.





      ANNUAL
    WASTE

    . QUALITY
ASSURANCE
SYMPOSIUM

                 PROCEEDINGS
           JULY 12-16, 1993
         HYATT REGENCY CRYSTAL CITY
            ARLINGTON, VA

-------
THE SYMPOSIUM IS MANAGED BY THE AMERICAN CHEMICAL SOCIETY
                         printed on recycled paper

-------
                                    TABLE  OF CONTENTS
Paper                                                                                                  Page
Number                                                                                                Number


QUALITY ASSURANCE

1       Current Trends and Uses of Performance Evaluation Materials in the Superfund Contract Laboratory Program.
        J. Barron                                                                                              1

2       Improved Laboratory Performance Through Matrix-Analyte Specific Proficiency Testing.  M. Caruso, K.W.
        Jackson                                                                                               11

3       Commerical Laboratory Position on National Laboratory Accreditation. J. Farrell, ffl, J.R. Hall, B.R. Hill,
        M. Moore, J.L. Parr, J. Ploscyca, P. Sleevi, D. Speis, S. Tucker, C. Vinson                                        18

4       Cost Effective Project Planning:  An Alternative Approach to Initial RCRA Facility Investigations. D. Loring,
        S. Chapnick                                                                                           29

5       Matrix Spikes and Surrogates, Do We Need Both? G. Robertson, M. Stapanian, F. Garner                          37

6       Status of Performance Based Method. I. Deloatch                                                            38

7       Quality Assurance Guidance for Environmental Sampling. N. Adolfo, A. Rosecrance                               39

8       Automated Real-Time Project Level  Laboratory Quality Assurance. A. Gladwell, C. Bruce Godfrey, K. O'Brien,
        M. Schmick                                                                                           47

9       Field Auditing Procedures for EPA Contractor Personnel. G. Janice, O.B. Douglass                                56

10      The Field Chemical Data Acquisition Plan: A Data Quality Objectives Approach for the Use of Real Time Field
        Analytical Data in Environmental Programs. P. Greenlaw, R.J. Bath, R.D. Spear, D. Lillian                         65

11      A New Cost-Effective Strategy for Environmental Reporting Software. E.A. Haley, L.A. Richardson                  66

12      Local Area Network Data Validation Software System. A. Davis, J. Vernon                                       71

13      Quality Assurance and Quality Control Measures for the Determination of Hexavalent Chromium from
        Stationary Sources. T.P. Dux, K. Hall, G. Jungclaus, A. Williams                                               72

14      Development of Technology Performance Specifications for Volatile Organic Compounds. M.D. Erickson,
        S.C. Carpenter, P.V. Doskey, A.D. Pflug, P.C. Lindahl, C. Purdy, W.E. Schutte                                    83

15      The QAPjP Qaugmire.  M.A. Kuehl                                                                        89

16      Evaluation of Methods for Determinations of Instrument Detection Limits. B. Organ, D.E. Dobb, TJ. Meszaros,
        J.C. Converse, R. McCallister                                                                             98

17      Technical Information and Management  for the Pit 9 Project Idaho National Engineering Laboratory.
        J. Owens, D. Forsberg,  D. Macdonald                                                                      99

18      Quality Assurance Plans for Basic Research and New Concept Studies. E J. Poziomek, A. Cross-Smiecinski          100

19      Data Validation Guidance for Conventional Wet Chemistry Analyses. A. Rosecrance, L. Kibler                     115

-------
20      Developing a Performance-Based Approach to Environmental Analytical Testing. B. Thomas, G.M. Mong,
        M. McCulloch, S.K. Fadeff, D.S. Sklarew, R.G. Riley, S.C. Goheen, J. Poppiti                                     13°

21      Quality Assurance and Data Collection: Electronic Data Transfer.  L. Tomczak, W. Lohner, E. Ray, J. Salesky,
        H. Spitz J.C. Gore                                                                                         136

22      An Integrated Approach to Maintaining Quality Assurance During the RI/FS Process. A. Tracy, W. O'Brien,
        W. Mills                                                                                                 146

23      Effluent Emissions Monitoring at the DOE Hanford Site.  L.W. Vance                                            155

24      Quality Assurance Audits of Laboratories. R.T. Winward                                                      166


SAMPLING AND  FIELD

26      Practical Suggestions to Improve the Quality of Field Work A Sampler's Perspective. T. Diebold                     179

27      Evaluation of the Hydro Punch to Access Groundwater Contamination by Volatile Organics. C. Van Sciver,
        E.Wallace                                                                                               180

28      An Integrated Application of Field Screening to Environmental Site Investigations: A Case Study. W. Mills,
        A. Tracy, T. Cline-Thomas                                                                                 181

29      Divergence of Field and Lab Results in the Ponca City Investigation. S. Dutta                                     193

30      Field Screening for Hazardous Materials in Soil and Groundwater at a Maintenance Building.  A.M. Larson,
        H.R. Kleiser, P.M. Curl                                                                                    206

31      Field and Laboratory Methods in Ecological Risk Assessments for Wetland and Terrestrial Habitats. G. Linder,
        M. Bollman, C. Gillet, R. King, J. Nwosu, S. Ott,  D. Wilborn, G. Henderson, J. DalSoglio, T. Pfleeger                 210

32      Increasing the Sensitivity of Field Headspace Analysis for Volatile Organic Compounds.  C. Van Sciver,
        R. Fowler                                                                                                222


ORGANICS

33      Regulatory Aspects of RCRA Analyses. B. Lesnik                                                             236

34      Stepwise Development Process for Regulatory Analytical Methods. P. Marsden, S.F. Tsang, M. Roby                240

35      Recycling Program for Laboratory Organic Solvents - Methylene Chloride and Freon-113. C.A. Valkenburg,
        W.T. Brown                                                                                              245

36      The Technology  and Performance of Several New Immunoassay Methods and a New Supporting Instrumentation
        System. S.B. Friedman, J.P.  Mapes, R.L. Allen,  T.N. Stewart, W.B. Studebaker, P.P. McDonald, R.E. Almond,
        T.A. Withers, S.P. Arrowood, D.P. Johnson                                                                   261

37      Screening of TCLP Extracts of Soil and Wastewater for 2,4-D by Immunoassay. M.C Hayes, S.W. Jourdan,
        T.S. Lawruk, D.P. Herzog                                                                                  262

38      A Rapid, On-Site Immunoassay for Detecting Petroleum Products in Ground Water. J.P. Mapes, P.P. McDonald,
        R.E. Almond, S.B. Friedman,                                                                                276

39      Appendix IX Extractions by Accelerated One Step Extractor/Concentrator.  M. Bruce, J. Carl, B. Killough,
        T. Zine, D. Burkitt                                                                                         283

40      Microscale Solvent Extraction Methods for Tarry Soils from Former Manufactured Gas Plant  (MGP) Sites.
        J. Schneider, I. Murarka, D. Mauro, M. Young, B. Taylor                                                      29R

-------
41      Development of an Environmental Method for the Analysis of Volatile Organic Compounds in Soils by Static
        Headspace GC/MS. T.J. Riga, G. O'Neil                                                                    313

42      In-situ Derivatization of Soil and Sediment Samples in an Supercritical Fluid Extraction Cell: The Extraction of
        Phenols and Chlorinated Phenols. D.R. Gere, L.G. Randall, C. Knipe, W. Pipkin, H. B. Lee                         326

43      Supercritical Fluid Extraction (SFE) of PCBs. M.Bruce                                                       335

44      Determination of Organic Compounds in Ground Water by Liquid-Solid Extraction Followed by Supercritical
        Fluid Elution and Capillary Column Gas Chromatography/Ion-Trap Mass Spectrometry.  J.S. Ho, P.H. Tang,
        J.W. Eichelberger, W.L. Budde                                                                             337

45      Comparison of Solid Phase Extraction with Salting-Out Solvent Extraction for Preconcentration of Nitroaromatic
         and Nitramine Explosives from Water. T.F. Jenkins, P.H. Miyares, K.F. Myers, E.F. McCormick, A.B. Strong        338

46      Dinoseb Analysis in the Field and the Laboratory. D. Anderson, S.F. Tsang, T. Jackson, P. Marsden                  347

47      Quantitation  of Polychlorinated Biphenyls Using 19 Specific Congeners. P. Marsden, S.F. Tsang,
        M. Kennedy, B. Lesnik                                                                                    360

48      Direct Determination of TCLP Phenols and Herbicides by HPLC. R.L. Schenley, W.H. Griest, J.E. Caton             372

49      Approaching the Sensitivity of an Electron Capture Detector (ECD) for the Analysis of Pesticides by Using
        GC/MS. L.C. Doherty, N. Low                                                                            373

50      Determination of Polynuclear Aromatic Hydrocarbons in Soil at 1 ug/kg Using GC/MS. B.N. Colby,
        C.S. Parsons, J.S. Smith                                                                                   379

51      Application of Lee Retention Indices to the Confirmation of Tentatively Identified Compounds from GC/MS
        Analysis of Environmental Samples. P.H. Chen, W.S. Keeran, W.A. Van Ausdale, D.R. Schindler, D.F. Roberts       385

52      A Quick Performance-Based HPLC Method for the Analysis of Polynuclear Aromatic Hydrocarbons (PAHs).
        M.W. Dong, J.M. DiBussolo                                                                               396

53      The Extraction and Analysis of Polychlorinated Biphenyls (PCBs) by SFE and GC/MS. Improvement of Net
        Detection Levels. D.R. Gere, L.G. Randall, C.R. Knipe, W. Pipkin, L.C. Doherty                                  397

54      Environmental Analyses-PAHs in Solid Waste:  Bridging the Automation Gap Between SFE and HPLC.
        D.R. Gere, L.G. Randall C.R. Knipe, W. Pipkin                                                              406

55      Analysis of PCB's in Soil, Sediments, and Other Matrices by Enzyme Immunoassay. R.O. Harrison,
        R.E. Carlson, A.J. Weiss                                                                                   419

56      Explosives Analysis of Atypical Matrices and Technical Enhancements to EPA Method 8330. D. Hooton,
        M. Christopherson, T.P. Dux, G. Jungclaus, H. Randa, R. Sauter                                                 430

57      Liquid Chromatography/Mass Spectrometry of High-Molecular-Weight Polycyclic Aromatic Hydrocarbons
        Using the Particle Beam Interface. M. Roby, L.D. Betowski, C. Pace                                            438

58      Quantitation of Pentachlorophenol (PCP) by a Rapid Magnetic Particle-Based Solid-Phase ELISA in Water and
        Soil. F.M. Rubio, T.S. Lawruk, C.S. Hottenstein, S.W. Jourdan, D. P. Herzog, Y. Zha                              439

59      Chromatographic Optimization  for the Analysis of an Expanded List of Volatile Organic Pollutants.
        D.M. Shelow, M. J. Feeney                                                                                452

60      Application of SW-846 Methods to the Identification of Unusual Brominated Compounds in High-Concentration
        Process Streams. D.V. Smith, L.P. Pollack, F.T. Varcoe                                                        453

61      Results of Laboratory Tests of the Accelerated One-Step Liquid-Liquid Extractor. R.K. Smith, R.G. Owens, Jr.,
         D.S. Geier, T.A. Shaw, D. Phinney, J. Carl                                                                   457

-------
62      A Matched Dual Capillary Column System for Pesticides. A. Spilkin, M.R. Hastings                                466

63      Performance Data for the Analysis of Phenols, Nitroaromatics, Cyclic Ketones, Haloethers, and Chlorinated
        Hydrocarbons. S.F. Tsang, N. Chau, P. Marsden, B. Lesnik                                                      481

64      Rapid Confirmation of Nitroaromatics and Nitramines Using UV Diode Array Spectral Comparison.
        B.A. Weichert, R.D. Baker, C.R. Campbell                                                                    482

65      The Effect of GPC Cleanup of Semivolatile Extracts on GC/MS  Analysis.  T. Willig, J.S. Kauffman                   483

66      Evaluation of the Ensys PAH-RISc Test Kit. R.P. Swift, J.R. Leavell, C.W. Brandenburg                            484

67      Results of Analytical Field Trials for PCBs Using an Immunoassay Technique. A. Weiss, A. Parsons                  500

68      Important Factors in Enhancing Supercritical Fluid Extraction Efficiencies for Environmental Applications.
        J.M. Levy, L.A. Dolata, M. Ravey                                                                           501


INORGANICS

69      Evaluation of a Rapid Steam-Distillation Procedure for the Extraction of Cyanide from Liquid and Solid Wastes.
        E.M. Heithmar, R. Herman, M.R. Straka                                                                      503

70      A Method Evaluation Study for the Analysis for Hexavalent Chromium in Solid Samples Using a Modified
        Alkaline Digestion Procedure and Colorimetric Determination.  RJ. Vitale                                        505

71      The Analysis of Reducing Soils for Hexavalent Chromium. D.P. Miller                                           506

72      Factors Affecting Sample Throughput for an ICP-OES System with a Segmented-Array CCD Detector.
        Z.A. Grosser, K.J. Fredeen, C. Anderau, D.A. Yates, K.M Barnes, T.J. Gluodenis, Jr.                                516

73      Design, Performance, and Environmental Applications of an ICP Array Detector Spectrometer.  D.D. Nygaard,
        F. Bulman, M. Almeida                                                                                     519

74      Variability in TCLP Metals Results from Stabilized Waste. G. Merewether, N. Shah                                520

75      Analgamation CVAA to Improve the Detection Limit of Mercury in Environmental Samples. S. Sauerhoff,
        Z.A. Grosser, S. Mclntosh                                                                                  521

76      The Development of EPA Total Mercury Method Using Cold Vapor, Fluorescence Mercury Detection Systems.
        The Mercury Method Will be Used for the Analyses of Mercury  in Water, Wastewater, Sea Water and Related
        Matrices. B.B. Potter, S.E. Long, J.A. Doster, R. Davis                                                         524

77      Determination of Tetraethyl Lead in Groundwater. B.N. Colby,  D.J. Bencivengo, C.L. Helms                         539

78      Use of a Telephone Service and Database to Provide Guidance and Increase Public Involvement in OSW Methods.
        D. Anderson,  R. Carlston, S. Hartwell, G. Hansen, K. Kirkland                                                   543

79      Extraction of Metal Ions from Solid Matrices by Complexation SFE.  W.F. Beckert, Y. Liu, V. Lopez-Avila,
        M. Alcaraz                                                                                                544

80      Microwave Sample Preparation for Mercury Analysis via Cold Vapor Atomic Absorption Spectrometry.
        W.G. Engelhart, S.E. Littau, E.T. Hasty                                                                      545

81      Study of EPA Method 300.0: Application to Hazardous Waste Analysis.  H. Mehra, D. Gjerde                       545

82      The Effects of Size Reduction Techniques on TCLP Analysis of Solidified Mixed Waste.  R.D. Thiel                  554

-------
AIR AND GROUNDWATER

83      Canister Analysis by Using Gas Chromatography/High Resolution Mass Spectrometer. J.P. Hsu, G. Miller,
        J.C Pan                                                                                              555

84      An Effective Moisture Control Solution for the GC/MS Determination of VOCs in Canister Samples Collected
        at Superfund Sites.  M.G. Winslow, D.F. Roberts, M.E. Keller                                                 556

85      Selection Criteria for Groundwater Monitoring Well Construction Materials.  J.R. Brown, L. Parker,
        A.E. Johnson                                                                                         566

86      Auto GC Design and Operation for Remote Unattended VOC Determinations. J.F. Ryan, I. Seeley                  567

-------
QUALITY ASSURANCE

-------
•J        CURRENT TRENDS AND USES OF PERFORMANCE EVALUATION MATERIALS IN THE
                           SUPERFUND CONTRACT LABORATORY PROGRAM

                           James W.  Barren, Quality Assurance Coordinator, Analytical
                              Operations Branch Hazardous Site Evaluation Division,
                               Environmental Protection Agency,401 M. St., S.W.,
                                            Washington, DC  20460
        Good day, I'm Jim Barton with the Analytical Operations Branch (AOB) of the Hazardous Site Evaluation
        Division. Our group manages some of the analytical services utilized by Superfund, including the contract
        laboratory program commonly known as the CLP.  This includes a comprehensive Quality Assurance
        program. One facet of this program is ensuring an adequate variety of Performance Evaluation Materials
        (PEMs), also referred to as Performance Evaluation Samples (PES). Historically, PES have been used by
        government agencies for several purposes.

                1. Multi-laboratory Validation Studies for validation
                  of analytical methodology.

                2. In the form of Blind PES for Laboratory Certification.

                3. In an on-going Quality Assurance (QA) Program.

        The CLP uses PES for all three purposes.  PES are prepared to support inter-laboratory studies to validate
        CLP Statements of Work (SOWs).  The Laboratories are required to participate in pre and post-award
        performance sample studies to maintain contract status, the so-called "de facto certification program."  We
        consider number 3, the PES we provide for routine assessment of data, the most important part of our
        program. I'd like to discuss some of the trends we see in the use of these materials in environmental work,
        and some of the approaches we are taking in our program. We feel we have addressed on a smaller scale
        many of the questions EPA will to consider as it looks toward a national laboratory certification program.

        The Analytical Operations Branch has established and operates a Superfund Performance Evaluation Sample
        Repository through it's Quality Assurance Technical Support Laboratory in Las Vegas. The Superfund
        Performance Evaluation Sample Repository was established in response to general and specific requests
        from Regions for an independent and relevant mechanism to monitor laboratory performance in the analysis
        of environmental samples from known and suspected hazardous waste  sites.  A wide range of performance
        evaluation samples is available  at no cost to the Regions to support their needs for independent and external
        monitoring of laboratory performance in the analysis of environmental samples in the Superfund Program.

        A key part of AOB's  program is  that it does not normally distribute standards or  Laboratory Control
        Samples, it distributes PES materials to monitor analytical work where  the government has a strong interest
        in determining the quality of the data it is paying for. Some exceptions are LCS'es such as the Interference
        Check Samples run to determine interferences in trace level metal analysis, and the multi-parameter LCS'es
        used in some Statements of Work. (24, 26)  Available performance evaluation sample types are water and
        soils matrices.  Performance evaluation samples for air analysis have been  under development, but work
        has stopped due to lack of regional interest.  Analytes incorporated into these samples include volatile and
        semi-volatile organic compounds,  pesticides and PCBs, metals, anions, cyanide, and chlorinated dioxins
        and  furans.   Stability studies are run are archived on PES to ensure  values are still within the statistical
        windows established for the particular PES. Regions select the type of performance evaluation samples

-------
most suited to their needs from a widely distributed catalogue.(20)

Through the Superfund Performance Evaluation Sample Repository, the Analytical Operations Branch also
provides related support services to the Regions.  A PC-based tracking system PEACTOOLS has been
developed, and is distributed to the Regions for use in assessing the analytical results for performance
evaluation samples  reported  by laboratories.   Using  this  system,  Regional data reviewers evaluate
laboratory performance on a particular set of environmental samples as soon as the data  package is
received.  These evaluations also provide important documentation of data quality, when the PE sample
is related to the site. Through the Superfund Performance Evaluation Sample Repository, Regions can also
obtain assistance in  establishing and operating their own performance sample repository which can serve
regional needs in monitoring analytical work done  for the Superfund Program.

AOB gets support and oversight from ORD at Las Vegas for it's program. In so far as possible we attempt
to obtain real world materials from sites to prepare PEMs, however we also purchase neat materials and
prepared sets from commercial vendors.

We prepare PES for contract compliance, in our quarterly blind program (Qbs) and for routine  monitoring
of data collection activities. In both types of samples, concentrations and recipes of the analytes support
CLP, and other Superfund work.  Our quarterly blind  samples are the typical single blind PES supplied
by most programs for certification.  Some of our Qbs are solids, but several years ago we went from full-
volume liquid samples to ampuled samples, because of problems in the samples, and to cut costs. It costs
about 10 more to prepare full volume samples compared to amputated samples. However AOB was aware
of the pitfalls of amputated PES already being used by other EPA certification programs.   The first
problem is the laboratories may run  the ampuled sample  concentrate undiluted, the second is calling around
to compare answers, and third expending much more time and effort on  these types of samples, then the
laboratory would on a routine sample. Some parties have recently come out of the closet, admitting the
excessive time spent on this type of sample, and  possibly running the ampule straight (  ).  The problem
we  found is  the labs do not  regard  this  as cheating.just an annoyance they must get through to stay
certified.  Calling around is probably more of a problem with government laboratories, where there is a
more collegia! atmosphere, than the commercial sector, where loss of certification by some laboratories,
means more business for the successful labs.

What  can be done about this  problem?  One  answer  is  to give less importance to the  do or  die
"Certification PES."  Another is to  send out different PESs to labs in the same study.  However as noted
we  do not consider answer comparing a problem  in the CLP,  since we have companies with multiple
laboratories, with one laboratory in the group  failing a  QB, while the rest pass.  In any event we do not
have sufficient  Laboratories to use two sets,  since our Qbs are scored on a statistical basis.  A third
approach AOB has  completed is to develop indicator compounds that can be added to the ampulated PES
samples.(l,2) This insures an  aliquot will be removed from the ampule, diluted to the proper volume, and
run according to the Statement of  Work  the laboratory is trying to be certified for.   To be useful as
indicator compounds, for contract compliance, analytes must possess these characteristics:

•  Must be Stable in the test solution matrix

•  Must be detectable by GC/MS when directly injected from ampule.

•  Good  Chromatography when directly injected and when run from normal  extract

•  Must be degraded or poorly recovered by the sample preparation technique used.

•  Must be a clearly detectable difference in the  results by direct injection and normal  preparation.

-------
Most of the people in our branch our experienced environmental chemists.  While a preliminary study to
pick candidate indicator compounds included VOCS, our experience we reasoned that probably no tries to
run the ampule for VOCs,  since this could produce a worse answer, that running it properly.  This leaves
Semi-volatiles, metals and  pesticides. The first group picked were the semi-volatiles.
Chemical
1-Hexanamine
Cyclohexalamine
2-Propenamide
1,3-Benzenediol
Phthalic Anhydride
Nicotine
Di-n-hexylamine
Simazine
Norflurazon
CAS number
111-26-2
108-91-8
79-06-1
108-46-3
85-44-9
54-11-5
143-16-8
122-34-9
27314-13-2
The above are some of the materials we decided to test as indicators, that were selected from a larger
list. (2)  Our reports are available to parties who wish to design their own Indicator program.  Since we
now use continuous liquid -liquid extraction at a low Ph we narrowed the field quickly.  Interestingly
enough the compound  we  may use  is nicotine.   This was accomplished since the CLP  requires
identification of tentatively identified compounds (TICs).  Since the laboratories did not identify all the
materials, it was necessary to run tape audits on the raw data.  EPA, EMSL-LV  helped us with tapes for
GC/MS Systems we did not have equipment for.  There were other logistics involved, such as picking
which internal standard to us with which indicator.  The next group we are going to work on are the
metals. The task will in effect reverse the work we did in developing interference check samples,(3,4))
by adding those interferents at levels that must be diluted to the proper volume, to  be able to identify target
analytes.  At present the Pesticides offer no easy  solution, since they are done by Gas Chromatography.
The course we are following here is to build on work done by the Office of Drinking Water, in developing
method 525, and try to achieve low enough quantitation limits, so all Target Compound list Pesticides could
be run by GC/Mass Spectrometry.  This would enable us to also develop indicator compounds for the
Pesticides.(34) Indicator compounds are not an absolute determinant of fraud, but they are one more tool
to ensure data integrity.

The CLP Quarterly Blind  PE Sample Study data does have other uses besides contract compliance:

•      CLP Laboratory Performance Measurement - This is the primary mission of the PE Study
Program.  All labs are measured against an independently prepared research mixture.

•      Surrogate Optimization- Compounds added to the PES are checked for  correlation to the Target
Compound List, to find new surrogates which can give better analyte recovery information.

•      Feasibility of Target Compound  List additions - The QB  studies have been used to determine
how well  SOW  perform with new  analytes  which have become  candidates  for addition to the Target
Compound List.

•      Method Performance Reports - Because the QB samples are uniform samples which are sent to
all laboratories, inter-laboratory precision and accuracy is determined.

-------
•       Extraction Method Comparison - QB sample results allowed a comparison between Separatory
funnel versus continuous extraction techniques for semi-volatiles, which are now used exclusively in the
CLP.
•       Laboratory Performance Database   The  QB  results are part of a laboratory performance
database to determine overall performance of a Contract Lab. (21)

Failure to analyze the Quarterly Blind sample  correctly, results in suspension of the laboratory until a
remedial PES is successfully analyzed. When statistically derived precision windows are used, as in this
case, a certain percentage of laboratories in the study will fail. This does not mean the data previously
produced  by the laboratory  is  in question,  unless  the   PES  results show a trend of unacceptable
performance.

In the near future the so-called national certification PES  may not be that important to the CLP.  CLP
Routine Analytical  Statements of Work (SOW), now being utilized, or  in the  testing phase (see table 1.)
shows, are more specialized, and may only have one laboratory supplying that service. For these SOWs

Table I.
Performance Evaluation Sample Use.

Document/Reference number.
CLP Low Concentration
Water/22
CLP Quick-Turn-Around
Methods/24
CLP Dioxin Furan/23
CLP High Concentration
Inorganic/28
CLP High Concentration
Organic/29
CLP Low-Medium Inorganic/30
CLP GeoTechnical/26
CLP Water and Soil
Characterizat./25
CLP Quick Turnaround
Dioxin/31
CLP Air Toxics./27
Types and Recommended Frequencies of PES
LCS (Zero-blind)
(1)
by the Sample Delivery
Group (SDG)
Special, multi-param
by the SDG or use >
by the SDG
by the SDG
by the SDG
Special, multi-parameter
PES
by the SDG

inorg., by SDG
Single or Double Blind (2)
by the SDG

by the SDG
by the SDG
by the SDG


by the SDG
by the SDG
By the SDG, All Org. could be
double-blind, on all types of
samplers, i.e. Tedlar bags, PUFs.
Contract
Action
for Poor
Performa
nee
on PES
yes
yes
yes
yes
yes
yes
yes
yes

yes
(1)  All liquid samples are ampuled, all soils homogenized
(2)  Single-blinds, all liquid samples are ampuled, all soils homogenized
    Double-blinds, all liquids are full volume, soils are not homogenized

-------
the project officer can have the laboratory run up to 5% PES.(35) He has the option of inserting them as
double-blind, that is appearing to be a routine sample.  However these  "by the case" PES have their own
set of problems as we are finding out with our Dioxin/Furan SOW.  Our Dioxin/Furan PES were our best
set of  PEMs, all  made from materials and hot spots obtained  from Dioxin sites,  with real world
intereferents such as PCBs present.  We've now gone through our entire set, and need to reblend the PES
to produce new concentration levels. Recent catalogues received from commercial vendors indicate they
are also addressing the problem of PES recognition. (32) Laboratories keep track of PES, and will compare
old results with the latest PES.
This last problem pits two schools of thought on PES, used on a daily basis, or by the case, to assess the
quality of environmental data.  The first group seems to feel that  PES should have extensive multi-
laboratory validation, producing somewhat tight windows for bias and precision, resulting in fairly static
recipes due  to cost.  The second, which AOB subscribes to, feels PES should be economical, readily
available,  and in a wide variety of realistic  materials, representing the matrices and analytes  being
monitored.  Recipes should be changed frequently.  While the PES should evidence sufficient statistical
control to assess the analytes of concern, it isn't necessary to have NIST grade materials.
As noted above, AOB distributes a variety of multi-Media PES For both CLP and non-CLP work.  Initially
PES are produced  to support a proposed SOW.  As is often  stated  PES should be similar to  samples
collected from a waste site, contain interferences relevant to the samples being analyzed, contain analytes,
and concentrations  approximate the range found in actual environmental samples. An initial workplan for
concentrations and recipes of analytes expected to be monitored under the proposed SOW is prepared. For
an interlaboratory  study the PES are then  made or purchased according to the workplan.   These PES
materials will continue to be made an utilized if the SOW goes into effect. The categories of PES AOB
distributes or is working on include:
       CATEGORIES OF PES DISTRIBUTED OR BEING DEVELOPED BY AOB
 CLP Organic Low/Medium, Low Con., High Con, Target Compounds, and TICs
 CLP Inorganic Low/Medium, Hicon, Target compounds
 SITE-CHARACTERIZED PES, Analytes typical of the site
 SITE CATEGORIZED (INDUSTRIALLY) PES, Analytes typical of that industry
 DIOXIN/FURANS
 QUICK TURN-AROUND PES, Really multi-parameter Laboratory Control Samples
 GEOTECHNICAL PES, Really multi-parameter Laboratory Control Samples
 FIELD ANALYTICAL METHODS
We don't plan to  cover  every  type of PES that could be produced, for  example a frequent Special
Analytical Services request are the traditional water quality parameters. We wouldn't stock these since they
have been available for a  long time commercially, with good quality.

PESs are one of the few external QA/QC measures for independent monitoring of
laboratory performance and demonstration of data comparability among laboratories. When the PE is site
related, it documents the quality of analytical results generated in the Superfund program, for that site.
Under the Superfund Accelerated  Cleanup Model  (SACM).  the  Office of Emergency and Remedial
Response is attempting to  reduce duplicative site investigations. However this may reduce the redundancy
of data making an external QC indicator such as PE samples more important.  Another Superfund initiative
has been a change in the delivery of analytical services, with more analytical resources being placed in the
region.  With the increase in the use of non-CLP laboratories, field analytical procedures, and mobile
laboratories,  the  availability of PES  validated by  fixed  laboratory  methods  would allow for  data

-------
comparability. (36)

When we first set up our PES repository at Las Vegas, the philosophy was to have good range of PES in
regard  to  matrices,  analytes present,  and  concentration ranges.   We  set  up  a  scoring program
PEACTOOLS,  which would score PES on a go/no basis.  This was to be a fairly static operation.  This
is no longer viable, and we are now planning to shift recipes and concentrations on a regular basis, with
more emphasis  on site characterized PES.

Double-blind PES like site characterized or site specific materials, are the optimal type of PES, if available.
They constitute a PES, which the laboratory  feels is a routine sample, and  does not put forth that  extra
effort associated with  single-blind PES. Some matrices lend themselves to double-blinds better than  other
matrices.  Solids are the most desirable matrix for double-blinds for Superfund, but they are also the most
difficult to produce.  Attempting to manufacture a solid where there is some statistical reliability of the
concentrations of analytes invariably leads to  homogenization, allowing the sample to be recognized as a
PES.

There does not  appear to be great difficulty in preparing water soluble analytes,  so metal samples and the
so-called Water Quality Parameters do not present a major problem. (7)  More extensive stability studies
are needed for this matrix to determine "shelf life."  Some work has been done in the regions with Volatile
Organic Compounds (VOCs).  AOB is recommending the regions follow the work done by Oak Ridge
National Labs in preparing double-blind VOC's in water.(14)  AOB's QATS laboratory has recently  been
able to  duplicate the results  of work  done by the US Army's CRREL  group.  This involves preparing
VOCs on soil.(15,16 ) Because of the shelf life of these materials we are recommending they be prepared
at the regional level, from materials in Regional PE repositories.  One problem noted in preparing double-
blinds using site-characterized materials for metals, was particle size. This was predicted from other work,
but verified by  analytical results. This has  spun off another project to determine  if sample sizes for metals
should be changed of be more flexible. This work will also have application to field analytical procedures,
where analytical results are also affected by particle size.

Air Samples lend themselves to double-blinds, as they are trapped in canisters, or bags or on solid matrices
such as  Poly-Urethane Foam (PUFs) or Tenax. In practice, real samples are indistinguishable from the
PES. Again, a  major need in this area is stability studies to determine shelf life. The only air parameter
we  get  requests for are  double-blinds prepared  on PUFs for pesticides and Dioxin.  Except in  very
specialized case we feel these types of PES should be prepared in the Regions.

The problem of double-blind solids PES is still difficult at this point. A sample that has been homogenized
by particle size  is immediately recognizable as a PE.  Samples which have been thoroughly mixed, but not
sized, require sample weights far above those specified in most methods to overcome the effects of particle
size on  concentration. The type of double-blind that seems to  hold the most promise at this time are site
characterized  materials, mentioned below, that is materials that are obtained from a site,   spiked and
homogenized, but not sized,  and returned to  the site, to be submitted blind as part of the regular sample
stream.

What has become apparent through workgroup meetings and conferences on PES and Site-characterized
materials, is that environmental personnel are  not concerned if they get a "NIST Grade" materials for this
type of sample.  The higher priority is being able to assess data quality in a timely manner at a reasonable
cost, and realistic time frames.

An  impediment to producing double-blind aqueous semi-volatile organic PES is more administrative than
technical.  As parties have commented, returning to one liter bottles instead of the two and four liter bottles
commonly used, with rinsing of the container to collect  materials that may have plated out on the glass,

-------
would allow greater accuracy in determining this type of PES.  As one commentator noted, this would also
correct an obvious negative bias in real samples.   Double-blinds should approximate the environmental
sample in water color,  and turbidity, or soil color, texture,  particle size range and even % moisture.
Containers should be the same as being used by the sampling  team, and obtained from them in advance.
The references, and anecdotal information indicate double-blind PES are being prepared and utilized on
a more frequent basis than in the past.

We have done a study comparing our Contract Required Quantitation limits for target compounds against
actual results in our analytical results database,  and with values decided on in Records of decision, and
the PRO cancer and non-cancer values  for those analytes.(5)  This has become part of our Pesticide task
mentioned above, to see at what levels we can get good quantitation for important analytes, and has led
us to re-evaluate our whole PES library.  While we will probably not change our CLP CRQLs, the PES
do  support other Superfund work where lower quantitation levels are needed.(6 )  This latter work
assignment actually complements our efforts to convert our CLP  pesticides to GC/MS.(34)

Whether single or double-blind samples, we try to prepare site-oriented materials.  AOB doesn't prepare
site-specific materials per se, only when its a large enough project to a have a surplus as site-characterized
materials.  Before we get cited for not explaining sites, our perception of these materials is:
"SITE PES"
SITE SPECIFIC PES: materials obtained from "hot spots" at a waste site,
additional typical contaminants being found at that site.
Site Characterized PES: may be site specific materials, but the surplus can
at sites with similar matrices.
possibly spiked
be used
with

SITE CATEGORIZED PES: Synthetic PES prepared using information obtained from CARD and
the NPL databases, according to the industrial category of that site.
We think the Site Categorized PES is one of the more interesting projects we have worked on is our effort
to develop industrially categorized site An other branches in our division has categorized waste sites on the
NPL by the industrial category they fall under, i.e electroplating. Searching our databases we are able to
identify the cases associated with those sites, and pull down the samples associated with those cases.  From
our CARD database we can get the sample results for the types of contaminants and their concentration
ranges.  This allows us to prepare synthetic single blind materials  that are typical of waste sites, and in
the absence of real  world materials allows a project manager some means to assess  the quality of a
laboratories data.  There  are 45 categories, but we may find enough overlap  in analytes lists between
related categories, that we would only develop one set of PES to use  for both sites.
Typical Information From the Databases for PES Design
Industry
Manufacturing, Gen.
Electro-plating
Fabricated Metal
Products
Sites on the NPL
683
79
136
Number of Cases
286
32
53
Number of Cases in
CARD
59
10
14

-------
A project we have provided assistance to in Region X, seems to indicate, at least for long term projects,
site characterized materials can be prepared economically, and  in a timely manner.(8) The advantage to
the National Program Office, is that one of the hardest parts of the task is to persuade field personnel to
fill up a few 55 gal barrels of dirt,  from hot spots if possible,  and ship them to the QATS Laboratory.
However when working on a set of site-characterized PES, one can insist on a surplus of materials as quid
pro quo.

A word of clarification is probably  needed to differentiate site-characterized materials from blind spikes
or matrix spikes.  In spiking, one simply squirts some analytes of known concentration onto or into the
sample.  Site-characterized PES are made from the real world matrix, but a liquid recipe is mixed with a
small volume of the soil, dried and blended  with a larger portion of soil.  Homogeneity is verified,
concentration levels, and acceptance limits are validated by multi-laboratory analysis.

We developed approaches to preparing PES for Field Analytical Methods such as field X-ray Fluorescence,
sometime ago, but with the renewed interest in Field Analytical Methods, we are going to implement the
study. (36) A survey of the Regions indicated the most commonly requested PES was  for VOCs in soil.
As noted above we feel this type of PES can now be produced with some assurance of statistical control
of the analytical  results.   Many of the types of specialized  materials,  such as wipe  samples,  are best
prepared locally.
To summarize, we think the trend in Performance Evaluation Materials is toward:

•  Away from "national certification"  PES to "by the case"  PES.

•  More emphasis on PES tailored to specific waste sites, Site Specific, Site Characterized materials.

•  Lower concentration levels, to approach as closely as possible, health effects criteria.

•  "Recipes" changed more  frequently to avoid PE  recognition.

•  Increased use of Double-blind PES.



References:

(1)     Work Assignment 1.7 Indicator Compounds for Evaluating Contract Compliance in Low/Medium
        and Low Concentration Semi-volatile Methods,
        Quality Assurance Technical Support Laboratory, March 15, 1993

(2)     Work Assignment 1.7, "Evaluationof Organic Indicator Compounds for Full Organic Multi-Media
        and Low Concentration PES Programs, Quality Assurance Technical Support Laboratory,
        Oct. 3, 1991

(3)     Technical Guidance  Document 1.6 "Development of an Interference Check Sample for Graphite
        Furnace Atomic Absorption Spectrophotometry," Quality Assurance Technical Support Laboratory,
        Dec. 28,  1990

(4)     Technical Guidance Document 1.6 "Development of an Interference Check Sample for Inductively
        Coupled Plasma Optical Emission Spectrophotometry," Quality Assurance Technical Support

-------
        Laboratory, June 26, 1990

(5)     Review of the Contract Laboratory Organic Target Compound List, prepared for the Analytical
        Operations Branch by VIAR and Co., June, 1992

(6)     Work Assignment 1.12   Re-Evaluation of Concentration Levels and Recipes Used In The
        Performance Evaluation Sample Inventory, Quality Assurance Technical Support Laboratory,
        Oct. 3, 1992

(7)     Carter, Mark, et. al. "How Good Is Your Lab?", Environmental Laboratory,

(8)     Preliminary Study on INEL Site "Site Characterized Materials Study",  On the Effect of Particle
        Size on Inorganic Results.

(9)     Work Assignment 1-9, Quality Assurance Technical Support Laboratory, Organic PE preparation,
        verification, Dec. 11, 1992.

(14)    Maskarinec, Micheal P., et al, "Preparation of Reference Water and Soil Samples for Performance
        Evaluation of Volatile Organic Analysis,"  JAOAC,  vol. 72,  no. 5,  1989, pg. 823.

(15)    Hewitt, Alan D., US Army CRREL, "Vapor Fortification:  A Method  to Prepare Quality
        Assurance Soil Samples for the Analysis of Volatile Organic Compounds", Nat. Symposium
        on Measuring and Interpreting VOCs in Soils, Las Vegas, NV, Jan.  12-14, 1993.

(16)    Hewitt, Allen D., et al, " Comparison of Analytical Methods  for Determination of Volatile
        Organic Compounds in Soil", ES & T, Vol. 26, No. 10, 1992, pg. 1932.
(17)    Proceedings of the Site-specific Soil QA Materials Workshop, Las Vegas, NV, April 13-16, 1992

(18)    Memo from Roy Jones, Region X to Jim Barren, AOB concerning differences  between PES,
        MS/MSD and Site-characterized materials.

(19)    Management Report, "Performance Evaluation Sample Repository and Related Services"
        available from  AOBs' Quality Assurance Technical Support Laboratory.

(20)    Quality Assurance Technical Support Laboratory, "Performance  Evaluation Sample Catalogue",
        1993.

(21)    Memo from Larry Butler, EMSL-LV to Jim Barron,  QAC,  AOB on  uses of the QB Program,
        Feb. 5, 1991.

(22)    OLCO2.0  Statement of Work, Low Concentration Water for  Organic Analysis.

(23)    DFLM01.0 Statement of Work, Polychlorinated Dibenzo-p-Dioxins and Polychlorinated
        Dibenzofurans, Multi-Media, Multi-Concentration.

(24)    Draft Statement of Work, "Quick Turn-Around Analysis"

(25)    WSCIN Statement of Work, Water and Soil Characterization in Multi-Media, Multi-Concentration,
        Nov., 1990

(26)    Draft Geotechnical Statement of Work, Aug. 1992
                                               9

-------
(27)    Statement of Work, Air Toxics

(28)    Statement of Work for High Concentration Inorganic Analysis

(29)    Statement of Work for High Concentration Organic Analysis

(30)    ILM03.0 Statement of Work For Inorganic Analysis, Multi-Media, Multi-Concentration

(31)    Statement of Work "Rapid Turnaround Dioxin Analysis, Multi-Media", 11/92 Rev.

(32)    Environmental Quality Control Standards, 1993 Catalogue, Environmental Resource Associates.

(33)    WA 1.3, Development of PEMs to Support Regional Quality Assurance Needs for CLP and
        non-CLP Superfund Work in Solid Media  Quality Assurance Technical Support Laboratory,
        continuation of WA, Feb. 26, 1993

(34)    WA 2.5, Evaluation, Improvement, Creation, and Standardization of the CLP Pesticide Procedures
        in the Organic RAS SOWs.  Improvement of Mass Spectrometry Quantitation Levels of Semi-
        Volatile Target Compounds, Jan. 30, 1993.

(35)    Analytical Operations Branch, EPA/540-R-93-062, Draft "Guidance For The Uniform Use Of
        Performance Evaluation Materials In The Contract Laboratory Program And Related  Superfund
        Activities, April, 1993

(36)    Development of Performance Evaluation Samples for Field Analytical Screening  Methods, Quality
        Assurance Technical Support Laboratory, Jan, 1991
                                              10

-------
         IMPROVED LABORATORY PERFORMANCE  THROUGH MATRIX-ANALYTE
                      SPECIFIC PROFICIENCY TESTING

M. Carusof Technical Director,and K.W. Jackson, Director,  Environmental
Laboratory  Approval  Program,  Wadsworth Center  for  Laboratories  and
Research, New York  State  Department  of Health, P.O.  Box 509 Albany, New
York 12201-0509

ABSTRACT
The New  York  State Dept.  of  Health's Environmental  Laboratory  Approval
Program  has  been   challenging  its participating   laboratories  with
proficiency tests  since  1985.  A  review of  the  proficiency test  data
demonstrates a distinct  improvement in laboratory performance, as measured
by percent recovery, for analytes that have  not been included in EPA's WP
series.   For example,  the  recovery for Non-potable  Water phenol  has
improved  from an average of  3.3% in 1987  to 45.1%  in  1991. A similar
improvement can be  demonstrated  for phenol  in  soil  where  first  round
testing yielded  30.7 % recovery and 1992  test data demonstrates an average
recovery  of 61.0%.   The production of Solid  Waste  proficiency  test
samples,  and  laboratory performance in   analyzing  these  samples  is
discussed.  The  data show  clearly that the  traditional  practice of  only
challenging   laboratories  with   sealed  ampules  containing   analyte
concentrates does not realistically assess their  ability to analyze  real
solid waste samples.

INTRODUCTION

New York State's Environmental  Laboratory  Approval  Program  (ELAP)  was
established in  the  Fall of  1985 as a successor to and  expansion of the
state's  Safe  Drinking  Water  Act  laboratory  certification  program.
Initially, ELAP  provided accreditation in the  categories of Potable Water
and Non-potable  Water  to  approximately  600  laboratories.  As originally
structured, ELAP produced  and  distributed its own proficiency test samples
for  both chemistry and  bacteriology.   ELAP  has since  grown   and  now
accredits  over   900 laboratories  in 30  other states  and  six  foreign
countries.  Each  of  the certification categories,  Potable  Water,  Non-
potable  Water,  Solid and Hazardous  Waste,  and  Air  and  Emissions,  is
proficiency tested  twice  yearly.   In 1992,  the program manufactured and
distributed over 40,000  samples.    These  consisted  of  concentrates  in
sealed  ampules   for the  water  chemistry  analytes,  full-volume  water
bacteriology  samples,  air  filter strips,  and full-volume  solid  waste
samples.  A list of the program's proficiency test samples  is  given in
Table 1.

The Solid and Hazardous  Waste and Air and Emissions categories were added
to the existing  water categories early in 1987. Due to the rapid expansion
of the program's  breadth, ELAP  initially accredited  laboratories in the
Solid and Hazardous Waste (SHW) category by  requiring  them  to  have and
maintain  accreditation  for  the  same analytes in the  Non-potable Water
                                    11

-------
category. This was  based on the theory that  both  the Non-potable Water and
Solid Waste methods are essentially the same.  Thus accreditation for SHW
could  be granted  by a  combination of  Non-potable  Waster  proficiency
testing and on-site inspection of SHW preliminary extraction procedures.
However, this did not take  into account the reduced recoveries and errors
that  can  be  introduced  during  the  additional  extraction  or  matrix
reduction processes required with a real-world  solid waste sample (such as
soil,  sediment  etc.).    This paper  will  examine   the  improvement  in
laboratory performance that is seen  through the use  of  proficiency test
samples  that  closely mimic  the real  samples  that  are being  analyzed
routinely.

PRODUCTION OF SHW PROFICIENCY TEST SAMPLES

In  1988 ELAP  started producing  it own  SHW proficiency  test  samples
starting with metals in sludge.  In general, production of other analyte-
matrix samples (listed in Table 1)  follows  this  same  procedure.

Initially a bulk  sample of apparently analyte-free sludge, sediment, silt,
sand or soil is collected and oven-dried at  105° C overnight.  This sample
is then ground, sieved to 100 mesh and  roller  mixed.  Sample aliquots are
now  randomly collected to  determine  background  concentrations  of  the
analytes of  interest. The remaining  dried sample is  then weighed  and
sufficient  appropriate solvent  is added  to  create  a  slurry.   Using  a
paddle blade mixer,  the analytes of interest are  spiked into the slurry at
concentrations calculated to produce the selected dry-weight targets.  The
slurry  is  then air or oven  dried (depending on  the solvent used)  and
roller mixed.  The  dried powder is then dispensed into screw cap bottles.
De-ionized water  is next added to each  bottle to produce a wet real-world
consistency. The bottles are  then  capped and  sealed with  stretch tape.
Several  samples  are  collected  from the  beginning,  middle  and  end  of
production and analyzed for product  homogeneity.   Additional  samples are
collected randomly and analyzed over time to determine product stability.
DISCUSSION

A question frequently asked of any  laboratory  certification  program is:
"How effective are you?".  One way  of demonstrating  effectiveness  is to
show improvements in recoveries  and  interlaboratory precision.  A cursory
review  of  statistical  summaries  over  the  years   demonstrates   that
improvement has  been made  in recovery and precision,  especially  for
analytes not  included  in other performance  testing  schemes.   However,
because of the rapid growth of  the  Program and change in  the  nature of
laboratories  approved,  from basically facility-types  to  basically  full-
service commercials, the  improved statistics may  be due to changes in
laboratory type rather than  laboratory ability.  An in-depth study of the
causes  of apparent improved performance appeared  appropriate.
                                    12

-------
The study protocol called for the selection of a set of laboratories (n ~
50) that participated in proficiency testing both in 1986 and the current
rounds.  As demonstrated in Table 2, for the analysis of water proficiency
test ampules, the  laboratories  that  analyzed  for purgeable aromatics  in
1986  had  an  average  recovery  of  97.1%.  In  the  current round  these
laboratories had an average recovery of 100%,  a marginal improvement.  It
is believed  that  since purgeable aromatics  are commonly  found  in other
performance  evaluation  schemes  in which  the  laboratories  may  have
participated, these laboratories were approaching their peak performance
in  1986.   Alternatively, if  one  examines  laboratory performance,  as
measured  by recovery,  for phthalate esters,  polynuclear  aromatics  and
priority  pollutant  phenols  a distinct  improvement  can be  seen.  Again,
comparison was made using a set  of  laboratories participating in both the
1986  and  current  rounds.   Average recoveries  for  phthalate  esters,
polynuclear aromatics and phenols went  from 10.7% to 91.4%, 6.5% to 79.1%
and 5.8%  to 62.4% respectively.

The data presented in Table 3 reflect the improved performance of a set of
laboratories (n = 50) that initially participated in proficiency testing
for metals  or  phenol in  solid  waste.  The  table also presents  the  same
group's  performance  in an aqueous  matrix for the same analytes.  These
laboratories were  only  able  to  produce  an average recovery of  55.4% for
cadmium,  nickel  and lead in a  solid waste,  while achieving an average
recovery  of 99.5%  for  these  same  metals  in an aqueous  matrix.   However,
through continuing proficiency testing the initial set of laboratories was
able  to  improve  their recovery  of metals from solid waste  to  a current
average  of 95.8%.   Similar  improvement can  be  noted for  phenol,  with
recovery  going from 30.7% to 61.0%.

As illustrated in Table 4, improved recoveries were not the only benefits
of  continued   matrix-specific   proficiency   testing.   Interlaboratory
precision for cadmium, nickel and lead markedly improved from an average
relative  standard deviation of 27.9% to 8.2%.

An additional factor  to be considered  is that several  laboratories  that
initially applied for certification  in the  SHW category withdrew as  soon
as  they  were  challenged  with  the full-volume  samples.    An  obvious
conclusion is that they doubted  their ability  to handle these samples and
pass  the  test.  However,  if  the program had continued  to  rely  on sealed
ampule concentrates to test these laboratories, they would have presumably
continued  to  participate and been  granted  certification for  Solid  and
Hazardous Waste analysis.
                                    13

-------
SUMMARY
The data Leave Little doubt that the traditional protocol of challenging
laboratories with proficiency test samples consisting of concentrates in
sealed ampules does not accurately assess  their ability to analyze real-
world solid waste samples.  Considerable resources are being provided for
the identification and clean-up of solid waste  sites,  and the additional
expense of research,  development and production of  realistic proficiency
test samples would seem to  be easily justified.
                                  14

-------
                                 Table  1

             NYS  ELAP:  CURRENT PROFICIENCY TEST SAMPLE TYPES
          Potable Water

Drinking Water Bacteriology:
      Total Coliform (& E.coli)
      Standard Plate Count
   Analytes represenitive of:
Chlorinated Acids
Trihalomethanes
Microextractables
SDWA Metals
 Methylcarbamate Pesticides
   SDWA Minerals
Organohalide Pesticides
   Volatile Aromatics
Volatile Halocarbons
   Miscellaneous:
      Asbestos
      Semi-volatiles

       Non-potable Water

Wastewater Bacteriology:
      Total coliform
      Fecal coliform
   Analytes represenitive of:
Chlorinated Hydrocarbon
Pesticides
Chlorinated Hydrocarbons
Atrazine and Carbaryl
Dioxins
Demand
Haloethers
Metals I:
      Ba, Cd, Ca, Cr, Cu, Fe, Pb,
      Mg, Mn, Ni, K, Ag and Na
Metals II:
      Al, Sb, As, Be, Hg, Se, V
      and Zn
Metals III:
      Co, Mo, Tl and Ti
Nitroaromatics and Isophorone
Minerals
Miscellaneous:
      B, CN, Phenols, MBAS, TOG,
      pH and Oil & Grease
Nitrosoamines
Nutrients
Organophosphate Pesticides
   Polynuclear Aromatics
Polychlorinated Biphenyls
   Phthalate Esters
Priority Pollutant Phenols
   Purgeable Aromatics
Purgeable Halocarbons
   Residue

    Solid and Hazardous Waste

   Analytes represenitive of:
Chlorinated Hydrocarbon
Pesticides
Metals I:
      Ba, Cd, Cr, Pb, Ni and Ag

Metals II:
      Sb, As, Hg and Se
Polynuclear Aromatic Hydrocarbons
   Polychlorinated Biphenyls
      Phthalate Esters

Priority Pollutant Phenols
Miscellaneous: Asbestos

      Air and Emissions

Asbestos
Fibers
Formaldehyde
Lead
Nitrate
                                    15

-------
                    Table 2




IMPROVED RECOVERIES THROUGH PROFICIENCY  TESTING

Analyte
Benzene
Chlorobenzene
Bis (2-ethyl-
hexyl)
phthalate
Benzo butyl
phthalate
Dimethyl
phthalate
Phenol
Pentachloro-
phenol
Benzo (b)
fluoranthene
Dibenzo (a,h)
anthracene
1986
Target
59.4
127.3
161.3
69.9
32.2
85.8
64.5
107.0
94.7
42.1
81.9
243.0
75.2
120.0
148.0
63.4
38.6
76.4
Mean
59.4
119.7
158.0
67.3
5.3
8.1
6.8
8.6
7.4
5.0
3.54
5.3
7.0
9.0
4.7
4.0
3.5
8.1
% Rec.
100.0
94.0
98.0
96.3
16.5
9.4
10.5
8.0
7.8
11.9
4.3
2.2
9.2
7.5
3.2
6.3
9.0
10.6
Current
Target
31.2
50.2
29.8
47.7
33.5
83.8
28.4
68.7
59.9
75.2
105.0
75.1
79.6
99.5
44.1
118.0
30.3
48.6
Mean
31.4
49.9
30.1
47.5
25.9
67.3
28.4
69.3
56.9
71.5
47.7
33.7
64.0
78.6
43.09
97.3
20.7
32.0
% Rec.
100.5
99.4
100.9
99.5
77.3
80.3
100.0
101.0
95.0
95.0
45.4
44.8
80.4
79.0
99.5
82.5
68.3
65.9
                       16

-------
               Table 3




MATRIX-SPECIFIC RECOVERY IMPROVEMENT

Analyte
Cadmium
Water
Solids
Nickel
Water
Solids
Lead
Water
Solids
Phenol
Solids
1988
Target
33.0
5.0
170.0
40. 0
220.0
113.0
183.0
Mean
33.2
3.6
168.8
17.6
217.1
56.2
56.2
% Rec.
100.6.
72.6
99.3
43.9
98.7
49.8
30.7
Current
Target
70.0
60.0
400.0
401.0
250.0
184.0
333.0
Mean
68.3
56.2
394.7
403.3
246.9
171.4
203.0
% Rec.
97.0
93.7
98.8
100.6
98.8
93.2
61.0
               Table 4




INTERLABORATORY PRECISION IMPROVEMENT

Analyte
Lead
Nickel
Cadmium

Mean
56.2
17.6
3.6
1988
8
9.8
5.8
1.2

RSD
17.4
32.9
33.3
Current
Mean
314.0
398.0
56.2
6
28.1
28.8
4.7
RSD
8.9
7.2
8.4
                  17

-------
  COMMERCIAL LABORATORY POSITION ON NATIONAL LABORATORY ACCREDITATION
John E. Farrell.  Ill  Vice  President,   Enseco,   a division of Corning
Laboratory Services Inc.,  2200  Cottontail   Lane,  Somerset, New Jersey
08873; Jack R. Hall, Technical  Director,  IT  Corporation, 9000 Executive
Park Drive, Suite A110,  Knoxville,   Tennessee  37923; Barbara R. Hill,
Director  of   Administration,    WMX   Environmental   Monitoring,  2100
Cleanwater Drive, Geneva, Illinois  60134; Marlene  0.  Moore, President,
Advanced Systems, Inc., Post  Office  Box 8090,  Newark,  Delaware 19714;
Jerry L. Parr, Director  of  Quality  Assurance  and Technology, Enseco,
Rocky  Mountain  Analytical  Laboratory,   4955  Yarrow  Street, Arvada,
Colorado  80002;  James  Ploscyca,   Director  of  Quality,  Industrial
Environmental Analysts, Post Office   Box  12846,  Research Triangle Park,
North Carolina  27709;  Peggy  Sleevi,   Director  of Quality Assurance,
Enseco, 2612 Olde Stone  Road,   Midlothian,   VA   23113;  David N. Speis,
Director Quality Assurance  and  Technology,   Environmental Testing and
Certification Corporation,  284  Raritan   Center  Parkway,   Edison, New
Jersey 08818;  E.  Scott  Tucker,   Director   of   Analytical Laboratory,
Clemson  Technical  Center,  100   Technology  Drive,   Anderson,  South
Carolina 29625; Craig 0. Vinson,  Senior  Laboratory Manager, CH2M Hill,
2567 Fairlane Drive, Montgomery, Alabama  36123

ABSTRACT

Environmental testing  laboratories   conduct   sample   analysis on soil,
air, water, and other materials  that may contain potentially hazardous
substances.  Services of the  environmental  testing industry are widely
used by the U. S.  Government,   state governments,  and private industry
because almost  every  environmental  decision  is   based on laboratory
data.  Reliable data are critical   to protect the public health and the
environment.

While the primary need for the   data generated by these  laboratories is
to comply with federal environmental  regulations,  there is no national
accreditation   process   to   ensure   uniform,    rigorous  standards.
Environmental laboratories currently are   accredited   under a system of
multiple state  and  private  programs  with   varying  quality assurance
criteria and excessive processing  costs.    Implementation of a national
accreditation program  is  a  cost-effective   method   of  ensuring data
quality.

International Association of  Environmental   Testing  Laboratories, Inc.
(IAETL) has assembled  a  committee   of   its   membership to discuss the
effect and feasibility of  a  national   program.   IAETL  represents over
150 commercial laboratories with an   estimated  45  to  50 percent of the
                                    18

-------
commercial laboratory  capacity.    IAETL  has  adopted  a  position in
support of  national  certification.    IAETL1s  committee has reviewed
various scenarios and implications of  a national program including the
recommendations  of  the   Committee   on   National  Accreditation  of
Environmental Laboratories (CNEAL)  Report  to  EPA.   lAETL's position
supports the  CNEAL  recommendations  and  emphasizes  the  adoption of
certain  critical  components:    data  assessment,  onsite assessment,
performance evaluation  testing,  a  clearly  defined  process for both
laboratories and accrediting  organizations,  and uniform, well-defined
criteria.

The objective  of  our  presentation  is  to  share  information on the
implications of a national program  and the specific recommendations we
have developed on  standards,  organization,  and enforcement to ensure
that the national program will be effective.

INTRODUCTION

In late 1989,  the  International  Association of Environmental Testing
Laboratories  (IAETL)  recognized   the   critical  need  for  national
accreditation.      The   unwieldy,   expensive,   and   time-consuming
accreditation processes  currently  in  place  were adding considerable
cost to environmental lab  testing  with  little  or  no benefit to the
users of laboratory services.    IAETL  formed a National Accreditation
Committee to address the issues.

The National Accreditation Committee reviewed current federal programs,
specifically the CLP program,  USATHAMA,  the Navy HAZWRAP, A2LA, AIHA,
DOE, DOD, and the Drinking  Water  program.  Major categories common to
the programs  were  determined  and  a  questionnaire  was developed to
survey current state programs.

At this point, mailings on the  subject revealed a wider interest group
than just  environmental  labs  for  a  national accreditation program.
Customers of the environmental  laboratories such as engineering firms,
various industries, and regulatory  bodies  also  had a vested interest
because of concerns about environmental data quality and prices.  A new
coalition was formed from  the  various  interest groups called NELAC -
the National Environmental Laboratory Accreditation Coalition.

NELAC used information from its  participants along with the previously
gathered  information  on  federal   and  state  programs  on  standard
environmental   laboratory   practices   and   accreditation   programs
structures to  formulate  four  options  to  be  considered as possible
models  for  a  national   accreditation  program.    After  thoroughly
examining and debating the various strengths and weaknesses of the four
models, NELAC presented a single, comprehensive option.
                                    19

-------
The advantages of the  national  accreditation  program as described  in
the option developed by the IAETL/NELAC groups include:

     •  Self-sustaining structure

     •  Built-in controls to ensure continuing quality and uniformity  of
       the program.

     •  Structures already in place (state  agencies) may be used in the
       process.

     •  A mechanism for  making  necessary  changes through the Advisory
       Board exists.

     •  Regulatory control is retained.

     •  Third-party administration is provided for.

     •  Funding requirements from regulatory agencies is minimized.

     •  States would be reimbursed for their participation.

     •  Higher standards of  quality  could  be  applied consistently  to
       environmental data across the country.

     •  Duplication of  efforts  in  accreditation  could  be minimized,
       particularly for  audits,  performance  evaluation (PE) samples,
       and applications.

     •  QA requirements could be  centralized within a structure similar
       to existing programs within the  EPA.   There would be a dynamic
       mechanism of input into  these requirements through the Advisory
       Board.

     •  The EPA can influence states to adopt reciprocity as part of the
       acceptance criteria and through funding.

The key elements of  this  option  were  the accreditation process, the
accreditation criteria, data assessment,  onsite  assessment, performance
evaluation testing, accountability,  and enforcement.
                                   20

-------
ACCREDITATION PROCESS

The accreditation process concurred upon can be diagramed as follows:
                       EPA FEDERAL PROGRAM ADMINISTRATION
       ADVISORY BOARD
                             ACCREDITATION COUNCIL
                                 ACCREDITORS
                                  ASSESSORS
                    ACCREDITED  ENVIRONMENTAL LABORATORIES
The EPA FEDERAL  PROGRAM ADMINISTRATION  is  managed by a high-level EPA
official who  administers the Accreditation Council contracts.

The ADVISORY  BOARD  approves standards.    Initially it would be made up
of individuals from federal  and  state agencies, laboratories and data
users appointed  for two  years  from  a  volunteer list.  In subsequent
years, board  members would be elected from nominated individuals.

The ACCREDITATION COUNCIL is  an  EPA  contractor  hired with a minimum
contract of 3 to 5  years to administer the program, approve accrediting
agencies, and maintain records.

The   ACCREDITORS    are   state   agencies,   independent   third-party
organizations,   and federal  agencies  that  would  accredit  per  the
national program criteria.  This would allow existing infrastructure in
these organizations to remain intact and functional if they so desire.

The ASSESSORS are the trained  individuals hired by the accreditors and
required  to  undergo  competency  testing  to  perform  data  and site
assessments.
                                    21

-------
ACCREDITATION CRITERIA

The program must consist of clearly established, uniform criteria.  The
laboratories must have detailed requirements  that they are expected to
meet to be accredited.  Using these criteria, the laboratory may design
the appropriate systems and the  internal assessments to maintain these
systems.  The criteria  must  be  clearly  defined  so that they can be
applied uniformly from location  to  location  by  any accreditor.  The
accreditation council, the  accreditors,  and  the  assessors also must
have uniform criteria by which the program is administered and by which
accreditors and  assessors  are  selected,  trained, perform laboratory
assessments, and document the accreditation process.

Accreditation Criteria for Laboratories

Uniform criteria for the laboratory will enhance a laboratory's ability
to generate data of known  and  documented  quality.   A system must be
established by which laboratories can be evaluated to determine if they
are capable of performing competently and that their QA/QC programs are
functioning to ensure the production  of  reliable and usable data.  To
this end, the criteria for  evaluation of laboratories must address the
following.

Organizational Structure and Staff

The organizational  structure must  be  defined  such that the technical
functions can be  performed  adequately.    The  structure must not put
undue pressure  on  staff  to  influence  their  judgment.   The person
responsible  for  Quality   Assurance   (QA)   must  report  to  senior
management, and  the  QA  role  should  be  independent  from  the data
generation process.

Staff must have adequate education,  training, and experience for their
assigned functions.   Laboratory management should not rely on education
and experience alone to ensure that technical assignments are completed
properly.    Therefore,  the  laboratory's  training  program  must  be
comprehensive and  documented.    All  employees  should  have  a clear
understanding of their job  functions.   Adequate supervision should be
provided to ensure that work is properly completed.

Quality Systems

The criteria for accreditation  should dictate the minimum requirements
for documenting an internal  quality  system  appropriate to the size of
the facility and the  scope  of  the  analyses  performed.  The quality
system should be  well   documented  so  that  all  staff understand its
function and requirements.     The  accreditation program should provide
enough direction to  the  laboratory  to generate quality documents that
                                    22

-------
describe  the  management's  policies   and  values,  the  laboratory's
organization  structure,  responsibilities  of  management  and  staff,
general QA procedures, detailed  QC requirements, and corrective action
processes for the systems described.

Facilities and Equipment

Minimum criteria for the physical  plant  are delineated to ensure that
the environment does not  adversely  affect test results, that adequate
security is  available,  that  appropriate  equipment  is available and
employed for  the  offered  testing,  and  that  appropriate health and
safety programs are implemented.

Data Generation and Record Keeping

The  accreditation  criteria  must  address  minimum  requirements  for
documentation of the  procedures  used  in  the laboratory by requiring
standard  operating   procedures,   sufficient   documentation  of  the
analytical process to reconstruct  the  events, and adequate procedures
for storage of these data for an appropriate period.

Accreditation Criteria for the Accrediting Council and Accreditors

The accrediting council, the  accreditors,  and the assessors must meet
criteria similar  to  those  as  described  above.    The accreditation
council    (the  administrative  arm  of  the  program/directed  by the
advisory board)  and  the accreditors (agencies providing accreditation
under  the  direction  of  the  accrediting  council)  must  be clearly
defined.  The organizational structures  must  be known and provide for
an  environment  in  which   the   staff  are  free  from  compromising
influences.   Assessors  (those  people  performing  the  data and site
assessments hired by  the  accreditors)  must  have adequate education,
training, and experience for their assigned functions.

Additionally, adequate  training  must  be  provided  and documented to
ensure that the assessors  are  competent  to assess the laboratory and
can provide uniform assessments  to  all  laboratories that they visit.
Each of these  groups  should  abide  by  a documented quality program,
including corrective  action  processes.    The  process  by  which the
laboratory assessments take  place,  the  reporting  mechanism, and the
maintenance of these records must be clearly defined.

DATA ASSESSMENT

Data assessment must  be  conducted  as  part  of the quality assurance
review of the laboratory operation.    Data assessment can be conducted
in two parts:  (1) determination  of the data acceptability relative to
minimum data  quality  criteria  established  for nationally accredited
laboratories and (2)  determination  of  the  data  availability at the
laboratory's site.
                                    23

-------
Data can be  reviewed  in  advance  of  the  onsite  visit to determine
problem areas that  need  to  be  investigated  during  the visit.  The
periodic onsite review process will   determine if the laboratory's data
assessment process includes data handling procedures, control criteria,
and policies for data corrective actions.

The data assessment criteria should   be defined clearly by the Advisory
Board.   These  criteria  will  ensure  that  all  data  generated by a
nationally accredited  laboratory  will  meet  a  minimum  data quality
standard.   Project-  or  regulatory-specific  criteria  still  will be
needed  for  more  stringent  data  requirements;  however,  a national
program will set minimum standards for data quality criteria.

Current data  quality  often  is  based  on  the  contents  of the data
deliverable package.  For  example,   data deliverables presented in CLP
format are assumed to  be  of  better  quality than drinking water data
presented as a single  reported  concentration.   The data user assumes
that  quality  control  parameters   for   an  application  are  within
acceptable limits if the  data  are   presented  in CLP data deliverable
format.  The drinking water data assessment criteria are clearly stated
for every primary drinking water parameter, and each must be met before
reporting any data.  The  CLP data deliverables include documented data
assessment criteria as  part  of  the  report.    The  question of data
quality is not answered  by  how  much  of  the  data are presented and
verified in the report; the  question  of data quality is answered only
when the data are found to meet the  acceptance criteria.

All data from a laboratory must  be   documented as to its quality.   The
data assessment verifies that data  both meet the minimum data criteria
and are well documented.

ONSITE ASSESSMENT

A  single   national   program   for   accreditation  of  environmental
laboratories must include  the  key   element  of an independent,  onsite
assessment as part  of  the  accreditation  process.   This assessment,
along with the data  assessment  described  previously,  should form the
foundation to grant (or deny)  accreditation  to a laboratory.   We have
used the word "assess"  to  connote   that  the  primary function  of the
assessment is to observe and  evaluate,  not to correct deficiencies or
make recommendations.

The onsite  assessment  should  address  all   aspects  of the facility,
staff, and systems to provide confidence that the laboratory is capable
of providing the services for which  accreditation is requested.
                                   24

-------
After the initial assessment,  the  accreditor  should prepare a report
that lists the findings  and  documents  any  deficiencies that must be
corrected for the laboratory  to  obtain accreditation.  The laboratory
should be provided with an opportunity  to remedy any deficiencies.  If
the assessment shows that  accreditation  is justified, the accrediting
council then should issue  a  certificate authorizing the laboratory to
provide data for the parameters included in the scope of accreditation.
This information should be  maintained  by  the accrediting council for
dissemination to users of laboratory services.  The accrediting council
should  establish  a   procedure   by   which   laboratories  can  seek
accreditation for additional tests.

The accreditation should be valid  for  1 year.  An annual reassessment
must be performed by the accreditor  that should address any changes in
the laboratory based on  information  retrieval  and  the results on PE
samples.  Where  concern  is  warranted,  an  onsite  assessment may be
performed.   In  any  event,  additional  onsite  assessments should be
performed at least once every 2 years.

The assessment should address all aspects of the laboratory's facility,
organization, quality systems, and record  keeping.  A critical element
of this process is to  determine  if  the laboratory personnel  have the
necessary education, training,  technical  knowledge, and experience to
perform  their  assigned  functions.    It  is  the  responsibility  of
laboratory management to  ensure  that  laboratory  personnel meet this
requirement.  A separate  accreditation  of  analysts would not improve
the process or add to  the  assurances already established by requiring
comprehensive  in-house  training  and  documentation  programs.     The
program could establish  criteria  relative  to the expected education,
experience, and knowledge for analysts.

Assessors must have a thorough knowledge of the accreditation elements,
including detailed knowledge of the analytical procedures, and must act
in an  impartial  manner.    The  accreditation  system  must establish
procedures for ensuring the competency of assessors.

PERFORMANCE EVALUATION TESTING

Performance Evaluation or Proficiency Testing  is  a key element of the
program  in  confirming  the  laboratory's  capability  to  perform the
methodology and  identifying  areas  for  improvement.    A performance
evaluation program  is  therefore  a  critical  requirement of national
laboratory accreditation.
                                    25

-------
The accreditation program will require  completion of an initial series
of PE samples representing  the  parameters  and  matrices in which the
laboratory is seeking accreditation.    The  results  of the PE samples
will be compared to  the  certified  values and the expected acceptable
range for that parameter and matrix.

Because PE samples are not  available  for all  parameters in all matrix
types, the PE program  must  have  a  cross  section of PE samples that
covers representative parameters.    For  example,  a  volatiles PE may
contain 5 to 8 compounds to  represent  the 20  to 30 volatiles normally
analyzed in wastewater methods.   Matrices evaluation may be limited to
water, air, soil,  and  sludges  for  specific  inorganics and organics,
with other matrices available  for  only  select  items (PCBs in oil or
mercury in fish for example).    The areas of matrix accreditation must
be clear  to  all  participants  and  potential  users  of the national
accreditation program.  If, for instance,  a PE  for dioxins in fish does
not exist, does the program  accredit  a laboratory for dioxins in fish
if the laboratory is accredited in dioxins in water and soil; or should
the  approach  simply  be  that   unless  covered  by  PEs,   particular
parameters in matrices will not be accredited.

The fee structure will depend on the type  and number of PEs  required by
the laboratory.    As  new  PEs  or  PEs  in different matrices become
available,  previously    accredited   laboratories   as  well   as  new
applicants will be encouraged to participate in these studies.

Only laboratories generating acceptable  data  will be accredited for a
particular parameter or set  of  parameters  for  a  matrix  type by the
method certified  by  the  laboratory.    For  a  laboratory  failing a
particular PE, the accrediting council will work with the laboratory to
evaluate the program areas and  needed changes.  Laboratories,  at their
expense, have up to  three  chances  in  a  1-year  period to perform a
single PE  successfully.     After  initial   accreditation approval, the
laboratory  will   be  required  to  perform  PEs  annually  to maintain
accreditation  for  the  designated   PE  parameter(s)  in  the  matrix
provided.  Failing multiple  PEs  (two  or  more)  by a laboratory will
initiate a reassessment by the accrediting council.

A well-organized, broad-based  PE  program  will   be  the  heart of the
national accreditation program to  ensure   continued proficiency of the
member laboratories.
                                    26

-------
ACCOUNTABILITY AND ENFORCEMENT

A critical element of  any  accreditation  program is accountability on
the  part  of  the  accreditors  as  well  as  the  laboratories  being
accredited.  In addition,  the  accreditors  must  be  in a position to
enforce any and  all  accreditation  and/or revocation of accreditation
decisions that they make.

The accreditors will  evaluate  laboratories  through  a combination of
onsite assessments, data  assessments,  and  performance on proficiency
test samples  submitted  to  the  laboratories.    The  onsite and data
assessment activities will be used to grant or revoke accreditation for
the laboratory.  The accreditation  must  be  specific for the areas of
testing performed by the laboratory, to incorporate both parameters and
matrix.  As discussed  previously, unacceptable performance in analysis
of PE samples can  result  in  loss  of  accreditation for a particular
parameter.  In  this  context,  the  accrediting  council  must be held
accountable for developing the following:

     • The   accrediting   council   is   accountable   for   providing
       requirements  and  standards  for  conducting  onsite  and  data
       assessments.  The  accrediting  council  also is responsible for
       establishing a viable performance evaluation testing program.

     • Once assessors are  selected,  the  performance of each assessor
       must be evaluated annually,  at  a  minimum,  and allow for more
       frequent evaluation if  their  performance  is questioned at any
       time.  This review  should  be  based  on a combination of their
       accreditation record and evaluations from laboratories.
      • A hearing process must be  establishedstoMresolve disputes about
       assessor performance and  laboratory  performance  issues.  This
       process would provide  all  parties  involved the opportunity to
       express their points  of  view  before 'affinal determination is
       made.

      • Assessors who  have  demonstrated  unacceptable performance must
       not be allowed to continue in that capacity once a determination
       has been made.  The accrediting council would maintain an active
       list of acceptable accreditors and assessors.

      • Laboratories  that  have  demonstrated  unacceptable performance
       must not be  allowed  to  maintain  an  accredited  status.  The
       accrediting council would maintain  an active list of accredited
       laboratories, along with the areas of accreditation.

Enforcement of  decisions  about  acceptable laboratories, accreditors,
and assessors  would  be  controlled  through  the  maintenance  of the
approved listings.    A  real-time  listing  of  approved assessors and
accredited laboratories could be  maintained on a computerized bulletin
board system.
                                    27

-------
CONCLUSION

Implementing  a  national  accreditation  program  is  a cost-effective
vehicle for improving and  maintaining  data  quality.  Such a national
program will help ensure  that  decision  makers and the public receive
accurate  and  reliable  data   to  make  cost-effective  environmental
decisions.  Users of environmental testing services need an independent
and comprehensive verification that laboratories have the capability to
perform  competently  and  that  laboratory  Quality  Assurance/Quality
Control programs are functioning  to  ensure the production of reliable
and  usable  data.    It  is  critically  important  that environmental
monitoring data are reliable, reproducible, accurate, and defensible to
support regulatory compliance  and  enforcement  decisions  made by the
public and private sector.
A  single,  comprehensive   national   program
environmental laboratories must establish:
for  accreditation  of
     1.  Uniform,  rigorous  requirements   for  environmental  testing
         laboratories to generate data of known and documented quality.
     2.  Uniform,   rigorous   standards    under   which
         organizations can oversee the program.
           accrediting
The accreditation process  should  be  administered by a private/public
partnership with federal oversight.  Procedures would be developed by a
governing board of representatives  from  the  EPA and other interested
federal parties, state agencies, the laboratory community, and users of
environmental data.
A program designed and administered with these features
costly,  redundant  requirements  while  significantly
accuracy, reliability,  and  timeliness  of  decisions  that affect the
quality of environmental cleanup and restoration efforts.
       would eliminate
       increasing  the
 IAETL supports and  continues  to  work  toward a single, comprehensive
 national program that will provide  an effective means of improving the
 consistent generation of high quality data to comply with environmental
 laws and to protect public health and safety.

 REFERENCES

 Jeanne Hankins, Final Report of the Committee on National Accreditation
 of Environmental Laboratories.  United  States Environmental Protection
 Agency, Washington, DC September, 1992.
                                    28

-------
    COST EFFECTIVE PROJECT PLANNING:  AN ALTERNATIVE APPROACH
                  TO INITIAL RCRA FACILITY INVESTIGATIONS

D. Loring. Principal, Loring Environmental Associates, 10 Graham Terrace, West Roxbury,
Massachusetts 02132; and

S.  Chapnick,  Program  Manager,  Gradient  Corporation,  44  Brattle  Street,  Cambridge,
Massachusetts 02138.

ABSTRACT

In a recent report by the General Accounting Office,  the RCRA program is criticized for being
"excessively slow in nearly all areas, and charges EPA with failing to develop a viable RCRA
cleanup strategy."1  It is the conduct of the large-scale multimedia RCRA Facility Investigations
(RFIs) which this paper addresses.  We propose an  approach that combines stages in the RFI
process and redefines the current phased approach to RFIs.  This alternative approach can reduce
the time and cost associated with the various RFI process stages and improve the  responsiveness
of determining appropriate corrective measures.

Project planning is an essential part  of designing large-scale investigative programs.  Although
the initial planing process can be quite time consuming, it is vital in achieving data that are both
scientifically sound and legally defensible to support corrective action decisions.  However, many
investigative programs have been delayed for years in the planning and negotiation stages. This
delay  is inefficient,  costly,  and  can result in a continuing risk to public health and  the
environment.

A key element in the proposed approach is a Phase I scope that is defined through an iterative
process with the regulatory agency, owner, and engineering and  laboratory contractors.  The
scope is defined based on site-specific information (e.g., historical chemical and physical data,
hydrogeological information), applicable  and relevant or  appropriate requirements (ARARs),
health based criteria and other toxicological information.  A site-specific sampling scheme and
target compound list is defined.  For  example, if groundwater is not a relevant exposure pathway
at the site, then groundwater samples would  not need to be collected  in the initial  phase.
Recommendations are made to base  methodology on  both health-based criteria and background
levels  of contaminants of potential concern. This approach allows for the controlled  use of
alternative methodologies (i.e., field techniques) that may have  higher detection limits than
standard methods but are acceptable to define the site conditions in reference to risk calculations.
The initial object of the process would be to reduce  risk,  not to return the  site  to a "pristine"
level.

This approach for RCRA sites is similar, in goals, to  the Superfund Accelerated Cleanup Model
(SACM).   The use  of a simpler approach and a narrower focus to the Phase I RFI allows for
timelier risk reduction for people and the environment.
                                           29

-------
INTRODUCTION

RFI guidance  defines the RCRA  Corrective  Action Process as  having numerous stages.
Generally, the approach to RCRA  Facility  Investigations (RFI) has  included a  large-scale
sampling and analysis program (as Phase I of a phased strategy) encompassing a broad range of
target compounds such as the full Appendix IX.  As a result of the magnitude of the Phase I
investigation (in terms of number of parameters, analytical methods,  detection limits, etc.), the
program details are so voluminous that the planning process can take  years. Additionally, once
the RFI is underway, several other stages in the RCRA process follow, including: review of the
results  by  the  regulatory  agency  based  on  health and  environmental  assessments;
recommendations for  interim corrective measures,  and/or a Corrective Measures Study (CMS);
implementation of the CMS; evaluation of the CMS; recommendations of appropriate corrective
measures  (actions),  and Corrective Measures Implementation(CMI).

In a recent report by the General Accounting Office, the RCRA program is criticized for being
"excessively slow in nearly all areas, and charges  EPA with failing to develop a viable RCRA
cleanup strategy."1

It is the conduct of the large-scale multimedia RCRA Facility Investigations (RFIs) which this
paper addresses. We propose an approach that combines stages in the  RFI process and redefines
the current phased approach to RFIs. This alternative approach can reduce the time and cost
associated with the  various RFI process stages and improve the responsiveness of determining
appropriate corrective measures.

OVERVIEW OF THE RCRA CORRECTIVE ACTION PROCESS

The RFI is carried out under either a permit or enforcement order.  RFIs  can range from limited
investigations to large-scale investigations for a wide variety of target analytes and matrices.  The
RCRA Corrective Action Process has numerous stages.  A RCRA Facility Assessment (RFA)
should include a review of historical information at the site and a site  inspection by the EPA, to
identify solid waste management units (SWMUs),  collect existing  information on contaminant
releases, and identify suspected releases  needing further investigation. Depending on the results
of the RFA, it is possible to either proceed with a RCRA Facility Investigation (RFI) or, in cases
where  public health or  the environment  is  threatened, interim corrective measures may be
required.

The owner or operator then performs the RFI to either verify the contaminant release, if required,
or to characterize the nature, extent, and rate of  migration of potential future releases.   The
results of this stage of the process are submitted to the regulatory  agency  for review.  Further
steps in the RCRA corrective action process include evaluation of the data generated from the
RFI against established health and environmental  criteria.  At this  stage, if further action is
determined necessary, a  corrective  measures study (CMS) is  performed  which  recommends
appropriate measures  to  correct the  release.  Interim Corrective measures can be taken either
prior to or during the  CMS if a threat to health or  the environment is identified during  the RFI
stage.
                                           30

-------
Finally,  subsequent to the CMS, corrective measures are implemented (CMI).  This includes
design, construction, operation, and monitoring of the corrective measures decided upon by the-
regulatory agency.

ITERATIVE PROCESS IN DESIGNING THE PHASE 1 SCOPE

Evaluation of a RCRA site is extremely complex, and requires the application of a wide variety
of scientific disciplines, and diverse technical and regulatory knowledge. During the course of
an RFI, many individual contractors are hired to participate in the various phases of the project.
The contractors can include engineers, field samplers, hydrogeologists, laboratory personnel,
toxicologists, risk assessment specialists, and quality assurance consultants.

However, during the initial planning stages in which the Sampling and Analysis Plan (SAP) and
Quality Assurance  Project Plan (QAPP) are  developed, the expertise of a number of these
individuals is not utilized.   For many  programs, there is minimal communication between
members of the  "team."  It is crucial to the success and  expediency of the RFI process to
assemble the entire project team to develop the work plan  for the  specific site  that takes into
account all of the necessary information required to generate valid data upon which to support
corrective action decisions.

Project team members with risk assessment or toxicological expertise can influence the scope of
the RI at the planning stage in order to avoid the low-level analysis of certain analytes that do not
pose a threat to human health or in order  to  ensure low-level detection limits  in cases  of
ecological  risk.  Additionally, these  team members  will help define the possible exposure
pathways and media so that the SAP can be formulated to minimize the numbers and types of
samples  being collected  and analyzed while maximizing  the usable  data generated.  Also, this
planning process can avoid costly resampling efforts due to omission of a significant media (e.g.,
air) or parameter type.  At this stage, human-health based risk calculations can be performed
using historical data, if available, as an estimate of the potential risks at the site. Health-based
cleanup  levels  can also  be  estimated  from these  risk calculations  to help clearly define the
detection levels needed for the  analytes of concern specific to the  site.

Both the EPA and  the site owner are  in  a position to require and facilitate working meetings
during which an appropriate target analyte list is  developed and associated required method
detection limit are determined.   This requires a commitment from  both the regulatory agency
and the site owner.  From this information, methodology choice (in particular field vs. laboratory
methods), data quality objectives, and a relevant sampling scope can be developed.  Only from
these criteria can a valid SAP and QAPP be developed.   In terms of cost-effectiveness and
timeliness  it  is crucial to evaluate field  analytical methods (e.g.,  field gas chromatographic
methods for analysis of volatile organic compounds, X-Ray Fluorescence techniques for the  field
analysis  of metals, bioassay techniques for the field analysis of PCBs) as viable  alternatives to
laboratory analyses. Field method technology has become more sophisticated and rigorous and
allows for the rapid analysis of constituents of concern on-site.  In addition, laboratories can
develop fast, cost-effective methodology when presented with a site-specific TAL  and associated
detection limits (e.g., "Soxtet" extraction, Solid  Phase Extraction, GC-MS-SIM, GC-ITD).
Therefore, decisions that are based on chemical information can be made more quickly - speeding
up the RFI process.
                                            31

-------
The current RFI guidelines are certainly not in conflict with this recommended process. In fact,
some of these activities are required.   However, from  experience  (in both  authoring  and
reviewing numerous Work Plans, SAPs,  and QAPPs in various EPA  Regions and states)  it is
evident that an organized, effective planning process rarely occurs. Much of the information
which does exist about the site history, physical site conditions, and historical data is not taken
into account during the planning phase or, if it is included in the Work Plan, is not included as
relevant information upon which to define specific QA/QC criteria in  the QAPP  This results
in a program where the target analyte list and selected methodologies  are inconsistent with the
site history and expected contaminants and physical conditions at the site.

Coordination between the project team members needs to be maintained to generate a SAP and
QAPP that  the field personnel  and laboratory personnel can  follow as a protocol  and that is
sufficient to support usable data. However, the SAP and QAPP are often done as "boiler plate"
documents with  only minor changes due to site-specific information. This results in a waste of
both cost and time when, for example, strict QA/QC criteria are required that are not relevant
to the type of samples being collected or the expected chemical contamination. For example, a
QAPP for a site in Louisiana required CLP-like quality control, DQOs, and detection limits for
sludge samples  that were known to be highly contaminated  with  organics and  metals.   The
laboratory could not meet most of the QAPP required matrix  QC or the detection limits stated
because of the highly complex and contaminated matrix. The complex matrix and the level of
contamination expected should have been taken into account during the generation of the QAPP
Less strict QC would not have compromised the usability of the data generated for this site  due
to the highly  contaminated nature  of the samples  — in other  words, the  relative  extent of
contamination and the order-of-magnitude  were the necessary information.  Lack of coordination
between project team members caused an increase in cost (laboratories charge more for CLP-like
QC than for more standard analyses) and time (CLP-like  QC required more time  to analyze  and
generate laboratory reports).

GUIDELINES FOR PHASE I OF RFI

The following steps are  presented as guidelines for the coordinated project  team approach in
generation  of the project QAPP, with specific  emphasis on the usability of the chemical data.
Similar team approaches would be required in generating the SAP and work plan and would bring
together the engineering and field  personnel with  the  laboratory and toxicologists and  risk
assessment specialists:

1.      Determine Target Analyte List (TAL) based on Historical Site  Activities  &  Data

The traditional approach in planning an RFI has been to analyze for the full Appendix IX list of
compounds.  However, many of the analytes included in Appendix IX may not  be appropriate
given the  historical information which exists about operations at  a particular facility and the
results of previous  sampling and analysis.  The project team members involved in the initial
planning need  to have the site-specific information necessary to define the scope in terms of the
target  analyte list.  A  site-specific target  analyte list must then be  developed  using  this
information. The expertise of both risk assessment and analytical chemistry personnel should be
involved in defining the site-specific TAL.  Without this information and coordination of project
team members, the  result may be inflated sampling and laboratory costs, and also higher costs
                                            32

-------
to validate data which is not of any particular use in the final determination of required corrective
measures.  In the Superfund Accelerated Cleanup Model (SACM), this problem is addressed:

                "Many, if not most of these assessments start from scratch, — they do
               not necessarily take into consideration the information and data generated
               by the studies that preceded them. This happens not only because of the
               obvious financial incentives to the contractor community and the human
               inclination to distrust the work of others, but because each part of the
               program is gathering data to respond to its particular perceived need."2

In some cases,  the  directives to produce  redundant  data come directly from the regulatory
authorities in an effort to restore the.site to "pristine" conditions. However, the focus must shift
from this effort to an effort of "worst first"3 as described in the SACM model.

2.      Define  the Sampling  and Analysis Procedures Based on  Site Conditions, Historical
Information, and Exposure Pathways

With the input  of the engineering and risk assessment  personnel, project costs and time-frames
may be reduced, for example,  by deciding that a portion of the soil and/or groundwater samples
may be analyzed using field rather than laboratory  methods.  This would  allow for less strict
QA/QC, rapid generation of chemical results, and contribute to a more timely progression to the
corrective measures process.

Often, the project personnel writing the QAPP are given instructions to "write a QAPP for the
full  Appendix IX list for water and soil samples."  Although, for example, hydrogeological
information about the site is used  in defining the  sampling plan, samples collected and analyzed
from a crucial part  of the site (e.g., a known  hot spot or chemical release) are not treated
differently than others in terms of defining DQOs, methods,  quality  control criteria,  detection
limits, and validation requirements. The same DQOs and detection limits are generally applied
to all samples in this type of generic QAPP. By changing this approach, by generating a QAPP
specific to the real needs of the program, usable  data will be generated in a more cost-efficient
approach.

3.      Review SAP, QAPP,  and Target Analyte List in Terms of Both Ecological and Human-
Health Risk Potential

Once the site-specific TAL is identified for the site, risk assessment  estimates  (preliminary
calculations based upon existing data from the  site) must  be taken into  account to choose
appropriate detection limits required for each of the  compounds on the list.

Volumes of historical data on which risk calculations may be based exist for most sites which are
in the RCRA corrective action process.  However, risk assessment personnel are usually not
brought into the RFI process until after the Work Plan, SAP,  and QAPP have been developed,
and after the analyses have been performed.  The interactive process of the project team, to make
decisions on detection limit requirements and methodologies, must occur prior to the development
of these site documents that direct the activities for the RFI.
                                            33

-------
The  input  from  risk assessment  project team  members  is crucial  in  choosing  appropriate
methodologies, both  conventional and  alternative.   This information is  vital to adequately
determine required quantitation levels and DQOs specific to the project.  For example, recent
RODs (1991) in both Region 1 and Region 5 have required targeted cleanup levels of 1 M§/L for
vinyl  chloride and 1,000 ng/L for toluene.4  However, most QAPPs require a detection limit of
10 fig/L for both compounds and identical DQOs and QC criteria for both compounds.  In this
scenario, the toluene results were generated by  the  laboratory under far  more  stringent QC
requirements than necessary to support the usability of the data  in  a  cleanup level model.
Conversely, any not detected vinyl chloride  results generated at the  10 /xg/L detection limit are
not usable because these nondetects are reported at an order-of-magnitude higher level than the
level  of concern (cleanup level) for this compound.

If, at  the planning stage, it is known that both toluene and vinyl  chloride expected contaminants
at the site,  and that  toluene  is potentially  highly  contaminated (i.e.,  in  the  ppm range) as
compared with vinyl chloride (i.e., in the low ppb range), methodology can be selected which
will give relevant reporting limits for both compounds based on the targeted cleanup levels, as
listed above.

However, if the  historical information about toluene and  vinyl chloride at the site were not
reviewed, and the planning not performed as a project team process,  the routine approach would
be to  define QAPP requirements to analyze for the Appendix IX Volatiles list by Method 8240.
This would result in not detected vinyl chloride results reported  at a  high sample detection  limit
due to dilutions that would be required for high levels of toluene.  With appropriate planning,
incorporating  the  expertise  of an analytical chemist and risk assessment personnel, a  GC
screening method for  toluene could be defined for the project and a  GC/MS method with strict
QC (e.g., EPA method 524.2) could be defined for vinyl chloride so  that the low-level detection
limit  needed for usability of the data could be achieved.

4.      Establish A "Hot" List for Interim Corrective Measures

Interim Corrective Measures may be necessary at any point in the RFI process.  During the  RFI,
sources can  be  identified as hot-spots as those regions of the site with the highest concentrations
of contaminants.   Once these hot spots have been identified, they should be evaluated against
health-based  criteria  or cleanup  level  criteria,  as  appropriate.   Field  screening analytical
techniques can  be extremely useful in identification of hot spots  and  extent of contamination for
rapid  response.  Interim corrective measures are appropriate at this stage.

SUMMARY

One of the crucial differences in the strategy which is recommended is evaluation of the historical
data  against both health-based criteria and estimated (or historical)  cleanup goals at the initial
planning stages of the  RFI and prior to the formulation of the Work Plan, SAP, and QAPP.  This
evaluation should be the basis from which the target analyte list (TAL) is identified, the Sampling
and Analysis Plan (SAP) and Quality Assurance Project  Plan (QAPP) are written, and the  basis
from  which  the  Data Quality Objectives  (DQOs),  including  target quantitation limits  and
methodologies, are chosen. Without this step, many of the current RFI activities are superfluous
in that methodology which is costly and time consuming  is being used for analytes which are not
                                            34

-------
harmful to human health at the low detection levels required by the methods or not present at the
site being investigated.  This strains the budgets and timetables and results in taking resources
away from remediation.

EPA is presently in the process of establishing requirements for RCRA corrective action.  These
rules are to address various parts of the RCRA corrective action process that are presently either
not defined or subject to misinterpretation in implementation. These regulations will be published
as Subpart S of 40 CFR Part 264 and are intended to address  the following issues:

        •      Action levels based on health and environmental factors
               •       Methods for development of action levels
               •       Specific action levels for soil and surface water
               •       Adjustment of action levels based on additive toxicity
               •       Influence of quantitation limits on action levels
        •      Definition of "constituent"
        •      Definition of Solid Waste Management Unit (SWMU)
        •      Notification and reporting
        •      Semantics

The  approach proposed in this paper was intended to address the issue of health-based action
levels by involving risk and toxicology personnel up-front in the  planning  process.  A  key
element in the proposed approach is a Phase I scope that is defined through an iterative process
with the regulatory agency, owner, and engineering  and  laboratory contractors.  The scope  is
defined  based on  site-specific information (e.g.,  historical  chemical  and  physical data,
hydrogeological information), applicable and relevant or appropriate requirements (ARARs),
health-based criteria and other toxicological information.   A site-specific sampling scheme and
target compound list is defined. For example, if groundwater is not a relevant exposure pathway
at the site, then groundwater samples would  not need  to be collected in the initial phase.
Recommendations are made to base methodology on  both health-based criteria and background
levels of contaminants  of potential concern. This approach  allows for the controlled use of
alternative methodologies (i.e.,  field techniques) that may have higher detection limits than
standard methods  but  are acceptable to define the site conditions  in  reference to risk-based
calculations for cleanup levels.  The object of this phase of the RFI process would be to reduce
risk, not to return the site to a "pristine" level.

This approach for RCRA sites is similar, in goals, to  the Superfund Accelerated Cleanup Model
(SACM).  The use of a simpler approach and a narrower focus to the Phase I RFI allows for
timelier risk reduction for people and the environment.  Additionally, this approach focuses on
one  of  the  four primary goals set by EPA's Administrator, Carol  M.   Browner, to make
permitting and other EPA decisions more  prompt,  fair,  and definitive by using streamlined
reviews and greater reliance upon state and local agencies.

FOOTNOTES

1       Energy & Environmental Document Service Catalog  First Quarter 1993.
                                            35

-------
2      Superfund Accelerated Cleanup Model (SACM), US EPA Office of Emergency and
       Remedial Response, Pub. 9203.1-01, March, 1992.

3      Ibid.

4      ROD Annual Report, April 1992, US EPA Office of Emergency & Remedial Response,
       Publication 9355.6-05-1.

REFERENCES

RCRA Orientation Manual. 1990 edition, US EPA Office of Solid Waste, EPA/530-SW-90-036.

Interim Final RCRA Facility Investigation (RED Guidance - Volume I  and II. May 1989, US
EPA Office of Solid Waste, EPA 530/SW-89-031.
                                       36

-------
          MATRIX SPIKES AND  SURROGATES,  DO WE NEED BOTH?

G. Robertson,  U.S.  Environmental Protection Agency,  EMSL-LV, Las
Vegas, Nevada  89193,  and M.  Stapanian and F. Garner,  Lockheed
Environmental  Science and Technology Company, Las  Vegas,  Nevada
89119.

The users  of measurement data in environmental chemistry require
specific knowledge of the quality of environmental measurements.
Since samples  are known to vary considerably in  analytical
difficulty, it is best to have estimates of precision and
accuracy  for each analysis.   Conventional approaches,  such as
performing a replicate spiking study (e.g., standard additions)
on each sample would be prohibitively expensive  for the analysis
of organic samples.  By examining existing quality assurance
data, we  investigated the relative merits of two lower-cost
alternatives.   These include 1) spiking one sample from each
batch of  samples of the same matrix with a subset  of the analyte
list and  performing duplicate analyses of this fortified sample,
and 2)  spiking every sample with a known quantity  of surrogate
compounds (substitute or isotopically labeled compounds).  While
either  approach will yield estimates of precision  and accuracy,
we found  better information in the .estimates derived from
surrogate analysis.  Minor modification of the set of surrogate
compounds would further enhance the quality of these estimates.
Based on  these results and on cost considerations, we recommend,
contingent upon the execution of a properly designed confirmatory
analytical experimental study, the use of surrogate recoveries in
the estimation of precision and accuracy  in each semi-volatile
and pesticide analysis.  The analytical design of the analysis
for volatile compounds will require a new approach to obtain
meaningful estimates of precision and accuracy.
                             NOTICE

 Although the research described in this abstract has been funded wholly
 (or in part) by the United States Environmental Protection Agency, through
 Contract Number 68-C-CO-0049 to the Lockheed Environmental Systems and
 Technologies Company, it has not been subjected to Agency review.  Therefore,
 it does not necessarily reflect the views of the Agency and no official
 endorsement should be inferred.
                               37

-------
Abstract not Available

-------
          QUALITY ASSURANCE FOR ENVIRONMENTAL SAMPLING

Nick C.  Adolfo.  QA/QC  Coordinator, Core  Laboratories,  1250 Gene Autry  Way,
Anaheim, CA 92805; Ann Rosecrance, Corporate Quality Assurance Director, Analytical
Chemistry Division, Core Laboratories, 10205 Westheimer, Houston, TX  77042.

ABSTRACT

The concept of quality  assurance (QA),  in environmental  sampling and  analytical
activities, is most often applied to the environmental laboratory. However, to ensure that
data are both legally defensible and scientifically valid, the procedures, equipment, and
personnel for the sampling portion of the environmental measurement process must meet
requirements similar to those for the laboratory.  They must also be reviewed with equal
criticism, because the potential for error which can result in inadequate, biased, or invalid
data is just as great during sampling operations  as it is during laboratory analysis.

Though the QA requirements for field sampling are well documented in Environmental
Protection Agency (EPA) references, they are often contained in documents where
sampling is not the main focus, and are not always noted or understood.  Therefore, much
of what is known or practiced in QA for field sampling is obtained the hard way, through
trial and error.   It is essential that those performing the work have clear and specific
guidance on QA procedures for the sampling portion of the environmental measurement
process.

The QA activities associated with environmental  sampling can be divided into three areas:
pre-sampling preparation, collection of samples including quality control samples, and
post-sampling  preparation.    This   article  will  review these  areas  and provide
recommendations  on  the specific steps  needed  to produce legally defensible  and
scientifically valid data.

PRESAMPLING ACTIVITIES

Presampling activities include both organizational  and procedural planning.   A project
coordinator and an independent QA coordinator must be selected. Also,  to ensure that
the responsibilities for each different phase  of the project are clear and that critical parts
of the investigation are not be overlooked,  a project organization chart must be written
outlining the specific responsibilities of each project team member.

Once the  project organization is clearly defined, a QA project plan (QAPP) must be
written, which often happens in tandem with the writing of a work plan. Sometimes, the
two plans are combined into one document. Required elements that must be included in
the QAPP are listed in Figure 1. In addition, the workplan or combination QAPP/work
plan should also include:
                                       39

-------
FIGURE 1.   ELEMENTS OF THE QUALITY ASSURANCE
           PROJECT PLAN
1.   Project Description and Purpose

2.   Project Organization

3.   Data Quality Objectives

4.   Sampling Procedures

5.   Sample Custody and Preservation

6.   Calibration Procedures

7.   Analytical Procedures

8.   Data Reduction, Validation, and Reporting

9.   Internal Quality Control

10.  Systems and Performance Audits

11.  Preventative Maintenance

12.  Data Quality Assessment

13.  Corrective Action

14.  Quality Assurance Reporting
                              40

-------
       o      a site specific sampling plan which is based on geostatistical or historic
              information about the site;

       o      standard operating procedures  (SOPs) for sampling for every  type of
              matrix involved including, groundwater, wastewater, soils, sludges, air, and
              composited samples;

       o      a summary of the tests to  be  performed based on historical data, the
              regulation of concern, and the data quality objectives; and,

       o      the number of samples, analytical  methods,  bottles, preservatives,  and
              sample holding times for each sampling location;

The analytical laboratory should be  selected prior to beginning the sampling phase of the
project.  The selection should be based on the laboratory's reputation for quality, the
results of EPA performance evaluation (PE) sample studies, the analysis of PE samples
for the specific tests to be performed, and an on-site laboratory audit.  The laboratory
should be kept fully  informed of  the anticipated sampling  schedule,  the number of
samples to be taken, and the time the samples  will arrive at the laboratory.  This allows
the laboratory  to ensure that both holding  and turnaround times can be met.  Sending
samples to a laboratory that is not informed and not prepared to receive them can result
in the samples being given  less attention or priority then samples which the laboratory
had already accepted for analysis.

Sample bottles,  preservatives, chain-of-custody  records, custody  seals,  and  other
documentation should be acquired  before  going into the field.  The sample shipping
method also should also be determined beforehand. Proper planning canjielp eliminate
common and detrimental sampling-related mistakes, such as the use of inappropriate
sample containers, not preserving samples adequately, and not shipping the samples to the
laboratory for  several  days because the  laboratory or the method of shipment was
unknown.

Project team members should be thoroughly trained for all equipment and procedures they
will be expected to use during the sampling  operation. They should also have been given
a 40-hour hazardous waste training course as required in the Code of Federal Regulations
(29 C.F.R. 1910.120) before entering potentially hazardous areas and collecting samples.
This training should be documented in individual training records for each project team
member.  If the training  documentation is procedure-specific and not project-specific,
these training records can be used for future projects that involve the same procedures.
A project-specific safety plan must be written and an initial project meeting should be
held to make certain that all project team members are familiar with the procedures to be
                                        41

-------
used. The project safety plan should outline the specific chemical and physical hazards
expected to  be encountered during  the sampling operation, and  provide measures for
protection against those hazards. It  is advisable, especially for larger projects, to select
a project safety officer who is either an industrial safety or industrial hygiene professional
to develop and implement  the  safety plan.  It is  also  advisable  to  conduct personnel
exposure monitoring and safety audits during the course of field operations.

The  need to have quality considerations written into the formal plans before the actual
operation begins is documented in the EPA's Guidance on Remedial investigations Under
CERCLA: "Because the primary aim of the quality assurance/quality  control  program is
to ensure that the data are reliable, rather than to ensure that a poorly conducted program
is adequately  documented,  the QA/QC aspects should be planned  in  advance as an
integral part of the investigation."

COLLECTION OF  SAMPLES

Quality considerations in the collection of field samples  include the use of sampling
equipment and standard procedures,  documentation of field activities, collection of field
QA samples, and performance of field QA audits.

Standard cleaning procedures for sampling equipment should  be used between samples
to ensure that cross contamination does  not occur.   If historical  data  about the
concentration levels of analytes of interest in the samples  to be taken is available,  it is
often advisable to take the samples in order of increasing concentration to help prevent
contamination  of the sampling equipment  after a high level sample is taken.

SOPs for use of sampling equipment, collection of samples, and the handling  of samples
after collection should be available in the field.  Copies of the QAPP and/or workplan
also  should  be available to  project members in the field as reference.  The  most best-
written QAPP  and the most detailed SOPs can become ineffective if the sampling crew
is not provided with copies.

Bound field  notebooks should be used to  record such items as the general sequence of
events,  the  number of samples  taken,  any  changes  in sampling  plans,  geological
observations, atmospheric conditions, difficulties in obtaining samples, and sample dates,
times,  locations, identifications, and descriptions (e.g.,  odors and colors).   Any other
unusual circumstances also should be recorded.  For projects involving drilling  operations,
boring  logs should be used.

Bottles, labels, preservatives, and chain-of-custody records  should be prepared according
to specific instructions before any samples are placed in the containers.  Bottles should
be washed and prepared  using the  appropriate EPA procedures,  and they  should be
labeled with a sample identification that is traceable to field  notebooks or to a master
                                        42

-------
sample-tracking log.  The unique sample identifications, sampling location, date and time
of sampling, sampler's name, tests to be performed on the sample, and the laboratory to
which the sample is sent should be recorded in the field notebook or master tracking log
should be . This process can be computerized in projects where computer capabilities are
available; the sample information can be entered into the electronic logbook and sample
labels and chain-of-custody records can be generated electronically.

QA  samples,  including trip blanks, equipment blanks,  duplicates, PE samples and
background samples, should be collected in the field.  Consideration should also be given
to the collection of additional sample volumes for laboratory duplicates and matrix  spike
analyses.  To provide an independent check of the primary laboratory, a percentage of the
samples can also be split and sent to two or more different laboratories.  A summary of
the types  and recommended frequencies for field QA samples is presented in Table 1.

On-site sampling operation audits should be conducted and documented by the project QA
coordinator to verify conformance with the QAPP and project  SOPs, and to ensure that
the field  notebooks, chain-of-custody, and  other documentation are complete.  The
representativeness of field and QA samples should also be monitored.

POST-SAMPLING ACTIVITIES

The first  post-sampling activities include the decontamination of clothing,  sampling
equipment and the outside  of sample containers.  The decontamination procedures are
outlined in the 40-hour hazardous waste training course.  Care should be taken to ensure
that the decontamination of the outside of sample containers does not  affect  sample
results.  For example, it would not be advisable to use benzene,  toluene, or other solvents
to clean the outside of a sample container holding sludge that will be analyzed for volatile
organic compounds.  All waste from the decontamination, including disposable clothing,
should be disposed of using approved hazardous waste procedures.

Before leaving the sampling location,  field  logbooks and other documents should  be
rechecked to ensure that the information recorded is complete.  Information that can  be
retrieved easily at this time may not be recoverable once the sampling crew has left the
site, or once  the site is altered in anyway after the sampling event (such as by  earth
moving equipment).

Samples that require cooling as part of the preservation should  be placed in cold storage
at 4°C as soon as possible after sample collection.  It  is important that containers  be
prepared adequately prior to shipment to ensure that the temperature is maintained until
the samples arrive at the laboratory.   Insulated coolers should be filled with  sample
containers and ice so as  to prevent breakage of samples  during shipment.   Individual
containers can be placed in plastic bags to avoid sample cross-contamination. Custody
seals should be placed on the cap or lid of each container and on the lid of the shipping
                                        43

-------
     TABLE 1.  RECOMMENDED TYPES AND FREQUENCIES FOR
               FIELD QUALITY ASSURANCE SAMPLES
    TYPE
      PURPOSE
   RECOMMENDED
    FREQUENCY
Field  (Equipment
Blank
Trip Blank
Decontamination
Blanks
Replicates
Splits
Field PE Samples
Measure background
level in sampling
equipment and
containers

Measures background
level from sample
transport and
containers

Measures background
level from field
decontamination
procedures

Measure variability
in results from
representative field
locations or
sampling procedures

Measure variability
in results between
laboratories

Measure accuracy of
overall measurement
system
 5%  or  one  per
 sampling event
One per  sample
shipment
One per each
decontamination
procedure
5% or one per
batch
5% or one per
sampling event
One set per
project
                             44

-------
container to ensure that samples are not tampered with during shipment.

Unless circumstances prevent it, the samples should be shipped the same day they are
taken, and should be received the following day at the laboratory.  This helps to ensure
that the samples reach the laboratory at the proper temperature, and that the laboratory
has ample  time to analyze the samples within the  holding and turnaround times.  If
samples are not shipped immediately after sample collection, they should be put in secure
storage areas that are accessible only by appropriate project personnel.  Samples should
never be left unattended.  Before samples are shipped, it is advisable to record the airbill
or shipping bill number on the chain-of-custody record. Chain-of-Custody, which  must
be completed before the sample is shipped, are placed in  a sealed  plastic bag inside of
the shipping container.  The chain-of-custody record should never be shipped separately.

CORRECTIVE ACTION AND DATA EVALUATION

Corrective  action should be taken  immediately  for any deficiencies noted  by a  field
sampling operations audit or field data evaluation.  However, unlike laboratory QC
samples, which are used as an ongoing control of the measurement process, field QA
samples can only be used for assessing the  quality of the data after  they  have  been
collected. Since this feedback is not immediate, the use of standard field QA procedures
is the best way to minimize errors that can result in  inadequate, biased, of invalid  data.
After the project is completed, the field documentation, results of field QA samples, and
reports of  sampling operations audits should be reviewed, along with the laboratory
documentation  and  results, to determine if the requirements originally outlined in the
QAPP have been met.  A summary of this  review should be included  in  any report
written about the study, and where  appropriate, the data should be qualified  to indicate
any
limitations  there may be in fulfilling the original purpose of the study.

ACKNOWLEDGEMENTS

This paper was previously published in the May/June 1993 issue of Environmental Testing
and Analysis.

REFERENCES

1.     U.S. EPA. Data Quality Objectives for Remedial Response Activities. PB 88-
       131370.  March 1987

2.     U. S. EPA. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods.
       SW-846, 3rd edition.

3.     U.S. EPA. Preparation of Soil Sampling Protocol Techniques and Strategies. PB
                                        45

-------
      83-206979. May 1983

4.    U.S. EPA. Guidance on Remedial Investigations Under CERCLA. EPA/540/G-
      85/002.  June 1985

5.    U.S. EPA. Interim Guidelines and Specifications for Preparing Quality Assurance
      Project Plans. QAMS-005/80. December 29, 1980.

6.    U.S. EPA.  RCRA Ground-Water Monitoring Enforcement Guidance.  August
      1985.

7.    U.S. EPA. Field Sampler Training Course Manual. PB86-217635.  May 1986.

8.    Kontopanos,  K.  N.  and Williams,  E.  S.  Quality Control in Field Sampling
      Methods.  Waste Testing and Quality Assurance:  Third Volume.  ASTM  STP
      1075. American Society for Testing and Materials. 1991.
                                    46

-------
AUTOMATED  REAL-TIME  PROJECT  LEVEL  LABORATORY   QUALITY
ASSURANCE

A.  Gladwell.  QA Director,  Curtis  &  Tompkins,  Ltd., 2495  Da  Vinci,  Irvine,
California,  92714; Kathy O'Brien, QA Director, Mark Schmick, Systems Manager,
Dr. C.  Bruce Godfrey,  President,  Curtis  & Tompkins, Ltd.,  2323 Fifth  Street,
Berkeley, California, 94710

ABSTRACT

Automated,   real-time   laboratory  data  management   systems  can be  used  in
conjunction with Data Quality Objectives  (DQO's) based on client needs, objectives
and historical site information  to design a relational  database management  system
that provides control beyond the typical analysis and quality control database.  With
appropriate  client/laboratory  communication and interaction,  such a  system will
provide clients with a complete project results database.

Curtis  & Tompkins has designed a real-time laboratory data management  system
that acquires  data directly from  instruments  and automatically checks data for batch
quality  control and calibration  as well as consistency with  historical project data.
This system  provides maximum automation  without compromising  data  security
yielding improved efficiency while minimizing response  time in correcting  errors.
The system provides Real-Time  Quality Control (RT-QC)  at these levels:

•Data Generation
          Calibration
          Batch Quality Control
          Completeness

•Monitoring
          Current sampling  event trends  and anomalies
          Historical  trends and anomalies
          RCRA statistical data reduction

Analytical  applications   of  this  system  can  also  be  automatically   performed.
Examples of this are anion/cation balances, ICP interfering  element  checks, BTEX
ratios (found vs. theoretical),  and carbon balance  (BOD/COD/TOC).
                                      47

-------
Such a system  has  benefits  not only for the laboratory  but for the  client  as well,
including:

•  Real-time feedback  on data anomalies,
•  Reduced costs of QC  compliance and data review,
•  Automatic  statistical data reduction and graphical reporting, and
•  Timely delivery  of complete and defensible monitoring  data.

This system results  in  unprecedented   improvements  in efficiency,  accuracy,  and
process  control to the  data  generation  phase of monitoring projects.  Consultants
using these  data  can be more  confident in data  quality  and historical  consistency.
They  are free  to concentrate  on higher  tasks  such as  trends  determination  and
environmental   risk  assessment.

INTRODUCTION

Why Real-Time Quality Control

There are many  sources of data associated with environmental  projects including
sampling, geological, geographical, analytical, meteorological,  analytical QA/QC  and
industrial hygiene, project data  quality objectives  and available historical data. Any
of the elements available  to laboratory  personnel can be entered into the database
management  system and used in producing the client or project results  database.

Using the Real-Time  Quality Control (RT-QC)  application  of  this system, we are
able to provide a detailed  review of summary and raw laboratory data including the
evaluation  of measurable  data quality parameters  such as holding times, instrument
calibration and batch quality control (including  blank,  laboratory  control  sample,
surrogate,  matrix  spike/matrix  spike duplicate  and  sample/sample   duplicate
information).  The  data are assessed  to ensure that  the  raw data do indeed  reflect
what has been  presented  in  the  summary  format (such  as  CLP-type forms).  The
overall goal being a complete report on the overall usability of  the laboratory data
specific  to a particular  job or  project.  The benefits of this automated  screening
system are tremendous.

   •      Diskette  deliverables  are evaluated for compliance with the measurable
          parameters
   •      Data  may be checked  quickly  and easily  to ensure  compliance  with
          project-specific  requirements
   •      A comprehensive  and thorough  data compliance  report  is generated
   •      Highly trained,  experienced personnel  are not required  to perform  this
          operation
                                       48

-------
There  are parameters which cannot be checked  using this system.  These include
pattern recognition  of chromatograms  (e.g.,petroleum  hydrocarbons and PCB's),
mass spectral  interpretation of TIC identity, ICP and AA spectral interferences,  and
masking interferences in the  case  of matrix interference  (e.g., with surrogates).
Therefore, the system cannot perform all functions and does have  some drawbacks
including:

    •     Complete  data validation cannot be performed
    •     Many laboratories  cannot produce diskette  deliverables
    •     Many data users do not understand the technology

By utilizing the system fully, a long-term monitoring application can be applied which
will evaluate  historical  data  in relation  to the  most  recent  sampling  event  and
provide trending data or anomalous data points.

Designing the System

Certainly, one of the core concepts driving C&T's LIMS project is the recognition
that the  screening  of laboratory  data  is  most effective  and  appropriate  when
performed at the point of data  generation.  However, this is only the first of many
decisions  in the LIMS  development  process.   Concerns  other  than data quality
control impact the  design, development  and implementation   of RT-QC.   The
generation of technically and legally defensible data requires the collection  of all the
data, whether or not it is to be reported to the client, and the future accessibility of
this data must be  assured.  A client may not want a data validation package today,
but  they may need  it in the  future and  all the information necessary to  generate
the  package   must be  retrievable.     Laboratory   size,  capital  available  for
improvements, software and hardware  systems currently in use, and the  complexity
of client requirements faced by the laboratory were our major considerations.  C&T's
experience is used as an example to illustrate some of these points.

System Requirements

RT-QC requires the availability of LIMS  throughout the lab at every instrument  and
workstation where  data  are  generated.   C&T has invested heavily in computer
hardware  and software to support our vision of easy data access throughout the  lab.
(See Figure 1). C&T's RT-QC system builds on  top of the data  sharing backbone,
but  only a few components are really  critical for a basic implementation  of Real-
Time QC.
                                        49

-------
Instrument  Interfacing:   there  must  be  a  means  of  gathering  analytical  data
electronically and "piping"it directly into the IIMS.  The majority of instruments at
C&T are  connected  to LIMS via serial, or "RS-232" lines.  Other instruments (gas
chromatographs,  for example) that are controlled by PC's send their data into LIMS
over the local area network.

Relational  Database Management  Software (RDBMS):  in our opinion a relational
database is crucial to flexibly manage  the large data  feeds generated by a typical lab.
C&T uses Oracle  but selection  of the RDBMS  should be  based on the actual
database  structures required by the laboratory, size of the facility, and  the  dollars
available  to allocate to such a system.

Multi-processing  Operating  System: the  basic  LIMS operating  system  must  be
capable of processing  many streams of data at once, in real-time.  A robust, multi-
user operating  system  such as UNIX or Novell Netware  should be considered.

Software  Toolkit: it is most  prudent  to  begin  by examining the LJMS presently  in
place in the laboratory to determine it's ability to proceed  with development  of RT-
QC.  C&T started with  a LIMS package  from Automated  Compliance  Systems,
which  provides  much  out-of-the-box LIMS functionality, as  well as  a  toolJkit for
customizing the software  according to the  needs of a particular  lab.  C&T RT-QC
system  was built  in-house  using this toolkit.

People: RT-QC  development requires   a high level of understanding  of RDBMS
technology in order to design the database  structures required  to efficiently screen
data and  provide immediate  feedback.

When  selecting a programmer (either as an in-house  hire or a contractor)  particular
attention  should be paid to how the  person  will fit  within the  culture  of the
laboratory.   The programmer  must be capable  of drawing  information  out  of
chemists and technicians  regarding the  processes they are running and introducing
new technology  and concepts.

To  maximize the success of any LIMS project,  you must get buy  in from the
analytical  staff to answer the  question, "What's in it for me?" They must be involved
with the project from the initial planning stages and should  clearly see the benefits
of its complete  implementation.  Adequate   training  of all staff will require  time and
resources  and must be included by management  in any development   schedule.
                                        50

-------
The C&T Scenario

C&T has  completely automated the procedure for completing  the screening of EPA
CLP metals data.  We selected this process for testing our RT-QC ideas because:
no  data  system  was in place, which  eliminated  the need  to determine  how to
integrate  it into RT-QC;  the CLP graphite furnace requirements  are  complex  and
confusing which often result in errors  requiring  rework  (and extensive analyst
training);  the short duration of the analysis means that immediate  feedback  will
result in timely correction and prevent generation of unacceptable  data; and we had
an  immediate need to improve the laboratory's ability to deliver the product in a
timely manner.

RT-QC has completely automated  the process of assuring that data which are  in the
final   results   database  meet  project-specific   requirements  (often  requiring
implementation  of CLP SOW).  The  data  are  handled  electronically  from  initial
generation,  and  deliverables  are  automatically  generated  to paper  or diskette.
Diskette results invariably match the paper report  because they are generated from
the same  database.  (See Figure 2.)

From the graphite  furnace analysts's standpoint,  RT-QC provides the answer  to the
question,  can I go  on?  It does this by reviewing the analytical result against all the
criteria specified in the CLP SOW; calibrations,  blanks, QC samples,  post digestion
spikes, duplicate  burns, and  all the other criteria for acceptable  data.  For example,
suppose  we've taken a GFAA reading  whose injection  RSD is > 20%, the  SOW
specified  limit.  About 5 seconds after the reading is taken, LJMS parses the data
file and reports  back to the  analyst that they will need to complete  another  analysis
of the sample (along with the reason why). The analyst completes the next run, to
which the LIMS responds again: this time, noting that  the data have been  marked
with  the  appropriate  CLP  RSD flag.  The  analyst  then continues  the analytical
sequence  using subsequent   feedback in the same  manner.

At  this point, the data are ready to report.  A full-screen  reporting  utility is used to
queue up and  generate  client-ready reports  in proper  CLP format.   Note that
nowhere  in this process were data  typed in by hand.  Note also that no runs had to
be  repeated unnecessarily; under RT-QC, the analyst knows immediately when QC
criteria are out of specification, and can adjust  or abort  the  run as needed.   And
only compliant data were stored in the final results database.
                                        51

-------
SUMMARY

Laboratory Benefits

The  benefits  of RT-QC  and associated  data  screening  and evaluation are reduced
turnaround time, increased  productivity (both through rapid feedback and  reducing
rework), and improved data quality.  Real-Time  Quality Control uses automation  to
complete  the  tasks which people find tedious and error prone  thereby producing
these benefits.

Although  RT-QC is rapid, thorough and many times more complete  than traditional
manual  method  of "after the fact" compliance  screening, it cannot  replace aspects of
laboratory  data  evaluation  which require "intelligent" review  such  as spectral and
chromatographic  interpretations.  These activities require analytical judgement based
on experience  which is most efficiently provided  by a chemist.  In an indirect manner
RT-QC  gives the  chemist  the  time necessary to do thorough  chromatographic
interpretation,  turning this  limitation into an  overall  advantage.

Client Benefits

Regulatory  requirements continue  to  drive  the quality improvement   process  in
environmental laboratories.   Commercial  environmental  laboratories  have another
incentive   we have to do it faster, cheaper, and better to remain competitive.   This
scenario provides a bright and challenging future for RT-QC.

The  CLP example presented  here is a practical place for a commercial  laboratory
to begin developing  RT-QC, but many possibilities  exist.   Projects which involve
more stable,  consistent sample  types  will  encounter  fewer  obstacles  in  the
development   stage.   It  may be advantageous  to  begin by providing  immediate
feedback  concerning  compliance  with  NPDES specified  methods.  The laboratory
could easily extend  the advantages of RT-QC  by developing  client, site, and sample
specific  requirements.    Data  could  be submitted  for internal  laboratory  QC
requirements  designed   in the same manner  as the  example  provided  above  but
tailored   to the   laboratory's  own  QA/QC   plan.    Data  which meet  laboratory
requirements could then be fed  into a statistics software  package and screened for
trending against historic data,  regulatory limits, or client supplied  data quality
objectives.  Reanalysis   of outliers could  be  initiated immediately  (literally  within
seconds of generating the outlier) which would reduce the possibility of reporting an
error, such as a false positive, to a client.  Client  reports  could automatically indicate
if regulatory  limits were exceeded.   The  possibilities for improving service  to the
client is tremendous.
                                        52

-------
RT-QC and the Future

At C&T we believe that RT-QC is a necessary step in meeting  the challenges of the
'90s for commercial  environmental  laboratories.   In  the  future  as more clients
develop project-specific data quality objectives, require electronic deliverables, and
push for ever more rapid turnaround times laboratories  must respond by automating
the data screening process.  The water pollution  control community  will continue to
feel detection  limit pressure  with ever increasing possibilities of false positives and
associated  costs.  Data trending against historic data points will become increasingly
important  to your clients.  Extensions of RT-QC will allow transfer of results directly
into statistical  analysis software  and provide immediate  feedback  to  the analyst
regarding  the status  of the  data  point.  Such advantages  are going to be key in
meeting the challenges of the future.
                                         53

-------
                   FIGURE 1.
LIMS NETWORK BLOCK DIAGRAM
                        INORGANICS LAB
       UNIX
       RDBMS
     Figure 1: Complex multi-operating system
     computer network used to collect and evaluate
     data from instruments and data systems.
                    54

-------
                       FIGURE 2.
   REAL TIME QUALITY CONTROL

                             DATA GENERATION
   ENTER QC
   SPECIFICATIONS
    LIMS
    RDBMS

     QCDATA
     SPECIFICATION
     TABLES
                "REALTIME" ^,
                QC SUMMAR7**
EVALUATE RESULTS
              AGAINST SPECIFICATIONS
                  (IDXL)
Figure 2: Process flow chart for real time QC. Data evaluated
against QC specifications enetered into structures on LIMS
file server are summarized directly to the analysts at their
workstation within seconds of data generation.
                         55

-------
   FIELD AUDITING PROCEDURES FOR EPA CONTRACTOR PERSONNEL

Greg Janiec. Quality Assurance Officer, and Owen B. Douglass Jr., Ph.D., CIH, Health
and Safety Officer, Zone I Technical Assistance Team Contract, Roy F. Weston, Inc.,
(WESTONo) West Chester, Pennsylvania  19380
ABSTRACT

Traditionally, the major focus  of quality assurance auditing in the hazardous waste
remediation field has been directed at laboratory monitoring and performance.  Field
sampling and data collection activities have been identified as significant components in
evaluating data validity and error.  A comprehensive field auditing program should
ensure that those who perform hazardous waste and environmental sampling are doing
so in accordance with established  Standard  Operating Procedures (SOPs),  Quality
Assurance Objectives and site-specific scopes of work.

A procedure has been developed for performing internal audits of contractor performance
at CERCLA Removal Actions within EPA Regions I through IV.  This procedure is
implemented through the use of a brief but thorough Site Audit Survey which directs the
auditor's attention to all critical phases of hazardous waste site operations support.  The
prime  focus  areas   include  sampling methods including  field screening  (hazard
categorization)  techniques,  use and calibration of direct-reading  field monitoring
instrumentation, site documentation,  site data management, compliance with health and
safety requirements,  and adherence to specific contract requirements.

INTRODUCTION

For  the  Zone  I  Technical  Assistance Team  (TAT)  contract,  WESTON and  its
subcontractors  employ over 200 multi-disciplinary environmental professionals in 11
office locations. These professionals provide technical and logistical support to the U.S.
Environmental Protection Agency (EPA) at CERCLA emergency response and hazardous
waste sites.  The  scope of work includes site assessment and evaluation, multi-media
environmental sampling (including biota), treatment system design and evaluation, air
monitoring, and general technical and engineering support within EPA Regions I through
IV, ERT and Headquarters.  All work assignments must be performed at the highest
levels of professional quality and in accordance with the  safety and health standards
established by OSHA. Periodic visits from the regional and program management staff
are necessary to ensure the quality of field work and client deliverables, and to identify
procedural refinements to augment the quality of the TAT services. Audits of field/site
operations are confined to the TAT contractor's scope of work and do not encompass any
aspect of the EPA's decision-making process or management techniques.
                                      56

-------
DISCUSSION

Detailed Program Goals

The site audit format was designed and implemented as a positive and constructive means
to 1) ensure adherence to EPA, TAT and WESTON quality assurance  policies and
procedures; 2) evaluate field screening and sampling capability and proficiency of field
personnel;  3) identify regional and/or program-wide deficiencies or trends; 4) exchange
innovative field techniques; and 5) obtain direct input from EPA On-Scene Coordinators
and Site Managers regarding TAT contractor performance.

Program Preparation

The TAT  auditing procedure required  a strong commitment of both program and
corporate resources.  Initially, auditing objectives and protocols had to be developed and
implemented which could evaluate and present an accurate picture of the proficiency and
capability of personnel assigned  to hazardous waste site operations and related support
activities.  WESTON chose to evaluate all aspects of field work, including technical,
health and safety and administrative aspects.  WESTON has embraced the concept of
Total Quality Management (TQM) and therefore structured its auditing procedures to
evaluate each component of a  project and examine its performance and relationship to
meeting the client's expectations.

Subsequently, qualified auditors were identified and  trained.  All participants were
required to complete a 40-hour Hazardous  Waste Site Worker Health and Safety Training
Program required by OSHA in 29 CFR 1910.120.  Completion of this training phase
permits the auditors to freely move about  hazardous waste sites to observe  and evaluate
worker compliance with specific sections of the site health and safety and QA sampling
plans.  Consistent with all other field personnel, the auditors are enrolled in a corporate
medical  monitoring program which evaluates and certifies each participant's physical
ability to participate in site activities including the ability to use respiratory  protective
devices.  Medical and respirator certifications enable the auditors to participate in site
entry, sampling, and monitoring operations located in the hot zone.

Auditing Procedures

The centerpiece of the current site auditing program is the Site Audit Survey Form
(Figures 1-1, 1-2 and 1-3).   This form provides a  framework for the site auditor
including a brief listing  of areas  which  should  be examined during  the auditing
procedure. The following protocols are used as guidance when performing a site audit:
                                        57

-------
                                                                      ZONE I
                                                       TECHNICAL ASSISTANCE TEAM (TAT)
                                                                SITE AUDIT SURVEY
              SITE:
        LOCATION:
       SITE PHONE:
              OSC:
                                                                             REGION:
                                                                          SITE LEAD:
                                                                       TATM ONSITE:
in
09
              EVALUATION CRITERIA

SCOPE OF WORK
 TDD w present in file
 TAT is performing tasks requested

SITE DOCUMENTATION
 Site log is in proper format
 Site files are organized
 POLREPs are factual and regular
 Personnel sign-in sheets are present
 Site entry logs are maintained
 Photographs are labeled and orderly
 Off-site TAT costs are documented

COST TRACKING
 TAT costs are current
 Project costs are current
 RCMS is being used

EQUIPMENT USAGE
 TAT equipment is controlled
 TAT personnel are trained in operation and
    maintenance of monitoring equipment
 Instrumentation calibration and maintenance is
    documented
                                                                                       COMMENTS
                   Y=Yes
                           N=No
M=Most of the Time
O=Occasionally     N=Never    N/A=Not Applicable

-------
               EVALUATION CRITERIA                                                 COMMENTS

SITE MANAGEMENT (TAT)
  Site lead is designated                                	      	
  Individual duties ate assigned                         	      	
  Morning safety meetings are held                      	      	

TECHNICAL INPUT
  TAT is monitoring removal                           	      	
  TAT assisted with scope of work                      	      	
  TAT performed spill containment                      	      	
  Innovative treatment technologies were evaluated        	      	
  Disposal options were evaluated                       	      	

SAMPLING AND MONITORING
  Sampling plan is current                             	      	
  Applicable SOPs are present and followed              	      	
  Sampling equipment is decontaminated properly         		
  Field screening is used appropriately                   	      	
  Air monitoring is conducted as per SOPs               	      	
  Analytical Data is present                            	      	
     Validation conducted by                                          	
     Funding source  utilized                                          	

HEALTH AND SAFETY
  Safety plan is complete  to date and signed              	      	
  Directions to the hospital are present                   	      	
  Emergency phone numbers are posted                  		
  Site Emergency Plan exists                           	      	
  Safety Plan has been implemented                     	      	
  Site is controlled                                   	      	
  Decontamination systems are in place and
     effective                                        	      	
  Overall safety precautions are in effect                 	      	
C:\WP31\DOCSl»OSC_DOC«IKBaTAT.aTB.AXID
               Y=Yes       N=No        M= Most of the Time        O= Occasionally      N=Never    N/A=Not Applicable



                                                                              1~Z

-------
                                                                                                                             Page 3 of 3




                                                     OVERALL PERFORMANCE COMMENTS
     Describe Site Conditions:
     Describe Activities Observed:
o>
o
     EPA Performance Comments:
     Other Comments:
                                      TAT AUDITORS:     	               DATE:



                                                         	               DATE:




                                                                 FIGURE 1-3

-------
1) If an audit  is to be conducted  by the TAT Zone Program Management  Office
(ZPMO), a member of the ZPMO contacts the appropriate TAT regional office manager
to schedule a mutually agreeable date to perform the site audit. The office manager may
designate himself/herself or a senior member of the regional  management team to
accompany the  ZPMO representative to the site on the established date.

2) If the audit is to be performed solely by the regional TAT office, the office manager
or selected senior regional management member is responsible for the completion of the
audit and corresponding reporting procedure.

3) The EPA On-Scene Coordinator or  Site Manager is  notified by  .the TAT office
manager of the pending audit at least 24 hours in advance. At this time, the EPA site
manager may be requested not to inform the on-site contractor personnel of the upcoming
audit.

4) Upon arriving on site, the auditor identifies himself/herself to the EPA and contractor
personnel, produces an identification and health and safety certification card, completes
the site sign-in sheet, reads/acknowledges the  site safety  plan, and briefs appropriate
personnel on the purpose and procedures associated with the audit process.

S) The auditor evaluates site activities  and records in  a manner to minimize the
interference with normal site activities. The auditor uses the Site Audit Survey Form to
evaluate the following major areas as applicable to the site scope of work:

        •  Work assignment documentation
        •  Field and personal log books including supporting documentation
        •  Site QA sampling plan
        •  Site health and safety plan including amendments
        •  Equipment (monitoring instrumentation) use proficiency of field personnel
        •  Field calibration and maintenance logs for all field instrumentation
        •  Storage and maintenance of equipment,
        •  Site reports and client work product deliverables
6) A hot zone entry may be made by the auditor to observe site operations (e.g., soil
sampling) and ensure adherence to relevant sections of the site scope of work, QA
sampling plan and established EPA and TAT standard operating procedures. If properly
trained, the auditor may participate in the respective site activity as a team member to
more fully appreciate the actual working conditions and related operational and weather-
related problems.

7) During a normal break period  or  after completion of the daily site activities,  the
auditor  conducts  a  briefing with the contractor site lead to identify areas in need of
                                        61

-------
improvement; establishes a  corrective  actions plan to improve  site operations; and
reviews relevant program policies or procedures.

8)  At the EPA Site Manager's convenience, the auditor solicits a candid evaluation of
general contractor performance from the EPA manager. The auditor briefs the EPA Site
Manager regarding the identified items to improve site operations and any technical
advice.

9) The auditor subsequently drafts  a  report using a standardized outline (Figure 1-4)
identifying the audit findings and corrective suggestions.  This is  sent to the EPA Site
Manager and TAT regional manager for review and comment. After a ten day review
period, the auditor initiates the final site report. These reports are then distributed to the
TAT Zone Program Manager (ZPM),  regional EPA Deputy Project Officer (DPO), and
EPA Project Officer (PO) in Washington, DC.

Audit Frequency

The target frequency for regional field audits is three per quarter per each region.  The
ZPMO goal is to conduct three audits per quarter across the Zone.

SUMMARY

The current TAT site audit program has been in effect since June of 1989. Informal site
auditing was performed using non-standardized formats prior to this time. A concerted
effort  was  made to keep the Audit Survey Form brief despite the tendency to  make
continued revisions and  lengthen the  document.   Finalized audit  reports are typically
from two to three pages in length. The simplicity of this auditing program is one of the
primary causes for its longevity and acceptance by contractor and  client  personnel.
Although  the paperwork and personnel  time commitments  are  minimal,  this  audit
program has demonstrated its ability to maintain and improve the delivery of high quality
services on a consistent basis. In terms of cost/benefit analysis, the return is significantly
greater than the cost.

This auditing program has achieved numerous positive effects within the Zone I TAT
contract including 1) the on-site evaluation of program field personnel; 2)  the candid and
timely feedback regarding contractor performance from the on-site EPA client;  3) the
evaluation of the effectiveness and practicality of current standard operating procedures;
and 4) the identification of regional or program-wide opportunities to improve training
programs, purchase new field monitoring/support equipment, or revise standard operating
procedures  to accommodate new technologies.

Unanticipated benefits  regarding this project are substantial.  The on-site nature of this
audit program requires the "hands-on" involvement of management, quality assurance and
                                       62

-------
                                Figure 1-4


                      SITE AUDIT REPORT OUTLINE
A. Introduction
      1.    Site Name
      2.    Location
      3.    Date of Audit
      4.    Auditors
      5.    Time Spent on site

B.    Observations of Activities

      1.    Review assignment scope of work
      2.    Documentation
      3.    Equipment calibration, maintenance and use
      4.    Site management
      5.    Technical support
      6.    Health and Safety

C.    Recommendations for Improvement
                                     63

-------
health and safety personnel. These individuals, many of whom had not performed actual
site work for one or two years, regained their "real world" perspective on field work.
This experience is extremely valuable  to  ensure  that new or revised policies  and
procedures have a positive impact on field operations.

The  technology transfer benefit of this auditing  program  also  proved  to be very
beneficial.  The auditors from the ZPMO were able to observe state-of-the-art techniques
and innovative approaches to resolving unique site problems in a real world situation.
Additionally, the auditors were able to participate in the assessment of the success or
shortcomings of each application. Bimonthly program management meetings served as
the forum to present these first hand observations to all the regional TAT office managers
to augment their capabilities to serve their clients.

Another unplanned benefit was a marked improvement1 in the morale of field personnel
including EPA  Site Managers.  The on-site presence of senior regional or  program
management auditors illustrated a commitment to field work and the front-line personnel.
The site audit visit proved to be an excellent opportunity for field personnel to candidly
interact with management to discuss issues and ideas in a non-office setting.

In conclusion,  WESTON's audit program  has  been  extremely effective  during  the
execution of the Zone I TAT contract. It is our belief that this program can serve as a
model for other hazardous waste site operations contracts.
                                       64

-------
10                  THE FIELD CHEMICAL DATA ACQUISITION PLAN:
                                A DATA QUALITY OBJECTIVES
         APPROACH FOR THE USE OF REAL-TIME FIELD ANALYTICAL DATA IN
                                ENVIRONMENTAL PROGRAMS

    Raymond J. Bath Ph.D., Waste Policy Institute, Germantown, MD 20874; Richard D. Spear Ph.D. United States
    Environmental Protection Agency, Region 2, Edison, NJ 08837; Daniel Lillian Ph.D, United States Department of
    Energy, Environmental Restoration and Waste Management, Germantown, MD 20874; and Pamela Greenlaw, Lockheed
    Environmental Systems & Technologies Co. Edison, NJ 08837

    ABSTRACT:

    The acquisition and use of data from on site real-time sample analyses has conceptually always
    been the preferred approach for environmental programs. An examination of some field programs
    has shown that a fUll capability field analytical service  yields a variety of project  benefits
    including major decreases in project time; substantial cost savings in analytical services; increases
    in  data quality;  and decreases in health  and safety concerns for sampling crews  and nearby
    residents. Specialized vehicles for field analyses, e.g., vans and other mobile facilities, have been
    developed and  manufactured.  A variety of high quality portable and transportable analytical
    instruments have been developed  to analyze  soil, water, and  air. Formalized research and
    developmental analytical programs exist in sensor technologies for industry and government. The
    question then  arises,  why are such field analytical services  not routinely used?   A possible
    explanation is in the lack of an up front planning guide for the acquisition of field chemical data.
    While the current data quality objective (DQO) guidelines present the up front planning framework
    for the entire remediation work plan process, the sampling and analysis plan has not folly explored
    the use of real-time field data.  This oversight has restricted field analytical services. With the
    addition of a field chemical data acquisition plan to the work plan process, real-time field analytical
    data acquisition can assist in strearnling the investigation  and remediation process.  This paper
    examines the state of environmental field analytical  services today for the collection of real-time
    data  and  presents  an approach using  the  current DQO  methodology for preparing  and
    implementation of a field data acquisition plan as part of the sampling and analysis plan.

    This paper is the opinion of the authors. It does not reflect the current position/policy of the USEPA or USDOE. Any
    manufacturers noted are not recommendations of equipment but are examples of current equipment trends.
                                               65

-------
•J 1                            A NEW COST EFFECTIVE STRATEGY
                        FOR ENVIRONMENTAL REPORTING SOFTWARE


       Leigh  A.  Richardson, Director of Technical Services, and E.  Alan Haley. Director  of Marketing,
       TELECATION, INC., 19423 N. Turkey Creek, Suite C, Morrison, CO 80465 (303) 697-8080
       ABSTRACT

       The prices for CLP reporting software have skyrocketed in the past year, reflecting the costs incurred by
       software vendors in attempting to stay current with the changing requirements and screening standards
       for CLP deliverables.  The higher prices have put a new burden on environmental analysis laboratories,
       who are required to produce CLP and CLP-like computer generated reports. In an attempt to control the
       costs of environmental reporting software and reduce this burden on laboratories, TELECATION, INC.
       has spent the last  year developing a new "software  system",  which takes advantage  of efficient
       programming strategy to produce a variety of environmental deliverables at far less cost.  In addition to
       the cost savings, the reporting options are more flexible, allowing both vendor and user to readily make
       sophisticated  changes to address special variations  in requirements.   This approach has been  used to
       generate reports modeled after the requirements of both CLP and RCRA/SW846, and other protocols are
       being examined.  This paper will describe  the techniques used to deliver both cost effectiveness and
       versatile software application for environmental reporting.
       INTRODUCTION

       The cost of computer software is controlled roughly by the ratio of the complexity of development to the
       market size.  The market size for laboratory software applications is infinitesimally small compared to the
       general  business software market place.  This, along with the  fact  that many of the requirements of
       laboratory software make  it inherently complex to develop, serve to inflate the final price of software
       designed specifically for laboratory use.  The size of the market is a factor which cannot be controlled by
       a software vendor.  However, the complexity of development can be controlled by applying large scale
       efficiencies in the process of software development.

