined by the monitor
and that determined by the reference method, plus a 2.5 percent uncertainty confidence coefficient based on a
test series.
Calibration Drift: the difference in the monitor output reading from the established reference value after a stated
period of operation. The reference value is established by a calibration standard which has a concentration
nominally 80 percent or greater of the full scale reading capability of the monitor.
Zero Drift: calibration drift when the reference value is zero.
Response Time: the time interval between the start of a step change in the concentration of the monitored gas
stream and the time when the output signal reaches 95 percent of the final value.
Detection Limit: three times the standard deviation of nine repeated measurements of a low-level (near blank)
sample.
Acknowledgement
The author thanks Dr. Arthur T. Poulos, President of Poulos Technical Services, Inc., for his significant
contributions to this research. The author would also like to thank the U.S. Environmental Protection Agency -
National Center for Environmental Research and Quality Assurance for its financial support of this work (Grant #
R-825380-01-0), and grant project officer Mr. William Stelz for his administrative guidance.
References
Barat, R.B. and Poulos, A.T., "Detection of Mercury Compounds in the Gas Phase by Laser
Photo-Fragmentation/Emission Spectroscopy," Applied Spectroscopy (accepted -1998).
Dodd, J. N., Sandle, W. J., and Williams, O.M., J. Phys. B: Atom. Molec. Phys. 3, 256 (1970).
Finkelstein, N., Gambogi, J., Lempert, W.R., Miles, R.B., Rine, G.A., Finch, A. and Schwarz, R.A., Proceedings
of 32nd Aerospace Sciences Meeting, Jaunary, 1994, Reno, NV, Paper #A1AA 94-0492.
Gowenlock, B. G. and Trotman, J., J. Chem. Society, pt. 2, 1454 (1955).
Mandl, A. and J.H. Parks, Appl. Phys. Let. 33 (6), 498 (1978).
Miles, R.B., "Absorption Line Filter Window and Method for Velocity Measurements by Light Scattering," U.S.
Patent #4,988,190 (1991).
Tekran, Inc., 1-132 Railside Road, Toronto, Canada.
Von Burg, R. and Greenwood, M.R., in Metals and
Their Compounds in the Environment, ed. by E.
Merian, VCH, Weinheim (1991).
Whitehurst, C. and T. A. King, J. Phys. D: Appl.
Phys. 20, 1577(1987).
to
manifold
Figure 1. Research Apparatus #1: Static Cell for
Atomic and Photo fragment Fluorescence
134
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Figure 2b. PFFS of Hg(CH3)CI
193 nm excitation - focused
Figure 2a. PFFS of Hg(CH3)CI
193 nm excitation - not focused
sen
1.2 -
I. I -
1 -
0 S -
08 -
07-
06-
01-
0 ^ -
03-
02-
0 I -
0
Ar, Air,
orN2
Frits
Heated
Tube
Solid
Samples
NOT TO SCALE
Metal
compound
vapor
w
-*
Mono.
MPC
W
Heated
Optical
Chamber
Mono-
chromator
-i ro
i I vent
^" U
Trap
Figure 3. Research Apparatus #2: Flow
Cell for PFF
Beam Stop
M ซ mirror
L-lens
MPC multi-pass cell
135
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
4 -
1 3
w
u.
t
2 -
1 -
Figure 4. Photofragment Fluorescence (PFF) Signal
vs. HgBr2 Concentration
2000 4000
HgBr2 Concentration (
6000
8000
Figure 5. Apparatus #3: Atomic Fluorescence and
Compound PFF in Vacuum
NARROW
LINEWIDTH
LIGHT
(on resonance)
dilution
nr ฐ" \
HO
Saturates
Diffusion
pimp
nozzle
\
Hg
Compound .
Saturator
prซoture
valve
III
VB.jm
Stop
"--|- - ~"
1
r
i
i
i
to
r
r*w
P^1
mchMkal
1 1 vacuum
cold PUTOP
Vacuum
Chamber
j !__.] windows
1 Lฑb
UV source t 1
vent 1 1
cold
trap
Claaer) ฃ-
Mhe.m
-W i
^f 1
n
Hg Atomic Fitter
(Princeton)
lor$ pass
ptpcul fitter
CCDor
to data
acquisition
L.J
j
PMT
BACKGROUND
SCATTERING
FROM WALLS
AND WINDOWS
ATOMIC
FLUORESCENCE
1 DOPPLER-SHIFTED
TRANSMISSION PROFILE
OF ATOMIC Hg FILTER
Figure 6. Doppler-shifted Resonant Fluorescence
FLUORESCENCE
SEEN BY DETECTOR
FREQUENCY
136
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
ORTHOGONAL BACKGROUND SUPPRESSION TECHNIQUE
FOR EPA'S FIELD INFRARED DATA PROCESSING
Blaterwick
NO ABSTRACT AVAILABLE
DEVELOPMENT OF A CONTINUOUS MONITORING SYSTEM
FOR PM10 AND COMPONENTS OF PM25
Morton Lippmann. Judy Q. Xiong and Wei Li
New York University Medical Center, Nelson Institute of Environmental Medicine,
57 Old Forge Road, Tuxedo, New York 10987
ABSTRACT
While particulate matter with aerodynamic diameters below 10 and 2.5 urn (PMio and PM25) correlate with
excess mortality and morbidity, there is evidence for still closer epidemiological associations with sulfate ion, and
experimental exposure-response studies suggest that the hydrogen ion and ultrafine (PM015) concentrations may
be important risk factors. Also, there are measurement artifacts in current methods used to measure ambient
PMio and PM25, including negative artifacts because of losses of sampled semivolatile components (ammonium
nitrate and some organics) and positive artifacts due to particle-bound water. In order to study such issues, we
are developing a semi-continuous monitoring system for PMio, PM25, semivolatiles (organic compounds and
NH4NO3), particle bound water, and other PM2s constituents that may be causal factors. PMซ is aerodynamically
sorted into three size-fractions: 1) coarse (PMi0-PM25); 2) accumulation mode (PM2.s-PM0is); and 3) ultrafine
(PMo.-is). The mass concentration of each fraction is measured in terms of the linear relation between
accumulated mass and pressure drop on polycarbonate pore filters. The PMois mass, being highly correlated
with the ultrafine number concentration, provides a good index of the total number concentration in ambient air.
For the accumulation mode (PM25-PMois), which contains nearly all of the semivolatiles and particle-bound water
by mass, aliquots of the aerosol stream flow into system components that continuously monitor sulfur (by flame
photometry), ammonium and nitrate (by chemiluminescence following catalytic transformations to NO), organics
(by thermal-optical analysis) and particlebound water (by electrolytic hygrometer after vacuum evaporation of
sampled particles). The concentration of I-T can be calculated (by ion balance using the monitoring data on NO3,
NHV, and SO4=).
OBJECTIVES
Background
Particulate matter (PM) is an ambient air criteria pollutant that does not, as listed, have any specific
compositional definition. When initially defined (in 1971) as total suspended particulate matter (TSP), it had no
specific particle size distinction either. The TSP inlet cut-size was determined by the inlet aspiration efficiency,
whose upper 50% cut-size (20-50 gm) varied with ambient wind speed and direction. The 1987 PMio National
Ambient Air Quality Standard (NAAQS) revision was defined in terms of the mass concentration of PM aspirated
by a non-directional and wind speed insensitive inlet with a 50% sampling efficiency at -10 urn aerodynamic
diameter. The PMio cut approximates that of the normal human upper respiratory tract during oral inhalation, so
that the sampled particles represent those that can penetrate to the thoracic airways (tracheobronchial tree and
more distal gas-exchange airways). Particles depositing along the tracheobronchial tree (lung conductive
airways) can exacerbate asthma and cause bronchitis and bronchial cancer, while those depositing in the
gas-exchange airways can cause lung fibrosis, emphysema and peripheral lung cancers. Particles that are
retained in the lung airways can cause persistent local irritation and/or dissolve and be translocated to more
distant organs via the bloodstream.12
Neither the PMio nor the TSP NAAQS made any distinction as to the chemical composition of the particles
sampled. In terms of composition, there is a relatively clear distinction between coarse and fine particles in the
ambient air, as illustrated in Figure 1.
137
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The coarse mode of the ambient aerosol is generally composed of soil and soil-like dust and ash particles
generated and/or dispersed by mechanical forces. It is typically dominated by basic mineral oxides in particles
larger than ~1.0 urn. By contrast, the fine mode particles are generally derived from gas-phase precursors, which
form in the atmosphere as ultrafme (nuclei mode) particles after chemical transformations. The transient nuclei
mode particles rapidly coagulate and coalesce to form larger accumulation mode particles in the light-scattering
range (0.1 to 1 pm), are persistent in the air for
many days, and contain almost all of the sulfate,
ammonium and hydrogen ion, as well as most of
the nitrate ion and carbonaceous particles.
Size-selective ambient air samplers have been
developed to collect the coarse mode and fine
mode particles separately, typically with a
relatively sharp size-cut at 2.5 urn in aerodynamic
diameter (PM25). When using a 2.5 urn cut, the
fine mode includes essentially all of the sulfates
and organics, but also includes the lower tail of
the coarse mode.
0.01
0.1
DP (urn)
Nuclei Mode Accumulation Mode
ป< ^^_
10
Coarse Mode
100
Fine Particles
-x-
Coarse Particles
TSP, Total Suspended Paniculate Matter
PMz.5
>
Coarse Fraction
Figure 1. Measured mass size distribution
showing particles in nuclei and accumulation
modes of fine particles. Also shown are
transformation and growth mechanisms (e.g.,
nucleation, condensation, and coagulation).
From some epidemiological studies, there is suggestive evidence that the excess mortality and morbidity
associated with elevated PM was more closely correlated with PM25 than with PMio.3 The fine particle mass in
the eastern U.S. is dominated by sulfates (SCv), and the epidemiological evidence suggests that excess
mortality and morbidity are as well or better correlated with ambient SO4= than with PM25.4 The SO4= is, in turn,
highly correlated with aerosol acidity ((-T), a more likely causal factor than SOC on the basis of controlled
exposure studies in humans and animals.4
At this time, the actual causal factor(s) for the excess mortality and morbidity are not clearly established.
However, in 1997, the EPA adopted both 24 hr and annual average PM25 standards on the basis that PM25 is the
best currently available surrogate index for the health effects associated with ambient air PM.5 At the same time,
EPA retained the PMio NAAQS, with some relaxation in stringency, because of residual concerns about the
health risks from the coarser particles that deposit on lung conductive airways, and may cause or exacerbate
asthma, bronchitis, or bronchial cancer.5
Currently available monitoring methods, which are based on filter sampling of ambient PMio or PM25, with
subsequent determination of sampled mass, have significant limitations for the accurate determination of
ambient PM mass concentrations. Some of the ambient PM is semi-volatile, especially ammonium nitrate and
some of the organic constituents. Sampled PM mass can be lost, especially when the temperature is elevated
during the sampling and/or prior to analysis. Nitrate (NO3-) can also be lost from the sampling filter after
collection of acidic sulfates as the H+ combines with NO3 to form nitric acid vapor (HNO3) that is carried off by air
passing through the filter. There can be positive artifacts as well, most notably due to water of hydration at high
ambient humidities. Their associated water may contribute to the measured fine particle mass without adding to
the health risk. In view of these considerations, there is a need to be able to determine the concentrations of
several key species within the accumulation mode particles that may be either candidate causal factors or likely
sources of sampling artifacts. The most important of these aerosol components are SO4=, NO3, NH4+, H+,
semi-volatile organics (SVOC) and water vapor.
There is also some concern about the health effects of ultrafine particles (those less than 0.1 um diameter).
Some animal studies indicate that extremely small mass concentrations of ultrafines 25 ug/m3) can cause excess
mortality and pathological changes after brief exposures67 Evidence for an important role for ultrafines is
reported by Peters et al.,8 who found somewhat closer associations between reduced pulmonary function in
nonsmoking adult asthmatics with number concentration than with PM25 mass concentration. Finally, it remains
138
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
possible that the effects of PM25 may really be due to its total mass concentration and not to any of its specific
chemical constituents. Thus, it is important to accurately determine the overall mass concentration.
Design Objectives
These considerations lead to our design objectives for the continuous monitoring of PMio and PM25. The
concentrations of components to be measured separately are: 1) coarse mode particle mass (i.e., PMi0-PM25); 2)
fine mode sulfate (SO4=); 3) fine mode nitrate (NO3~); 4) fine mode ammonium (NH4+); 5) fine mode organic
carbon (OC) and elemental carbon (EC); 6) fine mode water (H2O); 7) fine particle mass (PM25); and 8) ultrafine
mode mass (PM015).
The sum of mass components 1 and 7 equals PMio mass. I-T, a PM parameter of interest can also be derived
from the monitoring data. As shown in Figure 2, the milliequivalent sum of the accumulation mode anions (SCv
and NO3~) is equal to the sum of the cations (NH4+ + H+). Thus, we can estimate H+ concentration by net ion
difference. For ultrafines, the parameter of interest may be number concentration rather than mass
concentration. However, as indicated in Figure 3, the mass and number concentrations for particles below 0.1
urn are highly correlated. Thus, the ultrafine number concentration can be reliably estimated from the
corresponding measured mass concentration. Alternatively, the number concentration can be directly measured
using a condensation nucleus counter (CMC).
1-300.
E
^f 200.
z
ซ:
a) Albany. NY ''
(n=140)
f
ttf 1:1 line
*2''
JJV* f=0.94
r8
300 .
200 .
100 -
0 J
b) Buffalo, NY ,**
(n=178) *'
*
*
4
* <ป)*'
j|/P *
J^ r=x0.96
400
300.
200.
100
c) White Plains, NY
(r*=164)
0 '
*
jr*
-0.94
100 200 300 400 0 100 200 300 400 0
Measured H + (nmol/m3)
100 200 300 400
Figure 2. Comparison of estimated H+ (via ion difference) to directly measured H+ (via pH) for data collected on
Teflon filters at 3 NYDEC sites (June-August, 1988 and 1989). Unpublished data from Dr. G.D. Thurston, NYU.
160,000
140.000
1W.OOO
100,000
80.000
60,000
40*00
20,000
n
0 '
'
m
Qj? "" *"
n
3 a D n
\ ฐa$S?f ff 0
U ,_ , . . .
2.00 4.00 fi.00
Volume i.1nm (|ซn3cm-3)
SO 100 ISO
Volum* < 2.5 iim (jim3cm-3)
200
Figure 3. Relationships between particle number and volume concentrations: Lett panel for ultrafine particles
(smaller than 0.1 urn); right panel for fine particles (smaller than 2.5 M). From: USEPA (1996).
Design Concept
The overall design concept is illustrated schematically in Figure 4.
The PM10 inlet limits access to those particles that can penetrate into the human thorax. This is followed by a
virtual impactorwith a 2.5 urn cut-size. The coarse particle mode is directed onto a spot on a polycarbonate pore
ucleporeTM) filter tape using the filter resistance method developed by Koutrakis et al.9 After a suitable sampling
interval and determination of particle mass collected, the tape spot is mechanically advanced for sample storage
139
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
and presentation of a fresh filter surface for the next sampling interval. The fine particle fraction (suspended in
29 Ipm of the inlet flow) is carried into a second virtual impactor with a 0.15 urn cut-size (lowest practical
cut-size). The smaller (ultrafine) particles are collected on a filter spot on a sequential filter sampler for periodic
mass concentration analyses in a manner similar to that used to measure the coarse particle fraction. The
extension of the method of Koutrakis et al.9 to ultrafine particles is described in detail later in this paper. The
accumulation mode particles (0.15 to 2.5 urn), suspended in 2.9 Ipm of the inlet air, are directed into: 1) a stream
of 1.8 Ipm leading to the aerosol water detector; and 2) a stream of 1.1 Ipm leading to the inlets of the continuous
detectors for accumulation mode aerosol components of primary interest, i.e., SCV, NO3-, NH4+, OC, and EC. A
third filter tape sampler draws off 0.2 Ipin of the fine particle (PM25) stream into a filter tape sampler for the
determination of the overall mass concentration of PM2.5.
Virtual Impactor
(djO = 15 Mm)
2.9 Ip
Sulfate
Analyzer
1
Nitrate and
\mmonium
Analyzer
Organlcs
Analyzer
Particle-bound
Water Mass
Analyzer
Direct Readout of
(PM |0-PM2.9)
Mass Concentration
Direct Readout
of PM 2.5 Mass
Concentration
Direct Readout
of PM0 u Mass
Concentration
Direct Analysisof PMj 5-PM 0.15 Components
Figure 4. Overview of the system for continuous measurement of PM 10 and components of PM25.
With the exception of the aerosol water and paniculate mass detectors, each of the continuous monitors uses
well established detection methods available in widely used commercial instruments. Similarly, the PMio inlet
and 2.5 urn virtual impactor are widely used and commercially available. A 0.15 urn virtual impactor was
designed and tested by Sioutos et al.10
Monitoring System Elements Based on Commercial Available Equipment
1) Flame-Photometric Detector (FPD) for Aerosol Sulfate: The Meloy Model 285 FPD can be used to measure
total concentration of sulfur in the aerosol by using PbO diffusion denuders at the inlet to remove ambient
vapors, such as SO2, H2S, and mercaptans. The sulfur in the aerosol is, with rare exceptions, due to its presence
in sulfates (H2SO4, NI-UHSCX,, and (NH4)2SO4). Thus, the sulfate ion mass concentration is essentially equivalent
to three times the measured sulfur concentration. This application of the FPD has been described by Cobourn et
al.11 and Allen et al.12 The instrument detection limit is 1 ppb (4 ug/m3 SCv) with a sampling flowrate of 180
ml/min. Since the sample is preconcentrated 10 times by means of a 0.15 gm virtual impactor before analysis,
the detection limit is ~0.4 ug/m3 for accumulation mode particulate SO4 in ambient air.
2) Thermal-Optical Technique for Measurement of Aerosol Organic Carbon (OC) and Elemental Carbon (ECli
For measurement of aerosol organic and elemental carbon, we plan to adopt an in situ aerosol carbon analysis
method developed by Turpin et al.13 The method combines the sampling function of a two-port parallel filter
sampling technique with the analytical function of a thermal-optical carbon analyzer,14'15 and was employed in the
Carbonaceous Species Method Comparison Study (CSMCS) in Glendora, California, in the summer of 1986, for
side by side measurement of sub-2.5 gm aerosol carbon concentrations with other conventional sampling and
analysis methods.16 The detection limit of the method was reported to be as low as 0.2 ug carbon with a precision
of about 3%.
140
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
3) PMio Inlet and PM?5 Virtual Impactor: We use 16.7 liter/min (Ipm) inlets and 2.5 um virtual impactors from
Series 241 Graseby-Andersen PMio Manual Dichotomous Samplers. They separate the 16.7 Ipm of inlet flow
into: 1) a 1.7 Ipm stream containing the PMio coarse-mode fraction along with 10% of the fine fraction; and 2) a
15 Ipm stream containing 90% of the fine PM fraction.
Monitoring System Elements Under Development and/or Undergoing Evaluation Detail in this Research
1) Particulate Mass Concentration Monitors: We are using the newly developed method of Koutrakis et al.,9 in
which the mass accumulated on a polycarbonate pore (Nuclepore) filter can be shown to be directly
proportional to the pressure drop across that filter. The basis for the method is that the particles are collected at
the entries to and within the pores of the Nuclepore filter by interception or Brownian diffusion, rather than on
the surfaces between the pores by impaction. The flow through the pores is restricted by the presence of the
collected particles in proportion to their volume.
For the coarse particles (PMio-PM25), a Nuclepore filter with 10 urn pores is being used, while for the fine mass
2 urn pores are used. In order to measure the mass concentration of ultrafine particles, we use a modified
system that uses a filter with 0.2 um pores. The 0.2 urn pore filter
has a relatively high baseline flow resistance, and very small mass 1
increments due to collected ultrafines will markedly increase the h -J-
resistance, providing a very sensitive measure of the mass
concentration of ultrafine particles.
Design and Validation of the Ultrafine Mass Monitors
The ultrafine, mass monitor (CPMM-U) consists of two parallel
channels, four capillary pore filters (N1 ... N4), and two HEPA filters
(Figure 5). In addition, needle valves and flow meters are used to
control and monitor the flow rates in each channel. Two sensitive
pressure transducers (T1 and T2) are used to measure the change in
pressure drop at two locations along each channel as shown in
Figure 5. The measured pressure drops can be related to the mass
loading of the first capillary pore filter of the measurement channel.
The left channel, which has a capillary pore filter exposed to ambient
particles, is the measurement channel. The right channel, with both
capilliary pore filters behind the HEPA filter, is the reference
channel.
Figure 5. Schematic layout of the basic elements of the system for
the measurement of the mass concentration of the ultrafine function
of ambient air particulate matter. The increased flow resistance
across the N1 Nuclepore filter is proportional to the accumulated
mass and number of ultrafine particles.
The pressure drop across a capillary pore filter is affected by relative humidity, temperature, flow rate, and the
static pressure at the entrance of the filter. The effect of any one of these factors can exceed the change in
pressure drop due to particle loading. While N2 serves as a reference to eliminate the fluctuations in relative
humidity, temperature and flowrate, a one-channel design has serious limitations. First, since the range of the
change in pressure drop across N1 is very small (<5%) in comparison to the overall pressure drop of the filter, it
is difficult to accurately measure the change in pressure drop due to particle loading. Also, it cannot be assumed
that N2 operates at the same conditions as N1. This requirement is approximately satisfied for humidity,
temperature and flow rate. However, the pressure at the entrance of N2 may be different if N1 causes a
significant pressure drop, as in the case of using a capillary pore filter with a small pore size. A two-channel
design greatly improves on these limitations.
To measure the mass of ultrafine particles, a very small pore size is needed in order to match the size of
particles of interest and optimize the sensitivity of the instrument. This results in a very high baseline pressure
drop The typical pressure drop in the Koutrakis9 CPMM design using a 2 um pore size was -12 inches of water,
while the pressure drop in CPMM-U using a 0.2 um pore size is -70 inches. Two problems arise: 1) the difficulty
141
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
of balancing both pressure transducers as the sampling time increases (use of a pressure transducer with a
broad measurement range limits the sensitivity); 2) the difficulty of maintaining a leak-free system increases (a
small leak in the system will cause errors in mass measurement).
Only those particles depositing at the entrances to, or inside of the pores of, a filter can contribute to the increase
in pressure drop Therefore, the flow rate was selected to minimize the impaction of particles on the surface of
the filter and to maximize the diffusion and interception of particles inside of the pores. Other factors important
in the selection include: 1) ensuring that the change in pressure drop across N1 can be measured in a reasonably
period of time for the typical ambient particle concentration; 2) achieving a pressure drop that is not too high
(causing operational problems). A capillary pore filter with 0.2 urn pore size and a face velocity of 5 cm/sec was
selected.18
Koutrakis et al9 have shown that, for the CPMM, the increase in pressure drop of N1 can be calculated by
2T1-T2 One of the assumptions is that the pressure drop of N2 is the same as that of N1. This condition is
satisfied when using capillary pore filter of larger pore size (2.0 urn). However, this condition is not satisfied when
0 2 urn capillary pore filters are used, because the pressure drop across a capillary pore filter depends upon the
flow rate and pressure at the filter face. Because of the pressure drop across N1 (-70 in. of water), the pressure
drop across N2 is only about 85% of that across N1. Generally, the pressure drop across N1, AP, can be
expressed as, Ar_ H-K-n
_ APft/2
a - APW1
where T1 and T2 are the pressure differentials recorded by pressure transducer 1 and 2. APNi and APN2 are the
pressure drop across N1 and N2 at a given flow rate, respectively. For a face velocity of 5 cm/sec, the a has a
value of 0.85.
As aerosol enters the CPMM-U, it is dried by passing through a diffusion dryer. The flow then splits. On the left
path, particles deposit on N1 , causing a increase in flow restriction across N1. Any particle that penetrates N1
will be removed completely by the HEPA filter that is located further down the line. Therefore, the flow restriction
of N2 will not change due to particle loading during sampling. On the right path, particles are immediately
removed by the HEPA filter. The flow restrictions of both the first and second capillary filters will not change due
to particle loading during sampling because: 1) the capacity of the HEPA filter is much larger than that of a
capillary filter; and 2) the pressure drop across the HEPA filter is less than one hundredth of that across a
capillary pore filter under our experimental conditions. Thus, the increase in flow restriction of HEPA filter is
negligible in comparison to that of the capillary pore filter.
As particles deposit on N1, the balance of the system is
self-adjusted to accommodate the change in flow restriction
of N1. Thus, the pressure drop across the N1 increases and
the T1 reading increases. Assuming that the flow rates of
both lines are not significantly changed during the period of
sampling, the T2 reading also increases. For the calibration
of the ultrafine particle mass monitor (CPMM-U) we
employed a method using an Ultrafine Condensation
Particle Counter and monodisperse particles to calibrate the
CPMM-U. The system used is described in a separate
paper.19 The results of the calibration tests for various sizes
of monodisperse ultrafine particles are shown in Figure 6.
1
i
"3
a
z
1
o
s
Q.
Q
p
1
ฃ
.g
8)
/
6
5
4
3
2
1
n
. i i | < i i | . i . | i . | . i i | i i l i i y^
7 ]12 nm Q-'
y-0.136+0.0492ซ, R. 0.992 f.- J
.'*
j,-'
^.O 95.0 nm "
L ..- ' -|
: ."" :
ft"
: f.-''o 753 nm :
; -t5 66.8 nm '.
51.3 nm ..' 66'8 "m "
ฐ.-'' "
_." _:
O" 42.0 nm
: P 35.3 nm
"0 31.7 nm, ..,...,... | ... | -
20 40 60 80 100
Mass Concentration (ug/m3)
120
140
Figure 6. The results of the calibration tests for various
sizes of monodisperse ultrafine particles.
Since the pressure drop across the capilliary pore filters in the CPMM-U is 5-7 times higher than that in the
CPMM, the filter tape advancement system for CPMM was considered unsuitable for the CPMM-U. Koutrakis et
al.20 have shown that a small leak in the CPMM can cause a significant error in the measurement of mass
142
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
concentration. To avoid leakage problems we use a multi-head filter holder that can provide a tight seal for each
filter. Instead of a filter advancement system, we use a sequential sampling approach. Figure 7 shows the
systematic design of the multi-head filter holder. The multi-head filter holder is machined from an aluminum
block. Recessed filter holders (the number depends on the optimization of the CPMM-U) are machined into the
block. Each filter holder includes a stainless steel screen backup and a gasket. The solenoid valves, which are
controlled by a computer, select one sampling channel at a time. The capilliary pore filters are loaded into the
filter holder block in the laboratory, and the multi-head filter holder is replaced as a whole in the field. Since the
linear range of the instrument extends to a mass concentration up to 20 ug/m3, one filter may be used for long
period sampling if the concentration of particles is relatively low. For example, if the mass concentration of the
ultrafine fraction is 1 ug/m3, the linear response
range will not be exceeded until 20 one-hour
consecutive samplings are made on one filter.
Therefore, a multi-head holder containing 14 filter
holders should be sufficient for a one week
sampling under almost all ambient conditions in
Filter holder block the U.S.
Out
Solenoid valve
Figure 7. Systematic design of the multi-head filter
holder.
In
2) Particle-Bound Water Detection System: The basic problem that must be overcome is the extremely high
background water vapor concentration in the air compared to the water bound by particles at normal
environmental conditions. Due to the rapid equilibrium between water vapor and particle surface water
(milliseconds time-scale), there is no conventional method for separating the particle-bound water and its
coexisting vapor without disturbing the phase equilibrium. The innovative concepts of our method are: 1) to
collect particles over a relatively short time period (compared with the time scale of environmental variation);
and 2) to collect the particles without disturbing the water equilibrium between the particle and gas phases (by
maintaining the sampling system at the condition of the ambient environment); and 3) to minimize the sample
cell volume in which the associated air remains. The particle-bound water can then be readily detected above the
background in air by means of a highly sensitive moisture detector, such as a P2O5-Pt electrolytic hygrometer.
The electrolytic hygrometer was chosen as a water detector because: 1) it has the lowest detection limits
currently available; 2) it is relatively inexpensive; and 3) it is convenient to operate.
A schematic diagram of our semi-realtime analyzer
for particle-bound water in accumulation mode
aerosol is illustrated in Figure 8. The basis for the
technique is the accretion of PM25-PM015 particles
by means of a filter over a preset period of time.
Two identical sample cells, each one consisting of a
13 mm Teflon membrane filter (2 urn pore size) and
a small enclosure, are connected in series. The
Teflon membrane filter was selected due to its
excellent particle collection efficiency, low moisture
uptake, and low trace background. The upstream
sample cell collects and preconcentrates the
particles in the sampled air from the 0.15 pm virtual
impactor, while the second cell analyzes the
backgrounds from the air and filter.
Aerosol
Pressure Sensor
Vacuum
A 13mm Teflon filter
(2 Jim pare size) with
Teflon support is
mounted between two
Teflon washers
ฃ Vacuum Valve
Ijjl Three-way Solenoid Valve
Two-way Solenoid Valve
Filler
Figure 8. Schematic diagram of the system for
continuous measurement of particle-bound water
within PM2s
A pair of parallel sampling lines are used for alternating the sampling and analysis processes. They are
143
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
controlled by 2 three-way solenoid valves, the sample inlet valve (S1) and the exhaust valve (S2). When S1 and
S2 switch to line A, the line A starts sampling and the line B starts analysis of the samples collected in the prior
sampling period. At the end of the sampling period, both S1 and S2 switch, and the functions are reversed.
During Line A sampling, solenoid valve SA opens. When switched to analysis, SA closes, and the two sample
cells will be separated and analyzed in order. The air background goes first and the aerosol sample next. All the
solenoid valves are controlled by an interfacing computer with a preset program.
The water detector consists of a 1000 cm3 Vapor Dilution Cell (VDC), an Electrolytic Hygrometer, a flow
controller, a pressure sensor, a vacuum line, and a helium gas line. The procedures for analysis of the water
contents collected by each sample cell are: 1) evacuate the VDC; 2) close the vacuum line and open the VDC to
the cell to be analyzed; 3) extract the water bound in the particles and absorbed on the filter and draw it into the
VDC along with the air remaining in the cell by vacuum; 4) fill the system with helium gas to a pressure of about
900 torr (slightly higher than atmosphere); 5) open the system to the hygrometer, and allow a small flow (10
cm3/min) to pass through the electrolytic sensor; and 6) start one measurement. The output signal is recorded
and processed by an interfacing data acquisition system. The particle-bound water is determined from the
difference between the measurements of the two cells. The filter of the aerosol sample cell is changed after each
run.
A P2O5-Pt electrode is used to sense water vapor concentration in the system. It consists of a quartz tube wound
with two platinum (Pt) wires. The winding is coated with a phosphorus pentoxide (P2O5) film, which has a high
affinity for H2O. A voltage is applied to the platinum winding so that the water molecules adsorbed on the P2O5
film are electrolyzed to H2 and O2 and a current flow is generated as described below:
Cathode reaction:
Anode reaction:
2 H2O+2e- -> 2 OH + H2
2 OH -+ H2O + 1/2 O2 + 2 e
Since every H2O molecule electrolyzed produces two electrons (based on Faraday's Law of Electrolysis) the
current is directly proportional to the concentration of the H2O molecules in the gas stream. This correlation is
insensitive to gas pressure and mass flowrate. The sensor is commercially available for sampling moisture in gas
streams and at normal atmospheric pressure. The working range is 0-1000 ppm with an accuracy of 2%. The
lowest reported detection limit is 10 ppb.21 The sample cell volume is 2 cm3 (See Figure 8). For a flowrate of 1.8
Ipm and a sampling time of 60 minutes, the concentration of particle-bound water is elevated by a factor of
5.4x10s in relation to its carrier air stream (including a preconcentration factor of 10 provided by the 0.15 virtual
impactor). Therefore, the particle-bound water is measured despite the associated vapor in the air stream. A
detection limit of the system for measuring particle-bound water is ~5 pg/m3 of total particle mass concentration
with a water composition of 15% at RH above 40%, and 5% at RH below 40%, in a sampling period of 60
minutes.
3) Tungstic Acid Technique - Chemiluminescent NO* Detector (TAT-CLD) System for Measurements of
Particulate Nitrate (NCv) and Ammonium (NH/): A system for continuous measurement of particulate NO3 and
NH4+ is shown in Figure 9. The system combines a two-channel Chemiluminescent NOx Analyzer (Monitor Labs,
Model 8840) with the Tungstic Acid Technique (TAT) developed by Braman et al.22 The TAT was used by
Braman et al.22 for preconcentration and determination of gaseous and particulate nitrate and ammonia, based
on the principle that nitrate and
Exhaust ammonia are quantitatively chemi-
sorbed at the tungstic acid-coated
er"1 surfaces at room temperature, and
thermally desorbed at high tempera-
ture (350ฐC).
He
Tungstic
Acid
Denuder
Tungstic
Acid
Converter
Makeup
Air
i
fc 1-
Molybde
Convefl
Remove particles
350 ฐC
NO3Xp)*~ NOjIg)
3iOฐC
ป NH3(g)
T
Makeup
Air
Exhaust
Figure 9. Schematic diagram of
tungstic acid denuder and converter-
chemiluminescent NOX detector sys-
tem for continuous measurement of
nitrate and ammonium concentrations
in accumulation mode particles.
144
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The advantage of the TAT is that there is essentially no interference of aseous NO* in either the NO3- or NH4+
analyses. The weakness of the Braman et al.22 TAT is the limited particle collection efficiency of the TAF. For a
34 cm long three-section TAF, the particle penetration was as high as 22% at a sampling flowrate of 1 IpM.22
Large packed tubes can be used to improve efficiency, but this limits the sampling flowrate attainable. In our
system, the sample air is preconcentrated 10 times prior to entering the analysis system. The sampling flowrate
of the TAT-CLD may be as low as 0.1 - 0.2 Ipm without increasing the detection limit of the original TAT. Thus a
much higher particle collection efficiency is achieved. A two-channel CLD is used for parallel detection of NO3
and NH4+ in our system, which simplifies the processes of sample separation. It also minimizes the sampling
artifacts reported by Braman et al.22
The instrument detection limit of CLD is 2 ppb with a sampling flowrate of 250 ml/min for each channel. The
detection limits of TAT-CLD system for ambient accumulation mode particulate NO3 and NH4+ are below 0.1
gg/m3 for a 30-minute sampling period.
4) Data Acquisition and System Control: To automate the detection system and the data acquisition, we use a
computer interfaced data acquisition and instrument control system. As shown in Figure 10, it consists of an
IBM-PC computer, an Analog-to-Digital Converter Board (ADC), a four-channel Signal Conditioner, a Digital
Interface Board, and a Multichannel Relay Board (MRB). The Signal Conditioner converts the signals from
sensors to standard signals (0 - ฑ 5V, or 0 - ฑ 10V), which can be accepted by ADC. The signals received from
ADC are recorded and stored in the
computer at specified intervals. To control cp cp cp cin senors
the solenoid valves, a MRB and a Digital ~ "
Interface is used.
Figure 10. Schematic of the data acquisition
and instrument control system.
