The 14th Annual Waste Testing & Quality Assurance Symposium
               PROCEEDINGS
                                            Sponsored by the
                                          American Chemical Society and the
                                          U.S. Environmental Protection Agency
      July 13-15, 1998 • Crystal Gateway Marriott • Arlington, VA

-------
           Proceedings

         The Fourteenth Annual

Waste Testing & Quality Assurance
           Symposium
            (WTQA '98)
           July 13-15, 1998

        Crystal Gateway Marriott
            Arlington, VA

-------
                                          HIGHLIGHTS

                     14th Annual Waste Testing & Quality Assurance Symposium
                                            (WTQA '98)
                      Using A Performance-Based Measurement System (PBMS)

WTQA '98, cosponsored by the American Chemical Society (ACS) Division of Environmental Chemistry and the
U.S. Environmental Protection Agency (EPA) Office of Solid Waste and Emergency Response (OSWER), will be
held at the Crystal Gateway Marriott in Arlington, VA, July 13-15,  1998. This year's symposium continues the
partnership between the regulated community and their supporting laboratories and state and federal regulators
from the Research Conservation and Recovery Act (RCRA) and the Comprehensive Environmental Response
Compensation and Liability Act (CERCLA) programs.

Highlights
EPA is actively  working to implement the President's program for "reinventing" government and reforming
regulatory policy. As part of this endeavor,  EPA has been trying to break down barriers to using new monitoring
techniques. One barrier that OSWER is tackling is the requirement to use specific measurement methods or
technologies  in complying with  Agency regulations.  EPA's Environmental Monitoring Management  Council
(EMMC), members of the regulatory community, and Congress all  agree that  EPA needs to change the way it
specifies  monitoring  requirements  in  regulations and permits.  There  is broad  acceptance  for use of  a
performance-based measurement system (PBMS). The EPA Office of Solid Waste and Emergency Response
(OSWER) strongly supports  this position  and is committed to using  this  approach in both the RCRA and
CERCLA  monitoring  programs. This  year's plenary session  speakers,  including   Fred  Hanse,  Deputy
Administrator of EPA; Brad Campbell, Associate Director for Toxics & Environmental Protection; Steve  Koorse,
Hunton and Williams; Elizabeth Cotsworth,  Acting Director,  Office of Solid Waste; and Larry Keith, Waste Policy
Institute, will focus their remarks on various aspects of PBMS.

PBMS Implementation Session
This session, organized by the International Association of Environmental Testing Laboratories Section of the
American Council of Independent Laboratories (ACIL-IAETL), will include a presentation of EPA Program Office
PBMS Implementation Plans, followed by speakers presenting the laboratory community, regulated community,
and state regulatory agency perspectives on PBMS. The focus of the talks from the regulated community will be
on the barriers they expect to face when trying to obtain the benefits of PBMS and what we all need to do to
overcome the problems.

PBMS Methods Validation Session
Now that the  EPA  has moved  towards implementing a  PBMS approach to environmental monitoring, the
regulated  community needs  to  know how to validate methods  under this  new system. This session,  also
organized by ACIL-IAETL, will focus on how to tailor measurement system validation to the required data quality;
how validation criteria should be specified in order that it not serve as a barrier to using alternative measurement
technologies; and what is the minimum methods validation data set that one needs for the data obtained from
the analysis to be of known and documented quality.

EPA's Environmental Monitoring Research Program
This session will focus on environmental monitoring methodology research  supported under EPA's Science to
Achieve Results (STAR) competitive, extramural grant program. The program  has funded a number of research
projects whose goal it is to develop unique or novel techniques for monitoring  pollutants in the environment.
Methodology to monitor air, water, soil and other media will be presented.  The  session will review the results
from projects funded in prior years, and discuss the objectives and approaches to be  undertaken in  research
studies that received funding this  past year.

Superfund Session
The keynote for this Superfund session will be "Times Are Changing." Planned highlights include (1) a description
of the  Superfund pipeline over  time,  emphasizing  the near and longer term future and how that relates to
analytical service need; (2) information on how Superfund is working with other EPA program offices to enhance
our operation;  (3) the trend to encouraging economic redevelopment at Superfund sites; (4) how Superfund is
implementing a  PBMS approach;  (5) initiatives  related  to data  quality and minimal Quality Systems; (6)


                                                iii

-------
electronic data delivery and validation; (7) usefulness of the National Environmental Laboratory Accreditation
Conference (NELAC) accreditation for Superfund analytical work; (8) future direction for the Contract Laboratory
Program (CLP); (9)  Contracts 2000, Performance-Based Contracting, and what this means to the  laboratory
community, and (10) laboratory perspective on working for Superfund (a panel discussion).

SYMPOSIUM-AT-A-G LANCE
Sunday, July 12
8:00 a.m. - 9:30 a.m.
9:00 a.m. -12:00 noon
9:00 a.m. -4:00 p.m.
1:00 p.m. - 5:00 p.m.
Registration for Short Courses
Short Course: Analytical Strategy for the RCRA Program
Short Course: An Introduction to Practical Ethics for Environmental Laboratories
Short Course: Clean Chemistry for Trace Elemental Analysis
Monday, July 13
7:00 a.m. - 4:00 p.m.   Registration Open
8:00 a.m. -12:00 noon SW-846 Workgroups (closed)
8:00 a.m.  12:00 noon Short Course: Quality Systems, PBMS and NELAC: Putting It All Together
11:00 a.m. -12:00 noon Trial Exam for Environmental Analytical Chemists
2:00 p.m. - 4:30 p.m.   Opening Plenary Session
                             David Friedman, Office of Research and Development
                             Gail Hansen, Office of Solid Waste
                             Fred Hansen, Deputy Administrator of EPA
                             Brad Campbell, Associate Director for Toxics & Environmental Protection
                             Steve Koorse, Hunton & Williams
                             Elizabeth Cotsworth, Acting Director, Office of Solid Waste
                             Larry Keith, Waste Policy Institute
5:00 p.m. - 7:00 p.m.   Opening Reception
Tuesday, July 14
7:00 a.m. - 5:00 p.m.
8:15 a.m. -12:00 noon
8:15 a.m. -12:00 noon
10:00 a.m. -10:30 a.m.
12:00 noon -1:00 p.m.
1:15 p.m. -5:00 p.m.
1:15 p.m. - 5:00 p.m.
3:00 p.m. -3:30 p.m.
7:00 p.m.  8:00 p.m.

Wednesday, July 15
7:00 a.m. -5:00 p.m.
8:15 a.m. -12:00 noon
8:15 a.m. -12:00 noon
10:00 a.m.  10:30 a.m.
12:00 noon -1:00 p.m.
1:15 p.m. -5:00 p.m.
1:15 p.m. -5:00 p.m.
3:00 p.m. - 3:30 p.m.
7:00 p.m. - 8:00 p.m.

Thursday, July 16
9:00 a.m. -12:00 noon
9:00 a.m. -4:00 p.m.
Registration Open
Organic I Session
Inorganic Session
Break
Lunch Break
Organic II Session
Environmental Monitoring Research Session
Break
Trial Exam for Environmental Analytical Chemists
Registration Open
PBMS Implementation Session
QA Session
Break
Lunch  Break
PBMS Validation Session
General Session
Break
Trial Exam for Environmental Analytical Chemists
Short Course: Analytical Strategy for the RCRA Program
Superfund Session
                                                 IV

-------
                                        CONTENTS
ORGANIC

   Paper                                                                                  page
  Number                                                                                Number
     1      Overview of Current Status of the RCRA Organic Methods Program. B. Lesnik             3
     2      Appropriate Ways to Modify Existing Methods for New Applications. B. Lesnik              3
     3      Questionable Practices in Organic Laboratories. J.F. Solsky                             3
     4      Modifications to SW-846 HPLC Methods 8330 and 8310. S. Weisberg, M.L. Ellickson        4
     5      PCB Separations Using 2 Dimensional GC for Confirmation: Use of a Heart-Cutting          4
           Switching Valve. D.R. Gere, A. Broske, R. Kinghorn
     6      Estimating the Total Concentration of Volatile Organic Compounds in Soil at Sampling       5
           Locations: Field Trials. A. Hewitt, M.H. Stutz
     7      Final Evaluation of Method 3546: A Microwave-Assisted Process (MAP™) Method for       11
           the Extraction of Contaminants Under Closed-Vessel Conditions. J.R.J. Pare, J.M.R.
           Belanger, C. Chin, R. Turle, B. Lesnik, R. Turpin, R. Singhvi
     8      Phenoxyacid Herbicide Screening. M. Bruce, K.L. Richards, R.M. Risden                 17
     9      Field Demonstration of a Portable Immunosensor for Explosives Detection. A.W.           26
           Kustarbeck, P.T. Charles, P.R. Gauger, C. Patterson
     10     Environmental Applications of a Fiber Optic Biosensor. L.C. Shriver-Lake, I.B.             26
           Bakaltcheva, S. van Bergen
     11     Development and Validation of an Improved Immunoassay for Screening Soil for           27
           Polynuclear Aromatic Hydrocarbons. T.S. Fan, B.A. Skoczenski
     12     A New Dioxin/Furan Immunoassay with Low Picogram Sensitivity and Specificity           27
           Appropriate for TEQ Measurement: Applications Development. R.O. Harrison, R.E.
           Carlson
     13     Development and Validation of an Immunoassay for Screening Soil for Polychlorinated      28
           Biphenyls. T.S. Fan, B.A. Skoczenski
     14     Gasoline Range Aromatic/Aliphatic Analysis Using Pattern Recognition. S.E. Bonde        28
     15     Bioremediation Assessment Using Conserved Internal Biomarkers. P. Calcavecchio,       29
           E.N. Drake, G.C. VanGaalen, A.  Felix
     16     Method 8270 for Multicomponent Analyte Analysis. E.A. LeMoine, H. Hoberecht           35
     17     Benzidine? Really? R-K. Smith                                                     40
     18     Comparison of Volatile Organic Compound Results Between Method  5030 and Method      43
           5035 on a Large Multi-state Hydrocarbon Investigation. R.J. Vitale, R. Forman, L.
           Dupes
INORGANIC

   Paper
  Number
    19
    20

    21
SW-846 Inorganic Methods Update. O. Fordham
Direct Mercury Analysis: Field and Laboratory Applications, H.M. Boylan, H.M.
Kingston, R.C. Richter
Mercury in Soil Screening by Immunoassay. M.L. Bruce, K.L. Richards, L.M. Miller
 Page
Number
  53
  53

  55

-------
INORGANIC (continued)
   Paper
  Number
    22
    23

    24


    25


    26
    27
Utilization of a Field Method for the Semiquantitative Detection of Silver in
Environmental Samples in the 0-50 ppb Range. D. Kroll
Diagnosing Errors in Species Analysis Procedures Using SIDMS Method 6800. H.M.
Kingston, D. Huo, Y. Lu
The Use of 210Pb Dating Stratigraphy to Determine the Significance and Fate of
Chromium in Sediments near a Hazardous Waste Site. R. Rediske, G. Fahnenstiel, C.
Schelske, M. Tuchman
Air-Force Wide Background Concentrations of Inorganics Occurring in Ground Water
and Soil.  P.M. Hunter
New Tools for Liquid Sampling. J.D. Hoover, S.R. Somers
Analysis of Chemical Warfare Agent Decontamination Brines for Lewisite Degradation
Products  Using Gas Chromatography with Atomic Emission Detection. K.M.
Morrissey, T.R. Connell, J. Mays, H.D. Durst
 Page
Number
  60
  64

  65


  73

  77
  88
EPA'S ENVIRONMENTAL MONITORING RESEARCH PROGRAM SESSION
   Paper
  Number
    28
    29

    30
    31


    32



    33


    34


    35


    36

    37


    38

    39


    40
 Introduction, Session Scope and Purpose. W. Stelz
 Bioavailabilty and Risk Assessment of Complex Mixtures. K.C. Donnelly, W.R.
 Reeves, T.J. McDonald, L.-Y. He, R.L. Autenrieth
 Field Determination of Organics from Soil and Sludge Using Sub-critical Water
 Extraction Coupled with SPME and SPE. S.B. Hawthorne, C.B. Grabanski, A.J.M.
 Lagadec, M. Krappe, C.L. Moniot, D.J. Miller
 A Field Portable Capillary Liquid/Ion Chromatograph. T.S. Kaphart, C.B. Boring, P.K.
 Dasgupta, J.N. Alexander IV
 Rapid Determination of Organic Contaminants in Water by Solid Phase
 Microextraction and  Infrared Spectroscopy. D.C. Tillotta, D.C. Stahl, S.A.
 Merschman, D.L. Heglund, S.H. Lubbad
 Intrinsic Stable Isotopic Tracers of Environmental Contaminants. S.A. Macko, P.J.
 Yanik, N.A. Cortese, M.C. Kennicutt II, Y. Quin
 Recent Developments in Immunobiosensors and Related Techniques for the
 Detection of Environmental Pollutants. M. Masila, H. Xu, E. Lee, O.A. Sadik
 Multiplexed Diode Laser Gas Sensor System for In-situ Multi-species Emissions
 Measurements. R. Hanson
 Overview/Future of NCERQA Research Program. B. Krishnan
 Advanced Analytical Methods for the Direct Quantification and Characterization of
 Ambient Metal Species in Natural Waters. J.G. Hering, J.H. Min
 Radical Balance in Urban Air. R.J. O'Brien, L.A. George, T.M. Hard
 Environmental Applications of Novel Instrumentation for Measurement of Lead
 Isotope Ratios in Atmospheric Pollution Source Apportionment Results. Keeler
 Remote Sampling Probe with Fast GC/MS Analysis: Subsurface Detection of
 Environmental Contaminants. A. Robbat, Jr.
 Page
Number
   93
   93
   93



   94


  101



  102


  110


  117


  117

  117


  118

  119


  119
                                                VI

-------
ERA'S ENVIRONMENTAL MONITORING RESEARCH PROGRAM SESSION (Continued)
Paper
Number
41

42

43
44

45
46

47

48

49

50

QUALITY
Paper
Number
51

52

53

54

55

56

57

58

59

60
61


An Integrated Near Infrared/Spectroscopy Sensor for In-situ Envrionmental
Monitoring. R.A. Levy, J.F. Federici
A Nitric Oxide and Ammonia Sensor Array for Fossil Fuel Combustion Control. B.T.
Marquis, J.F. Vetelino, T.D. Kenny
Relevance of Enantiomeric Separations in Environmental Science. D.W. Armstrong
Development of Aerosol Mass Spectrometer for Real Time Analysis of PAH Bound to
Submicron Particles. J.T. Jayne, D.R. Worsnop, C.E. Kolb, X. Zhang, K.A. Smith
Real-Time Trace Detection of Elemental Mercury and Its Compounds. R.B. Barat
Orthogonal Background Suppression Technique for EPA's Field Infrared Data
Processing. Blaterwick
Development of a Continous Monitoring System for PMio and Compounents of PM2s,
M. Lippmann, J.Q. Ziong, W. Li
A Real-Time Sampler, Rams, for the Determination of PM25, Including Semi-Volatile
Species. F Obeidi, E. Patterson, DJ. Eatough
Dynamic Nuclear Polarization (DNP): A New Detector for Analysis of Environmental
Toxicants. H.C. Dorn
Partitioning Tracers for In-situ Detection and Measurement of Nonaqueous Liguids in
Porous Media. Brusseau
ASSURANCE


CD-R Archive and Catalog of Hewlett-Packard Low Resoltuion Format Data from
Networked GC/MS Systems. R.G. Briggs, H. Valente
The Use of Acceptable Knowledge for the Characterization of Transuranic Waste in the
Department of Energy Complex. R.V. Bynum, R.B. Stroud, R.D. Brown, L.D. Sparks
Performance Evaluation Soil Samples for Volatile Organic Compounds Utilizing
Solvent Encapsulation Technology. J. Dahlgran, C. Thies
U.S. EPA and U.S. A.F Interagency Agreement for Field Analytical Services. R.A.
Flores, G.H. Lee
Several Organic Parameters on Underlying Hazardous Constituents List Can Not Be
Measured at the Universal Treatment Standards. H.C. Johnson, C.S. Watkins
Ignitability Performance Evaluation Study - Are Your Waste Streams Being Correctly
Characterized? L.J. Dupes, R.J. Vitale, D.J. Caillouet
Techniques for Improving the Accuracy of Calibration in the Environmental Laboratory.
D.A. Edgerley
Quality Control Protocol for Analysis of Drugs and Explosives Using Ion Mobility
Spectrometry. J. Homstead, E.J. Poziomek
Reference Materials for the Analysis of Metals in Sludge. S.J. Nagourney, N.J.
Tummillo, Jr., J. Birri, K. Peist, B. MacDonald, J.S. Kane
Perspectives on Dioxin Analysis. S. Wilding
Interpretation of Ground Water Chemical Quality Data. G.M. Zemansky
Page
Number
124

125

131
131

132
137

137

146

147

148


Page
Number
151

152

161

165

165

173

181

187

188

191
192
                                      vii

-------
GENERAL

   Paper                                                                                 Page
  Number                                                                              Number
    62     Sample Introduction Techniques for Fast GC Analysis of Organochlorine Pesticides.      205
           C.E. Boswell, M.S. Clark, R.M. Woodard
    63     lonspray LC/MS Method for Quantitation of Major Degradation Products of Benomyl.      209
           C. Eilcone, T. Xuan, S.M. Reddy
    64     Methods Optimization of Microwave Assisted Solvent Extraction Techniques for          214
           Various Regulatory Compounds. G. LeBlanc, M. Miller
    65     New Developments in Closed Vessel Microwave Digestion Technology for Preparation    215
           of Difficult Organic Samples for AA/ICP Analysis. G. LeBlanc, B. Haire, B. Fidler
    66     Evaluation of ICP-OES and ICP-MS for Analysis fo the Full TCLP Inorganic Target       215
           Analyte List. M. Paustian, Z. Grosser
    67     Clean Metals Sampling. R.E. Stewart II                                            219

 Author Index                                                                            227
                                                VIII

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
   ORGANIC

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
             OVERVIEW OF CURRENT STATUS OF RCRA ORGANIC METHODS PROGRAM

                                             Barry Lesnik
                     US EPA, Office of Solid Waste, 401 M ST., SW, Washington, DC

                                      A/0 ABSTRACT AVAILABLE
           APPROPRIATE WAYS TO MODIFY EXISTING METHODS FOR NEW APPLICATIONS

                                             Barry Lesnik
                     US EPA, Office of Solid Waste, 401 M ST., SW, Washington, DC

                                      NO ABSTRACT A VAILABLE
                     QUESTIONABLE PRACTICES IN THE ORGANIC LABORATORY

                                           Joseph F. Solsky
           US Army Corps of Engineers (CENWO-HX-C), 12565 W Center Road, Omaha, NE 68144
                               Phone: (402) 592-9542, Fax: (402) 592-9595
                                 e-mail: joseph.f.solsky@usace.army.mil

SW-846 is a collection of performance-based methods used by the US Army Corps of Engineers (USAGE) for
execution  of environmental restoration projects. These methods provide  guidance  for the running of various
organic and inorganic protocols that can be used for the analysis of samples from a variety of sample matrices.
During recent laboratory audits conducted by the USAGE, certain 'questionable  practices' have been observed,
especially in the organic analysis areas.

Most  people have  a  relatively good  idea of what constitutes a fraudulent activity  today. The  concepts  of
'dry-labing,' 'peak shaving,' 'peak enhancing,' or 'time-traveling1 are well  understood. These practices clearly
involve the deliberate direct manipulation and/or alteration of data, often to achieve or meet method QC criteria.
However,  there are  a  new group of 'questionable practices' now being observed that often involve the selective
exclusion  of data to  achieve or meet method QC criteria.

This presentation will  focus on several of these 'questionable practices'. Examples of some of these practices
include the following:  (1)  One such practice involves the determination of initial  calibration curves. Laboratories
have been observed running eight or more standards for methods that state 'a minimum of five points should be
used to establish the initial calibration curve.' Up to three points are then discarded, even from the middle of the
curve  until the appropriate QC criteria can be met. No technical justification existed  for the deletion of these
points  other than to meet the method QC  criteria. (2) Another such practice  involves the evaluation of the
continuing calibration  verification solution. Laboratories have been observed averaging the %  difference or %
drift across all  target analytes even when several of  the target analytes  exceed the criteria  by a significant
amount such that the average still meets the criteria as stated in the method. (3) Another such practice involves
the reporting of acceptance  ranges for surrogates or laboratory control  samples.  Laboratories  have been
observed  reporting a very tight range indicating that they have good method control. However, an examination of
the actual control charts maintained  by the laboratory shows a significantly  wider  range. (4) Another such
practice involves the determination of the  method detection  limit (MDL).  Laboratories have been observed
running ten or  more standards and then discarding points  to achieve a lower  MDL.  No technical justification
existed for the deletion of these points other than to achieve a lower MDL.

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Should the above 'questionable practices' be considered as examples of fraudulent activities? Some of the
laboratories have described these practices as 'the common approach used by everyone.' Yet when described to
people within EPA, the clear response is that these approaches were never intended. The background history will
be discussed as to how and why these practices have developed and what can be done to overcome them.
                     MODIFICATIONS TO SW-846 HPLC METHODS 8330 AND 8310

                                     C.A. Weisberg. M.L. Ellickson
                  U.S. EPA, Region III, Office of Analytical Services and Quality Assurance,
                                839 Bestgate Road, Annapolis, MD 21401

Method 8000 of SW-846 allows for modification of the existing  methodology,  provided that the analyst has
documented the ability to generate acceptable results  by successfully performing an initial  demonstration  of
proficiency with the altered conditions prior to  sample analysis. Analysts may vary the extraction procedures
and/or HPLC parameters (mobile phase composition, elution program, injection volume, etc) in order to improve
sensitivity and chromatographic resolution, or  to reduce interferences resulting  in coelution with analytes  of
interest. The use of analytical columns different from those specified  in the methods may require that different
HPLC conditions and flow rates be used.  Generally, the HPLC methods 8330 and  8310  have been found  to
perform acceptably as written.  However, some  method modifications  need to be employed in order to improve
analyte resolution,  achieve desired quantitation limits and obtain definitive confirmation of the primary column
results. Some of the modifications to Method 8330 and Method 8310 reported  in this paper include: nitrogen
blow-down to adequately concentrate the final  acetonitrile extracts after the salting-out liquid-liquid extraction
procedure or the solvent exchange procedures; reduced flow rates and gradient elution schemes for the C-18
primary columns;  and, the use of an  acetonitrile/water mobile phase for the cyano  secondary column
confirmations.