       Most laboratory software applications fall into the general categories of:

           DATABASE APPLICATION - for data storage, retrieval, report generation, or;

           SPREADSHEET APPLICATION - for calculations, data relationships, and graph generation

       Since these are readily available tools, laboratories have always had  the option  of addressing their
       software needs  by  using  one or more of the  numerous,  low  cost  generic  database and spreadsheet
       products on the market and configuring the tools around each required application.  Many laboratories,
       however, are not staffed with personnel who have the background, training, or time to configure and/or
       program the general purpose tools into a final  application solution.   Further, for laboratories  who do
       have the programming resources, the additional costs of applying these resources to basic laboratory
       software needs largely offset the original cost savings realized in the general purpose software purchase.
                                                     66

-------
A software vender who develops turnkey software solutions specifically for the laboratory market place
must also incur these initial development costs. However, unlike the laboratory developing their own
software, the software vendor can amortize the initial development cost over the entire client base for the
product, providing a turnkey solution to a very specific task at a lower cost than  it can be developed
in-house, even if the programming resources exist.  Even so, if each software application is developed
from the "ground up"  as a stand-alone  product, the initial development cost for that product must be
absorbed, only by the purchasers of that individual product.  The next application will undergo a similar
level of development, adding to the cost of that product, which again must be ultimately borne by the
purchasers of that single product.  The challenge for the software vendor seeking to reduce the prices of
finished laboratory products is to address the repetitive nature of overhead development costs associated
with each individual product.
A STRATEGY FOR PROGRAMMING EFFICIENCY

Approximately one year ago, TELECATION, INC., a company specializing in PC-based  software for
the laboratory,  initiated a  study  into  ways to dramatically  increase the efficiencies  of application
programming with a corresponding reduction in development time and final software price.  In doing so,
TELECATION systematically analyzed the software needs of the laboratory community, looking for the
software functions  common to many different laboratory  applications.  The  software  company then
embarked on a project to develop a base product which would contain most, if not all, of the software
functions needed to implement almost any laboratory application.  Additionally, the base  product was
specified to contain a variety of tools to give the user control over such things as database design, report
definition,  formula  definition, etc.  The goal for this product, in and unto itself,  was that it should be the
single-most powerful  and  flexible  tool  ever  designed  to specifically address laboratory software
applications.

This product was to be different from the generic database and spreadsheet products on  the market, in
that it was to directly  address specific laboratory needs.  The system was designed to address general
purpose laboratory  data handling with little or no configuration by the final user of the product.  Special
laboratory-specific  functions, like analytical  quality control procedures, instrument interfacing, even
export to LIMS, were part of the basic product.

To allow for extension of the capabilities of this product beyond the basic design, tools were provided for
extensive tailoring  for any specific task.  A  typical computer-oriented chemist could use the  tools to
automate very specific  laboratory tasks, which could be routinely executed by anyone in the laboratory.

While user configurability of the system  was  to be a major  feature of the  software product,
self-configuration of software solutions is, as previously stated, not necessarily in the best interest of
every  laboratory.   Therefore,  TELECATION expanded its  strategy  for the "universal"  laboratory
software tool to include its own, turnkeyed applications, running on the base software product. By using
the configuration tools designed into the basic system, TELECATION did not have to design, code, test,
validate and implement the basic functions, for which these processes had been  previously conducted on
the base product. In other words,  the development overhead incurred in the base product,  did not have
to be repeated for each and every software solution operating on the basic system.  TELECATION's goal
of providing complete  and specific software applications for the laboratory at breakthrough prices was,
therefore,  realized  by  implementing a  systematic strategy for future software  development, which
dramatically reduces the product development overhead.

A block diagram of the concept is illustrated in Figure 1.
                                                67

-------
                     INSTRUMENT
                         DATA
KEYBOARD
 ENTRIES
                                          Input
                                  Basic laboratory tool
                               "DATA MANAGER 2000"
                                         Output
1
GRAPHS
CHARTS

1
REPORTS

1
USER
APPLICATIONS

1
"PLUG-IN"
LABORATORY
APPLICATION
"2000 Series"

1
LIMS
 Figure 1.  An Efficient Strategy for Laboratory Software Development
 PLUG-IN APPLICATIONS FOR A GENERIC LABORATORY TOOL

 The "base product" referred to above is called  "DATA MANAGER 2000".  The add-on applications,
 each of which takes advantage of the software development which went into DATA MANAGER 2000,
 are collectively referred to as  "2000 Series Applications."  All "2000 Series" applications literally "plug
 in" to DATA MANAGER 2000 through a series of program "hooks" designed  to send and/or  receive
 data from special data handling modules. The  magnitude of effort which is required for each  "2000
 Series" application varies from literally  a few hours for a simple application-specific database (like the
 Chemical Inventory Database) to many weeks for a more complicated application (CLP-like reporting),
 which may involve the automation of many relationships, calculations, and report formats.  In all cases,
 the prices for "2000 Series" applications are far less than software products developed from the  ground
 up with a specific purpose in mind.
A side by side comparison of software prices can be made in the case of CLP report generation software.
TELECATION, INC.  offers software  products (the "ENVIROFORMS"  ®  product  line)  designed
specifically to automate as much as possible the deliverable data requirements of the USEPA's Contract
Laboratory Program  (CLP).  While  the  "2000 Series"  products  do  not address all  EPA-specific
requirements with  the same rigor  as the "ENVIROFORMS"  products,  they provide an environmental
deliverable which  will meet the needs  of most laboratories  at a cost  savings of approximately 85%
compared to the CLP-specific alternative.   Indeed,  some of the price difference is associated with  the
development overhead associated with meeting EPA details  in the "ENVIROFORMS" products,  but
much  of the difference accrues from  the programming efficiencies associated  with the "2000 Series"
                                             68

-------
concept.  (NOTE: TELECATION, INC. continues to offer and update the "ENVIROFORMS" software
products for strict CLP applications, as well as the "2000 Series" products for those desiring to produce
CLP-like deliverable at reduced prices.)

Table I lists  the "2000 Series" applications currently available or under development.   Given the
company's  extensive history with environmental software and the variety of environmental reporting
protocols which  have  been promulgated, it is not surprising that many of these  packages address
environmental issues. However, of the current and planned applications, many address other laboratory
tasks, as well.  Specifically, a series of ISO 9000 and "Good Automated Laboratory Practices'1 (GALP)
oriented tools are currently available and many more are under development.
                                          Table I


             EXAMPLES OF "2000 SERIES" PLUG-IN SOFTWARE APPLICATIONS

        o   CLP-like reporting for Inorganics, VOA, SNA, PEST

        o   RCRA/SW846 reporting for Inorganics, VOA, BNA, PEST

        o   Chemical Inventory System

        o   Hazardous Waste Disposal Tracking System

        o   Detection Limit Calculator & Reporter

        0   SOP Filing System

        o   Personnel Training Log

        o   R-Bar Chart

        o   Standards/Reference Material Log

        o   Computer Hardware/Revision Log

        0   Instrument Usage, Calibration & Maintenance Log

        o   Raw Data Archive Log

        o   Multi-Phasic TCLP/Inorganic Reporting System

        0   Storm Water Discharge Monitoring
                                             69

-------
 ADDITIONAL BENEFITS OF THE STRATEGY

 It was previously stated that market size, in addition to development complexity, plays a significant role
 in software pricing.  In a very real sense, the product strategy described above increases the market size
 while reducing development complexity.  Since each plug-in application uses the same base product, the
 development costs of the base  product can be amortized over the market for all compatible products.
 which is  significantly  larger than for any  single product.   In this way, the development  strategy is
 controlling both factors which affect software prices.

 Price is not the only benefit from the strategy, however.  Personnel training is  a major concern and
 expense for a laboratory.  In today's laboratory,  a major  portion of  the training  dollar goes toward
 training software operators.   Since all applications in the "2000 Series" product line operate through a
 common  base  product, the learning process is greatly  simplified.  Once  a user learns to  use one
 application, they already have  the background to use any  other.  The strategy  thereby maximizes a
 laboratory's return on personnel training, by applying  the experience gained with one application to all
 other system-compatible applications.
SUMMARY

Software prices for laboratory applications have traditionally been high due the small market size and
complexity  of development.   Software  prices  can be  significantly  reduced  through a  software
development strategy based on plug-in, application-specific routines which depend on the functions of a
common base  product  for operation.   In this way, the  development effort which went  into the base
product can be used over and over again, thereby reducing the effort required to develop each compatible
product.  Additionally,  the  costs of the basic product can be  amortized over  the  market size  for aH
compatible products, which is significantly larger than that for any single product.
                                              70

-------
^ 2                   Local Area Network Data Validation Software System
      A.  Davis.  Quality Assurance Manager, BCM Engineers Inc.,  One Plymouth Meeting,
      Plymouth Meeting, Pennsylvania,  19462,  and J, Vernon, Approach,  Inc., 100 Summit
      Lake Drive Valhalla, New York, 19595.

      ABSTRACT

      The extensive data validation procedures required tjo review Contract Laboratory Program
      (CLP) data packages  generated for both regulatory  and private parties  environmental
      investigations are  extremely labor intensive and usually  do  not consider previous
      investigation data.   To  address these concerns, we  have developed  DataVal,  a data
      validation software that reduces the review effort and enables the validators to access the
      full investigative database.  DataVal translates CLP diskette  deliverable formats (DDF)
      and automates  validation procedures.  DataVal  enables the  validation  specialists to
      concentrate their review efforts in the  interpretive areas of the data package.  Using this
      software's  capabilities  to access the database generated over the entire  course of a large
      investigation, the validation specialists and data users can better interpret the viability and
      usability of the analytical program results.

      The DataVal software is a Client/Server application implemented on Windows™ based
      personal computers accessing a Relational Data Base Management System (RDBMS) that
      is a part of a Local Area Network (LAN).  The Client component was developed using
      Visual Basic™ and the RDBMS component chosen was SQL  Server. The  software itself
      is modular in design, permitting the validators the flexibility to  choose a "path" through
      the validation protocols.  These "paths"  allow assessment of complex  data relationships
       and hierarchical review.   The modular  software design lets the validator set separate
       criteria for regional data validation modifications or set criteria for previous or future data
      validation  requirements.  The ability to  access the previously validated database in the
       SQL Server allows validators to evaluate individual data sets in relation to the historical
       data.

       Our company's  project management system  allows the transfer of laboratory information
       management system (LEVIS) data directly  into  a larger project management historical
       database.  DataVal is being integrated into this project management system.   Upon
       completion of data validation, the data  qualifiers are electronically transferred into the
       project management historical  database,  ensuring consistency of data qualification.  The
       integration of DataVal into the project  management data system combines a powerful data
       interpretation tool with a high capacity database management system. The data users are
       supplied with a a  project  database that  is  better evaluated, cost effective,  and highly
       accurate.
                                               71

-------
4 Q              QUALITY ASSURANCE AND QUALITY CONTROL
 10                  MEASURES FOR THE DETERMINATION
             OF  HEXAVALENT CHROMIUM  FROM  STATIONARY  SOURCES

              Thomas  Dux. Quality Assurance  Coordinator,
               K. Hall, G.  Jungclaus, and A. Williams,
             Midwest  Research  Institute,  425 Volker  Blvd,
                      Kansas City, Missouri  64110

     ABSTRACT

     Midwest Research Institute has been collecting and analyzing
     stack gas samples for hexavalent chromium for the last 4
     years utilizing the method from the Methods Manual for
     Compliance with BIF Regulations (EPA/530-SW-91-010).   MRI
     has enhanced the methodology with the addition of matrix
     spikes, audit samples,  reporting limit procedures, and
     validation of the integrity of sample preservation.  This
     paper discusses these enhancements and associated data
     quality objectives plus the results from quality control
     samples.

     In addition to performing hexavalent chromium analyses, MRI
     has also reviewed numerous trial burn plans and trial burn
     results under contract to the EPA.  Through this experience,
     MRI has found that some firms conducting trial burns do not
     fully characterize the sample results in terms of the
     quality of the data.   Problems include the contamination of
     collection reagent with hexavalent chromium,  the lack of key
     information on the integrity of the collection media
     following sample collection and the lack of precision and
     accuracy data for the analysis results.  These difficulties
     have resulted in trial burn plans which were deficient and
     trial burn results which could not be validated.  This paper
     also addresses these issues.

     1.0  INTRODUCTION AND BACKGROUND

     1.1  Introduction

     Midwest Research Institute (MRI) has been collecting and
     analyzing stack gas samples for hexavalent chromium for the
     last 4 years utilizing the method from the Methods Manual
     for Compliance with BIF Regulations (EPA/530-SW-91-010).  In
     addition to performing hexavalent chromium analyses,  MRI has
     also reviewed numerous trial burn plans and trial burn
     results under contract to the EPA.  Through this experience,
     MRI has found that the method as written does not fully
     characterize the quality of sample results in regards to the
                                   72

-------
accuracy of the determinations, the detection limit and the
validation of sampling activities.  These difficulties have
resulted in trial burn plans which were deficient and/or
trial burn results which could not be validated due to a
lack of information.

1.2  Background

Besides the typical Method 5 quality control (QC) procedures
(e.g. calibrations, leak checks, etc.) the method for
hexavalent chromium has 5 major quality control procedures:

1.   Sampling Train Validation:  Hexavalent chromium is an
     analyte which is very subject to reduction since it is
     a good oxidizer.  To prevent loss of Cr+6 the pH of
     sampling train media (aqueous potassium hydroxide) must
     be maintained above 8.5, the train must be purged with
     nitrogen following sampling and the contents of the
     impingers must be filtered immediately following
     recovery.

2.   Instrument Calibration;  The instrument must be
     calibrated with at least four standards over one order
     of magnitude.  A curve is run before and after samples
     and the relative percent difference (RPD)  of the
     instrument response for the two analyses must be less
     than 10 %.  In addition, the calibration standards must
     exhibit a relative accuracy of 93 to 107 % when
     calculated versus the linear regression curve.

3.   Duplicate Analysis;  Each sample must be analyzed in
     duplicate and the RPD of the instrument response must
     be less than 10 %.  If RPD is not less than 10 %,  the
     samples are reanalyzed until this criteria can be met.

4.   Audit Sample;  An EPA audit sample must be analyzed as
     an independent check on the accuracy of the
     determinations.

5.   Blanks;  Field blanks must be analyzed and can be used
     to correct emission data for blank contributions.

The method does not address the following three quality
control subjects:

1.   Sample Holding Times;  Holding times are not given in
     the method.  The only applicable RCRA holding time
     currently available indicates a 24 hour holding time
     (Chapter 2, SW-846, Revision 2, 11/90), but does not
                              73

-------
     require preservation at a pH above  8.5.

 2.   Accuracy:  The method does not require any
     determination of the accuracy of the analysis  as  it
     pertains to the stack gas matrix (i.e. spikes).

 3.   Detection Limits;  The method does  not address
     detection limits or give a procedure for determining
     them.

 The remainder of this paper will discuss some results  for
 the method specified QC procedures and MRI's additional  QC
 measures.
2.0  QUALITY CONTROL PROCEDURES AND RESULTS

2.1  Holding Times and Validation of Sampling

As previously indicated, holding times are not specified  in
the BIF method.  The most current available RCRA holding
time is 24 hours (Chapter 2 SW-846).   However, most firms
use a 14 or 28 day holding time with preservation requiring
a pH greater than 8.5 and a temperature of 4 °C.   The
EPA/ASTM method for hexavalent chromium in water indicates a
holding time of 24 hours with preservation at pH 9 to 9.5
and a temperature of 4 °C ("Determination  of  Dissolved
Hexavalent Chromium in Drinking Water, Ground Water and
Industrial Wastewater Effluents by Ion Chromatography:
Collaborative Study"  K. Edgell, et.  al.,  Bionetics
Corporation, April 6, 1992.)   This document also indicates
that dilute standards should be prepared daily but that
stock standard solutions are stable for at least 3 months.

In planning trial burns or stack testing,  it is best to be
conservative and use the minimum acceptable holding time.
Most EPA regions and states will approve a 14 to 28 day
holding time with preservation at a pH greater than 8.5 and
a temperature of 4 °C.   The MRI  field  sampling and  analysis
team uses a 14 day holding time and preservation as
indicated.   However,  when sampling is not performed by MRI,
the laboratory meets a 14 day holding time from laboratory
receipt.

During reviews of trial burn plans and reports, it has been
noted that  some researchers are not aware of holding times
and preservation considerations for these samples.   The
following table gives results for the same samples, from the
same stack,  analyzed 6 and 19 days from receipt.
                              74

-------
Table 1 * Results for Hexavalent Chromium Determinations in
          Stack Samples Analyzed 6 and 19 Days from
          Laboratory Receipt
Sample
1
2
3
4
5
pH
14
14
14
13
13
Result
(ug/L) 6
Days From
Receipt
7.8
10.6
3.7
2.3
15.3
Result
(ug/L) 19
Days From
Receipt
3.4
3.0
2.0
0.74
8.1
% Loss of
Analyte
From 6 to
19 Days
56
72
46
68
47
These results are only indicative of hexavalent chromium
stability in impingers from that specific stack.  It is
interesting to note that the average loss is 58 % with a
standard deviation of 12 indicating good agreement for the
stability data despite the varying concentrations of the
samples.  These data clearly indicate that holding times are
an issue and sample results can drop within the time frame
accepted by regulatory agencies.

The BIF method has three validation steps which are specific
to Cr+6  (as opposed to Method 5 requirements).  These are a
nitrogen purge of the train following sampling, the
filtering of the samples following train takedown and each
impinger must have a pH greater than 8.5 at the end of an
analysis run.  Many Quality Assurance Project Plans (QAPjP)
for trial burns give descriptions of the sampling procedures
but do not address these guality control measures.  They
should be discussed in the QAPjP and compliance must be
documented during sampling to validate the collection and
preservation of the samples.

Of critical importance is the pH of the impinger solutions.
Many trial burn plans and reports do not address this topic.
Some field sampling crews have adopted the practice of
testing the stack before the trial burn and increasing the
normality of the potassium hydroxide solution if needed for
highly acidic stack gases to maintain a high pH.  In
addition, a significant number of trial burn reports and
associated raw data do not have the pH documented in the
field at the conclusion of sampling or upon receipt at the
laboratory.  Without evidence of a pH greater than 8.5, the
                              75

-------
hexavalent chromium results cannot be validated.

Some permit applicants challenge the rejection of hexavalent
chromium results in the absence of a pH check, and insist
that the check was done even if it was not documented.  MRI
tracks the pH of samples received for hexavalent chromium
analysis and in the past year about 10 % of the samples had
a pH less than 8.5.  The pH ranged from 5 to 7, and only one
of these low pH samples had Cr*6 above the detection limit.
(All samples collected by MRI personnel met the pH
requirement.)  Without documentation of sample preservation,
the sample results can not be validated.

MRI measures the pH of all Cr*6 samples before analysis and
reports the pH with the sample results in a manner similar
to inorganic, volatile and cyanide samples which also
require pH adjustment for preservation.  Accreditation
organizations, customers and regulators should demand that
sampling firms document the pH of the impingers in the field
and the pH should also be verified by the laboratory.
Laboratories should not report Cr*6 results without also
reporting the pH of the sample.

2.2  Precision Determinations

The method requires that each sample be analyzed in
duplicate and that the RPD of the instrument response must
be less than  10 %.  If RPD is not less than 10 %, the
samples are reanalyzed until this criterion can be met.
There are three reasons why this is not a very practical way
to handle precision.  First, this requirement does not take
into account the decrease in precision as the response
approaches the instrument detection limit.  Second, without
a technical reason, the first analyses that do not meet 10 %
RPD cannot be validly rejected and should be included in an
average result.   Third, this requirement assumes manual
injection allowing for a repeat of analysis until the
criteria is met.  This is labor intensive and not practical
given the availability of autosamplers.

MRI usually uses an autosampler for unattended,  overnight
analysis and injects all samples in duplicate.  The
precision criteria is rarely exceeded (<5 % of all
analyses).   In the cases where there are problems, it is
usually due to the response being at the detection limit or
there is significant interference from the stack gas matrix
which is evident in the chromatogram (e.g. extraneous
peaks.)   If manual injections are made, the RPD is tracked
and if samples do not meet the criterion it is analyzed a
                              76

-------
third time and the average of all three analyses is
reported.

In reviewing trial burn data for various EPA regions and
states, MRI has noted that precision data are often not
reported with trial burn results.  Some laboratory reports
do not address precision at all and others indicate only
single analysis of each sample.  These samples do not have
any preparation beside filtering; therefore duplicate
analysis is the only measurement of precision.  The
chromatograms are relatively short (<10 minutes) and
duplicate analysis is not a burden.  Permit writers need to
be aware of this issue in reviewing trial burn plans and
results.

2.3  Instrument Calibration

The method requires that the instrument be calibrated with
at least four standards over a concentration range of one
order of magnitude.  All standards are run before and after
samples and the relative percent difference (RPD) of the
instrument response for the two analyses must be less than
10 %.  In addition, the average of the calibration
standards, analyzed before and after samples, must exhibit a
relative accuracy of 93 to 107 % when calculated versus the
linear regression curve.  The calibration curve is verified
with an EPA audit sample.

This method of calibration is not amenable to automation and
routine analysis of large groups of samples.  Since the
standards are run before and after samples, the success or
failure of calibration cannot be determined until after
samples are analyzed.  This not very practical.  Instrument
drift is usually negligible, however the relative accuracy
of 93 to 107 % is sometimes hard to achieve with standards
being run before and after samples.  In addition, a curve of
one order of magnitude covers a relatively small
concentration range.  Calibration criteria can be met with a
25-fold range and with a slight relaxation of criteria (i.e.
90 to 110 % relative accuracy) can be extended over a 100-
fold range.

Currently MRI's clients want the method followed exactly as
written to provide defensible data without any  method
modifications.  However, the method could be updated and
made more suitable for routine analysis by incorporating the
following changes:

     Do not limit the calibration range.
                              77

-------
     Require a curve at the beginning but not at the end.
          Initial calibration criteria should be 90 to 110  %
          relative accuracy.
     Require continuing calibration standards which must
          meet a 10 % drift criteria (measured using
          instrument response.)

These changes will increase laboratory efficiency and make
the analysis method more like other SW-846 methods.  It is
hard to control instrument calibration when the calibration
occurs before and after sample analysis.   If criteria are
not met; all samples must be reanalyzed,  however the
modified calibration criteria will require reanalysis for
only those samples analyzed after the last successful
continuing calibration. Of course, using the average of the
curve run before and after samples provides marginally
better data, but if the calibration criteria are set
appropriately, data quality will not be affected.

These changes will not significantly affect data quality-
When put in the context of the data quality objectives of
the trial burn and other measurements of  equal importance
(GC/MS criteria are significantly wider)  these changes will
not affect regulatory decisions.

2.4  EPA Audit Samples

The method requires an analysis of an EPA audit sample and
results must be within 90 to 110 % of the true value.   The
purpose of this sample is to serve as an  independent check
on the accuracy of the analysis; however  it is not currently
available from the EPA.  Therefore, some  method of
verification of the accuracy of sample analysis is needed.
The analysis method does not require an independent standard
for calibration verification,  thus there  is no verification
of acceptable calibration.   For VOST and  SVOST there are EPA
audit samples and for metals analysis there is a requirement
for calibration verification with an independent standard.

At MRI, we have QA personnel prepare an independent audit
sample that is submitted for analysis with each batch of
samples.  The sample is prepared from NIST reference
material in dilute potassium hydroxide to mimic the
collection reagent and provide stability.  The concentration
of this sample is not known by analysis personnel.   The
results of the analysis must be within 90 to 110 % of its
true value.   The results for about 6 months of audit samples
is presented below in Table 2.
                              78

-------
Table 2 * Accuracy of MRI Hexavalent Chromium Internal Audit
          Samples
Audit sample True
Concentration (ug/L)
10.2
11.5
9.5
13.7
9.1
6.9
8.8
12
8.8
8.6
Average (standard deviation)
% Accuracy
102
110
98
101
101
98
110
97
101
100
102 (5)
Currently most firms do not provide this independent check
on the accuracy of sample data; however, a similar check can
be done by field samplers by submitting a reference
solution.  The preparation of such a standard requires
purchasing reference material and the time takes about 1.5
hours (solution standards are stable for at least 3 months).
Permit writers and trial burn reviewers should be aware that
some independent check on the accuracy is needed.

2.5  Spike Sample Analysis

The analytical method does not require spikes of samples to
determine the accuracy of the analysis given the stack gas
matrix.  However SW-846 requires spike samples, and the
previously mentioned study using the ASTM and EPA methods
for hexavalent chromium in water also require matrix spikes.
MRI requires a determination of the accuracy and precision
of every analysis to be evaluated in terms of the sample
matrix.  MRI does at least one matrix spike and sometimes a
duplicate matrix spike for a set of samples from an
incinerator, boiler or industrial furnace.

To prepare spikes, an aliquot of the sample is fortified
with a very small amount of a high concentration standard to
                              79

-------
achieve a concentration level in the middle of our
calibration curve.  This takes about 15 minutes of a
laboratory technician's time.  Our objective is for recovery
to be between 80 and 120 %; however other objectives for
trial burns, such as GC/MS surrogate recovery, are as wide
as 50 to 150 % recovery.  The data for the past year is
summarized in Table 3.

Table 3 * Recovery of Stack Gas Impingers Spiked with
          Hexavalent Chromium
Initial Sample
Concentration
(ug/L)1
13.1
6.39
5.96
7.08
7.29
3.81
<1.0
<1.0
<1.0
1.93
Overall Average
(standard
deviation)
Average % Recovery
from Matrix Spike
Duplicates
101
109
104
40
61
85
92
92
89
94
84 (22)
Relative Percent
Difference Between
Duplicate
NA2
16
4
4
2
3
7
3
NA
NA
NA
1  Spike level  was  about  10 ug/L  for all samples
2  NA = Not  applicable; a single  spike was prepared  instead
     of duplicate spikes.

The spike recovery data  indicates that  most samples give a
recovery well within the 80  to 120 %  range; and all but one
sample was within 50 to  150  % recovery-   The precision of
the results is excellent.  Even though  recovery is generally
good, these data indicate that recovery is  variable and
without any spike determinations, the accuracy of the
emission measurement is  unknown.  Usually samples with lower
recovery show poor peak  symmetry or extraneous peaks in the
chromatogram,  and these  difficulties  are  consistent for
                              80

-------
duplicate analyses of the matrix spike and between matrix
spike duplicates.  This indicates a truly different matrix
from calibration standards and supports the claim that low
recovery is due to the sample matrix and not to laboratory
operations.  In addition, good results for the audit sample
provides evidence that the recovery problem is due to matrix
effects and not a problem with calibration.

What is the impact of this low recovery?  That depends upon
the emission results.  For example, the apparent problem
indicated by the low recovery (40 to 65 %) in the previous
table did not translate into a regulatory problem.  Stack
emissions for hexavalent chromium were very low and if
emission rates were adjusted for the low recovery, risk
assessment indicated acceptable incinerator performance.
However, in cases where the emission results are borderline,
recovery data are needed to support the regulatory decision.

2.6  Detection Limit Determinations

The method does not specifically define detection limits nor
require their derivation.  MRI has done detection limit
determinations using multiple analyses of low standards (0.5
ug/L) and determined a detection based upon precision of
about 0.2 ug/L.  This 0.2 ug/L level is only slightly above
the 5 times signal to noise ratio for most samples.

In analyzing calibration curves, MRI has determined that
there exists a small amount of Cr*6 as background in our
standards.  The amount of this background level is between
0.2 and 0.4 ug/L and is evident as a positive intercept in
the calibration curve by linear regression.  This background
level is due to the potassium hydroxide (0.1N) used to
prepare calibration standards, since 0.1N potassium
hydroxide has an observable Cr*6 peak and straight deionized
water does not have any observable Cr*6.  However, this
background is insignificant when compared to stack gas
samples and the levels of regulatory interest.  In addition,
this background level is lower than or equal to that found
in field blanks of collection reagent.

Due to this background, MRI has adopted a calibration curve
that extends down to 1 ug/L and uses this level as a
reporting limit.  Since regulatory use of the data often
employs the detection limit to prove low or no risk to the
environment when the analyte is not detected, the use of the
lowest calibration standard provides a conservative
reporting limit.  However, in reviewing trial burn plans and
reports, MRI has noted that many firms doing these analyses
                              81

-------
present detection limits without describing the derivation
of the limit.  In most cases this is not a difficulty due to
the presence of hexavalent chromium in the samples, but in
cases where the analyte is not detected, and emissions rates
are based upon the detection limit, the detection limit must
be supported with a technical justification.

2.7  Blank Correction

The method allows for correction of emission data by
subtracting the amount found in the blank; however, a
statistical justification for blank correction is not
required.  For example, to blank correct VOST results, the
permit applicant must prove that the blanks are
statistically different from the sample results.  In
reviewing trial burn reports, the blank correction option is
often employed; however, its use is neither statistically
justified nor is its impact upon the regulatory decision
discussed.  Often this blank correction comes from the
analysis of a single field blank.  In one case, the blank
correction lowered emission levels by 25 to 50 %.

The general policy for blank correction in trial burns
presented in Handbook of QA/QC Procedures in Hazardous Waste
Incineration (EPA/625/6-89/023,  January 1990)  is to provide
a statistical justification of the blank correction and to
present emission data both corrected and uncorrected.  Given
the high toxicity of hexavalent chromium, this policy should
be incorporated in future revisions of the method.  In
addition, permit writers have considerable latitude in
requiring trial burn information, and they should accept
blank corrected emissions only with clear justification and
the emission rates presented as corrected and uncorrected.

3.0  CONCLUSION

The hexavalent chromium analysis method is rigorous and
produces relatively precise and accurate data suitable for
regulatory decision making.  However,  it does have a few
short-comings in QA/QC procedures for holding times,
accuracy, detection limits and handling blank data.  In
addition, calibration and duplicate analysis requirements
could be modified to allow a more cost effective automation
of the analysis.   It is hoped that the topics brought up in
this paper will be used when the method is updated and that
the enhancements will be required by trial burn plan
reviewers to monitor and demonstrate the quality of the
sample results.
                              82

-------
14         DEVELOPMENT OF TECHNOLOGY PERFORMANCE
        SPECIFICATIONS FOR VOLATILE ORGANIC COMPOUNDS

      C. Purdy and W.E. Schutte, Office of Environmental Development and Waste
      Management, U.S. Department of Energy, Germantown, Maryland; M.D. Erickson
      (Environmental Research),, S.C.Carpenter (Chemical Technology), P.V. Doskey
      (Environmental Research), P.C. Lindahl (Chemical Technology), and A.D. Pflug
      (Chemical Technology), Argonne National Laboratory, Argonne, Illinois.

      INTRODUCTION

      The Office of Technology Development (OTD) within the Office of Environmental
      Restoration and Waste Management of the Department of Energy has a mission to deliver
      needed and usable technologies to its customers. The primary customers are individuals
      and organizations performing environmental characterization and remediation, waste
      cleanup, and pollution prevention at DOE sites. DOE faces a monumental task in cleaning
      up the dozen or so major sites and hundreds of smaller sites that were or are used to
      produce the U. S. nuclear weapons arsenal and to develop nuclear technologies for national
      defense and for peaceful puposes (1,2). Contaminants and waste materials include the
      radionuclides associated with nuclear weapons, such as plutonium and tritium, and more
      common pollutants and wastes of industrial activity such as chromium, chlorinated
      solvents, and polychlorinated biphenyls (PCBs). Quite frequently hazardous wastes
      regulated by the Environmental Protection Agency are co-mingled with radioactive wastes
      regulated by the Nuclear Regulatory Commission to yield a "mixed waste," which
      increases the cleanup challenges from several perspectives.

      To help OTD and its investigators meet DOE's cleanup goal, technology performance
      specifications are being implemented for research and development and DT&E projects.
      Technology performance specifications or "performance goals" describe, quantitatively
      where possible, the technology development needs being addressed.  These specifications
      are used to establish milestones, evaluate the status of ongoing projects, and determine the
      success of completed projects. All too common within proposals, statements of work, and
      project milestones is a vague statement that the technology being developed will be "better,
      faster, safer, cheaper." A comparison to available technologies and quantitative definitions
      of the anticipated improvements are frequently lacking in proposals. Program managers
      and customers have little indication of whether they are funding a needed improvement in
      measurement technologies until the technology is actually developed, commercialized, and
      implemented.

      This pilot project has focused on chemical measurement technologies for volatile organic
      compounds (VOCs) in soil and ground water. The project team has developed a model for
      the development and implementation of performance specifications and has applied this
      model to VOCs. The customers and their needs define the problems to be solved, a review
      of relevant available technologies defines the baseline from which development must
      progress, and quantitative performance specifications are identified on the basis of unmet
      monitoring needs. These general performance specifications are then translated into
      project-specific performance specifications in consultation with the principal investigator.

      PERFORMANCE SPECIFICATION APPROACH

      The general approach is to document what currently exists or is nearing completion and
      compare that baseline to the customers' needs to identify the unmet requirements.  These
                                              83

-------
unmet requirements then are the basis for the technology development needs by the
investigators, to be addressed. The process needs to be quantitative where appropriate, to
steer project goals away from vague generalities like "faster" and toward specifics like
"reduction of labor requirements for Step A from 4 hr to 0.5 hr."

Performance specifications constitute a program management tool and must fit within an
overall program management structure. During the planning phase of a program,
performance specifications are needed to identify the research or demonstrations requiring
support. During the execution phase, performance specifications can help in the evaluation
of both proposals and ongoing projects.  The success of completed projects can be
evaluated against the established performance specifications to provide a basis for
considerations about further funding, technology transfer, or implementation.

For an individual project, technology performance specifications will be used to establish
milestones, evaluate the status of ongoing projects, and determine the success of
completed projects. Preliminary performance specifications will be required in proposals
and will be evaluated as part of the proposal evaluation.


I.      PERFORMANCE SPECIFICATION PROCESS

       A. Baseline Technologies for an Environmental Research or
          Waste Management Need

          1.   State the need of the customer. What problems need to be solved? What
               are the boundaries of the need? (Limit the scope to a specific need to keep
               the process focused).  What are the current deficiencies? What are needs
               versus desires? This information should be supplied in consultation with
               customers  where possible; interactions need to both quantify the need and
               distinguish truly needed features and performance specifications from
               desired,  but not required, properties.

               Discuss the potential impact of improved technologies. Where possible,
               quantitatively estimate what can be done better, faster, safer, or cheaper.

               This preliminary statement, prior to other work, is refined in Section I.

          2.   Describe the relevant technologies  that can be applied to the customer's
               need. The focus is on the customers' need, not the technique, so address
               all viable alternatives,  not only directly competing alternatives.  (For
               example, characterization by gas chromatography/mass spectrometry,
               describe  all VOC characterization technologies, not just the competing
               related technologies).  In addition to available technologies,  address
               current worldwide technology development and assess the probabilities of
               addressing unmet needs and the schedule for availability of the technology.
               For upcoming technologies, present only a brief overview of technologies
               that are still at the basic research stage,  will not be available within the
               customer's time frame, or have a low probability of successful
               implementation. If no technologies currently exist to solve this problem
               state that

              The text needs to include descriptions of technical alternatives, their status,
              their availability (with  a list of vendors), and any barriers to
              implementation.
                                        84

-------
   3.   Describe how the technology will be applied. What are the conditions of
        the operating environment (environmental, space, health and safety,
        operational constraints)? What are the time requirements (setup time,
        measurement time, time to data availability, need for 24-hour operation.)
        What are the user attributes (education/training, number of operators).

   4.   Tabulate quantitative performance specifications for current technologies
        (including technologies under development that appear to be on a path
        toward successful and timely implementation and availability to DOE; i.e.,
        don't fund new research and development if something is already in the
        pipeline).  Since it is possible to consider specifications in dozens of areas
        (technical performance, cost considerations, applicability, availability,
        etc.), restrict the effort by limiting the list to the ten most important
        specifications, cost should be one of the ten.

   5.   Document all of the above and provide statements about the quality of the
        information (e.g., average of five vendors' specifications with similar
        applications to DOE needs versus educated guess with no firm
        information).
B. Customer Requirements

   Revise and refine the customer need statement developed in Section LA
   Address the following issues in more detail:
    1.  Identify problems that need to be solved. See Section LA for more
       details.

    2.  Identify the customer(s). Name sites, sub organizations, individuals,
       operable units, etc. where available.

    3.  Define quantitative requirements  (e.g., regulatory drivers for detection
       limits).

    4.  Define functional and operating requirements .

    5.  Assess the priorities of the need (e.g., critical versus desired for
       convenience).

    6.  Assess the urgency of the need (e.g., a start date for projects, by which
       commitments to technologies must be made).

    7.  Document all of the above.
C.  Unmet Requirements

    Compare baseline technologies versus customer needs and list unmet
    requirements.
                                         85

-------
       D. Technology Development Needs

          1.   Discuss the unmet requirements that can realistically be met by the
               proposed technology; discuss the quantitative improvements that can be
               anticipated.

          2.   Tabulate quantitative performance specifications for the proposed
               technology (i.e., improvements over baseline).

          3.   Assess the quality of estimates  (e.g., proven in other applications; within
               theoretical limits of techniques; best scientific judgment).  Provide or
               reference documentation where available.

          4.   Recommend priorities  where appropriate, for addressing requirements.
               These priorities should be based on the customers' priorities, the
               probability of success, and the time frame for completion. Discuss all
               considerations in the prioritization.

          5.   Identify expertise (organization, name, address, and phone) that can be
               tapped to confirm the validity of the effort.

       After these steps, the needs baseline is established, and the performance
       specifications can be used as a program management tool to establish milestones,
       evaluate the status of ongoing projects, and determine the success of completed
       projects.

II.    PERFORMANCE SPECIFICATION PROCEDURE

      The development of performance specifications is a program management function.
To minimize bias in their preparation, the specifications must be prepared by a disinterested
individual(s); on the other hand, persons preparing performance specifications need good
technical skills and current awareness in the area. These  skills are precisely those needed
by potential investigators, so conflicts are to be expected.  After internal management
review, the draft should be critiqued by selected principal investigators both to provide
additional technical input and  to educate the principal investigators on the process.


HI.  USE OF PERFORMANCE SPECIFICATIONS

          1     The following guidelines should be applied to the use of performance
               specifications: The principal investigator should use the performance
               specifications as a goal in developing work plans, quality assurance project
               plans,  and  experimental plans and in assessing the progress of work.

          2.   The program manager  should use the performance specifications as a
               criterion for evaluating progress on the project. This will be especially
               critical at reviews where decisions on continued funding are made.

               The program  manager should also refer to the performance specifications
               in communications about the projected value of the technology to potential
               users.
                                        86

-------
          3.   Customers can use the performance specifications to assess the
               applicability of the technology to their needs and can plan to transfer the
               technology to their site application as it nears completion of development.


          4.   All involved parties (principal investigator, program manager, and
               customer) must remember that performance specifications are subject to
               change as new information becomes available (new research results,
               change in programmatic goals or changes in the customer's needs). All
               involved parties must initiate change whenever new information becomes
               available that will change the basis for the performance specifications.

           Changes in performance specifications can constitute a change in the statement
           of work and thus necessitate a contractual change and possibly adjustments in
           cost and schedule.

IV. APPLICATION TO VOC MEASUREMENT

Reviews of currently available technologies and ongoing research projects were prepared
on the following topics:

   Gas Chromatograhy
   Gas Chromatography/Mass Spectrometry (GC/MS)
   Electrochemical sensors
   Raman spectroscopy
   Infrared spectroscopy
   Fiber optics (ultraviolet, fluorescence, etc.)

As an example, the GC/MS review found the following:


            Two manufacturers currently market field GC/MS instrumentation,
     Bruker and Viking. Both systems are capable of operation under either line or
     battery power. Both are relatively insensitive to environmental conditions, but
     neither is light enough to be carried by a person.  VOCs can be introduced to the
     system by liquid injection, thermal  desorption, direct air sampling (bypass GC),
     air sampling by trapping on sorbent thermal desorption and purging and
     trapping.

       An older version of the Bruker instrument has been extensively used in the
       field for VOC analysis with results generally both more accurate and more
       precise than those from parallel fixed-base laboratory analysis, because loss
       of analyte during storage and transport is minimized. Substantial cost
       savings have been documented with field VOC analysis using field-portable
       GC/MS.

       Current research on instrumentation focuses on developing a customized
       purge-and-trap/temperature-programmable GC for the instrument (Leibman
       et al. Los Alamos National Laboratory); development of a microprocessor to
       control the sampling system and software designed to integrate the sampling
       and ion trap functions (Hemberger et al. Los Alamos National Laboratory);
       and development of direct-sampling ion trap mass spectrometry technology
       by Guerin and co-workers at Oak Ridge National Laboratory. On the last
       project, no gas chromatography is provided for separation of compounds;
                                        87

-------
       the mass spectral discrimination is judged sufficient  Sample introduction
       systems have been developed for air and water. The Los Alamos and Oak
       Ridge instruments have been field tested and are in various stages of
       commercialization. Additional work is being conducted on development of a
       person-portable (backpack) instrument by Meuzelaar and co-workers at the
       University of Utah.

Work is in progress to integrate the initial findings of the reviews and the customer needs to
determine the unmet needs and thus the technology development requirements.

V. CONCLUSION

Performance specifications formalize an informal process to screen potential areas of
research and match technology development with customers' needs.  Although the steps
outlined here may appear daunting, the process provides a structure for program
management. The benefits include early development of a linkage between ongoing
research and the customer's needs and also minimization of replication of research.

Much work remains on this project as we attempt to implement performance specifications
on technology development for VOC measurement.  Application of this methodology to
other characterization technologies and remediation technologies is contemplated.

ACKNOWLEDGMENT

Work supported by the U.S. Department Of Energy, Assistant Secretary for Environmental
Restoration and Waste Management, Office of Technology Development, under contract
W-31-109-Eng.-38.

BIBLIOGRAPHY
(1)  U.S. DOE Environmental Restoration and Waste Management Five-Year Plan, Fiscal
Years 1993-1997, DOE/5-0090P, August  1991.

(2)  U.S. Congress, Office of Technology  Assessments, Complex Cleanup: The
Environmental Legacy of Nuclear Weapons Production, OTA-O-484 (Washington, DC:
U.S. Government Printing Office, February 1991).
                                       88

-------
•J5                        THE QAPjP QUAGMIRE

     Marcia A.  Kuehl.  President/Owner, M.A. Kuehl  Company,  3470
     Charlevoix Court, Green Bay, Wisconsin 54311

     ABSTRACT

     The  Quality  Assurance  Project  Plan  (QAPjP)  has  become
     quicksand for environmental investigations.  As a regulator,
     I reviewed them,  as a contractor I wrote them, as a laboratory
     director I  tried to comply with  them.   As a  taxpayer  I  am
     appalled by how what was originally intended to be a document
     used to ensure that data of adequate quality is obtained has
     degenerated  into a pound  of  paper that  takes too  long  to
     write, review, and revise.  This paper will present the sticky
     points  in  the  writing of  QAPjPs and how good science  is
     sacrificed  for ease of  QAPjP approval.   The  substance  of
     regulatory agency comments and concerns about QAPjP contents
     and how they  can  be  addressed will  be  relayed via horror
     stories.  The viewpoint of  the laboratory  and it's often lack
     of involvement in the QAPjP will be offered. The new industry
     of data validation  will be reviewed and  it's  impact on the
     QAPjP assessed.   Suggestions  for  a  creative overhaul of the
     QAPjP  elements themselves  and ideas  for  streamlining  the
     development and approval processes will be proposed.

     INTRODUCTION

     December  29,  1980  marked  the birth  of  the  term  Quality
     Assurance Project Plan  (QAPjP). The EPA Office of Monitoring
     Systems and Quality Assurance created  QAMS-005/80  "Interim
     Guidelines and Specifications  for  Preparing Quality Assurance
     Project Plans".  And, like so many other "guidance" documents,
     QAMS-005/80  soon became defacto  regulation. All QAPjPs,  if
     they  are  to be approved, must  follow these specifications.
     Earlier, the EPA Administrator in May, 1979 required that "QA"
     must be done, smarting  from the lessons learned  at Love Canal.
     QAMS-005/80 prescribed that the primary way that QA was going
     to be assured for environmental measurements was through this
     QAPjP document.   QAMS-005/80  painstakingly describes how to
     number the  pages  but glosses  over how specifically to relay
     precision,  accuracy,  representativeness,  comparability  and
     completeness (PARCC) parameters and the mysterious process of
     data assessment. The QAPjP  was originally  intended to address
     both  the  laboratory and field  activities as  part  of  the
     Sampling and Analysis Plan  (SAP).  The field activities have
     since been  segregated  in a stand alone  document,  the Field
     Sampling Plan (FSP). Unfortunately, the PARCC parameters are
     not carried over into  the  FSP  as they relate  to  the field
     investigation.  Subsequent   documents  covering  Data  Quality
     Objectives  (EPA-540/G-87-003)  and Category IV QAPjPs (EPA-
                                   89

-------
600/8-91-006) were developed to provide additional aid.  All
of these documents are intended to help in the production of
an "approvable" QAPjP  so that the investigation  can begin.
The  quest   for   the  approval   signatures   requires  some
floundering in the quagmire.

WRITING THE QAPiP

Despite the  fact that the  format  of  the QAPjP is relatively
rigid,  some creative writing is required to get it through the
approval process.  Avoiding controversial points often results
in  good  science  being  sacrificed  for  ease of  approval.
Specific  target analytes or indicators  are  avoided because
they are  non-routine and often the project team settles for
CLP Contract Required Quantitation Limits (CRQL) because it is
easier than finding/validating a method with a lower detection
limit.  Use of an already approved QAPjP as a template with the
names changed (project,  organizational chart, laboratory) is
helpful,  but  this  practice perpetuates the  QAPjP  as  a one-
size-fits-all pantyhose exercise that is not tailored to the
specific project needs.

The  major  issues  that  delay  approval  of   the  QAPjP  are
identifying non-CLP laboratories to conduct the analyses, and
including non-routine analytes and/or methods.  Any way that
these issues can be presented to soften their deviation from
the usual CLP boilerplate,  such as citing other EPA accepted
references and internal guidance, can aid in quicker approval.

To CLP or not to CLP

The Contract  Laboratory  Program  (CLP),  first implemented in
1980, was never intended to be the  defacto "certification" for
environmental laboratories  that  it  has  evolved  into  today.
This "certification" perception was  a major  finding  of the
1991 EPA management review  of the CLP.  The perception is that
if a laboratory is participating  in the  CLP,  it  has  jumped
through enough documentation,  facilities audits, and QC hoops
to "guarantee" error free  data. This perception  is not only
unfair to those  laboratories that  for economic reasons choose
not to  pursue CLP  contracts, it is untrue.  Granted, the data
package documentation that has taken 13 years to refine has
withstood  legal  and  independent data  validator  scrutiny,
however non-CLP laboratories are also capable  of providing the
same defensible data packages to meet requirements.

It appears that when a non-CLP laboratory is named in a QAPjP,
the degree of scrutiny by the approving  agency increases.  An
on-site  audit  is triggered  and often  blind  performance
evaluation  (PE) samples  will be  sent prior  to  the on-site
audit.  It is ironic that despite this increased surveillance
                              90

-------
of non-CLP labs cited in QAPjPs,  only CLP  labs have ever been
suspended or disbarred by the EPA. The QAPjP author can help
head off  an unnecessary round of  comments  by  supplying the
following information in the first draft:

            - Lab certifications/state approvals
            - Dates of independent on-site audits
            - Blind PE studies lab participates in
            - Experience of lab with Level IV data packages

Offers  to provide  written audit  reports from  other state
and/or federal agencies and results of the last blind PE study
should be made at the pre-QAPjP meeting.

To SAS or not to SAS

Active CLP labs are  not immune to scrutiny, especially  if non-
traditional methods  (those not included in the CLP Statement
of Work) or lower detection limits are needed for the project.
These conditions require a Special Analytical Services (SAS)
request  if the  project  is  under  the auspices  of  an  EPA
contract.  The SAS necessitates that the author supply  a fully
documented  Standard  Operating   Procedure   (SOP)   with  QC
specifications  for  number, type, and  limits of  internal QC
samples.   This  is  usually a  Catch 22 proposition,  as  the
analysis  is "special"  because  there  is  no SOP,  let alone
standardized limits.  Even if the QAPjP is not for a  federal
lead project, the proposed laboratory must supply the  details
of what will constitute the analysis method and acceptance
criteria.

It is imperative that the lab  method SOP submitted for non-
traditional  analyses  contain  the  following  items,  or  an
additional  submittal will  likely be requested:

            - Method  specific QC  (frequency and limits)
            - Basis  for limit determination
            - How detection limit was determined
            - Supporting data for detection limit

One  of  the  most frustrating  issues  in  QAPjP  authoring is
centered around  the  SAS.   Often, reviewers with little or no
laboratory  experience will be  tasked with  determining  the
appropriateness of the proposed SOP submitted. The shortage of
bench chemists who desire to switch gears to  review paperwork,
and the dynamic  nature of  environmental analytical chemistry
contribute  to  the lack of QAPjP reviewers with hands-on lab
experience.  The most successful tactic to deal with  this is
to ask  to  have access to the  agency reviewer and agency lab
staff set up a conference with the proposed lab's analytical
team and draft an acceptable SOP together.
                              91

-------
OAPiP REVIEW COMMENTS

Written comments are received  from the agency reviewers and
assessed by the project team for whether the QAPjP would need
a substantive change affecting project scope, budget and/or
schedule.  Over the course of 13 years of authoring QAPjPs and
responding  to  comments,   I have  compiled   some  requested
revisions that illustrate the concept of losing the forest for
the trees by focusing on minute technical details that do not
make the project data more defensible:

The Case of the 5 Point Curve

The  reviewer  insisted   on  an initial  5  point curve  for an
indicator analyte that  was  expected  at high concentrations.
The lab SOP detailed a  4 point  curve  and an independent check
standard at a fifth concentration analyzed every 20 samples.
Heated 2 hour  conference calls over the linearity of the curve
near  the detection  limit   and the  semantics of the  SW846
terminology of "minimum  calibration" requirements  against
agency "requirements" ensued while  leachate from the site ran
into the river.  The gap between this PRP QAPjP submittal and
start of the  field work was 15 months.  The lab rolled over
and  rewrote their  SOP  to include  an additional  calibration
curve concentration near the detection limit and dedicated a
project  "babysitter" to ensure that  the analysts  added the
extra curve concentration.  The detected concentrations in the
samples were  so high that  a 3  point  curve  would have easily
sufficed as the range  component  of the calibration function
was  the  driving  issue  for  the  project  data,  not  the
sensitivity.

The Case of the Best Professional Judgement Holding Time

A SAS request  was prepared  using a  USGS reference method.  As
no holding time was established by the USGS in their method,
the  agency reviewer  disregarded  the 28  day holding  time
suggested  for the  same analyte in an EPA method,  and used
"best professional judgement"  to require that the sample be
analyzed within 48 hours of collection. The  process for SMO to
solicit  labs  that  would even  bid  on  the work took months.
After the project was finally underway, the analysis could not
be consistently accomplished within the 48 hour constraints,
and  resampling  was  done.    Out  of  intellectual  curiosity,
samples were  analyzed  again within 28 days  and results were
not significantly different.  The mythical holding time cost
the project 3 extra  days field time  at an  estimated cost of
$6,000 and delayed the project start by 2 months.
                              92

-------
Top Secret SOPs

Some  laboratories feel  that their  analytical SOPs  are so
revolutionary   and   innovative  that   they   consider  them
confidential,  proprietary business  information.   Evidently
there are many complex nuances to these  analyses that another
lab learning a helpful hint would affect competitive pricing.
As proprietary information, these labs will only submit their
SOPs directly  to the reviewing agency  and not to the QAPjP
author.   The perception is that there  is  a  lucrative black
market for lab SOPs  and the QAPjP  author is  selling them on
the street corner to other  labs too  lazy to write their own,
let alone follow the  reference method.  What is ironic is that
for  any  federal  lead  project,  a citizen  can request  all
pertinent  laboratory  data and  SOPs under  the  Freedom of
Information  Act.  Being caught between  the  reviewing agency
questioning  the details of  an SOP one is not allowed to see,
yet being held responsible  for revising the SOP and using to
assess the project data for compliance is a dilemma for the QA
professional.

OAPiP COMMENT  HALL OF  FAME

QAPjPs   have   been   revised  to    satisfy   the   following
"substantive"  comments:

1.  References to data or locations should be changed to read
    "in" Table X/Figure Y.

2.  In the organizational chart,  the  line from the Project QA
    Officer  to the Project  Manager should be a dotted line,
    and the  agency QA Officer box should be equal in size and
    location to the agency  Project Manager's box.

3.  The procedure for calibration and maintenance of the field
    trailer  thermometer/barometer should be included.

4.  Sample container cleaning protocols are not adequate to
    ensure contaminant-free bottles.  (The SOP  attached to the
    QAPjP was  the agency's  Scope of  Work from  it's currently
    awarded  bottle contract,  only retyped.)

Once a revision to a QAPjP is made and resubmitted for review,
it is not unusual to have a different reviewer assigned to it
and the "do  not pass go,  do not collect $200"  review process
starts all over again.

THE MISSING  LINK

More often than not,  the lab that must follow the QAPjP is
left out of  it's creation entirely,  or  consulted at the last
                              93

-------
minute before submittal.  An approval signature from the lab's
QA and managerial personnel was never included in QAMS-005/80.
For those  QAPjPs  involving the  federal CLP  lab  pool,  this
would  be  burdensome,  but for  PRP and  state lead projects,
review and approval involvement is a  necessity, especially if
the QAPjP  author  is  from  an  independent  engineering firm.
This  review/approval  also   is  necessary   for the   lab  to
appropriately price the analyses  and  advise whether they have
the capacity  to produce the deliverables  in the  turnaround
time allotted in the project schedule.

Once a candidate lab  is  selected and preliminary contact is
made,  the QA representative from the lab should be  involved in
project scoping and DQO negotiations. Attendance  at the pre-
QAPjP  meeting is  usually  invaluable.    This  time   is  not
perceived as profitable  as  lab analyses in the bottom line,
but the angst this  upfront  planning  will save later is well
worth it.

Being placed in the middle of philosophical arguments between
the agency  chemist/reviewer  and  the lab on  issues like the
real  utility of  matrix spikes   and  lab duplicate  control
statistics  is  not  a  pleasant  experience.    Trying to relay
chemical  "technospeak"  is  frustrating  and  direct  contact
between the factions with the QAPjP author as mediator is the
most efficient way to resolve these issues to get  the QAPjP
written and submitted.

DATA VALIDATION

It  is  not enough  to  specify the methods,  QC, calibration,
reporting  forms,   analytical   standards,  and data  package
contents/order.   The resultant  data package  must often be
assessed  by  an  "independent"  data  reviewer  to look  for
analytical  errors   or take  issue with "best professional
judgement" calls.   The new industry of data validation boomed
after the reported discovery of fraud in  several laboratories,
yet it  is  doubtful that without  an  on-site  inspection  by a
skilled computer guru and chemist with lab experience,  that
the anomalies would have been caught in data packages during
independent validation.

Data validation itself is an inexact science.   Most validators
defer  to the  National Functional Guidelines, but even these
documents allow for interpretive dancing by  the validator in
the areas  of the  effects of  out of compliance  calibration
compounds, matrix  spikes  and field precision data.  One of the
most disconcerting positions to be in is to be  the laboratory
QA  fall  guy when  the data validators  assess the data using
criteria you were  never made aware of -  and nonpayment of the
analytical  bill  is  threatened  by  the  QAPjP author.   The
                              94

-------
criteria to be used in the validation process should be stated
in the  QAPjP,  along with who/what  firm will accomplish it.
The validator's authority, responsibility, and communication
protocols should be clearly defined to control requests to the
lab  for  resubmissions  and  additional  information.  Data
validators should be involved early in the project planning to
become educated in the project goals and expected analytical
challenges.  Validating  analytical  data in a vacuum without
regard  to  balance between the intended use  of  the  data and
definitive precision and accuracy of each number  is only an
exercise in accounting.

OAPiP MAKEOVER

The 16  element  QAPjp contains  at least 9 sections that have
evolved into boilerplate.  The meat  of  the QAPjP is the table
of proposed samples, analytes and methods generally found in
Section  5.0.   Listed below  is each section  of  the current
QAPjP format and  suggested changes to each to streamline the
writing and approval processes:

Section 1.0   Title Page:  Add provisions for signatures for
Laboratory   Director,   Laboratory  QA  Manager,   and  data
validation firm.

Section  2.0     Table   of  Contents:    Change  to  reflect
consolidated sections as described below.

Section 3.0   Project Description:  Expend effort to organize
clearly the available historical data and analytical methods
used, indicate historical data quality  assessment and issues.

Section 4.0  Project Organization and Responsibility: Add the
authority and responsibility of the data validator including
reporting, lines and methods of communication and the process
for interaction with the laboratory.

Section 5.0   QA Objectives  for Measurement Data in Terms of
Precision,  Accuracy, Representativeness,  Comparability and
Completeness:  Reduce this section to  the table  of samples,
analytes, methods and internal/external QC samples and limits.
Highlight   any  special  detection  limits   and/or  method
restrictions.  Delete canned text boilerplate definitions of
PARCC.  The comparability descriptions  and objectives are not
quantifiable, and it remains  to be  seen what magic limit of
percent  completeness constitutes  an acceptable  level.  The
whole  issue of  field  data quality  objectives needs  to be
recognized and defined for the project in this section.

Section 6.0  Sampling Procedures:  This section usually defers
to a separate site specific sampling plan, which is generally
                              95

-------
reviewed separately-.  The sampling procedures, especially the
proposed   sample   collection   containers,   volumes,   and
preservatives should be reviewed by the lab to ensure that the
specifications are compatible with the lab analyses.

Section  7.0   Custody:  No earth  shattering revolutions  in
chain-of-custody protocols has  transpired over  the  last 13
years.  This  section is filler boilerplate and can be handled
by reference  to  the sample collection   plan and laboratory
custody SOP-

Section  8.0    Calibration  Procedures  and   Frequency:  This
section usually refers to the specific lab methods and should
be deleted and contained in section 9.0.

Section 9.0  Analytical Methods:  Lab specific SOPs should be
attached or  the  CLP SOW cited.   Any  known  project specific
deviations or contingencies should be mentioned.

Section  10.0  Data  Reduction, Validation and Reporting: The
data  validation  firm  should supply  their proposed  SOP for
validation and aid  in the  preparation of this  section.   The
usual  rehash  of  how to  calculate  RPD,  % recovery,  etc.  is
unnecessary filler.

Section  11.0  Internal QC Checks:  This section is not needed
as    it   should   be    included    in   the   table    of
samples/analytes/methods/QC in section 5.0.   Any lab specific
terms  used should have been defined  in  footnotes or text in
section  5.0.

Section  12.0  Performance and Systems Audits: This section is
more  canned boilerplate  unless a  site specific  field and/or
lab audit  is  proposed.   Delete  this  section and include in
section  4.0.

Section  13.0  Preventive Maintenance: Delete this section as
the analytical method  or  field SOP should cover it.  Any field
meter  maintenance  should be  included in  the field sampling
plan.

Section  14.0  Specific  Routine Procedures  to  Assess  Data
Precision, Accuracy, and Completeness:  This  section repeats
the table  in section 5.0 and generally contains  pages  of
definitions and  verbiage on  how to calculate items  such as
mean and standard deviation.  Back  in  1980, QAMS-005/80 noted
that  "Recommended  guidelines  and  procedures to  assess  data
precision, accuracy and  completeness are  being developed."
We're still waiting.

Section  15.0  Corrective Action:   Who  can  stop/restart lab
                              96

-------
analyses under what conditions is a major area of concern to
project budgets and schedule.  Section 4.0 could include the
flow of action to out of control situations.

Section 16.0  QA Reports to Management: Usually an elegantly
written  section that  says  little,  it  should contain  the
specifics of the project QA summary and who will write it.

SUMMARY

In the current system, for less than a year approval time from
QAPjP  submittal  to approval, authors  should  use  an active,
upstanding  CLP  lab  to do  only routine  CLP  SOW regulated
analyses on nice clean soil and water samples.   The utility of
full Level  IV  data packages  being  reviewed  by a third party
data validator firm should be evaluated for the project.

An overhaul of  the QAPjP  purpose and process  is needed.   An
injection of the lab personnel  and data validators early in
the  project planning  and evidence  of their  commitment  by
signatory approval of the QAPjP  is  critical to the success of
the  project.     The  perceived  dismal  record of  Superfund
investigations completed  and  Record  of Decisions  signed and
cleanup accomplished is testimony to this  need for change.  A
task  force  consisting  of lab personnel,  engineering firms,
data validators and risk assessors  should  be initiated by the
EPA Quality Assurance Management  Staff.  As Congress now calls
for  Superfund  to be "fixed", the  logical place to start is
it's infrastructure, the QAPjP.

REFERENCES

U.S.   EPA,   December  29,  1980.   Interim  Guidelines   and
Specifications for Preparing Quality  Assurance Project Plans,
QAMS-005/80.

U.S. EPA, March 1987.  Data Quality  Objectives for Remedial
Response Activities;  Development Process, EPA 540/G-87-003.

U.S. EPA, February 1991.  Preparation  Aids  for  the Development
of Category IV Quality Assurance  Project Plans. EPA-600/8-91-
006.

U.S. EPA, Draft  December  1990,  Revised June,  1991. National
Functional Guidelines for Organic Data Review.

U.S. EPA, October  1989  Revision. Laboratory Data Validation
Functional Guidelines for Evaluating Inorganics Analysis.

U.S.   EPA,   October  1991.   "Management  Review,  Contract
Laboratory Program".
                              97

-------
16
               EVALUATION  OF METHODS  FOR DETERMINATION OF
                       INSTRUMENT  DETECTION LIMITS


Meszaros,  T.J.,  Technical  Support  Supervisor,  Dobb,  D.E.,  Scientific
Supervisor, Organ.  B.D..  Senior  Scientist,  Converse,  J.C.,  Data Audit
Technician, Lockheed  Environmental Systems & Technologies Company, Las
Vegas, Nevada;  Russell McCallister, U.S.  Environmental Protection Agency,
(EPA) Environmental Monitoring Systems Laboratory, Las Vegas, Nevada.

ABSTRACT

Instrument  detection  limits (IDLs)  that are used  in  the  EPA Contract
Laboratory  Program  (CLP)  are defined  in the current  Statement  of Work
(SOW),  and are  reported  by  CLP   laboratories  each quarter.    The SOW
requires that  CLP  laboratories  perform seven analyses  for each analyte
(except  cyanide)  on three  nonconsecutive  days;  determine  the standard
deviation  for  the analyses  on  each  of those days;  and sum  the three
standard deviations.   This procedure  is  labor-intensive, requiring a full
week before results are obtained.  Questions  have been raised regarding
applicability  of  the  results over an  entire  quarter.   In  an  effort to
obtain more current IDL information,  the authors evaluated the daily IDLs
obtained over a 3-month period and compared their results to the contract
required  detection limit  (CRDL)   and  the  quarterly  IDL.   Using  data
submitted  by  CLP  laboratories,  IDLs  were calculated  using  continuing
calibration blanks  (CCB)  and standards  at  two  times  the  CRDL  (CRA for
atomic  absorption  analytes and   CRI   for  inductively  coupled  plasma
analytes).  The daily  method calculates  IDLs from existing QA/QC data and
analyses dedicated  to just  IDL determinations  are not necessary.   We
compare the results from the analyses using the alternate daily methods to
results from the analyses  using  the quarterly method defined in the SOW.
These data  may be useful  for determining  the applicability of  the SOW
method of IDL calculation.

Notice:   Although the  research described in this  article has been funded
wholly or  in part  by  the  United  States  Environmental  Protection Agency
through contract number 68-CO-0049 to Lockheed Environmental Systems and
Technologies Company,  it  has not been  subjected  to  Agency review and
therefore does not  necessarily  reflect  the views  of  the Agency  and no
official endorsement should be  inferred.
                                          98

-------
17               TECHNICAL  INFORMATION MANAGEMENT FOR THE PIT 9 PROJECT
                           IDAHO NATIONAL ENGINEERING LABORATORY

      J- Owens, Senior Scientist, Chemical Services, EG&G Idaho, Idaho Falls, ID 83415,
      D. Forsberg. Senior Engineering Specialist, Buried Waste Program - Pit 9 Project,
            EG&G  Idaho,  Idaho Falls, ID  83415,
      D. Macdonald. Program Manager Buried Waste Program, Department of Energy Idaho
            Falls Field Office, Idaho  Falls,  ID 83415.

      ABSTRACT

      This paper  describes the  development  of an access controlled data system that
      provides group-wide accessibility of laboratory and field information  in support
      of the Pit  9 cleanup at  the  Idaho National  Engineering Laboratory.   Pit 9 was
      selected for remedial  action  in accordance  with the Comprehensive Environmental
      Response, Compensation, and Liability Act (CERCLA)  pursuant to the  Idaho National
      Engineering  Laboratory  Federal   Facility Agreement  and Consent Order.   Pit 9
      contains approximately  150,000  ft3  of buried  waste,  of  which  approximately
      110,000 ft3 of transuranic waste  was generated at the Rocky Flats Plant.

      The  Pit 9 remedial action consists of three phases:  (1) Proof-of-Process Test
      (POP),  (2)  Limited  Production Test (LPT), and  (3)  Final  remediation.   Data
      produced during POP will  be used  to evaluate subcontractor's performance as part
      of the procurement process, and  to provide the Environmental  Protection Agency
      (EPA)  and the Idaho Department of Health and Welfare (IDHW)  with data justifying
      use  of the  proposed process for remediation of Pit  9.  It is  in the best interest
      of the subcontractors and EG&G Idaho  that  POP data be  identified as reliable,
      accurate, and verifiable by the  U.S. Department of Energy,  EPA, and IDHW.   The
      LPT  phase would demonstrate that  all  integrated systems function as proposed to
      give a high degree of confidence  that all systems are reliable before full-scale
      remediation would  begin.  The data generated during full  remediation will be used
      as evidence that applicable or relevant and appropriate requirements described
      in the Pit 9 record  of decision  had been met.

      EG&G  Idaho  feels  the  credibility  and  legitimacy  of data  produced during  the
      remedial action can  be greatly improved by  independent  data  management.   EG&G
      Idaho will  provide another layer  of data legitimacy through electronic links to
      subcontractor's laboratory instruments  and by  using  an  independent  laboratory
      information management system (LIMS).  EG&G Idaho will  independently store  the
      laboratory instrument  signals, the signal  processing software,  and  the quality
      assurance records/documentation  associated with sampling and/or analysis  to an
      optical disk for future analysis,  if  required, to verify objectives  have  been
      met.

      Data and information that  supports the  3  phases of the Pit 9  remedial  action
      will be received in many forms  such  as photos,  images, graphs, video-text, audio-
      text,  field  notebooks, and  electronic media.  The  system  was developed on  a
      VAX/WMS using Standard Query Language (SQL) Multimedia  for rdB/VMS which  is  a
      solution  for  manipulating,   storing,  and  retrieving  objects   from  client
      applications.  Storing multimedia objects in a database provides data integrity.
      concurrent  access, security,  standard storage format,  and  the ability to  share
      data across applications.  The SQL/services provide the ability to interface to
      the  following  environments:   VMS,  MS-DOS, Microsoft  Windows,  and  Macintosh
      systems.  The system will  provide for storage on optical media using  Write Once
      Read Many (WORM) technology to provide long term  storage,  and  will  provide  the
      necessary tools for rapid  manipulation of the data to controlled user  community
      long term.    Ultimately,  the  purpose  of this system is  to provide  a  secure,
      historical  data repository for multimedia technical information for  the Pit  9
      project.


                                          99

-------
1 p                    QUALITY ASSURANCE PLANS FOR BASIC RESEARCH
 ' °                                AND NEW CONCEPT STUDIES

          EJ. Poziomek, project manager, and A. Cross-Smiecinski, quality assurance officer, Harry
          Reid Center  for  Environmental Studies1, University of Nevada, Las  Vegas,  4505 S.
          Maryland Pkwy., Box 454009, Las Vegas, Nevada, 89154-4009.

          ABSTRACT

          Writing quality assurance project plans (QAPjPs) for basic research and new concept studies
          presents a major challenge to scientists and engineers. Due to the nature of basic research
          studies, few functions are followed repetitively in exactly the same manner, at least in the
          initial stages of the study and sometimes throughout the study. This often precludes the use
          of standard methods, operating procedures, the ability to designate appropriate corrective
          action,  even planning complementary  quality control checks, and other topics  usually
          addressed in a QAPjP.  Recently we were challenged with writing just such a QAPjP for a
          set of  studies.  It was clear that we  needed to maximize the flexibility  of EPA QA
          requirements, and so we chose to try the Category IV guidelines prepared by Guy Simes of
          the Risk Reduction Engineering Laboratory in Cincinnati.  Because the studies all fell into
          the realm of field screening methods involving proofs  of principle and new  concepts, we
          addressed them together in the first version of the QAPjP.  We  soon found  that many of
          the QA topics in  those guidelines were still  difficult to address.  We resolved that the
          QAPjP be written  with statements of commitment that,  as progress is made in the research
          studies, the "holes" in the QAPjP would be filled and eventually, as appropriate, the studies
          would  be broken  out  into  individual QAPjPs.  We also put a major emphasis on
          recordkeeping in order to ensure that each experimental  step, observation,  thought, and
          action be recorded in such a way that  the research can be thoroughly understood, reported,
          and repeated. The purpose of our poster presentation is to provide a detailed explanation
          of how we  prepared the QAPjP  discussed and to give examples of how other research
          studies can  address QAPjP requirements.

          INTRODUCTION

          The Harry Reid Center for Environmental Studies (HRC) has a number of basic research
          and new concept studies, the sponsors of which include the U.S. Environmental Protection
          Agency (EPA)  and the Advanced Research  Projects  Agency (ARPA).   These require
          preparation of Quality Assurance Project Plans (QAPjPs).  If the QA requirements are not
          1  Although the information in this paper has been funded wholly or in  part by the U.S.
       Environmental Protection  Agency under Cooperative Agreement  No. CR  818353 with the
       University of Nevada-Las Vegas, it does not necessarily reflect the views of the Agency, and no
       official endorsement should be inferred. Mention of trade names or commercial  products does not
       constitute endorsement  or  recommendation for use.  Partial support was also provided by the
       Advanced Research Projects Agency under Contract DAAD05-92-C-0027.
                                              100

-------
specified, HRC still has a policy of the need to define data quality objectives (DQOs) for
its research and new concepts studies.  In the process of trying to meet standard QAPjP
requirements (U.S. EPA, 1983) to these types of basic research and new concept studies,
it was discovered that many questions and objectives could not be established beforehand;
they could  not necessarily be anticipated. The QAPjP requirements were difficult and in
some instances not possible to apply to research studies. In searching for Agency-accepted
QAPjP  guidance,  some of the clearest, most concise, reasonable guidance devised was
discovered in the form of the graded approach (Simes, 1991) written by Guy Simes at the
U.S. EPA Risk Reduction Engineering Laboratory.  This groundbreaking set of guidance
documents introduced the concepts that 1)  not all projects can meet quality needs in the
same manner, and  2) QAPjP requirements are more flexible and therefore more supportive
of study objectives than often realized.

This paper describes HRC's adaptation of the graded approach of QAPjP requirements to
basic proof-of-principle and new concept research studies.  It is meant to provide a basic
template to assist investigators in writing QAPjPs.

SUMMARY

The graded approach to composing QAPjPs has four levels of requirements based on the
category of study  or project represented.  Level I requirements are  the most  rigorous
because the projects they address are critical to Agency goals and must withstand legal
challenge (Simes,  1991). In comparison, Level IV requirements, the least rigorous of the
four levels, address studies whose results are used to assess suppositions (Simes, 1991). A
comparison of QAPjP requirements is listed.

        Standard QAPjP requirements                    Level IV QAPjP requirements
        (U.S. EPA, 1983)                                (Simes, 1991)
              Project Description                              Project Description
              Organization and Responsibilities
              QA Objectives                                  QA Objectives
              Site Selection and  Sampling Procedures
              Sample Custody
              Calibration Procedures and Frequency
              Analytical Procedures                            Sampling and Analytical
              Data Reduction, Validation, and Reporting
              Internal QC Checks                              Approach to QA
              Performance and Systems Audits
              Preventive Maintenance
              Calculation of Data Quality Indicators
              Corrective Action
              Quality Control Reports to Management
              References                                      References
              Other Items                                     Other Items

Difficulty in specifying QA Objectives, parts of the Project Description, and a complete
Approach to QA/QC presents problems due to the unavailability of data in new concept
                                        101

-------
and basic research studies.  The solution to this problem lies in the commitment of the
research team to document requirements as answers are obtained and to state intents within
the QAPjP. The attached EXHIBIT is a condensed version of how HRC met the Level IV
QAPjP requirements for one set of basic research studies and commits to the requirements
for which there currently are no answers. Hopefully, research investigators can benefit from
the information and our experience.
                                       102

-------
EXHIBIT

For an explanation of what is needed for each requirement of a Level IV QAPjP, refer
directly to Simes  (1989)  for guidance.  Presented here  is a description  of  how HRC
composed the QAPjP entitled "Proof-of-Principle and New Concepts Studies in Chemical
Sensors, Biosensors, and In Situ Methods for Field Screening." The italicized print indicates
guidance and notes for consideration in composing the missing portions of the  QAPjP.
                                   SECTION 1.0

                             PROJECT DESCRIPTION
1.1    GENERAL OVERVIEW

Field screening methods for hazardous waste site investigations need to be rapid and low
cost to support on-site monitoring and characterization activities.  A variety of in situ and
field portable analytical methods are needed for the analysis of contaminants in all types of
environmental media.  Needs for environmental sensors also go hand-in-hand with needs
for sampling devices and methods.  This project is meant to address these requirements.

There are major barriers which, if overcome, would provide new capabilities in selectivity
and sensitivity of devices for field screening. This applies to both point and remote sensing
applications. For example, the major technology barrier to the development of fiber optic
and other sensors based on chemical/biochemical reactions and/or molecular association
effects is the proper selection of the sensor coating materials.  In most  cases there is also
a challenge to marry the chemistry to the particular device being used.  There is a need to
develop small and low power systems while maintaining selectivity and  sensitivity. There
are a variety of opportunities to exploit including the use of solid phase sensor chemical and
biochemical coatings, self-indicating films, solid-state extraction disks in combination with
spectroscopy, ion mobility spectrometry as an in situ well monitor, spectroelectrochemistry,
and technology integration, to mention a few.

1.1.1   The Process

There are two parts to this project.  The first involves performing applied research on field
screening and measurement methods relevant to  the type of pollutants most frequently
encountered at hazardous waste sites. This is an investment for the long term and is meant
to address the needs for basic information for proof-of-concepts. The second part is for the
short term and involves integration  of existing technologies in the search for new concepts.
The strategy also includes reviewing emerging research and technology, and transferring
appropriate aspects to development of methods for field screening.

Meeting field screening needs presents a major challenge because of the many scenarios
which could be encountered involving any combination of air, water,  or soil.   In addition,
the possible analytes, specifically their number and nature, can present  a major challenge
                                         103

-------
 even to the most experienced chemist. Organic chemicals alone - volatile, semi-volatile, and
 non-volatile - could range from several to hundreds in a particular sample.

 Field screening systems, whether they use microelectronics, electrodes, or fiber optics as
 transducers,  i.e., the physical probe, require chemistry to be  worked out for the specific
 chemical or class of chemicals to be monitored and measured.  Many times not only are
 there no guidelines for practitioners for  a particular application, but  there is also  no
 information base on the fundamental chemistry. For example, major barriers still remaining
 to be solved  in recognition coating selection and evaluation include:

        •      Lack of an information base on reactions and molecular association effects
               in vapor-solid and  liquid-solid  phases  which can be  drawn upon for
               development of sensor coatings.

        •      Lack of guidelines on choosing selective coatings for use in conjunction with
               sensors.

        •      Lack of  standard  or reference methods for  screening  and  evaluating
               candidate coatings in sensor applications.

 1.1.2   Statement of Project Objectives

 The general objective of this project is to design and develop new field screening methods,
 including those based on in situ processes. Priority is given to systems which offer simplicity,
 selectivity, sensitivity, and low cost.  One focus will be to overcome the technology barrier
 posed by nonselective  recognition coatings in sensor monitoring and measurement needs.
 The output can be  applied to a variety of sensors utilizing molecular recognition events
 based on reactions or molecular association effects.  Another focus will be to integrate
 existing technologies toward the development of new concepts in field screening.  Listed
 below are specific objectives for some of the studies under this project taken from research
 and study plans. New and updated project objectives will be documented in revisions of this
 QAPjP and/or in study-specific reports/plans as the objectives develop.

 Potential Use of Ultrasound  in Chemical Monitoring--(Research Plan,  M.S. Thesis in
 Environmental Chemistry, University of Nevada,  Las Vegas:   Grazyna Orzechowska,
 November 1992)
       The major objective of this study is to examine the potential of combining sonication
 with  existing multiparameter monitors such as  chloride  specific  electrodes  in detecting
 organochlorine pollutants in water.  This  is an example of technology integration.

       The approach in using sonication  is obviously applicable to compounds other than
 organochlorine ones; these could contain inorganic elements (other halides,  phosphorus,
nitrogen, and sulfur).  Anions specific to  the inorganic components would be  produced in
sonication. Changes in anion concentrations before and after sonication would be used in
monitoring for the pollutants.  The choice of organochlorine compounds was made because
these compounds are the most common pollutants found at hazardous waste sites. Success
                                        104

-------
with compounds such as trichloroethylene, chloroform, etc., will serve as proof-of-principle
and form a base for expanding the research to other pollutant classes.

Surface Acoustic Wave (SAW) Biosensor for Detection of Environmental Pollutants: Kim
Rogers, 1992
       Objectives of this study are to:  1) demonstrate the feasibility of a receptor- or
antibody-based biosensor  which can  detect a  single compound or several classes of
compounds of environmental concern; and 2) develop  protocols which will explore and
define  the applicability   (i.e.,  limits  of detection,  sample matrix and preparation
requirements, etc.) of this biosensor for field assays. In this case the  recognition coating
may be a chemical one or a biologically-derived material.

1.2    EXPERIMENTAL DESIGN

       Experimental  designs for many  of  the studies  under this  project are  under
development and will be further documented in  revisions of this QAPjP and/or in  study-
specific reports/plans  (as the Project Manager deems necessary) and as the designs are
finalized. Listed here for the benefit of the reader/reviewer are current approaches for the
specific studies addressed in the project objectives section.

Potential Use of Ultrasound in Chemical Monitoring-(Research Plan, M.S. Thesis in
Environmental Chemistry, University  of Nevada,  Las Vegas:   Grazyna Orzechowska,
November 1992)
       As mentioned  in the project objectives section, the study is meant to examine the
potential of combining sonication  with existing multiparameter monitors  in detecting
organochlorine pollutants in water. Success will establish a proof-of-principle for obtaining
more detailed data and for using the concept with other organic pollutant classes. There
are a number  of  factors  which influence the effectiveness  of sonication in a  chemical
reacting system. These will be taken into account in designing the experiments. Examples
follow.

       Choice of the solvent—Factors such as molecular cohesion and surface tension  of the
solvent can affect cavitation.  There is no solvent choice in the proposed research since the
application is in water. This will be the solvent.

       Reaction temperature—It is common practice in sonochemistry to use the lowest
possible  reaction  temperature consistent with reasonable  reaction time.   Keeping
temperature low (thus reducing solvent vapor pressure) minimizes "cushioning" effects which
could reduce the temperatures and pressures in cavitation. The temperature will be kept
close to room temperature using a cooling bath.

       Irradiation frequencies-Frequencies between 20 and  50 kHz are normally used in
sonochemistry.  A Branson Sonifier - Cell Disrupter model 101-063-200 with 20 kHz and an
output power of 400 W will be used for this work.

       Choice of disruptor horns-Disruptor horns (probes) transmit ultrasonic energy into
a solution.  The dimensions of the horn and the ultrasonic output power setting determine
                                         105

-------
 the amount of amplitude (tip movement) and degree of ultrasonic activity in the liquid.
 Generally, the smaller the tip diameter, the higher the amplitude. Larger tip diameters have
 less amplitude but can accommodate larger volumes.
        Three horns have been chosen for this work, a 3/4" diameter high gain horn with
 solid tip, a high intensity 1" diameter cup horn, and a tapered 1/8" diameter microtip horn.
 This will allow some flexibility in designing the experiments.  The tapered microtip horn, for
 example, will  have an amplitude three and a half times greater at its end than a standard
 horn.  It can be used with small solution volumes (1-2 mL).

        Other sonication equipment--A stainless steel continuous flow attachment with a
 water cooling jacket will be available for continuous processing of test solutions.
        Though ultrasound is above the audible range of the human ear, mechanical noise
 occurs when liquids are treated ultrasonically.  A soundproof box will be available to reduce
 this noise.

 Surface Acoustic Wave (SAW) Biosensor for Detection of Environmental Pollutants--(2%
 Initiative Study Description: Kim Rogers, 1992)
        A number of basic  parameters will be considered in the research approach to
 develop sensor recognition coatings. These include choices of the coatings and the test beds.
        Choice of candidate recognition coatings-Several considerations have to be taken
 into account when choosing a recognition coating for sensor applications. First, one must
 decide whether a reversible or irreversible effect is to be sought, then the chemistry of the
 coating has to be considered, for example:

        •      chemical reaction vs. molecular association mechanisms
        •      covalent bond formation vs. weak Van der Waals interaction
        •      hydrogen bond vs. charge-transfer vs. other donor-acceptor mechanism
        •      bimolecular vs. catalytic reaction mechanisms

       The strategy for biosensors is to seek coatings based on known reaction chemistry
 that occurs in solution. However,  a different approach may be needed for gas-solid and
 vapor-solid configurations  in sensor reaction chemistry.  Factors that  may affect sensor
 performance include:

              coating adhesion to  the surface
              mass vs. mechanical sensor coating mechanism
              surface vs. bulk coating sensor effects
              gain or loss  in coating mass with analyte sorption
              boiling point of the coating material
              crystalline solids vs.  amorphous

       Factors such as solubility and molecular weight will be considered in selection of the
candidate coating material.  Though the physical and chemical properties of a coating may
seem  ideal, the material may not  form a  film  conducive to  sensor applications.   Some
experiments may need to be designed by trial and error.
                                        106

-------
       Choice  of test bed (physical transducer)--Test beds for examining the candidate
recognition  coatings will  utilize SAW probes and quartz crystal microbalance (QCM)
sensors.  These devices monitor mass changes.
       The  equipment for measuring SAW sensor effects include a 158 MHz electronic
system, a 4-channel SAW data acquisition system, and a VG-400 automatic vapor generation
system. This sensor is able to detect low ppb levels with a selectivity of 1,000:1 or more.
The QCM is 12 MHz and somewhat less sensitive. However, it is useful for preliminary
experiments.

       Temperature  of the reaction-The  test  system will  be maintained  at constant
temperature, close to ambient conditions for the initial experiments. Different temperatures
using the same coating may be considered to increase selectivity.

       Analytes to be tested—In the biosensor research organic analytes such as PCBs will
be chosen.  The results of the study with PCBs will be extended to other organic analytes.
The biosensor study will address two main challenges:  first, to immobilize an active receptor
or antibodies to piezoelectric devices and, second, to develop methodology which will allow
for sensitive and reversible detection of environmental pollutants.

1.3    SCHEDULE

       The technology integration  activity utilizing sonication is scheduled to run for two
years with a goal to demonstrate proof-of-principle.
       The biosensor effort is also a two-year effort with a goal in the first year to construct
a flow cell and to develop assay protocols for detection of the target compounds. At the end
of two years, it is expected that standard assay protocols will be established and detection
limits determined for target analytes. Matrix effects and sample preparation requirements
for "real world" samples will be determined.

1.4    PROJECT ORGANIZATION AND RESPONSIBILITIES

       This subsection, since it is  self-explanatory, is eliminated to save space in these
proceedings.
                                   SECTION 2.0

                       QUALITY ASSURANCE OBJECTIVES

2.1    QA OBJECTIVES

       Project objectives have  been discussed in Section 1.0.  Because the research
performed under this project is basic proof of principles and new concepts investigations,
specific QA objectives are not specified at this time.  Initially the data will be evaluated
qualitatively or semi-quantitatively.  The data will be used to make decisions whether to
proceed further in evaluating the field screening methods developed  by this project and to
evaluate proofs and concepts.  A future goal of the project is the development of complete
                                         107

-------
 QA objectives for specific activities once proofs-of-principles have been determined and
 concept validation is initiated involving environmentally-related measurement data.

 2.2    QUANTITATIVE QA OBJECTIVES: PRECISION, ACCURACY,  METHOD
       DETECTION LIMIT, AND COMPLETENESS

       See Section 2.1. Specific quantitative QA objectives are not specified at this time.
 As they are developed, they will be included in reports and/or revisions of this document.
 Listed below in italics  are self-notes of intention and guidance to be considered when the
 answers  are  available.  Much  of  this guidance has been  taken  verbatim from a Simes
 document  (Simes, 1989, 1991).

 2.2.1  Precision

       It is believed that it is possible to estimate precision from duplicate determinations when
 the project gets to the stage where measurements are being made.

 2.2.2  Accuracy

       Likewise accuracy may be determined from  measurements of a known.  For example,
 if the field screening method is applied to an SRM and gets  an 80% recovery,  this is an
 indication of accuracy.

 2.2.3  Method Detection Limit

 2.2.4  Completeness (Simes, 1991)
       In  addition, it must be explained how the  QA  objectives are to  be  interpreted in a
 statistical sense.  Often they are interpreted in  a sense that all  data must fall within  certain
 windows; for this type of specification, any data that fail to satisfy QA objectives are rejected and
 corrective action  is  taken.  However, other interpretations are possible.   For example, the
 requirements may be satisfied if the average recovery is within the objectives.  This is where we
 describe how tabulated  QA objectives will be interpreted.

 2.3    QUALITATIVE   QA    OBJECTIVES:  COMPARABILITY   AND
       REPRESENTATIVENESS

       See Section 2.1.

 2.3.1  Comparability

       This is the degree to which one data set can  be compared to another.  For example, to
 evaluate a  cleanup process, analyses of the feed and discharge streams must be comparable.
Another example, to perform a nationwide environmental survey, methods used at different
 locations must be comparable, probably by the use of consistent methods and traceability of
standards to the same or a reliable source (Simes, 1991),
                                         108

-------
       / think that for this project,  comparability would be achieved by use of the same
evaluation methods for each test phase or mechanism. For example, comparability would be
achieved by using the same analytical methods to evaluate each sensor coating.

2.3.2   Representativeness

       This is the degree to which a sample or group of samples is indicative of the population
being studied. An environmental sample is representative for a parameter of interest when the
average value obtained from a group of such samples tends towards the true value of that
parameter in the actual environment, as the number of representative samples is increased.
Representativeness is normally achieved by collecting  a sufficiently large number of unbiased
samples (Simes,  1991).

       Because this study does not involve a sampling exercise, I think that representativeness
is associated with the applicability to the "real world" in two ways: 1) testing a field screening
device on spiked water or sand is not as representative  as testing it in a real field situation; and
2) approaching the true value of the parameter is gained by testing the device a sufficiently large
number of times  in a real field situation.

2.4    OTHER QA OBJECTIVES

       Mass changes for SAW and QCM sensors for "screening experiments" are sought to
be ±  10% based on frequency changes.
2.5    EFFECTS OF NOT MEETING QA OBJECTIVES

       Not meeting QA objectives may give false indications as to the efficacy of candidate
recognition coatings being screened.
                                   SECTION 3.0


                  SAMPLING AND ANALYTICAL PROCEDURES

3.1    SAMPLING PROCEDURES

       This requirement does not apply for proof-of-principle and new-concept studies.
These will be developed as part of each activity.

3.2    PROCESS MEASUREMENTS

       Manufacturers' instructions for any needed process instruments will be followed.

3.3    ANALYTICAL PROCEDURES AND CALIBRATION
                                        109

-------
3.3.1   EPA-Approved or Other Validated Standard Methods

       This requirement does  not apply for proof-of-principle and new-concept studies.
Methods will be developed as part of each activity.

3.3.2   Nonstandard or Modified Methods

       This requirement does  not apply for proof-of-principle and new-concept studies.
Methods will be developed as part of each activity.

3.3.3   Calibration Procedure and Frequency

       Manufacturers' instructions for instruments will be followed as recommended and
necessary.  An example is the calibration protocol for the Microsensor Systems Inc. Vapor
Generator (Appendix A).
                                   SECTION 4.0

                             APPROACH TO QA/QC

4.1    DATA RECORDKEEPING

       Due to the very basic nature of the research  performed  for  this project, few
functions will  be  followed repetitively in exactly the same manner,  at least in the initial
stages of the studies.  Therefore, it  is important that each experimental step, observation,
thought, and action be recorded to document the entire study in such a way that it can be
thoroughly understood, reported,  and repeated.   Accurate  and legible records of the
research are noted in permanently-bound scientific notebooks.  The scientific notebooks are
prepared, controlled, and reviewed using the HRC standard operating procedure, scientific
notebooks (Appendix A).

4.2    CALCULATION OF RESULTS (Simes, 1991)

       This section is primarily reserved to show how analytical results will be manipulated to
prove or disprove  a hypothesis.  TJiis section should provide formulas and summarize any
statistical procedures for reducing the data, including units and definitions of terms. Also define
procedures that will be employed for determining outliers or flagging data.

       If mass balance calculations are required, provide  exact formulas relating the mass
balance to the individual measurements that will be made. Specify data reporting requirements
or plans at this point.  Indicate units,  matrices, and wet or dry intentions, etc. List deliverables
and data deliverables for the report.  Will the data package include raw data? What type of QC
data will be reported?
                                         110

-------
       The final report should also include a summary of the original QA objectives and a
statement of whether they were met or further developments in this arena.  If project/QA
objectives were not met, explain the effect on the study or project.

4.2    INTERNAL QUALITY CONTROL (Simes, 1991)

       Describe the nature and frequency of all QC methods. The QC procedures should relate
to the QA/project objectives. They should be designed to support the objectives.  Examples of
QC checks to consider follow:

• Samples

       Collocated, split, replicate

• Spikes

       Spiked and duplicate spiked samples
       Spiked blanks
       Surrogate spikes and internal standards

• Blanks

       Sampling, field, trip, method, reagent,  instrument
       Zero and span gases

• Others

       Standard reference materials
       Mass tuning for mass analysis
       Confirmation by second column (for GC) or instrument
       Control charts
       Independent check standard
       Determinations for detection  limits
       Calibration standards
       Proficiency testing of analysts
       Any additional checks required by special needs of study

       The use of ambiguous  terms  is a common problem in  this section.   The terms
"duplicate" and "replicate" are examples. Explain the exact point in the study where replication
occurs. Does "replicate" refer to samples collected simultaneously or sequentially in the field;
to samples collected at the same sampling point but at different times; to samples that are split
upon receipt in the lab; or to samples collected under yet another splitting protocol?

       Similarly,  there are numerous types of blanks; the exact procedure for preparing these
blanks must also be described.

4.3     CALCULATION OF DATA QUALITY INDICATORS
                                         111

-------
4.3.1   Precision - (Simes, 1989)

       If calculated from duplicate measurements:
               RPD  =
                                       l - C2) x  100
                                       (C, - C2)/2
       RPD  =  relative percent difference,
       Cj     =  larger of the two  observed values, and
       C     =  smaller of the two observed values.
       If calculated from three or more replicates, use relative standard deviation (RSD)
       rather than RPD:
                                RSD = (s/y) x 100  ;

              RSD   = relative standard deviation,
               s     = standard deviation, and
               y     = mean of replicate analyses.

       When s is defined as follows:
y
n
                                 s =
                                               2
                             (y,-y)
                             n-1
                       standard deviation,
                       measured value of the i-th replicate,
                       mean of replicate measurements, and
                       number of replicates.
4.3.2   Accuracy - (Simes, 1989)

       For measurements where spikes are used:
                              %R = 100  x
                                             S-U
                                        112

-------
              %R    = percent recovery,
              S      = measured concentration in spiked aliquot,
              U     = measured concentration in unspiked aliquot, and
              C^     = actual concentration of spike added.
       For situations where a standard reference material (SRM) is used instead of, or in
       addition to, spikes:

                               %R = 100 x
                                            C
                                             srm
              %R= percent recovery,
              Cm=  measured concentration of SRM, and
              Csrm= actual concentration of SRM.

4.3.3   Completeness (sampling and analytical) - (Simes, 1989)

                                             \V~\
                                %C  = 100 x \ — \  ;
                                             \T\
              %C    = percent completeness,
               V     = number of measurements judged valid, and
               T     = total number of measurements.
              (statistical) - (Simes, 1989)

       Defined as follows for all measurements:

                                            \
                                %C =  100 x \ — \  ;
              %C    = percent completeness,
              V     = number of measurements judged valid, and
              n      = total number of measurements necessary  to achieve a specified
              statistical level of confidence in decision making
4.3.4   Method Detection Limit

       Defined as follows for all measurements:

                              MDL=tn.1> !.„=„.„) xS
                                        113

-------
             MDL  = method detection limit

             S      =      standard  deviation   of   the   replicate
                           measurements

             Vi i-«=099)     = Students' t-value appropriate to a 99% confidence level
                           and a  standard  deviation  estimate with n-1 degrees of
                           freedom.
                                  SECTION 5.0

                                 REFERENCES


                                 APPENDIX A

                                 PROCEDURES


ACKNOWLEDGEMENTS

Discussion  with William H. Engelmann, Kim  R.  Rogers, and Eric N.  Koglin of EPA,
Environmental  Monitoring Systems Laboratory-Las  Vegas, Grazyna  Orzechowska,
University of Nevada-Las Vegas, and James Petrowski, Naval Explosive Ordnance Disposal
Technology Center, Indian Head, Maryland, is gratefully acknowledged.

REFERENCES

1.      Simes, Guy F.  Preparation Aids for the  Development of Category IV Quality
       Assurance Project  Plans.   EPA/600/8-91/006.   U.S.  Environmental  Protection
       Agency, Cincinnati, OH, 1991.

2.      Simes, Guy F. Preparing Perfect Project Plans. A Pocket guide for the Preparation
       of Quality Assurance Project  Plans.  EPA/600/9-89/087.   U.S. Environmental
       Protection Agency, Cincinnati,  OH, 1989.

3.      Interim Guidelines and Specifications for Preparing Quality Assurance Project Plans
       QAMS-005/80.   EPA-600/4-83-004.   U.S.  Environmental Protection  Agency,
       Washington, DC, 1983.
                                       114

-------
19
DATA VALIDATION GUIDANCE FOR CONVENTIONAL WET CHEMISTRY ANALYSES

 Ann Rosecrance. Corporate Quality Assurance Director, Analytical Chemistry Division, Core
 Laboratories, and LaDonna Kibler,  Quality Assurance  Administrator,  Analytical Chemistry
 Division, Core Laboratories,  10205  Westheimer, Houston, Texas 77042

 Abstract:  Data validation is the process of determining the compliance of analytical data with
 established method criteria and project specifications.  The validation of environmental data
 assesses the quality of the data generated for environmental measurements and determines if the
 applicable analytical method requirements and project data quality objectives were met.  This
 paper provides guidance for the validation of data for conventional parameters obtained by wet
 chemistry analysis methods.  Summaries  of method  specified quality control (QC) criteria  for
 select wet chemistry methods are provided, in addition to general recommended QC guidelines.
 Examples of data review checklists  to use in conducting and documenting data validation  are
 included.   A standard approach is presented for  the review  and validation  of  data  for
 environmental samples analyzed for conventional parameters by wet chemistry analysis methods.

 INTRODUCTION

 Analytical methods from the U.S.EPA, Standard Methods, and  other sources are available  for
 the determination of conventional parameters in various sample  matrices by wet chemistry and
 instrumental techniques.  Each method and associated QC sections define specific requirements
 associated with application of the method; additional requirements may be  further specified by
 the governing regulatory agencies or defined in the quality assurance project plan and associated
 data quality objectives.  Laboratory analysts and data reviewers need to be familiar with  the
 requirements of the analytical methods that are routinely used  in order to ensure that  the
 appropriate procedures are followed and that the required criteria are achieved.  While  the
 available methods for organic and  inorganic  environmental analyses provide detailed and
 extensive specifications on QC and other technical requirements, the wet chemistry methods do
 not always include the same level of detail or specifications.  Relevant guidance for wet
 chemistry analyses is needed in order to validate data for conventional parameters.

 Data validation  activities  determine  if analytical data are in compliance  with the analytical
 method requirements and project specifications.  Data validation procedures developed by EPA
 for specific programs are used as standards for data validation.1'2'3  A previous publication
 provided method comparisons and data validation guidance for  EPA organic  and inorganic
 analysis methods for the determination of volatiles, semivolatiles, pesticides/PCBs, and metals
 in ambient air,  drinking water, wastewater, solid waste,  and hazardous waste.4  In this paper,
 an overview of QC requirements and recommended guidelines for select wet chemistry methods
 are provided.  A standard approach to data validation is presented  that can be applied to  the
 validation of data for conventional parameters obtained from wet chemistry analysis methods.

 The wet chemistry analyses covered include methods from EPA Methods for Chemical Analysis
 of Water and Wastes, EPA SW-846 Solid Waste methods,  EPA Contract Laboratory Program
 (CLP)  Statement of Work (SOW)  for Inorganic Analysis, and Standard  Methods  for  the
 Examination of Water and Wastewater where EPA methods are not specified.  The conventional
                                             115

-------
parameters covered  include  inorganic and organic non-metallics analyzed  by colorimetric,
gravimetric, titrimetric, turbidimetric, spectrophotometric, and other instrumental techniques.

DATA VALIDATION DEFINITION

Data validation has been defined by EPA as a systematic process, consisting of data editing,
screening,  checking, auditing,  verification, certification,  and  review in  order  to provide
assurance that data are adequate for their intended use.5  Data validation can be considered a
question and  answer  process  to determine if the data meet both  the  analytical method
requirements and the associated project specifications.  The four major questions to assess in
validating data are the following: (1) were the required technical and QC analyses  performed,
(2) were they included at the required frequency, (3) were the required acceptance criteria met,
and (4) were the project specifications or data quality objectives met.

DATA VALIDATION PROCESS

The data validation process  follows a step by step procedure  for  reviewing the data for
completeness, correctness,  and acceptability.  Specific procedures for data validation may be
found  in EPA data validation  documents  or  project specified  data  validation procedures.
Presented here is  a  generalized procedure for the  data validation process as applied to  wet
chemistry analyses.  After defining the project requirements and obtaining the data package, the
first step in the data  validation process should be a data completeness check to determine if all
of the required data and documentation are present.  If any of the required data or documentation
are missing, then the needed information should be obtained, if possible, before conducting the
data validation. The next recommended step in the process is a review of the QC data associated
with the analysis of the sample batch, referred to here as method QC. Next is a review of the
sample data and related QC information for each sample, followed by an overall review of all
of the data and associated documentation.  Any data that does not meet the method requirements
or project specifications at any of these steps should be qualified appropriately.  The final step
should be the preparation and  distribution of a data validation report.  A recommended sequence
for the data validation process that includes each of these steps is provided in Figure 1.

DATA COMPLETENESS  CHECK

In the first step of the data validation process, the data package should be reviewed to determine
if all of the required  data and associated documentation are present.  This check should include
a review of the analysis specifications to ensure that the required method, target analytes and
their reporting limits, and data quality objectives are defined. The next step should be a review
of the  documentation on sample custody and condition  in order to determine if custody was
maintained and if the sample was received and maintained in the proper condition.  The data
report, QC results, and associated raw data for each analytical parameter should be reviewed to
ensure that all required information is present. Documentation for sample preparation, standards
preparation,  standards  and  sample analysis,  data  interpretation  and  calculations,  and
correspondence should be reviewed to ensure that all of the required information is present. An
example  checklist to use for performing the data completeness check is provided in Figure 2.
                                        116

-------
DATA VALIDATION APPROACH

The recommended approach for performing data validation is to first summarize the analytical
method requirements, associated reference QC  requirements, and  the project specifications.
Summaries of the QC and other technical requirements are prepared for reviewing and validating
data for each target analyte. An example of a summary chart comparing the QC requirements
for select cyanide analysis methods is provided in Table 1. The data validation process proceeds
by following closely to the sequence of the analysis and the procedures established by EPA for
data validation. Data generated from any of the specified methods are reviewed for compliance
with each requirement and applicable criteria using general or method  specific data review
checklists.  Examples of a general checklist for wet chemistry analyses and a method specific
checklist for cyanide analysis  by the EPA CLP SOW for Inorganic Analysis are provided in
Figures 3 and 4, respectively.  Using summary charts that provide the required criteria and
checklists that  record compliance with the applicable criteria, a standard approach to data
validation can be performed effectively and efficiently for wet chemistry analysis methods.

QUALITY CONTROL REVIEW

The types of analyses that are subjected to data validation are method specified quality control
and sample specific quality control.  Method quality  control refers to the analyses  that are
necessary for initiating sample analyses and that are common to the sample analysis batch.  This
includes instrument calibration, calibration verification standards, blanks, laboratory control
standards,  spikes, and duplicates.  Also included in method QC are other analyses  that are
necessary to assess the field and laboratory procedures related to the sample data.  This includes
container certifications,  field blanks, field replicates, detection limit determinations, precision
and accuracy determinations,  and performance  evaluation analyses.  Sample quality control
refers  to the criteria that are  specific to  each  sample.  This includes holding times, sample
preparation and analysis, and the identification and quantitation of target analytes.

The following  section provides a summary of items to review during the validation of wet
chemistry  data.  Comparisons  of the  QC requirements for cyanide analysis  methods and
summaries of the method specified QC requirements for other wet chemistry analyses are
provided in Tables 1 and 2, respectively. Because QC requirements for wet chemistry analyses
are often undefined or not specified, general QC guidelines are provided in Table 3 that may be
considered when method or project specified criteria are  not  available.   These general QC
guidelines are not intended as  a replacement for specific requirements defined in the reference
method or project specifications, but are presented as guidelines for performing  data validation
when QC requirements are not specified.  Further details on QC requirements are provided in
the reference methods and  EPA or project specific data validation procedures.

Instrument Calibration.  Instrument calibration data are evaluated to ensure that the analysis and
the associated instrument were set up properly and that acceptable quantitative and qualitative
data were generated at the beginning of the analysis.   Initial calibration data are reviewed for
the analysis of the required analytes, at the required number of levels and concentrations, at the
required frequency, and within the required or recommended acceptance criteria (i.e., a linear
calibration  curve  with a correlation coefficient of > 0.995).
                                        117

-------
Calibration Verification. Calibration verification analyses are evaluated to ensure that acceptable
quantitative and qualitative data were generated during the course of and at the end of the
analytical run.   Initial and continuing calibration verification data are reviewed for the analysis
of the required analytes, from the correct source, at the required concentrations, at the required
frequency, and within the required acceptance criteria (i.e.,  % recovery).

Blanks.   Data  for  calibration  blanks and  preparation  blanks are evaluated  to  check  the
background from the analysis and preparation procedures, respectively. Blank data are reviewed
for the analysis of the correct source of material, at the required frequency, and within the
acceptable background levels for the target analytes.

Laboratory  Control  Standards  CLCS).   LCS analyses are  evaluated  to check the  overall
performance of the laboratory, including both sample preparation and analysis procedures, on
each method utilized.  LCS data are reviewed for the analysis of the correct type of standard,
at the required frequency,  and within the required acceptance criteria (i.e., % recovery).

Spikes.  Data for  spikes are evaluated to check if the analytical performance was within the
accuracy specifications that have been established for the method.  Blank spikes  determine the
laboratory performance for recovery of analytes in blank matrices and matrix spikes assess the
effect of the sample matrix on analyte recovery.  Spike data are reviewed for analysis of the
correct type of spike  (blank or matrix spike and pre-preparation or post-preparation), at  the
required frequency, with  the correct  analytes at the required concentration,  and within  the
required acceptance criteria (i.e., % recovery).

Duplicates.  Data for duplicates are evaluated to check if the analytical performance was within
the precision specifications that  have been established for the method.  Duplicate sample data
are reviewed  for analysis  of  the correct type of duplicate (sample duplicate or matrix spike
duplicate), at the required frequency,  and within the required acceptance criteria for relative
percent difference or other specified duplicate evaluation criteria.

Other  Quality  Control.   Other quality control refers to additional  analyses associated with
evaluation of the  acceptability  of the sampling and analysis procedures.  Quality  control
measures  associated with sampling include the evaluation of field blanks and replicate samples
to determine background contamination and sampling precision, respectively. Other laboratory
quality  control measures include instrument detection limit determinations and  performance
evaluation analyses.  Data for each quality control  analysis are reviewed to determine if the
required analyses were performed at the required frequency and if the results were within the
required acceptance criteria.

Sample Quality Control. Sample data are evaluated to ensure that holding times were met, that
preparation  and analysis procedures were performed correctly, that the analytes were reported
correctly  (both  identification and quantitation), and that reported  values were within  the
calibration range or linear range.  The adherence to holding times should be reviewed for each
parameter as holding  times are variable and depend upon the analyte and method specification.
Raw data for all samples should be reviewed and recalculated to determine if the reported results
correlate with the raw data.
                                        118

-------
DATA VALIDATION DOCUMENTATION

Data validation procedures should be documented on standardized forms such as the example
data review checklists provided in Figures 3 and 4.  The checklist or other form should report
the adherence or lack of adherence to each of the method requirements or project specifications.
The agreement of the raw data and the data report should be determined and documented.  Any
major or minor  deficiencies  identified  should be documented  in a  data  validation report
describing each deficiency  and its potential impact on the sample results.  Qualifiers used for
data in question should be in accordance with the project specifications and they should be
clearly defined.  Examples of qualifiers used in EPA data validation procedures are: (R), the
results are rejected due to serious deficiencies in quality control criteria; (J), the associated
numerical value is an estimated quantity because certain quality control criteria were not  met;
(N), presumptive evidence of presence of material; (U) the material was analyzed for but not
detected; and (UJ), a combination of U and J.1  The data validation  report should include an
overall assessment of the data, in addition to any recommendations for further action.

SUMMARY

Data validation is an integral part of the environmental  data generation process and in order to
be efficient and effective, the data validation process must be versatile and widely applicable.
With  the large number of analytical methods that are available for  environmental sample
analyses,  laboratory analysts and data reviewers must be familiar with the QC requirements for
each method and project.   The information  provided in this  paper summarizes  the QC
requirements for  several wet chemistry analysis methods and provides  general QC guidelines
when method specifications are not available.  Data validation  guidance, including  example
checklists for performing and documenting data validation, is provided so the data reviewer can
use a standard approach when validating data from various analytical  methods.  This guidance
provides an effective and efficient approach for addressing the specific requirements of each
method  utilized and for determining  if the applicable QC  specifications  were met.  This
information is not intended as a replacement for the analytical methods or EPA data validation
procedures, but is a guide on QC requirements and data validation for wet chemistry analyses.

REFERENCES

1. U.S.EPA Quality Assurance/Quality Control Guidance for Removal Activities. Sampling QA/OC Plan
   and Data Validation Procedures. EPA 540-G-90 004, April 1990.

2. U.S.EPA Laboratory Data Validation Functional Guidelines for Evaluating Inorganics Analyses. 1988.

3. U.S.EPA Laboratory Data Validation Functional Guidelines for Evaluating Organics Analyses. 1991.

4. A.E.Rosecrance, "Data Validation Guidance for EPA Organic and Inorganic Analytical Methods",
   Proceedings of the EPA Eighth Annual Waste Testing and Quality Assurance Symposium. Arlington,
   VA, July  1992.

5. U.S.EPA Interim Guidelines and Specifications for Preparing Quality Assurance Project Plans.
   QAMS-005-80, December 1980.
                                        119

-------
Figure 1.  DATA VALIDATION PROCESS
       Obtain Project
   Specifications and Data
     Quality Objectives
Obtain Data Validation
   Requirements
                                        J
/ wuicuil vtua r«iui\aye rr
1
Perform Data
Completeness Check
- Use Checklist
N
                Perform Method QC Review
                     - Use Checklist
                      Was Method
                      Criteria Met?
                Perform Sample QC Review
                      Use Checklist
                         Qualify Data
                      Was Sample
                      Criteria Met?
                Perform Overall Evaluation
-^ ^-\_
Were Project Specifications ^"~-~^( . .
and Data Quality Objectives Mef ^
^^ ^^^^
^^T^^
Yes
*
/ Prepare and Distribute

7
Qualify Data


                  Data Validation Report /
                            120

-------
               Figure 2.  Data Package Checklist
REVIEW ITEM
ANALYSIS SPECIFICATIONS



Project Specifications
Method Reference
Target Analytes and DL/QL
Specifications
SAMPLE CONDITION / CUSTODY


Chain of Custody Records
Sample Receipt/Log-ln Records
DATA REPORT









Case Narrative
Client ID/Lab ID
Parameters
Matrices
Dates of Preparation and Analysis
Method Reference
Required Forms
Sample Results/DLs
QC Results
RAW DATA





Organic
Metals
Wet Chemistry
Radiochemistry
Other
LOGS /RECORDS





Sample Preparation
Sample Analysis/Run Logs
Standards Preparation
Calculations
Correspondence
INCLUDED





























NOT INCLUDED





























NA





























COMMENTS





























OTHER
                               121

-------
              Figure 3.  Wet Chemistry Data Review Checklist
REQUIREMENT
A. CALIBRATION CURVE
1. Frequency: Daily?
2. Levels: Blank +• £ 3 or
standards?
3. Range: Appropriate?
4. Correlation coefficient 2 0.995?
B. LABORATORY CONTROL STANDARD
1 . Frequency: Each batch?
2. Source: Independent?
3. Reference: Certified?
4. % Recovery: 90- 110% or ?

C. BLANK
1 . Frequency: Each batch?
2. Matrix: Matrix specific?
3. Preparation: Entire procedure?
4. Analytes concentration: 

-------
Figure 4.  DATA REVIEW CHECKLIST FOR CYANIDE ANALYSIS DATA BY EPA CLP SOW
REQUIREMENT
A. CALIBRATION
1 . Frequency: Daily?
2. Levels/Range: Blank + 3 stds (1 at CRDL)?
3. 'Correlation coefficient: £ 0.995?
B. CONTINUING CALIBRATION
1. Frequency: Beginning/end/every 10 samples-2 hrs?
2. Level: Mid-range for CCV?
3. % Recovery: 85-115%?
C. OTHER STANDARDS (DISTILLED)
1 . Frequency: 1 per batch?
2. Level: Mid-range?
3. %Recovery: 85-115%?
D. CALIBRATION BLANK
1 . Frequency: Beginning/end/every 1 0 samples-2 hrs?
2. Analyte concentration: 
-------
         Table I.   Comparison of QC Requirements for EPA Cyanide Analysis Methods
Requirement
Water and Waste-
water Analysis

Method 335.2
RCRA Solid
Waste Analysis

Methods 9010/901OA
Sunerfund Hazardous
Waste Analysis

CLP SOW ILM03.0
Method Detection Limit
Titration: 1mg/L
Colorimetric: 0.02 mg/L
Titration: 0.1 mg/L
Colorimetric: 0.02 mg/L
CRDL: 10ng/L
Holding Time
14 days (24 hours when       14 days
sulfide is present)
                            12 days from sample receipt
Initial Calibration'1'
    Frequency
6 standards and a blank
Daily
6 standards and a blank

Daily
3 standards and a blank
(one standard at the CRDL)
Daily
Calibration Verification121
Frequency
Criteria
Other Standards (Distilled)
Frequency
Criteria
Calibration Blanks
Frequency
Criteria
Preparation Blanks
Frequency
Criteria
Laboratory Control Standard
Frequency
Criteria
Matrix Spike Samples
Frequency
Criteria
Duplicate Samples
Frequency
Criteria
NS
NS
NS
High and low standard
1 each per batch
90-1 10% Recovery
Colorimetric: 1 per batch
Use in initial calibration
Titration: 1 per batch
Colorimetric: Not specified
Titration: Use in calculation
Colorimetric: Not specified
NS
NS
NS
1 per batch to check
distillation efficiency
NS
NS
NS
Mid-range standard
Every 15 samples
85- 11 5% Recovery
High and low standard
1 each per batch
90-110% Recovery
Colorimetric: 1 per batch
Use in initial calibration
Titration: 1 per batch
Colorimetric: Not specified
Titration: Use in calculation
Colorimetric: Not specified
Independent check standard
1 per batch
85-115% Recovery
Matrix spike and matrix
spike duplicate per batch
NS
1 matrix spike duplicate
per batch
NS
CCV: Mid-range standard
Beginning, end, and every
10 samples or 2 hours
85-115% Recovery
Mid-level standard
1 per batch
85-1 15% Recovery
Colorimetric: Beginning, end and
every 10 samples or 2 hours
£CRDL
Titration: 1 per batch
Colorimetric: 1 per batch
Titration: Use in calculation
Colorimetric: <, CRDL
Distilled independent standard (ICV)
1 per batch
85-1 15% Recovery
1 per matrix per concentration
level per batch
75-1 25% Recovery
1 per matrix per concentration
level per batch
< 20% RPD for values > 5x CRDL
Other Method Criteria
 Verify sample pH £ 12;
 Check for oxidizing agents
 and sulfides
Verify sample pH > 12;
Check for oxidizing agents
and sulfides
Verify sample pH £ 12;
Check for oxidizing agents
and sulfides
(1) Calibration standards must be distilled tor EPA Methods 335.2 and9010/9010A when sulfides are present In the samples.
(2) CLP SOW specifies that the initial calibration verification standard (ICV) be distilled and analyzed as the laboratory control standard (LCS).
   For complete Information, refer to the method reference.
Acronym Definitions
    CCV = continuing calibration verification
    CRDL = contract required detection limit
    ICV = initial calibration verification
                   LCS = laboratory control standard
                   NS = not specified
                   RPD = relative percent difference
                                                        124

-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES
PARAMETER (METHOD NO.)
Acidity (305.1)'
Alkalinity (310.1)'
Anions (300.0)'
by ion chromatography
BODandCBOD (405.1)'
by electrode (5210 B)2
Boron (4500-B)2
by colorimetry
Bromide (320.1)'
by titration
Carbon (TOO (415.1)'
by combustion or oxidation
Chloride (325.3)'
by Hg(N03)j titration
Chloride (9252)3
by Hg(N03)2 titration
Chloride (4500E-CI- B)2
by AgN03 titration
COD (HACH 8000)'
by colorimetry
COD (Mid-level) (410. 1)1
by titration
COD (Low Level) (410.2)'
by titration
Color (110.2)'
by platinum-cobalt
Corrosivity (1110)3
toward steel
CALIBRATION CURVE
Frequency
pH Calib.
Daily/2
pH Calib.
Daily/2
Daily
Daily
Daily

Daily



Daily/
Range




Acceptance
Criteria
pH Meter
Specification
pH Meter
Specification
± 10% of
True Value
DO Meter
Specification
Linear

Instrument
Specification



Instrument
Specification




BLANKS
Frequency


1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Acceptance
Criteria


< Detection
Limit
£ 0.2 mg/L
Zero
Instrument
Blank
Correcting
< Detection
Limit
Blank
Correcting
Blank
Correcting
Blank
Correcting
Zero
Instrument
Blank
Correcting
Blank
Correcting
< Detection
Limit
Blank
Correcting
SPIKES
Frequency


10% or
1 /Batch

5% or
1 /Batch



5% or
1 /Batch
5% or
1 /Batch





Acceptance
Criteria


Not
Specified

80-120%



Not
Specified
80-120%





DUPLICATES
Frequency



10% or
1 /Batch
5% or
1 /Batch



10% or
1 /Batch
5% or
1 /Batch




5% or
1 /Batch
Acceptance
Criteria



See Table
1020:l2
See Table
1020:l2



Not
Specified
See Table
1020:l2




Not
Specified
OTHER
Type
pH Check
pH Check

GIG Std.;
Seed
Control;
LCS
LCS



LCS
LCS





Frequency
1 /Batch
1 /Batch

2/Batch;
2/Batch;
Daily
1 /Batch



Every 15
Samples
1 /Batch





Acceptance
Criteria
± 0.05 of
True Value
± 0.05 of
True Value

198 ± 30.5
mg/L;
0.6-1.0
mg/L;
90-110%
90-110%



Not Specified
90-110%






-------
                            Table II.  METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES  (continued)
PARAMETER (METHOD NO.)
Phenol (9065)'
by spectrophotometry
Phenol (420.1)'
by spectrophotometry
Phenol (420.2)'
by automated 4-AAP
Phos. Total «. Ortho (365.2)'
ascorbic acid, single reagent
Pho«. Total & Ortho (365.3)'
ascorbic acid, two reagent
Sulfate (375.4)'
by turbidimetry
Sulfida (376.2)'
by colorimetry
Sulfite (377.1)'
by titration
Surfactants, Anionic (425.1)'
as MBAS (5540C)2
Total Petroleum (418.1)'
Hydrocarbons (TPH) by IR
Turbidity (180.1)'
CALIBRATION CURVE
Frequency
Daily
Daily
Daily
Daily
Daily
Daily
Daily

Daily
Daily
Daily
Acceptance
Criteria
Linear
Linear
Linear
± 2% of
True Value
Linear
Linear
Linear

Linear
Linear
Instrument
Specification
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Solvent
Blank:
1 /Batch

Acceptance
Criteria
< Detection
Limit
Zero
Instrument
Establish
Baseline
Zero
Instrument
Zero
Instrument
Zero
Instrument
Use in
Calibration
Blank
Correcting
Use in
Calibration
Zero
Instrument

SPIKES
Frequency
10% or
1 /Batch







5% or
1 /Batch


Acceptance
Criteria
Not
Specified







80-120%


DUPLICATES
Frequency








5% or
1 /Batch


Acceptance
Criteria








See Table
1020:l2


OTHER
Type
LCS




LCS


LCS

Check
Standard
Frequency
Every 1 5
Samples




Every 3 or
4 Samples


1 /Batch

1 /Batch
per Range
Acceptance
Criteria
Not Specified




Not Specified


90-110%

Instrument
Specification
ro
O)
           CCV Continuing Calibration Verification
                                                  G/G  Glucose/Glutamic Acid Standard
LCS Laboratory Control Standard
                                                                                                                         RPD Relative Percent Difference
             Methods for Chemical Analysis of Waster and Wastes, EPA-600/4-79-020, U.S. EPA, Cincinnati, Revised March 1983.



           ' Standard Methods for the Examination of Watar and Wastewater, APHA-AWWA-WPCF, 17th. Edition, 1989.



           ' Test Methods for Evaluating Solid Waste. Physical/Chemical Methods. SW-846, U.S. EPA, Washington D.C., Third Edition, September 1986.



           ' Gibbs, C. R., "Introduction to Chemical Oxygen Demand", HACH Company, Technical Information Series-Booklet No. 8, 1992.



           1 Statement of Work for Inorganics Analysis. Multi-Media Multi-Concentration. U.S. EPA Contract Laboratory Program, ILM03.0.



           ' O93-8O. "Ta*t Methods for Rash Point by Pansky-Martans Closed Tester", American Society for Testing and Materials, Philadelphia.

-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES (continued)
PARAMETER (METHOD NO.)
HaxavalentCr (7197)3
by chelation and FLAA
Iodide (345.1 )'
by titration
Nitrogen - Ammonia (350. 1)1
by automated colorimetry
Nitrogen - Ammonia (350.2)'
by colorimetry
Nitrogen - Ammonia (350.2)'
by ion electrode (350. 3)1
Nitrogen - Nitrate (352.1)'
by colorimetry
Nitrogen - Nitrite (354.1)'
by colorimetry
Nitrogen - Nitrate & Nitrite
by hydrazine reduction (353.1)'
Nitrogen - Nitrate & Nitrite
by auto. Cd reduction (353.2)'
Nitrogen - TKN (351.3)'
by titration
Oil and Grease (413.2)'
by IR
Oil and Grease (413.1)'
by gravimetry
pH in Soil (9045)3
pH in Water (150.1)'
CALIBRATION CURVE
Frequency
Daily

Daily
Daily
Daily
Daily
Daily
Daily
Daily

Daily

pH Calib.
Daily/2
pH Calib.
Daily/2
Acceptance
Criteria
Linear

Linear
Linear
Instrument
Specification
Linear
Linear
Linear
Linear

Linear

pH Meter
Specification
± 0.05 of
True Value
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Solvent
Blank:
1 /Batch
Solvent
Blank:
1 /Batch


Acceptance
Criteria
< Detection
Limit
< Detection
Limit
Establish
Baseline
Zero
Instrument
< Detection
Limit
Zero
Instrument
Zero
Instrument
Establish
Baseline
Establish
Baseline
Blank
Correcting
Zero
Instrument
Blank
Correcting


SPIKES
Frequency
10% or
1 /Batch













Acceptance
Criteria
85-115%













DUPLICATES
Frequency
5% or
1 /Batch











5% or
1 /Batch
Every
Sample
Acceptance
Criteria
Not
Specified











Not
Specified
± 0.01 pH
units
OTHER
Type
LCS;
Method of
Standard
Addition


Distilled
Standards
Distilled
Standards


Efficiency
Check
Efficiency
Check
Distilled
Standards


Check
Standard
pH Check
Frequency
15%;
All EP Tox
Extracts


High&
Low/Batch
High&
Low/Batch


1 /Batch
1 /Batch
HighS,
Low/Batch


1 /Batch
1 /Batch
Acceptance
Criteria
Not
Specified;
Linear
Curve


Comparable
to Undistilled
Standards
Comparable
to Undistilled
Standards


100%
Reduction
100%
Reduction
Comparable
to Undistilled
Standards


Not Specified
± 0.05 of
True Value

-------
Table II. METHOD SPECIFIED QUALITY CONTROL FOR WET CHEMISTRY ANALYSES (continued)
PARAMETER (METHOD NO.)
Cyanide, Total (335.2)'
by titration
Cyanide, Total (335.2 CLP-M)'
by titrntion
Cyanide, Total (335.2)'
by colorimetry
Cyanide, Total (335.2 CLP-M)*
by colorimetry
Cyanide, Total (335.3)'
by automated UV
Cyanide, Total (335.2 CLP-M)'
by semi-automation
Rash Point (1010)' (D93)'
by Pensky-Martan Closed Cup
Fluoride (340.2)'
by ion electrode
Fluoride (340.3)'
by automated colorimetry
Halides (TOX) (9020)'
Hardness (130.2)'
by titration
Hexavalent Cr (218.4)'
by chelation and FLAA
CALIBRATION CURVE
Frequency


Daily
Daily
Daily
Daily

Daily
Daily
Daily in
Duplicate

Daily
Acceptance
Criteria


Linear
Linear
Linear
Linear

Instrument
Specification
Linear
± 3% of
True Value

Linear
BLANKS
Frequency
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch

1 /Batch
1 /Batch
1 /Batch
1 /Batch
1 /Batch
Acceptance
Criteria
Blank
Correcting
Blank
Correcting
Use in
Calibration
S Contract
Required
Detection
Limit
Use In
Calibration
s Contract
Required
Detection
Limit

< Detection
Limit
Establish
Baseline
< Detection
Limit
< Detection
Limit
< Detection
Limit
SPIKES
Frequency
1 /Batch
5% or
1 /Batch
1 /Batch
5% or
1 /Batch

5% or
1 /Batch



10% or
1 /Batch


Acceptance
Criteria
Distillation
Check
75-125%
Distillation
Check
75-125%

75-125%



Not
Specified


DUPLICATES
Frequency

5% or
1 /Batch

5% or
1 /Batch

5% or
1 /Batch



Every
Sample

Aspirate
Samples in
Duplicate
Acceptance
Criteria

:S 20%RPD

S 20%RPD

S 20%RPD



Not
Specified

Report
Average
OTHER
Type
Distilled
Standards
LCS;
Distilled
Standard
Distilled
Standards
LCS;
CCV;
Distilled
Standard

LCS;
CCV;
Distilled
Standard
p-Xylene
Standard


LCS;
Nitrate
Blank;
Efficiency
Check
Standard


Frequency
High &
Low/Batch
1 /Batch;
1 /Batch
Highi
Low/Batch
1 /Batch;
10%;
1 /Batch

1 /Batch;
10%;
1 /Batch
1 /Batch


15%;
Every 8
Samples;
1 /Batch


Acceptance
Criteria
90-110%
85-115%;
85-115%
90-110%
85-115%;
85-115%;
85-115%

85-115%;
85-115%;
85-115%
81 ± 2°F


Not
Specified;
Blank
Correcting;
±5% of True
Value



-------
                Table III.  General QC Guidelines for Wet Chemistry Parameters*

Calibration
QC Check Sample/LCS
Blank
Spike
Duplicate
Type
Multiple standards + blank
Independent source
Preparation blank
Matrix spike, pre-digestion
Sample duplicate,
pre-digestion
Freqency
Daily
Batch
Batch
5-10%
5-10%
Criteria
r > 0.995
90- 110% recovery
£DL
75 - 125% recovery
£ 20% RPD for values > 5 x DL;
± 1 x DL for values < 5 x DL
1 Recommended guidelines for QC analyses where method QC requirements are not available.
 Refer to the method reference lor method specific requirements.
Acronym Definitions
   DL = detection limit
   LCS = laboratory control standard
   RPD = relative percent difference
                                                 129

-------
Of)     DEVELOPING A PERFORMANCE-BASED APPROACH TO ENVIRONMENTAL
                                 ANALYTICAL TESTING

      Berta L. Thomas. Dr. Robert G. Riley, Gary M. Mong, Margaret McCulloch, Sandra K.
      Fadeff,  Dr.  Deborah S. Sklarew, and Dr. Steven C.  Goheen.  Pacific Northwest
      Laboratory^), Richland,  Washington 99352;   Dr.  James  S.  Poppiti, Laboratory
      Management Division, U.S. Department of Energy, Germantown, Maryland 20874

      ABSTRACT

      Documented analytical  methods are  prescriptive.  However, they can be generated,
      selected, or modified based upon performance criteria. Freedom to select and use methods
      based on performance criteria allows laboratories flexibility to generate data that meet well-
      defined data quality objectives (DQOs) cost effectively.

      To  support the U.S.  Department of  Energy Waste Management and Environmental
      Restoration (DOE-EM)  programs, we suggest using DQOs and a performance-based
      approach for selecting methods to analyze environmental and waste samples.  Over  the
      long-term, a performance based approach is viewed as a generally cost-effective approach
      for  producing analytical data of the needed quality. Adopting such an approach will
      encourage development of innovative and efficient analytical technologies that address  the
      complexities of waste and environmental characterization at DOE sites while ensuring
      comparable data are produced among the many laboratories supporting DOE-EM programs.

      A primary mission of DOE is the environmental restoration  of its  waste sites.
      Accomplishing this mission will require  that a vast number of environmental and waste
      samples be analyzed for the presence of inorganic, organic, and radionuclide constituents.
      Many of the needed analyses cannot be  performed according to existing published methods
      (e.g., ASTM or EPA) because of constraints that accompany sample characterization in
      radiological  environments.   Selecting  and applying  analytical  methods within the
      framework of  a performance-based  program would improve effectiveness of DOE
      analytical support programs.

      A guidance document, DOE Methods for  Evaluating Environmental and Waste
      Management Samples (DOE Methods), is being developed in  support of DOE-EM
      programs. This document contains general guidance for characterization of contaminants in
      radioactive waste and environmental samples as well as recommended analytical methods
      necessary to meet the broad range of analytical characterization needs existing across the
      DOE complex.  DOE Methods and other recognized documents (e.g., EPA's  SW-846)
      should be consulted when selecting methods to perform analyses in support of DOE-EM
      programs;  however, these methods may be modified as long as the modified methods are
      demonstrated to meet DQOs.  In addition, alternate sources of  methods may be used if
      those methods are cost-effective and again, are demonstrated to meet project DQOs.


      (a)Pacific Northwest Laboratory is operated for the U.S. Department of Energy by Battelle
      Memorial Institute under Contract DE-AC06-76RLO 1830.
                                            130

-------
INTRODUCTION

This paper presents concepts associated with implementing a performance-based approach
to waste and environmental analytical testing within United States Department of Energy
(DOE) programs. The approach employs a data quality objectives (DQO) process that leads
to the establishment of project specific DQOs, rather than imposing prescriptive analytical
method requirements on analytical laboratories to meet program goals. For the purposes of
this paper, we define DQOs  as statements that define the analytical-testing-method-
performance requirements for a program based upon a consensus that is reached by the
regulatory community, DOE Program Management, and analytical laboratory management
and staff (EPA 1986).

This paper describes 1) the basic elements of the approach, 2) strengths versus perceived
weaknesses in applying a performance-based approach to analytical  testing, and 3)
arguments that implementing such an approach over the long-term will reduce costs and
improve analytical data quality. The goal is to stimulate responsible debate regarding the
adoption of a performance-based approach to analytical testing.

BACKGROUND

Currently, analytical testing in support of environmental monitoring, restoration, and waste
management programs  is based upon required laboratory implementation of standard,
usually promulgated, analytical methods.  Thus, laboratory personnel  responsible for
supplying data in support of these programs are required to adhere strictly to analytical
methods specified in program specific plans.  This approach is prescriptive in nature;
laboratory personnel are given very little flexibility to perform analytical testing according
to methods they believe would best meet project specific DQOs or to deal with problems
associated  with highly complex sample matrices.   Many  problems associated with
characterization of DOE environmental and waste sample matrices are not addressed by
promulgated methods.

As a framework for the performance-based approach, the DQO process converges the
thoughts of interested parties (i.e., program managers, analysts, regulators, clients) toward
a consensus opinion regarding project goals, DQOs to meet project goals, and analytical
methods to be used. A decision to use promulgated methods, unpromulgated methods, or
some combination of both to meet DQOs will be a product of mis DQO process.

ELEMENTS OF A PERFORMANCE-BASED APPROACH

Once an analytical method has been selected through the DQO process, responsibility for
documenting validation of the method lies with the organization that will be implementing
the  method, the analytical laboratory.   Validation is the process of establishing  the
suitability of a method for authorized use in the laboratory (Dux, 1990). Guidance can be
found (Dux 1990; ASTM 1986) relating to the validation process. One  aspect of the
validation process consists of demonstrating method performance using relevant sample
matrices and addressing sample logistics (e.g., manipulation of samples  containing high
levels of radioactivity  in a hot cell or glovebox).  The  following  quality control
documentation should accompany validation of an analytical method.

   •  The method's bias and precision should be documented by analyzing spiked and
      replicate samples of the matrix of interest.  When it is not possible to perform this
                                      131

-------
       test on a portion of the actual sample matrix, the test should be performed using a
       matrix that exhibits as many of the physical and chemical characteristics of the
       actual sample matrix as possible. When applicable, test data can be used to develop
       method warning and control limits.

   •   Analysis of method and matrix blanks should be documented.

   •   The working range of the method should be documented. Parameters for which
       calculated data should be provided include minimum detection limit, limit of
       detection, and limit of quantitation.

   •   The ruggedness of the  method should be  documented.   Youden presents
       information relating to ruggedness testing (Youden and Steiner, 1975).

   •   Method interferences and limitations as they apply to the matrix in question should
       be identified and documented.

   •   The method should be controlled so that changes to the method subsequent to its
       selection and validation are documented and receive concurrence from the
       appropriate regulators and DOE project management

   •   Method comparability tests should be documented. Where applicable, the analytical
       laboratory should consider documenting comparability of the selected method with
       a standard, published method to obtain data credibility with peers. "Round robin"
       interlaboratory comparison studies can be helpful in demonstrating that a method
       produces comparable data. Periodic requalification of validated methods should be
       performed by laboratories using those methods.

STRENGTHS OF  A  PERFORMANCE-BASED APPROACH-COST AND DATA
QUALITY BENEFITS

The most important strengths of a performance-based approach to  analytical testing of
waste and environmental samples fall into two broad categories:  improved cost  and
improved data quality over time.

       Cost-A performance-based approach allows methods that are tailored to provide
       data of the quality needed to meet program objectives (part  of the DQO process
       decision) to be selected. The performance-based approach makes a larger pool of
       analytical methods available for consideration during the DQO process. Design of
       the most effective analysis program may be based upon cost, as long as assurances
       are made that DQOs will be met. For example, analytical  data generated from
       standard methods are often based upon minimum detection limits (MDLs), limits of
       detection (LODs), or limits of quantitation (LOQs) that are, in general, lower than
       those required to meet project DQOs.  In contrast, the performance-based approach
       allows the flexibility to adapt or chose alternative analytical methods that provide
       data of only the quality needed, while minimizing additional costs associated with
       methods that produce data of higher quality than project DQOs require.  In
       evaluating cost, factors to consider include time required to perform the method,
       equipment and facilities required to assure worker safety, training of analysts, and
       waste minimization and disposal.
                                      132

-------
      In applying the DQO process within the performance-based approach, analytical
      methods can be selected that are tailored to the resources available (e.g., personnel
      and equipment) rather than requiring purchase of additional equipment, making
      expensive modifications to existing facilities, or hiring specialized personnel to
      perform the analytical testing according to standard/promulgated methods.

   •  Data Quality-A performance-based approach to analytical testing offers a means to
      ensure enhanced data quality by selection/application of the best available analytical
      methodology. In doing so, weaknesses associated with standard methods can be
      avoided.  Currently, methods accepted for  use in support of regulatory
      requirements present a very conservative approach to laboratory analysis. Little
      room for adoption of innovative approaches exists in the current prescriptive
      environment.  This prescriptive approach has the effect of stifling creativity in
      analytical testing methodology.

      Standard methods, such as those published in SW-846,  were not developed to
      accommodate the variety of complex sample matrices encountered in DOE programs
      and frequently cannot be easily  adapted to such matrices.  In a program requiring
      prescriptive methods, the process of data production can over shadow the data upon
      which decisions  are based.  In this situation, the process can become more
      important than  the product. This can be dangerous because data quality may be
      compromised to comply with  standard methods.  Such situations may be avoided
      using a performance-based approach.

PERCEIVED WEAKNESSES OF A PERFORMANCE-BASED APPROACH

The following are perceived as weaknesses of the performance-based approach:

   •  Approach Acceptance-The performance-based approach is currently untested.
      Historically, promulgated methods are adopted to secure regulatory  "buy-in" by
      offering some assurance that data of high quality are produced and are defensible in
      the event that they are introduced into litigation.  Because of entrenched habits, it
      will take time for regulators and  the regulated community to accept a performance-
      based approach to analytical testing.  After the performance-based approach
      achieves success, the rigid, conservative approach to method selection will yield to
      the more flexible performance-based approach.

   •  Cost  of Documentation—Implementing a performance-based approach  to
      environmental and waste analysis could require a greater "up front" investment of
      time and dollars to gain acceptance.  Method validation by a laboratory could
      require that more quality control data be generated and documented in the near-term
      than might be required to support standard/promulgated methods. These added "up
      front" costs could create cash flow problems for small laboratories. However, the
      QC data produced will be more pertinent to program DQOs and to the matrix of
      interest.  Data will, therefore, be more meaningful.

   •  Training Costs—The flexibility  built into a performance-based approach places
      greater responsibility for management of unpromulgated or modified promulgated
      methods (i.e., selection, validation and application) on laboratory personnel.  The
      additional technical responsibility could lead to the  need for more education and
      training for laboratory analysts and managers to assure they are able to make sound
                                       133

-------
       scientific judgments during the method selection and application process.  This
       increased training for laboratory analysts and managers should, however, improve
       the quality of the data.

       Similarly, auditors providing program oversight might need more extensive training
       to effectively audit programs developed according to a performance-based
       approach.  The emphasis in auditing prescriptive-based programs  is placed on
       monitoring for compliance to program requirements. "Monitoring for compliance
       type auditing" is generally thought to be much simpler (i.e., requiring less extensive
       training)  than monitoring a performance-based program  for its effectiveness.
       However, this change in emphasis of the auditing process  should produce value
       added over time.

       Contractual Costs-More planning could be required for acquiring and monitoring
       subcontracts for analytical services  in programs developed  according  to a
       performance-based approach. The flexibility inherent in the performance-based
       approach could require DOE to spend more time preparing a request for proposal
       and the subcontracted organization to spend more time preparing the proposal than
       might be required using the more traditional approach that requires the use of
       standard/promulgated analytical methods.  However, the resulting proposal is more
       likely to better meet the needs of the program.

CONCLUSION

A primary mission of DOE is the environmental restoration of its waste sites.  Decisions
regarding site  remediation require the use  of legally defensible  data of high quality.
Because projected costs associated with remediation efforts are so high, DOE, the
regulatory community, and analytical laboratory staff need to consider how best to reduce
analytical costs while still providing sound data to serve as the bases for decisions.

The DOE  is endeavoring to create  an environment where continuous improvement is
fostered. This continuous improvement atmosphere  extends to organizations outside DOE
that have been contracted to provide analytical  testing support for DOE programs.  A
performance-based approach to analytical testing allows the flexibility to implement the
continuous improvement philosophy by allowing the DQO process to drive the selection of
sampling and analytical methods.

Standard/promulgated methods currently available for analytical testing do not address the
problems associated with measuring analytes in complex waste and environmental matrices
found at DOE sites, nor do they provide the data quality required for sound site remediation
and waste management decisions.

To  bridge this gap, the Laboratory Management Division of DOE is  supporting
development of DOE Methods  for Evaluating Environmental and Waste Management
Samples (DOE Methods).  DOE Methods contains guidance and  methods that are not
currently provided by existing guidance manuals for DOE and DOE contractor laboratories
to use for radioactive waste  and environmental sampling and analysis at DOE sites. DOE
Methods recommends  that analytical methods contained within the document  be
considered, through the DQO process, for use in support of DOE programs.  However,
encouragement is given to use alternate methods if they can be demonstrated to  provide a
                                      134

-------
cost effective alternative to methods published in DOE Methods or other recognized
methods documents (Goheen et al. 1992)

This paper describes the elements of a performance-based approach to analytical testing and
the strengths and weaknesses of the approach compared to a strictly prescriptive approach.
Readers of this  paper  are encouraged to  provide  constructive  comments  on the
performance-based approach that incorporates the DQO process.  All feedback will be
given serious consideration as this subject is addressed in forthcoming revisions of DOE
Methods.  All comments should be addressed to Dr. Steven C. Goheen, at  Pacific
Northwest Laboratory, P.O. Box 999, MS P8-08, Richland, WA 99352.

ACKNOWLEDGMENT

This work is funded by the Laboratory Management Division of the Department of
Energy's Office of Technology Development.

REFERENCES

American Society for Testing and Materials (ASTM) Standards. 1986. "ASTM Standard
Guide for Qualification of Measurement Methods by a Laboratory Within the Nuclear
Industry." ASTM C 1068-86. Annual book, Vol 12.01.

Dux, J. P.  1990. Handbook  of Quality Assurance for the Analytical Chemistry
Laboratory, Second Edition. Van Nostrand Reinhold, New York 10003, p. 62.

Goheen, S. C., M. McCulloch, S. K. Fadeff, D. S. Sklarew, B. L. Thomas, R. M. Bean,
G. K. Ruebsamen, S. M. Anantatmula, and R. G. Riley.  1992.  "A Guidance and
Methods Compendium:  DOE Methods  for Evaluating Environmental and Waste
Management Samples."  Proceedings, 8th Annual Waste Testing & Quality Assurance
Symposium, pp 105-124.

U.S. Environmental Protection Agency  (EPA).  1986. Development of Data Quality
Objectives: Description of Stages I and II.  Quality Assurance Management Staff.
Washington, D.C.