L
_L
1 Plug-In ADC
_ \ Board
Cable
Signal Conditioner
H Power
supply
Plug-In Digital
Interface
Cable
Multichannel
Relav Board
DISCUSSION
An ambient PM monitoring package (prototype monitor) has been designed to be capable of continuous
operation at our laboratories in New York City and in Tuxedo, NY, a location 50 miles north-northwest (NNW) of
New York City. It will provide records of concentration data as a fimction of time for PMio and eight of its
components, including coarse mode particles (PMio-PM25), fine particles (PM25), ultrafine particles (PM015) and
the constituents of the accumulation mode aerosol (PM25-PMois) of primary interest, including SCv, NO3~, NH4+,
l-f, H2O, OC, and EC. Using three different pore sizes for the three mass fractions helps insure that interception
will be the dominant collection mechanism and that each fraction will be collected within the optimum range for
linear response. Using this monitoring system, researchers will be able to measure the mass concentrations of
PMio and PM25 in near real-time, and without sampling artifacts due to sample volatilization and particle-bound
water. They will also be able to develop data bases for the concentrations of specific components of PM25 that
may be causal factors for the PM-associated health effects of concern (e.g., H+, SCv , OC, EC, and ultrafines),
thereby providing opportunities for more definitive epidemiological studies. Such studies could provide the basis
for future NAAQS for specific PM components and thereby more rationale design and implementation of source
controls of PM and/or PM precursors.
ACKNOWLEDGMENTS
This manuscript is a condensed version of a paper that has been submitted for publication in Applied
Occupational and Environmental Hygiene, and premission to reproduce parts of it has been graciously granted
by its Editor-in-Chief.
This research was supported by Grant R825305 from the U.S. Environmental Protection Agency, and is part of a
Center Program supported by Grant ES 00260 from the National Institute of Environmental Health Sciences.
The consultation and support of Dr. Petros Koutrakis and his colleagues for our application of the filter resistance
method for PM mass concentration measurement developed at the Harvard School of Public Health is also
gratefully acknowledged.
LITERATURE CITED
1 Lippmarm, M.; Yeates, D.B.; Albert, R.E.: Deposition, Retention and Clearance of Inhaled Particles. Brit. J.
Ind. Med. 37:337-362(1980).
145
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
2. Lippmann, M.: Size-Selective Health Hazard Sampling. In: Air Sampling Instruments, 8th Ed., pp. 81-119.
S.V. Hering, Ed. Am. Conference of Governmental Industrial Hygienists, Cincinnati, OH (1995).
3. U.S. Environmental Protection Agency: Air Quality Criteria for Particulate Matter. EPA/600/P-95/001F.
USEPA, Washington, DC (1996).
4. Lippmann, M.; Thurston, G.D.: (1996). Sulfate Concentrations as an Indicator of Ambient Particulate
Matter Air Pollution for Health Risk Evaluations. J. Exposure Anal. Environ. Epidemiol. 6:123-146 (1996).
5. U.S. Environmental Protection Agency: National Ambient Air Quality Standards for Particulate Matter.
Federal Register 62:38762-38896 (July 1997).
6. Oberdorster, G.; Gelein, R.M.; Ferin, J.; Weiss, B.: Association of Particulate Air Pollution and Acute
Mortality. Inhal. Toxicol. 7:111-124 (1995).
7. Chen, L.C.; Wu, C.Y.; Qu, Q.S.; Schlesinger, R.B.: Number Concentration and Mass Concentration as
Determinants of Biological Response to Inhaled Irritant Particles. Inhal. Toxicol. 7:557-588 (1995).
8. Peters, A.; Wichmann, E.; Tuch, T.; et al.: Respiratory Effects are Associated with the Number of Ultrafine
Particles. Am. J. Respir. Grit. Care Med. (In press).
9. Sioutos, C.; Koutrakis, P.; and Olson, B.A. Development and Evaluation of a Low Cutpoint Virtual
Impactor. Aerosol Sci. Technol. 21:223-235 (1994).
10. Koutrakis, P.; Wang, P.Y.; Sioutas, C.; Wolfson, J.M.: U.S. Patent No. 5571945 (1995).
11. Cobourn, W.G.; Husar, R.B.; Husar, J.D.: Continuous In Situ Monitoring of Ambient Particulate Sulfur
Using Flame Photometry and Thermal Analysis. Atmos. Environ. 12:89-98 (1978).
12. Allen, G.A.; Turner, W.A.; Wolfson, J.M.; Spengler, J.D.: Description of a Continuous Sulfuric Xcid/Sulfate
Monitor. In: Proceedings of the National Symposium on Recent Advances in Pollutant Monitoring of
Ambient Air and Stationary Sources, Raleigh, NC, pp. 140-151. USEPA, EPA-600/9-84-019 (1984).
13. Turpin, B.; Gary, R.; Huntzicker, J.: An In Situ Time Resolved Analyzer for Aerosol Organic and Element
Carbon. Aerosol Sci. Technol. 12:161-171 (1990).
14. Johnson, R.L.; Shah, J.J.; Gary, R.A.; Huntzicker, J.J.: An Automated Thermal Optical Method for the
Analysis of Carbonaceous Aerosol. In: Atmospheric Aerosol: Source/Air Quality Relationships, pp.
223-233. E.S. Macias and P.K. Hopke, Eds. American Chemical Society, Washington, DC, ACS Symp.
Ser. No. 167(1981).
15. Huntzicker, J.J.; Johnson, R.L.; Shah, J.J.; Gary, R.A.: Analysis of Organic and Elemental Carbon in
Ambient Aerosols by a Thermal-Optical Method. In: Particulate Carbon: Atmospheric Life Cycle, pp. 79-88.
G.T. Wolff and R.L. Klimisch, Eds. Plenum, New York (1982).
16. Hering, S.V.; Appel, B.R.; Cheng, W.; et a].: Comparison of Sampling Methods for Carbonaceous Aerosol
in Ambient Air. Aerosol Sci. Technol. 12:200-213 (1990).
17. Marple, V.A.; Rubow, K.L.; Turner, W.; Spengler, J.D.: Low Flow Rate Sharp Cut Impactors for Indoor Air
Sampling: Design and Calibration. JAPCA 37:1303-1307 (1987).
18. Spurny K.R.; Lodge, J.P.; Frank, E.R.; Sheesley, D.C.: Aerosol Filtration by Means of Nuclepore Filters:
Structural and Filtration Properties. Environ. Sci. Technol. 3(5):453-464 (1969).
19. Li, W.; Xiong, J.Q.; Lippmann, M. The Development of a Continuous Particle Mass Monitor for Ultrafine
Ambient Particles (In preparation).
20. Koutrakis, P.; Wang, P.Y.; Wolfson, J.M.: Private Communication (1996).
21. McAndrew, J.J.: Moisture Analysis in Process Gas Streams. Solid State Technol. 35:55-60 (1992).
22. Braman, R.S.; Shelley, T.; McClenny, W.A.: Tungstic Acid for Preconcentration and Determination of
Gaseous and Particulate Ammonia and Nitric Acid in Ambient Air. Anal. Chem. 54:358-364 (1982).
A REAL-TIME SAMPLER RAMS, FOR THE DETERMINATION OF PM25,
INCLUDING SEMI-VOLATILE SPECIES
F Obeidi, E. Patterson and D.J. Eatough
Department of Chemistry and Biochemistry, Brigham Young University, Provo UT 84602 USA
The RAMS, Figure 1, is a real-time ambient monitor for the determination of fine paniculate mass, including the
volatile components (Eatough 1998). The RAMS has a particle concentrator, followed by diffusion denuders to
remove gas phase compounds which can be absorbed by charcoal, a Nafion dryer to remove water, and a
"sandwich filter" containing a Teflon coated filter to collect particles and a charcoal impregnated filter to retain
146
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
volatile components which can be lost from the particles during sample collection. Semi-volatile fine particulate
material retained on the "sandwich filter include ammonium nitrate and semi-volatile organic compounds. The
"sandwich filter" is located at the tip of the tapered oscillating element of a TEOM monitor and mass retained on
the "sandwich filter" is measured as a function of time.
The results obtained with the RAMS have been validated by comparison with results obtained from diffusion
denuder integrated samples, to determine the mass of fine particulate material retained on a filter and the
semi-volatile organic material and ammonium nitrate lost torn the filter during sampling. This has included
comparisons with sampling periods for the denuder samplers as short as 1 hour. Results obtained with RAMS
and denuder samplers for samples collected in Riverside CA in the summer and Bakersfield, CA in the winter
show that semi-volatile fine particulate species are accurately monitored with the RAMS.
Research is currently underway to validate the measurement of volatile constituents of fine particles with the
RAMS in chamber experiments using well characterized particles of ammonium sulfate, ammonium nitrate,
glycerol and a carboxylic acid.
ACKNOWLEDGMENTS
The research reported here was supported by the U.S. Environmental Protection Agency STAR grant R825367
to Brigham Young University. Some support to the program has also been provided by the Electric Power
Research Institute, and Rupprecht and Patashnick, Inc.
REFERENCE
Eatough D.J., Obeidi F., Pang Y., Ding Y., Eatough N.L. and Wilson W.E. (1998) "Integrated and real-time
diffusion denuder samplers for PM25 based on BOSS, PC and TEOM technology," Atmos. Environ., submitted.
DYNAMIC NUCLEAR POLARIZATION (DNP):
A NEW DETECTOR FOR ANALYSIS OF ENVIRONMENTAL TOXICANTS
Harry C. Dorn
Department of Chemistry, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061
The objective of this study is the development of a new analytical instrument designed for environmental
monitoring applications. Specifically, this will consist of dynamic nuclear polarization detection with either
direct-coupling continuous-flow supercritical fluid chromatography (SFC/DNP) or recycled-flow 13C (DNP)
analysis of toxicant mixtures. The DNP detector is a variant of the well known nuclear magnetic resonance
(NMR) phenomena. A salient feature of NMR is the chemical shift parameter which provides a very sensitive
probe of the local electronic environment about a given atom in a molecule. Thus, the DNP detector could have
wide ranging applications for specific monitoring of various organic toxicants mixtures (e.g., chlorocarbons,
organophosphates, pesticides, petroleum pollutants, etc.). A major limitation of NMR for most environmental
monitoring applications has been sensitivity constraints. The DNP approach helps alleviate the sensitivity
limitation of NMR by transfer of polarization from an electron spin to the nuclear spin of interest (1H, 13C, 31P,
etc.). The corresponding DNP signal enhancements are proportional to the electron-to-nuclear magnetogyric
ratio (YB/YH) which is on the order of 103-104 for most nuclides.
In this presentation, LC/DNP and SFC/DNP results for chlorocarbon mixtures will be presented, In addition, the
results for continuous monitoring of a mixture of benzene and several chlorocarbons with a recycled-flow 13C
DNP instrument will also be presented. Finally, progress towards development of a routine SFC/DNP instrument
will be reported.
147
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
PARTITIONING TRACERS FOR IN-SITU DETECTION AND MEASUREMENT
OF NONAQUEOUS LIQUIDS IN POROUS MEDIA
Brusseau
ABSTRACT NOT AVAILABLE
148
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
QUALITY
ASSURANCE
149
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
150
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
CD-R ARCHIVE AND CATALOG OF HEWLETT-PACKARD LOW RESOLUTION
FORMAT DATA FROM NETWORKED GC/MS SYSTEMS
Robert G. Briggs and Herman Valente
The Wadsworth Center, New York State Department of Health,
Empire State Plaza, Box 509, Albany, New York 12201-0509
Phone (519)473-4871
robert. briggs@wadsworth. org
ABSTRACT
(1) Taped GC/MS data are not efficiently accessible; (2) Cataloging of archived data with descriptive information
requires retyping information entered by the operator at the time of data acquisition. We have addressed these
two problems, in our laboratory by archiving all of our data, generated on various Hewlett-Packard GC/MS
systems, on CD-R (Compact Disk, Recordable) media. Data are initially sent to a Mylex Level 5 RAID array; then
a CD is prepared using the Pinnacle Computing RCD-202 and associated RCD-PC software. A Microsoft Visual
Basic program developed in-house (Arch-CD) extracts sample information entered at analysis time from each
data file selected for archive and generates a Microsoft Access database (.MDB) file. This MDB file is included
with the data files on the CD and also merged with an integrated database. The database along with the CDs on
a changer provide instant access to five years of GC/MS data.
INTRODUCTION
Our laboratory applies EPA CLP (Contract Laboratory Protocol) volatilcs and semivolatiles methodology to water
and soil samples or their extracts.
We have four GC/MS systems dedicated to this work, three Hewlett Packard MSD systems and one Varian
Saturn Ion Trap system.
Submission rate for samples of all types averages a modest 750 per year. With daily calibration and other
QC, which proceeds irrespective of sample submission, this translates to about 3000 GC/MS runs per year.
The average input rate belies the fact that submission is episodic (figure 1) and that samples for a given
instrument are received at a rate up to 30 per day, rather than the more prosaic one per day. Turnaround
requirements cannot be adjusted to sample submission rate so we must be ready to handle the maximum.
Prior to 1993, we were limited in disk space, even with a network of PCs, and data was archived on tape. In
a major project where non-target analytes are of concern, it is desirable to reevaluate previously processed
samples when a new class of substances is detected. The report time is inordinately extended because of the
need to move data in and out of storage. Taped GC/MS data are not efficiently viewed or accessed.
Once data are sent to long-term archive, key information should be put in a database to facilitate future
access. A catalog of archived data should contain, at minimum, the descriptive information entered by the
operator at the time of data acquisition, but retyping this information into a database is tedious and should be
unnecessary.
SUMMARY
We have addressed these two problems of short-term data access and long-term accessibility in our laboratory.
For five years we have been archiving all of our data, generated on various Hewlett-Packard GC/MS systems, on
CD-R (Compact Disk, Recordable).
Data are initially sent to a 1.2 Gigabyte partition on a Mylex Level 5 RAID array.
Then they are mastered using RCD-PC software for the Pinnacle Computing RCD-202 onto a 0.8 Gigabyte
partition on the RAID array.
CD-R disks are prepared and transferred to SCSI chained Pioneer 604X/624X six disk changers.
CDs beyond the twelve accessible disks are kept in 6-disk cartridges for insertion in the changer and
reasonably rapid access.
A Microsoft Visual Basic program developed in-house (Arch-CD) extracts sample information entered at
analysis time from each data file selected for archive and generates a Microsoft Access database (.MDB)
file. This is included with the data files on the CD.
A master database is built up from individual MDB format files.
The success of this system is illustrated for analysis of volatile organics in water (figure 2). Although the
percentage of total time from sample receipt to the analytical report dedicated to data processing has
remained a constant 80% - 90%, the total time necessary for analysis and report has fallen significantly since
151
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
implementation of this system, from 40 days (or more) to 10 -20 days. During this time, the rate of
submission for samples in this category actually increased (figure 1).
The long report time events of 1995 (figure 2) were due to data glut on our ion trap system, one of the two
systems used for water volatiles. This low-level ion-trap system was only recently connected to the network,
has limited hard disk storage and has depended on tape backup. We have avoided recurrence of this
problem by networking the instrument and archiving its data with those from the HP systems. We have not
yet developed information-extraction methodology to incorporate these Saturn data into the database.
Hewlett Packard publishes the structure of data files for users whereas Varian does not.
CONCLUSION
Accumulated sample information from a five-year-period is now accessible for HP systems and the data files are
readily retrieved for use.
THE USE OF ACCEPTABLE KNOWLEDGE FOR THE CHARACTERIZATION
OF TRANSURANIC WASTE IN THE DEPARTMENT OF ENERGY COMPLEX
R. Vann Bynum
Science Applications International Corporation, 2109 Air Park Road S.E., Albuquerque, NM 87109
R. Butch Stroud and R. Denny Brown
Department of Energy, Carlsbad Area Office, 4021 National Parks Highway, Carlsbad, NM 88220
Laurie D. Sparks
Waste Isolation Division, Westinghouse Electric Co., a Division of CBS,
P.O. Box 2078, Carlsbad, NM 88221
ABSTRACT
The Resource Conservation and Recovery Act (RCRA) regulations codified in 40 CFR Parts 260 through 265,
268, and 270, authorize the use of acceptable knowledge as a method which can be used in appropriate
circumstances by waste generators, or treatment, storage, or disposal facilities to make hazardous waste
determinations. Acceptable knowledge, as an alternative to sampling and analysis, can be used to meet all or
part of waste characterization requirements under RCRA.
One example of the use of acceptable knowledge within the U. S. Department of Energy (DOE) complex is the
waste characterization requirements for the Waste Isolation Pilot Plant (WIPP), a deep geologic repository for
the disposal of transuranic (TRU) waste. The WIPP will be disposing of TRU mixed and non-mixed waste from
various generators within the DOE complex in accordance with the provisions of the DOE Carlsbad Area Office
(CAO) Quality Assurance Program Plan, the EPA Final Certification Decision, an anticipated RCRA Permit, and
the WIPP Waste Acceptance Criteria. Part B of the WIPP RCRA Permit application contains a Waste Analysis
Plan which describes the measures that will be taken by the DOE/CAO to assure that mixed wastes received at
the WIPP repository are characterized appropriately. To satisfy the characterization requirements, acceptable
knowledge is confirmed by radiography, drum headspace gas sampling and analysis, and solidified waste
sampling and analysis. Acceptable knowledge is primarily used in TRU waste characterization activities to
delineate TRU waste streams, to determine if TRU debris wastes exhibit a toxicity characteristic (40 CFR
261.24), and to determine if TRU wastes are listed (40 CFR 261.31). The physical form and the increased health
and safety risks associated with obtaining a representative sample of TRU debris wastes, clearly justify the use
of acceptable knowledge in making hazardous waste determinations.
DOE complex waste generators apply knowledge of their waste based on the materials and processes used to
generate the waste. Acceptable knowledge includes information regarding the physical form of the waste, the
base materials composing the waste, the nature of the radioactivity present, and the process(es) generating the
waste.
The DOE/CAO audits the TRU waste generators to grant TRU waste certification authority to the generators. The
DOE/CAO conducts audits at least annually thereafter to verify ongoing compliance with approved plans and
procedures including those for waste characterization using acceptable knowledge. The information resulting
152
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
from DOE/CAO audits is used by the DOE/CAO to help sites across the DOE complex prevent or solve problems
associated with the compilation and use of acceptable knowledge for TRU waste characterization. Past and
current issues with respect to the compilation, use, and defensibility of acceptable knowledge for TRU waste
characterization along with the solutions to those problems and their prevention will be addressed.
INTRODUCTION
Waste Isolation Pilot Plant (WIPP)
The WIPP is a deep geologic repository designed to permanently dispose of transuranic (TRU) radioactive waste
generated from the research, development, and production of nuclear weapons in an evaporite salt formation.
The mission of the WIPP Project, as authorized by the U.S. Congress in 1979 (Public Law 96-164), is to provide
a research and development facility to demonstrate the safe disposal of TRU waste generated as a result of
United States defense activities. WIPP is located in Eddy County in southeastern New Mexico about 26 miles
east of Carlsbad, New Mexico, in an area known as Los Medanos - a relatively flat, sparsely inhabited plateau
with little surface water. WIPP encompasses a 16-square mile area under the jurisdiction of the U.S. Department
of Energy (DOE) pursuant to the Land Withdrawal Act.1 The site boundary was established to ensure that at least
1 mile of intact salt exists laterally between the waste disposal area and the accessible environment and to
ensure that no permanent residences will be established in close proximity to the facility.
The DOE's objective is to operate and maintain the WIPP free of both chemical and radiological contamination.
Therefore, all waste sampling and analyses will be conducted by the DOE generator/storage sites in accordance
with the requirements of the WIPP Waste Analysis Plan (WAP, Chapter C of the RCRA Part B Permit
Application)2 and the WIPP Compliance Certification Application (Chapter 4)3 as allowed by 20 New Mexico
Administrative Code (NMAC) 4.1, Subpart V, Paragraph 264.13, and consistent with joint U.S. Environmental
Protection Agency (EPA) and U.S. Nuclear Regulatory Commission guidance.
Transuranic Waste
The transuranic (TRU) elements have atomic numbers greater than that of uranium (92). Examples of
transuranic elements include neptunium (Np), plutonium (Pu), americium (Am), curium (Cm), etc. Each element
typically has several isotopes and is produced during nuclear reactions. These man-made elements are
radioactive and provide the key components for building nuclear weapons.
Transuranic waste consists of clothing, tools, rags, and other items contaminated with trace amounts of
radioactive TRU elements - mostly plutonium. TRU waste is defined1 as "waste containing more than 100
nanocuries of alpha-emitting transuranic isotopes, per gram of waste, with half-lives greater than 20 years,
except for a) high-level radioactive waste, b) waste that the Secretary has determined, with concurrence of the
Administrator, does not need the degree of isolation required by the disposal regulations; or c) waste that the
Nuclear Regulatory Commission has approved for disposal on a case-by-case basis in accordance with Part 61
of Title 10, Code of Federal Regulations."
TRU-rmixed waste is waste that is commingled with hazardous materials, such as lead or organic solvents. It is
regulated by both the Atomic Energy Act and RCRA (as defined in 20 NMAC 4.1, Subpart VIII, Paragraph
268.35(d) and in the Federal Facility Compliance Act.4 TRU-mixed waste has physical and radiological
characteristics similar to TRU waste. The majority of TRU-mixed waste contains relatively small quantities of
spent halogenated solvents, which were used in cleaning and degreasing of equipment, glassware, and
components. Based on sampling of gases within TRU waste drums, the most common volatile organic hazardous
constituents are methylene chloride, carbon tetrachloride, and 1,1,1-trichloroethane.5 TRU-mixed waste also
contains various RCRA-regulated metals. These metals are usually associated with solid materials, such as lead
shielding. Lead, chromium, and cadmium are the most prevalent hazardous metals in TRU-mixed waste.
TRU-mixed waste constitutes approximately 60 percent of the DOE's TRU waste.6 The hazardous components of
TRU-mixed waste to be managed at the WIPP are designated in Part A of the RCRA Permit application.2
Henceforth, the term TRU waste used in this paper includes both TRU and TRU-mixed waste.
Transuranic Waste Generator/Storage Sites
TRU waste has accumulated over the past 50 years as a result of weapons development and production at U.S.
defense facilities. Since 1970, DOE has segregated TRU waste from other radioactive waste and stored it in a
manner that allows it to be retrieved. TRU waste to be emplaced at WIPP has resulted primarily from the
following: (1) nuclear weapons development and manufacturing, (2) plutonium recovery, (3) defense research
and development, (4) environmental restoration, (5) decontamination and decommissioning, (6) waste
153
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
management, and (7) testing at facilities that are under DOE contract.
Table 1 has been reproduced from the National TRU Waste Management Plan.7 It lists the volumes of TRU
waste currently in storage and the volumes of TRU waste projected to be generated by ongoing and new
missions during the life of the WIPP Estimates from environmental restoration, decontamination and
decommissioning and future missions are also included. The nuclear weapons complex consists of ten major
facilities, including those at large reservations listed in the upper half of Table 1.
Table 1. TRU Waste Storage Locations and Pre-treatment Volumes (cubic meters)7
Contact-Handled
TRU Waste
Site
Argonne National Lab.-East
Hanford Reservation
Idaho National Engineering and
Environmental Lab.
Lawrence Livermore National Lab.
Los Alamos National Lab.
Mound Plant
Nevada Test Site
Oak Ridge National Lab.
Rocky Flats Environmental
Technology Site
Savannah River Site
Small-Quantity Sites
Ames Laboratory
ARCO Medical Products
Company
Babcock & M/ilcox-NES
Battelle Columbus Laboratories
Bettis Atomic Power Laboratory
Energy Technology Engineering
Center
General Electric-Vallecitos
Nuclear Center
Knolls Atomic Power Lab.
Lawrence Berkeley Lab.
Missouri University Research
Reactor
Paducah Gaseous Diffusion Plant
Sandia National Laboratories
U.S. Army Material Command
Total Waste Volumes***
Location
Argonne, IL
Richland, WA
Idaho Falls, ID
Livermore, CA
Los Alamos, NM
Miamisburg, OH
Nevada
Oak Ridge, TN
Golden, CO
Aiken, SC
Ames, IA
West Chester, PA
Lynchberg, VA
Columbus, OH
West Mifflin, PA
Santa Susana, CA
Pleasanton, CA
Niskayuna, NY
Berkeley, CA
Columbia, MO
Paducah, KY
Albuquerque, NM
Rock Island, IL
Remote-Handled
TRU Waste
Stored* Projected Stored* Projected
through through
2033** 2033**
94
16,127
64,575
297
8,255
241
618
917
1,505
11,725
0
<1
20
0
0
7
6
0
<1
<1
2
7
2.5
104,400
109
7,305
15,009
835
8,544
6
19
180
6,988
17,811
<1
<1
0
0
114
0
3
0
4
1
0
44
0
56,972
0
200
86
0
101
0
0
1,268
1,268
1
0
0
0
0
0
0
8
<1
0
0
0
1
0
1,666
0
1,592
53
0
128
0
0
100
100
21
0
0
0
369
2
1
5
5
0
0
0
3
0
2,268
Volumes Prior to treatment and repackaging.
** Projected volumes include estimates from environmental restoration, decontamination and decommissioning,
and future Departmental missions, for example, the disposition of weapons-useable plutonium at the
Savannah River Site. Estimates will change based upon future compliance actions under environmental law.
*** Totals reflect rounding of numbers.
Continued temporary storage of TRU waste at these and other sites across the nation poses potential problems.
For example, some of the metal drums used to store TRU waste are showing signs of corrosion, and the contents
of these drums eventually will have to be repackaged. Not only would additional storage facilities be needed at
the generator/storage sites, but also additional worker exposures to penetrating radiation would occur due to
154
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
repackaging and inspection of waste containers. New treatment capacity would also be needed because much of
the TRU waste is subject to RCRA Land Disposal Restrictions and cannot be placed in or on the land unless it is
treated to satisfy those restrictions. Sound environmental practice requires that TRU waste be permanently
isolated to prevent human exposure for many generations to come.
TRANSURANIC WASTE CHARACTERIZATION
The process of identifying and classifying the chemical, physical, and radiological constituents of each drum of
waste is a critical aspect of waste characterization. TRU waste characterization is a subset of the waste
certification process and is based on sampling and analysis combined with acceptable knowledge of each waste
stream in accordance with the requirements of the TRU Waste Characterization Quality Assurance Program
Plan8 (QAPP). The QAPP utilizes a performance-based approach to allow individual sites to have the flexibility
to employ analytical and examination methods that meet the quality assurance objectives specified in the Waste
Analysis Plan (WAP)2 and implemented by the requirements of the QAPP
Retrievably stored1 TRU waste will be characterized by the generator/storage sites as the waste is retrieved.
Newly generated" TRU waste will be characterized as it is generated. Waste characterization requirements for
retrievably stored and newly generated wastes are slightly different and are discussed in the WAP Waste
characterization activities at the generator/storage sites include the following, although not all of these
techniques will be used on each container:
Acceptable Knowledge: Compilation of documented characterization and/or process knowledge into an
auditable record
Headspace-gas sampling and analysis: Used to determine volatile organic compound (VOC) content of
gases in the void volume of the containers
Sampling and analysis of homogeneous solid waste forms: Used to determine concentrations of hazardous
waste constituents and toxicity characteristic contaminants of waste in containers
Radiography: an x-ray technique used to determine physical contents of containers
Visual examination: Used to verify radiography results
Radioassay: Used to identify isotopic inventory and associated activity
The origins of these requirements in the WAP are traceable to applicable regulatory requirements and
commitments made to regulatory authorities such as the NMED and the EPA.
NMED Basis for Waste Characterization
The Waste Analysis Plan (WAP), Chapter C of the RCRA Part B Permit Application,2 describes the measures
that will be taken to assure that TRU waste received at the WIPP facility is within the scope of the RCRA permit
as established in and with unit-specific requirements of Title 20 of the New Mexico Administrative Code, Chapter
4, Part 1, Subpart V, Paragraph 264.13 (20 NMAC 4.1). The WAP establishes waste characterization
requirements that are referenced in the Waste Acceptance Criteria (WAC),9 the QAPP, and the TRU-waste
generator/storage sites Quality Assurance Project Plans (QAPjPs). It includes descriptions of waste parameters,
rationale, and characterization methods; waste sampling and analysis strategies; waste shipment screening and
verification processes; and a quality assurance (QA)/quality control (QC) program.
The WIPP underground disposal unit is classified as a "miscellaneous unit" subject to regulation under 40 CFR
Part 264, Subpart X. Permit applications for miscellaneous units must describe the wastes to be managed and
assess the potential environmental impacts associated with the proposed waste management activities. A listing
of the EPA Hazardous Waste Numbers that may be associated with the waste to be emplaced in the
miscellaneous unit is contained in Part A of the WIPP RCRA Permit Application.2 This listing was determined by
a survey of the generator/storage sites' TRU waste inventories and includes such RCRA-regulated constituents
as:
toxic characteristic contaminants listed in 20 NMAC 4.1, Subpart II, Paragraph 261.24, Table 1
(corresponding to 40 CFR 261, Subpart C, Paragraph 261.24) except for pesticides,
1 Retrievably stored waste is defined as waste generated after 1970 and before implementation of the QAPP
characterization requirements.
" Newly generated waste is defined as waste generated after implementation of QAPP characterization
requirements.
155
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
F-listed solvents (F001, F002, F003, F004, F005, F006, F007, and F009) found in 20 NMAC 4.1, Subpart I,
Paragraph 262.31 (corresponding to 40 CFR 261, Subpart D, Paragraph 261.3 1), and
hazardous constituents included in 20 NMAC 4.1, Subpart H, Paragraph 261, Appendix VII (corresponding to
40 CFR 261, Appendix VIII).
Waste is characterized on a waste stream basis. A waste stream is defined as waste material generated from a
single process or from an activity that is similar in material, physical form, and hazardous constituents. Wastes
are initially categorized into three broad Summary Category Groups that are related to the final physical form of
the wastes. These groups include homogeneous solids (Summary Category S3000), soil/gravel (Summary
Category S4000), and debris wastes (Summary Category S5000). Waste streams are grouped by Waste Matrix
Code Groups related to the physical and chemical properties of the waste. Generator/storage sites must use the
characterization techniques described in the WAP to assign appropriate Waste Matrix Code Groups for WIPP
disposal. The Waste Matrix Code Groups are solidified inorganics, solidified organics, salt waste, soils,
lead/cadmium metal, inorganic nonmetal waste, combustible waste, graphite, filters, heterogeneous debris
waste, and uncategorized metal.
A statistically selected portion of waste containers from waste streams of homogeneous solids and soil/gravel will
be sampled and analyzed for total volatile organic carbons (VOCs), semi-volatile organic compounds (SVOCs),
and metals. TRU waste classified as debris wastes will be characterized based on acceptable knowledge.
Acceptable knowledge refers to applying knowledge of the hazardous characteristic of the waste in light of the
materials or processes used to generate the waste. The use of acceptable knowledge is outlined in a guidance
manual10 wherein the EPA has specifically referred to the characterization of radioactive mixed waste as a
situation where the use of acceptable knowledge is appropriate.
Since waste containers will not be opened to perform confirmatory sampling at the WIPP site, waste
characterization data produced at the site is reviewed at three levels to ensure it meets Program needs and
objectives. At the data generation level, data are reviewed, validated, and verified. Data packages are submitted
to the project level for validation and verification. The third and final level is the CAO level at which data from
the project level are verified.
Data review determines if the raw data have been properly collected and ensures raw data are properly reduced.
Data validation confirms that the data reported satisfy the requirements defined by the user (e.g., quality
assurance objectives and data quality objectives) and is accompanied by signature release. Validation at each
level ensures that certain aspects of characterization and quality assurance have been met. Data verification
authenticates that data are in fact that which is claimed.
EPA Basis for Waste Characterization
An estimate of each generator/storage site TRU waste inventory was compiled in the TRU Waste Baseline
Inventory Report,11 modeled by the performance assessment,3 and evaluated using a sensitivity analysis to
determine the impact of each waste component on the long term performance of the repository. The results of
the sensitivity analysis identified several parameters that must be monitored and tracked to assure the validity of
assumptions used in the performance assessment. These parameters are listed in an Appendix to the
Compliance Certification Application (CCA)3 entitled "Waste Component Limits" (WCL). Being that Appendix
WCL identifies the radionuclide content of the waste as one of the components to be monitored and tracked,
radiological characterization of the waste by radioassay is of concern to the EPA.
ACCEPTABLE KNOWLEDGE
The use of acceptable knowledge is discussed in the EPA document, "Waste Analysis: EPA Guidance Manual
for Facilities that Generate, Treat, Store and Dispose of Hazardous Wastes."10 This document points out that
there are situations where it may be appropriate to apply acceptable knowledge to characterize hazardous waste.
One of these, which is applicable to the type of waste that will be accepted for disposal in the WIPP, is that the
physical nature of the waste does not lend itself to the acquisition of a representative sample. Additionally, the
RCRA regulations codified in 40 CFR Parts 260 through 265, 268, and 270, authorize the use of acceptable
knowledge as a method which can be used in appropriate circumstances by waste generators, or treatment,
storage, or disposal facilities to make hazardous waste determinations. Acceptable knowledge, as an alternative
to sampling and analysis, can be used to meet all or part of waste characterization requirements under RCRA.
156
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Acceptable knowledge refers to applying knowledge of the waste based on the materials or processes used to
generate the waste. Acceptable knowledge includes information regarding the physical form of the waste, the
base materials composing the waste, the nature of the radioactivity present, and the process generating the
waste. Acceptable knowledge is used to assign matrix parameter categories and EPA hazardous waste numbers
to waste streams and to determine the waste material parameters and radionuclides present in waste streams.
The collection and use of acceptable knowledge information applies to both retrievably stored and newly
generated waste streams.
To satisfy the characterization requirements of the WAP, acceptable knowledge is confirmed by radiography,
drum headspace gas sampling and analysis, and solidified waste sampling and analysis. Acceptable knowledge
is primarily used in TRU waste characterization activities to delineate TRU waste streams, to determine if TRU
debris wastes exhibit a toxicity characteristic (40 CFR 261.24), and to determine if TRU wastes are listed (40
CFR 261.31). The physical form and the increased health and safety risks associated with obtaining a
representative sample of TRU debris wastes, clearly justify the use of acceptable knowledge in making
hazardous waste determinations.
DOE complex waste generators apply knowledge of their waste based on the materials and processes used to
generate the waste. A generator site can establish the characterization of a waste stream by demonstrating an
understanding of the materials which are introduced into the process, and the process(es) which those materials
undergo. Understanding the process(es) which a material may undergo is very important, particularly with
respect to toxicity characteristic wastes, because some chemical processes may result in a change in
concentration of the RCRA constituents. Assignment of F-listed wastes also depends on knowledge of the
process that produces the waste. A change in concentration of the RCRA constituents during a process could
result in a new waste stream containing constituents with concentrations that are above the regulatory threshold.
In the case of certain compounds which may be either listed or toxicity characteristic, the use of acceptable
knowledge is the only viable route to determine whether a substance was utilized for its solvent properties or not.
If a constituent was used for its solvent properties, the waste stream would be assigned an F code. If the
constituent was not utilized for its solvent properties, it would be evaluated for the assignment of a D code, if
appropriate.
The Land Withdrawal Act1 included a number of requirements and restrictions on the wastes that can be
disposed of at the WIPP Among these requirements and restrictions are that the waste be generated by atomic
energy defense activities and that it is neither high-level waste nor spent nuclear fuel. DOE can verify
compliance with these conditions only through the use of historical information about the processes that
generated a particular waste stream. This historical information is a component of the acceptable knowledge
record that is assembled and assessed by the waste generators.
Consistency among sites in using acceptable knowledge information to characterize TRU waste involves a three
phase process: 1) compiling the minimum acceptable knowledge documentation in an auditable record; 2)
confirming acceptable knowledge information using radiography, and headspace gas and solidified waste
sampling and analysis; and 3) auditing acceptable knowledge records. The consistent presentation of acceptable
knowledge among sites in auditable records will allow WIPP personnel to verify the completeness of acceptable
knowledge and determine that the accuracy of acceptable knowledge has been documented for TRU waste
characterization during the audit process.