When reporting quantitated results for these analyses, it  is more important to  positively  confirm the primary
column results than it is to completely  resolve all of the target  compounds. The use of analytical columns
different from those specified  can produce coeluting  pairs  which  may differ  from those  mentioned in the
methods.  This  paper will also discuss the reporting  of coeluting compounds as mixtures,  and the  use  of
confirmatory techniques (e.g. dissimilar secondary column analyses, spectral library  matching, LC/MS, GC/MS
and GC-ECD).
                 PCB SEPARATIONS USING 2 DIMENSIONAL GC FOR CONFIRMATION:
                            USE OF A HEART-CUTTING SWITCHING VALVE

                                    Dennis R. Gere and Allen Broske
                       Hewlett-Packard, 2850 Centerville Rd., Wilmington, DE 19808
                                         Phone (302) 633-8162
                                      email dennis_gere@hp.com
                                           Russell Kinghorn
                SGE International Pty. Ltd., 7 Argent Place Ringwood, Victoria 3134 Australia

 Environmental gas chromatography analysis continues to be a difficult but necessary effort for the detection and
 identification of PCBs. One approach is the use  of a very good separation column, which would resolve most if
 not all of the 140-150 key congeners from one another. Another approach uses a more user-friendly approach to
 the congener specific analysis of PCBs. This approach is not a new or novel approach, but a revisit of a variation
 of  a general column-confirmation column method.  The  use  of  two  different  stationary  phase  columns
 (orthogonal) allows a geometric gain in separation  power compared to the algebraic gain by making the column

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
longer or more efficient.
The approach we used involves a heart cutting switching valve, one injector and two detectors. Some discussion
will be given concerning the choice of the stationary phases, the choice of carrier gas and especially the judicious
combination of temperature programming rate and carrier linear velocity. Focus will be upon key critical PCB
congeners as described by the European Union protocols, that is, congener specific detection and confirmation.

The configuration uses two columns and two detectors, with a valve between the two detectors and  a second
selective column before the second detector. In the initial analysis, the sample is  nominally separated on a
general-purpose column. As sets of unresolved PCBs elute, they are cut out of the first separation scheme and
sent to the second more selective column. The front of the second column may be cooled cryogenically (we will
also demonstrate what happens when the cooling of the second column is not used) to hold the analytes at the
front of the second column until the analysis of the first column is complete. Then a second temperature program
is begun to resolve previously unresolved pairs.

The first column is a 5 % phenyl phase and the second column we used was either a 35 % phenyl or a 50 %
phenyl phase of intermediate polarity. Either column is useful, and a specific choice may be made depending
upon the sensitivity level (the 35 % phenyl is a  much lower bleed column) or whether additional selectivity is
needed. There are a great many permutations of this approach possible, and we will outline what we did and
what can be done if even greater resolution is required.
           ESTIMATING THE TOTAL CONCENTRATION OF VOLATILE ORGANIC COMPOUNDS
                          IN SOIL AT SAMPLING LOCATIONS: FIELD TRIALS

                                            Alan D. Hewitt
    U.S. Army Cold Regions Research and Engineering Laboratory, 72 Lyme Road, Hanover, NH 03755-1290
                                            (603) 646-4388
                                            Martin H. Stutz
                U.S. Army Environmental Center, Aberdeen Proving Ground, MD 21010-5401
                                            (410)612-6856

ABSTRACT
This report describes a method for estimating the total concentration of volatile organic compounds  (VOCs) in
soil relative to a site-specific 0.2-mg/kg working standard and presents the results from three separate field trials.
This method was developed to provide a decision tool for field or laboratory personnel so they can implement the
appropriate soil sample preparation procedure for the selected method of instrumental analysis. Coupling a rapid
method for estimating the total VOC concentration with sample collection, handling, and preparation procedures
that limit substrate dissaggregation and exposure complements efforts to achieve site-representative estimates
for vadose zone contamination.

INTRODUCTION
Since the beginning of the Superfund and the Resource Conservation and Recovery Act (RCRA) programs, gas
chromatography/mass spectrometry  (GC/MS) (via Methods 8260 and  8240) has served as the major laboratory
instrument for identifying  and  quantifying VOCs in soils1  The principal reason for the selection of this analytical
detection system is that it provides  an  unambiguous identification of analytes present. Unfortunately, this very
desirable quality comes with the limitation that for quantification purposes the individual analytes must fall within
a concentration range of 2 to  3 orders of magnitude. High analyte  concentrations can degrade the performance
of the  MS detection system,  which  interrupts scheduled runs  and may  lead to  expensive instrument repairs.
Therefore, one of the challenges when  using an MS is how to couple it with a sample collection, handling, and
preparation protocol when analyte concentration can range over 7 orders of magnitude (percent levels to the
current levels of instrumental detection, approximately 0.005 mg/kg). To cope with this concern, samples thought
to be contaminated with VOCs at levels greater than 0.2 mg/kg are prepared by extraction (and perhaps further
dilution) with methanol (MeOH), i.e., the high-level method. In contrast, samples thought to have concentrations
less than  0.2  mg  VOC/kg  are  analyzed  directly, which is  referred  to as the low-level method. Many other
commonly used laboratory instruments and their respective methods for VOC detection (e.g., Methods 8015 and

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
8021B) also benefit from using these two approaches to sample preparation.

A second challenge is that VOCs in soils fail to maintain their concentration integrity if they are not collected and
handled  with  limited  disruption and  exposure and  if preventive measures are not taken to limit biological
degradation of aromatic compounds. Today it is generally recognized that the sample collection and handling
guidance, provided in the past by Method  5030, often resulted in a greater than 90% loss of the VOCs from soil
samples prior to laboratory analysis2-6. To  minimize losses due to volatilization and biodegradation,  new sample
collection and analysis  protocols were included  in the  third  update  of SW-8461:  Method  5035, "Modified
purge-and-trap and extraction  for volatile organics  in  soil and waste samples,"  and Method 5021, "Volatile
organic compounds in soils and other solid matrices using equilibrium headspace analysis."

The two most effective collection and handling protocols that can be used with these new methods for preventing
the loss of VOCs are 1) the on-site, rapid transfer of discrete samples with  a small coring tool to a vessel that
hermetically seals and already  contains the appropriate dispersion/extractant solution for the chosen method of
analysis7, or 2) obtaining and temporarily storing  (two  days at 4°C)  a sarqple in  an En Core™ (En  Novative
Technologies,  Inc.,  1241  Bellevue  St.,  Green  Bay,  Wise.  54302)  sampler  before  transferring it  into an
appropriately prepared vessel8.  In addition, it should be recognized that, if the sample is to be held for more than
two days before analysis, then  some form of chemical preservation may be necessary in addition to storage at
4°C. For example, acidification can be  used for low-level sample  preparation procedures when carbonates are
not present6

Because there is often no a priori  knowledge of the VOC concentrations at a given location, the data quality
objectives  for site characterization activities often  require that samples be collected and prepared for both the
lowand high-level analysis procedures. To avoid collecting and processing samples through  both  of these
preparation procedures for every location, it has been suggested that a rapid screening analysis be performed to
establish an estimate for the total VOC concentration9  This screening indicates the levels of  VOCs to expect,
before the sample is prepared for analysis, and thus whether collocated sample(s) taken for laboratory analysis
should be  prepared using the low- or high- level procedures, or both. The  method developed is based on the
comparison of responses of a hand-held photoionization detector (PID) to a sample relative to a 0.2-mg VOC/kg
site-specific working standard.  Recognition of the potential effort and cost savings by using screening as a
decision tool are two reasons why this method is being  considered for inclusion in the fourth update of SW-846
(proposed  Method 3815). This  paper briefly outlines this screening method  and  presents the results from three
case studies. Additional information concerning the development of this  method for screening  is  available
elsewhere9

SCREENING METHOD
Materials
The necessary equipment and reagents are as follows:
   1) Modified VOA vials (40 or 44 ml), Teflon-lined septa with 5- to 6-mm  hole
      punched through  the  middle  and  3-  x 3-cm  squares  of  light-gauge
      aluminum foil for temporary covers (see Fig. 1).
   2)  Coring tool for the collection and transfer of  discrete soil  samples, e.g.,
      disposable 10-mL plastic  syringes with the Luertip and rubber plunger cap
      removed, or an equivalent metal tube and plunger.
   3) A portable photoionization detector (PID) analyzer with a 10.6-eV or greater
      electrode discharge tube, digital display, inlet flow  rate greater than  200
      mL/min,  and sample inlet tube of 3 to 4  mm o.d. and at  least  3  cm  in
      length.
   4) A 10-mL liquid syringe.
   5) Reagent-grade water (i.e., water with no detectable  VOCs),  polypropylene
      glycol (PPG, or similarly low-vapor-pressure organic solvent), and principal
      VOC(s)  of site interest.
   6) A cylinder of calibration gas for the PID, e.g., 100 ppm isobutylene.
      Figure 1. Modified VOA vials for rapid total VOC screening of soil samples.
Screw Cap
Teflon-lined septum
with access hole
                                                                                            Standard or soil sample

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Standards
A stock standard is prepared by transferring the VOC  of interest into PPG.  The stock standard concentration
should be based on the density of the analyte of interest, so that a 1- to 3-uL volume transferred to a 40-mL VOA
vial containing 10 mL of reagent  water and 10 g of the site-specific soil matrix results in a 0.2-mg-VOC/kg
working standard. For example,
             Stock standard: 1.34 g/mL* x 2.0 ul_/2.5 mL -1.1  mg TCE/mL
             Working standard: 1.1  mg TCE/mL x 1.8 uL/10 g  soil = 0.2 mg TCE/kg
             *nfinsit\/ nf TCP
             *Density of TCE

Immediately after spiking, these working standard vials are covered with a single sheet of aluminum foil that is
held tightly in position with a septum  with a hole punched in the middle and a screw cap (Fig.  1). The vial
contents of the working standards should be thoroughly mixed by handshaking, then transported to the location
of the sampling activity, stored out of  direct sunlight, and allowed  to equilibrate for 1  hr prior to use. Working
standards should be prepared daily.

The PID response to the working standard should be at least 10x greater than its response to a blank (reagent
water, contamination-free site-specific  matrix, and appropriate volume of PPG). For analytes with high vapor
pressures or low octanol water partition coefficients, or both, and soil matrices with low organic carbon contents,
it  may not be  necessary  to include  the site-specific soil matrix in  the working  standards. This should be
established on a site-by-site  basis by comparing the means of triplicate working  standards with and without the
soil matrix. As a general rule, if the means differ by more than 20%,  it is recommended that the soil matrix be
included in the working standards.

Sample Collection and Analysis
Before field sampling, 10 mL of reagent water is added to the modified VOA vials. Once prepared, the VOA vials
for screening samples should be transported to the sampling location and stored with the working standards until
they are used.  The native structure of the  material being sampled for screening  should be kept  intact,  thus
experiencing as little disaggregration as possible during the collection and transfer process. This can often be
accomplished with a coring tool designed to obtain  a discrete sample.  For example,  a modified 10-mL syringe is
a practical tool for obtaining up to a 10-g soil sample. If 10 g cannot be easily obtained in a single transfer, more
than one corer  can  be  used, or a couple of transfers with a single corer can be made. This coring device is
transparent and comes with gradient markings so the volume/weight relationship for a given  material can easily
be established with a portable balance. The location of samples  taken both for screening purposes and for
laboratory  analysis  should be as close  as possible to each other (generally within a  10-cm radius) and from the
same  stratum. Before preparing (or exposing) a fresh sampling surface, for instance,  opening a split spoon  or
scraping away the top layer of a material, the cap and aluminum foil should be removed from  the screening VOA
vial. After retrieving a discrete sample, the core barrel should be inserted into the mouth  of the screening VOA
vial  and the  sample extruded.  Once the  sample  has  been extruded,  the aluminum foil  and cap should
immediately be  replaced on the vial. This collection and transfer process should take less than 10 seconds, and
the sample weight only has to approximate 10 g (plus or minus 2 g).

Before a working standard or sample is analyzed, the VOA vial should  be shaken by hand for 10 to 15 seconds.
Cohesive materials, such as silts and clays, do not break apart rapidly upon shaking and may require more than
15 seconds for  complete dispersion. The vial is then  visually checked both  for the complete dispersion of the
sample matrix and  for particles adhering to the aluminum foil cap liner (knock large particles off the aluminum
foil if present). Then the inlet tube of the PID is pushed through the foil liner to a set position about 3 cm below
the rim.  A maximum response will be achieved within 2 to 3 seconds of punching through the foil liner. The
maximum response for each sample screened and for the  analysis of each working standard should be recorded.

Daily Operating Procedure for VOC Screening
The PID should  be initially calibrated with a cylinder of standard gas (e.g.,  100 ppm isobutylene) at the beginning
of each day. This task can be performed before going to the sampling location. However, both the analysis of
site-specific working standards and the screening of a sampling location  should be performed under the same
conditions, thereby  normalizing meteorological influences on the performance of the PID. Site-specific working
standards should be prepared daily and in sufficient quantity to satisfy the study's objectives. At a minimum, one
working standard should be analyzed for every hour of site activity.

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Collection of samples for VOC analysis should always be the first operation performed after a surface to be
sampled has been exposed to the atmosphere.  This  includes samples both for screening and for laboratory
analysis. To establish how to handle and prepare the discrete sample for laboratory analysis (low, high, or both
procedures), a total VOC screening analysis should be performed at each sampling location. Therefore, before
opening a split spoon, scraping a fresh surface on a pit wall, removing surface vegetation and the appropriate
amount of top soil for a surface grid location,  or removing the first several inches of some other type of waste
material, the PID of  choice should be operating. Furthermore, if a  working standard  is being  utilized to verify
performance of the PID for the  sampling location, the analysis of a working standard should be completed before
exposing a fresh sampling surface.
Once a fresh surface  has been  exposed,  a sample
should be quickly obtained, transferred to a screening
VOA  vial, dispersed, and analyzed.  If the  maximum
response is greater than  the working standard  (or the
running average), the  sample or samples taken for
laboratory analysis should  be  prepared  using  the
high-level procedure (i.e.,  MeOH extraction). If the
maximum response is below the working standard, the
laboratory sample(s)  should  be prepared using  a
low-level procedure. The total elapsed time between
exposing a fresh  surface,  screening  a  sample, and
obtaining samples for  laboratory analysis  should be
less than 2  minutes.  As a precaution  against false
positive  and false negative  screening  estimates
relative  to   the   decision  point,  locations  where
screening results  are between 0.5 and 2x the working
standard response should have samples prepared by
both high- and low-level procedures.

Method Limitations
For this method of sample location screening to work,
the  VOC(s)   of  interest  must  be  detectable by
photoionization. If more than one analyte is of interest,
and  there  are large  discrepancies  (greater than  a
factor of  2)  in  photoionization  potentials,  then  the
range around the decision point where samples are
prepared  by  both  high- and  low-level procedures
should  be  increased   proportionally. That  is,  if the
responses for the VOCs of interest differ by  a factor of
3x, and the analyte with the highest response is  used
to make the working standard, then laboratory samples
from  locations where screening  results are only 0.3x
the working  standard  should  be  prepared by  both
procedures. However, this often will not be  a problem
for sites contaminated  with  common chlorinated and
aromatic  compounds  because   they  have  similar
photoionization potentials. This approach may  not be
effective for sample  matrices that are not  readily
dispersed in water (e.g., some clays and cementitious
materials).
Table 1. Field screening measurements and sample
preparation procedures.
Screening (response)*
No.
Site 1
S1-1
S1-2
S1-3
S1-4
S1-5
S1-6
S1-7
S1-8
S1-9
S1-10
Site 2
S2-1
S2-2
S2-3
S2-4
S2-5
S2-6
S2-7
S2-8
S2-9
S2-10
SiteS
S3-1
S3-2
S3-3
Working Std

6.5
6.5
—
—
—
7.0
8.0
—
—
6.3

2.3,2.5
3.2
4.2
3.9
—
3.9
—
—
3.8
4.8

5.6
4.8
5.3, 5.5, 5.2, 4.9
Sample

0.0
0.0
0.4
790
480
1400
26
2.2
12
4.4

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

0.0
0.0
0.0
Sample
Preparation
Method

5021 , LL**
5021, LL
5021, L
5035, HLf
5035, HL
5035, HL
5035, HL
5021, LL
5035, HL
5021 , LL

5035, LL"
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL
5035, LL

5021, LL
5021 , LL
5021, LL
FIELD TRIALS                                      t
This   method   for  rapidly   estimating  the   total
concentration  of  VOCs  was  tested  during  three   tt
different sampling  activities  performed  under  the
supervision  of personnel from EPA Region 1.  At the
sites visited, samples were obtained from near the surface with the aid of hand tools and from split-spoon core
barrels. All samples, whether collected for on-site screening or for off-site analysis, were transferred using a
   PID  field  screening:  Working  Std-Results  of
  analyzing a site-specific working standard 0.2 mg
  TCE/kg for Site 1  and 0.2 mg PCE/kg for Sites 2
  and  3.  Sample  --Results from  rapidly (<30 s)
  screening 10±2 g soil.
   5021,  LL-Sample  placed in 22-mL  VOA  vial
  containing 10 mL water.
  5035,  HL--Sample  placed  in 40-mL  VOA  vial
  containing 5 mL MeOH.
   5035,LL--Sample  placed in  40-mL  VOA  vial
  containing 5 mL water.

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


modified  syringe.  Samples collected  for  off-site analysis  were placed  into vials  containing methanol  or
organic-free water, as appropriate for the intended method  of sample  preparation (Method 5035 or 5021) and
analysis (Methods 8260, 8015, 8021), and analyzed within 48 hours.

The results of the screening analysis for both the working standards and samples are shown in Table 1, and the
results of the laboratory analysis are shown in Table 2. During these  field trials, the screening  results were only
used to decide whether to prepare samples by a low- or high-level procedure. With the possible exceptions of
sampling  locations  S1-3  and  S1-10, the  samples were  prepared  appropriately for the intended method  of
analysis. That is, the  analysis  system was not exposed to an unexpectedly high analyte concentration. Indeed,
this statement applies to all  of the samples, since the concentrations were not much greater than 0.2 mg/kg for
the individual analytes found in these two samples. Therefore, a scheduled run would not have been delayed nor
would the detector have been damaged; however, there may have  been individual analyte responses greater
than the highest calibration standard.
Table 2. Laboratory results.

 Sitel.
                                              HS/GC-PID/FID (mg/kg)
No.
S1-1
S1-2
S1-3**
S1-4**
S1-5
S1-6
S1-7
S1-8
S1-9
S1-10
Site 2.

No.
S2-1
S2-2
S2-3
S2-4
5
6
7
8
9
10
CDCE TCE Tol
—
—
0.021 0.15
5.2 29
—
250 19
—
0.040 0.020
—
—

PT/GC/MSf
Trichlorofluoromethane
—
—
—
0.004
—
—
0.001
0.002
—
—
PCE
<0.003*
<0.003
0.008
—
140
240
7.4
0.17
1.0
0.24

(mg/kg)
PCE
—
0.001
0.001
—
—
—
—
—
—
...
EBen
—
<0.003
0.074
15
—
33
...
...
...
—

p/m-Xyl
—
<0.003
0.076
15
—
76
...
—
—
—
Site 3.
o-Xyl
—
<0.009
0.082
3.5
—
68
...
—
—
—

Total
<0.003

0.41
68
140
690
7.4
0.23
1.0
0.24

HS/GC-PID/FID
Total
ND
0.001
0.001
0.004
ND
ND
0.001
0.002
ND
ND
(mg/kg)
No.
1.
2.
3.






PCE
<0.003
—
—






Total
<0.003
ND
ND






  <0.003-Peak identified but below quantation.
** Unidentified peaks present in chromatogram.
T Samples analyzed at EPA Region 1 Laboratory in Lexington, Mass.

In the case of sample S1-10, the recommendation that samples be processed through  both procedures when
sample  screening  results  are within  a factor of 2 to  0.5 of the working standards would  have provided the
necessary precaution  However, in the case of S1-3,  where the screening results were well below 0.5x the
working standard, this same logic would not have succeeded. Samples S1-3 and S1-4 were taken within 5 cm
(vertical) of one  another, and both the screening and  laboratory  results (Table 2) showed that this area had a
large vertical gradient  in VOC concentrations. In this  case, a review of  the data and perhaps the site history
would alert the sample collectors to a potential problem and therefore  the need to  implement an additional
precaution so as not to overload the analytical system. Two potential  solutions that would have worked at the
site where'samples S1-3 and S1-4 were taken  are 1)  to have taken screening samples on either side  of Oust
above and below) the sample taken for laboratory analysis, 2) or alternatively, to have automatically prepared

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


samples by both low- and high-level procedures based on knowledge of where the source regions were on this
site. Because of this experience,  an additional  recommendation is to use one of these two procedures (e.g.,
bracketing laboratory samples with screening samples or taking laboratory samples for both low- and high-level
preparation procedures) when  sampling  near known  or suspected source regions. Whenever  samples are
prepared by both a low- and high-level procedure, the sample prepared by the high-level  method should be
analyzed first. Furthermore, although not  reported here, it has become evident that when a  screening exceeds
the scale of the PID, which  is typically greater  than 2000  ppm, further dilution  of a sample processed  by the
high-level procedure is most likely  warranted prior to analysis.

SUMMARY
The problems of underestimating the concentration of  VOCs in  samples taken from  the vadose zone has
facilitated the acceptance of new sample collection, handling, and preparation protocols (e.g., Methods 5035 and
5021). These changes not only present challenges to field sampling teams but to the laboratories responsible for
sample  analysis as well. For this reason both  parties must  be  involved in the initial  design of the sample
collection  plan  and remain in contact throughout the  project. To assist in deciding  how samples should be
prepared for instrumental analysis, a simple total VOC  screening  procedure has been developed.  The main
purpose of this screening  method is to  provide a decision tool during the sampling activity to help establish
whether samples taken for laboratory analysis should be prepared by a low-level or a high-level procedure, or by
both. This screening process is, however, not foolproof. Likewise, neither are any others that must contend with
the possibility  of a heterogeneous analyte  distribution. For this reason, there are additional precautions that
should be taken when using this method. One is for the case when screening results for samples are within a
factor of 2 to 0.5 of the working standard results,  another is for the case when sampling near known or suspected
source regions, and the third is for when the PID's response to a screening sample is over range (greater than
2000 ppm). As demonstrated here, this screening procedure has the potential to greatly reduce the number of
samples that would have to be collected and processed  during a site investigation.

ACKNOWLEDGMENTS
Funding for this work was provided by the U.S. Army  Environmental  Center, Martin  H.  Stutz, Project Monitor,
and from the  Advanced Site  Monitoring System AF-25 program,  Dr. James Brannon, Project  Monitor. The
authors thank Louise Parker and Thomas Ranney for critical review of the text.

This publication reflects the  view  of the author  and does not suggest or reflect policy, practices, programs, or
doctrine of the U.S. Army or of the government of the United States.