Youden, W. J., and E. H. Steiner. 1975. Statistical Manual of the AOAC (Association
of Official Analytical Chemists). Washington, D.C.
                                     135

-------
21         QUALITY ASSURANCE AND DATA COLLECTION:  ELECTRONIC DATA TRANSFER

      L.  M.  Tomczak Senior Staff Scientist, J.  C.  Gore, Technologist  II, W. G. Lohner,
      Engineer I, E. C. Ray,  Meteorologist, and J. A.  Selasky,  Technologist  II, Fernald
      Environmental Restoration  Management  Corporation,  Radiological  Environmental
      Monitoring group,  P.O. Box 398704,  Cincinnati, Ohio 45239-8704 and H. B. Spitz
      PhD,  University  of Cincinnati, College of Nuclear Engineering, Cincinnati, Ohio
      45242-0072

      ABSTRACT

      The  Radiological   Environmental   Monitoring   (REM)  group   at   the  Fernald
      Environmental Management  Project  is involved  in  an Electronic  Data Transfer
      practice that will result in the improved  quality  assurance of collected data.
      This  practice focuses  on  electronic data transfer from the recording instrument
      to reduce the manpower normally  required for  manual data entry and improve the
      quality of the data transferred.

      The application  of this practice can enhance  any data collection program where
      instruments  with   electronic  memories  and   a  signal  output  are  utilized.
      Organizations employing this practice can strengthen the quality and efficiency
      of their  data collection program.   The use  of these practices  can  assist  in
      complying with Quality Assurance  requirements under ASME NQA-1,  RCRA, CERCLA, and
      DOE Order activities.

      Data  from Pylon  AB-5 instrumentation is typically configured to print data to a
      tape.  The REM group has developed  a process  to electronically transfer stored
      data.  The data  are  sent from  the  Pylon  AB-5 field instrument  to  a Hewlett-
      Packard portable hand  computer,  model  HP95LX.   Data are recorded and stored  on
      a 128 K-byte RAM card  and later transferred to  a  PC database  as an  electronic
      file  for  analysis.  The advantage  of this system is twofold:  (1)  Data  entry
      errors are eliminated  and  (2) considerable data collection and  entry  time  is
      eliminated.  Checks can  then  be conducted for data  validity  between recorded
      intervals due to light leaks etc.  and  the  detection of outliers.

      This  paper/presentation will discuss the interface  and connector components that
      allow this transfer of data from  the  Pylon  to the  PC to take  place  and  the
      process to perform that activity.

      INTRODUCTION

      Fifteen continuous radon monitors have been installed at the DOE Fernald
      Environmental Monitoring Project (FEMP)  at both on-site and off-site locations
      to measure what,  if any,  contribution  to the  natural radon background is made
      by sources of radon located on the  Fernald project.  On-site outdoor locations
      for each of the  samplers have been  selected to provide representative
      measurements  of radon  that are either  close to the sources of emission (K-65
      silos that contain concentrated  radium byproducts) or near buildings and areas
      where production occurred.  Several monitors  have  measured radon at the FEMP
      fenceline and from two background locations a minimum of 13 miles from the
      project including  one  located in the opposite direction of the prevailing
      wind.

      Each  monitor  prints the results  of  hourly measurements and also stores the
                                         136

-------
data i.n the instrument computer memory.  Each result consists of four fields
of data: sequence number, hour (24 hour clock), instrument response
(counts/hour), and radon concentration (pCi/1).  Data is retrieved from each
instrument by environmental data collection personnel on a weekly basis.
Data collection personnel initial each data paper tape and reset the
instrument sequence number to initiate another run.  An identification number,
calibration factors, and other data constants are printed on the data tape
whenever the instrument is reset.  Data are also collected/transferred
electronically onto a hand held computer at each location.  The instrument
response is also conducted using a check source on a weekly/monthly basis per
established protocol.

The radon collection instruments are housed in environmental enclosures to
protect them from the direct adverse effects of the environment but are not
heated or sealed in any manner.  They operate 24-hours a day and are expected
to function properly in all types of weather and generate 168 hourly readings
each week.  The instruments sample radon in the ambient air based upon the
diffusion principle and, thus, have no moving parts or pumps.  Other than
weekly inspection by the data collection personnel to retrieve data, reset the
instrument, and insure that the instrument is still operating, no other
operator attention is provided or required.  Instrument re-calibration is
scheduled on an annual basis contingent upon the satisfactory instrument
source check results.

OVERVIEW

Collection of data, albeit an important function, comprises only half of the
environmental monitoring activity.  Sometimes it can be just as cumbersome to
appropriately deal with the data that has been collected, as it can be to
collect the data.  The Radiological Environmental Monito-ing group at the DOE
FEMP facility has identified a solution, to at least take some of the dread
and monotony out of the voluminous environmental data reduction and reporting
activity, without sacrificing quality.

OBJECTIVE

Obtaining a workable electronic file for generating reports from data tapes
requires considerable labor intense quality assurance activities.  These steps
are depicted in Figure 1.  To relieve the burden of work related to the data
collection and reporting, a method has been devised to transfer the data
electronically.  These steps are depicted in Figure 2.

There are essentially two steps in the electronic data transfer process.  The
first step is the transfer of an electronic data file from the Pylon AB-5 unit
to RAM card on a handheld computer in the field.  The second step is the
transfer of field data from the RAM card to an electronic data file back at
the office on the hard drive of a PC.

This data is then transferred via disk along with the hard copy data tape of
the Pylon AB-5 data to a data previewer, and then to an independent group that
produces the environmental reports.  This activity should eliminate data
errors and data entry form.
                                    137

-------
MATERIALS NEEDED

The following is a list of equipment that is needed to complete the electronic
data transfer from the AB-5 Pylon to a personal IBM disk operating system
computer.

         .   Pylon AB-5 Radon Monitor
         .   Pylon Model PPT-1 Printer
         .   Pylon Model CI-55 Computer Interface
         .   CPRD (radon detector) or a 300A Lucas cell
         .   DB15 Female A/B Switch Box
         .   DB15 3 foot M.M. Transfer Cable
         .   DB15 3 foot M.F. Transfer Cable
         .   DB25/9 M.M. Gender Changer
         .   DB15 M.M. Gender Changer
         .   HP95LX Computer
         .   HP F1001A Connectivity Pack
         .   HP F1002A 128 K-byte RAM card
         .   128 K-byte RAM card drive linked to PC

METHODOLOGY

Data collection personnel have procedures for performing instrument checks and
retrieving data from each of the instruments.  They observe instrument
operation and determine that the data output tape appears normal.  They also
initial the data tape, electronically transfer the data, and reset the
instrument.  Should the instrument appear to be malfunctioning, they will
notify management and initiate a service request from the instrumentation
group.  Periodically, after electronically transferring and collecting the
data, the data collection personnel perform an instrument response check by
removing the radon sampler and attaching a check source.  Following the
performance check, the data collection personnel re-attach the radon sampler
to the monitor, and reset the instrument.

The following are the steps that are needed to transfer the electronic data
file from the Pylon AB-5 unit to a personal IBM disk operating system
computer.  Each step of transfer will be presented in detail in the Appendix.

QUALITY ASSURANCE OF TRANSFERRED DATA

To ensure that the data are successfully transferred from the Pylon AB-5 to
the RAM card, a quality control check is performed each week.  During the
weekly inspection one data tape from those collected is randomly selected and
10% of the data points are randomly chosen and compared to the same data
points on the electronic file.  Once each quarter, one data tape is randomly
selected and all data tape data points are compared with the electronic file
data points.

SUMMARY

Using the electronic data transfer scheme described in this paper, radon data
can be easily converted into an electronic file for further processing.  Once
the data are transferred to a disk, the data can be previewed by the
individual  assigned the task of quality assurance and data validation.   It
                                    138

-------
will be necessary to parse the data using a Lotus function to be able to work
with each of the columns.  This data translation process would also be
required if other software programs (any software capable of handling ASCII)
were used in the transfer process.  A check of data will then need to be
conducted to confirm that the data was correctly transferred and that all data
points were valid observations.

By utilizing the electronic transfer of data, considerable time can be saved
in both the areas of manual data entry, and the checks that are required to
ensure that the data is correct.  Activities that had previously taken more
than a day can be performed in much less than a day without sacrificing data
quality.

REFERENCES

HP 95LX User's Guide, Hewlett-Packard Co., 1991, Corvallis, OR
                                    139

-------
  Data Collection
  Personnel
QC/ Invalid Data
Discrepancy Control
                        Review Data for
                        Completeness/
                        Readability
                        Perform Double
                        Key Entry
                               Yes
Perform
Maintenance/ Repair
                          s Data
                          Complete
                          and
                          Readable?
Generate
Reports
t
w
Statistical
Analysis
                       FIGURE 1
    Manual  Data  Collection and Data Entry
                              140

-------
    Attach HP
    Connection Pack
    to AB-5 Unit
     Initiate HP
     BSLX for Data
     Transfer
    Place A/B Switch
    to "B" Position
      Return A/B
      Switch to "A"
      Position
     Disconnect HP
     Connection Pack
Import Raw Data
Files into Lotus
123
Save Data Files
in Lotus 1 23
Copy Files to
RAM Card in HP
9BLK
                                 Remove RAM
                                 Card from HP
                                 95LX
                                Insert RAM Card
                                into PC RAM
                                Drive
                                Review Data
                 FIGURE 2
Data Transfer from  AB-5 to  HP 95LX
Data Transfer from  HP 95LX to  PC
                        141

-------
                                   APPENDIX

      TRANSFERRINS DATA FROM THE PYLON MONITOR TO THE HP95LX COMPUTER

1) Hook up the HPF100A1 Connectivity Pack to the HP95LX and the CI-55
   Computer Interface
         Connect the CI-55 to a DB-15 M.M. Transfer Cable.

2)    Prepare the HP95LX for use

      Turn on the HP95LX
         Press:   ON/OFF key

         Press: BLUE FILER key
                Select PRN (DIR)
                Press: CURSOR DOWN key
                  Note: Delete all PRN files to start off the day.  This will
                        ensure that sufficient memory will be available on the
                        unit to collect the data.

         Press: F9 key to tag the old file to be deleted

         Press: F3 key
                Y  key (to delete)
                MENU key

                Press: Q key to quit
                Press: COMM key
         Press: MENU key
                Press S key for SETTINGS selection
         Press: U key for USE selection

         Press:Cursor Arrow —> key over to File (PYL.DCF)
                Press: ENTER key
                Check the following setup information:
Port:
Inter.
Baud
Stop
Parity
Char
Dial:
Type

1
3
1
none
8

Puls
         Press: Q key for QUIT selection

         Hold CTRL key and press F-5 key simultaneously

         Backspace to DAT and delete by using the <— (Break) key
                Type PRN\name.prn
                Press:  ENTER key
                                      142

-------
                  Note: 8 characters max. name

         Press:  MENU key

         Press:  C key for CONNECT selection
                (at this point set up the pylon to send data)

3) Transferring  data from the pylon unit to the HP95LX:

      Pylon Monitor
            Press: START/STOP key to stop current run

      Switch Box
           Place A/B switch to B position

      Pylon AB-5 unit
           Press: RECALL key

           Press: PROG STEP key, this will show the most current run.
            If it is necessary to collect a different run use the PUMP key to
            change the first digit and the START/STOP key to change the second
            digit. After the RUN is selected press the PROG/STEP key 3 times.

           Press and hold: STATUS key
            Display will show the run, cycle, and interval.

           Press: PROG/STEP key while holding the STATUS key
            This will start the data transfer.

           Release: PROG/STEP and STATUS key.
           When  the data transfer is complete the Pylon AB-5 display will
                 display the word "ready" .

      Switch Box
           Reposition A/B switch to the A position
           Disconnect CI-55 Computer Interface cable from transfer
           cable on switch box.

      HP96LX Computer
           Press: MENU key to exit program
           Press: Q key for QUIT selection

4) Checking data transfer
         Press:  FILER key
         Select  the PRN (dir) ENTER key
                (check if file and data exists)

         Press:  MENU key
         Press:  Q key for QUIT selection
         Turn off instrument
            Press: ON/OFF key
            Note: Prepare for next station.  At the next location it is not
                  necessary to repeat step #2, but step #1, the hookup is
                  still applicable.
                                     143

-------
      DATA TRANSFER FROM THE INTERNAL DRIVE OF THE HP95LX TO THE  RAM CARD

1) Transfer data from the HP95LX capture file to the HP95LX Lotus 123
   Directory on the RAM Card
         Turn on the HP95LX
         Press: OFF/ON key
         Press: LOTUS 123 key
         Press: MENU key
         Press: F key for FILE selection
                I key for IMPORT selection
                T key for TEXT selection
         Arrow to —> _PRN\ enter
         Search for file of interest
            Press: ENTER key

2) Saving the file
         Press: MENU key
                F key for FILE selection
                S key for SAVE selection

         Press ESC key two times
         Type in A:\file name
         Press: ENTER key
         Press: R key for REPLACE selection if necessary to write over
                previous data
         Press: ENTER key
            Repeat this step to transfer the files to the LOTUS (dir  Drive A
            RAM card)
         Transfer data from the 128 K-byte RAM Card to the PC via the RAM Card
            Drive

         Press: MENU key
         Press: Q for QUIT selection
         Press: Y for YES selection
         Turn off the HP95LX
            Press: ON/OFF key
                                     144

-------
THE FOLLOWING ARE DISCLAIMERS FOR ENVIRONMENTAL PAPERS/REPORTS:

THIS ABSTRACT/PAPER/REPORT WAS PREPARED AS AN  ACCOUNT OF WORK SPONSORED BY AN
AGENCY OF  THE UNITED  STATES GOVERNMENT.   REFERENCE  HEREIN TO  ANY SPECIFIC
COMMERCIAL PRODUCT,  PROCESS,  OR SERVICE BY TRADE NAME, TRADEMARK, MANUFACTURER,
OR OTHERWISE DOES NOT  CONSTITUTE  OR IMPLY  ITS  ENDORSEMENT, RECOMMENDATION, OR
FAVORING BY THE UNITED STATES GOVERNMENT OR ANY AGENCY THEREOF.  THE VIEWS AND
OPINIONS OF AUTHORS EXPRESSED HEREIN DO NOT NECESSARILY STATE OR REFLECT THOSE
OF THE UNITED STATES GOVERNMENT, OR  ANY AGENCY  THEREOF OR FERNALD ENVIRONMENTAL
RESTORATION MANAGEMENT CORPORATION,  ITS AFFILIATES OR ITS PARENT COMPANIES.
NOTICE  FOR OTHER  THAN  ENVIRONMENTAL REPORTS  (INTERNAL  SPECIALS,  SPECIALS,
SUBCONTRACTOR, TOPICAL REPORTS):

THIS REPORT WAS  PREPARED AS AN ACCOUNT OF WORK SPONSORED  BY  AN AGENCY OF THE
UNITED STATES GOVERNMENT.  NEITHER THE UNITED STATES  GOVERNMENT OR ANY AGENCY
THEREOF, NOR ANY OF THEIR EMPLOYEES,  NOR ANY OF ITS CONTRACTORS, SUBCONTRACTORS
NOR THEIR EMPLOYEES MAKES ANY WARRANTY, EXPRESS OR IMPLIED,  OR  ASSUMES ANY LEGAL
LIABILITY OR RESPONSIBILITY FOR THE ACCURACY, COMPLETENESS,  OR  USEFULNESS OF ANY
INFORMATION, APPARATUS,  PRODUCT, OR PROCESS DISCLOSED,  OR REPRESENTS THAT ITS USE
WOULD NOT  INFRINGE PRIVATELY  OWNED RIGHTS.  REFERENCE HEREIN TO ANY SPECIFIC
COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY TRADE NAME,  MANUFACTURER OR OTHERWISE,
DOES NOT  NECESSARILY  CONSTITUTE OR  IMPLY  ITS  ENDORSEMENT,  RECOMMENDATION,  OR
FAVORING BY THE UNITED STATES GOVERNMENT OR ANY AGENCY THEREOF.  THE VIEWS AND
OPINIONS OF AUTHORS EXPRESSED HEREIN DO NOT NECESSARILY STATE OR REFLECT THOSE
OF THE UNITED STATES GOVERNMENT OR ANY AGENCY THEREOF, OR FERNALD ENVIRONMENTAL
RESTORATION MANAGEMENT CORPORATION,  ITS AFFILIATES OR ITS PARENT COMPANIES.
DOE-FERMCO PRIME CONTRACT NO. IS DE-ACOR-920R21972.
                                   145

-------
22           AN INTEGRATED APPROACH TO MAINTAINING QUALITY
                        ASSURANCE DURING THE RI/FS PROCESS
                            Alex Tracy, William O'Brien, William Mills
                        Woodward-Clyde Federal Services, Rockville MD

       Abstract

             Recently much attention has been given to the applications  and requirements of
       Laboratory  Information Management Systems (LIMS),  including maintenance  of data
       integrity, implementation of automated scripts for the processing and reporting of data and
       data security.   However,  data management  and tracking during the field investigation
       phase of a project have not received the same scrutiny.  The case study described in this
       paper focuses on two potential problem areas in the data collection/reduction process: the
       standardization and documentation of sample tracking,  numbering, and  the  collection
       process; and the automation of electronic data transfer, analysis and review.

       Introduction

             Interest in the project began when an unusual  number  of cases of Hodgkin's
       disease were reported in  the  area of a former army hospital.   Public concern over a
       biomedical waste incinerator operated at the army hospital prompted the U.S. Army Corps
       of  Engineers  (USACOE) to launch  a site investigation.   The purpose  of the site
       investigation was to provide a baseline risk assessment, a contamination assessment and an
       ecological risk assessment for the areas of concern.   Due in  part to the high risk of
       litigation, Woodward-Clyde Federal Services  (WCFS)  augmented  its existing sample
       tracking program and  implemented strict sample handling and documentation procedures.
       This system was designed to track all samples from collection through  the reporting and
       data analysis phases with  a high  degree of confidence.  All samples  were systematically
       numbered and logged  upon collection. The analytical laboratories supplied documentation
       detailing sample receipt and status in the analytical process at specified time intervals.
       Both analytical results and supporting QA/QC  information were uploaded into a WCFS
       database using a data-parsing program that converts ASCII text files into database format.

             Once the data from the analytical laboratories  was uploaded into the WCFS
       database, programs were developed to compile statistical information in support of the on-
       going human and ecological risk assessments.  Programs were  also  developed  to review
       and qualify data, and provide  output suitable for the final report. By training the field
       team and maintaining  strict documentation, samples could be accurately tracked from the
       field through the lab until results were reported. By utilizing data-parsing programs along
       with batch  processing  to automatically upload data to the  database system, WCFS
       eliminated manual entry of data without having to write complex computer programs that
       could not offer the flexibility required by the data analysis.  A  diagram that outlines the
                                              146

-------
                                                 Data Flow Diagram
Field Information on Samples
          Sample Log-books
          COCs
Lab Info on Samples
  	Sampling database from lab
    Data from lab
COC Check to  ensure
 data completeness
          Sampling Database
Data extraction using
     Monarch
                                                                                                                                              O  <
                                                                                                                                              IP
                                                                                                                                              3  ^^
                                                                                                                                              JQ  o.
                                                            Paradox
O  pj
£?.  O.
°  £|

   OQ
    r-t-
    0
   •a
   ^.

    S-
    v
    a.
    en
    O
                                                                                                                                                 ft>
                                                          Quality Control
Sample log-books



1 Ecological Risk


Evaluate holding
times
1


1
Risk Assessment
Stat Tables
Data Distribution
Graph. Representation
Assignment of CT &
RME Values
Calculate Risk
Method Blank
Contamination
1


Surrogate
Recoveries




Contamination Assessment
Hits Tables
Background Assessment

                                                                                                                                                 fD

                                                                                                                                                 I
                                                                                                                                                 o>
                                                                                                                                                 en
                                                                                                                                                 O
                                                                                                                                                 a.
                                                                                                                                                 c
                                                                                                                                                 o
                                                                                                                                                 s.
                                                                                                                                                 en'
                                                                                                                                                 T3
                                                                                                                                                 en
                                                                                                                                                 O
                                                                                                                                                 el-
                                                                                                                                                 s'
                                                                                                                                                 O>

-------
       Roughly 520,000 records were manipulated in support of this project, and every
milestone for deliverables was  met on time.  Through documentation and automation,
WCFS was able to maintain the integrity of the data during the course of the project. This
presentation will discuss the systematized methods used to manage data and their impact
on the successful completion of the project.

Systematic Numbering Procedure

       In order to glean as much information as possible from the sampling event, all
sample numbers were coded to provide information about the date of sampling, the sample
matrix, whether the sample was a QC sample, etc.  To make the numbering system as
simple as possible, the date was used as a key element of the sample number. The site
identification or spatial location were not part of the coded sample number because all of
the sites would have to be selected and coded for sampling well in advance of the sampling
event.  The basic format of the sample number was:

       AAYYMMDD-##XX
       where AA refers to the installation or site code
       YYMMDD is the date (in the format year-month-day)
       ## is the number of the sample collected for that day
       and XX  is an extension that tells the matrix being sampled whether the sample is a
       QC sample.

       This method of coding information into the sample numbers provided field
personnel with a fairly simple method of numbering the environmental samples in a
systematic fashion.  It also provided the staff members checking the sampling data with
additional information that allowed for further QC  of the dataset.

Field Sampling Documentation

       Sample documentation was another key element in the quality control of the RI:
chains of custody were used to document sample transport, but a sample log-book, along
with extended documentation provided from the laboratory, made it easier to track and
QC samples from collection to analysis and invoicing. Chains of custody were
sequentially numbered in the same fashion as bank checks. If a  COC was voided, it was
retained to ensure that all COCs were accounted for.

       The sample log-book contained the coded field sample number, the sample's spatial
location and depth, the analyses requested, the  COC number used to document sample
transport and the initials of the individuals performing the sampling.  The labs were
required to confirm receipt of samples by a written statement listing the WCFS field
sample number, the corresponding lab sample number, the analyses requested for that
                                       148

-------
sample and a copy of the chain of custody for those samples. Every two weeks the labs
were required to submit a status report listing all samples still within the analytical process
at the laboratory as well as their status within the lab (logged, prepped, analysis complete,
etc.)

       At major milestones in the project the lab was required to submit an ASCII file
correlating WCFS field sample numbers with lab sample numbers. This allowed a check
of the WCFS sample log-book and all of the information provided by the lab, data upload
and analysis could begin.

Data Extraction from Diskette Deliverables

       Contract Laboratory Protocol (CLP) was used as the primary analytical
methodology for the RI/FS  (herbicides and dioxins were analyzed using SW-846
methods). This protocol is  characterized by very specific quality control, documentation,
and deliverables.  The CLP  diskette contains a wealth of information seldom used in its
entirety.  However, by using data-parsing and pattern-scanning programs,  WCFS was able
to extract both the analytical results and all of the supporting quality control information
that allowed WCFS to perform an automated data review.

       Templates to extract data from diskette deliverables were developed using
Monarch, a data-parsing program. This program allows the user to read ASCII files,
define traps that key on selection criteria to determine which records to extract, name and
format all fields for export into a database format.  Templates could be developed in
roughly a half-day, and then all data could be extracted without any data entry.  Gawk,  a
pattern-scanning language originally developed for the UNIX operating system, was used
to create the batch files that would run Monarch. It was used to extract the information
from the diskettes for roughly 150 separate sample delivery groups. This totaled nearly
520,000 records for six  different analytical methodologies.

Quality Control

       Master database files were created for each of the six analytical methods. This was
done by appending individual data files, and the files were then checked against the
information contained in the sample log-book to ensure that all samples were accounted
for and that the proper analysis had been performed and reported. A percentage of the
results in the database were checked against hard copies supplied by the lab and no errors
were found. For the entire  project, only three errors were found in the entire data set, and
all were attributable to data entry errors at the lab.

       Once the content of the database files had been checked to ensure  accurate upload
had been performed, results were qualified by the chemistry staff. Queries were created to
evaluate hold times, method blank contamination, and surrogate recoveries for all samples.
                                         149

-------
Sample results whose hold times had been exceeded were qualified, and method blanks
were correlated with their associated samples to perform a blank evaluation.

      Under CLP methodology, the "B" flag is used to denote a compound which has
been detected both in an environmental sample and its associated blanks.  If there is a
blank contamination, the EPA has issued guidance to qualify those results: for compounds
that are considered common lab contaminants to be considered as detects (not artifacts of
the sampling/analysis process), the concentration in the  environmental sample must exceed
ten times the concentration found in any associated blank for those compounds. For all
other compounds (those not considered common lab contaminants), the concentration in
the environmental sample must be five times greater than the concentration found in any
associated blank. The table shown below lists the compounds found in method blanks:

                   Lab Contaminants Found in Method Blanks
Compounds
(Units are ug/L)
Volatiles
2-Hexanone
Acetone
Methy Ethyl Ketone
Methylene Chloride
Semivolatiles
3-Nitroaniline
4-Nitroaniline
Carbazole
Di-n-Butylphthalate
Di-n-Octyl Phthalate
Diethylphthalate
Phenol
bis(2-ethylhexyl) pthalate
Avg.
Value

2.56
7.26
4.03
1.91

40.00
46.00
34.00
40.50
82.00
3.00
1200.00
203.09
Minimum
Value

1.00
3.00
1.00
1.00

40.00
46.00
34.00
1.00
82.00
3.00
1200.00
1.00
Maximum
Value

5.00
18.00
19.00
14.00

40.00
46.00
34.00
80.00
82.00
3.00
1200.00
580.00
No. of
Detects

9
46
33
35

1
1
1
10
1
1
1
11
No. of
Samples

90
90
90
90

66
66
66
66
66
66
66
66
       Roughly 293 detects in environmental samples had the "B" flag as a qualifier:  after
blank assessment and qualification, only two detects with the "B" flag were considered
actual detects and not artifacts of the sampling/analysis process.

       Surrogates were analyzed using Shewart charting techniques commonly used for
process control.  Samples were grouped by matrix, analysis date, and sample delivery
group to check for trends resulting from either matrix effects, or instrument or operator
error. The standard deviation was calculated for each group of samples and warning and
control limits were calculated by multiplying the standard deviation by a factor of two or
three, respectively. The data for those samples with unacceptable surrogate recoveries
                                       150

-------
was reviewed, the chromatograms were analyzed and if necessary, the results were
qualified.  Some Shewart charts for sediment samples are shown on the following page:

                        Surrogate Recovery for Sediment Samples: Volatiles
     130
      70 -1-
	 • 	 Recovery 	 '
n 	 UWL
— * 	 LWL
— * 	 UCL
—* 	 LCL
                         Surrogate Recovery for Sediment: Semivolatiles
	 • 	 Recovery 	 '
Q 	 UWL
— * 	 LWL
— o 	 UCL
— * 	 LCL
       Both blind and designated duplicates were submitted to the lab. They were
evaluated for the precision associated with both the sampling and preparation (sub-
sampling) methods. For samples with detects above the contract-required detection or
quantitation limits, the relative percent difference between a sample and its duplicate was
typically below 20%.  Matrix spike and matrix spike duplicate results were also evaluated
and were found to have similar precision. Accuracy of the methods was evaluated using
                                         151

-------
matrix spike/matrix spike duplicate results, laboratory control samples, and surrogate
recoveries.

Risk Assessment Calculations

       In support of the risk assessment, samples were first grouped by exposure area for
each group of receptors. The samples were further subdivided by  matrix. Contaminants
of concern were determined using automated queries to calculate minimum, maximum,
minimum non-detect, maximum non-detect, average, standard deviation, and number of
detects vs. number of samples for every chemical analysis by exposure area and media.
Data was log-transformed and the process was repeated. Both normal and log-
transformed data were evaluated with respect to background samples to determine if the
data set being analyzed was statistically distinct from the background results.  Based on
that statistical information and frequency of detection,  chemicals to be included in the risk
assessment were selected.

       Plots similar to the one shown below were used as a graphic means to determine if
the background samples were part of an statistically distinct population as compared to
on-site samples.  The graph shows the average plus or minus the standard error of the
mean (standard error of the mean^hree times standard deviation) for arsenic (AS),
beryllium (BE), cadmium (CD), chromium (CR), nickel (NT) and selenium (SE).  On-site
results for a particular element are designated by that element's symbol, while background
results are shown as the symbol plus "-BG."  Based on this technique, none of the
elements shown have statistically distinct populations between on-site and background
samples.  It must be emphasized,  however, that this test was merely one of several
techniques used to evaluate the presence or absence of on-site contamination.

                    Evaluation of Inorganic Analytes with  Respect to Background
60 i

50

40

30

20
10
0

-10

c






c


ft
]
, I , ? , - , 1 , I ,
]

D





1
' 1 ' • ' ' • ' • '
S AS-BG BE BE-BG CD CD-BG CR CR,

C




1 1


c


' ' i
l



BG Nl NI-BG

                        C-BG" refer* to a value obtained from background samples)
                  Avg-Std Error of Mean  ° Avg + Std Error of Mean  • Average
                                         152

-------
       Once the chemicals of concern were selected, data distribution was evaluated to
determine the proper treatment of the data for risk assessment purposes. For this site,
calculations using log-transformed data proved to be more conservative, so the log-
transformed version of the data was employed for all further risk assessment calculations.

       A number of scenarios were established based primarily on the number of samples
of a particular matrix for an exposure area.  Those scenarios are listed below:

                    Scenario Summary for Risk Assessment Data

1 0 or more samples per matrix per exposure area
Less than 10 same
No detects per mat

UCL > max.
UCL > avg., UCL < max.
es per matrix per exposure area

avg. < max.
max < avg.
rix per exposure area

Values used for RA
CT

UCL
Average

Average
Maximum

Omitted
RME

UCL
UCL

Maximum
Maximum

Omitted
 Note:
Maximum indicates the maximum detected value
Average is the average of all samples
UCL is the log-normally distributed UCL95 value
       Given the scenarios presented above, the data was uploaded into a spreadsheet
application that had the compatibility to easily convert files from the database application.
Three fields were added to each record; they included the name of the scenario for which
the data belonged, and the Central Tendency (CT) and Reasonable Maximum Exposure
(RME) values that should be used with that scenario. The CT and RME values were
modified by uptake factors for the dermal and ingestion portions of the risk assessment.
For well and surface water, the CT and RME values were modified by a value Kp, which
is the permeability constant for dermal uptake of compounds from water.  The CT and
RME were multiplied by an absorption factor for dermal uptake of compounds in direct
contact with the skin for all sediment, surface soil, and soil boring data.  These
calculations were again automated through the use of macros to only calculate if the
situation was correct for that scenario. The final tables  contained either the calculated
value or an "NA" if the data should not be modified by an uptake factor.
                                        153

-------
       Given the now reduced number of chemicals of concern with detects, the data was
uploaded into the Woodward-Clyde Risk Assessment application, Assess, which was used
to calculate risk. Following all calculations, the hard-copy reports generated were
checked for validity against the spreadsheet tables and the source database files.  The
spreadsheet macros proved helpful in classifying the data into the correct scenario and
then choosing the values to be used in the risk assessment.

Conclusion

       Methods can be easily employed during the field investigation phase of a project to
ensure levels of data integrity that are comparable to a Laboratory Information
Management System.  The first phase, which deals primarily with the  sampling event,
requires thorough planning, appropriate documentation and training of all field personnel.
Planning provides a forum for all project staff to arrive at a workable  solution and
documentation provides a means of ensuring that sufficient information is gathered
regardless of the individual performing the task. Typically, field personnel are typically
given only the minimum training required by their job: proper training and complete
involvement in the project allows the staff to realize the intended uses of the data.  This
translates into more accurate data collection.

       The second phase of data management should  be designed to minimize any manual
entry of data and should also allow sufficient information to be uploaded into a database
to permit automated review and analysis of all data. By using Monarch, a software that
will extract of information from ASCII files, it was possible to make use of all of the
information contained in a CLP diskette.  By using a relational database, it was possible to
easily perform the ad  hoc queries required to track all  samples from the time of collection
and to provide output for analysis.  This project made use of a variety of software
packages, including spreadsheets and risk assessment programs to minimize the chance for
human errors.  These elements of WCFS1 program allowed the staff to complete the RI/FS
on-time and on-budget.
                                        154

-------
23            EFFLUENT EMISSIONS MONITORING AT THE DOE HANFORD SITE

                    LaDell W. Vance. Principal Quality Engineer
                    Environmental  Management  Systems  Integration

                           Westinghouse  Hanford  Company
                                   P.O. Box 1970
                             Richland,  Washington 99352
      ABSTRACT

      There  are  numerous  regulatory requirements  controlling  the  effluent
      emissions  monitoring at a  U.S. Department of Energy  site.   This  paper
      defines  how these regulatory effluent emissions monitoring  requirements
      and the Quality Assurance oversight of these requirements were implemented
      by  Westinghouse Hanford Company, the  operations  contractor, at the  DOE
      Hanford Site.

      INTRODUCTION

      Westinghouse  Hanford Company (WHC)  struggled with the  implementation  of
      DOE  and  other  regulatory  effluent monitoring  Quality  Assurance  (QA)
      requirements.  There was little guidance or direction provided on how this
      should be  done.   By using the  QA organization  from the  earliest planning
      stages, a  more effective compliant process has evolved.

      These   initial   monitoring   requirements   were   prescribed   by   the
      implementation  of  DOE  Order  5400.1  which  specified  that  the   sites
      implement  plans  to  monitor  regulated hazardous and radioactive  effluents
      by  November 1,  1991.   Included in  this  order were the environmental  QA
      requirements  to be used in  this effort,  but there was little  direction
      given for  their  implementation.

      The  DOE   environmental  QA  requirements  are  very   similar   to   U.S.
      Environmental  Protection Agency (EPA)  Quality Assurance Management  Staff
      (QAMS) requirements that WHC had been using for well sampling and analysis
      activities.   These  plans had been  effective in monitoring well  water  so
      WHC decided to use  these as guidance for preparing plans for the  liquid
      effluents.

      The radioactive air requirements from 40CFR61,  Appendix B, Method  144,
      Section  4,  include the DOE requirements  plus  specific  items related
      directly  to radioactive air measurements, items  that  would normally  be
      considered  Quality  Control  (QC) activities.   It was decided  to  prepare a
      QA  Program Plan  (QAPP)  in  accordance with the EPA QAMS requirements  and
      include the stack specific measurement requirements as  appendices to this
      document.
                                          155

-------
These effluent monitoring activities  are  currently being performed with
added enhancements as the effort continues.

HANFORD SITE

The Hanford Site  covers  approximately  560 mi2 (1,450 km2)  of semiarid land
that is owned by  the U.S. Government and managed by  the U.S. Department of
Energy, Richland Operations  Office  (RL).   In  early  1943,  the U.S. Army
Corps of Engineers selected the Hanford Site as the location of reactor,
chemical  separation,  and  related  activities  for  the  production  and
purification of  special nuclear materials  and  other nuclear activities.
Since 1988, the mission of the Hanford Site has focused on environmental
remediation and restoration.

Activities on the Hanford Site are centralized in numerically designated
areas (see Figure 1).  The reactors are located along the Columbia River
in the 100 Areas. The reactor fuel processing units are in the 200 Areas,
which are on a plateau approximately 7 mi (11  km)  from the Columbia River.
The 300  Area,  located  adjacent to  and north  of the City  of Richland,
contains  the reactor  fuel  manufacturing  plants  and  the research  and
development laboratories.   The 400 Area, 5  mi (8  km) northwest of the 300
Area, contains the Fast  Flux Test Facility that  is used for testing liquid
metal reactor systems.  Other  areas and buildings  designated on the map
are for administrative purposes and do not produce regulated effluents.

Hanford Site  facilities are no longer producing  nuclear  materials  for
defense purposes, rather operations are continuing for cleanup purposes.
The air and liquid effluent streams are being generated during maintenance
or shutdown of existing facilities.
EFFLUENT MONITORING REGULATIONS AND QUALITY ASSURANCE REQUIREMENTS

Figure 2 depicts the roles of the DOE,  EPA, Washington State Department of
Ecology  (Ecology), Washington State Department  of  Health  (DOH),  and the
sources of the environmental QA requirements specified in the regulations.
This  paper  is  limited  to major  effluent  monitoring efforts  of  the
operating  facilities  that  require regulatory  control.   The  effluent
monitoring environmental  QA regulations are identified by the double-lined
boxes on Figure  2.  The  RCRA/CERCLA efforts required  for  cleanup,  waste
analysis, and efforts  specified  only for  DOE purposes are not included.

A major factor in almost all  environmentally regulated activities at the
Hanford Site is the Hanford Federal Facility Agreement and Consent Order
more  commonly  referred  to as the  Tri-Party Agreement  (Ecology,  et.al.
1992).  This agreement among DOE, EPA, and Ecology specifies requirements
and schedules for most environmentally  regulated monitoring,  operating,
and  construction activities.    As noted   on  Figure  2,  the  Tri-Party
Agreement relies on the  EPA QAMS documents for effluent environmental QA
requirements.
                                    156

-------
The DOE  effluent  monitoring QA requirements, DOE  Order 5400.1, General
Environmental Protection Program, Chapter  4,  Section 10, do not provide
direction on implementation. WHC decided to use the QAMS requirements for
all liquid  effluent monitoring activities.   These requirements specify
similar requirements while providing implementation guidance.

The other sources of QA effluent monitoring requirements found were:

      National Emission Standards for Hazardous Air Pollutants  (NESHAP),
      40CFR61, Appendix B, Method 114, Section 4, QA Methods

      National   Pollutant   Discharge   Elimination   System   (NPDES),
      40CFR122.44,  Establishing Permit Conditions

      Toxic  Substance  Control Act  (TSCA),  40CFR763.121,  Appendix  A,
      Quality Control Procedures,  and Appendix B, Detailed Procedures for
      Asbestos Sampling and Analysis.

There is only one  effluent stream, associated with the N Reactor, that has
an  NPDES permit.   Since  all  monitoring  activities are  specified and
controlled  by this  permit,  the N  Reactor stream  will not  be further
addressed in this paper. The TSCA controls are also not addressed in this
paper.

The remainder of this  paper will  discuss how the  regulated  liquid and
radioactive air streams were identified and the environmental QA controls
were applied.

INITIATION OF EFFLUENT  MONITORING PROGRAM

Regulated effluent  monitoring was  instigated in response  to  DOE  Order
5400.1,  which required the preparation of effluent  monitoring plans by
November 1, 1991.   The  Order states that

      "An implementation plan shall be prepared for each facility or
      group of  facilities,  the purpose of which  is to  ... comply
      with environmental regulations and DOE  policies."

WHC responded to this by preparing a management plan for facility effluent
monitoring plan activities to direct the identification and implementation
of these monitoring efforts.  The management plan specified the regulatory
criteria that would require monitoring, the facilities that  would need to
be  addressed,  the  basic   information  needed  in  a  Facility  Effluent
Monitoring   Plan   (FEMP),   and   the  schedule   for   completing   the
determinations.

The criteria, as noted in the management plan are taken from the following
regulatory requirements:

(1)   For Radioactive Air Emissions,  40CFR61.93,  "Radioactive emission
      measurements  ...  shall be made  at all release points which have a
                                     157

-------
      potential to discharge radionuclides into the air ... in excess of
      1% of the standard.'   As the EPA allowable dosage is 10 mrem/yr all
      facilities  that  could  cause  radiation doses  in  excess  of 0.1
      mrem/yr  to  the maximally  exposed  member  of the  public  must  be
      monitored.

(2)   For discharges of nonradioactive materials,  monitoring is required
      of  all  facilities with  the  potential  to  release quantities  of
      hazardous materials exceeding the reportable  quantities  listed in
      40CFR302.4,  "Designation  of Hazardous Substances."

(3)   All  potential  radioactive  discharges  to   the  ground  must  be
      monitored.

The plan preparation required that all  facilities be  reviewed to determine
if  any  of the  above  requirements  had  been  exceeded  and thus  require
monitoring.

FACILITY EFFLUENT MONITORING PLAN DETERMINATIONS

Each facility had  to be thoroughly examined to  identify each effluent, and
an analysis was done on each stream to  see  if any  of  the above criteria
were exceeded.  The determinations of the  streams  and  the preparation of
the FEMPs took over one year to  complete.

To put this in perspective,  there were  17 geographical areas that included
approximately 110  facilities.  Of these, 15 facilities, with 22 hazardous
and/or  radioactive  liquid  streams  and  7  radioactive emissions  stacks,
exceeded  the  criteria.   Discussions are ongoing  with  the  Washington
Department  of Health  that  may  add other  radioactive  stacks to  this
program.  There is also  another ongoing program (Air Emissions Inventory)
to identify the hazardous air emissions.

FACILITY EFFLUENT MONITORING PLANS

Each facility with a regulated stream(s)  was required to prepare a FEMP to
address the requirements specified  in the management  plan.   To expedite
this effort  and  to ensure   consistency  between  facilities,  a  guide for
preparing Hanford Site  facility effluent  monitoring plans  was  prepared.
This document expanded on the management plan requirements with guidance
on what to include in  each  of the following sections including:

      Purpose and  scope
      Applicable regulations
      Facility/process description,  source terms,  release pathways

      Characterization  of   the  airborne  and liquid  discharges  to  be
      monitored,  providing  potential effluent concentrations
      Characterization of effluent points, providing design criteria and
      technical  specifications  of monitoring/sampling  systems
                                    158

-------
      Historical monitoring/sampling data and a comparison of normal and
      upset conditions
      Sample analysis and laboratory procedures
      Notification and reporting requirements
      Interface with operational reporting requirements
      QA Plan
      Internal and external plan review requirements
      Compliance assessment describing conclusions reached, if monitoring
      was required and will system need upgrading.

The fifteen  FEMPs  were  prepared to these criteria  with  the appropriate
reviews and approvals.

QUALITY ASSURANCE INVOLVEMENT

The QA organization was  invited to  join  in  this  effort shortly after it
was started when the developers encountered problems trying to implement
the DOE Order 5400.1 QA  requirements.   The QA representative was familiar
with the QAMS document with its  defined method of implementation and this
was used  to implement the  DOE  requirements.   The  QA  organization also
reviewed the implementation documents to ensure that these were reviewed
and controlled in accordance with DOE Order 5700.6C, Quality Assurance.

The  facility  effluent  monitoring  QA Project  Plan that was  prepared
described the existing facility procedures and how they would be used to
implement the QA requirements.  However, this effort  was not successful as
the existing procedures  did not  have the  needed custody or control checks
required  for regulatory environmental QA  controls.   In  addition, each
facility had their  own  procedures with little  correlation between them.
There was  a need for some  standard operating procedures  (SOP)  with the
needed QA/QC controls.

The FEMP documentation provided the background information  that was needed
to prepare for monitoring the facilities, but other programs were used for
QA verification.  This  was done through  the implementation  of two other
regulatory requirements, the Liquid  Effluent Characterization Program and
National Emission Standards for Hazardous Air Pollutants (NESHAP).

LIQUID EFFLUENT CHARACTERIZATION PROGRAM

At the time of the signing of the Tri-Party Agreement,  the DOE agreed to
monitor  and  analyze  33  specified  liquid  streams, called the  Liquid
Effluent Study (LES), in accordance  with  the  EPA QAMS QA criteria.  There
had been  ongoing  monitoring of these  streams since 1985  in response to
congressional requests and  the regulators requested that this be continued
and that EPA QA requirements be applied to the activities.  These were the
same  streams as  those   identified  in the  FEMP determinations;  except
however,  as the  site  continues to  shut down  facilities,  some  of the
streams are no longer in use.
                                    159

-------
QUALITY ASSURANCE OVERSIGHT OF LIQUID EFFLUENT MONITORING ACTIVITIES

The initial LES sampling and  analysis activity was done shortly after the
signing of the Tri-Party Agreement  in 1989 without QA support and with no
QA/QC protocols.  Since information could not be verified or validated it
was  not  acceptable  to  Ecology.   After  this  was completed,  the  QA
organization was asked  to  perform  a post-completion surveillance of the
sampling   and   analysis  activities.  The   intent   was   to  verify  the
implementation of the required EPA  QAMS requirements.  After finding this
uncontrolled situation, the QA organization  was invited to be an integral
part in setting up the QA Program to ensure the validity of future data.

It was agreed initially that  a QA Project Plan (QAPjP)  would be prepared
in accordance with the  requirements of EPA  QAMS-005/80.   This  was later
expanded to include management control  and  changed to a QAPP.   There was
also consensus  that  all sampling  and  analysis  activities could  not  be
included  in  a single  plan.   Those activities that were common  to all
streams were  included  in  the  sitewide  QAPP.  Individual  Sampling  and
Analysis Plans  (SAP)  were prepared for stream-specific  information.

The QAPP provides the general  framework for sample collection,  laboratory
analysis,   and  data reporting  with the common QA requirements  for the
characterization of liquid  effluent streams. These common QA requirements
include  program description,  Data  Quality Objectives  (DQO),  program
organization,  sample control, laboratory  analysis,  and  data processing.

The  SAPs   are   wastestream-specific  documents that describe  how  QAPP
requirements are implemented.  The  SAPs  include  a discussion  of stream-
specific DQOs, a description  of the liquid  effluent, and a justification
for sample location  and frequency.   The  SAPs also  identify  any stream-
specific exceptions to the  QAPP.

An effluent monitoring  organization was responsible for  identifying the
regulatory  requirements,  preparing   the   QAPP,   and   specifying  the
implementing requirements and criteria  to include in the SAPs.   The SAPs
format  and content  were  based  on Guidance  for  Conducting  Remedial
Investigations  and Feasibility Studies under CERCLA  (EPA/540/G-89/004,
October 1988)  for a field sampling plan (FSP).

The SAPs are prepared by each  facility with  input from the common sampling
organization.    This  coordination  of  sampling  and  analysis  preparation
provides the control  needed to achieve  environmental QA implementation.

The environmental  and  facility QA  groups  work  together to perform the
stream surveillances.   The environmental  QA group reviewed  and approved
the QAPP and SAPs.  The QA group's familiarity with these documents allows
them to prepare the stream  surveillance plans.  The facility QA personnel
are familiar with the facility and thus better able  to  witness  the field
sampling activity.  Environmental QA is responsible for the surveillance,
but both groups coordinate  and approve  the  surveillance findings.
                                    160

-------
RADIONUCLIDE AIR MONITORING

The Clean Air Act was amended in December 1989 to  include 40CFR61, Subpart
H, National Emission Standards for Emissions of RadionucTides Other Than
Radon from Department of Energy Facilities.   This  subpart requires that a
QAPP for radionuclide  air  monitoring  be prepared in accordance with the
requirements  of 40CFR61,  National Emission  Standards  for  Hazardous Air
Pollutants (NESHAP), Appendix B, Method 114, Section 4, QA Methods.

The Washington Department  of Health was satisfied that the seven stacks
required monitoring, but did not  feel  that  the FEMPs adequately covered
the continuous monitoring requirements.   The DOH  required that regulated
stacks and analytical  laboratories be Point  by  Point  (PBP)  compared to
40CFR61, Appendix B, Method 114 to be certain that all requirements were
satisfied.  This major PBP exercise compared the existing air monitoring
equipment and program with that specified in the requirements, including
Section 4, QA Methods.

It was  decided to use the  QAPP as the binding  document to  control the
NESHAPs  activities.    All  stack monitoring sampling  and  analytical
requirements  are included  in the document.

NESHAPs QUALITY ASSURANCE  PROGRAM  PLAN  (QAPP)

There was a concern as  to how to address the QA requirements of Method 114
for all the various activities  at  the  stacks.  It was initially proposed
that an overall QAPP be prepared with  separate QAPjPs for each stack.  It
was finally agreed that a QAPP with the stack PBPs appended would address
all QA program requirements.

The QAPP was prepared in  accordance  with the guidance provided  by EPA
QAMS-004/80  and is  organized  as  shown on  Figure  3,  NESHAPs  Table of
Contents.  The  major  responsibilities and actions needed to control the
NESHAPs program are defined in  Sections 3.0,  6.0, and the appendices.

Section 3.0,  Quality Assurance Management, defines the various oversight
organizations that are responsible to  verify that the NESHAPs activities
are being  properly done.  The  body of the  QAPP  describes  these various
organizations and the  procedures they  use  to  perform  these oversight
activities.   WHC  is  organized with  a central  QA  group,  Environmental
Services  Quality Assurance.   This central  group reviews  and approves
documentation and provides guidance  to facility QA organizations, who
assist  with   field overviews.   The same  relationship exists  as  in the
liquid  effluent  monitoring where  actions are combined,  but the central
environmental QA group is  responsible  for the surveillances.

Section  6.0,  Radioactive  Air  Emissions  Measurement  Quality Assurance
Project  Implementation,  specifies the organizations that are performing
the   NESHAPs  activities.     This   section  describes  organizational
responsibilities  and  interfaces.   Included  are  responsibilities for
specifying NESHAPs requirements, sampling and sample transport, laboratory
                                     161

-------
analysis, flow data information, systems calibration, regulatory reports,
procedures preparation,  and data verification.  Section 6.0 was difficult
to  prepare  but  defining   these  responsibilities  and  organizing  the
activities was crucial to properly completing the NESHAPs activities.  In
preparing this document it  was determined  that  some activities were not
being  properly addressed.    These were  corrected  and  included  in  the
document.

The PBP comparisons for the seven stacks,  appended  to  the  QAPP,  provide
the total overall  program  control needed for NESHAPs  activities.   Items
that  are  specific  to  a stack  are   included  in the  PBP,   but standard
operating procedures are included, or referenced, in  the body of the QAPP.
There is an added advantage to this  in that if  additional  stacks have to
be registered  a  stack-specific PBP  can be  prepared and appended  to  the
existing QAPP.  There will  be  no need to prepare a QA Plan for each added
stack.

The QAPP provides the needed control of the NESHAPs activities but allows
the flexibility to add  regulated  stacks as  needed.

SUMMARY

The preparation  and  implementation  of the  QAPP was crucial  to properly
organizing the activities  and  obtaining verifiable data.

Both liquid and air monitoring started out without Environmental Services
QA  help  and  the  responsible  facilities were  at a loss  as  to  how  to
implement  these  new unique sampling  and  analysis  requirements.   The
Environmental QA person was  able to define what information was needed for
the QAPPs and help the responsible organizations in their preparation and
implementation.

It is important to have the  implementing organizations prepare the plans.
In the case of the FEMP  activities,  outside consultants  prepared  most of
the documentation without  facility input.   The  facilities were not aware
of the commitments made and were hesitant to implement these.  As noted in
the document,  the  LES  program is used  to  verify the control  of  liquid
effluents.

In  both  the  liquid and radioactive  air  monitoring  existing  facility
procedures were initially  used;  however,  it became imperative to  prepare
SOPs.  In  the FEMP program the existing procedures were not definitive
enough nor did they have the needed  QC controls.   In the  NESHAPs  program
it was determined  that it would be easier  to have common procedures  and
these have been prepared.

In closure, someone familiar with  Environmental  QA requirements,  but not
part of the assessment organization, should be a working partner with the
implementing   organizations  in   preparation   and   implementation   of
environmental  QA  requirements.     This  relationship  will  allow  for
compliance and will ensure  that the  data gathered is verifiable.
                                    162

-------
Radioactive
Mixed Waste
Storage Facility
                                                                                3000 Area
                                                                               H930S018.2
                                   Figure  1
                                       163

-------
                                                                                                                    40 CFR 763.121
                                                                                                                    App. A and B
o
                                                                                  i  40 CFR 61 App. B
                                                                                  1	Method 114
                   RCRA/CERCLA
                   Closure Plans
                           Ch.4,
                           Section 10
Ch.4,
Section 10
    CAA
   CWA
    DOE
    DOH
 Ecology
    EPA
  FEMP
    LES
NESHAP
 NPDES
  OEMP
  QAMS
    TPA
  TSCA
   WAP
EPA Authorizes State Agencies to Regulate
= Clean Air Act
= Clean Water Act
= Department of Energy
= Washington State Department of Health
= Washington State Department of Ecology
= Environmental Protection Agency
= Facility Effluent Monitoring Plan
= Liquid Effluent Study
= National Emission Standards for Hazardous Air Pollutants
= National Pollutant Discharge Elimination System
= Operational Environmental Monitoring Plan
= EPA Quality Assurance Management System
= Tri-Party Agreement                         H93osoi8.i
= Toxic Substances Control Act
= Waste Analysis Plan
                                                                    Figure 2

-------
           Figure 3.    National  Emission  Standards  for Hazardous  Air
                              Pollutants Table of Contents.
                                             CONTENTS
  1.0  INTRODUCTION  	    1

  2.0  QUALITY ASSURANCE POLICY STATEMENT   	    1

  3.0  QUALITY ASSURANCE MANAGEMENT   	    1
      3.1  ENVIRONMENTAL ASSURANCE   	    3
           3.1.1  Environmental Protection  	    3
           3.1.2  Environmental Compliance Verification 	    3
      3.2  QUALITY ASSURANCE  	    3
           3.2.1  Environmental Services Quality Assurance   	    3
           3.2.2  Plant Quality Assurance Engineers 	 . 	    4
           3.2.3  Audit Program Administration  	    4

  4.0  DOCUMENT CONTROL AND RECORDS   	    5
      4.1  ENVIRONMENTAL ASSURANCE	 .    5
           4.1.1  Environmental Protection  	    5
           4.1.2  Environmental Compliance Verification 	    5.
      4.2  QUALITY ASSURANCE  	    5
           4.2.1  Environmental Services Quality Assurance   	    5
           4.2.2  Plant Quality Assurance 	    5
           4.2.3  Audit Program Administration  	    5

  5.0  PERSONNEL QUALIFICATIONS  	    6

  6.0  RADIOACTIVE AIR EMISSIONS MEASUREMENT QUALITY ASSURANCE
      PROJECT IMPLEMENTATION  	    6
      6.1  INTRODUCTION 	    6
      6.2  ORGANIZATION AND RESPONSIBILITIES  	    7
           6.2.1  Environmental Protection  	    7
           6.2.2  Facilities	    7
           6.2.3  Occupational  Health and Safety  	   10
           6.2.4  Central Support Services  	   10
           6.2.5  Laboratories   .	   11
           6:2.6  Regulatory Analysis 	   12
           6.2.7  Facility Compliance 	   12
           6.2.8  Other Support Contractors 	   12

  7.0  PERFORMANCE AND SYSTEM AUDITS  	   12

  8.0  CORRECTIVE ACTION 	   13

 9.0  QUALITY ASSURANCE REPORTS 	   13

10.0  REFERENCES	   13


APPENDICES

A  METHOD 114 COMPARISON FOR STACK 291-A-1   	 A-1
B  METHOD 114 COMPARISON FOR STACK 291-8-1   	 B-1
C  METHOD 114 COMPARISON FOR STACK 291-Z-1   	 C-
D  METHOD 114 COMPARISON FOR STACK 296-A-22 	 D-
E  METHOD 114 COMPARISON FOR STACK 296-A-40 	 E-
F  METHOD 114 COMPARISON FOR STACK 340-NT-EX   	 F-
G  METHOD 114 COMPARISON FOR 222 LABORATORY 	 G-
H  METHOD 114 COMPARISON FOR 325 LABORATORY 	 H-
                                              165

-------
24                   QUALITY ASSURANCE AUDITS OF LABORATORIES

                     R. Terry Hinward. Senior Quality Engineer
                      Environmental  Services  Quality Assurance

                            Westlnghouse Hanford Company
                                   P.O. Box 1970
                             Richland,  Washington 99352
      ABSTRACT

      Experience gained  through audits  of numerous  laboratories across  the
      United  States  has helped  to  focus  on  some  common  problems  found  in
      laboratories.    Various  regulatory requirements  and  programs  help  to
      complicate the  audit process.  Many laboratories are almost overwhelmed by
      the requirements,  and the quantity and diversity of  audits  imposed upon
      them.    Occasionally  auditors  add  to this confusion  as they  sometimes
      arbitrarily  require   the  laboratories  to  follow   set  protocols   of
      documentation  instead of focusing on the end products  of the systems  in
      place.  Such actions can result  in the laboratories  developing  multiple
      parallel  systems  to  produce  the  same  end results.    This process  of
      multiple systems can be very costly and  a  source of numerous errors within
      the laboratory.  This presentation focuses on common pitfalls found within
      Laboratory Quality Assurance/Quality Control Systems and in the process of
      auditing such  systems and their performance outcomes.

      INTRODUCTION AND  BACKGROUND

      Requirements from  numerous  regulatory or governmental  agencies seem never
      to stop growing and expanding  in both scope and complexity.   Guidance for
      interpretation  and application of these requirements  is not  always clear
      or maybe even  available.   This can add  more confusion as to  exactly what
      is required and how to  approach the various situations.

      To  compound  these requirements,  numerous companies,  associations  and
      boards publish or provide  guidance and or requirements to  be  followed.
      Throughout the years, constant change and  additions  to requirements  has
      produced a tremendous burden  on both  the  laboratories and those  auditing
      the laboratories.   Even the terminology  and approaches used to  evaluate
      the laboratories  has  become very diversified and confusing.  At  present
      the author evaluates  some laboratories  within  the  DOE System and  several
      Commercial  Laboratories.    Samples  sent  to  these  laboratories   by
      Westinghouse Hanford  Company  may  be analyzed  under  several  different
      regulations.
                                         166

-------
Examples of these regulations are
            the Safe Drinking Water Act (SDWA)
            the Comprehensive  Environmental  Response,  Compensation and
            Liability  Act  (CERCLA)  and it's  Superfund  Amendments and
            Reauthorization Act (SARA)
            the Resource Conservation  and  Recovery Act  (RCRA)  and it's
            Hazardous and Solid Waste Amendments (HSWA)
            the Clean Air Act  (CAA)  and the  National  Emission Standards
            for Hazardous Air Pollutants (NESHAP)
            the Clean Water Act (CWA) and the  National Pollutant Discharge
            Elimination System (NPDES)
            the Water Quality Act of 1987 (WQA)
            the Toxic Substances Control Act (TSCA)
            the  Federal  Insecticide,  Fungicide  and  Rodenticide  Act
            (FIFRA), and
            the Solid Waste Act (SWA).
Sample types include:
            radiochemical
            mixed waste
            hazardous waste
            water quality
            air quality, and
            can consist of several matrixes.

These matrixes include:

            water
            air
            solid waste
            animal tissue, and
            vegetation.

These samples may or may  not  be subject to the Hanford Federal Facility
Agreement and Consent Order of  May  1989.   This agreement is between the
Washington State Department of  Ecology,  the United States Environmental
Protection Agency  and  the United  States Department  of Energy.   It is
commonly  referred  to as  the Hanford  Tri-Party Agreement.   Additional
requirements from this agreement are contained in most of the Statements
of Work  (SOWs)  to  contract laboratories.   It  is  extremely difficult to
provide  all  exacting  guidance  in  this fashion  due  to  the  enormous
diversity of work.

A  discussion  of lessons  learned  and  some  common   pitfalls  will  be
presented.  It is not the  intent  to single out laboratories or types of
laboratories but to discuss things found throughout several laboratories
in the nation.   No attempt will  be made to quantitate the extent of such
problems or to statistically treat this subject.
                                    167

-------
AUDIT. SURVEILLANCE. ASSESSMENT OR INSPECTION?

In evaluating  these laboratories each of  these  terms has  been  used  in
different  notifications  to  laboratories  about  upcoming  activities.
Several laboratories have expressed concern about what to expect based  on
what term is used.   EPA in general does different types of audits (system
and performance as an example), whereas the NRC calls similar activities
inspections.     The   DOE  sometimes  breaks   these  activities   into
surveillances, inspections,  assessments and audits.   In  this paper, the
term audit will be  used  to  cover all  these activities.   Each agency can
readily define what is  meant by the terms it uses  but  often fails to do  so
in a clear way to  the laboratories it  associates with.  What is important
to the  laboratory  is to  know  the extent  of the activity  (audit);  time
required  at  the  site;  requirements  for  availability of  personnel,  the
facility and records; what topics and  areas will be covered; the focus of
the audit in some detail.

The author's experience has  shown that sharing  the topics to  be covered by
the audit checklists with the laboratory, a few weeks  before  the audit has
in general been very successful.  This enables the  laboratory to review
requirements  (often finding  areas  of miscommunication  or  incomplete
understanding), allows time for  the  laboratory management  to  perform a
focused self assessment (management audit), allows the laboratory time to
gather the needed  evidence and yes, even time to get everything caught up.
The philosophy used here  is  a win-win approach.

A mistake auditors can  make  is  to equate the success of an audit with the
greater the number  of  findings,  the better  the audit.  Large numbers of
findings  often  result  from  poor  communication and  produce  a  lose-lose
situation.  If a company  can correct mistakes before the audit, much time
is saved  in getting  the  laboratory into compliance.   The  audit trail  is
still  there  to show when the  mistakes were  corrected  and  the  auditor
should require the laboratory to  still assess what potential  impacts were
to the data,  and to correct  or flag the affected data  if needed.  Much can
be learned  about   the  attitudes,  philosophy  and  general  operations  of
laboratories  by  this  approach.    When  other purposes  are  paramount,
checklist topics can be withheld from the  laboratory.

REQUIREMENTS

Great  care  needs   to be  taken  to identify  and  evaluate  based on  the
established requirements.   When  auditing a  laboratory's work done  under
specific regulatory requirements, those requirements need  to be complied
with  unless  a written  waiver  has  been  obtained  from  the  applicable
regulatory  body(ies).     Auditors  sometime impose   certain  formats  or
requirements which are  not based on  the regulations  or the contract with
the laboratory.  These are  not enforceable and can be a  source  of both
contention and loss of  credibility.

Sometimes, laboratories  sign contracts in  which  additional  requirements
have been imposed by the client.  These laboratories sometimes try to hide
                                   168

-------
behind the  fact that  the  area of  noncompliance  is not  regulatory and
therefore is not  important.   Any such written agreement  may be just as
important as the regulatory requirements to the client.

GIVE ME A CHECKLIST!

A critical  mistake some companies make is to develop checklists to be used
on audits and  then send  just anyone trained in auditing  to perform the
audit.  First,  of  all,  it is very difficult to produce a generic checklist
that  can  be used  in  various  laboratories.   Second,  the  available  of
commercially generated checklists is decreasing as the industry reacts to
suits or threatened  suits  from companies.  These companies have  used a
generic checklist  and  have missed critical items that  have resulted  in
heavy fines or loss of money from poor investments.

The most significant single point to be emphasized in this paper follows.
The  author  fully  believes  that  an auditor  should  be  recognized  as
technically competent  in the area of the audit being dealt with.   This
means that teams will often be used to perform the audit.   The checklist
is only  a  guide  and  a reference  point  and extensive  knowledge  of the
subject can not be  substituted for by making the checklist more extensive.

Knowledge  gained  through   audits  is  vital   to  contracts  and  strong
consideration  should  be  given  to who will  conduct  it.    Auditors
inexperienced  in a given area tend to do one of two things.   Either they
don't act in areas of concern because they aren't sure of the situation,
or they  have multiple findings which are not  valid.   This misuses the
laboratory's time and can cause needless confusion  and conflict.  Everyone
makes mistakes but technically competent auditors tend to make fewer and
less critical ones.

COMPLIANCE VERSUS  PERFORMANCE

Significant confusion exists in both laboratories  and with auditors as to
what  is  the  governing  consideration  between  good  performance  and
compliance to  requirements.  This  is not an easy  question to answer and
the correct answer depends upon individual situations.

Some agencies  have been more traditionally  performance  oriented whereas
others have  been  more compliance oriented.  All  have  incorporated both
elements but have  emphasized one  to a greater  degree.   For  example, EPA
has Performance Evaluation (PE) and Round Robin Sample Programs in which
it can  be  determined  in  a given  matrix if the  laboratory can produce
results that meet certain accuracy,  precision and detection requirements.
Good scores on these samples are mandatory to receive work or to continue
to receive work.  On the other hand other agencies often have checklists
based on compliance to very specific parts or steps of procedures and may
or may not have quantitative means of assessing the final  outcome.

The author, given  the  regulatory  bodies  that will  assess  the laboratory
results in this case, takes the stance that both of these considerations
                                    169

-------
must be fully satisfied.  Until  all  regulatory bodies who overview this
work agree as to acceptance of a different approach, both considerations
must  be  fully  satisfied.    An example  is presented  in  the  following
discussion of the potential  different approaches.

PROCEDURAL COMPLIANCE

A good example of the conflict between compliance  versus performance comes
in analytical procedural  compliance.   Some companies and auditors contend
that regulatory  procedures must always be exactly adhered to; while others
say  that,  if  the  performance  can  be  shown  to  be  equivalent,  a
significantly modified procedure is acceptable.  The author's experience
has been that most  companies take it for granted  that  all laboratories
follow the approved procedures.

While performing audits,  quite the opposite has been found  in many cases.
While  reviewing  RCRA water pH procedures  of  some 22  laboratories,  the
author found on  his first  visit that  21 out  of  22  procedures  were
noncompliant in following the  national  standard  procedure.   A procedure
found in  SW-846  and one found in another  approved  Standard Method were the
two procedures allowed to be  used  in  these contracts.  Both required that
separate aliquots of the  water sample be repeatedly measured until the pH
measurement agreement was within 0.1 or 0.05 pH  unit respectively.  Twenty
one of 22  laboratories were  performing  sample measurements  on  a single
aliquot.   Some  laboratories  had  other nonconformances  to  the procedures
also.  To the author's knowledge all  22  laboratories's pH procedures are
now compliant in this regard.

The  point  that  is  emphasized is  that  pH  is a very straight  forward
measurement and  if there  is this much  noncompliance to national procedures
with  the  pH analysis, what  about  all the other analytical  procedures.
There are several significant noncompliances  to regulatory procedures in
well  know  laboratories today.   Laboratories  have the  responsibility to
report all  potentially quality  affecting  modifications  found  in  their
procedures that  are reported to be "equivalent to" or reformatted national
standard procedures.   Auditors can  not afford  to  take  quick  looks  at
procedures if their contract  requires compliance.  Companies need to know
if strict procedural  compliance is mandatory  on the samples submitted to
laboratories and, if so,  hold laboratories to the responsibility of being
compliant.

The pH measurement  example  has  been attempted  to  be refuted  with  the
argument  that  through  Performance  Evaluation   (PE)  samples and  other
evidence that there  is a sound scientific  basis  that  the  correct answer
was obtained despite  the nonconformance.   This argument stands it's ground
if performance alone  is an acceptable end product. When the potential for
litigation exists concerning the conclusions  drawn based on these sample
analyses, the question of procedural compliance becomes overpowering.  The
bottom line is to  know the requirements for the  samples in question and to
have written waivers  if  necessary.   In  most cases  such  waivers  are very
hard to obtain and involve  significant cost and time.
                                    170

-------
TRAINING

This Is an area  in which most laboratories have trouble understanding what
is required and  in presenting proper evidence.  Training requirements will
vary  based on  company,  regulations and  situations encountered.   Some
companies require less strict training  documentation than those for which
these audits were performed.   The requirement  categories presented here
are:

            Qualification requirements  including  educational and training
            Job or position descriptions
            Evidence of training to policies and procedures
            On the job training or other performance indicators
            Documentation as  to the required frequency  of training  for
            specific tasks
      •     Evidence of training and approval  before sample analysis

Common deficiencies found include:

      •     Not having specific job descriptions
      •     Not having signed  qualification statements  listing specific
            qualifications required and demonstrated
      •     Procedural  or  policy  reading  lists  that   do not  require
            training to revisions
      •     Training to revisions up to several  months after put in use
      •     No  indication if this item is a one time requirement or what
            the requalification period is
      •     Copies of test scores or QA sample results with no indication
            in  writing  what a  passing score  is  and  what happens  if a
            person does not pass, or
      •     No record of what actions have been performed by the analyst
            before they are signed off as qualified.

A common mistake of auditors which causes  laboratories time and financial
burdens is to insist where or in what format this evidence  is kept.  What
is needed is for the laboratory to be able to produce the documentation.

DOCUMENT CONTROL AND CHANGE CONTROL

The major aspects of document and change control  that appear to be lacking
in some laboratories will  be discussed here.  Laboratory personnel need to
work to controlled copies of procedures.  Controlled copies means that a
system of accountability  has issued and assigned specific procedures that:

            are identifiable as to review and signature responsibility
            are traceable to issue and effective date
            provide clear indication as to what types of samples, matrixes
            and analytes are covered
            reference standard procedures this procedure meets
            clearly indicate what QA/QC accompanies the procedure  and that
            the copy is identifiable as a controlled copy.
                                    171

-------
The laboratory should develop a system to assure that when new revisions
of procedures are issued, training documentation  is  in place before the
procedure is used on client  samples; that all copies of old revisions are
removed from the work  place;   that the master copy of both  old and new
procedures are securely stored and maintained.

RECORDS MANAGEMENT AND SECURITY

Records must be  stored  in a certain manner to meet requirements. These are
not  universal  requirements  and  potential   laboratory  shortcomings  are
offered  as  things  to consider.   Laboratories  were  required,   by  the
contracts associated  with  these observations, to  store records  in  the
equivalent of an one hour lockable fire safe or cabinet.  These containers
needed to be either locked at all times when  not in actual  sight and use
by the records custodian (or delegated representative) or they needed to
be in a locked restricted access room to which only identified persons had
access.   Storage in  two separate  locations that  meet other  specific
requirements could substitute for the  fire  safe  requirement.

Even though  several  laboratories signed contracts with these requirements,
it  took  several  months  to  have most  come   into  full   compliance.
Inexperience as  well as incomplete checklists significantly contributed to
the delay in compliance.

CORRECTIVE ACTION AND NQNCONFORMANCES

Laboratories should  have systems to identify  nonconformances.  They also
need  systems to  identify,  track,  respond to, and correct conditions  or
items identified for corrective action.  Nonconformances are conditions in
which material or action does  not match  the  requirements  of  procedures,
contracts or  policies.   If such nonconformances  are identified  to  be
potentially quality affecting  or  violate the terms of  a  contract, they
need  to  enter  the corrective  action  system.   The most common  problems
encountered in  the  audits  of these laboratories ranged from not  having
systems  in  place  to having  systems  that  only  partially  addressed
corrective actions.

Conditions were often found  not to be  formally addressed in the first and
sometimes in follow up visits to laboratories,  such as

            loss of temperature control
            missed holding times
            missed reporting times
            poor results on round robin or  PE samples
            loss of  document or change control
            loss of  chain of custody or  internal chain of custody control
            loss of  or mishandled samples
            training noncompliances
            procedural nonconformances
            procurement noncompliances
            using standards or chemicals past their expiration date
                                   172

-------
            using noncompliant  standards (or  lack  of documentation  on
            standards)
            incomplete labeling of (standards,  reagents or samples)
            problems with balances
            not maintaining proper calibration  of instruments
            problems with water systems
            data calculation or reporting problems
            lack of proper control charting where applicable
            not meeting laboratory data specifications or contractual data
            quality objectives, and
      •     improper documentation of (standard, reagent or sample) make
            up.

The other  common situation was  to find that  follow  up and  closure  of
identified conditions (especially internally identified conditions) were
either not being done or  were  not  being documented.  Few laboratories had
developed trending systems and a  major weakness was that  the effects  of
deficient conditions on past work were not being evaluated adequately.

INTERNAL CHAIN OF CUSTODY

The contracts  audited against,  required documented  internal chain  of
custody.   This meant that at  any  given  time that the samples could  be
accounted for  and could  be traced  to  where  they were.   Personnel  had  to
sign out  the  samples from storage areas, maintain custody requirements
while they used the samples and signed unused portions back into storage.
Several  laboratories initially  had  difficulty  in  complying with  this
requirement.    Once   systems  were  developed  the level  of  compliance
dramatically increased.

PE AND ROUND ROBIN PROGRAMS

Most laboratories audited  were participating in more than  one PE  and/or
round robin program.  Most laboratories  fared  satisfactorily to well  in
the programs.   The major  problem  identified  in  this area was that many of
the laboratories  were not taking advantage of the  information  obtained
from these studies.   First, when passing scores  were obtained, those areas
in which the results were outside the  control limits were not followed up
on  to   identify  whether   a systems  problem existed  for   that  analyte.
Second,  samples  that could  have had  incorrect results  reported  for a
certain analyte as indicated by poor PE results  were not checked to see if
they were affected.

TEMPERATURE CONTROL

Refrigerator and  freezer  storage  requirements  for samples  and standards
were checked in the audits. Standard  storage requirement differences for
CERCLA and RCRA programs  were not accounted for  in many  laboratories.  The
requirements for  semi volatile organic  compounds (semiVOAs) illustrates
several points.  The two  cited programs require some of the standards to
                                    173

-------
be stored  at  different temperature  ranges  than for  the  other program.
Most of the same standards for semiVOAs can be used for both programs.

The  dilemma  for  the  laboratory  is  to  maintain two separate  sets of
standards in two different freezers or  store the  standards under the most
stringent requirements  and meet the intent  of both  programs.   Here is
where experience and knowledge are essential for the auditor.   In the case
of semiVOAs the critical factor is to  prevent volatilization and escape of
the analyte and no ill affects will  be presented to the standards stored
at the  slightly lower temperature range of the  more  stringent program.
The author's personal  opinion and policy is to not  present  a finding if
the laboratory  takes this stance.   It can be argued that  this  is not
strict  compliance  and  a slight  possibility exists  that the  practice is
quality affecting  but the author stands by  his opinion  until  directed
otherwise.

The narrow storage range for  inorganic  samples  is much more critical.  If
one  stores  at  lower  temperatures the  possibility  of freezing  quickly
increases.    The  freezing  process  can  not  be  tolerated.    Several
laboratories  were  found   to have   informal   temperature  documentation
programs during the first visit  to  them.    Problem  areas  included the
following:

      •     measurements done in  air,  not  liquid, which  caused frequent
            excursions from the  acceptable  temperature range,
      •     frequent excursions  from  acceptable ranges with no corrective
            action,
      •     no  documentation  as to  which  samples  were associated  with
            which refrigerators  or freezers,
      •     incorrect  postings  on refrigerators and/or  freezers as  to
            temperature requirements,
      •     no  procedure   or  policy  as  to  how to  handle  temperature
            excursions,
      •     use of non-calibrated thermometers,
      •     use of calibrated thermometers  passed their expiration date,
      •     not measuring  and/or properly documenting measurements and not
            documenting temperatures  of incoming samples.

Formal  documentation  and  compliance  improved  after the first  visit and
with each subsequent visit.  The most persistent problem remaining is the
formal documentation of problems and addressing  types of potential impacts
to specific samples.

INTERNAL QA/QC PROGRAMS

The extent of  the need for  this  program varies significantly  with the
needs of individual companies.   Specific requirements were given to these
laboratories in the regulations  and with the contracts governing the work
discussed  here.   Several  laboratories did  not  meet the  intent  and
requirements of this program until after the  author's first  audit.  Some
                                   174

-------
then came  into  partial  compliance and it was  after  the next audit that
more progress was made.

Internal QA/QC is as much for the benefit of the laboratory as it is for
the customer's benefit.  Here is where a good  idea of how the laboratory
systems  are  functioning  is  obtained on  a very  frequent basis.   Good
programs are  very cost effective and help avoid a  lot of repeat work.
Laboratories  were encouraged  to take  full  advantage  of opportunities
provided by internal QA/QC programs but were audited to the requirements
set forth.  Several  root causes of types  of deficiencies  found were traced
to not having an adequate internal QA/QC program.

INTERNAL LABORATORY AUDITS (SURVEILLANCES ETC.)

Internal audits of two types will be described.  EPA considers audits as
either system audits or performance  audits. By system audit, it is meant
to assess  if  all  portions  of a needed system  or activity were developed
and were auditable.  Being auditable is defined here as being able to be
traced to meeting requirements and in  being adequately documented.  Often
the  activities  involve  looking at  documentation  and  handling  much
paperwork.  Visual  inspections of processes can also be included.

The term Performance Audit is commonly used in  two different ways.  First,
this term  is used to describe assessments of activities as they actually
happen.  The performance of  processes or individuals is measured against
procedural or contract requirements,  in the real time mode. In this case,
a system audit could also contain a  performance portion.  The second use,
and  the  more  common  CERCLA  or  RCRA  definition  is  to  measure  the
performance  outcomes  against  the  requirements.    For  example   is  the
required   detection  limit,   accuracy,   precision,   holding   times  and
completeness being obtained.

Either  type  of audit can  range  in  scope from  informal  to formal,  from
minutes to several days or even weeks, from focused to  a narrow scope to
being very broad.   The number of auditors can vary significantly also.
The  scope  for  external   audits  is  even  more  variable.   For  example,
external performance audits  involving more than one laboratory may focus
on shared  Performance Evaluation (PE) and/or Round Robin samples, and on
the results of split samples performed by all   involved  laboratories.


Few laboratories were meeting the internal  audit requirements before this
audit program was  started.   These requirements had been imposed by both
the regulations  and the  contracts but were not fully  understood by the
laboratories  and  in  most  cases were  being   handled  informally.   The
planning,  scheduling, completion percentage, quality, quantity and scope
of internal audit activity dramatically  increased  in most laboratories as
audits continued.
                                    175

-------
LABORATORY MANAGEMENT ASSESSMENTS

These are activities that have met with varied requirements and success.
The actual requirements  are  somewhat  vague and as such  a  wide range of
laboratory management assessment practices have been experienced.

The principal  reasons for this process is to keep management in  touch with
what is actually happening in the laboratory and to provide  an independent
evaluation of the  quality  of  work being  done.    Actually  seeing  and
evaluating laboratory operations on  a real  time basis helps management to
stop  noncompliances   and   shortcut  processes  that  may  creep  into
laboratories.   It is surprising how managers were sometimes not aware of
current conditions in portions of their laboratories.

QUALITY CONTROL CHARTS

CERCLA type analyses do not specifically require control charts. Here the
limits are preset  and specific  instructions are  given as  to  what to do
when the quality control  limits  are  exceeded.  Control charts can be used
in CERCLA work for the purpose of trending  data.  For RCRA  and most other
work, control  charts are necessary.

The three most common types of control charts are standard recovery charts
that indicate  the laboratory bias for that analyte, spike recovery control
charts that  demonstrate accuracy for  a certain  matrix  for  a specific
analyte  and duplicate or replicate  control charts that  demonstrate  the
precision for a certain analyte.   Control charts need to be kept up at or
near  real time.    Their use,  besides  defining  working  accuracy  and
precision limits, is for warning  of marginal or unsatisfactory conditions
needing  immediate  attention  and which  might  lead to  formal  corrective
action.   They are also  used for  trending  conditions.   The  most  common
problems encountered in this area were:

            not having control  charts
            not having all  three types  of charts
            not keeping the  charts on a  real time or near real time basis
            not acting or documenting actions when limits were exceeded
            not tying noncompliances to specific  samples,  and
            not recognizing that  seven consecutive data  points on the same
            side of the mean constituted an out of control  situation.

LABORATORY DATA SPECIFICATIONS VERSUS DATA QUALITY OBJECTIVES

Each  laboratory  needs to determine  and provide  to  it's  customers  the
laboratory data specifications.   This helps the customer evaluate if the
laboratory can meet the customer's needs.  Accuracy, Precision, Detection
Limits, Instrument Detection Limits,  Quantitation  limits (these  are called
by  other names  also),  and  other  types of laboratory quality  control
parameters are these specifications.
                                   176

-------
Each customer needs to identify what it's data quality objectives (DQOs)
are that apply to the laboratory analysis.  Examples of these include:

            precision
            accuracy
            representati veness
            completeness and
            comparability (the PARCC parameters).

Specific  method requirements,  detection  limits,   instrument  detection
limits, quantitation limits  and other types of laboratory quality control
parameters are also specified.

The auditor needs to determine if the above specifications match.  In the
observations discussed  in  this paper,  it was not  possible  to  properly
evaluate this  information as it was not  sufficiently  available  in  most
cases.   It  appears  that some clean up  projects  are struggling  to first
develop  adequate  DQOs   and   then  to  properly  pass  those  on  to  the
laboratories.  It also appears that the laboratories have in general not
defined their data specification capability in enough detail  to be fully
auditable.

RADIOCHEMICAL PROCEDURES

The regulations and requirements for radiochemical and mixed waste samples
are the  most complex and incomplete of any  involved.   There are  a few
methods  or  procedures  that  have  national  recognition.    It  is  this
auditor's experience that compliance with these  is  less  than with other
procedures.

Several  types  of  analysis   have  individual  procedures  that have  been
developed by various  laboratories for their own use.   These procedures
often employ different separation and/or  analysis techniques and are not
easily  comparable.    In  the  contracts  involved  in  these  observations,
multiple radiochemical laboratories were  used.

A  PE  sample program is very difficult  to  develop for  many of  these
analytes  and  the unavailability  of  this tool has  severely  handicapped
evaluating  the comparability  of the  results  from  these  laboratories.
Extensive work is  needed in  this  area.  A final  observation  regarding
radiochemical laboratories is that in general they appear less focused on
procedural compliance and on  adapting to  changes than do hazardous waste
laboratories.

SUMMARY AND CONCLUSIONS

Requirements that laboratories must comply with are  extremely complex and
variable.  These same requirements impose much difficulty to those who do
audits  (various types  of  assessments)  of  the  laboratories.    Clear
direction as to how to  interpret  and  impose these  requirements  is not
always  obtainable.   There  remains significant  controversy  even  within
                                    177

-------
regulatory bodies as  to  how to apply some of  these  requirements.   This
situation adds stronger argument  to the position that auditors need to be
well informed both technically and with the quality assurance process.

Information obtained through performing audits of  a  variety of types of
DOE System and Commercial Laboratories has been shared.   No attempt has
been made  to  identify  these laboratories  or to quantitate  the problems
involved.   Examples  of  some  types  of problems  encountered  have  been
presented.  It must also be  recognized that this  is a living system with
constant change and audit programs must adapt to  the changes as must the
laboratories.

Repeat  rounds  of  audits   have  been  accomplished  and  compliance  to
requirements has been  evaluated.  It is concluded,  that when  requirements
and failures  to meet  these  requirements  were documented  in  the  audit
process, the amount  of time  to  come into compliance decreased.  Increased
frequency  of audits  from  yearly  to  twice  a  year  also   sped  up  the
compliance process.   The amount of improvement  has  not  been  quantitated.
If improvement in  performance results can be tied to the audit process was
not evaluated.
                                   178

-------
SAMPLING/FIELD

-------
PRACTICAL SUGGESTIONS  TO IMPROVE THE QUALITY  OF FIELD WORK
                   - SAMPLER'S  PERSPECTIVE

T. Diebold,  U.S. Environmental Protection Agency, Region IX,
San Francisco, CA  94105.

ABSTRACT

Before  you  attend  another  meeting on  field  sampling  and
quality assurance, consider the sampler's perspective on the
work they perform.

Several conferences  have been held recently  to discuss the
quality of  environmental monitoring data.   The discussions
have expanded beyond an initial focus on laboratory analysis
to include debates on the quality of field sample collection
activities.   Increasing emphasis is being put on assessing the
errors associated with  sample collection, which may be carried
through  the  entire  site  evaluation   process.    In  these
discussions,  samplers  can  provide  a valuable perspective on
improving the quality of field work.  To  date, their potential
to contribute in this area has not been fully realized.

In this presentation, some of the root causes of problems in
sample  collection  activities  will  be  discussed  from  a
sampler's point of view.  Examples  from  actual  field sampling
projects will be  used  to illustrate some of these problems,
which can occur in every stage of a project from planning to
sample handling.

Common pitfalls to be avoided in the planning and scheduling
stages will be  highlighted,  and  workable solutions to these
problems will be proposed.  Other  practical  suggestions to
improve the  quality of field work will  be  presented  in the
form of "Simple Things You Can Do  To Succeed in Sampling".
Sampler training and certification  will  be discussed, and the
benefits of cross-training of field and laboratory personnel
will be emphasized.  Lessons  learned  in the laboratory that
can be applied to the field will be shared.   The application
of Total Quality Management principles to field work will also
be discussed.
Notice:  Although the information discussed  in this paper has
been  funded  wholly   or   in  part  by  the  United  States
Environmental Protection Agency,  it has not  been subjected to
Agency review and does not necessarily reflect  the views of
the Agency and no official endorsement should be inferred.
                              179

-------
07      EVALUATION OF THE HYDROPUNCH**1 TO ASSESS GROUNDWATER
£ /                r«r»ijrp»»f TWfcTTniJ T*V VOT.aTTT.T3 ORRANTCS
  CONTAMINATION BY VOLATILE ORGANICS

          Charles Van Sciver
            Erica Wallace
New Jersey Department of Environmental
        Protection and Energy
  Environmental Measurements Section
          25 Arctic Parkway
     Trenton, New Jersey   08625
     ABSTRACT
     Hydrogeologic   investigations   have   traditionally   been
     performed with the use of monitor wells. Their disadvantages
     include  the  high  cost  to   install,  develop,  sample  and
     dispose of generated  cuttings,  development and purge water.
     Data  is  not  available  from  a  well  until  it  has  been
     installed,  which  can  lead   to  numerous  misplaced  wells,
     providing useless  data. This  can  lead to  an investigation
     requiring months  to  complete,  as  the process  is  repeated,
     until wells are properly located to yield the required data.

     The HydroPuncht111 II is a groundwater sampling tool developed
     for  rapid  and  cost  effective  collection  of  groundwater
     samples  from  any  depth in the  saturated  zone  without the
     installation  of  a monitor well. The tool was  evaluated to
     determine  a  cost  effective   decontamination  procedure  to
     prevent cross  contamination between  sample locations and if
     representative groundwater samples  could  be  obtained  when
     operating in the "hydrocarbon" sampling mode.

     Decontamination  procedures were  tested  on a  HydroPunchtm
     unit exposed  to  pure gasoline. After  decontamination blank
     water was run through the tool and  analyzed.  The procedure
     which provided non-detects for the  target  analytes  was:  a
     soap and  water wash  (laboratory grade  glassware detergent)
     followed  by  a  steam  wash   and  distilled/deionized  water
     rinse.

     Samples of groundwater  were collected  from the HydroPunchtm
     II in the hydrocarbon sample  mode and  from stainless steel
     well points.   The units were  placed so the  screen  of each
     tool would sample the same cross sectional interval within a
     5 foot radius  of  each other.  Twenty  two paired samples were
     collected and analyzed by  EPA method 602 plus  xylenes.
     Statistical  results using a  paired  T-Test  and the analysis
     of plots  generated from  logs of the  sample concentrations
     indicate  no  significant differences  between the  tools and
     good correlation  of data. This evaluation demonstrates the
     HydroPunchtin  II   in  the  hydrocarbon  mode  will provide  a
     representative sample of volatile organics in groundwater.
                                   180

-------
28              AN INTEGRATED APPLICATION OF FIELD SCREENING TO
                ENVIRONMENTAL SITE INVESTIGATIONS:  A CASE STUDY

                           Tina Cline-Thomas, William Mills, Alex Tracy
                         Woodward-Clyde Federal Services,  Rockville MD

      Abstract

            A base  in the Washington D.C. area was slated to undergo facility expansion.  This
      expansion was  to include construction of a commissary and parking lot, along with movement
      of an existing sports field and a playground to new areas.  Immediately prior to the start of the
      construction, information became available which indicated that a  landfill  had  existed in the
      general area slated for construction activities. The exact location and extent of the landfill were
      not known.  Woodward-Clyde Federal Services (WCFS) was retained by the Baltimore District
      of the U. S. Army Corps  of  Engineers (COE) to perform an investigation  of the area to
      determine the extent of the landfill and the health risk to construction workers and playground
      users.  Due to  the construction schedule, the project work had to be completed in an 8-week
      time period.  Normally this type of project would require up to 6 months. In order to meet both
      time and budgetary constraints, an intensive field sampling effort was undertaken in conjunction
      with the use of field screening methods.  Several field screening methodologies were employed
      to more fully characterize the site during the time between field  sampling and lab analysis: 22
      metals by portable X-Ray  Fluorescence Spectrometry  (XRF) and  Polychlorinated Biphenyls
      (PCB) and Aromatic Hydrocarbons (AH) by immunoassay. The field screening was used to:

               •  direct field sampling efforts  by delineating contaminated areas
               •  prioritize samples for laboratory analysis
               • provide the laboratory with information on  the expected range of contaminants

         In total, approximately 130 samples were screened in the field and approximately 30 of those
      samples had results verified by lab  analysis.  The metals results  for XRF and lab analysis
      generally corresponded to  each other, provided  samples were  thoroughly  homogenized.
      Although some false positives  were  observed  by field screening for PCB  and  AH,  no false
      negatives were observed.  This presentation will  discuss time and budgetary savings, QA/QC
      procedures and comparability of the field screening and lab results.

      Introduction

         Woodward-Clyde Federal Services (WCFS) was charged with the task of clearing a site for
      construction activities within 8 weeks while ensuring that  sufficient  samples had been taken to
      characterize the site and that those sample results were accurate. In addition to the historical
      information which indicated the presence of a landfill whose size and contents were unknown,
      there was the possibility that the landfill  area contained  burn pits where  PCB  transformer
      carcasses had been  disposed.  Some preliminary  work  indicated  elevated levels of PCBs,
      miscellaneous other organics and metals in the areas where the present playground was located
      and where the athletic field was to be relocated (Figure 1).
                                            181

-------
                    Figure 1:  Site Map Showing Proposed Construction
    As a result of health and safety concerns for construction workers and future residents a site
clearance  was undertaken.   Since the construction contract had already been awarded and the
government would face penalties if construction was delayed the time frame available for the
investigation was very short.

    Lab analysis takes 3-4 weeks  at normal  turnaround times.   Faster  turnaround times are
possible at premium  rates but even the fastest analysis requires 24-48 hours before results can
be reported.  The approach that was developed for this project involved field screening and lab
analysis with five day turnaround.  Lab analysis was  required to verify field screening results
and  to  provide the  quantitative information required  for the  health-based  risk  assessment
samples.  Field screening was performed on soil samples for PCB and aromatic hydrocarbons
(the Petrorisc"  immunoassay  which  was chosen is  designed  to provide  total  petroleum
hydrocarbon data but  is very sensitive to bi and tri-cyclic  polyaromatic  hydrocarbons) using
immunoassay  technology,  and  22 metals  using a  field portable X-Ray  fluorescence  (XRF)
spectrometer.   The field screening methods  were  used to  prioritize samples for lab analysis,
provide information  to the lab  on the  approximate concentration range expected  to minimize
reanalysis and provide extent of contamination information for the areas being investigated.
                                        182

-------
Methods

Immunoassay technologies are well established within the medical lab industry where they have
been used provide rapid, accurate test results for medical professionals.  In recent years this
technology has also started to emerge in the environmental analysis field.  The two parameters
analyzed by immunoassays for this project were PCB, using the Envirogard"" immunoassay by
Millipore and aromatic hydrocarbons using Ensys' Petrorisc11" immunoassay kits.  Soil samples
were analyzed for 22 metals using the Spectrace 9000 field-portable XRF.  The Spectrace 9000""
XRF uses a mercuric iodide (HgI2) detector along with a fundamental parameters algorithm to
qualitatively and quantitatively identify the metals.

    A lab facility was set up on base for sample log-in and analysis. All samples were labeled,
logged into a sample tracking system on and screened at this location. A portable computer was
used for sample tracking as well as to store both the results and the spectra produced by the
XRF.   All immunoassay data,  including  balance calibration,  extraction  weight  and the
absorbencies of both the samples and  standards were recorded in  bound lab notebooks.

 Immunoassays-General

    As both immunoassay kits  used methanol as their extraction solvent and immunoassay tests
are quite specific for their target compound(s), one extraction was performed on each sample
and the extract was refrigerated in a labeled screw-top vial.  This ensured any re-analysis would
be performed on the same  aliquot of each environmental sample.  The analysis reagents  were
added according to each manufacturer's instructions1'2'3'4 and all samples were run immediately
following calibration of the test kit. A spectrophotometer set to 450 nm was used to record the
absorbencies of both the standards and samples.

Aromatic Hydrocarbons (AH)

    For the Petrorisc"11 immunoassays,  sample absorbencies were  determined relative to a low
standard (0.7 ppm m-Xylene, which is equivalent to 100 ppm gasoline) which served as the
threshold of detection.  Two  aliquots of the methanol  extract were analyzed relative to this
standard:  the first represented the sample without  any  dilution and the second was the same
extract  at  a ten-fold dilution.   In  this  manner,  approximate concentrations  of  petroleum
constituents can  be determined with relative ease.  While  the  petroleum kits were calibrated
using m-Xylene, they were sensitive  to a variety of compounds  found in petroleum products
including bi- and  tri-cyclic aromatics.1'2  WCFS  utilized  the Petrorisc"" kits'  sensitivity to
aromatic hydrocarbons to indicate burn areas where these aromatic hydrocarbons remained as
products of incomplete combustion.  Because the Petrorisc"" kits  are sensitive to a variety of
compounds, the immunoassay results  correlated well with the hot spots  as defined by lab
analysis. Data on  the correlation between the two methods is shown in Table 1.
                                        183

-------
                Table 1:  Comparison of Lab  and Field Values for Petrorisc1
Sample Number
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Sample 9
Sample 10
Sample 1 1
Sample 12
Sample 13
Sample 14
Sample 15
Sample 16
Sample 17
Sample 18
Sample 19
Sample 20
Sample 21
Sample 22
Sample 23
Sample 24
Sample 25
Sample 26
Sample 27
Sample 28
Sample 29
Pelrorisc™ Value
ND
ND
ND
ND
ND
Detect
Detect
ND
Detect
Delect
Detect
ND
ND
ND
ND
ND
ND
Delect
ND
ND
ND
ND
ND
Detect
ND
ND
ND
ND
ND
Sum of PAH Values*
ND
ND
ND
ND
ND
ND**
> 1 ppm
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
> 1 ppm
ND
ND
ND
ND
ND
> 1 ppm
ND
ND
ND
ND
ND
*Sum PAH = Sum of all detects for compounds listed in SW-846 Method 8100.
**Dilulion at lab prevented proper quanlilalion.
                                               184

-------
   The protocol for performing analysis dictated that the difference between duplicate standards
(Delta Std) could not exceed 0.2 absorbance units (a.u.) or the calibration would be considered
invalid and the samples would be re-analyzed.  Although the precision data for the petroleum
kits was acceptable,  the use of a   repeat pipettor rather  than dropper bottles would have
improved the precision. (All figures showing precision data are scaled to equal size, so a visual
comparison may be made.)  Ensys will supply the reagents either in dropper bottles or in bulk
(for use with a pipettor), but for this project the dropper bottles were used.  A Shewart plot of
Delta Std for aromatic hydrocarbon analysis is  shown in Figure 2.
                  Figure 2: Precision Data for Petrorisc"" Immunoassay
                     1  2  3-4  3 G 78 9 10

                     _a_DeKo Std.         _^_Ux»r tornlngL1.1t

                      X upper control Llnrt    Q Lower control Linn
      3 24 23 26 27 ZB


Low«r MirnlnQ Limit
Fob/chlorinated Biphenyls (PCS)

    For the PCB kits3'4 (Envirogard*"1), the concentration of the two calibration standards (2 and
10 ppm Aroclor 1248) were used and a linear dose-response was assumed between those two
points in order to calculate an approximate Aroclor 1248 concentration.  As Aroclor 1260 was
the PCB  found at the site, Aroclor 1248 concentrations  were converted  to Aroclor  1260
concentrations  using relative response data for the two Aroclors provided by  Millipore3'4.  The
PCB kits were  used to delineate the volume of the burn pits  being investigated and the samples
taken for lab analysis in those areas was directed by the field  screening results. Correlation data
is presented in  Table 2.
                                         185

-------
Table 2: Comparison of Lab and Field Values for PCB
Sample Number
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Sample 9
Sample 10
Sample 1 1
Sample 12
Sample 13
Sample 14
Sample 15
Sample 16
Sample 17
Sample 18
Sample 19
Sample 20
Sample 21
Sample 22
Sample 23
Sample 2-4
Sample 25
Sample 26
Sample 27
Sample 28
Sample 29
Sample 30
Sample 3 1
Sample 32
Immunoassay Value
ND
ND
ND
ND
Delect
ND
ND
ND
ND
Delect
ND
ND
ND
ND
Delect
Delect
ND
Delecl
ND
ND
Detect
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
PCBs by Melhod 8080
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND*
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
                     186

-------
Sample Number
Sample 33
Sample 34
Sample 35
Sample 36
Sample 37
Sample 38
Immunoassay Value
ND
ND
ND
ND
Detect
ND
PCBs by Method 8080
ND
ND
ND
ND
ND
ND
   'Dilution at lab prevented proper quantitation

   Reagents were added using an Eppendorf repeat pipettor, which allowed for rapid analysis
with good accuracy  and precision.  Typically a total of twenty tubes were analyzed per run:
assuming analysis performed in duplicate, three standards and seven samples could be analyzed
in one run. A  Shewart plot (Delta Std) for both the low and high standards is provided in
Figures 3 and 4.
                      Figure 3:  Precision Data for PCB Low Std.

D.2





-0.2
'


-



^ "~^A / \/
-----^^-^t^*^**^


m Delta 9td. + Ibper Vtarolno Ll»u jfc-- Low»r Warning Limit
X upper control Limit _^_ LOWW control Liim










                                         187

-------
                      Figure 4:  Precision Data for PCB High Std.
                          control LFmU
                                          fcrning L)»lt

                                          Control U>1T
                                                          Vbrnlng Unit:
Field-Portable XRF

    All samples were first air dried and sieved through a 10-mesh sieve and sample descriptions
were recorded for all  samples.  If a significant amount of material would not pass through the
sieve, the material remaining in the sieve was retained and labeled as the exclusion products of
that environmental sample.  The dried samples  were then placed in 32mm  sample cups and
covered with Mylar film.5'6 Each sample cup was labeled with the field  sample number and
retained for re-analysis, if necessary.  At the end  of each day, both the results and spectra were
downloaded to a laptop computer for storage and data processing.

    A mid-range standard reference material (SRM) and a quartz blank  were run daily prior to
any samples,  after every 10 samples and at the end  of the analytical run  (10%  frequency). This
standard  was an  environmental  sample which had been  certified using traditional wet prep
techniques followed by GFAA and ICP analysis.  Both the SRM and blank were used  to confirm
instrument stability during  the project. The  standard was  less homogeneous than was initially
assumed at the start of the project: approximately  a week after  the XRF screening had begun,
particles where discovered in the sample cup which would not have passed through  a 10-mesh
sieve.  As there was a week of data on  the standard it was not re-prepped.  Particle-size effects
from  these large  particles  were believed  to  introduce some  variability in the standard  as
illustrated  in  the Shewart  plots for  the standard.   The  samples  were believed to be  more
homogeneous because all samples were dried and  sieved  prior  to analysis.   A plot of the Pb
results (Pb was one of the elements which had poor precision relative to most of the analytes)
for the standard over the course of the project is  shown in Figure 5 and a table of the accuracy
and precision data is shown in Table 3.
                                        188

-------
                       Figure 5:  Precision Data for XRF Std (Pb)
                      Conoentralon

                     per Control Limit
- 4>p*r wrnlng Liruft

.Lower Control Llrift
                                                          1H  19  ZO 21 22 23  24

                                                         . Lower Warning Lfmit
      Table 3:  Accuracy and Precision Data for XRF Standard Reference Material

Average (mg/kg)
True Value1
Percent Recovery
Std. Dev.
Relative Std. Dev.
Cr
184170
160287
114.9%
2263
1.2%
Ni
16014
13105
122.2%
423
2.6%
Cu
2787
2946
94.6%
144
5.2%
Pb
115
141
81.6%
19
16.9%
Cd
387
292
132.6%
65
16.9%
'All true values obtained by acid digestion followed by GFAA/1CP analysis

Conclusion


Method-Specific:  Immunoassay

The staff of two chemists performing analysis in the field lab was able  to screen approximately
20 samples per day for metals, PCBs and aromatic hydrocarbons.  The use of a repeat pipettor
is recommended both  to  speed immunoassay analysis  and to achieve better precision  and
accuracy. The correlation between lab and field data was good, but the difference in detection
limits and sample heterogeneity sometimes make it difficult to directly compare immunoassay
and lab data. However, the regions indicated as contaminated by field screening correlated very
well with the areas indicated as contaminated by lab analysis, historical data,  and PID/OVA
results of samples taken in the field.
                                         189

-------
Method-Specific:  XRF

The  use of an  independent  standard  which  was certified by  traditional  metals  techniques
(GFAA/ICP) gave the data produced  by the XRF  an additional level  of confidence.   The
instrument showed good stability  over the  course of the  project and  16.9%  was the worst
relative standard deviation for any  of the certified analytes (Cr, Ni, Cu, Pb, and Cd).  Had the
standard reference material  been completely  homogeneous, the standard  deviation would
probably have been considerably lower:  Shewart plots showed only  a few values which  were
close to the control limit (Average+/-[3*Std. Dev.]).  If these values were considered outliers
and removed, the standard deviation decreased markedly.XRF analysis of the standard reference
material correlated very well with its certified values:  the average percent recovery (defined as
[XRF value/True Value]*100) was 109.2% with a high of 132.6% for Cd and a low of 81.6%
for Pb.

OA/OC Issues

   Field screening can provide either Level  I or Level II data9:   for  this project the  field
screening data was regarded as Level I data and the  laboratory analysis  was used to make all
final decisions regarding site contamination. Specific guidelines for producing Level II data may
vary from  site to  site, and the sampling and analysis program must address the problems of
sample heterogeneity,  matrix effects,  interfering compounds, and sample contamination as  a
result of improper handling or preparation7.

   Although field screening can present additional challenges to the field  team, there are many
instances  where the additional data produced from  the lower-cost  field screening tests  can
significantly  reduce the  sampling error in  site investigations.    Analytical error  (bias  and
variability  introduced in the lab) typically accounts for only 15% of  the total error introduced
in the  site investigation process. The remaining 85%  of the error in  site investigations results
from insufficient samples or samples which do not accurately represent the contamination at the
site8.  Field screening allows for  rapid analysis following sample collection,  which  reduces
problems in sample handling,  preservation and  transport, and gives the field team the flexibility
to employ an iterative sampling strategy to fully characterize the contamination.

Effect  on Sampling & Lab Analysis

   Through the  integrated use of field  screening WCFS completed the site clearance on-time,
better delineated the extent of contamination and helped to direct the activities of the field crew.
In addition to the data provided by the lab, historical information such as aerial photos were used
to identify the area occupied by the former landfill (Figure 6).   While XRF  has  been  used
successfully in site investigations in the past, new advances  in detector technology will provide
field teams with  an instrument which is both  portable and sensitive. At the  time of this project
(April-May 1992),  none  of the immunoassay techniques had been recognized as methods by
EPA.     Subsequently  EPA  has  granted   SW-846  third   update  numbers  of  4010
(pentachlorophenol), 4020 (PCB), and 4030 (Total Petroleum Hydrocarbons) for immunoassay
screening techniques.
                                        190

-------
   In keeping with the DQO development process defined by EPA9,  the project should be
planned with field screening in mind from the outset, and a chemist familiar with the technology
to be employed should be involved during the planning stage. The overall effectiveness of field
screening will depend on project-specific needs.  It is recommended that the actual screening
analysis be carried out by,  or  under the supervision  of, a  qualified chemist to minimize
resampling and  reanalysis and  to ensure  that  results  are  not used inappropriately.    Both
immunoassay and XRF are mature screening technologies, which,  when used properly  can be
very cost-effective tools in the site investigation process.

Cost savings were realized in two ways on this project by utilizing field screening;  1) a reduced
number of samples were analyzed by the laboratory since the  field screening did not result in
very many false negatives, and 2) contamination information  was transferred to  the project
manager in a much faster  manner than for a normal project where one has to wait 2  4 weeks
for a lab result.  This meant that the project could proceed with decisions in a faster manner.
    Figure 6:  Site Contamination as Delineated by Field Screening and Lab Results
                                                                  woodwer a- CI y a
                                        191

-------
                                     References
'PETRO RIScm User's Guide, Ensys Inc,  1992.

2"Soil Screening for Petroleum Hydrocarbons by Immunoassay," Draft Method 4030, USEPA
SW-846 Third Update, July 1992.

3Envirogard Tests Kits User's Guide, Millipore Corporation, 1992.

""Soil Screening for Polychlorinated Biphenyls by Immunoassay," Draft Method 4020, USEPA
SW-846 Third Update, July 1992.

5Spectrace 9000 User's Guide. TN Technologies

6Donald E. Leyden,  Fundamentals of X-Ray Spectrometry as Applied to Energy Dispersive
Techniques: Tracer Xray, 1984.

7Kevin J.  Nesbitt, "Application and QA/QC  Guidance  USEPA SW-846  Immunoassay-Based
Field Methods 4010, 4020 & 4030;" Ensys Inc, 1992.

8Francis Pittard, Principles of Environmental Sampling:  A Short Course Presented Prior to the
8th Annual Waste Testing & Quality Assurance Symposium, July  11-12, 1992.

9USEPA, Data Quality Objectives for Remedial Response Activities. EPA/540/G-87/003, March
1987.
                                     192

-------
               DIVERGENCE OF FIELD AND LAB RESULTS IN
                     THE PONCA CITY INVESTIGATION

   Subijoy Dutta. Environmental Engineer, U.S. EPA, OSW/PSPD/CAPB, 5303W;
                    401 M St. SW, Washington, D.C. 20460.

ABSTRACT

This case study focuses on a Benzene-Toluene-Xylene (BTX) pollution of soils and water
in a  small hydrogeologically  secluded pocket of  Ponca City, Oklahoma.  Despite
complaint by  the residents,  State and Federal officials could not  find significant levels
of contamination at the site and recommended no action at this site. At the request of the
State treasurer, an independent hydrogeological investigation was conducted by the author
identifying high levels of BTX in the perched water table in an isolated pocket.

During an independent investigation, conducted by  the author,  very useful help  was
rendered by the community in data collection.  Samples from  the storm water runoff at
the site produced flame for a very short duration when lighted.  Other field tests,  such as,
oil emulsion/layering of the samples proved that the contamination was of significant
level. Laboratory results, however, did not show very high levels of contamination.  The
hydrogeological information from the site uncovered the mystery behind this inordinately
high levels of BTX in a  pocket of the perched water table.

The feasibility of a commercial hydrocarbon recovery from the spillage has been looked
into  as a cleanup option. However, CONOCO has rehabilitated  some of the residents
from the site and started a cleanup action at this site in  1990  with  full  community
cooperation.   The divergence of the  field and  laboratory results during  initial
investigation at this site  caused major problems and  dissentment  between the residents
and the State. The results from this study call for emphasizing the need to rely upon
field observations to locate and identify contamination and hazards associated with  it.

INTRODUCTION

Ponca City is located in the north central part of Oklahoma.   The CONOCO  refinery
borders southwest part of Ponca City. The site location is shown on Figure 1.  The area
of investigation for this assessment encompassed the boundaries of the CONOCO refinery
in Ponca City and  the Circle Drive area of south Ponca City adjacent to the eastern
boundary of the refinery. The study area is bounded by South Avenue on the north, by
Waverly  Street on the west, by a county road on the south, and  by State Highway 77
(14th Street) on the east. The horizontal extent of the study area is shown on Figure 2.
Refinery area designations and the watershed boundary of the study area are also shown
in this Figure.  The Circle Drive portion of the study area  extends from the eastern
Refinery boundary formed by South Third Street, South Fourth Street, and Seneca Drive
                                       193

-------
Figure 1  The Study Area in Ponca City, Oklahoma

-------
                .
~^      •  '      'I >*^

i:  •     •  • .   -\Jr«
 Figure 2 The Watershed Boundary and Sampling Locations in the Study Area
                                   195

-------
eastward to the  14th Street.

The vertical extent  of the  study area extends to the base of the uppermost aquifer
underlying the study area.  This alluvial aquifer ranges in depth from approximately 30
feet to 60 feet  and  is highly variable across the area thinning eastward towards the
Arkansas River.

Ponca city refinery was developed in early 1900's. After sixty years there were reports
of contamination of ground water. Hydrocarbon  in ground water was observed in the
1950's and 1960's in the Circle Drive residential area of Ponca City adjacent to the east
boundary of the Refinery.  In 1968 the Oklahoma Corporation Commission  received
reports from two houses who had water with high hydrocarbon odor in their basements.
Significant hydrocarbon accumulations have  not  occurred since the  late-1960's when
CONOCO, Cities  Service, and Ponca City jointly installed and operated a hydrocarbon
recovery well network across the East and South Refinery areas and  the Circle Drive
area of Ponca City.   CONOCO has taken some  measures to prevent and contain any
potential releases which might migrate off-site.

Ground  waters  containing dilute  concentrations  of  volatile  organic  petroleum
hydrocarbons were detected seeping  into Hoover  Ditch from  adjacent stream banks in
late-1986 and early-1987. This seepage was a direct result of heavy rainfalls in excess
of 20 inches which occurred in October and November 1986.  The heavy rainfalls and
accompanying recharge caused ground water elevations to rise to surface discharge points
in the Circle Drive area of Ponca  City.  Due  to serious public outcry,  Mr.  Ellis
Edwards, state  treasurer  at that time,  asked the author to  conduct  an  independent
hydrogeological investigation of the ponca city pollution problem.

The objectives of the independent hydrogeological investigation, conducted in  August,
1988, were to:

       (1)     Define the sources and extent of hydrocarbon contamination in the area;
              and

       (2)     Develop  recommendations  for remedial  actions for  controlling and
              remediating the hydrocarbon contamination problem,  taking  all viable
              alternatives into consideration.

The scope of the investigation included:

       Hi     Study  of the geological setting of the study area.

       HI     Physical   and  chemical   testing   of   water  and   water   bearing
              formations/soils.
                                        196

-------
       H     Recommendation of possible steps to reduce severity of the problem.

       H     Recommendation of recovery  of hydrocarbons from  the contaminated
              water.

Nature and Extent of the Problem:

       The initial assessment of the ground water and hydrocarbon conditions underlying
the CONOCO's Ponca City Refinery and Circle Drive area indicates that there are two
general types of hydrocarbon accumulations in the subsurface [1]:

       (1)  Free product - petroleum hydrocarbon mounds; and

       (2) Low concentration, soil-absorbed hydrocarbon residuals from historical leaks
and spills.

       The  initial study [1] determined that there were possibly four potential free
product    hydrocarbon  mounds which may  exist  under  the  CONOCO  Refinery.
Hydrocarbons in these areas are refined products, such as gasoline, kerosine, or heating
oils, and can be chemically identified as such.  Hydrocarbons observed in ground water
seepage  in  the  Circle Drive area  are dissolved  residuals and cannot be chemically
characterized  as a  unique petroleum  product.   This  hydrocarbon  appears to be
significantly older, as is indicated by the low concentrations of benzene, toluene, and
xylene.

       A major limitation of the study by Downs [1] was the la~k of adequate ground
water observation points, as well as concerns for the construction and completion history
of old hydrocarbon monitoring wells from the late-1950's and  1960's.  The 1987 field
program for filling the information gaps identified in the initial study  was implemented
in  the summer and  fall of 1987.   During this  program,  fifty-six (56)  new cone
penetrometer soil borings were constructed to develop detailed information on geological
maps for the  site.   In addition,  twenty-nine (29) new ground water/hydrocarbon
monitoring wells were installed to develop new data on geology, ground water elevations,
and potential hydrocarbon accumulations in the study area.

       The hydrogeologic study [2] was focused on the quality and quantity of water in
the surface and groundwater systems in the study area.  The estimates were based upon
the available information.  A description of the baseline hydrologic  conditions and a
prediction of the effects on these baseline conditions in the study and adjacent area due
to the pollution problem were amalgamated to form an integral part of the study.

       The conclusion from the study [2] are based upon the available database and field
measurements.   Inadequate information on groundwater quality  and quantity in the
                                        197

-------
adjacent area is a limiting factor in arriving at unequivocal conclusions concerning the
groundwater of the adjacent area of the study.

       The divergence of field observations and the laboratory results for the study area
caused major problem in the hazard assessment for the site.

SUMMARY

       The principal purpose of the study undertaken by the author was to assess the
hydrogeological impact of an underground spillage problem on a tract of land located in
Section 34-26N-2E of Indian Meridian in Kay  County,  Oklahoma.  This  pollution
problem has surfaced near the Circle Drive area of Ponca City, Oklahoma.

       The scope of the study also included investigation of the existing climatic and
hydrogeological conditions prevailing in and around the study area.

       Climate:

       A humid climate prevails in the study area and its vicinity.  Annual precipitation
on the general area based  on a 10-year average is found to be 32.34 inches.  Seasonal
precipitation is found to vary from 17%  to  31% of 10-year average annual rainfall.
Spring is the wettest season  when 31% of the precipitation occurs  from April through
June.  During the Fall season from October through December,  minimum precipitation
is observed to occur.

       The total pan evaporation based on a 10-year average data  is determined to be
56.2 inches.   Some of  the  local  meteorological  parameters  having  effect on surface
runoff,  air quality, particulate transport,  and  intensity of photochemical smog were
studied for a 10-year period.

       Wind velocity varies  considerably and can cause adverse effect due to increase
local concentration of dust and other particulates. Prevailing wind direction is from south
to north. The 10-year annual average wind velocity for the study area is 10.4 miles per
hour per month with a standard deviation of 1.12 and a  variance of 1.25 during the 12
month period.  There are periods of prolonged wind  movements up to 15-20 miles per
hour.

       The 10-year mean  of annual temperature is  found to be 60  degrees  F.  The
minimum mean  monthly temperature  occurred in January,  and the maximum  mean
monthly  temperature was  reached in  the month of July.  Drastic  variations  within a
certain month are not uncommon.

       The total pan evaporation is highest during  June, July, and August, ranging from
                                       198

-------
8.23 inches to 9.35 inches per day. Mean annual evaporation is 56.3 inches.

       Geology:

       Geology of the Study Area [2]:

       The pivotal area of study is section 34-T26N-R2E within the south and eastern
city limits of Ponca City, Oklahoma. The area is between the Conoco Oil Refinery Plant
on the west and the Arkansas River on the east. The general study area consist of nine
sections being sections 2,3 and 4 T25N, R2E and sections  26, 27, 28, 33, 34, and 35
T26N, R2E  in Kay  County, Oklahoma.  The area is  covered  on the  Ponca City
Quadrangle Oklahoma, 7.5 minute series, U.S. 6.5 Topographic map.

       Topography:

       The surface elevation varies from the highest point of 1020 feet in middle of
section 27 to around 900 feet in the stream bed of the eastward flowing Arkansas River
found  in SW/4 of section 2, T25N, R2E. This is a north to south differences of 120
feet. Elevation differences from west to east of approximately 110 feet.  In the area the
surface slopes south and east and all surface water runoff is drained into the Arkansas
River.

       The study area sits atop of a portion of a northeast-southwest trending anticline.
This anticline is expressed in the surface topography where the surface rocks dip slightly
either eastward or to the west depending on the location of outcroppings. The anticlinal
axis bisects section 34 slightly east of its northeast corner to the east of the southwest
corner of section 34,  T26N,R2E.   The  outcropping of strata from a regional aspect is
characterized by broad expanses of strata of benches or alternating  parallel escapements
with abrupt slopes facing eastward.  Older strata outcrops toward the east. The regional
dip is to the west less than 50 feet per mile.

       Stratigraphy and Lithology:

       The surface rocks are sedimentary. With the exception of the terrace sands, the
surface and the upper subsurface formations are of lower permian age. They embrace a
approximately 300 feet of sediments and consists of four limestone formations alternating
with same number of shale bud.  Locally thin  sandstones are present in  the  shale
formations. The Permian and subsurface Pennsylvanian beds have been  tilted gently
westward, and subsequent erosion has  beveled the section, exposing progressively older
beds to the east. Resting unconformably on the truncated edges of the Permian sediments
are local deposits of unconsolidated Quaternary alluvium and terrace materials that have
been deposited by the Arkansas River and its tributaries as it slowly worked its way
westward down the dip slopes of the limestone formations.  The key beds of the study
                                        199

-------
area are listed in a descending order.

QUATERNARY- Alluvium and terrace deposits.
PERMIAN- System
LEONARD SERIES:  Summer Group, Wellington shale
WOLFCAMP SERIES: Chase group of formations

       Alluvium and Terrace Deposits:

       Along both sides of the Arkansas River a considerable thickness of unconsolidated
sands,  slit, clay, and gravel occur overlaying the several Permian formations. Many of
the alluvial and terrace deposits contain sand and gravel layers that are highly porous and
permeable, and these deposits typically contain groundwater resources.  Recharge areas
for groundwater resources in alluvium and terrace deposits themselves become almost all
ground water contained in these deposits as a result of downward percolation of water
from the land surface.

       Remedial Investigation:

       As noted earlier, ground  water seepage and the minor concentrations of organic
and inorganic constituents that have been adsorbed and transported by rising ground
water levels have discharged into various areas  of Circle Drive.  This  includes natural
ground water discharge into Hoover Ditch and surrounding ditches and natural ground
water discharge into basements and foundation areas which intersect the rising ground
water table. The manifestation of natural ground water seepage in the Circle Drive area
has caused citizens  to become concerned about both the occurrence of ground  water
seepage and what  this  ground  water  might  be transporting.   The hydrogeologic
investigation conducted by the author  involved soil and water sampling from the study
area. Sampling locations are shown on Figure 2.  Water  samples were collected from
locations A and B and soil borings were conducted at location #1, #2, and #3.

       Testing of Soil &  Water:

       Chemical analyses were carried out on soil and water samples from the study
area.

       A columnar  section  of the alluvial formation from the study area is shown in
Figure 3.
                                       200

-------
ZONE #    DEPTH LITHOLOGIC DESCRIPTION
 A.   6-8'
 B.   9-45"
 C.   46-80"
Topsoil
Grey, Sandy, Silty
Loam (Highly Porous)
                    Reddish Brown Clay
                    (Very tight)
  Figure 3 Columnar Section of the Alluvial Formation from the Study Area
      Experiments were run to calculate porosity and permeability.
A highly porous zone (#B) was present on top of a tight clay Zone (#C). Table 1 Shows
the experimental results.
                                     201

-------
 TABLE 1 EXPERIMENTAL RESULTS OF SOIL SAMPLES
          Porosity                   Dispersion
                                    Coefficient
 ZONE #     (%)    Permeability   (sq. cm/sec.)
                        (mD)                        Sample#/Remarks
  B      45           2000            0.037        # (1-2)7(2-2)


  C      52           0.022            0.00         # (3-2B)
       The grey alluvium soil containing 19% of moisture showed a porosity of 45% and
a permeability of 2000 millidarcy. This zone was followed at the bottom by a tight clay
having a permeability of 0.022 millidarcy. The Zone C thus  served as a confining layer
for the fluids to move through Zone B of in the study area.

       The dispersion coefficient of the porous zone, as estimated roughly [3] , was
found to be 0.037 sq.  cm/sec, and the dispersion coefficient of the underlying clay zone
was found to be  0.0  sq. cm./sec.

       The result of the hydrogeologic investigation of the study area clearly identified
a zone of high permeability sand channel at about 3-10 ft. from the ground which carried
all the recharged  water from the watershed which was contaminated with hydrocarbon.
The very tight clay  underneath the sand channel served as a  confining layer.

       Field Observations:

       During an initial field visit in July 1988 the basements of two houses in the study
area were visited. Very strong hydrocarbon odor was present in both the basements and
oily liquids were visible. An oily surface discharge, oozing out of the ground,  near the
Mclnley  School was also noticed. Hydrocarbon odor  was also present at this location.
The worst  situation was observed where a  24" pipe  was discharging high volume of
water, with very  strong odor of hydrocarbon,  from  underneath a road  into  the storm
drainage  system.  The source of this water is from a subsurface system from the south
side of the site. The discharge  at this point could be visibly  identified to carry
hydrocarbon vapors.
                                       202

-------
       A 1/2 gallon jar was filled up with this water and was taken to a dark room with
the lid tightly closed. An open flame was introduced at the top of the jar upon immediate
opening of the lid.  Spontaneous blue flame was observed for a very short time at the
mouth of the jar. This field test confirmed the presence of volatile hydrocarbon in the
perched water table of the study area.

       Lab Results:

       Samples from the same discharge point (A) was analyzed in the lab which showed
a flash point of 220° F and 100 ppm Benzene, 200 ppm Xylene, and 300 ppm of heavy
hydrocarbons.

       Samples  from the  basement of the house  (B) showed  1.49 ppm toluene, and
< .0002 ppm both benzene,  and xylene. Faint  smell of hydrocarbon was noted by the
lab. These lab results were very similar to the earlier laboratory results as furnished in
the following paragraphs.

       An earlier report on Ponca City water analysis which was  done on January 20,
1988  showed that the basement water had total hydrocarbon levels which ranged from
121 to 4,725 part per billion. The concurrent benzene levels in basement water ranged
from  32 to 2,000 parts per billion.  This water was not used for drinking or bathing.
The results showed that there was volatilization potential for the various hydrocarbons
contained in the basement water. The air samples taken at the time of water sampling
were  below 0.5 ppm for benzene using the dragger tube calorimetric technique.

       The analysis by  the Oklahoma  State  Department  of  Health Environmental
Laboratory of basement water samples taken  from 6 homes identified by the citizens as
the "worst possible cases"  in the Circle Drive  Area of Ponca City indicated  total
hydrocarbon levels from 121 to 4,725 parts per billion.  Of those total hydrocarbons, the
concurrent benzene levels in  basement waters ranged from  32 to 2,000 parts per billion.

       These samples in and of themselves only address the potential for human exposure
through direct contact with the basement water. This water was not used for drinking
or bathing. These results do show that there  was volatilization potential for the various
hydrocarbons contained in the basement water.  Entry surveillance air samples taken at
the time of water sampling were below 0.5  ppm  for benzene  using the Dragger  tube
calorimetric technique.

       The analysis by  the Oklahoma State  Department  of Health Environmental
Laboratory of drinking water samples taken at 8 sites in the Circle Drive neighborhood
and adjacent areas  indicated 1  sample with a level of  4  parts  per billion of  total
hydrocarbon.  The remainder of the  drinking water samples were  below the  total
hydrocarbon detection limit of less than 2  ppb.  Additionally, none of the  specific
                                       203

-------
hydrocarbons including  benzene,  xylene,  or toluene were  found in drinking  water
samples at the 1 ppb detection limit for specific hydrocarbons.

       The drastic difference between the field observation and the laboratory results in
this study could not be accounted towards any noticeable sampling or other procedural
errors.  High volatilization of the samples might have caused the discrepancy.

       Conclusions and Recommendations:

       The dispersion characteristic of  the strata from the subject  study area were
investigated and the results show that the estimated dispersion coefficient of the highly
permeable zone is very high compared to the dispersion coefficient of the underlying
tight clay layer.

       In this study the calculations of porosity and permeability of different layers show
that a  highly permeable  zone lies  between two clay layers  in the affected area. This
channeling effect of  the highly permeable  sand  is partially responsible  and may be
considered as a natural calamity causing the  seepage of the contaminated water into the
basements of local residents.  The recovery of free hydrocarbons from this porous zone
could be done by using a pumping system.  This will largely reduce the  hydrocarbon
percentage of the contaminated groundwater.

       The chemical test report indicates that the concentration of Benzene, Toluene, and
Xylene (BTX) is less  then the standards set by Suggested No Adverse Response Levels
(SNARL). However, when this discharge seeps through the basements of local residents,
it creates uncomfortable and unhealthy living condition. The field observation from this
study strongly supports the need for a remedial action at this  site.  On the contrary, the
laboratory results do not call for such action. This divergence in the field and lab result
is highly controversial, and more so in a residential area.  The author recommends that
an emphasis should be put on the field observations in such cases so that some remedial
measures could be initiated at these sites.  Also, it is recommended that some refinement
in sampling and analytical procedures might be able to converge the differences between
the field observations  and laboratory results to an  acceptable  degree.
                                        204

-------
REFERENCES

1.  Downs, C. E., "Ground Water Hydrocarbon Assessment Ponca City Refinery And
      Vicinity Conoco Inc. ", Report on Hydrogeologic Study conducted in Feb. 1988.

2.  Dutta,  S., "Hydrogeological Investigation of the Pollution Problem in Section 34
      -T26N-R2E. in Kay County. Oklahoma." , Phase I & Phase II Report, July 5,
      and  August 16, 1988.

3.  Menzie, D. E., et. al., " Dispersivity as an Oil Reservoir Rock Characteristics ",
      Annual Report for Period Oct. 1, 1986 to Sept. 30, 1987, submitted to U.S. DOE
      for publication, Dec. 1987.
                                      205

-------
                     FIELD SCREENING FOR HAZARDOUS MATERIALS
               IN SOIL AND GROUNDWATER AT A MAINTENANCE BUILDING
Anita M. Larson.  Senior Hydrogeologist, Harding Lawson Associates,  2400 ARCO Tower, 707 17th
Street, Denver, Colorado 80202; Patricia  M. Curl, Senior Hydrogeologist, Harding Lawson Associates,
3247 Ramos Circle, Sacramento, California 95827;  Harry R. Kleiser, Project Manager,  Installation
Restoration  Division, U. S. Army  Environmental  Center,  ENAEC-IR-B,  Building E4480, Aberdeen
Proving Ground, Maryland 21010.
ABSTRACT

A combination of field screening techniques were  used to assess  the nature and horizontal extent of
hazardous materials  in soil and groundwater at a maintenance building area at the Sierra Army Depot,
California. Soil-gas surveys and groundwater sampling directly from a hollow-stem auger were employed
to study a maintenance building area for which little information regarding past waste management handling
and disposal practices was available.  Although one maintenance building was identified as the potential
source area for the generation  of hazardous materials, eight  other maintenance buildings and eleven
warehouses in the area were discovered to be additional potential source areas  of hazardous materials.

A series of field screening tools were used to design a cost-effective remedial investigation program and
a groundwater monitoring network at this site.  Since the use  of solvents was documented, an extensive
soil-gas survey program was designed to identify the potential horizontal extent of solvents in the soil and
groundwater.  A grid  of 300 soil-gas sampling locations over 40 acres was established.  The soil-boring
and groundwater sampling program was focused in areas where soil-gas contaminant concentrations were
highest.  Since  the groundwater gradient at the site was unknown, a series of piezometers were installed
to serve as both groundwater level monitoring locations and as one-time groundwater sampling points.
Additional one-time groundwater sampling points were determined on the basis of the groundwater gradient
and on a preliminary estimate of the horizontal extent of any detected contamination, thereby allowing for
monitoring wells to be installed down-gradient of detected contamination.

INTRODUCTION

Harding Lawson Associates (HLA) was contracted by the U. S. Army Environmental Center (USAEC),
formerly the U. S. Army Toxic and Hazardous Materials Agency, to perform a Remedial Investigation and
Feasibility Study (RI/FS) for nine of the twenty-two sites identified in the Sierra Army Depot (SIAD)
Federal Facilities Site Remediation Agreement (FFSRA) between the  State of  California-Environmental
Protection Agency (Cal-EPA) and SIAD. As a result of the Preliminary Assessment performed by the U.S.
Army Hygiene Agency in  1987, the  scope of work for the remedial investigation of these nine sites was
referenced  in the FFSRA.  On the basis of initial site reconnaissance, the Field Sampling Design Plan
(FSDP) and other associated work plans for the RI/FS were written with a two-stage field investigation to
meet the scope  of work requirements of the FFSRA and the overall objectives of the RI/FS in a cost-
effective manner.

To facilitate this two-staged investigation, routine communication with Cal-EPA was required for FFSRA
deadlines to be met and cost-savings to be  realized.  Upon completion of the first stage of  the field
investigation, sites were  eliminated  from further study while the  scope of work for other sites  was
expanded.  Data from  the first stage  of the RI allowed for data gaps to be identified and the scope of the
                                              206

-------
second stage of the RI to be determined.  Through documented negotiations with Cal-EPA, the second
stage of the RI was performed prior to the issue of a RI report.  This allowed for the writing of an RI
report that drew conclusions more meaningful than "that more data was needed." The staging of the field
investigation and continuous communication with the Cal-EPA allowed for a flexible  field screening
program to be implemented resulting in significant cost savings for the field investigation at the Building
210 Area  which is the topic of this paper.

SUMMARY

Stage one of the field program at the Building 210 Area included an extensive soil-gas survey and limited
soil-boring sampling.  Soil samples were obtained from areas where there  were high detections of
trichloroethene (TCE), freon-11, and trichloroethane (TCA) in the soil-gas.  The results  of soil samples
from these  areas revealed either very  low or no detections of these compounds indicating that the
groundwater was the likely source of the detections in the soil-gas although the depth to groundwater is in
excess  of 85 feet below the ground surface.

The  stage two investigation designed for the Building 210 Area was a combination of soil-gas surveying1
and  Hydropunch®2 sampling in conjunction  with  the  installation  of piezometers  and  groundwater
monitoring wells.  The objective of the stage two soil-gas survey was to reach an area west of the site
where the soil-gas results were non-detections.  It was important for an area of non-detections in the soil-
gas data to be identified to verify that the source of TCE in the soil-gas was not from another site west of
the Building 210 Area since groundwater  flow direction was  unknown and groundwater contamination
approximately one-half mile from the Building 210 Area was documented.

The  objective of the Hydropunch®  sampling  was to identify areas  in the groundwater of both  high
contaminant levels and no contamination for the installation of groundwater monitoring wells. In addition,
piezometers were installed prior to the Hydropunch® sampling to determine the local groundwater gradient.
Piezometer locations were selected for their usefulness in determining the groundwater gradient and for
verifying whether areas of suspected contamination were contaminated. Stage one soil-gas and soil boring
data and stage  two groundwater flow data from the piezometers were employed to determine optimal
Hydropunch® locations.

Due to weather and geologic conditions, the field program was implemented differently than planned.  Due
to record snowfalls, the soil-gas survey and drilling efforts became out-of-phase and a local laboratory was
contracted to do 24-hour turnaround organic screening via Environmental Protection Agency analytical
method 624 (EPA Method 624) for the Hydropunch® groundwater samples.  Another modification to the
program was required when it was discovered at the first boring that the sands at the site  were too dense
for the Hydropunch® tool to penetrate.  At that time it was determined that groundwater  samples would
be collected directly from the hollow-stem auger.

The hollow-stem auger was advanced five feet below the water table and a stainless steel bailer was used
to obtain  groundwater samples.  There was concern that the  aquifer would be disturbed and that the
groundwater samples would not be representative. Table 1 presents a comparison of TCE  in groundwater
samples collected in the hollow-stem auger to samples obtained in colocated wells.  The analytical results
in Table 1 are reported in micrograms per liter (ug/1); analytical methods used were EPA Method 624 for
the screening (bailed) samples and USATHAMA Method UM 20 for the well samples.  The detection limit
for Method UM 20  is 0.50 ug/1.
                                               207

-------
Table 1 Comparison of TCE Detections in Groundwater
Sample Identification
B21-3-MW
B21-4-MW
B21-5-PZ
B21-6-MW
Screening Results (ug/1)
3400
<1
210
<1
Well Results (ug/1)
3496
0.75
180
<0.50
Figure  1  presents the locations where groundwater screening samples (noted by  "HP" in the  sample
identification) were  taken and where piezometers and monitoring wells were  installed.  Figure 1 also
demonstrates how the screening  data was used to outline potential  contaminant plumes by contouring
screening data concentrations.  Table 1 and Figure 1 demonstrate one of the successes of this screening
program,  the close correlation between the groundwater screening data and the monitoring well data.  One
of the problems encountered implementing this program was the discovery of an extremely flat local and
regional groundwater gradient and the project team's delayed ability to determine the local gradient due
to fluctuations in the water table that were not stabilized until the wells and piezometers were developed.
Once the groundwater gradient was determined, it was confirmed that contaminants had not only migrated
to the northeast, as early field data supported, but that contaminants had also migrated to the southeast as
the last field screening data set had indicated.  Continued water  level monitoring data showed that a
groundwater mound, which was not anticipated, was present at this site.

In conclusion, the field screening techniques employed at the Building 210  Area met or exceeded their
intended objectives although unanticipated groundwater conditions did not allow for  all  necessary  field
investigation activities required at this site to be performed during stages one and two.
1.       Soil-gas surveying is the collection and analysis of air from the vadose zone for the analysis of
        target analyses.

2.       Hydropunch® sampling is a technique where a stainless steel probe is either driven into the soil
        or into an existing soil boring to collect groundwater.
                                                208

-------
                                                    Explanation
                                             Isoconcentration contours In microgn
                                             Building, permanent
                                             Building, semipermanent
                                             Pavement, roads, and parking
                                             Earth road
                                             Trail
                                             Railroad
EnvMWNTwnUI Serwai
 	  Fence
— •• — •  Reservation boundary
   A     Stage 2 sampling location
   A     Proposed sampling location
  ug/l     Micrograms per liter
                                                                    Prepared for:
                                                                    U.S. Army Environmental Center
                                                                    Aberdeen Proving Ground, Maryland

                                                                    Sierra Army Depot, Lassen County, California
Figure 1
Building 210 Area -
Trichloroethene Concentrations Detected in
Groundwater and Proposed Sampling
Locations
                                                                         209

-------
31          FIELD AND LABORATORY METHODS IN ECOLOGICAL RISK
            ASSESSMENTS FOR WETLAND AND TERRESTRIAL HABITATS

      Greg Linder. Mike Bollman, Chris Gillett, Ricky King, Julius Nwosu, Suean Ott,
      David Wilborn, ManTech Environmental Technology; Gray Henderson, University of
      Missouri; Environmental Research Laboratory, 200 S.W.  35th  Street,  Corvallis OR,
      97333;  Julie DalSoglio,  US  EPA/Region 8,  Helena,  MT,  59626,  and  Thomas
      Pfleeger, US EPA, Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis
      OR, 97333.

      ABSTRACT

      Field surveys are frequently included as part of an ecological assessment for hazardous
      waste  sites.  In  addition,  biological evaluations in both field and  laboratory have
      increasingly been considered critical components in the ecological risk assessment process
      for Superfund.  Depending upon habitat type, field methods have  been developed which
      lend themselves directly to the Superfund ecological risk assessment process. Freshwater
      wetlands, for example,  are frequently impacted by various anthropogenic chemicals, and
      field methods have been developed to help focus chemical and biological  tests that have
      been identified for laboratory studies. Here, we compare and contrast the field methods
      used  in  evaluating wetlands  that occurred in markedly  different  habitats and were
      impacted by two different, but relatively simple contaminant sources.  Both field studies
      were completed as part of  the baseline ecological risk assessment process.   One site,
      located in western Montana on the Clark Fork River, was Milltown Reservoir wetlands,
      and had  been impacted by heavy metals and sedimentation as a result of nearly ninety
      years of upstream mining activities.  The other  site occurred in the shrub-steppe of
      eastern Oregon,  which  had  been a storage and disposal area for over 25,000 barrels of
      chlorophenoxy herbicides and chlorophenols.  In both case studies, assessment methods
      had to be applied in lacustrine wetlands as well as the surrounding upland habitats.  As
      part of the ecological assessment for each site, a variety of field methods (e.g., terrestrial
      and aquatic tests] were critical to the evaluation, yet despite differences in contaminant
      sources,  each field site considered ecological effects potentially associated with:

             •      Preliminary food-web contamination;

             •      Phytotoxicity and contaminant uptake by plants;

             •      Adverse  biological effects expressed by soil macroinvertebrates;

             •      Adverse  biological effects expressed by sediment macroinvertebrates; and

                   Adverse  biological effects expressed by terrestrial and wetland vertebrates.
                                             210

-------
In conjunction with chemical analyses of soil, sediment, and water  samples,  these
biological and ecological evaluations yielded an integrated evaluation of ecological effects
and exposure for the baseline ecological risk assessments.  Overall, the field activities at
each site illustrate  the role that preliminary field  screening plays in the ecological
assessment process.

INTRODUCTION

Ecological  risk assessment has recently gained a more prominent role in hazardous waste
site evaluations (US EPA 1991;  Suter 1993),  and this increased role has been reflected
in technical guidance to assure that reliable, yet cost-effective methods  are used in the
process (Warren-Hicks, et al 1989; US EPA 1991; Under, etal. 1992). Field methods
for chemical analysis, particularly for screening purposes, have been developed and are
relatively routine in their application  in waste site assessment (Fribush and Fisk 1991).
But, biological methods are not nearly as widely used, despite biological tests and survey
methods being readily available,  well developed and standardized (Warren-Hicks, et al.
1989; ASTM 1992).  As previously  summarized, both chemically-based and toxicity-
based approaches may  make  significant contributions to ecological assessments for
hazardous  waste  sites (Parkhurst,  et al. 1989).  When  used together, chemical and
biological test and survey methods can directly evaluate ecological endpoints pertinent to
risk assessment.  From an ecotoxicological perspective, ecological  effects and exposure
assessments are complex  interrelated functions  which yield estimates of hazards, and
potentially  risks, associated with environmental contaminants in various matrices sampled
at a waste site.  Here, using two wetland habitats as examples, we illustrate an integrated
approach to ecological risk assessment, and in particular we focus on field and laboratory
activities that occurred early in the ecological risk assessment process.

Montana Field Site

Milltown Reservoir is on the Clark Fork River in western Montana, six miles east of
Missoula, Montana.  The reservoir was formed  in 1907 following  the construction of a
hydroelectric facility located on the Clark Fork River immediately  downstream from its
confluence with the Blackfoot River. Since construction of the dam, a  wetland habitat
has been created.  Because of the upstream copper mining activities in  the Clark Fork
River watershed, Milltown Reservoir  has accumulated  a large  volume  of sediment
enriched with heavy metals and metalloids, including arsenic, cadmium, copper, lead and
zinc.    The  Milltown Reservoir  wetland was initially  identified under CERCLA
(Comprehensive Environmental Response, Compensation, and Liability Act) in 1981 after
community well-water samples were found to have arsenic levels that ranged from 0.22
to 0.51 mg/L; the EPA recommendation for potable water supplies suggested that arsenic
not exceed 0.05  mg/L.   Within an ecological  context, however,  the impact of the
contaminated wetland soils and sediments on the indigenous wildlife  and vegetation
characteristic of the site was unclear.
                                        211

-------
Oregon Field Site

Alkali Lake, Oregon is located in south central Oregon (elevation ca 1300m).  When
seasonally recharged or during wet years, Alkali Lake  is characteristic of many Great
Basin lakes, being relatively shallow and highly alkaline.  During dry years, the lake
recedes and the exposed  lake bed  presents extensive evaporative  deposits of sodium
carbonate.  Between 1969 and 1971 over 25,000 drums containing herbicide formulations
(primarily,   2,4-dichlorophenoxyacetic   acid   [2,4-D],    MCPA  [4-methyl-2-
chlorophenoxyacetic  acid], and related chlorophenoxy herbicides) were hauled from a
Portland, Oregon production facility to a storage facility located immediately west of
Alkali Lake.  In November, 1976 the storage facility was converted to a disposal site, and
the drums were placed into shallow  (1 m), unlined trenches, crushed,  and buried on the
playa near Alkali Lake. Unfortunately, in addition to releasing the materials to the  soil,
groundwater contamination  resulted from these disposal activities, since  a relatively
shallow aquifer  (1 to 3 m below ground level) existed in  the closed-basin around the
Alkali Lake playa.  Prior  to these preliminary studies, the  ecological  effects associated
with this contaminated soil and aquifer were poorly understood.