AUDITING ACCEPTABLE KNOWLEDGE RECORDS
The DOE/CAO audits the TRU waste generators to grant TRU waste certification authority to the generators. The
DOE/CAO conducts audits at least annually thereafter to verify ongoing compliance with approved plans and
procedures including those for waste characterization using acceptable knowledge. The information resulting
from DOE/CAO audits is used by the DOE/CAO to help sites across the DOE complex prevent or solve problems
associated with the compilation and use of acceptable knowledge for TRU waste characterization. Past and
current issues with respect to the compilation, use, and defensibility of acceptable knowledge for TRU waste
characterization along with the solutions to those problems and their prevention are addressed in this section.
The Audit Process
CAO conducts an initial audit of each site to evaluate waste stream and program documentation prior to
certifying the site for shipment of TRU waste to the WIPP facility. The initial audit establishes a baseline that will
be reassessed annually. The audits are used to ensure the consistent compilation, application, and interpretation
157
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
of acceptable knowledge information throughout the DOE complex and to evaluate the completeness and
defensibility of site-specific acceptable knowledge documentation related to hazardous waste determinations.
Acceptable knowledge audit checklists typically include, but are not limited to, the following elements for review
during the audit:
Documentation of the process used to compile, evaluate, and record acceptable knowledge is available and
implemented
Personnel training and qualifications are documented
All of the required acceptable knowledge documentation has been compiled in an auditable record
Procedures exist for:
assigning EPA hazardous waste numbers to waste streams
assigning a matrix parameter category to a waste stream
determining waste material parameters present in a waste stream
determining the radionuclides present in a waste stream
resolving inconsistencies in acceptable knowledge documentation
confirming acceptable knowledge information through: a) radiography or visual examination, b)
headspace gas sampling and analysis, and c) solidified waste sampling
Results of other audits of the TRU waste characterization programs at the site are available in site records
Auditors assess all documents associated with the evaluation of the acceptable knowledge documentation for at
least one debris waste stream and one solidified waste stream during the audit. For these waste streams,
auditors review all procedures and associated processes developed by the site for: documenting the process of
compiling acceptable knowledge documentation; correlating information to specific waste inventories; assigning
EPA hazardous waste numbers; assigning matrix parameter categories; determining waste material parameters;
determining the radionuclides; and identifying, resolving, and documenting discrepancies in acceptable
knowledge records. The adequacy of acceptable knowledge procedures and processes is assessed and any
discrepancies in procedures are documented in the audit report.
Auditors review the acceptable knowledge documentation for selected waste streams for logic, completeness,
and defensibility. The criteria used by auditors to evaluate the logic and defensibility of the acceptable knowledge
documentation include completeness and traceability of the information, clarity of presentation, degree of
compliance with the requirements of the QAPP and WAP with regard to acceptable knowledge confirmation
data, nonconformance procedures, and oversight procedures. Auditors evaluate compliance with written site
procedures for developing the acceptable knowledge record. A completeness review is done to evaluate the
availability of the minimum required TRU waste management and TRU waste stream information. Records are
reviewed to assess the correlations to specific waste streams and to assess the basis for making waste
determinations. Auditors verify that sites include all required information and conservatively assign all potential
EPA hazardous waste numbers indicated by the acceptable knowledge records. All deficiencies found in the
acceptable knowledge documentation are included in the audit report.
Auditors verify and document that sites use management controls and follow written procedures to make waste
determinations for newly generated and retrievably stored wastes. Auditors review procedures used by sites to
confirm acceptable knowledge information using radiography or visual examination, headspace gas sampling
and analysis, and solidified waste sampling and analysis. Procedures to document changes in acceptable
knowledge documentation, EPA hazardous waste number assignments, matrix parameter category assignments,
waste material parameter determinations, and radionuclide determinations to specific waste streams are also
evaluated.
After the audit is complete, CAO prepares a final audit report that includes all observations and findings identified
during the audit. Sites are required to respond to all audit findings and identify corrective actions. If acceptable
knowledge procedures do not exist, the minimum required information is not available, or findings of
158
-------
WTQA '98 - J 4th Annual Waste Testing & Quality Assurance Symposium
noncompliance associated with waste determinations are identified, CAO will not grant the site characterization
and certification authority for the subject waste. Waste stream characterization and certification authority may be
revoked or suspended if findings during subsequent annual audits indicate a lack of compliance with approved
acceptable knowledge procedures.
Acceptable Knowledge Issues
To date, the Department of Energy, Carlsbad Area Office has certified three generator/storage sites: Los Alamos
National Laboratory (September 1997), Rocky Flats Environmental Technology Site (March 1998), and Idaho
National Engineering and Environmental Laboratory (April 1998). These sites have undergone initial and final
audits of their certification and acceptable knowledge processes and have successfully met all WAP criteria for
characterizing their TRU waste via acceptable knowledge. Issues that the audit team encountered during audits
of the adequacy, implementation, and effectiveness of the sites' acceptable knowledge procedures/processes are
summarized below.
Adequacy
Procedural inadequacy was a common problem encountered by the audit team. The WIPP requires the
"proceduralization" of every aspect of the acceptable knowledge process, from compilation and evaluation of the
acceptable knowledge documentation, to reconciliation of discrepancies in the documentation, to waste stream
designation and assignment of EPA hazardous waste codes. The following procedural inadequacies were noted
by the audit team:
The responsible party(ies) for some tasks identified in the procedures was not specified
A list of the documents generated as a result of implementing the procedure was not specified
Proper approvals for procedures were not obtained prior to implementation
Approvals for procedures were not adequately documented
Various requirements were omitted from procedures, such as a specification to look for prohibited waste
items
Necessary or required procedure steps, or forms, or references to other related procedures were not
specified
Implementation and Effectiveness
If the acceptable knowledge process procedures are adequate and complete, the satisfactory implementation of
these procedures results in accurate and defensible characterization of the waste by the generator, i.e. an
effective process. When procedures are not implemented correctly, the result is an incorrect determination, or an
inability to characterize the waste using acceptable knowledge, or an ineffective process. The audit team noted
several deficiencies with respect to the implementation and effectiveness of acceptable knowledge process
procedures:
Various processes, such as that for amending acceptable knowledge records, were not established
Some procedures did not reflect current practice
Some procedures were not being property implemented
The acceptable knowledge documentation was not traceable to source documents the acceptable knowledge
documentation was not compiled into an auditable package or record
Some sites that subcontracted the acceptable knowledge process work did not include requirements for
qualification/training of personnel or a requirement to prepare and train the site personnel in the statement of
work for the subcontractor
The required analytical data and/or acceptable knowledge documentation was not available or insufficient to
support the conclusions or EPA hazardous waste code assignments
Retrieval of records (source documents, etc.) was difficult, time consuming, or impossible
Documentation of training for various acceptable knowledge processes (review of acceptable knowledge
documentation, etc.) was non-existent or insufficient
EPA hazardous waste codes were removed without sufficient documentation that a discrepancy process was
followed
Acceptable knowledge documentation contains statements that are not corroborated by the source
documentation
159
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The process for the assignment of EPA hazardous waste codes was not consistent with QAPP/WAP
requirements
The acceptable knowledge summary report did not list all potential constituents as shown in supporting
documentation
Documentation to justify that the waste was generated as a result of atomic energy defense activities was
insufficient
Acceptable knowledge documentation for radionuclide distribution was not adequate to support radioassay
data to confirm the radionuclide inventory and obtain total activity in TRU waste
The EPA hazardous waste code for a constituent identified in the acceptable knowledge documentation as
being present in the waste was not assigned
Inconsistencies or discrepancies between acceptable knowledge documentation and/or analytical data were
not reported (packaging configuration, volume of waste, etc.)
CONCLUSIONS
Acceptable knowledge provides a reasonable and appropriate means to characterize nuclear waste for disposal.
Acceptable knowledge is the only route by which certain characterization decisions (e.g. the origin of the waste
being from defense activities, the use of solvents) may be made. The DOE and the waste generator sites have
developed and implemented an adequate system for the collection, assessment, and utilization of acceptable
knowledge for characterizing waste to be disposed of at the Waste Isolation Pilot Plant.
ACKNOWLEDGEMENTS
Portions of this work were conducted under contract to Sandia National Laboratories. Sandia is a multiprogram
laboratory operated by Sandia Corporation, a Lockheed Martin Company for the United States Department of
Energy Contract DE-AC04-94-AL85000. This work was conducted under the Sandia WIPP Quality Assurance
Program which is equivalent to NQA-1, NQA-2 (Part 2.7), and NQA-3 Standards.
REFERENCES
1. Waste Isolation Pilot Plant Land Withdrawal Act, Public Law 102-579, 1992.
2. "Resource conservation and Recovery Act Part B Permit Application for the Waste Isolation Pilot Plant," U. S.
Department of Energy (DOE), 1997, DOE/WIPP 91-005, Revision 6.4.
3. Waste Isolation Pilot Plant Compliance Certification Application, 40 CFR Part 191/194, DOE/CAO-2056,
Carlsbad, New Mexico, U. S. Department of Energy.
4. Federal Facility Compliance Act, Public Law 102- 386, Title 1,'3021(d).
5. "Waste Isolation Pilot Plant No-Migration Variance Petition," DOE/WIPP-89-003, Revision 1, Carlsbad, New
Mexico, U. S. Department of Energy, 1990.
6. "Waste Isolation Pilot Plant Disposal Phase Draft Supplemental Environmental Impact Statement,"
DOE/EIS-0026-S-2, Carlsbad, New Mexico, U. S. Department of Energy, 1996.
7. "National TRU Waste Management Plan. DOE Carlsbad Area Office: Leadership in Safe and Efficient
Cleanup of Transuranic Waste," DOE/NTP-96-1204, Revision 1, Carlsbad, New Mexico, U. S. Department of
Energy, 1998.
8. "Transuranic Waste Characterization Quality Assurance Program Plan," CAO-94-1010, Revision 0, Carlsbad,
New Mexico, U. S. Department of Energy, 1995.
9. "Waste Acceptance Criteria for the Waste Isolation Pilot Plant," WIPP-DOE-069, Revision 5, Change Notice
#2, Carlsbad, New Mexico, U. S. Department of Energy, 1991.
10. "Waste Analysis at Facilities that Generate, Treat, Store, and Dispose of Hazardous Waste; A Guidance
Manual," EPA-530-R-94-024, Washington D.C., Office of Solid Waste and Emergency Response, U.S. EPA,
1994.
11. "Waste Isolation Pilot Plant Transuranic Waste Baseline Inventory Report," CAO-94-1005, Revision 3,
Carlsbad, New Mexico, U. S. Department of Energy, 1995.
160
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
PERFORMANCE EVALUATION SOIL SAMPLES FOR VOLATILE ORGANIC COMPOUNDS
UTILIZING SOLVENT ENCAPSULATION TECHNOLOGY
James Dahlgran
U.S. Department of Energy, Idaho National Engineering and Environmental Laboratory,
850 Energy Drive, NIS 4149, Idaho Falls, ID 83401
Curt Thies
Thies Technology, 3720 Hampton, Suite 207, St.Louis, MO 63109-1438
Abstract
A mixture of volatile organic compounds (VOCs) was encapsulated and mixed with soil to produce a product
suitable for use as a double blind source of VOCs in soil performance evaluation sample. Two independent
laboratories analyzed the standard encapsulated VOC/soil mixture for benzene, toluene, ethylbenzene and
xylene by using U.S. EPA SW-846 Method 5030 in conjunction with SW-846 Method 8020. One laboratory
received the sample as a single blind standard while the other laboratory received the sample as a double blind
standard. The percent relative standard deviation (%RSD) for triplicate analyses ranged from 2.4 to 7.7. The
lowest %RSD was for meta/para-xylene (2.4%) from the sample analyzed as a double blind sample. Analytical
results from these pilot studies indicate that it is possible to prepare standard soil samples contaminated with
known amounts of VOCs, which unlike current market technology, will enable soil samples to be submitted to
environmental analytical laboratories as a truly blind sample.
Introduction
For most environmental analytical procedures, demonstrating proficiency of an analytical method is
accomplished utilizing known spiked samples, blanks, surrogate spikes and appropriate performance evaluation
standards. For analysis of volatile organic compounds (VOCs) in soils, the performance evaluation standards are
primarily methanolic solutions that contain target analytes and are spiked into a soil sample immediately prior to
analysis.
Demand for precise performance evaluation samples for VOCs in soil matrices has stagnated due to lack of
sample preparation technology. Current technology for preparing volatile performance evaluation samples utilize
solvent spiking procedures1 or vapor fortification methodologies developed by J. Hewitt1 of the U.S. Army Cold2
Regions Research and Engineering Laboratory. Private sector companies typically provide analysts a dilute
solution of analytes in a solvent, usually methanol. These methanol solutions are introduced either into the
analytical technique (purge and trap or headspace analyses) or placed onto sand3 (used to simulate soil)
immediately prior to instrumental analysis. These practices do not adequately replicate soil sample handling
procedures in the analytical laboratory. Liquid standards also inappropriately provide the analysts with an
opportunity to analyze the spiking solution.
Although vapor-fortified soils provide a means of examining spiked soils that are analogous to soils isolated from
the environment, such standards are difficult to disseminate to analytical laboratories as a double blind quality
control standard. An ideal VOCs in soil standard should provide analyte concentrations across the concentration
range of 5 ug/kg to 100 ug/g. Vapor fortification methodology can be customized to decrease concentrations
below 100 ug/g, but it is less amenable to water soluble analytes such as acetone or 2-butanone and some target
analytes are lost due to the varying absorbtivity of vastly differing soil matrixes. Early attempts to spike,
homogenize and transfer soil performance standards were unsatisfactory.4
To create a true volatiles in soil performance evaluation standard:
1. Volatile target analytes of interest must be unknown to the analyst.
2. Target analytes must be provided to the analyst in the soil matrix.
3. Volatile components must be protected from potential soil biological activity.
4. The standard must be stable over an extended period of time.
5. The target analytes must provide the laboratory with a wide range of VOC concentrations that are accurate
and reproducible independent of analytical methodology applied.
In order to meet these specifications, a novel method of spiking native soil samples is required. The objective of
this work is to demonstrate that microencapsulated VOCs provide a novel method of spiking soils and offer an
improved means of assessing precision and accuracy of VOC analyses of contaminated soils reported by
different analytical laboratories.
161
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Experimental
Microcapsules loaded with a VOC mixture (1:1:1 (V/V/V) toluene, p-xylene and ethylbenzene) were formed by
complex coacervation5. Figure 1 is a diagram of a typical complex coacervation encapsulation protocol. The
shell of the microcapsules formed is primarily gelatin cross-linked with glutaraldehyde.
Standard contaminated soil samples were produced by dispersing a known weight of microcapsules loaded with
the VOC mixture into a known weight of soil. Dry subsurface soil taken from the Idaho National Engineering and
Environmental Laboratory (INEEL) was sieved (60-80 mesh) and mixed with microcapsules to form a uniform
soil/microcapsule mixture. Figure 2 shows the steps envisioned for the preparation of a standard performance
evaluation soil sample that unitizes microencapsulated VOCs.
VOCs in soil standards were characterized by gas chromatography. A Hewlett Packard 5880 Series II flame
ionization gas chromatograph was used. For toluene, ethylbenzene and p-xylene mixture, the column was a 30
meter X 0.53 mm J&W Scientific DB-624 capillary column. For gasoline, the column was a 30 meter X 0.32 mm
J&W Scientific DB-1 fused silica capillary column.
In order to analyze the VOC content of capsules and soil containing capsules, a known weight of sample was
equilibrated in methanol at room temperature for a finite period. The samples were gently swirled to assure
complete wetting of the soil with the solvent. The methanol solution obtained was assayed directly (in the case of
gasoline in soil) or diluted with acetone and subsequently assayed (in the case of 1:1:1
toluene/ethylbenzene/p-xylene). Equilibration time was 30 min. for all samples.
Results and Discussion
In order to have a stable standard soil sample contaminated with VOCs, it is necessary to be able to prepare in a
reproducible manner VOC-loaded microcapsules that are stable for prolonged periods when mixed with a soil
sample. For a microcapsule loaded with VOCs to be stable, the capsule shell or coating must have essentially
zero permeability to the VOCs encapsulated. The shell must also be susceptible to water and/or methanol, since
the analytical methodology used to characterize VOC contaminated soil involves these solvents. Dry
microcapsules shells formed by complex coacervation are able to retain VOCs for prolonged periods as long as
they are not subject to high humidity storage conditions. In the presence of water or water/methanol mixtures, the
predominately gelatin shells become permeable thereby enabling release of the VOCs.
The first objective of the study was to demonstrate that a mixture of pure solvents could be encapsulated and
retained. The solvent mixture encapsulated was a 1:1:1 (V/V/V) mixture of toluene, ethylbenzene and p-xylene.
Three batches of capsules were made in order to examine lot-to-lot reproducibility. The capsules produced were
clean, uniform free-flowing powders suitable for incorporation into a soil sample. Triplicate analyses of 2 gram
samples of each capsule batch established that the capsules were 95.8 weight % solvent (relative percent
standard deviation; 1.0%). When 5.08 grams of one capsule sample, fraction passing 60 mesh screen, was
combined with 100.2 grams clean INEEL subsurface soil, the contaminated soil produced contained
approximately 56.5 milligrams (estimated by calculation) of toluene/ethylbenzene/p-xylene mixture per gram of
soil.
In order to examine homogeneity and accuracy this standard contaminated soil sample, three one-gram and
three five-gram samples of contaminated soil were assayed. Table 1 summarizes the replicate analytical data
obtained. Overall variability at one standard deviation was about 5% for all components at either the 1.0 gram or
5.0 gram aliquot size. This single analysis indicates that the concept of preparing an accurate performance
standard by utilizing microcapsules is realistic. It demonstrated adequate homogeneity in the sample at various
sample sizes as well as relative agreement with the amount of material expected (estimated 56.5 mg/gram vs.
found 48.6 mg/gram).
A second pilot involved the encapsulation of gasoline and preparation of soil sample contaminated with gasoline.
Four batches of capsules loaded with gasoline were prepared. The mean gasoline loading of these four samples
was 73.2% (standard deviation 6.2). Three ~2 gram samples of gasoline-loaded capsules were subjected in
triplicate to analysis and found to 3274 ug/gram toluene (std. dev. = 248, %RSD = 7.6), 4347 ug/g ethylbenzene
(std. dev. = 296, %RSD = 6.8) and 14922 ug /g p-xylene (std. dev. 931, %RSD = 6.2).
Ten grams of gasoline-loaded capsules were combined with 193.0 grams of soil to produce a pilot performance
evaluation sample. Three samples of this contaminated soil sample placed into individual 40 milliliter,
162
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposia
precleaned sampling vials each containing approximately 20 grams, were shipped to an
laboratory for analysis for benzene, toluene, ethylbenzene and xylenes (BTEX) Each
triplicate utilizing U.S. EPA SW-846 Method 5030 in conjunction with SW846 Method
results of these analyses.
environmental analytical
sample was analyzed in
8020. Table 2 contains
Another sample was sent as
a double blind sample to a
second analytical laboratory
for analysis. Table 3) sum-
marizes results of these
analyses. Other than a dis-
agreement concerning the
quantitation of benzene,
results obtained indepen-
dently by the two labora-
tories agree well.
Figure 1. Coacervation of
Solvents for Addition to Soil
Sample
core particle dispersed in solution of
polymer by agitation
Coacervation visible as droplets of colloid-rich
phase induced by one or more agents and Deposition
of coacervation droplets on surface of core
Mergence of Coacervation droplets to form the coating
KEY:
o
- Coacervation
ฐo droplets
O
o
Coating
Hardened coating
Shrinkage and crosslinking of the coating to rigidize it as necessary.
Raf: Deasy. Patrick B . Microencapsulation and Related Drug Processes Marcel Dekker, Inc
Mix core material (solvents) to appropriate
ratios
\
The analytical data obtained to date demonstrates using microcapsules to prepare soil performance evaluation
standards is attainable. The performance evaluation soil standard has the advantage of being as close to a real
world soil sample as technologically feasible and has demonstrated the ability to measure the quality associated
with the entire analytical methodology. This technology is capable of producing a performance evaluation sample
containing volatile target organic analytes within well-defined concentration ranges. The encapsulated process
can provide variable
concentration ranges
for numerous target
volatile analytes. The
resulting performance
evaluation sample will
be amenable to many
of the current US EPA
methodologies for the
analysis of environ-
mentally significant
volatile organics in
soils including analy-
sis carried out by
thermal desorption
techniques.
Emulsify to form small droplets
Form coating or interfacial polymerization
Mieroeapsule suspended in polymer forming phase
i
Filer, wish and recover microcapsulea
T
Mix capsules with diluvnl soil In appropriate
ratio to generate required standard
Figure 2. Typical
Steps of Standard
Preparation Process
Current efforts are focused on reducing the concentration of target analytes to the 5 - 200 ppb range and to
examine product storage stability. Once through this pilot examination, the carrier solvent will be spiked with
other analytes including; acetone, carbon tetrachloride, chloroform, methylene chloride, trichloroethane,
tetrachloroethylene. The author has submitted soil samples that contain microencapsulated VOCs to the Mixed
Analyte Performance Evaluation Program, operated by the U.S. Department of Energy, for round robin study
utilizing contracted environmental analytical laboratories.
163
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1. Data from Simulated Contaminated Soil Sample that Contains Microcapsules Loaded with a 1:1:1
(V/V/V) Toluene/ Ethylbenzene/p-Xylene Mixture, (units are milligrams per grain of soil)
Sample Weight (g.)
Toluene
Ethylbenzene
p-Xylene
1.0
5.0
Average
Std. Dev.
%RSD
Average
Std. Dev.
%RSD
11.96
0.57
4.8
12.00
0.62
5.1
22.21
1.05
4.7
22.78
1.16
5.1
22.41
1.08
4.8
23.02
1.17
5.1
Table 2. Gasoline in Soil Pilot Performance Evaluation Standard - Single Lab Analysis, (units are milligrams per
gram of soil)
Benzene
Toluene
Ethylbenzene
o-Xylene
m & p-Xylene
Sample 1
Sample 2
Sample 3
Average
Std. Dev.
%RSD
444
406
407
367
402
416
387
384
406
402.1
20.5
5.1
143
134
130
131
130
133
131
124
131
131.9
4.7
3.6
187
206
168
192
198
205
197
191
200
193.8
10.9
5.6
341
306
303
267
286
280
298
261
287
292.1
22.5
7.7
898
846
841
792
793
828
814
766
836
823.8
36.3
4.4
Table 3. Gasoline in Soil Pilot Performance Standard Sent as Double Blind to Second Lab for Analysis, (units
are milligrams per gram of soil)
Benzene
Toluene
Ethylbenzene
o-Xylene
m & p-Xylene
Sample 1
Average
Std. Dev.
%RSD
Sample 1
Confirmation
Average
Std. Dev.
%RSD
2.5 U
2.5U
2.5 U
2.5 U
NA
NA
NA
NA
120
120
99
100
109.8
10.2
9.3
120
120
120
120
120
0
150
150
120
110
130
16.7
12.9
200
210
220
210
210
7.1
3.4
270
270
230
250
255
16.6
6.5
350
360
380
350
360
12.2
3.4
750
760
610
680
700
60.4
8.6
900
940
960
920
930
22.4
2.4
Literature Cited
1. Conversations with RTC, ERA and APG.
2. Hewitt, A.D.; Clarence, L.G., Environ. Sci. Technol. 1995, 29, 769-773.
3. Environmental Research Associates Product Catalogue, 1997, 16.
4. Maskarinec, M.P.; Johnson, L.H.; Bayne, C.K., J Assoc. Off. Anal. Chem. 1989, 72, 823-827.
5. Deasy, Patrick B., "Microencapsulation and Related Drug Processes", Marcel Dekker, Inc., 1984.
164
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
U.S. EPA AND U.S. A.F. INTERAGENCY AGREEMENT FOR FIELD ANALYTICAL SERVICES
Raymond A. Flores
US EPA Region 6, Houston Community College, 10625 Fallstone Road, Houston TX 77099-4303
281-983-2139 ext. 139
FLORES.RAYMOND@EPAMAIL.EPA.GOV
George H. Lee
Analytical Services Division, AL/OEA, Armstrong Laboratory, Brooks AFB, TX 78235-5114
210-536-6166
GLEE@GUARDIAN.BROOKS.AF.MIL
The interagency agreement (IAG) between the parties above will provide for more timely analyses of EPA
samples while at the same time making more efficient use of the federal government's resources. At the time of
the writing of this abstract the IAG has just reached implementation phase and consequently data is limited. We
are confident that with the data generated in the next four months, we will be able to show that time and money
are saved by the use of this IAG. As the Project officers for the EPA-AF IAG, we will use the data from each field
project in making the cost savings comparison. We will discuss project startup and implementation. The onsite
analytical capabilities of the Air Force, both field screening techniques and field confirmation methods, will be
examined. USAF's Armstrong Laboratory located at Brooks Air Force Base, San Antonio Texas, will provide
onsite analyses for USEPA Region 6's Superfund Division. The elements of the Superfund Program that can
request work are the Brownfields, Removal, Emergency Response, Remedial, and Site Assessment Teams. The
Air Force Responders will be onsite in a time frame dependent upon the nature of the site activity. Most site
activities are non-emergency. Activities which are not time critical such as Brownfields and Remedial for
example, will allow sufficient time for a significant amount of preplanning for onsite activities. Other activities
which are more time critical require immediate action. Air Force responds to those activities in which there is a
threat or possible threat of risk to human health in less than 24 hours dependent on site location. In the near
future, AF is projected to have a mobile laboratory which can be driven to the site immediately upon notice.
Onsite field screening and confirmation techniques have been projected to be a more cost effective way to
perform portions of the chemical analysis evaluation of a site as early as 1988. Air Force's trained personnel and
specialized equipment can be used for EPA projects in EPA Region 6 and of course for Air Force's needs
worldwide. Air Force benefits in a number of ways such as becoming familiarized with EPA's SOPs for site
activities. EPA benefits by having access to a cost-effective service with unique capabilities and world wide
experience. This IAG provides for technology transfer and more efficient resource utilization.
SUMMARY
This agreement between Federal Agencies pursuing common goals allows for greater efficiency and cost
effectiveness for all and provides an opportunity for technology transfer. We recommend the use of interagency
agreements in similar cases. An IAG can be a commensal relationship with mutual advantages.
SEVERAL ORGANIC PARAMETERS ON UNDERLYING HAZARDOUS CONSTITUENTS LIST
CAN NOT BE MEASURED AT THE UNIVERSAL TREATMENT STANDARDS
Howard C. Johnson
Sample Management Risk Technologies
Clifford S. Watkins
Sample Management Risk Technologies, Idaho National Engineering and Environmental Laboratory,
Sample Management Office, Lockheed Martin Idaho Technologies Company,
P.O. Box 1625-3960, Idaho Falls, Idaho 83415
ABSTRACT
The Idaho National Engineering and Environmental Laboratory (INEEL) has several permitted treatment, storage
and disposal facilities. The INEEL Sample Management Office (SMO), operated by Lockheed Martin Idaho
165
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Technologies Company (LMITCO), conducts all analysis subcontracting activities for Department of Energy
Environmental Management programs at the INEEL. In this role, the INEEL SMO has had the opportunity to
subcontract the analyses of various wastes (including ash from an interim status incinerator) requesting a target
analyte list equivalent to the constituents listed in 40 Code of Federal Regulations (CFR) ง 268.48. Per 40 CFR ง
268.40, these analyses are required to ensure that treated wastes do not contain underlying hazardous
constituents (UHC) at concentrations greater than the universal treatment standards (UTS) prior to land disposal.
The language in 40 CFR ง 268.40 (d) (3) states that, "The treatment or disposal facility may demonstrate
compliance with organic constituents if good-faith analytical efforts achieve detection limits for the regulated
organic constituents that do not exceed the treatment standards specified in this section by an order of
magnitude."
The INEEL SMO has conducted this "good-faith effort" by negotiating with several commercial laboratories to
identify the lowest possible quantitation and detection limits that can be achieved for the organic LTHC analytes.
The results of this negotiating effort has been the discovery that no single laboratory (currently under subcontract
with the INEEL SMO) can achieve a detection level that is within an order of magnitude of the UTS for all
organic parameters on a clean sample matrix (e.g., sand). This indicates that for a typical waste sample, the
chances of the order of magnitude requirement not being met for many more than just the "problem analytes" is
likely. This does not mean that there is no laboratory that can achieve the order of magnitude requirements for
all organic UHCs on a clean sample matrix. The negotiations held to date indicate that it is likely that no
laboratory can achieve the order of magnitude requirements for a difficult sample matrix (e.g., an incinerator
ash). The authors suggest that the regulation needs to be revised to address the disparity between what is
achievable in the laboratory and the regulatory levels required by the UTS.
INTRODUCTION
The INEEL SMO conducts all analysis subcontracting activities for the Department of Energy Environmental
Management programs at the INEEL. Contracted analyses are primarily in the following analytical disciplines;
radiological, inorganic, organic and physical properties testing. Within each discipline numerous analytical tests
are requested on a large variety of sample matrices. Analytical test requirements range from field screening or
processing information to data required to satisfy U.S. Environmental Protection Agency (EPA) and Idaho
Division of Environmental Quality (ID-DEQ) requirements. The INEEL SMO subcontracts analytical work on a
variety of wastes (including ash from an interim status incinerator). Analytical requests have included a target
analyte list equivalent to the constituents listed in 40 CFR ง 268.48. Per 40 CFR ง 268.40, these analyses are
required to ensure that treated wastes do not contain UHCs at concentrations greater than the UTS prior to land
disposal. The language in 40 CFR ง 268.40 (d) states:
"Notwithstanding the prohibitions specified in paragraph (a) of this section, treatment and disposal
facilities may demonstrate (and certify pursuant to 40 CFR ง 268.7(b)(5)) compliance with the treatment
standards for organic constituents specified by a footnote in the table "Treatment Standards for
Hazardous Wastes" in this section, provided the following conditions are satisfied:
(1) The treatment standards for the organic constituents were established based on incineration in units
operated in accordance with the technical requirements of 40 CFR part 264, subpart 0, or based on
combustion in fuel substitution units operating in accordance with applicable technical requirements;
(2) The treatment or disposal facility has used the methods referenced in paragraph (d)(1) of this section
to treat the organic constituents; and
(3) The treatment or disposal facility may demonstrate compliance with organic constituents if good-faith
analytical efforts achieve detection limits for the regulated organic constituents that do not exceed
the treatment standards specified in this section by an order of magnitude."
The INEEL SMO has sought this "good-faith effort" by negotiating with several commercial laboratories to
identify the lowest possible quantitation or detection limits that can be achieved for the organic UHCs. The
primary emphasis has been on the nonwastewater standard. At this time, the wastewater standard for the organic
UHCs has not been performed by all of the INEEL SMO contracted laboratories. The results of this negotiating
effort has been the discovery that no single laboratory (currently under subcontract with the INEEL SMO) can
achieve the detection level of the UTS for all organic parameters of the UHC list on a clean sample matrix (e.g.,
sand). This indicates that for a typical waste sample, the chances of the order of magnitude requirement not
being met for many more than just the "problem analytes" is likely. This does not mean that there is no laboratory
that can achieve the order of magnitude requirements for all organic UHCs on a clean sample matrix. The INEEL
SMO continues to seek laboratory capability to analyze for the UHCs to the UTS in a cost-effective manner.
166
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Additionally, the regulated community should be aware that the "order-of-magnitude" provision might not be
allowed on incinerator ash from an interim status incinerator. In the September 19 1994 final rule the
"order-of-magnitude" provision is only applicable where incinerated wastes were treated in permitted (Part 264
Subpart O) units.
Within the capabilities of the current INEEL SMO subcontracted laboratories, no two laboratories analyze the
UHCs to the UTS the same way. In the guidelines of USEPA SW-846 a variety of approved methods can be
used to analyze for the various constituents. The SW-846 methods being used to analyze for the complete UHC
list by the laboratories currently contracted through the INEEL SMO are:
Method 1311 Toxicity Characteristic Leaching Procedure
Method 8015A Nonhalogenated Volatile Organics by Gas Chromatography
Method 8015M (modified) Nonhalogenated Volatile Organics
Method 8080A Organochlorine Pesticides And Polychlorinated Biphenyls By Gas Chromatography
Method 8081 Organochlorine Pesticides, Halowaxes And PCBs As Aroclors By Gas Chromatography:
Capillary Column Technique
Method 8081A Organochlorine Pesticides (PCBs) By Gas Chromatography
Method 8082 Polychlorinated Biphenyls By Gas Chromatography
Method 8140 Organophosphorus Pesticides
Method 8141 Organophosphorus Compounds By Gas Chromatography: Capillary Column Technique
Method 8150 Chlorinated Herbicides By Gas Chromatography
Method 8151 Chlorinated Herbicides By GC Using Methylation Or Pentafluorobenzylation Derivatization:
Capillary Column Technique
Method 8260A Volatile Organic Compounds By Gas Chromatography Mass Spectrometry (GC/MS):
Capillary Column Technique
Method 8270B Semivolatile Organic Compounds By Gas Chromatography/Mass Spectrometry (GC/MS):
Capillary Column Technique
Method 8310 Polynuclear Aromatic Hydrocarbons
Method 8316 Acrylamide, Acrylonitrile And Acrolein By High Performance Liquid Chromatography (HPLC)
The UHC list is extensive and there are many "problem analytes" to analyze. The methodology is left to the
capability of the laboratory and the laboratory's discretion. It would not be practical or reasonable for the INEEL
SMO to specify to a laboratory a required method for every constituent. Under the fixed contracts, obsolete
methods are still listed but current methodology is used whenever possible, incorporating the June 13, 1997
promulgated methods from the Third Edition of the EPA-approved test methods manual "Test Methods for
Evaluating Solid Waste, Physical/Chemical Methods."
To procure UHC organic analyses, it becomes necessary to negotiate with each contracted laboratory concerning
achievable detection and quantitation limits for the entire UHC list. The negotiations held to date indicate that it is
likely that no laboratory can achieve the order of magnitude requirements for a difficult sample matrix (e.g., an
incinerator ash). When negotiating with commercial laboratories that are under fixed price subcontracts
government contractors like LMITCO have difficulty authorizing additional financial support, for the "research
project" samples. Analytical difficulties arise when dealing with complex matrices such as an incinerator fly ash.
For example, alternate solvent system extractions may be more effective for the semivolatile organic
compounds. Laboratories currently contracted through the INEEL SMO routinely extract semivolatiles using
acetone/hexane or acetone/methylene chloride systems. They have found fewer fly ash matrix interferences
when they substituted a methylene chloride only extraction system. The alternate solvent system also requires
additional calibration with a resulting burden on the laboratory under a fixed price contract.
The organic UHC list is extensive and standards are hard to locate for the entire list. For example, one of the
contracted laboratories had difficulty initially procuring a standard for 4,4'-Methylene-bis-(o-chloroaniline) CAS#
101-14-4. No laboratory contracted through the INEEL SMO has determined quantitation limits or calibrated for
the 40 additional analytes that will become "underlying hazardous constituents" August 26, 1998. Commercial
laboratories contracted through the INEEL SMO have expressed concerns about the effort and expense that will
be required to obtain standards, conduct MDL determinations, etc. in order to analyze for the added constituents.
The 40 additional constituents are identified in Table 1 UTS.