REFERENCES
1   U.S. Environmental Protection Agency (1986) Test Methods for Evaluating Solid Waste. Vol. 1B. SW-846.
2.  Urban, M.J., J.S. Smith, E.K. Schultz, R.K. Dickinson  (1989) "Volatile organic analysis for a soil, sediment or
    waste sample," in Proceedings of the 5th Annual Waste Testing & Quality Assurance Symp., U.S.
    Environmental Protection Agency, Washington, DC, pp. 11-87-11-101.
3.  Illias, A.M., C. Jaeger (1993) "Evaluation of sampling  techniques for the analysis of volatile and total
    recoverable petroleum hydrocarbons (TRPH) by IR, GC, and GC/MS methods," in Hydrocarbon
    Contaminated Soils, Chelsea,  Michigan: Lewis Publishers. 3: 147-165.
4.  Lewis, T.E., A.B. Crockett, R.L. Siegrist, K. Zarrabi (1994) Soil sampling and analysis for volatile organic
    compounds. Environmental Monitoring and Assessment, 30: 213-246.
5.  Hewitt, A.D., T.F Jenkins, C.L. Grant  (1995) Collection, handling and storage: Keys to improved data  quality
    for volatile organic compounds in soil. American Environmental Laboratory, 7(1): 25-28.
6.  Liikala, T.L., K.B. Olsen, S.S. Teel, D.C. Lanigan (1996) Volatile organic compounds:  Comparison of two
    sample collection and preservation methods. Environmental Science and Technology, 30:3441-3447.
7.  Hewitt, A.D., N.J.E. Lukash  (1996) Sampling for in-vial analysis of volatile organic compounds in soil.
    American Environmental Laboratory, 7(8): 15-19.
8.  Hewitt, A.D. (1998) A tool for the collection and storage  of soil samples for volatile  organic compound
    analysis. American Environmental Laboratory, 9(10): 14-16.
9.  Hewitt, A.D., N.J.E. Lukash  (1997) Estimating the total concentration of volatile organic compounds in soil: A
    decision marker for sample  handling. Special Report 97-12. U.S. Army Cold Regions Research and
    Engineering Laboratory, Hanover, New Hamsphire.
                                                   10

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


  FINAL EVALUATION OF METHOD 3546: A MICROWAVE-ASSISTED PROCESS (MAP™)* METHOD FOR
            THE EXTRACTION OF CONTAMINANTS UNDER CLOSED-VESSEL CONDITIONS

                             J.R. Jocelyn Pare and Jacqueline M. R. Belanger
                                 Microwave-Assisted Processes Division
                                     Chung Chin and Richard Turle
                          Air Analysis and Quality Division, Environment Canada,
                     Environmental Technology Centre, Ottawa, ON, Canada K1A OH3
                                             Barry Lesnik
 US EPA, Office of Solid Waste, Method Section, 401  M. St., SW, Washington, District of Columbia 20460, USA
                                    Rodney Turpin and Raj Singhvi
    US EPA, Emergencies Response Branch, 2890 Woodbridge Ave., Edison, New Jersey 08837-3679, USA

ABSTRACT
Microwave-assisted extraction (MAP™) has been the subject of enhanced interest from the environmental sector
in the past year as a  result  of the need for  methodologies that will  improve sample preparation without
compromising the quality of the data while being sustainable environmentally. Liquid-phase microwave-assisted
extraction (MAP™) offers such advantages: it is  a very fast extraction technique, it consumes less solvent and
energy,  and  it is  cost  effective. A preliminary validation  study  involving closed-vessel  apparatus and
contaminants  such as PAHs, PCDDs/PCDFs,  chlorinated  pesticides,  and PCBs  was performed. Excellent
performance and precision were achieved for these analytes. In order to fully evaluate the method for the  range
of analytes an inter-laboratory study was  performed.  A round-robin study was  performed  with five laboratories
and  involved  thermally  labile  RCRA  target  analytes  such  as phenols,   phenoxyacids  herbicides and
organophosphorus pesticides. Three split samples were used along with a single standard operational procedure
(SOP). All analyses were performed by a single laboratory in order to minimise the variability of the results due
to the determinative procedure. Clean up was performed using  standard  procedures and analysis was done
according to the appropriate SW-846 methods. The broad range of applicability, the reduced sample preparation
time and the reduced amount of solvent used all  contribute to reach sustainable environmental protection goals.
Furthermore, the reduced operational costs associated with the protocol compared to conventional Soxhlet for
example - are significant and will prove valuable in these times where the "greening" of the laboratory usually
gives rise to higher operating costs.  Further work involving open-vessels apparatus is under way.

INTRODUCTION
The microwave-assisted process (MAP) is a  technology patented by Environment Canada1'3 The most widely
used  applications to date make use of microwave energy to extract soluble materials from different matrices,
mostly using organic solvents45. Microwave energy has been used in various ways to extract organic compounds
from a variety of matrices6. For example, the technology has been applied to organochlorinated pesticides from
sediments and  PCBs  from  water7,  petroleum  hydrocarbons from  soil8,  and  to  herbicides  from soils910.
Lopez-Avila et a/, used a MAP-approach to extract several groups of pollutants such as PAHs, PCBs, pesticides,
phenols  and base/neutral  compounds in  soils  and  sediments1113.  In all  these studies, microwave-assisted
extraction proved to be similar or more efficient than methods based upon the use of Soxhlet or ultrasound.

More recently, we reported on a preliminary validation of a draft method for inclusion into US EPA Test Methods
for Evaluating Solid Waste Physical/Chemical Methods (SW-846)14 Table  1 presents the types of compounds
that have been subjected to that preliminary work.

The method  reported herein makes  use  of partially microwave-transparent solvents  (or a  mixture of such
solvents) contained into a closed vessel.  Although not as elegant and efficient as methods using open-vessel
microwave-transparent solvents, it provides the possibility of combining the benefits of heat (enhanced solubility
and diffusivity) to the action of the microwaves on the  matrix.  Work is currently underway to validate open-vessel
methods and the results will be reported elsewhere in due time along with the parameters to be controlled to
effect even  more efficient  extraction  procedures using MAP. This paper reports on the final evaluation of a
closed-vessel  microwave-assisted extraction  procedure for environmental pollutants from soils and sediments
that recently met with the  US EPA approval and will be referred to as Method 3546 under SW-846. It is a
procedure for extracting water insoluble or slightly water soluble organic compounds from soils, clays, sediments,

*  MAP is a Trade-mark of Her majesty The Queen in Right of Canada as represented by the Minister of the
Environment.


                                                  11

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
sludges, and other solid wastes. The method is applicable to the extraction of semi-volatile  organic compounds,
organo-phosphorus  pesticides,  organo-chlorine pesticides, chlorinated herbicides, phenoxy acid herbicides,
substituted phenols, PCBs, and PCDDs/PCDFs which may then  be analyzed  by a variety of chromatographic
procedures.

TABLE 1. List of target analytes used in preliminary validation package	
 Base Neutral and Acid             PAH                             Organo-chlorines
 Hexachloropentadiene               Acenapthylene                    PCB-I
 Dimethylphthalate                   d10-Acenaphtene (surrogate)        PCB-2
 Diethylphthalate                    Fluorene                          Hexachlorobenzene
 Di-n-butylphthalate                  dio-Phenantrene (surrogate)        Simazine
 Butylbenzylphthalate                Phenanthrene                     Atrazine
 £>/'s(2-Ethylhexyl)adipate             Anthracene                       Lindane
 b/s(2-ethylhexyl)phthalate           Pyrene                           PCB-3
                                    Benzo(a)anthracene               Alachlor
 Pentachlorophenol                  dn-Chrysene (surrogate)           Heptachlor
                                    Chrysene                         PCB-4
                                    Benzo(b)fluoranthene              Heptachlor epoxide
                                    Benzo(k)fluoranthene              PCB-5
                                    Benzo(a)pyrene                   gam/na-Chlordane
                                    d12-Perylene (surrogate)            a/prta-Chlordane
                                    lndeno(123-cd)pyrene              frans-Nonachlor
                                    Dibenzo(ah)anthracene             PCB-6
                                    Benzo(ghi)perylene                Endrin
                                                                     PCB-7
                                    di4-Terphenyl (int. std)             Methoxychlor
	PCB-8	

This  method  has been validated  for solid  matrices  containing  50 to 10,000 ug/kg of semi-volatile  organic
compounds, 250 to 2,500 ug/kg of organo-phosphorus pesticides, 10 to 5,000 ug/kg of organo-chlorine pesticides
and chlorinated herbicides, 50 to 2,500  ug/kg of substituted phenols,  100 to 5,000 ug/kg of phenoxy  acid
herbicides, 1  to 5,000 ug/kg of PCBs, and 10 to 6000 ng/kg of PCDDs/PCDFs.

EXPERIMENTAL
The experimental procedures for the preliminary validation work has been presented elsewhere, hence all the
text presented herein  refers exclusively to the inter-laboratory work and is relevant to Method 3546 as approved.

This method  is applicable to solid samples only with small particle sizes. If practical, soil/sediment samples may
be air-dried and ground  to a fine powder prior to extraction. Alternatively, if worker safety or the  loss of analytes
during drying is a concern, soil/sediment  samples may  be mixed with anhydrous sodium sulfate or pelletised
diatomaceous  earth.  The total  mass  of material  to  be prepared depends on  the specifications  of the
determinative method and the sensitivity required for the analysis, but 2 - 20 g of material are usually necessary
and can be accommodated by this extraction procedure.

Safety
The  use  of  solvents combined  with the  operational  parameters associated  with this  method  will raise
temperatures and pressures in the extraction vessels to values that can be a safety concern in the laboratory.
Only equipment designed for laboratory use  and manufactured under legitimate rights should be used to ensure
that proper safety devices are built into the apparatus. Common sense laboratory practices can  be employed to
minimize this concern. For example, the following sections describe some additional steps that should be taken.

The extraction vessels are at elevated temperatures and pressure after the extraction stage. Allow the vessels to
cool  before  opening  (the  use of a water  bath is recommended  for this  purpose) and always  monitor the
temperature and pressure by re-connecting the control vessel to the apparatus prior to opening the vessels.

During the heating step,  some solvent vapors  may escape through the vessel  liner/seal cover.  Follow the


                                                   12

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


manufacturer's directions regarding the vessel assembly  and  instrument set up to prevent release of solvent
vapors to the laboratory atmosphere.  The instrument may contain flammable vapor sensors and should be
operated with all covers in place and doors closed to ensure  proper operation of the sensors. If so equipped,
follow the manufacturer's directions regarding replacement of extraction vessel seals when frequent vapor leaks
are detected.

Extraction
Decant  and discard any water layer  on a sediment sample. Mix the sample thoroughly, especially composite
samples. Discard any foreign objects  (sticks, leaves, rocks, etc.). Air dry the sample at room temperature for 48
hours in a glass tray or on hexane-rinsed aluminum foil. Alternatively,  mix the sample with an equal volume of
anhydrous sodium sulfate or pelletised diatomaceous earth until a free-flowing powder is obtained.

If multiphase waste samples are used, then they must be  prepared by the  phase separation method in Chapter
Two  of SW-846  before extraction. Dry sediment/soil and  dry waste samples  amenable to grinding. Grind or
otherwise reduce the particle size of the waste so that it either passes through a 1-mm sieve or can  be extruded
through a 1-mm hole. Disassemble  grinder between samples,  according to manufacturer's instructions, and
decontaminate with soap and water, followed by acetone and hexane rinses.

Gummy, fibrous,  or oily materials not  amenable to grinding should be cut, shredded, or otherwise reduced in size
to allow mixing  and  maximum  exposure of the  sample  surfaces for the extraction. The  analyst may add
anhydrous sodium sulfate, pelletised  diatomaceous earth, sand, or other clean, dry reagents to the sample to
make it more amenable to grinding.

Grind a sufficient weight of  the dried sample to yield the  sample weight needed  for the determinative  method
(usually 10-30 g). Grind the sample  until it passes through a 10-mesh  sieve. Prepare a method blank using an
aliquot of a clean solid matrix such as quartz sand of the approximate weight of the samples. Add the surrogates
listed in the determinative method to each sample and method blank.  Add the  surrogates and the matrix spike
compounds appropriate for the project to the two additional aliquots of the sample selected for spiking.

A volume of about 30 ml_ of the appropriate solvent system is added to the vessel and sealed. The extraction
vessel containing the sample and solvent system is heated to the extraction temperature and extracted for 10
minutes. The solvent systems used for this procedure vary with the analytes of interest and are listed below. The
mixture is allowed to cool. The vessel is opened and the contents are filtered. The solid material is rinsed  and the
various solvent fractions  are combined. The  extract may be  concentrated,  if  necessary,  and,  as needed,
exchanged into a solvent compatible with the cleanup or determinative step being employed.

Six vessels were always placed in  the  microwave oven at any one time to  standardise conditions. After
extraction, the sample carousel was removed from the microwave and cooled in a water bath. To ensure that it
was safe to proceed  with the filtration step the control vessel was returned to the microwave oven and the
temperature was  monitored before opening. Solvent loss was checked randomly in  some instances and found to
be below 1%.

Interferences
Refer to Method  3500 of SW-846. If  necessary, Florisil and/or sulfur cleanup procedures may be employed. In
such cases, proceed with Method 3620 and/or Method 3660 of SW846.

Apparatus and Supplies
CEM   Corporation  (Matthews,   NC) MAP™  solvent   extraction   systems  equipped  with  appropriate
microwave-transparent extraction  vessels should  be transparent  to microwave energy  and  capable of
withstanding  the  temperature and pressure requirements (200°C  and 200 psi) for  this procedure.  Models
MES-1000 or MSP-1000 have been used for the present work.

Solvents Systems and Reagents
All solvents should be pesticide quality or equivalent. Solvents may be degassed prior to use.

Organo-chlorine   pesticides,  organo-phosphorus  pesticides,  semi-volatile  organics  may be  extracted  with
acetone/hexane (1:1, v/v) or acetone/methylene chloride (1:1, v/v).
                                                  13

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium



                                        e extracted with acetone/hexane  (1:1,  v/v), or acetone/methylene
          <
chloride (1:1, v/v), orhexane.

Phenoxy acid herbicides and phenols may be extracted with acetone/hexane (1-1 v/v) and the phosphate buffer
solution.

Reagent grade chemicals shall be used in all tests. Organic-free reagent water should be used. Sodium sulfate
(granular anhydrous), Na2SO4 and pelletised diatomaceous earth can be used  as desiccant.  They should be
purified by heating at 400°C for 4 hours in a shallow tray, or by extraction with methylene chloride. If the latter
approach is used,  then a reagent blank should be prepared to demonstrate that the drying  agent is free of
interferences.

Phosphate buffer solution - for  use in  extraction  of  phenols  and phenoxyacid herbicides. Prepare a  0.1 M
phosphate buffer solution by adding 1.2 g reagent grade sodium phosphate into a 250-mL beaker, add 100 ml of
reagent water and thoroughly mix. Adjust the solution pH to 2 with the addition of reagent grade phosphoric acid

Quality Control
Chapter  One and  Method  8000 of SW-846 should be  followed  for specific Quality Control  procedures and
Method 3500 should be followed for sample preparation quality control procedures. Surrogate standards should
be added to  samples when listed  in the appropriate determinative method.

RESULTS AND DISCUSSION
Reference 14 presents a large body of information and specific data on a number of analytes. It provides the
basis for a major portion of the  performance work associated  with this procedure. References 12 and 15 are
reports of similar, more specific studies. References 16 to 18 deal specifically with phenols. Representative data
sets are presented in Tables 2 to 6. They are not exhaustive and are reported here as they are new data. Other
data can be  found in the references cited herein including references 19 and 20.

Chlorinated  pesticides: Single-laboratory  accuracy data were obtained for chlorinated pesticides using natural
soils, glass-fiber, and sand matrices. Concentrations of each target analyte ranged between 0.5 to 10 ug/g. Four
real-world split samples contaminated with  pesticides and creosotes were also used (obtained from US EPA
ERT, Edison, NJ). The latter were extracted by an independent laboratory using standard  Soxhlet procedures
and results compared to those obtained with this procedure. Extracts were analyzed by the appropriate method.
Method blanks  and five  spiked  replicates  were included. Work was  also carried  out to assess the level of
degradation  of thermally labile pesticides and it was found that  no  significant degradation takes place under the
procedure described herein. The  data are reported in detail in Reference 4. Data summary tables are included in
Method 8081 .

TABLE 2. Single-laboratory organochlorine pesticides analysis data from a real contaminated soil _
 Compounds          Average        Std. Dev.           RSD               n           REAC value
                        (ppb) _ (%) _ (ppb)
DDE+Dieldrin
Endrin
*DDD
*DDT
*Methoxychlor
a-Chlordane
Y-Chlordane
3380
21500
40000
62670
16500
730
720
340
2290
5750
8430
1980
100
90
10.06
10.66
14.38
13.45
12.03
13.37
12.47
3
3
3
3
3
3
3
7100
22000
45000
62000
16000
750
910
"(dilution 1:5); Soil samples obtained from US EPA Emergency Response Center archive bank through their
contract laboratory REAC (Edison, NJ). The standard Soxhlet extraction procedures were performed by REAC
three years earlier; this long storage period is believed to account for the low DDE+Dieldrin recovery data in the
present study. DDE+Dieldrin is the sum of the compounds since they were not resolved by chromatography.

Semivolatile orqanics: Single-laboratory accuracy data were  obtained for semivolatile  organics  using  natural
soils, glass-fiber, and sand samples. Concentrations of each target analyte was about 0.5 ug/g. Extracts were
analyzed by  the appropriate  method. Method blanks and five  spike  replicates were included. The  data are
reported in detail in Reference 14. Data summary tables are included in  Method 8270.
                                                   14

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


PAHs: Single-laboratory accuracy data were obtained for PAHs using five reference materials comprising marine
sediments (HS-3, HS-4,  and HS-5,  all from the National Research Council of Canada),  lake  sediments
(SRM-1491, from the  National Institute  of Science  and Technology), and a  natural soil contaminated with
creosote (SRS103-100, from Fisher Scientific, Fairlawn, NJ). Natural soils, glass-fiber, and sand samples were
also used  in spiked matrices work.  Concentrations varied  between 0.1 and 2000 ug/g. One real-world split
sample contaminated with creosote and pesticides was also used (obtained from  US EPA ERT, Edison, NJ). The
latter was extracted by  one laboratory using standard Soxhlet procedures and results compared to those obtained
with this procedure. Extracts were analyzed by the appropriate method. Method blanks,  spikes and five spiked
replicates were included.  Surrogates were used  in real-world split  sample. The data  are reported in detail in
Reference 14. Data summary tables are included in Method 8270.

PCBs: Single-laboratory accuracy data were  obtained  for PCBs using  three reference  materials  EC-1, EC-2,
EC-3 (from Environment Canada). Natural soils, glass-fibre, and sand samples were also used in spiked matrices
work. Concentrations varied between 0.2 and 10 ug/g (total PCBs). Extracts were analyzed by the appropriate
method. Method  blanks, spikes  and spike duplicates were included for the low concentration spikes; matrix
spikes were included for all other concentrations. The data are reported in detail in Reference 14. Data summary
tables  are included in Method 8082.

TABLE 3. Single-laboratory PCB recoveries data from certified Great Lake sediment materials	
 Sediment                  Aroclor       Std Dev.        RSD           nCertified value
                             (ppb)	(%)	(ppb)
EC-1
EC-2
EC-3
1850
1430
670
0.07
0.09
0.02
3.78
6.60
3.12
3
4
3
2000 ± 54
1160 ±70
660 ± 54
Sample size = 2 g extracted into a final volume of 4 ml; EC-2 and EC-3 certified values were provisional values
only, at the time the work was conducted. The data presented herein were part of the validation data package
used to confirm the certified values. Real samples were also tested when fortified with mixtures of native Aroclor
(1242, 1254, and 1260) to a 600 ppb level. Recoveries were in the 88% range with a reproducibility of 2% RSD.

Chlorinated herbicides (phenoxyacid herbicides): Multi-laboratory accuracy data  were obtained for chlorinated
herbicides spiked at 100 ng/g in one soil type. A certified spiked material was used (obtained from ERA, Arvada,
CO). Extracts were analyzed by Method 8151. Method blanks and three  replicates from five laboratories were
included. Data summary tables are included in Method 815 1.

Phenols: Single-laboratory accuracy data were obtained for phenols using  a number of spiked natural soils and a
number of real-world split soils. Concentrations varied  between 0.2 and 10 ug/g. Extracts were analyzed by the
appropriate method. The data are reported in detail in References 14 to 18. Data summary tables are included in
Method 8041. Multi-laboratory accuracy data were obtained for phenols spiked at 250 ug/kg in one soil type. A
certified spiked material was used  (obtained from ERA, Arvada,  CO). Extracts were  analyzed by Method 8041.
Method blanks and three replicates from five laboratories were included. Data summary tables are included in
Method 8041.

TABLE 4. Multiple-laboratory phenoxyacid herbicides recoveries from certified spiked material	
 Compounds                      Average                 Recovery                   RSD
                                                              (%)
2,4-D
2,4-DB
2,4,5-T
2,4,5-TP (Silvex)
Dicamba
Dichlorprop
Dinoseb
81
122
74
68
50
87
118
81
122
74
68
50
87
118
13.0
15.1
11.8
17.9
17.6
20.7
29.4
Material spiked at 100 ug/kg. Number of participating laboratories = 4. N = 3
                                                  15

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


TABLE 5. Multiple-laboratory phenols recoveries from certified spiked material
Compounds
2,4-D
2,4-DB
2 4 5-T
2',4',5-TP (Silvex)
Dicamba
Dichlorprop
Dinoseb
Average
(M9/kg)
81
122
74
68
50
87
118
Recovery
(%)
81
122
74
68
50
87
118
RSD
13.0
15.1
11.8
17.9
17.6
20.7
29.4
Material spiked at 250 ug/kg. Number of participating laboratories = 4. N = 3

Omanoohosphorus pesticides and  chlorinated herbicides:  Multi-laboratory accuracy data were  obtained for
organophosphorus pesticides spiked at 250 ug/kg in one soil type. A certified spiked material was used (obtained
from ERA, Arvada, CO). Extracts were analyzed by Method 8141. Method blanks and three replicates from five
laboratories were included. Data summary tables are included in Method 8151.

TABLE 6. Multiple-laboratory organophosphorus pesticides recoveries from certified spiked material	
 Compounds                                Average             Recovery              RSD
                                              (M9"
-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


ensure even shorter extraction times as well as higher extraction efficiencies.

ACKNOWLEDGEMENTS
The split samples were provided by Mr. Yi-Hua Lin and Vinod Kansal by Roy Weston Laboratory, REAC, Edison
through arrangements of Mr.  Phil Campagna of the US  EPA. Some  supplementary funding from the
Environmental Protection Service of the Department of Environment of Canada  is acknowledged. We are
especially indebted to the laboratories,  other than our own, that contributed to the inter-laboratory work, namely
APPL (Fresno, CA), US  EPA Region VI (Houston, TX), Emergencies Science Division of Environment Canada
(Ottawa, ON, Canada), and the US EPA/AED (Narragansett, Rl). The support provided by SAIC (Reston, VA) is
also  acknowledged.  The use of  trade names and Trade-marks throughout this  report does not imply an
endorsement or recommendation of use. The processes described herein may be subject to certain patents rights
and this paper does not in any way suggest or condone their use without proper rights to do so.