METHODS

As part of the  ecological  assessment at each  site, a  variety  of terrestrial  and aquatic
laboratory tests or field survey methods were critical to each wetland  evaluation (Table
1). While this overview must limit detailed descriptions of laboratory and field methods,
those methods applied at each site reflected site-specific contingencies whenever possible.
For example, amphibian work at the Montana site used standard methods (ASTM 1992),
but also tested surrogate species (bull frog, Rana catesbeiana) that were more relevant to
ecological interpretation (Linder, et al. 1991).  In contrast, if amphibian  work were
identified as critical  to the high  desert site,  an alternative surrogate (spadefoot toad,
Scaphiopus  intermontanus) would be appropriate.  Despite differences in contaminant
sources, each assessment considered ecological effects potentially associated with:

       •     phytotoxicity and contaminant uptake by plants;

       •     adverse biological effects expressed by  soil macroinvertebrates;

       •     adverse biological effects expressed by  sediment macroinvertebrates;

       •     adverse biological effects expressed by terrestrial and wetland vertebrates;
             and

       •     preliminary food-chain contamination.
                                        212

-------
            Table 1.   Summary  of field  survey  and  test methods  used  at  study  sites  (from Linder,  et
            al.  1994).
              Ecological assessment component
              Phytotoxicity and contaminant
              uptake by terrestrial and
              wetland plants, and field
              surveys for in situ effects and
              vegetation communities
to
_L
CO
Candidate field survey or test
methods
Standard laboratory phytotoxicity
tests for seed germination, root
elongation,  and plant vigor
Field tests for seed germination and
plant vigor
                                               Field surveys for vegetation
                                               communities
                                               Field and laboratory measurements of
                                               soil physicochemical properties
Methods potentially applicable to
Oregon field site, Alkali Lake
disposal area

Seed germination, root elongation,
and plant vigor tests with lettuce
Alternative site-specific tests
species,  including alkaline
tolerant  plants and native species
such as Sacrobatus vermiculatus
(greasewood)

Field testing limited in
preliminary studies, and seed
collections gathered for laboratory
testing

Field surveys for evaluating in
situ effects (e.g., plant health)
and vegetation mapping for baseline
evaluations

Soil texture, organic matter, pH,
cation exchange capacity, base
saturation, extractable acidity,
extractable nutrients, H-P,
electrical conductivity, water
retention, bulk density,
infiltration capacity, hydraulic
conductivity
Methods potentially applicable to
Montana field site, Milltown
Reservoir wetlands

Lettuce seed germination and root
elongation; plant  vigor tests
with lettuce and Hydrilla
verticUlata

Plant vigor tests  using
alternative species, including
native species (Potamogeton
pectinatus. or sago pondweed)
In-field seed  germination tests
with standard  laboratory test
species and plant vigor tests
using sago pondweed

Wetlands delineation and survey
of facultative and obligate
wetland plants
                                                                          Soil texture, organic matter, pH,
                                                                          cation exchange capacity, base
                                                                          saturation, extractable acidity,
                                                                          extractable nutrients, N-P,
                                                                          electrical conductivity, water
                                                                          retention, bulk density,
                                                                          infiltration capacity, hydraulic
                                                                          conductivity

-------
               Ecological assessment component
               Biological effects in soil
               macroinvertebrates, and field
               surveys of in situ effects and
               soil communities
                                   Candidate  field  survey or test
                                   methods
                                   Standard  laboratory  tests using
                                   earthworms,  nematodes, ants, or
                                   isopods
                                                  Field  tests using earthworms, ants,
                                                  or  isopods

                                                  Field  surveys for soil coranunities
                                        Methods potentially  applicable  to
                                        Oregon field site, Alkali  Lake
                                        disposal area

                                        Laboratory  tests with  nematodes,
                                        harvester ants  or isopods;  soil
                                        connunity tests (raicrobial  or
                                        nematodes)  using soil  cores

                                        Field tests using harvester ants or
                                        isopods

                                        Field collections of frequently
                                        captured surface- and  subsurface-
                                        dwelling invertebrates and
                                        vegetation-inhabiting  invertebrates
                                      Methods potentially applicable to
                                      Montana field site, Mllltown
                                      Reservoir wetlands

                                      Laboratory tests using nematodes
                                      or earthworms; soil community
                                      tests (microbial or nematode)
                                      using soil cores

                                      In-field testing with earthworms
                                                                                                                Field collections of frequently
                                                                                                                captured surface-dwelling and
                                                                                                                vegetation-inhabiting
                                                                                                                invertebrates
IS}
Biological effects in sediment
raacroinvertebrates, and field
surveys of in situ effects and
benthic communities
Standard laboratory tests using
benthic raacroinvertebrates
                                                 Field tests using benthic
                                                 macroinvertbrates
Laboratory tests using alkaline
tolerant sediment- or water column-
dwelling invertebrate (e.g., brine
shrimp)

Ho field tests currently available
for alkaline sediments or surface
water
Laboratory tests using Hyallela
azteca and Chironomus riparius
                                                                                                                 In situ  testing  with aquatic
                                                                                                                 macroinvertebrates  such as
                                                                                                                 Daphnia  maona
                                                 Field surveys for benthic coranunities
                                                 residues evaluation
                                                                          Field survey collections of
                                                                          sediment-dwelling  larval
                                                                          invertebrates and  brine shrimp
                                                                          community and contaminant
                                                                              Field collections of benthic
                                                                              communities for sediment quality
                                                                              triad analysis

-------
              Ecological assessment  component
               Biological effects  in
               terrestrial and wetland
               vertebrates, and  field surveys
               of in situ effects  and wetland
               communities
               Preliminary food-chain
               contamination evaluation
to
tn
Candidate field survey or test
methods
Standard laboratory tests using small
mammals, birds, or amphibians
Field tests using terrestrial
vertebrates

Field surveys for terrestrial
vertebrate communities
Designed feeding trials using
pertinent test species (e.g.,
terrestrial or wetland vertebrates)
Methods potentially applicable to
Oregon field site, Alkali Lake
disposal area

Candidate methods include
laboratory tests using small
mammals, birds, or amphibians

Candidate field methods using small
mammals, birds, and amphibians

Field surveys for small manraals and
other terrestrial vertebrates;
collection for analysis of gut
contents

Candidate studies address sediment-
dwelling benthic invertebrate and
shorebird food-chain contamination
evaluation
Methods potentially applicable to
Montana field site, Militant
Reservoir wetlands

Laboratory testing with
amphibians (or terrestrial
vertebrates)

In situ testing with amphibians
                                                                                                                              Terrestrial vertebrate surveys
                                                                                                                              and collections for contaminant
                                                                                                                              residues analysis
Candidate studies address
vegetation consumption by small
Baranalian herbivores

-------
Overall, the field activities for the ecological assessment at each site illustrate the role that
preliminary scoping activities play in the ecological assessment process.

Synopsis of Biological Methods at Montana Field Site

In the preliminary ecological  survey  for  Milltown Reservoir  wetlands,  terrestrial
vegetation  assessments were accomplished  through a  wetlands  delineation (Federal
Interagency Committee 1989), and standard plant tests for soil contamination evaluations
(seed germination, root elongation, and plant vigor;  see Warren-Hicks, et at.  1989;
Linder, et al. 1992).  In emergent zone habitats, plant and animal tests were  conducted
in laboratory and field (Warren-Hicks, et al. 1989; Linder,  et al.  1992;  ASTM  1992).
In situ tests were conducted using amphibians, and submerged aquatic vegetation was also
evaluated in emergent zone habitats (Linder, et al. 1992).  In upland areas earthworm and
seed germination tests were completed on-site and in the laboratory.  Root elongation
tests using  groundwater were also completed.  Additionally, surface soils were sampled
for metals  analysis (US EPA 1986) and characterizations of soil  quality (Page,  et al.
1982). In  order to determine food source  contamination for site-specific receptors like
small mammals, samples of native vegetation and terrestrial invertebrates were collected
and analyzed for target analytes (cadmium, copper, zinc, lead, and arsenic). For aquatic
food-chains, samples of aquatic biota, including fish, aquatic insects, and plants, were
collected and analyzed for target analytes.  Sediment samples were also collected from
emergent zone habitats and analyzed to characterize the extent of metal contamination in
depositional areas in the wetland.

Synopsis of Biological Methods at Oregon Field Site

As part  of the preliminary ecological survey  for Alkali Lake, terrestrial  plants and
invertebrates were sampled  in various seasons. Aquatic invertebrates were also sampled
at various times of the year, and representative species were identified and analyzed for
tissue residues of target analytes (2,4-D  and chlorophenols).  Vegetation  measurements
taken during the survey determined cover and qualitative  evaluations of plant health.  In
conjunction with survey samples for vegetation, terrestrial invertebrates  were sampled
using a combination of pit-fall traps and net surveys;  for terrestrial vertebrates,  small
mammals were trapped and gut contents were analyzed to  evaluate food sources.   In
addition to  small mammal surveys, habitat use by birds was also noted; shore birds were
surveyed and seasonal patterns in habitat use were recorded.  With these preliminary field
results target species were identified, and conceptual food-chain models were  developed
for risk assessments for terrestrial vertebrates.  Vegetation patterns were  also evaluated
for their  future role in long-term site monitoring.
                                        216

-------
SUMMARY

Montana field site.  Milltown Reservoir wetlands represents a single operable unit
within the Clark Fork River valley, but the wetlands must be viewed as a whole if they
are to be considered within ecological contexts.  Nonetheless, these preliminary studies
have  been incorporated into the  baseline  risk  assessment being developed  for the
wetlands, and suggested future work in the continuing ecological risk assessment process.
For example, soils and sediments at Milltown wetlands may be characterized as having
greater metal loads than the surrounding watersheds, although the historic and current
release and movement of metals may be quite dissimilar.   At Milltown wetlands, the
measures of exposure-point concentrations suggest that metals, and specifically target
analytes ~ arsenic,  cadmium, copper, lead, and zinc -- occur at greatest concentrations
in deposition zones located throughout Milltown wetlands.  On the basis of total metal
concentrations, it may be said that soils and sediments are both metal-enriched,  though
biological assessments suggest that only a small,  and variable, fraction of that metal is
biologically available and associated  with  subtle  biological effects.   Clearly, these
deposition areas may warrant additional study, or at a minimum should be monitored to
ensure that adverse effects do not develop with time.

As potential sources of metal, the soils and sediments, as well as the  associated water
column,  must be considered in light of the potential biological and ecological receptors
inhabiting  Milltown wetlands.  While diverse in  their  selection,  the  biological  tests
applied during the  field and laboratory efforts at Milltown wetlands fall  short  of
anticipating all the receptors potentially at risk. For example, while terrestrial vertebrates
such  as  small mammals have been considered,  the comparative-toxicity data base is
clearly incomplete.  Evaluations of vegetation completed at Milltown wetlands may have
missed soil microbial effects.  In summary,  an understanding of the potential effects of
metals on  communities and  populations at Milltown  wetlands  must be developed by
inference, based upon these quantitative estimates of effects made during the ecological
assessment.

The biological assessments at Milltown wetlands suggest that little,  or no acute toxicity
or adverse biological effects are occurring now. Consistently, and regardless of the field-
or laboratory-test methods used, biological assessments at Milltown wetlands could not
characterize acute toxicity; however, evidence of subtle biological effects was noted in
samples collected from  depositional areas.   Future  remediation plans should weigh the
potential biological  and  ecological  impacts associated with remediation efforts against of
the current impacts associated with elevated metals in soils and sediments.  While the
current subtle effects and  potential  effects  should  not be understated in the risk-
assessment, currently any widespread physical alteration of wetland habitats  may not be
justified on a technical basis.  Future site monitoring should address problems that may
occur with  remediation at Milltown wetlands and  at  upstream operable units, and in
particular should consider the long-term effects of metal exposures on vegetation in
                                        217

-------
depositional areas. Also, indirect effects associated with future management plans (e.g.,
habitat alterations associated with "no action" or sediment dredging alternatives) should
be considered as a potential consequence of the metals that presently occur in the soils
and sediments at Milltown wetlands.

The data gathered from laboratory and field work at Milltown wetlands has reduced the
uncertainty associated with the baseline ecological risk assessment, and has suggested that
biological and ecological effects, if present in the wetland, were subtle.  The uncertainty
associated with historic data or data collected in the current assessment at Milltown
wetlands must be stressed, e.g, were surrogate test species (in either laboratory or field)
sufficient for the  biological assessment?   Questions like these must be considered
proximate sources of uncertainty in the ecological risk-assessment process at Milltown
wetlands, and in any management decision derived in part  from these  studies.   At
Milltown wetlands:

       •      The  preliminary survey of  linear  food-chain contamination showed
              evidence of metal  bioaccumulation  in  small  mammalian  herbivores.
              Bioconcentration of metals was also occurring in emergent vegetation in
              some  reaches  of the  deposition zones within  Milltown wetlands,  but
              biomagnification of metals in herbivores did not appear to be a problem
              for the endpoints considered.  Field surveys at Milltown wetlands did not
              contradict these  assessments,  but  the sparse comparative-toxicity data
              available,  particularly for chronic endpoints, must be considered as a
              source of uncertainty.

       •      Vegetation tests, especially those used in evaluating water collected from
              the  rhizosphere, suggested that no acute effects  were  associated with
              groundwater and surface water, but subtle  growth-related  effects were
              occasionally observed  in samples collected from deposition zones at
              Milltown wetlands.  These subtle indications were noted in both laboratory
              tests using emergent  vegetation and  in  standard  root-elongation tests.
              Again, field surveys found no sign of altered vegetation  patterns  and
              reduced cover  was evident  only  in those areas  that had previously been
              physically manipulated.

       •      Earthworm evaluations  in  both  field and  laboratory  were consistently
              negative for acute toxicity, suggesting that soil macroinvertebrates may not
              be  at  great risk  as  long as the  current soil  conditions  exist.  (This is
              provided that earthworms are good sentinel species for assessing  soil
              health.)  Soil  microbial communities,  however,  were not adequately
              described and should be evaluated when methods are available.
                                        218

-------
       •     Preliminary  studies using amphibians  suggested that  subtle biological
             effects may be found at Milltown wetlands, but qualitative field surveys
             did not support any conclusions that those effects would be prominent nor
             adverse.

       •     Preliminary studies using bacterial tests yielded results consistent with the
             balance of biological-tests used at Milltown wetlands, but may not be
             representative of the soil community there.

       •     Plant-uptake studies  suggest  that garden  crops  bioconcentrate metals.
             Bioconcentration of metals in plants differs depending upon  the plant
             species and anatomical  feature considered (e.g., root versus leaf).

       •     Characterization and metal analysis of Milltown wetlands soils, sediments,
             and biota indicated that  metals have accumulated in various environmental
             matrices, and that Milltown wetlands is spatially quite heterogeneous with
             respect to metal deposition.  Within-sample unit variation was  relatively
             less than across Milltown wetlands variation, and metal deposition was
             consistent within topographic settings.   Soils within sample units were
             relatively homogeneous.

Oregon field site.  While the present  work was designed for data collection and survey,
its role within an ecological risk assessment for Alkali Lake has become essential.  The
critical early phases (problem formulation, and pilot investigations for exposure and
ecological  effects  assessment) of the ecological risk assessment process have been
addressed,  and  existing data  needs have been identified, which should allow more
adequate characterizations of uncertainty during the risk assessment process.  Baseline
exposure and ecological effects assessments can now be developed and based upon site-
specific data, and any future data needs may be clearly identified.  For example, within
the ecological  effects  assessment, field  surveys and  literature  reviews have been
considered, but  by  design little toxicity assessment has been  completed.   Future work
could  be developed to address this  data gap,  and additional comparative  literature
regarding contaminant effects could then be gathered for evaluating site-specific toxicity
data.   Whether  future work regards these specific data needs,  the baseline ecological
assessment at Alkali Lake provides empirical information upon  which decisions can be
based.

Numerous technical recommendations can be identified as a result of these baseline data
collections.  While the studies at Alkali Lake were preliminary and not developed as a
risk assessment,  the information should be considered in developing management plans
for Alkali  Lake, including any future  ecological risk assessment.   On  the basis of
historical information and the completed field reconnaissance, the following may  be
relevant to future management decisions regarding Alkali Lake:
                                        219

-------
       •      With the baseline information gathered during these preliminary studies,
             efforts should be continued to evaluate vegetation pattern changes at Alkali
             Lake.   The biological  information subsequently  generated in  such  a
             monitoring program could potentially benefit future evaluations of the
             effects of groundwater quality on vegetation.  These should be completed
             using,  for example, plant vigor tests modified  to  use representative
             species, e.g., Sacrobatus vermiculatus.

       •      Periodic soil sampling should be included in a monitoring program to
             evaluate soil contamination and  the  changing spatial patterns  of the
             groundwater plume.

       •      Additional  aquatic  invertebrate  sampling  should  be  considered  in
             conjunction with an evaluation of shorebird diet and foraging activities at
             Alkali Lake, particularly during the nesting season (e.g., spring and early
             summer field  activity).   Modified standard tests  (ASTM 1992) using
             nestling shorebird exposures may be beneficial to evaluations of risks to
             this terrestrial vertebrate.

Future work must reflect management and policy related issues relevant to site managers,
and field and laboratory efforts can directly address issues regarding risk characterization
and uncertainty.  Ecological risk assessments then could be developed more soundly and
site management more adequately implemented.

ACKNOWLEDGEMENTS

Developing  field and  laboratory methods  for  ecological  risk  assessment has  been
supported in part by U.S. Army Biomedical Research and Development Laboratory, Fort
Detrick, Maryland.  Portions of these proceedings have also been reported at the Third
International Symposium on Field Screening Methods for Hazardous  Waste Sites  held in
Las Vegas, Nevada (February 1993), and will be considered in more detail in Linder, et
al. (1994).

REFERENCES

ASTM. 1992.  Annual book of ASTM standards, Volume 11.04, American  Society tor
Testing and  Materials  (ASTM), Philadelphia, PA.

Federal Interagency Committee for Wetland Delineation.   1989.  Federal  manual for
identifying and delineating jurisdictional wetlands. U.S. Army Corps of Engineers,  U.S.
Environmental Protection Agency, U.S. Fish and Wildlife Service, and U.S.D.A. Soil
Conservation Service,  Washington, D.C. Cooperative technical publication, 76pp.
                                       220

-------
Fribush, H. and J. Fisk.  1991.  Field analytical methods for Superfund. In Proceedings
of Second International Symposium on Field Screening Methods for Hazardous Waste and
Toxic Chemicals, Las Vegas, Nevada, pp. 25-30.

Linder,  G.,  J. Wyant,  R. Meganck, and B. Williams.  1991.   Evaluating amphibian
responses in wetlands impacted by mining activities in the western United States, in Fifth
Biennial Symposium on Issues and Technology in the Management of Impacted Wildlife,
Thome Ecological Institute, Boulder, CO. pp. 17-25.

Linder, G., E. Ingham, C.J. Brandt, and G. Henderson.  1992. Evaluation of terrestrial
indicators for use in ecological assessments at hazardous waste sites, EPA/600/R-92/183,
U.S. Environmental Protection Agency, Environmental Research Laboratory, Corvallis,
OR.

Linder,  G.,  Bollman, M., Gillett,  C.,  King,  R.,  Nwosu, J., Ott, S.,  Wilborn, D.,
Henderson, G.,  Pfleeger, T., Darrow,  T.,  and   Lightfoot,  D.   1994.  The role of
preliminary field and laboratory studies  for ecological risk assessments in wetland and
terrestrial habitats: two case studies.  la Environmental Toxicology and Risk Assessment -
Third Volume, ASTM STP 1218, Jane S. Hughes,  Gregory R. Biddinger, and Eugene
Mones,  Eds., American Society for Testing and Materials, Philadelphia.   In Press.

Page, A.L.,  R.H. Miller, and D.R. Keeney,  Eds. 1982.  Methods of soil analysis. Part
2: Chemical and microbiological properties,  Second Edition.  American  Society of
Agronomy Monograph 9. Madison, WI.

Parkhurst, B., G. Linder, K. McBee, G. Bitton, B. Dutka, and  C. Hendricks.  1989.
Toxicity tests. In W. Warren-Hicks, B. Parkhurst, and S. Baker,  Jr. (eds.).  Ecological
assessment of hazardous waste sites.  EPA/600/3-89/013, U.S.  Environmental Protection
Agency, Environmental Research Laboratory, Corvallis, OR,  pp. 6-1 - 6-66.

US EPA.   1986.  Testing methods for evaluating solid waste.   U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response, Washington, D.C.,
SW-846.

US EPA. 1991.  Ecological assessment of Superfund sites:  An overview; ECO Update,
Vol. 1,  Number 2.   U.S. Environmental Protection Agency, Office of Solid Waste and
Emergency Response, Office of  Emergency and Remedial Response, Hazardous Site
Evaluation Division (OS-230), Washington, D.C., Publication 9345.0-051.

Suter, II, G.  1993.  Ecological risk assessment.  Lewis Publishers, Inc.   Chelsea, MI.

Warren-Hicks, W., B. Parkhurst, and S. Baker, Jr., Eds.  1989.  Ecological assessment
of hazardous waste sites. EPA/600/3-89/013.  U.S. Environmental Protection Agency,
Environmental Research Laboratory, Corvallis, OR.
                                      221

-------
       INCREASING THE SENSITIVITY OF FIELD HEADSPACE
          ANALYSIS FOR VOLATILE ORGANIC COMPOUNDS

                     Charles Van Sciver
                       Robert Fowler
           New Jersey Department of Enviromental
                   Protection and Energy
             Environmental Measurements Section
                     25 Arctic Parkway
                 Trenton, New Jersey  08625


ABSTRACT

The  use  of  portable  gas  chromatographs  for  on  onsite
analysis of volatile organics using the headspace analytical
technique has rapidly  increased. The  increased use of these
methods  is  in response  to  the significant savings  in both
rapid turnaround  time  of results and  savings  of analytical
costs as compared to conventional  laboratory  analysis. This
and other field screening  analytical  methods  have proven to
be a  valuble asset toward  achievement of more  rapid,  cost
effective strategies for performing site investigations.

This  study   evaluated  the  ability  of several  methods  to
increase the  sensitivity of  headspace analysis.  The methods
included  the addition  of  salt,  heating  and agitation.  A
Photovac  10S50  field  portable  gas  chromatograph  equipped
with  a  photoionization  detector   (PID)  was   used  for  the
analysis  of  the   samples.   The   target   analytes   were
trichloroethylene  (TCE),  perchloroethylene  (PCE),  benzene,
toluene, ethylbenzene,  meta and ortho xylene.  Three  vials
were prepared, one without  salt, one  with 5g  NaoSO4 and one
with  lOg   Na2S04.   The   vials  were   spiked  to   yield  a
concentration of  50 ppb  for each target analyte. Triplicate
analysis was performed on each spiked sample.  The vials were
then placed  in a  hot water bath at  65°C  and  analyzed,  then
shaken  and  analyzed.  The  percent  increase in response was
determined  for each method  and compound.  Response increases
of BTEX, TCE and  PCE were greatest with  the  combination of
adding  salt  and   heat  to  the  sample.  Individually,  the
addition  of  heat to  the  sample   produced   the a  greater
response compared to the  addition  of  salt alone,  although
both produced responses greater than 100% for all compounds,
except for benzene, compared to the reference solution. This
evaluation  demonstrates  several methods  which  can be  used
seperately or in  conjunction to  increase  the  sensitivity of
headspace analysis under field conditions.


INTRODUCTION

Soil  and ground  water  contamination  by volatile  organic
compounds (VOC)  is frequently  encountered at  industrial and
                              222

-------
hazardous  waste  sites.   The  physical  properties  of  VOCs
allow  for  their  rapid  movement in  the  vadose zone to the
groundwater.   Once in  the  ground  water,  VOCs may rapidly
impact potable wells causing  immediate health concerns to
receptors.

The  concern   for  the  rapid   assessment   of  groundwater
contamination  has in turn created  a need  for rapid methods
to  assess   the  extent  of  a   problem.   The  use   of  field
screening  analytical  methods for  the analysis of volatile
organic  compounds  in   groundwater  offers   a   means  for
real-time  decision  making   during  groundwater  contaminant
studies.

Compound specific  and  quanitative  field  static  headspace
analytical   techniques   l\ave   been  developed    for   gas
chromatographs  (GC)  I1/2/3).  The static  headspace method is
based  upon  the partitoning  of volatile organics between the
aqueous and vapor phases when a  sample is enclosed in sealed
container.  An  aliquot of the headspace is then withdrawn and
analyzed by the GC equipped with a photoionization detector
(PID).


THEORY

The  principle  of  all  quantitative  headspace analysis  is
based  on the  thermodynamic  conditions  of the phases.  These
conditions  require that equilibrium be  established between
the liquid  and gas phase.  The concentration of a VOC in the
headspace of  a liquid under static conditions is based upon
contaminant concentration in the liquid and its distribution
coefficient or Henery's Law.

                    C1   =   KG11
           Where:
                    C1   =   Initial Concentration
                    K    =   Distribution Coefficient
                    cH  =   Final Concentration

The   distribution   coefficient   is   a   measure    of   the
partitioning   of   a  compound  between the   liquid   and  gas
phases.  The sensitivity of the  static headspace method will
vary between  VOCs  due  to their  extremely  wide range  of K
values.   Therefore,  to  increase  the   sensitivity of  the
static headspace  technique,  the value of  the distribution
coefficient must be decreased.  This will increase the  mass
of VOCs driven into the headspace.

The  techniques   used  to   accomplish  this   under  field
conditions  include the  addition of  salt or  "salting out"
which  reduces  the solubility of  a compound  and  drives it
into the headspace. The use of heat and agitation applied to
a  sample  can  also drive  additional  amounts of  volatile
                              223

-------
compounds into  the gas  phase,  thus becoming  availible  for
detection.
OBJECTIVE

The objective of  this  study was to evaluate  the ability of
various  methods  to  increase the  sensitivity  of headspace
analysis for  selected  compounds.   The methods  included the
addition of  sodium  sulfate  (Na2SO4),  heating the sample to
65°C, and physical agitation.

The   target   analytes  were  common  aromatics,   benzene,
toluene, ethylbenzene,  meta-xylene, and ortho-xylene  (BTEX).
Also   targeted   were   two  common   chlorinated  aliphatic
hydrocarbons, trichloroethylene  (TCE)  and perchloroethylene
(PCE).


METHODOLOGY

Two groups of  40  ml. screw  cap septum vials  were prepared.
Each group had one vial without salt  (number 1), one with 5g
of  Na2SO4   (number  2)   and  the  third  with  lOg  of  Na2SO4
(number  3) .   The solutions were  prepared  by  filling each
vial  with   laboratory   grade ASTM  Type  II water with  no
headspace.  The  target  analytes were  added  to  each  vial by
piercing  the  septum  with  a  syringe  and  injecting  the
compounds in  sufficient quanity to yield  a concentrationof
50  parts per  billion   (ppb)  for  each  target  analyte.  The
vials  were  shaken intermittantly for ten minutes to ensure
the compounds were mixed thoroughly-

The headspace was created by withdrawing 10 ml of solution
from  the sealed  vial  with  a  10 ml  syringe.  To  allow air
entry,  a separate syringe needle with an  attached charcaol
tube  to prevent  the  entry of  any  air contaminants,  was
placed through the septum and into the vial.

Each of the three vials in Group 1 were spiked with benzene,
toluene,   ethylbenzene,    meta-xylene    and    ortho-xylene
resulting in an approximate concentration of 50 ppb for each
compound.  Each of  the vials  in  Group II  were spiked with
TCE and PCE resulting in a concentration of approximately 50
ppb.   All  vials  were  allowed   to   equilibrate  at  room
temperature (20°C) for one hour.

A  Photovac  10S50 field gas chromatograph  equipped  with  a
10.6  ev photoionization detector  (PID)  and a  CPSil  5 wide
bore  capillary  column  under isothermal  conditions  of 40°C
was used to perform all analyses.  All samples were manually
injected with gas  tight  syringes.    Syringes   were  purged
between  samples with ultra high  purity air (Ultra Zero).   A
                              224

-------
one point  calibration was  performed using sample  number 1
(without Na2SO4).

Each  vial   was  analyzed   in  triplicate   under  ambient
temperature  (20°C)  static headspace conditions.   Each vial
was then placed inverted in a hot water bath at 65°C for one
hour  for temperature  equilibration.  Each  vial  was  again
analyzed in triplicate, returning  the  vial to the hot water
bath  after  each  injection.    Care was   used  to  minimize
agitation of the vial  during  handling.   Sample  Nos. 1 and 3
from each group were removed from the hot water bath, shaken
for 60  seconds and  immediately analyzed.  The  results for
each sample were averaged and compared with Sample 1 of each
group.


RESULTS

The results  are included  in Tables  I and  II  which give the
percent  increases  in  response  due  to  the  addition  of
Na2SO4, heat, agitation and a combination  of all as compared
to sample number 1.


DISCUSSION

The results  revealed the sensitivity for  BTEX  and selected
chlorinated   aliphatic  hydrocarbons  increased   with  the
addition of heat, salt and agitation.

For TCE  and  PCE the addition  of lOg of NaoSO4 resulted in a
larger   response   compared  to  the vial  containing  5g.
Additional  literature  states mineralization  over  20%  would
only  result in slight increases in sensitivity,  and  would
lead to excessive waste of the  reagent.I1)

Heating the sample to  65°C resulted  in sensitivity increases
which  were  comparable to  adding  lOg  of Na2SO4.   Heated
headspace  for PCE  gave  a 14%  greater  response to  the PID
compared to the addition lOg Na2SO4.

When heat was applied  to sample 3A,  the increase in response
was  greater   than   the   combined  responses  of  heat  and
mineralization.   This was  most pronounced  for TCE with a
488% increase in response.

Benzene   and  toluene  were   not   as  responsive   to  the
mineralization  as TCE  and PCE.  They were  more responsive to
the addition of heat,  which gave on  an average, a 1.5-2 fold
increase  in  response compared  to  the  lOg mineralization of
the  sample.   The  addition of  heat and  mineralization for
the BTEX samples gave an  increase  in  response  greater than
the sum of the  seperate  increased  responses for  both heat
and mineralization.  This result was similar to the response
                              225

-------
                         TABLE I
          PERCENT INCREASE IN RESPONSE TO BTEX
Sample #
Benzene
Toluene
Ethyl -
Benzene
Meta-
Xylene
Ortho-
Xylene
2
Salt
5g
70.4%
103.9%
112.3%
127.1%
148.8%
3
Salt
lOg
99%
124%
110%
131%
198%
1
HWB
138%
223%
287%
312%
277%
2
Salt
HWB
242%
342%
419%
458%
634%
3
Salt
HWB
345%
391%
4,44%
483%
752%
1
HWB-SHK
149%
250%
347%
372%
413%
3
Salt
HWB-SHK
534%
570%
699%
758%
1255%
Salt  -   Sodium Sulfate.
HWB   -   Sample placed in hot water bath @ 65°C.
SHK   -   Sample shaken for one minute prior to analysis.
                          226

-------
                        TABLE II
       PERCENT INCREASE IN RESPONSE OF TCE AND PCE
Sample #
TCE
PCE
2A
Salt
5g
125%
124%
3A
Salt
lOg
171%
137%
1A
HWB
161%
151%
2A
Salt
HWB
252%
228%
3A
Salt
HWB
488%
377%
1A
HWB-SHK
225%
273%
3A
Salt
HWB-SHK
947%
572%
Salt -    Sodium Sulfate.
HWB  -    Sample placed in hot water bath @ 65°C.
SHK  -    Sample shaken for one minute prior to analysis.
                          227

-------
of the  chlorinated  hydrocarbons under the  same conditions.
The  greatest  increases  in  response  were   with the  later
eluting  and   heavier  compounds.    Ortho-Xylene   had  the
greatest increase in response (752%).

To further increase the response of the  select compounds to
the PID, the vials were shaken vigorously for one minute and
an  aliquot  of  the  headspace  immediately   injected.   This
physical action moves additional  VOAs out  of  solution into
the headspace.  Only  one sample was injected,  but responses
over  ambient  temperature  headspace  for  the BTEX  compounds
ranged from a 534%  increase  for benzene  to  a 1255% increase
for ortho-xylene.   Similar increases in responses were noted
for TCE and PCE with 947%  and 572%  respectively.


Figures I  and II show the  comparison of chromatographs  of
the  BTEX,   TCE  and  PCE   samples   which  were  at  ambient
temperature, mineralized,  heated,  and agitated.

There was  one problem encountered with the mineralization
technique.   When  adding  the water  sample  to the  vial with
Na2SO4,  a  bubble  results w  1-3  mis.  This  is due  to  air  in
the pore  spaces  of the Na2S04.  When 10 mis   of  water  are
removed to create the headspace varying volumes of headspace
are created  which can affect  the  comparability of results
between samples.   When the water sample is added to the vial
slowly  and the  vial  is  gently  agitated  to   saturate  the
Na2SO4,  a loss of VOAs will result.

To  eliminate  the problem  of varying  headspace, all  vials
should be  marked  at a volume of 30 mis or graduated vials
can be  used.   When  the  headspace   is  created,  a  volume  of
water is  removed  to  the  30 ml mark.   Therefore  all  vials
will have an equal volume  of headspace.

Caution  should  be  taken  when  using  heated headspace  and
shaking the vials with analysis by  GC.   The  GC column should
have  a  packing which  is  hydrophobic to eliminate  moisture
buildup  in the  column.   Excessive  moisture  can   clog  the
column and possibly damage the detector.
SUMMARY

This  limited  study has  shown  that  increased  analytical
sensitivity can be obtained when employing any number of the
techniques tested. The greater sensitivity to the PID can be
used to  detect lower concentrations  of compounds  in water
with  less  sensitive instrument  gains  or  less  sensitive
detectors such as an FID.   This also reduces the problems of
syringe contamination since at  less  sensitive settings, any
                              228

-------
                                  FIGURE I
                          CHROMATOGRAMS FOR BTEX
 PHOTOUAC
 PHOTOUAC
 PHOTOUAC
STPRT


STOP B  60B.0
SffllPLE LIBRflRT 1 PEC 3B 1332 13= 1
WflLTSIS »   4  uofl TEST
INTERWflL TEnP 28  2SB UL
GPIN      2  1-STP

CDnPOUNP MRTE PEflK R.T.  PRERXPFn
STOP 8  600. a
SffllPLE LIBRPRT 1  JBN 7 1333
BMflLTSIS «  IB  UOfl TEST
INTERNfiL TEHP 23  2SB UL
GBIN      2  2-STP
                             conpourip
STOP a  575.0
SM1PLE LIBRflRT 1  Jffll 7 1333  13JM
flNPLTSIS a  13  UDfl TEST
INTERNAL TEnP 23  23B UL
GfilN      2  3-STP

     NflnE PECK f«.T. RREP/PPn
                                      PEBK R.T. PREP/PPM
UNKNOUN
BEN2EHE
TOLUENE
UNKND'JN
ETHTLBEN2ENE
n-p XTLENE
o XTLEHE
i
"2
•3
1
5
6
7
25.
88.
J86.
317.
33B.
123.
5BB.
8
3
1
3
8
6
1
82.3
16.
«.
311
1B.
33.
11.
55
16
.8
61
11
B8
HiUS
PPB
PPB
nUS
PPB
PPB
FFB
UNKNOUN
TOLUENE
ETHTLBENZENE
n-p XTLENE
0 XTLENE
J
2
3
1
5
6
26.7 21.7 nUS
83.7 71.12 PPB
186.3 83. 12 PPB
331.5 82.37 PfB
121.1 8B.B6 PPB
5B2.2 31.72 PPB
UNKNDUN
BENZENE
TOLUENE
ETHTLBENZENE
n-P XYLENE
0 XTLENE
J -26. B 10. 3 
-------
                                   FIGURE  I CONTED1
                               CHROMATOGRAMS  FOR  BTEX
IPHOTOURC
                     PHOTOUAC
                   IPHOTOUflC
                                SIRRI.,
STOf *  S7S.B
sunruE LIBRPRT i
        17  UOO TEST
IMTERWL TEW 23
Oflf       2
                    STOP 8  S25.0
                    SW1PLE LIBRflpT 1  JW< 7 1333 H =23
                    PMOLTSIS »  13  UOP TEST
                    IHTERWL TEHP 23  23B UL
                    WlIM       2  2-STD NJB
                  STOP 8  575.0
                  SWIPLE LIBPWT 1
                  BMflLTSIS o  21
                  INTERNAL TEHP 28
                  SPIN       2
J«N 7 1333
UOft TEST
23B UL
3-STD HUB
BENUfE
UNM-iuUfl
TDLUtW
          PEfx R.r.
26
83
1 11
IB*
262
318.5
                                conpouNO
                                          PEPX P.T.
                                                               CDWDUMO NW.E  PEW F.T.
n-P XYLENE
D XTLEKE
    73.7 nUS
    I IB.2 PfB
    31. 2 «*>S
    112.1 PPB
    23. 3 P>US
     2. 1  US
                                TOLU£M£
33J.6 1«5.7 PPB
173.6 jei.2 PfB
SB I . 3 136. 8 PPB
                    n-P XTl_EH
                    0 XTLENE
 •26. I 15.6 fUS
 83.7 155.0 PPB
111.3 37.3 *US
J85.7 131.5 PFE
76?.3 31.5 pUE
33B.6 •JIS.l PPB
123.6 217.2 PPB
5BI.3 311.8 FPB
                                                               TOLUENE
                                                               P-P
                                                               0
  75.8  16.8 r<-'S
  83.I 201.7 PCS
  111.J  50,B rUS
  JP6.1 113.7 PPH
  262.3  26.5 <•<-'!.
  33B.B 223.B PPB
  Itt.f 217. 3 rfB
  jfc 1 . 3 311. I PPB
       SAMPLE  1
         HWB
                          SAMPLE  2
                      5g  Na2SO4/  HWB
                       SAMPLE  3
                   10g Na2SO4/ HWB
                                           230

-------
                                           FIGURE  I  CONTED'
                                      CHROMATOGRAMS  FOR  BTEX
 PHOTOUAC
ETPRI
STOP a  575.0
SAHPLE LIBRARY 1  JAN 7 1333 16=
ANALYSIS *   26   UOA TEST
INTERNAL TEflF 23   250 UL
GAIN         2   1-STD HUB-SHAKE

cnnPOUND NAPE  PERK  (5.T.
 UNKNCUN
 BENZENE
 UMKNOUN
 TOLUENE
UNKNOUN
EWYLBENiENE
n-P XYLENE
0 XYLENE
 26.B 17. 0 fiOS
 83.5 112.7 FFB
111.3 33.7 mVJS
186.1 152.1 FPB
•262.3 11.5 <"US
313.1  1.8  KS
331.S 183.1 FPS
121.1 183.2 FFE
5B1.3 217.6 PPB
PH01
rounc
                                         STPRT
                                        STDP 8   575.0
                                        SW1PLE LIBRPRT 1 JflM  7 1333  16:26
                                        BMPLTSIS »   28  UDfl TEST
                                        INTERNAL TEPP 23  25B UU
                                        GAIM        2  3-STD HUB-SHAKE

                                               NWIE  PEflK I?.T.
                                        UNKNOUN
                                        BEN1EWE
                                        UNKfiDUM
                                        TDLUENE
                                        UNKMOUIM
                                         n-P XYLEME
 •26,1 10.2 mUS
 83.3 287.2 FPB
J11.3 38.2 piOS
187,3 231.S FFB
262.8 61.8 jnUS
331.5 316.5 FFB
121.1 333.0 PFB
                                                      8 502,2 57^.6
          SAMPLE  1
          HWB-SHK
                                                SAMPLE  3
                                       109  Na2SO4,  HWB-SHK
                                                        231

-------
                                      FIGURE  II
                         CHROMATOGRAMS  FOR TCE AND PCE
': or *   3eo. o
innpLE LIBRKRT -> JRT» B 1333 >i--5i
flNflLTSIs »  37  UDP TEST
INTERWL TEHP 27  25B UL
GpJN       2  IP STP

COI^ L'ij;f.5 :07.7 ffil
            "J  270. -3 135.5 PPS
PHOTOURC
6JDP »   30B.0
SfmPLE LIBRflRT 3 JPM 8 133U I3'.2l
flnftLTSIS »  3l  UOP TEST
IMTEIWL TEtV 28  230 UL
GOIH       2  2P STD HUB

conpou^o rwiE  FEW P.T. PREP^PPH

UNKr^DUrv       J 27.6 117.2 rnUS
TCE         7 10B.3 182.7 FPB
FCE         3 722.i 171.3 FP8
                                   PHOTOUAC
                                                                  "C1 V"iT
                                 STOP •  3B2.0
                                 SWVLE LJBKPRT 3  JWC  8 1333 11 'II
                                 PMftLTSIS «  SB  UOB TEST
                                 INTERNOL TEW 28  23« UL
                                 anil*       2  3p STD HUB
                                                                  CDHPOU^D
                                                                            PEW R.'.
                                                                  TCE
                                                                  PCE
                                              28. e )33. 3
                                              33.3 17.6
       SAMPLE 1
          HWB
      SAMPLE 2
  5g Na2SO4/ HWB
                                       SAMPLE  3
                                  10g  Na2SO4/  HWB
                                          232

-------
                               FIGURE II CONTED*
                        CHROMATOGRAMS  FOR TOE AND PCE
 PHOTOUflC
wi,..r-
STOP «  3B0.a
SPW1.E UBIWT 3
flNPLTSIS «  S5
IHTEKWU- TEnP 28
Gf»IM       2
           JBH 8 1333 13=53
           UOB TEST
           2SB UL
           1« STP HUB SMKM
     MPHJ PEF'K f. •
UTIKMOUM
TCE
PCE
          1  7^.5 66.1 enUS
          2  33.5 163.B PPB
          3 228.I 133.3 PPB
                                PHOTOUAC
                                  -s-r
STOP i  30B.O
SfinPLE LIBRflRT 3 JflU 8 1333  11=33
BHPLTSIS «  S3  UDft TEST
INTEPMW. TEMP 28  25B UL
GAIN       2  3P STD HUB SHKN
                                CDnPDUND NPHE
UNKND1JN
TCE
                                         P£PX 15. T.
             36.6 112.1 FFB
            156,8 It.5 nUS
            213.3 3J53.1 PPB
      SAMPLE  1
      HWB-SHK
                                    SAMPLE 3
                              10g Ma2SO4/ HWB-SHK
                                        233

-------
small amount of contamination under normal conditions is not
detected.   In  situations where  on  site  field  analysis of
water samples is implemented in the absence of line power to
heat the samples, mineralization and shaking the vial can be
utilized    to    increase    the   analytical    sensitivity
dramatically.

Further evaluation  of  these techniques is needed  to assess
their application  to other compounds,  detectors,  headspace
volume  ratio,   Na2SO4   concentrations   and  other  types  of
mineral salts.
ACKNOWLEDGEMENTS

The  authors  would  like  to  thank Ken  Glasser and  William
Lowry  for their  assistance and  technical  review of  this
paper.
                             234

-------
                     REFERENCES
1.   loffe, B.V.; Vitenberg, AG.
     Head-Space Analysis and Related Methods in Gas
     Chromatography.  Wiley - Interscience  p62


2.   Dietz, E.A.; Singley, K.F.
     Determination of Chlorinated Hydrocarbons in Water
     by   Headspace   Gas   Chromatography.   Analytical
     Chemistry,  Vol.  51,  No.   11,  pp.  1809-1814  Sept.
     1979.
3.   Kepner, R.E., Maarse, H., Strating, J.
     Gas Chromatographic Head Space Techniques for the
     Quanitative Determination of Volatile Components
     in  Multicomponent  Aqueous  Solutions.  Analytical
     Chemistry, Vol. 36, No. 1, pp. 77-82. Jan. 1964.


 4.    Buchmiller,   R.C.   Screening  of   Ground  Water
     Samples  for  Volatile  Organic Compounds  Using  A
     Portable    Gas   Chromatograph.     Ground    Water
     Monitoring Review, Vol. IX No. 2  pp. 126-130.
                              235

-------
ORGANICS

-------
33
          REGULATORY ASPECTS  OF RCRA ANALYSES

by Barry Lesnik, USEPA, Office of Solid Waste, Methods
Section (OS-331), 401 M St.,  SW, Washington, DC  20460
     What is RCRA?:
           RCRA is the Resource Conservation and Recovery Act of 1976 which regulates
           the management of hazardous waste under Subtitle C. Administration of the
           program has  been passed down  to  the states via  authorization  of  State
           Programs in most States.
     How is a waste classified as hazardous waste under RCRA Subtitle C?

           A waste is classified  as hazardous  under  RCRA Subtitle C if it exhibits
           any of the characteristics  of ignitability.  corrosivity,  reactivity,  or
           toxicity,  or  if it contains  specific  constituents which are  listed  as
           hazardous.
     What is SW-846?

           SW-846 is the compendium of analytical  and test methods approved by EPA's
           Office of Solid Waste (OSW) for use in determining regulatory compliance
           under the Resource Conservation and Recovery Act (RCRA).


     What  are  the  drivers  that  determine  which environmental  analyses must  be
     performed?

           Specific regulations under RCRA Subtitle C are the determining factors as
           to which analyses need to be performed,  and which specific target analytes
           need to be  identified  in order to be  in compliance.   Examples of these
           specific regulations include:

           1)    monitoring  of  leachates  from hazardous waste  landfills  (Appendix
                 IX),

           2)    determination of whether a waste is hazardous by characteristic,

           3)    compliance with boiler, incinerator,  furnace (BIF)  rules,

           4)    compliance with land disposal restrictions (LDRs),

           5)    permit  compliance  for surface  impoundments,  storage  facilities,
                 etc.,  and

           6)    corrective action.
                                         236

-------
Are SW-846 methods  required to be used for  all  analyses performed under  RCRA
Subtitle C?

      Except for the specific cases listed below, where use of SW-846 methods is
      mandatory,  SW-846  functions  as  a  guidance  document   setting  forth
      acceptable, although not required, methods to be implemented by the user,
      as  appropriate,   in  responding  to  RCRA-related  sampling  and  analysis
      requirements.  The RCRA applications  listed in 40 CFR Parts 260  through
      270 where the use of SW-846 methods is mandatory are the  following:

      (1)   § 260.22(d)(l)(i)  -  Submission  of  data in support  of  petitions  to
            exclude a waste produced at a particular  facility (i.e..  delisting
            petitions) ;

      (2)   § 261.22(a)(l)  and (2)   Evaluation of waste against  the corrosivity
            characteristic;

      (3)   § 261.24(a)   Leaching procedure for evaluation of waste against the
            toxicity characteristic;

      (4)   §§ 264.190(a),  264.314(c), 265.190(a), and 265.314(d)   Evaluation
            of waste to determine if free liquid is a component of the  waste;

      (5)   § 266.112(b)(1)    Certain analyses in support of  exclusion from the
            definition of a hazardous waste  of  a  residue which was  derived  from
            burning hazardous waste in boilers and industrial furnaces;

      (6)   § 268.32(i)    Evaluation of a waste to determine if it is a liquid
            for purposes of certain land disposal prohibitions;

      (7)   §§  268.40(a),  268.41(a),  and 268.43(a)   - Leaching procedure for
            evaluation  of  waste  to  determine  compliance with Land Disposal
            treatment standards;

      (8)   §§  270.19(c)(l)(iii)  and (iv),   and  270.62(b)(2)(i)(C)  and (D)  -
            Analysis  and    approximate   quantification   of   the  hazardous
            constituents identified  in  the  waste prior  to  conducting  a trial
            burn in support of an application for a hazardous  waste  incineration
            permit; and

      (9)   §§  270.22(a)(2)(ii)(B)  and  270.66(c)(2)(i)   and (ii)  -  Analysis
            conducted in support of a destruction and removal efficiency (ORE)
            trial burn  waiver for boilers and  industrial furnaces burning low
            risk wastes, and analysis and approximate quantitation conducted for
            a trial  burn in  support  of an  application  for a  permit  to  burn
            hazardous waste in a boiler and industrial furnace.

      In Hazardous  Waste Programs in  RCRA-authorized States,  the States can
      require the use of SW-846  methods.  A number of States have regulations
      that require the use  of SW-846  methods for  hazardous waste analysis under
      their RCRA  Programs.    Some  of these  States  require  the use of  Second
      Edition methods,  while  others  require the  use of Third Edition methods.
                                    237

-------
      State policies on the use of draft methods vary from blanket approval of
      some methods to methods review on a case by  case  basis.   In addition to
      the mandatory applications specified above,  EPA Regions  can request the
      use of specific Second Edition, Third Edition, or draft SW-846 methods for
      other regulatory applications.


What  is   the  current  version  of  SW-846  that  must be  used  for  mandatory
applications?

      The  use  of  the  SW-846  Second  Edition,  including Updates,  is  still
      mandatory for these  applications.   Also,  the Final Rule replacing the
      Second Edition with the Third Edition of SW-846 (including Update 1) for
      these mandatory applications is passing  through the  final stages of the
      regulatory process.   We expect to promulgate this  Final Rule early in
      1993.
What is an "EPA-approved" method?

      This  is  a  term that  has been  bantered  about  quite  a bit  recently,
      sometimes rather loosely,  particularly  with respect to  some of  the  new
      analytical technologies.   From the RCRA  point of view,  "EPA  Approved"
      means that a method has been incorporated by reference  in  a Final Rule
      which has been published  as  a Federal Register Notice  (FRN) either  into
      SW-846 or directly  into the  RCRA regulations.  In  short "EPA  Approved"
      methods  are promulgated  methods  which  can  be  used  without  special
      permission  for  RCRA applications  where  the  use  of  SW-846 methods  is
      mandatory.  Therefore,  until  a method is promulgated by an FRN,  it is  not
      an approved method  for  these mandatory  applications, no  matter where it
      may be in the regulatory process.

      EPA  Regions  may  grant a general Regional approval  to certain  draft
      analytical methods for  use within a Region for specific applications.   In
      addition to this general Regional approval, Regions  and States  may grant
      facility-specific  approval  for  the  use  of draft  methods  for  permit-
      specific applications.   Examples of facility-specific approval include the
      use of Method 8330 by some Regions for the  analysis of explosives residues
      by High Performance  Liquid Chromatography and Method  8290  for the analysis
      of dioxins by High Resolution GC/MS.  Both Method 8330 and Method 8290  are
      currently draft methods which will be proposed in  the Second Update.


How does a method become "EPA-Approved"?

      This is a very brief overview of the EPA  regulatory process, through which
      SW-846 regulatory packages must pass.   SW-846 methods are  not published in
      the Federal Register, but  are incorporated by reference in the appropriate
      RCRA regulations.   The  process  roughly follows the following  steps  and
      takes about 18  to 30 months to complete:

      1)    After methods are  approved  by the SW-846  Technical Workgroups,
                                    238

-------
            technical  comments  are  addressed,  and a  regulatory  package  is
            prepared.  This regulatory package consists of a proposed  rule, a
            preamble  to   the  proposed  rule,  and  the  necessary  transmittal
            documentation to pass it through the system.

      2)    After completion of  the  internal  EPA review, the regulatory  package
            is sent to the Office  of Management and Budget  (OMB)  for its review
            and approval.

      3)    After  OMB approval,  the  proposed  rule  is  signed  by the  EPA
            Administrator, is published  in the  Federal Register, and  becomes
            available for  public comment.  (The proposed methods are provided to
            SW-846 subscribers by the Government Printing Office  for comment.)

      4)    After the public comments are addressed, the  regulatory package goes
            through the same  internal EPA and OMB reviews a  second time,  and the
            rule is promulgated through a second Federal Register notice.


Can "draft" methods be used for analyses performed under RCRA Subtitle  C?

      Yes, since most  analyses  performed for compliance with  RCRA Subtitle  C
      regulations fall outside of  the scope of the mandatory applications of SW-
      846 methods,  for which promulgated methods must be used.    "EPA Approval"
      of  methods  is a  factor  only  in  the  mandatory  applications  of  SW-846
      methods under RCRA,  or  in State Programs  where use of  SW-846 methods is
      mandated.   For all other non-mandatory applications under RCRA, draft
      methods  can,  and,  in many  instances where they  provide  improved data,
      should be used.
Do SW-846 methods have to be performed exactly as written?

      Monitoring  requirements  under  RCRA Subtitle  C specify  only  that  the
      analyst must demonstrate  that  he can determine  the analytes of concern in
      the matrix of concern at the  regulatory  level  of concern.   Since SW-846
      methods are written as guidance for a wide variety of matrices, it is up
      to the individual analyst to optimize a particular method to his specific
      needs.   Allowable modifications  include adjustment  of sample  size  or
      injection  volumes,  dilution   or   concentration of   the   sample,   and
      modification or replacement of  equipment.   These method changes must be
      documented, as I stated previously the analyst must demonstrate that his
      method can meet the previously-stated analytical requirements.

      However,  methods are inflexible in some aspects, e.g.  generation and use
      of  an appropriate  calibration  curve,  determination  of recoveries  and
      precision,  and demonstration of  applicable detection limits in the matrix
      of concern.

      Philosophically, unlike other  EPA Programs, RCRA specifies "what" needs to
      be determined, and leaves the  "how" up to the analyst.
                                    239

-------
34
     STEPWISE DEVELOPMENT PROCESS FOR REGULATORY ANALYTICAL METHODS

Paul Marsden.  Methods  Project Manager,  Siu-Fai Tsang  and  Mark Roby,
Science Applications International Corporation (SAIC),  92121
     ABSTRACT

     It is unusual for an EPA Program Office to fund the development of new or
     untested  technology for  regulation  analysis.    In  fact,  most  methods
     development  projects  are  better  described  as  the  translation   of  a
     procedure practiced in a  limited number of laboratories to a written SW-
     846 method  that may be performed  by many analysts.   Although  the same
     analytical  techniques  employed in  commercial analyses may  be  employed
     during  method  development,  there  are significant  differences  in the
     approach  to method  development compared  to  commercial analysis.   For
     example,  there  is no single approach  to  method development; scientists
     must make professional judgements during  the conduct of a  study.   The
     quality  assurance  program  is  often  a   reflection  of  the  analyst's
     commitment  to  measurement  quality  and familiarity  with  the analytical
     system  rather  than  a   set  of  defined  acceptance  criteria.     This
     presentation  describes a  ten  step  process  for method  development and
     method  validation  conducted   for   the  OSW:    (1)   Identification  of
     need/definition  of  performance  objectives,   (2)    Delineation  of  QC
     procedures,  (3)  Selection of analytical  approach,  (4) Determination of
     instrument  sensitivity,  (5)  Method  optimization,   (6)  Generation  of
     performance data for a clean matrix,  (7) Documentation of interferences,
     (8) Demonstration of matrix suitability,  (9)  Calculation of quantitation
     and detection limits,  and (10) Workgroup review.  These ten points will be
     illustrated using examples of method development efforts conducted for the
     OSW.     Mechanisms  that   facilitate   laboratories   and  manufacturers
     participation in the method development effort will also be discussed.
     INTRODUCTION

     The Methods Section of the Office of Solid Waste  (OSW) has responsibility
     for  developing  and  promulgating  reliable  analytical  methods  for  the
     determination  of chemicals on Appendix  8  or  for  waste characteristics
     (e.g.,  toxicity,  flammability  and corrosivity).     These  methods  are
     designed for use  in  a  variety of laboratories and  have embedded quality
     control  (QC)  requirements that  facilitate the  intercomparison of data.
     Each method must be evaluated  for suitability for the use in environmental
     or waste matrices. Central to the validation process is single laboratory
     testing, which establishes the quality of data that can be obtained using
     a method and its suitability for  the soil matrix.  Validation of potential
     SW-846 methods is an  important element in OSW's quality assurance program.
     Following method validation,  a SW-846 method and  method validation report
     is reviewed by an OSW workgroup.
                                         240

-------
DEFINITION OF NEED

Each method  development/evaluation  project begins with  a  definition of
need.  This need is identified by the EPA Program Office responsible for
the regulation driving the measurement.   The statement of need generally
includes  the target  analyte  list,  potential  sample matrices,  target
detection  levels,   and  QC  requirements.    At  a  minimum,  method  QC
requirements must be compatible with SW-846 Chapter 1 as well as any other
applicable general  methods (e.g.,  Method  8000    Chromatography).   The
definition of need is developed into the project data quality objectives
and the target performance criteria.
ANALYTICAL APPROACH

The analytical approach for a method development or evaluation project is
based   on   measurement  sensitivity,   stability/repeatability   of  the
measurement,  and  bias.   In addition, potential  SW-846 procedures  are
evaluated on the basis of several qualitative criteria including the cost
requirements for analytical hardware/equipment, any additional facilities
requirements,  ease  of use/operator skill,  cost  per  determination,  and
analyses per day.

In  many cases,  the  analytical  approach  has  been used  for a  similar
application.  Program  Offices method studies  are generally initiated after
a  technology  has   been  demonstrated  as  useful   for  the  analysis  of
environmental or waste matrices.  Some of the sources for this information
are provided below, along with the period  of time  that might be required
to initiate a method evaluation study.
SOURCE
Instrument Vendors
EPA Regions and States
DOE National Laboratories
ORD
USATHAMA/USDA
Commercial Laboratories
Chemical Literature
Available
now
X
X
X
X
X
X

6-12 months
X

X
X
X
X
X
2-5 years
X

X
X


X
Program Offices often re-test methods obtained from other sources in order
to  establish  that   the  sensitivity,  precision,   ruggedness  or  matrix
suitability are appropriate for their regulatory needs.
                                   241

-------
INITIAL EVALUATION

Initial evaluation of a method generally starts with establishing that the
technique is operating under statistical control.   The detector  linearity
and dynamic range are  then established for determinative methods using a
calibration using at least five  concentrations of  standards.  The analyst
will determine whether there are interferences in a clean control matrix
(i.e., no false positives) for sample preparation methods.

Critical measures of  data performance  will be established during single
laboratory testing. These include the accuracy, precision, and  limits of
detection/quantitation that can be achieved in sample matrices of concern
to  the RCRA  program.    In addition,  single  laboratory testing  will
establish the probability of obtaining false positives and false negatives
with  the  technique.    Initial laboratory  testing  is generally conducted
using  soil  and  water spiked with known  concentrations of  target and
interfering analytes.
OPTIMIZATION AND RUGGEDNESS TESTING

In many cases, additional method optimization may be conducted to provide
the target sensitivity,  accuracy, precision or matrix suitability required
by the Program  Office.   This  optimization is accomplished by systematic
alteration  of  method  conditions  in  order  to  achieve  the  performance
targets.

Ruggedness  testing may also be  conducted on candidate  methods  if they
perform erratically or when it is important to specify control limits for
critical  method parameters.   AOAC-type  ruggedness testing assumes that
each  method  parameter  varied  in  the  test  is  independent.    Limited
factorial   designs   (e.g.  Planket-Herman)  are  more  appropriate  when
dependent method variables are tested.   In either approach, previously
identified  method  parameters   are varied  between  two  conditions  and
statistical  evaluations  of  the  results  are   used  to   interpret  the
performance changes.
PERFORMANCE TESTING

Method  performance  is first  established  using  a  spiked  clean  soil.
Aliquots of one soil type prepared at three concentrations: one half the
target detection limit, two times the target detection limit  and 10 times
the target detection limit.   Method bias (accuracy)  is determined at each
concentration  by  calculating  the  mean  recovery   of  the  spiked  (or
characterized) analytes for  the seven replicates.  Method repeatability is
established by running the method over  several weeks.
                                    242

-------
                                       x 100%
            X = Mean value for the seven replicate determinations

            X = Spiked or characterized concentration
      Method precision is determined by calculating the percent relative
      standard deviation of the spiked  analyte  recoveries  for the seven
      replicates at each concentration.
                       precision = —- x 100%
                                       X
            X = Mean value for the seven replicate determinations

            o = Standard deviation for the seven replicates
After  demonstrating  suitable performance  for  the  analysis  of  a  clean
matrix, performance testing  of more challenging matrices  of  interest to
the Program Office  is conducted.  Most  current SW-846 methods provide data
on  three   different  types  of RCRA  matrices.     Those  samples  may  be
characterized  reference  materials or  spiked  matrices  containing  known
amounts of target  analytes.   Analyses of  these  samples  also  provides
information on the  effects of matrix interference.  However,  it would be
nearly impossible to generate data on all matrix/analyte  interactions.  As
a  result,   Chapter  1  of SW-846  provides  guidance  on evaluating  the
suitability of methods for specific samples.
WORKGROUP REVIEW

Once methods are  evaluated, they are submitted for approval by the Program
Office.    In  the  case  of  OSW,  workgroups  meet   to  discuss  organic,
inorganic, and miscellaneous methods as well as quality assurance issues.
Workgroup members are provided methods in SW-846 format, which includes:

•Scope and  Application  including  analytes  and matrices  for which  the
      method is recommended
•a method summary (scope and principles)
•a list of required apparatus  and reagents
•instructions for the preparation of standards
•sufficient detail and clarity in any sample preparation procedures
•detailed analytical procedure
•Calibration procedure
                                    243

-------
•Method interferences
•calculation procedures

The workgroup is also provided performance data, including:

•the accuracy and precision of recoveries
•detection and quantitation limits
•Example output
•Analytical QC data and control charts
•results of confirmatory analysis
•Results of ruggedness testing, if performed
•Discussion of selectivity and sensitivity of the method
•Discussion of limitations of the method

The workgroup  reviews  the method narrative and the performance review.
They have  the  options  of approving the method,  rejecting  the method or
requesting additional information.
CONCLUSION

The process  of  method development and evaluation does not  seem to be a
growing  activity  in  our  industry.    This  is unfortunate  given  the
analytical challenges  that  we  face in producing scientifically credible
data that is useful to decision makers.   Chemists need to educate clients
about what is the most appropriate type of analysis.  Regulators need to
show greater flexibility  in allowing alternates to standard methods.
                                   244

-------
35         RECYCLING PROGRAM FOR LABORATORY ORGANIC SOLVENTS
                              Methylene Chloride and Freon-113

       Cornelius A. Valkenburg, Senior Scientist,  and William T. Brown, Laboratory Manager,
       Energy Laboratories Inc.,  1107 South Broadway, Billings, Montana  59107

       ABSTRACT

       Due to hazardous waste disposal costs and associated environmental concerns, programs for
       hazardous waste reduction and recycling of laboratory chemicals are of increasing importance.
       A  large fraction  of  the hazardous wastes generated  by environmental testing laboratories
       originate from the solvents used to extract  organic hazardous constituents from aqueous, soil,
       and waste samples.  These  solvents include methylene  chloride  (MeCI2),  MeCI2/acetone
       mixtures,  diethyl-ether, hexane,  and Freon-113.  Glassware rinsing  and  chromatographic
       extract clean-up  techniques such as gel permeation chromatography  also produce significant
       quantities  of waste solvents.

       Minimization of  solvent use is attained by selection/recommendation of the EPA  method,
       methods development and modification, and instrument and glassware selections.  Modified
       procedures must meet or exceed all  Quality Assurance/Quality Control (QA/QC) parameters
       specified by the method. Continuous liquid-liquid extraction techniques for EPA method 8270
       are currently developed to use less than 200 mL of MeCI2 per sample.  Similar solvent use
       reductions  have been  applied to pesticide/PCB and herbicide  extractions.   Solid  phase
       extraction  techniques such as those developed for the USEPA drinking water program use
       much  less  solvent and are recommended to clients when appropriate to the sample matrix and
       the desired  analytical  objective.   A  solvent  recycling  program  can  minimize  solvent
       consumption by processing waste solvents for re-use within the laboratory. An on-site solvent
       recycling program using spinning band distillation to  purify waste solvents is presented for
       MeCI2 and Freon-113.  Detailed work with MeCI2 and preliminary results for Freon-113 are
       discussed.

       Our solvent recycling program begins with  the recovery, collection, and segregation of waste
       solvents.  Recovered solvents are segregated according to the type of laboratory  operation.
       Recovered MeCI2 is  divided into GPC waste, extractor waste, general waste, and recovered
       condensate from solvent concentrations.  Each type of waste receives different handling and
       processing prior to purification by spinning band distillation.

       Purification of MeCI2 by spinning band distillation produces 3 or more fractions. Handling and
       quality testing of each collected fraction  are determined by the original segregated  waste
       source.   Quality testing  includes specific gravity,  GC/FID, GC/MS,  GC/ECD, and  FT/IR.
       Recycled solvent is  returned to suitable  laboratory operations based on testing results.
       Recycled MeCI2 is demonstrated to be suitable for re-use in sample preparation for semivolatile
       analysis by EPA  Method 8270.  Estimates of  labor and material  costs demonstrate the cost
       effectiveness of the recycling  program.

       INTRODUCTION

       In  EPA pollution prevention strategies, source reduction is the preferred method for reducing
       waste.  However,  for  those  wastes that cannot be reduced  at the source, recycling is
       considered the next  best alternative.
                                                245

-------
A large portion of the hazardous wastes generated by analytical laboratories orgiginates from
use of organic solvents in laboratory  operations.   Methylene  chloride,  Freon-113, methyl
alcohol,  hexane,  acetone,  and  xylene,  constitute  the bulk  of the  solvents  used  by
environmental testing laboratories.

Laboratory procedures which generate organic solvent waste include: glassware cleaning and
rinsing, sample extraction  and preparation, and liquid chromatographic techniques used  for
extract clean-up and/or analysis. Collected condensate from evaporative extract concentration
techniques are a  source of additional waste solvent. Many laboratories vent  solvent vapors
from  extract  concentration  procedures  directly to the atmosphere.   Recycling solvents
encourages the condensation and recovery of these  solvent emissions.

Land disposal of solvents was banned by the EPA in  1986. Laboratories typically handle their
solvent wastes by off-site incineration. Halogenated solvent wastes are normally segregated
from non-halogenated as they require more costly incineration at facilities specifically designed
to handle these materials.

Laboratory solvent wastes  are well suited for recycling programs using distillation technology.
Solvent recycling can be performed  either on-site or off-site.  On-site recycling refers to the
laboratory performing  all aspects of the  recycling  program.  Off-site recycling is recycling
performed by a  commercial solvent recycling facility.  After off-site recycling,  processed
material is either returned to the original generator using a custom or toll processing, or is sold
on the open  market.   In either  on-site or  off-site  recycling  programs,  appropriate  solvent
collection and segregation, is required in order to produce optimum recycled solvent quality.

Advantages to recycling on-site  include:  generally  favorable economics; more control over
waste  management  (feedback  loops  for   appropriate  solvent  segregation);   reduced
transportation risks; and reduced reporting requirements (currently RCRA treatment permits are
not required to recycle on-site).  Analytical testing laboratories  are well suited to an on-site
recycling  program.  They can perform the quality testing  necessary for their recycled solvents
and can utilize the recycled material appropriately, based on the quality testing performance
data.  Analytical testing laboratories also should have the technical  expertise to handle these
hazardous wastes.  Disadvantages to on-site recycling include the initial capital costs, safety
hazards associated with handling solvent wastes,  permitting requirements imposed by some
states, and the additional labor, technical, and  managerial concerns.

Advantages to off-site  recycling are the reduced technical and  managerial demands, and  in
limited cases-improved economics for small solvent  volumes. Currently, economical  off-site
recycling  of laboratory  solvent waste is limited to certain geographic areas primarily due to
hazardous waste transportation costs. Off-site  recycling  commonly  recycles non-halogenated
solvents (flammables) as a fuel source.

Energy Laboratories Inc (ELI)., of Billings, Montana is committed to reducing the generation of
hazardous wastes and the consumption of solvents. This paper describes the  ELI recycling
program which is fully operational for methylene chloride, the solvent used in largest volumes.
Preliminary results with Freon-113 recycling efforts are also discussed.
                                           246

-------
EXPERIMENTAL

Materials

Table 1 provides  a list of materials and their uses in the recycling laboratory.

Methods

MeCI2 Waste Processing

The handling of methylene chloride waste streams is illustrated in Figure 1. The MeCI2 waste
stream processing flow chart lists the procedures associated with recycling four categories of
MeCI2 waste.  Throughout the solvent collection and recycling process, segregation of waste
and distillate fractions is maintained.

Methylene chloride wastes are segregated according to the types of contaminants expected.
Wastes with percent levels of other solvent contaminations are handled separately. Generally,
higher boiling  trace  contaminants  present  few  problems  for  purification  by distillation.
However, waste  with high levels of higher boiling contaminants,  such  as phthalates, may
require  pre-distillation,  or  double-distillation  in the  high-purity distilling  system.   Solvent
contaminated  with  high levels of low boiling compounds (other solvents), particularly those
with similar boiling points, or which form azeotropes with MeCI2, require specialized handling.
Methyl alcohol and water are known to form  azeotropes with MeCI2.

Phase separation is performed by decanting or by transfer using separatory funnels.  Bulk
filtration (if required) uses sediment settling, decanting, and vacuum filtration through 10 inch
diameter Whatman #4 filter paper.   Pre-distillation  is used  to remove  non-volatile soluble
residues.  Pre-distillation is performed with custom manufactured glassware which allows the
teflon spinning band  distilling column of the ABC Integrity 2000 system to be by-passed.
Automation of pre-distillation is achieved by using the electronic control systems of the ABC
Integrity 2000. Solvent drying is done by mixing calcium chloride pellets  (5-10 g/L) with the
pre-distilled waste  solvent  for 24  hours.   In-line vacuum solvent filtration (see Table  1,
Materials list) is used to separate the  drying agent from the solvent.  In-line filtration is favored
since solvent vapor emmisions are controlled and recoverable.

Distillation is  performed using the ABC Integrity 2000  spinning  band  distillation  system.
System operating parameters are per the manufacture's recommendations with modifications
made relative to waste characteristics.  The system  is operated with manual switching
between the collection of the forecut and the main fraction.  Handling  of forecut and still-
bottom fractions  is  discussed later.

Collect fractions are preserved with cyclohexene to a level of 0.01% (v/v). The preservative
prevents the accumulation of hydrochloric acid  (HCI) in the methylene chloride.  Cyclohexene
or amylene are commonly  used for this purpose.  These compounds do not prevent the
formation of acid  in methylene chloride,  but undergo an addition reaction  with HCI across the
double bond.  Proper solvent storage minimizes decomposition.

GC/FID, GC/MS, and other screening procedures, are used to  determine the collected fraction
quality.  Based on screening results, the material is segregated for re-distillation or appropriate
use.  Segregated  material is composited  with material of similar quality to generate larger lots
of material.  Composited material is further tested and distributed for appropriate use.
                                           247

-------
Freon-113 Waste Processing

A flowchart for recycling Freon-113 Waste Streams is shown in Figure 2.  The procedures for
handling  Freon-113 are comparable to those for methylene chloride.  Recycled  Freon-113
quality is characterized according to intended use. Recycled IR grade Freon-113 (IRF113) has
no significant absorbance at the C-H stretching wavelength 2950 (cm"1) used to quantitate
Total Petroleum Hydrocarbons by IR as per EPA Method 418.1.  EPA method 413.2 for oil and
grease by IR also has similar requirements. Trace levels of contaminants may be acceptable
in IRF113 since quantitation by IR is performed by measuring the differences between the
absorbance of the extracted sample against the reference solvent.  Levels of acceptable purity
have not yet been determined. Non-IR grade recycled Freon-113 (NIRF113) is  contaminated
with part per hundred or  less (determined  by GC/FID) of volatile solvents and  exhibits a
significant interfering  IR absorbance.

Complete segregation of recyclable  IRF113 from recyclable  NIRF113 is critical.  Certain
compounds/solvents  cannot  be  removed effectively from   Freon-113  by  spinning-band
distillation.  These compounds (yet to be identified) either azeotrope  with, or have similar
boiling points to Freon-113. These co-distilling compounds  are considered contaminates when
their concentration  and/or interfering  absorbance  is significant.   Co-distilling/interfering
compounds may persist regardless of the number of re-distillations, and, if mixed with IRF113
would  irreversibly contaminate the  IRF113.  Unleaded gasoline  is known to contain non-
removable compounds of this type.

NIRF113 is suitable for use in determining oil and grease by gravimetry since the contaminants
are determined by the mass of extracted residues after evaporation.  NIRF113 has insignificant
residue after drying. Evaporated solvent condensate collected from procedures using NIRF113
is suitable for recycling only as NIRF113.

Distillate Fractions Handling

Handling of distillate fractions for MeCI2 and Freon-113 is shown in Figure 3. Segregation of
all materials is  maintained  by  batch and  fraction  type.  Pre-distillations are performed
automatically using the modified ABC Integrity 2000 system. Distillations are performed using
the ABC  Integrity 2000 teflon spinning band distillation system. This distillation system has
been tested by the manufacturer to have 50 theoretical plates or more with cyclohexane and
methylcyclohexane.

High contaminant levels, regardless of boiling point, can effect the quality of any separation.
Experience has shown that increasing the volume of the forecut and still-bottom fractions can
improve the quality of the collected fraction. The number of collect fractions can be subdivided
manually if desired.

The amount of forecut  collected  varies according  to the waste type  and system operating
conditions. Azeotropes and lighter boiling materials are collected in the forecut. Water forms
azeotropes with both MeCI2 and Freon-113, and is always collected  in the forecut. The solvent
drying  step prior to distillation reduces the amount of forecut needed to remove trace water
from a  Freon-113 or MeCI2 solvent waste. Often, free water is observed as an immiscible layer
in the initial portion of the forecut.  Head temperature  and observation of moisture condensate
at the head of the distilling column are used to  determine the presence or absence of water.
Forecut is not complete until all moisture is removed. Collected forecuts should be dried and
re-distilled with other similar materials.  Due to  possible percent levels of other solvents in a
                                          248

-------
re-distilled forecut, recycled forecut may be suitable only for glassware washing and  other
similar laboratory operations. Specialized handling techniques can be used to remove certain
solvents prior to re-distillation.

Depending on the quality of starting material, still-bottoms (raffinate) from pre-distillations and
distillations are either disposed of as hazardous waste or are retained and combined for further
recycling  with  other similar still-bottom  material.   The amount of forecut collected when
reprocessing them should be minimal.

QA/QC Procedures

Testing procedures are selected  to characterize material suitability  based on intended use of
the recycled  product.  Recycled  MeCI2 and MeCI2 waste solvents were prepared for GC/MS
analysis using the solvent concentration procedure normally used for EPA method 3520/8270
analysis.  J.T.  Baker "Capillary  Analyzed"  MeCI2 is used  as the control solvent.   Recycled
solvents are concentrated from 200 ml to a 1 mL final volume in a Turbovap prior to analysis.
Solvent concentrates are analyzed according to EPA method 8270 on a Hewlett Packard 5970
MSD.  The selected list of 8270  analytes tested for represents most of the analytes from the
CLP Semivolatile Target Compound List. A five point calibration ranging from 20 //g/ml to 150
/yg/mL  was  used.  Internal  standards fortification levels  were 40 fjg/ml.    The  gas
chromatograph was operated in  the splitless injection mode with a 0.5 /A. injection on  a 30
meter 0.25 mm i.d., 25 //m film thickness, XTI-5 capillary column.  The GC was temperature
programmed: 40° C hold 4 minutes, 20° C/min. to  120° C., hold  1 minute, 10° C/min to
250° C hold 2  minutes, 10° C/min to 310° C  hold 6 minutes.

Additional QA/QC testing for MeCI2 such  as density, acidity, solvent purity by GC/FID and
GC/ECD (concentrated  then hexane solvent exchanged) are performed, but not reported here.
Freon-113 is tested for  impurities by  GC/FID, and also by Infrared (IR) according to  EPA
Method 418.1.

RESULTS AND  DISCUSSION

Analytical data  which demonstrate the  effectiveness of recycling by distillation are shown in
Figures 4 through 8. Chromatograms of solvent method blanks for these analyses are given in
Figures 7  and  8.   The  material presented illustrates  the results from  recycling a highly
contaminated MeCI2 solvent waste. Figures 4, 5, 6, 7,  and 8 are EPA method 8270 GC/MS
reconstructed ion chromatograms (RIC) of the primary starting material, first distillation, second
distillation, solvent method blank for Figures 4  and 5, and solvent method blank for Figure 6,
respectively.   Individual RICs are normalized to the largest peak present. Internal  standards
are at the same level in all RICs presented (40 ^g/ml) and can be used to compare  results
between  the RICs  shown.  Internal standards  were added  after  concentration  and  thus
concentration factors need  to be accounted  for.  Target and  non-target analyte  GG/MS
identifications and quantitations are discussed during the presentation.

Figure  4 is the  RIC of  the original starting  material  concentrated 20:1.   The source  of this
material was  general laboratory MeCI2 waste.  This waste required phase separation and bulk
filtration to remove water and sediments.  The resulting filtered product  had a strong  yellow
tint and smelled of gasoline.  The filtered material was pre-distilled to  remove soluble  non-
volatile  residues.   The  1.5  L  of still-bottom from the pre-distillation  was disposed of as
hazardous waste.
                                          249

-------
Figure 5 is the RIC of the collected fraction of the first distillation. The collected material was
concentrated 200:1  for analysis by GC/MS. Total batch size was 19 L to give: 2 L of forecut;
15.5 L of collected fraction; and 1.5 L of still-bottom.  No analytes on the EPA method 8270
list were detected above method reporting limits. However, the material was not considered
suitable for use in EPA Method 8270 or other laboratory operations due a series of early eluting
peaks observed in the RIC. The collected fraction  was re-distilled  and the forecut and still-
bottom were recycled.

Figure 6 is the RIC  of a second distillation which contained 15 L of the material shown in
Figure 6, and 4 L of  a similar material also  requiring a second distillation.  The distillate
collected was concentrated 200:1 for GC/MS analysis.  The distillation produced: 1.8 L of
forecut;  13.5 L of  collect; and 3.5  L of still-bottom.  The forecut and  still-bottom were
recycled.  The collected material is considered suitable for use in EPA Method 8270 and was
also determined to be suitable for  use in EPA method 8080 (data not shown).

Figures 7 and 8 are the solvent method blanks associated with the solvent concentration and
analysis.  Differences in GC/MS instrument sensitivity and  the chromatographic peak shapes
reflect analyses performed over two different time periods.

SUMMARY

Solvent  recycling for  analytical  testing  laboratories  is  environmentally  responsible and
economically viable.  Recycling  not  only reduces  disposal costs, but provides a reusable
product at potentially less cost than the virgin solvent. Estimates of labor and material costs
are given during the  presentation.

The material presented here evaluates the potential for solvent recycling using purification  by
spinning band distillation.  The purity of the recycled MeCI2 solvent  is evaluated relative to
probable intended use. Quality assurance and quality control requirements for EPA method
8270, and other similar procedures,  specify that target  analyte cannot exist in the method
blanks at concentrations above the  reporting  limits.   Therefore,  solvents  used in sample
preparations  for EPA Method  8270 must be free of target analytes and interferences  after
equivalent solvent concentration techniques. American Chemical Society (ACS) specifications
for solvent purity are also recommended guidelines for determining solvent quality.  The quality
of the recycled MeCI2  product produced in the second distillation discussed here met these
requirements. Some recycled materials not suited for use in sample extraction may be suitable
for use in other laboratory operations  such as glassware cleaning and rinsing.

Methylene  chloride is particularly  well suited for recycling  by distillation since it is primarily
used  in  analysis procedures  for  semivolatile compounds.  Trace  levels  of other solvent
impurities do  not interfere in the analysis of semivolatile compounds.  EPA Method 418.1 using
Freon-113  extraction and  analysis by IR is affected by trace levels of solvent impurities.
Because  trace levels of other solvents can interfere in IR analysis, recycling by distillation is
more difficult and requires very careful waste solvent segregation.
                                           250

-------
ACKNOWLEDGEMENTS

The authors would like to thank the following people for their contributions:  Gary Warren,
Jack Hanewald, Jim Rasmussen, TJ Parker, Dan Dalbey, Bruce Williams, John Standish, and
other chemists and staff of Energy Laboratories for their support and on-going cooperation in
solvent collection, segregation, and testing;  Jim Stunkel, Kevin Kelly, and Mark Orr of ABC
laboratories  for their  technical  assistance;  Judy Gebhart  of  ICF-QATS for  her editorial
suggestions; and Roger Rourk Jr. and Paul Van Trieste of B/R Instruments for the pre-distilling
glassware.
                                        251

-------
                                     TABLE 1
  SOLVENT RECYCLING LABORATORY - MATERIALS and EQUIPMENT LIST
               FACILITIES
Solvent preparation area
Distilling apparatus area
Fume Hood or equivalent vapor control system
Bulk solvent storage area
Two, 110 volt 20 amp circuits
Deep well sinks

             APPARATUS
ABC Integrity 2000 automated spinning band
distillation system - 19 Liter option
Custom Glassware to fit ABC Integrity 2000
19 Liter pot flask (ABC)
2 Liter separatory funnels
5 inch powder funnels (glass)
Oilless solvent resistant vacuum pump
   (Cole Farmer teflon coated diaphragm pump)
   Cryotrap
    Ice bath
4 Liter vacuum flask
    10 inch diameter buchner funnel
Whatman #4 filter paper, 24 cm
Alltech In-line filter degasser
    Whatman #41 filter paper, 4.7cm
Anhydrous calcium chloride pellets
Cyclohexene
Hydrometers
Solvent storage containers
  (Clean empty 4 Liter amber glass solvent jugs)
  (Clean empty  100# Freon-113 Cans)
Sampling containers
  (40 ml VOA vials, 4 Liter amber glass jugs)
Hazardous waste storage containers
Analytical testing laboratory
    MISCELLANEOUS APPARATUS
Aluminum foil
Tools
Teflon tape
Nitrogen gas supply (clean and dry)
Drum  cart
Cleaning brushes
Glassware cleaning detergents
1:1 Sulfuric acid
Methyl alcohol
Transfer pipets (5, 15 and 25 mL)
Safety equipment
Teflon tubing
Teflon squeeze bottle
Rubber stoppers
             FUNCTION
Site for solvent handling
Site for distillations
Solvent vapor removal
Storage of solvents
System power
Glassware washing

             FUNCTION
Distillation and Pre-distillation

Pre-distillation
Pre-distillation
Phase separation
Solvent transferring
Vacuum filtration and solvent transferring

Vacuum filtration vapor recovery
Vacuum filtration vapor recovery
Bulk filtration
Bulk filtration
Bulk vacuum filtration
In-line vacuum filtration and transferring
In-line vacuum solvent filtration
Solvent drying
Methylene chloride solvent preservative
Density determinations
Solvent fraction storage
Sampling

Storage of un-recyclable solvent wastes
Materials testing
             FUNCTION
Clean work surface, vapor control
Setup and maintenance
Fittings and sealing solvent containers
Evacuating solvent vapors from containers
Drum handling
Glassware cleaning
Glassware cleaning
Glassware cleaning
Glassware cleaning
Sample collection
Safety
Solvent transfer
Solvent handling
Vacuum system connections
                                       252

-------
                         Figure 1
METHYLENE  CHLORIDE  WASTE STREAMS
                Processing Flowchart
1 GENERAL &
EXTRACTOR
Contaminants include:
Solvents
Semivolatiles
Water
Particulates
1

1

SEGREGATION
1
PHASE SEPARATION
|
BULK FILTRATION
1
PRE-DISTILL
*1) Collect Fraction
2) Still Bottom
|
SOLVENT DRYING
I
I

WASTE CATEGORY
GPC
Contaminants include:
Semivolatiles
Trace Solvents





| SEGREGATION





PRE-DISTILL
(as indicated)


SOLVENT DRYING
(as indicated)

-» Pll TPATinM ,-.
CONDENSATES FROM
TURBOVAP or
KUDERNA DANISH
Contaminants include:
Solvents
Methyl Alcohol
Hexane
Acetone
Water

| TESTING

SEGREGATION J

| PHASE SEPARATION






L SOLVENT DRYING



                        DISTILLATION
                      1) Forecut
                      *2) Collect Fraction
                      3) Still Bottom
                       PRESERVATION
                      SCREENING TESTS
                          (pass )
(fail)
                       SEGREGATION
                       COMPOSITING
                     COMPOSITE TESTING
                        DISTRIBUTION
•Collect Fraction is processed towards final product. See Figure 3 "Distillations" for
 handling procedures of other fractions.
                               253

-------
                           Figure 2
           FREON-113  WASTE  STREAMS
                  Processing  Flowchart
     (NIRF113)
  NON-IR GRADE
WASTE CATEGORY

 NON-IR or IR GRADE
     (IRF113)
    IR GRADE
Contaminants include:
   >2000 ppm TPH
Unleaded gasoline
Recyled Non-IR Freon
       Water
    Participates
        T
 Contaminants include:
  200-2000 ppm TPH
 Diesel Range Organics
       Water
     Particulates
                            TESTING
                          SEGREGATION   |
                               I
                       PHASE SEPARATION

                         BULK FILTRATION
                          PRE-DISTILL
                         *1) Collect Fraction
                         2) Still Bottom
                        SOLVENT DRYING
                           FILTRATION
                            TESTING
                          DISTILLATION
                          1) Forecut
                         *2) Collect Fraction
                          3) Still Bottom
                            TESTING
                            (Pass)
                          SEGREGATION
                          COMPOSITING
                       COMPOSITE TESTING
Contaminants include:
   <200 ppm TPH
       Water
    Particulates
        I
                       (Fail)
                      |    DISTRIBUTION   |


 ''Collect Fraction is processed towards final product. See Figure 3 "Distillations" for
 handling procedures of other fractions.
                                 254

-------
in
01
                                                                 Figure 3
                                                          DISTILLATIONS
                                                           FRACTIONS HANDUNG
i. PRE-DISTILLATrqN]

          I	
          I	,    ,	I	,
   COLLECT (80-95%)  |    fSTILL-BOTTOM (5-20%]
  (Handle as per solvent               |
  processing flow chart)               |	
          I              	I	,        I
                              FILTRATION    H  I    WASTE
                          	I	,
                           PRE-DISTILLATION |
          I
    DISTILLATION
          !          ,
          i
          i
  	I	,    ,__
   COLLECT (60-80%)  |    |  FORECUT (10-20%) |                  | STILL-BOTTOM (10-20%)
  (Handle as per solvent               |                                       |
  processing flow chart)               |	                      j	
                                         PHASE SEPERATE
                                                 WASTE
                                         SOLVENT TESTING
                                          DRYING AGENT
                                            FILTRATION
FILTRATION
WASTE
                                                                 PRE-DISTILLATION
                                           DISTILLATION

-------
to
s
            Figure 4.  Reconstructed Total Ion Chromatogram of General Methylene Chloride Waste concentrated 20 : 1 and analyzed per

                     EPA Method 8270. Internal Standards at 40ug/L

-------
ro
tn
6.4-
6.2:
t.O-
5.8-
5.6-
5.4-
5.2-
5.0-
4.8-
4.6-
4.4-
4.2-
4.0-
3.9-
3.6-
33.4-
S3.2-
X
^3.0-
2.8-
2.6:
2.4:
2.2:
2.0:
1.6:
1.4-
1.2:
1.0:
0.8-
0.6-
0.4-
0.0 ,
4






















i
li






















t
1
1






















L
i























I
6






















i
U
f
























B
S. g ^
O> LJ U1
5 ^ o
1 xx r~t
£ ™ I
1 1 I
I i ,J
-




















Wj [__





















\1




















1

J W ^
co c^' J5
o
-p
§
1





















V
I ?
1












1





,
1

	 ^.]
10 12 14 16 18 20 22 24 26 28 30
Hln
                                                                                                             32
                                                                                                                   34
                                                                                                                         36
                                                                                                                                38
           Figure 5.  Reconstructed Total Ion Chromatogram from first distillation of General Methylene Chloride Waste concentrated 200:1

                     and analyzed per EPA Method 8270.  Internal standards at 40ug/L

-------
                  1.6


                  1.5
to
01
09
 1.3-


 1.2-


 1.1


 1.0


 0.9


So.8-
X
*s
>-



 0.6-


 0.5


 0.4


 0.3


  0.2


  0.1
                    0.0-,
                       4
                        10
                               12
                                                                      00
                                            16
                                                  18
20
Hin
                                                               22
                                                                     24
          Figure 6. Reconstructed Total !on Chtomatograms from second distillation of General Methylene Chloride Waste concentrated
                    200:1 and analyzed per EPA Method 8270.  Internal standards at 40ug/L

-------
8.8-
8.5-
8.2-;
8.0-
7.8-
7.5-
7.2-
7.0-
6.8-
6.5-;
6.2-:
6.0-!
5.8-:
5.5-:
5.2-i
5.0-:
4.8-:
iN
34.2-:
^4.0-:
3.8-1
3.5-:
3.2^
J.o-i
2.8-1
2.5-:
2.2-;
2.0^
1.8-:
1.5-i
1.2-!
i.o-:
O.B-:
0.5-
0.2-
-S -*
ro ^ « i-
^ ^ tn
•^f r^. ^
•p ^ "^
1 i' 1
I 1 ]
i \ T
r-t















F





































L..JI .. I
cj in
•^) OJ
(T» hO
T3J \fl
s
ID
fe
t-
>-i
1
























r-j
a
TJ
k
c
j>
i























\i .. —
A Q i/ * I, i^ li.J»t..^.LL | | , M — r 	 i. 	 . — - - - — 	
4 6 8 10 12 14 16 18 20 22 24 26 28 30
Hln
                                                                                        32
                                                                                                    36
                                                                                                          38
Figure 7.  Reconstructed Total Ion Chromatograms of Solvent Method Blank for Figures 4 and 5.
         Solvent concentrated 200:1 and analyzed per EPA Method 8270.  Internal standards at 40ug/L

-------
    2.4-i
    2.3i
    2.2-:
    2.1-i
    2.0-i

    1.8-;
    1.7-i
    1.6-:
    1.5-;
    1.4-
     i.o-:
     0.9-:
     0.8-;
     0.7-i
     0.6H
     0.5-:
     0.4-;
     0.3-i
     0.2-i
     0.1-'
                                          14
                                                16
                                                       18
                                                              20
                                                              Hln
                                                                    22
                                                                           24
                                                                                                                    36
Figure 8.  Reconstructed Total Ion Chromatograms of Solvent Method Blank for Figure 6.
          Solvent concentrated 200:1 and analyzed per EPA Method 8270.  Internal standards at 40ug/L

-------
3R   The Technology and Performance  of Several New  Immunoassay Methods  and the
      RISc 2000 Instrumentation  System.

      Stephen B. Friedman, James P. Mapes, Randy L. Allen, Thomas N.  Stewart, William B.
      Studebaker, Patricia P  McDonald,  Richard  E. Almond,  Tracy A. Withers, Shannon P.
      Arrowood,  and Darrel P. Johnson

      EnSys Inc.
      Data generation is often  the  rate-limiting  element that influences  the  overall
      efficiency of the environmental  clean-up process.  New immunoassay methods are
      providing a faster and  less expensive way  to detect the pathogenic  compounds  that
      are present  in the environment.

      We have developed a number of immunoassay  methods that expedite site mapping,
      remediation  and  monitoring  activities by  providing reliable  results immediately
      and for less cost.   The immunoassay methods  for Pentachlorophenol, PCB's and
      Total  Petroleum  Hydrocarbon detection have been previously reviewed and have
      been accepted for use by the EPA.

      Additional methods, and a RISc 2000 instrumentation system, have  been developed
      and evaluated.  These  new methods will '.steel  PAH's,  explosives \,.e. TNT, ^Dx),
      and petroleum contamination  in a variety of matrixes.  The PAH  and Petroleum
      Contamination tests use monoclonal antibody reagents  in a chromogenic enzyme
      immunoassay format.   The tests  for explosives use the chromogenic  chemistry  of
      the recently accepted EPA method  8510.   All  of the tests permit the simultaneous
      testing  of multiple  samples at multiple detection levels  and provide results
      within  30 minutes.  The PAH,  Explosives  and  Petroleum Test for Water all retain
      our conservative approach to environmental testing  and have  demonstrated a  >95%
      confidence  in  their ability to detect contamination  when  used in accordance with
      the recommended  protocols.

      The EnSys RISc 2000  Instrumentation System  has been developed to support  the
      expanding  library  of immunoassay  methods that are currently available or that
      will be introduced.   The  instrument rapidly analyzes the kinetics of the
      immunoassay's chromogenic  reaction and  uses this data in the  calculation of
      results.  The RISc 2000  instrument will read and record sample  results at the  rate
      of one every  10 seconds.  System characteristics include faster results, additional
      quality  control,  and a  printed  report that summarizes the data obtained and  the
      interpretation of results.   This caper will  review the  applications,  chemistry and
      performance characteristics of the  methods and  instrumentation developed.
                                          261

-------
37           SCREENING OF TCLP EXTRACTS OF SOIL AND WASTEWATER
                              FOR 2,4-D BY IMMUNOASSAY

      M.C. Haves. S.W. Jourdan, T.S. Lawruk and D.P. Herzog, OHMICRON, 375 Pheasant
      Run, Newtown, Pennsylvania 18940

      ABSTRACT

      A commercially available enzyme immunoassay  kit has been adapted to screen TCLP
      extracts of environmental samples for the regulatory action levels of the commonly used
      herbicide, 2,4-Dichlorophenoxyacetic Acid (2,4-D).   The immunoassay was  originally
      designed to quantitate very low concentrations of 2,4-D in water and includes calibrators
      at 1, 10 and 50 ppb.  A simple one-step  1:1000 dilution of a  TCLP extract brings the
      detection range of the assay in line with the 2,4-D toxicity characteristic established by the
      EPA at 10 ppm. The method which combines the standard TCLP extraction with dilution
      and subsequent analysis by  a magnetic particle-based ELISA has been evaluated for
      sensitivity, matrix effects, accuracy and susceptibility to interferences.  The results of the
      evaluation studies are described here.

      INTRODUCTION

      The Methods Section of the EPA Office of Solid  Waste has stated a need for'more rapid,
      less expensive field screening  procedures that  do  not  compromise the accuracy  of
      pollutant  evaluation.  Availability of reliable, sensitive, inexpensive screening of  large
      sample loads of potentially contaminated  samples coupled with subsequent quantitative
      analysis of positives by established instrumental methods is an approach that addresses this
      need.   The Methods Section of OSW has provided an outline of the type and quantity of
      work needed to demonstrate  whether a proposed  procedure is suitable for inclusion in
      SW-846 as a screening method.   Using this  outline as an  experimental  model,  a
      demonstration of the performance characteristics of an immunoassay screening method for
      2,4-dichlorophenoxyacetic acid (2,4-D) has been completed and is presented here.

      Currently approved SW-846  methods for analysis of 2,4-D in  soil  or water samples
      require the  performance of  difficult  solvent extractions and  chemical  derivatization
      reactions  by  highly trained chemists and the use of expensive analytical instruments
      (Methods #8150 and #8151).  The use of the Toxicity Characteristic Leaching Procedure
      (TCLP/Method 1311) as the method of choice for preparing extracts  of liquid  and  solid
      wastes for subsequent analysis of the original fourteen priority pollutants, including 2,4-D,
      heavily influenced  the  design  of  the study.    All  wastewaters and soils  used  in
      demonstrating the capability of the immunoassay method were  subjected to the Method
      1311 extraction procedure prior to analysis.  The 2,4-D RaPID Assay® kit, a magnetic
      particle-based  immunoassay for 2,4-D developed and  manufactured  by Ohmicron, has
      been commercially available since 1992 for the determination of part per billion levels of
                                              262

-------
the herbicide in water. This study shows that the immunoassay can be readily adapted for
use as a screening method for analysis of TCLP extracts for  the presence of 2,4-D at the
regulatory level of 10 ppm.

MATERIALS AND METHODS

Equipment and reagents:  2,4-D RaPID Assay kits, 12 x 75 mm disposable plastic test
tubes,  2,4-D  TCLP  Sample  Diluent, precision  pipets,  magnetic  separator,  RPA-I
photometer and vortex mixer are available from Ohmicron, Newtown, PA.  2,4-D powder
for gravimetric preparation of spiking stock solutions was obtained from Reidel-de Haen
(Hanover, Germany).  All stock solutions were prepared in acetonitrile (pesticide grade).
The accuracy of preparation of the stock solutions was assessed by comparison with an
EPA  certified standard solution  (5000 ug/mL in acetonitrile)  purchased  from NSI
Environmental Solutions, Research Triangle Park, NC.

Environmental Samples.  Several water samples were  studied for matrix effect. They
included municipal tap water from Newtown, PA and surface water  ("runoff") collected
from a stream that flows through eastern Pennsylvania farmland. Wastewaters ("effluent")
were collected from two separate drainage or discharge pipes in an agricultural community
in south central New York State.  Two soil types were evaluated:  a Sassafras sandy loam
(<5% organic matter) from New Jersey and a muck soil (46% organic  matter) also from
New Jersey.

Procedures.  TCLP extractions were performed on soils and wastewaters as described in
Method 1311. Spikes were made volumetrically into the final extract after filtration. The
detailed TCLP extract screening procedure is published and available from Ohmicron as an
"Application Procedure."   All immunoassay results are converted to ppm 2,4-D in the
waste extract for evaluation versus the regulatory level.  Detailed directions on use of the
immunoassay kit  are found in the package insert.

Reference Method.   Selected blank and spiked matrices were split and sent to a local
reference lab for analysis by SW-846 Method #8150 for chlorinated herbicides.

RESULTS AND DISCUSSION

Establishing a cutoff concentration for the screening procedure.  In order to  use the
quantitative results  produced by the immunoassay in a qualitative  way, a cutoff was
established that could reliably discriminate the regulatory action level (10 ppm) from one-
half that level (5  ppm). The TCLP extraction buffer (sodium acetate/acetic acid pH 4.9)
was diluted with sample diluent and spiked with 2,4-D  at 5, 7.5,  10 and  15 ppm. The
spiking stock solutions were verified for accuracy before use by comparison with an EPA
certified reference preparation.  Results of this comparison are given in Table 1.  Thirty
replicates of each solution were then tested in three immunoassay batch runs.  Figure 1
shows  that use of a  cutoff concentration of 7.5  ppm provides  perfect  discrimination
                                        263

-------
between 10 ppm and 5 ppm.  This cutoff concentration falls near the mid-point on the
calibration curve where precision is optimal.   In fact, the distribution of results observed
for a 7.5 ppm solution overlapped neither the 5 nor the 10 ppm distributions.  In a similar
manner the TCLP extract of a 2,4-D-free organic soil was spiked and tested  by the
immunoassay.  In the case of the soil extract, ten replicates were tested at each level of
interest in one assay run.  A similar pattern of results distribution was observed when a
TCLP extract of this muck type  soil was used  as the background matrix.   Figure  2
demonstrates no overlap of the regulatory level with half its concentration.  When the
definitions of "positive" as greater than 7.5 ppm and "negative" as less than 7.5 ppm are
applied to these data (Figure 3), the utility of the method as a screening procedure can be
seen.  A clear distinction is observed for the buffer and the soil extract matrices between 5
and 10 ppm.

Cross reactivity with chlorophenoxy compounds and other pesticides.  The compounds
listed in Table 2 were added to the immunoassay system until a response equal to 10 ppm
2,4-D was observed.  When an effect equivalent to 10 ppm 2,4-D could not be found at a
concentration of 5,000 ppm, cross-reactants were tested once at 10,000 ppm. Table  2
shows that most of the 2,4-D esters tested reacted with more potency in the assay than
2,4-D itself.  However, these  derivatized compounds are not ordinarily  significant for
environmental samples.  Many commonly used pesticides that are structurally unrelated to
2,4-D showed no reactivity in the screening procedure up to 10,000 ppm.  Structurally
related compounds of regulatory interest, 2,4,5-T and 2,4,5-TP (Silvex), reacted much
more weakly than 2,4-D.   Greater than ten times more 2,4,5-T and approximately 135
times more Silvex would be required to produce a positive result with the screening
method. In other words, a sample load of 130 ppm 2,4,5-T or 1375  ppm Silvex must be
present  to produce  a result  equivalent  to  a 10 ppm  concentration of  2,4-D.
Silvex was studied in greater detail because  it is considered a priority pollutant  and  a
regulatory action level for the TCLP Toxicity Characteristic has been published for this
compound.  Increasing amounts of Silvex were mixed with a 5  ppm 2,4-D solution in
order to evaluate the amount of cross-reacting Silvex required to produce a response
above the 7.5 ppm cutoff.  Results shown in Table 3  indicate that between 100 and 200
ppm of Silvex would be required to elevate a marginal 2,4-D sample concentration (5 ppm
spike) into the positive range.  The regulatory action level  for Silvex is 1 ppm.   This
concentration of Silvex by itself or in combination with negative or even borderline levels
of 2,4-D in the immunoassay will not influence the baseline 2,4-D result whatsoever.

Matrix specific performance of the screening method.  A variety of water and soil matrix
types were studied to determine their interference, if any, with the immunoassay screening
method. All matrices except the municipal water were first treated as solid or liquid waste
according to the procedures given in Method 1311 (TCLP Procedure). Prior to spiking,
each extract was diluted,  tested in the immunoassay and determined to contain a non-
detectable concentration of 2,4-D.  Aliquots of the extract were then spiked volumetrically
with 5, 10 and 15  ppm 2,4-D prior to  1:1000 dilution  and subsequent immunoassay.
                                         264

-------
Municipal drinking water was not extracted or filtered but was spiked directly with 35, 70
and  140 ppb 2,4-D prior  to  1:7 dilution and immunoassay.  This dilution brought the
drinking water concentrations into the same range of the calibration curve for analysis that
was  used for the waste extract testing.  These lower spiking levels were chosen because
the USEPA has proposed a MCLG of 2,4-D in drinking water of 70 ppb. Each of the 28
blank and spiked matrices were tested five times by the screening method.

Table 4 shows that the proposed screening method accurately discriminated negative and
positive regulatory levels in all matrices tested.  All results on matrices spiked at one-half
the regulatory level (5 ppm)  were less than 5.9 ppm 2,4-D showing no  false positives
(results >7.5 ppm). Results on matrices spiked at the 10 ppm regulatory level ranged from
a low of 10.2 ppm to a high of 14.5 ppm with no false negatives  (results <7.5 ppm).
Samples spiked above the  regulatory level at 15 ppm gave quantitative results from 15.5
to 19.1 ppm 2,4-D, again  showing no negatives.  The slight  positive  bias  in quantitative
results at the 10 and  15 ppm spike levels may be due to the fit of the calibration curve in
this concentration range or may have been introduced in the spiking or  dilution steps. The
nature of the matrix appears to have no influence on the results. The results on the spiked
drinking water demonstrates  that the 2,4-D immunoassay  can be applied  in a similar
manner (with a lesser dilution) to screening of drinking water matrices.  Although the five
replicates run here  do not  show  statistical discrimination of  the  concentrations  of
regulatory interest demonstrated above, the drinking water results were calculated from
the same portion of the 2,4-D calibration curve as for the waste extracts so the precision
of the  measurements should  be  identical  to  the buffer  system shown in  Figure  1.
Undiluted drinking water  from a variety  of sources has  been shown to  produce no
interference in the 2,4-D RaPID Assay (Lawruk, et al, 1993). Excellent discrimination of
negative from positive results on drinking water samples in  the 35 to 70 ppb range can
therefore be expected.

Correlation with a quantitative SW-846 reference method.   Selected samples from the
spiking study were submitted for analysis by the SW-846 Method  #8150 for chlorinated
herbicides.  The results of the analyses are given in Table 5.  100% method agreement was
seen at 1/2 the regulatory level and at 1.5 times the regulatory level - an important goal for
new screening methods  as  stated by EPA/OSW.  The quantitative immunoassay results
agreed with the spiked 2,4-D concentrations better than the GC results did. However, in
cases where the spike level  was  frankly positive (15  ppm) or negative (5  ppm), the
interpretations of the two methods were  identical.   The  qualitative status agreement
between the screening method  and  the reference method was more variable when the
spiking concentration was at or near the cutoff of both methods (10 ppm).  This was true
in part because we are using different criteria to make the status judgment (7.5 ppm cutoff
for the immunoassay vs. 10 ppm cutoff for the reference method).  In addition, the GC
method is expected to show a normal distribution of results around 10 ppm in repeated
analyses.  If the GC  result is used as the  criteria for correct assessment  of the sample
status, then only three results  (#7,  19 and 23) show opposite status interpretations.  In  all
                                        265

-------
three cases, the immunoassay would give a positive when GC was negative (<10 ppm).
This indicates that there may be a greater tendency toward false positive interpretation
than false negative interpretation when the screening method is evaluated in this way.  The
correlation of the quantitative results of the two methods on eleven spiked samples in the
5 to 15 ppm range  shown in Figure 4 confirms this observation.  Regression analysis of
this limited data set predicts a 10 ppm GC concentration would be read at about 13 ppm
by immunoassay.  The  apparent higher  concentration predicted by immunoassay could
result from the fit of the calibration curve in the 10 to 20 ppb range giving results that are
slightly higher than expected or lower results by the  reference method could be due to
some loss of analyte during  sample extraction, concentration and or derivatization steps.
A larger sample correlation study across a broader concentration range would provide a
better  characterization  of  the  correlation  of these two methods,  but  quantitative
comparison of methods was not an important objective of this study.  In another study
done with 56 water samples in the 10 to 500  part per billion range collected from various
locations across the United States, the 2,4-D RaPK) Assay was shown to correlate very
well (r = 0.970) with a GC/MS method (Lawruk et al, 1993).

SUMMARY

A magnetic particle based immunoassay for 2,4-D has been successfully adapted for use as
a screening method with TCLP extracts of some commonly encountered  forms of liquid
and solid waste. The method is inexpensive compared to GC methods, is easy to perform
and can provide results  on as many as 50 samples in less than an hour. The quantitative
accuracy of the  immunoassay results  in the ppm  range has been demonstrated by
comparison of selected  spiked samples with the currently  accepted SW-846 method. A
screening cutoff concentration has been established that discriminates between the 10 ppm
regulatory level and half that  level with no false negatives or false positives based on
spiking studies.  Cross-reactivity  with  environmentally relevant compounds has  been
shown to be insignificant and masking  agents are of little consequence  because of the
thousand fold dilution made prior to the immunoassay.  This test should now be applied to
field testing situations where it is expected to be a valuable tool for producing rapid results
for site mapping of contaminated soils, for effluent monitoring in agricultural applications
and in remediation activities.

REFERENCES

Lawruk, T.S., Hottenstein, C.S., Flecker, J.R., Hall, J.C., Herzog, D.P., Rubio, P.M.,
Quantitation of 2,4-D and Related Chlorophenoxy Herbicides by a Magnetic Particle-
Based ELISA, 1993, (manuscript submitted for publication).
                                        266

-------
Table 1. Calibration of spiking solution with EPA reference standard. Two hundred
fifty milligrams of 2,4-D (free acid) was dissolved in acetonitrile in a 25 mL volumetric
flask.  Aliquots from 0.5 mL to 3.5 mL were diluted in one liter volumetric flasks to
prepare solutions containing 5, 10,  15 and 35 ppm of 2,4-D. A 2,4-D EPA reference
standard (NSI Environmental Solutions, Research Triangle Park, NC) with a certified
concentration of 5000 jig/mL (in acetonitrile) was diluted volumetrically in a similar
manner in order to produce solutions at the same concentrations.  lOOjiL aliquots of each
solution were then diluted in 100 mL of TCLP Diluent prior to immunoassay.

                              concentration of 2.4-D (ppm)
source of 2.4-D                  predicted       measured            % difference

spiking solution                     5              5.05                  101
                                   10            11.33                  113
                                   15            17.40                  116
                                  35            37.30                  106

reference standard                   5              5.05                  101
                                   10            10.70                  107
                                   15            17.14                  114
                                  35            38.91                   111
                                     267

-------
TABLE 2. Cross-reactivity of chtorophenoxy compounds and structurally unrelated
pesticides in the 2,4-D RaPID Assay. The compounds shown below were prepared in
the TCLP Sample Diluent and added to the immunoassay in increasing amounts until a
result equal to that seen in a sample containing 10 ppm 2,4-D was seen.
Compound

2,4-D

2,4-D propylene glycol ester
2,4-D ethyl ester
2,4-D isopropyl ester
2,4-D methyl ester
2,4-D sec-butyl ester
2,4-D butyl ester
2,4-D butoxyethyl ester
2,4,5-T methyl ester
2,4-D isooctyl ester
2,4-D butoxy-propylene ester

2,4-DB
MCPA
2,4,5-T
Silvex methyl ester
4-chlorophenoxyacetic acid
MCPB
Silvex (2,4,5-TP)
Dichlorophenol
Dichlorprop

Triclopyr
MCPP
Mecoprop
Pentachlorophenol
Picloram

Alachlor
Aldicarb
Aldicarb sulfate
Aldicarb sulfoxide
Atrazine
Benomyl
Butylate
Captan
Captofol
Carbaryl
Carbofuran
Dicamba
1,3-Dichloropropene
Dinoseb
Metolachlor
Metribuzin
Simazine
Terbufos
Thiabendazol
   Concentration that
causes a 10 ppm 2.4-D result

      10 ppm

     0.52 ppm
     0.54 ppm
     0.96 ppm
      1.09 ppm
      1.40 ppm
      1.60 ppm
     2.00 ppm
      12.0 ppm
     20.0 ppm
     20.6 ppm

       95 ppm
      110 ppm
      130 ppm
      665 ppm
      815 ppm
      980 ppm
      1375 ppm
     2380 ppm
     5000 ppm

     > 10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm

     > 10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     > 10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     > 10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
     > 10,000 ppm
     >10,000 ppm
     >10,000 ppm
     >10,000 ppm
                                           268

-------
Table 3. Effect of silvex on 2,4-D results near the positive/negative cutoff point.
Solutions of silvex alone and silvex/2,4-D mixtures were prepared in TCLP buffer to
demonstrate the potential effect of a structurally similar, environmentally significant cross-
reactant on the immunoassay results. Silvex alone produced non-detectable 2,4-D results
in concentrations as high as 100 ppm. When increasing amounts of silvex
with a 5 ppm 2,4-D solution no
change in the status
of the result was seen
were mixed
until 200 ppm
of silvex was present in the mixture.

Silvex concentration 2.4-D
(ppm)
0
0.5
1.0
2.0
100
200
0
0.5
1.0
2.0
100
200
nd = non-detectable

concentration
(ppm)
0
0
0
0
0
0
5.0
5.0
5.0
5.0
5.0
5.0

2.4-D RaPID
Assay result
(ppm)
nd
nd
nd
nd
nd
0.8
4.46
4.99
4.48
5.03
7.09
9.17


Interpretation

negative
negative
negative
negative
negative
negative
negative
negative
negative
negative
negative
positive

269

-------
Table 4.  2,4-D spiking results on TCLP extracts of environmental waste matrices.
2,4-D was spiked into buffer, water or a TCLP extract of the matrices shown after the final filtration step. Each matrix was
diluted 1:1000 and tested by immunoassay five times. Each result was then compared with the 7.5 ppm cutoff to
determine if its status was positive or negative.
                                              2.4-D concentration bv Immunoassav
          Matrix/Spike
    1     TCLP buffer
    2     TCLP buffer + 15 ppm
    3     TCLP buffer + 10 ppm
    4     TCLP buffer + 5 ppm

    5     Sandy soil extract
    6     Sandy extract + 15 ppm
    7     Sandy extract + 10 ppm
    8     Sandy extract + 5 ppm

    9     Organic soil extract
    10     Organic extract+15 ppm
    11     Organic extract + 10 ppm
    12     Organic extract + 5 ppm

    13     Effluent #1
    14     Effluent #1 + 15 ppm
    15     Effluent #1 + 10 ppm
    16     Effluent #1 + 5 ppm

    17     Effluent #2
    18     Effluent #2 +15 ppm
    19     Effluent #2 + 10 ppm
    20     Effluent #2 + 5 ppm

    21     Runoff
    22     Runoff+15 ppm
    23     Runoff+10 ppm
    24     Runoff + 5 ppm

    25     Municipal water
    26     Municipal water + 140 ppb
    27     Municipal water +70 ppb
    28     Municipal water + 35 ppb
Determination #
cone, units
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppb
ppb
ppb
ppb
1
nd
15.5
10.6
5.0
nd
16.0
11.1
5.8
nd
16.1
11.2
5.2
nd
19.1
11.7
5.1
nd
16.3
12.1
5.1
nd
16.7
11.5
5.1
nd
162
75
38
2
nd
16.9
12.0
4.7
nd
15.8
12.4
5.9
nd
17.4
11.2
5.1
nd
18.2
12.6
5.7
nd
18.1
12.5
5.8
nd
19.0
11.2
5.4
nd
168
71
42
3
nd
16.3
10.2
5.0
nd
17.5
11.8
6.4
nd
16.3
11.5
5.0
nd
18.9
14.5
5.1
nd
17.0
10.4
4.5
nd
18.8
11.1
5.7
nd
155
75
37
4
nd
16.6
10.9
4.9
nd
16.5
10.8
4.9
nd
16.5
10.4
4.5
nd
17.9
11.1
4.8
nd
17.9
10.2
5.0
nd
17.5
11.8
5.2
nd
156
63
40
5
nd
17.9
10.4
4.3
nd
16.0
12.7
5.5
nd
17.5
11.0
5.1
nd
18.8
13
5.6
nd
15.8
12.6
4.9
nd
19.3
11.4
5.8
nd
165
78
33
mean
.
16.6
10.8
4.8
-
16.4
11.8
5.7
.
16.8
11.1
5.0

18.6
12.6
5.3
-
17.0
11.6
5.1
-
18.2
11.4
5.4
-
161
72
38
% POS
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
N/A
N/A
N/A
N/A
%Ntt
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
100
0
0
100
N/A
N/A
N/A
N/A
          N/A = not applicable to wastewater regulatory limit
                                                 270

-------
Table 5. Comparison of screening results with quantitative 2,4-D results obtained with SW-846 Method #8150.
Selected samples described in Table 4 were split and sent to a local reference lab for analysis. The mean of five immunoassay results and the
status of that mean with reference to the 7.5 ppm cutoff is shown. Quantitative GC results by method #8150 are given. Some samples were
submitted in duplicate. The status of the GC result is reported as positive if greater than or equal to the 10 ppm regulatory limit, negative if less
than 10 ppm. Agreement of the screening result with the GC result is assessed .
immunoassay
LD.#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Marrix/SDtke
TCLP buffer
TCLP buffer +15 ppm
TCLP buffer +10 ppm
TCLP buffer + 5 ppm
Sandy soil extract
Sandy extract +15 ppm
Sandy extract +10 ppm
Sandy extract + 5 ppm
Organic soil extract
Organic extract +15 ppm
Organic extract + 10 ppm
Organic extract + 5 ppm
Effluent #1
Effluent #1 + 15 ppm
Effluent #1 + 10 ppm
Effluent #1 +5 ppm
Effluent #2
Effluent #2 +15 ppm
Effluent #2 + 10 ppm
Effluent #2 + 5 ppm
Runoff
Runoff + 1 5 ppm
Runoff + 1 0 ppm
Runoff + 5 ppm
Municipal water
Municipal water + 140ppb
Municipal water +70 ppb
Municipal water + 35 ppb
cone, units
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppb
ppb
ppb
ppb
mean
nd
16.6
10.8
4.8
nd
16.4
11.8
5.7
nd
16.8
11.1
5.0
nd
18.6
12.6
5.3
nd
17.0
11.6
5.1
nd
18.2
11.4
5.4
nd
161
72
38
immunoassay
status
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
negative
positive
positive
negative
N/A
N/A
N/A
N/A
2,4-D result by
Method #8150
nd
13.0
11.0
5.6
nd
*
5.9,52
*
nd
*
10.0, 9.5
*
•
*
11.0, 7.8
3.6
*
11.0
8.8, 9.5
*
nd
*
9.7, 8.6
5.5
nd
*
58, 59
*

GC status
negative
positive
positive
negative
negative
•
neg, neg
*
negative
*
pos, neg
*
*
*
pos, neg
negative
*
positive
neg, neg
*
negative
*
neg, neg
neg
N/A
N/A
N/A
N/A
Agreement of
screen and GC">
yes
yes
yes
ye*
yes
•
no
•
yes
*
equivocal
*
•
•
equivocal
yea
*
yes
no
•
yes
•
no
V"
N/A
N/A
N/A
N/A
nd = non-detectable
N/A = not applicable to wastewater regulatory limit
• = no analysis with method #8150
                                                              271

-------
    15
                  Cutoff
0)
c
a
a:
_c

.3
 0)
 0)
    ">
            Negative
                       Positive
                                         1/XJ5.0 ppm Spike
                                         l\\|7.5 ppm Spike
                                         H010.0 ppm  Spike
                                              15.0 ppm  Spike
                                10
                                                    15
20
                           ppm 2,4—D by Immunoassay
Figure 1. Demonstration of the sensitivity of the selected cutoff for 2,4-D in the
TCLP buffer matrix. 2,4-D was spiked into TCLP extraction buffer at concentrations of
5, 7.5, 10 and 15 ppm. Each spiked solution was then diluted 1:1000 in Ohmicron 2,4-D
TCLP Sample Diluent. Thirty immunoassay determinations were then made on each
diluted solution. The frequency of occurence of the results is plotted against the range of
2,4-D concentrations measured.
                                     272

-------
    8
                 Cutoff
 CD
 cn
 c
 o
 V)
 0>
 O
=**=
    6  -
    4  -
    2  -
    0
5.0  ppm Spike
7.5  ppm Spike
10.0 ppm  Spike
15.0 ppm  Spike
                              10                  15
                         ppm  2,4—D by Immunoassay
                      20
Figure 2. Demonstration of the sensitivity of the selected cutoff for 2,4-D in the
TCLP extract of an organic soil. 2,4-D was spiked into the TCLP extract of an organic
soil at concentrations of 5, 7.5, 10 and 15 ppm. Each spiked solution was then diluted
1:1000 in Ohmicron 2,4-D TCLP Sample Diluent.  Ten immunoassay determinations were
made on each diluted solution.  The frequency of occurence of the results is plotted
against the range of 2,4-D concentrations measured.
                                    273

-------
 3  100

 E
 Q.
 Q.

in   80
 Q)

 O
 Q.
 E
 o
 o

 tn
 en
 0)
 0)
 tn
 O
Q_
60
40
     20
 0
O TCLP  Buffer

V TCLP  Extract of

    Organic Soil
                         5               10              15

                          ppm 2,4-D  Spiked into Matrix
                                                                     20
Figure 3. Break point of the 2,4-D screening test in buffer and soil extract. Spiked

buffer or soil extracts described in Figs. 1 and 2 were graded for positivity or negativity
based on the 7.5 ppm cutoff.  In both matrices the 10 ppm regulatory level is

distinguishable from 5 ppm (one-half the regulatory level) 100 % of the time.
                                    274

-------
     20
     15
 O
 0
  Q.
  Q.
 v«x
 Q
  i
 ^J-
 C\j"
       r =  .836

   slope =  0.589
intercept =  2.05
     10
      0
        0
         5              10             15

          2,4-D (ppm) by immunoassay
20
Figure 4. Quantitative correlation of 2,4-D immunoassay with method #8150. TCLP
extracts of various matrices that were spiked at 5, 10 or 15  ppm 2,4-D were split and
submitted for  quantitative analysis by method #8150  for chlorinated herbicides (GC
detection.)  The GC result (average of duplicates in some cases) was plotted against the
corresponding  immunoassay  result  (average of  five  replicates).   The  correlation
coefficient,  slope and intercept was calculated by regression analysis for results on these
eleven samples.
                                      275

-------
     PETRO RIS5® WATER - A RAPID, ON-SITE IMMUNOASSAY FOR

       DETECTING PETROLEUM PRODUCTS IN GROUND WATER.

Patricia P. McDonald, Richard E. Almond, James P. Mapes and Stephen B. Friedman
EnSys Inc. P.O. Box 14063, Research Triangle Park, NC 27709

ABSTRACT

There is a recognized need for innovative and cost-effective field methods for identifying
and assessing contamination of ground water with petroleum products. A rapid, on-site
enzyme immunoassay was developed for the detection of low levels of gasoline, diesel
fuel and other petroleum components in contaminated ground water samples.  The test
sensitivity for gasoline is 165 ppb, for diesel is 245 ppb and for m-xylene is  100 ppb.
The test can also be used to measure other petroleum fuels such as  fuel oil and jet fuel.
The immunoassay was developed as a  complete kit to be used on-site. The test is rapid,
taking less than 30 minutes to complete the immunoassay.

Validation studies established that the  test is specific for refined petroleum hydrocarbon
(PHC) constituents. Ground water samples which were PHC-free by GC analysis tested
negative in the immunoassay. Minimal matrix interferences were observed from ground
water obtained from different regions of the United States.  The evaluation of field
samples demonstrated good correlation between the results obtained  by the immunoassay
and by standard analytical methods for petroleum hydrocarbons. The test serves as an
accurate  field based alternative  to traditional lab methods for  screening contaminated
waste sites.

INTRODUCTION

The contamination of ground water by petroleum hydrocarbons frequently occurs during
the processing  and storage  of  refined petroleum  products.   A rapid  procedure for
evaluating contaminated sites  has been  developed using  competitive immunoassay
methodology.  Immunoassay technology has been successfully applied to environmental
testing (1-6). To  make it easier and less expensive to test for petroleum  fuels  in water,
we have developed and validated an on-site immunoassay for the detection of gasoline at
165  ppb and  diesel  at 245 ppb.   In   many states,  regulations  concerning water
contamination  focus on the  BTEX components. This test was  designed to specifically
detect the significant components in  gasoline and  diesel contaminated  samples.   The
balanced response to both aliphatic and aromatic compounds gives the Petro RIS5® Water
Test wide applicability to testing petroleum fuels at low detection levels.

This test has an advantage over traditional analytical methods because it is an inexpensive
means of evaluating samples on site.  For this reason, some of the variability which is
                                       276

-------
observed in PHC  analysis,  due to problems with sampling  and handling of volatile
analytes, is reduced.

EXPERIMENTAL METHODS

This paper presents data that was obtained using the Petro R1S5® Water Immunoassay kit
(EnSys, Inc., cat no. 70410) and the procedure outlined  in the kit instructions.  The kit
contains all the components necessary to process the water sample and conduct the assay
in the field.

The test consists of 3  principal steps.  The first step is to prepare the ground water
sample for  use in the immunoassay. The sample is collected in a VOA vial containing
buffer salts which are dissolved in  the sample by mixing.  The sample is briefly stored
in ice to allow the sample to  cool and the sediment to  settle.  The second step is to
perform the immunoassay.  The sample is transferred to a tube containing an enzyme
conjugate pellet (in a stable lyophilized form).  The reconstituted conjugate and sample
mixture is poured into an antibody-coated tube.  After incubating for 10 minutes to allow
the competitive assay to proceed, the tube is washed and the color producing substrates
are added.   The final step  is the  interpretation of results which compares the color
intensity of the sample to  the kit standard which is run in parallel with the sample.  A
lighter color for the sample indicates that the sample is contaminated.

RESULTS  AND DISCUSSION

The Petro RIS5® Water Test was validated by evaluating several performance criteria
including the  immunoassay  sensitivity, specificity, and  precision.  The test was also
examined for correlation to a reference method using field samples and gasoline spikes.
The kit storage stability was evaluated at both ambient and elevated temperatures.  All
of these performance characteristics are important to assure that the test is accurate and
field compatible.

The Petro RIS5® Water Test was configured to give less than 5% false negative results
for samples containing concentrations of petroleum fuels  at or above the detection level.
This was accomplished by setting the m-xylene standard at 55 ppb. At this concentration,
95% of samples containing 100 ppb m-xylene, 165 ppb gasoline or 245 ppb diesel will
be detected as positive. Other compounds were tested in the assay to determine a cross-
reactivity profile. The concentrations of PHC needed to give a positive result in the assay
are listed in Table 1.   The test recognizes several petroleum hydrocarbon constituents
with the greatest sensitivity to short chain aliphatic and small  aromatic compounds.

The assay specificity was examined using two approaches. First, PHC-free water samples
were evaluated and found  to be negative in the immunoassay compared to the standard.
Second, the matrix interference was examined in the assay using PHC-free water samples
                                       277

-------
from different geographical locations across the United States.  When the signals of the
PHC-free samples were compared to the signal from a PHC-free laboratory (deionized
> 18MQ with an organic removing filter) water control, the mean signal ratio  (BSAMPLE
OD/ BLAB OD) was 0.96 ± 0.07.  The absence of matrix effect would yield a ratio of 1.
 This data demonstrates that there was minimal matrix interference in the immunoassay.

The precision of the test was evaluated by measuring intra-assay repeatability and inter-
assay reproducibility. Replicate determination of assay signals for samples containing 100
ppb m-xylene spikes were obtained. The intra-assay percent coefficient of variation was
< 8% which was equivalent to 12 ppb m-xylene. The inter-assay percent coefficient of
variation was < 12% which was equivalent to 19 ppb m-xylene.

The storage stability of the test kit was evaluated by testing the assay performance in real-
time and temperature-accelerated (storage at 37 °C) studies for 3 months (Figure 1).  The
regression analysis of assay signal decay and sensitivity (BSTANDARD/BOPPM CONTROL) currently
suggests that long-term stability will be attained for at least 6 months when the kits are
stored at room temperature (about 22 °C).

Field trials were used to establish a correlation between the Petro RIS2® Water Test and
the reference USEPA method 5030/modified 8015.  In these studies, there was an overall
correlation of 84% with individual site correlation ranging from 71 % to  100%, Table 2.
No false negative results  were observed which is consistent with the test configuration.

CONCLUSION

This paper has  presented an alternative method for quick field screening of  petroleum
hydrocarbon contamination of ground water samples.  We have validated a competitive
enzyme immunoassay for the detection of gasoline and diesel at low concentrations in
water.  The rapid screening of multiple samples is  a significant advantage to traditional
analytical methods.

REFERENCES

(1)    Van  Emon, J. M., & Lopez-Avila, V. (1992) Anal. Chem. 64,  79-88
(2)    Van  Emon, J. M. (1990) ACS Symposium Series 442, American Chem. Soc.,
       Washington, D. C., p.  58-64
(3)    Van  Emon, J. M., R. W. Gerlach, R. J. White, & Silverstein, M. E. (1991) in
       Field Screening Methods for Hazardous Wastes and Toxic Chemicals, p. 815-818
(4)    Mapes, J. P., McKenzie, K. D., McClelland, L. R., Movassaghi, R., Reddy, R.
       A., Allen, R. L., & Friedman, S. B. (1992) Bull. Environ. Contam.  Toxicol. 49,
       334-341
(5)    Allen, R. L., Manning, W. B., McKenzie, K. D., Withers, T. A., Mapes, J. P.,
       & Friedman, S. B. (1992) in Contaminated Soils, Diesel Fuel Contamination, P.
                                       278

-------
       T. Kostecki & E. J. Calabrese (Eds), Lewis Publishers, Chelsea, M, pp. 37-46
(6)     Goh, K. S., Hernandez, J., Powell, S. J., Garretson, C., Troiano, J., Ray, M.,
       & Greene, D.D. (1991) Bull. Environ.  Contam. Toxicol. 46, 30-36
                                      279

-------
Table 1.      Petro RIS2® Water Test Sensitivity
          Compound or Substance
Cone. Needed to Give
   a Positive Result
        (ppb)
          Petroleum Fuel Products
             Gasoline
             Diesel fuel, #2
             Jet A fuel
             Jet fuel, JP-4
             Kerosene
             Fuel oil, #2
         Formulated Petro  Products
             Mineral spirits
            Aromatic Compounds
             Toluene
             Ethylbenzene
             o-Xylene
             m-Xylene
             p-Xylene
             Styrene
             Naphthalene
             Acenapththene
           165
           245
           280
           185
           215
           210

           490

           740
            65
           100
           100
           590
            65
             8
            10
Methanol  was spiked  individually with the listed compound.   Standard curves were
performed with each compound and  compared to the m-xylene  standard.   The
concentration of cross-reactant necessary to give a positive result was calculated from the
relative ED^ (estimated dose at 50% binding) of the compounds and m-xylene.
                                      280

-------
Table 2.      Correlation of Petro RIS5* Water Test Field Trial Results with EPA GC-
             FID Method  (5030/modified 8015) Laboratory Results
Trial No.

1
2
3
4
Total
Correlation of
Immunoassay Results with
GC-FID Results
93 %•
100%«
71 %•
75 %•
84%
False Pos.
Results
(FP/Total)
1/14
0/10
4/14
3/11
7/49
Four field trials were conducted  with the Petro RIS5® Water Test  at former service
stations.  Contaminated  groundwater samples were collected in duplicate.  One set of
samples was analyzed on-site and the other set was sent to the laboratory for analysis by
GC-FID.  Gasoline field spikes were made in blank water at the time of collection.

•      The laboratory underestimated the gasoline field spikes by >  25%.
                                      281

-------
     1.8


     1.6
        I

     1.4
     1.2


      1


     0.8


     0.6


     0.4
        0
0.5
   1.5        2         2.5
Time (months)
3.5
Figure 1.     Storage stability of the Petro RIS5® Water Test
             The test kit was stored at 37°C.  At each of the indicated time points the
             kits were allowed to cool to ambient temperature and were used for
             analysis.  Bo  is the  absorbance  at 450 nm from an unspiked methanol
             sample.  Bstd is the absorbance  obtained with the kit standard, and the
             Bstd/Bo is the ratio of the two absorbance values.  A change in the ratio
             would indicate a change in the test sensitivity.
                                       282

-------
39                          APPENDIX IX EXTRACTIONS
                                              BY
           ACCELERATED ONE STEP™ EXTRACTOR / CONCENTRATOR

       Mark Bruce. Director of Research and Development, Enseco-Wadsworth/ALERT Laboratories,
       a Division of Corning Lab Services, 4101 Shuffel Dr. N.W., North Canton, Ohio 44720; James
       Carl, Product Development Manager, Bruce Killough, Product Development Technician, Tony
       Zine, Product Development Specialist, Science Products Division, Corning Incorporated, 673
       County Rt. 64,  Big Flats, New York 14814; David Burkirt, President, Burkitt Consultants, Inc.,
       2 Captain's Way, Exeter, New Hampshire 03833.

       ABSTRACT
       The Accelerated One Step™ system shortens extraction time from 18 to 5.5 hours.  The solvent
       requirement is  reduced from 500 to 100 mL.  This new liquid-liquid extraction system was
       validated by spiking all Appendix DC organochlorine pesticide and semivolatile (BNA) analytes
       into 3  aqueous  matrices at 2 concentration levels.  Ground water, TCLP buffer #1 and waste
       water were spiked near the expected method detection limits and at levels 20 to 50 times higher.
       Organochlorine pesticide and BNA accuracy and precision are within the limits published in
       SW-846.  The average recovery for all Appendix EX pesticides and BNAs is 77% with 13%RSD.

       The Accelerated One Step™ system reduces solvent requirements  by eliminating the solvent
       pool at the bottom of the extraction chamber,  reducing the volume of the boiling flask and
       minimizing solvent vapor losses.  The extraction is accelerated  by  shortening the analyte
       transfer time from extraction chamber to boiling flask and maintaining a high solvent boiling
       rate. The membrane resists plugging by most common environmental samples.  Only very oily
       samples have impeded proper solvent cycling.

       The Corning  Accelerated One  Step™  extractor/concentrator  system is  equivalent to
       conventional continuous liquid-liquid extraction while using '/5 the solvent and '/3  the
       extraction time. The accuracy, as measured by  percent recovery, is equivalent to conventional
       continuous liquid-liquid  extraction. Precision, as measured by %RSD, is  better with  the
       Accelerated One Step™ system.  The extracts contain fewer interfering polar compounds, this
       improves analysis (particularly for TCLP). The new system also reduces labor requirements and
       increases lab extraction capacity.  The Accelerated One Step™ system will contribute to more
       timely and cost effective environmental analyses while reducing laboratory use, exposure and
       disposal of hazardous solvents.

       INTRODUCTION
       Solvent reduction has become  one of the battle cries of the Environmental Protection Agency
       for the 1990's.  The EPA has received considerable environmental and political pressure to
       reduce the  amount of pollution  generated by analytical methods which are  required to
       demonstrate compliance with EPA regulations.  The largest volume of analytical waste is from
       the  organic solvents required by current organic  extraction methodologies.   Both SW-846
       methods,  3510  (separatory funnel) and 3520 (continuous liquid-liquid),  are used for aqueous
       samples and require large  amounts (300-500 mL) of organic solvents such as methylene
       chloride. In addition, continuous liquid-liquid extraction times are typically 18 hours, plus setup
       and  cleaning time.

       Two technologies  in particular are being touted by the EPA as the "solution" to the organic
       solvent waste  disposal  problem.  Solid  Phase Extraction (SPE) and Supercritical Fluid
       Extraction (SFE) offer much promise to reduce solvent use and shorten  extraction time.  The
       largest impediment to full-scale environmental application of SPE and SFE is the lack of
                                              283

-------
methodologies rugged enough to handle the great diversity of environmental matrices and target
analytes.  Solid phase extraction has been successful for selected analytes in clean water
matrices such as drinking water.  However, some chemical and mechanical problems remain.
Extraction of very water soluble analytes (such as phenol) from 1L samples has usually yielded
recoveries less than 20%.  Particulates from waste water samples are prone to plug both SPE
cartridges and disks. Prefilters and larger disks have reduced the plugging problem but many
environmental samples are still too problematic to be handled routinely by SPE. Current SPE
technology cannot replace conventional liquid-liquid extraction for all  analytes and matrices.
However, this does not mean that solvent reduction efforts must be postponed.

The Accelerated One-Step™ extractor/concentrator system  is an excellent interim solvent
reduction solution that is available now.  The solvent volume reduction from 500 mL to 100 mL
(or less) is very attractive.  In  addition, reducing the extraction time from 18 hr to 5.5 hr is also
beneficial.  The extraction chemistry of the new One-Step™ system is the same as conventional
liquid-liquid extraction. Thus, the data are equivalent.

This new extractor was evaluated  in several areas to test its viability for routine organic
extractions. Analyte,  can the  target  compounds covered by  the  Appendix  IX list be
quantitatively extracted? Matrix, can all usual water matrixes be extracted without mechanical
problems from emulsions or paniculate plugging? Accuracy & precision, are method bias  and
reproducibility equal to or better than conventional liquid-liquid extraction? Ruggedness, is the
method and glassware durable enough to tolerate misuse and still produce acceptable results? Is
the extract "dry" enough that drying with sodium sulfate is no longer needed?

In short, the goal was to develop an apparatus which  would extract as well as conventional
continuous liquid-liquid extraction for all common environmental water matrixes yet be safer,
faster and more cost effective.

EQUIPMENT AND SUPPLIES
Hardware
    Accelerated One-Step extractor /  concentrator systems (see Figure 1)
    Neslab refrigerated circulator, CFT-75
    Tecator heated circulators, 1046
Reagents
    Methylene chloride, hexane, methanol, acetone, sulfuric acid, sodium hydroxide

RESULTS & DISCUSSION
Glassware
The accelerated One-Step™ apparatus  differs from conventional continuous liquid-liquid
extraction (and the current One-Step™ extractor) in four key areas.  1) The solvent pool at the
bottom of the conventional extraction chamber has been eliminated. The solvent is returned
from the bottom of the extraction chamber to the distillation flask (or  K-D) via gravity feed
rather than siphon action.  A hydrophobic membrane is placed across the bottom of the
extraction chamber. Organic solvent is dripped through the sample in the conventional manner.
However, the solvent  passes through the membrane  at the bottom  and runs back to  the
distillation flask. No pool of solvent is required at the bottom of the extraction vessel for siphon
purposes.   Thus, less solvent is required.  The extraction time is  also shorter since it is not
necessary to transfer analytes  from the solvent pool to the distillation flask via the solvent pool
dilution process of a conventional continuous liquid-liquid extractor.  Figure 1 shows the flow
of solvent.  2) The hydrophobic membrane effectively excludes water from the solvent thus
eliminating the need for a sodium sulfate drying step. 3) The solvent volume in the boiling flask
has also been reduced from 300 mL to 100 mL.  4) Evaporative solvent losses are reduced by
the shortened extraction time and well-sealed joints. Employee and environmental  exposure to
solvent vapor are reduced.  Also, initial solvent volume can be reduced, which saves on solventcost
                                          284

-------
         Solvent flow
         indicated by
         arrows  —&
                                        Condenser
                   tj;
    Extraction Chamber
                              Pyrex®
Membrane —fc
                   Stopcock
                                         t
                    Boiling Flask/ —^
                    Concentrator Tube
                                    Snyder Column
                                         u
Figure 1. The Accelerated One Step™ Extractor/Concentrator System.
                                285

-------
Extraction Method
Most extraction parameters are the same as described in SW-846 Method 3520.  The slight
differences were glassware assembly and use, solvent volume, extraction time and acid / base
extraction order.  The disposable hydrophobic membrane was sealed in place above the
stopcock.  The stopcock was  closed and 100 mL of methylene chloride was added to the
extraction chamber. The water sample was poured into the extraction chamber.  Surrogate and
matrix spiking were performed  normally, as well as pH adjustment. Once the boiling flask was
hot, the stopcock was opened allowing the solvent to run through the membrane into the boiling
flask. Solvent cycled through the extraction system. The solvent boiled off at a rate of 15
mL/min.  The stopcock was closed to concentrate the extract after the extraction was complete
in 5.5 hours.  The acid fraction of the semi volatile BNAs was extracted first, followed by the
base fraction.  Acid and neutral analytes remained in the boiling flask during the basic
extraction. Nitrogen blow-down was performed in the combination boiling flask-concentrator
tube.  The pesticide fraction extract was exchanged to  hexane after the extraction was complete.
The hydrophobic membrane excludes water during the extraction so sodium sulfate drying was
not required.

METHOD VALIDATION

Three aqueous  matrices were spiked at two  concentration  levels  with Appendix IX
Organochlorine pesticide and BNA analytes.  Analyte percent recovery (accuracy, bias) and
percent relative standard deviation of recovery (precision) were calculated using the results of
the appropriate analysis techniques.  Each high level matrix spike was extracted in triplicate,
while there were 7 to 8 replicates of the low level spike.  Method Detection Limits (MDL) and
Reliable  Quantitation Limits (RQL) were calculated from the low level data. MDL = standard
deviation x 3.143 (for seven replicates).  RQL = MDL x 4.

Ground water, TCLP buffer #1  and waste water were spiked with the analytes in Tables 1 and 2.
These tables  contain most of the semivolatile analytes from Appendix IX. The low and high
concentration spike levels are also shown. SW-846 analysis methods 8270 and 8080/8081 were
used as appropriate. Also, several samples with "native" analytes were extracted.  The results
were compared with data from the current liquid-liquid extraction method.

Test Matrices
Three aqueous matrices were studied; ground water, TCLP buffer #1 and industrial waste water.
The  ground  water was  taken from  a residential  drinking  water  well, which  had  high
concentrations of calcium, magnesium and iron. TCLP buffer #1 is a sodium acetate - acetic
acid buffer solution used by the Toxicity Characteristic Leaching Procedure to simulate landfill
leaching. Its  pH is 4.93 ± 0.05. Industrial waste waters have very diverse matrices. Therefore
no single waste water can be selected which completely represents all of the possibilities.
Nevertheless, a paint stripper effluent waste water sample was selected that represents many
common waste waters.

Matrix Spike Study
The  three matrices were spiked with the compounds listed in Tables  1 and 2 just prior to
extraction. The BNA extractions were acidified first to extract the acid-neutral fraction. Next,
the sample was basified and  the basic analytes extracted (note: the  acid-neutral analytes
remained in  the boiling flask).  After the basic extraction was completed, the  extract was
concentrated  to 1 mL by closing the stopcock to perform the macro concentration  followed by
N2 blowdown. Pesticide extraction was at a neutral pH as described by the current liquid-liquid
extraction methods. The high level matrix spikes were extracted first followed by the low levels
to test for analyte carryover. No carryover from one extraction to the next was observed.
                                          286

-------
Mechanical Performance
The membrane's resistance to plugging was tested with very fine inorganic solids, dissolved
solids, dissolved organics and settleable organic muck. The membrane did not plug or restrict
proper solvent cycling under  any of these conditions. The plugging effect of inorganic solids
was tested by adding 50 g of Kaolin to 1 L of tap water.  Dissolved solids effects were tested
with TCLP buffer #1.  Dissolved organics were examined with a pond  water sample.  The
settleable organic muck sample was prepared by taking 50 g of muck from the bottom of a pond
at a local hog farm.

Two types of  contaminants have reduced membrane performance.  High concentrations of
surfactants and very oily samples may generate enough emulsion to consume the entire 100 mL
of solvent.