167
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
SUMMARY
The current INEEL SMO contracted laboratories operate within the highly competitive commercial laboratory
industry. With laboratory capacity available to DOE (i.e. possessing a NRC license to receive radioactively
contaminated samples) limited nationwide, extensive "non-routine" analytical requests on complex matrices are
not desired by the laboratories. Obtaining the full suite of UHC analyses can be difficult. The inability to
determine if a treated waste has achieved the required treatment standard concentrations can place programs in
a position subject to violation of regulatory requirements. The authors' suggest that the regulation needs to be
revised to address the disparity between what is achievable in the laboratory and the regulatory levels required
by the UTS.
The authors' poster delineates the achievable detection limits for organic UHC parameters at several commercial
laboratories. The intent of the poster is to bring this issue to the attention of the regulators and regulated
community. Another benefit is to discuss analytical approaches used by other laboratories that may achieve
greater sensitivities for the difficult organic analytes on the UHC list.
Table 1 UTS identifies the organic hazardous constituents, along with the nonwastewater and wastewater
treatment standard levels. 40 CFR ง 268.48 also states: "For determining compliance with treatment standards
for underlying hazardous constituents as defined in ง 268.2(i), these treatment standards may not be exceeded.
Compliance with these treatment standards is measured by an analysis of grab samples, unless otherwise noted
in the following Table 1 UTS". Included in the table are the negotiated quantitation limits for the nonwastewater
standard from three INEEL SMO contracted laboratories (A, B, and C). These limits can be compared to the
listed regulatory nonwastewater standard concentration. Presently the most difficult ash samples are requiring
five to ten fold dilutions -in addition to alternate solvent system extractions- which further limit the ability of the
laboratory to achieve quantitation limits at (or below) the UTS for all of the UHCs. The negotiated quantitation
limits are the laboratories "ideal" quantitation limits on a clean matrix (e.g., sand) and do not represent the
current achievable limits on difficult sample matrices (e.g., incinerator fly ash). When the laboratories
quantitation limit is at the treatment standard concentration, the INEEL SMO requires the laboratory to have the
instrument detection limit for that analyte less than 0.33 times the treatment standard. This limit is also specified
in the task order statement of work between the laboratory and the INEEL SMO.
Table 1. UTS - Universal Treatment Standards
(Note: NA means not applicable.)
Regulated
constituent/common
name
Organic Constituents
A2213(6)
Acenaphthylene
Acenapthene
Acetone
Acetonitrile
Acetophenone
2-Acetylaminofluorene
Acrolein
Acrylamide
Acrylonitrile
Aldicarb sulfuric (6)
Aldrin
4-Aminobiphenyl
Aniline
Anthracene
Aramite
alpha-BHC
beta-BHC
delta-BHC
gamma-BHC
Barban (6)
Bendiocarb (6)
Bendiocarb phenol (6)
Benomyl (6)
Benzene
CAS(1)
Number
30558-43-1
208-96-8
83-32-9
67-64-1
75-05-8
96-86-2
53-96-3
107-02-8
79-06-1
107-13-1
1646-88-4
309-00-2
92-67-1
62-53-3
120-12-7
140-57-8
319-84-6
319-85-7
319-86-8
58-89-9
101-27-9
22781-23-3
22961-82-6
17804-35-2
71-43-2
Wastewater
standard
Concentration
in mg/l (2)
0.042
0.059
0.059
0.28
5.6
0.010
0.059
0.29
19
0.24
0.056
0.021
0.13
0.81
0.059
0.36
0.00014
0.00014
0.023
0.001
0.056
0.056
0.056
0.056
0.14
Nonwastewater
standard
Concentration in
mg/kg (3)
unless noted as
"mg/l TCLP"
1.4
3.4
3.4
160
38
9.7
140
NA
23
84
0.28
0.066
NA
14
3.4
NA
0.066
0.066
0.066
0.066
1.4
1.4
1.4
1.4
10
Negotiated
Nonwastewater
Quantitation Limit
INEEL SMO
contracted
Laboratory A
undetermined
0.33
0.33
0.01
0.02
0.33
0.33
0.01
50
0.01
undetermined
0.005
0.33
0.33
0.33
0.66
0.005
0.005
0.005
0.005
undetermined
undetermined
undetermined
undetermined
0005
Negotiated
Nonwastewater
Quantitation Limit
INEEL SMO
contracted
Laboratory B
undetermined
0.67
0.67
0.01
0.05
0.67
0.67
NA
23
0.05
undetermined
0.066
NA
0.67
0.67
NA
0.002
0.004
0.006
0.002
undetermined
undetermined
undetermined
undetermined
0.005
Negotiated
Nonwastewater
Quantitation Limit
INEEL SMO
contracted
Laboratory C
undetermined
3.4
3.4
160
38
9.7
140
NA
23
84
undetermined
0.066
NA
14
3.4
NA
0.066
0.066
0.066
0.066
undetermined
undetermined
undetermined
undetermined
10
168
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Benz(a)anthracene
Benzal chloride
Benzo(b)fluoranthene
(difficult to distinguish
from benzo(k)
fluoranthene
Benzo(k)fluoranthene
(difficult to distinguish
from benzo(b)
fluoranthene)
Benzo(gji,i)perylene
Benzo(a)pyrene
Bromodichloromethane
Bramomethane/Methyl
bromide
4-Bromophenyl phenyl
ether
n-Butyl alcohol
Butylate (6)
Butyl benzyl phthalate
2-sec-Butyl-4,6-
dinitrophenol/Dinoseb
Carbaryl (6)
Carbenzadim (6)
Carbofuran (6)
Carbodfuran phenojJ6}
Carbon disulfide
Carbon tetrachloride
Carbosulfan (6)
Chlordane (alpha and
gamma isomers)
p-Chloroaniline
Chlorobenzene
Chlorobenzilate
2-Cloro-1 ,3-butadiene
Chlorodibromomethane
Chloroethane
bis(2-chloroethoxy)
methane
bis(2-Chloroethyl)ether
Chloroform
bis(2-Chloroisopropyl)
ether
p-Chloro-m-cresol
2-Chloroethyl vinyl ether
Chloromethane/Methyl
chloride
2-Chloronaphthalene
2-Chlorophenol
3-Chloropropylene
Chrysene
o-Cresol
m-Cresol (difficult to
distinguish from
p-cresol)
p-Cresol (difficult to
distinguish from
m-cresol)
m-Cumenyl
methylcarbamate (6)
Cyclohexanone
o,p'-DDD
p,p'-DDD
o,p'-DDE
p,p'-DDE
o,p'-DDT
p,p'-DDT
Dibenz(a,h)anthracene
Dibenz(a.e)pyrene
56-55-3
98-87-3
205-99-2
207-08-9
191-24-2
50-32-8
75-27-4
74-83-9
101-55-3
71-36-3
2008-41-5
85-68-7
88-85-7
63-25-2
10605-21-7
1563-66-2
1563-38-8
75-15-0
56-23-5
55285-14-9
57-74-9
106-47-8
108-90-7
510-15-6
126-99-8
124-48-1
75-00-3
111-91-1
111-44-4
67-66-3
39638-32-9
59-50-7
110-75-8
74-87-3
91-58-7
95-57-8
107-05-1
218-01-9
95-48-7
108-39-4
106-44-5
64-00-6
108-94-1
53-19-0
72-54-8
3424-82-6
72-55-9
789-02-6
50-29-3
53-70-3
192-65-4
0.059 I
0.055
0.11
0.11
0.0055
0.061
0.35
0.11
0.055
5.6
0.042
0.017
0.066
0.006
0.056
0.006
0.056
3.8
0.057
0.028
0.0033
0.46
0.057
0.10
0.057
0.057
0.27
0.036
0.033
0.046
0.055
0.018
0.062
0.19
0.055
0.044
0.036
0.059
0.11
0.77
0.77
0.056
0.36
0.023
0.023
0.031
0.031
0.0039
0.0039
0.055
0.061
3.4
6.0
6.8
6.8
1.8
3.4
15
15
15
2.6
1.4
28
2.5
014
1.4
014
1.4
4.8 mg/l TCLP
6.0
1.4
0.26
16
6.0
NA
28
15
6.0
7.2
6.0
6.0
7.2
14
NA
30
5.6
5.7
30
3.4
5.6
5.6
5.6
1.4
0.75 mg/l TCLP
0.087
0.087
0.087
0.087
0.087
0.087
8.2
NA
0.33
1.65
0.33
0.33
0.33
0.33
0.005
0.01
0.33
0.5
undetermined
0.33
1.65
undetermined
undetermined
undetermined
undetermined
.005 mg/l TCLP
0.005
undetermined
0.005
0.33
0.005
0.33
0.01
0.005
0.01
0.33
0.33
0.005
0.33
0.33
0.01
0.01
33
0.33
0.02
0.33
0.33
Reported as
p-cresol
0.33
undetermined
0.1 mg/l TCLP
0.01
0.01
0.01
0.01
0.01
0.01
0.33
0.01
0.67
0.05
0.67
0.67
0.67
0.67
0.05
0.05
0.67
2.6
undetermined
67
1.3
undetermined
undetermined
undetermined
undetermined
4.8 mg/l TCLP
0.005
undetermined
0.009
0.67
0.005
NA
0.05
0.005
0.05
0.67
0.67
0.005
0.67
0.67
NA
0.05
0.67
0.67
0.05
0.67
0.67
Reported as
p-cresol
0.67
undetermined
0.75 mg/l TCLP
0,087
0.087
0.087
0.087
0.087
0.087
0.67
NA
3.4
6.0
68
6.8
1.8
3.4
15
15
15
2.6
undetermined
28
2.5
undetermined
undetermined
undetermined
undetermined
4.8 mg/ TCLP
6.0
undetermined
0.26
16
6.0
NA
0.28
15
6.0
7.2
6.0
6.0
7.2
14
NA
30
5.6
5.7
30
3.4
5.6
Reported as
p-cresol
5.6
undetermined
0.75 mg/l TCLP
0.087
0.087
0.087
0.087
0.087
0.087
8.2
NA
169
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
1 ,2-Dibromo-3-
chloropropane
1 .2-Dibromoethane/
Ethylene dibromide
Dibromomethane
m-Dichlorobenzene
o-Dichlorobenzene
p-Dichlorobenzene
Dichlorodinuoromethane
1 , 1 -Dichloroethane
1 ,2-Dichloroethane
1 , 1 -Dichloroethylene
trans-1 ,2-
Dichloroethylene
2,4-Dichlorophenol
2,6-Dichlorop henol
2,4-Dichlo phenoxyacetic
acid/2,4-D
1 ,2-Dichloropropane
cis-1 ,3-
Dichloropropylene
trans-1 ,3-
Dichloropropylene
Dieldrin
Diethylene glycol,
dicarbamate (6)
Diethyl phthalate
p-Dimethylamino-
azobenzene
2,4-Dimethyl phenol
Dimethyl phthalate
Dimetilan (6)
Di-n-butyl phthalate
1 ,4-Dinitrobenzene
4,6-Dinitro-o-cresol
2.4-Dinitrophenol
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Di-n-octyl phthalate
Di-n-propylnitrosamine
1 ,4-Dioxane
Diphenylamine (difficult
to distinguish from
diphenylnitrosamine)
Diphenylnitrosamine
(difficult to distinguish
from diphenylamine)
1 ,2-Diphenylhydrazine
Disulfoton
Dithiocarbamates (total)
(6)
Endosulfan I
Endosulfan II
Endosulfan sulfate
Endrin
Endrin aldehyde
EPIC (6)
Ethyl acetate
Ethyl benzene
Ethyl cyanide/
Propanenitrile
Ethyl ether
96-12-8
106-93-4
74-95-3
541-73-1
95-50-1
106-46-7
75-71-8
75-34-3
107-06-2
75-35-4
156-60-5
120-83-2
87-65-0
94-75-7
78-87-5
10061-01-5
10061-02-6
60-57-1
5952-26-1
84-66-2
60-11-7
105-67-9
131-11-3
644-64-4
84-74-2
100-25-4
534-52-1
51-28-5
121-14-2
606-20-2
117-84-0
621-64-7
123-91-1
122-39-4
86-30-6
122-66-7
298-04-4
137-30-4
959-98-8
33213-46-9
1031-07-8
72-20-8
7421-93-4
759-94-4
141-78-6
100-41-4
107-12-0
60-29-7
0.11
0.028
0.11
0.036
0.088
0.090
0.23
0.059
0.21
0.025
0.054
0.044
0.044
0.72
0.85
0.036
0.036
0.017
.0.056
0.20
0.130
0.036
0.047
0.056
0.057
0.32
0.28
0.12
0.32
0.55
0.017
0.40
12.0
0.92
0.92
0.087
0.017
0.028
0.023
0.029
0.029
0.0028
0.025
0.042
0.34
0.057
0.24
0.12
15
15
15
6.0
6.0
6.0
7.2
6.0
6.0
6.0
30
14
14
10
18
18
18
0.13
1.4
28
NA
14
28
1.4
28
2.3
160
160
140
28
28
14
170
13
13
NA
62
28
0.066
0.13
0.13
0.13
0.13
1.4
33
10
360
160
0.02
0.02
0.01
0.33
0.33
0.33
0.01
0.005
0.005
0.005
0.005
0.33
0.33
10
0.005
0.005
0.005
0.01
undetermined
0.33
0.33
0.33
0.33
undetermined
0.33
1.65
1.65
1.65
0.33
0.33
033
0.33
0.33
0.33
0.33 cannot be
separated from
Dipheylamine
0.33
0.02
undetermined
0.005
0.01
0.01
0.01
0.01
undetermined
0.01
0.005
0.25
0.01
0.05
0.05
0.05
0.67
0.67
0.67
0.05
0.005
0.005
0.005
0.005
0.67
0.67
10
0.005
0.005
0.005
0.01
undetermined
0.67
NA
0.67
0.67
undetermined
0.67
1.6
3.3
3.3
0.67
0.67
0.67
0.67
0.5
13
1 3 reported as
Dipheylamine
NAfor
nonwastewater,
reported as
decomposition
product
azobenzene for
wastewater
0.03
undetermined
0.066
0.13
0.13
0.13
0.13
undetermined
33
0.005
360
160
15
15
15
6.0
6.0
6.0
7.2
6.0
6.0
6.0
30
14
14
10
18
18
18
0.13
undetermined
28
NA
14
28
undetermined
28
2.3
160
160
140
28
28
14
0.1
13
1 3 reported as
Dipheylamine
NA
6.2
undetermined
0.066
0.13
0.13
0.13
0.13
undetermined
33
10
360
160
170
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
bis(2-Ethytthexyl)
phthalate
Ethyl methacrylate late
Ethylene oxide '
Famphur
Fluoranthene
Fluorene
Formetanate
hydrochloride (6)
Formparanate (6)
Heptachlor
Heptachlor epoxide
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclo-
pentadiene
HxCDDs (All Hexa-
chlorodibenro-p-
dioxins)
HxCDFs (All Hexa-
chlorodibenzo-furans)
Hexachloroethane
Hexachloropropylene
Indeno (1 ,2,3-c,d)
pyrene
lodomethane
Isobutyl alcohol
Isodrin
Isolan (6)
Isosafrole
Kepone
Methacrylnitrile
Methanol
Methapyrilene
Methiocarb (6)
Methomyl (6)
Methoxychlor
3-Methylcholanthrene
4,4-Methylene
bls(2-chloroaniline)
Methylene chloride
Methyl ethyl ketone
Methyl isobutyl ketone
Methyl methacrylate
Methyl methansulfonate
Methyl parathion
Metolcarb (6)
Mexacarbate (6)
Molinate (6)
Naphthalene
2-Naphthylamine
o-Nitroaniline
p-Nitroaniline
Nitrobenzene
5-Nitro-o-toluidine
o-Nitrophenol
p-Nitrophenol
N-Nitrosodiethylamine
N-Nitrosodimethylamine
N-Nitroso-di-n-
butylamine
N-Nitrosomorpholine
N-Nitrosopiperidine
N-Nitrosomethyl-
ethylamine
N-Nitrosopyrrolidine
117-81-7
97-63-2
75-21-8
52-85-7
206-44-0
86-73-7
23422-53-9
17702-57-7
76-44-8
1024-57-3
118-74-1
87-68-3
77-47-4
NA
NA
67-72-1
1888-71-7
193-39-5
74-88-4
79-83-1
465-73-6
119-38-0
120-58-1
143-50-0
126-98-7
67-56 -1
91-80-5
2032-65-7
16752-77-5
72-43-5
56-49-5
101-14-4
75-09-2
78-93-3
108-10-1
80-62-6
66-27-3
298-00-4
1129-41-5
315-18-4
2212-67-1
91-20-3
91-59-8
88-74-4
100-01-6
98-95-3
99-55-8
88-75-5
100-02-7
55-18-5
62-75-9
924-16-3
59-89-2
100-75-4
10595-95-6
930-55-2
0.28
0.14
0.12
0.017
0.068
0.059
0.056
0.056
0.0012
0.016
0.055
0.055
0.057
0.000063
0.000063
0.055
0.035
0.0055
0.19
5.6
0.021
0.056
0.081
0.001
0.24
5.6
0.081
0.056
0.028
0.25
0.0055
0.50
0.089
0.28
0.14
0.14
0.018
0.014
0.056
0.056
0.042
0.059
0.52
0.27
0.028
0.068
0.32
0.028
0.12
0.40
0.40
0.40
0.40
0.013
0.40
0.013
28
160
NA
15
3.4
3.4
1.4
1.4
0.066
0.066
10
5.6
2.4
0.001
0.001
30
30
3.4
65
170
0.066
1.4
2.6
0.13
84
0.75 mg/l TCLP
1.5
1.4
0.14
0.18
15
30
30
36
33
160
NA
4.6
1.4
1.4
1.4
5.6
NA
14
28
14
28
13
29
28
2.3
17
2.3
35
2.3
35
0.33
0.33
10
0.1
0.33
0.33
undetermined
undetermined
0.005
0.005
0.33
0.33
0.33
0.0005
0.0005
0.33
0.33
0.33
0.01
0.1
0.1
undetermined
0.33
0.05
0.02
5 mg/l TCLP
0.33
undetermined
undetermined
0.05
0.33
1.65
0.005
0.01
0.01
0.33
0.33
0.02
undetermined
undetermined
undetermined
0.33
0.33
1.65
1.65
0.33
0.33
33
1.65
0.33
0.33
0.33
0.33
1.65 _j
0.33
0.33
0.67
0.05
NA
1.3
0.67
0.67
undetermined
undetermined
0.002
0.056
0.67
0.67
0.67
0.001
0.001
0.67
0.67
0.67
0.05
5
0.66
undetermined
0.67
1.3
0.05
Analyzed as total
10mg/kg,
equivalent to 0.5
mg/l TCLP
0.67
undetermined
undetermined
0.12
0.67
30
0.005
0.01
0.01
0.05
NA
0.06
undetermined
undetermined
undetermined
0.67
NA
3.3
3.3
0.67
0.67
0.67
3.3
0.67
67
0.67
0.67
1.3
0.67
0.67
28
160
NA
15
3.4
3.4
undetermined
undetermined
0066
0.066
10
5.6
2.4
0.001
0.001
30
30
3.4
65
170
0.066
undetermined
2.6
0.13
84
0.75 mg/l TCLP
1.5
undetermined
undetermined
0.18
15
30
30
36
33
160
NA
4.6
undetermined
undetermined
undetermined
5.6
NA
14
28
14
28
13
29
28
2.3
17
2.3
undetermined
2.3
undetermined
171
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Oxamyl (6)
Parathion
Total PCBs (sum of all
PCB isomers, or all
Aroclors)
Pebulate (6)
Pentachlorobenzene
PeCDDs (All
Pentachlorodibenzo-
p-dioxins)
PeCDFS (All Penta-
chlorodibenzo-furans)
Pentachloroethane
Pentachloronitrobenzene
Pentachlorophenol
Phenacetin
Phenanthrene
Phenol
o-Phenylenediamine (6)
Phorate
Phthalic acid
Phthalic anhydride
Physostigmine (6)
Physostigmine salicylate
(6)
Promecarb (6)
Pronamide
Propham (6)
Propoxur (6)
Prosulfocarb (6)
Pyrene
Pyridine
Safrole
Silvex/2,4.5-TP
1 ,2,4,5-Tetrachloro-
benzene
TCDDs (All
Tetrachlorodibenzo-p-
dioxins)
TCDFS (All Tetra-
chlorodibenzofurans)
1,1,1,2-
Tetrachloroethane
1,1,2,2-
Tetrachloroethane
Tetrachloroethylene
2,3,4,6-
Tetrachlorophenol
Thiodicarb (6)
Thiophanate-methyl (6)
Tirpate (6)
Toluene
Toxaphene
Triallate (6)
Tribromomethane/
Bromoform
1 ,2,4-Trichlorobenzene
1,1,1-Trichloroethane
1 ,1 ,2-Trichloroethane
Trichloroethylene
Trichloromonofluoro-
methane
2,4,5-Trichlorophenol)
2,4,6-Trichlorophenol
2,4,5-Trichlorophen-
oxyacetic acid/2, 4,5-T
23135-22-0
56-38-2
1336-36-3
1114-71-2
608-93-5
NA
NA
76-01-7
82-68-8
87-86-5
62-44-2
85-01-9
108-95-2
95-54-5
298-02-2
100-21-0
85-44-9
57-47-6
57-64-7
2631-37-0
23950-58-5
122-42-9
114-26-1
52888-80-9
129-00-0
110-86-1
94-59-7
93-72-1
95-94-3
NA
NA
630-20-6
79-34-5 1
127-18-4
58-90-2
59669-26-0
23564-05-8
26419-73-8
108-88-3
8001-35-2
2303-17-5
75-25-2
120-82-1
71-55-6
79-00-5
79-01-6
75-69-4
95-95-4
88-06-2
93-76-5
0.056
0.014
0.10
0.042
0.055
0.000063
0.000035
0.055
0.055
0.089
0.081
0.059
0.039
0.056
0.021
0.055
0.055
0.056
0.056
0.056
0.093
0.056
0.056
0.042
0.067
0.014
0.081
0.72
0.055
0.000063
0.000063
0.057
0.057
0.056
0.030
0.019
0.056
0.056
0.080
0.0095
0.042
0.63
0.055
0.054
0.054
0.054
0.020
0.18
0.035
0.72
0.28
4.6
10
1.4
10
0.001
0.001
6.0
4.8
7.4
16
5.6
6.2
5.6
4.6
28
28
1.4
1.4
1.4
1.5
1.4
1.4
1.4
8.2
16
22
7.9
14
0.001
0.001
6.0
6.0
6.0
7.4
1.4
1.4
0.28
10
2.6
1.4
15
19
6.0
6.0
6.0
30
7.4
7.4
7.9
undetermined
0.02
0.8
undetermined
0.33
0.0005
0.0005
0.33
1.65
1.65
0.33
0.33
0.33
undetermined
0.02
3.3
0.66
undetermined
undetermined
undetermined
0.33
undetermined
undetermined
undetermined
0.33
0.33
0.33
7.9
0.33
0.0005
0.0005
0.01
0.005
0.005
0.33
undetermined
undetermined
undetermined
0.005
0.5
undetermined
0.005
0.33
0.005
0.005
0.005
0.005
1.65
0.33
7.9
undetermined
0.03
10
undetermined
0.67
0.001
0.001
0.01
0.67
3.3
0.67
0.67
0.67
undetermined
0.02
28 reported as
Phthalic
anhydride
28
undetermined
undetermined
undetermined
0.67
undetermined
undetermined
undetermined
0.67
0.67
0.67
7.9
0.67
0.001
0.001
0.05
0.005
0.005
0.67
undetermined
undetermined
undetermined
0.005
0.16
undetermined
0.005
0.67
0.005
0.005
0.005
0.005
3.3
0.67
7.9
undetermined
4.6
10
undetermined
10
0.001
0.001
6.0
4.8
7.4
16
5.6
6.2
undetermined
4.6
undetermined
undetermined
undetermined
undetermined
undetermined
1.5
undetermined
undetermined
undetermined
8.2
16
22
7.9
14
0.001
0.001
6.0
6.0
6.0
7.4
undetermined
undetermined
undetermined
10
2.6
undetermined
15
19
6.0
6.0
6.0
30
7.4
7.4
7.9
172
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
1 ,2,3-Trichloropropane
1,1,2-Trichloro-1,2,2-
trifluoroethane
Triethylamine (6)
trls-(2,3-Dibromopropyl)
phosphate
Vernolate (6)
Vinyl chloride
Xylenes-mixed isomers
(sum of o-, on, and
Pj*ylene
96-18-1
76-13-1
10-144-8
126-72-7
1929-77-7
l~~ 75-01-4
133020-7
0,85
0.057
0.081
0.11
0.042
0.27
0.32
30
30
1.5
0.10
1.4
6.0
30
0.01
0.01
undetermined
3.3
undetermined
0.01
0.005
0.05
0.05
undetermined
0.33
undetermined
0.05
0.005
30
30
undetermined
undetermined
undetermined
6.0
30
Footnotes to Universal Treatment Standards Table:
(1) CAS means Chemical Abstract Services. When the waste code and/or regulated constituents we described as a combination of a chemical with
it's salts and/or esters; the CAS number is given for the parent compound only.
(2) Concentration standards for wastewaters are expressed in mg/l and are based on analysis of composite samples.
(3) Except for Metals (EP or TCLP) and Cyanides (Total and Amemable) the nonwastewater treatment standards expressed as a concentration
were established, in part, based upon incineration in units operated in accordance with the technical requirements of 40 CFR part 264, subpart
O, or 40 CFR part 265, subpart O, or based upon combustion in fuel substitution units operating in accordance with applicable technical
requirements. A facility may comply with these treatment standards according to provisions in See 268.40(d). All concentration standards for
nonwastewaters are based on analysis of grab samples.
(4) Both Cyanides (Total) and Cyanides (Amenable) for nonwastewaters are to be antlyzed using Method 9010 or 9012, found in "Test Methods for
Evaluating Solid Waste, Physical/Chemical Methods". EPA Publication SW-846, as incorporated by reference in 40 CFR 260.11, with a sample
size of 10 grams and a distillation time of one hour and 15 minutes.
(5) These constituents am not "underlying hazardous constituents" in characteristic news, according to the definition at See. 268.2(i).
(6) Between August 26,1997, and August 26, 1998, these constituents are not "underlying hazardous constituents" as defined at Sec. 268.2(i).
REFERENCES
Protection of the Environment, 40 Code of Federal Regulations.
Test Methods for Evaluating Solid Waste, Physical/Chemical Methods SW-846 Third Edition, United States
Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, D.C.,
December, 1996.
RCRA Regulations and Keyword Index, Elsevier Science Inc., New York, New York, 1997 Edition.
IGNITABILITY PERFORMANCE EVALUATION STUDY ARE YOUR
WASTE STREAMS BEING CORRECTLY CHARACTERIZED?
Lester J. Dupes. Rock J. Vita I e
Environmental Standards, Inc., 1140 Valley Forge Road, Valley Forge, Pennsylvania 19482-0911
Debra J. Caillouet
Stationary Environmental & Energy, Chrysler Corporation,
800 Chrysler Drive E, Auburn Hills, Michigan 48326-2757
ABSTRACT
Thirteen commercial laboratories were evaluated for consideration of providing waste stream characterization
support. An initial single-blind performance evaluation (PE) study was performed to determine the accuracy of
the laboratory-reported results when compared to known values.
The initial PE study was conducted for constituents typically analyzed for hazardous waste characterization.
These constituents included Toxicity Characteristic Leaching Procedure (TCLP) volatiles, TCLP semivolatiles,
TCLP metals, and ignitability. A review of the ignitability tests performed by SW-846 Method 1010 and 1020
exhibited a wide range of both positive and negative bias from the known flashpoint temperature of the pure
compound used as the PE sample. Results were compared to reproducibility criteria generated from American
Society of Testing Materials (ASTM) Method D93-96.
Because of the erratic ignitability results observed for many of the participating laboratories, the laboratories
were requested to perform a self-audit inspection of their current SOP and actual laboratory procedures. These
self-audit results were used to ensure inter-laboratory consistency and compliance to SW-846 Method 1010 and
Method 1020.
173
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
A second round of PE samples were subsequently submitted to each of the thirteen participating laboratories for
analysis. These results showed some improvement for several of the laboratories, whereas other laboratories
again reported results with a significant bias. The results of this study exhibited a wide range of ignitabilities
which would have represented incorrect waste characterization if the PE samples were actual waste stream
samples. This paper will focus on discussion of the PE sample results, the laboratory self-audit results, and
method compliance issues.
INTRODUCTION
A comprehensive laboratory evaluation process was performed to identify commercial laboratories to provide
waste characterization testing services. This evaluation included an initial technical survey of approximately 40
laboratories. The survey results were used to further reduce the potential candidates to 13 laboratories.
Laboratory audits specifically focused on waste characterization analyses and performance evaluation (PE)
samples for specific parameters of interest for the characterization of complex wastes was used to further
evaluate the 13 short-listed laboratories. The parameters included ignitability, TCLP volatiles, TCLP
semivolatiles, and TCLP metals. An initial set of single-blind PE samples were prepared as whole volume
samples by a reputable commercial PE provider.
A single-blind PE study sample is defined as a "test" sample in which the laboratory is aware that the sample
submitted is a PE sample but does not have knowledge relative to the analytes or true concentrations contained
in the PE. A single-blind sample permits the data user to better understand a laboratory's analytical accuracy and
by inference, draw conclusions on the accuracy of actual waste sample results.
The matrix of the ignitability PE sample was a pure compound. The TCLP volatile sample was an aqueous
matrix, and the TCLP metals sample was a soil matrix. The PE samples for all 13 laboratories were prepared
from the same lot number to further reduce variance and permit a direct and meaningful comparison between
laboratories. Whole-volume samples, which require no further dilution or preparation by the laboratory prior to
analysis, were chosen to further reduce variables caused by different technicians or chemists. The PE samples
were prepared and shipped on ice under Chain-of-Custody procedures to each laboratory directly from the
provider.
In addition, an actual complex industrial waste stream sample was also used in this PE study. This waste stream
was selected to provide the laboratories with a representative "real-world" sample that the laboratories may
receive as an actual sample for waste testing. In particular, a paint purge solvent was selected for testing. This
paint purge solvent is relatively homogeneous and has previously been characterized as hazardous due to a
flashpoint temperature below the regulatory limit and volatile concentrations above the regulatory limit.
To collect this sample, the PE provider furnished bottles to the industrial facility for packaging this solvent. This
sample was collected and shipped back to the PE provider for homogenization and repackaging for delivery to
the 13 laboratories. This PE sample was shipped by ground carrier and was not refrigerated. Since it was not
certified, this sample was to be evaluated only for interlaboratory comparison purposes and was analyzed for
ignitability, TCLP volatiles and TCLP metals.
Performance by the individual laboratories was evaluated not only for the results proximity to the certified values
but also for communications, data packaging and reporting, method compliance, and timeliness of deliverables.
These latter issues were evaluated using techniques and procedures addressed in a previous manuscript (Dupes
and Rose, 1995). The issues are not discussed any further herein.
INITIAL PERFORMANCE EVALUATION SAMPLE RESULTS
In general, the laboratories reported acceptable results for the TCLP volatile organic compounds and metals,
relative to the program requirements. However, ignitability results ranged from 114ฐF to 153ฐF, with only four
laboratories meeting acceptance criteria for this parameter. The ignitability PE sample was certified by the
provider of the pure compound to have a flashpoint of 140ฐF ฑ 5ฐF. This certified value represents the flashpoint
of the compound as determined by various manufacturers of this compound. The temperature of the PE sample
was specifically chosen to determine variance around 140ฐF This temperature is also defined as the regulatory
limit for characterization of the waste as hazardous. Wastes that flash at a temperature of <140ฐF are considered
hazardous and wastes that flash at >140ฐF are considered non-hazardous by the definition for ignitability (40
CFR Part 261).
174
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Ten of the laboratories reported the temperatures higher than the regulatory limit of 140ฐF. resulting in a
classification as non-hazardous. Results for the three laboratories classifying the PE sample as hazardous
reported results of 129ฐF, 114ฐF, and 120ฐF. Those temperatures are all well below the lower acceptance limits
of 135ฐF for the PE sample. Six other laboratories were slightly above the 145ฐF acceptance criteria. A summary
of the results is presented in Table 1.
The "real world" purge solvent sample submitted to the laboratories was consistently identified by all laboratories
as hazardous based on flashpoint. The reported flashpoints ranged from 66-76ฐF (several laboratories reported
the sample flashed at less than their reporting limit of room temperature).
The variance in the certified (pure compound) PE results may have been due to the use of thermometers which
have not been adequately calibrated against a National Institute of Standards and Technology (NIST) standard
thermometer (this is routinely noted in a number of audits recently performed by the authors), the lack of
correction for barometric pressure, or analytical variance of the methodologies used by the laboratories. Each of
the laboratories management or quality assurance staff were required to conduct an intensive self-audit review of
the PE sample analysis, general analysis conditions, and quality control procedures to determine potential
causes for the wide range of flashpoint temperatures obtained for the PE sample.
SELF-AUDIT QUESTIONS AND FINDINGS
Upon completion of the initial PE study, a self-audit questionnaire was generated for completion by each of the
participating laboratories. The audit questionnaire consisted of 26 questions obtained from a technical review of
SW-846 Methods 1010 and 1020A; ASTM Methods D93-80, D93-90, D93-96, D3278-96, and E502-84. These
questions included verification of the method used and referenced for analysis of the initial PE study sample,
sample storage conditions, instrument analysis conditions, analysis procedures, thermometer calibrations, quality
control measures, acceptance criteria, and corrective actions.
The questionnaire was provided to each of the laboratories for completion by laboratory management or quality
assurance personnel. In addition to providing written responses to the questions, a copy of the laboratory
standard operating procedure was requested for review.
A review of the thirteen laboratory responses and SOPs were conducted to determine possible reasons for the
significant variation observed. A summary of issues identified in the self-audit checklists by the authors is
presented below.
According to the original request for analysis, all laboratories were requested to perform Method 1020A.
Several laboratories did request a change to Method 1010, which was granted and documented. However,
many laboratories did not request this change. Whenever a method change is necessary, data users (viz.,
the client) should be contacted prior to implementing the method change.
Several laboratories used compounds (i.e. acetic acid, mixed xylene, and kerosene) that were not stated in
the method for quality control purposes. The laboratories were directed to use p-xylene conforming to the
ASTM specifications, as required by Method 1010. The acceptance criteria should be 81ฑ2ฐF as required by
ASTM D93-90 or 81ฑ1ฐF as required by ASTM D93-96.
Several laboratories did not store the ignitability PE sample at 4ฑ2ฐC. Although it could be argued that a pure
single-component compound used for the PE would be unaffected by volatile constituent loss, actual waste
samples are typically complex mixtures, and lighter fractions can be lost due to volatilization. All laboratories
were requested to store future samples scheduled for ignitability at 4ฑ2ฐC to reduce the possibility of loss of
lighter fraction volatile constituents.
Five laboratories had not calibrated or could not demonstrate that the thermometer used for ignitability
determination had been calibrated with a NIST-standard thermometer within the last year. One laboratory
noted a 2-5ฐF difference upon calibration of three different thermometers. In addition, a laboratory also noted
that after the recalibration a mercury gap in one thermometer was noted and that the gap increased as the
temperature increased. All laboratories were directed to recalibrate the thermometers on an annual basis. A
multi-point calibration should be performed to determine accuracy at several temperatures, and correction
factors must be taken into account when measuring flashpoint temperatures.