REFERENCES
1. J.R.J. Pare, M.  Sigouin, and J.  Lapointe, U.S. Patent 5,002,784,  March 1991  (various international
   counterparts).
2. J.R.J. Pare, U.S. Patent 5,338,557, August 1994 (various international counterparts); U.S. Patent 5,458,897,
   October 1995 (various international counterparts).
3.   J.R.J.  Pare,  U.S.  Patent  5,377,426,  January  1995 (various  international  counterparts); U.S. Patent
   5,519,947, May 1996 (various international counterparts).
4.  J.R.J. Pare, J.M.R. Belanger, and S.S. Stafford, Trends Anal. Chem. 13 (1994) 176.
5.  L.G. Croteau, M.H. Akhtar, J.M.R. Belanger, and J.R.J. Pare, J. Liquid Chromatogr.
    17(1994)2971.
6. K. Ganzler, A. Salgo, and K.  Valko,  J. Chromatogr. 371 (1986) 299.
7. F.E. Onuska and  K. A. Terry, Chromatographia 36 (1993) 191; J. High Resol. Chromatogr. 18 (1995) 417.
8. E. Hasty and R. Revesz, American Laboratory (2) (1995) 66.
9. R. Steinheimer, J. Agric. Food Chem. 41, 588 (1993).
10. S. J. Stout, A.R. daCunha, and D.G. Allardice, Anal. Chem. 68 (1996) 653.
11. V. Lopez-Avila, R. Young, J. Benedicto, P Ho, R. Kim, and W.F. Beckert, Anal. Chem. 67 (1995) 2096.
12. V. Lopez-Avila, R. Young, and W.F. Beckett, Anal. Chem. 66 (1994) 1097.
13. V. Lopez-Avila, J. Benedicto, C. Charan, R.  Young, and W.F Beckert, Environ. Sci. Technol. 29 (1995) 2709.
14. K. Li, J.M.R. Belanger, M.P Llompart, R.D. Turpin, R. Singhvi,  and J.R.J. Pare, Spectres. Int. J. 13 (1996) 1.
15. R.  McMillin,  L.C.  Miner,  and L. Hurst. Abbreviated microwave extraction of pesticides and PCBs in  soil.
   Spectros. Int. J. 13 (1), 41-50 (1997).
16. M.P.M.P Llompart, R.A. Lorenzo, R.  Cela,  and J.R.J.Pare. Optimization of a microwave-assisted extraction
   method for phenol and methylphenol isomers in soil samples using a central composite design. Analyst, 122,
   133-137(1997).
17. M.P Llompart, R.A. Lorenzo,  R. Cela, J.R.J.  Pare, J.M.R. Belanger,  and K. Li.  Phenol and methylphenol
   isomers determination in  soils  by in-situ microwave-assisted extraction and derivatisation. J. Chromatogr. A
   757, 153-164(1997).
18. M.  P. Llompart, R. A. Lorenzo, R. Cela, K. Li, J.M.R. Belanger, and J. R. J. Pare. Evaluation of supercritical
   fluid extraction,  microwave-assisted extraction and sonication  in the determination  of  some phenolic
   compounds from various soil matrices. J. Chromatogr. A, 774, 243-251  (1997).
19. C. Chiu, G. Poole, Y. Shu, R. Thomas, and R. Turle. Microwave vs. Soxhlet for the extraction of dioxins and
   furans from solid samples. Organohalogen Compounds 27, 333-338 (1996).
20. C.  Chiu, G.  Poole, M.  Tardif,  W.  Miles, and  R. Turle.  Determination  of PCDDs/PCDFs  Using
   Microwave-Assisted Solvent Extraction. Organohalogen Compounds 31, 175-180 (1997).
                               PHENOXYACID HERBICIDE SCREENING

                       Mark L. Bruce. Kathleen L. Richards and Raymond M. Risden
                      Quanterra Inc., 4101 Shuffel Dr. NW, North Canton, OH 44720
                                       brucem@quanterra.com
                                                 17

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Abstract
EPA SW-846 Methods 8150 and 8151 have been traditionally used for phenoxy acid herbicide analyses. These
methods require the use of hazardous derivatization reagents  and a highly flammable solvent. The sample
preparation process is very long, complex, labor intensive and time consuming. Immunoassay, as used in EPA
SW-846 Method 4015, uses aqueous  based chemistry with minimal solvent volume and much simpler and faster
sample  preparation.

Quanterra has demonstrated that EPA SW-846  Method 4015 (2,4-D)  can  be extended  to include 2,4,5-T and
silvex by using both the 2,4-D and silvex kits from Strategic Diagnostics Inc. It is useful for screening both water
and  soil samples  for phenoxyacid  herbicides.  Non-detect  samples from the immunoassay  (IA) screen are
reported as nondetect at the applicable reporting limit. Samples which have responses greater than the threshold
(i.e.  positive response) are confirmed by the traditional methods. The overall  false negative rates of 0.5% for
waters and 1.0% for soils were well below the EPA Office of Solid  Waste  criteria of 5% at the reporting limits
listed below. The false positive rates were 12.5% and 11.5% for waters and soils respectively. Water reporting
limits were 2 ug/L, 10 ug/L and 10 ug/L for 2,4-D, silvex and 2,4,5-T  respectively. Soil  reporting  limits were 1
mg/kg, 1.5 mg/kg and 1.5 mg/kg respectively.

Switching to  IA  improves  laboratory  safety, reduces  organic  solvent  usage  and  disposal,  improves
turn-around-time and reduces analytical costs.

Introduction
Many herbicide analysis  requests target three main analytes of interest: 2,4-D,  2,4,5-T and silvex (2,4,5-TP).
Also, >95%  of these samples are reported as "non-detect" for these analytes.  Thus,  it would be very useful to
screen samples submitted for the. analysis of these herbicides.  Only those samples with herbicides above the
reporting  limit  would  be subjected  to quantitation  by the  more  exhaustive  traditional  analysis  methods.
Immunoassay is capable  of providing the screening information with an aqueous chemistry based method that is
simpler, faster, safer and  less expensive.

Currently  Immunoassay Method 4015 is only applicable to 2,4-D. However, this validation study demonstrates
that  it is possible  to extend the method by including a Strategic Diagnostics Incorporated (SDI) analysis kit
developed specifically for silvex by Ohmicron. Both the 2,4-D and silvex immunoassay kits have cross reactivity
for 2,4,5-T because of structural similarities. This cross reactivity enables the  combined use of the two kits to
effectively screen for all  three compounds, The  cross reactivity is shown  below expressed  as least detectable
dose (LDD).

                                Table 1. Cross Reactivity
Compound
2,4-D
2,4,5-T
Silvex
2,4-D Assay
LDD (ug/L)
0.70
3.0
170
Silvex Assay
LDD (ug/L)
100
1.0
1.4
The Ohmicron (now part of SDI) 2,4-D kit has been validated following Office of Solid Waste (OSW) guidelines.
It was approved by the OSW Organics workgroup and incorporated into Method 4015. Method 4015 was part of
the Update III package for SW-846 promulgated in mid 1997. Ohmicron had established threshold test levels of
10  ug/L in water and 150 ug/kg  in soil with their magnetic particle based assay kit. The sensitivity difference
between the two matrices is mostly due the  large dilution  of the methanol soil extract needed to reduce antibody
exposure to methanol.

The Ohmicron Silvex kit was validated  similarly to the 2,4-D kit but the results were not submitted to the OSW
workgroup. The silvex kit has comparable sensitivity to the 2,4-D  kit. Ohmicron has documented that the silvex
assay responds similarly to 2,4,5-T (i.e. cross reactivity).

Immunoassay Method Summary
The immunoassay form applied in these two kits is enzyme linked immunosorbent assay (ELISA). Information is
available  from  SDI product literature and  the EPA - Las  Vegas (web  site http://www.epa.gov/crdlvweb/asb/
immuchem/forum.htm) describing IA basics.
                                                   18

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


For these  SDI kits a subaliquot (250  uL)  of aqueous sample is combined with 250  uL of enzyme conjugate
solution. One portion of the conjugate molecule resembles the analyte of interest. Antibody which is bound to
small magnetic particles is added to this  analyte/conjugate  mixture. The  analyte and  conjugate compete for
antibody binding sites.  The higher the analyte  concentration the less conjugate is bound to the antibody. The
lower the analyte concentration the more conjugate is bound to the antibody.

A magnet is  used to  retain  the antibody  particles and  the  analytes and  conjugates that  have bound to the
antibodies.  All other analytes,  conjugates  and matrix components are washed away.  A color development
reagent  is added to the  antibodies. This reagent reacts with another portion of the conjugate molecule and
develops color proportional  to the amount of bound  conjugate  present. Thus, the  color (i.e.  absorbance
measured  on the filter photometer) is highest (~1) for blanks and clean samples and the  least absorbance is
produced for high concentration standards and samples. This  inverse relationship between analyte concentration
and absorbance response has caused some confusion.

Thus, when using this IA format for threshold testing the results scenarios are:

a)  if sample absorbance > threshold  standard  absorbance then analyte concentration is < threshold (e.g. non
detect at 10 ug/L), b) if sample absorbance < threshold standard absorbance then analyte concentration is >
threshold (e.g. positive analyte > 10 ug/L)

Since the color response will always have  some variability associated with  both the threshold standard and the
sample,  it is  customary to prepare the  threshold  standard at a concentration slightly below the reported test
threshold.  This "low bias" on the standard  favors the assay toward producing false positives in order to reduce
the false negative rate. The following 10 ug/L report threshold  example illustrates:

Table 2. Example Assay Batch
Assay
threshold standard #1 (7 ug/L)
threshold standard #2 (7 ug/L)
LCS(10ug/L)
sample #1
sample #2
sample #3
sample #4
sample #5
sample #6
sample #7
NIS of ND sample #1(>10 ug/L)
MS of ND sample #2(>10 ug/L)
Absorbance
0.8
0.88
0.6
0.97
0.92
1.02
0.86
0.4
0.55
0.7
0.57
0.86
Comment

use the avg std response 0.84 for cutoff
pass
ND<10
ND<10
ND<10
ND<10
positive >10
positive >10
positive >10*
pass
fail - false negative**
   sample #7 may be a false positive at the test threshold since it falls between the. threshold standard and the
   LCS. In other words it may have herbicide present at 8 ug/L. It would be sent for 8150 confirmation.
** This MS illustrates a documented false negative which would trigger corrective action.

The threshold standard (7 ug/L) is intentionally lower than the reported test threshold (10  upg/L) to reduce the
false negative rate due to normal statistical variability and minor preparation losses and matrix interferences.

Quanterra intended to improve the sensitivity of the soil assay by reducing or eliminating the need for the large
dilution  used to reduce the impact of methanol on the antibodies. Evaporating the methanol from a small aliquot
of sample extract, followed by reconstitution in water was effective at removing the methanol.  However, this
process also concentrated the interferences in the soil  matrices and raised the level of non-specific antibody
binding.  Unfortunately,  this  raised the false  positive rate  to an unacceptable level.  Thus,  this preparation
modification was not used in the validation study.

The silvex kit has not  been "validated" for 2,4,5-T but  the SDI  cross reactivity data  suggest that the silvex
antibodies work similarly for 2,4,5-T threshold testing. 2,4,5-T was investigated along with 2,4-D and silvex for
which the IA kits were designed. In fact, 2,4,5-T was detected by both the 2,4-D and silvex IA kits during the
validation study. At least one  kit produced an acceptable 2,4,5-T response  for each sample included in the study.
                                                   19

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Sympo:
mum
This validation study builds on top of the existing studies  and knowledge discussed above.  Quanterra has
demonstrated that the two IA herbicide kits are appropriate to screen out water and soil samples that have no
detectable levels of 2,4-D, 2,4,5-T or silvex at the applicable  reporting limits. This is not intended to replace
Method 8151 for quantitation but to focus its use where it is most applicable. Since the extended IA method is not
intended for quantitation and  much validation work  has already  been completed for the IA kits very  little
additional direct comparison between IA and GC data is necessary to establish that the IA method is capable of
screening out  non-detect samples. We intended to utilize matrix spikes of well characterized matrices, matrix
spikes of real  samples (ND for herbicides by 8150/51) and a few real samples with herbicides that have  been
previously analyzed  by 8150/51. Hereafter,  herbicides  that entered  the sample by environmental transport
mechanisms and have undergone aging and weathering  are referred to as native herbicides, Unfortunately, the
native herbicide samples that were available at the time of the validation study could not be used  as  intended
because the analyte concentrations were much lower than the final test thresholds. To compensate,  many of
these samples were fortified with known amounts  of each herbicide and included in the study as matrix spiked
samples.

Validation Goal
The main goal was to demonstrate that the IA method responds to herbicide levels known to be at or above the
IA  reporting limit (threshold)  for each analyte.  Since  samples with native  herbicides above the final  test
thresholds were not available, the number of matrix spiked samples was significantly increased over the original
validation plan.  Water matrix spiked samples were increased  from 6 to 18. Soil matrix spiked  samples  were
increased from 7 to 12. Also, using matrix spikes  allowed us to more accurately document performance at the
critical area near the reporting limit. It was not necessary to analyze  the matrix spiked samples with 8150/51
since the herbicide  concentrations were already known.  The matrix spiked soil samples were stored at 4°C at
least overnight and usually for several days between spiking  and extraction  in order to age them and more
closely mimic native analytes.

Validation Protocol
Matrix spiked samples were prepared using techniques previously employed for SW-846 methods development
work. Separate spike solutions containing known amounts of each of the three target analytes were prepared at
appropriate  concentrations in acetone. These high  concentration standards were further diluted in reagent water.
Water samples were spiked with these aqueous solutions, homogenized by shaking and immediately  assayed.
Soil samples were spiked, acetone solvent evaporated at room temperature and tumbled overnight in a rotary
mixer. The  spiked soil samples were  stored at least overnight  at 4°C and usually for several days prior to
extraction in order to "age" the samples and thus more closely mimic native soil samples. Obviously  this process
does not duplicate the extensive weathering which can occur in real environmental samples, but the combined
use of short term "aging" and fuller's earth that is  known to be  difficult to extract should simulate many difficult
samples. Each  of the following sample-analyte  combinations in Tables 3 & 4 was assayed at  least once.
Replicate analyses indicated in Q.

Analytical Procedure Summaries
Water preparation
Allow particulates in water sample to settle, filter (0.45  urn PTFE) if sample is cloudy with suspended solids,
aliquot 250  uL of sample into plastic assay tube. Aqueous matrix  spike solution added as appropriate. TCLP
buffers and  samples were spiked as appropriate in  a small vial. A 25 uL aliquot was transferred to the assay tube
and 225  uL  of diluent was added.

So/7 preparation
Weigh 10 g of soil sample into 50 ml plastic centrifuge tube. Add acetone matrix spike solution as  appropriate.
Allow solvent to evaporate. Add 2-3 ball bearings.  Tumble overnight (note: some wet clay samples formed large
clumps during tumbling and were not tumbled when matrix spiking was required for additional analyses). After
"aging", add 30 mL  of extraction solvent (75% methanol, 23% reagent water, 2% acetic acid). Recap centrifuge
tube, rotate  onto side  and mechanically shake at about 200 cycles/minute for 30 minutes. Stand centrifuge tubes
up  and allow soil to settle for 1 hour and/or centrifuge for 3-5 minutes. Filter (0.45 urn PTFE) a few milliliters of
extract. Transfer 5 uL of extract into plastic assay tube and add 245 uL of diluent.
                                                   20

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                 Table 3. Water Sample List and Matrix Soike Levels

Sample

ground water
industrial waste water
TCLP buffer #1
TCLP buffer #2
TCLP Buffer #1 samples t
X3
02
43
01
ZX
Analyte
No spike

(10)
(4)
(13)







Water samples t
HN
HV
QK
5G
DA
W2
4JL
(6)
(7)
(5)
(4)
(5)
(6)
(5)
Performance Evaluations
Wise. PE
WS038 PE
(3)
(3)
2,4-D
high
10
10
100
100(7)

100
100
100
100
100

10
10
10
10
10
10
10




low
2(4)
2









2
2
2
2
2
2(8)
2



si I vex
high low
10 5
10 5
100(2)
100(7)

100
100
100
100
100

10 5
10 5
10 5
10 5
10 5
10 5
10 5



2,4,5-T
high low
10(27) 5
10 5
100(27)








10 5
10 5
10 5
10 5
10 5
10 5
10 5



Additional water samples screened for false positives t
8H
AH
HK
we
5V
65
6
K1
K2
W3
W6
W7
WA
JW
KO

(4)
(4)








































































t Samples previously analyzed by 8150 but no herbicides detected.

Immunoassay
Allow all IA reagents  (particularly the enzyme  conjugate and antibodies) to warm to room temperature. Add
sample or standard aliquot (250 uL final volume). Add 250 uL of enzyme conjugate. Add 500 uL of suspended
antibody coupled magnetic particles  and vortex mix for 1-2 seconds. Incubate at room  temperature for 30
minutes. Apply magnetic rack base for 2 minutes to separate magnetic particles from bulk fluid in tubes. Pour out
tube contents while the magnetic particles are retained at the bottom of the tubes. Rinse the antibodies twice with
wash solution. Remove  magnetic rack base. Add 500 uL of color development reagent and incubate for 20
minutes. Add 500 uL of sulfuric acid stop solution and read absorbance at 450 nm.

The assay tubes were arranged in a 3 X 10 layout in the SDI magnetic rack. Often two batches of 30 assays were
performed simultaneously in the rack which holds a maximum of 60 tubes. A typical  30 tube layout is shown
below. It was quite common in the early batches of the validation study to assay standards prepared by both SDI
                                                  21

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


and Quanterra® (QES) as a means of verifying the standards prepared in-house.
Table 4. Soil Sample List and Matrix Spike Levels
Sample
sandy soil
loam soil
fuller's earth* (dry)
fuller's earth* (50%
moisture)
Soil -samples t
03
04
2N
36
20
20
22
25
1V
NO
spike
(5)
(4)
(4)
(3)
(4)
(4)

high
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0

2,4-D
ug/kg
low
0.3 (2)
0.3 (2)
0.3 (2)
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
Analyte
silvex
ug/kg
high low
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5(8)
1.5
1.5
1.5

2,4,5-T
ug/kg
high low
1.5(2)
1.5(2)
1.5(2)
1.5
1.5
1.5
1.5
1.5
1.5
1.5(8)
1.5(8)
1.5
1.5
* Fuller's earth is an absorptive clay known to challenge the efficiency of solid extraction procedures.
t Samples previously analyzed by 8150 but no herbicides detected above (or near) spike concentrations.

Raw Response Data
The average difference in response between duplicate SDI and QES standards was 4%. The response difference
between silvex and 2,4,5-T was small (average 12.5%) and 2,4,5-T consistently produced less response. Thus,
2,4,5-T was used as the threshold compound for all silvex kit assays  in order to reduce the false negative rate,
particularly for 2,4,5-T containing samples. The test threshold standards were prepared at 70% of threshold
concentration  (0.7 x expected  reporting level) in order to reduce false negatives for water matrices. Threshold
standards were prepared at 50% of the threshold concentration for the  soil assays.
Table 5. Example IA Batch Layout for 2,4-D kit assays - 10 ug/L Reporting Level and matrix spikes
1
2
3
ABODE FGH I J
QES
"10"std
W2
SDI
"10"std
GW
W2+
D
W6
GW+
D
W3
W6+
D
TCL
P
W3+
D
W7
TCLP j WW
_jUL^A
QES LCS
atlORL
W7+
D
SDI LCS
atlORL
Wise
PE
WW
+ D
W4
Wise
PE
W1
W4+
D
WS038
PE
W1 +
D
W5
WS038
PE
SDI
"10"std
W5+
D
QES
"10"std
 Key:
QES "10" std = 7 ug/L 2,4-D standard prepared by Quanterra
SDI "10" std = 7 ug/L 2,4-D stock standard prepared by SDI
QES LCS at 10 RL = 10 ug/L 2,4-D standard prepared by Quanterra
SDI LCS at 10 RL = 10 ug/L 2,4-D stock standard prepared by SDI
D = 2,4-D spiked sample (10 ug/L)
GW = Ground water
TCLP = TCLP buffer #1  [10 X dilution]
WW = Industrial waste water
W1, W2, W3, W4, W5, W6, W7 = real water samples without native herbicides
PE = Performance evaluation samples from Wisconsin and EPA WS038 programs.

Results and Discussion
The key questions when evaluating the reliability of this screening method are:
1) If an analyte is present  in  a sample at  a concentration greater than or equal to the threshold  (i.e. reporting
                                                  22

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


limit), what are the chances of the assay generating a false negative response?

2) If no analyte is present in a sample at or above the threshold concentration, what are the chances of the assay
generating a false positive response?

                             Table 6. Threshold Standards
Threshold Level
(reporting limit)
2.0 ug/L
10. ug/L
1 .0 mg/kg
1.5 mg/kg
Compound
water samples
2,4-D
silvex or 2,4, 5-T
soil samples
2,4-D
silvex or 2, 4, 5-T
actual
standard concentration
1 .4 ug/L 2,4-D
7.0 ug/L 2,4,5-T
3.3 ug/L 2,4-D
5.0 ug/L. 2,4,5-T
False negatives are undesirable since they would report a sample as "clean" with regard to the target herbicides
when it was not and potentially increase environmental health risks. False positives are undesirable since they
would  needlessly "trigger" a batch of traditional 8151  analyses which would increase  organic solvent  usage,
analyst exposure to hazardous reagents, turn-around time and cost.

Since the same number of replicates were  not run for each sample it is not appropriate to simply divide the
number of false negatives or positives by the total number  of assays. The false negative or positive rate was
determined for each sample - analyte  combination  or  unspiked  sample. These individual rates were  then
averaged to determine the overall false negative and positive rates. LCS results were not included in these
calculations because they do not contain real sample matrix. The false negative rate must meet the  normal
Office of Solid Waste criteria, <5%. The false positive rate was expected to be <10%.

The matrix spike results from samples  producing  consistent false  positive responses were not  included when
calculating the  false negative rates for either water or soil samples.

Water Samples
Production threshold levels of 2 ug/L for 2,4-D and 10 ug/L for silvex and 2,4,5-T were selected after evaluating
the initial water sample data. Table 7  summarizes the percentage of false positives and false negatives for each
sample - analyte combination.