VALIDATION STUDY RESULTS

The Accelerated One Step™ extractor/concentrator system showed equivalent accuracy and
better precision than conventional liquid-liquid extraction  for  semivolatile BNAs  and
organochlorine pesticides.  The average recovery for all compounds and spike levels was 85%
for pesticides and 69% for BNAs.  The average %RSD was 10  and 16 for pesticides and BNAs,
respectively. Calculated MDLs were consistent with published EPA data.  Since the chemistry
of the extraction has not changed, extraction efficiency should not change either. In short, those
compounds that extract well with current methodology are "well behaved" with the new system.
Those analytes that were "problem compounds" are still "problem compounds". Precision is
improved because sample extract handling is reduced. Space  restrictions limit presentation of
all validation results.  The full validation study results are available upon request.

Organochlorine Pesticides
Accuracy  and  precision were excellent for all ground water and full strength TCLP buffer
extractions. The high level waste water extractions were also good. As expected, the low level
waste water extractions were  not as good as the others because of the high levels  of non-target
compounds present in the sample.  Pesticide recovery and %RSD data for each Appendix DC
compound are  listed in Table 3.  No data are available for endosulfan n because it coelutes with
other target analytes on both the RTX-5  and RTX-1701 GC columns used in this study. Non-
target compounds in the waste water prevented the analysis of  several analytes at the low spike
level.

Figure 2 shows a frequency plot of recovery for the  high level  spikes and the data published in
method 8080B. The number of compounds with a particular recovery for each matrix is shown.
For example, 13 pesticides from the  EPA data had recoveries between 80 and 90%.  The shape
of the distribution plots show that overall analyte recovery is  the same between the EPA  data
using conventional  liquid-liquid  extraction and the new One Step system.  The recovery
frequency plot in Figure 3 clearly shows that low level analyte recovery for ground water and
TCLP buffer compares well with the high level spikes. Recovery from low level waste water
spikes decreased and was more varied. The third replicate of the high level TCLP buffer spike
was invalidated because of a  lab error.  Thus, the recovery and %RSD for this data subset are
based on 2 replicates.

Figures 4 and 5 show the corresponding frequency distributions for %RSD.  The EPA data show
typical %RSDs between 10-20%.  Most %RSDs for the Accelerated One Step™ system were
less than 10%.

Table 4 compares overall average recovery and %RSD from this study with the  method
proficiency data published in  8080B.  All average recoveries were  well within the range
specified in 8080B and compared well with the published EPA recoveries.  All average %RSDs
                                         287

-------
were lower than both the published EPA %RSD limit and the single analyst %RSD. This is
very encouraging since several extractionists participated in this study.

Pesticide MDL and RQL data are summarized in Tables 5 and 6 and Figure 6. All MDLs from
ground water and TCLP buffer are lower than the published reagent water MDLs from 8081
except for a-chlordane and methoxychlor.  The difference for a-chlordane is not significant.
The higher MDL  for methoxychlor is probably attributable to the higher spike level for this
compound  in the commercially prepared mixtures.  As expected, waste water MDLs  are
elevated because of non-target compounds.  This serves as a reminder that MDLs are matrix
dependent.

Semivolatile - BNAs
Accuracy and precision were excellent for the high level ground water and full strength TCLP
buffer extractions.  The high level waste water extractions were also good.  As expected,
accuracy and precision decreased in the low level extractions because  of the lower spiking level.
BNA recovery and %RSD data for each Appendix IX compound are listed in Table 7.

Figure 7 shows  a frequency plot of recovery for the high level spikes and the data published in
method 8270. The number of compounds with a particular recovery for each matrix is shown.
For example, 16 BNAs from the EPA data had recoveries between 80 and 90%.  The shape of
the distributions show that overall analyte recovery  is the same between  the EPA data using
conventional liquid-liquid extraction and the Accelerated One Step™ system.  The recovery
frequency plot in Figure 8 shows that analyte recovery for low level  spikes decreased and was
more varied though quite acceptable.

Figures 9 and 10  show the corresponding frequency distributions for %RSD. The EPA data
show typical %RSDs between 10-30%.  Most %RSDs for the new One Step system were less
than 10% for the high level spikes.  Low level spike %RSDs were about 20%.

Table 8 compares overall average recovery and %RSD from this study with the single analyst
method proficiency data published in 8270. Over 95% of the average recoveries were within the
single analyst range specified  in 8270 and compared well with the published EPA recoveries.
This is very good since multiple extractionists and GC/MS analysts performed this study.  The
average recovery for 2 compounds did not fall within the EPA single analyst limits.  The
average high level recovery for 4,6-dinitro-2-methylphenol was 74%, which is within the EPA
guidelines. The low pyrene recovery may be related to an analysis limitation in 8270.  The
internal standard for pyrene does not consistently mimic pyrene's chromatographic performance.
The average pyrene recovery of 59% is within the overall performance range of 52-115%.  All
average %RSDs (except 4 analytes) were lower than both the published EPA  %RSD limit and
the single analyst %RSD. The 4 analytes which exceeded the single analyst %RSD window did
pass the %RSD limit criteria.  Those compounds were: 2-chloronaphthalene, 4,6-dinitro-2-
methylphenol, 2,4-dinitrophenol, and pentachlorophenol.  This indicates the Accelerated  One
Step™ system has very good precision since several extractionists participated in this study.

BNA MDL and RQL data are  summarized in Table 9 and Figure 11.  All MDLs are consistent
with previous internal MDL studies.  No EPA MDL data was available for comparison. As
expected, ground  water MDLs were the best, demonstrating once  again that MDLs are matrix
dependent.

Extract Cleanup
This new extraction system is chemically equivalent to conventional continuous liquid-liquid
extraction with 2 exceptions.  First, extract drying  with sodium  sulfate is not necessary for
pesticides/PCBs or BNAs because the membrane effectively excludes water from the methylene
chloride extract. Second, full strength TCLP buffers  can be extracted without transferring large
                                         288

-------
quantities of acetic acid to the extract that interferes with GC and GC/MS analysis. Thus, the
RQLs are at least one order of magnitude lower than current Toxicity Characteristic regulatory
limits.  The membrane does not exclude interferences such as diethylene glycol from the extract

CONCLUSION

Extraction with the Corning Accelerated One Step™ extractor/concentrator system is equivalent
to conventional  continuous liquid-liquid extraction while using  l/5 the solvent and l/j the
extraction time. The accuracy, as measured by percent recovery, is equivalent to conventional
continuous liquid-liquid extraction.  Precision,  as measured by  %RSD, is  better with the
Accelerated One Step™ system.  The extracts contain fewer interfering polar compounds, this
improves analysis (particularly for TCLP).  The membrane resists plugging for all common
environmental samples.  The new system also reduces labor requirements and increases lab
extraction capacity. The Accelerated One Step™ system will contribute to more timely and cost
effective environmental analyses while reducing laboratory use, exposure and disposal of
hazardous solvents.

ACKNOWLEDGMENTS

Many people have contributed to  the team effort which made this project a success.
Corning, Inc.: Tim Mitchell, Joe  Speciale, Dave Patrick, Mike Oakley, Tim  Hertlein, Pete Sisti
and Bob Sell.
Hoover Company: Tony Cassanta, Bob Sack.
Enseco-Erco Laboratory: Nancy Rothman, Peter Kane.
Enseco-Rocky Mountain Analytical: Diane Lowry.
Enseco-Wadsworth/ALERT Laboratories: Betty Winiarski, Dave Counts, Jennifer McCrory,
Jeff Guiler, Tom Hula, Mark Ulman, Don Kirstead, Angel Heist, Velma Carr, Doug Joy,  Susie
Dempster,  Laura Himes, Rita Tomayko, Chuck Jacobs, Ray Risden,  Steve Jackson, Lou
Mancine, Tom Fausnight,  Steve Sablar, Todd Benenati, Dorthy Leeson, Craig Hackett, Jim
Horton, Tammy Tokos, Carl Roth, Edward Brunner, Joe Grant, Ralph Byrd, Rick Danford, John
Gruber, Rhonda Kuster, Mike Sekel, Mark  Nebiolo, Mike Paessun,  Judy Schrock,  Chris
Daugherty, Bob Scafate, Brain Haueter, Marv Stephens and Bob George.f

REFERENCES

SW-846, 3rd Edition, Method 3520, Revision 1, Dec., 1987 and Revision 2, Nov.,  1990.
                        EPA = 80SOB data
                        GW = Ground Water
                        TC = TCLP Buffer »1
                        WW = Waste Water
                        %Recovery
                                                20-i
                        GW a around Water
                        TC > TCLP Buffer »1
                        WW = Waste Water
                       %Recovery
Figure 2. High Level Pesticide Recovery.
Figure 3. Low Level Pesticide Recovery.
                                        289

-------
    30i
        ww
            TC
               GW
                           6PA = 8080B data
                           GW = Ground Water
                           TC = TCLP Buffer «1
                           WW = Waste Water
                                     50
                                   40
                           10
                                  %RSD
                   EPA
Figure 4. High Level Pesticide %RSD.
                                                           ww
                                                      GW = Ground Water
                                                      TC = TCLP Buffer *1
                                                      WW = Waste Water
                                                                                   %RSD
                            Figure 5.  Low Level Pesticide %RSD.
  20-1
      WW
          TC
                             EPA = 8081 data
                             GW = Ground Water
                             TC = TCLP Buffer #1
                             WW = Waste Water
             GW
     a1    Method
   o.o3     Detection
0.01          Limit
                 EPA
                                                        40-|
                                                       EPA.6270 
-------
Table 1.   Organochlorine Pesticides and Spike Concentrations
Compound
Aldrin
a-BHC
P-BHC
5-BHC
Y-BHC
a-Chlordane
y-Chlordane
p-Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate (I & O)
Dieldrin
Endosulfan I
Endosulfan II
Endosulfan sulfate
Endrin
Endrin aldehyde
Endrin Ketone
Heptachlor
Heptachlor epoxide
Isodrin
Kepone
Methoxychlor
Methyl parathion
Parathion
Low (|ig/L)
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.10
0.10
0.10
0.10
1.0
0.10
0.05
0.10
0.10
0.10
0.10
0.10
0.05
0.05
0.05
0.5
0.5
0.05
0.10
High (|ig/L)
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5.0
5.0
5.0
5.0
50.
5.0
2.5
5.0
5.0
5.0
5.0
5.0
2.5
2.5
2.5
25.
25.
2.5
5.0
Surrogates
Decachlorobiphenyl
Dibutylchlorendate
Tetrachloro-m-xylene
0.20

0.10
10
5.0
5.0
                           291

-------
     Table 2.  Semivolatile Compounds (BNAs) and Spike Concentrations (|ig/L)
Compound
Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite
Benzo[a]anthracene
Benzofblfluoranthene
Benzofklfluoranthene
Benzo[g,h,i]perylene
Benzofajpyrene
Benzyl alcohol
Bis 2-chloroethoxy)methane
Bis 2-chloroethyl)ether
Bis 2-chloroisopropyl)ether
Bis 2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzvl phthalate
Carbazole
4-Chloroaniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Dimethoate
p-(Dimethylamino)azobenzene
7,12-Dimethylbenz[a]anthracene
Low Hiah*
10 500
10 500
10 500
20 1000
10 500
10 500
10 500
20 ns
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
20 1000
10 500
10 500
3,3'-Dimethylbenzidine (accidentally left out of spike)
a, a-Dimethyl-phenethylamine
2,4-Dimethylphenol
Dimethyl phthalate
1 ,3-Dinitrobenzene
4,6-Dinitro-2-methylphenol
2,4-Dinitrophenol
2.4-Dinitrotoluene
2,6-Dinitrotoluene
Di-n-octyl phthalate
Dinoseb
50 2500
10 500
10 500
20 1000
10 500
10 500
10 500
10 500
10 500
10 500
Diphenylamine (coelution problem, not spiked)
Disulfoton
Ethyl Methacrylate
Ethyl methanesulfonate
Famphur
Fluoranthene
Fluorene
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadiene
Hexachloroethane
10 500
20 1000
10 500
10 500
10 500
10 500
10 500
10 500
10 500
10 500
Compound
Hexachlorophene
Hexachloropropene
lndeno(1 ,2,3-cd)pyrene
Isophorone
Isosafrole
Methapyrilene
3-Methylcholanthrene
Methyl methanesulfonate
2-Methylnaphthalene
2-Methylphenol
Low
50
10
10
10
10
10
10
10
10
10
Hiah*
2500
500
500
500
500
500
500
500
500
500
3-Methylphenol (coelution problem, not spiked)
4-Methylphenol
Naphthalene
1,4-Naphthoquinone
1-Naphthylamine
2-Naphthylamine
2-Nitroaniline
3-Nitroaniline
4-Nitroaniline
Nitrobenzene
2-Nitrophenol
4-Nitrophenol
4-Nitroquinoline 1 -oxide
N-Nitrosodi-n-butylamine
N-Nitrosodiethylamine
N-Nitrosodimethylamine
N-Nitrosodiphenylamine
N-Nitrosodi-n-propylamine
N-Nitrosomethylethylamine
N-Nitrosomorpholine
N-Nitrosopiperidine
N-Nitrosopyrrolidine
5-Nitro-o-toluidine
Pentachlorobenzene
Pentachloroethane
Pentachloronitrobenzene
Pentachlorophenol
Phenacetin
Phenanthrene
Phenol
p-Phenylene diamine
Phorate
2-Picoline
Pronamide
Pyrene
Pyridine
Safrole
1 ,2,4,5-Tetrachlorobenzene
2,3,4,6-Tetrachlorophenol
Tetraethvl dithiopyrophosphate
o-Toluidine
1 ,2,4-Trichlorobenzene
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
O,O,O-Triethyl phosphorothioate
1 ,3,5-Trinitrobenzene
10
10
10
10
10
10
10
10
10
10
10
44
10
10
10
10
10
10
10
10
10
20
10
50
10
10
20
10
10
50
10
10
10
10
10
10
10
10
10
20
10
10
10
10
44
500
500
500
500
500
500
500
500
500
500
500
2200
500
500
500
500
500
500
500
500
500
1000
500
2500
500
500
1000
500
500
2500
500
500
500
500
500
500
500
500
500
1000
500
500
500
500
2200
Surrogates
Nitrobenzene-ds
2-Fluorobiphenyl
p-Terphenyl-d-|4
Phenol-ds
2-Fluorophenol
2,4,6-TribromoDhenol
50
50
50
100
100
100
500
500
500
1000
1000
1000
                                                 ns - not spiked
* High level spike concentration for ground water is 40% of the listed value and the high level spike concentration for waste
water is 50% of the listed value.
                                            292

-------
Table 3. Pesticide Accuracy and Precision

Compound
Aldrin
a-BHC
S-BHC
Y-BHC
a-Chlordane
Y-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate II
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Ground Water
High
%R %RSD
97 5
86 7
91 6
90 8
82 7
99 5
98 5
94 6
89 6
92 5
97 6
70 5
74 5
98 5
97 5
91 4
97 5
91 6
93 6
Low
%R %RSD
92 9
85 6
95 7
89 8
83 6
105 7
102 8
101 10
91 7
92 7
101 9
84 7
84 6
100 7
102 7
95 9
100 7
106 7
97 8
TCLP Buffer
High
%R %RSD
90 5
87 11
82 5
89 6
78 6
91 3
90 6
80 3
79 4
83 5
89 4
62 5
59 7
92 7
93 4
91 4
91 4
69 34
85 7
Low
%R %RSD
93 10
92 12
88 11
94 8
80 12
99 5
93 7
105 12
88 4
83 11
94 15
84 12
69 17
92 10
93 9
92 11
96 9
91 12
90 11
Waste Water
High
%R %RSD
78 8
79 7
81 7
69 9
76 8
83 9
86 9
106 7
88 9
79 9
102 10
79 6
88 6
89 8
89 9
79 9
89 10
42 3
81 10
Low
%R %RSD
110 22
interference
65 7
64 22
51 10
74 22
45 14
interference
81 23
53 14
94 30
70 21
44 51
55 11
74 12
41 28
59 14
36 28
60 20
Table 4. Average Pesticide Accuracy and Precision vs EPA Data
Compound

Aldrin
a-BHC
P-BHC
S-BHC
y-BHC
a-Chlordane
Y-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate II
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Accelerated One-Step™
Extractor/Concentrator
Average Average
%R %RSD
93 10
86 9
84 7
83 10
75 8
92 9
86 8
97 8
86 9
80 9
96 12
75 9
70 15
88 8
91 8
82 11
89 8
73 15
85 10
8080B
%R %R
Range
81 42-122
84 37-134
81 17-147
81 19-140
82 32-127
82 45-119
82 45-119

84 31-141
85 30-145
93 25-160


90 36-146
97 45-153
89 26-144
89 30-147

85
%RSD Limit %RSD Single
Analyst
21 16
24 13
32 22
36 18
23 12
20 13
20 13

28 20
28 13
36 17


38 12
25 10
27 13
37 20

29 16
                            293

-------
             Table 5. Pesticide Method Detection Limits

Compound

Aldrin
a-BHC
P-BHC
5-BHC
Y-BHC
a-Chlordane
y-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate H
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Ground Water
Low Level
MDL RQL
0.012 0.049
0.008 0.031
0.010 0.039
0.011 0.044
0.008 0.032
0.012 0.047
0.013 0.052
0.032 0.127
0.018 0.074
0.020 0.079
0.026 0.105
0.169 0.678
0.159 0.637
0.020 0.080
0.010 0.042
0.025 0.098
0.021 0.085
0.024 0.094
0.038 0.152
TCLP Buffer
Low Level
MDL RQL
0.014 0.057
0.016 0.066
0.014 0.057
0.012 0.046
0.014 0.057
0.007 0.029
0.010 0.039
0.038 0.152
0.010 0.041
0.027 0.107
0.044 0.174
0.301 1.206
0.343 1.373
0.028 0.112
0.013 0.050
0.030 0.120
0.027 0.110
0.032 0.130
0.053 0.213
Waste Water
Low Level
MDL RQL
0.036 0.143
interference
0.007 0.027
0.022 0.088
0.008 0.032
0.026 0.106
0.010 0.040
interference
0.056 0.225
0.022 0.089
0.084 0.337
0.435 1.739
0.676 2.703
0.018 0.071
0.015 0.060
0.036 0.142
0.027 0.110
0.027 0.109
0.083 0.332
Tabie 6. Average Pesticide Method Detection Limits vs. EPA Data
Compound
Aldrin
a-BHC
b-BHC
g-BHC
d-BHC
a-Chlordane
g-Chlordane
Chlorobenzilate
4,4'-DDD
4,4'-DDE
4,4'-DDT
Diallate I
Diallate H
Dieldrin
Endosulfan I
Endosulfan Sulfate
Endrin
Endrin Aldehyde
Average of all pesticides
Accelerated One-Step™
Extractor/Concentrator
Average Average
MDL RQL
0.021
0.012
0.010
0.010
0.015
0.015
0.011
0.035
0.028
0.023
0.051
0.302
0.393
0.022
0.013
0.030
0.025
0.028
0.055
0.083
0.048
0.041
0.040
0.059
0.061
0.044
0.140
0.113
0.091
0.205
1.207
1.571
0.088
0.051
0.120
0.101
0.111
0.221
8081
MDL
0.034
0.035
0.023
0.025
0.024
0.008
0.037

0.050
0.058
0.081


0.044
0.030
0.035
0.039
0.050
0.043
                               294

-------
Table 7. BNA Accuracy and Precision

Compound

Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite 1
Aramite 2
B enzo [ a] anthracene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[g,h,i]perylene
Benzo[a]pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzyl phthalate
Carbazole
4-Chloro aniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate 1
Diallate 2
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichloro benzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Average of all BNAs
Ground Water
High
%R %RSD
88 9
87 8
97 6
95 11
75 15
69 7
87 7
not spiked
not spiked
88 7
108 5
97 7
81 5
86 7
83 8
108 9
78 7
69 8
89 7
104 10
91 5
35 8
49 9
84 4
77 7
88 10
70 7
102 8
93 9
113 12
114 8
83 7
97 8
100 7.
69 4
67 4
69 4
109 19
74 11
101 10
100 6
86 9
Low
%R %RSD
70 7
60 6
67 9
36 12
41 30
72 11
57 6
92 19
70 26
57 9
45 14
60 8
35 13
33 8
58 13
91 7
72 9
86 15
23 17
77 14
42 20
25 10
62 9
40 19
55 7
97 6
45 8
90 10
71 10
72 16
61 9
29 16
78 6
60 21
56 13
55 13
60 9
25 33
45 12
49 20
72 13
55 17
TCLP Buffer
High
%R %RSD
91 4
89 5
81 5
115 3
140 10
92 8
76 4
not spiked
not spiked
80 4
86 7
79 5
77 2
87 5
100 1
113 4
80 5
118 5
94 1
87 4
92 5
43 5
87 8
71 3
110 5
131 5
70 7
104 4
83 4
83 4
78 4
80 4
96 4
97 4
65 9
60 8
65 7
117 2
83 5
93 0.4
106 7
84 8
Low
%R %RSD
74 11
66 8
67 17
30 39
57 28
75 20
61 7
101 17
87 21
62 11
56 13
66 26
48 20
40 19
59 21
94 17
68 25
85 18
156 49
82 6
47 16
28 9
71 19
46 20
55 8
100 13
45 31
93 10
69 16
80 14
66 10
40 22
82 7
81 16
55 24
50 27
56 28
23 43
47 18
55 21
84 12
59 25
Waste Water
High
%R %RSD
69 6
72 5
83 4
102 10
114 16
125 12
60 5
not spiked
not spiked
71 5
66 7
81 4
71 3
74 8
97 12
112 4
91 7
131 6
76 17
63 4
68 14
39 6
92 7
56 11
113 4
108 10
81 7
78 9
73 1
64 7
58 10
70 5
77 11
70 12
68 2
64 4
70 3
94 9
86 8
96 9
87 10
76 10
Low
%R %RSD
54 15
52 15
86 13
71 14
29 21
89 13
41 21
77 22
74 18
37 30
37 30
37 31
21 37
34 29
68 12
93 13
68 17
57 15
111 22
44 16
42 22
35 18
55 21
39 19
59 15
88 14
52 22
56 15
39 30
58 17
53 12
19 45
64 12
87 20
51 19
50 20
52 19
32 7
54 13
77 15
92 11
56 23
               295

-------
Table 8. Average BNA Accuracy and Precision vs EPA Data
Compound

Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite 1
Aramite2
Benzo[a]anthracene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[g,h,i]perylene
Benzo[a]pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzyl phthalate
Carbazole
4-Chloroaniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate 1
Diallate 2
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Average of all BNAs
Accelerated One-Step™
Extractor/Concentrator
Average Average
%R %RSD
74 9
71 8
80 9
75 15
76 20
87 12
64 8
90 19
77 22
66 11
66 13
70 14
55 13
59 13
77 11
102 9
76 12
91 11
91 19
76 9
63 14
34 9
69 12
56 13
78 8
102 10
60 14
87 9
71 12
78 12
72 9
53 17
82 8
82 13
61 12
58 13
62 12
67 19
65 11
78 13
90 10
69 16
8270 Rev 2 1990
%R Range
%R
96 60-132
89 54-126




80 43-118


88 42-133
93 42-140
87 25-145
98 D-195
90 32-148

112 49-165
86 43-126
103 63-139
84 29-137
91 65-114
66 D-140



84 41-128
89 65-114
78 36-120
91 38-145
93 44-140


88 D-200

59 8-111
80 49-112
86 17-154
73 37-106
123 8-213
87 53-122

43 D-100
83
%RSD %RSD
Limit Sfrirje Analyst
28 15
40 24




32 21


28 15
39 22
32 19
59 29
39 22

35 16
55 35
46 24
41 26
23 13
23 18



37 23
13 7
29 18
33 20
48 28


70 30

17 13
31 20
42 25
32 24
71 28
26 15

27 28
36 21
                        296

-------
Table 9. BNA Method Detection Limits


Compound

Acenaphthene
Acenaphthylene
Acetophenone
2-Acetylaminofluorene
4-Aminobiphenyl
Aniline
Anthracene
Aramite 1
Aramite 2
B enzo [a] anthracene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[g,h,i]perylene
Benzo[a]pyrene
Benzyl alcohol
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
B is(2-chloroisopropy 1 )ether
Bis(2-ethylhexyl)phthalate
4-Bromophenyl phenyl ether
Butyl benzyl phthalate
Carbazole
4-Chloroaniline
p-Chlorobenzilate
4-Chloro-3-methylphenol
2-Chloronaphthalene
2-Chlorophenol
4-Chlorophenyl phenyl ether
Chrysene
Diallate 1
Diallate 2
Dibenz[a,h]anthracene
Dibenzofuran
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
3,3'-Dichlorobenzidine
2,4-Dichlorophenol
2,6-Dichlorophenol
Diethyl phthalate
Average of all BNAs
Ground
Water
Low Level
MDL RQL
1.5 6.1
1.1 4.5
1.8 7.3
2.6 10.5
3.6 14.5
2.3 9.3
0.9 3.8
10.4 41.4
11.1 44.4
1.5 6.1
1.9 7.5
1.4 5.4
1.4 5.7
0.8 3.1
2.3 9.1
2.0 7.9
2.0 8.0
3.9 15.7
1.1 4.5
3.3 13.1
2.5 9.9
0.8 3.0
1.6 6.4
2.3 9.1
1.1 4.5
1.7 6.9
1.0 4.1
2.8 11.0
2.2 8.7
3.4 13.7
1.6 6.3
1.4 5.6
1.3 5.2
3.8 15.1
2.2 9.0
2.2 8.9
1.6 6.5
2.4 9.7
1.6 6.5
3.0 11.8
2.8 11.0
3.3 13.2
TCLP
Buffer
Low Level
MDL RQL
2.5 9.9
1.6 6.2
3.4 13.5
7.0 28.0
4.8 19.2
4.5 17.9
1.2 4.8
10.1 40.6
10.8 43.1
2.1 8.4
2.2 8.7
5.2 20.9
2.9 11.7
2.2 8.9
3.7 14.7
4.9 19.4
5.0 20.1
4.5 18.2
22.9 91.8
1.4 5.8
2.2 8.7
0.8 3.0
4.1 16.5
2.8 11.1
1.4 5.6
3.8 15.1
4.1 16.6
2.9 11.7
3.2 12.9
3.2 13.0
1.9 7.7
2.7 10.8
1.8 7.1
3.8 15.0
3.9 15.8
4.0 16.0
4.6 18.5
3.0 11.8
2.5 10.1
3.5 14.1
3.0 12.1
5.3 21.3
Waste
Water
Low Level
MDL RQL
2.5 9.9
2.3 9.2
3.3 13.1
6.0 23.9
1.8 7.2
3.3 13.3
2.6 10.2
10.0 40.0
7.8 31.3
3.3 13.3
3.3 13.1
3.5 13.8
2.3 9.3
3.0 11.9
2.4 9.4
3.7 14.7
3.5 14.2
2.6 10.4
7.4 29.6
2.1 8.3
2.8 11.2
1.9 7.7
3.5 14.0
2.3 9.1
2.6 10.5
3.8 15.3
3.4 13.6
2.6 10.5
3.5 14.0
2.9 11.6
1.9 7.4
2.5 10.0
2.3 9.4
5.3 21.3
2.8 11.4
3.1 12.2
2.9 11.7
0.7 2.6
2.1 8.5
3.6 14.3
2.9 11.7
4.1 16.3
Overall
Average

MDL RQL
2.2 8.6
1.7 6.6
2.8 11.3
5.2 20.8
3.4 13.6
3.4 13.5
1.6 6.3
10.2 40.7
9.9 39.6
2.3 9.3
2.4 9.8
3.3 13.4
2.2 8.9
2.0 8.0
2.8 11.1
3.5 14.0
3.5 14.1
3.7 14.8
10.5 42.0
2.3 9.1
2.5 10.0
1.1 4.6
3.1 12.3
2.4 9.8
1.7 6.9
3.1 12.4
2.9 11.4
2.8 11.1
3.0 11.9
3.2 12.8
1.8 7.2
2.2 8.8
1.8 7.2
4.3 17.2
3.0 12.0
3.1 12.4
3.1 12.2
2.0 8.1
2.1 8.4
3.3 13.4
2.9 11.6
4.5 18.0
                297

-------
40              MICROSCALE SOLVENT EXTRACTION METHODS
                         FOR TARRY SOILS FROM FORMER
                     MANUFACTURED GAS PLANT (MGP) SITES
     David  Mauro  (Vice  President), June Schneider (Project Manager), Michael Young
     (Senior Chemist), and  Barbara  Taylor  (President),  META Environmental,  Inc., 49
     Clarendon Street, Watertown, Massachusetts 02172.

     Ishwar  Murarka  (Senior  Program  Manager),  Land and Water  Quality  Studies,
     Environment Division, Electric Power Research Institute, Palo Alto,  California 94303.
     ABSTRACT
     This paper discusses two microscale solvent extraction (MSE) methods developed for the
     Electric Power Research Institute (EPRI)  for use at former Manufactured Gas Plant
     (MGP) sites:  one method is for the determination of monocyclic aromatic hydrocarbons
     (MAHs), particularly  benzene, and the other  method is for  the determination of
     polycyclic aromatic hydrocarbons (PAHs).

     The MAH MSE method was developed for a soil sampling and analysis project in which
     it was important to be able to quickly and accurately  determine the concentration of
     benzene in soil samples in the field, and to use the soil  concentrations to predict TCLP
     benzene results.  This method consists of a solvent extraction using methylene chloride
     (DCM) that is carried out quickly over dry ice using refrigerator-chilled glassware and
     tools.  The extraction is followed  by analysis using gas chromatography with flame
     ionization detection (GC/FID).

     The PAH MSE method was developed as a rapid method for use in the field to help aid
     in the placement of borings or monitoring wells and for screening  contaminant levels
     during remedial activities.  This method is a modification of EPA Methods  3550 and
     8100.  It consists of solvent  extraction with DCM/acetone (1:1) followed by analysis
     using GC/FID.

     Both methods were compared with standard  EPA laboratory  methods in different studies.
     The results of these comparison studies  show that the MSE methods  can provide
     accurate, rapid,  and  cost-effective results that  are comparable to those  of standard
     laboratory methods.  Although the methods were developed primarily for field use at
     former MGP  sites, they can  be used at most  sites where  soils are contaminated with
     MAHs or PAHs, and can be used either in  a field or standard laboratory setting.
                                           298

-------
INTRODUCTION
In recent years there has been an increasing interest in potential environmental or human
health problems resulting from contaminants remaining at former manufactured gas plant
(MGP) sites (Moore, 1989).  As a result of this concern, the Electric Power Research
Institute  (EPRI) initiated a program to study several aspects of former  MGP  site
contamination problems.  One portion of the research sponsored by EPRI has been the
development of analytical  chemistry  methods for  the determination of  the  major
constituents in  tarry materials derived from MGP processes.

In a recent study (META, 1993), tarry residues from eight former MGP sites across the
country were collected  and analyzed.  Although the tarry residues  from former MGP
processes are known to be complex  mixtures of hundreds of compounds (EEI, 1984),
concentrations  as high as 328,000 mg/kg (or 32.8% of the total mass of the  sample) of
19 polycyclic aromatic hydrocarbons (PAHs) and 25,200 mg/kg (or 2.52% of the mass
of the sample) of 7  monocyclic  aromatic hydrocarbons  (MAHs) were observed among
the eight samples tested.  These high concentrations clearly  show the likely  abundance
of MAHs and PAHs in  MGP-related tars and tar-contaminated soils.

As a result of  the significance of the MAH and PAH content in residues from former
MGP sites, META Environmental, Inc. (META) developed, under EPRI sponsorship,
two methods for analysis of tarry MGP residues and  soils.   One method  is for the
determination of MAHs and the other method is for the determination of PAHs  in MGP
tars and soils.  Both  methods use microscale solvent extraction (MSB) methods followed
by analysis using gas chromatography  with  flame ionization detection  (GC/FID), and
both methods are designed for easy adaptability for field use  to provide accurate, rapid,
and cost-sensitive analysis of samples during site investigations  or remedial activities.

MICROSCALE SOLVENT EXTRACTION (MSE) METHODS
Both of the MSE methods presented here were developed for EPRI for use at former
MGP sites, although the  circumstances were different.  The PAH MSE  method was
developed  first, for  use in  rapidly  characterizing  subsurface  soils  during site
characterization or remedial activities at MGP sites.  Initially, the PAH results were used
to help define the horizontal and vertical extent of contamination in subsurface  soils, as
well as to determine the most accurate placement of groundwater monitoring wells. The
PAH MSE method  has gone through several revisions in the years since it was first
developed, and  has been successfully used at approximately  a dozen MGP sites across
the country.

The  MAH  MSE method was developed more recently, primarily for the accurate
determination of benzene in the field. Because benzene  is a known constituent of MGP
tarry residues and is included among the Toxicity Characteristics Leaching  Procedure
(TCLP) constituents, high concentrations of benzene in soils or residues will cause them
to be classified as characteristically hazardous when the TCLP benzene concentration is
                                       299

-------
greater than 0.5 mg/L.  If a relationship between the total benzene level and the TCLP
benzene level in MGP samples can be established, then monitoring for total benzene
levels in the field can be used to determine the probability that a given  sample will fail
the TCLP test for benzene.  Being able to predict TCLP failure rapidly in the field can
save considerable time and money during a variety of site activities.

Currently, there are several field methods for determining the concentration of benzene
in soil samples, including purge and trap,  portable photoionization detector (PID), and
heated headspace methods.  However, purge and trap analysis in the field is problematic,
and in addition there has been growing concern  that standard purge and trap techniques
may significantly underestimate levels of volatile constituents, such as benzene, in soils
and other solid matrices (Sawhney, 1988 and Steinberg, 1987). Field portable PIDs are
not really appropriate because they monitor total volatile compounds and are not selective
for the compound benzene. Heated headspace has been shown to be significantly less
effective for soils that are high in organic content, such as soils containing MGP residues
(Voice, 1993).  Thus, there was a need to develop an alternate method for analysis of
benzene in soils at former MGP sites.

As well as providing rapid turnaround of results either in a field or in-house laboratory
setting, the MSB methods are cost-effective, with savings usually in the range of 20 to
50%  of standard laboratory  analysis  fees.  In addition, the MSB  methods are more
environmentally sound because of the reduced  volumes of solvent used and solid and
solvent waste produced.  Furthermore, the smaller sample size necessary for the analysis
significantly increases the options for available sampling techniques.

MAH MSE Method
In the MAH MSE method, all samples, solvents,  glassware, and equipment are kept
chilled to minimize the amount of volatilization that occurs during the  sample preparation
procedure.  The MAH MSE method consists of the following steps:

       1.      approximately 2 grams  of sample are placed in a 15 ml glass centrifuge
              tube that is in a holder over dry ice;
       2.      approximately 2 grams  of anhydrous sodium sulfate are added;
       3.      surrogate standard and 5 ml of methylene chloride (DCM) are added and
              the tube is tightly capped;
       4.      the sample is shaken vigorously for at least 2 minutes, then centrifuged for
              5 minutes; and
       5.      the extract is filtered through glass wool and sodium sulfate.
The extract  is  analyzed by  gas chromatography using flame  ionization detection
(GC/FID), and positive  results are quantified using the external standard method  of
calculation.  Extraction efficiency is monitored by the recovery of the surrogate standard
                                        300

-------
compounds fluorobenzene (FB) and difluorotoluene (DFT). The detection limit for this
method is 0.250 mg/kg (ppm).

PAH MSE Method
The PAH MSE method is a modification of EPA Methods 3550 and 8100 and consists
of the following steps:

       1.     approximately 2  grams  of  sample are  placed into  a 20 ml  glass
             scintillation vial along with 2 to 4 grams of anhydrous sodium sulfate;
      2.     surrogate standard and 10 ml of DCM/acetone (1:1) are added;
      3.     the sample is disrupted using an ultrasonic microtip probe for 2 minutes;
      4.     the extract is filtered through glass wool and  sodium sulfate into a
             Kuderna-Danish (K-D) concentrator tube;
      5.     the soil sample is extracted twice more by shaking briefly with 2 X 5 ml
             of DCM/acetone; and
      6.     the extracts are combined and the volume reduced to 0.5 ml by K-D.

The  final extract is analyzed by GC/FID and positive results are quantified using the
internal  standard  method of calculation.  Extraction efficiency is monitored by the
recovery of the surrogate standard  compound 2-fluorobiphenyl (2-FBP).  The detection
limit for this method  is 0.025 mg/kg.

RESULTS AND DISCUSSION
The  results of several studies comparing the MSE methods to standard  methods are
presented and discussed in  the following paragraphs.  It should  be noted that graphical
representations of the data  are sometimes presented in units of /xg/kg (ppb)  rather  than
the standard units of mg/kg (ppm)  used throughout this report.  The use of jig/kg units
in some figures is for convenience and visual simplicity only, and does not affect the
discussion of the results.

MAH MSE Method
The MAH MSE method was compared to standard EPA methods in two studies. In one
small study with  4 MGP soil samples,  ranging in total MAH concentration from non-
detect to 1,500 mg/kg, the MAH MSE method showed significantly higher concentrations
for all MAHs than EPA Method 8020 (when concentrations were greater than the MAH
MSE detection limit of 0.250 mg/kg).  The results of this small study  are presented in
Table 1.  As shown in this  table, most of the Method 8020 results are 60 to 70% of the
values obtained using the MAH MSE method.

In a  second, larger study, 92 MGP soils were collected from ten different sites around
the country. The samples were extracted and analyzed by META's laboratory using the
MAH MSE method, and by a second laboratory using Methods 5030 and 8240 for the
purge and trap extraction of volatile organic compounds  followed by GC/MS analysis.
                                      301

-------
MAH MSB results for benzene ranged from non-detect (0.250 mg/kg) to 2,500 mg/kg
while the Method 8240 benzene levels ranged from non-detect (0.005 mg/kg) to 1,100
mg/kg.  The results are shown in Figure 1, which is a graphical representation of the
base ten log of the MAH MSB benzene results (in jtg/kg) plotted versus the base ten log
of the Method 8240 benzene results (in /*g/kg).  Figure 1  also includes a 1:1 ratio line,
and any points located to  the left of this line indicate results where the benzene value
obtained by the MAH MSB method was greater than the result obtained by Method 8240.
As can be seen in this figure, the majority (88%) of the MAH MSB benzene results were
greater than the Method 8240 benzene results. The correlation coefficient of the best fit
line  is 0.912, indicating that the results are comparable, and  visual inspection of the
graph indicates that the relationship is probably linear except for concentrations near or
below the MAH MSB detection limit.  As can also be seen in this figure, the greatest
variability  in the data occurs at the left hand side of the graph, when the MAH MSB
benzene results are in the range of 1.0 mg/kg (ppm).  This is not surprising since those
measurements are made near the MAH MSB method detection limit of 0.250 mg/kg, and
greater variability near the detection limit is common.

The  ratio of the MAH MSB benzene results to the Method 8240 results has an average
value of 34.7, indicating that measurements made using the MAH MSB method are an
average of 35 times greater than measurements made using Method 8240. However, the
standard deviation of the ratios is 103, indicating a large range of ratios, with a few very
high ratios skewing the mean upwards.   The median  ratio  is 3.4,  and is a better
indication of the actual difference observed between the methods.

Soil  samples  extracted by Method 5030 may be prepared in one of two ways, and the
large differences between the analytical results can be accounted for regardless of the
sample preparation technique. Low level soil samples are extracted by weighing 5 grams
of soil into a tube, adding 5  ml of water, and bubbling  an inert gas through.  Studies
performed  by Steinberg (1987) and Sawhney (1988) indicate that low level purge and trap
analysis only removes volatile compounds that are located in the soil pore spaces, and
does not remove compounds that have diffused into the internal micropores of the soil.
Since the MAH MSB method  is a solvent extraction technique, it should be able to
penetrate the soil particles and liberate the substances trapped inside, resulting in higher
observed volatile compounds levels.  In Method  5030, high level samples are  first
extracted by  shaking 4 grams of soil with 10 ml of methanol, then spiking 5 ml of water
with an  aliquot of the methanol and bubbling  inert  gas through.   Some of the soils
analyzed for  this project contained visible quantities of tarry residues, and would have
been prepared using the high  level technique.  MGP tar residues are not very  soluble in
methanol, so the interaction between the volatile compounds and the extraction solvent
is significantly reduced, resulting in impaired extraction efficiency.  It has  also been
shown that methanol extraction of soils yields highly variable results and poor analytical
precision (Voice, 1993).   In contrast,  the MAH  MSB  method utilizes DCM as an
extraction solvent, which is capable of solubilizing a much greater proportion of the tar
                                        302

-------
matrix constituents than methanol, again resulting in an increase in the observed levels
of volatile compounds.

The MAH MSB method for solvent extraction of soils for volatile components may be
problematic  in that the sample preparation procedure must be carried out quickly and
carefully, preferably in a cool environment using chilled tools. However, the results of
this study clearly indicate that it is a feasible alternative, and yields data that are a better
representation of the true concentrations of volatile compounds in soil.

PAH  MSE Results
The PAH MSE method was compared to standard EPA methods in two different studies.
The first study was conducted as an interlaboratory study, with  splits of 73 MGP soil
samples extracted at  META's laboratory by the  PAH MSE method with analysis by
GC/FID, and  at a second laboratory  by  the EPA Method 3550  for  the  sonication
extraction of soils with analysis by GC/MS  (Method 8270).  The total PAH  results for
these  soils ranged from non-detect to 100,000 mg/kg by Method 8270, and  from non-
detect to 110,000 mg/kg by the PAH MSE method.  Figure 2 is a graphical comparison
of the total PAH results, with the base ten log of the PAH MSE total PAH results plotted
versus the base ten log of  the Method 8270 total  PAH results.  As can be seen in this
figure, the PAH MSE data compared very well with the  Method  8270 data.  The
relationship  is  nearly 1:1 (slope of best fit  line = 1.01), indicating that the magnitude
of the individual results from the two  methods are very similar.   Also, the correlation
coefficient of the data is 0.996, again indicating that the results are very comparable, and
that the relationship is linear.

One of the major concerns regarding  the use of  microscale modifications of standard
methods  has been the issue  of sample size and  the  representativeness of samples
(Kratochvil and Taylor, 1981). It has been  speculated  that the decrease in sample size
(from 30 grams to 2 grams) may mean that a representative subsample would be difficult
to collect given the observed heterogeneity of soil samples.  If this were truly a problem,
then the results of this study would have shown outlying points where non-representative
subsamples were analyzed. A wide range  of soil types (from clay to coarse sand and
gravel) and  concentration  ranges (from visibly clean to visibly  tarry)  were examined
during this study, so any problems with non-homogeneous sample matrices should have
been readily apparent.  The excellent correlation of the data points shown in Figure 2
indicates that such  outliers were not present, so representative subsampling  was not a
problem in this study.

The second study was also  conducted as an interlaboratory study,  with splits of 49 MGP
soil samples extracted  at META's laboratory by the PAH MSE method with analysis by
GC/FID, and at a second laboratory by the EPA Method 3540 for the Soxhlet extraction
of soils followed by GC/MS analysis (Method 8270).  The total PAH results ranged from
non-detect to 1,310 mg/kg  by Method 8270, and from non-detect  to 3,130 mg/kg by the
                                       303

-------
PAH MSB method.  Figure 3 shows a graphical comparison of the total PAH results,
which is similar in format to Figure 2.  The results of this study were somewhat different
than those of the first study.  Figure 3 includes a 1:1 ratio line, and any points located
to the left of this line indicate results where the total PAH value obtained using the PAH
MSB method was greater than the result obtained using Method 8270.  As shown  in this
figure, the majority (85 %) of the results are to the left of the 1:1 line, indicating that the
PAH  MSB method yielded consistently higher total PAH results than Method  8270.
However, the correlation coefficient of the best fit line is 0.990 indicating that the results
are still comparable, and the relationship appears linear, as it did in the first study.

There are two main differences between the first and second studies which could account
for the different results.  First,  a different laboratory was used for the Method 8270
analyses in each study, whereas  the same laboratory performed the PAH MSB method
for both studies.  Second,  in the  first study the samples were extracted by Method 3550
for the sonication extraction of soils and in the second study the samples were extracted
by Method 3540 for the Soxhlet extraction of soils.   Since the PAH MSB method is a
modification of Method 3550, it is very similar to that method.  The excellent correlation
between the results in the  first study may be a reflection of the similarity of the sample
preparation  procedures employed by the two laboratories.   In the  same  way, the
consistent under-recovery of analytes  in the second study may be  a reflection on the
dissimilarity of the sample preparation procedures, or it may simply be a result of some
kind of operational error by the  second laboratory.

Operational factors that could result in a systematic bias in the sample results  (i.e.,
consistent over- or under-recovery of analytes) include holding times (if exceeded),
storage conditions (if the refrigerator is too warm), extraction efficiency (as monitored
through the use of surrogate standards), and calibration of the analytical system (if old
or degraded standard solutions are used).  Operational factors that could result in non-
systematic bias (i.e., only some  sample results are affected) include sample matrix and
heterogeneity (reflected in matrix spike recovery and duplicate RPD results), holding
times (if only some samples exceed them), extraction efficiency (if only some surrogate
recoveries  are outside criteria), and  improper handling of  samples (left  at  room
temperature,  extract  volume reduced too  much).   There was  not enough information
available for the second study to allow full evaluation of the source of error,  although
the fact that the PAH MSB and Method 8270 results correlate well certainly indicates a
systematic source of bias rather than a non-systematic source.

In an attempt to learn more about the differences between the PAH MSB data and the
Method 8270 data in the second  study, the PAH  results were summed according  to the
number of rings in each compound, then graphed in a fashion similar to the total PAH
results.  The graphs  for the three to six ring compounds were very similar to the total
PAH graph (Figure 3), and are not presented here. However, the graph for the two ring
compounds was interesting, and  is presented as Figure 4.  Figure 4 illustrates that the
                                       304

-------
difference between the results for the two ring compounds was even greater than the
difference between the total PAH results.  Over 91 % of the values obtained by using the
PAH  MSB method  were  greater  than those obtained  by  using  Method  8270.
Furthermore, the average ratio of the MSB to 8270 results was 5.5, indicating that the
MSB results were an average of five times greater than the 8270 results.  (Whereas the
average ratio for total PAHs was only 2.3).  The two ring compounds are the most
volatile of the PAH compounds, and as such are the most likely to be lost during the
sample preparation procedure.   The Soxhlet extraction procedure involves continuous
heated solvent extraction of the soils for 16 to 24 hours. Soxhlet extraction is generally
believed to be  a more thorough extraction technique for soils than sonication because of
the length of time involved.   However, if not closely controlled it can lead to lower
recoveries of the more volatile components. Volatile components may also be lost during
the concentration procedure if the extract volume is reduced too much.  It is probable
that some combination  of these factors resulted in the lower PAH results, particularly the
two ring compound results, for the samples analyzed by Method 8270.

It should be noted that  the PAH MSB analysis of the samples for the second comparison
study was  complicated  by the presence of large quantities of interfering compounds.  The
peaks for  those compounds sometimes coeluted with  the target PAH peaks,  making
quantitation difficult.   At sites containing  interfering compounds,  it is possible to
introduce a cleanup or  fractionation step to the PAH MSB method that would reduce or
remove the analytical interference.  Another possibility would be to  analyze the  PAH
MSB extracts by GC/MS (if available) rather than GC/FID, which would also serve to
remove the analytical interference.

QUALITY CONTROL
Standard Quality Control (QC) measures were employed for the MSB methods in all of
the studies described in this paper.  Parameters including surrogate standard  recovery,
internal standard performance, matrix spike recovery, and duplicate sample performance
were investigated to evaluate the overall efficiency,  precision, and accuracy of  the
methods.  In addition,  instrument detection limits and linearity were monitored through
the analysis of initial and continuing calibration   standards.   Holding  times  were
established to preserve sample integrity and representativeness. As many aspects of the
standard EPA  CLP Methods QA/QC program were emulated as was feasible for each
MSB method.  Some of the QC summary results of the MSB and standard methods are
discussed here.

MSE Methods
Table 2 provides a summary of the QC data for the MSE methods. In this table, average
recovery values for matrix spike samples, and average RPD values for duplicate samples
refer to the average calculated for all compounds in all samples.  The range given is  for
the maximum  and  minimum calculated  values, considering  all individual compound
results  in all samples.
                                      305

-------
As can  be seen in Table 2, all EPA CLP criteria  were met overall, although some
individual measurements did exceed the stated criteria. Some fluctuation in QC results
is normal and expected, even under the best operating conditions. The results in Table
2 depict a method that is  in control, and performing  satisfactorily.

Standard Methods
Table 3 provides a summary of the QC data for the  standard methods. QC results for
MAH MSB comparison  study number 1 were not  available for review,  and are not
included in Table 3.  Table 3 includes QC results only for the matrix spike compounds
that were directly comparable to the results  of the MSB methods (i.e., MAHs and
PAHs). Similarly, recoveries were evaluated only for the surrogate standards associated
with the MAH and PAH  compounds of interest.

As can  be seen in Table 3, all EPA CLP criteria  were met overall, although some
individual measurements did exceed the stated criteria. The results in Table 3 indicate
that the laboratory analytical system was in control during the analysis of the samples for
these studies.

SUMMARY
Two microscale solvent extraction (MSB) methods for the analysis of MAHs and PAHs
were introduced.  Use of the methods  is particularly suited  to the  analysis of  soils
containing MGP-related residues, in support of investigatory or remedial activities. Also,
the methods are easily adaptable to  a  field laboratory  setting, and are capable of
providing rapid turnaround of analytical results.

The  MSB methods  were compared  to their  EPA  GC/MS  counterparts,  and  the
comparisons were favorable.  The MSB methods were  shown to correlate well  with
results obtained using standard methods. Furthermore,  it was demonstrated that QC
results  from the MSB methods  were acceptable under the  EPA criteria,  and  that
acceptable precision and accuracy were possible.

It was suggested that use of the MSB methods would greatly benefit site  investigation and
remedial activities at former MGP sites, both in   terms of reductions in costs and
generated wastes and the efficiency and accuracy of the sampling/remedial program.
Although the MSB methods have not been tested at other sites, it is speculated that their
use is applicable at any site containing similar types of organic contamination, provided
that significant analytical  interferences are not present or are minimized appropriately.
                                       306

-------
                           Table 1

              MAH MSE Comparison Study No. 1
                          (in mg/kg)
Compound

Benzene
Toluene
Ethylbenzene
Total Xylenes

Total MAHs
Sample 1
MSE

ND1
ND
ND
ND

ND
8020

ND2
ND
ND
ND

ND
Sample 2
MSE

ND
ND
ND
ND

ND
8020

ND
0.002
ND
0.003

0.005
Sample 3
MSE

1.0
ND
0.94
2.8

4.7
8020

0.32
0.19
0.57
1.7

2.8
Sample 4
MSE

62
330
110
1,000

1,500
8020

42
220
77
630

970
The detection limit for the MAH MSE Method is 0.250 mg/kg.

The detection limit for Method 8020 is 0.0013 mg/kg.
                            307

-------
                              Figure 1
                MAH Method Comparison Study No. 2
                     Benzene Results (in /yg/kg)
Log (MAH MSE)
  8T
                       234
                          Log (Method 8240)
                        Data points
1:1 Line
                              Figure 2

                PAH Method Comparison Study No. 1
                    Total PAH Results (in mg/kg)
           Log (PAH MSE)
            5
            3
            1
i '^^T
i
3
5
                          Log (Method 8270)
                        Data points
1:1 Line
                                308

-------
                            Figure 3

                PAH Method Comparison Study No. 2
                   Total PAH Results (in mg/kg)
Log (PAH MSE)

  4 T
                         Log (Method 8270)

• Data points

1:
1 Line
                             Figure 4

                PAH Method Comparison Study No. 2
                Two Ring Compound Results (in /sg/kg)
 Log (PAH MSE)

   8
0
                       234
                          Log (Method 8270)
                        Data points
                                     1:1 Line
                               309

-------
                            Table 2
             Summary of MSE Methods QC Results
Audit
Matrix Spike
%Recovery
Duplicate Sample
RPD1
Surrogate
Standard I2
% Recovery
Surrogate
Standard 23
% Recovery
Average
Range
Average
Range
Average
Range
Average
Range
EPA CLP
Criteria4
M: 66 - 139
P: 35 - 137
50%
M: 59 - 113
P: 43 - 116
59 - 113
MAHMSE
Study No. 1
84%
83-85
9%
6- 11
107%
93 - 131
101%
90- 119
Study No. 2
82%
41 - 105
37%
0- 120
85%
38- 119
82%
41 - 127
PAH MSE
Study No. 1
79%
37 - 120
38%
0.4 - 180
91%
52 - 123
NU
NU
Study No. 2
85%
45 118
34%
0-90
95%
56 - 141
NU
NU
Relative Percent Difference.

Surrogate standard 1 is fluorobenzene for MAH MSE, and 2-fluorobiphenyl for
PAH MSE.

Surrogate standard 2 is difluorotoluene for MAH MSE, and is Not Used (NU) for
PAH MSE.

EPA CLP criteria are taken from SOW OLM01, and are approximated using the
most stringent criteria for compounds that are chemically  similar to  those
analyzed for by the MSE methods.  M and P refer to criteria for the MAH and
PAH MSE methods, respectively.
                              310

-------
                                   Table 3
                  Summary of Standard Methods QC Results
Audit
Matrix Spike
Compound 1'
Matrix Spike
Compound 22
Duplicate Sample
RPD
Surrogate Standard
Compound 1'
Surrogate Standard
Compound 24
Ave %Rec
Range
Ave RPD
Range
Ave %Rec
Range
Ave RPD
Range
Average
Range
Ave %Rec
Range
Ave %Rec
Range
EPA CLP
Criteria5
V: 66 - 142
S: 31 - 137
V: 21
S: 19
V: 59 139
S: 35 - 142
V: 21
S: 36
50%
V: 84 138
S: 30 115
V: 59- 113
S: 18- 137
Method 8240
Study No. 2
96%
45 125
5%
2- 14
97%
46- 134
7%
0- 18
32%
0 100
105%
81 141
90%
62-112
Method 8270
Study No. 1
91%
42 - 141
13%
1 -85
145%
0-440
32%
1 - 100
12%
0-67
92%
42 139
119%
38 220
Study No. 2
94%
34 186
24%
0.8 - 83
125%
52 228
34%
1.7 77
10%
0-52
NA
NA
NA
NA
1      Matrix spike compound 1 is benzene for Method 8240, and acenaphthene for
      Method 8270.

2      Matrix spike compound 2 is toluene for Method 8240, and pyrene for Method
      8270.

3      Surrogate standard 1 is toluene-d8 for Method  8240, and 2-fluorobiphenyl for
      Method 8270.

4      Surrogate standard 2 is bromofluorobenzene for Method 8240, and p-terphenyl-
      d 14 for Method 8270.

5      EPA CLP criteria are taken from SOW OLM01.  V and S refer to criteria for
      Methods 8240 and 8270, respectively.

NA   Information not available.
                                    311

-------
                          ACKNOWLEDGEMENTS

The research work reported here was sponsored by the Electric Power Research Institute
under contracts RP2879-01 and RP2879-12, directed by Dr. Ishwar Murarka.

META would like to gratefully acknowledge the  support and interest  of the utility
company personnel who participated in this research effort, particularly Mr. Jean Pierre
Moreau and Dr. Edward Neuhauser of Niagara Mohawk Power Company.  META
would also like to thank Atlantic Environmental Services, Inc. for their assistance.
                                REFERENCES

       Moore, T.,  "Managing the Gaslight Legacy", EPRI Journal, July/August 1989,
pp. 22-31.

       META Environmental, Inc.,  Chemical and Physical Characteristics of Tar
Samples  from  Selected Manufactured  Gas  Plant  rMGP) Sites. EPRI Report,  in
publication, 1993.

       Edison Electric Institute (EEI), "Handbook on Manufactured Gas Plant Sites",
September 1984.

       Sawhney, B.L., Pignatello, J.J., Steinberg, S.M., J. Environ. Qual., 1988, Vol.
17, pp. 149-152.

       Steinberg, S.M., Pignatello, J.J., Sawhney, B.L., Environ. Sci. Technol., 1987,
Vol. 21, pp. 1201-1208.

       Voice, T.C., Kolb, B., Environ.  Sci.  Technol.,  1993, Vol. 27, pp. 709 - 713.

       Kratichvil,  B.,  Taylor,  J.,  "Sampling for Chemical  Analysis", Analytical
Chemistry, July 1981, Vol. 53, No. 8, pp. 924-938.
                                      312

-------
41           DEVELOPMENT OF AN ENVIRONMENTAL METHOD FOR THE
               ANALYSIS OF VOLATILE ORGANIC COMPOUNDS IN SOILS
                             BY STATIC HEADSPACE GC/MS

      Tammv J. Riga. Applications Chemist, Greg O*Neil, Marketing Manager, Tekmar
      Company, P.O. Box 429576 Cincinnati, Ohio 45242-9576

      There are several methods available for analyzing volatile organic compounds in soils.
      The purge and trap technique is widely applied to the analysis of these compounds in
      soils. An analytical method is being developed using static headspace as an alternative
      approach to purge and trap.

      Static headspace is a discontinuous gas extraction where the analytes partition between
      the sample and the headspace. Various soil types were analyzed using a GC/Saturn n
      Mass Spectrometer and conditions were optimized for separation, precision, detection
      limits and linearity.  Standards were prepared by diffusing analytes through the soil (i.e.
      under zero headspace conditions) in order to simulate plume contaminated soils in the
      field. This was compared to the traditional technique of spiking the surface of the soil.

      This method also studies the effects of pre-mixing of soils and the addition of organic
      modifiers to increase the gas phase concentration within the headspace in order to in-
      crease overall method sensitivity. Internal standards were added to a matrix modifying
      solution for quantitation.
                                           313

-------
INTRODUCTION

The most problematic concern of soils analysis is how to deal with the varying
characteristics of the soil matrix. EPA methodologies which focus on the analysis of
Volatile Organic Compounds (VOC's) in water have been developed around the Purge
and Trap technique (1).  Variables such as moisture content, density, % organic matter,
amount of sand and silt, mineral content, and ionic interactions do not pose a major
concern in water analysis as they do with soils.

Soil characteristics need to be identified because each region of the country will have a
particular type of soil profile. These characteristics depend upon the native geology,
vegetation, the presence of the agriculture, and the qualities of the local climate. Soil
types may be grouped on the basis of their characteristics into two groups: Pedocals and
Pedalfers. Pedocals are distinguished by the accumulation of carbonates, calcium, or
magnesium in all or part of the soil. Pedalfers are distinguished by the absence of
carbonates and usually by an accumulation of iron and aluminum compounds (2). Once
the soils have been assigned to either of these two groups, they are further sub-typed into
hundreds of classes based upon ten distinct characteristics describing the soil and its
horizon (i.e. color, texture, structure, chemical composition including % organic matter,
% clay, % sand, type, and geology, etc.).

All of these characteristics indicate that soils exhibit highly variable matrices. The
reduction of variability in density, mineral content, and ionic interactions are some of the
problems addressed in this paper.

Two types of soils were analyzed in this study: topsoil, and sand. These soil types
represented extremes in % organic matter, mineral content, moisture, and densities. Each
soil type has a different density ranging from 0.5 gm/mL (topsoil) to 3.00 gm/mL (sand).
These densities cause the volume distribution within the headspace vial to vary. A way to
reduce this variability within the vial is to add  a Matrix Modifying Solution (MMS). The
MMS is made up of a 100% saturated aqueous salt solution that is pH buffered below 2.

Utilizing a 100% aqueous salt solution eliminates the variability  in mineral content and
helps to offset the sodication process. The sodication process is a natural occurrence in
areas of the country where precipitation is less than 20 inches annually. As the water
evaporates in the atmosphere, the salts are left  behind in the soil. The process of
increasing the sodium saturation of the soils exchange process is called sodication. The
hydrolysis of the Na+ ions or Na2C03 compounds can cause a strong alkaline reaction
changing the pH of the soil also (3). With the addition of salt saturated MMS each sample
has the same Na* concentration prior to analysis. The MMS also increases the ionic
strength of the sample which leads to increased sensitivity in the headspace analysis.

Addition of acidified MMS to soil samples aids in the maintenance of constant pH
                                        314

-------
conditions prior to analysis (resistance to strong alkaline reactions such as those
occurring in the sodication process). A low pH also helps to prevent
dehydrohalogenation of compounds and serves as a preservative against microbial
activity in the sample which could ultimately lead to degradation of analytes.

EXPERIMENTAL

The two types of soils analyzed in this study consisted of topsoil (very high organic
content) and sand (very low organic content). Each of these soil types has a different
density ranging from 0.5 gm/mL (topsoil) to 3.00 gm/mL (sand). These densities cause
the volume distribution within the headspace vial to vary. An average 2g sample of each
soil type used in this study occupied 4.0 mL (topsoil) and 0.67 mL (sand) volume within
the vial. The resulting phase ratios (V/Vm) produced by a 2g sample of each soil type
within a 22 mL vial was 4.5 (topsoil) and 31.8 (sand).  This large difference in phase ratio
produces significant differences in headspace measurements (4).

To reduce this volume and phase ratio variability, 10 mL of matrix modifying solution
(MMS) was added to each 2g sample for a final volume of 14 mL (topsoil) and 10.67 mL
(sand). The resulting phase ratio differences were significantly reduced (0.57  for topsoil
and 1.06 for sand).  The MMS is made up in 500 mL of reagent grade water. The solution
is then pH buffered below 2 with Phosphoric Acid (H3PO4) and saturated to 100% with
170g Sodium Chloride (NaCl).

The analytes chosen for this study were the 502 A/B standards purchased from
AccuStandard. A key factor in developing this method was finding an appropriate
internal standard/surrogate mixture so that the analytes could be tracked throughout the
entire chromatogram. The molecular weights, boiling points, and relative reaction
chemistries (polarities) were studied for each of the 502 A/B analytes. The Method 8260
mix was selected because it contained four internal standards and three surrogates with
similar properties of the 502 A/B analytes. Each internal standard was responsible for
tracking approximately 13 compounds in each section of the chromatogram (see Table 1
for listing).

Infernal Standards          RT         Surrogates                  RT
Pentafluorobenzene          17:50        Dibromofluoromethane      16:30
1,4-Difluorobenzene         19:53         Toluene-d8                24:19
Chlorobenzene-d5           27:08         4-Bromofluorobenzene      29:16
l,4-Dichlorobenzene-d4      31:37

Table 1.  The relative retention times for the Method 8260 internal standards and
surrogates. Each internal standard is responsible for tracking approximately 13 analytes.
                                       315

-------
Once the analyte list has been established, the next step in the development process is to
select the proper detector and column for the analysis. Due to the volatility of the
analytes, the detector should be very sensitive and the column should possess relatively
high retention characteristics to give good chromatographic separation. The following
GC conditions and 7000 Headspace Autosampler parameters used in this analysis are
listed below.
Tekmar 7000 Headspace Autosampler

     Platen temperature:
     Sample equilibration time:

     Loop Size:
     Vial size:
     Pressurize time:
     Pressure equilibration time:
     Loop fill time:
     Loop equilibration time:
     Mix:
     Mix power:
     Injection time:
     Line temperature:
     Loop temperature:
     Transfer line:
     Transfer line back pressure:
     Vial pressurization setting:
     Vial needle flow rate:
85°C
10 min. (FET mode)
50 min. (samples)
250 uL (EFT)
22 mL
0.20 min.
0.08 min.
0.10 min.
0.08 min.
10.00 min.
8
1.00 min.
85°C
85°C
0.32 mm fused silica
15 psig
15 psig
40 mL/min.
Varian 3400 Gas Chromatograph/Saturn II Ion Trap Detector
Injector temperature:
Manifold temperature:
Transfer line temperature:
Oven temperature program:
       Initial temperature:
       Ramp 1 temperature:
       Ramp 2 temperature:
       Final temperature:
Electron multiplier:
Target value:
Scan range:
200°C
265°C
220°C

10°Chold7min.
4°C/min. to 100°C
10°C/min. to 200°C
200°C hold 4 min.
1400 mV
17,000
48-260 amu
                                      316

-------
J&W DB-VRX (J&W Scientific, Folsum, California)
       60 M x 0.32 mm ID x 1.8 udf
       Carrier gas: Helium @ 1.0 mL/min.
       Phase ratio: B=44 (High K= High retention)

The transfer line from the 7000 Headspace Autosampler was connected to the analytical
column via a zero dead volume union.

RESULTS and DISCUSSION

I. Gas Sampling and Injection (PET}

Full Evaporation  Technique (FET) is a process by which an aliquot of methanolic
standard containing the volatile analytes of interest (usually 1-20 uL) is taken and placed
into a headspace vial. The vial is then sealed and the analytes are fully vaporized. This
technique is used to establish relative retention times and chromatographic separation.
Figure  1 is an example of chromatographic separation from an FET injection. All
compounds exhibit good chromatographic separation and response. The reproducibility
study showed that 95% of the analytes fell below 5% RSD.
195,000-
Figure 1:   Chromatographic Separation from an FET injection
                                     317

-------
II. Static Partitioning/Mass Distribution- 2 Phase Equilibration

Once the retention times and chromatography were established, the reproducibility of a
two phase gas/liquid headspace system was evaluated. For this experiment, 10 mL of the
MMS containing an analyte concentration of 100 ppb and 50 ppb internal standards/
surrogates was added to a headspace vial. The analytes in solution were equilibrated at
85°C for 50 minutes.  This standard represents a baseline for 100% recovery and serves
as a comparison for soils analysis.  It is a reference point to establish how much com-
pound is being absorbed when a soil matrix is analyzed. The chromatogram in Figure 2
shows good response and chromatographic resolution. Four analytes and two surrogates
were selected from each section of the chromatogram for reproducibility studies. Table 2
shows the %RSD's for 10 mL MMS based upon six replicate runs.
     220,000
                 ...      I  II   ll
 Figure 2:  10 mL MMS Spiked with 100 ppb 502 A/B Standard (IStd = 50 ppb)
              Trichloroflouromethane
              1,2,2-Trichloroethane
              o-Xylene
              1,2-Dichlorobenzene
              Dibromoflouromethane (surr)
              Toleuene-dg(surr)
RSD

2.4%
6.1
2.8
2.0
3.0
0.5
 Table 2. Reproducibility of a two phase gas/liquid headspace system (10 mL MMS)
                                       318

-------
HI. Static Partitioning/Mass Distribution- 3 Phase Equilibration

    A. Spiked sand vs. Diffused sand

The next experiment looked at the addition of 2 g of soil to the headspace vials for analy-
sis. The addition of soil creates a three phase equilibrium system (solid, liquid, gas). The
analytes must migrate from the soil into the MMS, and then partition between the MMS
and the headspace. The first sample chosen for analysis was 2 g of sand. This represented
the best case scenario because it has low adsorptivity and very low organic content.
These characteristics indicate that sandy soils exhibit low matrix interferences.

Figure 3 illustrates an example of a 2 g spiked sand sample versus a 2 g diffused sand
sample. Spiked sand represents a technique where the analyst spikes the analytes, internal
standards, and surrogates directly into the sand. The 10 mL of MMS is then added to the
sand and the vial capped. In a diffused sand sample, the analytes,  internal standards, and
surrogates are added directly to the MMS. The MMS (10 mL) is added to the 2 g soil
sample and mixed on a rotator for one hour. The purpose of the diffused technique is to
simulate real world plume contaminated samples.
           75,000 I
           509,000
Figure 3: Spiked Sand technique VS. Diffused Sand technique
                                        319

-------
Table 3 shows the area counts and %RSD's comparing the two types of samples. The
same four analytes and surrogates selected in the above study were used as reference
points again. The area counts of the diffused sand sample are between six to twenty-two
times greater than the spiked sand. This is because the analytes in the spiked sand have
not been evenly diffused throughout the sample. The reproducibility over the average of
six replicate runs indicated that 87% of the analytes had RSD's less than 10% for both
samples.
                                       Spiked   Sand      Diffused    Sand
                                       Area     RSD      Area        RSD

    Tricholoroflouromethane            15715    11.5       147974      2.5
    1,1,2-Trichloroethane                7324     8.2        160146      2.0
    o-Xylene                           37728    3.6        236264      2.5
    1,2-Dichlorobenzene                10905    5.0        82523       8.5
    Dibromoflouromethane (surr)        5352     4.8        45694       5.1
    Toluene-d-                         17048    2.1        385076      2.2
             5

Table 3. Area counts and % RSD's comparing spiked VS. diffused technique (sand)
    B. Spiked topsoil vs. Diffused topsoil

The next sample consisted of 2 g of topsoil. This represented the worst case scenario
because it has high adsorptivity and very high organic content. These characteristics
indicate that topsoils exhibit a high degree of matrix interferences.

Figure 4 illustrates and example of a 2 g spiked topsoil versus a 2 g diffused topsoil. The
technique for spiking and diffusing the analytes into the soil was the same as that used in
the above sand study. When the analytes are spiked directly into the soil, the high organic
content of topsoil results in localized adsorption. The localized spike has a limited
surface area for analyte exchange and migration from the soil into the MMS. In a field
sample, where contaminants have "plumed" through the soil, the surface area for analyte
migration is much greater in comparison to a spiked sample.
                                        320

-------
            10,000
             70,000
                     Spiked Topsoil
                                        JJJL
                     Diffused Topsoil
                        Illl..
Figure 4: Spiked Topsoil technique VS. Diffused Topsoil technique
Table 4 shows the area counts and % RSD's for the comparison of the two types of
samples. The same four analytes and surrogates selected in the above study were used as
reference points again. The area counts of the diffused topsoil sample are once again
greater than that of the spiked topsoil. The reproducibility over the average of six repli-
cate runs indicated that 99% of the analytes in the diffused topsoil had RSD's less than
20%. Only 68% of the analytes in the spiked topsoil had RSD's less than 20%.
                                      Spiked Topsoil
                                      Area      RSD
Diffused Topsoil
Area       RSD
    Tricholoroflouromethane
    1,1,2-Trichloroe thane
    o-Xylene
    1,2-Dichlorobenzene
    Dibromoflouromethane (surr)
    Toluene-dg

Table 4.  Area counts and % RSD's comparing spiked VS. diffused technique (topsoil)
22704
1896
4796
1112
2355
4025
14.6
9.3
35.1
16.9
3.6
3.1
145993
15998
33956
8058
15765
23333
7.5
5.0
16.2
3.4
7.9
4.8
                                       321

-------
IV. Effects of Increasing Organic Content

Soil samples that contain a very high organic content tend to very adsorptive. The
analytes tend to adhere to organic sites in the soil which reduces the release of contami-
nants. The addition of the MMS helps to reduce some of this variability because the
analytes are able to migrate more rapidly from the soil to a liquid solvent (MMS).

Figure 5 is the resultant chromatogram illustrating a diffused sand sample to a diffused
topsoil sample.
          509,000
           70,000
 Figure 5: Effects of increasing organic content.

 Table 5 shows the difference in peak area counts when comparing soils that have varying
 amounts of organic content. In both soil samples, Trichlorofluoromethane (gas) exhibits
                                       Topsoil             Sand
                                       Area                Area
    Tricholoroflouromethane
    1,1,2-Trichloroethane
    o-Xylene
    1,2-Dichlorobenzene
    Dibromoflouromethane (SUIT)
    Toluene-dg
145993
15998
33956
8058
15765
23333
147974
160146
236264
82523
45694
385076
Table 5. Area counts illustrating the effect of increasing organic content
                                       322

-------
very similar area counts. This is because Trichlorofluoromethane is a relatively volatile
compound with a high diffusion coefficient. The other five compounds in table 5 show
between a seven to sixteen fold decrease in area counts with the topsoil sample. The
topsoil tends to hold onto the analytes, especially those heavier compounds found in the
middle and very end of the chromatogram.

V. Diffused Topsoil- 14 Day Time Study

Many EPA laboratories today have a period of up to 14 days before the sample has to be
run. The purpose of this next study was to examine the relative degradation of analytes
over a period of 14 days. Three replicates of samples for each day were prepared using 2g
of topsoil and 10 mL of MMS. The MMS was spiked with an analyte concentration of
100 ppb (internal standards and surrogates were at 50 ppb) and the samples were all
mixed for one hour to simulate the diffusion technique. Each day, for 14 days, three
replicates were pulled from a 4°C refrigerator and run by headspace.

Figure 6 is a graph of the area counts from early eluting compounds vs time
(Pentafluorobenzene as the internal standard). Four additional compounds were selected
from the front end. The graph shows that over a period of 14 days, there is some degrada-
tion of the analytes. The important point to notice is that the internal standard degraded at
the same rate as all the other analytes. This indicated that the internal standard was
properly tracking the analytes of interest. As long as the internal standard falls at the
same rate as the analytes, quantitive results may be expected over the time period tested.
                                               Pentaflourobenzene (IS)
                                               Chloromethane
                                          	1,1-Dichloroethene
                                               1,2-Dlchloroethane
Figure 6:  Area Counts of early eluting compounds VS. time
                                       323

-------
Figure 7 is a graph of the area counts of late eluting compounds vs time (1,4-Dichloro-
benzene-d4 as the internal standard). Once again, all analytes are properly being tracked
and degrading at the same rate as the internal standard.
                                       —   1,4-Dichlorobenzene-d4  (IS)
                                       	tert-Butylbenzene
                                       	p-lsopropyltoluene
                                       	1,2-Dlchlorobenzene
                                            Naphthalene
Figure 7: Area counts of late eluting compounds VS. time

VI. Carryover

A common problem associated with soil analysis today is carryover from very highly
contaminated samples. Samples in the high ppm range can shut down an instrument and
halt sample processing due to high level contamination and carryover. The major advan-
tage of headspace analysis over purge and trap is that there is no carryover or contamina-
tion problems associated with this technique. The relative carryover percentage is less
than 0.2% for all analytes run after a 10 ppm sample. The Saturn Mass Spectrometer and
the column also showed no major effects of contamination.

CONCLUSION

The soils analysis method by static headspace is very clean, sensitive, and reproducible. It
also eliminates many variables found in current soil methodologies. Future work will
focus on practical field sampling techniques. One example of this is to use disposable
syringes that will deliver a volume of soil to a screw top headspace vial. The MMS with
the internal standards and surrogates will automatically be added to the sample in the
field and the vial sealed.  This mode of field sampling is simple, less prone to error, and
eliminates losses associated with sample transfer prior to analysis.
                                       324

-------
Another focus will be on the addition of a non-volatile organic modifier to the MMS in
order to increase the sensitivity of compounds analyzed from highly organic matrices.
Final detection limits and dynamic ranges will then follow.

REFERENCES

1.  U.S. Environmental Protection Agency. Methods for the Determination of Organic
Compounds in Drinking Water, EPA-600/4-88-039; 1988.
2.  Finkl, Charles W. Jr., 1982, Soil Classification, Benchmark Papers in Soil Science/1,
pp. 145-168
3.  Tan, Kim H., 1982, Principles of Soil Chemistry, Dekker, New York, pp. 187-189.
4.  B.V. loffe, A.G. Vitenberg, "Head-space Analysis and Related Methods in Gas Chro-
matography" 1984 John Wiley and Sons, Inc., pp. 30-37
                                      325

-------
42   IN-SITU  DERIVATIZATION  OF  SOIL  AND  SEDIMENT  SAMPLES IN  A
      SUPERCRITICAL  FLUID  EXTRACTION  CELL;  THE  EXTRACTION  OF
      PHENOLS AND CHLORINATED PHENOLS

      DENNIS GERE.  LENORE G. RANDALL, CHARLES KNIPE, WILLIAM PIPKIN
      Hewlett-Packard, Little Falls Site, 2850 Centerville Rd., Wilmington, DE 19808

      H. B. LEE, Environment Canada, P.O. Box 5050, Burlington, Ontario, Canada LTR 4A6

      ABSTRACT

      This manuscript will describe  a series of experiments which allow the  simultaneous derivatization of
      analytes in solid and semi-solid samples containing phenolic compounds  including the more intractable
      chlorinated  species. The  derivatizing reagents acetic  anhydride and triethyl amine were added   by
      admixing the separate reagents to the extraction cell before  the cell was placed in the SFE apparatus.

      Further experiments with respect to the reaction and extraction efficiency for methylphenols and
      nitrophenols as well as the chlorophenols and permutations of mixed moieties is in progress, collectively
      in the Little Falls and National Water Research Institute, Environment Canada labs.  Preliminary results
      suggest that  dinitrophenols  do  not  react  under  the  conditions  described previously, although
      mononitrophenols appear to react as phenol and chlorinated phenols do.  This work is continuing and will
      be described in a later publication.

      INTRODUCTION

      The supercritical fluid extraction (SFE) of phenols and chlorinated phenols from soils and sludges is an
      analyte/matrix pair receiving significant attention with respect to getting an acceptable, robust SFE method
      developed  and ultimately formalized as a registered EPA method.  The group of H.B. Lee (1) has
      published a method which employs in situ derivatization and post-SFE clean-up of the extract to remove
      the unreacted derivatizing agent.

      Experimental

      SFE Conditions:

       Hewlett-Packard Models 7680T and 7680A were used without modification.

      SFE Method The following conditions are grouped according to function.

      Modifiers/reagents added  to the thimble; 100 uL acetic anhydride, 100 uL triethyl amine (sample size is
      usually 1 gram).

      1. Extraction — pressure, 5250 psi;  extraction  chamber temperature, 80 C;  density, 0.80 g/mL;
      extraction fluid composition, CO2; static  equilibration time,  5 minutes; dynamic extraction time,  10
      minutes; extraction fluid flow rate, 2.0 mL/min; resultant thimble-volumes-swept, 2.

      2. Collection (during Extraction) - trap packing, Hypersil ODS; trap temperature, 15 C; nozzle (variable
      restrictor) temperature, 45 C.

      3. Reconstitution (of collected extracts) - rinse solvent, 50/50 (v/v) isooctane; collected fraction volume,
      1.2 mL; trap temperature, 40 C; nozzle (variable restrictor) temperature,  45 C; rinse solvent flow rate,
      1.0 mL/min; fraction destination, vial #n.
                                                  326

-------
GC Conditions:
Carrier gas =  hydrogen, 15 psi head pressure .splitless injection, injection volume =  1.0 uL; column =
HP 5  fused silica capillary 0.200 mm x  25 meters; column temperature program , initial  temperature
= 70 °C, initial program rate  = 30 °C/mm (to 120°C) 2nd program rate = 2°C/min,  final  temperature
= 200 °C.

Results and Discussion

This is a study (FIGURE 1) which was started by the lab in Canada, and is being continued towards some
interesting possibilities.   It was stated earlier that phenols and substituted phenols are  reasonably polar
compounds, and are particularly  difficult  to extract out of solid matrices.  Another problem that causes
them to be more intractable,  or  more difficult to extract  and analyze, is that they often occur in solid
wastes in ratios of high-low concentrations.  This is seen in a chromatogram in a later diagram (FIGURE
5). The  phenols  are usually distributed in a broad range of phenolic constituents—chloro, nitro,  methyl
and every mix that might be possible.

There arc at least  three possibilities to the extraction of phenolic compounds. From solid  waste, one
approach would be to use carbon dioxide by itself.  Carbon dioxide is a relatively non-polar  solvent, and
for these  relatively  polar analyses,  does not yield particularly good results  in  real samples.  Another
approach might be to  add a bquid to the CO2 stream-a modifier or additive.  The approach which was
done by  Bill Lee  and  his co-workers was actually to do an in-situ derivatization   in the extraction vessel
and  then extract the derivatized materials.   We also will take  a  brief look at  supercritical fluid
chromatography to suggust a third interesting approach for the extraction.

FIGURE 1. Outline of SFE extraction Study	
 Phenols and substituted phenols are polar

      phenols occur in high/low concentration ratios

      broad range of substituents

       *  chloro
       *  methyl
       *  nitro
       *  all permutations of above substituents mixed
 3 Possible avenues of structuring the SFE method

      CO ? only

      in-situ denvauzation with  CC>2 only

       CO;> + modifier + additive ( as in packed column SFC)

 Hard look al Canadian method and any limitations

 Automation

      PrepStation to automate  manual cleanup steps

      SFE- GC Bridge to speed throughput and minimize operator error
                                               327

-------
            Loading the Thimble for In-satu Derivatizaticsi
      30 id Acetic Anhydride
        50 li Standards

        30 ul TrietMamine



        FQter paper disk
200mgCelite

  1 gram of solid sample
  mntfomng Phenols
   200mgCelite
FIGURE 2. Loading the extraction vessel with sample and reagents
                                        328

-------
There are fewer manual steps as you automate, therefore, less hands-on time per method, and time is very
costJy in environmental laboratories. Improved accuracy, precision and the overall method become less
dependent upon the skill level and the particular approach of a given technician or operator.
Derivatization coupled with analysis allows just-in-time delivery of each sample for analysis when the
analytical device is finished with the last sample and is ready for the next one.  Automated optimization of
reaction conditions will allow some interesting possibilities for unattended execution of designed
experiments.  Samples in report out in a very timely manner, and unattended, you can set a set of eight
samples into the extractor and come back in the morning and find your results unattended.

Figure 2 illustrates the basic approach that's used in the in-situ derivatization.  This is a schematic drawing
of the  extraction vessel .  The sample, reagents and other materials are placed in the extraction vessel
starting from the bottom. The end-piece is attached to the metal thimble first, then a  filter paper disk  is
placed  on the face of the end-piece frit. Then, 200 mg, of an inert  material such as celite  is placed on top
of the filter paper disk, followed by a 1 gram sample of solid waste or sludge. A liquid aliquot
 ( 30 uL) of triethyl amine (TEA) is added to the sample via a micro syringe. This serves to raise the
effective pH to an optimum level  for the reaction that will occur  between  the phenols and the acetic
anhydride. Another layer consisting of 200 mg of celite is then placed on top of the sample. Another Iquid
aliquot  ( 30 uL) of acetic anhydride is metered onto the celite with a micro syringe. This acetic anhydride
is  the derivatizing reagent to form acetyl derivatives of the various phenols. Finally, another disk of filter
paper is placed on top of this. The exit end-piece (fritted cap) is placed on the metal extraction vessel and
the entire sample vessel is placed in  the carousel of the SFE.

The reconstitution  solution resulting  from the SFE  step is removed from the extractor and a clean-up
procedure is carried out.  This  procedure is done manually (separate from  the SFE equipment). Work  is
underway to incorporate this into automatic PrepStation SPE manipulations to facilitate automating clean
up these steps.  The reconstituted extract solution   ( derivatized phenol acetates dissolved in hexane)  is
mixed  with 3 ml of a  one percent solution of potassium carbonate and is agitated for one minute.  That
potassium carbonate neutralizes the excess TEA reagent, which was added to obtain an effective optimum
pH for the derivatization reaction in the supercritical  CO2. The extract solution  is then passed through  a
small bed of magnesium sulfate (this can be packed  into a Pasteur pipet) to remove excess water.  That
resulting filtrate  is then passed through  another short bed of 5% deactivated silica.The  analytes  are
eluted  in a 10 ml  solution,  consisting of a  hlmixture of petroleum ether/methylene  chloride.  This
solution  is evaporated  and re-exchanged to  5 ml of isooctane.  A  luL aliquot is  injected  into the gas
chromatograph.  The SFE method used a temperature of 80 degrees at a constant density of 0.8 g/ml.  with
a 5 minute static period for the derivatization and soaking of the sample. This was followed by a 10
minute period  of dynamic extraction  at a volume flow rate ( measured as liquid carbon dioxide  at  4
degrees centigrade) of 2 ml per  minute.

Results from Lee (1) and co-workers  with a commercially available reference soil 130-100 ,  comparing
steam distillation to supercritical fluid extraction in-situ derivatization are seen in Figure 3.

FIGURE 3.  Recovery and  Precision from Reference Soil 130-100 ( Ref.  1)

                   Steam Distillation    SFE               Steam Distillation    SFE
                   Recovery mg/kgm  Recovery mg/kgm   Precision mg/kgm   Precision mg/kgm
2,3,5    Tri  Cl          0.40              0.36                 0.01               0.01
2,3,4,5 Tetrad           14.4              13.9                  0.40               0.30
2,3,4,6 Tetrad          20.6              20.2                  0.40               0.30
2,3,4,5 Tetrad            1.9               1.8                  0.10               0.10
Pentachiorophenol      1,499.0          1,483.0                67.0                93.0

The recoveries for the  supercritical fluid extraction  correspond very well with the steam distillation for all
of  the  compounds.  Steam distillation  requires many additional  manual manipulations—most  are quite
                                               329

-------
operator- dependent. In comparison,  the SFE  equipment used in this study is operated in the same
manner by all operators, because all steps are computer controlled and recalled from a stored method.

This table ( Figure 3.) shows the results of a representative phenol containing  sample. This particular
sample is from a site in Canada adjacent to  stored railroad ties (treated with creosote to promote long
lifetime of the wood). It is normally expected mat there will be a large concentration of many chlorinated
phenols.  In addition, as mentioned above, there was a very small concentration of trichloro phenol and a
very high  concentration of the pentachlorophenol.

In Figure 4, we see the chromatogram which yielded  the information presented in the table  (Figure
3.).The first peak annotated is 2-3-5 trichlorophenol, the next three annotated (small peaks) are tetrachlor
phenols and the large  off-scale peak  at the end of the chromatogram is pentachlorophenol.  As noted
above, the small peak,  the trichloro phenol, was found at a concentration of 0.4 mg/kgm and the largest
peak, pentachlorophenol was found  at 1483 mg/kgm.  This  illustrates the large  range of phenolic
compounds to be expected, greater than three orders of magnitude range.
FIGURE 4.  GC/ECD Chromatogram of Reference Soil Containing Chlorinated Phenols
                                                        3   1
                                                         '
                                                                             8:
                                                                           .A...
                                                                                 'll
                                               330

-------
FIGURE 5.  Recovery from Spiked Sample - 6 replicates (Ref. 1)

1 Id
100
90
80
70
GO
50






-





n











•i .








2.G 3.
2.1


^













"i











M








-,












-L


—







-|










n























-i -,

•^




: - :
.


5 3,4 2.3.G 2,4,5 2.3,5,6 2,3,4.5
2.:! 2 1.»J 2,:i..r) 3, 1,5 2,3. ,(! 2,3, 1.5.G
• Rf-c'U'erj
0 5 mg/kgrr


Ket
5 0

oven
rag/kgm



Figure 5 shows the recovery of a number of the substituted chlorophenols ranging from the di-substituted
up to the penta chlorinated substituents.  The solid (darkest)  bar graphs  represent recovey of  spiked
levels of 0.5 mg/kgm ( parts per million on a weight-per-weight basis).  The lighter patterned bar graphs
represent recoveries for the same compounds at 10 times higher concentration, approximately 5 mg per kg
level.  It is relevant to note the significance of the recovery across this variety of  chlorinated phenols at
the two different levels.   As expected, there is a slightly  lower recovery at the sub- part per million level
compared to the 5 part per million level, however, the average recovery for six replicates is greater than
87 % in the worst case  and greater than 104 % in the best case, with an  overall grand mean recovery
greater than 95 %.
                                               331

-------
FIGURE 6. % Relative Standard Deviation from Spiked Sample - 6 replicates (Ref. 1)
             10
                 2.6     3,5     3,4    2,3,6    2,4,5  2,3,5,6  2,3,4,5
                    2,1    2.3    2,4,6   2,3.5   3,1,5   2,3,4,6   pep Grand mean
                                   |%rsd      I
                                   0.5 ing/kgm I
I %rsd
I 5.0 mg/kgni
 In Figure 6 we see that in addition to good recovery, the SFE in-situ derivatiztion/extraction yields very
 acceptable precision. The solid (darkest)  bar graphs show that precision for  the  compounds at the 0.5
 (parts per million ) mg/kg level ranges from 3% rsd to 8  % rsd. The (lighter) patterned bar graphs show
 the precision  at 5  mg/kgm (ppm).  As one would normally expect,  in  trace analysis, the precision
 improves significantly as the detected concentrations are 10 times greater. That is, the grand mean % rsd
 for the 14 compounds at the 0.5 ppm level is 6.5 % while at the 5.0 ppm level it improves to 5.0%.

 Because  the derivatization method as described here does not derivatize dinitro phenols, we are examining
 another possiblity for the broader set of phenols. We are following some experiments that our colleagues
 at HP are doing with packed column SFC. When analyzing mixtures with  packed column SFC, the
 conditions in many respects  are analogous to supercritical fluid extraction.  If one can find experimental
 data on packed column SFC, this may  suggust good conditions for SFE with solid matrices.   Jerry Deye
 and Terry Berger of our laboratories at Hewlett-Packard in Wilmington, DE have an SFC chromatogram
 performed the separation of many of the same phenols described in this  study, as well as  dinitro phenols
 Figure 7 ).
                                                332

-------
             Figure  7
SFC of un-derivatized Phenols
               EM MUM (MAM*
                         RTflw
                        1 014 2-Nfcphml
                        2(177
                        3QN
                        41.23
                        5143
                        I t.4f 2,4J)ttfa9ftml
                        71*
                        i in PM*
                        IU9
                        10 W
                        11 146
                                 20 l
                         MoHl phlHi
                                 ouM
                           IMfflpKlM
                         Cwwer
                      333

-------
The total elution time for the eleven phenols, is  only three and a half minutes.  That is a fast , high
resolution    separation of the  eleven  phenol  compounds.  These  include     2-nitro,  the  4-6
dinitromethylphenol and phenol itself. This is a packed  SFC DIOL-bonded silica column.  The mobile
phase was supercritical CO2 with a  mixture of  trichloro acetic acid  ( 0.05 %) in methanol .  The
methanol/trichloro acetic acid mixture was  then introduced as a  5% volume of the carbon dioxide. This
represents an additive  in methanol, then added to CO2~it is actually a ternary mixture.  The pressure
was  very low,  2000 psi, at 37 degrees  centigrade. The phenols are not derivatized.  We are taking a
look at this simultaneously with the derivatization steps and the automation.

Summary
We are evaluating simplification of running  the proposed method routinely by automating the post-SFE
fraction clean-up using automated ( HP PrepStation) instrumentation.

We have further looked at the reaction and extraction efficiency  for methyl phenols and nitro phenols as
well as the chlorophenols described in the original study. We find the in-situ method very good for die
chlorinated phenols. For dinitrophenols,  we are considering un-derivatized extraction with co-solvents and
additives to enhance the extraction and recovery. One or more of these approaches will be offered to the
USEPA SW846 committee in the form  of "deliverables" (3) for consideration as methods for extracting
phenols and substituted phenols. This normally would result in  a round robin study  to test the ultimate
robustness in a wide variety of laboratories.
 Acknowledgment

 The authors would like to thank all of the people who are helping with these SFE phenol studies and
 especially Barry Lesnik,  Robert Hong-You, Tom Peart, Pat Castelli, Joel Cheng, William Pipkin and
 many others who helped in their own way

 References

 1. H.B.Lee , T.E. Peart, R.L. Hong-You, J. Chromatogr. 605 (1), 109-113 (1992).

 2. D.R. Gere, C.R. Knipe, P. Castelli, J. Hedrick,  L.G. Randall, J. Orolin, H. Schulenberg-Schell, R.
 Schuster,  H.B. Lee,  and L. Doherty  "Bridging the Automation Gap between Sample Preparation and
 Analysis:   SFE, GC, GC/MSD and HPLC Applied to Environmental Samples" manuscript in  press,
 J.Chromatog.  Sci.

 3. Personal communication from Barry  Lesnik, Office of Solid Waste, US Environmental Protection
 Agency, 401 M Street, SW, Washington, DC  20460 USA.
                                              334

-------
43          SUPERCRITICAL FLUID EXTRACTION (SFE) of PCBs

      Mark L. Bruce. Director of Research and Development, Enseco-Wadsworth/ALERT
      Laboratories, a Division of Corning Lab Services, 4101 Shuffel Dr. N.W., North Canton,
      Ohio 44720


      ABSTRACT

      Supercritical fluid extraction parameters will be optimized using fractional factorial
      experimental designs.   Matrix modification with water will  also be  investigated.
      Preliminary results with Aroclor 1248 indicate more rigorous extraction conditions than
      those specified in Method 3560 may be necessary.

      INTRODUCTION

      Extraction of hydrocarbons from soil samples was one of the first  applications of
      analytical SFE with carbon dioxide. Similarly hydrocarbon extractions were the first SFE
      method from the  EPA  Office of Solid Waste. The commercial instrumentation has
      improved significantly in the past two years.  The environmental lab can  choose between
      several SFE instruments specifically geared for the high through-put lab.  The next step is
      to extend the current SFE methodology to  other environmental analytes.  Non-polar
      compounds such as PAHs and PCBs should be extractable under conditions  similar to
      hydrocarbons.  SFE of TPHs will dramatically reduce the volume of Freon-113® used by
      the Jab.  Extending this technology to other analyte fractions will reduce the  use of
      methylene chloride. This will reduce both environmental and employee  exposure to this
      suspected carcinogen.

      Work by Langenfeld,  Hawthorne, Miller and Pawliszyn (1) shows good PAH and PCB
      recovery using  high temperature (200°C) extractions.   Preliminary work at Enseco-
      Wadsworth/ALERT indicates that the parameters of  the TPH method may not be
      sufficient to achieve  good PCB recovery.  Two  approaches are possible, 1) use co-
      solvents  (modifiers) or 2) use higher temperature, pressure and  flow.   Modifiers have
      both  advantages and disadvantages.  We will focus on the second option.  If acceptable
      PCB recovery can be obtained without modifiers for off-line extraction then the eventual
      switch-over to on-line extraction will be simpler.

      INSTRUMENTATION. EQUIPMENT AND SUPPLIES

      Supercritical Fluid Extractor
       Suprex, PrepMaster, AccuTrap
        Restrictor, prototype variable restrictor (manual)
       Suprex, AP44
       5 rnL extraction vessel
      Gas Chromatograph
       Hewlett Packard 5890
       Electron Capture  Detector
       Column RTX-5, 30 m, 0.32 mm ID, 0.5 [L DF
      Solvents
       Hexane, EM Science
       CO2, SFC grade with 1500 PSIA Helium headspace with dip tube, Scott Specialty
                                            335

-------
DISCUSSION

There are several supercritical fluid extraction parameters to optimize.   They are
extraction time (static & dynamic), CO2 pressure, CO2 flow rate, extraction temperature,
analyte trap, modifier type, and amount.  Solid matrices such as loam, ash and various
clays (kaolin, Fuller's earth, montmorillonite)  will be  examined.  Fractional factorial
experiment designs will be used to investigate and optimize this  large number of
variables.

Water affects organic  extractions in several interesting ways.  Too much water is
detrimental and reduces analyte recovery.  However, too little water can also reduce
analyte recovery.  Extracting TPHs from dry clays with Freon-113® was facilitated by
adding water to the sample (2).  A similar effect was noted for supercritical fluid
extraction of TPHs in draft Method 3560 (3). However, too much water was a detriment.
Very wet clay samples must be partially dried before good TPH recovery can be achieved
(4,5).  These water effects are inconsistent if one views water in the role  of a CC>2
solubility modifier.  Water as a polar  solvent should only decrease the  extraction
efficiency of non-polar analytes such as hydrocarbons. However, Hawthorne has noted at
several conferences (6, 7, 8) that co-solvents should also be viewed as matrix modifiers.
McNally observed physical changes in soil matrices when water was added (9).  These
physical changes  may explain in part how water affects the extraction process.  This
raises the possibility of using water as an SFE modifier even for non-polar analytes. If
effective, water would be the least expensive and safest modifier.

Preliminary work with PCBs indicates that the SFE conditions specified in draft Method
3560, December  1992  may not be sufficient for good PCB  recovery.  Aroclor 1248
recovery from Kaolin was 51% using these parameters.  The final goal is an extraction
method that yields acceptable PCB recovery in a short time frame under production
laboratory conditions.

SUMMARY

Fractional factorial experimental  designs will be used to optimize the key variables
affecting PCB recovery from various solid matrices.  The use of water as matrix modifier
will be investigated.

REFERENCES

(1) Analytical Chemistry 1993, 65, 338-344.
(2) American Environmental Laboratory, 1990,12,25-28.
(3) SW-846 draft Method 3560, December 1992,4-5.
(4) Interlaboratory Evaluation of an Off-line SFE/IR Method for Determination of
Petroleum Hydrocarbons in Solid Matrices Project Report, EPA/600/R-93/056 Apr, 1993.
(5) Proceedings 8th Annual Waste Testing & Quality Assurance Symp., 1992,427-436.
(6) PitCon, 1991
(7) 4th Internal. Symp. on Supercritical Fluid Chromatography and Extraction, 1992.
(8) American Petroleum Institute Workshop, Colorado Springs, Feb., 1992.
(9) Proceedings  Internal. Symposium on  Supercritical Fluid Chromatography and
Extraction, 1992, 35-36.
                                       336

-------
44   Determination  of Organic Compounds  in Ground  Water by Liquid-Solid  Extraction
      Followed   by   Supercritical   Fluid   Elution   and   Capillary   Column  Gas
      Chromatography/Ion-Trap Mass Spectrometry

      By J.S. Ho.  P.M.  Tang*, J.W. Eichelberger, and W.L. Budde
      Environmental  Monitoring  Systems  Laboratory,  U.S.  Environmental   Protection
      Agency, 26 W.  Martin Luther King Dr. Cincinnati, Ohio 45268
      * Technology Applications,  Inc.  c/o address above

            Methylene  chloride (MeCl2) has been widely used  in RCRA/CERCLA  monitoring
      as an extraction  solvent for extraction of organic  contaminants  in ground water
      and solid waste  samples. The testing protocol  using methylene chloride  liquid-
      liquid extraction is wasteful  in time for preparation  of analytical samples and
      also costly in disposing  the waste solvent.  Environmental  Monitoring  Systems
      Laboratory-Cincinnati (EMSL-Cincinnati) of USEPA has initiated a research project
      to develop  alternative extraction methods to replace methylene chloride.  One
      alternative method for extraction  of organic pollutants from water  is a  liquid-
      solid extraction  (LSE).   This method  uses  solid  sorbents  (i.e.  C18) to trap
      analytes  and uses  solvents to elute the adsorbed analytes.  The LSE method offers
      low solvent consumption (i.e. 15-60*mL of MeCl2 compared to  >400 mL of MeCl2 in
      the  liquid-liquid extraction) and  less  waste.   The EMSL-Cincinnati  is also
      investigating a novel  technique, analytical scale supercritical  fluid  extraction
      (SFE), to replace the methylene  chloride  solvent  elution.  The SFE uses  a fluid
      (i.e. C02) that  is  non-toxic, non-combustible, chemically  inert,  and easy to
      discard.    For the  LSE method,  water samples  are  first  passed through 47-mm
      diameter  Empore™  C18 disks.   Then,  different amounts of MeCl, are used to elute
      the adsorbed analytes from the disks.  An  amount of 15  mL MeCT2 is used for clean
      samples without particulates, while at  least 60 mL of MeCl2  are  required for
      dirty  samples containing  particulates.    These  MeCl2  extracts  need  further
      concentration before  injecting  into a capillary  column  gas chromatograph for
      trace analysis.   The  use  of SFE  as  an  alternative  elution  technique to the
      solvent elution of LSE is more attractive.  After filtering the analytes from the
      water by  using the C18 disk,  the  adsorbed  analytes  are  eluted with supercritical
      CO, (about 30  mL)  ana collected  in  a small amount of solvent (1-2 mL).   The SFE
      method combines  the elution and  the concentration  procedures into  one step.  In
      comparison  with   the   solvent elution  of  LSE,  the  SFE  reduces  the  solvent
      consumption by 90% and generates much less waste.   Detailed discussions  of the
      recovery  efficiencies, relative  standard deviation  (RSD) and detection limits as
      well as the effects  of various solvent modifiers  on the SFE  extraction  will be
      presented at symposium. The disk liquid-solid extraction (LSE) coupled with the
      supercritical  fluid  elution (SFE)  provides  clear  advantages  in   terms  of
      extraction  time,  solvent used,  and solvent  disposal especially  for  samples
      containing particulates like  ground water sample.
                                         337

-------
45                    COMPARISON OF SOLID PHASE EXTRACTION WITH
              SALTING-OUT SOLVENT EXTRACTION FOR PRECONCENTRATION OF
                 NITROAROMATIC AND NITRAMINE EXPLOSIVES FROM WATER

      Thomas F. Jenkins and Paul H. Miyares, U.S. Army Cold Regions Research and Engineering
      Laboratory, Hanover, New Hampshire 03755-1290, and Karen F. Myers, Erika F. McCormick,
      and Ann B. Strong, U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi
      39180


      ABSTRACT

      Residues of high explosives are one of the most significant pollution problems at DOD facili-
      ties, Recently, the EPA has lowered the concentration at which these compounds are thought to
      be harmful to human health. Because TNT, RDX, and HMX, as well as several manufacturing
      impurities and environmental transformation products, are quite mobile in the soil and have
      caused groundwater pollution, there is an increasing demand for low-level analysis of these
      compounds in groundwater from installation boundary wells.

      RDX and HMX are very polar, and normal liquid-liquid extraction/preconcentration techniques
      result in poor recovery. Two innovative preconcentration techniques have been reported that
      appear  to offer improved recovery and adequate preconcentration: solid-phase extraction
      (SPE), and salting-out solvent extraction (SOE). This paper compares cartridge-SPE, mem-
      brane-SPE, and SOE using a series of reagent-grade water samples fortified with low concentra-
      tions of  11 nitroaromatics  and nitramines and a set of groundwater samples from an explosives-
      contaminated DOD facility.

      Results  indicated that  the three methods were comparable with respect to low-level detection
      capability, which ranged from 0.05 to 0.30 (J-g/L. Percent recoveries were generally greater than
      80% except for HMX and  RDX by membrane-SPE, which were consistently lower. Interferences
      were found in extracts from about half of the groundwater samples preconcentrated using the
      two SPE procedures but were not found in any of the extracts from the SOE. These interferences
      were traced to matrix interaction of the groundwater with the polymeric resins.


      INTRODUCTION

      Batch liquid-liquid extraction (LLE) has often been used to extract organic analytes from water
      prior to  analysis. The success of this approach depends upon favorable solvent/water partition
      coefficients (K ) for the analytes of interest. Since many organic pollutants are relatively non-
      polar, batch extraction  with an immiscible organic solvent has been successful.

      TNT (2,4,6-trinitrotoluene)  and RDX (hexahydro-l,3,5-trinitro-l,3,5-triazine), along with
      several of their manufacturing impurities and environmental degradation products, have been
      observed in groundwater  at a number of Army installations [1-5]. Health advisories have been
      issued by the EPA for many of these compounds (Table  1), with drinking water criteria at sub-
      part-per-billion concentrations. For this reason, the Army has developed analytical methods
      to determine these compounds at trace levels in water. To detect these compounds at the sub-ppb
      level, the first step has  been preconcentration of the analytes using an extraction process.
                                                338

-------
                                               Table 1. Proposed drinking water criteria
                                               for nitroaromatics and nitramines.
                                                            Proposed  drinking
                                                               water limit
                                               Compound
Reference
Classical LLE of these analytes has been report-
ed, but poor extraction efficiency has been found
for RDX and HMX. Octanol/ water partition co-
efficients (Kou.) for HMX and RDX are 1.15 and
7.24  [6], respectively, indicating the compounds
are quite polar and will be difficult to extract
from water using nonpolar solvents.

For this reason, several alternative procedures
were developed. The most successful of these
have been the  use of  solid-phase  extraction
(SPE) [13, 14]  and salting-out liquid-liquid  ex-
traction (SOE)  [15-17].

The objective of the work described here was to
obtain a direct comparison of the cartridge SPE
[13]  and membrane SPE methods [14] with  the
SOE method [17]. Aqueous test solutions of nitro-
aromatics and nitramines prepared in ground-
water without  the use  of organic solvents allow-
ed a realistic comparison at  concentrations of
analytes  below  the proposed drinking water
criteria  (Table 1). In  addition,  all three  pro-
cedures were tested with a set of groundwater samples from a military toxic waste site known to
be contaminated with explosives.
HMX
RDX
TNT
2,4-DNT
2,6-DNT
1,3-DNB
400*
2.0*
2.0*
50"
0.1
40**
0.007
1.0*
171
18]
P]
[10]
[H]
[10]
[11]
[12]
                                                   EPA Lifetime Health Advisory Number
                                                   EPA number for increased cancer risk of
                                                   1.0x10-6.
EXPERIMENTAL

Salting-out solvent extraction/nonevaporative preconcentration procedure

This method was based on the salting-out solvent extraction procedure [17] in combination with
the nonevaporative preconcentration technique [18] reported elsewhere. A 251.3-g portion of
reagent-grade sodium chloride was added to a 1-L volumetric flask along with a 770-mL sam-
ple of water. A stir bar was added, and the contents were stirred at maximum speed (rpm) until
the salt was completely dissolved. A 164-mL aliquot of ACN was added and stirred for 15 min-
utes. The stirrer was turned off and the phases allowed to separate. The ACN phase (about 8
mL) was removed, and 10 mL of fresh ACN were added. The flask was again stirred for 15 min-
utes, followed by phase separation. The ACN was removed and  combined with the initial
extract. The combined extract was placed in a 100-mL volumetric flask and 84 mL of salt water
(325 g  NaCl per 1000 mL of water) was added, and the contents were  stirred for 15  minutes.
After allowing the phases  to separate, the ACN phase was carefully removed and placed in a
10-mL  graduated cylinder. An additional 1.0-mL aliquot of ACN was then added to the volu-
metric  flask and the contents were stirred. Again the phases were allowed to separate, and the
resulting ACN phase was added to the 10-mL graduated cylinder. The resulting extract, about
5-6 mL, was than diluted 1:1 with reagent-grade water prior to analysis  [19].
                                          339

-------
Cartridge solid-phase extraction

Porapak R (80/100 mesh) was obtained from Supelco and precleaned by Soxhlet extraction using
ACN [19]. After precleaning, the material was air-dried briefly and then oven-dried at 105°C
for 2 hr.

Empty 3-mL extraction cartridges were packed with 0.5 g of the precleaned Porapak R. The car-
tridges were placed on a Visiprep Solid-Phase Extraction Manifold (Supelco) and conditioned
by eluting with 30 mL of ACN followed by 50 mL of reagent-grade water. A 500-mL aliquot of
each water sample was pulled through a cartridge at about 10 mL/min.

The cartridges were eluted by passing a 5-mL aliquot of ACN through the cartridge at about 2
mL/min, and the eluate was collected in a 10-mL graduated cylinder. The resulting extract,
about 5 mL, was diluted 1:1 with reagent-grade water prior to analysis [19].


Membrane solid-phase extraction

The 47-mm Empore styrene-divinyl benzene disks were precleaned by soaking them in 50 mL of
ACN [19]. Each membrane was soaked for four 24-hour periods with fresh ACN. Each disk was
rinsed with ACN, then centered on a 47-mm vacuum filter apparatus and leached with a 20-mL
portion of ACN and a 50-mL aliquot of reagent-grade water. Just before the last of this water
was pulled through the membrane, the vacuum was stopped and the reservoir was filled with
sample. The vacuum was then turned on again and a 500-mL aliquot of a water sample was
pulled through the membrane. This took from 5 to 7 minutes, with resulting flow rates ranging
from 70 to 100 mL/min. Air was then drawn through the membrane for 1 minute to remove excess
water. A 5-mL aliquot of ACN was then used to extract the analytes from the disk, and the
extract was diluted 1:1 with reagent-grade water prior to analysis [19].


RP-HPLC analysis

All analyses were conducted using reversed-phase HPLC as described elsewhere [19,20].


Preparation of standards and samples

Since we felt  it was important that test samples be completely free of organic solvents, we pre-
pared all test samples in a totally aqueous matrix. This was done by preparing individual
aqueous  analyte stock solutions by placing a few hundred milligrams of Standard Analytical
Reference Material (SARM) of  each specific analyte in individual 4-L brown glass bottles, fill-
ing the bottles with reagent-grade water, adding a stir bar, and  stirring for two days at room
temperature.

Each solution was filtered through a 0.45-mm nylon-66 membrane (Supelco) into a clean brown
glass bottle. Aliquots of each solution were then analyzed against standards prepared in ace-
tonitrile to estimate the concentration of analyte in each aqueous stock solution.
                                           340

-------
 RESULTS AND DISCUSSION

 Certified reporting limit test

 Spiked test solutions were used to
 conduct a four-day Certified Report-
 ing  Limit Test  (CRL) as described
 elsewhere [19]. This test allows a
 comparison of low-concentration de-
 tection capability, percent recovery,
 interferences, and overall precision.
 A discussion of the CRL and how re-
 sults compare with the EPA Method
 Detection Limit are presented else-
 where [21]. Chromatograms for sam-
 ples containing  target analytes at
 about 0.2 ug/L and preconcentrated
 by the three procedures are present-
 ed in Figure 1.

 The CRLs obtained are  shown in
 Table  2.  Overall,  the CRLs  for  a
 given  analyte are quite  similar for
 all  three preconcentration  tech-
 niques. None of the procedures is
 consistently superior to the others in
 low-concentration detection capabil-
 ity. CRL values range from a low of
 0.032 ug/L for DNB using cartridge
 SPE to a high of 0.83 ug/L for tetryl
 using membrane-SPE. All values for
 HMX,  RDX, TNT, 2,4-DNT, and 1,3-
 DNB are below the proposed drink-
 ing water limit for these compounds.
 While  2,6-DNT was not tested, it is
 very unlikely that a CRL as low as
 the proposed value of 0.007  ug/L
 would be obtained. The only  CRL
 value that appears to be out of line
 is the 0.83 ug/L value of tetryl using
 membrane-SPE. Inspection of the da-
 ta indicates this  high CRL is due to
 low recovery on one of the four days.
                                Salting-out
                                2X Sample
                         Membrane SPE (Empore SDVB)
                               2X Sample
                           Cartridge SPE (Porapak R)
                                2X Sample
                      Retention Time (min)

Figure  1.   LC-18  chromatograms  for  2X  sample
preconcentrated  by  salting-out,  cartridge-SPE,  and
membrane-SPE methods at CRREL.
The regression line obtained from the plot of found vs. spiked concentrations was also examined
for curvature using lack-of-fit testing.  Linear relationships described the data at the 95%
confidence level. The slopes of these linear regression lines are measures of the overall percent
recoveries of these analytes  using each preconcentration method. Recoveries (Table  3), in
general, are quite good (near a theoretical value of 100%). Measured recoveries for the salting-
out procedure range from 93-119%. Likewise, recoveries for cartridge-SPE and membrane-SPE
range from 83-133% and 81-116%, respectively.
                                          341

-------
Table 2. Certified reporting limits for various  Table 3. Overall percent recovery and relative
preconcentration techniques.                    standard deviation (RSD) from Certified Re-
                                              porting Limit Test.
Analyte
          CRL (fig/L)
Salting-    Cartridge- Membrane-
  out         SPE        SPE
                                                                 % recovery* (% RSD)
                                                          Salting-    Cartridge-  Membrane-
                      Analyte
                                                            out
                                    SPE
SPE
HMX
RDX
TNB
DNB
Tetryl
TNT
2-Am-DNT
2,4-DNT
o-NT
p-NT
m-NT
 0.19
 0.13
 0.052
 0.081
 0.20
 0.086
 0.10
 0.083
 0.13
 0.22
 021
021
027
0.042
0.032
0.24
0.068
0.046
0.085
0.10
0.12
0.13
0.33
0.12
0.051
0.036
0.83
0.13
0.055
0.044
020
0.23*
0.37
HMX
RDX
TNB
DNB
Tetryl
TNT
2-Am-DNT
2,4-DNT
o-NT
p-NT
m-NT
106 (10.5)
106 (8.7)
119 (7.6)
102 (6.6)
93 (16.4)
105 (7.6)
102 (9.1)
101 (5.8)
102 (9.1)
96 (18.1)
97 (12.4)
107 (9.6)
116 (22.0)
133 (8.7)
115 (2:6)
83 (32.8)
111 (7.5)
113 (4.1)
109 (6.8)
107 (8.1)
104** (6.6)
100 (7.3)
81 (14.0)
116(11.1)
116 (10.3)
103 (6.5)
83 (46.4)
97 (10.5)
103 (8.9)
94 (6.6)
92 (15.6)
89t (18.0)
86 (17.2)
   One outlier removed for this analyte/method
   combination.
                                      Slope of regression line of spiked concentration
                                      vs. found concentration x 100.
                                      Lack-of-fit test indicates data not adequately
                                      described by linear relationship at the 95%
                                      confidence level.
                                      One outlier removed for this analyte/method
                                      combination.
 Comparison using groundwater samples
 A further test was conducted using 58 groundwater samples from the Rockeye site at the Naval
 Surface Warfare Center.  All of the samples were analyzed by the direct method and after pre-
 concentration as well using the three procedures (SOE, cartridge-SPE, and membrane-SPE).
 Table 4 summarizes the  results for samples where the concentrations of at least one nitroaro-
 matic or nitramine analyte was high enough to be obtained using the direct method. Since the
 concentrations obtained by the direct analysis are subject to fewer sources of error than those
 obtained using preconcentration, we are treating these values as "true" values for purposes of
 comparison. We can then compare  the results from the various preconcentration techniques rela-
 tive to these "true" values.

 Examination of Table 4 indicates that all three preconcentration procedures did a fairly good
 job of recovering these  analytes (generally greater than 80%). The membrane-SPE method,
 however, recovered less  of HMX and RDX than the other two procedures. Recovery of HMX,
 RDX, and TNT by the cartridge-SPE and salting-out methods was greater than 80% in all cases,
 with a slightly better recovery for  the cartridge-SPE method for HMX and RDX.

 For about half of the samples, the chromatogram for the salting-out method is blank with
 respect to target analytes  and interferences, but the chromatograms for both the two solid-
 phase methods, at the same attenuation, show large peaks at a number of retention times across
 the entire chromatogram (Figure 2).  Second-column confirmation  indicated that none of these
 peaks resulted from the presence of nitroaromatic or nitramine explosives. Their presence, how
                                           342

-------
Table 4. Comparison of results for direct analysis of ground-
water samples from the Rockeye site at the Naval Surface
Warfare Center, Crane, Ind., with the three preconcentration
methods.
                           Concentration  (ng/L)

Sample  Method*   HMX   RDX    TNB    TNT   4A
ever, interfered with the abil-
ity to detect nitroaromatic and
nitramine analytes at concen-
trations well above the CRLs.
These large interference  peaks
were not found  when  these
groundwater  samples  were
analyzed without preconcen-
tration and hence appear to be
introduced by the solid phases
themselves. Since the Porapak
R and  the SDVB membranes
were cleaned  separately and
all the Porapak R cartridges
were  packed  from material
cleaned in the same batch, we
do  not believe these peaks
were a result of poorly cleaned
material.  Rather  it appears
that some component of these
samples interacted with the
solid phases to either degrade
the polymer or release contam-
inants  from within the  poly-
mer by either swelling or col-
lapsing the polymer matrix.


CONCLUSIONS AND
RECOMMENDATIONS

c,-.c   j   .. • j      j         * Membrane-SPE (SPE-M), cartridge-SPE (SPE-C), and salting-out
SOE and cartridge- and  mem-   (CQP)
brane-SPE were  compared
with respect to their ability to
preconcentrate nitroaromatic and nitramine explosives from water prior to RP-HPLC analysis.
Both fortified reagent-grade water and contaminated groundwater samples were used in this
assessment.

Low detection capability and overall precision were comparable among the three procedures.
Recoveries of HMX and RDX are better using cartridge-SPE and SOE than membrane-SPE. Re-
covery of the nitroaromatics was acceptable for all three procedures.

At present the major problems associated with the use of the SPE procedures are the inadequacy
of the current cleaning procedures. While use of the Soxhlet procedure on a batch basis for the
Porapak R material appears to be adequate, cleaning must be accomplished on-site just before
use or the contamination  reappears. The cleaning procedure we used for the membranes is cum-
bersome and wastes solvent.

Problems with interferences were encountered using both the cartridge-SPE and membrane-SPE
procedures for a number of actual groundwater samples. These interferences appeared to be iden
20649



20650



20662



20663



20667



Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
Direct
SPE-M
SPE-C
SOE
151
98
156
161
119
60
107
98
26
19
17
22
281
153
214
232
318
199
356
319
135
121
147
138
82
64
85
66
160
176
138
154
94
89
109
90
618
488
666
558
33
32
34
38
9.0
7.6
92
10.3
42
51
34
46
21
22
26
26
19.2 284
19.5 317
19.6 328
18.6 320
9.6
112
122
13.7








65
75
78
78
166
216
239
217
           343

-------
                               Sa»ing-out
                               Cartridge-SPE
                               Membrane-SPE
                    a        12
                   Retention Time (min)
                                       16
                                                20
Figure 2. LC-18  chromatograms for sample 20641 from
Rockeye site, Naval  Surface  Warfare  Center, precon-
centrated by salting-out, cartridge-SPE, and membrane-
SPE methods at WES.
tical to the compounds released from
the solid phases during cleaning, but
were apparently released from the
SPE phases due to a matrix interac-
tion with a number of groundwater
samples. The nature of  this inter-
action  is still unclear. These  com-
pounds would interfere with deter-
mination of nitroaromatic and nitra-
mine explosives even at reasonably
high concentrations.

The elimination of the need to use
evaporative preconcentration  with
the salting-out procedure is a major
improvement.  We believe the new
procedure will be more precise and
less subject to error in routine use
than the initial method, which uti-
lized a Kudena-Danish evaporator.

In  summary,  the SOE,  cartridge-
SPE, and membrane-SPE preconcen-
tration techniques are all  capable of
providing adequate analyte precon-
centration of nitroaromatics and ni-
tramines prior to RP-HPLC deter-
mination. Of  the three,  the  SOE
method appears to be least prone to
interferences.  The  membrane-SPE
method requires  the least sample-
processing time, but its recovery of
HMX is the poorest  of  the three
methods. The cartridge-SPE method
requires the least solvent per  sam-
ple, but its routine use  with the
currently available processing mani-
fold appears prone to problems with
cross contamination [19].


ACKNOWLEDGMENTS

The authors would like to acknowl-
edge Dr. C.L. Grant, Professor Emeri-
tus, University of New Hampshire,
and M.E. Walsh of the U.S. Army
Cold Regions  Research  and Engi-
neering Laboratory for useful  com-
ments and suggestions on the manu-
script. In addition, Roy Wade of the
                                         344

-------
U.S. Army Engineer Waterways Experiment Station (WES) is acknowledged for groundwater
samples collection, and S.  Paige Pitts and Allyson H.  Lynch of WES for assistance In
groundwater extractions. The authors also  thank Dr. E.S.P. Bouvier of the Waters Chromatog-
raphy Division, Millipore, Inc., for useful discussions regarding the nature of the interferences
observed  for the solid-phase extraction materials. Funding for  this research was provided
jointly by the U.S. Army Environmental Center, Aberdeen Proving Ground, Maryland, Martin
H.  Stutz, Project Monitor,  and the U.S.  Army Engineer Waterways  Experiment  Station,
Vicksburg, Mississippi, Ann B. Strong, Project Monitor.


LITERATURE CITED

[1]   D.L. Pugh, Milan Army Ammunition Plant Survey, U.S. Army Toxic and Hazardous Mate-
     rials Agency, Report DRXTH-FS-FR-82131, Aberdeen Proving Ground, MD, 1982.
[2]   D.H. Rosenblatt, "Contaminated Soil  Cleanup Objectives for Cornhusker Army Ammuni-
     tion Plant," U,.S. Army Medical Bioengineering Research and Development Laboratory
     Technical Report 8603,  Fort Derrick, MD, 1986.
[3]   R.F. Spaulding and J.W. Fulton, "Groundwater munition residues and nitrate near Grand
     Island, Nebraska,  USA," Journal of Contaminant Hydrology, 2:  139-153, 1988.
[4]   M.P. Maskarinec, D.L. Manning, and  R.W. Harvey,  "Application of Solid  Sorbent Collec-
     tion Techniques and High Performance Liquid Chromatography with Electrochemical
     Detection to the Analysis of Explosives in Water Samples," Oak Ridge National  Labora-
     tory, Report TM-10190, Oak Ridge, TN, 1986.
[5]   D. Layton, B. Mallon,  W. Mitchell, L. Hall, R. Fish, L. Perry, G. Snyder, K. Bogen, W.
     Malloh, C. Ham, and P. Dowd, "Conventional Weapons Demilitarization: A Health and
     Environmental Effects Data Base Assessment," Lawrence Livermore Laboratory,  CA,
     1987. AD-A220 588.
[6]   T.F. Jenkins, "Development of an Analytical Method for the  Determination of Extract-
     able Nitroaromatics and Nitramines  in Soils," University of New Hampshire, Durham,
     Ph.D.  dissertation, 1989.
[7]   Environmental Protection Agency,  Health Advisory for HMX, Criteria  and Standards
     Division, Office of Drinking Water, Washington, DC, November, 1988a.
[8]   Environmental Protection Agency,  Health Advisory for RDX, Criteria  and Standards
     Division, Office of Drinking Water, Washington, DC, November, 1988b.
[9]   Environmental Protection Agency,  Trinitrotoluene Health Advisory, Criteria and Stan-
     dards Division, Office of Drinking Water, Washington, DC, January, 1989.
[10]  Environmental Protection Agency,  Dinitrotoluene Health Advisory, Criteria and Stan-
     dards Division, Office of Drinking Water, Washington, DC, 1992.
[11]  E.L. Ernier, "Water Quality Criteria for 2,4-dinitrotoluene and 2,6-dinitrotoluene,"  Oak
     Ridge National Laboratory Report AD-ORNL 6312,  Oak Ridge, TN,  1987.
[12]  Environmental Protection Agency, Health Advisory for  1,3-dinitrobenzene, Criteria  and
     Standards Division, Office of Drinking Water, Washington, DC, January,  1991.
[13]  M.G. Winslow, B.A. Weichert, and R.D. Baker, "Determination of Low-Level Explosive
     Residues in Water by HPLC: Solid-phase Extraction vs. Salting-out Solvent Extraction,"
     in Proceedings of the EPA 7th  Annual Waste Testing and Quality Assurance Symposium,
     July 8-12, 1991, Washington, D.C., 1991.
[14]  D.F. Hagen, C.G.  Markell, and G.A. Schmitt, "Membrane  Approach  to  Solid-phase
     Extraction," Analytica  Chimica Acta, 236:  157-164, 1990.
[15]  D.C. Leggett, T.F. Jenkins and  P.H. Miyares, "Salting-out Solvent  Extractions for  Precon-
     centration  of Neutral Organic Solutes from Water," Analytical Chemistry, 62(13): 1355-
     1356, 1990.
                                          345

-------
[16]  P.H. Miyares and T.F. Jenkins, "Salting-out Solvent Extraction Method for Determining
     Low Levels of Nitroaromatics and Nitramines in Water," USA Cold Regions Research
     and Engineering Laboratory, Special Report 90-30,1990.
[17]  P.H. Miyares and T.F. Jenkins, "Improved Salting-out Extraction/Preconcentration Meth-
     od for  the  Determination of Nitroaromatics and Nitramines in Water," USA Cold
     Regions Research and Engineering Laboratory, Special Report 91-18,1991.
[18]  T.F. Jenkins and  P.H. Miyares, "Non-evaporative Preconcentration Technique for
     Volatile and Semivolatile Solutes  in Certain Polar Solvents," Analytical Chemistry,
     63(13): 1341-1343,1991.
[19]  T.F. Jenkins, P.H. Miyares, K.F. Myers, E.F. McCormick, and A.B. Strong, "Comparison of
     Cartridge and Membrane Solid-phase Extraction with Salting-out Solvent  Extraction for
     Preconcentration of Nitroaromatic and Nitramine Explosives from Water," USA Cold
     Regions Research and Engineering Laboratory, Special Report 92-25,1992.
[20]  T.F. Jenkins, M.E.  Walsh, P.W. Schumacher, P.H. Miyares, C.F. Bauer, and C.L. Grant,
     "Liquid Chromatograph Method for Determination of  Extractable Nitroaromatic and
     Nitramine Residues in Soil," Journal of the Association of Official Analytical  Chemists,
     72(6): 890-899,1989.
[21]  C.L. Grant, A.D. Hewitt, and T.F. Jenkins, "Experimental Comparison of EPA and USA-
     THAMA Detection and Quantitation Capability Estimators," American Laboratory, pp.
     15-33, February, 1991.
                                          346

-------
            DINOSEB ANALYSIS IN THE FIELD AND THE LABORATORY


D.  Anderson.   APPL,  Fresno CA   93722;  F.  Tsang,  T.  Jackson,  and  P.
Marsden, SAIC, San Diego, CA  92121


ABSTRACT

Dinoseb,  6-sec-butyl-2,4-dinitro  phenol,  is  a herbicide  and  a plant
desiccant.    The  multiple  functional  groups  of  Dinoseb  provide  both
opportunities  and challenges  to  the  analytical chemist.    Dinoseb   is
generally measured using Method 8150/8151; unfortunately,  the method  is
time consuming and requires  the use  of diazomethane (an explosive and a
carcinogen).   In addition,  several Dinoseb-specific difficulties  are not
fully explained  in Method 8150 (which was developed as  a multianalyte
method  for phenoxyacid herbicides).    As  a result,  engineers and field
staff have had the  experience of seeing the characteristic yellow  color of
Dinoseb  on site  while  the  laboratory receiving  samples  reports  non-
detects.   The performance of  alternate,  more  specific,  techniques  for
analyzing  for Dinoseb  are  described,  including  a colorimetric field
screening  procedure.   This  screening procedure produces  a distinctive
yellow  color  on  Florisil™  and  has  been  used  to  measure  Dinoseb
concentrations of  5  ppm in  soil.   Techniques  for  improving laboratory
analysis of Dinoseb  such as  HPLC  are  also discussed.   Optimized sample
preparation procedures for Dinoseb are presented including derivatizing
samples  using diazomethane  and alternate  extraction techniques  allowed
under Method  8150.  Finally,  the conversion of Dinoseb to reduced (amino
substituted)   analogs will be  discussed.  Dinoseb  analyses offer a chemist
a  tremendous  opportunity to  design a  site-specific analytical program.
This presentation  will provide guidance  and performance data  to those
interested in designing  the  most  appropriate and cost-effective Dinoseb
monitoring program.
INTRODUCTION

Dinoseb,  2-(1-methylpropyl)-4,6-dinitrophenol  is  a plant  desiccant and
defoliant.   Prior  to its ban by  EPA,  dinoseb  was used on  a variety of
crops including potatoes and cotton.  Dinoseb residues are often visible
at  a  site because  of its  intense  yellow color  and its  propensity to
crystalize out of soil.   Standard analytical practice is to send samples
to  a  laboratory for  herbicide  (Method 8150)  or  GC/MS analysis  (e.g.,
Method 8270).  This  can  greatly  increase  the time to make a remediation
decision and does not necessarily improve  the data available to the site
manager

This  paper  describes alternate,  more  rapid  analytical  techniques for
dinoseb.  These  include a  field  screening procedure,  a  smaller  volume
                                   347

-------
laboratory extraction procedure,  direct GC  (non-derivatized) analysis and
an HPLC procedure for Dinoseb.  In addition, some guidance is offered in
extracting and derivatizing  samples  when Method 8150/8151 are employed.
The  results  were  generated  by Agricultural  and  Priority  Pollutant
Laboratory  of   Fresno,   CA   and  Science  Applications  International
Corporation in a number of studies for commercial and government clients.
Support for comparison of the swirling methanol extraction with Soxhlet
extraction (Method 3540)  was  provided by the Methods Section of the US EPA
Office of Solid Waste (OSW) and the Risk Reduction Engineering Laboratory
(RREL).   The  OSW study  involved generation of performance data  for 27
substituted phenols from spiked soils using both extraction techniques and
is reported in a  poster  presentation at  this symposium.   The  RREL study
determined the  recovery  of spiked dinoseb  and  a  surrogate (bromoxynil)
from  soil and  compared  in   situ contamination of  Dinoseb  using  both
extraction techniques.
DINOSEB SCREENING PROCEDURE

Dinoseb has  a bright yellow  color that can  be  used to  screen  for the
compound.  Dinoseb  is extracted from soil  using  methylene chloride in a
scintillation vial.  The sample is agitated and the solids are allowed to
settle and the methylene chloride  is transferred to a second scintillation
containing a small amount of Florisil.  The Florisil adsorbs the dinoseb
and presence of dinoseb  is indicated by yellow color.
PREPARATION OF DINOSEB AS A PHENOL

Dinoseb data was  developed  during a performance study of several SW-846
methods including Method 8041  (phenols).  Although a total of 27 phenols
were evaluated,  data presented in this report  is  limited to mixture #1
which contains dinoseb.

Each of the  phenols were spiked into a clean,  sandy  loam soil  from San
Diego.  Each  5  g aliquot was  spiked  separately,  then extracted using a
Soxhlet  apparatus   (Method  3540)  or  the  aqueous  methanol  leaching
procedure.  Methylene chloride was used as  the  solvent  in Method 3540;
acidic aqueous methanol  (1:1 methanol/pH  2  water)  was used to leach the
phenols from soil on a orbital shaker (150  rpm).  Following extraction,
methylene chloride  extracts  (Method 3540) were  exchanged to methanol so
that all  sample  extracts could be treated  in the  same fashion.  . Sodium
chloride,  distilled water and hydrochloric acid  was added to  each extract
so that it was primarily aqueous.  Phenols were extracted using 3 X 3 mL
portions of methylene chloride.  The methylene chloride was  concentrated
under nitrogen and analyzed by GC/FID.

A subsequent study for RREL evaluated the  use of HPLC  for  the analysis of
dinoseb.  Reversed phase HPLC is  an attractive complement  to  the methanol
                                   348

-------
leach procedure because no solvent exchange is required prior to analysis.
Five replicates of sample spiked at 1, 5 and 25 /^g/mL were  extracted using
pH 2 methanol to demonstrate recovery using the methanol  procedure.   The
suitability of spiking dinoseb into soil on the same day  of  analyses  was
supported by  comparison  of  recoveries  1 and  2 days  after spiking.   The
recovery  of dinoseb  was virtually  the same  when spiked  samples were
extracted and analyzed 1 and 2 days after fortification.

The RREL  study also  included  a comparison  of Soxhlet extraction and  the
methanol  leaching procedure  using  a   single  sample  contaminated with
dinoseb.  Five 10.0 g aliquots of  the sample were placed in an Erylenmyer
flask.    The   surrogate  bromoxynil  was  spiked  into  each  sample   and
thoroughly mixed.  A volume  of 10  mL pH  2 methanol was  added  to the flask
to extract the phenols.  Flasks containing the  five replicate samples  and
the lab blank were placed on a shaker table and swirled for 25 minutes at
150 RPM.  The 10 mL methanol was pipetted off  and an additional volume of
5 mL pH 2 methanol was added to  the sample.   The shaker table extraction
was  repeated.    The  two portions of  methanol were  combined for HPLC
analysis.   Five  additional  aliquots of  the  sample  were extracted by  the
Soxhlet procedure  (Method 3540)  using methylene chloride.   The methylene
chloride was exchanged to methanol prior to analysis.
PREPARATION OF DINOSEB USING METHOD 8150

Dinoseb is an analyte for Method 8150/8151 (phenoxyacid herbicides).  In
some  cases,  the  acid/base  partitions  specified  in  the  method  may  be
required to achieve adequate cleanup  or  to meet project DQOs.   However,
analysts are cautioned that  dinoseb may be lost when a 30 g sample  is used
because of the increased sample handling and solvent transfers required.
Loss of dinoseb can be minimized by reducing the  sample size, the number
of  solvent  transfers,  and completing the  preparation as quickly as  is
prudent (i.e., there  is no  convenient place  to stop Method  8150/8151  is
used  for  dinoseb) .    Acid washing  of all glassware will  also   improve
recovery of dinoseb.
GC ANALYSIS

After  acquisition  of  the  target  chemicals  for   the  OSW  study,  gas
chromatographic  (GC)   conditions  were  evaluated  in order  to  achieve
separations of  the  target analytes and  to  establish nominal  low level
calibration  concentrations.     GC  analyses   were   performed   on  both
underivatized phenols and phenols derivatized  using diazomethane.   Data
for the underivatized  phenols  only  are reported here as  this  report is
focused  on rapid  techniques  for  the  analysis  of  dinoseb.    The  GC
conditions recommended for  the direct analysis  of  substituted phenols,
including Dinoseb,  are:
                                   349

-------
                  GC CONDITIONS

Column:                 DB-5 30m x 0.53 mm id
Initial Temperature:                80°C
Program:                      Hold 1.5 minute
                              6<>C/min to 230
                              10oC/min to 275
Final Hold:                   4.5 rain
Run time:                     35 min
Detector:                     FID, 300°C
Carrier gas:                  Nitrogen, 6 mL/min
Hydrogen:                     30mL/min
Total nitrogen:         30 mL/min (carrier and makeup)
Injector:         1/4 inch Packed w/ megabore liner, at  200°C
Initial five-point  GC/FID  calibration established the linear portion of
instrument  response for  each of  the phenols  including Dinoseb.   The
concentrations  of the  calibration solutions,  retention  times  (Rt) and
relative standard deviations of the individual  response factors  (RSD) for
the mixture containing Dinoseb are  provided in Table 1. The later-eluting
analytes in mixture  1  generally had higher RSDs  (18-21%) and more narrow
calibration ranges  (10-20  fold) than  the other phenols.  The problems in
using GC to measure  underivatized  dinitro-, trichloro-, tetrachlor-, and
pentachlorophenols  are evident in  the IQL (Table 2) and recovery values
(Table 6 and 7).

Examination of the GC chromatograms of underivatized phenols reveals broad
and non-Gaussian peaks.   The use  of  a  temperature programmed injector
(Varian SPI) did narrow peaks  somewhat, methylation is the only reliable
technique for producing narrow phenol  peaks on GC.  Therefore,  if there is
significant  co-extracted  interferences  or  lower detection  limits are
required  for  the  GC  analysis of dinoseb,  sample  extracts  should be
derivatized.    Diazomethane  should  be used  for  derivatization,  boron
triflouride - methanol does not perform reliably  with dinoseb.
                                    350

-------
           TABLE 1   GC/FID Calibration of Phenols   Mix 1
Analyte
Phenol
2-Cresol
3-Cresol
2 , 4-Dimethylphenol
2 , 6-Dimethylphenol
2 , 3 -Dimethylphenol
3-Chlorophenol
4-Chloro-3-
methylphenol
2 , 3 , 5-Trichlorophenol
2 ,4 , 5-Trichlorophenol
2 , 5-Dinitrophenol
2 , 4-Dinitrophenol
2,3,5,6-
Tetrachlorophenol
4,6-dinitro-o-cresol
Dinoseb
Calibration
solutions /ig/mL
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
5,25,50,100,200
10,40,50,100,200
40,50,100,200,400
5,20,25,40,50
10,40,50,100,200
5,20,25,50,200
Rt) min
6.37
8.17
8.65
9.63
10.54
11.32
11.68
14.07
15.47
16.05
18.37
19.29
20.42
21.72
25.71
%RSD
2.5
3.4
4.9
3.1
5.8
6.1
12.3
14.8
18.2
18.3
15.3
20.8*
17.6
14.9
15.6
may require non-linear calibration
                                 351

-------
TABLE 2 - Phenol GC/FID IQL - Mix 1
Analyte
Phenol
2-Cresol
3-Cresol
2 , 4 - D ime thy Ipheno 1
2 , 6-Dimethylphenol
2 , 3 - Dime thy Ipheno 1
3 - Chlorophenol
4-Chloro-3-methylphenol
2 , 3 , 5-Trichlorophenol
2,4, 5-Trichlorophenol
2, 5-Dinitrophenol
2 ,4-Dinitrophenol
2,3,5,6-
Tetrachlorophenol
4, 6-dinitro-o-cresol
Dinoseb
Rt , min
6.37
8.17
8.65
9.63
10.54
11.32
11.68
14.07
15.47
16.05
18.37
19.29
20.42
21.72
25.71
cone,
/jg/mL
10
10
10
10
10
10
10
10
10
10
20
20
10
20
10
std dev
(s)
0.35
0.36
0.47
0.36
0.46
0.47
0.69
0.82
0.96
1.03
3.63
2.82
1.96
3.18
1.56
IQL 10s
Mg/mL
3.5
3.6
4.7
3.6
4.6
4.7
6.9
8.2
9.6
10.3
36.3
28.2
19.6
31.8
15.6
                352

-------
HPLC ANALYSIS

HPLC/UV  is  employed for the analysis  of dinoseb.   It  is an attractive
technique for rapid laboratory analysis because of the low detection level
for dinoseb, the compatibility with the methanol  leach procedure and  the
relatively poor chromatographic performance of underivatized dinoseb using
GC/FID.  A  study was  conducted  for  RREL to evaluate the practicality of
HPLC for the  analysis of bioremediation samples.   Dinoseb was analyzed
using  a Hewlett-Packard model  1090A  series  II Liquid Chromatograph
equipped with  a diode  array UV/VIS  detector  and a computerized data
system.  Dinoseb concentrations  in water, soil,  sediment  and-treated soil
were determined by  HPLC/UV with  a PSP-Spherisorb ODS2  column (5 ftm,  250 x
4 nun i.d.).   Dinoseb  and a surrogate  (bromoxynil) were  detected by the
diode array UV/VIS  detector at 254 nm.

A gradient  solvent  program was employed:  30% methanol and 70% 0.1% acetic
acid buffer for 2  minutes,  and  then the methanol  was increased  to 100%
over the  next 6 minutes and held  at  that  level for  2 minutes.   The
methanol was then decreased to 30% over  4 minutes  and held for 6 minutes.
The solvent flow rate  was 0.5 mL/minute  and the  column temperature was at
40°C.   The  pH  of  the mobile phase  was critical  as  the  chromatographic
behavior of Dinoseb and bromoxynil are very  dependent  on the ionization of
these phenols;  this  is  particularly true for the surrogate bromoxynil.
Total run time was  20 minutes.

Tables  3 and 4  provide six-point initial calibration data for dinoseb and
the surrogate  bromoxynil.   The  tables  provide  the concentration  of the
calibration solutions, the retention times of the target analytes,  peak
area counts from the data system, response factors,  and the percent RSD of
the response factors.
                                   353

-------
    TABLE 3 - Calibration of Dinoseb
Cone . /zg/mL
0.50
1.0
5.0
10.0
50.0
100.0
200.0
RT
14.47
14.48
14.47
14.48
14.52
14.55
14.55
Area Count
20.0
39.2
186
374
1862
3663
7199
Mean =
Std =
%RSD =
RF
0.02500
0.02551
0.02688
0.02674
0.02685
0.02730
0.02778
0.02658
0.00098
3.7
TABLE 4   HPLC Calibration of Bromoxynil
Cone . /ig/mL
0.50
1.0
5.0
10.0
50.0
100.0
200.0
RT
12.95
12.96
12.96
12.96
12.97
12.98
12.98
Area Count
21.8
44.3
221
447
2215
4394
8653
Mean =
Std =
%RSD =
RF
0.02294
0.02257
0.02262
0.02237
0.02257
0.02276
0.02311
0.02271
0.00025
1.1
                    354

-------
FIELD SCREENING RESULTS

The  dinoseb  screening procedure  provides a  rapid means  of monitoring
dinoseb in soil.   It can provide a detection level of  5  ppm and can be
used to establish the presence or absence of in situ dinoseb in the field.
The technique has been demonstrated to perform reliably at several sites
and has  reduced the  equipment  costs associated  with soil  removal.   A
normal practice for  the  removal of  dinoseb at  a contaminated site is to
"chase the color".   That  is,  to  remove soil until  all  of the yellow color
is gone.   Unfortunately,  this technique is quite  subjective  and is subject
to problems  when dinoseb  is  bound  within soil  particles.   A  site may
appear to be  clean at the end of the day but show color the  next day.  The
dinoseb screening procedure reduces the potential for mis-interpretation
of soil  color  by providing a  reliable  test  for  dinoseb  extracted from
soil.
PHENOL STUDY

Dinoseb can be  leached from soil using a variety  of solvents including
methanol,  methylene  chloride  and ether.   Studies  to  date  indicate that
leaching is as  effective  as  the  standard  extraction procedures (Methods
3540 and 3550)   However,  the leaching procedure may need to be tailored
to specific soil  types.   Studies at SAIC have  demonstrated that acidic
methanol provides the  best recoveries  from  Idaho  soil;  while studies at
APPL  demonstrate  that  an aqueous  sodium  sulfate/ether leach provides
better recoveries from California soil.  Additional studies will required
to establish the ruggedness of leaching procedures for dinoseb and other
substituted phenols in additional sample matrices.