175
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Six laboratories did not correct for barometric pressure, as required by the ASTM D93 methods. Several of
the laboratories noted that the correction would be very small (< 0.2ฐF); however, one laboratory noted the
flashpoint temperature would increase by 0.9ฐF All laboratories were directed to correct for the ambient
barometric pressure of the laboratory, as required by ASTM methodologies referenced in Method 1010 and
Method 1020A.
Many of the laboratories did not report a duplicate result for the ignitability analysis by Method 1010.
Although not specifically stated in Method 1010 or ASTM D93, duplicate analyses of all samples should be
performed to comply with ASTM E502, which indicates that the average results for flashpoint should be
reported. This will also provide a consistent approach with respect to Method 1020A, which requires duplicate
analyses. Sufficient volume (>150 ml) should be collected or provided in order to conduct duplicate analyses
by Method 1010.
All laboratories indicated the Pensky Martens apparatus used for analysis by Method 1010 is located in a
hood to reduce surrounding drafts and for health and safety considerations. However, several laboratories
indicated that the hood was turned on during the actual testing, which can cause loss of volatiles through
drafts. Several laboratories indicated that problems have occurred when hoods containing the apparatus are
turned on during analysis. Laboratories were directed to limit the amount of draft around the apparatus.
Three laboratories did not meet p-xylene acceptance criteria of 81ฑ2ฐF (ASTM D9390) or 81ฑ1ฐF (ASTM
D93-96) prior to analysis of the PE sample. One laboratory reported an acceptance range for p-xylene of
80-91 ฐF. Two laboratories did not report p-xylene results on the logbook pages. These practices are not
acceptable. All p-xylene calibrations must meet acceptance criteria and be recorded prior to sample analysis.
Several laboratories have implemented the use of an additional laboratory control sample (LCS) near the
regulatory limit of 140ฐF This is not a method requirement but is recognized as a good laboratory practice.
One laboratory was increasing the rate of temperature of the sample by 2ฐF/minute ; prior to application of
the test flame. The laboratories were directed to follow the rate of temperature increase, as specified in the
ASTM D93 Method as 9-11ฐF/minute.
One laboratory reported that the sample was stirred during the actual application of the test flame to the
sample. Samples must not be stirred during the application of, the test flame to the vapor space, as required
by the ASTM method.
Several laboratories indicated the introduction of the test flame into the vapor space is maintained from
0.5-1.5 seconds. Analysts should attempt to consistently introduce the test flame into the vapor space for 1
second, as required by the ASTM protocol referenced by Method 1010.
One laboratory reported the ignitability results to the nearest tenth degree. One laboratory reported the raw
results on a stenopad, which did not include, the analysis test name, method number, analyst name,
laboratory name, or QC results. The laboratories were directed to report result to the nearest 1ฐF Adequate
documentation of quality control (QC) results, barometric pressure, analysis date, etc. must be maintained by
all laboratories through the use of formal logbooks.
Several laboratories had not reviewed their SOP within the last year to verify method compliance and actual
laboratory procedures. All laboratories were directed to review and update laboratory SOPs on an annual
basis to reflect the actual procedures performed and to monitor for continuing method compliance.
SECOND ROUND IGNITABILITY PE STUDY
Upon completion of the self-audit review, laboratories were notified of the findings and requested corrective
actions prior to performing the second ignitability PE sample study. Three ignitability PE samples were selected
for the second round study. The flashpoint of the PE samples was specifically designed by the authors to provide
a range of flashpoints. The flashpoint-certified values of Flashpoint Sample #1, Flashpoint Sample #2, and
Flashpoint Sample #3 were 140ฐF, 120ฐF, and 170ฐF, respectively. These three flashpoints provide a sample that
would be considered hazardous (120ฐF), a sample at the hazardous regulatory limit (140ฐF) and a sample that
would be considered non-hazardous (170ฐF). The actual compounds used for the PE samples were chosen and
prepared by the same commercial PE provider used for the initial study. The expected flashpoint temperature
176
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
represents the flashpoint for the compound as determined by various manufacturers of this compound. All PE
samples were pretested prior to use by the PE supplier by Pensky Martens closed-cup procedures (US EPA
SW-846 Method 1010) and were corrected for barometric pressure of 620 mm. All PE samples were contained in
glass bottles with Teflon-lined caps and were shipped on ice under Chain-of-Custody to the 13 participating
analytical laboratories. Sufficient volume (150 ml) was sent to each laboratory to permit duplicate analyses of the
PE sample.
ANALYTICAL RESULTS AND CRITERIA FOR EVALUATION
Reproducibility criteria were calculated based on ASTM D93-96, which provides precision and bias data for the
ASTM method. Reproducibility in this ASTM method is defined as "the difference between two single and
independent results obtained by different operators working in different laboratories on identical material that
would, in the long run, in the normal and correct operation of the test method, exceed the following values only
one case in 20" (ASTM D93-96).
where:
R=BX
B = 0.078
X = the reported result in ฐC
R = reproducibility
Since the ASTM method is unclear as to which temperature should be used in calculating the reproducibility
criteria, criteria for the three flashpoint PE samples were generated by using the expected flashpoint (converted
to degrees Celsius) in the equation listed above. The criteria calculated should represent a range encompassing
the expected value of the PE sample. The following criteria were calculated for the PE samples:
Sample Identifier Expected Flashpoint Reproducibilily Criteria
Flashpoint Sample #1 140ฐF ฑ 8ฐF
Flashpoint Sample #2 120ฐF ฑ 7ฐF
Flashpoint Sample #3 170ฐF ฑ 11 ฐF
According to the previous ASTM Method revision D93-90, reproducibility criterion of ฑ 6ฐF should be used for
evaluation of the analytical data. According to the ASTM D93-90 method, this criteria applies to all liquid
samples with flashpoint temperatures less than 220ฐF Evaluation of the data using the more stringent criteria in
D93-90 excluded only two more results when compared to the criteria calculated by Method D93-96. A summary
of the results are presented in Tables 2, 3 and 4.
All participating laboratories except one analyzed the three PE samples by SW-846 Method 1010. One
laboratory analyzed the samples by SW-846 Method 1020. Three provided results after the due date, missing the
specifically stated date as a requirement in the documentation accompanying the PE samples.
DISCUSSION OF SECOND ROUND IGNITABILITY PE STUDY RESULTS
Upon review of the PE sample results, the inter-laboratory accuracy of the results obtained from the laboratories
was observed to be variable. Four laboratories reported results which, when compared to the known flashpoint,
were greater than the Method D93-96 criteria for all analyses (except for the analysis of Flashpoint #1 sample by
one laboratory, which was within criteria). Reported results, when compared to the known flashpoint, ranged from
10ฐ to 70ฐF absolute difference from the known value of the PE samples. Results reported by one laboratory
were biased very low (-27ฐ to -70ฐF below the known flashpoints of the materials). Results reported by the
laboratories are unacceptable and may indicate a recurring problem relating to the analyses being performed by
these laboratories.
Two laboratories reported two of the three results outside the Method D93-96 criteria. When compared to the
known flashpoint, reported results ranged from 8ฐ to 12ฐF absolute difference from the known value of the PE
samples.
Five laboratories reported one of the three results outside the calculated Method D93-96 criteria. When
compared to the known flashpoint, reported results ranged from 7ฐ to 13ฐF absolute difference from the known
177
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
value of the PE samples. One laboratory reported duplicate results, but did not report and average result for
each PE sample. Finally, two laboratories reported results for all three of the PE samples within the Method
D93-96 criteria.
In general, a majority of the laboratories reported a positive bias when compared to the known flashpoints of the
samples. Furthermore, several laboratories did not follow the instructions provided after the self-audit. Two
laboratories did not report the results to the nearest 1ฐF Instead, the laboratories reported results to one decimal
place. In addition, two laboratories did not analyze all samples in duplicate as requested. Finally, one laboratory
sent the PE samples from the facility being considered for program inclusion, where ignitability is apparently no
longer performed, to an affiliated network laboratory. This transfer occurred without notification to the authors.
This network facility was not previously audited for this program.
SUMMARY
Based upon a review of the initial ignitability study, self-audit review, and subsequent PE study, the use of
ignitability data for actual waste testing should be highly scrutinized. The regulatory-approved methods require a
significant amount of operator experience in conducting tests that are consistent and method compliant.
Evaluating ignitability through audits and frequent double-blind PE studies should be implemented to verify
acceptable performance. Without continuous review of this method, inaccurate ignitability results could be
causing your waste to be disposed of improperly.
ACKNOWLEDGMENTS
Mr. Chuck Wibby and Environmental Resource Associates for providing technical assistance in identification and
evaluation of compounds for the ignitability performance evaluation study, providing assistance in collection,
homogenization, and repackaging of the Chrysler waste stream sample and supply of all PE samples in this
study.
Mr. Brian Miller, of Chrysler Corporation for collection, packaging, and shipment of the waste stream sample
used in this PE study.
REFERENCES
American Society for Testing and Materials. D93-96 Standard Test Methods for Flash-Point by Pensky-Martens
Closed Cup Tester. Philadelphia, PA:ASTM, 1996.
American Society for Testing and Materials. E502: Standard Test Method for Selection and Use of ASTM
Standards for the Determination of Flash Point of Chemicals by Closed Cup Methods. Philadelphia,
PA: ASTM, 1994.
American Society for Testing and Materials. D93-90: Standard Test Methods for Flash- Point by Pensky-Martens
Closed Cup Tester. Philadelphia, PA:ASTM, 1990.
American Society for Testing and Materials. D3278-81 Test Method for Flash-Point of Liquids by Setaflash
Closed Tester. Philadelphia, PA:ASTM, 1981.
American Society for Testing and Materials. D93-80: Standard Test Methods for Flash-Point by Pensky-Martens
Closed Cup Tester. Philadelphia, PA:ASTM, 1980.
Dupes, L. J. and G. Rose. "Conducting a Performance Evaluation Study -It's Not Just Analytical Results."
Eleventh Annual Waste Testing & Quality Assurance Symposium. Washington, DC, 1995.
United States Environmental Protection Agency. Code of Federal Regulations, EPA Regulations for Identifying
Hazardous Waste, 40 CFR Part 261, Subpart C 261.21. Continuously Updated.
United States Environmental Protection Agency. "Chapter Seven: Characteristics Introduction and Regulatory
Definitions." Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846, Third Edition,
Update III, Office of Solid Waste, December 1996.
United States Environmental Protection Agency. "Method 1020A Setaflash Closed-Cup Method for Determining
Ignitability." Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846, Third Edition,
Update I, Office of Solid Waste, July 1992.
United States Environmental Protection Agency. "Method 1010: Pensky-Martens Closed-Cup Method for
Determining Ignitability." Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846,
Third Edition, Office of Solid Waste, September 1986.
178
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1. Initial Ignitability Performance Evaluation Study
56789
Number of Participating LataoratortM
10
12 13
Acceptance Limits
Table 2. Ignitability Performance Evaluation Study PE Flashpoint Sample #1
56789
Number of Participating Laboratory*
10
11 12
Acceptance Units
179
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 3. Ignitability Performance Evaluation Study PE Flashpoint Sample #2
140
100
56789
Number of Participating Laboratories
10
11
12
Table 4. Ignitability Performance Evaluation Study PE Flashpoint Sample #3
120
56789
Number of Participating Laboratories
10
11 12
Acceptance (Junta
13
180
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
TECHNIQUES FOR IMPROVING THE ACCURACY OF CALIBRATION
IN THE ENVIRONMENTAL LABORATORY
Dennis A Edaerlev
Quanterra Environmental Services, 880 Riverside Parkway, West Sacramento, California 95605
Abstract
Consistent and reliable procedures for generating calibrations are essential to accurate laboratory results.
Unfortunately the interpretation of acceptable practice is often based on misunderstanding or derived from
practices commonly utilized in non-environmental methods, and therefore does not provide a reliable means for
maintaining data quality. This paper presents a demonstration that some common practices used in the
calculation and evaluation of calibration factors, including the use of unweighted regression and the associated
correlation coefficient, are inappropriate for environmental analysis due to high relative errors which result at the
low end of the curve. Alternate criteria for evaluation of calibration curves are proposed based on the Relative
Standard Error (%RSE). Statistical derivations and examples are presented to demonstrate how this approach
provides an improved measure for the evaluation of calibration data based on weighted regression. Other related
considerations for assessing acceptability of calibration data are also presented.
Introduction
Any analytical measurement must employ reference elements to ensure traceability to relevant basic quantities.
The quality of a calibration depends on the uncertainty of the reference, the appropriateness of the reference and
how well the calculation procedures match the requirements of the analysis. A majority of the methods employed
in the environmental laboratory are based on a relative calibration where standards of known content and
concentration are analyzed by a suitable detector. The responses of samples analyzed under the same
conditions are then used to calculate concentrations by numerical interpolation to a response curve from the
calibration standards.
The only criterion for initial calibration in SW-846 method 8000A1 reads, "If %RSD is less than 20 an average
calibration factor can be used otherwise data should be fitted to a curve." There are many examples of well
defined and reproducible calibrations which either due to nonlinearity, or a non-zero intersection of the axis will
not meet this 20% criterion for acceptability. Recognizing this, update III to SW-846, Method 8000B2 provides
additional direction on use mid evaluation of least squares regression, adding criteria for higher order curves.
However, several critical issues are not sufficiently considered and overall the current guidance remains
incomplete with regard to error weighitng and evaluation of acceptability.
Linear Calibration
The most commonly adopted option for handling calibration data, which is linear but does not meet the criteria
for averaging, involves the calculation of coefficients for a linear equation of the form:
y = Ax + B
Equation 1
4.0 -,
35 -
ง 3'ฐ '
X 2.5
3fc
720
w
ง 1.5
a
0.5 -
0.0 -
c
Average
%RSD = 15.0 ^ 9
^
.' Linear fit
L ,' ^ = 0.928
S
y
S-*
12345
Concentration (x)
Concentration is defined here as the independent variable (x) and
response as the dependent variable (y) in compliance with method
8000B. A least squares regression is employed and the value of
the correlation coefficient (r) or the coefficient of determination (r2)
is evaluated as a measure of acceptability. A value of 1.00
represents a perfect correlation. Generally in practice, a value of r2
greater than 0.990 is considered satisfactory.
The practical difficulty encountered in this approach is displayed in
Figure 1. For the example data set, the line based on an average
calculation (shown as a dotted line) easily meets the 20%
acceptance criteria for %RSD, yet r2, does not meet the criteria for
acceptability with a value of 0.928.
Figure 1
181
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
To understand this inconsistancy we must examine more closely the statistical difference between an average
calculation and the regression line. The average calibration factor is determined according to equation 2.
C = -n ฃ
Equation 2
Where:
C = Average Calibration factor
C, = Calibration factor for calibration level i (y/x,)
n = number of calibration levels
This and all following equations may also be adapted to internal standard methods by substituting the relative
response calculated as in Equation 3 for the measured response (y/).
./Relative _ ,/
Yi -yi
v/s
Equation 3
y, = response of target analyte
x/s = concentration of internal standard
y/s = response of internal standard
Graphically the average and associated error limits based on ฑ1
standard deviation are shown in Figure 2. This type of normalized plot
makes visual examination of the data more straight-forward3 as
values at the low end of the calibration are shown at the same relative
scaling as those toward the high end.
The average represents the value which will minimize the variance in
the calibration factors (s2^ for all calibration points4 For reference the
calculation of variance is shown in Equation 4.
Figure 2
1.1 1
15 0.85
" 0.9
0 0.85
"ง 0,6
& 0.7S
0 0.7
0.65
0.0
' -k +1 Std Dev
- -1 Std Dev - ^
"ซ
0 1.00 2.00 3.00 4.00 500
Concentration (x)
2 _
SC -
Equation 4
If y, is defined as the expected response for calibration level i from the relationship y, = x,C, by substitution the
variance can also be expressed as:
c2 -
Sc-
n-1
Equation 5
Using the method of least squares we can determine a mathematical relationship between the dependent and
independent variables which minimizes the residual variance. By definition the residual variance represents the
variability due to experimental error5 and does not include that contribution to variance which is attributable to
differences in the independent variable. The residual variance of yon x is defined as:
n-1
Equation 6
A comparison of equations 5 and 6 demonstrates that the average value is the same as a coefficient derived
from least squares regression if a weighting of 1/x2 is applied. The average calibration factor gives a result which
is identical to that produced by the method of least squares using (1/concentration2) weighting and intercept
182
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
forced through zero.
Alternatively a generalized equation for calculating the coefficient which fits the simple relationship y = Ax and
minimizes the residual variance without weighting is determined by substituting Ax for y in Equation 6 and setting
the derivative with respect to A equal to zero. This gives for the calibration factor:
Equation 7
For the example data set, calibration coefficients and residual variances are compared in table 1. The
coefficients determined from these two approaches are quite different and it is obvious that an average
calibration coefficient does not minimize the residua! variance of y on x.
Table 1
Calculation type CF
s2
Average
Linear fit
0.916
0.784
1.335
0.768
Transforming the vertical axis of Figure 2 to a non-normalized form
gives the plot shown in Figure 3. A comparable plot for the unweighted
regression line using the same set of example data is presented in
Figure 4. Both figures represent graphically, with the dashed lines,
effective error weighting based on ฑ1 standard deviation.
Average
12345
Concentration (x)
Linear Fit
1234
Concentration (x)
Figure 3
Figure 3 demonstrates that 1/x2 weighting gives a relationship that
emphasizes precision at the low end of the calibration range. In
environmental analyses frequently the objective is to ensure that
target analytes do not exceed defined regulatory or action limits,
hence reducing quantitation error at low concentrations is especially
important.
Figure 4
Deriving the best form for a calibration curve must include consideration of the weighting factors which are
appropriate to the requirements of the analysis. As was shown empirically for the single factor calibrations and
can be proven for the general case6, the value of the coefficients is sensitive to the weighting. Method 8000B
requires that at least three replicates at a minimum of 5 concentration levels are used to derive weighting factors
which are defined as the inverse of the standard deviation squared for each concentration. In practice this is an
especially burdensome requirement, both in the amount of data required and the computational difficulty which
results when using individually derived weighting values for each calibration level. Faced with the options in
method 8000B, most laboratories will probably choose curve fitting without any consideration of weighting
regardless of the potential negative impact on data quality.
For many methods in the environmental laboratory errors in measurement are proportional to the magnitude of
the parameter measured. Likewise it is common to consider percentage errors relative to the concentration as we
have demonstrated in the case of the average calibration calculations. Rather than perform multiple replicates
each time a calibration is run, the laboratory should make an initial determination of the relationship between
concentration and the standard deviation throughout the calibration range, Where a direct proportionality is
found, all subsequent calibrations should be based on the 1/concentration2 weighting which is consistent with
average calibration curves. Conversely, if standard deviations are of equal absolute magnitude throughout the
concentration range, calibrations should be unweighted and average calibration factors should not be used.
183
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
When the weighting is expressed as a function of the concentration most data systems are capable of perforating
the calculation.
Percent Relative Standard Error
The magnitude of the residual variance with error weighting applied will provide a measure of the experimental
error for the derived curve. Figure 5 shows the relationship between variance and calibration coefficient for a
fitting equation of the form: y = Ax.
The value of relative standard deviation defined according to Equation 8 has widespread acceptance as a
measure of the error associated with an average calibration.
1 ฅ*-c')'
%RSD = 1OOx^rxV ^ -,
c v ""1
Equation 8
\
Weighting = 1/xs
Weighting = 1
0.7
0.75
0.8 O.B5 0.6
Calibration (actor
Because standard deviation is equal to the square root of variance, it
can be shown that %RSD is also equal to the square root of the
weighted residual variance of y on x, calculated as a percentage,
using a derivation similar to that for Equation 5.
Likewise the coefficient of determination (r2) is normally, recognized
as an indication of error associated with regression curves.
Figure 5
Equation 9
( \2
Since for any set of data the I,(y-9) term is not dependent on the form of the calibration curve function or the
coefficients, the value of (1-r2) will be directly proportional to the unweighted residual variance as defined in
Equation 6.
It becomes apparent that for an average calibration, the coefficient of determination calculated will not provide
an optimum measure of error as r2 applies only to an unweighted least squares determination. The %RSD on the
other hand is only meaningful as a measure of error when applied to an average calibration. A generalized
indicator of error, Percent Relative Standard Error (%RSE), which can be applied to any form of weighted
regression function is derived similarly to %RSD making adjustment for the degrees of freedom in the
relationship by replacing the n 1 factor with n - p where p is an integer equal to the number of coefficients as
defined in Equation 10.
%RSE is equivalent to %RSD when calculated for the average. In Figure 6 the least squares line derived using
(1/concentration2) weighting is compared to the average line and the unweighted least squares line for the same
data set used in the previous figures. In this example the elimination of the zero point is responsible for an
increase in the %RSE for the weighted line (p = 2).
Equation 10
Where:
y, = Actual response of calibration level
y, = Calculated response from curve
p = number of terms in the fitting equation
(average = 1, linear = 2, quadratic 3)
n = number of calibration points
184
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
1234
Concentration (x)
The improved low-end accuracy for this weighted line compared to the
unweighted regression line is shown in Table 2. Concentrations
measured close to the lowest calibration point would be reported with
values approaching 100% lower by using an unweighted linear
calibration as compared to the weighted curve.
Table 2
% Error (calculated - True)
Concentration
0.2
0.5
1.0
2.5
5.0
Weiahted
1.7
1.2
2.7
16.7
28.7
Unweiahted
109
29.9
2.8
25.1
8.9
Figure 6
Only recently, in the update to SW-846 method 8000B and also in a draft of guidance for development or
modification of water methods7 has the EPA recognized the importance of considering weighting in calibration
calculations. These documents do not, however provide options for evaluating weighted regression fits. The
coefficient of determination (COD) used in method 8000B for evaluating polynomial curves is "weighted" only to
adjust for the degrees of freedom in the fitting equation and does not provide a measure of error which is suitable
for regression curves derived with error weighting.
Non-Linear Calibration
While it is reasonable to use the simplest mathematical relationship, which gives acceptable accuracy, it should
not always be assumed that an average or linear curve is preferred. Some detector systems commonly used in
the environmental laboratory are inherently non-linear. As an example the
electron capture detector (ECD) commonly utilized for its sensitivity to
chlorinated pesticides and herbicides can for most analytes provide
calibrations which are linear over a 20X concentration range. It is often
difficult to optimize detector conditions for multiple analytes with widely
varying electron affinities and even under ideal conditions many ECD
detectors show linearity within 20% RSD only over a narrow concentration
range. The calibration data plotted in Figure 7 was curved for enhanced
accuracy over a 32X range by applying a quadratic fit. While the %RSD for
the average was not acceptable at 44.9% the quadrati curve gives a
%RSE of only 3.7%.
Figure 7
MCPP (ECDI
Awrage
16 32
Concentration (x)
In lieu of higher order curve fitting the approach often defaults to reducing the calibration range, which normally
requires re-analysis of some or all calibration points as well as increasing the number of sample dilutions, both of
which will add to the cost of the analysis without necessarily providing significant improvement in quality. A
simple recalculation of the curve parameters based on a quadratic fit allows the full concentration range of the
calibration data to be utilized with improved accuracy. The use of second or higher order curves should not be
applied simply to achieve minor improvements in the %RSE that will allow an otherwise unacceptable curve to
be used. Justification for higher order calculations should be based primarily on an understanding of the
performance characteristics of the detector or the method. The use of higher order curve fitting should also not
be substituted for proper instrument maintenance, nor should an effort be made to extend calibrations beyond
the detector saturation level.
Evaluation of Calibrations
Whenever possible, data evaluation should be performed against rigid criteria that will prevent any tendency for
analyst bias to affect reported results and allow automated data validation processes to be implemented. In a
production-oriented laboratory setting, visual examination of every calibration curve is not routinely performed.
185
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Rather curves are evaluated against calculated criteria only. Modern data systems allow regression parameters
to be calculated with very little analyst effort. Although the %RSE calculation is not commonly provided it can
usually be implemented either in a user function or by exporting the results to a spreadsheet.
Toluene (PID)
10 20 30 40
Concentration (x)
14
0.4 0.6
Concentration (x)
The behavior of the calibration curve near the reporting limit should
also be a primary consideration for environmental methods.
According to the proposed update to SW-846 method 8000B, data
should be considered unreliable for instrument response less than 3
times the y-intercept from the curve if the value is positive, or less
than the concentration calculated from zero response if the y-intercept
is negative. The use of weighted curve fitting will greatly reduce the
possibility that the intercept exceeds these limits. The unweighted line
in Figure 8a appears to provide a very good fit to the data with r2 =
0.999. The expanded view near the low end of the calibration shown
in Figure 8b clearly demonstrates the improved fit for the weighted
line.
Figure 8a
Recommendations for the minimum number of data points needed to
prepare a calibration vary from one to fifteen depending on the
method and the order of the fitting equation. Normally five points are
considered adequate for linear curves. Some problems may be
corrected by elimination of data points from the calibration
calculation. This should only done if they are either the highest or
lowest concentration and the number of points remaining meets the
minimum requirements of the method. The analyst should also be
aware that removing the low point might adversely affect the reporting
limit, as quantitation must not be performed outside the concentration
range of the calibration.
Figure 8b
When second order curves are evaluated, the acceptability should
include an evaluation of the curve inflection points. In Figure 9, the
quadratic curve (solid line) provides a significant improvement in
%RSE over the weighted least squares line (dashed). However, the
y-value of the curve inflection for this line is below the response of the
highest data point, thus quantitation near the upper end of the
calibration range could give erroneously high results.
Figure 9
Freon113{ELCD)
12-
10 -.
3
5
64
2 -
5 10
Concentration (x)
15
As a rule, all points in the calibration curve should demonstrate a consistent relationship between concentration
and response (response increases with increasing concentration). According to SW-846 method 8000B "... the
curve must be continuous, continuously differentiable and monotonic over the calibration range."
Summary
The correlation coefficient as a criterion for evaluating regression curves does not apply to weighted regressions
that are necessary for accurate low concentration reporting of environmental data. The correlation coefficient is
also not consistent with the %RSD criterion used to evaluate average curves, often leading to calibration data
that is not acceptable for a linear regression based on the correlation coefficient but does meet %RSD criterion
of an average calibration.
The Percent Relative Standard Error (%RSE) provides an improved criteria for evaluation of calibration curves in
environmental laboratory methods. The advantages of this approach are as follows:
186
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
All curve types can be calculated with the same (1/concentration2) weighting applied, This places emphasis
on relative error with improved accuracy at concentrations near the reporting limit.
The %RSE criterion is consistent for evaluation of all curve fitting types. Interpretation of acceptability for
calibration curves is simplified with the same criterion applied to all curve types and no conflicting criteria.
The most appropriate curve fitting model can be applied to each set of calibration data with evaluation
criteria between different curve types directly comparable.
The evaluation of all calibration data should include, as a minimum, the following checks:
%RSE < maximum limit
concentration level of low standard < reporting limit
low point intersection values < reporting limit
Number of calibration levels meets method requirements
All points must be monotonically increasing
Also for second or higher order curves inflection points should not be within the calibration range
With the widespread availability of powerful computer hardware and software in the laboratory, it is unnecessary
to sacrifice data quality for the sake of simplification. Analysts should be familiar with the productivity and quality
benefits of least squares curve-fitting algorithms. Clients and regulators must understand the importance of
weighted curve fitting and the need for complete and consistant evaluation criteria, which will provide high quality
results for the lowest cost.
References
1. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW846, 3rd Edition, Final Update 1,
July 1992, Section 8000A
2. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW846, 3rd Edition, Revision 2,
December 1996, Section 8000B
3. R. Cassidy and M. Janoski, LC-GC 10(9), 692 (1992)
4. Allen L. Edwards, An Introduction to linear Regression and Correlation, W.H. Freeman and Company, 1976
5. Maurice G. Kendall and William R. Buckland, A Dictionary of Statistical Turns, Hafner Publishing Co., 1972
6. J.C. Miller and J. N, Miller, Statistics for Analytical Chemistry, John Wiley, 1984
7. Guide to Method Flexibility and Approval of EPA Water Methods, U.S. EPA, Office of Water, Draft:
December 1996.
8. Grant T. Wernimont, Use of Statistics to Develop and Evaluate Analytical Methods, Association of Official
Analytical Chemists, 1985
9. G. Bulmer, Principles of Statistics, Dover Publications, 1979
QUALITY CONTROL PROTOCOL FOR ANALYSIS OF DRUGS AND
EXPLOSIVES USING ION MOBILITY SPECTROMETRY
Juliana Homstead and Edward J. Poziomek
Old Dominion University, Department of Chemistry and Biochemistry, Norfolk, VA 23529
(757)440-4005
email: epoziome@ODU.EDU
Ion mobility spectrometry (IMS) is widely used as a screening tool in the detection of contraband drugs and
explosives. It is also gaining popularity as a semiquantitative instrument used in the research of these and other
compounds, including hazardous environmental contaminants. We have successfully used IMS in studies
involving cocaine and TNT, and have developed a quality control protocol to assess and ensure the quality of our
data. This protocol employs cocaine hydrochloride as the reference standard in the positive mode and
trinitrotoluene (TNT) as the reference standard in the negative mode. A five point calibration curve was
187
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
generated for each of the reference compounds in order to determine a concentration level suitable for quality
control (QC) check solutions. We have established peak amplitudes and reduced mobility constant (Ko) for the
QC check solutions that must be met each day before proceeding with analyses. Any deviation from these
criteria requires assessment of the problem and appropriate corrective action. We have found this procedure
helpful in maintaining data quality and in providing an early indication of potential problems.
REFERENCE MATERIAL FOR THE ANALYSIS OF METALS IN SLUDGE
Stuart J. Nagourney. Nicholas J. Tummillo, Jr., John Birri, Kenneth Peist, Bruce MacDonald and Jean S. Kane
INTRODUCTION
New Jersey Department of Environmental Protection (NJDEP) regulations limit the levels of discharged toxic
substances in sludge effluents by sewage treatment plants. Once maximum contaminant levels are established,
they become part of the facility's operating permit. When chemical analyses indicate that these permit levels are
exceeded, the NJDEP has statutory authority to assess significant monetary penalties.
Sludge samples vary widely in their physical and chemical composition, ranging from liquids with low dissolved
solids content and small quantities of organics and metals, to multi-phase samples and cakes with solids
contents often greater than 50 % and concentrations of metals and other constituents at percent levels. Since
permit limits for metals are based on the amount leached during mineral acid digestion, these acid extractable
concentrations, rather than total concentrations, are the values of interest. With several options for methods of
sample preparation and measurement being proposed by State and Federal agencies, comparability of data
among methods is a critical issue. Reference materials of similar matrix and composition to the various sludge
matrices, and having known metal levels with defined uncertainties, are essential to insure that an accurate
assessment of the data being generated is made,
To expand the utility of existing environmental SRMs, NIST has initiated efforts provide data on the
acid-extractable levels of metals in selected new solid sample environmental SRMs. A recent paper by these
authors described a collaborative project among the NJDEP, USEPA, Region II Technical Support Branch and
NIST to develop sludge reference materials from domestic and industrial sources having reference values for
their acid-extractable metals content. The study was successful for the domestic material, resulting in the
derivation of reference values for EPA acid-extractable methods as part of SRM 2781, Domestic Sludge. The
addendum to SRM 2781 contains leachable mass fractions for 14 metals and compares the leach recoveries to
total metal concentrations obtained separately. This paper discusses the analysis of the industrial sludge
material, NIST SRM 2782.
EXPERIMENTAL
Reference Samples
The candidate industrial sludge reference material SRM 2782, was prepared from more than 100 kg of
electroplating waste supplied by AT&T Bell Laboratories, Murray Hill, NJ. This material was shipped to NIST,
wherein it was freeze-dried, processed, radiation sterilized and homogenized by contractors according to
procedures used to prepare United States Geological Survey and NIST geological reference materials. Samples
were aliquotted and shipped to the participating laboratories, the NJDEP Bureau of Radiation and Inorganic
Analytical Services (BRIAS) and USEPA, Region II, Technical Support Branch, for chemical analysis.
Sample Preparation
The NJDEP used open vessel hot plate digestion (NJDEP Method 100) for all of their acid digestions. The
USEPA used open vessel hot plate digestion (USEPA Method 3050) for the industrial sludge and Method 3050
and closed vessel microwave digestion techniques (USEPA Method 3051) to prepare the domestic sludge for
measurement. EPA prepared two sets of samples using microwave digestion, one without and one with HCI.
Instrumentation
The NJDEP used a Perkin-Elmer (PE) Model 5000 atomic absorption spectrometer (AAS) for its flame atomic
absorption (FAAS) metal measurements. A similar unit, equipped with a PE Model 500 furnace and a PE AS-50
188
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
autosampler, was used for the graphite furnace atomic absorption (GFAAS) measurements. The NJDEP also
employed a Thermo-Jarrell Ash Model 25 sequential inductively coupled plasma emission spectrometer
(ICPOES) for some of its metal determinations. The USEPA used a CEM Model MDS-2000 microwave digestion
system for some sample preparations and a Thermo Jarrell-Ash Model 61 Simultaneous ICPOES and a PE 5100
GFAAS, equipped with a model AS-600 furnace and AS-60 autosampler for its metal determinations. The
USEPA performed the metals measurements by either simultaneous ICPOES and GFAAS, using one technique
per metal. The NJDEP employed more than one technique to measure most of the metals.
Sample Processing
Each laboratory analyzed its own digests by one or more instrumental methods. Once completed, the two
laboratories exchanged extracts and conducted another series of measurements using the same techniques.
RESULTS AND DISCUSSION
Developing "reference" values for the composition of real environmental samples requires inclusion of random
errors of measurement and any systematic bias which might be inherent in the method(s) employed to estimate
the true value. Issues such as the method of measurement, the complexity and composition of the sample matrix
and the concentration and heterogeneity of the analytes all contribute to the overall uncertainty. While relative
uncertainties of -20% to + 20% are expected when EPA methods such as FAAS, GFAAS and ICPOES are
applied to nonaqueous media, reference values based upon these methods are still useful as confirmation of
results of environmental monitoring such as the analysis of sludge effluents from treatment plants.
Current Results for Industrial Sludge
The initial set of analyses performed on NIST SRM 2782 yielded poor agreement between the two laboratories.
For the most recent series of tests, the precision of the 6 aliquots of NIST SRM 2782 ranged from 0.2 % to 2.0 %
for all elements measured by both participating laboratories. There is no discernable change in the level of
precision of the replicate measurements obtained in the earlier and later studies Table 1 compares the results
obtained by both laboratories in the initial ('94) and most recent ('96) studies. There is no discernable pattern in
the NJDEP data from 1994 to 1996; the more recent series of USEPA results are generally higher than those
obtained previously. Differences between the grand means of all results obtained for each element by the two
laboratories for the more recent set of measurements were within 6% for 6 elements and 10% for 12 of the 17
metals studied; 2 others were within 12 %. Cd, Cr and K were the only elements outside this range,
Analysis of Variance (ANOVA) can be utilized to evaluate whether the individual element means obtained by the
two laboratories for each element, generated by different methods of sample preparation and analysis, are
statistically significant. As shown in Table 2, for the 14 metals where agreement between the grand means was
within 12 %. ANOVA analyses, as shown by the F value (ratio of variance between data sets to the variance
within a data set) and P value (measure of the probability the actual sample set fell within hypothetical
frequencies for infinitely large data populations) showed that the means obtained by the two laboratories agreed
within specified tolerances for 11 of the 14 metals; the exceptions being Ba, Cr and Mn. This serves as further
corroboration to the extent of the agreement between the results.