The overall false negative rate of 0.5%  for water samples was excellent. Sample W2  had one false negative
among 8 replicates at the 2 ug/L  threshold for 2,4-D for a false negative rate of 12.5% for this sample -  analyte
combination. The ground water  (GW) and  TCLP  buffer #1  produced  many false negatives for 2,4,5-T when
assayed with the silvex kit at the 10 ug/L (GW) and 100 ug/L  (TCLP) levels. Keeping the spike levels and  analyte
the same  but  switching to the 2,4-D kit significantly improved the assay reliability.  The false  negative rates
dropped to 16.7% and  5.9% for  these two matrices respectively. Since 2,4,5-T is not a normal TCLP  analyte
these false negatives have little  practical impact on assays  of TCLP samples, although this does indicate  that
2 4 5-T may be more susceptible  to false negatives than the other two analytes. The ground water false negative
rate was reduced  by  assaying  for  2,4,5-T with  the 2,4-D  kit as  noted above. This  indicates that either
immunoassay kit will respond to 2,4,5-T  and  that at least one kit is likely to produce a positive response at the 10
ug/L threshold when 2,4,5-T is present.

The false positive rate was determined by dividing the total number of "non-detect"  samples included in the study
into the number of samples that produced false positives for a false  positive  rate of 12.5%. Replicate assays
were  performed on two of the samples that produced false positives. This confirmed  that a matrix interference
existed which produced the false  positive. One of the false positive sample responses was near the response of
the threshold standard and probably would not produce a false positive in all instances  if replicate assays were
performed   It appears that false positives are primarily generated by matrix interferences. Thus, positive matrix
interferences are likely to be site  specific and the false positives that do occur in actual production use of the IA
kits should  be  clustered together in a limited number of sample lots.  This means that most of the sample lots
assayed are expected to be free of false positives.
                                                   23

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium



Table 7. False Negative and False Positive Rate Calculations for Water Samples
Matrix or sample

ground water
waste water
TCLP buffer #1
TCLP buffer #2
X3
02
43
01
zx
HN
HV
QK
5G
DA
W2
JL
Wise PE
WS038 PE
Matrix or sample

8H
AH
HK
we
5V
63
65
6
K1
K2
W3
W6
W7
WA
JW
KO
Average rates

No spike
% false positives
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0


No spike
% false positives
0
100
100
0
0
0
0
0
0
0
0
0
100
0
0
100
12.5%

2,4-D
% false negatives
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1/8 = 12.5
0
0
0
2,4-D
% false negatives
















0.7%
overall false
Silvex
% false negatives
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Silvex
% false negatives
















0%
negative rate 0.5%
2,4,5-T
% false negatives
3/18=16.7
0
1/17 = 5.9






0
0
0
0
0
0
0


2,4,5-T
% false negatives
















2.3% |

 Soil samples
 Attempts to improve the sensitivity of the soil method were unsuccessful.  Initial information from SDI indicated
 that the 50X extract dilution included in the SDI soil preparation method was designed to remove the deleterious
 effect that methanol has on the immunoassay. We intended to remove the  methanol interference by evaporating
 a small aliquot of the  extract to dryness then redisolving  the analytes  in diluent solution or water.  Analyte
 recovery  appeared acceptable and there was no visible methanol residue, but there was still some small positive
 interference  which probably would have prevented reliable  assays at the 30 pg/kg target threshold. Real soil
 extracts showed very large positive interferences when 250 uL of extract were concentrated for the assay. Thus,
 it was  necessary to restrict  to  soil  preparation  to the  original  SDI/Ohmicron  soil  prep method.  This  was
 accomplished by diluting 5 pL of extract with 245 \iL of diluent. When this SOX dilution was combined with higher
 analyte spike levels the interference problems were reduced to an acceptable level.

 The following preparation  method  was used for  most soils:  10 g soil +  30 ml  extract solvent, shake for 30
 minutes,  settle or centrifuge, filter, 5 pL of extract + 245 |jL diluent. Thus, the final concentration of analytes
                                                    24

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
presented to the assay were 150X more dilute than in the original soil sample. Quantities of some samples were
limited so the preparation amounts were scaled back to 5 g soil and 15  mL of extraction  solvent. The assay
threshold standard concentrations reflected the dilution  built into the sample prep. Most spike levels were 1.0
mg/kg for 2,4-D and  1.5 mg/kg for silvex and  2,4,5-T. Some low level (0.03 and 0.3 mg/kg) work was attempted,
but the false positive rate was  unacceptably high.

The concentration of the test threshold standard concentration was also  reduced  from 70% to 50% of threshold
for soil assays. This decreased the false negative rate without unacceptably raising the false positive rate. Table
8 below shows 3 soil - analyte combinations had false negative results for  silvex  or 2,4,5-T when  spiked initially
at 1.5 mg/kg. Replicate sample aliquots were  spiked, extracted and assayed. No additional false negative results
were  produced. Thus,  the overall  false negative rate was 1.0%, which is excellent. False positive results were
reported for two samples at the  1.0  mg/kg 2,4-D or 1.5  mg/kg silvex/2,4,5-T assay levels. The overall false
positive rate for soil  samples was 11.5%. It is expected that false positives will be primarily caused by  specific
constituents in the soil samples and  are thus more likely to  be related to specific sites rather than be evenly
distributed through all sample lots.

    Table 8. False Negative and False Positive Rate Calculations for Soil Samples
Matrix or sample
ground water
loam soil
fullers earth dry
fullers earth wet
03
04
2N
36
2E
20
22
25
1V
Average rates
No spike
% false positives
0
0
0
0
50
0
0
0
100
0
0
0
0
1 1 .5%
2,4-D Silvex
% false negatives % false negatives
0
0
0
0
0
0
0
0
0
0
0
0
0%
average false negative
0
0
0
0
0
0
0
0
1/8= 12.5
0
0
0
1 .0%
rate = 1 .0%
2, 4, 5-T
% false negatives
0
0
0
0
0
0
0
0
1/8= 12.5
1/8 = 12.5
0
0
2.1%
The soil extraction procedure did not appear to suffer any serious extraction  efficiency problems despite its
simplicity and short time frame. Even though the IA results were not quantitative, the low false negative rate
indicates that analyte recovery was > 50% and many recoveries were > 70%. Previous Quanterra work with wet
fullers earth with nonpolar analytes and solvents, showed very low (<20%) analyte recovery for hydrocarbons.
Good extraction efficiency of the phenoxy acid herbicides with the simple shake extraction used with these IA
kits was probably due to the following reasons: 1) The polar extraction solvent  (methanol / water / acetic acid)
readily permeated the wet clay matrix. 2) The polar solvent molecules could effectively displace polar analyte
molecules from the polar sorption sites on the matrix. 3) The polar analytes were readily soluble  in the polar
extraction solvent.

Conclusion
The validation study results are summarized in Table 9 below. The false  negative rates easily meet the normal
EPA Office of Solid Waste criteria of <5%. The. false positive rates although slightly higher than  the target 10%,
are still acceptable. The water reporting limits are in the low part-per-billion range and meet the validation plan
minimum  objectives.  In   particular,  analyses  from  TCLP  buffers  and  samples  demonstrated excellent
performance  for 2,4-D and silvex at levels  well below the regulatory limits. The soil reporting limits although
higher than originally expected should still be useful for many types of herbicide samples.

           Table 9. Summary of Performance Results and Reporting Limits
Matrix
water
2,4-D
2ug/L
1 mq/kg
Silvex
10ug/L
1 .5 mg/kg
2, 4, 5-T
10ug/L
1 .5 mg/kg
False Negative
0.5%
1 .0%
False Positive
12.5%
1 1 .5%
                                                   25

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Acknowledgments
The authors would like to thank Allen Debus from US EPA, Region 5, Craig Kostyshyn and Timothy Lawruk from
Strategic Diagnostics and Susie Dempster, Carolynne Roach, Chris Lee and Richard Burrows from Quanterra
Inc.

References
1. Strategic Diagnostics Incorporated Product Literature - 2,4-D Rapid Assay
2. Strategic Diagnostics Incorporated Product Literature - Silvex Rapid Assay
3. SW-846 Method  4015 - Screening for 2,4-Dichlorophenoxyacetic Acid by Immunoassay.
      FIELD DEMONSTRATION OF A PORTABLE IMMUNOSENSOR FOR EXPLOSIVES DETECTION

               Anne W. Kusterbeck1, Paul T. Charles1, Paul R. Gauger2 and Charles Patterson1
    1Centerfor Bio/Molecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375
                                phone: 202 404-6042, fax: 202 404-8897
                                   akusterbeck@cbmse.nrl.navy.mil
                2GeoCenters, Inc., 1801 Rockville Pike, Suite 405, Rockville, MD 20852-1633

Environmental biosensors are being developed at the Naval Research Laboratory for detection of the explosives
TNT and RDX in groundwater and  monitoring of cleanup progress for these compounds at remediation  sites.
Based on a displacement immunoassay, the portable sensor, known as the FAST 200, has been engineered by
Research International (Woodinville, WA) to quantitate water samples with no sample preparation or reagent
addition. Analysis is complete within five minutes. The sensor, along with a fiber optic biosensor, recently was
extensively tested in field trials at several U.S. EPA Superfund sites to validate sensor performance.  Results of
these studies and application of the technology will be described.
                  ENVIRONMENTAL APPLICATIONS OF A FIBER OPTIC BIOSENSOR

                    Lisa C. Shriver-Lake1, Irina B. Bakaltcheva2, and Saskia van Bergen3
    1Center for Bio/Molecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375
                               phone: (202) 404-6045, fax: (202) 404-8897
                                     email: lcs@cbmse.nrl.navy.mil
                2GeoCenters, Inc., 1801 Rockville Pike, Suite 405, Rockville, MD 20852-1633
                 3George Mason University, 4400 University Drive, Fairfax, VA 22030-4444

Detection and remediation  monitoring of the explosives TNT and RDX on-site requires a sensitive and preferably
portable method. The fiber optic biosensor is based on a competitive fluoroimmunoassay being performed on the
core of an optical fiber probe. A portable version of the sensor was engineered by Research  International
(Woodinville, WA) and is known  as the Analyte 2000. With this sensor, four  optical probes can be monitored
simultaneously and relatively Adirty@  samples can be employed. Analysis takes 16 minutes for the four probes.
This sensor,  along with the  FAST 2000,  was extensively evaluated at three  on-site trials to validate sensor
performance. Results of these studies and other applications for the fiber optic biosensor will be described.
                                                 26

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


   DEVELOPMENT AND VALIDATION OF AN IMPROVED IMMUNOASSAY FOR SCREENING SOIL FOR
                            POLYNUCLEAR AROMATIC HYDROCARBONS

                                 Titan S.  Fan and Brian A. Skoczenski
                Beacon Analytical Systems, Inc., 4 Washington Ave., Scarborough, ME 04074

Polynuclear aromatic hydrocarbons (PAHs) are a group of fused ring compounds most typically found as
combustion by-products. Over the past 4 - 5 years a number of immunoassay kit manufacturers have developed
and commercialized soil screening method for PAHs. Briefly, there methods involve a rapid extraction of soils by
shaking with methanol followed by analysis of the filtered sample extract using competitive enzyme
immunoassay. We have developed an improved method for screening of PAHs in soil samples. The new method
utilizes a modified sample extraction step that  results in improved extraction efficiency compared to earlier
methods. In addition, the specificity of the antibody used in the method allows for a better estimate of the total
PAHs present. Immunogen design and antibody specificity will  be described in detail. Results of concordance
study with a gas chromatographic method will be presented.
     A NEW DIOXIN/FURAN IMMUNOASSAY WITH LOW PICOGRAM SENSITIVITY AND SPECIFICITY
              APPROPRIATE FOR TEQ MEASUREMENT: APPLICATIONS DEVELOPMENT

                                         Robert O. Harrison
                    CAPE Technologies, L.LC, 3 Adams St., South Portland, ME 04106
                                         Robert E. Carlson
                    ECOCHEM Research, Inc., 1107 Hazeltine Blvd, Chaska, MN 55318

Since 1990 the  commercial  development of immunoassay kits his  opened a new market in environmental
analysis. The US EPA has approved more than  10 immunoassay screening methods under the 4000 series of
Field Screening  Methods within SW-846. These tests are now  widely used in  assessment and remediation of
hazardous waste sites.

This recent success has not included the  development of a useful immunoassay method for polychlorinated
dibenzo-p-dioxins and   polychlorinated  dibenzofurans (PCDD/Fs).  The  development  and  application of
immunoassays for PCDD/Fs  pose unique challenges not found in immunoassays for other analytes.  First, the
sensitivity requirements of PCDD/F analysis are  typically in the  ppt range rather than the high ppb to mid  ppm
range of the existing 4000 series screening methods. Second, if the test is to provide useful data, the EIA
response  must  correlate to  the  relative  toxicity of the  17  most  toxic  PCDD/F congeners. No  previous
immunoassay  has  demonstrated  the  necessary  combination  of  sensitivity  and specificity required  for
measurement of toxic equivalency (TEQ) at ppt  levels. No immunoassay specific sample preparation methods
have been developed because of the obvious lack of commercial potential demonstrated by all previous PCDD/F
immunoassays

A new enzyme immunoassay (EIA) for PCDD/Fs has been developed using novel chemistry. The sensitivity of
this test is  approximately 4 pg of 2378-TCDD, which is more than an order of magnitude better than previous
PCDD/F  immunoassays. Based on typical sample size, this sensitivity is  sufficient to  measure low ppt TEQ
levels in  solid samples  or 0.1 ng/m3 TEQ in stack gases using only a small fraction of the prepared sample
extract. This sensitivity allows detection of 2378-TCDD in a 10 uL sample aliquot at the  same concentration as
the lowest calibration solutions typically used for HRGC/HRMS  based PCDD/F methods such  as EPA Methods
1613, 8290, and 23. These sensitivity comparisons indicate that the EIA is capable of screening samples prior to
HRGC/HRMS analysis without consuming an unacceptably large  proportion of the sample.

The dioxin/furan  congener cross-reaction profile of this EIA is suitable for TEQ measurement. The test is most
sensitive to the three most toxic congeners, 2378-TCDD, 12378-PnCDD, and 23478-PnCDF EIA specificity data
plus  HRGC/HRMS data from previously analyzed samples have been utilized in a simple, additive  response
model to predict  the EIA response for each sample. The resulting correlation between predicted EIA response
and TEQ validates the concept of TEQ screening  by EIA for a variety of samples.


                                                27

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
A muliti-laboratory collaborative effort is now in progress for evaluation of kit performance and development of
sample  preparation methods.  Results for fly ash  and soil validate the additive response  model for  TEQ
measurement  by  EIA.  These  results also validate the use of the kit for screening fully cleaned samples.
Extension of this validation to partially cleaned samples is in progress, with positive initial results for fly ash and
soil. Work on immunoassay specific sample preparation methods, including rapid extraction of soil and fly ash, is
also  in progress. The ultimate  goal of this coli program is to develop sample preparation  protocols which will
maximize throughput and cost-effectiveness of immunoassay based PCDD/F screening.
           DEVELOPMENT AND VALIDATION OF AN IMMUNOASSAY FOR SCREENING SOIL
                                FOR POLYCHLORINATED BIPHENYLS

                                  Titan S. Fan and Brian A. Skoczenski
                Beacon Analytical Systems, Inc., 4 Washington Ave., Scarborough, ME 04074

Polychlorinated Biphenyls (PCBs) were commonly used in  electrical applications due to their properties of high
thermal transfer and low conductivity. Since being identified as toxic substances their  use has been banned in
the US. Due to the once widespread use and their stability  in the environment  a large number of sites are
contaminated  with  PCB residues.  A  number of  immunoassay kit  manufacturers have  developed and
commercialized soil screening methods  for PCBs. Briefly, these methods involve  a rapid extraction of soils by
shaking with  methanol followed by  analysis  of  the  filtered  sample  extract  using  competitive  enzyme
immunoassay. We have developed an improved method for screening of PCBs in soil samples. The new method
utilizes  a  modified sample  extraction step that  results in  improved extraction  efficiency compared to earlier
methods.  The immunoassays performed on the sample  extract  and yields qualitative  results at  1, 5, 10 or 50
ppm. The test can be used for measuring Aroclors 1016, 1242, 1248, 1254 and 1260. Results of a concordance
study with a gas chromatographic method will be presented.
         GASOLINE RANGE AROMATIC/ALIPHATIC ANALYSIS USING PATTERN RECOGNITION

                                           Steven E. Bonde
               Petro Star Inc., 201 Arctic Slope Ave., Suite 200, Anchorage, Alaska 99518-3030

 Finding an analytical technique for the analysis of aromatic and aliphatic compounds in the gasoline range of
 hydrocarbons is of great interest to not only the laboratory but also the regulated community. In many states,
 including Alaska, great pressure to apply risk based cleanup standards is driving the need to separate "high  risk
 compounds" from a given matrix in an economical manner with high confidence. False positives can mean an
 unneeded costly cleanup, while false negatives can mean non-compliance and possible fines if found. Analysis
 of aromatic and aliphatic compounds in the gasoline range has been accomplished using pattern recognition
 algorithms based on PCB matching criteria. Using existing methodologies for the analysis of gasoline range
 hydrocarbons in multimedia samples by GC-FID-PID this algorithm has been developed using widely available
 software to recognize and quantify aromatic hydrocarbons in  a given sample. Soil, sediment, and water samples
 were analyzed using standard Alaska Department of Environmental Conservation methodology
 (AK101/EPA8021) and the results of pattern matching showed a significantly reduced number of false positives
 for the aromatic portion of the analysis.
                                                  28

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


            BIOREMEDIATION ASSESSMENT USING CONSERVED INTERNAL BIOMARKERS

                        P. Caleavecchio. E.N. Drake, G.C. VanGaalen and A. Felix
                Corporate Research Laboratories, Exxon Research and Engineering Company,
                                Rt. 22 East, Annandale, New Jersey 08801
                                            (908) 730-2308

ABSTRACT
The lack of homogeneity of field samples is often a primary concern in the analytical assessment of hydrocarbon
bioremediation treatment efficacy. One approach  to handling the sampling variability is to take a statistically
significant number of samples to adequately represent the distribution of the contaminant.  This approach  is
limited in that small  biodegradation changes often cannot be seen within the inherent sampling variability. An
additional drawback of this approach is the high cost of analyzing sufficiently large numbers of samples to draw
conclusions.

An alternative approach, which can minimize the effects of sampling variability,  is to use naturally occurring
molecules, which resist biodegradation  as  conserved internal  "biomarkers".  There are a number of marker
classes commonly found in petroleum products, which can be analyzed  by  GC/MS, these  include hopanes,
steranes and isoprenoids. These markers can be used as the basis for relative comparisons of target analytes
before and after treatment. The ratios  of two or more markers can also be used as a fingerprint to help identify
the petroleum product source.

This poster demonstrates the application of biomarker normalized GC/MS data to evaluate the  biodegradation of
petroleum products  from well-mixed  refinery sludge waste. A comparison  is shown tracking the  treatment
progress using absolute target analyte quantities and relative  amounts normalized to biomarkers.  The results
show that the biomarker normalized data provided a good  basis for determining the percent biodegradation of
total oil as well as the molecular components, similar to the results obtained using absolute quantities.

The conclusions of the above study indicate that although significant amounts of 2, 3, 4 and 5 ring  polynuclear
aromatic hydrocarbons (PAHs) were removed during treatment the biomarkers used to normalize the data were
conserved, resisting significant biodegradation. The implication of this work is that the  use of biomarkers can be
effective in  situations that are less homogeneous  to assess  biodegradation of petroleum constituents from
products with similar sources.

INTRODUCTION
The objective  of the following study was to remediate a refinery process sewer  sludge, through aerobic
biodegradation,  using composting to  accelerate  this process.  Composting is a modification of the "biopile"
process,  which includes the addition of biodegradable ammendments (straw, wood chips, manure, etc.) to help
aerate and supply energy for microbial growth. The laboratory scale evaluation was conducted  in  three-liter
insulated glass beakers designed to minimize the loss of heat generated by the compost.

To minimize the effects of sludge heterogeneity  on analytical results,  a number of steps were  taken which
included  mixing the sample well. An additional approach was to use naturally  occurring "biomarker" molecules,
which resist biodegradation, as internal standards. The  latter approach was taken considering that future field
applications would probably have much higher sampling variability due to the difficulty in homogenizing large
volumes  of compost.

The sludge used in these tests had significant amounts of two biomarkers typically found in petroleum products,
C30 17a,  21B(H)-Hopane ("hopane") and  C29 25-Nor-17a(H)-Hopane ("norhopane"). The above molecules are
related by a mechanism proposed by  Peters and Moldowan that suggests the origin of norhopane through the
biological demethylation of hopane (see figure 1). Both biomarkers were found by selective ion (SIM) GC/MS  at
m/z 191,  norhopane also has a characteristic peak at m/z 177.

The progress  of the composting  experiments were tracked by periodic sampling during  the  period of peak
biological activity as  indicated by naturally occurring  elevated temperatures within the  cells. The analytical tests
performed on  the extracted hydrocarbon included Total Petroleum Hydrocarbon  (TPH, modified New Jersey
method OQA-QAM-025-10/91) and GC/MS analysis for priority pollutant (PP) polynuclear aromatic hydrocarbons
(PAHs). The  TPH data  was quantified by total ion  GC/MS response using an  external calibration of free oil


                                                  29

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                          Aerobic
                       Blodegradatton?
    2» 24


   C3017a(21/J(H)-HoPnne
C,g 2S-Nor-17a(H)-Hopan«
 O0-De»me«hylhopane")
Proposed   origin  of   25-norhopanes  by
bacterial demethylation of 17a (H)-hopanes.
The  methyl  group  attached to  the  C-10
position in  the C3o 17a (H)-hopane (left) is
removed to produce the C2g 25-norhopane
(right).

Figure  1.  Proposed  origin  of  norphopane
from hopane
                             Figure 2. Biomarker ratio of hopane/norhopane

collected from the sludge.  The PAHs  were quantified  using  internal standards spiked  into the extracts as
prescribed in EPA SW846 method 8270.

The "biomarker normalized" data was obtained by dividing the GC/MS response of the analyte by that of hopane.
The degradation  measured  relative to hopane was then  compared to the "absolute" values obtained using the
traditional approach described  in the above paragraph.  The conservation of hopane throughout  the test, was
monitored by comparing its GC/MS response relative to that of norhopane.

SUMMARY
The results show that rapid degradation of  both TPH  and PAHs occurred  over  the six-week  period  of the
composting experiments,  using both the absolute measurement and the hopane normalized  approach.  During
this period there  was no indication  to suggest that significant hopane degradation  occurred as shown by little
change  in the ratio of hopane to norhopane (see figure 2). Also qualitative examination of the biomarker's ion
chromatograms at m/z 191  and 177 show little change before and after treatment (see figures 3 and 4). In
contrast, the total ion chromatogram of the extracted hydrocarbon shows significant removal of  the oil  during
treatment (see figure 5).