GC/FID analysis of underivatized  phenols  demonstrate  that there  is  a
positive bias for the recoveries of the higher molecular weight species.
This was  particularly evident for  the samples extracted  using  Soxhlet
(Method 3540)  which are reported  in  Table 5  (i.e. , a recovery of 200% was
reported for  dinoseb).   Perhaps  the co-extracted  interferences  in soil
contributed  to  the  quantitation   difficulties  observed  using  GC/FID
analysis.

The difficulties of  analyzing underivatized phenols  was also observed with
the alternate  extraction  procedures.   The recoveries from  the methanol
leaching procedure  in Table  6  also reveal  a  consistent high bias  for
dinoseb and other phenols.  However, the recoveries of using the leaching
procedure  are generally closer to 100%  than those  obtained with  Soxhlet
extraction.    This  observation could support the supposition  that  a co-
extracted  material is  interfering with the GC analysis of phenols.  In any
case,  there appears  to be  the  potential for significant difficulties when
dinoseb and other phenols  are analyzed by GC/FID without derivatization.
                                   355

-------
TABLE 5   Recovery of Underivatized Phenols by GC and Method 3540
Analyte
Phenol
2-Cresql
3-Cresol
2 ,4-Dimethylphenol
2 , 6-Dimethylphenol
2 , 3 -Dime thy Iphenol
3 - Chlorophenol
4-Chloro-3-
methylphenol
2,3, 5 -Trichlorophenol
2,4,5- Trichlorophenol
2 ,5-Dinitrophenol
2 ,4-Dinitrophenol
2,3,5,6-
Tetrachloro'phenol
4, 6-dinitro-o-cresol
Dinoseb
Spiking cone
/*g/g
20
20
20
20
20
20
20
20
20
20
40
40
20
40
20
Recovery
(%)
93
95
98
93
101
106
116
128
136
139
177
157
236
201
210
% RSD
16.9
13.6
10.3
11.5
8.1
7.1
6.7
3.8
4.1
3.0
5.1
7.3
3.5
3.8
4.9
n
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
                                356

-------
            TABLE 6   Recovery of Phenols by Leaching and GC
Analyte
Phenol
2-Cresol
3-Cresol
2 , 4-Dimethylphenol
2 , 6-Dimethylphenol
2 , 3-Dimethylphenol
3 - Chlorophenol
4-Chloro-3-methylphenol
2,3, 5 -Trichlorophenol
2,4, 5 -Trichlorophenol
2 , 5-Dinitrophenol
2 , 4-Dinitrophenol
2,3,5 , 6-Tetrachlorophenol
4 , 6-dinitro-o-cresol
Dinoseb
Spiking
cone /ig/g
20
20
20
20
20
20
20
20
20
20
40
40
20
40
20
Recovery
(%)
94
97
99
89
82
101
111
119
127
134
123
131
204
166
169
% RSD
2.6
2.8
2.8
3.8
4.0
2.6
2.7
2.4
2.8
2.9
6.6
6.8
3.3
4.3
4.2
n
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
HPLC STUDY

HPLC was  evaluated  as  an alternate procedure for the  rapid  analysis of
underivatized dinoseb.   Use of HPLC analysis  allows  the  analysis of the
methanol leachate without derivatization  or solvent exchange.  The strong
UV  absorption  of dinoseb  provides  an opportunity  for a  low  detection
level.   An initial evaluation of the methanol leaching procedure combined
with HPLC analysis was  conducted by measuring the recovery of dinoseb and
proposed surrogate, bromoxynil,  from spiked samples.   The performance of
the method seems reliable based on the results provided in Table 7.

The methanol leach/HPLC technique was  tested against Method 3540/HPLC by
extracting  aliquots of  the same  contaminated  samples  using  methanol
leaching (Table 8)  and Soxhlet extraction (Table  9).  The results obtained
for dinoseb using  either extraction procedure are very  similar;  higher
recoveries  for  bromoxynil  were  obtained  using  Soxhlet  extraction.
Additional work may be  required to provide a more suitable surrogate.
                                   357

-------
   Table 7.   1 Jtg/g dinoseb spiked soil extracted with pH2 methanol
1 /zg/g spiked soil
RT
14.26
14.24
14.24
14.22
14.22
Area
37
38
36
31
32
Mean =
Std =
%RSD =
Cone .
1.1
1.1
1.1
0.9
0.9
1.0
0.091
8.9
% Rec.
108
111
106
91
94
102
9.1
8.9
5 ppra Bromoxynil
RT
11.51
11.49
11.48
11.45
11.47
Area
151
140
152
151
147
Mean =
Std =
%RSD =
Cone .
5.5
5.1
5.5
5.5
5.3
5.4
0.18
3.4
%Rec.
110
102
110
110
107
108
3.6
3.4
TABLE 8   Methanol Leach/HPLC - Five Replicates of a Contaminated Sample
sample #21 -Dinoseb cone. /ig/mL
RT
14.44
14.46
14.47
14.44
14.44
Area
1333
1758
1607
1722
1795
Mean =
Std =
%RSD =
Extract
Cone.
30.3
39.9
36.5
39.1
40.8
37.3
4.2
11.4
Sample
Cone.
45.4
59.9
54.7
58.7
61.1
56.0
6.4
11.4
5 ppm Bromoxynil recovery
RT
13.03
13.02
13.02
13.02
13.01
Area
116
194
198
148
146
Mean =
Std =
%RSD =
Cone .
2.3
3.9
4.0
3.0
3.0
3.2
0.7
21.8
%Rec.
46.9
78.5
80.1
59.9
59.0
64.9
14.1
21.8
                                   358

-------
  TABLE 9    Method 3540/HPLC    Five  Replicates of a Contaminated Sample
sample #21-Dinoseb cone. /ig/mL
RT
14.50
14.49
14.50
14.49
14.49
Area
1468
1912
1887
1530
1580
Mean =
Std =
%RSD -
Extract
Cone .
33.3
43.4
42.9
34.7
35.9
38.0
4.7
12.4
Sample
Cone.
50.0
65.1
64.3
52.1
53.8
57.1
7.1
12.4
5 ppm Bromoxynil recovery
RT
12.95
12.95
12.95
12.94
12.95
Area
111
266
267
239
256
Mean =
Std =
%RSD =
Cone .
5.6
5.4
5.4
4.8
5.2
5.3
0.5
5.5
%Rec.
112
108
108
96.7
104
106
5.8
5.5
CONCLUSIONS

In  our  experience  the  screening  procedure  described  for Dinoseb  can
provide   reliable  evidence  that  Dinoseb  is present  at  a  site.   The
performance data presented here that a methanol leach can provide adequate
extraction of dinoseb from some soils particularly when combined with HPLC
analysis.    The  lower  IQL  and larger  linear range of  HPLC  vs.  GC/FID
indicate that HPLC may be a more appropriate technique than underivatized
phenol  analysis  for providing rapid  laboratory measurements.   Project
managers should  consider the  use of screening  or  simplified  laboratory
procedures (HPLC)  for the  analysis of Dinoseb  when Project DQOs  do  not
require the use of Methods  8150/8151 or 8270..
                                   359

-------
QUANTITATION OF  POLYCHLORINATED BIPHENYLS  USING 19 SPECIFIC CONGENERS

Siu-Fai Tsang and Paul Marsden. SAIC,  San Diego,  CA  92121, Mike Kennedy,
Mesa College, San Diego 92111, and Barry Lesnik, Organic Methods Program
Manager, OSW/EPA 22403
ABSTRACT

Polychlorinated biphenyls (PCBs) are ubiquitous  and persistent pollutants
that have  been banned for  over  20 years.   Despite  our  long history in
measuring  PCBs  in  the  environment,  the  interpretation  of analytical
results  for  PCBs are often a problem during monitoring  activities.   A
major analytical  difficulty results from the  fact that most PCB methods
specify the use Aroclor standards  as the basis of quantitative analysis.
Aroclors are complex mixtures of chlorinated biphenyls  (e.g., analysis of
an Aroclor 1260 standard using capillary GC/ECD should produce more than
50  peaks).   While  most  of the  PCBs  found in  the environment  were
originally discharged as Aroclors,  the PCBs found  in  sediments today have
been altered by environmental weathering processes.  As a result,  they may
not produce the same pattern of peaks as the original  Aroclor  contaminant.
In  order to improve  the  reliability  of  PCB  analyses, a  mixture  of 19
specific  PCB  congeners  is  proposed  as  an   alternative  calibration
standard.   These individual congeners can be resolved on an SE-30 type
capillary column  (e.g., DB-5), are available  through commercial vendors,
and represent the major peaks of  the  individual Aroclors.   Spiked  soil and
reference materials have been analyzed using quantitation based on Aroclor
and congener  calibration.   Analysts have  demonstrated that mercury and
sulfuric acid  clean up  do  not effect the recovery of these congeners or
Aroclors.
 INTRODUCTION

 PCB's were produced  in the United States under the trade name Aroclor by
 Monsanto until  they  were  banned in  1972.   Aroclors are mixtures of many
 individual chlorinated biphenyls (congeners) which were sold as products
 on the basis  of their weight percentage of chlorine.   For example, each
 molecule  in Aroclor  1254 contains  12  carbons  and an  average chlorine
 content  (w/w) of the mixture was 54%.   The exception to this system was
 Aroclor 1016, which  is similar  to Aroclor 1242 except that 1016 contains
 only 0.4%  pentachloro congeners while 1242 contains 4.2%.   The lack of
 higher chlorinated congeners allows Aroclors  1016 to be  combined with
 Aroclor 1260 in a single calibration standard.  All other Aroclors must be
 analzed as individual standards.

 PCBs are generally analyzed using gas chromatography with electron capture
 detection (GC/ECD) (e.g.,  SW-846 Methods 8080/8081 or Method  608).  These
                                    360

-------
methods   specify   that  analyte   identification   and  quantitation   be
accomplished by comparing the peak pattern observed in a  sample  with  the
peak patterns from commercial Aroclor mixtures.  This approach of patten
matching  is appropriate  when  the  pattern in the sample is similar  to  an
Aroclor standard and when the  analyst is  experienced in PCB measurements.
However,  the practice of pattern matching can be difficult when  a sample
contains  a heavily weathered or an incompletely treated Aroclor.   Use  of
Aroclor pattern matching is also a questionable practice  when the sample
contains  more  than one  Aroclor  or  PCBs  created  as  an  industrial side
product (e.g., chlorination of a still keeper).

Capillary GC analysis increases the complexity of PCB quantitation based
on  pattern matching  while  it provides  opportunities  to  improve PCB
measurement based  on congener analysis.   The  complexity has increased
because 40-100  peaks may  be  observed for  an  individual Aroclor  using
capillary GC  (Figure  1)  in contrast to the  seven  to  ten peaks  observed
using  packed  column.    Laboratories  often  utilize  poorly  documented
quantitation procedures  for capillary  data  that are  highly dependent  on
the  training  of  the analyst.   Differences  in quantitation practices
probably  contribute to the  interlaboratory  variability observed  for PCB
analyses  (i.e., 36% of the  1991 WP PCB values were incorrect2).   At the
same  time,  capillary  GC  provides  a  tool  to resolve   individual PCB
congeners which can be exploited  to improve analyses3'*.   This  can be a
significant effort,  chlorination  of biphenyl  (C12H10) produces  209 PCB
congeners.  The IUPAC has adopted a numbering system for the PCB congeners
which is used in this paper.

While Method  8081B  provides some  guidance  on the  use of congeners for
quantitation  of  PCBs  in Method  8081,  no  sample  performance data are
provided.   In addition, the recommended mixture of congeners needs  to be
refined   by   adding  higher   chlorinated  species.      Calculation  of
concentrations using  a  mixture of 12 congeners  provides  results very
similar to Aroclor-based quantitation for Aroclors  1016,  1232, 1242, and
1248.  Unfortunately, calculation  of concentrations using the congeners
produces values that are  only 45-60% of the values obtained using Aroclors
1254 and 1260.5 Use of a longer list of congeners suggested by Draper of
California's  Hazardous  Materials   Laboratory  (HML)   provides   better
correlation (76-85%) for Aroclors  1254 and  1260.6  However,  HML's list
indues forty congeners.

This  study  was  designed  to  provide  the Office  of  Solid Waste  with
performance data  for  a  congener-specif ic  analysis  of  PCBs  in  solid
matrices.    One of  the  considerations in designing  this study  is the
requirement for a five-point calibration  of  all target  analytes.   It is
also hoped that these data will assist the  Methods Section  in establishing
whether congener-specific analysis of  PCBs has  application for  the RCRA
program.
                                   361

-------
 1 .
        : cs
 SOOO-j


 6OOO-
        ]

 4 O O O 4.
                          1 O
                                                                 3O

                          1 O
                                                                30
2.
1  O<
         co
        - C5
        -' SO
5OOO-
       O
                         1 O
                                            20
       Figure  1.  Chromatograms of Aroclors 1242, 1254 and  1260
                                 362

-------
EXPERIMENTAL PROCEDURES

SELECTION OF CONGENERS   During the first stage of this study, candidate
quantitation congeners were  selected  and chromatographic separations of
the  compounds  were established.   Each  congener solution  was analyzed
individually to establish the retention time.  The retention times  of the
congeners were  compared  with the  retention times of  the  major peaks of
Aroclors 1242,  1254 and 1260.  These results were compared with Draper's
findings6.   A group of 19 congeners were  selected which represented the
major peaks in Aroclors.   Table 1  lists the congeners and a chromatogram
of the standards is provided in Figure 2.

Comparison  of  the  response  factors  generated  in this  study with the
relative  response  factors   reported  by  Mullen*   shows  considerable
differences between the two  data sets.  The response factors generated in
this  study were  based  on   standards  in  hexane purchased  from  Ultra
Scientific.   Mullen's  relative response  factors were  obtained  using
standards  from  S.  Safe  of  Texas  A&M.   Because  of  the  difficulties  in
preparing pure congeners, it  is recommended  that additional analysis  of
standards be completed to establish  response  factors for  the selected
congeners.    It is  also  recommended  that  relative  response  factors  be
included in any specific-congener  method as a quality control measure.

CHROMATOGRAPHIC CONDITIONS    Some  analysts recommend 60  m  columns with
very  slow  temperature ramps  (e.g.,  l°C/min,  Mike Mullen)  in order  to
separate almost all the PCB  congeners.  This approach requires run times
from 90 to  120  minutes, results  in short column lifetimes and allows only
6-7  analyses in a 12 hour  shift.   As the  object of this  study was  to
develop a practical and routine analytical procedure,  use of a 30 m DB-5
column seemed more appropriate for this procedure.  The  carrier  gas was
nitrogen to demonstrate  that  these  separations could  be  achieved  in
commercial  laboratories.

                   GC Conditions
      Initial temperature
      Temperature ramp
      Final temperature
      Run time
150°C, hold 0.5 min
5°C/min
275°C, hold 9.5 min
 35 min
      Carrier
      Makeup

      Injector
      Detector
      Column
  6 mL/min, nitrogen
 54 mL/min, nitrogen

packed, 200°C
electron capture, 300°C
30m X 0.53mm id, DB-5
                                   363

-------
                         TABLE 1 - CONGENER MIX
Congener
2-Chloro
2,3-Dichloro
2,2' ,5-Trichloro
2,4' ,5-Trichloro
2,2' ,5,5'-Tetrachloro
2,2',3',5'-Tetrachloro
2 , 3 ' , 4 , 4 ' -Tetrachloro
2,2' ,4,5,5'-Pentachloro
2,2' ,3,4,5'-Pentachloro
2,3,3' ,4' ,6-Pentachloro
2,2' ,3,5,5' ,6-Hexachloro
2,2' ,4,4' ,5,5'-Hexachloro
2, 2', 3, 4, 4', 5' -Hexachloro
2,2' ,3 ,4, 5, 5 '-Hexachloro
2,2' ,3,4' ,5,5' ,6-Heptachloro
2,2' ,3,4,4' ,5'-, 6-Heptachloro
2,2',3,4,4',5,5' -Heptachloro
2,2' ,3,3' ,4,4' ,5 -Heptachloro
2, 2', 3,3', 4, 4', 5, 5', 6 -Nonachloro
Decachlorobiphenyl (surrogate)
IUPAC#
1
5
18
31
52
44
66
101
87
110
151
153
138
141
187
183
180
170
206
209
Rt
6.52
10.07
11.62
13.43
14.75
15.51
17.20
18.08
19.11
19.45
19.87
21.30
21.79
22.34
22.89
23.09
24.87
25.93
30.70
32.63
Column flow rate - 6 mL/min
                                   364

-------
2 . Oeo-
                                                             30
          FIGURE 2   Chromatogram of Specific Congeners
                               365

-------
Five-point  calibrations  were  performed  using  the  congener  mixture
according  to  the  criteria specified  in Method  8000 of  SW-846.   The
calibration mixtures and the RSDs are reported in Table 2.

                   TABLE 2 - INITIAL CALIBRATION DATA
IUPAC*
1
5
18
31
52
44
66
101
87
110
151
153
138
141
187
183
180
170
206
209
(surrogate)
Rt
6.52
10.07
11.62
13.43
14.75
15.51
17.20
18.08
19.11
19.45
19.87
21.30
21.79
22.34
22.89
23.09
24.87
25.93
30.70
32.63
Calibration solutions ng/mL
25, 100, 250, 500, 1000
25, 100, 500, 1000, 5000
25, 100, 500, 1000, 2500
25, 100, 500, 1000, 5000
25, 100, 500, 1000, 2500
25, 100, 500, 1000, 5000
25, 100, 500, 1000, 5000
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 2500
12.5, 50, 250, 500, 1250
25, 125, 250, 500, 1250
12.5, 50, 125, 250, 500
25, 50, 250, 500, 2500
25, 50, 250, 500, 2500
25, 125, 250, 500, 2500
25, 125, 250, 500, 2500
5, 25, 250, 500, 2500
5, 25, 250, 500, 2500
% RSD
14.1
4.6
10.5
3.1
9.1
4.2
6.1
6.4
9.4
7.0
4.8
9.2
13.4
7.8
5.9
5.3-
19.6
23.0*
5.6
7.3
*reduce calibration range
Column flow rate - 6 mL/min
Sulfuric acid and sulfur cleanups were evaluated to ensure that they did
not  result  in  the  loss  of Aroclors  or  congeners.    Although  these
techniques  are  in common  use in  environmental laboratories,  there  is
                                   366

-------
Little documentation  that  they  can not destroy any PCBs7.   A mixture  of
congeners,  Aroclor  1242,  Aroclor 1254,  and  Aroclor  1260 were  each
subjected  to both  cleanup procedures.   Peak  areas  of  the individual
congeners  were  compared before  and after cleanup  to  determine congener
recoveries.  The sum of the areas  of six or seven peaks from each  of the
Aroclor  standards  were compared before  and after  cleanup  to determine
Aroclor  recoveries.   The  results  for the  cleanup of Aroclor  1254 are
presented  in Table  3 as  illustration  of the fact  that  neither cleanup
destroys PCBs.
                    TABLE 3    Cleanup  of Aroclor 1254
Rt
14.26
15.01
16.65
18.92
20.82
21.82
total
Isomer #
52
44
66
110
153
138

Area of std
22824
9401
41318
35594
33339
34882
192790
Area after
cleanup
25170
10149
42677
36032
33964
34839
197521
% Recovery

103
Column flow   7 mL/min
RESULTS

Performance testing of congener quantitation was conducted using two EPA
freeze dried soil samples provided by  the  RT Corporation and two spiked
soils.   PCBs  were  quantitated using the specific-congeners  and Aroclor
standards.  The EPA Soil Group 1 sample was contaminated with a reported
concentration of 44.1 mg/Kg of Aroclor  1242.  The EPA Soil Group 2 sample
was contaminated with a  reported concentration  of  2.08  mg/Kg of Aroclor
1254.   San Diego soil  sample spiked with  Aroclors  1254 and  1260  were
spiked at  the  500  ppb level.  A aliquot of 5 grams of  each sample was
mixed  with  anhydrous  sodium sulfate  and  extracted  using a  Soxhlet
apparatus (Method 3540) with methylene chloride as a solvent.  A duplicate
determination of the spiked Aroclor 1254 was also prepared and analyzed.
The extracts were exchanged to hexane  prior to  analysis,  no  cleanup was
employed.  Three of  the soils were also extracted using a Dionex Model 703
supercritical fluid extractor with carbon dioxide at 100°C, 300 atm for 10
minutes static and 45 minutes dynamic.   The restrictor was maintained at
150°C and the receiving  vial  at 4°C.   The  vial  was filled with 15 ml of
hexane and decachlorobiphenyl as an  internal  standard.   The  recovery of
                                   367

-------
internal  standard was  used to  correct for  the  loss  of  solvent that
occurred during the extraction.   Each  extract was analyzed using a 30 m
DB-5 column using the conditions described  above.   Tables 4 and 5 present
the recoveries  from the  EPA  soils determined using  Soxhlet extraction
based on congener and Aroclor  quantitation  as  well as supercritical fluid
extraction (SFE) using Aroclor quantitation.   Table 6 provides the amount
of individual congeners determined in duplicate soils spiked with Aroclor
1254 as well as  recoveries for Soxhlet and SFE. Table 7 provides recovery
data for soil spiked with Aroclor 1260.

In all cases, the concentrations based  Aroclor standards were larger than
those obtained using congener analysis.  The recoveries based on congeners
were  lowest  for  the  EPA soils  in part because  interference  prevented
measurement of congener #66.  It is worth noting  that while the congener
values  are  also  lower  in the  spiked  samples, they  are closer  to  the
nominal values than were obtained using Aroclor quantitation

             TABLE 4 - Group 1 Soil (44.1 ppm Aroclor 1242)
Recovery
category
TOTAL (ppb)
TOTAL (%)
Recoveries
Quantities based
on congeners
using Soxhlet
22744 ppb
51.6%
Quantities
based on
Aroclors using
Soxhlet
49800 ppb
113%
SFE (by
Aroclor)

79.9%
column flow rate 6mL/min
SFE recovery = apparent recovery/(surrogate recovery*100)


             TABLE 5   Group 2 Soil (2.08 ppm Aroclor 1254)
IUPAC*
TOTAL (ppb)
TOTAL (%)
Recoveries
Quantities based
on congeners
using Soxhlet
1363
65.5%
Quantities
based on
Aroclors using
Soxhlet
2202
106%
SFE (by
Aroclor)

73.2%
column flow rate 6mL/min
SFE recovery = apparent recovery/(surrogate recovery*100)
                                   368

-------
         TABLE 6   Duplicate Spiked Soil (500 ppb Aroclor 1254)
IUPAC*
52
44
66
101
87
110
151
153
138
141
187
183
180
170
206
209
TOTAL (ppb)
TOTAL (%)
Rt
14.83
15.59
17.24
18.15

19.53
19.96
21.45
21.88
22.41


24.96
25.04

32.63


Recoveries
Concentration
of congeners
by Soxhlet
36.5, 40.0
10.5, 11.2
82.6, 88.2
66.3, 71.4

107.5, 109.8
3.7, 4.6
59.4, 59.4
3.6, 3.9
54.0, 59.4


3.7, 4.4
3.2, 3.9

102, 101
431
86.2%
Aroclor
cone
X
X
X
X



X

X






661
132%
SFE by
(Aroclor)
X
X
X
X



X

X





218

83%
column flow rate 6mL/min
SFE recovery = apparent recovery/(surrogate recovery*100)
                                   369

-------
              TABLE 7 - Spiked Soil (500 ppb Aroclor 1260)
IUPAC*
66
101
87
110
151
153
138
141
187
183
180
170
206
209
TOTAL (ppb)
TOTAL (%)
Rt
17.25
18.17

19.44
19.96
21.41
21.88
22.41
22.97
23.17
24.95
26.03
30.85
32.63


Recoveries
Concentration of
congeners
by Soxhlet
15.7
11.7
-
82.8
33.2
70.4
18.9
40.8
34.1
9.4
85.9
18.8
2.4
99%
424.1
84.8%
Aroclor cone
X




X
X
X
X
X




687
137%
column flow rate omL/min
CONCLUSIONS

      The approach of using a mixture  of  19 specific congeners appears to
be a viable approach to measuring PCBs in the environment.  Recoveries of
soils spiked with  Aroclors 1254 and  1260 determined  using the specific
congener mixture  were between  80 and 90%.   Aroclors  in environmental
samples were recovered at 51-66% of the certified values.

      While  this   approach  is  promising,  it  is not yet  ready to  be
distributed as  formal method.   At  least one of  the  selected congeners
(#153) coelutes with another Aroclor congener.   In addition, the correct
response   factors   for   the   congeners  relative   to   the   surrogate
decachlorobiphenyl needs to be established.  Finally  the  list of congeners
                                   370

-------
should be further evaluated in terms  of common matrix  interferences.   For
example,  isomer  # 66 was  buried beneath a  chlorinated pesticide peak,
possibly  DDE  or  DDT,   in   the  EPA  soil  samples.    Despite  these
shortcommings,   congener  specific  analysis  could  be  applied  to more
environmental and waste studies.
REFERENCES

1.  Brinkman, U.A. and A. De Kok. In: Halogenated Biphenvls. Terphenyls.
Naphthalenes. Dibenzodioxins and Related Products. R.D. Kimbrough  (ed.),
Elsevier. North Holland, Amsterdam (1980).

2.  Rohrlich,  T.,  "Waste Pile of Data  on  Pollution."Los  Angeles Times.
p.l, September 13, 1992.

3.  Alford-Stevens,  A.,  T.A. Bellar,  J.W.  Eichelberger and  W.L.  Budde.
"Method  680.    Determination  of  Pesticides  and  PCBs  in  Water  and
Soil/Sediment by Gas Chromatography/Mass Spectrometry," EMSL-Cincinnati,
U.S. EPA, November 1985.

4.  Mullin, M.D.;  C.M.  Pochini;  S.  McCrindle; M.  Romkes;  S.H.  Safe;  and
L.M. Safe, "High-Resolution PCB Analysis:   Synthesis and Chromatographic
Properties of all 209 PCB Congeners,"  Environ. Sci. Technol.. 18, 466-476,
June 1984.

5.  Marsden,  P.,  "Evaluation of Methods for Analysis of Polychlorinated
Biphenyls  (PCBs)"  EPA/600/X-90/222,  Environmental  Monitoring  Systems
Laboratory   Las Vegas,  9/1990.

6.  Draper,  W.M.  and  D.  Wijekoon.  "A  New and  Improved  Technique  for
Speciation and Quantitation of Aroclors  in Hazardous Wastes," Proceedings
of  the  Seventh Annual  Waste  Testing and  Quality Assurance  Symposium,
Office of Solid Waste and Emergency Response, U.S. EPA,  Washington D.C.,
July 1990.

7.  Copeland,  G.B. and C.S.Gohmann. "Improved Method  for Polychlorinated
Biphenyl Determination in Complex Matrices," Environ. Sci. Technol. 16(2):
121-124 (1982).
                                   371

-------
AQ                       DIRECT DETERMINATION OF TCLP PHENOLS AND
H0                                      HERBICIDES BY HPLC
                               R. L. Schenley. J. E. Caton, and W. H. Griest
                                       Analytical Chemistry Division
                                      Oak Ridge National Laboratory
                                     Oak Ridge, Tennessee 37831-6120
      A method for determining TCLP acidic semivolatile compounds (cresols, chlorophenols, and herbicides)
      is being developed to permit radioactive mixed wastes and their leachates to be analyzed with a minimum
      of operator exposure to sample radioactivity and very little laboratory waste generation versus current
      GC-ECD methods.  The HPLC method features direct injection of a 25 /tL aliquot of aqueous waste or
      leachate (with filtration and acidification, if needed) onto a 150 mm X 4.1 mm ID (5 urn particles) PRP-1
      polystyrene column. The mobile phase is a gradient of 0.1 M acetic acid in 10/90 (v/v) acetonitrile/water
      and acetonitrile.  Quantitation is by the method of external standards at wavelengths of 235 and 280 run
      using a diode array  UV detector.  Compound confirmation is by UV spectrum or retention tune on an
      ODS column. The detection limits meet TCLP Regulatory Limits, and run tunes are 16 min per sample,
      including mobile phase reequilibration. Matrix spike recoveries at or below the Regulatory Limits in
      TCLP Fluid No. 2 ranged from 88.6 to 128.2 %, but were much lower in alkaline, high-nitrate wastes
      simulating nuclear wastes.  Surrogate standard recoveries determined over the course of 12 months for
      2,4,5-T spiked into 61  TCLP leachates  at 5.0 to  7.5 mg/L  averaged 96.7 ± 14.8%.   A  better
      confirmation column phase and mass spectrometric confirmation  will be added in the future.
      Research sponsored by the Office of Technology Development, Laboratory Management Division under
      U. S. Department of Energy contract DE-AC05-84OR21400 with Martin Marietta Energy Systems, Inc.
                                    Th«  uAfritttd  mmmavt hM  bun
                                    authored by a omu»cior  of th» U.S.
                                    Govwmwnt  mtar  contract  No. OE-
                                    AC06-840R21400. AecorrJnoty. th« U.S.
                                    GoMmmmt   rettira  •  noraxcfcam.
                                    royiltv-frM totraa to pubfah or reproduce
                                    th* puMMwd form of ttn ctmutoiuon, or
                                    •low ottwr* to do to. tor U.S. Govwnrram
                                    purpoam."
                                                 372

-------
49    Approaching the Sensitivity of an Electron Capture Detector (BCD) for
                     the Analysis of Pesticides by using GC/MS
      Linda C. Doherty and Norman Low, Scientific Instruments  Division  Hewlett-Packard
      Co., 1601 California Ave., Palo Alto, CA 94304
      ABSTRACT
      Pesticides are analyzed routinely by GC/ECD with dual column confirmation.  The ECD
      has excellent sensitivity and can be used to determine some pesticides at levels in the
      low femtogram range.  Unfortunately, it is easily contaminated and lacks the linearity of
      other detectors.  Also the chromatography must be optimized, to eliminate any co-
      eluting analytes.

      We decided to approach the  analysis of pesticides using  a selected ion  monitoring
      (SIM) based GC/MS method. The  MS tolerates dirty samples much more readily than
      the ECD and is easier to clean.  SIM has the benefit of giving up to 4  orders of
      magnitude  linearity, much better than an  ECD.   It  also  accommodates  co-eluting
      analytes by allowing for their separation based on mass.  This speeds up method
      development time.

      The  USEPA Method  60S1  pesticides were used  for these  experiments.   (PCBs,
      toxaphene and chlordane were excluded.)  Initially a standard at the 100 pg level was
      analyzed by electron impact (El) GC/MS in the scan mode (60-430 a.m.u.). One or two
      significant ions were chosen for each pesticide from the mass spectrum.  Because
      pesticides have  unique El  fragmentation patterns, the specific  ions chosen were
      indicative of the  pesticide.  The SIM method was developed  using those ions.  The
      pesticides were separated into elution groups to get the best results.   Four  groups of
      ions were used.  Standards were injected over a 30 fold concentration range to best
      mimic  the range required by the Contract  Laboratory Program  3/90 Statement of
      Work2  All pesticides were linear over this  range.  The hexachlorocyclohexanes
      standards (BHCs) were analyzed from 0.1  pg to 1000 pg.  Standards were then spiked
      with reference gas oil and analyzed in both SIM and scan mode. Solutions were spiked
      at both the 0.1 and  1.0% levels.

      INTRODUCTION
      Historically the ECD has been used for the identification and  quantitation of chlorinated,
      electrophilic pesticides.  As mass spectrometers improve in sensitivity other approaches
      to the analysis of pesticides need to be tried.  Mass spectrometry affords the analyst
      the ability to obtain  a third dimension of information to confirm the presence or absence
      of a  certain pesticide. The following work explores the use  of selected ion  monitoring
      (SIM) for the analysis of chlorinated pesticides.
                                            373

-------
EXPERIMENTAL
The  GC/MS system  was composed of an  HP  5890 Series II GC with Electronic
Pressure Control (EPC). EPC allows the user to control the pressure or flow of GC
carrier gas accurately.  The autosampler  was an  HP  7673B ALS.   The  mass
spectrometer was the HP 5972A. The GC was installed with an HP-5MS column (30 m,
0.25 mm ID, 0.25 \w\ film).  A single tapered, deactivated glass  liner with glass wool
(HP P/N 5062-3587) was installed in the injection port.
250° C
85° C, 1.5 minute hold
30°C/minto190°C
3.6° C/min to 240° C
1.5 minutes
15  psi for  1.5  minutes, then
constant flow
280° C
2 stops (1 \L\, 5 nl syringe)
GC:
Injection port
Initial oven temperature
Purge on
EPC program

Transfer Line
Injection size

MS:
SIM (4 groups)
Group 1 start time 6.7 minutes
Group 1: m/z= 180.9, 108.95

Group 2 start time 8.7 minutes
Group 2: m/z= 100, 66, 240.8, 196.9

Group 3 start time 12.7 minutes
Group 3: m/z= 246, 79.05, 81, 67, 194.9

Group 4 start time 15 minutes
Group 4: m/z= 234.9, 271.8

max sensitivity autotune
EM = tune + 700 V
Group 1 included the four hexachlorocyclohexane pesticides (BHCs). Group 2 included
Heptachlor, Aldrin, Heptachlor Epoxide and Endosulfan I.  Group 3 included 4,4'-DDE,
Dieldrin,  Endrin, Endosulfan II, 4,4'  ODD and Endrin  Aldehyde.  Group 4 included
Endosulfan Sulfate and 4,4'-DDT.

Some standards were prepared  from a 100 ppb pesticide solution (HP Pesticide
Evaluation Sample P/N 5062-3589). Solutions for the standard curves were made with
a stock solution diluted from a Method 8080 standard  (HP P/N 8500-6011)  .  All
                                         SCAN

                                         60-430 amu
                                         A/D =2A2
                                         Threshold = 0
                                         max sensitivity autotune
                                         EM = tune + 700 V
                                      374

-------
solutions were prepared in iso-octane.  ASTM Reference Gas Oil #1  (HP P/N 5060-
9086) was spiked into standards for the "dirty matrix" experiments.

RESULTS
Scan Mode:
All pesticides were analyzed, as standards, down to 10 pg injected. The data collected
in  scan mode  gave quality library searches at 25-50 pg  depending on the pesticide.
The  standard curves were linear from 25-1000 pg  for all of the pesticides.  Certain
pesticides  gave three orders  of magnitude linearity.   The  sensitivity level  was
dependent  on  the pesticide.  Some of the compounds fragmented easily.  Figure 1
shows the  mass spectrum of Endosulfan I  versus DDE.  DDE can be seen at much
lower concentrations than Endosulfan I.  This is because the ion current for Endosulfan
I is spread over many masses and not concentrated into one or two masses.
                                                         Endosulfan I
                                               .",V,Tt..lll,,.ll'r, ..;..;,,	
                                               300 320 340 380 380  400 420
           150000


           100000


            50000
2-


178
105 |
87
82 J \ 1
i. i.fAtL
7 i«
11 1 I 233
IL. L .ill JL ,. i,l . I

4,4'DDE
31 a
I


281
I 	 fc 	



	
                      100 120  140 160 180 200 zfo' 2^0 ' 2efo idti ' ido 320  340 3^0 ' '3j(o' ' 4yd VjV
Figure 1.
A comparison of the spectra for two pesticides to show the amount of
fragmentation. Notice the difference in maximum abundance counts.
The largest ion for Endosulfan is about SOX smaller than m/z= 246 for
DDE.
The  most prominent ion  (the base peak)  in the spectrum also affects the ultimate
detection limit.  The base peak is generally used as the  quantitation ion.  Figure 2
shows the spectrum of DDE versus Dieldrin. The base peak in DDE is m/z= 246 were
the base peak for Dieldrin is m/z= 79.  The higher the mass of the ion, the  less likely
that there will be co-eluting background at the same mass. In Figure 3  a chromatogram
of a standard at 100 pg is displayed.
                                       375

-------
UUAfalttt
150000

100000 .

50000

0

200000 .

100000

0.
ntz->






62 j
J... I.
'bo" 'si
7


66
Vr "
60 8I




10
87
I
.jl Ajl
J 100



1
94

Vob




5
12

J,
Vi



D8

1-H



178
I
^ 21

L, A j. - l,ll r 1" ,
' ' V4b' ' 'ieb' ' 'ido ' '20^0 '



143 173 193 207

>' i4b ido' is'o 2ob
2«







Ho ' 24ld



237
•n Ti'lVV
220 240


™ 4, 4'-DDE





34*0' 3efd M'O 4clo' 42"o' '

Dieldrin

^277
1 L h 291 316 345 334
-(|l)j A - ,. ... ...
260 2SO 300 320 340 360 380 400 420
Figure 2.     Spectra for two pesticides where the base peak differ in mass. The
             higher the mass of the base peak, the less background interference.
             Signal-to-noise improves at higher masses.  Notice the abundance
             counts for the base peak are similar.
                         100 pg 808 peehodes 15 pei ml; Run al
                         02 57 PMPSTcn Kfcn FebOI. 1093
                         V»]IW
Figure 3.     The total ion chromatogram of a 100 pg pesticide standard analyzed in
             the scan mode.

One benefit of mass  spectrometry is the ability to separate co-eluting peaks without
extensive chromatographic method development.  Figure 4 shows an example of this.
Dieldrin and DDE are co-eluting but still easily quantifiable.
                                       376

-------
                  3
                  r, -i
                                 DDE
                                  Dieldrin
Figure 4. Separation of co-eluting DDE and Dieldrin by mass.

Standards were then spiked with reference gas oil #1 (RGO #1).  A standard 100 pg
solution was spiked at the 0.1 % (1 |il to 1000 |il) and 1% levels (10 pJ to 1000 |il). DDE
and ODD gave high quality library searchable  results at the 0.1% spiked level.

SIM
SIM experiments were  run  to get the lowest detection  limits.  SIM  also gives better
quantitation  reproducibility  over scan.  Since SIM allows for a higher  rate of data
acquisition, more points are collected across a GC peak and the GC peak  is better
defined and  integrated.    For  dirty samples,  SIM will  reduce  background  noise,
increasing the sensitivity.

Standards were prepared  and  run  from 5  pg to 1000  pg for all  pesticides.   The
hexachlorocyclohexane pesticides were analyzed down to 100 fg.  Figure 5 shows the
calibration curve for Lindane. The RGO spiked standards were  also analyzed in SIM.
Figure 6 gives a SIM/scan comparison for DDE in the presence of the  RGO matrix.
                              Response = 1 76ft+OCB ' Ami - 8 OSe+003
                               COT Coef = 0 998  Cufw FT Lkwar
Figure 5.
The calibration curve for Lindane from 100 fg to 1000 pg injected
splitless.
                                      377

-------
               Abundance
               1000-g
                900-=

                800^
                700 \
                600^
                500 \
                4001
                3004
                20QJ
                   P.P-DDE
                             SIM
                800-;

                700-E
                600.;

                500^

                400^

                300-j

                200-1
JJ
                             Scan
Figure 6.     A comparison between SIM and scan for the 0.1% spike standards. The
             concentration for DDE is 100 pg.

SUMMARY
Results  look favorable for substituting a  mass spectrometer for an  ECD.  Detection
limits of MS approach those of an ECD and give a much less ambiguous confirmation
for the pesticides.  By using one or two unique ions per pesticide for confirmation, false
positives and negatives are much easier to eliminate.
1USEPA Method 608, Document Number EPA-600/4-82-057, July 1982
2USEPA Contract Laboratory Program  Statement of Work for Organics Analysis, Document
 Number OLM01.8, August 1991
                                      378

-------
50            DETERMINATION OF POLYNUCLEAR AROMATIC
              HYDROCARBONS IN SOIL AT 1 ^g/kg USING GC/MS

      Bruce  N. Colbv  and C. Steve Parsons,  Pacific Analytical, 6349 Paseo  Del Lago,
      Carlsbad, California 92009, and James S. Smith, Trillium, 7A Graces Drive, Coatesville,
      PA 19320

      Abstract: A GC/MS method has been developed which is capable of determining PAHs
      below  1 pig/kg in soil samples.  It uses a 100  g initial sample, soxhlet extraction using
      toluene, extensive sample cleanup including alumina column fractionation, and a 300 jiL
      final extract volume. Measurements are made using selected ion monitoring techniques
      on a quadrupole GC/MS.

      INTRODUCTION

      Because several polynuclear aromatic hydrocarbons (PAHs)  are potent carcinogens, it is
      appropriate to determine them at relatively low concentrations in environmental samples.
      At present there are two accepted ways to measure PAHs  in soils, Method 8270 and
      Method 8310. Method 8270 is a reasonably robust GC/MS method with indicated PAH
      detection  limits of  660 |ig/kg.  Method 8310 is an  HPLC-fluorescence method with
      indicated  detection  limits ranging  from 9  to 1550 Hg/kg  depending on the analyte.
      Because of its generally lower detection limits, Method 8310 is frequently identified as
      the method of choice for field  studies. Unfortunately, Method 8310 is often unable to
      provide these lower detection limit values due to interference from non-target materials
      present in field samples. This is especially true in situations where alkyl-PAHs are
      present.
      The advent  of more sensitive GC/MS instrumentation has made it possible to detect
      PAHs  at considerably lower levels than those identified  in Method 8270.   When this
      more sensitive instrumentation is combined with extensive sample cleanup, it is possible
      to generate PAH detection limits which are less than 1 Hg/kg.

      METHOD SUMMARY

      A 100  g aliquot of soil sample is fortified with the labeled analogs (Cambridge Isotopes)
      of the target PAH analytes.  Moisture is removed from the sample by boiling the sample
      in toluene using a Barrett moisture trap attached to  a boiling flask.  The  dehydrated
      sample and toluene are then transferred  to a soxhlet apparatus to generate  the sample
      extract. The extract is subjected to alumina column fractionation (SW-846 Method 3611,
      Appendix B). Sulfur is removed using the tetrabutylammonium sulfite procedure if it is
      present in substantial quantities.

      The final extract is  spiked  with an  internal  standard  and analyzed by GC/MS (VG
      MD800) in the SIR  data acquisition mode.  Two mass descriptor groups are used which
                                             379

-------
incorporated the quantitation mass for each analyte. The dwell time on each mass is 60
msec and the cycle time is 1 sec. The instrument is calibrated with solutions containing
0.01, 0.1,  1, 10 and 100 ng/^iL of each PAH plus 2 ng/^iL of each labeled (deuterated)
analog of each PAH.  No labeled analog was available for indeno(l,2,3-cd)pyrene.

EXPERIMENTAL
Several experiments were undertaken during the  course of the study summarized here.
They  included a method detection limit (MDL)  study,  an analysis of the NIST SRM
1941,  plus  determinations  on  a variety  of field  samples many  of which were
contaminated with coal tar derived materials.  The MDL study  was undertaken using
Ottawa Sand as the substrate. Seven aliquots were spiked  to 2.0  |ig/kg with each PAH
plus the labeled analogs and then prepared and analyzed as described above.
In another experiment a 1 gram aliquot of the NIST SRM  1941 (river sediment) was
taken and processed as if it were a 10 gram sample.  Calculated concentrations were
multiplied by a factor of ten to account for the different sample size.
A wide variety of field samples were prepared and analyzed using the above procedure.
These samples were associated with waste sites involving petroleum and coal tar related
materials contamination.

RESULTS AND DISCUSSION
The GC/MS instrument was readily calibrated for the lower molecular weight PAH
compounds from 0.01 to 10 ng/jiL. For the higher molecular weight PAH  compounds
calibration was from 0.1 to 100 ng/(iL.  The reason for this was peak shape. The lower
molecular weight  PAHs resulted  in very  narrow peaks, so detector saturation was
encountered with the 100 ng/pL injections.  The higher molecular weight PAHs did not
yield  such narrow peaks and consequently did not saturate  the detector, but for the same
reason they did not produce reliable peak areas  for the 0.01 ng/pL injections.  SICP
traces for the low point standards for  chrysene and benzo(a)anthracene  are shown in
Figure 1 and a typical calibration curve for benzo(a)pyrene is shown in Figure 2.  After
these  experiments were performed, an improved transfer line was fitted to the GC/MS. It
greatly improved peak shape for the higher molecular weight PAHs, but no calibration
curves have been generated since it's installation.
The MDL study proceeded without incident and produced  the results shown in Table 1.
All of the  MDL values were less than 1 |ig/kg. Percent recovery and analytical precision
values were also calculated from the MDL study data. These are also shown in Table 1.
Analysis of NIST SRM 1941 resulted in generally excellent agreement with the NIST
provided values (Table 2). There may be a slightly positive bias in the measurements but
it is not consistent.  Also, the acenaphthylene value was quite different from the NIST
value. The rest, however, were reasonably close.
                                        380

-------
Initial experiments were performed without benefit of sulfur removal but several field
samples as well as SRM 1941 contained substantial quantities of sulfur.  With some of
the field samples there was so much sulfur that little else was evident on initial inspection
of the  total ion current trace.  While the sulfur did not seem  to create measurement
problems through ionization supression, the removal step was added to assure that it
would not become a problem. Figure 3 shows a full scan chromatogram for a high sulfur
sample. The broad peak centered at scan 1124 is for S8.  The baseline "hump" starting at
about scan 750 and continuing through 2500 is caused by the alkyl-PAHs which interfer
severely with Method 8310 measurements.
Finally, it is important to note that  problems were  encountered  occasionally  with
deuterium exchange on some of the labeled PAHs. This should not be surprising because
exchange is almost certainly the mechanism used to generate the labeled analogs.  Due to
the potential for exchange, a secondary set of quantitative references was identified for
those compounds having analogs which exhibit exchange.  Because this  is only an
occasional problem and it is one that can be monitored and solved through the use of the
secondary references,  it is not considered  a  serious problem.   Further,  Cambridge
Isotopes was  contacted with regard to this problem  and they have now  initiated a
program to develop 13C-labeled PAHs which will be  used in future experiments.

SUMMARY
By slightly increasing  sample size, using extensive  sample cleanup, slightly decreasing
final volume and making use of today's more sensitive GC/MS equipment, it is possible
to determine PAHs in soils at less than 1 ng/kg.  Care must be taken to monitor for label
exchange if deuterated  analogs are used for  quantitation.
             Figure 1 - SICP TRACES FOR 0.01 ng/jiL STANDARD

                               Pacific Analytical
    Sanple: 8.B1 ngAtL std
    feK15ft14
    y.fS
                                                          Instrument: VG86
                    Chrysene-dl2
                                                                      2638
Benz(a)anthracene-dl2
    y.FS
     96
    Scn669
                 67B
                       675   688   685    698   693   7flB   785   718   715  72B
                                       381

-------
        Figure 2 - BENZO(A)PYRENE CALIBRATION CURVE
      0.1
                                                         0)
                                                         CO

                                                         o
                                                         a
                                                         (0
                                                         0)
                                                         DC
 1               10


Concentration (ng/uL)
100
  Figure 3 - SOIL SAMPLE EXTRACT PRIOR TO SULFUR REMOVAL


                       Pacific Analytical
Sample: Soil Extract
                        Instrument: UGB4
                              382

-------
Table 1 - PRECISION. RECOVERY AND MDL RESULTS
Analyte
Acenaphthene
Acenaphthylene
Anthracene
Benz(a)anthracene
Benz(a)pyrene
Benz(b)fluoranthene
Benz(k)fluoranthene
Benz(ghi)perylene
Chrysene
Dibenz(a,h)anthracene
Fluoranthene
Fluorene
Indeno( 1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
Recovery
(%)
99
93
101
92
124
102
102
95
91
112
101
100
84
111
107
112
Precision
(%sd)
4.4
12.7
3.5
7.2
10.1
6.0
6.1
13.9
4.8
10.3
4.7
13.0
18.0
5.1
4.6
11.5
MDL
Gigftg)
0.27
0.74
0.22
0.42
0.79
0.39
0.39
0.83
0.27
0.73
0.30
0.82
0.95
0.36
0.31
0.81
                   383

-------
Table 2 - COMPARISON WITH NIST SRM 1941 RESULTS
Compound
Acenaphthene
Acenaphthylene
Anthracene
Benz(a)anthracene
Benz(a)pyrene
Benz(b)fluoranthene
Benz(k)fluoranthene
Benz(ghi)perylene
Chrysene
Fluoranthene
Fluorene
Indeno( 1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
PA
^g/kg
50
220
310
560
610
1100
1100
623
860
2000
110
570
1360
640
1600
NIST
Hg/kg
52
115
228
599
754
864
857
566
702
1401
104
559
1322
603
1238
%
Diff.
-4
63
30
-7
-21
24
25
10
20
35
6
2
3
6
26
                     384

-------
51      APPLICATION OF LEE RETENTION INDICES TO THE CONFIRMATION
               OF TENTATIVELY IDENTIFIED COMPOUNDS FROM GC/MS
                        ANALYSIS OF ENVIRONMENTAL SAMPLES

           Paul H Chen, W. Scott Keeran, William A. Van Ausdale, David R. Schindler, and Dwight F. Roberts

                      Analytical Services Division, Environmental Science & Engineering, Inc.
                                  P.O. Box 1703, Gainesville, FL  32602

       ABSTRACT

       In the EPA methods for the analysis of semivolatile organics by GC/MS, samples are generally
       analyzed for the target compound list (TCL) components by an automated data system and for
       non-TCL components by a library search.  The non-TCL components identified and reported as
       tentatively identified compounds (TIC) are rarely confirmed.  The retention indices (RI) of TICs
       may  provide a means of confirmation.   The purpose of this study was  to investigate the
       applicability  of  Lee  retention indices to the confirmation of TICs for several  classes  of
       compounds. The classes of compounds studied are PAHs, heterocyclic PAHs, n-alkanes, straight
       chain primary alcohols, and  fatty acid methyl esters. The RIs of these classes of compounds have
       been reported by Rostad et al. and were used for comparison with our data.  In addition, the RIs
       for Appendix IX compounds, free fatty acids,  and explosives and their related compounds were
       also determined by us in this study. The results show that high reproducibilities  between our RIs
       and the published data and between our data determined by different GC conditions (normal and
       slow ramps) were observed for PAHs and heterocyclic PAHs.  However, poor reproducibility
       was  observed  for  other classes  of compounds  studied.   It  should be noted  that high
       reproducibilities of RIs determined  under different GC conditions can only be achieved  if
       compounds of interest and RI standards are eluted during the temperature ramping period. The
       high reproducibility for PAHs can be explained by the fact that the four Lee RI  standards are all
       PAHs, the change in their  chromatographic retention behavior under different GC  conditions
       should be similar with other PAHs of interest.  Though the reproducibility of RIs for non-PAH
       compounds is not good for inter- and intra-laboratory comparisons, they provide  the elution order
       and position for isomers and homologs which are very useful  for compound  identification  or
       confirmation  purpose.

       INTRODUCTION

       In the EPA methods for the analysis of semivolatile organics by GC/MS,  samples are generally
       analyzed for the target compound list (TCL) components by an automated data system and for
       non-TCL components by a  forward library search  of a published mass spectral database.  The
       results of TCL components analysis are generally accurate.  However, non-TCL components
       identified and reported  as  tentatively  identified compounds (TIC) are rarely verified.  The
       retention  indices (RI)  of TICs may provide  a means of confirmation for certain types  of
       compounds.
                                                385

-------
Kovats retention indices  (1), based on  a  homologous series  of n-alkanes under isothermal
conditions, have been widely available for many organic compounds. Lee retention indices (RI),
based on a series of four polycyclic aromatic hydrocarbons (PAHs) as retention index standards,
have been reported by Lee et al. (2) and Vassilaros et al.  (3) for a large number of polycyclic
aromatic compounds. These Lee RIs are determined using capillary columns GC operated under
temperature programming conditions. Rostad  and Pereira (4) reported Lee RIs determined by
GC/MS for a large number of PAHs and other organic compounds of environmental interest.
The purposes of this presentation are:  [1]  to compare the Lee  RIs determined in our lab with
those published by Rostad et al. and other workers for several classes of compounds, [2] to
compare the Lee RIs determined in our lab  run at different GC conditions for several classes of
compounds, and  [3] to explain the differences between our RI values and published values and
their applicability for compound confirmation.  The classes of compounds  that are reported by
Rostad et al.  and our laboratories are PAHs (including N, S, and O-heterocycles), n-alkanes,
primary alcohols, and fatty acids methyl esters.  The RI values of additional compounds that will
be reported by us in this study are Appendix IX compounds, free fatty acids, and explosives and
their related compounds.
EXPERIMENTAL SECTION

Sample Preparation. Water samples were extracted at pH > 11 and then at pH < 2 with methylene
chloride according to EPA Method 625 (5).  Soil samples were extracted with methylene chloride
in Soxhlet extractors according to EPA SW-846 Method 3540/8270 (6).  Before extraction, each
sample was spiked with 1.0 mL of surrogate spiking solution which contains 100 /ng/mL each of
acid surrogates and 50 ng/mL each of base/neutral  surrogates.  Methylene chloride extract was
concentrated to 1 mL with Kuderna-Danish concentrator and analyzed by GC/MS.

GC/MS Analysis. Samples and standards were analyzed on an HP 59^0 MSD and an  HP 5988
GC/MS system.  For samples run on the HP 5970, the column used was a 25m  x 0.2mm i.d.
HP-5 (0.33  /im coating) fused  silica capillary column.   For the normal ramp, the column
temperature was held isothermal  at 40°C for one minute and then programmed at 10°C/min to
310°C  and held isothermal at this final temperature for 8-10 minutes.  Under this temperature
programming, chrysene elutes during the temperature ramping period, but benzo(g,h,i)perylene
elutes during the isothermal period. In order to have benzo(g,h,i)perylene (the last RI standard)
elute during the temperature ramping period, a slow ramp was also used. In the slow ramp,  the
oven temperature was programmed at 4°C/min instead of 10°C/min for the normal ramp.  For
samples run on the HP 5988 GC/MS, the column used was a 30m x 0.25mm i.d. DB-5MS (0.50
urn coating) fused silica capillary  column (J&W Scientific, Folsom, CA). For the  normal ramp,
the column temperature was held isothermal at 40°C for 4 minutes and then programmed at 10°C
per minute to 295°C, and held isothermal at this final temperature for 10 minutes.  For the slow
ramp,  the oven temperature was programmed at 4°C/min  to 315°C  and held at this final
temperature for 3 minutes.  The  mass spectrometer was scanned from 35 to 500 amu per half
second. The  extract or the standard was spiked with a mixture of six internal standards (which
contains naphthalene-d8, phenanthrene-dlO, and chrysene-d!2 among other compounds) before
GC/MS analysis according to CLP protocol (7). A forward library search was performed  for
                                          386

-------
non-TCL compounds on a Wiley/NBS data base which contains 139,000 different spectra (8).
Compounds were tentatively  identified by library search or by elucidation of the compound
structure from its mass spectrum if no match was found in the library.  The retention time of the
tentatively identified compound or the standard was used for the calculation of Lee RI according
to the equation described in the following paragraph.

Calculation of Lee Retention Indices. The Lee retention  indices are calculated according to the
following equation:

           RI = 100 Ox  - Tz) + 100 * Z
                  Tz+i  Tz

Where:     Tx is the retention time of compound of interest
           Tz is the retention time of the preceding RI standard
           Tz + i is the retention time of the following RI standard
           Z  is  the number of rings in the preceding RI  standard
The retention index standards used are naphthalene (RI = 200.00),  phenanthrene (RI = 300.00),
chrysene (RI = 400.00), and benzo(g,h,i)perylene (RI = 500.00).  When these compounds are not
found in the sample, their retention times are calculated by adding the differences in the retention
times between the first three RI standards and their corresponding deuterated internal standards
in the daily calibration standard to the retention times of the corresponding internal standards  in
the sample.  Determination of RIs for certain classes or categories of compounds (Appendix IX,
fatty acid methyl esters, primary alcohols, and some PAHs) were made from standard mixtures,
while others were made from actual sample extracts and, if possible, confirmed with standards.

RESULTS AND DISCUSSION

Retention  indices (RI) of PAHs determined in this study for both  normal and slow ramps are
listed in Table 1. RIs reported by Rostad et al. (4) are also shown in Table 1 for comparison.
When RIs are not available from Rostad for certain PAHs, the values from Vassilaros et al. (3)
are listed for comparison.  When RI values are larger than  400 (which means the compounds
elute after chrysene), the values from Vassilaros, or in some  cases,  from Lee et al. (2), are also
listed in Table 1  for comparison.

The reason that  the Rostad's RIs instead of Vassilaros'  and  Lee's RIs are primarily used for
comparison is that the former's instrumental  conditions are very similar to ours. Rostad's RIs
were determined by GC/MS using DB-5 column with He as the carrier  gas.  Vassilaros' and
Lee's RIs  were determined by GC using coated silica SE-52 column with H2 as the carrier gas
and glass capillary SE-52 column with He as the carrier gas, respectively.

Table 1 shows that the RI  values obtained by different GC conditions  (normal and slow ramps)
in our lab agree  very  well.  Our RI values also agree very well with  those reported  by Rostad
for PAHs  which  elute before chrysene.  For PAHs which elute  after chrysene, large deviations
                                           387

-------
are found between our values and Rostad's values. This discrepancy is likely due to the fact that
under the GC conditions we used for slow ramp, the last RI standard [benzo(g,h,i)perylene] is
eluted during the temperature ramping period, while this probably was not so for Rostad's work.
It should be noted that high reproducibility of the RI determined under different GC conditions
can only be achieved for those compounds having their upper bracketing RI standard elute during
the temperature ramping period.  This  was stressed by Lee but not by Rostad.  For the GC
conditions used by both Lee and Vassilaros, the last RI standard (picene) was eluted during the
temperature programming period.  Retention time of benzo(g,h,i)perylene was  not given in
Rostad's paper.  However, based on the GC conditions and the lower values of RI obtained for
compounds bracketed by chrysene and benzo(g,h,i)perylene, it appears that benzo(g,h,i)perylene
was eluted in the isothermal  period. For the PAHs which elute after chrysene, our RI values
agree  very  well  with  those reported  by  Vassilaros or  Lee   with  an exception  of
dibenz(a,h)anthracene.

Retention indices of heterocyclic PAHs determined by this study and  those reported by Rostad,
Vassilaros, and Lee are listed in Table 2. As shown in Table 2,  our values  agree very well with
those reported by the previous workers.  High reproducibility within our lab run at  different GC
conditions  and good agreement between  our data and  the  published  data for PAHs  and
heterocyclic PAHs  can be explained by the fact that the RI standards used  are all PAHs, the
changes in their chromatographic retention behavior under different  GC conditions should be
similar with other PAHs of interest.  The results of this study indicate that Lee retention indices
can provide a high degree of confirmation for PAHs and heterocyclic PAHs such as N-, S-, and
0-heterocyclic PAHs.

Table 3 lists the retention indices of n-alkanes determined in this study for both normal and slow
ramps.  RIs reported by Rostad are also listed in Table  3  for comparison.   As shown in Table
3, poor reproducibility was observed not only for RIs determined by different labs (ours and
Rostad's), but also for RIs measured at different GC conditions in our lab. This can  be explained
by the fact that the changes in chromatographic behaviors are different for n-alkanes  and the RI
standards  (which are PAH)  under  different GC conditions.   It should be noted that highly
reproducible RIs  can be obtained if the samples are run on the same column and at the same
conditions.  The RIs given in Table 3 can be used for identification of an individual n-alkane in
a series of homologous n-alkanes if the GC conditions are not drastically  different from those
used in this work.

In addition to  PAHs and n-alkanes, other classes of compounds were also investigated.  They
include straight chain primary alcohols and fatty acid methyl esters. The RIs values determined
by us and Rostad for these classes of compounds are listed in Tables 4 and 5. Tables 4 and 5
show that RIs of primary alcohols and fatty acid methyl esters run at different GC conditions are
not very reproducible.  The explanation  described above for n-alkanes can be applied here also.
Again, the RIs given in these tables can  be used for identification of an individual compound in
a series of homologs if the GC conditions used are not drastically different from those used in
this study.
                                           388

-------
Lee RIs of Appendix DC compounds, n-fatty acids, and explosives and their related compounds
were also determined.  They are shown in Table 6 to 8. The RIs for Appendix IX compounds
can be used to help the confirmation of these compounds in the samples if their standards are not
analyzed.  Poor reproducibility was observed for most of the Appendix IX compounds, while
good reproducibility was observed for those which had the chromatographic retention behaviors
similar to those of PAHs.

ACKNOWLEDGMENTS

We thank P. Dumas and his department for extracting the samples and C.  Diaz for her assistance
in GC/MS analysis.
REFERENCES

1.   Kovats, E. Helv. Chim. Acta.  1958. 44, 1915.

2.   Lee, M.L., Vassilaros, D.L., White, C.M., Novotny, M, Anal. Chem.. 1979. 51, 768.

3.   Vassilaros, D.M., Kong, R.C., Later, D.W., Lee, M.L., J.  of Chromatog.. 1982. 252, 1.

4.   Rostad, C.E., Pereira, W.E., J. High Resolution Chrom. and Chrom.  Commun.. 1986. 9,
    328.

5.   EPA 40 CFR Part 136 Fed. Regist.. 1984.  49 (No. 209).

6.   USEPA Test Methods for Evaluating Solid Waste:  Physical/Chemical Methods, SW-846,
    3rd Edition, 1986.

7.   USEPA Contract Laboratory Program Statement of Work for Organic Analysis, Multi-media,
    Multi-Concentration,  1988.

8.   McLafferty, F.W., Stauffer, D.B. The Wilev/NBS Registry of Mass Spectral Data. Wiley:
    New York, 1989.
                                         389

-------
TABLE 1. LEE RETENTION INDICIES OF POLYCYCLIC AROMTATIC HYDROCARBONS
    CAS
Compound
Rostad RI
 RI from this Study
Normal    Slow
Ramp[l]  Ramp[2]
496117 IH-Indene, 2,3-dihydro-
95136 IH-Indene
91178 Naphthalene, decahydro-
119642 Naphthalene, 1,2,3,4-tetrahydro-
447530 Naphthalene, 1,2-dihydro-
91203 Naphthalene
91576 Naphthalene, 2-methyl-
90120 Naphthalene, 1-methyl-
92524 l,l'-Biphenyl
939275 Naphthalene, 2-ethyl-
1127760 Naphthalene, 1-ethyl-
581420 Naphthalene, 2,6-dimethyl-
575417 Naphthalene, 1,3-dimethyl-
575439 Naphthalene, 1,6-dimethyl-
Naphthalene, 2-ethenyl-
581408 Naphthalene, 2,3-dimethyl-
571584 Naphthalene, 1,4-dimethyl-
571619 Naphthalene, 1,5-dimethyl-
208968 Acenaphthylene
573988 Naphthalene, 1,2-dimethyl-
Naphthalene, l-(2-propenyl)-
83329 Acenaphthene
644086 1 , 1'-Biphenyl, 4-methyl-
829265 Naphthalene, 2,3,6-trimethyl-
2245387 Naphthalene, 1,6,7-trimethyl-
86737 9H-Fluorene
1730376 9H-Fluorene, 1-methyl-
85018 Phenanthrene
120127 Anthracene
605027 Naphthalene, 1-phenyl-
Phenanthrene, 3-methyl-
Phenanthrene, 2-methyl-
613127 Anthracene, 2-methyl-
4H-Cyclopenta(def)phenanlhrene
Phenanthrene, 4-methyl-
832699 Phenanthrene, 1-methyl-
779022 Anthracene, 9-methyl-
35465715 Naphthalene, 2-phenyl-
Anthracene, 9-ethenyl-
206440 Fluoranthene
129000 Pyrene
168.87
170.83
173.31
195.47
195.85
200.00
221.57
224.53
236.59
239.27
239.88
240.89
243.76
244.06

247.04
247.15
247.62
248.75
249.67

254.98
256.12
264.99
267.43
270.77
289.49
300.00
301.75
312.74
319.19 [3]
319.93 [3]
321.47
321.77 [3]
322.81 [3]
323.79
329.52
330.73

344.68
352.77
169.04
171.05
173.01
195.21

200.00
221.51
224.42
236.34
239.06
239.30
240.42
243.42
243.96
245.81
247.04
246.55
247.00
248.88
249.53
253.94
254.87

264.98
267.94
270.57
289.27
300.00
301.84

319.08


322.52



330.85
334.13
344.78
352.54





200.00
220.98
224.16






245.39



248.48

253.53
254.36



270.28

300.00
301.59
312.91
318.88
319.76
321.16
322.42
322.87
323.67

330.76
334.63
344.83
352.72
  (Continued)
                                  390

-------
TABLE 1. LEE RETENTION INDICIES OF PAHs (Continued)
                                                     RI from this Study

CAS Compound
Benzo(a)fluorene
243174 Benzo(b)fluorene
Pyrene, 2-methyl-
Pyrene, 1 -methyl -
604535 l.l'-Binaphthalene
Benzo(ghi)fluoranthene
Cyclopenta(cd)pyrene
56553 Benz(a)anthracene
217594 Triphenylene
Benzo(c)phenanthrene
218019 Chrysene
1 -Phenylphenanthrene
Chrysene, 5-methyl-
612782 2,2'-Binaphthalene
1 2-Methylbenz(a)anthracene
7-Methylbenz(a)anthracene
Benzo(b)fluoranthene
Benzo(j)fluoranthene
207089 Benzo(k)fluoranthene
7, 12-Dimethylbenz(a)anthracene
Benzo(a)fluoranthene
Benzo(e)pyrene
50328 Benzo(a)pyrene
198550 Perylene
56495 3-Methylcholanthrene
193395 Indeno(l,2,3-cd)pyrene
215587 Benzo(b)triphenylene
53703 Dibenz(a,h)anthracene
Picene
191242 Benzo(ghi)perylene
Dibenzo(a,l)pyrene
Dibenzo(a,e)pyrene
Dibenzo(a,i)pyrene
Dibenzo(a,h)pyrene
[1] 40°C for 1 min., 40° to 310°C @ 10°C/min
[2] 40°C for 1 min., 40° to 310°C @ 4°C/min
[3] Value from Vassilaros
[4] Value from Lee
[5] Compound eluted during isothermal period

Rostad RI
366.72 [3]
369.17
369.40 [3]
373.45
385.23
389.92 [3]
396.55 [3]
398.77
399.45
391.24
400.00
421.66 [4]
420.20 [3]
421.81 423.91 [4]


439.51 443.13 [3]
443.13 [3]
440.04 444.02 [3]


452.29 [3]
448.69 454.02 [3]
451.27 457.17 [3]
462.09 468.44 [4]
489.49 493.24 [3]
490.13
491.01 496.20 [3]
500.00 [3]
500.00 500.29 [3]









Normal
Rampfl]
366.39
368.66
370.13
373.23

390.83
398.17
398.51


400.00




























Slow
Ramp[2]
366.48
368.93
369.91
373.27

390.82
398.41
398.53
399.94
390.89
400.00
421.17
420.40
424.01
421.24
425.90
442.34
442.45
443.05
443.90
446.30
452.22
454.06
457.02
468.88
492.36

493.42

500.00
540.01 [5]
554.76 [5]
560.34 [5]
563.59 [5]





                                 391

-------
TABLE 2. LEE RETENTION INDICIES FOR

CAS

230273
260946
85029

86748



4265252
132649

11095435
132650
239350
[1] Value from

Compound
Quinoline
Benzo(h)quinoline
Acridine
Benzo(f)quinoline
7-Azafluoranthene
9H-Carbazole
2-Azafluoranthene
Benz(c)acridine
9, 10-Anthracenedione
Benzofiiran, 2-methyl-
Dibenzofiiran
9-Fluorenone
Benzothiophene
Dibenzothiophene
Benzo(b)naphtho(2,l-d)thiophene
Vassilaros [2] Value from Lee
TABLE 3. LEE RETENTION INDICIES FOR

CAS
111842
124185
1120214
112403
629505
629594
629629
544763
629787
593453
629925
112958
629947
629970
638675
646311
629992
630013
593497
630024
630035
638686
630046
544854
630057




Compound
Nonane
Decane
Undecane
Dodecane
Tridecane
Tetradecane
Pentadecane
Hexadecane
Heptadecane
Octadecane
Nonadecane
Eicosane
Heneicosane
Docosane
Tricosane
Tetracosane
Pentacosane
Hexacosane
Heptacosane
Octacosane
Nonacosane
Triacontane
Hentriacontane
Dotriacontane
Tritriacontane
2,6, 10-Trimethylpentadecane
Pristane
Phytane
HETROCYCLIC PAHS


Rostad RI
210.32
301.50
303.99
307.30
350.50 [1]
309.22
347.39 [1]
393.41 392
330.53 [2]
184.50
259.75
293.88
201.84
296.03
390.12 389

n-ALKANES

Rostad RI
140.88
161.47
182.68
204.09
222.87
240.29
256.75
272.17
286.86
300.96
317.20
332.62
347.42
361.53
375.03
387.99
400.45
413.20
425.51
437.68
448.93
460.36
471.96
484.94
499.88










.60 [1]






.37 [1]

RIfrom
Normal
Ramp
138.27
159.66
180.44
200.12
218.83
236.21
252.61
268.29
282.99
296.84
312.43
327.63
342.11
356.14
369.44
382.35
394.59
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
NA[1]
275.46
283.84
298.42
RI from this Study
Normal
Ramp
210.37


306.86
350.37
308.63
347.15
392.06
330.26

259.74
293.64

295.85
389.64

this Study
Slow
Ramp
144.03
163.22
183.23
202.83
221.55
239.35
256.12
272.02
287.19
301.71
318.12
333.60
348.41
362.61
376.29
389.39
402.21
415.93
429.45
442.42
454.90
467.01
478.65
489.99
500.99
279.10
288.01
303.08
Slow
Ramp





309.24

392.11







































[1] Compound eluted during isothermal period
                                        392

-------
TABLE 4. LEE RETENTION INDICIES FOR SOME PRIMARY ALCOHOLS
                                    RI from this Study
CAS
111706
111875
143088
112301
112425
112538

112721

36653824
112925
629969
661198
Compound
1-Heptanol
1-Octanol
1-Nonanol
1-Decanol
1-Undecanol
1-Dodecanol
1-Tridecanol
1-Tetradecanol
1-Pentadecanol
1-Hexadecanol
1-Octadecanol
1-Eicosanol
1-Docosanol
Rostad RI
155.86
176.31
196.63
216.09
234.41
251.68

283.29

314.03
345.48
374.56
400.57
Normal
Ramp
154.52
175.63
195.98
215.20
233.42
250.63
266.83
282.29
296.98
312.09
342.84
371.04
397.16
Slow
Ramp
157.77
177.62
197.48
216.62
234.77
252.06
268.36
283.90
298.56
315.14
346.29
374.80
401.29
TABLE 5. LEE RETENTION INDICIES OF FATTY ACID METHYL ESTERS
                                                RI from this Study
CAS
111115
110429
111820
1731880
124107
7132641
1 12390
1731926
112618
1731948
1120281
Compound
Octanoic acid, methyl ester
Decanoic acid, methyl ester
Dodecanoic acid, methyl ester
Tridecanoic acid, methyl ester
Tetradecanoic acid, methyl ester
Pentadecanoic acid, methyl ester
Hexadecanoic acid, methyl ester
Heptadecanoic acid, methyl ester
Octadecanoic acid, methyl ester
Nonadecanoic acid, methyl ester
Eicosanoic acid, methyl ester
Rostad RI
187.83
226.04
260.05
275.62
290.32
305.23
321.34
336.74
351.38
365.43
378.87
Normal
Ramp
186.78
224.56
258.35

288.78
-
319.52
-
349.48
-
376.90
Slow
Ramp
188.24
226.16
260.37

291.18
-
322.66

352.84

380.64
                                393

-------
TABLE 6. LEE RETENTION INDICIES OF APPENDIX IX COMPOUNDS

                                                 RI from this Study
Compound
Pyridine
2-Picoline
N-Nitrosomethylethylamine
Methylmethanesulfonate
N-Nitrosodiethylamine
p-Benzoquinone
Ethylmethanesulfonate
Nitrosopyrrolidine
N-Nitrosomorpholine
o-Toluidine
3-Methylphenol
Acetophenone
N-Nitrosopiperidine
a,a-Dimethylphenethylamine
2,6-Dichlorophenol
HexacMoropropene
Resorcinol
N-Nitroso-di-n-butylamine
1 ,4-Benzenediamine
Safrole
Isosafrole
1 ,2,4,5-Tetrachlorobenzene
1 -Chloronaphthalene
1 ,4-Naphthoquinone
m-Dinitrobenzene
Pentachlorobenzene
1-Naphthylamine
2-Naphthylamine
2,3,4,6-TetracWorophenol
5-Nitro-o-toluidine
Diphenylamine/N-Nitrosodiphenylamine
Diphenylhydrazine/Azobenzene
1 ,3,5-Trinitrobenzene
Diallate
Phenacetin
Dimethoate
4-Aminobiphenyl
Pentachlorom'trobenzene
Pronamide
4-Nitroquinoline- 1 -oxide
Methapyrilene
Aramite
p-Dimethylaminoazobenzene
Chlorobenzilate
3,3'-Dimethylbenzidine
2-Acetylaminofluorene
7, 12-Dimethylbenz(a)anthracene
Hexachlorophene
3-Methylcholanthrene
Dibenz(a,j)acridine
1625-C [1]

112.18
116.23
125.97
135.55

146.43

178.41
177.76

175.81
188.31


207.47

215.58

219.97
230.20
228.25

241.72

260.55
263.47
265.58


276.62
276.62


288.47

294.81

299.19

339.60



430.13

453.37

Normal
Ramp
108.15
121.17
124.45
131.02
139.17
143.19
146.35
175.18
175.79
176.52
175.79
175.06
184.19
195.74
204.68
203.65
213.50
213.63
212.86
218.98
234.79
227.13
236.62
241.61
247.45
260.10
262.17
264.23
264.80
272.14
274.57
275.43
284.47
285.04
285.77
290.75
293.92
297.93
297.57
330.31
334.85
363.69
366.03
372.66
381.39
391.13
NA[2]

NA[2]

Slow
Ramp
121.47
130.12
132.65
137.85
144.23
147.33
150.77
176.52
177.39
177.86
178.68
176.57
185.84
195.47
203.40
203.60
216.79
215.35

219.88
235.68
227.19
236.15
241.61
248.35
260.09
262.20
264.62
265.40
272.81
276.00
276.42
287.18
287.02
288.72
292.33
295.26
298.71
300.49
330.82
337.56
364.71
368.20
371,32
380.70
390.32
443.82

468.69
487.56
[1] U.S.E.P.A. Industrial Technology Division Method 1625-C
[2] Compound eluted during isothermal period
                                     394

-------
TABLE 7. LEE RETENTION INDICIES FOR N-FATTY ACIDS

                                  RI from this Study
Compound
n-Heptanoic acid
n-Octanoic acid
n-Nonanoic acid
n-Decanoic acid
n-Undecanoic acid
n-Dodecanoic acid
n-Tridecanoic acid
n-Tetradecanoic acid
n-Pentadecanoic acid
n-Hexadecanoic acid
Hexadecenoic acid
Octadecenoic acid
n-Octadecanoic acid
Normal
Ramp
176.75
198.43
217.59
234.82
249.52
264.58
278.19
293.98
308.82
324.00
321.82
350.07
352.98
TABLE 8. LEE RETENTION INDICIES OF EXPLOSIVES AND
          THEIR RELATED COMPOUNDS

                                  RI from this Study
                                      Normal
             Compound                 Ramp
    Nitrobenzene                        180.05
    o-Nitrotoluene                       194.65
    m-Nitrotoluene                      202.07
    p-Nitrotoluene                       205.47
    1,3,5,7-Tetraazatri-
      cyclo[3.3.1.1(3.7)]decane            206.78
    1,3-Dinitrobenzene                   247.64
    2,6-Dinitrotoluene                    250.87
    2,4-Dinitrotoluene                    262.61
    1,3,5-Trinitrobenzene                 285.91
    2,4,6-Trinitrotoluene                  287.66
    4-Amino-2,6-dinitrotoluene            321.91
    2-Amino-4,6-dinitrotoluene            330.74
    2,4,6-Trinitrobenzenamine             337.56
    3,5-Dinitrobenzenamine               321.42
                                        395

-------
52  A QUICK PERFORMANCE-BASED HPLC METHOD FOR THE ANALYSIS OF
     POLYNUCLEAR AROMATIC HYDROCARBONS (PAHs)

     M.W. Dong and J. M. DiBussolo, The Perkin-Elmer Corporation, 761 Main Avenue,
     Norwalk, CT 06959-0250

     ABSTRACT
     A quick HPLC method for the analysis of polynuclear aromatic hydrocarbons (PAHs) in
     multimedia samples is described. The method is based on rapid extraction, direct injection
     of the diluted or concentrated sample extracts, fast HPLC separation, and selective
     programmed fluorescence detection. The elimination of sample cleanup reduces the assay
     time to one hour from sample extraction to data report.  A 16-min PAH separation was
     developed using a new HPLC column packed with 3-um polymeric CIS column
     materials. Method sensitivity was found to be 20 ppb levels for soil samples and low ppt
     levels for water samples. UV absorbance detection was used for the quantitation of
     acenaphthylene and for peak confirmation.  The method was applied to soil,  sediment,
     water, waste oil, and air particulate matter, and yielded excellent agreement with certified
     values from standard reference materials. This method is useful for routine monitoring of
     PAHs and for rapid sample screening in particular.  Method validation parameters in
     terms of precision, accuracy, sensitivity, selectivity, linearity, range, and ruggedness are
     documented and compared to those from official methods.  The proposed method
     provides a'quick turnaround method alternative to EPA Methods 550, 550.1, 610, 8310
     and TO-13. Advantages of this method include, cost-effectiveness, reduced solvent
     usage, and good recovery of all PAHs including volatile ones. Method limitations and
     precautions are also discussed.
                                           396

-------
53    The Extraction and Analysis of Polychlorinated Biphenyls (PCB's) by SFE and
        GC/MS.  Improvement of Net Detection Levels

        D. R. GERE.  L. G. RANDALL , C. R. KNIPE, W.  PIPKIN
        Hewlett-Packard, Uttle Falls Site, 2850 Centerville Rd. , Wilmington, DE 19808

        L. C. DOHERTY, Hewlett-Packard Company, Scientific Instruments Division, 1601
        California Ave., Palo Alto, California 94304
        Abstract

        Contaminants in soils,  sediments,  and solid wastes is  are   problem today  for  the general  public,
        government labs and private industrial labs. This paper examines applying supercritical fluid extraction
        (SFE) systems to the detection and analysis  of PCBs from an oil matrices,  solid wastes, sludges,  and
        other materials. In this work contaminants are extracted using SFE and analyzed by GC and GC/MSD.
        The PCB from oil matrix work  was done  to explore whether SFE would  decrease  coextractant
        interferences with the  GC/MSD and GC/ECD analysis relative to established manual  procedures -
        thereby, affording lower detection levels and improved quantitation.

        Introduction

        The detection and analysis of PCB's in soils, sediments, fish tissue and solid wastes is a common problem
        today for the general public, government labs and private industrial  labs. One of the necessary steps
        involves  sampling to assess what chemicals and products are present,  and to what quantitative extent.
        Supercritical fluid extraction, (SFE) is an expedient and cost effect means of removing chemicals from the
        solid waste samples as well as transformer oils, food, eggs and many other materials prior to analysis.

        Experimental

        SFE:  A  Hewlett-Packard  Model  7680T was used without modification.  The   conditions for  the
        experiments are described below.

        1. Extraction - pressure, 1218 psi ( 84 bar );  extraction chamber temperature, 40 ° C; density, 0.35
        g/mL; extraction fluid composition, CC^; static equilibration time, 2 minutes; dynamic extraction time,
        25 minutes; extraction fluid  flow rate, 3.0 mL/min; resultant thimble-volumes-swept, 28; 75 ml bquid
        CC>2  ( @ 4  ° C ) flowed during dynamic extraction.

        2. Collection (during Extraction) - trap packing, Hypersil ODS;  trap temperature,  40   ° C; nozzle
        (variable  restrictor) temperature, 45  ° C.
        3. Reconstitution (of collected extracts) — rinse solvent,  isooctane; collected fraction volume,  1.0 mL;
        trap temperature,  40  C; nozzle (variable  restrictor) temperature,  45 C; rinse  solvent flow rate,  2.0
        mL/min; fraction destination, vial #n.

        GC  : A  Hewlett-Packard Model 5890 Series  II GC/ECD with splitless injection,  and Hewlett-Packard
        HP-5 (30 m x 530 um, 0.88 urn film thickness) column was used.  The chromatographic system included
        the HP 7673 B injector.

                 A  Hewlett-Packard  Model 5890 Series  II GC with Electronic  Pressure Control (EPC),  a
                                                    397

-------
Hewlett-Packard 7673  Automatic Liquid Sampler with Auto  250/320 urn for automated  on-column
injection into a 250 urn ID column, and Hewlett-Packard Models 5972 and 5971  MSD were used. A
Hewlett-Packard HP-5MS column (30 m x 250 urn, 0.25 urn film thickness) with a 5 m retention gap
(250 urn ID) was used;  it is a low-bleed column designed specifically  for operation with the  MSD.
Deactivated glass press-fit connectors were used to connect the retention gap to the column and polyimide
resin was applied to either end of the retention gap for a firm seal.

Materials and Reagents

Solvents.    For the PCB's, the supercritical fluid extraction to remove PCB's from  the oil samples was
done at the Little Falls site and subsequently the extracts were sent to the  Scientific Instruments Division
for analysis.  Consequently, Pesticide grade iso-octane from Burdick and Jackson was used as the SFE
reconstitution solvent at the Little Falls  site; at the Scientific Instruments  Division (SID), iso-octane and
hexane, both OmniSorv grade from EM Science, were used in preparing standards.  Also  at the  SID
laboratory, ethyl acetate,  distilled in glass from Burdick and Jackson, was  used in the manual solid phase
extraction (SPE) clean-up of transformer oil.

Extraction fluid. Carbon dioxide, SFE/SFC ECD grade from Air Products, Allentown, PA,  was used as
the bulk extraction fluid in all the supercritical fluid extractions.

Samples.

The sample of Aroclor 1260 in transformer oil was obtained  from Environmental Resource Associates
(ERA), Arvada, Colorado. The Aroclor in the sample had a  certified concentration value of 49.2 mg/kg.
Sample aliquots of 100 uL were used for each conventional manual clean-up experiment carried out at the
Scientific  Instruments Division site using alumina solid phase extraction cartridges:  Extract-Clean ESP
obtained from Alltech (Stock No.  400825,  Alumina-N, 500 mg in 19 mL  reservoir). (The  'ESP'
cartridges, with teflon  reservoirs, are provided  especially for mass spectral analysis.)  Additionally, in
some of uie SFE work done at the Little Falls site, the oil was dispersed on activated alumina and/or
Chromosorb W-H.P. (80/100 mesh, Manville Products corp., Denver, CO).  The alumina was obtained
from Supelco, Inc., Bellefonte, PA (Alumina  F-l,  Catalog No. 2-0284, 80/100 mesh).  The procedure
outlined by Bellar and Lichtenberg ( 1  ) for activating the alumina was followed  prior to its use. Where
cited, filter paper disks, cored out of Whatman Qualitative filter paper, Catalog No. 1003-055, are placed
at bom ends of the sample.
 Results and Discussion

 PCB's are present throughout our environment, and cause concern from many perspectives ( 2 ).  The
 analysis  of samples containing PCB's is usually complicated by two factors (at least).  PCB's are present
 in trace  quantities, and  the  matrices are usually complex.  Usually  the matrices include  many  other
 compounds similar in polarity and chemical size.  It was observed during mis study  (which is still in
 progress) that PCB's have solubility characteristics in carbon dioxide that are very similar to transformer
 oil components and to a lesser extent similar to fats or lipids.  PCB's are very often present in the same
 mixture with such things as transformer oils, petroleum products and fats.

 In (his paper, we will present data from the selective fractionation of PCB's  via supercritical fluid
 extraction (SFE), and the subsequent detection and analysis with the HP 7680 T SFE and the  HP 5972 A
 MSD  .  The initial study involves the  optimization  of the SFE extraction conditions for PCB's in
 Transformer oil. Further work involved other matrices such as solid wastes and fish tissue. The SFE
 cleanup is compared to conventional cleanup such as the use of alumina cartridges and alumina admixing.
 Comparisons are drawn for sensitivity using both the  SIM mode and Scanning  modes with the  SFE
                                               398

-------
cleanup.   Preliminary  conclusions indicate  that SFE improves detection limits, and eliminates the
interferences of transformer oil.  Also, it appears that the  SIM mode works well with such samples, and
thus scanning would not be necessary.  The obvious advantages of the MSD over the use of the BCD with
GC will be discussed and summarized.

For some period of time, we have been selectively extracting (fractionating ) cholesterol from the fats
contained in  various food products via  SFE ( 3 ) as well as selectively extracting Vitamin A (beta
carotene) from cod liver oil ( Figures 1 and 2 ).  This experiment was usually optimized for maximum
selectivity and minimal time by judicious choice of temperature and density (pressure) ( Figure 3 ).

A characteristic of supercritical fluid extraction is its amenability to fractionation:  adjusting the solvent
power of the extraction fluid (through density, temperature, composition) can lead to selective solvation of
groups of compounds from a sample.  Another approach is to superimpose selectivity on the supercritical
fluid extraction process by mixing a chemically  active sorbent with  the sample so  that some of the
components  extracted from  the sample  are  preferentially  bound to the sorbent —  remaining in  the
extraction chamber  — while others are extracted and collected.  In either approach the net result is
somewhat analogous to what  is produced by applying large-scale column chromatography for clean-up
and fractionation of complex extracts; however, the whole process  is easily automated within an SFE
method.  Both approaches have been applied by other research groups to samples which are relevant to the
problem at hand —  e.g., the nearly selective extraction of PCB's relative to fats from sea gull eggs by
David et al ( 4 ) by employing an appropriately low density and the selective extraction of chlorinated
pesticides from chicken fat by mixing  the sample  to be extracted  with  alumina by King  et  al (16).
Similarly, in our work within Hewlett-Packard (), we have been mapping  extraction conditions at which
fats are largely  extracted and  those at which fats are largely not extracted (while other components like
cholesterol and Vitamin A  are).   These regions are outlined in Figure 8.   Note that  the "Fat Band"
demarcates the  division  into the regions  —  at  conditions  above  the  band (particularly  higher
pressures/densities) large quantities of fats coextract with other components, and  at conditions below the
band (lower pressures/densities) much smaller amounts of  fats coextract. Within the band, minor to bulk
quantities of fat can be extracted; the level seems to be heavily influenced by the matrix. The designations
of the regions outlined in Figure 9  are not technically rigorous but that diagram is a very useful empirical
tool to use in designing an SFE strategy — as demonstrated by this work on selectively extracting PCB's
from an oil matrix.
One of the  first published demonstrations of such experimental  fractionation by  supercritical carbon
dioxide was  by  Frank and Sandra ( 4 ) in a paper where they describe the fat free extraction of  PCB's
from sea gull eggs. They selectively extracted PCB's away from the associated fats. At one combination
of density and temperature the PCB's were extracted along with the fats (* 35 % by weight). However by
reducing the  temperature ( constant  density ) it was possible to obtain the PCB extract with only a 3 96
fat content.

We continued this type of exploration into the SFE of PCB's in transformer oil. Extraction of transformer
oil sample was first done at conditions roughly represented by point number " 1" in Figure 3  after having
mixed  the  oil  sample with alumina.  This experiment was  not successful,  possibly  due  to  poor
reproducibility of the procedure for activating the  alumina, which  is empirical at best,  involving  many
complex  variables such as the origin of the alumina, particle diameter and other parameters beyond the
scope of this study. This was followed by SFE with no alumina sorbent, and extracting the oil sample at
the milder conditions, point "2"  in Figure 3.   Point  2 was chosen since  a nearly fat-free extract of
cholesterol   had been obtained in routine experiments and demonstrations from  fatty matrices   with
similar conditions. This is illustrated in Figure 1 where cholesterol was selectively extracted away from the
trigrycerides ( as well as mono and di glycerides ).

A  comparative liquid solid cleanup of the PCB's from the oil matrix with  an acid clean-up  followed by
extraction with ethyl acetate was carried out.  The  extracts from the s SFE were compared to  the ethyl
                                               399

-------
acetate liquid solid extracts. Aggressive SFE conditions (point 1 in Figure 3),  did  not accomplish a
significantly better clean-up of die sample compared to the liquid-solid clean-up.  However, at the milder
SFE conditions (point 2 in Figure 3), the resultant SFE extracts are noticeably less complex than those
produced  by the manual clean-up.   This  is  shown  by Figures 4 and  5.   The results for  the
octachlorobiphenyl homolog series were quite similar to those of the heptachlorobiphenyl homolog series.
Notice  mat  mere  are still  some matrix  interferences  contributing  to  the  background  for  the
hexachlorobiphenyls. This leads men to carrying out the SFE experiment with conditions indicated as zone
3 in Figure 3.  One major goal will accomplished then if the SFE can selectively remove the PCB's from
a transformer oil matrix and effectively lower the mass spectrometer detector minimum sensing limits to
approach those of the clean standards. In this work an external standard curve was generated from  10 pg
to 500 pg.  A pattern for a standard solution of Aroclor 1260 could be recognized down to 5 pg injected.
Quantitation was done by choosing the three largest peaks in each homolog series and summing their areas
to give a total for each level of chlorination.  The work demonstrated that a S-ion SIM mode was fifteen
times more  sensitive man detection in scanning mode while  giving as much pattern  information as is
needed for identification ( 5 ). This sensitivity approaches that achieved by an  ECD.  The MSD, used in
the electron impact (El) mode, does not have a problem with response factor variation for the different
PCB congeners. Co-eluting electrophilic peaks mat make it difficult  to do ECD quantitation (e.g., DDT
and DDE) will not affect MSD quantitation because they are not of the same mass ( 8 ).
References


1. T.A. Bellar and J.J. Lichtenberg, "The Determination of Polychlorinated Biphenyls in Transformer
Fluid and Waste Oils," United States Environmental Protection Agency, Publication No. EPA-600/4-81-
045, September (1982).


2. Personal communication from  Barry Lesnik, Office of Solid Waste, US Environmental Protection
Agency, 401 M Street, SW, Washington, DC 20460.


3. D.R. Gere, " Selective extraction of Cholesterol from Fats in Butter Cookies and other Baked Food
Products" presentation at Midwest AOAC annual meeting, Champaign, 111., June 1992.


4. F. David, M. Verschuere, and P. Sandra, Fresenius J. Anal. Chem. 344. 479-485 (1992).


5. D.R. Gere, C.R. Knipe, P. Castelli, J. Hedrick,  L.G. Randall, J. Orolin, H. Schulenberg-Schell, R.
Schuster, H.B. Lee, and L. Doherty"Bridging the Automation Gap between Sample Preparation and
Analysis: SFE, GC, GC/MSD and HPLC Applied to Environmental Samples" manuscript in press,
J.Chronaatog.  Sci.


6. I.E. France, J.W. King, and J.M. Snyder, J. Ag. Food Chem. 39, 1871-74 (1991).


7. W. Pipkin,  D.R. Gere, L. Frank, and M. Fogelman,  "Optimization of Supercritical Fluid Extraction
Conditions for the Extraction of Fats from Food Products," Paper No. 737, The Pittsburgh
Conference.Georgia World Congress Center, Atlanta Georgia, March 8-12, 1993.


8. M.D. Erickson, Analytical Chemistry of PCBs. Butterworth Publishers, Stoneham, MA, 1986.
                                               400

-------
    Figure 1. Selective SFE of
     Cholesterol from Fats
D
(E
E
10;


 9;
  i

 8:


 7:


 6


 5


 4


 3:

  •
 2:


 1:
       LC fl 210,25  560,40
       LQ
      of NLE-CHOL.D
      of NLE-VITfl.D
Triglycerides/Fats
                   holesterol Peak
                    4       6
                    Time  (min.)
                                8
                     401

-------
Figure 2. Selective SFE of
 Beta Carotene from Cod Liver Oil
                    Triglycerides/Fats
                       Vitamin A Peak
                 4       6
                 Time (mlh.)
                 402

-------
    Figure 3. Density-Temperature Table
             for setting SFE Selectivity
40   50  60  70
                            Temperature  (C)
                          80   90  fOO IfO  120 130 140
   1.00
   0.95
   0.90
   0.85
   0.80
   0.75
j 0.70
E 0.65
   0.60
   0.50
§0.45
0 0.40
   0.35
   0.30
   0.25
   0.20
   0.15
526  618  711  804
383 1463  544  644
281 155] 420  489
211 269  32TU04
•164] 213  264  314
    175  218  261
             223
             097 990 1083
             680 764 665
             516 626 695
             447 500 565
             365p416 467      595
             305 348 3921436  510
                          690
             260  297  334 3721 425
             227  259  f290l 322
 93
 91  109  1
 89  104  122
                              284
                                  354
                              311
                              375
                              338 365
                              299 322
                              266 286
87: #2
84: #3
81
77
 IV
 60
75
63
flJB  120
101  111
 93  100
 82
 67  71
121
108  116
 94   99
 76   80
105 110 116
         91
He head
space line
                           403

-------
    Figure 4. Comparison of SFE to
Ethyl Acetate Extractoion for Hexachloro
        biphenyls in Aroclor  1260
       Hexachlorobiphenyl pattern for
       Aroclor 1260
      **
      m

	 Jill
4
11

50 pg standard
Extracted ion m/F
360
            I)    M    W
     tu
            HffP 7 T'T'T
            U    M
                JiuU-~*..
                            SFE of an oi sampte
                             Extracted ion mft=
                                 360
n't    u
                             OS sample elutedwth

                             Extracted ion m/z=
                                 300
                                 fll   Ml
                     404

-------
Figure 5.       Comparison of SFE to
      Ethyl Acetate Extraction for
 Heptachloro biphenyls in Aroclor 1260
     Heptachiorobiphenyi pattern for
     Aroclor 1260
I
. _ .
M | 'I M | M I I .1 ' •
                            60 pfl standard
                           Extracted ton mfe= 394
          II    II    tl    n't    •»
                           SFEofanoaumpli
                           Extracted ton mlSF 394
T-'I i-flrrtw*
                                  i j
                   «    Hi    u    ni
                        09 sample ekitedwiihEthy Acetate
                           Extracted ion mftF 394
                  - .  .. ,f * ,f—*~—m— ~
         *r i i i 'i i r i i i i i i i i i i » • i « i i
          u    u    t>   MI   ni
                                   MI
                       405

-------
54   Environmental  Analysis -  PAH's  in Solid Waste:  Bridging
     the Automation Gap  between   SFE and  HPLC


     D.R. Gere/  C.R.  Knipe,  William Pipkin and  L.G.  Randall
     Hewlett-Packard  Company,  Little Falls Site,  2850 Centerville Road,
     Wilmington,  DE,  19808,  USA

     ABSTRACT

     The detection and analysis  of  contaminants in soil,  sediments,  and solid wastes
     are a common problem today for the general public,  government  labs  and private
     industrial  labs  (   1).   This  work  with  polyaromatic   hydrocarbons  is  being
     undertaken  as the first  step  towards  deriving an  EPA  SFE method similar  to  the
     recent  draft method 3560  for  total petroleum hydrocarbons  (TPH)  (  2  ) .  This
     present study has been  in progress for  some time  at Hewlett-Packard.  The study
     focused on  the  extraction and  analysis  of PAH's  (polyaromatic  hydrocarbons,
     sometimes   called   polynuclear   aromatic   hydrocarbons).      There  are   two
     significantly different  parts  of the study.  The primary element of  the  study
     has been to provide the  experimental data  to validate  a tentative  alternative
     to the  liquid solid extraction of  solid  waste PAH's by  SFE.  It is relevant  and
     helpful  to  slightly modify   the standard EPA  method 8310  for a  couple  of
     reasons.   One simple need is  to  frequently use  an internal  standard,  and often
     times a surrogate as relative  measure of  the  extraction  efficiency.     Second,
     it was  helpful to operate  the  HPLC column at a higher linear velocity than that
     called  out in the  EPA method  8310.   This  modification  allowed increased  speed
     and resolution.  Also,  with regard to the  analysis part  of  the overall method,
     we made an  adaptation of  the tentatively optimized (final)   SFE steps to better
     accommodate the  analysis by GC and GC/MS  with  capillary GC columns.

     INTRODUCTION

     A  considerable   effort  was   directed  towards    the   optimization   of   the
     supercritical fluid extraction  steps. The primary  parameters  that  were   the
     temperature,  the density  or  the pressure  and  the  combination of  density  and
     temperature.   Another parameter   studied  in detail  was  the  use of modifiers  or
     co-solvents added  to the  flowing  extracting  stream of  carbon  dioxide.  Liquid
     modifiers  can be added  in an  almost infinite number of  permutations  of  types
     and amounts.

     Figure  1 is  an  outline  of the  study yielding  the  data in  this  manuscript. A
     brief initial study re-iterated  some of the HPLC method  steps  included  in  EPA
     Method  8310.  These  iterations resulted  in some minor changes  to  optimize  the
     time and resolution for sample throughput and  robustness.

     A second study (in  much  greater  detail)  reviewed the SFE extraction in light of
     previous studies and  experimental apparatus.  The  optimization in  this  study
     involved two major  areas in order to provide recoveries greater  than  95  %  for
     all 16  PAH's described  in EPA Method 8310. The first part  involved  use  of  sub
     ambient solid trapping  to ensure  the   recovery  of  the  relatively  volatile
     compounds.  Second  part   of  the   study  involved  the  optimization  of the  co-
     solvents or  modifiers  useful in  recovery of  the  relatively  large  and more
     intractable (from solid waste)
     PAH compounds.

     Experimental
                                       406

-------
Materials and Reagents

Reconstitution solvents.

The  reconstitution  solvents dispensed  by the SFE  instruments varied  according
to  the  particular  analyte  class  being extracted.    Furthermore,  these  same
solvents were used  to create the internal  and  external standard  solutions.  For
the     polyaromatic     hydrocarbons,     a     50/50     v/v     mixture     of
acetonitrile/tetrahydrofuran was  used:    both  were HPLC grade from Burdick  and
Jackson.

External calibration  standards.

For  polyaromatic  hydrocarbons,  a   standard  mixture  from   Supelco,   Inc.   in
Bellefonte  PA,  Polynuclear Aromatic Hydrocarbons Mixture  610-M (Catalog No.  4-
8743),   or  alternately, Hewlett-Packard  standard  mix  #  85-6035was used.    The
Supelco  standards are a mixture of 16 PAH's at  levels  ranging from 100  to  2000
ug/mL  in  a  50/50  mixture  of  methanol/methylene  chloride.  The  HP   standard
contains  the same  16  PAH's  at  the  500 ug/mL  level  in  acetonitrile.   These
standard samples were used in  the automatic preparation of  more  concentrations
by  serial  dilution corresponding to 1/2,  to  1/50 the original  concentration,
using a  50/50 mixture  of acetonitrile/THF as the diluent.

Extraction   fluid.

Carbon  dioxide,  SFE/SFC ECD grade  from Air Products,  Allentown,  PA,   was  used
as  the  bulk extraction  fluid  in all the supercritical fluid extractions.    In
the  PAH applications,  modifiers  were added to  the bulk  C02 extraction fluid.
In  the  PAH  work done at  the  Little  Falls laboratory,  these  were  methanol,
water,  and  methylene  chloride  (all,  HPLC  grade  from  Burdick  and   Jackson,
Muskegan,  Michigan),    forming  extraction   fluid  mixtures   of   95/1/4  (v/v/v)
CO2/methanol/water  in  one  method  and  95/1/4   (v/v/v)   CO2/methanol/methylene
chloride in  the other.

Mobile phase solvents  for  liquid  chromatography.

Acetonitrile  and water were  used;   their  source  and  purity are  noted above
elsewhere.

Samples.

For  the  work done  on  the PAHs  from soils,   spiked  samples were prepared in  our
laboratory  and  also   were  purchased  from Environmental Resource Associates
(ERA),  Arvada,  Colorado.  The   following  outlines  the  chemicals  and  procedure
used at  our  Little  Falls facility.   First,   soil  from a local yard  was prepared
by  blending  200 g soil  with an  equal  volume  of dry ice  prepared  from  SFE/SFC
ECD  grade  C02  (see above)  in  a  blender for  5  minutes,   manually  removing  the
remaining  large rocks  and sticks,  and then  sieving  the  cold mixture.    This
process  produces a  finely  divided soil  with its  moisture  content  still intact  (
3  ).   A PAH  standard  mixture,  Supelpreme-HC  PAH  Mix   (Catalog  No.   4-8905),
obtained from  Supelco, Inc. (Bellefonte, PA)  and having  16  PAH's at  levels  of
2000  ug/mL  was  diluted   by   a   factor  of  100  using  a   50/50  mixture   of
acetonitrile/tetrahydrofuran   (see  above).    One-hundred milliliters  of   the
resultant stock  solution was mixed with  100 g of the  finely divided  soil.  The
mixture  was  gently   stirred    for  3  hours   until   the   acetonitrile   and
                                    407

-------
tetrahydrofuran had  evaporated.   This  resulted in a  spike mix of  20 mg/kg of
each PAH in the soil.

Supercritical fluid extraction.

For SFE, Hewlett-Packard Models 7680T  and  7680A were used without modification.
In applications  requiring modifier  addition to  the bulk  CQ2,  Hewlett-Packard
Model 1050 pumps  (either  isocratic or quaternary models) were  used.   The pumps
were coupled to  the  SFE instrument  as  prescribed by  the  manufacturer.    A
contributed software program  controlled the SFE/modifier pump  system. Figure 2
depicts a block diagram of the combination  of  the SFE with  a modifier pump with
the modular HPLC to create the "bridge"  system used in an automatic sequence to
maximize overall sample  throughput and to  minimize  operator, intervention, thus
providing  reduced  overall  system error  (improved  sample  %  relative standard
deviation or % RSD).

Liquid chromatography.

A Hewlett-Packard Model  1050  LC  equipped with  an MWD diode array detector,  100
vial autosampler  and  quaternary  gradient  pump was  used without modification.
The separation was  carried out on  a Vydac C18  (25  cm x 4.6 mm,  5  um particle
size) column.

Analytical Procedure,Extraction

A  quantity of three  grams  of soil  was extracted  each time  in applying  the
method  to  the certified  spiked  sample  for the assay  of  PAH's  from  the soil.
When screening with  real samples  of high  concentration, sometimes  only  1 gram
of soil was  used,  admixed with an  equal amount of inert material  such as sand
or Chromosorb  W-H.P.   Two  similar SFE methods  have been developed — one  for
analysis by HPLC and one for analysis by GC.   In both  the  HPLC- and  GC-focused
SFE methods, there were  three extraction steps with  a  single fraction produced
for  the SFE recovery  of PAH's  in  the  final   method.   The  purpose  of  Step  1
(Figure 5  ) was  to collect more  volatile PAH's. The  purpose  of  Step 2 (  Figure
6 ) was to  collect  involatile PAH's. The purpose of  Step 3  (  Figure 7 )  was to
sweep modifier from the instrument  and raffinate before de  pressurization  and
to  reconstitute  the  involatile  PAH's  in  the  same vial  containing  the  more
volatile PAH's from Step 1.

PAH SFE Method for subsequent HPLC analysis.

The following conditions for Step 1  are grouped according to function.

1. Extraction  — pressure,  1749  psi;   extraction  chamber temperature,  120 C;
density,  0.22  g/mL;    extraction  fluid  composition, CO2;  static equilibration
time,  2 minutes;  dynamic extraction  time, 10  minutes;  extraction  fluid flow
rate, 2.0 mL/min;  resultant thimble-volumes-swept,  12.    2.  Collection  (during
Extraction)  —  trap  packing,  Hypersil  ODS;  trap  temperature,  5  C;  nozzle
(variable  restrictor)   temperature,   55  C.    3.  Reconstitution  (of   collected
extracts)   —  rinse  solvent,  50/50  (v/v)   THF/acetonitrile;  collected fraction
volume,  0.8   mL;   trap  temperature,   60  C;   nozzle  (variable  restrictor)
temperature, 45  C;  rinse solvent  flow rate,  1.0 mL/min;  fraction destination,
vial #n.

The following conditions for Step 2  are grouped according to function.
                                    408

-------
1.  Extraction —  pressure,  4899  psi;  density,  0.63  g/mL;  extraction  chamber
temperature,   120   C;      extraction   fluid   composition,   95/1/4   (v/v/v)
C02/methanol/water;  static  equilibration time,  1 minutes;  dynamic  extraction
time,  30  minutes;  extraction  fluid  total  flow  rate,  4.0  mL/min;  resultant
thimble-volumes-swept, 25.    2.  Collection (during Extraction)  —  trap packing,
Hypersil ODS; trap  temperature,  80 C;  nozzle  (variable  restrictor)  temperature,
45 C.  3.  Reconstitution  (of collected extracts)   —  none.

The following conditions  for Step  3 are grouped according to  function.

1.  Extraction —  pressure,  4899  psi;  density,  0.63  g/mL;  extraction  chamber
temperature,  120 C;   extraction  fluid  composition,  CO2;  static  equilibration
time,  5 minutes;   dynamic  extraction  time,   10   minutes;  CO2  flow  rate,  4.0
mL/min;  resultant thimble-volumes-swept,  8.    2.  Collection  (during Extraction)
—  trap  packing,   Hypersil  ODS;   trap   temperature,  80   C;  nozzle  (variable
restrictor)  temperature,  45 C.    3. Reconstitution  (of collected  extracts)  —
rinse solvent, 50/50  (v/v)  THF/acetonitrile;  collected fraction volume,  0.8  mL;
trap  temperature,  80  C;  nozzle  (variable restrictor)  temperature,  45 C;  rinse
solvent  flow  rate,  1.0 mL/min; fraction destination,  vial #n.

Method for subsequent GC  analysis.

The following conditions  for Step  1 are grouped according to  function.

1.  Extraction — pressure,  1749 psi;    extraction chamber  temperature,  120  C;
density,  0.22 g/mL;   extraction  fluid  composition,  CO2;  static  equilibration
time,  2  minutes;  dynamic extraction  time,  10 minutes; extraction fluid  flow
rate, 2.0  mL/min;  resultant thimble-volumes-swept, 12.    2.  Collection  (during
Extraction)  —  trap   packing,   Hypersil  ODS;  trap  temperature,  5  C;  nozzle
(variable  restrictor)  temperature,  55   C.    3.   Reconstitution  (of  collected
extracts)  -- rinse solvent,  1/1  (v/v)  methanol/methylene  chloride;  collected
fraction volume,  0.8  mL; trap  temperature,  30 C;  nozzle  (variable  restrictor)
temperature,  45  C;  rinse solvent  flow  rate,   1.0  mL/min;  fraction  destination,
vial  #n.

The following conditions  for Step  2 are grouped according to  function.

1.  Extraction —  pressure,  4899  psi;  density,  0.63 g/mL;  extraction  chamber
temperature,   120   C;      extraction    fluid   composition,    95/1/4    (v/v/v)
C02/methanol/methylene  chloride;  static  equilibration  time,  1 minutes;  dynamic
extraction  time,  30  minutes;   extraction fluid  total  flow  rate,  4.0  mL/min;
resultant  thimble-volumes-swept,   25.     2.   Collection (during  Extraction)  —
trap  packing,   Hypersil  ODS;  trap   temperature,   80  C;   nozzle   (variable
restrictor)  temperature,  45 C.   3.  Reconstitution  (of  collected extracts)
none.

The following conditions  for Step  3 are grouped according to  function.

1.  Extraction —  pressure,  4899  psi;  density,  0.63 g/mL;  extraction  chamber
temperature,  120  C;   extraction  fluid  composition,  C02;  static  equilibration
time,  5 minutes;   dynamic  extraction  time,   10   minutes;  CO2  flow  rate,  4.0
mL/min;  resultant thimble-volumes-swept,  8.    2.  Collection  (during Extraction)
—  trap  packing,   Hypersil  ODS;   trap   temperature,  80   C;   nozzle  (variable
restrictor)  temperature,  45 C.    3. Reconstitution  (of collected extracts)  —
rinse  solvent,  1/1   (v/v)  methanol/methylene  chloride;    collected   fraction
volume,   0.8   mL;   trap  temperature,   30  C;   nozzle  (variable   restrictor)
                                    409

-------
temperature, 45  C;  rinse solvent  flow  rate,  1.0 mL/min;  fraction destination,
vial ttn.

Liquid Chromatography

The  liquid  chromatographic  conditions  for  PAH's  were  the  following:    an
injection   volume   of   10   uL,   a   column   oven  temperature   of  30   C,   a
water/acetonitrile mobile phase which was  programmed isocratic  40 % water for 3
minutes   followed   by   3   different   ramps    to   increase  the  acetonitrile
concentration to a final value of 100 % for 3 minutes.

Results  and Discussion

The extraction  from solid  samples of  polyaromatic  hydrocarbons   (PAH's)  is  an
application  which  has  been  receiving  significant  attention  with respect  to
getting an  acceptable,  robust  SFE  method  developed  and ultimately formalized as
an  EPA draft method.

EPA Method  8310 describes  the HPLC  analysis  to be  used for  PAH's.    In this
work,  this  method was modified slightly for the consideration of  analysis time,
and  resolution  of   the  16  prescribed  compounds.     In   order  to  provide
chromatographic  resolution  of  both the internal  (  biphenyl or bromo  benzene )
and  surrogate   (m-quaterphenyl)  standards,  it  was  found  to  be necessary  to
change  the  programming  of  the mobile  phase and  to  operate  at  higher  linear
velocities.  A  representative  liquid  chromatographic separation is presented in
Figure 3.

In developing and optimizing the SFE  method,  personnel from the Hewlett-Packard
laboratory  in   Waldbronn,   Germany  and  the  Inland  Consultants   laboratory  in
Chicago,  Illinois, USA  have collaborated with  the Hewlett-Packard Little Falls
site laboratory.  Additionally, the group  of  Dr.  H.B.  Lee  of the  National Water
Research  Institute,  Environment  Canada  in Ontario,  Canada  further modified the
method outlined  here specifically  to use  with  gas  chromatography  and  GC/MS  as
the analytical  techniques  ( 4 ) .  Over the  last six months, our work  has been
directed towards optimizing the supercritical fluid extraction  method itself —
including studying the  effect  of temperature,  density  (or  pressure),  and, most
importantly, composition of  the  extraction  fluid.  Throughout  the course  of
developing  the method,  we utilized  three  different ways of exposing the sample
in an SFE system to  a modifier/CO2 solution:    addition  of  liquid modifier into
thimbles  containing  the samples  (e.g.,  1  mL  liquid to  10  g soil  sample),  the
use of premixed tanks  of liquids plus  CO2  (e.g.,  5 weight  % methanol  in CO2),
and the use of  an  external  pump  to meter modifier into  the  CO2 stream ahead of
the extraction  chamber.   The method described here  can be  carried  out with
either premixed tanks   or the  external  modifier pump.   Without  some  automated
means of  wetting the  sample  directly  with  the modifier part  way  through  the
method,  the  approach of addition  into the thimble often provides lower average
recoveries  for  the whole range of PAH's studied.

The  "PAH class  of  compounds"  (  listed  in Table  1   ) includes    a   range  of
compounds that  is  broad with  respect  to  volatility  and molecular  weight.  The
SFE method  development  described  here  focused on  two  (overlapping)  groups  of
PAH's,   with  the  result  that  each  group  is  extracted,  concentrated,  and
reconstituted best by  different  SFE  conditions.  We  defined these as  different
substeps making  up  a full  method.  These groups  are  considered  to  include  the
relatively  volatile,  smaller  PAH's   and  the less  volatile,larger  size  (  and
molecular weight)  PAH's.   In   Figure 4  the  individual  compounds are  sorted
according to size,  going from  smaller compounds at the left hand  side to larger
                                    410

-------
ones. The  recoveries  are low  at  both ends —  the most volatile  and the  least
volatile  (largest  molecules)  -- with  a  maximum in  recoveries  for intermediate
PAH's.  In  Figure  4 have plotted   earlier data from a  USEPA report  (where the
apparatus did  not  have the luxury  of  such items  as  sub-ambient  solid trapping
and  modifier  pumps )   and added  the  composite data  of the  three laboratories
participating in this methods  development  work.   The data  are outlined in  Table
2.

*
I
2
3
4
5
6
7
8
Compound
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
1
9
10
11
12
13
14
15
16
Compound
Benzo (a) anthracene
Chrysene
Benzo (b) fluoranthene
Benzo (k) fluoranthene
Benzo (a) pyrene
Dibenzo (a, h) anthracene
Benzo (g,h, i) perylene
Indeno (1,2, 3-cd) pyrene
Table 2. A Comparison of Recoveries of PAH's
obtained in Prior Work and in This Work
after Method Optimization (% Recoveries)
PAH
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
EPA
Method- 1
(11)
40.0
53.1
80.5
71.4
98.2
65.7
79.3
77.2
61.8
60.0
47.2
47.2
30.1



EPA
Method-
2(11)
50.3
56.4
87.1
77.3
90.8
68.2
83.0
67.1
58.6
53.5
38.7
38.7
22.6



Grand Mean
Mean of
EPA
Reports
45.0
54.5
83.8
74.2
94.5
66.9
81.1
72.1
60.2
56.8
43.0
43.0
26.5



61.7
This
Study -
Little
Falls
103.4
84.5
78.8
119.9
143.3
50.0
104 ..0
66.0
99.4
114.8
95.9
95.1
75.3
99.2
89.3
88.4
This
Study --
Waldbronn
124.0

72.0
117.0
127.0
1000.0
106.0
40.0
156.0
96.0
104'. 0
79.0
77.0
53.0
46.0
112.0
This
Study --
Inland
Env
80.0






97.0

105.0


95.0




Mean of
This
Study
102.4
84.5
75.4
118.4
132.6
75.0
105.0
67.8
127.8
105.3
100.0
88.0
87.0
76.1
67.7
100.2
94.6
                                   411

-------
In this  study,  the  PAHs,  have been  divided into  two groups—the  early
eluting  ones  (in  the  chromatography)  referred  to  as  the  relatively
volatile  PAH's  and  the remaining  PAH's  which are  larger  in  size  and
molecular weight  (and more intractable for extraction  in solid wastes).

The recoveries  at either  end  of  the PAH  list  are  improved.   What  we
believe  has  been happening  in prior  work  (  5,6  }is  that  conditions
appropriate  for extracting and collecting intermediate   and large  PAH's
result  in the  volatilization  of  the more  volatile  compounds  —  e.g.,
temperatures  are  too  high  and extraction  times  are too  long  thereby
resulting  in  too  much  expanded  CO2  sweeping   over  and  through  the
collected  extracts.   Conversely,  those same  extraction conditions  were
not aggressive enough to extract   the larger PAH's — particularly,  with
respect to temperature and composition.  These  problems were  overcome   by
going to  a multiple-step  procedure using an SFE  system  which   I.  has a
solid  trapping   system  capable  of  thermal   operation  at  sub-ambient
temperatures for  volatile  collection  and high  temperatures for operation
with modifiers and   2. also has the  capability of automatically changing
the extraction fluid composition from one step  to another step.

The details  were  outlined  in the  Experimental  section.  Briefly, part  of
the PAH's  were  extracted  and recovered in the  first  step  using pure  CO2
at moderately  low density and  temperature  and with cold  trapping on  an
ODS trap.   These were  reconstituted  into an autosampler  vial  with  just
0.80 mL collected fraction volume.  The rest of the PAH's were removed  by
a subsequent step using a mixture of CO2 with  water  and methanol as  the
extraction   fluid,  higher   operating  temperature  and  density  in   the
extraction region,  and  a  higher temperature in the  trapping region  with
the ODS.  The PAH's are not  reconstituted directly after the  second  step.
A short third step with pure CO2 (but with all  other conditions as in  the
second  step)  is  used  to  purge  the  system of  modifier  before   de
pressurization.   The analytes recovered in the  second  step  (and possibly,
any moved during  the beginning of the third step)  are  reconstituted  in
the same  autosampler vial containing the  first fraction — with another
0.8 mL  collected  fraction  volume.   Therefore,  all recovered  analytes  are
merged automatically into a  single fraction to  be analyzed.

The data  in Table  2 are plotted  in  Figure 4. The bar graphs  show  the
percent recovery  for the 16  EPA  PAH  compounds.   For  instance,  this  1st
compound  is  naphthalene,   the  last one  is  indeno  (1,2,3-cd)  pyrene  and
compound number 6 is anthracene.

The solid  bar  graphs start at a  relatively low  percentage and increase
and then  decrease  again,  we   Two  things,  a  solid  trap  which  can  be
operated  at  sub  ambient temperature  during  the  extraction  step  and  the
use of modifiers, have  allowed  2  significant advantages.  One allows  the
recovery of  the   relatively  volatile  PAH compounds.  The clear bar graphs
represent the mean  of the two  labs within  HP,  the private environmental
lab in  Skokie  ,  Illinois  (  7 ) and the  environmental  lab in Burlington,
Ontario  Canada.   It  can  be   seen  that  the  recovery  is   significantly
improved for all  16  PAH compounds.   This is  the  resultant   of  the  two
separate  phenomena.  First,  the volatiles    are  extracted  under    mild
conditions and  rinsed   to  the reconstitution vial   (  from  the  solvent
trap)  before the  later  step  with  more  aggressive extraction conditions.
Secondly, these more  aggressive conditions  with the modifier extract  the
larger  less  soluble (  in  CO2  only ) .  During  the  extraction step  these
                                    412

-------
compounds  are  precipitated on  the  solid  trap  (  which  is  now  at  a
substantially  elevated  temperature  )  and  then  rinsed   into   the  same
vial.  The   grand mean  recovery for the  16 compounds  (  multiple  sets  of
experiments  in  four  different labs  in  three different  countries  !   }  is
96'. As  a  test of   our  hypothesis about the  volatiles  and the modifier
aiding the large PAHs, we then  carried out experiments  which  divided the
3 steps such  that  the reconstituted  solutes from step one were  analyzed,
and  then  the extracts from steps  2  and 3  (combined)  were injected  into
the  HPLC and analyzed.  In Figure 8,  the results  from  step one are  shown
as solid bar graphs,  while the result from the combination of  steps  2 and
3 are shown as the open or clear bar graphs.
This data would appear to support the hypothesis very well.

 Three are  three  different  groups  of compounds  that  are detected by the
HPLC UV  detector, with  essentially  3  different  relative sensitivities.
The  first  sub group  is  represented by  naphthalene. At the wavelength  of
detection called  out  by  the EPA HPLC method  8310  (  254  nanometers),  the
minimum detectable quantity  (of the  net SFE and HP analytical method)  at
a  signal  to  noise ratio of 4:1  is  0.100  mg per kg  (ppm weight/weight
basis)  for naphthalene.   The  most favorable compounds are represented  by
anthracene  naphthalene.  Anthracene detected  at  254  nm,   is  being sensed
virtually the lambda maximum,  and so now the minimum detectable quantity,
at  the  same  signal  to  noise  ration is  0.005  mg/kgm.     Indeno(1,2,3-
cd)pyrene represents  the  most  typical  type of   the  PAH  compounds--has a
minimum  detectable  quantity  of  0.030  mg  per  kg.      All  of  the  PAH
compounds together meet the EPA solid waste criteria for sensitivity.

Finally,  it  is relevant to mention  an  interesting observation   seen  by
the  lab   at  the  Canadian  Center  for  Inland  Waterways  in   Burlington,
Ontario,   a   branch  of Environment  Canada  ,  H.B.  (Bill)  Lee   and his
coworkers.   Their experiments unlike   the   experiments  of  other  three
labs collaborating,  in  this study (the Wilmington,  DE HP site,  the   HP
Germany  site,  and Inland Environmental Consultants in Chicago)  involved
the  use  of  gas  chromatography  (  GC/MS)   for  the  analysis   step.   This
brought   some concern about the 4^  water  as one  of  the co-solvents.  If
any  of the  water  remains  in the  reconstitution solution  there might  be a
detrimental  effect  on  the  GC  fused   silica  capillary  column  and the
injection procedure.   This  lab modified the original  SFE method from   le
methanol, 4*  water to 1-  methanol, 4* methylene chloride  (with all  other
conditions   the  same) .   This is  work  which  we duplicated  then  in our
Wilmington  laboratories  and we found no  significant  difference  in   the
overall recovery. This work is described in a separate publication .

In   summary   ,  this   particular  method  appears  robust  enough  to  be
considered  for a  Round Robin   study  such  as occurred  with the EPA method
3560 TPH last year.


Acknowledgment

Many people have contributed to this work by sharing ideas and results  of
work in progress.  They include Barry Lesnik, Tom  Peart,  Robert Hong-You,
Pat  Sandra,  Frank David,  Jerry King, Viorica Lopez-Avila,  Lee Altmayer,
Werner Beckert, Steven Pyle, Robert Marsden, and Philip Wylie.
                                    413

-------
References
l.D.R.  Gere,  C.R.  Knipe,  P.  Castelli,  J.  Hedrick,    L.G.  Randall,  J.
Orolin, H.  Schulenberg-Schell, R.
Schuster, H.B.  Lee, and  L.  Doherty"Bridging  the Automation Gap  between
Sample  Preparation  and   Analysis:   SFE,  GC,  GC/MSD and HPLC  Applied to
Environmental Samples" manuscript  in  press,  J.Chromatog.  Sci.


2. Personal communication from Dr. Barry  Lesnik, Office of  Solid Waste,
US Environmental  Protection  Agency,   401  M Street,  SW,  Washington,  DC
20460 USA.


3.  Dr. Mary Ellen  MacNally, Agricultural  Products,  Experimental
      Station,   E402/3328B,   E.I.   du   Pont  de  Nemours  and  Company,
Wilmington, DE USA.


4.   H.B. Lee,  T.E. Peart,  R.L.   Hong-You,  and  D.R.Gere,  "Supercritical
Carbon  Dioxide  Extraction  of  Polycyclic  Aromatic  Hydrocarbons  from
Sediments," manuscript submitted for  publication.

  5. W.  Beckert,  "An Overview of the  EPA's Supercritical  Fluid  Extraction
 (SFE)  Methods  Development Program,"  ACS Symposium:  Supercritical Fluids
in Analytical Chemistry sponsored  by       the    Division   of   Analytical
Chemistry at the 201st National meeting of the American Chemical  Society,
Atlanta, Georgia, April 14-19, 1991.

6.  V.  Lopez-Avila,  N.S.  Dodhiwala,  and  J.   Benedicto,  Evaluation  of
Various  Supercritical Fluid  Extraction Systems for Extracting  Organics
from  Environmental  Samples,  Final  Report for Work     Assignment  1-1,  EPA
Contract 68-C1-0029, Environmental Monitoring  Systems Laboratory,  Office
       of Research  and Development, U.S. Environmental  Protection  Agency,
Las Vegas, Nevada   89119,  February, 1992.

7.  Personal communication from John   Orolin,Inland  Consultants,  Inc.,
Environmental  and  Engineering Services,     3921 Howard St., Skokie,  IL,
60076,  USA  .
                                    414

-------
     Figure  1 :
Outline of Study
Verify the HPLC method
- speed of analyRis.Reaolution
- Internal Standard (biphenyl or bromo
  bonzeno)
Surrogate (Quaterphenyl)
Optimize SFE parameters
  Temperature
-• Density (pressure)
  Modifier
- Volatiles .Large mw PAH'n
  Basic Parameters
 • Modifier
  — add to thimble
  - Pre-mixed tanks
  "• On-line modifier pump
  Recovery
  Reproducibility
Figure 2 : SFE-HPLC Bridge with Modifer
                       Pump
   jr¥5
                                                    LC
  Mod/fie*
    Pump
                          415

-------
      Figure 3 :  EPA PAH Compounds HPLC
Elle  graphics  integration Qeports  Abort Help
               13] 003-01 OLD: MWD A. Stu-254.16 Rcl-390.20
    Figure 4  : Comparing HP Method to Method Without
        Modifier and Without Solid Subambient Trapping




100







20

0



















—














—














..._











•;


\ 	








; .














,





._..








,














,














,





	 	 	 	 	 	 — 	




im
T 	 "•;;
_J -—51— 	 	
1 1
, 1 ; IV";;, , , ,





	








,
m MI
l AH

120

100
80

60

JO




1 AM
                56789
                  t.SEPA PAH Compounds (increasing size)
                      • EPA mean   HP mean
                             416

-------
   Figure 5   :   SFE Conditions Step 1
                 Sample - 3.0 g Soil
HP7680T SFE + HP1050 Pump: Step 1
Extiaction    "Pressure            1749 psi
             • Temperature        720 C
             • Composition         CO2
             • Static Equilib.Time   2.0 mm
             " Dynamic Extn. Time   10.0 mitt
             * Flow Rate           2.0 mL/min
< i,-[t:- ?!...•!•:     "Trap               OD5
             * Temperatures       5 C.trap; 45 C nozzle
Reconstitution • Solvent Fraction Vol.  THF/ACN, 0.8 mL
             • Temperatures       eOC, trap; 45 C noz
             • Flow Rate           2.0 mL/min
  Figure 6 :  SFE Conditions Step 2
                  Sample - 3.0 q Soil
HP7680T SFE + HP1050 Pump: Step 2
Kxtrn.'timi    -Pressure            4950 psi
             * Temperature        120 C
             « Composition         CO2 + 1 % MeOH +4% H2O
             * Static Equilib.Time    2.0 min
             • Dynamic Extn. Time   40.0 min
             • Flow Rate           2.0 mL/min
Collection     "TraP               ODS
             = Temperatures        80 C,trap; 45 C nozzle
 Reconstitution : THIS STEP DOES NOT REQUIRE A RINSE
                         417

-------
      Figure 7   : SFE Conditions Step 3
                   Sample - 3.0 g Soil
  HP7680T SFE + HP1050 Pump: Step 3
  Extraction
» Pressure
" Temperature
» Composition
» Static EquilibTime
« Dynamic Extn. Time
» Flow Rate

» Trap
« Temperatures
•4950 psi
120 C
CO2
0.5 min
10.0 mm
2.0 mL/min

ODS
80 C,trap; 45 C nozzle
  Reconstitution. Solvent, Fraction Vol. THF/ACN, 0.8 mL
               • Temperatures       eOC, trap; 45 C noz
               • Flow Rate          2.0 mL/min
Figure 8 : Separating the Variables: Recovery of the
      Volatile Fraction of PAH's vs. Larger PAH's






40 • 1 ^ • '' • | i • ' ' •

i! i i! 1

I
E . --tfiHy-

1 *
-Lflwi n i

i I
^ 1

1
1
j
H-l



i
1

i
j

j





1
1
1
1








JL



ffi









                            8  9 10 II 12  13 14 IS  16
                      \ \ Recovery [  |X Recovery
                      iStep)   I—I Steps 2& 3
                           418

-------
55        ANALYSIS OF PCB'S  IN SOIL, SEDIMENTS, AND
           OTHER  MATRICES  BY ENZYME IMMUNOASSAY


       Robert O. Harrison, ImmunoSystems, Inc., Scarborough, Maine  04074
       Robert E. Carlson, Ecochem  Research Inc., Chaska, Minnesota  55318
          Alan J. Weiss, Millipore  Corp., Bedford, Massachusetts  01730


      ABSTRACT

      A competitive inhibition E_nzyme ImmunoAssay (EIA) for the determination
      of polychlorinated biphenyls  (PCB's)  is  commercially available as the
      EnviroGard™ PCB Kit.  Soil sample preparation can be performed in the
      field using disposable kit components.   The test can analyze PCB's in the
      field in less than 30 minutes total,  using  no specialized equipment. Test
      specificity is restricted to PCB's, primarily Aroclors 1016,  1242,  1248,
      1254,  and 1260.  The kit can be  used to screen unidentified Aroclors at
      4 levels from 1 to 50 ppm in  soil.  This method has been  reviewed  by the
      EPA Office of Solid  Waste  and will be  proposed for  inclusion  in the
      4000  series  of  screening  methods in  the  next  SW846  update.
      Quantitative data  for soils  in this concentration range  can be  easily
      obtained by standardization  of  the kit with the same Aroclor as  in the
      samples (if known).  Quantitative  EIA  analyses of river  sediments were
      completed the same day in a  portable lab  on site,  with a tenfold  increase
      in sample  throughput and a tenfold  decrease in cost per sample compared
      to GC-ECD.   Data for  several methods and  matrices are presented.
      These results demonstrate that EIA kits  can effectively analyze for PCB's
      in many situations at a fraction of the cost  and in a fraction of the time of
      standard methods.

      INTRODUCTION

      Reagent and Method Development

      The  development of  the  EIA  for PCB's followed  these  steps:   1) PCB
      derivatives were  synthesized for conjugation to proteins;  2) one or these
      PCB derivatives  was conjugated to a  carrier  protein and the  resulting
      conjugate  was  used to immunize animals,  which then produced antibodies
      recognizing  both the  PCB derivative  and PCB's;   3) a PCB derivative
      was conjugated  to horseradish  peroxidase (HRP) to make a conjugate
      which  can  be  captured by anti-PCB antibodies;   4)  the  PCB-HRP
      conjugate was  used  to  screen and  select antibodies;  5) the  selected
      system  was optimized   for   PCB  sensitivity and solvent and  matrix
      tolerance, then  characterized for specificity;  6)  sample preparation
                                          419

-------
methods  were developed for  specific sample types;   7)  these methods
were validated using Held samples.

PCB EIA Procedure

Figure  1  schematically illustrates the procedure used  for the analysis of
samples  containing RGB's.  In summary:  1)  rabbit antibodies which
recognize the PCB structure are immobilized  on the walls of plastic test
tubes;  2)  PCB's in solvent, in  the form of either standards, calibrators, or
samples  prepared  as  described below,  are mixed with Assay Diluent in
tubes,  allowing PCB's to  be  captured  by  the  immobilized antibodies
[incubation  1).  PCB's are specifically retained on the solid  phase when
the rest  of the  sample  is washed  away  (wash  /);  3)    PCB-enzyme
conjugate is added to tubes and bound  in the same manner  as  in step 2
[incubation 2}. The unbound conjugate is washed away [wash 2} and the
amount retained by the immobilized  antibody  is  inversely proportional to
the amount of PCB bound in step 2;  4)  enzyme substrate and chromogen
are added  to  the tubes for  color development  by the bound enzyme
[incubation 3).  The  intensity of color  is proportional to  the amount of
captured  enzyme  and is inversely proportional  to the amount of PCB
bound  in step 2. Therefore, more color means less PCB.  Incubations
are  typically,  but  not always, 5 minutes.   Washes use  tap  water.
Sample and  calibrator volumes vary with the  sample  type and  protocol,
but are generally 5, 25, or 100 u,l.  Incubation 1  volume is 500  uJ.

Sample Preparation

Sample  preparation  methods are given  below  for  several  matrices.
Where EIA analysis is indicated, the extract is used as described in step
2  of the  EIA  Procedure above (incubation 1  of  Figure 1).
      Soil sample  preparation is summarized as follows:  Weigh 5  g soil
on portable balance and place in  polypropylene extraction bottle, extract
soil by adding 5 ml of methanol and shaking  vigorously for  two minutes.
Filter extract  and collect for storage or immediate EIA analysis.  Analyze
extract as  described in step 2  of the  EIA Procedure  above.   All
components  required  for this  method are commercially available  in kit
form.
      This study used two methods  of  sediment  sample preparation for
EIA analysis.  They are summarized  as  follows:   Method 1)  homogenize
wet sediment, weigh  4 g  onto solvent rinsed Al foil,  add   10-15  g of
anhydrous sodium  sulfate to form  a  friable mixture, extract  using 20 ml
isopropanol with 2 minutes of sonication and  15 minutes  of shaking, let
settle briefly  or centrifuge, then filter supernatant.  Some extracts were
prepared by variations of this  method,  all of which are similar  in speed,
simplicity, and ease of field use.  Analyze extract as described  in step 2
of the EIA Procedure above.  Method 2)  homogenize, Soxhlet extract (48
hours)  using  1:1 acetone:hexane, exchange to cyclohexane  (for GC-ECD
                                     420

-------
or storage),  exchange to isopropanol  for  EIA.   Analyze extract  as
described  in step 2  or the EIA Procedure above.
      Surface wipe sample preparation is summarized as follows:  Mark
surface  with  template,  remove wipe from methanol in extraction bottle,
wipe area to be analyzed, place wipe back in extraction jar and extract
by vigorous shaking for  1  minute.  Analyze extract as described in step 2
of the EIA Procedure above.  All components required for this method are
commercially  available  in kit form.
      Water  sample ^reparation is summarized  as follows:  condition
glass  barrel  SPE  column  with isopropanol  and reagent water,  extract
sample  by   drawing  through column,   air  dry  column, elute  with
isopropanol.  Analyze eluate  as described in step  2  of the EIA  Procedure
above.   All  components  required  for this  method  are  commercially
available.

RESULTS AND DISCUSSION

Test Specificity

The test response to Aroclors 1016,  1242,  1254,  and  1260 is within
twofold  of me response for Aroclor  1248.  The  broad specificity of the
EIA for the  common  moderately  chlorinated Aroclors  allows either
screening  or  Quantitative  analysis.    This specificity  is  relatively
independent of trie slight differences  in protocol among the  sample types
described  below.   Biphenyl and  several  chlorinated ring compounds
were  also tested for crossreactivity  in the EIA.  All  of these compounds
demonstrated less than 0.5% crossreactivity compared to Aroclor 1248:
3,3'-dichlorobenzidine,  1,2-dichlorobenzene, 1,3-dichlorobenzene,  1,4-
dichlorobenzene,  1,2,4-trichlorobenzene,  biphenyl, 2,4-dichlorophenol,
2,5-dichlorophenol,  2,4,5-trichlorophenol,  2,4,6-trichlorophenol,  and
pentachloropnenol.  These results mean that more than 200 ppm of any of
these  compounds would be required  to give the same  test response as  1
ppm of Aroclor  1248.  Also, more fhan 10,000 ppm  of  any of these
compounds would be required to give the  same test  response as 50  ppm
of Aroclor 1248.

Soil Screening

Soil screening results obtained using the kit with calibrators at 5, 10, and
50 ppm are  shown in  Table  1.   me respective actual concentrations of
these  calibrator solutions are 3,  5, and  22 ppm Aroclor  1248.  These
values were  chosen to provide  99% confidence that positive samples
will be detected regardless of the Aroclor present,  variations in  extraction
efficiency, or method imprecision for either GC or EIA.  This calibration
strategy biases the test slightly toward false positive results,  providing
greater  certainty  that  contaminated  samples  will  be  screened  out
effectively.    For  a  population  of samples linearly  distributed  in
                                     421

-------
concentration  from 1  to 100 ppm,  the  calculated false  positive  rate is
less than 6%, while the calculated false negative rate is  less than 0.1%
for  all three  decision points.   Several samples in Table  1  illustrate this
point.  For example, sample  3  is identified as > 50 ppm,  but is less than
50  ppm  by GC (29.6 ppm).  This result is expected because the actual
concentration  of the sample is  higher than the actual 22  ppm of  the 50
ppm calibrator.  Table  1  shows  that most  of  the false  positive results
occur in  the concentration range between the middle and  low calibrators,
due precisely to the bias designed  into  the  test.  Prudent site screening
demands that samples  in  this range  be  identified  for follow-up  GC
analysis  to ensure accurate  decision making.  The data of Table 1 show
that the slight inherent bias  of  the test accomplishes this effectively.  This
method has been reviewed by the EPA Office of Solid Waste and  will be
proposed for inclusion  in the  4000 series of screening  methods in the
next SW846 update.   Another version of this kit has been developed for
screening  at  I  and  5  ppm in  soil.   Test protocol  and  performance
characteristics are similar to the first  kit.

Quantitative Soil Analysis

If the  Aroclor present in the samples  is known, quantitative data  can
easily  be obtained  over a wide range of concentrations.   This is done by
substituting  for the calibrators a  standard curve of the known Aroclor.
Concentrations  of the  Aroclor  in the samples are  determined by
comparison to the standard curve.  Expansion of the range of quantitation
is possible by diluting high concentration extracts and by using a more
sensitive EIA protocol  for levels near or  below 1 ppm.  Table  2 illustrates
results of the latter procedure for a set of soil samples  containing  less
than 5 ppm,  compared to the GC results from  8080 analysis.  Because
Aroclor  1248  is recognized   slightly  better  than Aroclor  1260,
standardization on Aroclor  1260  will  yield slightly  higher values (as
demonstrated  in Table 2).  Correction for this difference in crossreactivity
between Aroclors  1248 and  1260 would  approximately  equalize the
slopes of the two regression lines in  Table 2.  Other projects using this kit
have  obtained quantitative  data  from field  samples covering a range 5
decades of concentration, from 0.1  to more  than  10,000 ppm.  Method
precision was tested in a 9  day validation study at 3 sites  using  4 field
samples  over  a concentration range of 5 to  35 ppm, resulting in  a 27%
concentration coefficient of  variation within  site and  across  day and kit
lot.

Quantitative Sediment Analysis

The PCB EIA kit  was  used for  quantitative sediment analysis  in  two
different situations. A comparison of EIA and GC-ECD was made using
sediment samples from a typical Great Lakes  industrial  contamination site,
prepared for  EIA by metnod 2.  For GC-ECD analysis, these samples
                                     422

-------
were  further cleaned  by  gel  permeation  chromatography and other
techniques,  including  sulfur removal.   Correlation between  the  methods
was good (Figure 2) over two decades of concentration,  from 0.2 to 20
ppm.   This result demonstrates  that the EIA  is able to  measure RGB's at
low levels in sediments  when  properly  presented to  the antibodies,
despite the  presence of sulfur compounds  and  heavy  metals  in these
sediment  extracts.   Sediments  which  have  undergone  significant
dechlorination  may  yield  inaccurate  results because  of decreased
recognition of less chlorinated congeners by  the antibodies; such  was not
the case here.    Precision  in   this  study  was good;  the  pooled
concentration coefficient of variation  (%Cv)  was  13%  for 7 pairs  of
duplicates,  2 sets of 6 replicates,  and  1  set of 7 replicates.
      Based on  the  above success of method  2/EIA  for quantitative
sediment  analysis,  method  1  was  developed  for simpler  sample
preparation  ana EIA analysis in the field.  Quantitative sediment  analysis
by  method 1/EIA was used at other Great Lakes industrial contamination
sites.   Sediment  cores were transported to  a lab trailer  on  shore for
sampling.   Same day analysis was performed using sample preparation
method 1/EIA.  Performance data for the EIA are summarized in tables 3
and 4.  Correlation of EIA results with GC-ECD  results was determined
for  a set of 41  samples containing 0.1  to  12  ppm  from  3  sites;  the
correlation coefficient was 0.84 and the regression equation was y =
0.63x + 0.47  ppm.  Timely on site analysis  of  total  PCB in  sediments
enabled  researchers  to  guide field sampling   to ensure  appropriate
samples were taken for GC-ECD analysis.  In addition, project  managers
were  able to conform confidently  to dredging guidelines and  to  guide
disposal  of  dredge spoil.   Further simplification  of sediment  extraction
method 1  is  currently under investigation.