The USEPA obtained good results for 14 metals where extractions were obtained by both Method 3050 (hot plate
digestion) and Method 3051 (microwave digestion), with analyses performed by ICP-OES. Agreement is within ฑ
6 % for all metals studied but Na and K; with a range in bias from - 4.8 % for Al to +5.2 % for Zn. These results
are shown in Table 3. The values for the alkali metals Na and K are approximately 30-40 % higher for the
microwave technique. When the t-test for means assuming unequal variance is applied, for 6 metals (Ba, Ca,
Cu, Mn, Ph and Zn), the differences between the calculated means obtained by each method were statistically
significant. However, these detected differences are relatively small, especially when compared to the control
limits assigned to acid-extractable analytes measured in non-aqueous media. All standard deviations are ฑ 2
sigma.
Comparison of Leachable Concentrations for Domestic Sludge and Industrial Sludge
The addendum to the Certificate of Analysis for NIST SRM 2781, domestic sludge compares the % leach
recoveries of selected metals with the values obtained for their total mass fractions. For elements where
leachable metal results were obtained from both the USEPA and NJDEP laboratories, a ratio of leachable metal
concentration to total metal concentration (obtained by NIST) was calculated. For the domestic sludge these
ratios were between 0.82 and 0.96 for 9 out of 11 elements; only Al (0.50) and Cr (0.71) were below this range.
Preliminary values supplied by NIST for the industrial sludge material shows a different pattern; for the 14
189
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
elements for which such data is available, only 6 have ratios of leachable to total that are greater than 0.80. Na,
K, Mg, Al have ratios below 0.20, the ratio for V is 0.21, Ba and Ni are 0.60 and 0.62 respectively, while Ca is
0.72. Only certain transition elements (Cu, Ph, Zn, Mn, Fe and Co) have leachable/total ratios for the industrial
sludge that are greater than 0.80. The results are summarized in Table 4. Note that for the industrial sludge all of
the ratios for the alkali and alkaline earth metals as well as Al are 0.72 or below. While the reasons for the
specific analyte-matrix interactions are uncertain, it is apparent that the more complex nature of the industrial
sludge plays a key role in how much metal is leached by the regulatorily-approved digestion methods, and that
the amount that leaches varies significantly by element and group. This issue must be considered when
evaluations of leachable metal concentrations are made for environmental assessment and policy
considerations.
Table 1. Industrial Sludge - Grand Means
Element
Al
Ba
Ca
Cd
Co
Cr
Cu
Fe
K
Mg
Mn
Na
Ni
Pb
V
Zn
Table 2. Industrial Sludge - ANOVA
Analyses
Mean DEP'94
1502 ฑ43
150 ฑ1
4687 ฑ 470
4.0 ฑ0.1
70.0 ฑ 6
79.8 ฑ 1
2485 ฑ 146
255020ฑ 17000
116ฑ1
508ฑ37
274 ฑ13
2430ฑ18
125ฑ2
581 ฑ18
23.2ฑ 1
1266 ฑ31
Mean EPA'96
1587 ฑ107
161 ฑ11
4950 ฑ 144
11. 4 ฑ3.0
51 .6 ฑ7
76.7 ฑ 4
2459 ฑ 37
256600ฑ 2300
121 ฑ2
498 ฑ69
274 ฑ7
2431 ฑ20
95.5 ฑ6
558ฑ9
17.5 ฑ2
1181 ฑ59
Mean DEP'94
1380 ฑ50
132 ฑ4
4320 ฑ141
15.4 ฑ0.8
54.4 ฑ 2
55 3 ฑ2
2270 ฑ 53
232000ฑ5600
58.9 ฑ6
441 ฑ17
224 ฑ7
2000 ฑ123
908ฑ2
51 9 ฑ30
20.6 ฑ1
1170ฑ21
Mean DEP'96
1528 ฑ46
144 ฑ4
4685 ฑ151
2.3 ฑ0.1
56.0 ฑ 3
58.3 ฑ1
2420 ฑ 42
253000ฑ14000
78.6 ฑ21
470ฑ15
247ฑ7
2573 ฑ307
96.1 ฑ4
553 ฑ22
15.9ฑ1
11 58 ฑ71
Al
Ba
Ca
Co
Cr
Cu
Fe
Mg
Mn
Na
Ni
Pb
V
Zn
F
0.82
33.9
1.73
0.25
3.67
0.51
4.09
2.25
69.6
0.06
4,20
3.52
2.32
0.07
P
0.38
0.00
0.22
0.62
0.00
0.49
0.06
0.14
0.00
0.81
0.09
0.09
0.14
0.79
DEP/EPA '96
1.038
1.118
1.056
0.921
1014
1.059
1.109
0.945
0.994
1.010
1.100
1.019
CONCLUSIONS AND RECOMMENDATIONS
Reference values have been derived for the leachable
concentration of 14 metals in an industrial sludge standard
reference material (NIST SRM 2782). Both open vessel hot-plate
acid digestion (NJDEP Method 100 or USEPA Method 3050) or
microwave digestion (USEPA Method 3051) are appropriate
methods for sample preparation of these materials and the
leachable concentrations may be measured by either FAAS,
sequential or simultaneous ICPOES. This information supports
the application of performance-based methodology since there
are several combinations of sample preparation and measure-
ment systems that can achieve similar results. Following
complete statistical review by NIST, sludge Standard Reference
Materials from both domestic and industrial sources-will be
commercially available. Reference values for their acid
extractable metals content and associated uncertainties will be
provided to support the quality assurance of sludge metal
measurements. Analyses of these SRMs should be required as
part of a POTW's compliance data submitted to regulatory
agencies Such as the USEPA and NJDEP- It is also
recommended that these materials become part of any future
laboratory certification program for sludge effluents.
190
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 3. Industrial Sludge - Method Comparison
Element Method 3050 Method 100 3050/100
Mean SD Mean SD
Al 1520 18 1540 8 0.987
Ba 148 2 144 1 1.027
Ca 4870 58 4720 34 1.032
Co 58.5 0.9 57.0 0.9 1.026
Cr 57.9 0.6 58.2 2.4 0.994
Cu 2500 37 2420 45 1.033
Fe 260000 4500 236000 1100 1.102
K 72.4 11.8 80.1 1250 0.904
Mg 470 5 490 7 0.959
Mn 256 3 246 3 1.041
Na 2300 39 2470 60 0.931
Ni 99.6 3.7 95.1 1.1 1.047
Pb 545 6 524 13 1.040
Zn 1210 15 1210 11 1.000
Table 4. Domestic Sludge -Recoveries and Comparisons w/Certificate
Element Domestic Sludge Industrial Sludge
Ratio: Leachable / Total Ratio: Leachable / Total
Ag 0.88
Al 0.50 0.11
Ba - 0.60
Cd 0.86
Ca 0.93 0.72
Co - 0.82
Cr 0.71
CU 0.96 0.94
Fe 0.87 0 95
K - 0.02
Pb 0.91 0.97
Mg 0.82 0.18
Mn 0.82 0.86
Na - 0.19
Ni 0.90 0.62
V 0.21
Zn 0.88 0.93
- Only NJDEP or USEPA leachable results were available, mean values
between the two laboratories was not computed.
PERSPECTIVES ON DIOXIN ANALYSIS
Stevie Wilding
USEPA, Region III, Office of Analytical Services and Quality Assurance,
839 Bestgate Road, Annapolis, MD 21401
Phone: 410-573-2733, FAX: 410-573-2771
Dioxin analysis is very complex. Many people do not understand how dioxin analysis is performed within a
laboratory or how to validate the data when they receive it. Through my technical assistance to internal and
external customers concerning dioxin analysis, it has become clear they do not understand and are afraid to
191
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
understand the methodology of dioxin analysis. Prior to validating PCDD/PCDF data, it is imperative to have
complete knowledge of the method used. The purpose of this paper is to try and simplify dioxin analysis. To
break the analysis into understandable pieces and to compare it to other organic analyses. The basis for the
Dioxin/Furan analysis is Selected Ion Monitoring (SIM) Gas Chromatography/Mass Spectrometry (GC/MS). The
objective of this paper is to explain SIM GC/MS and isotope dilution for the average nonchemist. The
advantages and disadvantages of using isotope dilution and SIM over the "routine" organic methods will be
discussed.
INTERPRETATION OF GROUND WATER CHEMICAL QUALITY DATA
G.M. Zemanskv
Compass Environmental, Inc., 3000 W. 19th Court, Lawrence, Kansas 66047-2300
ABSTRACT
Ground water is sampled to assess its quality for a variety of purposes. Whatever the purpose, it can only be
achieved if results are representative of actual site conditions and are interpreted in the context of those
conditions.
Substantial costs are incurred to obtain and analyze samples. Field costs for drilling, installing, and sampling
monitoring wells and laboratory costs for analyzing samples are not trivial. The utility of such expenditures can
be jeopardized by the manner in which reported results are interpreted as well as by problems in how samples
were obtained and analyzed. Considerable attention has been given to standardizing procedures for sampling
and analyzing ground water. Although following such standard procedures is important and provides a necessary
foundation for understanding results, it neither guarantees that reported results will be representative nor
necessarily have any real relationship to actual site conditions. Comprehensive data analysis and evaluation by a
knowledgeable professional should be the final quality assurance step, it may indeed help to find errors in field or
laboratory work that went otherwise unnoticed, and provides the best chance for real understanding of the
meaning of reported results.
The focus of this paper is on the interpretative part of the process. Although formal interpretation necessarily
comes late in a project, when data have been generated and the report is being written, it will be most useful if
relevant elements can be integrated into the project from the beginning. When this is done, it increases the
likelihood of achieving project objectives as well as understanding the data. To facilitate interpretation, the
following steps should be included:
1. Collection, analysis, and evaluation of background data on regional and site-specific geology, hydrology, and
potential anthropogenic factors that could influence ground water quality and collection of background
information on the environmental chemistry of the analytes of concern.
2. Planning and carrying out of field activities using accepted standard procedures capable of producing data of
known quality.
3. Selection of a laboratory to analyze ground water samples based on careful evaluation of laboratory
qualifications.
4. The use of appropriate quality control/quality assurance (QC/QA) checks of field and laboratory work
(including field blank, duplicate, and performance evaluation samples).
5. Comprehensive interpretation of reported analytical data by a knowledgeable professional. The analytical data
must be accompanied by appropriate QC/QA data, be cross-checked using standard water quality checks and
relationships where possible, and be correlated with information on regional and site-specific geology and
hydrology, environmental chemistry, and potential anthropogenic influences.
Application of this sequence of steps and their importance in interpretation of ground water quality data are
discussed in this paper. The discussion includes several illustrative case examples.
INTRODUCTION
There area number of reasons we might want to know about groundwater chemical quality. For example, we
192
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
might be interested in using an aquifer as a source of water today or we may want to ensure that current
practices haven't caused contamination so that ambient ground water quality remains legally acceptable whether
or not it is used in the future. At first glance, you'd think that obtaining and interpreting ground water quality data
would be a fairly straight forward exercise. Simply follow the usual scientific approach: (1) take a sample; (2)
analyze it; and (3) compare analytical results to a set of criteria. However, obtaining reliable data and properly
interpreting them turns out to be more complicated than that; even in the relatively simple case of surface water.
When it comes to ground water, everything seems more complex.
To begin with, obviously, with ground water you can't just go down to the stream and get a sample. Because
ground water is underground, you need to do more preparation in advance of sampling. This involves obtaining
background subsurface information and an access point. Such information will be helpful, both in planning how to
obtain a sample as well as in interpreting analytical results when they become available. Whether you use a
standard monitoring well, direct push technology, or something else, getting access to sample isn't always easy
and can influence the quality of the sample obtained. There are also a number of potentially confounding factors
with regard to the next step of the process, laboratory analysis. Sample quality can change between the time a
sample is obtained and the time it is analyzed and, even if it doesn't, the overall reliability of laboratory results is
not the sure thing many people assume it is. Finally, selection of appropriate criteria to compare data to may not
always be straight-forward. Relevant criteria may either not exist or be incomplete.
Comprehensive data interpretation by a knowledgeable professional should be the final quality assurance step of
any project involving ground water quality data. It may indeed help to find errors in field or laboratory work that
went otherwise unnoticed and provides the best chance for real understanding of the meaning of reported results.
Proper project planning should prepare for this final step by obtaining relevant information early on and including
relevant data collection into field segments of the project. The following steps must be integrated into and carried
out throughout the project to facilitate final interpretation:
1. Collection, analysis, and evaluation of background data on regional and site-specific geology, hydrology, and
potential anthropogenic factors that could influence ground water quality and collection of background
information on the environmental chemistry of the analytes of concern.
2. Planning and carrying out of field activities using accepted standard procedures capable of producing data of
known quality.
3. Selection of a laboratory to analyze ground water samples based on careful evaluation of laboratory
qualifications.
4. The use of appropriate QC/QA checks (including field blank, duplicate, and performance evaluation samples).
5. Comprehensive interpretation of reported analytical data by a knowledgeable professional.
The analytical data must be accompanied by appropriate QC/QA data, be cross-checked using standard water
quality checks and relationships where possible, and be correlated with information on regional and site-specific
geology and hydrology, environmental chemistry, and potential anthropogenic influences.
BACKGROUND INFORMATION
Background information serves several functions: (1) it facilitates obtaining access to sample groundwater; (2) it
provides guidance regarding selection of appropriate sampling methods and analytical variables; (3) it places
ground water quality data into context; and (4) it provides a quality assurance check. The following background
information is necessary to fulfill these functions: (1) regional and site-specific geology; (2) regional and
site-specific hydrology; (3) information on the environmental chemistry of variables of concern to the project; and
(4) broad information on potential anthropogenic influences including site conditions and possible contaminant
sources. The latter includes not only those conditions which may have impacted ground water, but those which
could have influenced sample quality as a result of installation and testing of monitoring wells or otherwise during
the sampling process. For example, shallow contamination can be carried down into a deeper aquifer during field
work and airborne chemicals may cause trace contamination of ground water samples if they enter an open
borehole, monitoring well, or sample being placed into a container.
FIELD PROCEDURES
-Access Points
The nature of subsurface conditions will influence the type of access point that is possible. Sample quality will
also be impacted by the type of access point selected. Commonly used ground water access points include: (1)
monitoring wells; (2) wells or piezometers installed for other purposes; and (3) direct push technology.
193
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The nature of the access point has, in particular, a relationship to the level of total suspended solids (TSS) likely
to be present in samples. TSS levels would be expected to be relatively high in samples obtained using direct
push technology and low in samples obtained from a water supply well. Assuming proper design and installation
(including development), TSS levels in samples obtained from monitoring wells will normally be low except
where fine-grained materials are screened. However, development is something that is often neglected or treated
in a pro-forma manner when monitoring wells are installed. Additional development, as opposed to routine
purging, may also be required when there are long periods between sampling events. This is illustrated in Table
1, showing results for inorganic variables in unfiltered samples from a well that was properly developed prior to
initial sampling but was not redeveloped when sampled again after two years.
Sampling Methods
Available sampling methods are often constrained by the type of access point utilized. Since they will have a
direct bearing on sample quality, the sampling methods used must be taken into account in planning sampling
events and in interpretation of the data obtained. This category includes consideration of both field equipment
and procedures. For example, regulatory agencies are more frequently requiring analysis of unfiltered samples.
This may introduce substantial variation into the process, particularly for inorganic variables.
Sample Handling and Preservation
The order of sampling, type of container, and sample preservation method utilized can affect the quality of
samples analyzed in the laboratory. As discussed further below, U.S. Environmental Protection Agency (USEPA)
specified sample preservation methods and maximum allowed holding times do not always ensure sample
quality will not change between the field and laboratory analysis.
Field Analysis/Observations
Reliable sample preservation methods do not exist for some water quality variables. In other cases, field
measurements carried out for other purposes (i.e., well purging) are routinely available and preferable (e.g., to
laboratory analysis for the same variable) or field observations can be made that provide useful information
otherwise lost if not recorded at the time. USEPA requires field analysis (by specifying immediate analysis) for
only five variables: (1) chlorine residual; (2) pH; (3) dissolved oxygen (by probe); (4) sulfite; and (5) temperature.
The validity of some USEPA maximum holding times and preservation combinations is questionable. For
example, although USEPA recognizes that pH and dissolved oxygen levels may change substantially if not
analyzed in the field and therefore requires immediate analysis without holding for them, substantial changes in
alkalinity may occur for the same reasons but 14 days holding time is allowed for this variable. Because it is
allowed and more convenient, samples are frequently analyzed in the laboratory for alkalinity instead of the field.
Similarly, USEPA allows maximum holding times of 48 hours and seven days for color and total suspended
solids (TSS), respectively. USEPA also allows seven and 14 days holding time for unacidified and acidified
volatile aromatic compounds, respectively. Both color and TSS may change substantially in samples over these
allowed time frames and research has shown both substantial loss of volatile aromatic compounds in less than
seven days in unacidified samples and that substantially greater holding times than 14 days are appropriate for
many compounds when samples are acidified.1
When sampling ground water, field analysis should routinely occur for the purge variables conductivity, pH, and
temperature, using appropriate equipment operated, calibrated, and maintained in accordance with manufacturer
recommendations. The following field observations related to sample quality may also be made: (1) color; (2)
odor; and (3) turbidity. Although this information is essentially "free" for the taking, it is often not recorded. In
particular, the latter observation can provide at least a qualitative indication of the presence of TSS and their
possible effect on inorganic variables. The difference that filtering makes in reported concentrations of many
inorganic variables when unfiltered samples contain high solids concentrations is substantial and readily
demonstrated with split samples.
LABORATORY ANALYSIS
Too often, the analytical laboratory for a project is selected on the basis of cost only. Such cost savings may
prove to have been very expensively purchased if quality isn't also delivered. Laboratory qualifications should be
carefully researched in advance of selection. This can be done by such things as reviewing historic results on
performance evaluation samples, auditing the facility, submitting performance evaluation samples, and checking
references. Checks should continue throughout the project with the appropriate use of blank, duplicate, and
performance evaluation samples.
194
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Unfortunately, no laboratory is perfect. Even the better laboratories often get some wrong answers when
analyzing USEPA performance evaluation samples. These are in a relatively clean matrix, without the kinds of
interferences that can complicate real world samples, and are being analyzed under test conditions. A laboratory
would be foolish not to make its best effort on these samples. You can expect somewhat lower quality with
regard to samples being routinely analyzed for the average client. Routine QC/QA and data validation practices
are not the complete answer to data quality. In particular, the latter usually don't include various data
cross-checks or take into account what is known about site conditions. In the worst case, laboratory reported
analytical results can be more artifacts of the sampling and analysis process than representative of ambient
ground water quality.
There are a number of factors which can influence laboratory results. But even when analytical laboratories are
performing well, it should be recognized that: (1) at best, they can only analyze samples in the condition
received; and (2) standard limits of precision and accuracy allow considerable variation. As noted above, a
number of factors may result in samples reaching the laboratory which are less than representative of site
conditions. Even when laboratory QC/QA requirements are met, the level of allowed variation limits the
conclusions which can be based on laboratory data. USEPA Superfund acceptance criteria for percent recovery
of matrix spikes (MS) and laboratory control standards (LCS) are shown in Table 2.23
Data reported by laboratories should be carefully reviewed or "validated" before being utilized. USEPA national
functional guidelines for data review under Superfund are an example of typical validation guidelines. These
specify a number of checks intended to assure that correct procedures were followed (e.g., holding times,
calibration, blanks, matrix spikes and spike duplicates, and laboratory control standards). Although this type of
review is useful, it is important to realize its limitations. The standard conclusion one consulting company uses
after a successful validation process contains the statement that "the data collected during this investigation are
valid as qualified for use in representing Site conditions and for use in risk assessments."4 Since the validation
process referred to doesn't take relevant site information or available data cross-checks into account, such a
statement goes too far. Neither is it necessarily correct.
CASE EXAMPLES OF INTERPRETING GROUND WATER DATA
Organic Contamination at a Superfund Site
Trichloroethylene (TCE) groundwater contamination was recognized at a Superfund site. The presence of other
halocarbons associated with TCE degradation, notably cis-1,2-dichloroethylene (c-1,2-DCE) and vinyl chloride,
was also recognized. However, the potentially responsible party (PRP) claimed the TCE originated elsewhere
and was part of a wider regional problem involving a number of sources and contaminants. Site work performed
by the PRP in 1996 identified a variety of other compounds in samples from relatively deep wells drilled using air
rotary equipment. The compounds reported in various samples fell within the following categories:
1. Volatile organic compounds (VOCs) -
a. Halocarbons including TCE and TCE degradation products.
b. Petroleum hydrocarbons, including aromatic compounds (e.g., xylenes).
c. Trihalomethanes, particularly chloroform and bromodichloromethane.
2. Semivolatile organic compounds (SVOCs) -
a. Petroleum hydrocarbons (benzoic acid and naphthalene).
b. Phenol and other phenolic compounds.
c. Phthalates, particularly bis(2-ethylhexyl)phthalate,
Review of these data indicated that other interpretations were much more likely. Among the reasons the data
were suspect were the following:
1. With the exception of halocarbons, most of the compounds involved had not been reported in samples from
shallow wells previously drilled using hollow-stem augers. Contamination in such cases generally moves from
surface or near-surface sources downward. Furthermore, the concentrations involved were generally low and
appeared to be randomly distribution rather than in a pattern suggesting any relationship to possible sources.
2. Air rotary drilling is a possible source of petroleum hydrocarbon and phenolic compound contamination.
3. Some of the boreholes had been left open to the atmosphere for substantial periods of time (i.e., on the order
195
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
of months) after drilling before monitoring wells were installed in them. The ones open the longest were also
adjacent to an Interstate Highway. Petroleum hydrocarbons have been identified in vehicle emissions and
ambient urban air.5 Research has also indicated the potential for petroleum hydrocarbons and phenolic
compounds in urban air to be transported into ground water.67
4. PRP consultants had not decontaminated at least some downhole equipment (particularly water level
indicators) when used between boreholes or between monitoring wells. This may cause cross-contamination.
5. THM contamination is common in chlorinated tap water from surface sources, but unusual in ambient ground
water. Thousands of gallons of chlorinated tap water from a surface source known to contain substantial
concentrations of trihalomethane compounds had been introduced into most of the boreholes as a part of the
testing program during field work.
6. Phthalates are recognized by USEPA as a common SVOC laboratory contaminant.8
7. Background regional information strongly indicated the PRP was the source of halocarbon contamination and
that there was no other more widespread regional problem involving halocarbons or other compounds.
Resampling in 1997 provided further confirmation of this interpretation. Data for the 19 wells sampled in both
years are shown in Table 3 (parts a and b). With the exception of a reduced number of low concentration hits for
phthalate compounds (which are recognized as common laboratory contaminants) and, in one case, a
trihalomethane compound, only halocarbons were reported in 1997 This indicates that the various other
compounds reported in 1996 samples were a transitory impact of drilling, testing, sampling, and analysis rather
than regional ground water contamination. This transitory impact dissipated over time as a result of natural
mechanisms such as flushing by ambient ground water flow and biodegradation. The change in apparent
distribution of halocarbon contamination between 1996 and 1997 apparent from these results may also indicate
that shallower contamination was carried downward by intrusive work performed in 1996 and that this also
produced a transitory impact on ground water samples.
Ground/Surface Water Relationship at a RCRA Site
Ground water monitoring at landfills and other Resource Conservation and Recovery Act (RCRA) facilities is
oriented towards ensuring that a release of contaminants, if it occurs, will be detected. In general, detection
monitoring requires that data from wells downgradient of the facility and subject to impact in event of a release
be statistically compared to data from upgradient background wells. If the comparison results in a statistically
significant difference with a downgradient increase, a release is assumed. However, other regional conditions
must be considered if such comparisons are to be useful.
In this case, there is a network of monitoring wells installed up and downgradient of a land treatment unit (LTU)
located at a refinery adjacent to a major river. Statistical analysis showed that concentrations of some variables
were elevated in downgradient wells and that the elevations were statistically significant. Did this mean that a
release had occurred?
A linkage between the river and the adjacent alluvial ground water aquifer would be expected based on general
principles. This was confirmed by analysis of two lines of evidence: (1) correlation of ground water elevations
with river flow (a surrogate for stage); and (2) statistical analysis of water quality data. Time series plots of
ground water elevations versus river flow showed an evident visual correlation, which was confirmed by linear
regression analysis. The correlation was best for those wells closest to the river (correlation coefficient of 0.80)
and decreased with distance from the river. A comparison of water quality data is supportive. Data indicating
central tendency for upgradient monitoring wells and the river are presented for six variables in Table 4. For the
three major ions in Table 4, including chloride, river concentrations far exceed those in upgradient ground water.
In these cases, statistical analysis shows that concentrations in samples from downgradient monitoring wells are
significantly higher than upgradient ground water. The reverse is true for the three elements listed. Their
concentrations are higher in upgradient ground water than the river and statistical analysis shows that their levels
in downgradient monitoring wells are significantly lower than upgradient.
Inorganic Water Quality at a Superfund Site
Consultants for a PRP at a Superfund site involving a limestone aquifer in the midwest US developed several
theories regarding the nature of the aquifer involved based on their interpretation of data for inorganic
196
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
constituents. Data for alkalinity, aluminum, calcium, iron, silica, sodium, and sulfate were prominent in their
interpretation. However, these consultants did not consider relevant site-specific information and admitted they
were unaware of the degree of variation that can occur in laboratory analysis of ground water samples that meet
acceptance criteria (see Table 2).
A potentially important piece of site-specific information not considered by the PRP's consultants was that most
of the borings involved had been drilled considerably deeper than the wells later installed in them. For the wells
being installed in such deeper borings, boreholes were first partially filled by pouring in sodium bentonite chips.
This occurred for about two-thirds of the relatively deep wells drilled. On the average, approximately one-third of
the borehole was filled (i.e., 71 of 221 feet). This process undoubtedly resulted in the introduction of chemicals
from the hydrating chips into the water (both as dissolved and suspended solids) as they fell through the water
column. Of the analytes relevant to this site, sodium bentonite chips are typically composed of silica and oxides
of aluminum, iron, calcium, sodium, magnesium, and potassium (in order of concentration). They also contain a
small level of water soluble nitrate. Given the chemistry of silica and calcium and the likelihood that calcium in a
limestone aquifer would be expected to already be near saturation, concentrations of these variables would
probably not be greatly affected by this. However, concentrations of aluminum, iron, sodium, magnesium, and
potassium could be and this appears to have been the case. The potential for this was increased by the fact that,
although consultants for the PRP purged three well volumes immediately prior to sampling, they did not develop
the monitoring wells after installation.
To evaluate the effect of filling boreholes with bentonite on inorganic ground water quality, monitoring wells
sampled both during 1996 (shortly after installation) and 1997 (nearly a year since last sampled) were divided
into two groups: (1) bentonite filled (BF); and (2) unfilled (UF). Median data for major cations, major anions, and
several other variables grouped into these two categories for both 1996 and 1997 sampling events are presented
in Table 5. The 1996 data clearly indicate impact from bentonite filling for most of the variables listed except
calcium and silica, BF:UF ratios for sodium, aluminum, and iron indicate nearly an order of magnitude or greater
level of enrichment for those variables as a result of bentonite filling. This is also evident in the STIFF diagrams
of median grouped data in Figure 1. The STIFF diagram for unfilled wells (Figure 1a) is typical of what would be
expected for a limestone aquifer.9 There are several relatively minor differences between the STIFF diagram for
bentonite filled wells (Figure 1b) and unfilled wells (Figure la), but by far the most notable difference is the
sodium "bulge" to the lower left of the diagram. The impact of the bentonite appears to have been transitory.
With the possible exception of nitrate, the enrichment appears to have been flushed away due to ambient ground
water flow by the time wells were resampled nearly a year later. STIFF diagrams for both sets of wells (UF and
BF) when resampled in 1997 were similar to each other and the one for unfilled wells in 1996 (Figure 1a).
The PRP's consultants pointed to two aspects of the data to support their interpretation that the aquifer involved
was an "open" system (i.e., rapidly recharged from the surface throughout the aquifer): (1) abnormally high
concentrations of aluminum and iron in samples from some relatively deep wells (exceeding solubility limits); and
(2) lack of any apparent ground water evolution (change in quality along a flow line) between wells at higher
elevations and those at lower elevations (a distance of roughly one mile). As discussed above and shown in
Table 5, aluminum and iron enrichment appears to have been related to filling boreboles with sodium bentonite
chips. Not knowing about this circumstance and since the overlying clay soils involved would be normally
expected to be rich in silica as well as aluminum and iron, the consultants interpreting inorganic water quality
data for the PRP felt compelled to provide another explanation why silica concentrations were not also enriched
when aluminum and iron were. Their explanation was an assertion, without any data, that the soils were lateritic.
Lateritic soils develop in hot, wet tropical climates subject to heavy rainfall when the intense chemical weathering
that occurs under those conditions removes both soluble materials and much of the silica. Lateritic soils are not
characteristic of the temperate climate midwest US.10 A much more likely explanation for the relative magnitudes
of aluminum, iron, and silica has to do with the manner in which samples were taken, preserved, and prepared
for analysis. Samples to be analyzed for aluminum and iron are acidified in the field and digested in the
laboratory. Since the samples involved were unfiltered, this meant that some particulate aluminum and iron
would be included in the measurement. In contrast, samples to be analyzed for silica are not acidified in the field,
but are filtered prior to analysis.
With respect to ground water "evolution," this phenomena has generally been documented on a regional rather
than a site scale. This may be in part because very substantial "evolution" must occur to be detectable against a
background of variation sampling and analysis introduced variation as well as environmental variation. For an
extreme example, using the acceptance criteria of 75 to 125 percent recovery for matrix spike samples (see
197
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 2), a sample having a true value of 100 mg/L of calcium could be reported to have either 75 or 125 mg/L.
Although the higher of these two numbers is 67 percent greater than the low number, either one would be within
acceptance criteria. With enough sampling events and statistical analysis of data, it might be possible to detect
an "evolutionary" change within acceptance criteria boundaries (e.g., a change from 50 to 65 mg/L along the flow
path); however, the possibility of seeing this change with a single set of samples appears slim when so much
variation is acceptable.
SUMMARY AND CONCLUSIONS
Comprehensive data interpretation by a knowledgeable professional should be the final quality assurance step of
any project involving ground water quality data. It may indeed help to find errors in field or laboratory work that
went otherwise unnoticed and provides the best chance for real understanding of the meaning of reported results.
Proper project planning should prepare for this final step by obtaining relevant information early on and including
relevant data collection into field segments of the project. The following steps must be integrated into and carried
out throughout the project to facilitate final interpretation:
1 Collection, analysis, and evaluation of background data on regional and site-specific geology, hydrology, and
potential anthropogenic factors that could influence ground water quality and collection of background
information on the environmental chemistry of the analytes of concern.
2. Planning and carrying out of field activities using accepted standard procedures capable of producing data of
known quality.
3. Selection of a laboratory to analyze ground water samples based on careful evaluation of laboratory
qualifications.
4. The use of appropriate QC/QA checks (including field blank, duplicate, and performance evaluation samples).
5. Comprehensive interpretation of reported analytical data by a knowledgeable professional.
The analytical data must be accompanied by appropriate QC/QA data, be cross-checked using standard water
quality checks and relationships where possible, and be correlated with information on regional and site-specific
geology and hydrology, environmental chemistry, and potential anthropogenic influences.
References Cited
1. USEPA holding times cited in this paragraph are from Table II of Part 136, Title 40, U.S. Code of Federal
Regulations (40 CFR 136).
2. Office of Emergency and Remedial Response. 1994. USEPA contract laboratory program national functional
guidelines for inorganic data review. Publication No. EPA 540/R-94/013, USEPA, Washington, DC, pp. 21-24
and 27-29.
3. USEPA Superfund contract laboratory program (CLP) statement of work (SOW) Form III VOA-1 and Form III
SV-1.
4. Burns & McDonnell Waste Consultants, Inc. 1996. Quality control evaluation report 96-072-4 dated
September. Kansas City, MO, p. 10-1.
5. Schleyer, R. 1991. Influence of the deposition of anthropogenic organic substances from the atmosphere on
ground-water quality. IN: Ground Water: Proceedings of the International Symposium, G.P, Lennon, ed.,
American Society of Civil Engineers, pp. 293-298.
6. Harley, R.A., et al. 1992. Respeciation of organic gas emissions and the detection of excess unburned
gasoline in the atmosphere. ES&T, Vol. 26, No. 12, pp. 2395-2408.
7. Pankow, J.F., et al. 1997. The urban atmosphere as a non-point source for the transport of MTBE and other
volatile organic compounds (VOCs) to shallow groundwater. ES&T, Vol. 31, No. 10, pp. 2821-2828.
8. Office of Emergency and Remedial Response. 1994. USEPA contract laboratory program national functional
guidelines for organic data review. Publication No. EPA 540/R-94/012, USEPA, Washington, DC, pp. 62-65.
9. Hounslow, A.W. 1995. Water Quality Data: Analysis and Interpretation, CRC Lewis Publishers, Boca Raton,
FL, pp. 86-87.
10. Tarbuck, E.J. and F.K. Lutgens. 1992. The Earth: An Introduction to Physical Geology, 4th Ed., Macmillan
Publishing Co., New York, pp. 130-131.
198
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1. Effect of Monitoring Well Development1
$jJiJJ!$il?:'-'': .;:.:. ,. , ,-.'.L>:^. '^d^;^
Calcium
Chloride
Iron
Magnesium
Manganese
Potassium
Sodium
Appearance
>i:-'ii!5v^ '^^p^HpUP^'''^" "'A ;;::^''v'i
92.
43.
0.08
2.4
0.069
ND
5.8
Clear
^m^w^mmm^1 : -j^;
104.
58.3
5.53
2.84
0.13
533
6.38
Turbid
1. Concentrations in mg/L. ND means non-detect at 1 mg/L.
2. New well developed and purged prior to sampling.
3. Well unused for two years. Three well volumes purged prior to sampling.
Table 2. USEPA Superfund Acceptance Criteria
'^IWl^^^lilt-'fBett^^1.'" '" '-.; ;7:'ixS
Elements (60 10)
VOCs (8240):
Benzene
Trichloroethylene
SVOCs (8270):
Pyrene
Phenol
:.S';x"-v- i^J^^ptw^ f?" ?;, vi-:
75-125
76-127
71-120
26-127
12-110
:. " wi*! '''r&WM*' "t"j; ^'~Kr}*i&ys&-~FfKf,; ;.." .;: J-. ',:,.ซ' rf'i
>: Vs'l.atriK, ?'-' .r,ta>*&i:S.i -'.. v' '-TV ..: .' :;..=
80-120
-
-
-
-
1. Identification of variable and USEPA analytical method with example volatile and semivolatile organic
compounds (VOCs and SVOCs, respectively).
2. Matrix spike (MS) and laboratory control sample (LCS) acceptance criteria in percent recovery. Inorganic
criteria from USEPA national functional guidelines. Organic criteria from CLP SOW Forms III VOC-1 and
SV-1.
Table 3a. Reported Organic Compounds
- " :*
ฐ0#ilfi^yf$ฃ i- ti'j-^ -ป .'.' '-
VOCS:
Halocarbon
Petroleum Hydrocarbon
Trihalomethane
SVOCS:
Petroleum Hydrocarbon
Phenolic
Phthalate
Highest Cortcentrafil^H * ^ , r
V 1998 s ป;: e--,
236.
29.6
26.3
44.
175.
72.3
/lVr '-; aff>97 ' sr iซ;.
172.
ND
5.08
ND
ND
43.2
1. Highest reported concentration of any compound in category out of 19 wells sampled in both years.
Table 3b. Reported Organic Compounds
:;fซapfiVft;|^:iA, - ^
'jft^RP5*1^'^''1 " "
VOCS:
Halocarbon
Petroleum Hydrocarbon
Trihalomethane
SVOCS:
Petroleum Hydrocarbon
Phenolic
Phthalate
NuBi^fDfWfeflsFfeporteain1 :^ ys Aj3r
*-'-4,r "\ ( 1-99(1 L, ~~ t ^ "':'' :?