The TPH removal during treatment is shown in figure 6,  which compares the absolute  measurement to the
hopane  normalized results.  These  results  show somewhat higher degradation estimates relative  to hopane,
ranging  from 4%  higher after one week to 30% higher after 6 weeks of treatment. The comparison  of a selection
of the more abundant 3, 4 and 5 ring PP PAHs using the absolute and hopane-normalized  approaches is shown
in figures 7  and 8  respectively.  This comparison shows good  agreement for  phenanthrene and   pyrene
measurements using  both  approaches,   with  a  somewhat  lower estimate  for the  hopane  normalized
benzo(a)pyrene degradation (~ 30% lower). The lower concentration  of benzo(a)pyrene (1 ppm) compared to
                                                  30

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


phenanthrene (27 ppm) and  pyrene  (15  ppm) may have attributed to the variability seen in its degradation
measurements.
                 Abundance	1911 191.00  (150.71) to  lt>1.70J :  2Ptt-H-I.l3
                     2000J

                            Norhopane
                     1500 J
                     1000 J
                      500 J
                 Time-->
              Hopane
               37.32
m/z!91
                                   37! do' ialdo' 'aslo'o' 40! do' '411 do' «loo' 43!do' 4* loo'
700 -1


600


500


400


300 J


200


100


  0
                                36
                            Norhop.
                                    Iqn 177.00  (176.70 to 177.70): ZDB-14-1.D
                                  46
                                                           m/z!77
                 ritne-->
                              so  3700 3800  39o 40o 4lo
                                                                    4300 4*00
                 Figure 3. Before treatment ion chromatograms with hopane and norhopane

The degradation of the C1, C2 and C3 alkylated homologs of phenanthrene and anthracene are shown in figures
9, 10 and  11  respectively. This comparison shows very close  agreement between the absolute and hopane
normalized approaches (less than 10% difference) over the duration of the composting experiments.

CONCLUSIONS
This work demonstrates a practical application using selective ion GC/MS to characterize the biodegradation of
petroleum contaminants relative to naturally occurring biomarkers. The hopane and norhopane biomarkers used
in the above composting experiments showed no sign of significant biodegradation over the tests duration, while
234 and  5 ring  PAHs as well  as TPH showed  rapid biodegradation. A  comparison of biodegradation
measurements using conventional quantification and biomarker normalized approaches showed good agreement
for the most abundant analytes with somewhat higher variability for those in low concentrations.

The biomarker approach was easy to  use and is independent of many sources  of analytical and sampling
variability associated with absolute measurements of analytes. Although this was a controlled experiment using a
relatively homogeneous sludge,  many pilot and field applications often deal with a wide range of contaminant
distribution resulting in sampling variability which is difficult to  handle. The large numbers of samples needed to
measure statistically  significant changes  can be  avoided using  biomarkers since analyte responses  are
normalized to stable (naturally occurring) internal standards.
                                                  31

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposiun
An additional benefit of biomarkers is in site characterization of petroleum spills where ratios of two or more
biomarkers can be used as a fingerprint to help identify the source of the contaminant. The degree of weathering
and natural attenuation in various areas of a spill can also be assessed using this approach.
Abundance
1800-
1600 .
1400
1200 -
1000 -
BOO -
600 -
400 .
200
lime--*
HOT






-^WjvV^-V\ftW
Ion 191. dO (190.70 b'6 m. 70) .- 30B-27-J.il 	 	
'
hopane
37.32





A-JLu\
opane
m/z!91
I
j
1
II
.UuJ 1 UJuLxvJ^^^^ — ^ — w— .
3e'.00 37100 38.00 39.00 4oldo 4l!6o 42lo'o 43ldo 44^00
Abundance
800 ,




Time-
700 .
500 .
500
400 -
300 -
200 -
100 -
0 -
- >
36
He



'^Wu
Ion. 177.00 U7S.70 to 177.70): 20«-av-3.B
45
rhopane
m/zl77


ft 1.

36^00 37^00 38.00 39^00 4o'.00 4l'.00 iizloO 43.00 44.00
                   Figure 4. After treatment ion chrmatograms with hopane and norhopane

 ACKNOWLEDGMENTS
 Dr. Roger C. Prince for his insights into the application of biomarkers to oil spills.

 REFERENCES
 The Biomarker Guide -  Interpreting Molecular Fossils in  Petroleum and Ancient Sediments. K. E. Peters and J.
    M. Moldowan, Prentice Hall, Englewood Cliffs, New Jersey 07632
                                                   32

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Abundance'

  200000.




  1SOOOO-




  100000-




  50000
                               TIC:  20B-I4-1.D
                                                   Before
 Time-->      ' 10 lob'  is lob  'a'oio'o'  as loo   '30 lob'  35 lob  4o!oo
Abundance
14000O -
120000-
100000-
eoooo .
60000 -
40000 -
20000 -
0 -
rime-->
J





	 '
10



. o'o '





. ,11, I-l—"-
TIC: 2U8-27-3.D





UU^L.
After




^u^-



^--—
1 1 	 I ' 1 1 1 1 	 1 	 1 	 r-^-1 	 1 	 1 	 1 	 T^-T 	 Y— ] 	 1" I 1 	 T J 1 1 1 1 "T
15.00 20.00 25.00 30 00 35.00 40.00
                                                                      Figure 5. Before and after
                                                                      treament total ion
                                                                      chromatograms
Figure 6. Comparison of
absolute and relative TPH (GC)
degradation
                                                         Tims (woks)
                                                 33

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
   100


   90


   SO


   70


   80
 *  40


    30


    20
                                                                          e  j
                                                                  • Pyrana
                                                                              Figure 7. Degradation of 3, 4
                                                                              and 5 ring PAHs absolute
                                                                              values
                                Time (wuki)
Figure 8. Degradation of 3, 4
and 5 ring PAHs relative to
hopane
                                                 i
                                                                a Absolute
                                                                • Relative to HopollB
                                                                              Figure 9. Comparison of
                                                                              absolute and relative
                                                                              C1-phenanthrene / anthracene
                                                                              degradation
                               Tun« [WMk»
                                                      34

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                                                          Figure 10. Comparison of
                                                                          absolute and relative
                                                                          C2-phenanthrenes /
                                                                          anthracenes degradation
Figure 11. Comparison of
asolute and relative
C3-phenanthrenes /
anthracenes degradation
                                                                                         a Absolute      {
                                                                                         • Relative to Hopane I
                                                          Tim* (wMk>)
                     METHOD 8270 FOR MULTICOMPONENT ANALYTE ANALYSIS

                                Elaine A. LeMoine and Herman Hoberecht
                The Perkin Elmer Corporation, 761 Main Avenue, Norwalk, Connecticut 06859

ABSTRACT
The identification and quantitation of multicomponent analytes (yielding more than one chromatographic peak)
can  be  an analytical and  productivity challenge. Multicomponent analytes such as Aroclors, Toxaphene,  and
technical Chlordane tentatively identified by another method may be confirmed using  SW846 method 8270.
Alternate confirmation  of  a tentative  identification may be made using an electron capture detector (ECD)
method such as  8081  or 8082  with a second  column. For instruments with  sufficient sensitivity, the mass
spectrometer and ECD can be used in parallel for a simultaneous tentative identification and quantification. This
paper will investigate the utility of a new mass spectrometer system for the quantitative identification of a mixture
of multicomponent analytes. The method will be evaluated  for detection limits, linearity, accuracy, and precision.
The  GC-MS method will be compared with the dual column method for analytical capability, productivity,  and
compliance.
                                                  35

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


INTRODUCTION
While the ability to positively identify sample analytes can be accomplished with the use of two columns, it is not
necessarily the most desirable of options. In many cases the confirmational column alone is not sufficient and
additional clean-up procedures need to be performed to eliminate co-eluting analytes. The additional equipment
and analysis time required places productivity burdens on a laboratory. Gas Chromatography/Mass Spectrometry
is widely used because its selectivity enables positive identification without additional sample processing. Along
with the ability to make qualitative determinations, GC/MS is an invaluable tool for providing quantitative results.
Mass Spec methods  however, are generally considered less sensitive than  conventional detector  methods,
although sensitive enough for most applications. The analysis of multicomponent analytes, such as Toxaphene
and the  Aroclors is more of a  challange.  EPA  Method  8270C  states, "In most  cases,  Method 8270 is  not
appropriate for the quantitation of multicomponent analytes, e.g., Aroclors, Toxaphene, Chlordane, etc., because
of limited sensitivity for those analytes. When these analytes have been  identified by another technique, Method
8270  is  appropriate  for  confirmation  of the  presence of these analytes when concentration in  the extract
permits."1 The development  of more sensitive quadrupole mass spectrometry technology along with innovative
sample introduction techniques,  allow for the quantitation of many of these analytes at levels previously  not
achievable.  The data  to  follow illustrates  the  ability  of  quadrupole  mass  spectrometry  to  quantitate
multicomponent analytes at these lower levels. The ability to accurately identify and quantitate using GC/MS can
eliminate the need for additional confirmatory analyses and reduce the amount of sample preparation required.

EXPERIMENT DESCRIPTION
Identical standards were  analyzed using two sets  of experimental conditions. A  50 uL Large Volume Injection
was used in both cases. One set of standards was analyzed using the GC/MS Full  Scan  mode  (FS-50) and  the
second using the Selected Ion Recording mode (SIR-50). Table 1 lists the chromatographic conditions used for
both experiments, while  Tables 2 and  3 list the Mass Spec conditions used for each set. The results  are
evaluated with respect to the accepted  standard analytical techniques.
Perkin-Elmer AutoSystem XL
Column:
Pre-Column:
Oven Temperature Program:
Programable Pneumatic Control (PPC):
Programable Split/Splitless (PSS) Injector:
Injection Volume:
PE-5MS 30 to x 0.25 mm; 0.25 urn film thickness
1 m x 0.32 mm deactivated fused silica
55°C for 5 min., 45°C/min. to 160°C; 6°C/min to 320°C
Helium 1.0 mL/min.
55°C for 4 min.; ballistic to 250°C; Solvent Purge Mode
50 uL
Table 1. Chromatographic Conditions
                                                FS-50
Perkin-Elmer TurboMass Mass Spectrometer
Mass Scan Range:
Scan Speed:
Filament Delay:
Ion Source Temperature:
Transfer Line Temperature:
lonization Mode:
50 - 350 m/z
2.0 scans/sec
5 min.
150°C
250 °C
El
Table 2. Full Scan Mass Spectrometer Conditions
                                                SIR-50
Perkin-Elmer TurboMass Mass Spectrometer
Selected Scan Masses:
Scan Speed:
Filament Delay:
Ion Source Temperature: ,
Transfer Line Temperature:
lonization Mode:
159,231,233
m/z
2.0 scans/sec.
5 min.
150°C
250°C
El
Table 3. Selected Ion Recording (SIR) Mass Spectrometer Conditions

RESULTS
Toxaphene standards at 0.10 ng/uL, 0.20  ng/uL, 0.50 ng/uL, 1.00 ng/uL, and 5.00 ng/uL concentrations were
                                                  36

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
analyzed using both methods. The chromatograms shown in Figure 1 were obtained using the SIR mode. All the
calibration standards clearly exhibit the characteristic Toxaphene pattern.
SO uL SUM
STD126....i3
1O°1 5. Ong Toxaphene
» %-j

STD12& 5

1 1 Q ng Toxaphene
H
STD128 4 ' ' ' ' ' "
rt
1 05 ng Toxaphene
STD12B 3

1 D.2 ng Toxaphene
% J
nl • •**
STDI28 2

1 0.1 ng Toxaphene


^f-i 23C)3
ill I 231 1
_-A ^^-v^AuMt>A^JSt'J\k^vVltjlj(>lji,_^,

20.8B
231 1 23,02
,11 i 231 1
^ _Aw- 	 oA.^A.H^X/MA<^wU^.^mi,, . 	
• • i • • • 	 	 i ... i • .,.,..,,,..,..,.,.
2O H7 21 76 23.02
231 .1 231 231
j^^^^iJ^^
' ' ' 	 ,....,..,,, 	 ....,.,, 	 , .
20.87;231
Jt_j^_u^JU^^^
••'••••'•"• 	 1 .... I ....,...., 	 , .
20.4S_ 21.76 2301
231 K I 231 231

KIR of 3 Cliannels EH
1SS.BO
1 .4667;

-
SIR of 3 Channels El-t-
159 OD


SIR of 3 Channels El-i-
159 DO.
1 S1e6
SIR of 3 Channels £!+
1SS.OD
6.7165
SIR of 3 Channels El-t-
1 S9 l)D

              10.00     12.0O    14.00     16.00    18.00     2000    22.00     24.00     26.00    28.00
                    Figure 1. Calibration Standards show recognizable pattern for all levels.

                        Full Scan Mode
  Compound 3 name: Toxaphene-3
 Correlation coefficient: r ^-- 0.99931 2, rA2 ••= 0.999624
 Calibration curve. 674.670 * x H- 79.0644
 Response type. External std. Area
 Curve t/pe: unear. Origin: txciude, Welgrtting: Null, Axis tran

     6.83e3-|
  Response-

           0.0      2.0     4.0
                                    6.0
                                            8.0
                                                   rrn ng
                                                    10.0
Four chromatographic peaks were selected and
determined    as   representative    of    the
multicomponent     analyte      Toxaphene.
Calibration Factors (CF) were calculated based
on the  integrated  peak areas and the known
standard  concentrations.  From  these results,
the Relative Standard Deviation (RSD) for each
multilevel concentration range was determined.
These results were averaged providing a final
Toxaphene RSD. Correlation coefficients were
calculated  in  a  similar  fashion   and   are
illustrated in Figures 2 and 3.

The  results  of all  the  calibration  data  and
acceptance criteria are listed in Tables 4 and 5.
Both  experimental  results easily  comply with
method performance specifications.
Figure 2. Toxaphene peak #3 Calibration Curve.
                                                    37

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                            SIR Mode
Compound 2 name: Toxaphene- 2 	 '.
CoeiTicient or Determination. 0.993732 Tne Method Dptprtinn Limits (MDLsl Iktprl in
caiioration curve 13SSS3 * x + -2213 52 i ne lvieinoa ueiecuon umiis JMUUS; usiea in
Response type: external std. Area Table 6 are the result of seven (7) replicate
Curve type:
6 7365-1
.





Response-


-2 21 e3-
O

Calibration


Peak #2
Peak #3
Peak #4

unear. origin: include, weighting: NUM. AXIS trans injections of a 0.10 ng/|jL standard using the
^ standard deviation and the t-statistic.
/' Integrated peaks representative of the entire
/ calibration range can be seen in Figure 4. The
S bottom chromatogram was obtained from a
/" 0.05 ng/^L standard which is below the lowest
s^ calibration standard of 0.10 ng/uL. The peaks
y are readily discernible above the noise and
/ can be easily integrated.
^ Figure 3. Toxaphene peak #2 Calibration
sX Curve

.O 1 .Q 2.O 3.Q A.O 5.0
Full Scan-50
Peaks






Toxaphene (Average of 4 peaks)
RSD
Actual

11.0
13.6
12.6
8.4
11.4
Acceptance
Limit
f\ " i^fi^lf^fef^^f&Mait^^^^^Ss




15.0
Correlation Coefficient
Actual

0.99934
0.99949
.99962
0.99948
0.9995
Acceptance
Limit



:'-M?iyi
0.99
Table 4. Comparison with Calibration Acceptance Criteria using Full Scan mode.

Calibration


Peak#1
Peak #2
Peak #3
Peak #4
SIR-50
Peaks






Toxaphene (Average of 4 peaks)
RSD
Actual

10.4
8.2
10.8
7.5
9.2
Acceptance
Limit
iiliSip|u||ilpi^^p



15.0
Correlation Coefficient
Actual

0.99936
0.99973
0.99967
0.99934
0.9995
Acceptance
Limit




0.99
Table 5. Comparison with Calibration Acceptance Criteria using Selected Ion mode.
Calibration Peaks


Peak #1
Peak #2
Peak #3
Peak #4
Toxaphene (Average)
Calculated Analytical Detection Limits
FS-50
(ng/ML)
0.073
0.089
0.105
0.035
0.07
SIR-50
(ng/ML)
0.065
0.009
0.021
0.014
0.02
Table 6. Calculated Detection Limits
                                                    38

-------
                      WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Toxaphene •
STD 1 28_S
100-
Oj
19.47 1 00 ng/uL
231
2G2BE
^A^,.J\^^^Vyv\J
STO12S 4 ' ' '
100-
Area •
Ul J
o-
12s17 ° 5° n9/UL
14546
--"^- X_^-
STD12R_L>
100-
Area
%
STD1
100-
0-

I1947 0.10ng/ul_
, 5642 p
2B 1
1947 0 Q5 ng/uL
231
3612 ^ •%_/"
19.00 1950 20. QQ
20.87
A ii^a ai-78=2J1-aM=4
\yV*xJk

»A-^/v^xJLWV

_20.87
A ifii^ 2175.231:49710
rA\ \ k A
" ~"^*

—•• "N,-/1^^7 ^^t_X^ V

ft 20.B6;231;)9258
/ ^ A i. 21. 75.231. 119B6
U ^-«k^,-\. 	 - - ^ '"^^
I i.i i • J.
20.87
JWi
20. 5D 21

21.75
231
8311
OD 21.50 22.00
SIR Of 3
^23.02
IT 231
„ 1175806
v,. -•^^•w^A-JLys.^
BIR Of 3
23.02:231 :3B978
k, *-•*' \.^-j"^\jf^^A^\__
3IR of 3
23.01:231:8966
Channels EI+
159.00,
2,5766
~^_ 	 —
''M'"
Channels EI+ ,
159.00'
1 43e6
-V . —-
Channels tl*
159 00
5.2165'
SIR nf 3 Channels EI+
23.02
231
5857
22.50 23.00
159 OO
6. OSes
23. 5O :
                                Figure 4. Integrated Toxaphene Peaks
STD209_18

 1OO-I
                                                                                             scan EM- ,
                                                                                                   66
di u^2ny_ i fct

 1QO-,

Toxapfnene
A - ^A.
^-^- 	 . 	 	 . 	 ^-- — '^.,s^-r^-^- ^- '•*•


A--,

i
A
>_^A/W

4

\/K,...^

          1 OO ng Toxapfiene and O.1 ng Pesticide Mix
                   1 cap      1 oso      i 1 op      i.i_s.Q     _^2oo      12so      1300

                            Figure 5. Aldrin and Toxaphene Extracted Ions
   l^Q9_1 8

1 DO-,
                                                                                                  Scan ei-t- .
                                                                                                        SSi
  a i Li:?uy _i a
                             oxaphene
            1OO ng Toxapriene arid O.1 r>8 Pesticide Mix
            366   '   ib'o'o   ' "  ib's'p   '   II'DO     .t.i.sp.     iapo     .1.25O      1?QQ      ISSD

                                 Figure 6. Library Searchable Spectra
                                                   39

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


What sets the  Mass  Spectrometer apart from other forms of detectors is the ability to  selectively-identify
individual masses. Figure 5 shows the Total Ion Chromatogram (TIC) of a mixture of 100 ng/uL Toxaphene and
0.10 ng/uL Pesticide Mix. The Extracted Ion (El) mass 159 is Toxaphene, and the Extracted Ion (El) mass 66 is
Aldrin which was confirmed by a NIST library search as seen in Figure 6. Aldrin is easily identified and integrated
without additional preparatory procedures.

The  Contract Laboratory Program (CLP) lists  0.2  ng/uL as the quantitation limit for Aroclor  1221 using an
Electron Capture Detector. Figure 7 shows Aroclor 1221 well above the noise level at the 0.20 ng/uL quantitation
level, using GC/MS in the SIR mode and large volume injection.
                        XM-Qcldr 1 221
                         O.2O nti/uL_
                                                                                 SIF? MortR
                                                                                  L.V1-5O
                                                                                   31.2231
                                                                                    326  3i
              21.IJU   '^2-ua   23!oU   :iM LJU   2S QO   i>6.OU   'i7.OO   2Q.OO   29 OO   3Q-Otl   31.PO  S

                             Figure 7. Quantitation Limit Pattern Recognition

SUMMARY
The ability of GC/MS to selectively  identify a component based on an extracted ion chromatogram from a
mixture of compounds not only assures  a  positive  identification, but also saves time by eliminating additional
cleanup and  analyses. Recent technological advances in quadrupole Mass Spectrometry have increased the
instrument's sensitivity. The use of Selected Ion Recording provides further sensitivity enhancements. In addition
to the detector and it's mode of operation,  the use of large volume injection with a programmable inlet system
allow for introduction  of larger sample volumes. The combination of these elements enhances the sensitivity of a
GC/MS system so multicomponent analytes  can be identified and  quantified in  an efficient and  productive
manner.

References
1. "Method 8270C, Semivolatile Organic  Compounds by Gas Chromatography/Mass Spectrometry (GC/MS)," in
    Test Methods for Evaluating  Solid Waste Physical/Chemical  Methods,  SW-846, Third Edition, U.S.
    Government Printing Office, Springfield, VA (1996).
                                        BENZIDINE? REALLY?

                                            Roy-Keith Smith
                  Analytical Services, Inc., 110 Technology Parkway, Norcross, GA 30092
                                   (770) 734-4200, Fax (770) 734-4201

ASI has been performing sample analysis using GC-MS Method 8270, with a variety of sample preparations from
SW-846 for many years. In 1995 soil samples were received from a manufacturing site for analysis. The samples
were part of a general site survey of the manufacturers facility to determine what may need remediation efforts in
the future. ASI performed the requested analyses, and found and reported benzidine in several of the samples at
levels approaching 1000 mg/kg. Unfortunately the data reports were simply filed by the manufacturer.

In late 1997, the  manufacturing site became  under consideration for sale to another company. As part of the
                                                  40

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


pre-sale investigation of the site, past analytical records were examined and the benzidine results came to light.
As there had never been any benzidine used or stored on the site and the manufacturing processes involved no
chemical syntheses, there were some questions about the validity of the reported results.

The raw data generated during the sample  analyses was  examined in detail.  The  initial calibration  was
acceptable, as were the daily calibration verifications and tunes.  The system performance criteria were being
met. The initial calibration was  used for quantitation, with retention times and user generated library spectra
being updated on a daily basis. Examination of the raw chromatograms (Figure 1) from the samples revealed a
hump-o-gram.  Random  MS scans  from  the  hump suggested a petroleum-based background  interference.
Although the benzidine hits were buried in the hump-o-gram, rather than being isolated well defined peaks, the
candidates matched up perfectly with the retention time and daily generated  mass spectrum of the benzidine
standard.

The client had not requested a library search for TIC with the original analysis.  By happenstance during the data
review, a library search of the questioned peak was performed. Quite surprisingly the search generated a match
for dibenzothiophene, rather than  the expected benzidine.  Spectra of both compounds  were pulled from the
database for examination.

It is rather startling how similar the mass spectra of benzidine and dibenzothiophene appear in  a fast  visual
comparison (Figures 2 and 3). Both have a dominant peak at m/z 184 and an assortment of low intensity smaller
m/z signals.  The two compounds have the same unit mass molecular weight (184), the  only difference  in the
molecular formulas being the two amino groups in benzidine and the sulfur in dibenzothiophene, Ci2H8(NH2)2 vs.
Ci2H8S. By coincidence the mass of the two amino groups (32) is the same as the sulfur (32).