Other  Matrices

Analyses of  PCB's on surfaces and in water are also possible using  the
same  EIA kit as for the previously described applications,  but with slightly
modified  protocols and  appropriate  sample   preparation  methods.
Decisions  about PCB contaminated surface remediation are often limited
by  the turnaround time for wipe sample analysis by GC.  The PCB EIA kit
has also  been  formatted  as a wipe test kit for  screening  use  in such
situations, analogous to the previously described soil screening  kit.  Data
from  64  field  samples  from  three  sites  were  used   to set actual
concentrations for the  10 and 100  u.g/100 cm2 calibrators.  The wipe kit
design includes the same false  positive bias as described above for the
soil screening kit, for exactly the same reasons.   Decisions are made in
the same  fashion as  soil  screening, as less  than  (or not less than)  the
decision  level.    For  a  population  of samples  linearly distributed  in
concentration from  1  to  100 u.g/100 cm2,  the calculated false positive
                                     423

-------
rate is 6 to 7%, while the calculated false negative rate is less  than 0.7%
for both 10 and 100 ^ig/100 cm2  decision points.
      PCB's  in water  adsorb strongly  to  even  small  amounts  of
particulate  material.   Therefore, analysis  of total PCB  in water requires
efficient recovery of this adsorbed  material.  This is accomplished by the
use of  high throughput CIS columns which efficiently capture  particulate
material.  These columns  allow the analysis of slightly  to moderately
turbid  field  water  samples  with  minimal  reduction  in  flow  rate.
Isopropanol is  used  to elute the PCB's  captured  by the  CIS from the
dissolved phase and simultaneously extract the adsorbed PCB's from the
trapped particles.  EIA analysis  is performed directly on the eluate without
volume  reduction.   Data on the recovery of  PCB  spiked into  water are
shown  in Table 5.  These results  demonstrate the ability of  the test to
detect 0.1  ppb total  PCB in a small (500 ml) sample in approximately
one hour using a simple sample concentration method.  Preliminary data
indicate a detection  limit below  25 ppt in 4000  ml  of water, and
improvement  may  be  possible with modest modifications of  the protocol.
The ultimate  sensitivity of this  method  will depend on several factors,
including sample volume, particle load, particle chemistry, etc.
CONCLUSIONS

1. Test specificity  is restricted to  PCB's, primarily Aroclors  1016, 1242,
   1248, 1254, and 1260.
2. The test is capable of analyzing for PCB's in soil in the field in  less
   than 30 minutes, using no specialized equipment.
3. Screening of soils containing unidentified Aroclors can be performed at
   50,  10,  5, or  1  ppm, with  99%  confidence of  detection  of
   contaminated samples.
4. Quantitation from 0.1  to  10,000  ppm  is possible for soil  samples
   containing an identified Aroclor.
5. Screening of surfaces contaminated with unidentified Aroclors can be
   performed  at  10 and  100|ig/100 cm2 with  99%  confidence of
   detection.
6. The design  of the EIA accommodates significant variations in sample
   type and protocol, allowing the analysis of PCB's in many matrices.
7. Ongoing work with this kit includes improved  sediment analysis, water
   analysis, oil analysis, and biological tissue analysis. Field testing  and
   validation are proceeding  for alf these applications.

ACKNOWLEDGEMENT

The initial phase of development  of this PCB immunoassay was  partially
supported by the US EPA through a sub-contract  to ECOCHEM from Mid-
Pacific Environmental  Laboratories,  Inc.
                                     424

-------
Table  1.  Screening of 32 soils using the EnviroGard™  PCB  EIA Kit.  Color was
read  by differential photometer and interpretation of EIA result was  based  on
comparison  of optical density value  to  calibrator  optical  density value.  The
correctness of  the  EIA interpretation  is  indicated  for  each calibrator level  as
follows:   normal =  correct; italicized = false positive, bold  = false negative.
Soil
Sample Tvoe*
2/
23
12
6
1
10
22
8
3
18
14
9
21
16
4
20
2
29
15
25
30
27
32
13
19
26
1 1
28
31
5
7
17
N/A
N/A
Loam/Clay
Sand
Clav
Clay
N/A
Clay
N/A
Sand
Clay
Loam/Sand
Clay
Clay
N/A
Clay
N/A
N/A
Clay
Loam
N/A
N/A
Loam
Clay
Clav
Clay
Loam/Clay
N/A
Clay
N/A
Loam/Clay
Clay
EnviroGard™ EIA Result
Comparison to Calibrators EnviroGard™
Aroclor <5 ppm <10 ppm <50 ppm EIA
Content Calibrator Calibrator Calibrator Interpretation
1254
1254/60
1260
1260
1260
1260
1242/54/60
1260
1254
1260
1260
1260
1242/54/60
1260
1254/60
1260
1254
1254/60
1260
1248/54
1254/60
1248/54
1248/54
1260
1260
1248/54
1260
1248/54
1254/60
1254
1260
1260
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
No
No
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
>50 ppm
<50 ppm**
>50 ppm
>50 ppm
>50 ppm
>50 ppm
<50 ppm
<50 ppm
>50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<50 ppm
<10 ppm
<50 ppm
<10 ppm
<1 0 ppm
<50 ppm
<50 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
<5 ppm
GC
Result
(ppm PCB
by EPA
8080)
1012
164
73
68
60
56
36.6
32
29.6
24
15
13
13
1 1
6.7
5
4.7
4.6
4
4
3.9
3
2.7
2
1.4
0.3
0.2
0.1
0.1
<0.1
<0.1
<0.1
  N/A = not available
  *  Further work  indicated  false negative response was  due  to  sample matrix  effect.
    Subsequent analysis indicated the  presence of an interfering substance that prevented
    PCB binding to the antibody.  This false negative at the  highest  calibrator level was not
    due to cross-reactivity or  incomplete PCB extraction.
                                        425

-------
Table 2.  Correlation of quantitative soil analysis by EIA and GC.
Sample
A
B
E
1
H
D
SSI
G
C
SS2
F
SS5
GC Result
oom bv 8080
0
0
0
0.2
0.3
0.6
0.6
0.6
1.0
1.1
1.4
3.1
Lab A
ppm by EIA
«0.15
«0.15
«0.15
0.3
1.0
0.4
0.7
1.0
1.2
2.3
1.6
5.0
Lab B
ppm bv EIA
«0.1
«0.1
«0.1
0.29
0.46
0.52
0.74
1.05
0.92
1.8
1.45
3.6
      Aroclor used for standard curve    1260
      Regression equation data (vs. GC)
                      r2
                      slope
      _ Y intercept
                                       0.93
                                       1 .56
                                       -0.03
                                                        1248
                        0.96
                        1.15
                        0.05
 Table 3.   Performance of sample preparation method  1  with  EIA for quantitative
 sediment analysis.
 Parameter
 Detection  Limit


 Precision
                Performance     Comment
                60 ppb         Determined by  repeat analysis  of  system blanks
                                (n = 7); 95% confidence

                                Concentration %CV, based  on duplicate analyses
                                of extracts of field sediment samples.   [%CV =
                                {(mean/standard deviation) xlOOJ]
                intraassay       7.7  %CV      n = 48 pairs of duplicates
                interassay       7.9  %CV      n = 52 pairs of duplicates

Spike recovery  90% of nominal  Based  on  detection of Aroclor 1248 spiked into
                                clean soil, n = 8
 Recovery
                73±9%
Recovery of total PCB from dried  SRM sediment,
relative to GC-ECD  result, n = 6
                                       426

-------
Table 4.  Comparison of sediment analysis by PCB EIA and GC-ECD.

                       Method 1/EIA    Method 2VGC-ECD
Cost per  Sample            $15              $150-300
Sample Throughput     10 per 2.5  hours      1  per 2.5 hours

* GC-ECD analysis requires Method 2 sample preparation procedure followed by gel
   permeation chromatography and sulfur removal.
Table 5.   Recovery of Aroclor  1248 Spikes from  Tap Water Samples.  This  water
contained  0.5  mg/liter particulate material,  determined gravimetrically  using a 0.45
Jim filter.   (* reagent water blank subtracted)

ng Spike       nn[           Final ng/ml       n         Recovery (mean  ± SD)
   50        500             0.1          3             73±29%
  250        500             0.5          3            1 14±50%
 1000        500             2.0          3             90±40%
  100      4000             0.025        1            125%*
FIGURE  LEGENDS
Figure 1. Principles of the Test.  Schematic depiction of the EIA for PCB's.
Figure 2. Correlation of  GC and EIA Results for Sediment Analysis.  Sediments were
analyzed by method 2 and EIA as described in  the methods section.
                                      427

-------
   ittiation 1:
   v =:  .. -.            %    %         ••
Dilution of sample or calibrator is incubated in
tube containing immobilized antibodies.
  = PCB
  = Interfering Material
  = Anti-PCB Antibody
 I**'.' •
 Sample matrix is washed away,
 leaving only RGB's bound to antibodies.
9&ubation2:
&
-------
  100
   10
EIA
ppm
         y = 1.23x - 0.12
         R = 0.97
         n = 33
               2
                     GC ppm
                                            100
                       429

-------
56
EXPLOSIVES ANALYSIS OF ATYPICAL MATRICES AND TECHNICAL ENHANCEMENTS TO
                          U.S.EPA METHOD 8330
                           Dennis Hooton. Senior Chemist
           M. Christopher son,  T.  Dux,  G.  Jungclaus,  H.  Randa, R.  Sauter
                            Midwest Research Institute
                                425  Volker Boulevard
                            Kansas City, Missouri 64110
     ABSTRACT

     The U.S. EPA Method  8330  from SW-846  is  used for  the  determination of
     nitroaromatics and nitramines (including HMX,  RDX,  TNT, etc.) in soil
     and water samples by high performance liquid chromatography  (HPLC).
     This method is being used for remedial investigation  programs at former
     ammunition facilities  throughout the  United States.   Variations of this
     method have been used  by  Midwest Research Institute (MRI) as the basis
     for analyzing more exotic matrices  not addressed  by the SW-846 method,
     including:  surface  wipe  samples, concrete chips, plant biota, and TCLP
     extracts.

     In order to improve  data  consistency,  technical enhancements have been
     incorporated into method  applications to improve  overall performance and
     quality.  These include:   "pre-mixing" soil for better homogenization
     and analytical subsampling,  performing multiple extractions  for trace-
     level analysis of water samples  using the "salting-out" technique  (U.S.
     Army Corps of Engineers draft procedure),  and enhancing chromatographic
     analysis with the addition of an internal standard  for better
     qualitative identification (relative  retention time)  and quantization
     (relative response factors).
      INTRODUCTION

      Since  1985, MRI has  performed explosives  analyses  for remedial
      investigation  of  sites  throughout the  United States using the EPA Method
      8330 "Nitroaromatics and Nitramines  by High  Performance Liquid
      Chromatography (HPLC)."   Although the  basic  method has remained the
      same,  MRI has  added  several  enhancements  to  the procedure which has
      improved overall  method performance.   In  addition  to analyzing soil and
      water  samples,  the method has been adapted for use with other matrices,
      including:  surface  wipes, concrete  solids,  TCLP extracts, and plant
      biota.  This paper presents  an overview of how this method has performed
      in  these different applications and  describes some techniques which have
      improved overall  performance.
                                         430

-------
MATRICES
Method 8330 is a very versatile method which can be adapted to a variety
of matrices.  However, special applications require special techniques
to achieve the most reliable results.   The following paragraphs describe
how this method has been applied to various matrices.
Soil and Sediment

Method 8330 specifies that soil and sediment samples be initially dried
in air at room temperature or colder and protected from direct sunlight.
Then the sample is ground in a solvent-rinsed (acetonitrile) mortar.
MRI has found that preparing a "pre-mix" of the soil before actually
taking analytical subsamples helps to assure sample homogeneity and
precision of the method.  The expanded procedure also uses the air-
drying technique, which is then followed by grinding the soil/sediment
using a solvent-rinsed laboratory burr mill to quickly reduce the
particle size of the soil.  Nearly all of the sample will pass through a
30-mesh screen.  This step is important for several reasons:  first, it
allows much larger samples or an entire sample to be easily prepared for
analysis, reducing the concern of inadequate composite mixing during
field sample collection, and second,  it provides a sample texture more
conducive to further homogenization.   The ground sample is collected in
a glass jar and rotary tumbled to further mix the sample prior to
analytical subsampling.

Table 1 shows a comparison of Method 8330 precision results for the
analysis of contaminated soils.   Data provided in the SW-846 methods'
"Intralaboratory Precision of Method for Field-Contaminated Soil
Samples," was based on a study where the method was tested by six
different laboratories.   These data are compared to field-contaminated
samples which were analyzed by one laboratory (MRI) three separate times
over an approximate one-month period.   Although these data are not
directly comparable because of the significant differences in analyte
concentrations, the data do indicate that comparable or better precision
can be achieved at trace contaminant levels using pre-mixed soil
samples.
Concrete Chips

The technique used for soil/sediment preparation has been successfully
used on hard clay soils and also very hard samples which would not
normally be amenable to standard grinding procedures (i.e.,  mortar and
pestle).  For example, this technique was used on concrete chip samples
collected from cinder block walls to determine if residual explosives
remained in the concrete.   The same technique, described above for the
soil samples, was used to reduce the concrete chips to a powdery texture
prior to analysis by Method 8330.  Although none of the Method 8330
analyte list were found in these samples, method recovery and precision
                                    431

-------
were determined using spiked concrete samples.  Table 2 shows precision
and accuracy for this matrix;  spiked soil samples, which were extracted
and analyzed concurrently with the concrete samples, are also shown  for
comparison.
     Table 1.   Comparison of SW-846 Interlaboratory Precision Values
             with "Pre-mixed" Results for Contaminated Soils

Analyte
Tetryl
RDX
TNT
TNB
SW-846 data
Concentration
(Mg/g)
2.3
104
877
7
669
3
72
Precision
(RSD)a
18
17
8
18
10
8
12
Pre-mixed sample data
Concentration
(Mg/g)
2.2
1.7
3.4
12.1
Precision
(RSD)a
8
7
9
1
 a Relative standard deviation.
         Table 2.   Matrix Spike Duplicate Recovery and Precision
                      for Concrete and Soil Samples

Matrix
Concrete
Sample No. 1
Concrete
Sample No. 2
Soil

Spike
level
(units)
2.5
Mg/g
2.5
Mg/g
2.5
Mg/g
Average recovery (range)
HMX
97
(±4)
99
(±4)
102
(±D
RDX
97
(±8)
102
(±4)
98
(±D
TNB
78
(±47)
106
(±4)
108
(±3)
Tetryl
37
(±1)
83
(±4)
94
(±10)
TNT
29
(±26)
95
(±3)
100
(±1)
24 DNT
69
(±27)
105
(±4)
108
(±D
Water

The current version of Method 8330 (November 1990),  is applicable to
relatively high-level determinations of explosives in water samples.
This method consists primarily of a "dilute-and-shoot" approach which is
effective and straight-forward for samples with analyte concentrations
ranging from approximately 4 fig/L (for DNB) to 44 /ig/L (for tetryl) .
Trace-level analysis of samples with analyte concentrations as low as
0.1 /zg/L require extraction and concentration of the resulting extract.
MRI uses a technique based on a U.S. Army Corps" of Engineer's (USAGE)
                                    432

-------
draft protocol in which 400 mL of a water sample is saturated with
sodium chloride salt, followed by liquid-liquid extraction of the
explosive compounds into acetonitrile.  This "salting-out" technique has
been successfully used for many years to analyze groundwater and surface
water samples from federal remediation sites.

MRI has modified and enhanced the original draft method in several ways
to improve overall performance or to reduce common reagent interferences
exhibited primarily at the trace levels.  These changes include per-
forming a second acetonitrile extraction of the salt-saturated aqueous
sample to recover any residual explosives not collected in the first
extraction and pre-cleaning the salt reagents (sodium chloride and
sodium sulfate) by solvent extraction and heat.   The results of these
changes are shown in the following table (Table 3) where recovery and
precision are shown for spiked control samples at trace level concen-
trations.  The method has proven to be reliable and reproducible over
hundreds of water samples.
               Table  3.   Summary  of Accuracy  and  Precision
                     Data for  Spiked Water  Samples8-13
Analyte
HMX
RDX
2,4-DNT
TNT
TNB
Tetryl
Average
recovery
(%)
96
83
53
78
85
76
Precision0
12
20
11
8
8
4
No. of
determinations
5
5
5
5
5
5
        Method performance was measured by spiking clean water
        samples with separate extraction batches.
        Spike levels for the target analytes in water were
        1.25 Mg/L for all analytes except tetryl at 5
        s = Standard deviation of the average.
TCLP Extracts

The "salting-out" extraction technique has also been successfully used
for TCLP extracts in evaluating the leachability of explosive compounds
from soil.   In one study, Method 8330 was used to extract highly
contaminated soil samples.  The same soil samples underwent the TCLP
extraction procedure and a portion of the TCLP extract was analyzed
using the "salting-out" procedure.
                                    433

-------
These results show that the method adequately' characterized both the
original soil and the resultant leachate for explosives contaminants.
The extract concentrations were consistent with the original soil
contamination, producing 25% to 100% of the total possible compound that
would be expected based on the original soil analysis (assuming a factor
of 50 to convert /zg/g soil concentrations to /zg/L leachate
concentrations).  These results are shown in Table 4 below.
          Table 4.  Analysis Results for RCRA Characterization
                         Soils  and TCLP Extracts

Matrix
TCLP prep, blank
Soil
TCLP ext.
Soil
TCLP ext.
Soil
TCLP ext.
Soil
.TCLP ext.
Soil
TCLP ext.
Soil
TCLP ext.
QUANTITATION LIMITS:
Matrix:
Soil
TCLP extract

Units
/ig/L
fg/8
PS/L
/ig/8
pg/L
/*g/g
MS/L
Mg/S
M8/L
/*g/8
/*8/L
/*8/S
PS/L
Units
**g/8
MS/L
Analyte concentration
HMX
< QL
0.53
< QL
0.33
19
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
RDX
< QL
3.0
180
3.3
150
0.34
4.8
0.97
25
0.91
20
0.19
< QL
TUB
< QL
41
170
4.9
21
4.3
57
43
1200
28
860
< QL
< QL
Tetryl
< QL
< QL
140
< QL
< QL
< QL
< QL
< QL
76
< QL
65
< QL
< QL
TNT
< QL
2600
70000
180
3700
89
3000
49
1500
31
990
22
630
2,6-DNT
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
< QL
2,4-DNT
< QL
8.3
130
2.5
40
3.0
74
0.96
25
0.90
22
< QL
< QL
Concentration
0.15
1.1
0.15
1.1
0.50
3.3
1.0
6.7
0.15
1.1
0.50
3.3
0.15
1.1
Surface Wipes

Surface wipes were analyzed using an adaptation of the soil/sediment
procedure in which surface areas were wiped with acetonitrile-saturated
gauze.  The wipe samples were simply treated as a solid sample by adding
extracting solvent to the collection jar and processing the sample the
same as the Method 8330 procedure for soils.  Since no explosives were
found in the field samples, a spiked control was used to demonstrate
recovery as shown in Table 5.
                                   434

-------
              Table 5.  Control Spike Recovery For Surface
                           Wipe (Gauze) Sample
Analyte
HMX
RDX
TNB
Tetryl
TNT
2,4-DNT
Percent recovery
(*)
75
77
80
87
80
75
             Spiked concentrations were 5 /ig per gauze sample
             (representing a 100 cm2 surface area).
Plant Biota
Biota samples present a more unusual sample matrix in that the explosive
compounds may be chemically bound to the plant structure.  As expected,
the extraction technique is quite different than would normally be used
in the absence of biological transformation of these contaminants.  For
example, one technique used for plant studies is to acid-hydrolize the
plant sample, then extract the mixture with ether to recover the
explosive compounds, followed by a solvent exchange and a column
chromatography step.  Method 8330 analysis parameters were used to
measure explosive residues in the plant material.

Method performance for plants is presented in Table 6 below.   Although
chemical recoveries are lower than other matrices,  these are reflective
of the more compelx extraction procedure (which is  not part of
Method 8330) rather than the Method 8330 analysis procedure used to
analyze the extracts.   This is illustrated by the fact that similar low
recoveries were found for the control samples (spiked reagent taken
through the extraction procedure) reported with the biota data.   It
should also be noted that the plant extracts did not produce  any unusual
or problematic background interference for the Method 8330
chromatography.
                                   435

-------
           Table 6.  Blank, Control, and Matrix Spike Results
                   (Percent Recovery)3  for Plant Biota
Biota controls
Method blank
Control 1
Control 2
Matrix spike
terrestrial
Plant No. 1
Matrix spike
terrestrial
Plant No. 2
Root plant No . 3
Root plant No . 4
RDX
< QL
52%
60%
57%
45%
49%
54%
TNB
< QL
51%
59%
Interference13
21%
37%
43%
Tetryl
< QL
78%
79%
~ 37%°
~ 53%°
63%
42%
TNT
< QL
43%
45%
~ 35%d
39%
28%
29%
2,4-DNT
< QL
40%
44%
37%
43%
35%
38%
 a Spike levels in control and matrix spikes were 20 /tg/g f°r
   indicated analytes.
 b TNB recovery indeterminate due to chromatographic interference.
 0 Tetryl recoveries are estimated because the extract concentrations
   were near the quantitation limit.
 d Estimated recovery for TNT due to high native level.
METHOD ENHANCEMENTS

The most significant technical improvement MRI has made to the SW-846
method has been the use of an internal standard (a chemically-similar
compound which is spiked into sample extracts and standards just prior
to analysis).  This addition serves several important functions.  First
and foremost, it is used to establish relative retention time windows
for qualitative identification for the Method 8330 analytes; it is
relative in the sense that chromatographic retention times are
normalized to the internal standard marker within each chromatogram.
This technique provides a much more precise determination of close
eluting compounds.   For example, because the retention times for the
2,4-DNT and 2,6-DNT isomers vary by only a few seconds, normal time
variances between one injection to the next can make it difficult to
absolutely distinguish between the two compounds.  However, when an
internal standard marker is added to set relative retention time windows
for these two analytes, there is an easily discernable difference.  The
table below (Table 7) presents a summary of how precision is
dramatically improved by the use of relative retention time over normal
Method 8330 retention time identification criteria.  Relative percent
differences decrease for all analytes when using RRT values which
reduces the chance of reporting false positive results.
                                    436

-------
           Table 7.   Comparison of Absolute Retention Time vs.
               Relative Retention Time (Internal Standard)
                         for Explosive Standards

HMX
RDX
TNB
DNB
NB
TNT
2,4-DNT
Tetryl
2,6-DNT
2 -NT
4 -NT
3 -NT
Average RTa
2.8
4.1
5.4
6.5
7.6
8.8
10.5
7.3
10.2
12.8
13.8
14.8
RPDC
1.4
1.7
1.3
2.0
1.6
1.6
1.9
2.1
2.8
2.8
3.1
2.9
Average RRTb
0.137
0.199
0.264
0.317
0.370
0.428
0.509
0.359
0.495
0.618
0.667
0.715
RPDC
0.73
0.50
0.38
0.95
0.27
0.42
0.39
0.56
0.61
0.32
0.45
0.28
 a Average Retention Time (RT) in minutes.
 b Average Relative Retention Time (RRT) - analyte RT over the internal
   standard RT.
 c Precision expressed as Range Percent Difference over 20 injections
   and 7 different concentrations.
The internal standard also provides the advantage of improved
quantitation by establishing the linear relationship between an analyte
concentration and its corresponding relative response,  i.e., analyte
response vs. internal standard response.   This technique corrects for
minor fluctuations in injection volume and sample preparation volume
after the internal standard is added.   And finally,  the internal stan-
dard can be used to monitor injection precision during automatic (auto-
injector) analysis runs.
SUMMARY

Method 8330 can be adequately adapted to a wide variety of matrices,
including surface wipes, solid materials,  plants,  and TCLP extracts.
Method performance, especially precision,  can be improved through more
rigorous preparation of soil/sediment samples prior to extraction and by
double-extracting aqueous samples.   Addition of an internal standard to
sample extracts just prior to analysis provides more accurate qualita-
tive identification of target analytes and improves quantitative analy-
sis.  Method 8330 has proven itself to be  a rugged and reliable test for
determining trace-level explosives  in environmental investigations.
                                    437

-------
LIQUID CHROMATOGRAPHY/MASS SPECTROMETRY OF HIGH-MOLECULAR-WEIGHT
POLYCYCLIC AROMATIC HYDROCARBONS USING THE PARTICLE BEAM INTERFACE

L. Don Betowski, U.S. EPA, EMSL-LV, Las Vegas, NV, 89193
Chris Pace, Lockheed Environ. Sys. & Tech. Co., Las Vegas, NV, 89119
Mark Robv. SAIC, San Diego, CA, 92121

Most studies of polycyclic aromatic hydrocarbons  (PAH) in
environmental samples have focused on PAH containing five fused rings
or less  (MW < 300).  Because of analytical difficulties, including  low
vapor pressure, lack of analytical reference standards, lower sample
concentration, and the large number of isomeric possibilities, the
higher molecular weight PAH  (MW > 300) are seldom examined.  There  are
no standardized methods for the measurement of high-molecular-weight
PAH.  Despite analytical difficulties, the presence of high-molecular-
weight PAH has been demonstrated in many environmental samples
including carbon black, coal tar, petroleum, diesel and other air
particulates, and  in soils.  Although the number of toxicity studies
on high-molecular weight PAH are limited, several dibenzopyrenes  (MW
302) are known to be highly carcinogenic.

In this work, particle beam liquid chromatography/mass spectrometry
(PB LC/MS) was investigated as a means to measure high-molecular-
weight PAH.  Instrument performance was evaluated with 17 PAH covering
the molecular weight range 300 to 450 amu.  The PAH were separated  by
conventional high performance liquid chromatography  (HPLC) using a
polymeric octadecylsilica  (C-18) packing  (Vydac 201-TP) and gradient
elution with methanol/tetrahydrofuran  (THF).  Instrument detection
limits as measured by selected ion monitoring  (SIM) on the singly
charged molecular  ion of each PAH were found to be 0.15 to 0.6 ng on
column for PAH up  to 352 amu and 2 to 4 ng on column for PAH greater
than 352 amu.  Linear response was observed for those PAH with
molecular weight 300 to 352 amu over the concentration range 0.05 to
5.0 ug/mL. Non-linear response was observed for PAH with molecular
weight greater than  352 amu over the concentration range 0.25 to 25
ug/ml.   The PB electron ionization  (El) mass spectra of the PAH were
found to be variable with  the ion distribution ratio of the singly
charged  molecular  to the doubly charged molecular ion being dependent
on molecular weight, ion source temperature, and  on concentration.
Particle beam negative chemical ionization  (NCI)  of the PAH was also
examined.  Substantial signal enhancement  (10 to  600 fold) was
observed but found to be highly dependent on the  amount of THF  in the
mobile phase.

Notice:  Although  the research described  in this  abstract has been
funded wholly by the U.S.  EPA through contract #68-CO-0049 to
Lockheed,  it has not been  subjected to agency  review.  Therefore,  it
does not necessarily reflect the views of the  agency.
                              438

-------
58      Quantitation of Pentachlorophenol (PCP) by a Rapid Magnetic
              Particle-Based Solid-Phase ELISA in Water and SoiL


      Fernando M.  Rubio.  Timothy S.  Lawruk, Scott W.  Jourdan, Charles S.
      Hottenstein,  Van  Zha,  and  David  P.  Herzog,  Ohmicron  Corporation,
      Newtown, Pennsylvania 18940.


      ABSTRACT

      Traditional  testing   methods  for  pentachlorophenol   (PCP)  are  time
      consuming,   expensive   and    require    specialized   instrumentation.
      Immunochemical assays provide methods that are  sensitive, rapid, reliable,
      which are suitable for  lab and field  analysis.   A magnetic  particle-based
      ELISA for the  determination  of  PCP  in  water  and soil samples has  been
      developed.   Paramagnetic  particles, used  as the  solid  support, allow for
      precise antibody addition and superior assay kinetics.  The assay procedure
      and detailed  performance characteristics,  including correlation with GC/MS
      and HPLC, are discussed.  The method has a test range of 0.06 to 10 ppb
      in water, and typical within assay %CV of less than 10%.  Recovery studies
      averaged 105%.   The application of this  method  permits  cost-effective
      evaluation of samples without solvent disposal and can result in savings of
      time and money.

      INTRODUCTION

      Pentachlorophenol (PCP) is used worldwide in commercial wood treatment,
      paper  production, the leather industry  (as a preservative and fungicide), and
      agriculture (as an insecticide and  herbicide).  Its extensive  use has made
      PCP an  ubiquitous environmental  pollutant; the chemical has been  found in
      water, soil,  food,  and  air,  often in  high  concentrations.   Like  many
      organochlorine compounds, PCP has toxic effects  in humans.  It is rapidly
      absorbed through the skin, respiratory and gastro-intestinal  tract,  inducing
      skin rashes,  respiratory  diseases, increased hepatic  enzyme activity  and
      renal failure.

      Pentachlorophenol is  one of the compounds regulated by USEPA under the
      NPDWR with  a Maximum  Contaminant Level of 1  ppb in drinking water.
      Under   the  RCRA   Program  (Toxicity  Characteristic),  the  maximum
      concentration of PCP  allowable in solid waste is 100 ppm.

      The principles of enzyme  linked  immunosorbent assays  (ELISA) have  been
      described (Hammock  and Mumma, 1980).  Magnetic  particle-based ELISA's
      have previously been described  and applied to the  detection of pesticide
      residues (Itak et al, 1993;  Lawruk et al,  1993; Itak et al,  1992; Lawruk et
      al,   1992; Rubio et  al, 1991).   These  ELISAs eliminate the  imprecision
      problems that  may be associated with antibody coated plates and tubes
      (Harrison et  al,  1989;  Engvall,  1980)  through the covalent coupling of
      antibody to the magnetic  particle solid-phase.   The uniform dispersion of
      particles throughout the reaction  mixture allows for rapid reaction kinetics
      and precise addition of antibody.  The PCP magnetic-based ELISA described
                                        439

-------
in this paper combines  an  antibody specific for PCP with enzyme labeled
PCP.  The presence of PCP in a sample is visualized through a colorimetric
enzymatic reaction  and  results  are  obtained  by comparing  the color in
sample tubes to those of calibrators.

MATERIALS AND METHODS

Amine terminated superparamagnetic  particles of  approximately  1  um
diameter were obtained from Advanced Magnetics,  Inc. (Cambridge, MA).
Glutaraldehyde  (Sigma Chemical, St. Louis, MO).  Rabbit anti-PCP serum
and PCP-HRP conjugate  (Ohmicron, Newtown, PA). Hydrogen peroxide and
TMB  (Kirkegaard &  Perry Labs,  Gaithersburg, MD).   PCP  and related
compounds, as  well as  non-related cross-reactants  (Chem Service,  West
Chester, PA).

The anti-PCP coupled magnetic  particles were  prepared by glutaraldehyde
activation (Weston and Avrameas, 1971). The  unbound glutaraldehyde was
removed from the particles by magnetic separation and washing four times
with  2-(N-morpholino)  ethane  sulfonic  acid   (MES)  buffer.    The  PCP
antiserum and  the activated particles  were incubated overnight at  room
temperature with agitation.   The unreacted glutaraldehyde was  quenched
with  glycine buffer  and the  covalently coupled anti-PCP particles  were
washed and diluted with a Tris-saline/BSA preserved buffer.

Water samples (200 uL) and horseradish peroxidase (HRP) labeled PCP (250
uL) are incubated for 30 minutes with the antibody coupled solid-phase (500
uL) (step 1).  A magnetic field is applied  to  the magnetic solid-phase  to
facilate  washing and  removal  of unbound PCP-HRP  and  eliminate any
potential interfering substances (step 2).  The  enzyme substrate  (hydrogen
peroxide) and chromogen (3,3',5,5'-tetramethyl benzidine  [TMB]) are then
added and incubated for 20 minutes (step 3).  The reaction is stopped with
the addition  of acid and the final  colored product is analyzed using the RPA-I
RaPID Analyzer™ by determining the absorbance at 450 nm.  The observed
absorbance  results were compared to  a  linear  regression line using a log-
logit standard curve  prepared  from calibrators  containing 0,  0.1, 2.0, and
10.0  ppb of PCP.  If the assay is performed in  the field (on-site), a battery
powered photometer such as the RPA-III™ is used.

When analyzing soil  samples,  a simple extraction  is  performed prior  to
analysis: 10 g  of soil is shaken for 30 minutes (extractions as short as 1
minute could be performed)  with 20  mL  of  a solution  of  0.5% sodium
hydroxide/75%  methanol/25%  water  (w/v/v).   After settling, the sample
supernatant  is diluted at least 1:500 in pentachlorophenol zero standard and
assayed as in the case of water samples.

RESULTS AND DISCUSSION

Figure 1 illustrates the mean standard curve for the PCP calibrators collected
over  79 runs,  error  bars  represent one  standard  deviation  (SD).   The
displacement at the 0.1  ppb level is significant  (86.9% B/Bo, where B/Bo is
the absorbance  at 450 nm observed for a sample or standard divided by the
                                  440

-------
absorbance at the zero standard).  The assay sensitivity based on 90% B/Bo
(Midgley et al, 1969} is 0.06 ppb.

A precision study in which surface and groundwater samples were fortified
with PCP at 4 concentrations, and assayed 5 times in singlicate  per assay,
on five different days is shown in Table 1.  Coefficients of variation (%CV)
within and  between  day (Bookbinder and  Panosian, 1986) were less than
13% and 12% respectively.

Correlation  of twenty groundwater samples using values  obtained  by the
ELISA method (y) and GC/MS EPA Method 625 (x) method is illustrated in
Figure 2. The regression analysis yields a  correlation of 0.980 and a slope
of 1.08 between methods.

Table 2 summarizes the accuracy of the PCP ELISA. Added amounts of PCP
(0.50, 1.50, 3.0, an  8.0 ppb) were recovered correctly in all cases with an
average assay recovery of 105%.

Table 3 summarizes the cross-reactivity data using a variety of chlorophenol
compounds.  The percent cross-reactivity was determined as  the  amount of
analogue required  to  achieve  90% B/Bo (Least  Detectable  Dose).   Many
non-structurally related  agricultural compounds demonstrated no reactivity
at concentrations up to 10,000 ppb (data not shown).

Table 4 indicates  that  no  interferences  are  present from compounds
commonly found  in  groundwater samples at concentrations much higher
than  usually  found in  those  waters  (American  Public  Health Association,
1989).

Recovery data  from two  different soil types  (Piano  loam and  Sassafras
sandy loam)  fortified with pentachlorophenol  at 5,  10,  and 50  ppm is
summarized in Table  5.  To compare extraction efficiencies, soils were also
extracted with methanol using a procedure similar to the procedure listed in
materials and methods.  Average recoveries of PCP for the spiked  soils were
100% with NaOH/MeOH/H20 and 78%  with methanol.

To compare immunoassays results with HPLC, the same extracts prepared
in the above  soil recovery procedure were acidified with HCL and filtered
thru 0.7 um glassfiber microfilters. Samples were analyzed by HPLC at the
Air Pollution Lab of New Jersey Institute of Technology.  Figure 3  shows
good  agreement  (correlation  =  0.977,  slope   = 0.975)   between  the
immunoassay and  HPLC methods for the twelve  (12)  spiked soil samples
listed  in Table 5.

SUMMARY

This work describes a particle-based ELISA for the detection of PCP and its
performance  characteristics  using water and soil  samples.   The assay
compares favorably to GC/MS and HPLC determinations, and eliminates the
need  for expensive  instrumentation and  solvent disposal.   The  ELISA
exhibits  good  precision  and  accuracy which   can   provide   consistent
monitoring of environmental samples.   Using  this ELISA, fifty (50)  results
                                  441

-------
can be obtained in less than one  hour without the problems  of variability
encountered with antibody coated  tubes and microtiter plates  (e.g. coating
variability, antibody leaching, etc.).  This system is ideally suited for the
adaptation  to  on-site monitoring of PCP  in water,  soil,  and solid waste
samples.

REFERENCES

American Public Health  Association, Standard Methods for Examination of
Water and Wastewater. American Public  Health Association, Washington
DC, 1989.

Bookbinder, M.J.; Panosian, K.J. Correct and Incorrect Estimation of Within-
day and Between-day Variation. Clin.Chem. 1986,  32, 1734-1737.

Engvall,  B.  Enzyme Immunoassay  ELISA  and  EMIT.  In  Methods  in
Enzymology: Van Vunakis, H.; Langone, J.J.,  Eds.; Academic Press:  New
York, 1980; pp 419-439.

Hammock, B.D.; Mumma,  R.O. Potential of Immunochemical Technology for
Pesticide Analysis. In Pesticide Identification at the Residue Level;  Gould,
R.F.,  Ed.; ACS Symposium Series, Vol. 136; American Chemical Society:
Washington, DC, 1980; pp. 321-352.

Harrison, R.O.;  Braun,  A.L.;  Gee, S.J.;  O'Brien,  DJ.;  Hammock,  B.D.
Evaluation  of  an Enzyme-Linked  Immunosorbent Assay (ELISA)  for the
Direct Analysis  of   Molinate   (Odram®)  in  Rice Field  Water. Food &
Agricultural Immunology 1989, 1, 37-51.

Itak,  J.A.;  Olson, E.G.;  Flecker, J.R.;  Herzog,  D.P.  Validation  of a
Paramagentic  Particle-Based ELISA for the  Quantitative Determination  of
Carbaryl  in Water. Bulletin of Environmental Contamination and Toxicology
1993, 51,   in press.

Itak, J.A.;  Selisker, M.Y.;  Herzog,  D.P. Development and Evaluation  of a
Magnetic Particle Based Enzyme Immunoassay for Aldicarb, Aldicarb Sulfone
and Aldicarb Sulfoxide. Chemosphere 1992, 24, 11-21.

Lawruk, T.S.;  Lachman, C.E.;  Jourdan, S.W.; Flecker, J.R.; Herzog, D.P.;
Rubio, F.M. Quantification of Cyanazine in Water and Soil by a Magnetic
Particle-Based  ELISA. J. Agric. Food Chem. 1993, in press.

Lawruk, T.S.; Hottenstein, C.S.; Herzog, D.P.; Rubio, F.M. Quantification of
Alachlor  in Water by a Novel  Magnetic Particle-Based ELISA. Bulletin  of
Environmental Contamination and Toxicology  1992, 48,  643-650.

Midgley,  A.R.; Niswender, G.D.;  Rebar, R.W. Principles for the Assessment
of  Reliability   of  Radioimmunoassay  Methods   (Precision,  Accuracy,
Sensitivity, Specificity). Acta Endocrinologica. 1969, 63 163-179.

Rubio, F.M.; Itak, J.A.;  Scutellaro, A.M.;  Selisker, M.Y.;  Herzog,  D.P.;
Performance Characteristics of a  Novel Magnetic Particle-Based Enzyme-
                                  442

-------
Linked Immunosorbent Assay for the Quantitative Analysis of Atrazine and
Related Triazines in Water Samples. Food & Agricultural Immunology 1991,
3, 113-125.

Weston, PD, Avrameas, S.; Proteins coupled to polyacrylamide beads using
glutaraldehyde.  Biochem.  Biophys.  Res. Commun.  1971, 45, 1574-1580.
                                 443

-------
                        Table 1
Pool Number     Pool 1    Pool 2

Replicates          5         5
Days              5         5
N                 25       25
Mean(ppb)        0.51      1.67
% CV (within)     12.5       8.8
% CV (between)   11.4       8.6
                    Pool 3    Pool 4
                      5
                      5
                     25
                    3.16
                     7.7
                     1.8
              5
              5
             25
            8.63
             6.7
             3.2
                        Table 2
   PCP added
   (DDb)

   0.50
   1.50
   3.00
   8.00

   Average
PCP observed
    (ppb)

    0.49
    1.63
    3.34
    8.43
 SD
(ppb)

0.09
0.17
0.29
0.71
Recovery

   98
  108
  111
  105

  105
                         444

-------
                                       Table 3
Compound

Pentachlorophenol
2,3,5,6-Tetrachlorophenol
2,3,3,6-Tetrachlorophenol
2,3,6-Trichloropheno!
2,3,5-Trichlorophenol
Tetrachloro-hyoroquinone
2,4,6-Trichlorophenol
2,4,5-Trichlorophenol
2,3,4-Trichlorophenol
2,5-Dichlorophenol
2,6-Djchlorophenol
2,3-Dichlorophenol
2,4-Djchlorophenol
3,5 -Dichlorophenol
Hexachlorobenzene
Hexachlorocyclo-hexane
3,4-Dichlorobenzene
4-Chlorophenol
Pentachlorobenzene
Pentachloronitro-benzene
90% B/Bo
LDD (ppb)
0.06
0.21
0.91
2.44
1.52
8.7
15.1
21.5
53.2
62.9
266
611
887
1670
1560
5790
NR
NR
NR
NR
50% B/Bo
(ppb)
2.20
4.06
14.6
62.9
119
148
463
574
1730
7830
5990
> 10000
> 10000
> 10000
> 10000
> 10000
NR
NR
NR
NR
% Cross
Reactivity
100
54.1
15.1
3.4
1.8
1.4
0.5
0.4
0.1
0.03
0.04
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02

-------
                       Table 4
Compound      Max. Cone. Tested          Interference

Nitrate              250 ppm                  No
Copper             250 ppm                  No
Nickel              250 ppm                  No
Thiosulfate          250 ppm                  No
Sulfite              250 ppm                  No
Sulfate            10,000 ppm                No
Iron                250 ppm            No up to 50 ppm
Magnesium          250 ppm                  No
Calcium             500 ppm                  No
NaCI                1.0 M             No up to 0.65 M
Humic acid          250 ppm            No up to 10 ppm
Silicates            2500 ppm          No up to 1000 ppm
Zinc               250 ppm                  No
Mercury (+ 2)        250 ppm                  No
Phosphate          250 ppm                  No
Manganese          250 ppm                  No
                          446

-------
                                   Table 5
Extractant


Methanol




Methanol
 Soil Type


Piano loam
 Sassafras
sandy loam
PCP Spiked
  (ppm)
    0
    5
    10
    50

    0
    5
    10
    50
PCP Observed
   (ppm)
     nd
    2.6
    7.7
    40.1

     nd
    5.0
    8.6
    36.8

  Average
                                                                Recovery
                                            52
                                            77
                                            80
                                                                  100
                                                                  86
                                                                  74

                                                                  78
NaOH/MeOH/H20'
NaOH/MeOH/H20
Piano loam
 Sassafras
sandy loam
    0
    5
    10
    50

    0
    5
    10
    50
     nd
     5.5
     8.8
    47.2

     nd
     5.5
    10.5
    47.6

  Average
                                                                  110
                                                                   88
                                                                   94
                                                                  110
                                                                  105
                                                                   95

                                                                  100
nd = below LDD of assay
1 Methanol =100% methanol
2 NaOH/MeOH/H2O =0.5% NaOH/75% MeOH/25% H2O (w/v/v)

-------
   o—<

   «—*
    A
    D
               Magnetic Particle Immunoassay
Magnetic Particle with
   Antibody Attached
Pesticide Conjugated
   with Enzyme
Pesticide
Chromogen/Substrate
Colored Product
                                         1. Immunological Reaction
00
              2. Separation
                                 3. Color Development

-------
   90.0
o
CD
   80.0




   70.0



   60.0



m  50.0



   40.0



   30.0




   20.0
           0.1
                                            10
                  PCP Concentration  (ppb)
Rgure 1.  Dose response curve for PCP.  Each point represents the
mean of 79 determinations. Vertical bars indicate +/- 1 SD about
the mean.
                           449

-------
_Q
Q.
Q_
D
W
9
Q.
o
on
   140
   120 -
   100 -
    20 -
     0
             20    40     60    80    100    120

                GC/MS Method 625 (ppb)
Figure 2.  Correlation between PCP concentrations as determined
by ELISA and GC/MS method 625, in water samples, n = 20, r
=  0.980, y = 1.08 + 2.30
                         450

-------
     50
     40
   E
   D
   CO
   CO
  9  20
  Q_
  O
  Q:
      10
      0
        0
10
20      30


HPLC (ppm)
40
50
Figure 3. Correlation between PCP concentrations as determined
by ELISA and HPLC in soil samples,  n = 12, r = 0.977, y =
0.98 +  1.39.
                          451

-------
59  CHROMATOGRAPHIC OPTIMIZATION FOR THE ANALYSIS OF
     AN EXPANDED LIST OF VOLATILE ORGANIC POLLUTANTS
     DAVID M. SHELQW AND MICHAEL J. FEENEY
     Restek Corporation 110 Benner Circle, Bellefonte, PA 16823

     Volatile organic analysis (VOA) is one of the most common environmental test
     procedures for analysis of pollutants in air, water, and spil. Current methodology specified
     in EPA methods involves concentrating the trace volatile organics onto an adsorbent
     material and then thermally desorbing them onto a chromatographic column. The list of
     target compounds to be analyzed  are similar for air, water and soil matrices and there are
     indications that additional compounds will be added to the target compound lists in the
     future. When the target lists of EPA methods TO-14, 524.2, 624, 8240 and 8260 are
     combined, including various surrogates and internal standards, the list contains over 100
     compounds. Resolution of all of these compounds in a single chromatographic analysis can
     be difficult if not impossible, however, through proper column selection and
     chromatographic optimization most compounds can be resolved.

     This paper will describe a new technique for GC method development using
     thermodynamic retention indices and computer modeling of the chromatographic process.
     From the compound retention times using two different chromatographic analyses, it is
      possible to predict the best oven temperature and flow conditions for the target volatile
      compounds. This greatly simplifies the task of optimizing the GC method development,
      especially for complex mixtures. The application of thermodynamic retention indices will
      be demonstrated for the analysis of an expanded list of volatile organic compounds.
      Examples will be shown using the 502.2 and 624 stationary phases which are the most
      commonly used phases for these analyses.
                                            452

-------
     APPLICATION OF SW-846 METHODS TO THE IDENTIFICATION OF UNUSUAL
      BROMINATED COMPOUNDS  IN HIGH-CONCENTRATION PROCESS  STREAMS.

D.V.  Smith   and   L.P.  Pollack,  Science   Applications  International
Corporation,  7600A Leesburg Pike, Falls Church, Virginia 22043, and F.T.
Varcoe,   Triangle  Laboratories  Incorporated,   P.O.  Box  13485,  Research
Triangle Park, North Carolina, 27709.

Abstract

      This   presentation   reports   work   performed   to   characterize
manufacturing residuals containing brominated and other organic compounds.
SW-846  methods  were  used   for  analysis of  volatile and  semi-volatile
organics and  selected metals.   Methods developed for  the  TSCA testing
program were used for analysis of brominated dioxins and furans.  Metals
analysis  was performed  using  ICP/MS  instrumentation   to  utilize  its
superior dynamic range.  However,  it  was found  that selenium could not be
reliably determined by this method because  of  interference  from bromine
isotopes  of  similar  atomic mass.   GFAA  was  evaluated  as a  specific
confirmatory method for selenium in  this environment.  It was also noted
that the use of SW-846 methods  8260  and 8270 as  written may not provide
the  level  of QC  needed  for  complete  evaluation of  the volatile  and
semivolatile organic data.   To quantify  the predominant organic compounds
in the samples collected, significant dilutions  were required resulting in
loss of some spike recovery data. The paper presents options for sample
extraction and analysis which may assist in overcoming these difficulties.

Introduction

The purpose of this presentation  is  to  describe  some special challenges
that  were  encountered  in  analyzing   a  series  of  high-concentration
industrial wastes containing brominated compounds.

Background

A study  of  flame  retardant manufacturing resulted  in a need  to  sample
several process waste streams  containing  brominated compounds.   These
samples were  analyzed  for volatile and semi-volatile organics, metals,  and
brominated dioxins and furans.   Several  aspects of this work were unusual
and demonstrated  issues  not normally  encountered in analysis  of  other
industrial wastes or environmental media:

o     Brominated dioxins  and furans were suspected to be present in these
      samples and  newly developed  methods  were used  to identify  and
      quantitate these species.

o     It was  found that bromine present in the samples interfered with the
      determination of selenium by ICP-MS.

o     Many of the  samples contained elevated concentrations of one or two
      semivolatile organics  which made determination of other species more
      difficult.

o     Some samples consisted of  20%  NaOH, making field  adjustment  of pH
      impossible and complicating laboratory handling.
                                   453

-------
Dioxins/Furans

SW-846 Method 8290 addresses the determination of chlorinated dioxins and
furans in  solid wastes, but does  not include their brominated analogs.
Methods  have  been developed  for  these  compounds  [1-3]  to  meet  the
requirements of EPA's dioxin product  test  rule  (40 CFR §766).   Methods
development was sponsored by the Brominated Flame Retardant Industry Panel
(BFRIP).   In this work, we  applied these methods to  industrial  wastes.
Some samples contained  brominated  diphenyl  ether  flame retardants,  which
could interfere with the determination of the  furans.  The methods include
HPLC cleanup procedures to  remove  \these  compounds.

A typical mass of spectrum of 1,2,3,7,8-pentabromodibenzofuran is shown in
Figure 1.  Criteria ions used  for  identification  of  this compound are as
follows:
         Mass
 Composition
                Abundance'
559.608
561.606
563.604
C12H379Br481BrO
C12H379Br3Br20
C12H379Br281Br30
0.51
1.00
0.98
          (1)
            Theoretical abundance relative to peak at 561.606
Brominated  compounds  are  immediately  recognizable  in  a mass  spectrum
because  the  element consists of  two isotopes of almost  equal  abundance
 (50.54% 79Br  and  49.46% 81Br)  giving rise to series of doublets,  triplets,
etc.  Note that the major ions present have almost twice the mass of those
which  would be  seen  in the spectrum  of  the corresponding  chlorinated
dibenzofuran.


07
135 Er'^^^^Cr^^^'^Br

163 214 2f2
1 • i1, -r 1- ' "


A


1

295
1

323
t
       125   150   175   200   225   250   275  300   325  350
323
1] 374
404
,1 «5 «2
56]
It
     325   350   375
400
425   450   475
                                            500
                            525   550
  Figure 1.   Mass Spectrum of 1, 2, 3,7,8-Pentabromodibenzofuran (Courtesy
              of Mike  Re,  Radian  Corporation)
                                    454

-------
A unique problem encountered  in  the  application of these methods  is  the
fact that  methylene bromide,  the  extraction solvent of  choice,   is  not
available commercially in the  same high purity as methylene chloride.   The
analyst may have to purify the methylene bromide  and verify that it  is
free of contaminants before beginning sample extractions.

ICP/MS

All  of  the   metals  included  in  the  definition  of  the  Toxicity
Characteristic, except mercury,  were quantitated using Method 6020  (ICP-
MS) .  This  method generally gives superior detection  limits together with
broad dynamic range.  However, we found that  this method cannot be applied
to the determination of selenium  in samples  containing bromine because  of
mass-number interference.  Selenium has six stable isotopes which span the
same range as the isotopes of bromine:
Isotope
7
-------
                                                           Sample
                                                           17903
                                     100
       I   I
        125
                         As
                           Se
Ag
Calibration
Blank
                                            ,Cd
                            i  I   i
                                     100
        125
  Figure 2.    ICP-MS Spectrum of (top) a brominated -sample containing no
               selenium,  and (bottom) the calibration  blank.
Acknowledgement s

The authors  wish to thank Edwin Rissmann, U.S. Environmental Protection
Agency, Bruce Colby and the staff of Pacific Analytical Incorporated, Mike
Re, Radian Corporation,  and  Siu-Fai Tsang, SAIC,  for their assistance and
guidance during this project.

References

[1]   Y.   Tondeur,    "Analytical  Protocol   for   the  Determination  of
      Polybrominated  Dibenzo-p-dioxins  and  Dibenzofurans   by  High-
      Resolution    Gas   Chromatography/Medium   High-Resolution   Mass
      Spectrometry   in  Pentabromodiphenyl Oxide",  Prepared  by Triangle
      Laboratories   for  the  Brominated Flame  Retardant  Industry Panel,
      January   1991.     Similar   protocols   have   been   prepared  for
      decabromodiphenyl oxide,  octabromodiphenyl oxide,  tribromophenol,
      and tetrabromobisphenol A.

[2]   J.R.  Donnelly,  W.D.  Munslow,  T.L.  Vonnahme, N.J.  Nunn,  and C.M
      Hedin, "The Chemistry  and Mass Spectrometry for Brominated Dibenzo-
      p-dioxins  and Dibenzofurans",   Biom.  and  Environ.  Mass  Spect.,
      14:465-472  (1987).

[3]   J.R. Donnelly,  W.D. Munslow,  A.H. Grange, T.L Petitt, R.D. Simmons,
      K.S.   Kumar,   and  G.W.   Sovocool,  "A  Gas   Chromatographic/Mass
      Spectrometric Approach for Isomer-Specific Environmental Monitoring
      of the 1700 Bromo-, Chloro-,  and  Bromochloro-  Dibenzo-p-dioxins",
      Biol.  Mass Spec., 20:329-337  (1991).

Notice: This  work was funded by the U.S. Environmental Protection Agency under contract, no.
68-WO-0027   It has not been reviewed by  the Agency and the contents do not necessarily
reflect the views and policies of EPA. Mention of trade names, commercial  products,  or
publications  does not constitute1 endorsement or recommendation  for use.
                                    456

-------
R1    RESULTS OF LABORATORY TESTS OF THE ACCELERATED ONE-STEP™ LIQUID-
w '                                    LIQUID EXTRACTOR

           Roy-Keith Smith. Ph. P.. Robert G. Owens, Jr.- Denise S. Geier, Timothy Allen Shaw and
                                           Donna Phinney
                       Analytical Services, Inc, 390 Trabert Ave, Atlanta, GA 30309
                                           James B. Carl
                               Coming Incorporated, Corning, NY 14831

       ABSTRACT

       A series of experiments were conducted using the Corning Accelerated One-Step™ liquid-liquid
       extractor/concentrator for BNA and TPH analyte isolation on real samples.  The results  indicate
       the device gives comparable results to the traditional designs of continuous liquid-liquid extractors
       in a fraction of the time and solvent usage.  The device was evaluated as a replacement for the
       separatory funnel liquid-liquid technique, and for the most part offers distinct advantages.

       INTRODUCTION

       Continuous liquid-liquid extractors play a key part  in the laboratory analysis of extractable
       semivolatile organics such as PAHs, PCBs, chlorinated insecticides, acidic herbicides, phthalates,
       and a host of other pollutants from industrial wastewaters. Use of the extractors avoids formation
       of intractable emulsions, a common problem in the separatory  funnel liquid-liquid extraction
       methods. Further the extractors are not prone to unexpected selective retention of target analytes
       or plugging due to particulates, as  are the proposed solid phase extraction devices. However, the
       major drawback to the use of continuous liquid-liquid extractors is the 18-24 hours required for an
       extraction. This is particularly a problem when the sample  must be made acidic, extracted, then
       made basic and re-extracted. A second problem is the decomposition of sensitive target analytes
       in the strongly basic or acidic solution  during the lengthy extraction.  A third, and potentially the
       worst problem, is the relatively large volumes of chlorinated solvents which are used and which
       must be either disposed or recycled at the end of the extraction.
       In 1991 Corning Glass Works invented a fundamental modification of the continuous liquid-liquid
       extractor, by addition of a hydrophobic membrane  to the bottom of the extraction chamber.  The
       membrane allows passage of non-polar organic solvents and dissolved materials, while retaining
       polar and aqueous liquids. Use of the membrane eliminates the solvent pool  at the bottom of the
       extraction chamber and renders obsolete the  solvent siphon tube. The membrane  serves a second
       function in drying the solvent as it passes through, eliminating the need for sodium sulfate drying
       columns after the extraction process.
       Analytical Services,  Inc. began testing the Accelerated One-Step™ Extractor in January, 1993.
       Objectives of the test experiments  were to determine the amount of time necessary for extraction
       of analytes, determine the range of extractions which could be performed with the device, examine
       which solvents were compatible with the device and produce date comparing the effectiveness of
       the device with the liquid-liquid separatory funnel and more tradition continuous liquid-liquid
       extractors. The results of these test objectives and some observations of the device in use are the
       subject of this paper.

       METHODS AND MATERIALS

       BNA surrogates, matrix spike solutions and internal standards were purchased  as concentrates
       from Supelco (Belafonte, PA) or Ultra  Scientific (Kingston, RI) and were those specified in SW-
       846, 3rd Edition. Final Update. Methylene chloride, 1,1,2-trichlorotrifluoroethane (Freon 113™),
       carbon  disulfide, acetone and  hexane  were of the highest  quality available (Pesticide Residue
       grade or Nano-grade™) from Fisher Scientific. All volumetric measurements were performed in
       Class A volumetric glassware.  BNA analyses were performed by EPA method 8270A, 1990  on a
       Hewlett-Packard 5890A, Series II  GC with 5971 Mass Selective detector.  The capillary column
       used was a Hewlett-Packard HP5-MS column.  Daily QC acceptance criteria for DFTPP tuning,
                                               457

-------
SPCC, and CCC were met prior to analysis, as specified in the method.  TPH analyses were
performed by modified EPA method 418.1.
The extractor was cleaned prior to each use with detergent, brush and hot water, rinsed with hot
water, rinsed with chromic acid solution, rinsed with hot water, rinsed twice with DI (organic free)
water and allowed to air dry.  Immediately before use it was rinsed with the extraction solvent.
The membrane was rinsed immediately prior to use with the extraction solvent.  After assembly (3
to 5 Teflon™ boiling chips were added to the concentrator flask), 50 mL of solvent was added to
the extraction  chamber with  the stopcock closed.   The sample was  added to the  extraction
chamber, then the container rinsed with 50 mL of the extraction solvent which was added to the
chamber.  The surrogates and/or matrix spikes were added to the sample with a volumetric pipet
followed by pH adjustment with 1:1 sulfuric acid or 10 M sodium hydroxide.  pH was read with
wide range paper. The stopcock was opened and about half of the solvent was allowed to drain
into the concentrator flask.  Heat was applied to the concentrator flask.  Extraction times were
established from the first drops of solvent to fall from the condensor into the sample. At that time
the stopcock was opened and all the solvent allowed to drain into the concentrator. A distillation
rate of 15 to 25 mL per minute was maintained. At the end of the extraction the stopcock was
closed and the solvent concentrated to about 10 mL, at which time the heat source was removed.
The apparatus was allowed to cool, then the concentrator was removed. The solvent was removed
to 2 mL under nitrogen blowdown.  The concentrate was quantitatively transfered to a Class A 2
mL volumetric ground glass stoppered tube which had previously been calibrated at the 1.00  mL
level.  The sample was reduced to 1.00 mL with nitrogen blowdown,  internal standards were
added and the sample sealed and shaken followed by transfer to an autosampler vial for analysis.
TPH extractions were similar except there was no solvent concentration.

RESULTS AND DISCUSSION

The first set of experiments were performed by adding acid matrix spike (200 ng in methanol per
method 3500) and BN surrogate (100 ng in methanol per method 3500) compounds to DI water at
neutral pH and extracting them with methylene chloride for 1, 2,4 and 16 hours.  The results are
in Table 1 with example plots  shown in Figures 1 and 2 . The point  at time 30 hours (PES)
represents the results of a post-extraction spike and analysis of the surrogate mixture. Poor results
were obtained for the acidic  spikes (phenol, 2-chlorophenol, 4-chloro-3-methylphenol and 4-
nitrophenol) as expected for pH 7.  The BN surrogates (nitrobenzene -ds, 2-fluorobiphenyl  and
terphenyl-di4) were  recovered with above 90% efficiency  within  4 hours.   Surprisingly
pentachlorophenol was recovered in 90% efficiency in 4 hours at pH 7 (Figure 3).

       Time              1            2            4            18          PES
     Compound
       Phenol            35.1        61.3          58.8         69.1           194
   2-Chlorophenol        18.8        43.2          37.6         40.1           172
  Nitrobenzene d5       45.3        72.2          91.5         93.0          95.4
4-Cl-3-methylphenol     8.85        23.8          14.5                        168
  2-Fluorobiphenyl       56.9        75.8          90.9         90.6          99.8
   4-Nitrophenol        23.2        45.2          66.7         86.3           174
  Pentachlorophenol      52.9         123           175          192           195
   Terphenyldl4        82.2        74.9          86.4         86.6          82.1

Table 1.  Acid matrix spike compound and BN surrogate recoveries (ng) for extractions at various
times (hours) at pH 7.
                                         458

-------
                                                             PES Result
                 ng Phenol spike recovery at pH 7
                                 15      20
                                  Time Hours
Figure 1.  Phenol recovery (in ng) vs. time (hours) extracted at pH 7.  The PES represents the
amount of spike placed in the extractor as determined by neat analysis.
                                        -o
                                                           -O  PES Result
                           ng Nitrobenzene d5 recovery at pH 7
                           10
1 5      20
Time Hours
25
30
35
Figure 2.  Nitrobenzene ds recovery (ng) vs. time (hours) extracted at pH 7.  The PES represents
the amount of spike placed in the sample as determined by neat analysis.
                                         459

-------
                                                             PES Result
                      ng Pentachlorophenol spike recovery at pH 7
        25
                           10
             25
       30
  35
                                   15      20
                                   Time Hours
Figure 3. Pentachlorophenol recovery (ng) vs. time (hours) extracted at pH 7. The PES represents
the amount of spike placed in the extractor as determined by neat analysis.

      The experiment was repeated with initial acidification to pH < 2 with extraction times of 1,
2, and 4 hours.  Over 80% recoveries of all phenolic compounds (Table 2.), except 4-nitrophenol,
were obtained, as illustrated for phenol (Figure 4). The 4-nitrophenol results are illustrated in
Figure 5. These results indicated that 4 hours extraction for each pH change would give suitable
results.
       Tune               1
    Compound
      Phenol           105.7
  2-Chlorophenol        127.2
  Nitrobenzene d5        69.2
4-Cl-3-methylphenol     128.3
  2-Fluorobiphenyl        69.9
   4-Nitrophenol         94.9
 Pentachlorophenol      129.1
   Terphenyl d!4         72.5
131.9
144.8
85.7
165.8
82.0
146.3
187.1
80.2
165.2
165.2
92.3
190.9
90.1
162.4
205.7
92.6
 PES

188.4
185.1
 98.4
191.4
100.7
210.8
197.0
 89.1
Table 2. Acid matrix spike compound and BN surrogate recoveries (ng) for extractions at various
times (hours) at pH <2.
                                         460

-------
                                                               PES Result
                      ng Phenol spike recovery at pH <2
            0       5       10      15      20
                                   Time Hours
Figure 4.  Phenol recovery (ng) vs. time (hours) extracted at pH <2.  The PES respresents the
amount of spike placed in the extractor as determined by neat analysis.
                                                              PES Result
                        ng 4-Nilrophenol spike recovery at pH <2
                           10
25
30
35
                                   15      20
                                    Time Hours
Figure 5.  4-Nitrophenol recovery (ng) vs. time (hours) extracted at pH <2.  The PES respresents
the amount of spike placed in the extractor as determined by neat analysis.

The extraction of TPH samples for IR analysis with Freon 113™ (1,1,2-trichlorotrifluoroethane)
was examined.  Spikes of the calibration standard mix (isooctane, hexadecane and chlorobenzene),
were made into Dl water and extracted for various times. The results are presented in Figure 6.
The  first  three data points were collected and analyzed on one day and the 8 and 22 hour
                                         461

-------
extractions performed on another day.  The results are equivalent given the ±10% acceptance
criteria of the calibration curve and indicate that a 1 to 2 hour extraction time is suficient.
       120 -i
       100 -
                                                                -Q
                 2.5
7.5
                                   10   12.5   15   17.5   20   22.5
                                     Hours
Figure 6. THP standard recovery (percentage) from DI water vs. duration (hours) of extraction
using Freon 113 as the solvent.

The ability of the hydrophobic membrane to pass organic compounds was evaluated by placing 50
mL of a solvent in the extraction chamber and recording the time required to drain. The data are
presented in Table 3.  Most of the solvents tested passed through the membrane within the same
time window, Freon 113 being marginally faster.  As expected the hydrogen bonded solvent, n-
butanol was significantly retarded.  Carbon disulfide has been proposed as a possible Freon 113
replacement for the IR-TPH  test, however it  is not suitable for use with the Accelerated One-
Step™ membrane, as shown in Table 3.  Carbon disulfide is an opposite polarity compound
(carbon  is polarized  negative and sulfur positive) as compared to the other halogenated and
hydrocarbon solvents tested.  These  data support the observations of the lower extraction
efficiency of the  hydrogen bonding carbamates, amines and highly acidic phenols, such as the
mono- and di-nitrophenols, as reported by Bruce (1993).
                  Solvent
                 Freon 113
                  Hexane
                  Benzene
                Chloroform
             Methylene chloride
           tert- Butyl methyl ether
                  Acetone
                 n-Butanol
              Carbon disulfide
                   Water
                           Minutes. Seconds
                                 8.45
                                10.40
                                16.07
                                11.43
                                 9.55
                                12.42
                                10.17
                                 55.0
                           < 10 mL after 55.0
                            none after 120
Table 3. Time (minutes) for 50 mL of solvent to pass through the hydrophobic membrane of the
Accelerated One-Step™.
                                         462

-------
A number of real samples were then extracted using 4 hours extraction times each for the acid and
basic  extractions.   Two different samples, a TCLP extraction and a wastewater  sample were
performed as BNA matrix spike (MS) and matrix spike duplicate (MSD), as part of a  normal batch
of samples.  The MS was extracted using the separatory funnel technique of Method 3510 with
sodium sulfate drying column and Kuderna-Danish concentration, while the MSD was extracted
using  the One-Step™ extractor and  direct concentration, followed by nitrogen blowdown. The
results are presented in Table 4 and demonstrate that the One-Step™ extraction is at  least as good
as the separatory funnel technique of Method 3510 for normally encountered samples.

                                     TCLP Sample           Wastewater Sample
Surrogate Compounds            Sep Funnel    One-Step™   Sep. Funnel   One-Step™
2-Fluorophenol                      35           62            33            49
Phenol d5                          28           79            25            65
Nitrobenzene d5                     88           86            89            85
2-Fluorobiphenyl                    75           83            82            77
2,4,6-Tribromophenol                84           89            75            81
Terphenyldl4                      90           88            83            73

Matrix Spike Compounds
Phenol                             15           71            11            60
2-Chlorophenol                      31           73           29            65
1,4-Dichlorobenzene                 39           77            15            65
N-Nitrosodipropylamine              57           103          45            81
1,2,4-Trichlorobenzene               43           80           20            66
4-Chloro-3-methylphenol             35           86           26            85
Acenaphthene                       49           88           45            79
2,4-Dinitrotoluene                    42           74           42            80
4-Nitrophenol                       14           74            15            75
Pentachlorophenol                   48           97           39            78
Pyrene                             55           83           44            77

Table 4. Comparison of percent recoveries of BNA matrix spike and surrogate compounds on the
same  samples  using  Method 3510 (separatory  funnel) and  the  modified 3520  using  the
Accelerated One-Step™

An industrial wastewater sample from a industrial cleaner  formulator was analysed for acid
extractables as a sample and as a matrix spike with separatory funnel extraction. Past history of
the sample indicated it needed to be diluted 1:20  (50 mL sample to 950 mL Dl water) to avoid
swamping the GC/MS detector. The same sample  at 1:20 dilution was extracted on 4 consecutive
days as a BNA matrix spike with the Accelerated One-Step™. The results are presented in Table
5.  The One-Step™ consistently produces higher recoveries of the more volatile compounds such
as phenol, 1,2-dichlorobenzene, 2-chlorophenol and benzoic acid, while prerforming about as well
as Method 3510 on the rest of the compounds.
                                        463

-------
Compounds
Analytes
1,2-Dichlorobenzene
2-Methylphenol
4-Methyl phenol
Isophorone
Benzole acid
Naphthalene
Diethylphthalate
Di-n-butylphthalate
Bis(2-ethylhexyl)phthalate
Surrogates
2-Fluorophenol
Phenol d5
Nitrobenzene d5
2-Fluorobiphenyl
2,4,6-Tribromophenol
Terphenyl d!4
Matrix Spike Compounds
Phenol
2-Chlorophenol
1,4-Dichlorobenzene
1,2,4-Trichlorobenzene
4-Chloro-3-methylphenol
Acenaphthene
2,4-Dinitrotoluene
4-Nitrophenol
Pentachlorophenol
Pyrene
   Sep. Funnel
Sample     MS
  32       32
                Accelerated One-Step
          MSI      MS2     MS3
                         ,TM
   3
   6
  103
   4
   5
  28
   4

  60
  80
  75
  58
  177
  85
 2
 1
72
 3
 1
12
 3

54
74
92
47
153
63

67
119
65
62
162
53
63
48
130
47
 97
 15
 9
 5
173
 7
 12
 26
 6

 90
201
 95
 59
157
 71

205
197
102
 74
182
 68
 77
116
 88
 71
 97
139

 2
249
 8
 15
 12
 7

 63
176
 86
 49
148
 68

155
175
 95
 67
177
 57
 79
119
 85
 63
91
128
14
 3
189
 7
11
18
 5

72
182
88
55
146
72

160
172
82
63
173
67
82
132
94
73
MS4
 65

 10
  1
 120
  5
 10
 11
  1

 87
 168
 80
 35
 128
 57

 171
 151
 75
 52
 165
 44
 73
 131
 77
 47
Table 5.  Comparison of recoveries of target analytes and matrix spike compounds for a strong
industrial effluent processed with  a  separatory funnel by  method 3510  and processed on 4
consecutive days with the Accelerated One-Step™ extractor.  Acid compounds are spiked at the
200 ng level while BN compounds are at the 100 ng level.

A wastewater treatment plant sludge (6%  solids)  was extracted for BNA analysis using a
continuous liquid-liquid extractor and the Accelerated One-Step™ extractor. Another portion of
the same sample was diluted 1:30 with DI water and 1 L extracted using  Method 3510 with a
separatory funnel.  The  membrane of the Accelerated One-Step™  did not plug, however
channeling of the solvent through the sludge was noticed.  The surrogate recoveries for the three
extraction techniques are presented in Table 6. Although none of the recoveries were quantitative,
the Accelerated One-Step performed significantly better than the other two techniques.
                                        464

-------
                                      Sludge sample with 6% solids
Surrogate Compounds             Sep Funnel   Cont. Liq-Liq   One-Step™
2-Fluorophenol                       006
Phenol d5                            5              14            26
Nitrobenzene d5                      47            33            63
2-Fluorobiphenyl                     27            31            60
2,4,6-Tribromophenol                 2              3             49
TerphenyldH                       31            45            65

Table 6. Comparison of percentage suggogate recoveries of BNA separately funnel, continuous
liquid-liquid and Accelerated One-Step™ extraction of a sludge sample containing 6% solids.
The separatory funnel extraction was performed on a 1:30 dilution of the sludge.

Three other wastewater samples have been extracted for BN or BNA analysis.  The surrogate
recoveries are presented in Table 7. In general, use of the Accelerated One-Step™ extractor gives
recoveries which are comparable to those obtained using separatory funnel techniques in each
respective batch.

Surrogate Compounds              41082-1        40619        40598-1
2-Fluorophenol                                 -  65 (56)
Phenol d5                                        93 (62)
Nitrobenzene d5                   115(121)      101(108)       69(73)
2-Fluorobiphenyl                  105(88)        74(96)        69(74)
2,4,6-Tribromophenol                            105 (103)
Terphenyl d!4                     87(98)        74(79)        50(61)

Table 7.  Percentage  surrogate compound recoveries for wastewater  samples  extracted  and
concentrated using the Accelerated One-Step™  Samples 41082-1 and 40598-1 were extracted
for base/neutrals only.  The values in parentheses are the average recoveries for the batch.

Two instances of membrane plugging were encountered. The first was a sample which contained
a fine dense grit which coated the membrane and  was not displaced by  the methylene chloride
solvent.  The sample was  successfully prepared by separating the solid from the liquid layer,
separately extracting the solid with sonication and the liquid with the One-Step™, then combining
the extracts. The second was a sample which formed a tenacious emulsion as the solvent fell from
the condenser. The surface of the bubbles coated the membrane and stopped the flow of solvent
through it. The sample had to be processed by  continuous liquid-liquid extraction followed by
removal of the large amounts of water in the extract by slurrying with sodium sulfate.

SUMMARY

The Corning Accelerated One-Step™ liquid-liquid extractor/concentrator has been shown to offer
rapid extraction  of target  analytes  with  acceptable recoveries and minimal solvent usage as
compared to more traditional liquid-liquid extractors. When compared  to separatory funnel
liquid-liquid extraction, the Accelerated One-Step™  is shown to be a viable alternate with
demonstrated savings through reduced solvent usage and elimination of the drying  column and
Kuderna-Danish concentrator with no decrease in the recovery of most analytes.

REFERENCES

Bruce, M.L., 1993. Extraction of Appendix IX BNAs, Pesticides and PCBs by Accelerated One-
Step™ Liquid-Liquid Extractor/Concentrator with Analysis by GC/MS  & GC/ECD. A Single
Method Validation Report. Enseco-Wadsworth/ALERT Laboratories,  4101 Shuffel Dr  NW
North Canton OH 44720
                                        465

-------
A MATCHED DUAL CAPILLARY COLUMN SYSTEM FOR PESTICIDES

Mitchell R. Hastings, Alan J. Spilkin. J&W Scientific, 91
Blue Ravine Road, Folsom, California 95630. Phone  (916) 985-
7888

ABSTRACT

In early and  existing GC pesticide analyses methods, packed
columns were  and are  employed.   The use  of  packed columns
has, in many  cases,  given way to  the use  of  Megabore (0.53
mm  I.D.)  columns in  these  methods.  Many laboratories  and
regulatory  agencies  have realized  that, relative  to packed
columns,  Megabore columns  offer  greater  reproduceability,
longer   lifetimes,   increased  resolution,   and  decreased
overall analysis times.

Many laboratories, among them those doing pesticide analyses
for the EPA and  FDA,  have both a high sample throughput and
require quick  sample  turnaround  times.   It follows that any
enhancement to the  GC system that increases  resolution  and
decreases run  times will enhance the performance and profits
of  these  laboratories.   This  decrease  in run  time becomes
even more attractive  if  minimal  to no hardware changes need
be  made.    What  follows  is  an  experiment  converting  the
chromatography of  an EPA Contract Laboratory Program (CLP)
Chlorinated   Pesticide  analysis  currently  done  by  two
Megabore columns  to  an analysis utilizing two  0.32  mm I.D.
columns.   A reduction in run  time of over 25%  is realized
while  still maintaining the ability  to use  a  Splitless  or
Megabore direct injection inlet.

INTRODUCTION

BACKGROUND

Quantitative and confirmatory qualitative information can be
obtained for  a list of analytes by running the analytes  on
two columns with significantly dissimilar stationary phases.
Many state  and federal regulatory  agencies accept this data
as conclusive.  The benefit of single injection, dual column
analysis is a time savings by having both quantitative  and
confirmatory qualitative information available from one run.

For  example,  from analysis  of a  sample using  one column,
tentative identification  of the  analyte  of interest is made
if  its retention  time matches  the retention  time  of  the
standard run  on  that  column.   Quantitative confirmation the
presence of that component  on  the  column with the different
stationary  phase if  the  component's  retention  time  also
matches the standard's retention  time  on  that  column.   If
                              466

-------
components are  resolved  on both columns,  quantitation data
can  be  compared  to  insure  quantitative  accuracy.    This
technique becomes even more powerful if a selective detector
is employed.

In the  latest  available  version (OLMO1.3,  2/91)  of the CLP
Chlorinated  Pesticide Method,  two  30  m  x  0.53  mm  I.D.
columns with dissimilar  stationary phases are used (DB-608,
50%   phenyl   50%   methyl   polysiloxane;   DB-1701,   14%
cyanopropylphenyl, 86% methyl polysiloxane)  and a selective
detector (ECD)  is employed in the analysis of 20 chlorinated
pesticides.    It  was  the  purpose  of  this experiment  to
determine if two 0.32 mm I.D. columns of the same stationary
phase type could  be utilized in  order  to  decrease  the run
time, while still meeting  the required  resolution criteria,
and without changing hardware or increasing hardware costs.

INLET AND COLUMN CONNECTION CONSIDERATIONS

A number of different options were considered when designing
this single  injection,  dual column  system.   Five  criteria
were used  to evaluate certain  inlet and  column  connection
options.  The inlet and column connection design had to be:

1) equal or  lower  in cost compared to the  current  low cost
of a Megabore configuration--the least expensive mode,
2) adaptable to splitless or Megabore direct injectors since
both injection techniques are currently  in use,
3) highly  inert  (some analytes  degrade easily  when exposed
to active  sites)  by eliminating any metal  or  other  active
sites in the sample flow path,
4) easily,  reliabilly, and reproducibly  installed,
5) amenable to the use of two 0.32  mm I.D. columns.

Table I  is a report card  of how  each option  scored  in the
above categories.
Table I
Option
Qualified Unqualified
Graphpak Divider for Simul-
taneous Sampling to Two
Capillary Columns (Hewlitt
-Packard)

Injection Tee Kit for
Simultaneous Analyses
on Two Wide Bore Columns
(Supelco)
   4,5
1,2,3
   1,3
2,4,5
                             467

-------
Two Hole Ferrule                     1,2,3,5       4

Deactivated Megabore Guard
Column with a Deactivated
0.53/0.32/0.32 mm                   1,2,3,4,5
3-Way "Y" Union

To  reduce activity,  a deactivated,  splitless  liner would
have been used  if splitless injection mode  was  used.  This
particular experiment employed the Megabore direct injection
mode as it was thought that this technique could potentially
pose sample transfer  and  solvent front problems1 that would
not result  from the use  of the splitless mode  (see "Inlet
and  Column Flow Considerations"  below).    A  deactivated,
Direct  Flash  Vaporization liner  (J&W  Scientific)—the type
of  liner that allows  the  Megabore  guard column to seal into
a tapered region below an  expansion chamber—was  used with
the  Megabore  direct  injection mode.    The  configuration
decided upon is shown in Figure I.

COLUMN DIMENSION CONSIDERATIONS

Provided  that  temperature,  phase  ratio,  stationary phase
type,  linear  velocity, test  compound,  and  length  are kept
constant—smaller diameter columns yield an increased number
of  theoretical plates and sharper peaks than larger diameter
columns.

As  a demonstration  of  this,  two 30 meter,  DB-608 columns of
the same phase  ratio  but  with differing diameters (0.53 and
0.32  mm I.D.'s)  were tested  under identical  temperature,
carrier  gas,  and linear  velocity  conditions.    To simplify
the mathematics,  the  equation used to describe  the area of
the Gaussian  peak ideally obtained from a  chromatogram was
that of  a  triangle: Area = 1/2 Base x Height.   Theoretical
plates  were  determined using  the  same compound,  and width
was calculated using the following formula2:

w = 4t / n1/2
where

w = width at the base of the peak
t = retention time  of the peak
n = calculated theoretical plates

If  the same  amount  is  injected  on   each  column,  we  will
assume this amount  generates  the  same  number of area counts
whether  chromatographed  using the  0.53  or  0.32  mm  I.D.
column.   So  if the area  remains  the  same and  the width of
the  peak  decreases,  the   height  of  the  triangle  must
increase.
                              468

-------
So we can say that:

0.5 x Base0_53 x Height0>53 = 0.5 x Base0>32 x Height0-32

or
Base0_53/Base0_32  = Height0i32/Height0>53

The 0.53 mm  I.D.  column generated  40400  theoretical  plates
and the retention time of the test compound was 11.523 min.,
therefore the width  of  the base of this  peak  is 0.229 min.
The height was set at 1 unit  since  we are trying to see the
resulting ratio difference in height.

The 0.32 mm  I.D.  column generated  85100  theoretical  plates
and the retention time  of  the test  compound was 11.460 min.
which makes the base of the peak 0.157 min wide.  Therefore,
all of the above parameters remaining the same, the increase
in the  peak  height  resulting from the use of the  0.32 mm
column as opposed to the 0.53 mm I.D.  is approximately 1.5
for the same compound when using the triangle approximation.
This  figure  cannot  be  exactly  correlated  to  peaks  of
differing retention,  or peaks eluting on the ramp portion of
a  temperature  programmed  run,  but it does  demonstrate  a
benificial peak  height,  or signal  to noise  increase which
implies  a  resulting increase  in  system  sensitivity when
using a smaller,  rather than a larger diameter column.

INLET AND COLUMN FLOW CONSIDERATIONS

With  the  above  column  considerations  in  mind, why  not use
0.25  mm I.D. or smaller  diameter  columns  in  this   single
injection,  dual  column  setup  and really  minimize  run times
and system sensitivity?

In  order  for sample transfer  to  occur  efficiently  (in  a
short period of time and  with minimal band broadening) from
a  Megabore  direct  inlet  to  a  column(s),   their  generally
needs to be a minimum of 4-5 mL/min. of carrier flow through
the  inlet3.   Therefore,  the  summed  flow  through the two
columns must be at least 4-5 mL/min.

As demonstrated  by Van Deemter  curves4,  carrier gases have
flow  ranges  for  particular  length,  diameter,  and film
thickness  columns   in  which  they  will  yield  the  most
efficient    (greatest   number   of   theoretical   plates)
chromatography.   Table II shows what are considered general,
effecient flow  ranges  for  different diameter,  length, and
film  thickness columns  for components  of varying retention
using Helium or Hydrogen as a carrier gas5.
                              469

-------
Table II
                        He Efficient         H2 Efficient
Inner Diameter (mm^  Flow Ranae  (mL/min)  Flow ranae  (mL/min)
0.25
0.32
0.53
0.9-1.3
1.4-2.2
4.0-6.0
1.3-1.8
2.2-2.9
6.0-7.9
Since  most  analysts  doing  the  CLP  Chlorinated Pesticide
analysis  are  using  Helium  as  a  carrier  gas,   and  since
Helium's  efficient   flow  range  is  lower  than  that  of
Hydrogen, Helium was  considered  the limiting parameter.   It
was selected as the  carrier  gas in this  experiment as its
lower  efficient  flow range  could  potentially  cause  flow
problems  in  the Megabore  inlet,  or chromatography problems
if  it was  necessary to  operate  above the  efficient  flow
range to avoid  inlet  problems.

Allowing for the  flow limitations of the Megabore inlet and
the  Helium   carrier  gas,  it   becomes  apparent  that  the
smallest  diameter  column  that  could  be  most efficiently
employed in  a single  injection,  dual column analysis is the
0.32  mm  I.D.  column,  the  efficient  flow  range  for  two
columns summed  being  2.8-4.4 mL/min.

If  sample transfer problems  from  the inlet to  the column
exist, they  would  be  manifested  in two ways.  First, if the
solvent  front  interfered  with  the  chromatography of  the
first  eluting  peak,  it would  tail,  be  excessively broad,
and/or  possibly  rounded  at  the  top,  especially  on  the
thinner film (0.25 jum)  DB-1701 column.  Second, if the flow
through the  inlet  was inadequate, all  peaks might tail.

ADDITIONAL COLUMN  CONSIDERATIONS

There  are   a  number  of  advantages,   relavent  to  this
experiment,  that  smaller diameter  columns  have over larger
diameter  columns  of the same  length and  phase  ratio.   They
produce  less bleed because  there  is  less  stationary phase
available  to  degrade.   Under  proper flow  conditions,  as
demonstrated above, smaller diameter columns generally yield
sharper  peaks.     Additionally,  because  there  are  more
theoretical  plates generated  by smaller  diameter columns,
the same  resolution can be achieved in a  shorter period of
time.   This  means  compounds  spend  less time  in the column
and,  since  the width of a peak is  directly related to the
amount of time  the compound spends  in  the column  (the amount
of  time  it   has  to  diffuse   longitudinally  in  the mobile
phase), the  compound  suffers  less from band broadening when
it  elutes  earlier.    For  the  same  amount  of  analyte  on
column,   this  results  in  narrower, taller peaks.   All of
                              470

-------
these factors  translate directly  into  a greater  signal to
noise  ratio,   and,   more   importantly,   lower,  more  easily
reached detection limits.

There are, however,  relavent  disadvantages  of using smaller
diameter  columns  as opposed  to  Megabore columns.   Smaller
diameter  columns of the  same  phase  ratio are  generally
fouled more quickly by non-volatile sample  residues.   This
fouling can be minimized by the use of an adequate length of
guard column and/or  by  the use of  a plug of silanized glass
wool  in  the  inlet.    Smaller diameters  of the  same  phase
ratio  are also  thought  to   be  slightly more  active  than
larger diameter columns because the thinner stationary phase
allows the  analytes more  interactions  with the  surface of
the fused silica tubing than  with  the thicker film Megabore
columns.    Also,  assuming  analytes are well  separated  from
the solvent front on a  1  or 2/iL  injection,  up to  8p.L can be
injected on a megabore column with a standard film thickness
without  significant  negative  chromatogrphic  consequences
(retention time  shifts  from  run  to run  and noticeable band
broadening of  components) .    At  most,  only  3-4/iL  should be
injected  on  the 0.32  mm  I.D. dual column  setup.    So,  if
injection volumes greater than  3-4juL are normally used, the
greater signal to noise generated by using a 0.32 as opposed
to a 0.53 mm I.D. column would be somewhat negated.

MEETING THE RESOLUTION REQUIREMENTS

In the latest  available version  (OLMO1.3,  2/91)  of  the CLP
Chlorinated    Pesticide    Method,    certain    resolution
requirements must be met.   All components must be adequately
resolved, particularly  the targeted pesticides  in Table III
which are ordinarily difficult  to resolve  on  the Megabore
system.   Acceptable resolution  is defined  as  the  depth of
the valley between  two  adjacent  peaks as being  greater than
or equal to 60% of the height of the shorter peak.

Table III	Pesticides in CLP Resolution Check Mixture

     gamma-Chlordane          Endosulfan Sulfate
     Endosulfan I             Endrin Ketone
     p,p'-DDE                 Methoxychlor
     Dieldrin

SUMMARY

As shown  in Figure  2,  peak   shapes  and  widths  with a  2 /xL
injection  for  tetrachloro-m-xylene,  the  earliest  eluting
peak  on   both  columns,  and  decachlorobiphenyl,  the  latest
eluting peak  on  both columns, do  not  display symptoms that
                             471

-------
would  indicate  sample transfer  problems from  the  inlet to
the column.

Figure  3  contains  the  compound list  with  the conditions
under  which  the  chromatograms  in   Figures   2,4-6  were
generated.

Figure  4  demonstrates  the  chromatography of  the  list  in
Figure 3 on the  DB-608,  and Figure  5 shows the measured and
calculated resolution of  the only partially co-eluting pair
on this  column:  alpha-Chlordane  and  Endosulfan  I.   Figure 6
demonstrates the chromatography  for  the compounds listed in
Figure 3 on the  DB-1701, and Figure 7 shows the measured and
calculated resolution  between the only partially coeluting
pair on this column: Endosulfan  Sulfate and Methoxychlor.

The GC/FID instrumentation was used for  two  reasons:  1)  it
was considered  adequate to  determine  if the chromatography
of the dual column capillary system would or would not work,
and 2) a GC/dual BCD instrument  was not available.
The  above  example   chromatograms   demonstrate  acceptable
resolution with  on-column  amounts in the 2-5 nanogram range
by  GC/FID.   CLP  Chlorinated  Pesticide  analysis,  however,
uses  GC/ECD  and routinely  analyzes for  compounds  in  the
picogram  range.    Knowing  that peak  width   is  dependent
partially  on the amount of compound  present,  it is expected
that the peaks  should  be narrower  as  there is at  least 2
orders  of magnitude  difference  in  the  amount on  column.
This difference  should result in analyte resolution better
than that demonstrated, especially for critical pairs.

If  the system described  above were  to be  used for routine
sample analysis, it  would be  highly recommend  that  a guard
column of  adequate length  (2-5  meters)  be  used in order to
protect  the  analytical  columns from   non-volatile  sample
residues.

The dual column  system using 0.32 mm I.D. columns in the CLP
Chlorinated  Pesticide  analysis,  on the  surface,  appears to
be  a  viable   alternative  setup which  would  result  in
increased  sample throughput  and  greater profits.  Also, more
samples could be run between quality control checks as their
analysis  is  required  every 12 hours as  opposed to every so
many  samples.    Detection  limits   should be  more  easily
achieved  by  using the  smaller  diameter  columns   as  they
result in an increased signal  to noise ratio of analytes.

The overall  cost of  the dual  column setup  would be less as
column price is  directly  proportional to  column diameter.
The dual 0.32 mm I.D.   setup  could  be  used in  either  the
Megabore  or  splitless  injection mode.    In this particular
                              472

-------
CLP protocol, as with most  EPA methodology,  substitution of
columns   is   explicitly  allowed.     Section   2.2.1  says
"Equivalent  columns  may  be  employed  if  they  meet  the
requirements  for   resolution,  initial   calibration,   and
calibration verification..-"  If a capillary  system such as
this  is  proven  effective   for  pesticide   analysis,   PCB
analysis   should    follow   easily   as  identification   is
accomplished   on   a   pattern   recognition   basis,   and
quantitation   should   not    change    from   whatever   mode
laboratories are currently using.

Future  experiments  might include:  single injection sample
splitting  into  three dissimiliar  stationary phase  0.25  mm
I.D.  columns;  use of hydrogen as a  carrier as  opposed  to
helium on the system described above;  an investigation into
the significance  of using the thinner film  capillaries  on
sensitive compound degredation  at  the  10-100 picogram level
using GC/ECD;  and the  investigation  of  other  Megabore  and
capillary dual column methods  that could be  adapted to this
dual column setup, or dual column setups  with other types of
stationary  phases.      The  possibility  of   using  shorter
capillaries  that  would  still  generate the  same  number  of
theoretical  plates   as   a  Megabore  column  could  also  be
investigated.
1.  Jennings,   W.  Analytical  Gas  Chromatography;  Academic
Press, San Diego,  California,  1987;  p 10.

2. Rood,  D. A Practical Guide  to the Care,  Maintenance, and
Troubleshooting of  Capillary  Gas  Chromatographic  Systems;
Huthig Buch Verlag GmbH, Heidelberg,  1991;  p 109-113.

3. Rood,  D., Ibid.;  p 108.

4. Jennings, W. Ibid.; p 111-151.

5. Rood,  D., Ibid.;  p 32.
                             473

-------
              Figure 1
   INSTRUMENT CONFIGURATION
INJECTION PORT
                         ECD
                      ECD
ANALYTICAL
 COLUMNS
   GUARD
   COLUMN
 0.5 - t METER
 GLASS"Y"
CONNECTION
           GAS CHROMATOGRAPH

-------
 900C
 Lie*.
        5.60
1.0e

980Q

960Q

940Q

920Q
                    Figure  2
5.80     6.90
    Time  (min.)
                            22
.0
              24.5        25.0
                    Time (min.)
6.20     6.40
                    25.5
              26.0
                         475

-------
                                            Figure  3
  Column:              DB-608
               30m x 0.32 mm I.D., O.Sum
               J&W Scientific P/N 123-1730

  Column:              DB-1701
               30m x 0.32 mm I.D., 0.25 Mm
               J&W Scientific P/N 123-0732
  Carrier:        Helium at 48 cm/sec
               (measured at 150" C)
  Oven:         110*Cfor0.5min
               110-140*Cat20*/min
               140-235*C at 9*/min
               235* C for 4.5m in
               235-280*C at 15*C/min
               280* C forSmin
  Injector:        Megabore direct, 250*C, 2uL
               Direct Flash Vaporization Liner
               2-5 ng each component
  Conector:      2 m x 0.53 mm I.D. deactivated guard column
               0.53/0.32/0.32, 3-way "Y" union
  Detector:      FID, 280*C
               Nitrogen make-up gas  at 30 mL/min
 1. Tetrachloro-m-Xylene
 2 alpha-BHC
 3. gamma-BHC
 4. beta-BHC
 5. Heptachlor
 6. delta-BHC
 7. Aldrin
 8. Heptachlor Epoxide
 9. gamma-Chlordane
 10. alpha-Chlordane
 11. Endosulfan I
 12. p, p'-DDE
 13. Dieldrin
 14. Endrin
 15. p,p'-DDD
 16. Endosulfan  II
 17. p,p'-DDT
 18. Endrin Aldehyde
 19. Endosulfan  Sulfate
20. Methoxychlor
21. Endrin Ketone
22.  Decachlorobiphenyl
23.  Unknown
                                        476

-------
  1.3e
    .2e
a
   1.0e
  900(
     Figure 4
DB-608: Chlorinated
     Pesticides
                                              20
                   I	f
   Jij
                                              23
                                       15  17
                               10
                                11
                                 12
                                  13   14
                                       16
                     18
                      19
                   10          15
                         Time (min.)
                        20
25

-------
CO
9500

9400

9300

920O

9100

9000

890G

8800

87001

860
       :
   85004
           14.20
                        Figure 5
                                           77.3% RESOLUTION
                     14.40       14.60
                      Time (min.)
14.80

-------
  1.4e
  1.3e4
(O
  1.0e
  good
       5
                         Figure 6
                   DB-1701: Chlorinated
                         Pesticides
                          22
UJLJW
           15
           I
                       11
       10


       2
                          14
                                  20
                             16
                              17
                                      21
                                        23

10          15
      Time (min.)
                          20
25

-------
9600
950
940
9300
 920O
 9100
 9000
          16.00
                              Figure 7
       77.3% RESOLUTION
16.20
 Time (min.)
16.40
16.60

-------
53  PERFORMANCE  DATA  FOR  THE ANALYSIS  OF  PHENOLS,   NITROAROMATICS,
     CYCLIC KETONES,  HALOETHERS AND CHLORINATED HYDROCARBONS

     Siu-Fai Tsancr.  Nellie Chau, and Paul Marsden, SAIC,  San Diego, CA
     92121 and Barry Lesnik, Organics Methods Program Manager,  OSW/EPA
     22403

          The U.S. EPA, Office of Solid Waste (OSW)  is  responsible for
     providing  reliable,  robust  analytical  methods  with  documented
     performance. This mission requires that the OSW  regularly evaluate
     developments in  analytical technologies  that  could improve  the
     measurement of  chemicals regulated under the Resource Conservation
     and  Recovery Act  (RCRA).   Whenever  modifications  to  existing
     methods   are  proposed   by   research  scientists,   government
     laboratories, commercial laboratories, or instrument manufacturers,
     the Methods Section of the OSW compares them with  existing SW-846
     methods.  If a technique is promising,  single- or multi-laboratory
     method studies  are conducted to document method performance.

          This  report  provides performance  data  for  the analysis of
     phenols  [Method 8041],  nitroaromatics  and  cyclic  ketones  [Method
     8091],  haloethers [Method  8111],  and chlorinated   hydrocarbons
     [Method  8121A]  using  capillary gas chromatography  (GC).   Target
     analytes  were   extracted  from  spiked soil  using   Method  3540
     (Soxhlet)  and analyzed  with  0.53 mm id a  DB-5  capillary  column.
     Phenols  were  analyzed  with   and without  derivatization using
     diazomethane.     Initial   conditions   for   some   chromatographic
     separations  were  derived   from  reports   presented   by  the
     Environmental Monitoring Systems Laboratory at Las  Vegas (EMSL-LV)
     during the 1990 and 1991 Symposia.
                                  481

-------
      RAPID CONFIRMATION OF NITROAROMATICS AND NITRAMINES
           USING UV DIODE ARRAY SPECTRAL COMPARISON

Bradley A. Weichert, Manager, GC/HPLC Department, Robert D.
Baker, Supervisor, HPLC Group, Dr. Curtis R. Campbell, Sr.
Staff Scientist, HPLC Group.

ABSTRACT
    The present proposed SW846 method for the determination
of nitroaromatics and nitramines, designated EPA 8330,
requires a second column confirmation for any positive
identification from the primary analysis. Since this confir-
mation is sequential to the primary analysis it results in
additional turnaround time for an already complex determina-
tion. By comparing spectral data collected using a diode
array detector the confirmation can be accomplished with a
single analytical run.

    Spectra from the UV range of 200-400 nm are collected
for each peak within the retention windows determined by a
standard curve run prior to sample analysis. These spectra
are stored electronically for comparison to the standard
curve spectra in the method library. Peaks of interest may
also be checked for peak purity by comparing spectra from
the leading edge, apex, and downslope of each peak.  Criteria
for confirmation are established by comparing spectral
integrity over the range of the standards run for calibra-
tion.

     This technique presents a three-dimensional approach to
analyte determination and reduces the total time of  analy-
sis. Some of the remaining problems associated with a second
column confirmation, such as nonresolution of all the
analytes and interferences which do not appear in the
primary analysis, can also be eliminated or reduced.
                             482

-------
65
     Tom Williq,  Chemist and Jon S.  Kauffman,  Ph.D.,  Group Leader GC/MS
     Semivolatiles,  Lancaster Laboratories,  Inc.,  2425 New Holland Pike,
     Lancaster,  PA 17601-5994
     ABSTRACT
     Gel Permeation Chromatography (GPC)  is a size exclusion type of
     chromatography which efficiently removes high boiling materials that may
     interfere with the analysis of target compounds in semivolatile extracts.
     Currently, the use of GPC to clean up environmental soil samples is
     required under the USEPA Contract Laboratory Program (CLP)  Statement of
     Work.  The purpose of this study was to evaluate the effects of GPC clean
     up on GC/MS performance in order to justify the cleanup of all samples with
     difficult matrices by GPC prior to analysis.  Potting soil was chosen as
     the subject matrix due to the high level of organic interferences found in
     this matrix.  A sample was split into two portions both of which were
     extracted by USEPA SW-846 Method 3550.  One extract was then cleaned up
     using GPC.  Each extract was then injected four times on a Hewlett-Packard
     GC/MS.  In order to monitor the performance of the GC/MS system, a standard
     containing all of the semivolatile Priority Pollutant List compounds at a
     concentration of 50 ppm was injected before and after each sample.  The
     degradation of the system was monitored by measuring the decrease in
     response factors against the number of injections of the potting soil
     extracts.  Results were compared for the GPC extract versus the non-GPC
     extract.  The response factors dropped drastically for five acid compounds:
     2,4-dinitrophenol, 4-nitrophenol, pentachlorophenol,  2,4,6-trichlorophenol
     and 4,6-dinitro-2-methylphenol after only one or two injections of the non-
     GPC extract.  However, the response factors for these compounds were much
     more  stable even after four injections of the GPC extract.   A similar trend
     was observed for the internal standards, which are used to monitor system
     performance.  This loss of system sensitivity after injecting the non-GPC
     extract can be explained by an increase of active sites in the injector
     port  caused by the buildup of high molecular weight compounds in the
     injector liner and on the stainless steel seal.  Although the experiment is
     very  simple, the results make the utility of GPC cleanup evident and
     therefore justify the extra time and effort spent in adding this step to
     semivolatile sample preparation.
                                  483

-------
66
               EVALUATION OF THE ENSYS PAH-RISc  TEST KIT

R. Paul Swift. Ph.D.. Senior Chemist, John R. Leavell, Chemist, Chris W.
Brandenburg,  Chemist,  ESAT  Region 10, ICF Kaiser  Engineers,  Inc., 7411
Beach Drive East, Port Orchard, Washington, 98366

ABSTRACT

A  polynuclear aromatic  hydrocarbon  field  screening  method,  utilizing
enzyme-linked immunosorbent assay techniques, was recently evaluated.  The
validation study was conducted to  evaluate the performance and assess the
utility of this method for use  in  field screening activities at hazardous
waste sites.  Design criteria included analysis of field samples, analysis
of well-characterized  reference materials, test kit response to various
soil types  (including analyte-spiked soils), and  a general performance
evaluation using selected soil types with varying PAH concentrations.

INTRODUCTION

The authors  recently conducted a validation study  for an Enzyme-Linked
Immunosorbent Assay  (ELISA) field-screening test method,  PAH-RISc*.  This
semiquantitative chromogenic analytical method, developed by Ensys, Inc.,
Research Triangle Park, NC,  was evaluated for use in the determination of
Polynuclear  Aromatic  Hydrocarbons  (PAHs)   in  environmental  soil  and
sediment samples.  The experimental design and interpretation of results
obtained are described below.

BACKGROUND

Field Screening

Field screening  for contaminants  has applications  in  a  wide  variety of
situations  (1), from collecting real-time data relating to worker safety
to monitoring plume boundaries  resulting from materials spills.  Low cost,
rapid-turnaround sample  analysis  is  beneficial to  site characterization
and assessment,  as  well.   Sampling  locations of possible  interest  and
local "hot-spots" may be identified quickly,   aiding in the selection of
samples to be collected for subsequent CLP analysis.  Other considerations
such as remediation efforts and effluent compliance  may also be monitored.

The need for  rapid, reliable, semiquantitative  methods for  measuring
environmental contaminants has resulted in the introduction and develop-
ment  of new technologies.    In  addition,   laboratory-based  analytical
methods such as GC/MS and HPLC, among others,  are finding  increased use in
field applications  (2).

One novel field  screening method  gaining regulatory acceptance utilizes
immunoassay-based testing (3).   Immunoassay techniques  have been used for
more than 15 years in medical and clinical settings.  This technology has
received Agency  approval  for  a  few classes  of  analytes  (4),  and  is
currently being applied to the  detection of PCBs, petroleum hydrocarbons,
and pentachlorophenol.  One of the earlier environmental applications in
which this  technology was used was  in  the detection  of pesticides  and
pesticide residues  (5).

Enzyme-Linked Immunosorbent Assays

Immunochemical  analysis  is   a   well-established   clinical  diagnostic
technique which is gaining wider acceptance in other areas.  In the field
of environmental  analysis, antibodies have been or are  being developed for
                                         484

-------
a number of contaminants.   All  these  antibodies are of animal origin, but
advances  in monoclonal  preparation  have  yielded in-vitro  propagation
methods.

In order to obtain antibodies of the desired specificity, an immunological
response  must  first  be  generated in  the host.  Typically,   the target
analyte,  or hapten,  is  derivatized  to  produce  a  "handle"  for  use in
attachment  to  a  macromolecule.   Historically,  proteins such  as bovine
albumin have been used  as  the  carrier.  This step  is  essential because
direct injection of the analyte,  due to  its  small mass (on a biological
scale), would result in detoxification by the animal's liver.   In order to
be recognized as  an  antigen, or immunogen, by  the  immune system, a foreign
substance must have a mass which  is  large  on a molecular  scale (>10,000
Da).   When  injected  into  the  host  animal  these  immunogens  cause the
animal's immune system to generate antibodies in  response to the foreign
substance.   The  hapten  is  then bound to an  enzyme,  such  as  horseradish
peroxidase, to form  an enzyme-conjugate.  The conjugated compound is used
as the chromogenic reagent in this test.

After the antibodies are extracted from the host,  they are sorted based on
sensitivity and specificity to  the hapten.  In the polyclonal method, the
above steps are repeated until a desired quantity  of antibody is obtained.
The test kit used in this study utilizes  a monoclonal preparation of the
desired antibodies using hybridoma technology,  in which antibody-producing
(spleen) cells are fused with myeloma cells.  The resulting progeny cells
are able to produce relatively large  amounts  of  the  specified antibody.

In  a  typical test kit,  antibodies of  the desired specificity  are then
immobilized on a solid substrate such as  a small  test tube.  A predeter-
mined amount of sample,  possibly containing analyte,  is extracted with a
suitable  solvent.   An aliquot  of extract and a  fixed  amount  of enzyme-
conjugate  are  added  to  the antibody  tube and allowed to  incubate for a
given  period  of  time.    Competition  between  the  analyte and  enzyme-
conjugate  for  a  limited  number  of  antibody binding  sites  results  in
binding of  the haptens  in  proportion to their relative  concentrations.
Unreacted  haptens  are washed  from the tube  and two color  development
reagents  added.   These  reagents react with  the  bound  enzyme-conjugate,
producing a depth of color  proportional to the amount of enzyme-conjugate
bound.  The "intensity"  of color, compared photometrically to a calibra-
tion solution containing analyte,  is  inversely proportional to the amount
of analyte present in the original sample.

This test method  utilizes ten-fold serial  dilutions of the solvent extract
for comparison with a 1 ppm calibration  standard  prepared in  the same
manner as the  samples.  A comparative photometer  measures  the difference
in absorbance between the sample and the standard, and uses the difference
measurement rather than the parametric value  for  quantitation.

EXPERIMENTAL DESIGN

Four  main design parameters were selected  for  this validation  study,
including analyses of 30 samples for  comparison with CLP-generated data
analysis of  four  "worst-case"  Superfund-class reference  samples,  false-
positive  and  then false-negative  (1  ppm  spike  level)  reaction  to soil
types, and a general reliability/performance  test using a given soil type
spiked with  varying amounts  of analyte.    Given  a  fixed number  of test
kits,   the  key design  criterium  was  to  conduct  a  comprehensive  yet
conclusive set of experiments.
                                   485

-------
Polynuclear aromatic hydrocarbons are  a class of compounds which have one
or more aromatic rings,  generally in conjugation.  The majority of common
PAHs contain three, four or five  rings, may have  alkyl-  or  aryl- groups
bound  to  them,  and  the  rings  themselves may be  heterocyclic.    One
important three-ringed  PAH is phenanthrene, because  it  has a  physical
structure which resembles the skeleton of many other PAH  compounds.

        Table I  PAH RISc" Soil Test Sensitivity to PAH Compounds
Number of Rings
2 rings

3 rings





4 rings




5 rings




6 rings

PAH Compound
Naphthalene

Acenaphthene
Acenaphthylene
Phenanthrene
Anthracene
Fluorene

Benzo[a]anthracene
Chrysene
Fluoranthene
Pyrene

Benzo[jb]fluoranthene
Benzo[£]fluoranthene
Benzo|a]pyrene
Dibenzo[a,/)]anthracene

lndeno[/,2,3-c200

11
>200
* Samples with stated concentration will give positive result greater than 95% of the time when
tested at stated concentration level.
The  antibody  used by  this test  kit  was developed  to  target,  or  key,
phenanthrene.  According  to the manufacturer,  the kit  is  designed  such
that concentration of phenanthrene necessary to result in a positive test
is 1 ppm  (6).  With  the exception of  anthracene,  which  is  detectable at
0.81 ppm, the test is  less sensitive  to  other  PAHs.   The concentrations
necessary to produce a positive result range from  1.2 ppm for chrysene to
200 ppm for naphthalene.   A list of several PAH compounds (SW-846  Method
8310 Target List) and their respective detection limits has been provided
by the vendor  (Table I).   The reported  sensitivities have been used as
                                   486

-------
scaling factors in some sections  of this report.  For example, to produce
an equivalent response using naphthalene rather than phenanthrene as the
target analyte, 200-times more  naphthalene  (on a mass basis) is required.
This relationship will be further investigated in a later section.

Phase I:    Analysis of Field Samples

Several issues were  raised  during the initial design of  the validation
study, the most obvious of these being the evaluation of the test kit with
respect to the manufacturer's claim:

      This method correctly  identifies 95% of samples that are PAH-
      free and those containing I ppm or 10 ppm of PAHs.   A sample
      that develops less color than the  standard is interpreted as
      positive.   It  contains PAHs.   A  sample that develops more
      color  than  the  standard is interpreted  as  negative.   It
      contains less than 1 ppm or 10  ppm PAHs.

To substantiate this claim,  30 samples were selected from the Manchester
Environmental Laboratory  (MEL)  soils storage facility.  These samples were
analyzed previously using either  GC/MS (tentatively identified compounds
excluded but noted) or HPLC  methodology; total  PAH  values (16 analytes)
ranged  from  zero  (non-detect)  to >182  ppm.   Soil  types  ranged  from
weathered sand to dark loamy humus.

Phase II:   Analysis of Reference Samples

Four well-characterized reference samples were selected for  use  in this
study.   These  samples have  been  independently  analyzed  by  a  number  of
laboratories  and  are  representative of  materials present  at  various
preremediated Superfund sites.  The matrices range from marine sediment to
composited hazardous  waste.   All samples  contain a variety  of contami-
nants,  including   pesticides,  PCBs,   and   in one  case,   tributyl  tin.
Clearly, potential interferents are present in these  samples.  A scheme in
which three analysts conduct the analyses, with one analyst performing the
tests in duplicate, provides an empirical  estimate  of  both precision and
accuracy.

Phase III:  Reaction to Soil Types

Soil compositions  vary greatly on a  regional basis.  Extractable humic
components, among  many others, have the potential to negatively affect the
performance of the test kit.   Analysis of  soils of  various compositions
would provide information  necessary to estimate the effect  of soil type on
test kit reliability.  To that end,  11  samples  collected  from different
soil horizons  throughout  Washington  state  by the US Geological  Service
were analyzed.  The PAH content  of each of  these library  soils has been
previously determined to  be  below the detection limit of  the  test kit.
Therefore, any positive test result(s) could be attributed to what have
been  termed   "relatively  unremarkable"  contaminants,   which  possess
structural features similar to those  of  phenanthrene (7).

PAH compounds adsorbed onto soil  surfaces  may resist methanol extraction
and the degree of adsorption may  be dependent on the nature of the soil
In order to  investigate  this effect,  a  subset of the library  soils was
selected and  the  soils fortified, or spiked,  with 1 ppm phenanthrene
This  quality control  measure  was used  to  qualitatively estimate  the
distribution of phenanthrene between the  soil  matrix  and the methanol
extract.
                                   487

-------
Phase IV:   Reliability/Performance Evaluation

A set of samples  consisting  of  weathered,  sandy soil was composited for
use in  this phase of  the validation study.   Each of  the  samples were
previously  confirmed  to be  PAH-free.   Eight spiking  levels  were used,
ranging from  0.1  to  10.0 ppm normanlized  to the phenanthrene response.
For the spike, a three-component mixture containing a three,  four and five
ringed  PAH  - phenanthrene,  pyrene  and  benzo[.fc]fluoranthene -   was
prepared.  The relative amount of each component was scaled based on the
sensitivity of the test kit to the individual components.

The objective  of  this  set of measurements  was to provide an estimate of
error with respect to false positive  results  below the detection limit as
well as false negative results for samples with PAH concentrations above
the detection limit.

It was also hoped that these measurements would provide insight relating
to test kit response for mixtures of PAHs.   Because  the assay utilizes
competitive binding of the antibody-antigen for  detection of the analyte,
we were interested in  determining whether  the presence of  one PAH would
affect the  sensitivity of the test to another  PAH.   Rather than  having
only one  PAH competing with  the  enzyme conjugate  for  antibody  binding
sites,  there  could be two or more competitors.   There is no a  priori
reason to believe that the test  response may  be  calculated using a simple
weighted average concentration of  the individual PAHs and the stated test
sensitivities to each.

RESULTS AND DISCUSSION

This experiment was  conducted under  relatively  ideal  conditions.   Three
experienced chemists  performed  several  trial analyses  in order to gain
familiarity with the kit.  Work was performed at the laboratory benchtop
in  a  controlled  environment  (temperature,  humidity,  etc.).    Quality
control/quality assurance, as specified  through personal communications
and training  by the manufacturer,  included analysis of method  blanks,
regular  pipet  delivery  calibration  verification,  and  rejection  of
analytical  sequences  in which  the  relative  absorbance  of  replicate
calibration  standards  varied by  more than  0.2 absorbance  units.   The
analytical  instructions  supplied  with  the  kit  and  QA/QC  guidelines
provided by the manufacturer were followed without exception.

The laboratory results reported  for the GC/MS and HPLC data were generated
in accordance with full QA/QC requirements  following USEPA CLP and SW-846
protocols.  All  data reported have been validated.   To maintain  client
confidentiality,  any  unnecessary site-specific references  have  been
omitted.

One  set of  measurements was  made   to  evaluate  the  response   of  the
comparative photometer.   A  series of five  spiked-blank solutions were
prepared in duplicate,  with concentrations of 0.5, 0.75,  1.0, 1.25  and 1.5
ppm phenanthrene.    Within each  set,  the  1.0  ppm was  treated  as  the
calibration standard.   Aliquots of  these  solutions were withdrawn  and
processed as  if  they were actual  soil extract.  The  absorbance of each
solution was measured and  recorded.  The resulting "calibration" curve was
determined through a  linear least-squares fit of  absorbance vs. concentra-
tion.   The results are illustrated in Figure I.

From the data,  it appears as though the  instrument  response  is  linear
around the zero (difference)  absorbance.  However, the  slope  is relatively
small -  therefore variations in  absorbance affect the  apparent  sample
                                   488

-------
  Figure 1  Instrument Response
concentrations.
cautiously.
Instrument  readings  near  zero  must  be  interpreted
Phase I: Thirty  soil  samples  were  selected for ELISA-PAH analysis.   No
attempt was made to target the one  or ten  ppm concentration  levels; this
portion of the study was designed to emulate  characterization of samples
in the field.  These samples were previously  extracted (EPA  Method 3540)
and  characterized  for  base/neutral/acid   extractable  compounds  (BNAs,
including PAHs;  EPA  Method 625),  volatile organic  analytes  (VOAs;  EPA
Method  624),  metals   (ICP-AES,   EPA-CLP  Method  200.8),   and  pest-
icides/polychlorinated biphenyls  (pest/PCBs; GC/ECD,  EPA Method 8080).  In
addition to  quantifying  total PAH  content and  relative  amounts  of  the
individual compounds, the supporting analyses were  used  to  characterize
the soil matrices and identify potential interferences.   The results are
shown in Tables II  and III.

In Table II,  the  total  PAH concentration for each sample is compared with
the  ELISA  results.    False-positive  and  false-negative  results  are
indicated by  "+" or  "-",  respectively.   Results from GC/MS  analysis of
sample  2361  indicated  the   presence  methyl-phenanthrenes  at  3  ppm.
Although the  associated  data  were  qualified  "NJ" to  indicate  "there is
evidence that the analyte is present,  the  associated numerical result is
an estimate", the levels apparently present would be sufficient to yield
positive identification  at  1  ppm.    This  result slightly increases  the
accuracy rates (indicated in parenthesis).   The estimated accuracy for the
ELISA test,  based analysis of the thirty samples,  was 86.7% (90%) at 1 ppm
and 73.3% at  10 ppm.  The frequency  of  false-positive results  at  1 ppm was
13.3% (10%), with no false-negatives.  The frequency  of  false- positive
results  at  10 ppm  was  20%,  and  false-negatives,  6.7%.    One sample,
selected randomly,  was analyzed in  duplicate.
                                   489

-------
Table II  Analysis of Field Samples / Total PAH Content

Sample ID
4005
4006
4007
4010
4659
4300
4301
4302
4303
4640
2107
2358
2358D
2359
2360
2361
2362
2363
2364
2365
2366
2368
2369
2370
2371
2372
2373
2374
1 ppm Test
<1



ft
*





ft



*

*



*




ft

•ft
>1
*




ft


ft
ft





ft

ft
ft
*



ft
ft



10 ppm test
<10














*

*

#
»
*



«
*
*
«
>10
ft
*
ft


ft
ft
ft
ft
ft

ft
-
«



»



#
*
*





Lab Result (ppm)
0.2
12.2
16.0
0.0
0.5
8.7
147.7
182.3
4.4
0.2
0.0
85.4
85.4
28.5
0.3
0.6
0.0
1.8
3.4
6.7
0.9
43.2
72.8
1.3
0.3
0.4
27.9
0.0
False +/-
Evaluation
@ 1 ppm
+








+





+ '








+



Evaluation
@ 1 0 ppm
+




+


+
+







+





+




                          490

-------
   Table II  Analysis of Field Samples / Total PAH Content

Sample ID
2375
2376
2377
1 ppm Test
<1

*

>1


•
10 ppm test
<10
•
•
•
>10




Lab Result (ppm)
16.4
0.4
9.5
False
Evaluation
@ 1 ppm



+ /-
Evaluation
@ 1 0 ppm



' Tentatively identified compounds list indicates methyl-phenanthrenes present at 3 ppm
Table III Analysis of Field Samples / Normalized PAH Content

Sample ID
4005
4006
4O07
4010
4659
4300
4301
4302
4303
4640
2107
2358
2358D
2359
2360
2361
2362
2363
2364
2365
1 ppm Test
<1



•
•





•



•

•



>1
•




•


•
•





•

•
•
•
10 ppm test
<10














•

•

•
*
>10
•
•
•


•
•
•
•
•

•
•
•



•



Normalized Lab
Result (ppm)
0.1
8.1
9.0
0.0
0.2
5.2
56.9
73.2
0.1
0.0
0.0
47.3
47.3
11.5
0.2
0.5
0.0
1.2
1.7
3.6
False +/-
Evaluation
@ 1 ppm
+







+
+





+ '




Evaluation
@ 10 ppm
+
+
+


+


+
+







+


                            491

-------
       Table III  Analysis of Field Samples / Normalized PAH Content

Sample ID
2366
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
1 ppm Test
<1
*




*

*

*

>1



*
*





*
1 0 ppm test
<10
«



*
*
«
«
*
«
»
>10

*
•
tt








Normalized Lab
Result (ppm)
0.6
27.5
49.2
0.8
0.1
0.2
13.5
0.0
6.4
0.2
2.8
False +/-
Evaluation
@ 1 ppm



+
+






Evaluation
@ 10 ppm



+


-




' Tentatively identified compounds list indicates methyl-phenanthrenes present at 3 ppm
Table III shows the comparison of normalized PAH concentrations with the
ELISA results.   As  mentioned previously, the  sensitivity of  the  test
varies with different PAH compounds.  In order to account for this effect,
each PAH concentration was normalized by dividing its actual value by that
concentration required to produce a positive result at 1 ppm (phenanthrene
= 1).   This was done  to minimize  any bias introduced into the test by
anti-PAH  specificity.   A critical assumption is  that  the manufacturer-
provided  list of  PAH sensitivities are valid in  the presence  of  two or
more PAH  compounds.  Sensitivity to the methyl-phenanthrenes tentatively
identified in sample 2361 is not known, but was assumed to be similar to
that of phenanthrene.  As in the previous data set, estimated accuracies
including any contribution from the methyl-phenanthrenes are enclosed in
parenthesis.  The estimated accuracy was 80% (83.3%) at 1 ppm and 70% at
10 ppm.   The frequency  of false-positive results at 1 ppm was 20% (16.7%),
with no  false-negatives.   At 10  ppm,  the frequency  of  false-positive
results was 26.6%, and false-negatives, 3.3%.

The  results using  total and  normalized  PAH  concentration values  are
similar.  If samples  with PAH concentrations exclusively near 1  and 10 ppm
had  been used,  a  greater discrepancy  would have  been  observed.   The
majority of these samples contained levels of PAHs  separated from either
test kit detection  level.    This demonstrates  the need  to  apply  the
screening  test  judiciously,  with operator  intervention/interpretation
applied to those samples  near either limit.

Phase II:  Four  soil  reference samples were  analyzed by  three different
analysts, with one analysis performed in duplicate  ("D" in Table IV), to
illustrate the  variation introduced by different  analysts  and also the
duplicate precision.   Comparison of the data  is  shown in Table  IV.  There
was considerable variation in the  results,  apparently due to the presence
                                   492

-------
of interferents,  including  compounds to which  the ELISA test  exhibits
cross-reactivity.   The summary  in Table IV lists the total and normalized
PAH concentrations.

These samples were selected  because they  present a real challenge to the
ELISA test.  They  are  each representative of the types of materials found
at Superfund  sites.    Prior to analysis,  the  samples  were sieved  and
homogenized.  Non-representative material was  removed,  but an effort was
made to preserve the physical character of  the original sample.   A brief
summary of each soil follows.

Sequim Bay: This matrix is typical  of marine or  lower-wetlands  sediment,
      with  corresponding levels of marine  salts.   The consistency  was
      similar to  water-saturated  silt.   Although the concentration  of
      various contaminants were each less than  1  ppm,  a large  number of
      different compounds were  present.  This particular reference sample
      has been independently characterized as  many as 50 different times.
      Some of the  compounds  present were:

Phenol - 0.24 ppm; 4-methyl  phenol - 0.24 ppm;  pentachlorophenol - 0.42
ppm;  biphenyl  -  0.79  ppm;   methylnaphthalenes - 0.25 ppm; tributyl  tin
chloride - 0.11 ppm; halophenyl phenyl ethers - 0.4 ppm;  tetrachloroguicol
- 0.47 ppm

                    Table IV Analysis of Reference Samples

Sample ID
Sequim Bay
- marine sediment
- low levels of many
pesticides, chloro-
benzenes, phthalates
H.I. BSRM #XX
Composite soil
PCBs: 130 ppm
Lead: 5000 ppm
- contains slag/ash
RTC Sample #XXX
- PCP: 2190 ppm
- substituted
naphthalenes:
2300 ppm
Soils Bldg 0003
Isophorone: 4.4ppm
(dimethyl hexanone)
Trial
1
2
3
3D
1
2
3
3D
1
2
3
3D
1
2
3
3D
1 ppm test
<1
















>1

•
•
•

I
•
•

•
•
•

•
•
•
10 ppm test
<10
•



•
•



I


•
•


>10

*
•
•


•
•
"

*
•


•
•
100 ppm test
<100
•

•

•
*



•


«
•


>100

•

*


«
•
•

•
tt


•
•
GC/MS Results
Lab Result
(ppm)
1.73
1.73
1.73
1.73
18.3
18.3
18.3
18.3
8611
8611
8611
8611
24.8
24.8
24.8
24.8
Normalized Lab
Result (ppm)
<0.1
<0.1
<0.1
<0.1
12.1
12.1
12.1
12.1
5764
5764
5764
5764
12.4
12.4
12.4
12.4
                                   493

-------
      As indicated in Table IV, the actual PAH concentration was  less than
      2 ppm.   When  normalized based  on specificity to  individual PAH
      compounds, the  apparent  PAH concentration was  less  than  0.1 ppm.
      Three of four ELISA  analyses  indicated that the PAH concentration
      was greater than 10 ppm,  and two  tests  yielded results greater than
      100 ppm.  Clearly,  materials were present which  respond to the test
      in a similar manner as do PAHs, and exhibit a strong interference.
      Unfortunately,  no  clear pattern  was  apparent  in  the  spurious
      results.   The test did  appear to yield either correct  or false-
      positive results at the 1 ppm  level, depending  on which set of lab
      results were considered.

HI  BSRM #XX:  This bulk  site  reference  material composite sample was
      manufactured from soils collected at a Superfund site by the USEPA
      in  Region  10  and  known  to contain  PAHs  and  PCBs.   The  matrix
      consisted largely of marine sediments  and silt,  although there is a
      significant amount of ash, slag and related  fallout originating from
      the operation of a  secondary smelter for many years.  Other man-made
      debris and residue include cinders,  brick material and sandblasting
      waste from ship refinishing.  Local "hot-spots"  resulting from wood
      and  pole treating and  indiscriminate dumping  of  waste  oils  and
      electrical transformer/capacitor dielectric contributed  PCBs  and
      PAHs to the BSRM.

      The  reference  sample  exhibited a  fairly   high degree of  matrix
      inhomogeneity on  both the  macroscopic and microscopic level.   A
      concerted  effort   was made to  homogenize  the sample  prior  to
      analysis.  The  variability  is  reflected, in part,  by the  standard
      deviations associated with the analytical results:

Lead - 5480 ± 2620 ppm;  Arsenic - 37.8 ± 7.3 ppm; Total PAHs (7  carcino-
genic PAHs  identified in MTCA)  - 18.3 ± 4.6 ppm;  Total PCBs -  127 ± 39
ppm.

      A comparison of the sample results is  shown in Table IV.  All ELISA
      analyses at the 1 ppm level correlate with the laboratory data.  The
      first two trials produced identical results at 10 and 100 ppm (true-
      negatives), but disagree  with  the duplicate analyses performed by
      the  third  analyst.   The  duplicate  analyses were self-consistent.
      Because the duplicate analyses were performed using the same soil-
      extract, sample inhomogeneity  rather than ELISA precision may have
      contributed to  the disparity.

RTC Sample  #XXX:  This sample was prepared under  contract  to the  EPA as
      part  of  a RCRA study and made  available  for  use in  this  study.
      Laboratory analyses  were performed for both total metals and BNA
      compounds.  The levels of sodium, potassium, nickel and total PAHs
      were relatively high.  Some of the analytes present at significant
      levels were:

Chromium -  16100 ppm; Phenanthrene  - 2430 ppm;  fluoranthene -  1840 ppm;
pentachlorophenol - 2190 ppm; methyl-naphthalenes - 200 ppm; carbazole -
81 ppm;

      Of the  reference  samples, this was the most highly contaminated.
      All ELISA  results  indicated PAHs present at 1  ppm.   One  analysis
      yielded false-negative results at 10 and 100 ppm.  This analysis is
      clearly  in error; with  the  levels of contaminants  known  to be
      present,  it  seems  unlikely  that normal  experimental or  method
      uncertainties were the cause.   Since the 10  and  100 ppm results were
                                   494

-------
      generated using serial dilution  of  the 1 ppm extract,  we believe
      their inaccuracy may be attributed to dilution error.

Soils Building 003:  This  sample  was also a  composite  soil  manufactured
      from materials  which  had collected in  a  soils laboratory  over a
      period of  time.   A  large  number of  assorted samples  of varying
      origin were  blended,  with aliquots  withdrawn for waste  disposal
      characterization.   Some of  the material was  dated and  all had been
      stored at room temperature.   Microbial and oxidative decomposition
      were expected to have yielded a variety of degradation products of
      the PAHs and other contaminants.  Subsequent  analysis showed that
      with the exception of 4.4 ppm  isophorone  (dimethyl hexanone),  the
      composition of  contaminants  was due  predominately  to  PAHs,  with
      total and normalized concentrations of 24.8 and  12.4  ppm, respec-
      tively.   The  matrix of the  composite soil  sample included clay,
      silt, sand and loam.

      All ELISA results indicated PAHs present at the  1 ppm level.  Two
      replicates yielded  false-negative  results at  10 ppm.   Duplicate
      analysis yielded  correct results at  10 ppm,  but false  positive
      results at 100 ppm.   No simple explanation can be offered for this
      disparity.

Phase III:  Eleven soil samples collected from various soil horizons across
Washington state were analyzed using  the  ELISA  method.  The  soils were
previously determined to contain undetectable amounts  of PAH  compounds.
The purpose of  this phase of the experiment was to determine the effect of
soil type with respect to false-positive ELISA results.

Analysis of these soils yielded negative results in each case.   On this
basis, soil type did not appear to  negatively impact the accuracy of the
test with respect to false-positive results.

Four soils from this  set were spiked with PAH to provide an estimate of
false-negative  results  with  respect  to  soil  type.   These soils  were
primarily clay/silt with a  fair amount of humic or loamy material.  A 10-
gram aliquot of each soil was spiked directly with  phenanthrene to give a
final concentration of 1 ppm,  allowed to weather at  room temperature for
24 hours,  extracted,  and analyzed in duplicate.  This method was preferred
to that in which  the extract is spiked because adsorption of  phenanthrene
on the soil surface is a potential  physical  interference.  Resistance to
methanol-extraction would result in a negative bias and, therefore, false-
negative results.   Phenanthrene was selected for use as the  spike on the
basis of ELISA sensitivity, and to  reduce any possible cross-reactivity
questions which may arise  from the use of a mixed-PAH spike.   The results
are shown in Table V.

Results from  the  spiking study  show a high  frequency (75%)  of  false-
negative results,  indicative of the inability of the methanol extraction
fluid to completely liberate the  phenanthrene from the matrix.  Although
it is reasonable to expect that the methanol-based extraction is not 100%
efficient,  these  results  yield  an  unacceptable  percentage  of  false-
negative results.   The bias built into the ELISA test did not appear to be
able  to  provide a  large enough  margin of uncertainty in this  case.
Because a CLP-quality extraction and analysis was not  performed,  it was
not possible to estimate the actual  efficiency of the extraction usinq the
ELISA method.

These  limitations  should  be compared  with  those  generally encountered
during  laboratory-based  analysis.    Standard  CLP  soxhlet  extraction
                                   495

-------
involves refluxing the solid sample for a minimum of 18 hours in boiling
solvent.  Even then,  extraction  efficiency  may only approach 80% or so.
Matrix spike analysis generally involves adding the  spike compound to the
extraction  solvent  rather  than  to  the soil  itself.   Obviously,  no
weathering of the PAH/soil occurs.  However,  this does not change the fact
that the kit was unable to detect PAH at 1 ppm in soil.

                      Table V  Reaction To Soil Types

Sample ID
8114
8123
8129
8136
8142
8500
8500 S
8501
8501 S
8506
8512
8512 S
8513
8513 S
8107
1 ppm test
<1 ppm
*
*
*
*
*
*
*
*
«
*
*

«
*
*
>1 ppm











#



10 ppm test
<10 ppm
•
•
*
*
«
*
*
«
*
*
*
«
*
«
w
>10 ppm















Spike result
phenanthrene
1 ppm
n/a
n/a
n/a
n/a
n/a
n/a
-
n/a
-
n/a
n/a

n/a
-
n/a

Soil description
yellow brn sandy loam
pale brn silty loam
brn silty loam
pale brn sand
pale brn sand
drk brn silt loam
-
drk brn gry sandy loam
»
lake sediments
It brn coarse loam sand
••
drk brn stony silt loam
»
It grey silty clay
Two  papers previously published  by the manufacturer  described  spiking
studies performed  during validation of their  PCB  and pentachlorophenol
immunoassay methods  (8-9).   In each case the  extracts,  rather  than the
native  soils,  were  spiked.   A  similar  study reported  poor  extraction
efficiency  for PCBs  spiked  directly onto  the soil  (10).   We  felt  it
important and more procedurally valid to spike  and then weather the soil.

Phase IV: Another spiking study was performed to estimate reliability of
the ELISA method as a function of analyte concentration.  Sandy, weathered
soil previously determined to  be  PAH-free was  divided into 57 - 10 gram
aliquots, spiked directly, and allowed to interact for one hour.

A spiking  solution  was prepared  using a  three,  four and  five  ring PAH
compound - phenanthrene  (3), pyrene (4) and benzofk]  fluoranthene (5)  -
such that their relative, normalized contributions  on  a mass basis to the
spike cocktail were  equivalent.   Using the  1 ppm spike as an example,  a
solution was prepared such that 0.33 ppm of phenanthrene would  be added to
the  soil.   The same contributions were desired  from  the four and five-
                                    496

-------
ringed compounds, but the test is less sensitive to them.  Since  3.5
of pyrene is required to produce the same  response  as  1 ppm phenanthrene,
3.5 times more pyrene was used  (3.5x0.33).   Similarly,  the test is  9.4
times less sensitive to the benzo-compound, so 9.4x0.33 ppm benzo[k]fluor-
anthene was used.  Calculation shows that the total PAH concentration in
the spiked soil  was  actually  4.6 ppm.  A  similar scheme was used in  the
preparation of the soils at the other spiking levels.  With the exception
of  the  blank  sample,  which  was  analyzed  once   in duplicate,  seven
replicates were analyzed in duplicate at each of  0.1,  0.5, 0.8, 1.0, 1.5,
2.0,  5.0, and 10.0 ppm (normalized)  concentrations.   Seven extractions
were performed at each level, and each extract analyzed in duplicate.   The
results for normalized and  total PAH concentrations are shown in Tables
VI(a) and VI(b), respectively;  agreement of  estimated  error rates  for  the
different spike  levels is coincidental.

          Table Vl(a)  Reliability Test  Normalized Concentrations
Normalized Value
(ppm)
Estimated Rate of
False Positives (%)
Estimated Rate of
False Negatives (%)
0.0

0,0



0.1

0.0



0.5

78.5



0.8

78.5



1.0



21.5

1.5



0.0

2.0



0.0

5.0



0.0

10.0



0.0

            Table Vl(b)  Reliability Test  Actual Concentrations
True Value (ppm)
Estimated Rate of
False Positives (%)
Estimated Rate of
False Negatives (%)
0.0
0.0

0.46
0.0

2.32

21.5
3.71

21.5
4.63

21.5
6.9

0.0
9.3

0.0
23.2

0.0
46.3

0.0
Three  false-positive results were  observed  at  both the 0.5  and 0.8 ppm
levels.  The 0.5 ppm trial was repeated  (extraction and analyses), and the
results were identical  to those obtained in the  first trial.  Three false-
negative  results  were  observed at the  1  ppm level.   No  false-negative
results were seen above 1 ppm.

The total PAH results  showed a high rate  of  false-negatives.   This data
would  seem  to  indicate that normalizing the PAH concentration  based on
sensitivity  of  the anti-PAH to  the  various PAH  compounds  is  somewhat
valid.  Also, the  discrepancy between the  spike recoveries in Phase V and
Phase VI is attributed to the difference in  soil matrices.   More PAH was
extracted from the  sandy  soil  than from  the  loamy soil.    This further
supports the conclusion that adsorption may introduce a low bias into the
test results.

CONCLUSION

It was the intent of this validation study  to investigate and evaluate the
performance of the Ensys PAH-RISc* test kit.   The goal was  to  conduct a
comprehensive set  of experiments  while  working  under  the  constraint of a
predetermined number of available test  kits.
                                   497

-------
This immunoassay test  method  performed favorably,  although not quite as
well,  as claimed  by  the  manufacturer.   It  does  appear  to  have the
potential  to be  used  as  a  field  screening  tool,  provided  that the
limitations of its use be borne in mind.

The primary caveat is that this kit should be used by personnel capable of
proper  interpretion  of  the  results  from a scientific standpoint.   A
negative result at 1 ppm does  not necessarily  indicate that less than 1
ppm  of PAH  contamination  exists.    Relatively  high  concentrations  of
particularly carcinogenic PAHs, such as benzo[a]pyrene, may be present on-
site at levels which are well above 1 ppm but below the levels necessary
to generate a positive test response.   The sensitivity of the test kit to
the various compounds must be borne in mind when interpreting the results.
PAH-extractability may introduce  low-bias at 1 ppm; the implications of
this must be  considered  when  concentrations of PAHs  are  near  the 1 ppm
level.

Also, proper soil sampling  is  a scientific technique.   In order to obtain
accurate, representative results  using this test kit,  the operator should
follow appropriate soil collection methods.

Any samples which generate positive  results, as  well as a percentage of
those which test negative, should be submitted for subsequent confirma-
tional analysis by classical analytical methodology.

ACKNOWLEDGEMENTS

The authors  would  like to thank  the U.S. EPA for  supporting  the study
under the Environmental Service Assistance Teams Zone 2 Contract 68D10135.
We would also like to thank Ensys Inc for  providing  the test kits used in
this study, and for the training rendered in their proper use.   The USGS
soils and their characterization data were provided by Mr. Kenneth Ames,
USGS,  and Mr.  Dickey Huntamer, Washington State Department  of Ecology.
Mr.  Gerald  Muth,  USEPA, provided  initial technical  direction on  the
experimental design.

DISCLAIMER

Although the research described in this manuscript has been funded wholly
or in part by the EPA Contract 68D10135 to ICF Technology Inc., it has not
been subjected to the Agency's review and therefore does not necessarily
reflect  the views  of the  Agency;  no official endorsement  should  be
inferred.

REFERENCES

(1)   H.M. Fribush, J.F. Fisk, "Field Analytical  Methods for Superfund",
      Field Screening Methods for Hazardous Wastes and Toxic Chemicals,
      Proceedings  of  the  Second  International  Symposium,  ICAIR  Life
      Systems, Inc.,  Las Vegas, NV, pp 25-29,  February 12-14,  1991.

(2)   Office of Emergency and Remedial Response, Hazardous Site Evaluation
      Division, Field Screening Methods Catalog,  Users Guide,  EPA/540/2-
      88/005, U.S. EPA, Washington, D.C., July, 1990.

(3)   J.M. Van Emon and R.O.  Mumma, Immunochemical Methods for Environmen-
      tal Analysis, ACS Symposium Series 442,  American Chemical Society,
      Washington,  D.C., 1990.
                                   498

-------
(4)     U.S. EPA Office of Solid Waste,  SW-846  Draft Methods 4010, 4020,
      4030, October, 1992.

(5)    H.W. Newsome  and P.G.  Collins,  Assoc. Off. Anal.  Chem,  70, 1025,
      1987.

(6)    Personal communication, K.R.  Carter,  Ensys, Inc., 1993.

(7)    S.B. Friedman, CHEMTECH,  732,  1992.

(8)    J.p. Mapes,  et.a.Z.,  "Rapid, On-Site  Screening est for Pentachloro-
      phenol in Soil and Water-PENTA RISc*,  Superfund *91 Proceedings of
      the  12th  National  Conference, HMCRI,  Washington,  D.C.,  December
      1991.

(9)    J.p.  Mapes,   et.  al.,   "PCB-RISc* -   An  On-Site  Immunoassay  for
      Detecting PCBs  in  Soil", Superfund  *91  Proceedings  of  the  12th
      National Conference, HMCRI, Washington, D.C.,  December 1991.

(10)   M.  Chamberlik-Cooper, R.E. Carlson, R.O. Harrison, "Determination of
      PCBs by Enzyme Immunoassay", Field Screening Methods for Hazardous
      Wastes and Toxic Chemicals, Proceedings of the Second International
      Symposium,  ICAIR Life  Systems,  Inc.,  Las  Vegas,  NV,  pp  625-628,
      February 12-14,  1991.
                                  499

-------
67   RESULTS OF ANALYTICAL FIELD TRIALS FOR PCBs USING AN
      IMMUNOASSAY TECHNIQUE
      Andrew Parsons and Alan Weiss, Millipore Corporation, Bedford,
      Massachusetts 01730.