6.
2.
12.
3.
7.
12.
1 ' ' '~ f' - - '" -^'SSir9rM;ฃq. " -" ' ,"i - i -^"' i" -~~' ^fe
.f ' L? - . v "' '' .""-^ F* "f^f*-f'ry3 - J' i !.:- -|j~.-,!+^:^-.rir|jJi
3.
0.
1.
0.
0.
6.
1. Number of wells out of 19 sampled in both years in which any compound in the indicated category was
reported at any concentration.
199
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 4. Ground/Surface Water Relationship
Major Ions:
Chloride
Sodium
Sulfate
Elements:
Arsenic
Barium
Iron
6.96
6.76
323
6.85
725.
13.2
Sa^^^SilJaHiiKk'.-s3A~*-^ifai!i
300.
340.
150.
2.
130.
0.01
1. Site data. Mean for upgradient monitoring wells (1994-1997).
2. USGSdata. Median for river at nearby gaging station (1981-1995),
Table 5. Grouped Inorganic Data
1996 Data:
BF
62.2
4.75
1.04
35.55
UF
59.4
2.97
.5
0.5
BF:UF Ratio
1.04
1.60
2.06
8.46
1997 Data:
BF
58.66
4.47
.55
3.06
UF
58.48
4.02
.5
2.76
BF:UF Ratio
1.00
1.11
1.11
1996 Data:
BF
204.
7.31
12.8
33.5
UF
181.
5.82
6.69
16.6
BF:UF Ratio
1.13
1.26
1.91
2.02
1997 Data:
BF
177.
3.93
7.62
10.2
UF
174.
4.65
4.52
9.5
BF:UF Ratio
1.02
0.845
1.69
1.07
1996 Data:
BF
1.21
0.94
8.72
305.
UF
0.10
0.05
9.68
196.
BF:UF Ratio
12.1
18.8
0.901
1.56
1997 Data:
BF
0.31
0.10
9.2
231.
UF
0.29
0.06
3.8
226.
BF:UF Ratio
1.06
1.67
2.42
1.02
1. All concentrations are median values for grouped data in units of mg/L.
200
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Milliequivalents per liter
54321012345
Ca
Mg
Nป+K
ft
HCOj+COj
S04
Cl
NO3
Figure 1. Grouped Median Data STIFF Diagrams -1996
a. Wells in unfilled boreholes (UF).
Milliequivalents per liter
54321012345
Ca
Mg
Nป+K
Fe
HCOj+COj
SO4
Cl
NO,
b. Wells in bentonite filled boreholes (BF).
201
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
202
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
GENERAL
203
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
204
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
SAMPLE INTRODUCTION TECHNIQUES FOR FAST GC ANALYSIS OF ORGANOCHLORINE PESTICIDES
C. Eric Boswell. Michael S. Clark, and Rattana M. Woodard
National Air and Radiation Environmental Laboratory, 540 South Morris Avenue,
Montgomery, Alabama 36115-2601
ABSTRACT
Advances in GC instrumentation in recent years have allowed the analysis of organochlorine pesticides to
become more efficient. Pressure programable injectors, for example, have allowed analysts to reduce run times
as well as control injector phenomena such as vapor cloud expansion and thermal degradation. GC analysts in
environmental laboratories have many options available to them in configuring instrumentation, especially in the
techniques they choose for sample introduction. When attempting to maximize the benefits of modem GC
instrumentation, analysts must be prepared to adapt their working knowledge of classical GC techniques to
newer ones. At the National Air and Radiation Environmental Laboratory (NAREL) we learned that several of our
ideas about classical splitless injections had to be modified when we optimized SW846 Method 8081 for speed,
Endrin/DDT degradation, and chromatographic resolution.
In an attempt to reduce the run time for SW846 Method 8081, we configured a GC with dual pressure
programmable injectors, two dissimilar 0.53 mm capillary columns, and dual ECDs. While the biggest time
savings resulted from a relatively high initial oven temperature, having the ability to pressure program the
injectors also helped to reduce run times. Programming the injectors at a high initial pressure also reduced
Endrin/DDT degradation by insuring a short injector residence time. A high initial oven temperature and high
initial injector pressure presented a challenge since high pressure applications typically depend on low initial
oven temperatures to achieve solvent focusing. In this analysis solvent focusing is not practical because the
boiling point difference between the solvent, hexane, and the earliest elating compound is too large. We found
that although we wanted to use techniques that seemed to preclude each other we could empirically match our
starting oven temperature with a relatively high initial injector pressure and have a degree of cold trapping to
achieve both short GC run times, low Endrin/DDT degradation, and good chromatographic resolution. This
deviation from classical splitless injection used both the principles of solvent focusing and high pressure injection
to achieve the desired goal. The result was a GC method for twenty-two of the organochlorine pesticides on the
Method 8081 list with a run time of less than 20 minutes, very low Endrin/DDT degradation, good resolution on
difficult analyte pairs, and all the benefits of splitless injections.
INTRODUCTION
When attempting to maximize the benefits of modern GC instrumentation, analysts must balance their desire to
improve the efficiency of an analysis with certain pragmatic concerns: (1) ease of maintenance, (2) use of
standard consumables, and (3) ruggedness of new procedures. At the NAREL we wanted a GC method for
twenty-two of the organochlorine pesticides on the Method 8081 list with a run time of less than 20 minutes, very
low Endrin/DDT degradation, good resolution on difficult analyte pairs, and all the benefits of splitless injections.
We also wanted the resulting chromatographic system to be one that was easy to use by any reasonably trained
analyst. Most of the modifications employed in our design were in the sample introduction part of the
chromatographic analysis.
We configured a GC with dual pressure programmable injectors, two dissimilar 0.53 mm capillary columns, and
dual ECDs. Although most of the work was in choosing the appropriate sample introduction techniques, there
were other parameters to optimize. For example, we chose a relatively high initial oven temperature to reduce
the cycle time of the analysis We also adjusted the detector makeup gas flow to achieve acceptable sensitivity.
We wanted to take advantage of injector pressure programming which has become standard equipment on most
GCs. We chose splitless injections over direct injections because we believe the splitless technique to be more
rugged. In our dual injector system, we kept the injector pressures similar during injection and programmed them
differently after the split valve opened to achieve chromatographic resolution. We chose a single taper inlet
sleeves because they are easily obtainable and easy to maintain. We chose a compromise initial oven
temperature that was hot enough to keep the GC cycle time short and low enough to prevent the earliest eluting
compounds from tailing too much. This is different from classical splitless injection technique where the initial
oven temperature is kept low until most of the sample can slowly migrate from the inlet sleeve to the head of the
column. This deviation from classical splitless injection used both the principles of solvent focusing and high
pressure injection to achieve the desired goal. The result was a GC method for twenty-two of the organochlorine
pesticides on the Method 8081 list with a run time of less than 20 minutes, very low Endrin/DDT degradation,
205
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
good resolution on difficult analyte pairs, and all the benefits of splitless injections.
EXPERIMENTAL
A mix containing twenty pesticides, two surrogate standards, and three internal standards (Ultra Scientific) was
prepared from the dilution of certified standard mixes with pesticide grade hexane (Burdick & Jackson). Our
experimental results are best discussed in terms of injection technique, pressure programming, and column
selection.
We decided to use splitless injections rather than direct injections to both simplify routine maintenance and
increase system ruggedness. Using a 4 mm ID single taper inlet liner eased routine column maintenance
because the insertion distance of the column into the liner was the only critical step. A 1.0 ul_ volume of hexane
extract will expand to approximately 300 uL of vapor at 6.5 psig. This is easily contained within the 900 uL of
available volume in a 4mm ID liner. We found that a pressure increase after injection provided two benefits: (1)
the vapor cloud moved onto the analytical column in a smaller plug and (2) the compounds were pushed through
the column quicker. We chose 12.4 psig as the injector pressure which resulted in a column linear velocity of 102
cm/sec. We used 120ฐC as our initial oven temperature. These represent a compromise between classical
solvent focusing and high pressure injection. Moving the sample vapor cloud onto the column quickly has several
benefits. Along with keeping early eluting peaks from tailing, high pressure injections dramatically reduce
Endrin/DDT degradation. Combined Endrin/DDT degradation rarely exceeds 5.0% with this configuration.
Optimization of the opening and closing times for the split vent are still crucial for quantitative accuracy.
We chose analytical columns based on chromatographic separation and ruggedness. A 5% diphenyl 95%
dimethyl polysiloxane column such as Restek's Rtx-5, provided both the good chromatographic separation and a
high temperature limit. As a confirmation column we chose a hybrid column developed by Tammy Carey1 that
met our chromatographic needs. The GC oven temperature program is given in Table 1 along with the other
chromatographic conditions. The temperature program was optimized to separate critical pairs such as
Endosulfan I/alpha Chlordane and Dieldrin/DDE on the Rtx-5 column and Heptachlor epoxide/gamma-Chlordane
and Endosulfan II/DDT on the Hybrid column. Chromatograms of the twenty-two organochlorine pesticides are
shown in Figure 1 and calibration curve summaries are given in Tables 2 and 3.
Table 1. Chromatographic Conditions
GC Parameter
Carrier Gas
Injector
Pressure Program
Temperature Program
Detector
Value
Helium
Splitless, 1uL, Purge Delay 0.75 min.
Inlet Temperature 250ฐC
Initial Linear Velocity: 102 cm/sec. (ฃ
Rtx-5
12.4 psi Hold 0.50 min.
to 5.8 psi @ 99.0 psi/min. Hold 4.00
to 14.5 psi @ 0.75 psi/min.
to 19.0 psi @ 1.5 psi/min.
Hybrid
12.4 psi Hold 0.50 min.
to 7.5 psi @ 99.0 psi/min. Hold 6.00
to 14.0 psi @ 0.75 psi/min.
to 20.0 psi @ 1 .5 psi/min.
g 120ฐC
min.
min.
120ฐC Hold 0.5 min.
to150ฐC@30ฐC/min.
to 200ฐC @ 4ฐC/min.
to 285ฐC @ 20ฐC/min. Hold 1.0 min.
N2 Makeup @ 50 mL/min.
Anode Purge @ 5mL/min.
ECD@310ฐC
SUMMARY
At the NAREL we wanted a GC method for twenty-two of the organochlorine pesticides on the Method 8081 list
with a run time of less than 20 minutes, very low Endrin/DDT degradation, good resolution on difficult analyte
206
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
pairs, and all the benefits of splitless injections. We also wanted the resulting chromatographic system to be one
that was easy to use by any trained analyst. Although most of the work was in choosing the appropriate sample
introduction techniques, there were other parameters to optimize. We deviated from classical splitless injection
technique slightly to accommodate the use of pressure programmable injectors. This deviation from classical
splitless injection used both the principles of solvent focusing and high pressure injection to achieve the desired
goal. The result was a GC method for twenty-two of the organochlorine pesticides on the Method 8081 list with a
run time of less than 20 minutes, very low Endrin/DDT degradation, good resolution on difficult analyte pairs and
all the benefits of splitless injections.
REFERENCES
1. Carey, Tamela; Chlorinated Pesticide Analysis Using Hybrid Columns (unpublished), CH2M Hill, Montgomery,
AL
Table 2. Calibration Curve Summary of 5% Diphenyl Column
INITIAL CALIBRATION REPORT
INSTRUMENT:
CALIBRATION DATE: 03/09/97
GC#5/FRONT (5%DP)
METHOD:
SEQUENCE:
DATA FILES:
SW846/ 8081
E030997A
010F010-014F0101
COMPONENT
Pentachlorobenzene/IS#1
Tetrachloro-m-xylene
alpha-BHC
beta-BHC
Lindane
delta-BHC
Heptachlor
Aldrin
Heptachlor Epoxide
gamma-Chlordane
Endosulfan I
alpha-Chlordane
Dieldrin
DDE
o-p'-DDD/IS#2
Endrin
Endosulfan II
ODD
Endrin Aldehyde
Endosulfan Sulfate
DDT
Endrin Ketone
Methoxyclor
Pentabromobiphenyl/IS#3
Decachlorobiphenyl
5ppb
1.00
0.65
0.51
0.37
0.50
0.46
0.70
1.16
1.47
1.65
1.35
1.69
1.35
1.44
1.00
1.46
1.40
1.12
1.16
1.33
1.22
1.55
0.45
1.00
1.55
10ppb
1.00
0.66
0.54
0.37
0.53
0.47
0.67
1.28
1.46
1.65
1.37
1.71
1.29
1.35
1.00
1.40
1.48
1.05
1.17
1.22
1.12
1.51
0.46
1.00
1.53
25ppb
1.00
0.72
0.63
0.39
0.60
0.54
0.68
1.37
1.46
1.70
1.42
1.78
1.31
1.47
1.00
1.37
1.53
1.02
1.18
1.17
1.15
1.51
0.45
1.00
1.44
50ppb
1.00
0.74
0.75
0.40
0.69
0.65
0.67
1.40
1.43
1.70
1.41
1.77
1.32
1.59
1.00
1.37
1.50
1.07
1.10
1.17
1.19
1.50
0.45
1.00
1.35
100ppb
1.00
0.73
0.80
0.39
0.74
0.69
0.66
1.40
1.35
1.61
1.34
1.67
1.25
1.55
1.00
1.28
1.38
1.01
1.01
1.13
1.17
1.43
0.43
1.00
1.34
SD
0.00
0.04
0.11
0.01
0.09
0.09
0.01
0.09
0.04
0.03
0.03
0.04
0.03
0.08
0.00
0.06
0.06
0.04
0.07
0.07
0.04
0.04
0.01
0.00
0.09
MEAN
RRF
1.00
0.70
0.65
0.38
0.61
0.56
0.68
r 1.32
1.43
1.66
1.38
1.73
1.31
1.48
1.00
1.37
1.46
1.05
1.12
1.20
1.17
1.50
0.45
1.00
1.44
%RSD
NA
5.13
17.61
3.40
14.88
16.77
1.99
7.05
2.94
2.09
2.29
2.60
2.65
5.62
NA
4.26
4.01
3.60
5.79
5.93
3.05
2.56
2.36
NA
6.16
P/F
NA
NA
NA
COMMENTS:
207
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 3. Calibration Curve Summary of Hybrid Column
INITIAL CALIBRATION REPORT
INSTRUMENT:
GC#5/REAR (HYBD)
CALIBRATION DATE: 03/09/97
METHOD:
SEQUENCE:
DATA FILES:
SW846/ 8081
E030997A
060R010-064R0101
COMPONENT
Pentachlorobenzene/IS#1
Tetrachloro-m-xylene
alpha-BHC
Lindane
beta-BHC
Heptachlor
delta-BHC
Aldrin
Heptachlor Epoxide
gamma-Chlordane
alpha-Chlordane
DDE
Endosulfan I
o-p'-DDD/IS#2
Dieldrin
Endrin
ODD
Endosulfan II
DDT
Pentabromobiphenyl/IS#3
Endrin Aldehyde
Methyoxyclor
Endosulfan Sulfate
Endrin Ketone
DCB
5ppb
1.00
0.76
0.65
0.63
0.46
0.95
0.54
0.64
1.10
1.29
1.30
1.29
1.35
1.00
1.22
1.15
0.91
1.06
1.11
1.00
1.09
0.61
1.40
1.80
1.73
10ppb
1.00
0.79
0.65
0.64
0.39
0.94
0.55
0.63
1.10
1.44
1.48
1.28
1.25
1.00
1.15
1.12
0.85
1.03
1.03
1.00
1.04
0.58
1.36
1.73
1.80
25ppb
1.00
0.89
0.77
0.75
0.42
0.93
0.63
0.68
1.12
1.47
1.46
1.29
1.25
r 1.00
1.19
1.16
0.87
1.08
1.04
1.00
1.01
0.55
1.41
1.70
1.82
SOppb
1.00
0.80
0.82
0.79
0.42
0.90
0.70
0.70
1.07
1.43
1.39
1.28
1.20
1.00
1.22
1.12
0.92
1.07
1.04
1.00
0.93
0.51
1.23
1.63
1.69
100ppb
1.00
0.81
0.94
0.89
0.47
0.87
0.81
0.78
1.12
1.50
1.40
1.38
1.25
1.00
1.23
1.15
0.92
1.05
1.06
1.00
0.89
0.48
1.22
1.55
1.54
SD
0.00
0.04
0.11
0.10
0.03
0.03
0.10
0.05
0.02
0.07
0.06
0.04
0.05
0.00
0.03
0.02
0.03
0.02
0.03
0.00
0.07
0.05
0.08
0.06
0.10
MEAN
RRF
1.00
0.81
0.77
0.74
0.43
0.92
0.65
0.69
1.10
1.42
1.40
1.30
1.26
1.00
1.20
1.14
0.90
1.06
1.06
1.00
0.99
0.55
1.32
1.63
1.71
%RSD
NA
5.47
14.20
13.15
6.62
3.35
15.67
7.89
1.88
5.17
4.49
2.86
3.99
NA
2.46
1.48
3.06
1.59
2.55
NA
7.18
8.51
6.25
3.77
5.84
P/F
NA
NA
NA
COMMENTS:
208
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
.OซBS-
4.Oe4-
C OCF-estStci
1O
Figure 1. Chromatograms of twenty-two
Organochlorine Pesticides on a dueal injector
GC system.
C OCI=estSt
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
analytical chemistry is gaining importance because of their persistence and toxicity. Furthermore, many countries
have now passed legislation to ensure that pesticides are used safely and responsibly in order to protect the
public from unsafe levels of pesticide residues. Due to the high polarity, low volatility and thermolability of
carbamates, GC and GC/MS methods are not amenable to this class of compounds. HPLC with UV detection
has been the standard analytical method for analysis and quantitation of polar compounds. Benomyl, a
carbamate and a widely used systemic fungicide for disease control in crops, poses additional analytical
challenges to the environmental chemist due to its extreme instability in the environment. Benomyl decomposes
to three stable major degradation products: 2-aminobenzimidazole (2-AB), methyl 1-H-benzimidazol-2-
ylcarbamate (MBC), and 3-butyl-2,4-dioxo-s-triazino [1,2-a] benzimidazole (STB).
EPA Method 631 for the determination of benomyl and MBC in industrial and municipal wastewater restricts the
detection and quantitation to MBC only since benomyl is hydrolyzed to MBC in this method. This isocratic UV
method has a Method Detection Limit (MDL) of 8.7 ppb for MBC.
The UV method adapted in this report monitors the three major degradation products of benomyl by gradient
elution. The UV analysis has a method detection limit of 2 ppb for 2-AB, MBC, and STB.
METHOD
Materials
2-AB was obtained from Aldrich Chemical Company (Milwaukee, Wl). MBC was obtained from Chem Services
(West Chester, PA) and STB was obtained from E.I. du Pont Nemours & Co. (Wilmington, DE). Solvents used
for HPLC and extraction were HPLC grade acetonitrile, methanol and water from JT Baker (Philipsburg, NJ). The
formic acid used in the HPLC solvents, as well as the injection solvent, was from Sigma (St. Louis, MO). The
standard bore (4.6 x 250 mm) HPLC column was a C18 with a 5-um particle size by Whatman (Clifton, NJ). The
narrow bore column (2 x 50 mm) was a Monitor C18 with a 3-um particle size by Column Engineering (Ontario,
CA). A Waters Oasis (Milford, MA) 3-mL, 60-mg SPE cartridge was used for the solid phase extraction.
Preparation. Benlate Extraction Procedure
A Benlate mix spike solution was comprised of a mixture of 2-AB and STB in methanol at a concentration of 2
ppm. 200 mL of water sample was poured into a 400-mL beaker. HPLC grade water was used for blanks and
laboratory fortified blanks. The pH of all samples was checked and recorded using a pH meter. For any spiked
samples (laboratory fortified blanks and matrix spikes), 200 uL of Benlate mix spike solution was added and
mixed well. Two Normal NaOH was added to adjust the pH of all the samples to 10. The solid phase extraction
(SPE) cartridge was conditioned by aspirating 5 mL of methanol through the cartridge followed by 10 mL of
HPLC grade water without allowing the cartridge to dry. The sample was filtered through the conditioned SPE
cartridge under vacuum. The sample beaker was then rinsed with 5 mL of HPLC grade water adjusted to pH 10.
The rinse was passed through the SPE cartridge and vacuum dried for 3 min. Elution with 5 mL of methanol and
collection of the eluate into a 15 mL centrifuge test tube was completed. The eluate was dried down with nitrogen
in a water bath at 40ฐC to a volume of 1 mL. One mL of 0.2% formic acid was added and mixed well.
HPLC Conditions
A Shimadzu VP series HPLC (Columbia, MD) was used with the gradients, flow rates and columns listed in Table
1.0 and Table 1.1. The flow from the standard bore column was split down from 1 mL/min to approximately 200
ul_/min into the Turbolonspray source at ambient temperature. For the narrow bore column no split was
necessary and the entire effluent was sent into a Turbolonspray source set to 200ฐC.
Table 1.0. HPLC Gradient
Standard Bore Column
4.6 x 250 mm
Flow rate: 1 mL/min
Whatman C18 5 urn
Time (min)
0
10
20
20.1
30
Solvent A
0.1% Formic Acid
70
30
15
70
70
Solvent B
Methanol + 0.1% Formic Acid
30
70
85
30
30
210
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1.1. HPLC Gradient
Narrow Bore Column
2 x 50 mm
Flow rate: 0.2 mL/min
Column Engineering
Monitor C1 8 3 pm
Time(min)
0
0.85
2.25
4
6
7
7.1
10
0.1% Formic Acid
100
50
45
45
30
30
100
100
Acetonitrile + 0.1%
0
50
55
55
70
70
0
0
Formic Acid
Mass Spectrometer
A PE Sciex API 150 (Concord, Ontario, Canada) single quadrupole mass spectrometer was operated in positive
ion mode with a Turbolonspray source. The source temperature was ambient for the standard bore HPLC column
and was set to 200ฐC for the narrow bore column. The compounds were monitored using selected ion monitoring
(SIM) with the following m/z values:
m/z-134.1
The ion optics path was optimized for each compound separately in the separations solvent. The values for the
orifice, ring and QO are listed in Table 2.0. The mass spectrometer was operated at unit resolution with an
ionspray voltage of 4600 V.
Table 2.0. Ion Path Voltages
Compound
2-AB
MBC
ST13
OR
45
20
30
RNG
160
120
155
QO
-4
-3
-4
RESULTS
Initially the primary emphasis was on transferring an already validated UV HPLC method to a single quadrupole
MS method altering as few parameters as possible. Only two changes to the original procedure were necessary
for the method transfer. Formic acid was added to the mobile phase of the HPLC solvents, and the injection
solvent was changed to contain a final concentration of 0.1% formic acid in 50% methanol. All other parameters
remained the same. As a consequence of the change in pH, the elution order of the compounds changed but
chromatographic resolution was maintained and all compounds were baseline resolved. The eluant from the
4.6-mm column was initially split to 50 pL/min into the Turbolonspray source at room temperature. Occasionally
at a high percentage of organic, the 50-uL/min split into the mass spectrometer would lose flow resulting in a loss
of signal. This was corrected by decreasing the split ratio sending approximately 200 uL/min into the ion source
at the expense of sensitivity. Nevertheless, the limit of quantitation for the standards (0.01 ppm) was an order of
magnitude better than that of the UV analysis of standards (0.1 ppm). (See Figure 1.0 and Figure 1.1.)
The Method Detection Limit (MDL) was calculated with the following equation:
MDLcalculated = t x s x Vext/Vsampie
Vext = 2 mL t = Student Coefficient
Vsampie = 200mL s = Standard Deviation
Table 3 0 contains the solid phase extraction validation data and Table 4.0 contains the MDL calculations for
both UV and LC/MS analyses. While the calculated MDL for the UV and LC/MS methods are similar, the actual
211
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
MDL achieved by the two methods is an order of magnitude different.
Table 3.0. SPE Validation for Water
Recoveries
UV 4.6x250 mm
Compound Spjke Level
2-AB 5 ug/L
MBC 5 ug/L
STB 5 ug/L
MDL-1
ug/L
25
5
54
%
50
100
108
MDL-2
ug/L
2.5
53
5.5
%
50
106
110
MDL-3
ug/L
2.7
54
5.8
%
54
108
112
MDL-4
ug/L
2.7
4.6
4.7
%
54
92
94
MDL-5
ug/L
32
56
5.7
%
64
112
114
MDL-6
gg/L
3.3
5.3
5.5
%
66
106
110
MDL-7
yg/L
3.1
5.2
53
%
62
104
106
%Mean
57.1
104
108
SD
6.7
6.4
6.6
%RSD
12
6
6
LC/MS 4.6. 250 mm
2-AB 2 ug/L
MBC 2 ug/L
STB 2 ug/L
1 74
1.98
1 57
87
99
79
1.86
2.12
1 64
93
106
82
1 76
2.04
1.54
88
102
77
1.9
2.2
1.58
95
110
79
1.86
2.18
1.54
93
109
77
1.94
2.26
1.62
97
113
81
1.64
218
1 66
82
109
83
90.7
107
79.7
3.9
5.2
2
4
5
3
Table 4.0. MDL
Method Detection Limit
UV 4.6 x 250mm
Compound
2-AB
MBC
STB
Student
Coefficient
3.143
3.143
3.143
n
7
7
7
Std. Deviation
6.7
6.4
6.6
MDL
(calculated)
0.21 ppb
0.20 ppb
0.21 ppb
MDL (actual)
2 ppb
2 ppb
2 ppb
LC/MS 4.6. 250 mm
2-AB
MBC
STB
3.143
3.143
3.143
7
7
7
3.9
5.2
2
0. 12 ppb
0.16 ppb
0.063 ppb
0.2 ppb
0.2 ppb
0.2 ppb
In order to improve the setup and ease of use the LC/MS method was altered further. A 2 X 50-mm HPLC
column replaced the larger diameter column and the entire eluant flow was sent directly into the Turbolonspray
source. The source was operated at 200ฐC. This source temperature gave the best overall response for the three
compounds of interest while not significantly adding to the background signal of the final method. The initial
gradient was based on the same column volumes and used the same buffer system of the larger column. The
result was a fine separation of 2-AB and MBC but STB did not elute as a sharp peak. With the addition of heat
the methanol solvent system generated an additional problem. The background signal increased dramatically
making detection of the low standards difficult. All efforts to reduce the background with methanol as the organic
solvent failed. Methanol was replaced with acetonitrile, resulting in much lower background counts and a much
improved STB peak shape. The gradient was changed to accommodate the stronger solvent. In fact, the peak
shapes improved for all three analytes. 2-AB and MBC each elute in roughly 20 uL and are baseline-resolved.
STB elutes in approximately 100 uL. Because of this difference in elution volumes a period MS method
generated better S/N for each compound. This reduced elution volume from the narrow bore column improved
the limit of quantitation (LOQ) even further. While the reduction in column dimension should only result in a
four-fold gain in sensitivity, the above changes resulted in a 50-fold gain since only 10 uL were injected onto the
2-mm column versus 100 uL injected onto the 4.6-mm column. The small volume of the 2-mm column as
compared to the standard bore column also reduced the analysis time from 30 minutes to 10 minutes.
Figures 1.0, 1.1 and 1.2 are views of the low standards for each method. Each has a S/N greater than 10:1. The
LOQ for the standards is 0.1 ppm, 0.01 ppm, and 0.002 ppm for the UV, 4.6, and 2-mm methods, respectively.
The resultant calibration curves are shown in Figure 2.0. A linear 1/X weighting was used for the LC/MS analyses
while no weighting was used in the UV
analysis. Table 5.0 lists the percent
deviation of the mean from the theoretical
(%DMT) and the %RSD obtained over five
days for the various levels using the 2-mm
LC/MS method. Table 5.0 demonstrates the
high degree of accuracy and consistency
required for successful quantitative results.
', net:
ฃ - H
Figure 1.0. UV Low Standard (0.1 ppm;
100-uL injection volume).
212
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
STB
2-AB
V Mt
1 \
MBC
#;,'ซ#%:; U< k^'i^^vAy,
/W\T
Figure 1.1. 4.6 x 250 mrn Low Standard (0.01 ppm; 100-uL
injection volume).
Figure 1.2. 2 x 50 mm Low Standard (0.002 ppm; 10-uL
injection volume).
Table 5.0. %DMT Statistics
2-AB
MBC
STB
Five-Day Percent Deviation from Theoretical LC/MS 2 x 50 mm method
Level
1000ppb
200 ppb
1 00 ppb
20 ppb
10 ppb
2 ppb
1ppb
2-AB
%DMT
2-AB
96.19
111.15
110.49
106.65
104.38
94.05
78.97
% RSD
2-AB
1.12
0.94
2.35
1.95
4.04
2.51
13.26
MBC
%DMT
MBC
97.8
107.69
106.79
102.2
99.57
98.91
95.67
% RSD
MBC
0.44
1.87
1.98
2.81
2.38
4.67
9.29
STB
%DMT
STB
99.45
105.99
105.38
98.76
91.64
108.2
114.25
% RSD
STB
0.05
1.38
2.06
5.81
5.67
11.72
16.74
While the 0.001-ppm standard is detected with the 2-mm method, the %RSD for the accuracy of the three
compounds was deemed to be too high for the LOQ. The MDL-7 for the 2 x 50-mm LC/MS method needs to be
completed, but based on the previous results for the standards the MDL should be approximately 0.02 ppb.
CONCLUSIONS
A method was developed to quantitate the major degradation products of benomyl in a simple, single quadrupole
LC/MS assay. An existing UV HPLC method was slightly modified for analysis by a PE Sciex. API 150 mass
spectrometer. An immediate order of magnitude gain in sensitivity over the UV MDL resulted. This LC/MS
method obviates the additional step of hydrolysis described in EPA Method 631, allowing for direct quantitation
of the major degradation products of benomyl. Currently the benomyl LC/MS method can be used with extraction
protocols for water. Additional optimization of the HPLC and MS method produced further gains in sensitivity and
also simplified the system setup. The narrow bore method requires the validation process to be completed, but
this should not prove difficult based on the clean matrix provided by the SPE protocol and the consistency of the
standards analyzed. The increased specificity afforded by MS also allowed for a reduction in HPLC and data
analysis time. Future work will result in the development of extraction protocols for waste and sediment. Also,
implementation of an internal standard should improve the results obtained by this method.
213
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Figure 2.0. Calibration Curves for2-AB, MBC, and STB.
METHODS OPTIMIZATION OF MICROWAVE ASSISTED SOLVENT EXTRACTION
TECHNIQUES FOR VARIOUS REGULATORY COMPOUNDS
Greg LeBlanc, Mike Miller
CEM Corp., PO Box 200, Matthews, NC 28106-0200
(704)821 7015
We have seen a period of rapid growth in the separation sciences, with gas and liquid chromatography as well as
the hyphenated techniques becoming commonplace. There is a need to streamline the extraction techniques
preceding these chromatographic analyses. Time and solvent usage have become critical factors. In response to
this need several new sample preparation technologies have emerged. As a result analysts are faced with the
challenge of adapting current methods or developing entirely new ones depending on the extraction technology
they use.
Microwave Assisted Solvent Extraction is a new technique that works with existing solvent regimes with minimal
modifications and significant solvent reduction. It also has the benefit of reducing the extraction time. This study
will review the methods optimization process when using a closed vessel microwave assisted solvent extraction
214
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
technique. It will focus on extraction temperature, sample water content and ratios for solvent mixtures.
Recovery data will be presented for pesticides and fungicides from reference soils and vegetable samples.
NEW DEVELOPMENTS IN CLOSED VESSEL MICROWAVE DIGESTION TECHNOLOGY FOR
PREPARATION OF DIFFICULT ORGANIC SAMPLES FOR AA/ICP ANALYSIS
Greg LeBlanc. Bobbie Haire and Bob Fidler
CEM Corporation, PO Box 200, Matthews, NC 28106-0200
(704)821-7015
Microwave digestion techniques are well accepted as a means of preparing samples prior to AA and ICP
analysis. Pressurization of the sample vessel accelerates the dissolution process by permitting acids to attain
higher temperatures than under ambient conditions. However, it has been historically difficult to control the
microwave digestion process under ultra-high temperature and pressure conditions. As well, throughput of
difficult high-pressure samples using conventional microwave instrument technology has been limited, thus
inhibiting its use for preparation of some difficult organic and inorganic samples. This paper will demonstrate new
microwave instrument technology for optimization of ultra-high pressure and temperature conditions for
microwave preparation of difficult organic and inorganic samples for AA and ICP analysis. Data will be provided
to demonstrate the digestion performance under significantly elevated pressure and temperature conditions as
well as the impact of such conditions on the microwave preparation instrumentation.
EVALUATION OF ICP-OES AND ICP-MS FOR ANALYSIS OF
THE FULL TCLP INORGANIC TARGET ANALYTE LIST
Marc Paustian and Zoe Grosser
The Perkin-Elmer Corporation, 50 Danbury Road, MS-219, Wilton, CT 06897-0219
Abstract
Toxicity Characteristic testing is performed to assess the potential of a material to leach hazardous constituents
after disposal in a landfill. The Toxicity Characteristic Leaching Procedure (TCLP) extract solution is analyzed for
31 organic components and 8 metals and if the regulated limits are exceeded (Table 1) the material is
considered hazardous and must be treated appropriately.
Currently, ICP-OES and ICP-MS have the technical capability of determining the full suite of TCLP elements in
one run. EPA ICP-OES method 601 OB includes Hg in the target analyte list. EPA ICP-MS draft method 6020A
now includes Se and Hg in the target analyte list. This provides an opportunity to demonstrate the utility of TCLP
analysis using one sample preparation and one technique for analysis.
Compliance with the generally accepted requirements of the RCRA program will be demonstrated for the full
suite of TCLP analytes including mercury. ICP-OES and ICP-MS will be compared for the analysis.
The results will be validated with the analysis of the required Quality Control for performance-based methods and
will include detection limits, spike recoveries, and duplicate sample analytical data.
Experimental
Raw TCLP extracts were provided by Trinity River Authority in Dallas Texas. They were prepared by digestion
with EPA method 3015 using the Multiwave Microwave Digestion System (Perkin-Elmer Corp.). To evaluate the
recovery vs. hotplate digestion, for mercury, the samples were also spiked with mercury and digested on a
hotplate using method 3020. Gold was added to some samples to evaluate the effect on retention of mercury
215
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
during the digestion process and resulted in a concentration of 2 mg/L in the final solution.
The samples were analyzed using ICP-OES and ICP-MS, for comparison. The instrumental conditions are shown
in Tables 2 and 3.
Table 1.TCLP Metal Limits
Element
As
Ba
Cd
Cr (Total)
Pb
Hg
Se
MCL(mg/L)
5.0
100.0
1.0
5.0
5.0
0.2
1.0
Table 2. ICP-OES Instrumental Conditions
Optima 3000 DV
Parameter
RF Power
Nebulizer Flow
Auxiliary Flow
Plasma Flow
Sample Pump Flow
Plasma Viewing
Processing Mode
Auto integration
Read Delay
Rinse
Replicates
Background Correction
Settings
1 500 watts
0.4 L/min
0.5 L/min
15.0 L/min
1.5 ml_/min
Axial
Area
5 sec min-20 sec max
45 sec
45 sec
2
one or two points
Table 3. ICP-MS Instrumental Conditions
ELAN 6000
Parameter
RF Power
Nebulizer Flow
Auxiliary Flow
Plasma Flow
Sample Pump Flow
Sample/Skimmer Cones
Scanning Mode
Lens
Read Delay
Rinse
Replicates
Detector Mode
Settings
1500 watts
0.94 L/min
1.0 L/min
15.0 L/min
1.5 mL/min
Nickel
Peak Hopping
Lens Scan Enabled
10 sec
35 sec
3
Dual Mode
Results and Discussion
The instruments were qualified by measuring method detection limits using the procedures specified in the
Federal Register, part 1361 The method detection limits are shown in Table 4 and compared with the MCL
(shown here in ug/L for easier comparison) for the element. It is generally accepted that the MDL should be ten
times less than the MCL to ensure sufficient confidence at the decision-making point. The detection limits for
both the ICP-OES and ICP-MS are well below the limits set by the criteria specified.