Detailed peak mass  matching and  relative abundance comparisons of select peaks reveal definitive mass
spectral variances between the compounds that allows conclusive identification of either analyte. First off,  both
compounds have a M + I isotopic peak,  however only dibenzothiophene has a very prominent M + 2 signal due
to the sulfur. The  fragmentation patterns of the molecular ions of the  two compounds are nowhere near alike.
Both compounds  can loose a  hydrogen to give  M-1  signals,  however only benzidine  will continue to  lose
hydrogens giving the M-2 and M-3  peaks. Benzidine can also lose a -NH2 group to generate the M-16 peak at
m/z 168 or a NH3 group,  forming a benzyne at  m/z  167 (M-17). Dibenzothiophene has no fragmentation pathway
to generate either m/z 167 or 168. Extrusion of sulfur from dibenzothiophene gives  rise to a  prominent peak at
m/z152(M-32).

The stability of the molecular  ion  of benzidine is probably enhanced  through ring-expansion of one of the
aromatic rings to include a nitrogen in a seven-membered aromatic ring (aza-tropyllium ion). Concerted ejection
of a CNH4 unit (M-30) from this ring generates  m/z  154. A similar ring expansion, followed by concerted ejection
of SCH from dibenzothiophene forms m/z 139 (M-45) as a significant signal that is quite  undistinguished in the
benzidine spectrum. Other important differences are indicated  in Figures 2 and 3, and include m/z 65 and 77 in
the benzidine spectrum, while Dibenzothiophene exhibits 69 and 79.

Although the distinguishing  features of the two spectra are easy  to overlook by eye, it  was obvious that the
computer spectral matching algorithm was having no such problem, and further investigation focused upon the
user generated spectra. This is displayed in  Figure 4, along  with the spectra  from one  of the  challenged
identifications.  What  was in the sample matched up almost perfectly with what was stored as a spectrum of
benzidine from the standard. Using the identification criteria listed above to examine the library spectra led to the
inescapable conclusion that  the standard used  for initial and continuing calibration was dibenzothiophene rather
than benzidine.

The retention time of dibenzothiophene is slightly  less than that of benzidine, however not  so much as to be
really startling.  Benzidine itself exhibits shifts in absolute and relative retention  times as columns are changed in
the GC-MS. As it  is our  habit to replace columns with  recalibration, the slight shift in retention time from one
initial calibration to the  next was unexceptional. The quick visual  examination  of the  mass  spectra  that
accompanied the recalibration failed to detect any differences.

Dibenzothiophene  is a naturally occurring substance commonly found in high sulfur crude oils. Discussions with
the client revealed that the samples with "benzidine" all came from the soil underneath  a storage area where
several barrels of high sulfur Venezuelan fuel  oil #6 had been  placed. The high sulfur levels had made the oil


                                                   41

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Abundance"
                               TIC: BU27L.D
85OOOOB j
8000000 :
7500000^
1
1
7000000 1
6500000 \
5000000
S5000OO.
50000OO .
4SOOOOO .
40000OO-.
3500000 .
30OOOOO
2500000
200DOOO _
1500OOO 1
1000OOO .
soooao-
0


!




i









           5.00
                   10.00
                           15.00
                                    20.00
                                             25.00
                                                     30.00
                        unacceptable for use in the boilers at
                        the facility.  The GC-MS  chromato-
                        grams  contained  hydrocarbon hump-
                        o-grams  along with the  "benzidine"
                        and other sulfur-containing aromatics.
                        As   all  the   evidence  was  now
                        consistent, the  reports  were re-issued
                        deleting the benzidine hit.

                        The  investigation  was expanded  to
                        include examination of the  calibration
                        spectra both prior to and after these
                        particular samples were analyzed. It
                        was  found  that the problem began
                        several months prior to October, 1995
                        surprisingly   in  the   middle   of  a
                        supplier's lot  number.  The  problem
                        continued after October, 1995, through
                        the  next lot  number  of  benzidine
                        standard  purchased from the supplier.
                        It was  not until Summer, 1996 and a
                        further lot  number change that  the
                        spectra reverted  back  to the  correct
                        benzidine mass spectrum.
                        Figure 1. Chromatogram of sample
ASI contacted the supplier of the benzidine standard and presented them with the above evidence. The supplier
was  not making any spectral checks  upon  purchased stock standards  as part of their QA  program. Only the
technically outdated melting point determination was being verified, and it was not being performed as a mixed
melting point. There were  no ampules  of the particular lot numbers available for examination, however the
supplier offered to reimburse the cost of the standards.

We went  back through every sample analyzed during  the year  that the incorrect  standard  had been used.
Fortunately we found that no other samples had been  reported with hits for benzidine, thus there were no false
positives.  We are  still  in the process of doing a manual search  of  the tape archives for any false  negative
benzidine  hits in these samples. None have been found to date.
How could we have  caught the error and
prevent it's happening again in  the future?
The corrective actions we  have instituted
are  to:  1)  inject new GC-MS  standards
under old calibrations prior to changing out
the column and re-calibrating; 2)  use second
source  standards  to  check   each  new
calibration; and 3) closely examine the mass
spectrum  of  each new benzidine standard
that is purchased.

The lesson learned?  Don't trust anyone's
claims  or  documentation   of   purity  or
authenticity.  Their acceptance  standards
may well be different than your own.
Figure 2. Mass spectrum of benzidine from
Wiley-NIST library
1SB
            111
           ill
 i !'!'> ii i I'l'ii 111 iTI'i
 US  168 17B IBB
    Wafc1"*'
unr
                            Utn
• •lll'l'lll. I'I'll I I l'1'l'l I
 11B  its l^B 1«
                                                    C12H12N2
                                                   42

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
             in



 iti  ii i^11!^"^1^11^"^"^"^1 ^"^"^"y,;1^1  ^.
                                                            Figure 3. Mass spectrum of dibenzothiphene
                                                            from Wiley-NIST library
        zfc1 •• ak
oB  yB  OB  ?B
 D Lie troth lophene
               tft 119  \St
                                                     C1ZHBS
                          Abu'ndance
                              5000 .
Figure 4. Mass spectra
from sample and lab
generated library
                          Abundance
              41  55  69   92   11S9   I  "'

             ~rVJ,' ,*^ ,-VlA"l'VV"l'Jl'Jll-| .' W\- \	I'rvf'i' 'i ''l''i'
              40   60  80   100  120  140  II
                                                                         196
                                                                         '• ' '
                                                                              218232246  264278
                              5000
                           I   J ,   1U11.  ,    	-—.— «-«.»w  «u-K«.fU

                           160  180 '266  zio''aJo^^o^aS"OT
           Scan" TffTB~T21.333  minT:  BOX68TD" ~{- ",~*T
                                 184
3SIS
         79
            92
                                     139
                               113  13 2
          " -LT-T-T'IT i1 r 1 i rV ^I'V-l-'t"!''!1'i'•[•T-t~flVrT"r-rji'-'[ i'i "I j'i-i1 i "n i'V''i i I-T-T i i i i--i ' i i i i-r-T-r-T-, , , r-pi-
   Ti/Z-->     40   60  80  100  120  140   160  180  200  220  240  260  280
                                                    204
                                                250  268  286
        COMPARISON OF VOLATILE ORGANIC COMPOUND RESULTS BETWEEN METHOD 5030
           AND METHOD 5035 ON A LARGE MULTI-STATE HYDROCARBON INVESTIGATION

                              Rock J. Vitale, Ruth Forman, and Lester Dupes
          Environmental Standards, Inc., 1140 Valley Forge Road, Valley Forge, Pennsylvania 19482

ABSTRACT
With  the  promulgation of SW-846 Update  III  during  June  of  1997,  elimination of  Method  5030 and
implementation of  Method  5035  have  created  significant  challenges  for  the  regulatory and  laboratory
communities (US EPA, 1997). Based on historical data, the results for volatile organics in certain sample types
using the previously approved direct heated purge technologies were observed to be biased low (Hewitt, 1994).
The  loss of volatile organics  was not  observed to be the  determinative process of Method 5030  but of the
sampling,  preservation, and  preparatory aspects of the methodology (Hewitt,  1997,  Siegrist, 1992). The
promulgation of Method 5035  requires training of field samplers and a decision-tree approach to collecting and
analyzing samples.
                                                 43

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


As recent as the first quarter 1998,  for previously initiated on-going projects, various regulatory agencies have
been issuing directives to either implement Method  5035 or continue with Method 5030 (US EPA, 1997). In order
to understand the implications of implementing Method  5035 after several years of using "traditional" soil
sampling methods and analysis by Method 5030 on  a large multi-state hydrocarbon investigation, a 4-week study
was performed to ascertain the differences in field activities, documentation issues, and analytical results.

A description of the comparative study, inclusive of observations  between the traditional Method  5030 and the
three Method 5035 options (e.g., Encore®, sodium bisulfate, and methanol), will be presented. In addition, some
of the lessons learned from the study will be discussed.

INTRODUCTION
A two-phased study was performed to determine the comparability between investigative samples  collected and
analyzed  by  the  "traditional" sampling/analytical method  (SW-846 Method  5030) and investigative samples
collected  and analyzed  by the  recently promulgated sampling/analytical method (SW-846 Method  5035). Soil
samples included in  this study were collected from  four states with different soil  types and contaminant
concentrations. The soil samples were analyzed for a list of 18 volatile compounds by gas chromatography/mass
spectroscopy (SW-846 Method 8260A).

STUDY DESIGN
The study included the collection of soil samples in two phases. The first phase of sampling was conducted at
select locations in Ohio, West Virginia, Pennsylvania, and Maryland from January 27 through February 11,1998.
Soil samples were collected  at 56 sample locations. Nine field duplicates were collected during this phase. A trip
blank (one sodium bisulfate and one deionized water) was collected for each day of sample  collection. Based on
previous  analyses and  remedial activities, samples collected during phase one were expected to either be
"clean"  or require  low-level analysis. Accordingly,  methanol samples  were  not  collected for this  phase.
Specifically, for the first phase of sampling, samples were collected utilizing three techniques as follows:

•  The traditional method of sample collection (in a 125 ml wide-mouthed glass jar);

•  Utilizing  a  plastic syringe and  placing five grams of soil into a  40 ml glass vial pre-preserved (by the
    laboratory) with sodium bisulfate; and

•  Utilizing an Encore® sampler (the Encore® analyses for both phases were performed with the modifications
    recommended by the International Association of Environmental Testing Laboratories [IAETL]).

The second phase of sampling  was performed the  week of March  2, 1998. Samples were collected at 33 select
locations  (including two field duplicates) in West Virginia. The selection criteria for sample locations collected for
the second phase of the project were based upon available historical sample  concentration data. The sample
locations  were  selected to include  samples that contained low, medium, and high concentrations of volatile
compounds (based on the traditional sample collection historical data).

Trip blanks (one sodium bisulfate,  one deionized  water,  and one  methanol)  were collected for each day  of
sample collection. For the second phase of sampling, samples were collected utilizing four techniques as follows:

•  The traditional method of sample collection (in a 125 ml wide-mouthed glass jar;

•  Utilizing  a  plastic syringe and  placing five grams of soil into a 40  ml glass vial pre-preserved (by the
    laboratory) with sodium bisulfate;

•  Utilizing an Encore® sampler (the Encore® analyses for both phases were performed with the modifications
    recommended by the IAETL); and

•  Placing five grams of soil into a vial containing methanol.

All samples  were packed in coolers at  4°C under Chain-of-Custody and shipped via overnight courier to a
reputable commercial environmental laboratory. All samples were analyzed within a holding time of 14 days of
samples collection. In addition, each  sample type (i.e., traditional, Encore® sodium  bisulfate and methanol)


                                                   44

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


collected at a given sample location was  analyzed  within 24 hours of the other  sample types.  In order to
minimize possible confounding effects of analytical holding times,  all  sample types at a  given location were
analyzed within the sample 24-hour time period.

DATA REVIEW
Reduced data package deliverables were  prepared by the laboratory for all samples.  The reduced data package
deliverables included a summary of the reported analytical results for all field samples (including samples, field
duplicate samples and trip blanks), the associated method blank results, the associated laboratory control sample
recoveries, and the surrogate recoveries. The analytical data for both phases were reviewed for completeness of
the data package deliverables, compliance with the  SW-846 Methods  5035 and 8260A,  and usability of the
reported analytical  results  (Clark and Vitale,  1996). Although  important for comparability,  compliance with
methodologies often provides little information on data quality (Blye and Vitale, 1995).

The initial  and continuing  calibration criteria for Method  8260A were met  for all study samples. With  the
exception  of  one compound  in one laboratory  control  sample  (LCS), the LCS  recoveries  were within
study-specified limits (75-125%). The recoveries for one or more of the three volatile  surrogate compounds were
outside  the limits specified (varying limits for each surrogate with limits between 70-121%) for many (38) of the
study samples. Because the "true value" accuracy was not important for this study, the surrogate recoveries are
not expected to affect the operational definition for comparing techniques on respective sample aliquots, which
was most relevant in a comparative study. Analysis of the study trip blanks and laboratory method blanks did not
reveal the  presence of target analytes, with the exception of methylene chloride and acetone. The  positive
sample  results for these two compounds were rejected from further consideration.

Eleven  field  duplicate  samples were collected  for the study; approximately one field duplicate  sample was
collected per day of sample collection. Field duplicates provide  valuable information on precision and sample
representativeness when evaluated properly (Zeiner, 1994). Acceptable precision (<50% RPD, as defined for this
project) was noted between six field duplicate pairs. High RPDs (>50% RPD) were noted for five of the duplicate
pairs. Such sample variability appears to  have complicated a  meaningful comparison of techniques as separate
sample  aliquots were collected and analyzed for each of the study sampling techniques.

SUMMARY OF RESULTS
A  summary of the reported  analytical results is provided on  Table  1. All  results are reported on  a dry-weight
basis. Variations between sample-specific quantitation limits were evident due to the sample collection volume,
the percent moisture of the sample, and the sample-specific dilutions performed. The sample collection volume
at a given sample  location is  different for each sample  type due to the manner in which  the  sample was
collected. The quantitation limits for samples preserved with methanol were raised by the laboratory due to the
medium-level sample preparation.

Fourteen of the sodium bisulfate samples were not analyzed due to the observed concentrations of non-target
compounds in these samples (samples containing sodium bisulfate cannot be analyzed  using medium-level
protocol). The corresponding traditional and Encore® samples at these sample locations  were analyzed at a
medium-level due to the same reason. In addition, six of the sodium bisulfate samples could  not be analyzed due
to observed sodium bisulfate effervescence.

In total, 79  Encore® samples, 79  traditional  samples,  62 sodium  bisulfate preserved samples,  and 23
methanol-preserved samples were collected for the volatile pilot study.  Of the 79 sample  locations examined,
positive results were reported in 33  of the sample locations. Positive results were  reported for four aromatic
compounds (benzene,  toluene,  ethylbenzene,  and total  xylenes);  two  ketones (4-methyl-2-pentanone and
2-butanone);   and four  chlorinated   aliphatic  hydrocarbons  (1,1-dichloroethene,   1,1,2,2-tetrachloroethane,
1,1,1-trichloroethane, and tetrachloroethene).
                                                  45

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
   TABLE 1. SUMMARY OF ANALYTICAL RESULTS
Sample Number
SAMPLE 1
total xylenes
SAMPLE 2
benzene
ethylbenzene
total xylenes
SAMPLE 3
toluene
2-butanone
SAMPLE 4
toluene
SAMPLE 5
toluene
SAMPLE 6
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 7 (DUPLICATE OF SAMPLE 6)
tetrachloroethene
toluene
total xylenes
SAMPLE 8
total xylenes
SAMPLE 9 (DUPLICATE OF SAMPLE 8)
total xylenes
SAMPLE 10
total xylenes
SAMPLE 11
1,1-dichloroethene
1,1,1-trichloroethane
toluene
ethylbenzene
total xylenes
SAMPLE 12
1 ,1 ,2,2-tetrachloroethane
ethylbenzene
total xylenes
SAMPLE 13
toluene
SAMPLE 14 (DUPLICATE OF SAMPLE 13)
1 ,1 ,2,2-tetrachlorethane
SAMPLE 15
4-methyl-2-pentanone
SAMPLE 16
4-methyl-2-pentanone
SAMPLE 17
toluene
ethylbenzene
total xylenes
SAMPLE 18
4-methyl-2-pentanone
ethylbenzene
total xylenes
Encore™

6U

6U
6U
6U

6U
95U

6U

6U

6U
6U
6U
6U

5U
5U
5U

36

610

760

200J
2200
1200
21 OJ
930

310
200J
1200

5U

6U

31,000

13,000

1000
6200
30,000

2700U
2000
20,000
Traditional

6U

6U
6U
15

5U
110

6U

7U

6U
6U
6U
6U

6U
6U
6U

140J
0
440U
0
130U

320U
570
110J
320U
160J

330U
280J
2100

6U

6U

9700

8400

1600
12,000
63,000

6500U
1400
18,000
Sodium Bisulfate

6

6
12
14

8
110U

7

12

23
23
15
260

13
9
7

45
0
N/A
0
N/A

N/A
N/A
N/A
N/A
N/A

N/A
N/A
N/A

6

16

N/A

N/A

15
16
53

1,300
790
240
Methanol

N/A

N/A
N/A
N/A

N/A
N/A

N/A

N/A

N/A
N/A
N/A
N/A

N/A
N/A
N/A

N/A

N/A

N/A

N/A
N/A
N/A
N/A
N/A

N/A
N/A
N/A

N/A

N/A

N/A

N/A

N/A
N/A
N/A

N/A
N/A
N/A
                         46

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                    TABLE 1. (ConU
Sample Number
SAMPLE 19
toluene
ethylbenzene
total xylenes
SAMPLE 20
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 21
toluene
2-butanone
SAMPLE 22
1,1-dichloroethene
toluene
SAMPLE 23
benzene
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 24 (DUPLICATE OF SAMPLE 23)
benzene
tetrachloroethene
toluene
ethylbenzene
total xylenes
SAMPLE 25
benzene
toluene
ethylbenzene
total xylenes
SAMPLE 26
ethylbenzene
total xylenes
SAMPLE 27
benzene
ethylbenzene
SAMPLE 28
benzene
toluene
ethylbenzene
total xylenes
SAMPLE 29 (DUPLICATE OF SAMPLE 28)
benzene
toluene
total xylenes
SAMPLE 30
ethylbenzene
total xylenes
SAMPLE 31
benzene
ethylbenzene
total xylenes
Encore™

63
20
160

6300
5U
5U
11

7U
130U

5U
5U

250J
901
1300
470
4600

250J
80J
1200
570
5600

5U
5U
5U
5U

600
370

6U
6U

2000
7400
1500
14,000
6U
6U
6U

120J
60J

5U
5U
5U
Traditional

34
11
59

5000
6U
6U
6U

6U
120U

6U
6U

220J
300U
1100
550
5500

220J
300U
1500
430
4500

6U
6U
6U
6U

230J
110J

6U
15

6U
6U
6U
6U
7J
30
4600

120U
320U

13J
23J
33
Sodium Bisulfate

61
31
280

2900
14
9
24

7
120

16
6U

N/A
N/A
N/A
N/A
N/A

N/A
N/A
N/A
N/A
N/A

45
90
26
130

N/A
N/A

12
7U

N/A
N/A
N/A
N/A
6U
6U
23

N/A
N/A

6U
6U
7
Methanol

N/A
N/A
N/A

N/A
N/A
N/A
N/A

N/A
N/A

530U
100J

670
200J
3300
870
8400

400J
200J
1900
740
7200

570U
570U
570U
300J

1200
930

61 OU
610U

650U
200J
650U
600J
540U
100J
780
0
340U
80J

260U
60J
670
                            47

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium



                                           TABLE 1.(Cont.)
Sample Number
SAMPLE 32
toluene
ethylbenzene
total xylenes
SAMPLE 33
benzene
toluene
total xylenes
Encore™ Traditional
21 OJ 280U
540 280U
3300 280U
6U 6U
6U 6U
6U 6U
Notes: U - Not detected at or above the associated numerical value
N/A Not analyzed
J - Estimated value; result less than the quantitation limit
E Estimated value; result exceeded the calibration range and
Sodium Bisulfate
N/A
N/A
N/A
8
14
8
not reanalyzed
Methanol
470U
470U
550
61 OU
61 OU
610U

COST ANALYSIS
For the  reputable commercial laboratory that was utilized for this study, the cost of analysis of all four sample
collection types is identical, and there are not any cost implications for analyzing a volatile soil sample by any of
the four sample collection  methods. The cost of sample bottleware  preparation is different between the four
sample  collection types. The cost for traditional sample collection bottleware is built into the contract with the
laboratory. The cost for the Encore® sampler is roughly $10.00 per sampler. Three Encore® samplers are filled at
each sample location at a total of <$30.00 per sample location. The cost for a methanol-preserved vial (two vials
are sampled at a given location when the methanol preserved sample is not collected in tandem with a sodium
bisulfate preserved sample; otherwise, only  one vial is submitted) is <$35.00 per sample or <$60.00 per sample
location. The  cost for a sodium  bisulfate  preserved vial is approximately $50.00  per sample. Two sodium
bisulfate preserved vials are collected per sample location at a cost of approximately $100.00  per sample
location.

The labor costs for the collection  of the various sample types can vary and exact  costs could  not be precisely
calculated based on information available. The Encore® sampler and the traditional  method of sample collection
were found to  be the quickest and easiest sample types to collect. The sodium bisulfate samples were found to
take the longest time to collect and provided the most difficulty in the field.

FIELD OBSERVATIONS
Sodium Bisulfate and Methanol Sampling Techniques
The sodium bisulfate and methanol techniques are very similar and are reviewed  here jointly. These methods
were  by far the most cumbersome  of the selected sampling techniques. The necessity of using an analytical
balance under imperfect field conditions proved frustrating. The balance readings would fluctuate wildly with the
slightest amount of  wind or movement of the  sample  preparation  surface. In  addition,  unless  the sample
preparation surface was perfectly  horizontal, the scale could not calibrate nor zero  out. Furthermore, calibration
of syringes with soil  matrix was time consuming in comparison to the traditional and Encore® methods. Using
razor knives to cut syringe tips evenly required practice  and care. Finally, preservative solution was prone to
spillage unless great  care was taken while placing the soil in the  vials.

The methanol  technique has an additional drawback. When shipping samples via popular air couriers, samples
must  be packaged according to IATA regulations. The regulations regarding shipping of flammable compounds
are very strict  and may result in lost or delayed delivery if the sample containers are not properly packaged. In
addition, the appropriate  paperwork and package labeling must be completed in  a  precise manner,  or the
shipment will  be delayed  or returned to the sender.  From a field  sampling  perspective of the  methods used
during the pilot study, these were by far the most time consuming and frustrating.