           The use of immunoassays (IA) for field screening of environmental
      samples has increased dramatically over the past few years. Their
      utilization in remedial investigation and monitoring has gained widespread
      acceptance due in large part to the efforts of the US EPA to demonstrate
      the technology and begin the process of promulgating methods in which IA
      methodology is incorporated.
           As part of the Superfund Innovative Technology Evaluation (SITE)
      program, the EnviroGard™ PCB Test Kit (Millipore Corp., Bedford, MA)
      was evaluated as a field screening method for identifying soil samples
      containing less than 10 mg/kg PCB. As part of the field screening activities,
      the SITE demonstration contractor (PRC Environmental Management,
      Inc., Kansas City, Kansas) standardized the IA with Aroclor 1242 (the
      major PCB contaminant at the site) and assessed the lA's utility in
      quantitating the level of PCB contamination. The screening results, which
      have been published in the SITE Technical Evaluation Report (TER), were
      consistent with the assay's performance claim of less than 5% false
      negatives (0/96 in the actual demonstration).  To assess the accuracy of the
      EnviroGard PCB test kit, 89 samples that were quantitated using the IA
      with Aroclor 1242 standards, were compared to values obtained by
      GC/ECD analysis (SW846-8081).  The correlation coefficient (r2) for this
      data set was 0.45. Six of the 89 samples, all with high (> 300 mg/kg) PCB
      levels, skewed the data so that when they were removed (i.e., N = 83), r2
      increased to 0.87. However, when only samples with GC/MS values of <
      100 mg/kg were considered, r2 dropped to 0.33. Zero to 100 mg/kg, which
      encompasses most regulatory  contamination limits, is the approximate
      dynamic range of the IA: Quantitation of levels higher than 100 mg/kg
      normally requires sample extract dilution.  The high degree of scatter that
      was reported in the TER might be attributed to the imprecision associated
      with the field screening procedure: Analyses were performed in a non-
      laboratory environment, no reference (e.g., performance evaluation)
      samples were included, all results were based on single determinations, and
      the analyst, although competent, was only marginally trained in IA
      methodology. In order to determine if improving the method precision will
      improve its accuracy, the SITE demonstration samples will be reanalyzed
      under more rigorous, laboratory conditions.
           If no improvement in accuracy is seen, it would suggest that IA
      technology, at least in terms of PCB analysis, might be limited to screening
      applications. On the other hand, if upgrading the analytical data quality
      and running the method under controlled laboratory conditions improve
      the lA's accuracy (i.e., r2 > 0.8) in the 0 - 100 mg/kg range, it may be
      possible to bring many of the advantages of this innovative technology to
      quantitative analyses.
           Data from the reanalysis will be presented and compared to the
      results obtained during the field demonstration.  Based on the reanalysis,
      recommendations with regard to the scope and limitations of PCB
      immunoassay will be made.
                                       500

-------
58  IMPORTANT FACTORS IN ENHANCING SUPERCRITICAL FLUID EXTRACTION
     EFFICIENCIES FOR ENVIRONMENTAL APPLICATIONS

     J.M.  Levy.  L.  A.  Dolata,  R.  M.  Ravey,  Suprex Corporation,  125
     William Pitt Way,  Pittsburgh,  PA  15238  (412)  826-5200

     ABSTRACT - Poster Presentation

     Supercritical fluid extraction (SFE) has  a  broad range of
     applicability,  especially with regards to environmental matrices.
     SFE has achieved a significant amount  of  attention due to  the
     benefits of eliminating toxic,  liquid  solvent usage,  reduction in
     sample preparation time and  an increase in  the overall analytical
     reliability of determinations.   On-line SFE/GC-MS is  a powerful
     technique to accurately analyze and  quantitate environmental
     analytes.  In addition,  the  off-line transfer of SFE  effluents to
     collection vials adds a considerable amount of flexibility in
     characterizing complex matrices since  a full complement of
     analytical tools can be used (i.e.,  GC, LC,  IR,  NMR and UV) .
     Moreover, the advantages of  SFE can  be further augmented by the
     use of automation for greater  sample throughput which can  be
     especially important for environmental applications.

     Examples will  be presented showing the use  of  SFE/GC-MS and FID
     methodologies  for the determination  of different target analytes
     in environmental matrices, such as polynuclear aromatic
     hydrocarbons (PAHs),  total petroleum hydrocarbons and pesticides
     in the soil.   The discussion will also focus on the experimental
     verification of optimized SFE  variables to  achieve efficient and
     quantitative extractions of  the target analytes in environmental
     solids.   An example is shown in Table  1 where  an off-line  SFE/GC
     comparison was  made between  the extraction  of  PAHs from soil at
     different pressures,  indicating that higher pressures were
     necessary for  the complete recovery  of the  PAHs,  especially the
     four  and five  ring PAHs.
                                  501

-------
Table I
Compound

Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)Anthracene
Chrysene
Benzo(b,k)-
Fluoranthene
Benzo(a)Pyrene
EPA Method
8270
Acceptance
Range fppm)

24.2-40.6
14.7-23.5
527-737
414-570
1270-1966
373-471
1060-1500
744-1322
214-290
271-323

130-174
80.1-114.3
Concentration Levels (ppm)
250 atm   350 atm   450 atm
  23
  20
 566
 445
1682
 357
1028
 703
  74
  74
  23
   *
 601
 471
1978
 439
1459
1153
 235
 251

 107
  64
  25
  22
 614
 458
1911
 400
1571
1269
 284
 314

 155
  89
                              502

-------
INORGANICS

-------
69   EVALUATION OF A RAPID STEAM-DISTILLATION PROCEDURE FOR THE
           EXTRACTION OF CYANIDE FROM LIQUID AND SOLID WASTES

      Edward M. Heithmar. U. S. Environmental Protection Agency, P. O. Box 93478, Las
      Vegas, NV 89193-3478; Richard Herman and Michael R. Straka, Perstorp Analytical
      Environmental, 9445 SW Ridder Road, Suite 310, Wilsonville, OR 97070

      Cyanide in waste occurs as free cyanide (CN' or HCN) or in complexes with metal ions.
      Depending on the thermodynamic stability and kinetic lability of the complexed forms,
      they may or may not be toxic to aquatic organisms if leached into surface water.  For
      the  estimation  of  potential  water pollution,  two  operationally  defined  cyanide
      determinations are often performed on wastes.  Total cyanide is usually determined by
      air-purged refluxing of the waste acidified with dilute sulfuric acid in the presence of
      magnesium chloride catalyst (1). The volatilized HCN is trapped in an alkaline scrubber
      solution and determined by  one of a variety of methods. Strongly complexed cyanide
      (e.g., ferricyanide and ferrocyanide) is often determined by a second air-purged reflux
      after oxidation of other cyanide forms with alkaline hypochlorite. The presumed-toxic
      "cyanide amenable  to chlorination" is determined by difference (1).   An  alternate
      estimation of the more toxic cyanide species can be made by the "weak-acid dissociable
      cyanide" procedure, which utilizes air-purged refluxing to liberate HCN from  a sample
      containing zinc and buffered at pH 4.5.  Under these conditions, iron cyanide complexes
      do not generate HCN (2).

      All of the above procedures for the measurement of total cyanide and  the estimation of
      the more toxic cyanide fraction are problematic. In addition to being prone to numerous
      interferences, they all rely on  the air-purged reflux for the liberation of HCN.  The
      reflux requires a minimum of about 90 minutes per sample.  The procedures prescribed
      in SW-846 require a large reflux apparatus that precludes the simultaneous refluxing of
      several samples. None of the procedures have been validated for the  determination of
      cyanide in solid waste. An  alkaline extraction procedure for such samples is  provided
      in SW-846, but recoveries are poor (3).

      A steam-distillation procedure for the liberation of HCN from  samples containing free
      and complexed cyanide will be described in this presentation. The method is rapid (< 15
      minutes per  sample)  and provides good cyanide recovery from  synthetic  samples.
      Statistical analysis of the effect of various experimental parameters will be presented.
      Results for liquid and solid wastes using steam distillation will be compared to standard
      methods of analysis.

      REFERENCES

      (7)    "Total and Amenable Cyanide," Method  9010A, Test Methods for Evaluating
             Solid Wastes  (SW-846), U.S. Environmental Protection Agency, Washington,
                                            503

-------
       D.C., November 1990.

 (2)    "Standard Test Methods for Cyanide in Water:  Test Method C - Weak Acid
       Dissociable Cyanides," Standard Test Method D 2036-91, Annual Book ofASTM
       Standards,  Volume  11.02,  pp.  83-99,  American  Society  for  Testing  and
       Materials, Philadelphia, PA, 1992.

 (J)    "Cyanide Extraction Procedure for Solids and Oils," Method 9011, Test Methods
       for Evaluating Solid Wastes  (SW-846), U.S. Environmental Protection Agency,
       Washington, D.C., November 1990.
Notice:  Although the research described in this presentation was conducted in part by
the U.S. Environmental Protection Agency, it has not been subjected to Agency review.
Therefore, it  does not reflect the views of the Agency.  Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
                                      504

-------
70    ABSTRACT
       A  METHOD  EVALUATION STUDY FOR THE ANALYSIS  FOR  HEXAVALENT
       CHROMIUM  IN  SOLID  SAMPLES  USING A MODIFIED  ALKALINE DIGESTION
       PROCEDURE AND COLORIMETRIC DETERMINATION

       R.J. Vitale. Quality Assurance Specialist/Principal, Environmental Standards, Inc., 1220 Valley
       Forge Road, P.O. Box 911, Valley Forge, PA  19482.

       ABSTRACT

       The difference between the levels of hexavalent chromium (Cr(VI)) and trivalent chromium
       (Cr(m)) in environmental samples is important from a human  health concern  due to the
       significantly greater toxicity of Cr(VI) compared to Cr(m).  Consequently, knowledge of the
       chemical speciation and distribution of these two valence states of chromium in solid matrices
       is essential for adequate characterization and risk assessment of sites containing elevated levels
       of chromium. While current approved analytical methodologies exist to successfully differentiate
       between the 3+ and 6+ valence states in the liquid (aqueous) medium, there  is currently  no
       approved method for the determination of  total Cr(VT) in the solid medium.  An alkaline
       digestion method (Method 3060) appeared in the SW846 2nd Edition as a preparatory method
       for Cr(VI) analysis in soils.  However, subsequent research and evaluation of the method in
       1986  resulted  in  Method  3060 being  omitted from SW846 3rd Edition  due to a lack of
       predictability in environmental samples relating to oxidation of Cr(in) and/or reduction of
       Cr(VI).  More than a thousand field samples and the results of a recently completed method
       evaluation study (MES) have demonstrated significant improvement over Method 3060 (SW846
       2nd Edition). The MES which is the focus of this presentation, was divided into four distinct
       portions with the objective of showing that the modified Method 3060 is predictable and reliable
       within the  constraints  of the  sample types evaluated.    These  include:   (1) sample
       homogenization,  (2) spiking studies using Cr(HI) and Cr(VI), (3) a comparison of the modified
       Method  3060 to Method 3060 and (4) a mass balance study to determine the ultimate fate of
       Cr(VI) spikes with respect to chromium oxidation and/or reduction. Nine different types of solid
       matrices were selected for testing, ranging from quartz sand to chromite ore processing residues.
       All samples were characterized for auxiliary parameters including total organic carbon (TOC),
       total sulfides, pH, oxidation/reduction potential  (ORP), and percent solids.  The results of the
       study indicate that good matrix spike  recoveries  (76-115%) and duplicate precision  were
       achieved  in  non-reducing   samples  using   the  modified  alkaline  digestion  (Method
       3060)/colorimetric technique (Method 7196A).  In contrast, for highly reducing samples (e.g.,
       anoxic sediments), zero percent Cr(VI) matrix spikes were obtained.  For  these samples,
       auxiliary characterization parameters and mass balance results demonstrated that such samples
       are  not  capable of maintaining a  Cr(VI) matrix spike in the 6+ valence state either in the
       laboratory or in the environmental settings from which they were collected.   Spiking studies
       using soluble forms of Cr(IH) indicate that less than  1%  oxidation  of Cr(DI) to  Cr(VI) is
       possible on the freshly precipitated Cr(m) that forms  during the alkaline digestion,  which
       should not hinder an accurate characterization of Cr(VE) in non-reducing solid matrices. A  more
       difficult challenge for analytical chemists who use this method will be to avoid the interpretation
       that poor spike recoveries automatically mean that the method does not yield acceptable results.
                                             505

-------
71     THE ANALYSIS OF REDUCING SOILS FOR HEXAVALENT CHROMIUM

      Don P. Miller. Chemist, U.S. EPA Region VH Laboratory, 25 Funston Road, Kansas
      City, Kansas 66115.
      ABSTRACT

      The Second Edition of  SW-846  contained a  method (3060, 7196)  for the
      determination of hexavalent  chromium  in  soil.   The method  did  not  perform
      adequately for reducing  soils and was dropped from later editions of SW-846. A
      need to analyze soil samples for hexavalent chromium exists; this paper presents a
      method suitable for reducing soils. The Second Edition method worked under two
      conditions: 1) the soil contained hexavalent chromium, or 2) the soil did not contain
      hexavalent chromium and was non-reducing.  The Second Edition method failed the
      matrix spike recovery test if the soil did not contain hexavalent chromium and was
      reducing.  The method given in this presentation uses capillary electrophoresis (CE),
      passes the matrix spike recovery test for reducing and non-reducing soils, and has a
      method detection limit of 4 mg Cr(VI) per kg soil (dry weight  basis).  This CE
      method is  presently used  in our laboratory for  the determination of hexavalent
      chromium in soil.

      INTRODUCTION

      Previously presented methods and subsequent variations for the  determination of
      hexavalent chromium, Cr(VI), in soil (for example, SW-846, Second Edition, Method
      3060, 7196) have given acceptable results if Cr(VI) is present in the soil or the soil
      is non-reducing.  Many of the methods fail a spike recovery quality assurance, QA,
      test if  Cr(VI) is not present in the soil and  the soil is reducing.  Experience with
      actual  samples indicates that  the previous methods failed often  and that a  need
      existed for a new method to determine Cr(VI) in reducing soils.

      This method determines the concentration of Cr(VI) in soil samples.  The range is
      from 0.5 to 500 ppm (mg of Cr(VI) per kg of soil). This method  determines easily
      extracted Cr(VI) from friable soil and other similar solid or semi-solid  matrices, but
      is not acceptable for oily sludge, bricks, and other similar matrices.   This method
      tolerates colored interferences  in the soil and  passes QA tests for soils which contain
      reducing agents.

      The data  quality objectives are:   1)   The  result of the  measurement is the
      concentration of Cr(Vl) in the soil sample expressed in units of mg/kg on a dry
      weight basis.  2) The desired detection limit is 4 mg/kg (dry weight) or better. The
      actual  detection limit based on preliminary measurements for this method appears
                                       506

-------
in the Method Performance section.  3) The desired matrix spike recovery is 75%
or better. The control limits for the matrix spike based on preliminary measurements
appear in the Control Limits section.  4) A frequent use of the result of this method
is to decide whether the concentration of Cr(VI) in a soil sample is equal or greater
than an action level.

Summarizing the method: extract soluble Cr(VI) from the soil by using an alkaline
solution,  separate  Cr(VI)  from  interferences  in  the  extract  by  capillary
electrophoresis,  and  measure Cr(VI) directly  by absorbance  at  270 nm.   The
extracting solution is alkaline to inhibit reduction of Cr(VI) by any reducing agents
present in the soil.

Interferences are any soluble substances in the soil absorbing at 270 nm and co-
migrating with Cr(VI).  Any reducing agent in the soil capable of reducing Cr(VI)
in alkaline solution, assuming such an agent and Cr(VI) co-exist in the soil when the
soil was sampled, will be a negative interference. Any unfilled binding sites (mineral
or organic) in the soil, assuming such sites co-exist with  available Cr(VI) in the soil
at the time of sampling, will be a negative interference. The usual reducing agents
found in natural soils do not reduce Cr(VI) in alkaline solutions in the time frame
of this method.

APPARATUS AND MATERIALS

1.   Capillary electrophoresis (CE)  with electromigration injection, absorbance
detector at 270 nm, and a device to flush the capillary between runs.  A number of
instruments are commercially available. The instrument used  to gather the data in
the  Method Performance  section was:  ISCO Model 3850 Electropherograph;
uncoated silica,  60 cm, 50 pm ID capillary; detector  at 270 nm, 0.002 absorbance
range, 0.10 seconds rise time; voltage at 15  kilovolts, negative  polarity (the injector
end  is negative); injection  by electromigration, 5 seconds at 2 kV.

2.     Integrator or Strip Chart Recorder.

3.     40 mL sample vials with lids.

4.     Disposable  Pasteur pipettes with cotton or tissue wrapped tips.

5.     pH paper.

6.     Micropipettes, adjustable, 10- to 100-^L and 100- to  1000-//L.

7.     Volumetric flasks, 100 mL.
                                      507

-------
1.     Extraction  Solution:    0.1  gram  sodium  acetate,  0.04   gram  alkyl
trimethylammonium bromide (Sigma M-7635 or equivalent), 0.1 mL of 10 N sodium
hydroxide, all diluted to 100 mL with deionized water.

2.     10 N NaOH: Dissolve 40 grams sodium hydroxide pellets in 50 mL deionized
water.  Dilute to 100 mL with deionized water.

3.     1 N NaOH: Dilute 10 mL of above 10 N NaOH to 100 mL with deionized
water.

4.     Dry reagent grade potassium dichromate (K2Cr2O7).

5.     Cr(VI) Calibration  Stock  Solution,  2000 mg Cr(VI)  per liter of solution:
Dissolve 0.566 gram potassium dichromate and  0.1 gram sodium acetate in 50 mL
deionized  water. Dilute to  100 mL with deionized water.

6.     Cr(VI) Calibration  Standards:   Dilute  the  following volumes of Cr(VI)
calibration stock solution to 100 mL with Extraction Solution:

             Cr(VT).(m  CrVIL soln)
                   0                         0
                   1                         50
                   3                         150
                   5                         250
                   10                        500
                   20                        1000
                   30                        1500
                   50                        2500

7.   Dry sand known to be free of Cr(VI).

8.  Method Standard Stock Solution: Dissolve 0.283 grams of potassium dichromate
and 0.1 gram of sodium acetate in 50 mL of deionized water, dilute to 100 mL with
deionized water to make stock solution with a concentration of 1000 mg Cr(VI) per
liter of solution.   Prepare this  solution using a separate source of  potassium
dichromate than the Calibration Stock Solution in Step 5 above, and keep it as
independent as possible from the Calibration Standards.
                                     508

-------
PC SAMPLES

The definitions of recommended QC samples are:

1.     Method Blank:  a sample of sand known to be free of Cr(VI) put through the
same steps as a regular soil sample.

2.     Matrix Spike: an amount of the Cr(VI) Calibration Stock Solution added to
a replicate of a soil sample at the appropriate step (see Analytical Procedure section:
Sample Preparation) and analyzed as a regular soil sample.

3.     Matrix Spike Duplicate:  an amount (as above) of the Cr(VI) solution added
to  a second replicate of a soil sample to give the same increase in concentration as
the above matrix spike, and analyzed as above.

4.     Method Standard: an independent check sample of sand known to be free of
Cr(VI) spiked (as above) with Method Standard Stock Solution to give 40  mg
Cr(VI)/kg sand (or other amount as specified in a QA project plan) and put through
the same steps as a regular soil sample.

SAMPLE COLLECTION. PRESERVATION. AND HOLDING TIMES

Collect soil samples using standard field procedures.  Since the Cr(VI)  and  other
interfering substances may be unstable in the disturbed soil, analyze the samples as
soon as possible  after sample collection. Do not add  preservatives to the sample.
Holding time studies were not conducted.

ANALYTICAL PROCEDURE

1.     SAMPLE PREPARATION

Tare a 40 mL wide-mouth glass  vial (as commonly used for VGA  analysis) and
transfer approximately 20 grams of the sample into the vial.  Measure and record the
actual sample weight.  Add the extraction fluid to the sample in the vial, 15 mL or
enough to make a free-flowing slurry, and mix the solid and liquid phases by shaking.
Calculate and record the actual volume of extraction fluid using the measured weight
and the density of the fluid. Test the pH of the slurry immediately and, if necessary,
adjust the pH with 10  N NaOH to greater than 8.5. For matrix spike and matrix
spike duplicate  samples, add  the spike to the sample after  the pH is adjusted.
Experience showed that analysis should be completed within two hours after adding
the extraction solution to prevent  loss of Cr(VI).
                                    509

-------
2.     INSTRUMENT

Set up  the  CE instrument with  electromigration injection  according to  the
manufacturer's instructions.  Use extraction fluid to fill both reservoir cups.

3.     CALIBRATION

Analyze each of the calibration standards.  Flush  the capillary between standards
with three drops of extraction fluid.

Construct a calibration curve. Initial trials indicated the correlation coefficient was
0.995 or greater  over a range of zero to 50 mg/L.

4.     SPIKING PROCEDURE

Calculate the amount of spike (S) to add  to the sample as follows:

Let "D" equal the desired increase in concentration of Cr(VI) due to the added spike
to the soil sample, in units of mg Cr(VI)/kg soil, dry weight basis.
Then,
            S =    V  *  1.000.000
                 V'2000   -  1
                D *  W* F
where:
    S = the volume of calibration stock solution (which is
        2000 mg Cr(VI) per liter) added as a spike to the soil
        sample, in units of //L,

    V = the volume of extraction fluid added to the soil
        sample, in units of L,

    1,000,000 = the units conversion factor to convert the
        units of S from L to
   2000 = the concentration of the calibration stock solution
        in units of mg Cr(VI)/L,

   D =  the desired concentration increase, in units of
        mg Cr(VI)/kg, dry weight basis,
                                      510

-------
    W = the weight of the soil sample in the vial, "as is", in
        units of kg, wet weight basis, and

    F = the unitless fraction of dry material in the soil
        sample.

Add the spike, S  (jiL), to the sample at the appropriate step given above, and mix
thoroughly by shaking.

5.    ANALYSIS

Draw about 100 /*L of the liquid from the slurry through a tissue or cotton-tip into
a  Pasteur  pipet  and transfer the  clear liquid into the sample cup in the CE
instrument. Inject  the sample by electromigration into the instrument using a 5-
second injection time.  Apply the CE voltage and record the absorbance measured
at 270 nm with a  chromatographic integrator or  stripchart recorder.   For most
samples, the Cr(VI) absorbance peak occurs at about two minutes.  The retention
time depends on the concentration of the buffer, the flush history of the capillary, the
temperature, and the voltage.

Since other electrophoretic peaks appear in the electrophoreogram and the capillary
becomes coated with contamination when running regular  soil samples, flush the
capillary before every injection using three drops of  1.0 N NaOH,  followed by six
drops DI water, and finally six drops of extraction fluid. Indications of inadequate
flushing are shifts in retention time, ghost peaks, and broad peaks.  A small peak
observed at the correct time window should be confirmed with Another injection after
flushing the capillary.

CALCULATIONS

1.    Calculate the result, H, as follows:

      H (mg Cr(VI) /kg soil, dry) =
           C (mg Cr(VD / L soln) * V (L sohri
           W (kg soil, as is) * F (unitless, solids in soil)

where:

    H = the measured concentration of Cr(VI) in the soil, dry
      weight basis, in units of mg/kg,
                                     511

-------
    C  = the concentration of Cr(VI) measured in the extraction
       fluid, taken from the detector response and the
       calibration graph, in units of mg/L,

    V  = the volume of extraction fluid added to the soil sample
       in units of L,

    W = the weight of the soil sample in the vial, "as is" or
       "wet weight",  in units of kg,

    F  = the unitless fraction of dry material in the soil
       sample.

2. Calculate the  spike recovery for the matrix spike and the matrix spike duplicate
as follows:

       % Recovery (spike) = 100 * R/(H +  D)

where:

    % Recovery (spike) =  the matrix spike recovery in units of
                   percent, dry weight basis,

    H  = the measured unspiked sample result given above, in
       units of mg/kg, dry weight basis,

    D  = the increase in concentration of Cr(VI) in the sample due
       to the spike added, in units of mg/kg,  dry weight basis,

    R  = the measured spiked sample result, obtained using the
       same expression as H  above, in units of mg/kg, dry
       weight basis.

3. Calculate the method standard recovery as follows:

    %  Recovery (standard) = ( H / STD ) *  100

where:

       % Recovery (standard) = percent recovery of the method
                     standard in units of percent, dry
                     weight basis,
                                      512

-------
    H = the measured sample result given above, in units of
      m8/kg, dry weight basis, and

    STD = the actual or true value of the concentration of
        Cr(VI) in the standard sample, ip units of mg/kg dry
        weight basis.  The default value of STD was set
        to 40, but may be changed to meet project objectives
        specified in a QA project plan.

4.  Calculate the percent relative standard deviation of the matrix spike and the
matrix spike duplicate (the two replicate samples may have different weights, but
each must have the same increase in concentration due to the spike, see Analytical
Procedures section: Spiking Procedures) as follows:

        %RSD = sqrt(2) * abs( MS - MSD ) / (MS + MSD)

where:

        %RSD = percent relative standard deviation in units of
            percent,

        sqrt(2)  = the square root of 2,

        abs(X)  = the absolute value  of X,

        MS = the measured result for the matrix spike sample,
           obtained as H above, in units of mg/kg dry weight
           basis, and

        MSD = the measured result for the matrix spike
           duplicate sample, as MS above.

CONTROL LIMITS

Using the data given in the Results Section, the 95% confidence interval for the
control  limit for the percent recovery of the method standard and the matrix spike
was from 60% to 140%.  This assumes  the method  standard and the matrix spike
behave the same.

The 95% control limits of the percent relative standard deviation of the lab matrix
spike and the lab matrix spike duplicate were: lower limit =  5%, upper limit =
10%.
                                     513

-------
When conducting field spiking, note  that field  spiked  samples are potentially
unstable, since Cr(VI) and  reducing agents could co-exist indefinitely only in a
completely dry soil spiked with a dry substance.

METHOD  PERFORMANCE

The following is a summary of results for a number of different matrices using this
method.  All following values were  the result of seven separate replicate samples
analyzed in one batch.  All values are in units of mg Cr(VI) per kg soil, dry weight
basis.  All entries in the  following tables were calculated using the expressions
specified above, except the MDL which followed 40 CFR Part 136 Appendix B. The
naming and data reporting conventions  are:

    Av.Obs.   the average observed value,

    n = 7       indicates seven replicate samples were used
            to determine the reported  value,

    A%R      the average percent recovery,

    MDL      the detection limit,

    df       the degrees of freedom of the reported value,

    AL%RSD    the average laboratory percent relative
            standard deviation.
1.     Fertilite (tm) 'Top Soil", UPC# 7270140030, Hyponex Corporation,  14111
Scottslawn Road, Marysville, Ohio 43041. (This soil failed QA using the previous
methods.)

 Spike Added  Av.Obs.(n = 7>>  A%Rfn = 7^   MDL(df=6^ AL%RSDrdf=6^

    0           non-detect           -
    10.3            8.4             84.1          3.2         119
   102.8            96.9             94.3          -           4.5
                                    514

-------
2.  U.S. EPA Synthetic Soil Matrix, Risk Reduction Engineering Laboratory, Release
Control Branch, Edison, New Jersey  08837-3079.  (This soil  passed  QA using
previous methods.)

 Spike Added  Av.Obs.(n=7)  A%R(n=7^  MDUdf=6) AL%RSD(df=6^

    0            non-detect          ...
    4.0             4.3            106.5          4.3         32.5
   40.0             53.4            133.5           -          9.2
3.    The above Synthetic Soil Matrix fortified with 2% glucose by weight.  (This
soil failed QA using previous methods.)

  Spike Added  Av.Obs.(n=7^ A%R(n=7^  MDL(df=6^  AL%RSD(df=6)

    0           non-detect            -          -
    4.0            6.4            161.          4.3          21.9
   40.0            27.0              67.5          -            8.9
4. Loess, underlying material obtained from road cut on Interstate 70 between the
57th Street exit and the 61st Street overpass, Section 11, T11S, R24E, Wyandotte
County, Kansas.  (This soil passed QA using previous methods.)

  Spike Added  Av.Obs.(n=7^   A%R(n=7)  MDL(df=6)  AL%RSD(df=6)

    0           non-detect           ...
   5.52             4.2             75.9           0.55         4.8
  55.2            52.7             95.5            -           2.8
SUMMARY

The above text is an abridged version of the SOP used in our laboratory for the
analysis of soil samples for hexavalent chromium.  The  throughput for one person
using this method is 20 to 30 samples per day. This compares with a throughput of
2 to 3 samples per day using the older method. Analysts  familiar with both methods
prefer the CE method. Background reading for this method includes: 1) P. Jandik
et al., "Electrophoretic Capillary Ion Analysis: Origins, Principles, and Applications."
LC-GC 9 634-645, 1991, and 2) J. E. McLean & B. E. Bledsoe, "Behavior of Metals
in Soils." EPA/540/S-92/018.
                                     515

-------
79       FACTORS AFFECTING SAMPLE THROUGHPUT FOR AN ICP-OES SYSTEM
'                       WITH A SEGMENTED-ARRAY CCD DETECTOR

      Z.A. Grosser. K.J. Fredeen, C. Anderau, D.A. Yates, K.W. Barnes, T.J. Gluodenis, Jr.
      The Perkin-Elmer Corporation, 761 Main Ave., Norwalk, CT 06859-0215


      Introduction

      Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) is a fast,
      economical technique commonly employed for the determination of metals in a
      variety of environmental matrices. Advances in optics and detection systems for
      ICP-OES have  led to  instruments  that have  new  capabilities  and  better
      performance  (1,  2). In this work we examine the potential to increase sample
      throughput by evaluating the factors that may be varied versus the data quality
      objectives for the analysis.  We will characterize several of the ICP performance
      characteristics and evaluate several real applications for potential time savings.

      Data quality objectives (DQOs)  are critical in evaluating the performance that
      must be  achieved with an ICP-OES method.  DQOs  that will be particularly
      important  in  evaluating sample throughput  are detection limits,  accuracy
      requirements, precision, and dynamic range.  If these are well-characterized for a
      particular analytical situation then potential sacrifices  can be identified and
      judicious choices  made in pursuing sample throughput increases.

      Several ICP-OES instrumental factors will affect the DQOs and analysis time.
      These  factors were investigated to quantitate the magnitude of each effect.
      Integration time will affect both the detection limits and precision of an analysis.
      Fifteen elements  representing the full wavelength range were studied and showed
      that increasing the integration for each replicate will improve both detection
      limits and precision. The greatest improvement is for elements with transitions
      below  350 nm where  photon shot noise is  the predominant source of noise
      affecting  the signal and background.

      Background correction can remove spurious radiation from the desired signal by
      comparison with an off-line measurement.  If background correction is done
      simultaneously, rather than sequentially,  detection limits can be improved for
      elements with transitions above 250 nm and exceed detection limits without
      background correction above 350 nm. These transitions take place in a region of
      the spectrum  dominated by flicker noise, which can be  partially removed  by the
      correlation in  simultaneous background correction.

      In a complex sample, background correction may be even further improved by the
      use of  a  multivariate correction procedure  such as Multicomponent Spectral
      Fitting  (MSF) (3). MSF uses the full spectrum to create a model of peak shapes
      used in correcting the analytical signal.  Figure 1 demonstrates the magnitude of
      improvement  that can be expected with different types of background correction
      procedures. MSF improves the detection limit beyond that which can be achieved
      without background correction.
                                           516

-------
                                                         No BGC, 8 sec

                                                         Simult BGC, 8 sec

                                                         Seq BGC, 24 sec

                                                         MSF, 8 sec
           Zn Pb Cd Ni Mn Fe Cr V  Cu  Ti La Al  Sc  K

                         Element
        Figure 1. Effect of Background Correction on Detection Limits

Three applications were studied to demonstrate the practical application of the
decision-making  process. The first application is to increase  the  sample
throughput for a wear metals  trend analysis (WMTA) requiring moderate
detection levels for 16 elements. An oil additive and  trace analysis (OATA)
requires lower detection limits and precision better than 1%.  Finally a TCLP
analysis for 7  elements extracted into an acetic acid buffered matrix was
examined.  Table I lists the  data quality objectives for the  three analytical
situations.

                                 Table I
                         Data Quality Objectives of
                       Three ICP-OES Applications
                         WMTA
No. Elements                16
No. Reads/Sample            1
Detection Limits (mg/L)    0.01-0.5
Accuracy                  ± 10%
Precision (RSD)            3%
LDR (orders/elem.)          3
Rinse Out (orders)            3
   OATA
     16
      2
  0.001-0.1
     4-5
   TCLP
7 (14 lines)
     3
<0.1xMCL
   ±10%
    3%
    4-5
LDR to 3x DL LDR to 3x DL
Results

After careful evaluation of the various instrumental factors the WMTA was
speeded up to allow the analysis of a standard, blank, and 20 samples in 13.5
minutes.  Including a QC check sample every fifth sample, 98 samples can be
analyzed in 1 hour. The average precision obtained at concentrations above 50x
the detection limit were 1.4% RSD.
                                     517

-------
The OATA speed  was increased to analyze 35 samples per hour.  For low
concentration samples an analysis time of 2 minutes was necessary to meet the
DQOs,  but when concentrations approach 20 mg/kg the analysis time can be
reduced to 1.2 minutes/sample.

Two emission lines were monitored for each TCLP element to add confirmation
information to the analysis. The analysis was performed in less than one minute
with 65 samples per hour analysis rate, including the QC checks. If the ICP
periodic  table is  monitored  to evaluate  the  sample for other possible
contaminants the analysis is still performed in less than 2 minutes per sample.

Summary

Advances  in ICP-OES technology have provided the base for improvements in
sample throughput.  ICP-OES factors have been evaluated to demonstrate the
relationship with sample throughput and the concepts applied to three real-world
applications. In all cases the analyses were  speeded and the sample throughput
increased  significantly.  The  principles examined  can be  applied  to other
situations to evaluate the potential for time savings.

References

1.  T.W. Barnard, M.I. Crockett, J.C. Ivaldi, and P.L. Lundberg, Design and
Evaluation of an Echelle Grating Optical System for ICP-OES, Anal. Chem. 65,
pp. 1225-1230, 1993.

2.  T.W. Barnard, M.I.  Crockett, J.C.  Ivaldi, P.L.  Lundberg, D.A. Yates, P.A.
Levine, and D.J. Sauer, Solid State Detector for ICP-OES, Anal. Chem. 65, pp.
1231-1239, 1993.

3.  J.C. Ivaldi, D. Tracy, T.W. Barnard, and  W. Slavin, Multivariate Methods for
Interpretation  of Emission Spectra from the Inductively  Coupled Plasma
Spectrochim. Acta 478(12) pp. 1361-1371, 1992.
                                    518

-------
73   DESIGN, PERFORMANCE, AND ENVIRONMENTAL APPLICATIONS
      OF AN ICP ARRAY DETECTOR SPECTROMETER

      D. Nvgaard. ICP Product Manager, Frank Bulman, and Manny Almeida, Baird Corporation,
      125 Middlesex Turnpike, Bedford, MA 01730

      This paper will describe and evaluate a new kind of ICP-optical emission spectrometer — one
      utilizing a mega-pixel, two dimensional, solid state array detector.  With more than one million
      individual detector elements, the array captures a high resolution electronic photograph of the
      complete UV-visible emission spectrum.  All elements in the periodic table that emit UV-visible
      light in an ICP discharge can be detected and quantified at any number and any combination of
      wavelengths.  This capability enables  rapid qualitative  and  semiquantitative screening of
      unknown samples, an application that will be discussed in detail especially as it pertains to waste
      analysis. In addition, rigorous quantitative analysis of water and waste will be demonstrated to
      the quality control standards established by the Contract Laboratory Program.
                                            519

-------
74        VARIABILITY IN TCLP METALS RESULTS FROM
                              STABILIZED WASTES
       Gene  Merewether. Analytical Chemist, Nilesh Shah, Senior Analytical Chemist;
       Quality/Analytical Programs, Chemical Waste  Management,  Inc.,  7250  West
       College Drive, Palos Heights, Dlinois 60643.
       ABSTRACT
       The  purpose  of this  multi-laboratory  study  was  to identify variables in  Toxicity
       Characteristics Leaching Procedure (TCLP) and subsequent analysis and to recommend
       standard protocols to  improve inter-laboratory precision.   The TCLP, together with
       subsequent digestion and instrumentation methods, allow choices of laboratory practices.
       The choices made by individual laboratories contribute to inter-laboratory variability.  For
       instance, TCLP allows extraction of any particle size less than 3/8 inch, and any extraction
       time from 16 to 20 hours.  Additionally, choices made  about specific digestion  methods
       and instrument methods may also allow further inter-laboratory variation.

       For analyzing TCLP metals, a laboratory can choose from any appropriate digestion or
       instrument procedures.  For example, lead  can be analyzed either by graphite furnace
       atomic absorption (GFAA) spectroscopy  (Method 7421) or inductively coupled plasma
       (ICP) atomic emission spectroscopy (Method 6010).  The digestion procedure associated
       with each analysis is Method 3020 for GFAA or Method 3010 for ICP.  Inevitably, there
       are  differences between the  methods,  and there is  variability within each  method.
       Individual waste matrices can also add variability to the results.

       The goal of improving inter-laboratory precision therefore required  identifying  and
       quantitating variables that lead to differences in TCLP results. The approach taken in this
       study was to have each participating laboratory analyze a "tailor made" proficiency sample
       so  that inter-laboratory and intra-laboratory variability could be measured.   Existing
       QA/QC  data from  proficiency samples  and standard reference materials  were  also
       examined.  The results from round robin analyses and the recommendations for analytical
       protocols are discussed.
                                             520

-------
75   AMALGAMATION CVAA TO IMPROVE THE DETECTION LIMIT OF MERCURY
                              IN ENVIRONMENTAL SAMPLES

      Steven Sauerhoff. Susan Mclntosh, and Zoe Grosser
      The Perkin-Elmer Corporation, 761 Main Ave, Norwalk, CT 06859-0219


      Introduction

      Mercury is of continuing interest in the environment due to its persistence and
      toxicity. Mercury determination has traditionally been performed using cold vapor
      generation coupled with atomic absorption (CVAA) to achieve a detection limit of
      200 ng/L (ppt) in a prepared sample. Each sample is handled individually and the
      technique is very labor  and time intensive. Manual sample handling  may also
      expose the sample to additional sources of contamination, increasing the variability
      in the results obtained with the technique.

      Flow injection techniques (FI), where reagents are mixed with the digested sample
      automatically, provide  several advantages.  The samples  are  isolated from
      atmospheric and handling contamination.  The volume of sample necessary to
      generate the signal is greatly reduced, which also reduces the consumption of
      reagents and waste generated from the process. Interferences are reduced because
      the mercury vapor is separated from the mixture very quickly, limiting the time for
      adverse reactions to occur.  The sample throughput is increased and the procedure
      can run unattended (1).

      Although FI can significantly improve the quality and ease of a classical mercury
      determination, lower  detection limits would also be useful. Coupling FI with
      newer light sources, such as the EDL System 2 can lower the detection limit to 60
      ng/L for routine analyses. However, research into bioaccumulation of mercury and
      naturally occurring marine levels require the establishment of still lower detection
      levels.  This work describes the  development  of a FI system coupled with
      preconcentration of mercury by amalgamation to reduce mercury detection limits
      to 2 ng/L for an 8.5-mL sample size (2).

      Figure 1 shows the FI  system coupled to the amalgamation accessory for mercury
      preconcentration.  Mercury vapor is generated in the mixing manifold and
      separated in the gas-liquid separator. The mercury vapor is trapped on a
      gold/platinum gauze and held there until a sufficient amount is deposited. The
      deposition is halted and the gauze heated to release the mercury as a concentrated
      plug into the absorption cell of the atomic absorption spectrometer, where the
      absorbance is measured and quantitated.

      Results

      Figure 2 compares the signals obtained for 1000  ng/L mercury with and without
      amalgamation preconcentration. The method detection limit, measured  using the
      procedure in CFR, Part 136 Appendix B, was found to be 2 ng/L for an 8.5-mL
      sample and 60 seconds of deposition time.  This represents the detection of 17 pg
      of mercury on an absolute basis. Larger samples and longer deposition times might
      decrease the detection limit, but eventually reagent and handling contamination
      are going to limit further improvement.
                                          521

-------
                                        KSnCI,
Figure 1. FI Mercury Amalgamation System
The mercury vapor is channeled through the FI valve to the amalgamation system.
This controlled vapor deposition technique reduces sample carryover and improves
precision.  The  precision for replicate samples  was 2.5% RSD at  20 ng/L of
mercury.
          8.244
                                              	 1.88 ppb
                                              	 1.86 ppb Hg CU
                                      i<— 1.8 ppb Hg 68 sec. (Wilgo.
                                 <—1.8 ppb Hg Cold Va
                                  h^^^Mrib

                                  (nc)
2fl.ee
Figure 2. Mercury Signals With and Without Amalgamation
                                     522

-------
Summary

The system  described  in this  work provides  several  advantages for the
determination of mercury compared to conventional CVAA batch analysis. The
detection limit using amalgamation is reduced to  levels that  allow the study of
marine and estuary  ecosystems.  The addition of reagents  is  automatic and the
sample throughput is improved to allow the determination of 30 samples per hour
in unattended operation.


References

1. S. Mclntosh and B. Welz, Perkin-Elmer  Method 245.1A; Determination of
Mercury in Drinking Water and Wastewater by Flow Injection Atomic Absorption
Spectrometry (Cold Vapor Technique), ENV-12A,  available  from The
Perkin-Elmer Corporation, 761 Main Ave., Norwalk, CT, 06859-0012.

2. S. Mclntosh, The Determination of Mercury at Ultra-Trace Levels Using an
Automated Amalgamation Technique, Atom. Spectrosc., 14 (2), pp.47-49 (1993).
                                    523

-------
7fi     THE DEVELOPMENT OF EPA TOTAL MERCURY METHOD USING COLD VAPOR FLUORESCENCE
'      MERCURY DETECTION SYSTEMS.   THE  MERCURY METHOD  WILL BE USED FOR THE ANALYSE
             OF MERCURY IN WATER,  WASTEWATER,  SEA WATER AND RELATED MATRICES.


      Authors:  Billy B.  Potter.  Research Chemist,  Inorganic Chemistry Branch,
      Chemistry Research Division,  Environmental  Monitoring Systems Laboratory,  U.S.
      Environmental Protection Agency,  26 W. Martin Luther King Dr.,  Cincinnati Ohio
      45268.

      Stephen E. Long and Jane A.  Doster,  Technology Applications,  Inc.,  Cincinnati,
      Ohio.

      Robert Davis, Computer  Services Corporation,  Cincinnati,  Ohio.

      ABSTRACT:

      The U.S. Environmental  Protection Agency (USEPA)  is  developing a method for
      the determination of total  mercury found in water and sediment at the part per
      trillion (ppt) level.   The  total  mercury method  has  an estimated method
      detection limit (MDL)  of 5  ppt to 20 ppt of mercury.   The MDL is  made possible
      by digesting the sample using bromide/bromate reagent  followed by detection of
      elemental mercury by cold vapor atomic fluorescence  spectrometry at 253.7  nm

      INTRODUCTION:

      Methylmercury poisoning of humans  and cats in Japan known as "Minamata" disease
      (1950-1970) was caused by the consumption of mercury  (10 to 24 parts per million
      [ppm])  contaminated fish.   This environmental catastrophe marked the beginning
      of a world-wide concern that  mercury may be a global pollution  problem.   The
      history, hazards,  and concerns about mercury pollution  were documented by 19721.
      The Minamata Bay tragedy led to independent  investigations of  the  Great  Lakes
      region by  the  United  States  and  Canada.   These  studies  found high  levels  of
      mercury  in  fish.    From  these  early  studies,   it  became  apparent  that  a
      standardized mercury method was needed for the exchange of mercury data between
      the two countries.   Jacobs et  al2  in 1960 introduced the oxidation/digestion of
      biological samples by wet digestion.  The  wet digestion technique  of mercury
      samples using  potassium  permanganate  and  sulfuric  acid  gained  acceptance  in
      Canada and United States.   The Canadian  government made strides to  standardize
      this approach.   The credit for  developing wet  digestion  and the  cold  vapor
      (flameless) atomic absorption technique  is  credited  to the  Canadian team Hatch
      and Ott3   However,  their approach to digestion  of mercury  samples  varied from
      the  Jacobs'  approach.    This variation of  sample  digestion  and  a lack  of
      standardization of  methodology  created  data compatibility  problems  between
      countries. The Canadian government elected to standardize on one of these sample
      digestion techniques.   By 1970 the cold vapor technique and  the permanganate
      sulfuric acid oxidation/digestion was accepted by the  Fisheries Research Board
      of Canada and Federal Water  Quality Administration  (FWQA) as provisional methods
      with minor differences  in  sample digestion.  The United States reviewed the
                                         524

-------
results of the Canadian methods, then provided the Water Pollution Control Agency
(WPCA, 1970) a provisional method for mercury analysis.

The USEPA, established in 1971,  took over the responsibilities of the WPCA.  The
USEPA,  Environmental Monitoring  and Support  Laboratory    Cincinnati (EMSL-
Cincinnati) continued  the  research  on the  WPCA provisional cold vapor method.
The USEPA, EMSL-Cincinnati  proceeded to standardize the method.  The EPA's method
development  protocol  required  a  feasibility  study   and single   laboratory
validation study.  The  method feasibility study was conducted by J.F.  Kopp, M.C.
Longbottom, and  L.B. Lobring.A   These researchers used an experimental design
that included the analysis of known  interferences  (chloride, sulfide, organics)
and other variables (digestion, heat, time).  The result of this research in 1972
lead to the development of  the optimized "Cold Vapor" USEPA Mercury Method 245.1
for mercury-total5 that was  completed  in 1976  and revised in 1983.   The USEPA
adopted an extensive method validation policy and then proceeded to confirm the
single operator  precision  by expanding to  multi-laboratory confirmation.   The
USEPA,  EMSL-Cincinnati,  Quality Assurance  Branch conducted  a  nation-wide
validation  study.   This study  validated  the USEPA  Mercury Method 245.1  that
produces data of known  quality at the part per billion (ppb) level. However, new
regulatory demands are  now being made to detect mercury at the parts per trillion
(ppt)  level.   These regulatory demands are a  result  of  new consensus  among
researchers6-7-8'9 over  growing mercury pollution of the aquatic  food  chain.

There  are at  least  23 states  with existing fish  consumption advisories for
mercury and numerous USEPA Superfund sites with significant mercury contamination
of soils, ground and  surface waters.   The fish advisories and the need to monitor
trace level mercury migration from Superfund sites has prompted many requests for
a mercury method able to detect mercury at the ppt level.  To protect the aquatic
environment, many  states are  now implementing monitoring and enforcement  of
National Pollutant Discharge  Elimination System (NPDES) permits established by
water quality-based effluent  limitations.   The effluent limitation for mercury
uses a bioaccumulation factor in the equation that has resulted in permit values
being set below  the USEPA Method 245.1 detection limit for total mercury.   The
current USEPA Method 245.1 has a detection limit for mercury of  .2 /ig-Hg/L (.2
ppb) in water.  This current MDL is  too high  to satisfy monitoring requirements
as established by the USEPA and state water quality-based effluent limitations
requirements.   A mercury method with  a low detection  limit  (20  to  2  ppt)  is
desired for the determination of mercury in natural waters.  This would support
studies that are attempting to define probable sources of mercury contamination
in the  aquatic food chain.   There  is a study being proposed for the  Florida
Everglades that will require ppt detection limits.

The State  of  Florida has determined that many sport fish caught in the  South
Florida Everglades are contaminated with high levels  (0.5  to 4 ppm) of mercury.
This has caused  a  fish advisory to  be issued for the Florida Everglades.   The
search for the  cause of mercury contamination  is hampered by a mercury method
having a  detection limit  of  0.2 ppb.  This  detection  limit  is  too high for
ambient water analysis where  mercury levels  are  expected to be in the  low ppt
level.   The  State  of Florida  is   located  in  USEPA  Region  4.   The  USEPA,
                                   525

-------
Environmental  Monitoring Systems  Laboratory    Cincinnati  (EMSL-Cincinnati)
received a request to support USEPA, Region 4, by providing methods development
and quality assurance resources for the Everglades study.  Region 4 will provide
a plan of study for the Everglades ecosystem10. The study of the Everglades will
require  methods  that can measure  mercury below the  ambient levels  found in
nature.
The  current  EPA Mercury  Method 245.1  may be  applied to  many  sample  types
(industrial and  domestic).   Although it is  a  rugged method,  it  suffers from
interferences that  absorb UV radiation that  can cause a positive bias. ^There are
two ways to overcome these interferences and lower the detection limits:

(a)  The  atomic fluorescence  of  mercury  is   independent  of  UV  absorption
     interferences.  Atomic fluorescence detects mercury by emission rather than
     by absorption.   Since  emission is inherently more sensitive than atomic
     absorption, it  is expected that detection  limits  may be  lowered  to  parts
     per trillion  (ppt, ng/L) range.

(b)  Interferences  may be eliminated  by pre-concentration of mercury  on gold
     amalgam.   Gas  vapors containing interferences pass through the instrument
     before the  mercury  analysis.   The  gold/mercury amalgam trap  is  heated,
     releasing the  mercury vapor for analysis by atomic fluorescence detector.
     This increases mercury sensitivity by producing a highly concentrated peak.

EXPERIMENTAL:

A statistically-based experimental  design or chemometric  approach as described
by Deming  and  Morgan (1987)11 was  selected for the evaluation of the  mercury
method. The chemometric experimental approach was applied to this mercury method
to speed the  process of method evaluation.  The chemometric approach is dynamic
(modifiable)  and recursive  (experiments may be repeated).   During the execution
of the experiments  an evaluation of each "phase" of an experiment is required.
When a modification of the experiment was required,  it was strongly supported by
the statistical evidence.   The experimental design consisted  of  the following
phases:

Phase 1 -Familiarization Study.
Phase 2 -Automated  Instrument Optimization  Study.
Phase 3 -Automated  Instrument Linearity Study.
Phase 4 -Mercury Precision and Recovery Study.
Phase 5 -Instrument Stability Study.
Phase 6  Initial Interference Study.
Phase 7 -Sample Preservation Study.
Phase 8 -Single Laboratory Validation Study.
Phase 9 -Establish  Instrument Control  Charts.
Phase 10   Establish Clean Room Protocol.
                                   526

-------
The  instrument selected for  the study  is  a QuesTron "PSA" Automated Mercury
Analyzer  (Merc  22) .    This  instrument  offers a  complete  system  containing
Autosampler,   Gold  amalgam pre-concentrator,  fluorescence  detector,  an IBM
compatible  USEPA  Contract  Laboratory Program (CLP)  quality  control software
program and data acquisition  system.

The  automated instrument  is  generally  configured as shown below.    The gold
amalgam accessory for the instrument system is  not  shown in this configuration.
The  gold amalgam accessory will be  evaluated during  the ruggedness  testing
portion  of  the  method  development  and is not  part  of  the  scope of this
experimentation.
                                                         FLUORESCENCE
                                                          DETECTOR
             PERISTALTIC
               PIMP
                                                                       CARBON
                                                                WASTE    FILTER
                                                                      CVent to HcxxO
   AUTOSAIiPLER
GABBIER y


S
«
I
• 	 /
	 f
•/c
I
\LIQUID
WASTE
COLLECTION
vant to Hood}
GASM-IQUID
SEPARATOR
' If*
d u

                     ARSON GAS
              FIGURE 1: PSA AUTOMATED MERCURY FLUORESCENCE SYSTEM
                                     527

-------
RESULTS:

Instrument Optimization:

In the  familiarization  and optimization phase  of  the  experiments,  the mercury
analyzer was optimized for maximum sensitivity and/or signal-to-noise ratio.  The
use of Simplex optimization was investigated using the carrier gas and sheath gas
flow rates as selected variables.  A Simplex program written in BASIC was used
for algorithm computations.   The  optimized  flow rates calculated by the program
were 156 mL/min for the  sheath gas  and 329 mL/min for the carrier gas, although
there  was  some evidence  from the Simplex convergence  behavior that  the  two
variables were not interactive.   The timing sequences on the hydride generator,
which define the analytical cycle, were  also  optimized manually to obtain good
peak shapes.  Work  continued  on  the evaluation and optimization of instrument
performance and assessment of sample preparation  procedures.

Improvements were made on  the  system to increase precision and sensitivity.  The
rotameters  for  the  carrier,  sheath,  and dryer tube  gas were replaced  by  new
rotameters from PS Analytical.  The new rotameters for the carrier and sheath gas
flows  were  modified to a lower  and  more practical  range.    The  dryer  tube
rotameter was modified  to allow much higher flows  to  remove  moisture from the
sample more effectively.  The sensitivity  of  the  instrument was also increased
over 30% upon replacement of the  worn dryer tube  element.   After these changes
were made,  a check on the precision of the instrument was performed.  The system
was calibrated  once an  hour  for  seven  hours  and  the  slope  of the  curve  was
monitored as an indication of the instrument  drift.  The RSD of the slope over
the day was 3.6%, a noticeable improvement  over earlier drift that ranged from
about 10-20%.

Before conducting any detailed analytical benchmark studies,  it was decided to
optimize  the instrumentation  with respect  to maximizing  the  slope  of  the
sensitivity curve,  while  maintaining acceptable  precision, accuracy  and peak
shape dynamics.   As only two user-adjustable variables were considered to have
a critical effect on the  analytical performance of the  instrument,  namely  the
carrier and  sheath  gas  flow rates, it was not considered necessary to  use  a
simplex  optimization process.     The  experiment  consisted  of  running  two
repetitions  of  a 32-factorial design.   Each  repetition  was  conducted on  a
separate day to find out  the day-to-day reproducibility of the  system.  The order
of the experiments within each repetition was randomized.  The levels of carrier
gas flow chosen for the experiment were 200,  350 and 500 mL/min, while the sheath
gas flow rates  chosen were  0, 350 and  700 mL/min.  For each  experiment,  the
instrument  was  calibrated using  a calibration  blank  and three  calibration
standards (50,  100 and 150 ng/L).   Using this calibration curve,  the instrument
was then used to determine three test concentrations (25, 75 and  125 ng/L).  The
output variables recorded for each experiment  were the maximum peak height for
the 50 ng/L calibration  concentration,  the slope of the calibration line,  the
correlation of the calibration line and the  estimates provided by the system for
each of  the  three  test  concentrations.    The  results  of  the experiments  are
summarized in Table  1 below.
                                   528

-------
TABLE 1: DATA FROM OPTIMIZATION EXPERIMENTS
Carrier
mL/min
200
200
200
200
200
200
350
350
350
350
350
350
500
500
500
500
500
500
Sheath
mL/min
0
0
350
350
700
700
0
0
350
350
700
700
0
0
350
350
700
700
Day
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
Peak
Height
146.2
164.6
101.3
108.4
63.1
71.6
106.8
118.6
84.5
93.7
61.3
68.2
76.4
93.6
68.0
68.5
48.8
54.3
Slope
2.24
2.52
1.52
1.68
0.95
1.04
1.60
1.86
1.27
1.38
0.91
1.01
1.13
1.39
0.97
1.11
0.71
0.82
r
1.0000
1.0000
0.9999
1.0000
0.9998
0.9997
1.0000
1.0000
0.9998
0.9999
0.9999
0.9996
0.9999
1.0000
0.9998
0.9999
0.9999
1.0000
25
ng/L
24.4
24.1
26.4
24.6
25.1
23.6
25.1
24.3
24.8
23.2
22.9
25.8
24.2
23.5
21.5
24.4
24.5
25.8
75
ng/L
75.5
****
79.4
77.1
75.7
74.2
75.3
71.4
73.5
74.2
76.5
76.0
73.5
71.7
71.4
74.0
76.2
75.4
125
ng/L
****
****
118.7
****
124.8
125.6
****
****
122.8
120.9
127.5
127.1
126.5
119.8
118.6
125.1
126.4
125.0
****Test concentration signal off-scale r correlation coefficient

For the  (500,0) run  on day  1,  two duplicates of  each calibration  and test
concentration were run through the system.  However,  it was decided that running
duplicates for all factor setting combinations would be too time consuming.  All
of the remaining runs were based on just one sample of each of the calibration
and test concentrations.  To equalize the underlying variance, only the numbers
for the first  of  the  duplicates on the first run were  used.   The calibration
curve,' correlation coefficient  and  estimates  for  the  test" concentrations were
computed manually from the peak heights produced by the instrument.
                                   529

-------
To characterize the accuracy of the readings of the three test concentrations
the absolute percent deviation from the true concentration was used.  Originally
a multivariate analysis of variance (MANOVA) was proposed for the analysis or tne
data.   However,  the  missing  values  in  Table  1  meant  that  only  thirteen
observations could be used in the MANOVA.  Individual statistical analyses were
therefore,  run using the  response variables, calibration slope  and percent
deviation for  each of  the three  sample  concentrations.   Peak height was not
included because it provides similar information as  the slope of the calibration
line.  The correlation  coefficient was omitted because the calibration line was
almost always a perfect fit  (r values ranging from 0.9996 to 1.0000).  Type III
sums  of  squares were used because it could not be assumed  that interaction
effects were insignificant.  The day was included in the analysis because it was
of interest  to  know whether the performance of the  system  changed from day to
day.   Table 2  contains  the F-statistics  and  p-values  for  each of  the  four
response  variables  for  the  factors  carrier  gas  flow,   sheath  gas flow,  the
interaction  term between the two and the day.   This information shows  that the
accuracy  of  the  system does not change  significantly when  the flow levels are
varied.  However, the slope of the  calibration line differs significantly based
on all of the independent variables in the  model.

TABLE 2: OPTIMIZATION TEST STATISTICS
Dependent
Variable
Slope of
Calibration Line



Absolute
Percentage Error
(25 ng/L)


Absolute
Percentage Error
(50 ng/L)


Absolute
Percentage Error
(125 ng/L)


Source of Error

Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
Model
Carrier Gas
Sheath Gas
Carrier Sheath
Day
F-statistic

153.02
203.71
392.61
35.53
42.44
0.43
0.39
0.46
0.54
0.00
0.73
0.16
1.28
0.91
0.11
0.86
0.22
2.33
0.82
0.05
Degrees of
Freedom
9,8
2,8
2,8
4,8
1,8
9,8
2,8
2,8
4,8
1,8
9,7
2,7
2,7
4,7
1,7
7,5
2,5
2,5
2,5
1,5
p- value

0.0001
0.0001
0.0001
0.0001
0.0002
0.8867
0.6924
0.6471
0.7129
0.9714
0.6765
0.8566
0.3350
0.5083
0.7495
0.5876
0.8067
0.1928
0.4903
0.8262
                                    530

-------
Inspection of the data in Table 1 shows that the slope  of  the line increases as
either the carrier flow or sheath flow is lowered.  Unfortunately,  the peak shape
produced by the system at the lower gas flows is not of an acceptable shape for
routine measurements  and although higher slopes  would maximize  the detection
power of the system it also would compromise the  robustness of the measurement
process.  The experimental results clearly show that the calibration line slope
is inversely related to the carrier and sheath  gas  flow rates.   The flow rates
selected for continuing work were, therefore, based on  the lowest values of the
gas flows  that  were consistent with  the  provision of  reasonable peak shapes.
These experiments have indicated that adjustment of the  flow rates within the
limits employed here has no effect on the overall accuracy of the system.   The
flow settings for the system were selected based  on a  compromise between these
two parameters.   The  final  results  of the optimized instrument setting are in
Table 3 below.  These optimized  settings and procedures were held constant for
the remaining experiments.

TABLE 3: EXPERIMENTAL CONDITIONS FOR  SINGLE LABORATORY VALIDATION

Fluorescence Instrument Condition Instrument

PSA Merlin Series AFS
Flow rate, blank
Flow rate, SnCl2
Flow rate, sample
Carrier gas rate
Sheath gas rate
Drier tube gas rate
8.00 mL/min
3.00 mL/min
8.00 mL/min
0.35 L/min
0.70 L/min
3.00 L/min
Delay time
Rise  time
Analysis time
Memory time
10 s
25  s
30 s
60 s
INSTRUMENT PERFORMANCE BENCHMARKS

As part of the familiarization phase of the project, initial assessments of the
instrument detection limit (IDL) and analytical precision were made.  Precision
evaluated as relative standard deviation (n=10) was estimated at 100, 150, and
200 ng/L concentration levels  to be 4.2, 1.6 and 0.9%, respectively.  The IDL was
evaluated using  three  different variations in sample preparation.   All  three
initial assessments yielded an IDL of around 3 ng/L, where the IDL was defined
as  that  concentration giving an analytical signal equal  to three  times the
standard deviation of the blank signal.

Once the instrument settings were optimized and improvements to the system had
been made the instrument performance was assessed in more detail.  The IDL was
determined to be  0.8  ng/L.   Short term precision was measured  by  running ten
replicates of an inorganic mercury standard at  10,  50,  and 100X  the  IDL and
digested using the bromate/bromide technique.  The RSD of the instrument response
at each  concentration  was 2.3,  1.3,  and  0.79% respectively.   The short term
precision experiment was repeated using an organic mercury standard at the same
concentrations.    The RSD  at  10,  50,  and 100X  the IDL were  2.1,  1.1, and 1.0%
respectively.  Instrument stability was determined  by  measuring the long term
                                    531

-------
precision of the analysis  of a mid-level mercury standard over the course of one
work day.   The RSD of seven separate  analyses over six hours was 1.2%.  A control
chart that represents the day to day long term stability of the instrument was
constructed  by  plotting  the  sensitivity  slopes  for  twenty-five  calibration
curves produced between 8/13/92 and 10/14/92.   The  average slope  over the two
month period was 0.914 with  an RSD of  6.9%.   All  slopes were within three
standard deviations of the average,  indicating no excessive drift in the system.

COMPARISON OF SAMPLE DIGESTION METHODS

Two methods  for sample digestion were  assessed to  determine  which  should be
recommended for the proposed EPA Method.  The standard EPA type digestion defined
in Method 245.1  using KMn04/K2S208,  and a semi-automated method for total mercury
is being evaluated for acceptance as an official USEPA method.  This method is
used routinely by  the Yorkshire Water Authority (YWA)  in the United  Kingdom12.
In  the  YWA method, a  sample  aliquot  is  digested using  free  bromine reagent,
resulting in the break down of the commonly occurring organomercury species to
mercury (II )13'1/(.    Elemental  mercury  vapor  is  generated  from  the  digested
sample  by  reduction  with stannous  chloride and  is  purged from solution  by a
carrier stream of argon.   The mercury vapor is determined  by atomic fluorescence
spectrometry at  253.7  run15.   The  proposed method procedure  is simplified and
summarized as follows:

(1)  Transfer 35-40 mL of sample to  a  50-raL tared container.

(2)  Add 5 mL (1+1) hydrochloric acid and  1 mL 0.1N potassium bromate/potassium
     bromide.

(3)  Allow samples to stand for at least fifteen minutes before analysis.

(4)  If  the  yellow  color does not persist  after  fifteen minutes,   add  more
     KBr03/KBr solution.

(5)  Add 12% (w/v) hydroxylamine hydrochloride  (NH2OH-HC1) at a cone,  of 6.0 ftL
     per 10 mL of the sample, to eliminate  the excess bromine until yellow color
     disappears.

(6)  Turn on the automated instrument/detector  and allow to stabilize.

(7)  The sample enters gas/liquid  separator with SnCl2 to form mercury vapor.

(8)  The vapor is analyzed by cold vapor atomic fluorescence spectrometry.

Each digestion procedure  was optimized for reagent volume and final  dilution.
Sample containers for digestion and  analysis were also evaluated.   The original
EPA Method 245.1 utilizes  glass BOD bottles. Some initial studies for this work
were done in glass volumetric  flasks.  The difficulty in using glass containers
for sample digestion is that their capital cost can be prohibitive for disposable
use and they  must be cleaned using  a  rigorous and time  consuming wash regimen if
                                   532

-------
they  are  to be reused.   At ng/L detection  levels  any contamination from the
sample container becomes a crucial issue.  To reduce the risk of contamination,
disposable polypropylene sample containers for digestion and immediate analysis
were  introduced.   Initially, the digestion  procedures were experimented with
using  inorganic standards  in deionized water  to determine the  stability of
mercury in  each digestion matrix.  Early  studies showed  that both procedures
processed  the mercury  standards  equally well.   The  effects  of heat  on the
performance of the permanganate method were also initially  tested with inorganic
and organic standard solutions in  deionized water.  No significant  difference in
recovery was  observed between heated  and non-heated  samples  (Table  3) .   The
importance of the heating step for the  permanganate method became  apparent when
more  complex  matrices were  studied.   The current  methods  using permanganate
require that additional reagents be added to digest adequately samples high in
chloride  concentrations.    This  complicates  the  analysis procedure  since it
requires the preparation of a separate blank for samples treated with additional
reagents.

To compare the permanganate  method with  the bromate method in high concomitant
chloride conditions,  an artificial seawater with  an estimated chloride content
of 18,000 ppm was used as a sample  matrix.  Three aliquots of artificial seawater
were spiked to a concentration of 100 ng/L with either 100% inorganic mercury,
50:50 inorganic/organic mercury, or 100% organic mercury.   Two samples of each
spike were digested and analyzed for each digestion procedure.  The permanganate
method was tested both with and without heating.  The average recoveries for the
inorganic, 50:50,  and organic mercury spiked samples using  the bromate digestion
were 93.8, 92.1, and 90.4% respectively with close agreement between replicates.
For the permanganate digestion using heating,   average recoveries  were  95.3,
85.6, and  81.4%  respectively with a high  relative percent  difference between
replicates.   Repeating the  permanganate digestion  without  heating  yielded
recoveries of 108.6,  83.3, and 62.1% for the inorganic, 50:50, and organic spiked
samples also with high RPDs  between replicates.

     TABLE 3:   COMPARISON OF BROMATE AND PERMANGANATE DIGESTION RECOVERIES
    (100  NG/L INORGANIC, 50:50,  AND ORGANIC MERCURY)  IN ARTIFICIAL SEAWATER

                               % Recovery/(RPD)

               Bromate        Permanganate with Heat     Permanganate no Heat

Inorganic      93.8  (0.43)          95.3  (6.7)              108.6  (4.2)
50:50          92.1  (2.3)           85.6  (5.1)               83.3 (12.9)
Organic        90.4  (0.11)          81.4 (18.7)               62.1  (5.6)

Several advantages were found using the bromate digestion.  It provided better
accuracy and precision and increased sensitivity about the permanganate method.
The bromate method also is preferable because it requires  less reagents and can
be performed at ambient temperature in a shorter amount of time.   Specifically
for saline matrices,  the bromate digestion does not require further additions of
                                    533

-------
reagents or supplemental standard preparation  to  complete  the digestion, thus
eliminating this time consuming procedure that is necessary  for the permanganate
digestion.  From the evaluation of these data it was decided to incorporate the
bromate digestion into the experimental development of the  method.

INTERFERENCE STUDIES:

Studies of the affects of chloride on the recovery of mercury using the bromate
digestion and  the  permanganate digestion techniques  were  performed.   It was
noticed upon heating that  the  samples  lost the purple permanganate  color and
formed a brown precipitate.  This was not observed  in the  standards  that were
heated.   A  noticeable  difference  in  sensitivity  was  observed between the
digestion procedures.    Slopes  of  calibration standard  curves  prepared with
bromate digestion ranged from 0.9 to 1.0 and a 100 ppt sample had a peak height
of about 100,  whereas slopes for the permanganate  digestion ranged from 0.5 to
0.6  and  the peak height of the 100  ppt sample  was  about 60.   The  bromate
digestion appears to be superior to  the permanganate digestion in sensitivity,
accuracy and precision and is faster and  easier to perform.

Further interference studies (Table 4) were performed using the bromate digestion
only.  Potential interferences from sample  matrices  containing fluoride, sulfate,
nitrate,   phosphate,  gold,   silver,  or a mixture  of   calcium,  copper,  lead,
manganese, barium and iron  were evaluated.  Recoveries  for matrices of fluoride
up to 100 ppm  and phosphate up to 1000 ppm were over 93%.  The  composite sample
containing 1 ppm of Cu, Pb, Mn,  Ba,  and 10  ppm  of  Ca and  Fe had 100%  recovery.
The samples  used for preparing the sulfate and nitrate matrices were too high in
residual background  mercury to analyze.  The determination of mercury in a 1 ppm
gold matrix  had a pronounced interference  that was  thought to be from a reaction
of mercury vapor combining with gold from the sample matrix in  the gas/liquid
separator during analysis.   A  sample matrix containing  silver  at 1 ppm also
exhibited this effect, but to a lesser extreme.   Potential  interferences from
volatile  organic compounds were  also evaluated using the  bromate digestion.
Recoveries were 100% for sample  matrices  containing  up to 50 ppb chloroform or
500 ppb toluene.
                                   534

-------
                         Table 4:  Interference Study
Interferent
Fluoride
Sulfate1
Nitrate1
Phosphate
Gold2
Silver2
Composite:
Ca
Cu
Pb
Mn
Ba
Fe
Chloroform
Toluene
Level
SOppm
lOOppm
lOOppm
SOOppm
lOOppm
lOOppm
SOOppm
lOOOppm
Ippm
Ippm
lOppm
Ippm
Ippm
Ippm
Ippm
lOppm
5ppb
25ppb
SOppb
SOppb
2 SOppb
SOOppb
Recovery
98.2%
101.4%
115.4%
197.6%
151.5%
93.0%
97.0%
100 . 3%
76.1%
182.2%
101.6%
98.9%
102.1%
102 . 6%
100 . 9%
101.9%
99.1%
1    Sulfate and  nitrate  solutions obtained from an  outside  source  contained
     high levels of mercury.

2    Chemical Interference

ppm = mg/L     ppb
                                     535

-------
SINGLE LABORATORY VALIDATION:

The single lab validation study was completed.   Precision and accuracy data were
generated and the linear range  of  the method was  determined to be  25 Mg/L-
Recoveries for duplicate samples  fortified at 25, 50 and 100 ng/L and precision
for six replicate samples fortified at 50 ng/L were determined for each of the
six aqueous matrices selected for the  study.  The Trade Effluents (T&E) primary
effluent  matrix  yielded  lower   recoveries  than the  other matrices   and  the
recoveries  also had  a  higher standard  deviation.    Matrix  interference  was
suspected in this case,  so a standard additions  curve was constructed using this
matrix.   The curve generated showed  no  interference  from the matrix.   Some
laboratories  have claimed difficulty  when using  the bromate-bromide digestion
procedure for samples  high in organics.  Further studies with the T&E matrix are
planned using a modification of the digestion procedure similar to one for trade
effluents described by  Yorkshire Water Authority in the United Kingdom.  This
procedure uses more vigorous digestion conditions.

DISCUSSION:

The determination of total mercury by automated cold vapor atomic fluorescence
spectrometry has a linear range approximately 2  ng-Hg/L to  25 /ig-Hg/L.  The MDLs
as calculated are as follows:

             METHOD DETECTION LIMITS FOR MERCURY

               MATRIX	ng/L

               Reagent water              1.8
               Florida marsh water        3.3
               Synthetic seawater         2.6

The digestion procedures evaluated seem equivalent except  sea water where  the
bromide/bromate reagent  tends to perform better  than the permanganate procedure.
The statistical approach requires  that  the interferant studies should be resumed.
This will include field testing of many kinds  of sample matrix and ruggedness
testing.

The gold  amalgam accessory  may  be useful  when dealing with  certain  kinds  of
interferences (organics  and inorganics) not yet identified.   Interferences  may
be eliminated by preconcentration of  mercury  on gold.  Gas vapors  containing
interferences are passed through  the instrument before analyzing mercury.   The
gold/mercury amalgam trap is heated to release the mercury for cold vapor atomic
detection analysis.   The gold amalgam  accessory would be  inserted  between  the
mercury vapor generator and the  fluorescence  detector.    The future  of  this
accessory will be evaluated during short-term and long-term stability  studies.
These studies are needed for this accessory before its  inclusion in an official
USEPA method for Total Mercury.
                                   536

-------
ACKNOWLEDGEMENTS

The  following individuals  are  acknowledged  for  their contributions  to this
project: Professor Peter Stockwell, Paul Stockwell and Dr. Warren Corns, (P.S.
Analytical Ltd.,  Kent, UK)  and Jim Coates  (Questron Corporation, Princeton, NJ)
are thanked for their technical support.  Dr. Ron Jones (Florida International
University, Miami, FL) is gratefully acknowledged for providing the surface water
sample from the Everglades National Park, Florida.

REFERENCES

1.  Selikoff, I.J. (Editor-in-Chief); "Hazards of Mercury", Environmental
Research. An International Journal of Environmental Medicine and Environmental
Sciences, Vol. 4, No. 1, Mar. 1971.

2. Jacobs, M.B.;  Yamaguchi, S.; Goldwater, L.J.; Gilbert,  H.;  Amer. Ind. Hyg.
Assoc. J., Vol. 21, 1960.

3. Hatch, W.R.; Ott, W.L.;  "Determination of sub-microgram quantities or
mercury by atomic absorption spectrophotometry" Anal. Chem. Vol.40, pp.2085-
2087, 1968.

4. Kopp, J.F.; Longbottom,  M.C.; Lobring, L.B.; " 'Cold Vapor' Method for
Determining Mercury"; Jo. AWWA, Vol. 64. No.l, Jan.  1972.

5. U.S. EPA Method for Chemical Analysis of Water and Wastes.  EPA-600/4-79-
020, Revised March 1983, Method 245 (Total Mercury).

6. Fitzgerald, W.F.; Clarkson, T.W.; "Mercury and Monomethylmercury: Present
and Future Concerns", Environmental Health Perspectives, Vol.  96,  pp. 159-166,
1991.

7.  Slemr, F.; Langer E.; "Increase in global atmospheric concentrations of
mercury inferred from measurements over the Atlantic Ocean", Nature, vol. 355,
Jan. 30, 1992.

8.  Swain, E.B.,  Engstrom,  D.R.; Brigham, M.E.; Henning, T.A.; Breonik, P.L.;
"Increasing Rates of Atmospheric Mercury Deposition in Midcontinental North
America", Science, Vol 257, pp.784-787,  Aug. 1992.

9.  Sorensen, J.A.; Glass,  G.E.; Schmidt, K.W.; Huber, J.K.; Rapp  Jr.,  G.R.;
"Airborne Mercury Deposition and Watershed Characteristics in Relation to
Mercury Concentrations in Water, Sediments, Plankton, and Fish of  Eighty
Northern Minnesota Lakes";  Enviro. Sci.  Technol. Vol. 24,  No.  11,  1990.

10.  Stober, J.;  Hicks, D.; "MERCURY CONTAMINATION IN THE EVERGLADES
ECOSYSTEM", A Plan of Study for USEPA Region IV, Atlanta,  Ga.; Prepared by
Environmental Services Division, Ecological Support Branch,  Athens, GA.,
Draft-Copy 1992.
                                   537

-------
11. S.N. Deming, S.L. Morgan; "Experimental Design: A Chemometric Approach",
Elsevier, Amsterdam, 1987.

12.  Yorkshire Water Methods of Analysis, 5th Ed. 1988.
(ISBN 0 905057 23 6)

13. Farey,  B.J.; Nelson, L.A.;  Rolfe M.G.;  "A rapid technique for the
breakdown of organic mercury compounds in natural waters and effluents."
Analyst, 1978, 103.  656.

14. Nelson,  L.A.;  "Brominating  solution for the preconcentration of mercury
from natural waters.",  Anal.  Chem.,  1979, 51,  2289.

15. Stockwell P.B.  and R.G.  Godden,  J.  Anal.  At.  Spec, 1989,  4,  301.
                                  538

-------
77                DETERMINATION OF TETRAETHYLLEAD
                                 IN GROUND WATER

      Dante J. Bencivengo, Bruce N. Colby and C. Lee Helms, Pacific Analytical, 6349 Paseo
      Del Lago, Carlsbad, California 92009

      Abstract: A chemical analysis method for determining tetraethyllead (TEL) in ground
      water has been devised based on purge and trap GC/MS. It incorporates TEL-d^ as an
      internal standard and naphthalene-d8 as a surrogate. The method detection limit (MDL)
      has been calculated as 2.1 |ig/L and recoveries averaged 102 percent.

      INTRODUCTION

      Organic lead  compounds  constitute  the  largest  single  application of  industrial
      organometallic chemistry. Of those generated commercially, tetraethyllead (TEL), an
      octane-boosting agent, has been produced in the largest quantity. DOT classifies TEL as
      a Class B poison. TEL has a strong affinity for lipids and is known to affect the central
      nervous system causing symptoms such as fatigue, ataxia,  psychosis and convulsions.
      Due  to the  quantities produced  and its highly  toxic  nature, TEL  is a  potentially
      significant threat to the environment.
      At present there are no "official" TEL analysis methods. The most common approach to
      inferring TEL levels, especially those presumably  associated with  leaking underground
      gasoline tanks, is to determine  "organic lead" via  solvent extraction using xylene
      followed  by reaction iodine,  then  tri-capryl methyl ammonium  chloride in methyl
      isobutyl ketone (MIBK);  finally, elemental lead is determined using flame AA. The
      detection limit using this method is only about 100 (Xg/L and it is nonspecific.

      Because tetraethyllead has a vapor pressure exceeding that of naphthalene, it is amenable
      to analysis using a purge and trap gas chromatography based analysis system.

      METHOD SUMMARY

      A 25 mL aliquot of ground water is spiked to 50 (ig/L each with tetraethyllead-d20 and
      naphthalene-dg. It is then subjected to GC/MS analysis  using conditions equivalent to
      those identified in Method  8260. The GC/MS is calibrated with  standard solutions
      containing 10,  20, 50, 100 and 200 jig/L of TEL  and naphthalene-d8, plus 50 jig/L of
      TEL-d2o- All quantitation and quality control are performed per Method 8260 except that
            ao is used as the internal standard and naphthalene-dg is the surrogate.
      EXPERIMENTAL

      Data were acquired using an OI purge and trap interfaced to a Fisons MD800 GC/MS
      using a 20-to-l split where the transfer line connects to the head of the GC column. A
      60 m x 0.32 mm ID, 1.8 micron film DB-624 column was employed using a temperature
                                             539

-------
o
program which included a 4 min initial hold at 40 °C, then a program of 8 °C/min to 200
°C and a final hold of 10 min. This arrangement provided good performance without the
added complexity of cryogenic trapping. All  data reduction and report generation was
performed using the Lab-Base data system.
Neat TEL was  obtained from Ethyl Corporation, TEL-d20  was obtained from  Merck
Isotopes (Merck stock is now available  from  Cambridge Isotopes), and naphthalene-d8
was obtained from  Cambridge Isotopes. An MDL study was carried  out using seven
laboratory water samples spiked at 5 jig/L.  Approximately 60 field samples from a site
suspected  to involve TEL contamination  were analyzed.  Matrix  and matrix spike
duplicates were all spiked to 50 }ig/L.

RESULTS AND DISCUSSION
Mass chromatograms showing the elution of TEL, TEL-d20 and naphthalene-d8 are given
in Figure 1 for a 50 ng/L standard solution. Peaks shapes are good and retention times
are reproducible.  Mass spectra for TEL and  TEL-d20 are shown in Figure 2.  Masses
chosen of quantitation were 295 for TEL and 310 for TEL-d20.
In performing the  initial  analyses it was noted that both TEL and TEL-d20 were
apparently being lost over time as they sat in the  autosampler. Initial suspicions were that
the TEL was decomposing. This was difficult to  justify, however, because TEL could be
detected in ground water samples which most likely were many years  old. Subsequent
review of the data suggested  that ethyl groups were exchanging between labeled and
natural TEL. If a  TEL/TEL-d20 spiked sample was allowed  to stand for several hours,
spectra for TEL-dJ5i TEL-dio and TEL-d5 would start to appear. This exchange  process
was not significant when samples were allowed to stand less than an hour prior to
initiating the purge cycle. Consequently, all  samples were purged  immediately after
spiking.

A typical calibration curve is shown in Figure  3. Response factors are constant to within
25% RSD, but a  marked tendency for the  response factor to  increase with increasing
concentrations  of  TEL was noted.  Consequently, calibration via linear regression  is
recommended if accurate results are more  important than strict adherence to SW-846
protocol.

Matrix spikes  and  matrix spike duplicates using  field samples  yielded  an average
recovery of  102% and an average difference of 1.5% based on three sets  of data.  The
method detection limit study provided a calculated MDL of 2.1 (ig/L.

TEL was detected in field samples even  when they were very highly contaminated. The
example shown in  Figure 4 is for a well water sample with 32 jig/L of TEL. The primary
reason TEL can be  measured  interference-free in samples like these  is due to the high
masses used for quantitation.
                                       540

-------
        Figure 3 - TETRAETHYLLEAD CALIBRATION CURVE
          250
          200 -
          150 -
        on
        
-------
SUMMARY
                                                                             X
Purge and Trap GC/MS is a viable technique for determining tetraethyllead in ground
water. It can be both accurate and precise if care is taken to avoid ethyl group exchange
between labeled internal standard and natural material.

                 Figure 1 - TETRAETHYLLEAD STANDARD

                              Pacific Analytical
    Sample: USTD885
                                                         Instrument: HP8BB
    7K05A8?
    X.FS-
      a
    100
      0
    100
    X.FS
    Son
                          26,06
                                TEL-d2B
26.63
                                   TEL
                         30,53
                                          NaphthaIene-d8
                        2608
                                  2800
                                             38O0
                                                        3200
                                                                      33792
                                        24!
                                       8888
832
X
                                                                  3400
               Figure 2 - TETRAETHYLLEAD MASS SPECTRA
                              Pacific Analytical
Samp
7K8£
ioa-
X.FS
•«xx:
iaa-
XFS-
0-

lo: VSTDBBS
;A87'2&46 (26.852) REFINE
TEL-d28
21
206
« 153 |
i« 	 188 130 ' 200
TEL
21
206
a 100 i ^a **aa
J* 	 *y^ 	 13S 	 2QQ

,8 24fl
^18

2:
233
18
209

Instrument : HD8B8
12 35328
3 O
308
243 276 306 ?"•
i' |L -281 "I (
258 naa ' T;Q
7 22016
295
293
266 396
264-1-267 /
— 239 	 300 	 358
                                     542

-------
78  USE OF A TELEPHONE SERVICE AND DATABASE TO PROVIDE GUIDANCE AND
     INCREASE PUBLIC INVOLVEMENT IN OSW METHODS
     D. Anderson. R. Carlston, and S. Hartwell, Science Applications International Corporation,
     7600A Leesburg Pike, Falls Church, Virginia 22043, and G. Hansen and K. Kirkland, U.S.
     Environmental Protection Agency, OSW Methods Section, 401 M Street S.W., Washington
     DC 20460.
      ABSTRACT
            The Methods Section of EPA's Office  of Solid Waste (OSW) is  responsible for
      approving and issuing methods used to evaluate solid waste and determine whether the
      waste is hazardous under RCRA. These methods are contained in EPA's methods manual
      'Test Methods for Evaluating Solid Waste, Physical/Chemical Methods", also known as SW-
      846.  Individuals in the Methods Section have specific areas of responsibility for  these
      methods, including organic, inorganic, miscellaneous, and characteristic test methods.  Each
      member receives many inquiries  and comments regarding  method  status,  selection,
      performance, and detection capabilities.  The number of inquiries (particularly phone  calls)
      is often overwhelming and, in order to fully answer a question, more than  one member of
      the Methods Section is frequently involved.

            The Methods Section, therefore, identified a need to more efficiently respond to
      inquiries and to document the transmission of the information for future reference. In order
      to accomplish this task, the OSW Methods Section established a telephone service known
      as the Methods Information Communication Exchange (MICE) Service.  Using a voice-mail
      system, MICE is set up to contain informational messages of the most commonly asked
      methods questions and  record questions and comments from callers regarding technical
      difficulties or issues regarding the use of the SW-846 manual. After compiling the incoming
      messages for one day, the calls are sorted and distributed to appropriate technical staff
      members for a telephone response. Each response is summarized, categorized, and entered
      into a computer database on a periodic basis so that internal reports can be sent to the
      OSW Methods Section.

            Once categorized and entered into the database, the resulting call responses and any
      issues  raised are used internally at the Agency to document typographical errors in the
      manual and recurring technical problems with a particular method. The data are also used
      to record ideas and comments for possible directions of future analytical research and
      manual expansion.  The MICE Service voice-mail system has received up  to 821  calls and
      responded to 349 individuals requesting information in one 4 week period.
                                         543

-------
EXTRACTION OF METAL IONS FROM SOLID MATRICES BY COMPLEXATION SFE

W.  F. Beckert. U.S. Environmental Protection Agency, EMSL-LV, Las Vegas, Nevada 89119 and
Y.  Liu, V.  Lopez-Avila,  and M.  Alcaraz, Midwest  Research Institute,  California  Operations,
Mountain View, California 94043.

ABSTRACT
Complexation combined with solvent extraction has been widely used to extract metal ions from
various matrices.  Despite its good performance, this technique is usually time and labor intensive.
Since we have been working on SFE techniques for other classes of compounds, we  undertook s
feasibility study to determine whether SFE could be used to extract metal ions from solid matrices.
We knew from work published by Wai and coworkers that metal complexes formed with fluorinated
complexing agents such  as bis(trifluoroethyl)dithiocarbamate (FDDC) are soluble  in  supercritical
carbon dioxide;  thus, we were  expecting that  these complexes  could  be extracted  by  SFE.
Experiments were performed to derivaiize the metal ions with FDDC, extract the complexes by SFE,
and then analyze the extracts by gas chromatography with atomic emission detection.  Data will be
presented on the extraction of C\? +, Cd2+, Zn2+, and Co2"1" from spiked filter paper and spiked
soil by complexation SFE and then on  the analysis of the complexes by gas chromatography with
atomic emission detection.
                                         544

-------
80   MICROWAVE SAMPLE PREPARATION FOR MERCURY ANALYSIS VIA
      COLD VAPOR ATOMIC ABSORPTION SPECTROMETRY

      W.G. Engelhart. S.E. Littau and E.T. Hasty,
      GEM Corporation, Matthews, NC   28105

      ABSTRACT
      Mercury contamination in the environment has become an increasing concern due
      to the toxicity of this  element.   The volatility of  elemental mercury  and
      organomercury  compounds requires precautions be taken to avoid losses during
      sample preparation.

      Microwave acid digestion of environmental samples inside  sealed vessels is an
      EPA approved method for metals analysis by AA and ICP spectroscopy. This study
      will describe microwave digestion methodology required for mercury analysis via
      cold vapor AAS.

      Analytical results obtained following microwave digestion of standard  reference
      materials  and real world samples spiked with organomercury standards will be
      presented. Microwave sample preparation methods for SRM 1575 Pine Needles,
      US EPA Fish, US EPA Dried Sludge, effluent municipal wastewater, waste motor
      oil and mixed solvent fuel will be described.
                                       545

-------
                    STUDY OF EPA METHOD 300.0:
             APPLICATION TO HAZARDOUS WASTE ANALYSIS

Harish Mehra, Chemical Waste Management,  Inc., P.O. Box 4249,
Modesto, CA  95352  Tele: 209 491 7511

Douglas T. Gjerde, Sarasep,  Inc., 1600 Wyatt Dr.,  Suite 10, Santa
Clara, CA  95054, Tele: 408 492 1029


ABSTRACT:

One of the first rules of using chemical  analysis in the chemical
waste industry is that standard analytical methodology must be
employed.  The second rule is that even if a standard method is
used, it must be verified.  Analytical data are used to determine
proper disposal, secure information that  may then be used in
litigation, etc.   So it is important to  get the right answer and
be able to prove it.

This work is a study of EPA Method 300.0  ion chromatography
method for anions.  The method has been certified under ASTM
method D4327.  Two commercial columns were studied for
selectivity accuracy and precision.   The  effectiveness of EPA
Method 300.0 was studied using standard reference materials.

Several samples of hazardous waste were examined.   Of particular
interest are halogenated organic solvent  sample containing high
amounts of chloride, fluoride,  and bromide.  Prior to anion
determination,  these sample were digested using a combustion
bomb.  Residues were taken up with an aqueous bicarbonate medium
and the samples analyzed.  Halide values  are used in determining
the feasibility of hazardous waste disposal and recycling.
Furthermore,  the cost of disposal is directly dependent upon the
halogen concentration of the waste.   Bromide containing waste are
more expensive to incinerate than chloride containing waste.
                              546

-------
INTRODUCTION:

Characterization of a waste as hazardous is performed by
regulatory agencies, including EPA, OSHA, and DOT.  Generally
hazardous waste is either corrosive, reactive, toxic, or
flammable.  Once a hazardous waste is generated, it must be
properly managed.  Proper management includes the cost effective,
safe, and legal disposal of the waste.  However, to accomplish
this, the content of the waste must be known accurately-  For
example, whether a waste is incinerated and how it is incinerated
depends on the individual concentrations and total concentration
of the halogens in the waste.

Because of this and of possible litigation considerations, it is
important to use standard analytical methodology.  But even if
standard analytical methodology is used, it is still important to
be able to show that accurate results are being achieved.
Accurate results are gnerally obtained with standard methods such
as EPA method 300.

Ion chromatography is widely used for the analysis of
environmental samples  (Ref. 1-3).  The most common method for
anion determinations in ion chromatography is EPA Method 300.0.
The method has been certified as ASTM Method D4327.  The method
is based on ion exchange chromatography and therefore,  is largely
sample independent and rugged.  The ion exchange column acts as a
buffer to the sample as it is injected.  Thus, retention times of
sample peaks are much more constant (matrix independent)  than
other forms of chromatography when a variety of sample are
injected.

This work examines the accuracy and precision of the method for
organic solvent digests.  Two different analytical columns are
compared.

EXPERIMENTAL:

The ion chromatograph used in this work was a Dionex 4000i
equipped with a AMMS™  suppressor and conductivity detection.
The regenerant was 25 mM sulfuric acid at 2.8 mL/min flow.  A 25
uL loop was used.  Except for a standard digest material used for
column comparison, the normal Method 300.0 eluant concentration
and flow rate was decreased to increase resolution of the anions:
The eluant concentration was 0.9 mM Na2C03 / 0.85 mM NaHC03 and
the flow rate was 1.5 mL/min.  All reagents were analytical or
reagent grade.

Two different analytical columns were used in this study: Dionex
HPIC AS4A™, 250 x 4 mm and Sarasep AN300™ 100 x  7.5 mm.   The  two
columns are designed to operate under identical eluant
concentrations and flow rates.  No changes were made in the
instrumentation.  Each column was equipped with a 5 cm guard
column available from their respective companies.
                              547

-------
All samples and standards were prepared with bomb calorimeter
 (Ref. 4) digestion  (PARR Model 1261 calorimeter).  Approximately
0.5 g of a representative (homogenized) sample was accurately
weighed into a capsule.  Some samples are not highly combustible.
If necessary for combustion, 0.2 - 0.5 g of halogen-free kerosene
was added to the capsule.  All weights were recorded to 3 decimal
places.  To a clean bomb 10 mL of 5% aqueous NaHC03  was added,
coating the internal surface of the bomb.  The capsule was placed
in the bomb seat.  A Ni alloy wire was tied to the ignition
electrodes.  If the sample was liquid, the wire was placed close
the sample, and if solid, the wire touched the sample.  The PARR
cover was screwed tight and about 30 psi of oxygen was charged to
the bomb. After combustion,  the bomb was shaken.  The solution
and 3 rinses were made up to 50 mL.  The sample was filtered 0.45
micron filter prior to injection into the ion chromatograph.

RESULTS AND DISCUSSION:

Digest of sodium salt standards.  The selectivity of two
commercial columns were compared under identical separation
conditions specified by method 300.0.  A liquid standard
containing 10 - 50 ppm of sodium salts of fluoride,  chloride,
bromide, nitrate, and sulfate were digested and injected into the
chromatograph.  The results are shown in Figures 1 and 2.
Figure 1 shows that for the AS4A column, fluoride is non-retained
and elutes in the water dip.  Other materials such as non-ionic
organics and cation are also non-retained and elutes with
 fluoride.  Counter ions are suppressed by the AMMS and will not
affect quantification of fluoride provided the capacity is not
exceeded.  No study has been published on the effect of non-ionic
organics on quantification.

 Figure 2 shows exactly the same separation on an AN300 column.
The selectivity is identical expect that the peak window is
retained.  Non-retained material is elutes well before the
 fluoride peak.  Normally a water dip would elute first, but in
this case the digest contains material that gives a positive
peak.  A unknown positive peak eluting with fluoride will give a
 false positive result.  It is unknown what causes the positive
peak or whether it is present in all digests.

Digests of Standard Reference Materials.  Several organic
standard reference materials  (SRM) were obtained from the
National Institute of Standards and Technology  (NIST) and
Environments Resource Associates  (ERA)  (see Table 1) .  The
samples contains elemental bromine, fluorine, chlorine, and
sulfur; however ions generated in the digest were bromide
 fluoride, chloride, and sulfate.

The results of the ion chromatography analysis are shown in Table
2.  The data are a compilation of the injection of individual
SRMs listed in Table 1.  The data show that the two commercial
column gave virtually identical results.  The concentrations
obtained were all within the accepted range of elemental
                              548

-------
concentrations reported for the SRMs.

Note the precision of the method, measured over a eight day
period, was  < 5%.

Digests of Hazardous Waste Samples.  Three actual hazardous waste
samples were obtained from generator companies.  When the sample
is received, the generator also provides a profile of the
elemental composition.  The types sample received can range from
pesticides, oil, organic flux,  The samples were digested and
analyzed by ion chromatography using the two columns.  The
results, reported as elemental wt/wt percent, are shown in Table
3.  All of the samples were also spiked with a known
concentration of each ion.  The recovery of the standard spikes
ranged from 90 - 113%.  While the precision of the analyses were
< 5%.

The peak was sharper and higher for bromide with the AS4A column.
Therefore, under identical conditions, limit of quantification
(LOQ),  was 0.2% for the AS4A column and was 0.5% for the AN300
column.  On the other hand, the AN300 column produced sharper
peaks for fluoride  (LOQ 0.2%) and chloride (LOQ 0.2%).

All other results were comparable for the two columns.
Table 1.  SRM for Ion Chromatography.

Digest Ion Generated          SRM

Fluoride            Spex Lot tt W5-378
Bromide                       NIST #3184
Sulfate                       NIST #2684A
Chloride                      ERA #03042
                              549

-------
Table 2.
        ANALYTICAL RESULTS OF  STANDARD REFERENCE MATERIALS
                     ON TWO DIFFERENT COLUMNS
Sample ANIONS IN PPM


Standard
Reference
Materials
(NIST) from
Environment
Resource
Associates
cr
AS4A
255




246

AN300
236




246

Br~
AS4A
1012




—

AN300
1012




—

S042'
AS4A
8.41




8.20

AN300
8.32




8.32

F
AS4A
1.57




1.70

-
AN300
1.35




1.51

Table 3.  ANALYTICAL 1C RESULTS ON TWO  DIFFERENT COLUMNS
Sample ANIONS IN PERCENTAGE

Rinse water
containing
oil and DDT
% Error
DY-90
% Error
Organic Flux
% Error
Cl
AS4A
<0.5
*94%
—
<0.5
*92%
—
2.03
•"106%
2.3
AN300
0.38
*96%
3.8
0.437
*94%
—
2. 18
*112%
2.6
Br'
AS4A
<0.2
*107%
—
<0.2
*108%
—
<0.2
*108%
—
AN300
<0.4
*108%
—
<0.4
*109%
—
<0.4
*111%
—
F-
AS4A
<0.6
*97%
—
<0.6
*99%
—
<0.6
*102%
—
AN300
<0.2
*99%
—
<0.2
*105%
—
<0.2
*109%
—
SO <2~
AS4A
2.89
*95%
3.2
<0.6
*92%
—
<0.6
*113%
—
AN300
2.79
*94%
3.9
<0.6
*95%
—
<0.6
*108%
—
        % Accuarcy
                              550

-------
ACKNOWLEDGEMENTS:

The authors would like to  express  their  appreciation to Dr.
Nilesh Shah, Ghee Lee, and Gayle King  for  their  input in this
paper.
REFERENCES:

1.   D. T. Gjerde and  J. S.  Fritz,  "Ion  Chromatography"  2nd ed.,
Huethig: Heidelberg and New  York  (1987).

2.   R. J- Joyce and A. Schein,  "1C: A powerful  analytical
Technique for environmental  laboratories" Am.  Lab.,  11  (1989)  46,

3.   W. T. Frankenberger,  Jr.,  H.  C. Mehra,  and  D.  T. Gjerde,  J.
Chromatogr. 504  (1990) 211.

4.   ASTM D240-87.
                              551

-------
                         Cl
in
01
                               DIGEST STANDARD

                               Dionex AS4A
                               Br
                                                SO,
                  Fig. 1
                                                 8 minutes

-------
Ul
U1
w
                              DIGEST STANDARD

                              Sarasep AN300 ~* n
                           Cl
SO4
                                                 10 minutes
                  Fig. 2

-------
The Effects of Size Reduction Techniques on TCLP analysis  of
                     Solidified Mixed Wastes
R. D. Thiel, Research Engineer,  EG&G Rocky Flats,  Inc. P. 0. Box 464,  Golden, CO
80401.
The Rocky Flats Plant  (RFP)  is presently engaged in research  and
development activities for the stabilization of mixed wastes.   Two
stabilization   technologies   (microwave   melting   and   polymer
encapsulation) are being developed to process various waste forms
produced at  RFP.   At the Nevada Test Site storage  facility one of
the acceptance criteria  for  low level  mixed waste is passing  the
Toxicity Characteristic Leaching Procedure  (TCLP).

The microwave melting and polymer encapsulation processes rely, at
least partially, on a physical separation  of toxic  metals from  the
environment via formation of a monolith containing the mixed waste.
The most important step,  but least  specified,  in the  TCLP analysis
of these types of waste forms is size reduction.

This  paper  is a  study  of  the   effects  of several   different
approaches  for size  reduction of  the  products  of   the above
mentioned  waste stabilization  processes.    Inductively  Coupled
Plasma Atomic Emission  analysis  results for several TCLP metals  are
presented to prove how the  size reduction step drastically  affects
the TCLP analysis results.   Also presented  are results  that show
how the three stabilization techniques respond differently to  the
size reduction step of the TCLP analysis.
                             554

-------
AIR AND GROUNDWATER

-------
                            CANISTER ANALYSIS BY USING
         GAS CHROMATOGRAPHY/fflGH RESOLUTION MASS SPECTROMETER

Jong-Pvng Hsu. Director, Greg Miller and Joseph C. Pan,
Department of Environmental Chemistry
Southwest Research Institute
6220 Culebra Road
San Antonio, Texas 78228-0510

       Summa*-passivated canisters used for the collection of air samples have distinct advantages
over tenex and charcoal tubes since the whole air sample is collected with long storage stabilities and
allows for multiple and flexible analysis schemes to be used.

       When very  low  detection limits  are  required, pressurization of the canister is generally
required  which  must use specialized and expensive sampling devices  which will pressurize the
canister above atmospheric pressure. A simpler approach and the preferred one is to do the sampling
in evacuated  canisters using inexpensive flow controllers and sample up to atmospheric pressure.
Clean humidified air is when added in the laboratory to bring the pressure above atmosphere and the
sample is when concentrated cryogenically before analysis.  Addition of humidified air dilutes the
sample and raises the detection limits and therefore places sensitivity demands on the instrumentation
when low detection limits are required.

       A gas chromatography/high resolution mass spectrometer has always offered better sensitivity
and selectivity when low resolution quadrople instruments. Operating at resolution of 3000 in full
scan-mode has quite a number of distinct advantages in highly contaminated air samples since due
to the high resolutions used,  it is possible to detect target analysis even in the presence of high
background  contaminants.   We  have found that  several modifications are required in sample
introduction to  prevent  high voltage flashovers to occur. The sensitivity and selectivity of the
approach is indeed improved and the results will be presented.
                                        555

-------
 AN EFFECTIVE MOISTURE CONTROL SOLUTION  FOR THE GC/MS
 DETERMINATION  OF VOCS  IN CANISTER SAMPLES COLLECTED
                         AT  SUPERFUND  SITES

Michael G. Winslow. Manager, Organic Analytical Division, Dwight F. Roberts, Manager,
GC/MS  Department,   and  Michael  E.  Keller,  Supervisor, Air  Toxics  Group,
Environmental Science &  Engineering,  Inc.  (ESE),  P.O.  Box  1703, Gainesville, FL
32602.

ABSTRACT

EPA Compendium Method  TO-141 has become the preferred guidance method  for the
sampling and analysis of volatile organic compounds (VOCs) in ambient air samples .  A
Contract Laboratory  Program  (CLP)  protocol2 for  the analysis  of air  samples  at
Superfund sites, which is based on TO-14, has recently  been drafted by the USEPA .
Like TO-14, the CLP draft  protocol recommends cryogenic preconcentration of the air
sample followed by thermal desorption into a chromatographic  system.  Because
moisture in the air sample is cryogenically preconcentrated  along with the VOCs,  it will
also delivered into the chromatographic system during themal  desorption.  The moisture
can  adversely  affect   chromatography  and/or detector  sensitivity and reliability.
Therefore, unless the  moisture  in the sample is  removed  or  significantly reduced,
sample size  must  be limited with  consequent  limitations to method  sensitivity.
Compendium  Method TO-14 and the proposed CLP  analytical protocol  recommend a
Nafion  dryer for moisture removal.  However, this technique  limits the method  to  non-
polar analytes  and introduces potential carryover and  reproducibility problems.
Because many polar VOCs are of great interest to the EPA, a canister analysis procedure
that  allows these  analytes  to  be determined with acceptable precision  and accuracy
should  be preferred over  an  approach  that eliminates  that possibility.  This paper
decscribes a commercially  available system that effectively controls moisture in  large
volume air samples, thereby allowing both non-polar and polar VOCs to be measured at
very low concentration levels with acceptable precision and accuracy.

INTRODUCTION

In response to CERCLA and SARA mandates for assessment of potential air emissions and
air quality impacts of toxic contaminants before and during Superfund site remediations,
the  USEPA's Analytical Operations Branch (AOB) within the Hazardous Site Evaluation
Division (HSED) of athe Office of Emergency and Remedial Response (OERR) has drafted
a CLP SOW for the analysis of air samples collected at Superfund sites. One of the four
analytical protocols which has been drafted for the proposed  CLP air program involves
the  determination of VOCs  in whole air samples collected in  passivated  stainless-steel
canisters.  The CLP draft protocol is based on Compendium Method TO-14 and expands
the  target compound list from 41 to  61  analytes, including  several polar  VOCs.  Air
samples  are  collected in  6-L canisters, 500 mL  of which is preconcentrated by
collection on a cryogenically cooled trap, revolatilized  and backflushed onto a capillary
GC  column for separation, and quantitated with a mass  spectrometer operated  in the
full-scan  mode.   The full-scan mode is  required  in order to tentatively identify  non-
target or unknown responses.
                                      556

-------
The CLP draft protocol recommends removal of excessive moisture from the sample gas
stream with a Nafion® dryer,  if applicable.  Excessive moisture can cause the cryotrap
or GC capillary column  head to plug with ice, degrade chromatographic resolution,
especially of the early eluting gases, and cause MS source overpressure. However, use
of a Nafion® dryer as a  moisture reduction  technique limits quantitative determinations
to non-polar VOCs only, because polar VOC compounds are coincidentally removed, or
partially removed,  from the  sample stream.  Because the protocol requires a 500 mL
sample to be analyzed, a Nafion® dryer will have to be employed for most environmental
air samples.   Analysis of sample volumes greater than  100 ml  generally   requires
moisture reduction.   Unfortunately the proposed  target compound  list does contain
several polar  VOCs.  This apparent conflict is not clearly addressed by the protocol.

This paper  presents results of a  preliminary  evaluation of an  Entech  Model 2000
preconcentrator  interfaced  to  a  Hewlett-Packard  5890   gas  chromatograph
(GC)/Finnigan Incos 50 mass spectrometer (MS).  The analytical  system can reliably
measure VOC concentration  levels as  low as 0.1  ppbV while  avoiding the typical
problems caused by the introduction of excessive moisture into the analytical system.  At
the same time, the preconcentrator system allows both non-polar and  polar VOCs to be
determined for target and unknown compounds alike.  The components and operating
configurations of the system are described and the results of system performance  and
method detection limit (MDL) studies are also presented.

EXPERIMENTAL

Canister Cleaning and Leak Checking

To determine ambient air VOC concentrations with  accepable precision and accuracy at
ppbV and sub-ppbV concentration levels, it is imperative that the  laboratory  establish
the cleanliness and integrity  of each sampling canister.  ESE's laboratory in Gainesville,
Florida, utilizes  an  Entech Model 3000 Canister Conditioning System for cleaning  and
leak checking canisters prior to sampling and analysis.

Canister leak checking  is  performed  prior to  canister cleaning.  The canisters  are
connected to the system manifold via 1/4-inch stainless-steel tubing and  the  canister
valves are left closed. A high vacuum pump is engaged and the  vacuum pressure is
monitored using an in-line  pirani  gauge (0-2000 mtorr sensor).   If  a  significant leak
exists,  it will  manifest  itself by not allowing the system to quickly evacuate to
approximately 0.3 torr,  and/or the system pressure will quickly  rise once  the high
vacuum pump is disengaged.  Quick disconnects are situated between each canister and the
manifold to allow easy isolation of individual canisters which may have leaking valves.
After the repetitive fill/evacuate cycling is  completed, the high vacuum pump is used for
final canister  evacuation.  Leaks will not allow the canisters to achieve  a vacuum reading
of approximately 0.5 torr.

Four to eight 6-liter canisters are cleaned simultaneously over a 4-hour period through
automatic, unattended cycling between canister filling and evacuation  modes. Each of the
canisters is heated to 100°C for the duration of the  cleaning process.  A  rough-vacuum
oilless diaphragm pump  cycles  between  eight 8-minute fillings  and  eight  8-minute
evacuations.  Nitrogen vent  gas from a liquid nitrogen tank is used for the fill gas and a
humidification chamber containing  ASTM Type  II  HPLC grade water is used to  add
                                        557

-------
moisture to the fill gas in order to assist in the displacement of VOCs from the interior
surface  of the canisters.  In the fill mode the canisters are pressurized to about 25 psig.
After completion of eight cycles with the low-vacuum pump, the canisters are evacuted
with a high-vacuum oil-based pump for about 75 minutes to an  absolute pressure of 0.5
torr, which is measured by an in-line pirani gauge.  A cryogenic trap containing liquid
nitrogen is placed between the pump and the  system controller to keep oil vapor out of
the system.

At the completion  of the cleaning cycle, the canisters are pressurized  to 30  psig with
humidified nitrogen and analyzed by GC/MS. No target analyte should be detectable at or
above the lowest calibration standard (0.1  ppbV for this  study)..   After analysis, each
canister is re-evacuated to 0.5 torr.

Standards Preparation

Working calibration  standards were prepared  in canisters  at  a minimum of five
concentration  levels  in the range of 0.1  - 15 ppbV.  These working standards  were
prepared from stock  calibration mixtures containing approximately 100 ppbV of each of
the target analytes  in  dry nitrogen in high  pressure  (2000  psi)  cylinders.  These
calibration mixtures were purchased from Scott Specialty Gases   and are certified  to ± 5
% for each of the VOCs.

The working  calibration standards were  generated  by  dynamic flow dilutions of the
purchased mixtures  with cleaned,  zero grade nitrogen humidified  with ASTM Type  II
HPLC-grade water.  ESE  utilizes  an Entech Model 4560 Dynamic  Dilution System for
this process.   This system utilizes up to six mass flow controllers for  simultaneous
blending from multiple cylinders.  Canisters are filled  to 15-25 psig.  Newly  prepared
calibration standards  are allowed to equilibrate at least  24 hours.

The 40  compounds listed in Compendium Method TO-14 (ethyl toluene was not included)
were evaluated for this paper.

Instrumental  Analysis

Canister samples are analyzed with an Entech Model  2000 automated VOC cryogenic
preconcentrator equipped  with an  Entech  2016 16-position autosampler  manifold and
interfaced to a HP 5890 GC/ Finnigan INCOS 50 MS operated in the full scan mode (m/e
35 to 270 amu).  Since the preconcentrator system  is under software control  (IBM
compatible),  a QA/QC report  is a standard  feature  for documentation of actual run
conditions.

Canisters are attached  to the manifold with 1/8 inch stainless-steel tubing.  A leak check
of the system is then performed.   Sample flow is set under mass flow control to 150
mL/min. to give  an  integrated total volume of 1000  ml_.   Both subambient and
pressurized samples can be analyzed because  the system employs a mechanical pump to
draw samples across the  preconcentor.  The  internal  standards, bromochloromethane,
1,4-difluorobenzene,  and d5-chlorobenzene,  are  also added to each analysis at 150
mL/min. to a total volume of 100 ml.  All rotary valves  and transfer lines exposed to the
sample  matrix are heated at 70-100°C.
                                        558

-------
The  Entech preconcentrator is configured with a 3-stage trapping approach which allows
a large sample volume to be concentrated and the moisture and carbon dioxide removed
without significant loss of non-polar or polar VOCs.  Entech refers to the technique  as
microscale purge and trap because it is based on  classical purge and trap principles.
(15-20 u.L of water is purged with  30-40 ml of nitrogen  instead of 5000 jiL  being
purged with  400 mL of  nitrogen.)    The first  stage is  a high-volume cryogenic trap
(1/8-inch nickel tube containing glass beads) cooled during sampling to -150  °C with
liquid nitrogen. After cryotraping 1000 mL of air sample,  the first stage  trap is heated
rapidly to room temperature and slowly purged with about  mL of nitrogen to  transfer
the trapped VOCs to a second stage  trap (1/8-inch nickel tube) containing hydrophobic
Tenax TA  held at 0°C.  Only the amount of water vapor that can saturate 30-40 mL of
nitrogen  (.7 - .9 jiL) will  be delivered  to the second  stage Tenax trap. Carbon dioxide
and some the water will not be trapped by the Tenax. The second stage trap is then heated
to 170°C and backflushed with  helium  to a  third stage.megabore  focusing trap cooled
with  liquid  nitrogen to -150°C.  After heating and backflushing  of the  second stage trap
is complete,  the third stage trap is then very rapidly heated to above 100°C to allow a
rapid injection of the VOCs onto the GC analytical column.

The  target analytes are separated on  a DB-1 fused silica capillary column, 75 m x 0.32
mm  I.D., with a 1 micron film thickness.   The chromatographic  run  is started at 35°C
and  held for 5 min.   A subambient starting  temperature  is not required because of the
rapid transfer of VOCs from the preconcentrator's  third  stage megabore focusing trap  of
the Entech preconcentrator to the  head of the GC column.  Sharp peaks with 2-5 sec.
widths are maintained,  even  for  early  eluting  compounds.  The column  is  then
temperature  programmed at 6°C/min.  to  180°C and  then at 7.5°C/min. to 225°C.  The
scan rate is approximately  3 scans/sec.  The total  run time for preconcentration of the
sample and chromatographic separation  is  approximately 60  min.   Figure  1 shows a
mass chromaogram of a 1  ppbV standard of target VOCs.

The  MS  scan rate  is approximately 3 scans/sec.  The Incos 50 turbomolecular source
pump draws about 170 L/min.  This pumping rate effectively prevents source pressure
increases from any residual sample moisture.

System Performance and Method Detection Limit (MDL) Studies

The  reproducibility and linerarity of the analytical  system were evaluated over a  two-
week period.  Eight calibration standards and a blank were prepared as described above
in separate  6-liter canisters.  The blank  canister  was prepared with humidified  zero
grade nitrogen.  The calibration standards  contained the 40 compounds listed in  Method
TO-14 and were prepared at concentration levels of 0.1, 0.25, 0.5, 1, 2,  5, 10, and  15
ppbV.  Another canister containing the three internal standards was also  prepared.  All
canisters were attached to the autosampler manifold and analyzed, after instrument
tuning, as described above. Internal standards were  added to each analysis at 5 ppbV.
The  canisters were  analyzed three times, with a week between analytical runs.  Each
analytical sequence lasted about 9 hours after tuning.  Initial canister  pressures ranged
from 10 to15 psig for the first day's run, from 1  to 2  psig for the second day's run, and
from -1 to -2 psig for the last day's run.

In addition to the  reproducibility study  described above, an MDL study was performed.
Lower limits  of detection  for the target compounds were  estimated for each instrument
                                        559

-------
by  determining the  method detection limits  (MDLs)  as  specified  by  the U.S.
Environmental Protection Agency  (U.S. EPA).3 Nine canisters samples were prepared at
about 0.5 ppbV for each of the target compounds used the standards preparation system
described above.   The nine replicate samples were then analyzed and quantitated against
a calibration curve containg eight standards ranging from 0.1 to 15 ppbV.  The compound
responses divided  by the appropriate  internal standard  response were  plotted against
concentration.   The  MDL  for each target compound is calculated by  multiplying  the
standard deviation of the seven replicate concentration measurements by  the appropriate
one-sided  t-value corresponding to  n - 1  (8) degress of freedom.  The corresponding t-
value for seven measurements is 2.897.

RESULTS  AND DISCUSSION

The reproducibility, or precision,  of the analytical system was evaluated by comparing
the responses of the target compounds and their corresponding internal  standards over
the three  analysis days at  different calibration  standard concentrations.  In addition
compound retention time variability  was evaluated.  Table 1  summarizes response and
retention data for vinyl chloride,  benzene, and hexachlorobutadiene (HCBD) at 0.1, 1
and 10 ppbV.   These  compounds represent low boiling (vinyl chloride  = -13.4°C),
middle boiling (benzene = 80.1°C, and  high boiling (HCBD = 215°C) VOCs, respectively.
For  each  of the three compounds listed in  Table 1, the relative stardard deviations
(RSDs) of  the absolute  responses  for the  three analysis days  over  the three
concentration levels  averaged less than 6 percent .   The relative response factors
(RRFs) averaged less than 5 percent. The RSDs of the absolute responses of each of the
internal standards  for the  three analysis  days and within each  analysis day over  the
three concentration levels averaged  less than  5 percent.  Retention times had ranges no
greater than .03 minutes for the  three  compounds.   Response and  retention time
reproducibility for the other target VOCs were  also very good.

Table 2 summarizes calibration data for vinyl chloride, benzene, and HCBD on the second
analysis day. The compound responses were  very linear over the 0.1 to 2.0 ppbV range,
as measured by either a linear regression curve or the %RSD of the RRFs.

Table 3 lists the calculated MDLs for this study.   The calculated MDLs  averaged 0.11
ppbV for all  the TO-14 compounds.  Seven of the  calculated MDLs were less than 10% of
the 0.5 ppbV nominal concentration analyzed.  This indicates that the replicate analysis
for those compounds should have been targeted  at a lower concentration (probably 0.25
ppbV).  The 0.1  ppbV responses for  all  the  TO-14  compounds  were  well  above
background  noise,  in  the range of 10 - 80,000 area counts.  Compound responses were
very linear over the 0.1-2  ppbV range and generally approached a quadratic fit across
the full  calibration  range (0.1  - 15 ppbV).

Polar VOCs

Although this paper does not discuss experimental results for polar VOC determinations
using the analytical system described above, preliminary analysis of eight polar VOCs of
interest  have been good. The compounds that are currently being investigated are those
listed in the CLP draft protocol target compound  list.  They include acetone, acetonitrile,
acrolein,  acrylonitrile,  methyl  ethyl   ketone,  methyl  isobutyl  ketone,  methyl
methacrylate, and vinyl acetate.  Other  polar analytes will also be examined.
                                       560

-------
Several limited studies have been performed elsewhere that have examined the stability
of selected polar compounds in canisters.  There is sufficient evidence to indicate that
some polar VOCs are unstable in canisters if held over extended periods due to canister
and/or  matrix  effects.    However,  considerable  experimentation  is still  needed,
especially at low concentration levels.

SUMMARY

Routine measurement of non-polar and polar ambient VOCs at ppbV and sub-ppbV levels
by GC/MS-SCAN can be  acheived with acceptable precision and  accuracy if sample
moisture is effectively reduced.  This has been demonstrated with  a  analytical system
comprised of an Entech preconcentrator, a Hewlett-Packard GC, and a Finnigan MS.

ACKNOWLEDGEMENTS

The authors thank Norm Staubly for his  laboratory assistance.

REFERENCES
1.     W.T. Winberry, Jr., NT. Murphy and R. M. Riggin,  "Method TO-14,"
       Compendium of Methods for The Determination of Toxic Organic Compounds in
       Ambient Air.  EPA-600/4-89-017.  U.S.  Environmental Protection Agency,
       Research Triangle Park, NC, 1988.

2.     Statement-of-Work for the Analysis of Air Toxics at Superfund  Sites (Draft),
       "Volatile Organics Analysis of Ambient Air in Canisters,"  U.S. Environmental
       Protection Agency, Washington, D. C.,  December, 1991.

3.     Federal Register. "Definition  and procedure for the detemination of the method
       detection limit," Code of Federal Regulations. Part 136, Appendix B, Oct. 26
       (1984).

4.     M. G. Winslow, D.F. Roberts, and M.E. Keller, "Determination of VOCs in Ambient
       Air at 0.1  ppbV for the Clean Air Status and Trends Network (CASTNet)",
       presented in May 1993 at the EPA/AWMA Internation Symposium on the
       Measurement of Toxic and Related Air Pollutants held in Durham,  N.C.
                                       561

-------
Table 1:  Area Response and  Retention Time Reproducibility of
          Selected  VOCS

10 ppbV
Resp. (xlO3)
IS Resp. (x103)
RRF
Ret. Time
t ppbV
Resp. (x1Q3)
IS Resp. (x103)
RRF
Ret. Time
0.1 opbV
Resp. (xlO3)
IS Resp. (x1Q3)
RRF
Ret. Time
P
vc
2058
764
1.35
4.03

278
808
1.72
4.03
29.7
739
2.01
4.02
ressuriz
2/17/93
Bzn
2379
2078
0.57
10.16

302
2143
0.70
10.15
27.1
1968
0.69
10.15
ed
I
HC8D
883
1665
0.27
28.11

124
1652
0.38
28.11
13.5
1479
0.46
28.10
Atmospheric
2/23/93
vc
2067
702
1.47
4.02

271
762
1.77
4.04
27.9
724
1.93
4.02
Bzn
2256
1947
0.58
10.15

279
2000
0.70
10.16
26.1
1894
0.63
10.14
HC8D
910
1610
0.28
28.11

111
1577
0.35
28.11
13.8
1440
0.48
28.09
Sub-Atmospheric
3/04/93
vc
1996
732
1.36
4.03

290
788
1.84
4.03
31.3
748
2.09
4.02
Bzn
2195
1978
0.56
10.16

300
2094
0.72
10.16
29.8
1985
0.75
10.15
HCBD
843
1582
0.27
28.12

106
1614
0.33
28.11
14.8
1551
0.48
28.11
                                 562

-------
Table 2:  Calibration and  MDL  Results  for Selected VOCs
Calibration:
Standard VINYL CHLORIDE
(ppbV) Response RRF
2 515511 1.67
1 270598 1.77
0.5 141763 1.93
0.25 74681 2.03
0.1 27932 1.89
0 0
LR. Corr. Coeff. = 0.9995
y-lntercept = 6895
Average RRF = 1.86
%RSD Of RRFs = 7.5

BENZENE HCBD
Response RRF Response RRF
575498 0.700 219910 0.340
279430 0.699 111351 0.353
132341 0.672 51646 0.355
68616 0.733 27152 0.371
26112 0.652 13810 0.461
3248 254
0.9993 0.9997
-5397 264
0.691 0.376
4.5 13
                                 563

-------
Table  3:     Calculated MDLs  of TO-14  Compounds
             Compound	.	MDL   (ppbV)

             Dichlorodifluoromethane                             0.10
             Chloromethane                                      0.12
             1,2-Dichloro-1,1,2,2-tetrafluoroethane               0.15
             Vinyl chloride                                      °-07
             Bromomethane                                      0.09
             Chloroethane                                        0-04
             Trichlorofluoromethane                              0.12
             1,1-Dichloroethene                                 0.11
             Methylene chloride                                  0.40
             1,1,2-Trichloro-1,1,1-trifluoroethane                0.20
             cis-1,2-Dichloroethene                              0.07
             1,1-Dichloroethane                                 0.07
             Chloroform                                         0.06
             1,2-Dichloroethane                                 0.06
             1,1,1-Trichloroethane                               0.29
             Benzene                                            0.03
             Carbon tetrachloride                                 0.03
             1,2-Dichloropropane                                0.09
             Trichloroethene                                     0.04
             cis-1,3-Dichloropropene                            0.09
             trans-1,3-Dichloropropene                          0.12
             1,1,2-Trichloroethane                               0.10
             Toluene                                            0.07
             1,2-Dibromoethane                                 0.11
             Tetrachloroethene                                   0.06
             Chtorobenzene                                      0.04
             Ethylbenzene                                        0.06
             m,p-Xylenes                                        0.14
             Styrene                                            0.03
             1,1,2,2-Tetrachloroethane                           0.30
             o-Xylene                                           0.04
             1,3,5-Trimethylbenzene                             0.06
             1,2,4-Trimethylbenzene                             0.13
             Benzyl chloride                                     0.30
             1,3-Dichlorobenzene                                0.05
             1,4-Dichlorobenzene                                0.06
             1,2-Dichlorobenzene                                0.04
             1,2,4-Trichlorobenzene                             0.08
             Hexachlorobutadiene                                 0.09

                                            Average MDL =      0.11
                                     564

-------
Figure
Mass  Chromatogram of a Canister VOC Standard
       RIC                                 DATA: 02179387 #1        SCANS  556 TO 5580
       92/17/93 15:39:09                   CALI: 02179307 #3
       SAMPLE:  1.0PPBU
       CONDS.:
       RANGE:  G   1,6881  LABEL:  N  5,  5.0  QUAH:  A   5,  5.8 J   Q   BASE:  U  29..  3
 RIC
             1000
                2800
3000
17:13
4000
23:05
5000
23:52
                                                                                                                 SCAN
                                                                                                                 TIME

-------
SELECTION CRITERIA FOR GROUND-WATER MONITORING WELL CONSTRUCTION
MATERIALS

J.R. Brown. US Environmental Protection Agency, Washington, DC
20460, L. Parker, US Army Cold Regions Research and Engineering
Laboratory, Hanover, NH  03755-1290,  A.E.  Johnson, Science
Applications International Corporation, McLean, VA  22102

ABSTRACT

This presentation discusses a regulatory perspective on the
selection of ground-water monitoring well  casing materials for
hazardous and solid waste disposal facilities that must comply
with the Resource Conservation and Recovery Act requirements.
The primary focus will center on EPA's current recommendations
for selecting well casing materials,  with  special consideration
given to regulatory requirements,  data quality objectives, common
types of well casing materials and their physical and chemical
characteristics,  and geochemical considerations.   A table
summarizing the recommended selection criteria and a series of
case studies will also be presented.
                             566

-------
           AUTO-GC DESIGN AND OPERATION FOR
    REMOTE UNATTENDED VOC DETERMINATIONS

                   James F. Ryan. Environmental Marketing Manager
                                        and
                         Ian Seeley, Senior Technical Specialist
                            The Perkin-Elmer Corporation
                      50 Danbury Road, Wilton, Connecticut 06897
ABSTRACT

Recent emphasis on the design considerations for ozone precursor analytical systems using gas
chromatography has focused on the practical and logistical difficulties of using liquCid cryogen
for trapping volatile organics.  Conventional approaches to the problems of trapping species in
the range of C2 to C10 require liquid nitrogen for cold-trapping and for sub-ambient operation of
the chromatographic oven.  Liquid nitrogen is not  only expensive, but difficult to supply  in a
reliable fashion for long-term unattended operation at remote locations.

A system has been developed and evaluated for the on-line ambient detection of hydrocarbon
ozone precursors that does not use liquid cryogen.  This novel chromatographic approach has
been used to optimize the separation of C2 to C10 compounds using a pressure switching facility.
Such a chromatography system has the added benefit of stabilizing the column system resulting
in improved retention time reproducibility. This is consistent with the requirements of a robust,
stable system capable of long-term unattended operation.
INTRODUCTION

U.S. Environmental Protection Agency, namely carbon monoxide, sulfur oxides, nitrogen oxides,
ozone, particulate matter less than 10 microns in size, and airborne lead.  Though EPA has
promulgated these levels for some time, a number of urban areas in the United States continue
to be unable to bring their normal levels into compliance with EPA regulatory levels. This "non-
attainment" problem is especially acute with respect to ozone. There are 23 cities in the U.S.
that are considered non-compliant with ozone regulations.

In order to address this problem, the 1990 Clean Air Act Amendments (CAAA) required that EPA
develop an ozone abatement strategy based not only on the chemical species itself (i.e., ozone),
but based as well on hydrocarbon precursors of ozone that react with sunlight and nitrogen
oxides to form the unhealthful, ozone-rich, summer's haze so prevalent in urban areas.

In order to limit the release of low molecular weight hydrocarbon precursors into ambient air, it is
necessary to determine the hydrocarbon source. Therefore, an instrumental approach has been
developed permitting the automatic collection and chromatography of C2 to C10 hydrocarbons
from ambient air.

INSTRUMENTAL DESIGN

This system is based on the Perkin-Elmer ATD-400 Automatic Thermal Desorption instrument
that has a Peltier-cooled adsorbent trap. After being electrically cooled to -30 oC, ambient air is
                                        567

-------
drawn through the trap.  Following  a predetermined sampling time, the trap is automatically
connected to the Perkin-Elmer 8700 gas chromatograph via a 6-port valve, and the trap  is
programmably heated  at 40 oC/sec; hydrocarbons are flashed off the adsorbent and onto a
capillary GC column.

Chromatographic separation is mutti-dimensional, i.e., a combination of a PLOT column for the
extremely volatile low molecular weight compounds and a BP1 boiling point column for the
higher molecular weight species.  All the organic analytes from the trap are transferred to the
BP1 column. For the first 13 to 14 minutes, the effluent from the BP1  column is directed onto
the AI2O3.Na2SO4 PLOT column.  Thus early eluting poorly resolved components from the first
column are transferred for further Chromatographic separation on the PLOT column. After 13 to
14 minutes, the effluent from the BP1 column is switched (using the Deans system) directly to a
2nd FID.  Higher boiling, well-resolved components are analyzed on this second detector and are
prevented from contaminating the PLOT column. For the remainder of the analysis, components
elute simultaneously on both columns. However, as each peak only elutes on one  or the other
column, and not  on both columns, data  handling is not complex.  This system is illustrated in
Figure 1.
I                 Vacuum |_| M
                   Rump   | | C
                                                               Air
                                     Cr»rrl«»r
•»O_ Column
                                              Hpnoumotlc I
                                                Swttcti   I
Figure 1: Schematic diagram of ambient air monitoring system.
The resulting Chromatographic output comes from two GC detectors, one for the PLOT column,
and a second for the BP1  column.  Figure 2 shows a  typical GC output for a 55-component
standard.
                                          568

-------
Perkin Elmer Ozone Precursor
12-


10-

8-
6-j
14-
12-
1 OJ

8-
61
C
joo_O3O

C»bn 1 PLOT

< t
| 	 iLl_JJJj
i
. . . . , , . . , , , r i . | . . . . | i . . . , .
job_O3O
C.bnut2BP-l *^* **:
"\\ "\ »
XV *^^**JI \ ll i ||
27 | 2* | 1 I k 1 ll
, .IklJUJ^




5 5 1O 15 2O 25




*
LI-






«a 44


JL



A

3O
!•
13


,







System
20
-1« '
X^\


1


• i J

» 4
•



93
1


I


\

JJ




*»

2<
JL 	 ^


i
i
iu^__
35 AQ 45


Figure 2: Multidimensional chromatographic output from ambient air monitoring system.

Peak identities are shown in Table 1.
VOCS Chromatographed on PLOT Column
01
02
03
04
OS
06
07
08
09
10
11
12
13
14
Ethane
Ethytene
Propane
Propene
teobutane
n-Butane
Acetylene
trans-2-Butene
1-Butene
cis-2-Butene
Cyclopentane
teopentane
n-Pentane
2-Methyl-2-butene
15
18
17
18
19
20
21
22
23
24
25
26


Cyclopentene
Trans-2-pentene
3-MethyM-butene
1-Pentene
cis-2-Pentene
2,2-Dimethylbulane
3-Methytpentane
2-Metnylpentane
2,3-Dimethylbutane
Isoprene
4-Methyl-1-pentene
2-Melhyl-1-pentene


VOCs Chromatographed on BP1 Column
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
n-Hexane
trans-2-Hexene
cis-2-Hexene
Methylcyclopentane
2,4-Dimethylpentane
Benzene
Cyclohexane
2-Methylhexane
2,3-Dimethylpentane
3-Methylhexane
2,2,4-Trimelhylpenlane
n-Heptane
Methylcyclohexane
2 ,3,4-Trimethylpenlane
Toluene
42
43
44
45
46
47
48
49
50
51
52
53
54
55

2-Methylheptane
3-Methylheptane
n-Ociane
Ethyl benzene
p-Xylene
2-Methylheptane
3-Methylheptane
n-Nonane
I sopropyl benzene
n-Propyl benzene
a-Pinene
1 ,3,5-Trimethylbenzene
b-Pinene
1 ,2,4-Trimethylbenzene

               Table 1: Identification of chromatographic peaks in Figure 2.

The advantages of this instrumental system are that the 55 compounds of interest are separated
cleanly, and without the use of liquid nitrogen.  Neither the VOC collection (using the Peltier-
cooled trap) nor the chromatography (which starts at 40 oC) require liquid nitrogen.

In defining the limits of this system, questions naturally arise  concerning trap breakthrough
volume and overall detection limits.  Figure 3 below shows three of the key C2 compounds,
introduced to the system at 16 to 19 ng in absolute amounts.  After introduction of the three
hydrocarbons, zero grade bottled air was drawn through the Peltier-cooled trap and the organic-
laden adsorbent.  The trap was then heated to release the three VOCs for chromatography. As
shown, acetylene response began to fall  off only above 600 ml of air volume. Therefore, 600
mL of air was chosen as the sample size for analysis.
                                          569

-------
         400000 -i
              C2  HYDROCARBONS
COMPONENT  RECOVERY  v  VOLUME  SAMPLED
            SAMPUNG  TIME: 40min
           16 to  19ng each component
         300000^
         200000-
         100000-
          Ethan* (75«RH)
          EthyUn* (75«RH)
          AMfylww (75«RH)
               200
                FLOW  RATE (mL/mln)
                15           25
                1 ' I '  '.' '  '.' M'
                                                       35
                 600            1000
                SAMPLE  VOLUME (ml)
Figure 3: Determination ot air volume breakthrough values tor acetylene, ethylene and ethane.

That modest sample size notwithstanding, Figure 4 shows that this system achieves low level
quantrtation levels well within the part-per-trillion range. Methyl cyclohexane, the second labeled
peak, has been quantified at 0.06 ppb.
                  TYPICAL LOW LEVEL DETERMINATION
 |
  AlLJL
Figure 4: Illustration of low quantrtation levels achievable with air monitoring system.

As an example of how this system can be used, consider Figure 5.  In this chart, concentrations
of three VOCs, namely isobutane, n-butane and ethane (based on ppb of carbon), in laboratory
ambient air have been determined over a 24 hour period.  Note that the X axis is based on a
starting time of 9:00 AM, i.e., the first series of bars is at 9:00 AM, the second at 10:00 AM, etc.
                                      570

-------
At 6:00 PM, there is a dramatic increase in isobutane.  This is when the  evening laboratory
cleaning crew begins their work, and one can speculate that the isobutane and n-butane rise
might be due to the cleaning reagents they use, specifically if isobutane is a propellant in one of
their spray cans. The butane data, along with charts derived from other hydrocarbon data, is
presented in Figures 5 to 8.
     Diurnal Variations of Ozone Precursors
            HOUR from fcOOun 12/1/82
       Dumal Variations of Cfeone Precursors
                                                          Ham torn Ittam 12/1/n
Fig. 5: Concentration of butanes and ethane
       over 24 hour period.
Fig. 6: Concentration
       and toluene
of
benzene
          Diurnal Variation* of Ozona Praeuraore
              J J J iiJL
              Haora Iran IMim 11/1/tl
                                                      Diurnal Variationa of Ozona Pracuraon
        B^jalii Jj. ..•.
                                                          Ham fram Sfivdm llri/»
Fig. 7: Concentration of nonane and decane    Figure  8:
        over 24 hours.
           Concentration  of  isoprene  and two
           pinenes over 24 hours
It is interesting to compare the four charts, even if explanations for the data are not readily
apparent. As noted above, the butanes might be explained as a propellant used in a spray can.
However, no similar explanation for the presence of nonane, decane, isoprene and the pinenes
readily comes to mind, other than that the engineering laboratory in which this instrument is
located is adjacent  to  a major  roadway.   Investigations to determine the  source  of these
chemicals continue.

It should also be noted that an intrinsic part of this ambient air monitoring system is the ability to
not only collect data at a remote site, but also to transmit information on the  status of the
instrument operation and reports on the collected data  to a central operation.  Using personal
computers,  modem  telecommunications  software  such as  PC-Anywhere, and  Perkin-Elmer
Turbochrom 3 chromatographic data handling software, the status of the ambient air monitoring
system can  be  evaluated, raw data can  be examined, and summary  reports generated for
transmission to a central operation where they can be combined with reports from other stations.
Raw data files, because of their size, would likely be collected once per week during routine
visits to  the monitoring  site.  Chromatographic run and summary reports  could  be  collected
several times a day as needed.
                                            571

-------
SUMMARY

There are unique advantages to an ambient air monitoring system such as that described above.
First and foremost, the system routinely operates unattended.  The instrument used to collect
data for this paper has been in continuous operation for over 1500 hours, and has consumed
only 1 tank of hydrogen. No liquid nitrogen was required and zero grade air is generated using a
laboratory compressor. Secondly, even though the instrument operates unattended, through the
use of appropriate data handling and remote operation software, reports on instrument operation
and chromatographic results can be collected as often as necessary.
                                         572

-------
AUTHOR INDEX

-------
 Author Index
 Author
Paper Number    Author
Paper Number    Author
Paper Number
 Adolfo, N	7
 Alcaraz, M	79
 Allen, R.L	36
 Almeida, M	73
 Almond, R.E	36,38
 Anderau, C	72
 Anderson, D	46,78
 Arrowood, S.P	36
 Baker, R.D	64
 Barnes, K.M	72
 Barton, J	1
 Bath, R.J	10
 Beckett, W.F	79
 Bencivengo, D.J	77
 Berman.R	69
 Betowski, L.D	57
 Bollman.M	31
 Brandenberg, C.W	66
 Brown, J.R	85
 Brown, W.T	 35
 Bruce, M	39,43
 Budde, W.L	44
 Bulman, F	73
 Burkitt, D	39
 Campbell, C.R	64
 Carl, J	39, 61
 Carlson, R.E	55
 Carlston, R	78
 Carpenter, S.C	14
 Caruso, M	2
 Caton, J.E	48
 Chapnick, S	4
 Chau,N	63
 Chen, P.H	51
 Christopherson, M	56
 Cline-Thomas, T	28
 Colby, B.N	50, 77
 Converse, J.C	16
 Cross-Smiecinski, A	18
 Curl, P	30
 DalSoglio, J	31
 Davis, A	12
 Davis, R	76
 Deloatch, 1	6
 DiBussolo, J.M	52
 Diebold, T	26
 Dobb, D.E	16
 Doherty, L.C	49,53
 Dolata, L.A	68
 Dong, M.W	52
 Doskey, P.V	14
 Doster, J.A	76
 Douglass, O.B	9
 Dutta, S	29
Dux, T.P	13'56
Eichelberger, J.W	44
                  Engelhart, W.G	80
                  Erickson, M.D	14
                  Fadeff, S.K	20
                  Farrell, J	3
                  Feeney, M.J	59
                  Forsberg, D	17
                  Fowler, R	32
                  Fredeen, K.J	72
                  Friedman, S.B	36,38
                  Garner, F	5
                  Geier, D.S	61
                  Gere, D.R	42, 53, 54
                  Gillet, C	31
                  Gjerde, D	81
                  Gladwell, A	8
                  Gluodenis, T.J	72
                  Godfrey, C.B	8
                  Goheen, S.C	20
                  Gore, J.C	21
                  Greenlaw, P	10
                  Griest, W.H	48
                  Grosser, Z.A	72, 75
                  Haley, A.E	11
                  Hall, J.R	3
                  Hall, K	13
                  Hansen, G	78
                  Harrison, R.0	55
                  Hartwell, S	78
                  Hastings, M.R	62
                  Hasty, E.T	80
                  Hayes, M.C	37
                  Heithmar, E	69
                  Helms, C.L	77
                  Henderson, G	31
                  Herzog, D.P	37, 58
                  Hill, B.R	3
                  Ho, J.S	44
                  Hooton, D	56
                  Hottenstein, C.S	58
                  Hsu, J.P	83
                  Jackson, K.W	2
                  Jackson, T	46
                  Janiec, G	9
                  Jenkins, T.F	45
                  Johnson, A.E	85
                  Johnson, D.P	36
                  Jourdan, S.W	37,58
                  Jungclaus, G	13, 56
                  Kauffman, J.S	65
                  Keeran, W.S	51
                  Keller, M.E	84
                  Kennedy, M	47
                  Kibler, L	19
                  Killough, B	39
                  King, R	31
                  Kirkland, K	78
                  Kleiser, H.R	30
                  Knipe, C	42, 53, 54
                  Kuehl, M.A	15
                  Larson, A	30
                  Lawruk, T.S	37,58
                  Leavell, J.R	66
                  Lee, H.B	42
                  Lesnik, B	33,47, 63
                  Levy, J.M	68
                  Lillian, D	10
                  Lindahl, P.C	14
                  Linder, G	31
                  Littau, S.E	80
                  Liu, Y	79
                  Lohner, W	21
                  Long, S.E	76
                  Lopez-Avila, V	79
                  Loring, D	4
                  Low, N	49
                  Macdonald, D	17
                  Mapes, J.P	36, 38
                  Marsden,  P	34,46,47, 63
                  Mauro, D	40
                  McCallister, R	16
                  McCormick, E.F	45
                  McCulloch, M	20
                  McDonald, P.P	36, 38
                  Mclntosh, S	75
                  Mehra, H	81
                  Merewether, G	74
                  Meszaros, T.J	16
                  Miller, D.P	71
                  Miller, G	 83
                  Mills, W	22, 28
                  Miyares, P.H	45
                  Mong, G.M	20
                  Moore, M	3
                  Murarka, 1	40
                  Myers, K.F	45
                  Nwosu, J	31
                  Nygaard, D.D	73
                  O'Brien, W	22
                  O'Brien, K	8
                  O'Neil, G	41
                  Organ, B	16
                  Ott, S	31
                  Owens,	J-17
                  Owens, R.G	61
                  Pace, C	57
                  Pan, J.C	83
                  Parker, L	85
                  Parr, J.L	3
                  Parsons, A	67
                  Parsons, C.S	50
                  Pfleeger, T	31
                  Pflug, A.D	14

-------
Author Index
Author
Paper Number   Author
Paper Number   Author
Paper Number
Phinney, D	61
Pipkin, W.	42, 53, 54
Ploscyca, J	3
Pollack, L.P	60
Poppiti.J	20
Potter, B.B	76
Poziomek, E.J	18
Purdy.C	14
Randa, H	56
Randall, L.G	42, 53, 54
Ravey, R.M	68
Ray, E.	21
Richardson, L.A	11
Riga, T.J.	.. 41
Riley, R.G	 20
Roberts, D.F	51, 84
Robertson, G	5
Roby, M	'.	34, 57
Rosecrance, A.	7, 19
Rubio, F.M	58
Ryan, J.F.	86
Salesky, J	 21
Sauerhoff, S	 75
Sauter, R	56
Schenley, R.L	48
Schindler, D.R	51
                 Schmick, M	8
                 Schneider, J	40
                 Schutte, W.E	 14
                 Seeley. 1	86
                 Shah, N	74
                 Shaw, T.A	61
                 Shelow, D.M	59
                 Sklarew, D.S	20
                 Sleevi, P	3
                 Smith, D.V	60
                 Smith, J.S	50
                 Smith, R.K	61
                 Spear, R.D	 10
                 Speis, D	 3
                 Spilkin, A	62
                 Spitz, H	., 21
                 Stapanian, M	5
                 Stewart, T.N	36
                 Straka, M	69
                 Strong, A.B....	45
                 Studebaker, W.B	36
                 Swift, R.P	66
                 Tang, P.H	44
                 Taylor, B	40
                 Thiel, R.D	,	82
                 Thomas, B	20
                 Tomczak, L	21
                 Tracy, A	22,28
                 Tsang, S.F	'.	 34,46,47,63
                 Tucker,  S	;	3
                 Valkenburg, C.A:	35
                 Van Ausdale, W.A	51
                 Van Stiver, C	27,32
                 Vance, L.W	23
                 Varcoe,  F.T	 60
                 Vernon, J	12
                 Vinson,  C	3
                 Vitale.RJ	70
                 Wallace, E	27
                 Weichert, B.A	64
                 Weiss, A.J	55,67
                 Wilborn, D	31
                 Williams, A	13
                 Willig, T	65
                 Winslow, M.G	84
                 Winward, R.T	24
                 Withers, T.A	36
                 Yates, D.A	72
                 Young, M	40
                 Zha, Y	58
                 Zine, T	39

-------