216
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 4. Method Detection Limits
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
MCL (ug/L)
5000
5000
100,000
1000
5000
200
5000
1000
MDL (ug/L)
ICP-OES
6
2
0.3
0.1
0.6
8
2
3
MDL (|jg/L)
ICP-MS
0.007
0.09
0.01
0.03
0.05
0.34
0.006
0.18
The results were measured on the samples digested with the microwave procedure for the full suite of elements.
The results are shown in Table 5 and compared with results reported by the supplier. The digested samples
compared well with the results provided. The ICP-OES and ICP-MS results compared well with each other, The
results were all well below the maximum contaminant levels for all elements.
Table 5. Comparison of Sample Results
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
ICP-OES
Sample 1
(Mg/L)
8
8
253
2
15
<8
7
5
ICP-MS
Sample 1
(ug/L)
0.1
5
214
0.4
12
<0.3
4
4
TR-1
(pg/L)
<2
<5
232
<8
11
<0.5
4
<5
ICP-OES
Sample 2
(ug/L)
6
6
260
1
6
<8
5
2
ICP-MS
Sample 2
(ug/L)
0.05
3.4
218
0.08
7
<0.3
2
2
TR-2
(ug/L)
<2
<2
278
<8
5
<0.5
<4
<5
Predigestion spike recoveries were evaluated for samples digested with the microwave procedure. The samples
were spiked with 2.5 mg/L of As, Cr, Pb, Ag, 0. 5 mg/L of Cd, Se, 0.1 mg/L Hg, and 50 mg/L of Ba. The samples
labeled "Spike 2" also had gold added (2 mg/L). The spike recoveries are shown in Table 6 for ICP-OES analysis
and Table 7 for ICP-MS analysis.
Table 6. ICP-OES Spike Recoveries
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
Sample 1
Spike 1
%Rec
33
104
102
106
100
60
99
104
Sample 1
Spike 2
%Rec
2
107
102
108
101
62
102
107
Sample 2
Spike 1
%Rec
30
107
102
107
100
62
100
105
Sample 2
Spike 2
%Rec
2
108
101
108
101
63
100
106
Blank
Spike 1
%Rec
63
99
101
105
99
54
102
96
Blank
Spike 2
%Rec
2
100
101
106
99
60
103
96
The spike recoveries were excellent. The silver recoveries diminished with time, as is typical with a nitric-only
digestion. Since the ICP-MS analysis was performed first and the ICP-OES analysis was performed after several
days had passed the effect is more evident in the ICP-OES results. The addition of small amounts of chloride
with the gold solution in the second spike accelerated the precipitation of silver. The mercury spike recoveries
were less than expected and may be due to insufficient cooling of the vessels before the solution was
transferred. The solutions were not stored in Teflonฎ bottles, providing another possible source of loss. Since the
addition of gold improved the recoveries of mercury only slightly in the microwave digestion, better overall
performance may be obtained if the gold is left out. The effect of gold addition on the retention of mercury in a
traditional hotplate digestion was investigated. The recoveries compared to spikes digested with microwave are
217
-------
shown in Table 8.
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 7. ICP-MS Spike Recoveries
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
Sample 1
Spike 1
92
104
115
101
96
62
109
103
Sample 1
Spike 2
6
100
120
103
96
53
107
108
Sample 2
Spike 1
97
102
116
103
96
92
109
98
Sample 2
Spike 2
4
104
118
103
99
53
106
107
Blank
Spike 1
106
93
117
104
100
55
109
103
Blank
Spike 2
7
102
116
103
101
54
106
113
Table 8. Effect of Gold on Mercury Spike Recovery
Sample 1
Spike
(% Rec)
Sample 1
Spike Dup.
(% Rec)
Sample 1
Spike+Gold
(% Rec)
Sample 1
Spike Dup. +Gold
(% Rec)
Hot Plate Digestion
ICP-OES
ICP-MS
52.1
44.4
32.8
37.0
56.9
51.2
61.5
47.5
Microwave Digestion
ICP-OES
ICP-MS
60
62
62
62
62
53
63
53
The addition of gold only slightly improves the recoveries of mercury using the microwave digestion. The
hotplate digestion results show more dramatic recovery improvements. In addition, the duplicates are much more
consistent in the hotplate digestion when gold is added before the digestion process.
Summary
Much progress has made in increasing the scope of inorganic methods to include all the analytes of interest.
Analyses can be performed more economically when only one sample preparation and one analysis technique
are needed to fully evaluate a sample.
This work has demonstrated that mercury can be successfully determined in a digested TCLP extract matrix, as
a part of the full suite of elements traditionally measured. Both ICP-OES and ICP-MS show excellent method
detection limits for the elements measured and good predigestion spike recoveries with the exception of
mercury. Further work is needed to define the mechanism for mercury loss in the digestion and transfer
procedure used before the instrumental analysis.
Method 601 OB for ICP-OES and proposed 6020A for ICP-MS contain the full list of analytes for TCLP analysis.
This work validates the utility of these methods for this type of matrix. Microwave sample preparation procedures
are especially useful when mercury is included in the analyte list, coupled with careful sample transfer and
storage.
Acknowledgement
The authors would like to thank Bill Cyrus of the Trinity River Authority, Dallas, Texas for providing the TCLP
extracts.
References
1. US Code of Federal Regulations 40, Ch. 1, Pt 136, Appendix B.
218
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
CLEAN METALS SAMPLING
Roger E. Stewart II
Virginia Department of Environmental Quality, Office of Water Quality Assessment and Planning,
629 East Main Street, Richmond, Virginia 23219
ABSTRACT
Scientific evidence1 of aquatic organism toxicity to low concentrations of dissolved metals has led to the need for
accurate and precise measurements of these compounds in a variety of aqueous matrices. A practical method2
will be presented for the collection and analysis of freshwater and wastewater samples for trace metals. A hands
on demonstration of clean techniques will be presented to show the simplicity and ease of collecting scientifically
defensible samples for compliance monitoring purposes. Field sampling equipment specifications, collection
procedures, the importance of blanks, and preservation of samples, will be discussed as topics important to
designing a procedure for collecting water samples with metals concentrations in the range where toxicity is
significant.
INTRODUCTION
To be protective of aquatic organisms and human health the US Environmental Protection Agency has
established water quality standards3, which specify safe concentration levels in background waters, for toxic
metal species. Regulatory agencies must evaluate the concentrations of analytes in receiving waters and
discharges to determine "reasonable potential" for impact to an ecosystem. Reliable methods for measuring
trace concentrations of target compounds are essential to developing permit limits on discharges.
As regulatory levels for metals have decreased, sometimes several orders of magnitude, field collection
protocols which are practical and applicable to a wide variety of site conditions and personnel must be
developed.
The US Geological Survey (USGS) and the US Environmental Protection Agency (USEPA) have made progress
in the last two years towards developing widely applicable field and laboratory procedures. The USGS has
focused on internal procedures4 designed to be used by USGS personnel. As a regulatory agency the USEPA
has developed procedures56'789101112 which can be used by monitoring groups, public agencies, and the
regulated community.
The freshwaters appropriate for collection by this method include all surface waters and groundwaters with a
specific conductivity of approximately 1000 umhos/cm or less. Appropriate wastewaters include treated effluents
(conductivity < 1000 umhos/cm). Saltwaters, brackish waters, highly turbid wastewaters, i.e. landfill leachates,
are not appropriate as they require special laboratory preparation and analysis.
The protocols contained in this Standard Operating Procedure are applicable to the compounds listed in Table 1
Target Analytes, Analytical Test Methods, and Detection Limits on page 8.
Additionally this SOP is suitable for freshwater and treated final effluents with concentrations of toxic metals
below approximately 200 ug/L. The 200 ug/L threshold should be applied cautiously as this is only a
generalization of the effect of contamination. For example, because of well documented contamination problems
with copper and zinc, if a final effluent has historically had copper or zinc reported in the 0.2 mg/L range use of
this protocol may reveal that the actual concentrations are significantly lower. However if the historical numbers
for cadmium, arsenic, or mercury have been in the 0.2 mg/L range use of this protocol may not affect these
concentrations. Another factor to consider when determining the applicability of this SOP is that typically the
reporting level of historic data is much higher than the lowest concentrations when compared to Water Quality
Standards.
For concentrations above approximately 200 ug/L, existing 40 CFR 136 procedures are adequate and contain
the necessary Quality Controls (including the requirement to collect blanks) to make reliable measurements in
the mg/L range. The United States EPA Region III has prepared extensive guidance for existing and new data
which falls into this range.
METHODOLOGY
Improving on a combination of the more salient features from the USGS, USEPA, and others413 the Virginia
219
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Department of Environmental Quality (VADEQ) began a pilot project designed to 1) determine if current
technology could be used to accurately measure metal concentrations in water below 1.0 ug/L using clean
techniques, 2) develop protocols for wastewater treatment plant effluent collection and analysis which could be
incorporated into guidance and, 3) develop freshwater protocols for an ambient water quality monitoring
program.
The main design considerations were to develop a system to 1) collect field samples free of contamination and,
2) minimize the level of effort required by field technicians.
When measuring at trace concentrations small amounts of background contamination and interference can be
significant. The most likely sources of contamination are:
1. improper sample handling techniques,
2. improperly cleaned sampling equipment and sample containers, and
3. atmospheric dust and debris.
These ideas led to the design of a closed loop sample container which could be filled onsite, transported to the
laboratory, prepared, and digested all in the same container. Figure 1 Loop Sample Bottle shows the container
used to collect the samples.
Figure 2 Sample Collection Scheme illustrates the closed loop sample collection system. Sample collection is
accomplished by pumping water from the sampling zone through a teflon tube, peristaltic tubing, and a capsule
filter, directly into the loop sample container.
The entire process consists of the following steps:
1. Sample containers and tubing kits are cleaned and certified in a controlled laboratory environment.
2. The containers and kits are packaged into coolers and shipped to the field.
3. Blank ultra pure water from a blank sample container is processed through the tubing kit to condition and
clean the capsule filter.
4. Ultra pure water from the sample container is pumped through the tubing assembly into the empty blank
container to produce the field blank.
5. Sample is collected into the empty sample container using the tubing assembly.
6. Blanks and samples are shipped to the laboratory.
7. All samples and blanks are acidified, digested, prepared in the original container.
8. Samples are analyzed accordingly.
The advantage of this design is that field technicians receive the sample containers and equipment ready to use.
Other than filter conditioning1415 there are no cleaning steps required and no need to preserve the samples other
than icing. The sample containers include the water to be used for field blank collection with the blank collection
exactly duplicating the site conditions of the sample collection. All the plasticware is inexpensive and can be
discarded after single use.
DISCUSSION
Ambient Sample Collection Protocol
1. Locate an area where sample processing will occur. This should be an area free of falling debris and swirling
dust, flat, smooth, and protected from the wind. The tailgate of a vehicle or the back of a Suburban are good
locations.
2. Locate the equipment box and coolers containing the sample containers and kits in the area where sample
processing will occur.
3. Cover the work area with a large piece of plastic film. Set out the pump and connect the battery. Switch
pump on for a quick burst to check that it is working. Dial the pump speed to 5.
4. Remove a tubing kit, two sample containers, and a bridge bottle from the cooler and place on the plastic
near the pump.
5. Remove a pack of sample gloves from the storage container and place on the plastic.
6. Remove the plastic sample caddie from the storage box and place it on the sample processing area near the
pump. Secure the two sample bottles in the caddie.
220
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Bridge Bottle Filling
1 . Locate the sample weights for connection to the BRIDGE BOTTLE.
2. Locate the polypropylene sampling rope spool, cut a sufficient length of rope to allow for deployment, and
place on plastic.
3. Don one or two pairs of vinyl gloves using clean precautions.
4. Tie one end of the sampling rope to the five pound weight leaving approximately a V long end for connection
to the BRIDGE BOTTLE.
5. Untie or open by tearing the top of the outer plastic bag containing the BRIDGE BOTTLE.
6. Reach into the outer bag and untie or tear the inner bag near the handle connection. Check the configuration
of the tubing to ensure that proper filling will occur. While the bottle is still in the inner bag it is acceptable to
remove the top fitting to check the inner sipper tube. Adjust all fittings appropriately.
7. When the fittings have been properly secured and adjusted remove the BRIDGE BOTTLE from the inner bag
and lay on the plastic film. Tie the weighted end of the rope onto the handle of the bottle leaving about 6" of
line between the bottle and the weight.
8. Proceed to the sampling location with the BRIDGE BOTTLE apparatus. If appropriate carry several extra
pairs of gloves to the site to facilitate bridge bottle handling.
9. When deploying from bridges with moderate to low stream velocities collect the sample upstream of the
bridge by lowering the assembly into the water. Ensure that the assembly does not contact any structures or
other objects as it is lowered into the water.
10. Once in the water the weight will partially submerge the BRIDGE BOTTLE which will begin to fill. Check to
insure the air release tube is above the water level and not obstructed. When the bottle is first submerged a
good indication it is filling properly is a small slug of water may be expelled from the air vent tube. The bottle
will fill quickly if it has been properly adjusted.
11. Problems with filling from bridges can occur when stream velocities are high. Sampling on the downstream
side of bridges is acceptable to avoid the risk of losing the assembly due to the current sweeping it under a
bridge or other obstruction. When stream velocities are high an additional 7.5 pounds of weight will aid in
sample collection. The added weight will cause the container to sink lower when partially filled which may
submerge the vent tube. The vent tube can be extended past the bottom of the bottle to prevent filling with
water when the weight is heavy or the water is rough.
12. Other problems with filling can occur when the inlet tube is clogged, the vent tube contains a slug of water or
other obstruction, the vent tube is below the surface of the water, the weight is not positioned close enough
to the bottle, or the vent tube or inlet tube has become disconnected from the bottle.
13. When the BRIDGE BOTTLE is approximately 1/3 to 1/2 full retrieve the bottle and return to the sample
processing area. Ensure that the assembly does not contact any structures or other objects as it is retrieved.
14. When deploying while wading or from a small craft the BRIDGE BOTTLE can be submerged by hand without
the weights.
15. When the water level at the sample site is very shallow it may be difficult to submerge the BRIDGE BOTTLE
deep enough to begin siphoning. The alternative is to use the effluent sample configuration where the stream
sample is pumped directly into the loop sample container. Sampling in this manner requires the pump
assembly to be transported to the site. This is best accomplished by attaching the pump assembly to a
backpack.
16. Once the BRIDGE BOTTLE has been brought back to the sample processing area set it next to the pump
and remove the weight. With the inlet and vent tubing properly configured the BRIDGE BOTTLE can remain
on the plastic outside of a bag without any danger of atmospheric contamination.
Ambient Dissolved Grab Blank
1 . Refer to Figure 2 Sample Collection Scheme for the schematic of the field sampling equipment used to
process blanks and samples.
2. Determine which tech will be clean hands and which will be dirty hands.
3. Dirty hands and clean hands don one or two pairs of vinyl gloves. Dirty hands opens the sample bottles outer
plastic bag, clean hands opens the inner plastic bag.
4. Dirty hands opens the grab kit's outer plastic bag, clean hands opens the inner plastic bag and removes the
tubing assembly.
5. Clean hands disconnects one side of the sample loop on a sample container. Clean hands connects the end
of the tubing kit opposite the filter to the opened sample container. Remember the sample container is full of
clean water from the lab.
6. Dirty hands connects the peristaltic tubing at approximately the mid-point of the length to the field pump,
clean hands inverts the sample container, and dirty hands switches on the pump.
221
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
7. Process the entire contents, 1000 mis, of the sample container through the tubing and filter apparatus at a
flow rate of 500 mls/min. (pump setting of 5). At the beginning of the sample processing orient the filter
cartridge with the flow arrow pointing up. This will insure proper wetting of the filter. After the last noticeable
air bubbles have been expelled from the filter it can be oriented in any direction during the rinsing. When the
sample container is nearing empty orient the capsule filter with the arrow down. Once the sample container
is empty continue processing until the free water in the tubing and filter has been expelled.
8. Dirty hands turns off the pump. This step is a rinse of the filter which cleans and conditions the media. The
rinse can be pumped directly to waste.
9. Use the same flow and orientation scheme throughout all sample and blank processing.
Field Blank
1 . Clean hands removes the pump tubing from the empty bottle and connects the capsule filter to the empty
container via the sample loop tubing. Remove one side of the loop fitting from a second sample container
and connect the free end of the tubing to the container. Remember that this second bottle is full of clean
water from the lab. Invert the container, dirty hands switches on the pump. Process the entire contents,
approximately 1000 mis, of the sample container through the tubing and filter apparatus into the first sample
container.
2. Dirty hands switches off the pump when the filter has been emptied. Clean hands disconnects the outlet
tubing from the blank sample container and immediately reconnects the loop tubing, seals the inner bag and
places inside the outer bag. Dirty hands identifies this as a blank by recording the ULTRA # as the
CONTAINER ID, seals the outer bags, and places on ice in a sample cooler.
3. Clean hands immediately (immediately means the sooner the switch is made the less likely contamination
can adhere to the end of an open tube, immediately means less than one minute) disconnects the tubing
from the sample container and then connects the free end containing the filter, which was just removed from
the blank sample container, to the empty sample container. Clean hands then disconnects the vent tubing
from the BRIDGE BOTTLE and connects the pump tubing in place of the vent tubing.
4. The field blank collected in this manner is a comprehensive blank because it is collected in the same
equipment as the sample and it is processed like the sample through all steps of the protocol. This is the
most important check of contamination in the protocol.
Ambient Dissolved Grab
1. In a field notebook dirty hands records the sample container identification number, date, time. Clean hands
secures the sample container in the sample caddie.
2. While keeping the BRIDGE BOTTLE near level dirty hands holds the bottle at the midsection or lower and
shakes the vessel to ensure mixing. Dirty hands switches on the pump and clean hands collects a full sample
container while following the filter orientation scheme. It is acceptable to fill the sample container to
overflowing, however avoid filtering more than 1000 mis through the filter.
3 Once the sample container is full, dirty hands switches off the pump. Clean hands disconnects the outlet
tubing from the sample container and immediately reconnects the loop tubing on the sample container, seals
the inner bag and places inside the outer bag. Dirty hands identifies the sample by recording the ULTRA # as
the CONTAINER ID, seals the outer bags, and places on ice in a sample cooler.
4. The clean protocol is complete at this step and field parameters can now be taken from the remaining water
in the BRIDGE BOTTLE. Suggested field parameters include pH, Conductivity, Temperature, and Dissolved
Oxygen. Additional laboratory samples for the solid series and total organic carbon should be prepared;
group code SOLIDS, catalog number 190-243, and group code NME10. catalog number 190-25.
5. When finished with the BRIDGE BOTTLE seal the apparatus in a plastic bag and return to the cooler for
recycling back to DCLS. DCLS will reclean, certify, and repackage for reuse.
6. Rinse the rope and weights with ambient water to remove any visible dirt, place inside a plastic bag, and
store in the storage container. Rope may be reused several times if rinsed frequently.
Effluent Sample Collection Protocol
Equipment Setup
1 Locate an area near the final effluent sampling location where sample processing will occur. This should be
an area free of falling debris and swirling dust, flat, smooth, and protected from the wind. The tailgate of a
vehicle or the back of a Suburban are good locations.
2. Locate the equipment box and coolers containing the sample containers and kits in the area where sample
processing will occur.
3. Cover the work area with a large pick of plastic film. Set out the pump and connect the battery. Switch pump
222
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
on for a quick burst to check that it is working. Dial the pump speed to 5.
4. Remove a tubing kit and two sample containers from the cooler and place on the plastic near the pump.
5. Remove a pack of sample gloves from the storage container and place on the plastic. Refer to, for the
schematic of the field sampling equipment used to collect treatment plant grab samples.
6. Remove the plastic sample caddie from the storage box and place it on the sample processing area near the
pump.
7. Locate the sample wand used for positioning the teflon sample tubing in the effluent.
Effluent Dissolved Grab Blank
1 . Refer to Figure 2 Sample Collection Scheme for the schematic of the field sampling equipment used to
process blanks and samples.
2. Determine which tech will be clean hands and which will be dirty hands.
3. Dirty hands and clean hands don one or two pairs of vinyl gloves. Dirty hands opens the sample bottles outer
plastic bag, clean hands opens the inner plastic bag.
4. Dirty hands opens the grab kit's outer plastic bag, clean hands opens the inner plastic bag and removes the
tubing assembly.
5. Clean hands disconnects one side of the sample loop on a sample container. Clean hands connects the
teflon end of the tubing kit opposite the filter to the opened sample container. Remember the sample
container is full of clean water from the lab.
6. Dirty hands connects the peristaltic tubing to the field pump on the end of the pump tubing closest to the
connection with the teflon tubing. This allows for slack on the filter end of the pump tubing. Clean hands
inverts the sample container, and dirty hands switches on the pump.
7. Process the entire contents, 1000 mis, of the sample container through the tubing and filter apparatus at a
flow rate of 500 mls/min. (pump setting of 5). At the beginning of the sample processing orient the filter
cartridge with the flow arrow pointing up. This will insure proper wetting of the filter. After the last noticeable
air bubbles have been expelled from the filter it can be oriented in any direction during the rinsing. When the
sample container is nearing empty orient the capsule filter with the arrow down. Once the sample container
is empty continue processing until the free water in the tubing and filter has been expelled.
8. Dirty hands turns off the pump. This step is a rinse of the filter which cleans and conditions the media. The
rinse can be pumped directly to waste.
9. Use the same flow and orientation scheme throughout all sample and blank processing.
10. Clean hands removes the teflon tubing from the empty bottle and connects the capsule filter to the empty
container via the sample loop tubing. Remove one side of the loop fitting from a second sample container
and connect the teflon end of the tubing to the container. Remember that this second bottle is full of clean
water from the lab. Clean hands inverts the container, and dirty hands switches on the pump. Process the
entire contents, approximately 1000 mis, of the sample container through the tubing and filter apparatus into
the first sample container.
11. Dirty hands switches off the pump when the filter has been emptied as evidenced by air bubbles in the
system. Clean hands disconnects the outlet tubing from the sample container and immediately reconnects
the loop tubing on the sample container, seals the inner bag and places inside the outer bag. Dirty hands
identifies this as a blank, seals the outer bags, and places on ice in a sample cooler.
12. This is a comprehensive blank because it is collected in the same equipment as the sample and it is
processed like the sample through all steps of the protocol. This is the most important check of
contamination in the protocol.
Effluent Dissolved Grab
1 . In a field notebook dirty hands records the sample container identification number, date, time. Clean hands
secures the sample container in the sample caddie.
2. Clean hands connects the filter to the empty sample container. Clean hands presents a section of the Teflon
tubing just past the inlet to dirty hands who then attaches the tubing to the sample wand.
3. The entire assembly: sample caddie containing the empty sample container, sample tubing, pump/battery,
and sample wand are transported to the effluent sampling location.
4. Dirty hands places the sample wand sample collection zone taking precaution not to touch the tip of the
sampling tube on any items. Once the sample tube is located in the effluent take precaution not to let the tip
contact anything but water.
5. Dirty hands switches on the pump and clean hands collects a full sample container while following the filter
orientation scheme. It is acceptable to fill the sample container to overflowing, however avoid filtering more
than 1000 mis through the filter.
223
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
6. Once the sample container is full dirty hands switches off the pump. Clean hands disconnects the outlet
tubing from the sample container and immediately reconnects the loop tubing on the sample container, seals
the inner bag and places inside the outer bag. Dirty hands identifies the sample, seals the outer bags, and
places on ice in a sample cooler.
7. The clean protocol is complete at this step and field parameters can now be taken directly from the effluent.
Suggested field parameters include pH, Conductivity, Temperature, and Dissolved Oxygen. Additional
laboratory samples for the solid series and total organic carbon should be prepared; group code SOLIDS,
catalog number 190-243, and group code NME10, catalog number 190-25.
CONCLUSIONS
Using these clean procedures to collect more than two hundred samples the data indicate that collection of
contaminant free samples at trace concentrations is possible. State regulatory agencies should be encouraged to
move forward in promulgating guidance for trace element wastewater characterizations and ambient water
quality monitoring. The careful application of the techniques developed by the U.S. Geological Survey and the
U.S. Environmental Protection Agency can produce high quality data, satisfactory for regulatory decision making.
By adopting the system of field procedures proposed here, investigations of water quality can be performed
simply and cost effectively.
ACKNOWLEDGMENTS
This work was prepared under contract with the U.S. Environmental Protection Agency. The author would like to
thank the following people for their assistance in conducting this work: Jim Anderson, Lisa McMillan, and Ron
Gregory.
Table 1. Target Analytes, Analytical Test Methods, and Detection Limits
Parameter CAS Number
,;,.-,,.., ,,-'.-..:' . ,; ^;:,, ,
Aluminium
Antimony
Arsenic
Cadmium
Calcium
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Selenium
Silver
Zinc
7429-90-6
7440-36-0
7440-38-2
7440-43-9
7440-70-2
7440-47-3
7440-50-8
7439-89-6
7439-92-1
7439-95-4
7439-96-5
7439-97-6
7440-02-0
7782-49-2
7440-22-4
7440-66-6
*!<*- Method Detection Limits, ug/L t ! t; ;
ICPMS USN '* ICPMS MSN ICPM$ : ICP AS CVAA '
TR .' ,, .v.:. :..,!* USN {& ', . ;;*; ,,,;i!'
0.05
0.07
0.06
0.02
0.02
0.17
0.02
0.04
0.19
0.26
0.04
0.03
0.03
0.04
0.04
0.06
0.03
0.03
0.02
0.06
0.03
0.03
0.37
0.33
0.24
1.72
0.87
0.28
1.32
0.12
0.39
0.77
0.15
2.18
5.37
2.37
0.08
2.27
4.98
2.3
0
0.58
1.71
1395
0.07
ICPMS inductively coupled plasma mass spectrometry sample introduction by ultrasonic nebulization
USN . , - , " i ^ ]' ' " - , ;:; -
ICPMS inductively coupled plasma mass spectrometry sample introduction by ultrasonic nebulization
USNTR total recoverable
ICPMS inductively coupled plasma rrtas^ spectrometry i
ICP A6S inductively coupled plasma atorriic emission spectrometey sample introduction #y ultrasonic ! ,
USN nebulization
CVAA cold vapor atomic absorption spectrorrietry
REFERENCES
1. Quality Criteria of Water 1986. United States Environmental Protection Agency, EPA 440/5-86-001, May 1,
1986.
2. Standard Operating Procedure for the Collection of Freshwaters and Wastewater for the Determination of
224
-------
WTQA '98 - 14th Annual Waste Testing & Qualify Assurance Symposium
Trace Elements, Virginia Department of Environmental, 15 May 1997.
3. Federal Register, 60848, Vol. 57, No. 246, Tuesday, December 22, 1992.
4. U.S. Geological Survey Protocol for the Collection and Processing of Surface-Water Samples for the
Subsequent Determination of Inorganic Constituents in Filtered Water, Open-File Report 94-539.
5. Method 1631: Mercury in Water by Oxidation Purge and Trap, Cold Vapor Atomic Fluorescence Spectrometry,
EPA 821-R-95-027, April 1995 DRAFT.
6. Method 1632: Determination of Inorganic Arsenic in Water by Hydride Generation Flame Atomic Absorption.
EPA 821-R-95-028, April 1995 DRAFT.
7. Method 1636: Determination of Hexavalent Chromium by Ion Chromatography, EPA 821-R-95-029, April
1995.
8. Method 1637: Determination of Trace Elements in Ambient Waters by Chelation Preconcentration with
Graphite Furnace Atomic Absorption, EPA 821-R-95-030, April 1995.
9. Method 1638: Determination of Trace Elements in Ambient Waters by Inductively Coupled Plasma-Mass
Spectrometry, EPA 821 -R-95-031, April 1995.
10. Method 1639: Determination of Trace Elements in Ambient Waters by Stabilized Temperature Graphite
Furnace Atomic Absorption, EPA 821-R-95-032, April 1995.
11. Method1640: Determination of Trace Elements in Ambient Waters by On-line Chelation Preconcentration
and Inductively Coupled Plasma-Mass Spectrometry, EPA 821-R-95-033, April 1995.
12. Method 1669: Sampling Ambient Water for Trace Metals at EPA Water Quality Criteria Levels, U.S.
Environmental Protection Agency, 821-R-5-034, April 1995.
13. Benolt, Gaboury, Clean Technique Measurement of Pb, Ag, and Cd in Freshwater: A Redefinition of Metal
Pollution, Environ. Sci. Technol., Vol. 28, No. 11, 1994.
14. Horowitz, Arthur J., et.al., Problems Associated with Using Filtration to Define Dissolved Trace Element
Concentrations in Natural Water Samples, Environ. Sci. Technol., Volume 30, Number 3. Pages 954-963,
1996.
15. Horowitz, A.J. et. al., The Effect of Membrane Filtration Artifacts on Dissolved Trace Element
Concentrations, Wat. Res. Vol. 26, No. 6, pp. 753-763, 1992.
6" X 0.19" ID
size 25 C-Rex
Figure 1. Loop Sample Bottle
1 liter plastic
wide mouth
high-density
polyethylene
leakproof
1.5" X Q.I 9" ID
size 25 C-Flex
1.5"X1/4"ODFEP
polypropylene screw cap
with 2X19/64" holes
Peristaltic Pump
capsule filter
L
-r
4'CFLEX25tubing
2" white PVC
f
1Ltwo port
polyethylene bottle
Figure 2. Sample Collection
Scheme
225
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
226
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Author
Alexander IV, J.N.
Armstrong, D.W.
Autenrieth, R.L.
Bakaltcheva, I.B.
Barat, R.B.
Belanger, J.M.R.
Birri, J.
Blaterwick
Bonde, S.E.
Boring, C.B.
Boswell, C.E.
Boylan, H.M.
Briggs, R.G.
Broske,A.
Brown, R.D.
Bruce, M.
Bruce, M.L.
Brusseau
Bynum, R.V.
Caillouet, D.J.
Calcavecchio, R.
Carlson, R.E.
Charles, P.T.
Chin, C.
Clark, M.S.
Connell, T.R.
Cortese, N.A.
Dahlgran, J.
Dasgupta, P.K.
Donnelly, K.C.
Dorn, H.C.
Drake, E.N.
Dupes, L.
Dupes, L.J.
Durst, H.D.
Eatough, D.J.
Edgerley, D.A.
Eilcone, C.
Ellickson, M.L.
Fahnenstiel, G.
Fan, T.S.
Fan, T.S.
Federici, J.F.
Felix, A.
Fidler, B.
Flores, R.A.
Fordham, O.
AUTHOR INDEX
aper
o.
1
3
9
0
5
9
5
4
1
2
0
1
2
1
0
2
6
5
2
2
7
3
3
1
9
9
5
8
6
7
8
7
3
4
1
3
1
5
5
4
9
Page
No.
94
131
93
26
132
11
188
137
28
94
205
53
151
4
152
17
55
148
152
173
29
27
26
11
205
88
102
161
94
93
147
29
43
173
88
146
181
209
4
65
27
28
124
29
215
165
53
Author
Forman, R.
Gauger, P.R.
George, L.A.
Gere, D.R.
Grabanski, C.B.
Grosser, Z.
Haire, B.
Hard, T.M.
Hanson, R.
Harrison, R.O.
Hawthorne, S.B.
He, L.-Y.
Heglund, D.L.
Hering, J.G.
Hoberecht, H.
Homstead, J.
Hunter, P.M.
Huo, D.
Jayne, J.T.
Johnson, H.C.
Kane, J.S.
Kaphart, T.S.
Keeler
Kennicutt II, M.C.
Kenny, T.D.
Kinghorn, R.
Kingston, H.M.
Kingston, H.M.
Kolb, C.E.
Krappe, M.
Krishnan, B.
Kroll, D.
Kustarbeck, A.W.
Lagadec, A.J.M.
LeBlanc, G.
LeBlanc, G.
Lee, E.
Lee, G.H.
LeMoine, E.A.
Lesnik, B.
Lesnik, B.
Lesnik, B.
Levy, R.A.
Li, W.
Lippmann, M.
Lu, Y.
Lubbad, S.H.
Paper
No.
18
Page
No.
43
9
38
5
30
66
65
38
35
12
30
29
32
37
16
58
25
23
44
55
59
31
39
33
42
5
20
23
44
30
36
22
9
30
64
65
34
54
16
1
2
7
41
47
47
23
32
26
118
4
93
215
215
118
117
27
93
93
101
117
35
187
73
64
131
165
188
94
119
102
125
4
53
64
131
93
117
60
26
93
214
215
110
165
35
3
3
11
124
137
137
64
101
227
-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Author
MacDonald, B.
Macko, S.A.
Marquis, B.T.
Masila, M.
Mays, J.
McDonald, T.J.
Merschman, S.A.
Miller, D.J.
Miller, LM.
Miller, M.
Min, J.H.
Moniot, C.L.
Morrissey, K.M.
Nagourney, S.J.
O'Brien, R.J.
Obeidi, F.
Pare, J.R.J.
Patterson, C.
Patterson, E.
Paustian, M.
Peist, K.
Poziomek, E.J.
Quin, Y
Reddy, S.M.
Rediske, R.
Reeves, W.R.
Richards, K.L.
Richards, K.L.
Richter, R.C.
Risden, R.M.
Robbat Jr, A.
Sakik, O.A.
Schelske, C.
Shriver-Lake, L.C.
Singhvi, R.
Skoczenski, B.A.
Skoczenski, B.A.
Smith, K.A.
Smith, R-K.
Solsky, J.F.
Somers, S.R.
Sparks, L.D.
Stahl, D.C.
Stelz, W.
Stewart II, R.E.
Stroud, R.B.
Stutz, M.H.
Paper
No.
59
33
42
34
27
29
32
30
21
64
37
30
27
59
38
48
7
9
48
66
59
58
33
63
24
29
8
21
20
8
40
34
24
10
7
11
13
44
17
3
26
52
32
28
67
52
6
Page
No.
188
102
125
110
88
93
101
93
55
214
117
93
88
188
118
146
11
26
146
215
188
187
102
209
65
93
17
55
53
17
119
110
65
26
11
27
28
131
40
3
77
152
101
93
219
152
5
Author
Thies, C.
Tillotta, D.C.
Tuchman, M.
Tummillo Jr, N.J.
Turle, R.
Turpin, R.
Valente, H.
van Bergen, S.
VanGaalen, G.C.
Vetelino, J.F
Vitale, R.J.
Vitale, R.J.
Watkins, C.S.
Weisberg, S.
Wilding, S.
Woodard, R.M.
Worsnop, D.R.
Xu, H.
Xuan, T.
Yanik, P.J.
Zemansky, G.M.
Zhang, X.
Ziong, J.Q.
Paper
No.
53
32
24
59
7
7
51
10
15
42
18
58
55
4
60
62
44
34
63
33
61
44
47
Page
No.
161
101
65
188
11
11
151
26
29
125
43
173
165
4
191
205
131
110
209
102
192
131
137
228
------- |