Encore® Sampling Technique
The Encore™  sampling system was very straightforward in its approach  and implementation, although  some
minor problems were encountered during the pilot study. Problems were encountered when trying to place loose
and/or wet soils into the Encore® sampler. Soils of this type had to be  manipulated into the Encore® with another
sampling device (such as a spatula). The only other sample collection issue involved improper seating of the cap
on  the  plunger. However,  by pushing down  on  a hard  surface with the T-handle, the cap  could  be  seated
properly. Overall, the Encore® system appeared to be easy to use even under adverse field conditions.

                                                  48

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Traditional Sampling Technique
This sampling method is well-known to most field technicians and has been  employed under a wide variety of
field conditions. It was apparent during the field study that this technique was easily implemented and rivaled the
Encore® sampling  technique for ease of use.  However,  from a field perspective, some clays, silts  and other
tightly compacted soils are difficult to place into a sample container so that no head space is allowed. Breaking
up  soils  to  place into  sample  containers  may  result in loss of volatiles,  thereby  lowering  detectable
concentrations. However, from a field perspective, this technique has been historically easily implemented.

CONCLUSIONS
The following observations can be made from  the overall study. With the exception of the first bullet item, the
conclusions presented below should not be  interpreted to be applied to other sites, soil types, and concentrations
of analytes. These are general conclusions regarding this particular data set; there may not be an equivalent
trend noted in  other data sets. Our findings suggest that inherent  difficulties  associated  with analyzing soil
samples makes definitive states regarding data comparability difficult. Furthermore, the number of positive data
points  and the disparity observed for half of the collected field duplicates makes  statistical trend analysis
problematic at best.

•   At sample locations where methanol-preserved samples were collected and where the concentration of target
    analytes  was  within range  of the  medium-level analysis, the concentration of target  analytes  of  the
    methanol-preserved sample type was  greater than the concentration of the analytes in the other sample
    types  at the same sample location. This is consistent with other studies appearing in peer-reviewed literature.

•   At sample locations where methanol-preserved samples were not collected and where the concentration of
    target analytes was within range of  the medium-level analysis, the concentration  of the aromatic analytes
    appeared to be greater in the traditional sample collection type than the other sample collection types at the
    same  location.

•   At sample locations where methanol-preserved samples were not collected and where the concentration of
    analytes  was  within  range of the medium-level analysis,  the  concentration of the non-aromatic target
    compounds appeared to be greater  in  the Encore sample collection type than the other sample collection
    types  at the same location.

•   At sample locations where methanol-preserved samples were not collected and where the concentration of
    target analytes was not within range  of  the  medium-level  analysis, the  concentration  of the analytes
    appeared to be greater in the sodium bisulfate sample collection  type than the other sample collection types
    at the same location.

REFERENCES
Blye, D.R. and R.J. Vitale. "Data Quality - Assessment of Data Usability Versus Analytical Method Compliance."
    Proceedings from the Eleventh Annual  Waste Testing and Quality Assurance  Symposium. Washington, DC,
    23-28 July 1995.
Clark, M.A. and R.J. Vitale. "How to Assess  Data  Quality  for Better Decisions." Clearwater, New York Water
    Environmental Association (NYWEA), Vol. 26, No. 2 (1996).
Hewitt,  A.D.  "Losses of Trichloroethylene From  Soil  During  Sample  Collection, Storage  and  Laboratory
    Handling." SR  94-8,  US Army Corps of  Engineers, Cold Regions Research & Engineering Laboratory. 1994.
Hewitt, A.D. "Preparing  Soil Samples for Volatile Organic Compound Analysis."  SR 97-11. US Army Corps of
    Engineers, Cold Regions Research & Engineering Laboratory. 1997.
Siegrist, R.L. "Volatile Organic Compounds in  Contaminated Soils: The Nature and Validity of the Measurement
    Process." J. Hazard Mater Vol. 29 (1992):3-15.
US EPA.  "Determination of Volatiles in  Soil - Directive  for  Change."  Memorandum from  Mr. Norman R.
    Niedergang, Director of Waste, Pesticides and Toxics, US  EPA Region 5,  to Corrective Action Project
    Managers and  QA Staff.  1997.
US EPA.  "Test Methods for Evaluating  Solid  Waste." SW-846,  Update  III,   United States Environmental
    Protection Agency. 1997.
Vitale  R.J. "Balancing Regulatory Compliance with Technical  Validity."  Proceedings from Soil  Sampling for
    Volatile Organics Workshop. O'Hare  International Holiday Inn, Chicago, IL, 10  December 1997.
Zeiner,  ST.  "Realistic  Criteria  for the  Evaluation of Field Duplicate Sample  Results." SUPERFUND XV.
    Washington, DC, 29 November-1  December 1994.


                                                  49

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                50

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
 INORGANIC
        51

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                              52

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                               SW-846 INORGANIC METHODS UPDATE

                                            Oliver Fordham
                     U.S. EPA, Office of Solid Waste, 401 M. St., SW, Washington, DC

                                      NO ABSTRACT AVAILABLE
               DIRECT MERCURY ANALYSIS: FIELD AND LABORATORY APPLICATIONS

                       Helen M. Boylan. H.M. "Skip" Kingston, and Robert C. Richter
          Duquesne University, Department of Chemistry & Biochemistry, Pittsburgh, PA 15282-1503

 ABSTRACT
 EPA Method 7473 is designed for the determination of total mercury in solid and aqueous samples. This method
 is based on the instrumental methodology of the  Direct Mercury Analyzer-80 (DMA-80) (Milestone, Inc.) in which
 sample preparation and analysis are essentially integrated into a single analytical  step. The method's  unique
 capability for direct analysis allows for application  in either laboratory or field settings. Method 7473  has been
 validated by analysis of various  Standard Reference Materials (SRMs) in both the  laboratory and in  the field.
 This validation data has been presented1 Results from Method 7473 have also been confirmed by independent
 analysis using traditional methods. Method 7473  has been used on-site in conjunction with mercury remediation.
 Real-time analysis using this technique has provided an  accurate and cost-effective  risk assessment of mercury
 contaminated sites.

 INTRODUCTION
 There  are several analytical techniques  that may  be applied for the determination  of mercury in solid  waste.
 Existing EPA methods for the  analysis of  mercury  include  inductively coupled plasma atomic  emission
 spectroscopy (ICP-AES) (Method  601 OB),  cold vapor atomic  absorption  spectroscopy (CV-AAS)  (Method
 7471A), and  anodic stripping  voltammetry (ASV). Regardless  of  the method used, sample  preparation is
 required. "Soils, sludges, sediments, and other solid wastes require digestion prior to analysis"2. Method  7473 has
 an advantage over traditional mercury analysis because  it eliminates the need for a discrete sample preparation
 step. Direct analysis is performed by integration of thermal decomposition, amalgamation, and atomic absorption
 spectroscopy. While the fundamental theory for this type of  analysis has  been available  in the  literature, the
 DMA-80 is the firs instrumental implementation of these integrated concepts.

 A schematic of the DMA-80 is shown in  Figure 1.  The sample  is automatically  inserted into the  quartz
 deomposition tube, where it is first dried and then thermally decomposed. The gaseous decomposition products
 are carried by a flow of oxygen to the catalytic core, which is maintained at a temperature of 750 °C to  ensure
 complete thermal decomposition. The oxygen flow  continues to carry the gases to the gold  amalgamator, where
 mercury is selectively trapped. Continuous oxygen flow removes any remaining decomposition products. The
 amalgamator is subsequently heated,  releasing the  mercury vapor to the absorbance  cuvettes where peak height
 is measured at 253.7 nm as a function of ng of mercury.

 Calibration for Method 7473 can be performed in two ways. One method is by  the traditional analysis of aqueous
 standards. The  ability for direct analysis also allows for unique calibration using solid standards with a certified
 mercury content.  Method 7473 provides the option to perform calibration using solid  samples, "An alternative
 calibration using standard reference materials may be used..."3  This option is beneficial, especially for on-site
 analysis, when transport and storage of aqueous standards may be problematic.

 Subsequent to validation of Method 7473 in both laboratory and field settings, the  method was applied to the
 laboratory analysis of  a series  of contaminated  soils. Duplicate soil samples  were sent to  a commercial
 laboratory for independent analysis. Only those soils with a mercury  content less than 10 mg/kg were analyzed
directly due to the extreme sensitivity of the instrument. Soils  above  10 mg/kg were  leached using EPA  Method
3051A  prior to analysis. Mercury content in these  soils  ranged from 1-700 mg/kg.  Regardless of the mercury
content, results  using Method 7473 agree with results using the traditional cold vapor  Method 7471A  and show

                                                  53

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
greater precision. Average RSDs for Methods 7473 and 7471A were less than 10% and 15%, respectively.

In collaboration with a local engineering firm, Method 7473 has been used on-site to evaluate the remediation
efforts of a large national utility company. A site near Gettysburg, Pennsylvania, was first evaluated. A sample
was taken from each wall and the floor of the 124 ft3 excavated area. It was determined that remediation efforts
were successful, as the mercury content in all samples was below the action level of 20 mg/kg. These  results
have been confirmed by independent commercial laboratory analysis.

A second on-site evaluation was performed in Hocking County, Ohio. Real-time results were used to direct the
extent of  excavation. The original scope of work based  exclusively on  the site characterization was an
excavation area of 750 ft3  Use of Method 7473 on-site produced  real-time  information as to the level of
remediation required and  allowed remediatiors to reduce the planned excavation area by more than half to 250
ft3. Real-time Method 7473 results indicated that the reduced excavation was adequate  in all areas except one.
Further excavation in that location was thus performed, providing  a more accurate remediation and eliminating
the need for a return trip to the site. Approximately $9,000 in savings resulted from the reduced amount of soil
remediated and elimination of a return trip.

SUMMARY
A method for the  direct  determination  of total mercury in both  laboratory  and field environments has been
established.  Method 7473 has been validated by analysis of SRMs  as well as by independent analysis. Data
indicates that Method 7473 can achieve lab-quality results in a field setting, use of Method 7473 in the field can
lead to more accurate and cost-effective risk assessment of mercury contaminated sites. The next step will be to
investigate compatibility of Method 7473 with speciation technology.
                     CtMctor
                    I ami
   I
             rO
            •Hrix*
            mnot
CVKtaf








1
X—
t~-
^

"^

»
)
                                                                                  Figure  1.  Schematic
                                                                                  of the DMA-80
Table 1. Comparison of techniques use for method validation (n > 5).
EPA Method 7473 7471 A
Technique direct mercury analysis CV-AAS
Average RSD <10% <15<>/0
Agreement? V y

6020
ICP-MS
-15%
                                                  54

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


REFERENCES
1. Walter, P.J., Kingston, H.M.; Boylan,  H.M.; Han, Y. Presented at the Waste Testing and Quality Assurance
   Symposium, Arlington, VA 1997.
2.  US EPA Test Methods for Evaluating Solid Waste, Physical/Chemical  Methods;  SW-846,  2nd ed.; US
   Government Printing Office: Washington, DC, 1997.
3. US EPA Test Methods for Evaluating  Solid Waste, Physical/Chemical Methods; SW-846, Update IV ed.; US
   Government Printing Office: Washington, DC, 1998.
                          MERCURY IN SOIL SCREENING BY IMMUNOASSAY

                                 Mark L. Bruce and Kathleen L. Richards
                       Quanterra Inc., 4101 Shuffel Dr. NW, North Canton, OH 44720
                                        brucem@quanterra.com
                                           Lorraine M. Miller
                    Olin Corporation, 1186 Lower River Road NW, Charleston, TN 37310
                                        lmmiller@corp.olin.com

 Abstract
 EPA SW-846  Method 7471  (Cold Vapor Atomic Absorption,  CVAA) has been traditionally used for mercury
 analysis. This  method requires several hours  of labor and  total processing time to prepare soil samples  and
 analyze them. The sequential nature of the analysis  step limits  sample throughput and capacity.  Also,  the
 method is difficult to perform at a remediation  site. In contrast, immunoassay (IA), as used in proposed Method
 4500, prepares and analyzes the samples in parallel and can provide very large sample capacity in a 2-3 hour
 processing time. Also, the IA kit is much more field portable.

 Quanterra performed an  evaluation  of the  BiMelyze Soil Extraction and  Mercury  Immunoassay kit from
 BioNebraska for the Olin Corporation. Four different site soil samples were analyzed in quadruplicate by CVAA
 and IA following the  New York State Department of Health  (NYSDOH) Alternate Testing Procedure guidelines.
 Threshold standards were setup  at 1 mg/kg and 15 mg/kg. No false negatives were observed. Thefalse positive
 rate was 6%. The overall accuracy rate was 94%.

 Switching to immunoassay improves turn-around-time, sample capacity and field portability.

 Introduction
 Community, environmental and economic concerns are exerting pressure on the environmental market to reduce
 analysis turn-around  time and cost. At the same time,  some data  end users have realized that traditional  test
 methodologies and their QA/QC requirements are  not necessary  for all environmental decisions. There  is a
 growing need to provide reliable, quick turn-around field testing in order to expedite site remediation. Accelerated
 testing can be  instrumental in reducing the impact that remediation activities may have on  the local community.
 It also facilitates faster site closure, thus reducing  the time the excavation site  is  exposed  to the  effects of
 weathering.

 Immunoassay has been widely used in biochemical  testing for health services for decades. Several companies
 have  adapted  this  technology to test environmental pollutants  including  organic compounds  and metals.
 Immunoassay  has  proven  to be a  low cost, fast turn-around,  high capacity analysis for  the health  sciences.
 These same characteristics make it very attractive for the environmental market. Immunoassay (IA) for metals
 does not provide data that is identical to the traditional inductively coupled plasma or atomic absorption tests.
 The specificity of IA should make it less susceptible  to the interferences that limit spectroscopic  analyses.
 However, the biochemical  nature of IA may make it sensitive to new interferences. The QA/QC data normally
 available from IA can include replicates and matrix spikes.

This study compared the results from proposed US EPA  SW-846 Method 4500  (Mercury in  Soil Sample by
Immunoassay)  with  EPA SW-846 Method 7471 (Cold Vapor  Atomic Absorption, CVAA). Method  4500  was
performed using the BioNebraska BiMelyze Soil Extraction and Mercury Immunoassay kit. Actual field samples

                                                  55

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


from a remediation site containing the analyte of interest (mercury) were used for this study.

Immunoassay Method Summary (excerpted from Section 2.2 of Method 4500)
"Solid samples are prepared by extraction with a mixture of hydrochloric and nitric acids for ten minutes and then
buffered  prior to  analysis. The  sample is added to a tube (treated with BSA-glutathione)  and incubated at
ambient temperatures for five minutes. The mercuric ions bound to the sulfhydryl groups of the BSA-glutathione
are now  reacted with a reconstituted antibody specific  for mercury and incubated for five  more minutes. A
peroxidase conjugate is added to the sample, reacting with any mercury specific antibody. The substrate is then
added forming a color that is in proportion to the amount of mercury originally present in the sample. The color
produced is then spectrophotometrically compared with the control standards."

Experimental
This project evaluated the BioNebraska BiMelyze Soil Extraction and Mercury Immunoassay kits (BN-IA-Hg),
proposed Method 4500 by comparison to the traditional mercury analysis method (cold vapor atomic absorption,
SW-846 7471). NYSDOH Alternative Testing Procedure guidelines (4/1/86) were followed.

The NYSDOH Alternative Testing Procedure specifies that limited approval for a "new" test method requires the
following analytical work:

1) Three samples from the soil source.
2) Analyze four aliquots of each sample with the approved method (7471).
3) Analyze four aliquots of each sample with the alternative method (4500).
4) Analyte concentrations should range from the detection limit to 20% greater than the regulatory limit.
5) The result from the alternative method  must fall within  the confidence  interval of the approved method (based
   on two times the published standard deviation).

An additional sample from the soil source which  had  a mercury concentration  previously determined to be well
below the lowest  action threshold was evaluated to test the likelihood  of the IA kit producing false positives.
False positives occur when interferences or analytical error produces a test result that is greater than the control
threshold  when the  analyte concentration  is actually  below the threshold. A false  positive  can  result in
unnecessary remediation work.

Also, it is common to  intentionally bias the results of semi-quantitative field tests slightly high. This reduces the
false negative rate. Many organic  immunoassays use a bias of 30%. A  20% bias had been used previously by
others working with the BioNebraska mercury test kit. A  30% bias on the threshold standards was selected for
this study.

The NYSDOH requirements described above were met as follows:

1) Four samples  from the soil sample source were  selected based  on previous CVAA analysis to cover the
   appropriate concentration ranges. Approximately 150 g of each sample was frozen and then homogenized
   cold to reduce analyte  losses. The samples were stored at 4°C  until appropriately sized aliquots  were
   removed from the sample containers for each test method.  No elemental mercury was visible  and the
   samples appeared visually to be homogeneous before sub-aliquots were removed.
2) Four aliquots of each of the 4 samples were analyzed by Method 7471.
3) Four aliquots of each of the 4 samples were analyzed by Method 4500 following the procedure described in
   the BioNebraska test  kit product literature. The two  control thresholds were  1  and 15 mg/kg since those
   correspond to the two action limits fortheapplicable site.
4) Ideally one sample would have been available for the middle of the following ranges: x < 0.5 ppm, 1 ppm < x
   < 5 ppm, 10 ppm < x < 15 ppm, 15 ppm < x <  20 ppm
   The quadruplicate CVAA results show that 3 of the 4 ranges were covered by the samples selected.
5) Since the alternative method is  semi-quantitative, the range  determined  for each  individual sample was
   compared to the corresponding  average quantitative result from Method 7471

Results and Discussion
Accuracy
The mercury field test kit is semi-quantitative for this site (e.g. x < 1 mg/kg or 1  mg/kg < x <15 mg/kg or 15
mg/kg  <  x). The  concentration  range determined by method 4500 (IA) was compared to  the concentration


                                                  56

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
measured by 7471 (CVAA). IA results which placed the sample in the correct range were considered "accurate".
Results that were too low were labeled "false negative". Results that were too high were labeled "false positive".
Normal US EPA criteria for field test kits specify the false negative rate should not exceed 5%. NYSDOH criteria
were not available.

                           Table 1. Comparison of Average Results, CVAA vs IA
Sample ID
B41 -24-1 02297
B1 1-02-1 021 97
B44-02-102197
B1 -24-1 021 97
CVAA
avg
mg/kg
0.12±0.02
9.4 ±0.88
35.2 ±9.2
16.9±14
IA
avg
abs
0.21
0.43
1.20
0.77
IA interpretation (mg/kg)
x<1 1
-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The average IA results agreed with the average CVAA results for each of the four samples.

As expected  the  individual results  showed  greater variability and slightly less agreement between the two
analysis types  One  immunoassay  result from B41-24-102297 (x <  1  ppm) was a false positive since its
absorbance equaled that of the  1 ppm threshold standard.  Occasional results of this type  are to be expected
when  using low biased threshold standards  to reduce the  false negative  rate. The IA  and CVAA results for
samples B11-02-102197 (1  < x < 15  ppm) and B44-02-102197 (15 ppm < x) showed excellent agreement.

Sample B1-24-102197 had a concentration near the 15 ppm threshold and was apparently more heterogeneous
than the other samples (despite the extensive homogenization that had been performed).  Both the IA and CVAA
results were  more variable and were evenly  split on either side of the 15 ppm threshold. As a general data set
the results from the two methods are in agreement. Please note that since the individual IA and CVAA analyses
were performed on separate aliquots of the same sample, it  is not appropriate to directly compare the individual
test results in order. Rather the quadruplicate  results from each method must be considered as a set.

Given the results discussed above  and  shown in the table  below,  no false negatives (0%) and only one  false
positive (6%) were observed in the 16 sample assays. The agreement rate between the two tests was 94%.

The average  IA kit absorbances  correlated well with the CVAA average results. This indicates it may be possible
to use the IA kit for quantitative analysis, particularly if standards were prepared in or made from soil samples
                                                    on-site. The chart  below shows IA absorbances plotted
                                                    relative to the CVAA  results. The IA standard  soils
                                                    (from BioNebraska) are also shown  relative to  their
                                                    actual  prepared concentrations.  The  least squares
                                                    linear equation for the 4 soil samples was :

                                                    IA  abs = 0.02879 (Hg cone mg/kg) + 0.2089 with an R2
                                                    value of 0.9848.
                                                    Figure 1. Comparison of IA Absorbance vs CVAA
                                                    Concentration
in
5


1 .
0.6 •
0.4 •
0.2
0 •
(
X
X
*
X soil samples
o soil standards

) 10 20 30 40
Hg mg/kg (CVAA)
 Precision
 The standard deviation of the absorbances produced by the IA test were useful when interpreting results near a
 control threshold. It  is not practical to calculate the standard  deviation  of the concentration  since this is a
 semi-quantitative test in its current configuration.

 The percent relative  standard deviation (%RSD) of the result (cone for CVAA and absorbance for IA) for each
 method was comparable. This indicates the two methods have the same precision for this group of samples.
 Most likely sample homogeneity was the factor which limited the precision of both methods.

                         Table 3. Comparison of Precision, CVAA vs IA
Sample ID
B41 -24-1 02297
B1 1-02-1 021 97
B44-02-102197
B1 -24-1 021 97
CVAA
Std Dev
0.01
0.44
4.6
7.1
%RSD
8.6
4.7
13
42
IA
Std Dev
0.02
0.02
0.23
0.18
%RSD
10
4.9
19
24

average %RSD

17


 Information from a BioNebraska representative indicates that the use of volumetric pipettes instead of the "eye
 dropper" volume  measurements described  in the product literature may have  improved precision by  a few
 percent.

 Sample Turn-Around Time
 Complete kit preparation, extraction and analysis of a batch of soil samples took approximately 3 hours for a
                                                  58

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


 batch of 19 assays (16 samples). Kit preparation (including reagent prep and sample container labeling) took
 about 0.5 hours. Extraction and filtration of the soil samples took 1.75 hours. Performing  the immunoassay took
 0.75 hours. Two analysts were working together performing the soil filtrations. If a single analyst was performing
 all the filtrations with the filters supplied in the kit,  add about 0.5 hours to the total time Obviously smaller
 batches of samples would reduce the time required for kit preparation, extraction and filtration but the assay step
 would be only slightly shorter. Also, autopipettes and repeating pipettes were used (at client request) to improve
 accuracy and precision. This equipment (not included in the field kit) also shortened the time necessary for some
 of the liquid handling steps in the immunoassay.

 Sample Capacity
 The field portable version of the test kit  is designed to process 13 soil samples in each batch. Smaller batches
 can be processed but the cost per sample increases because the test kits are used less efficiently and additional
 control soil samples must be purchased. The  laboratory  version of the test can be configured for much larger
 batch sizes.  Using laboratory pipettes and previous experience with  immunoassay, two field kits were combined
 to process all 16 tests (4 samples in quadruplicate) in the same batch. Laboratory kits are available to process up
 to 96 assays (-80 samples) in a batch.

 Analyst Skill
 The field test kit was designed for use by field technicians with average manual dexterity and attention to detail
 but limited experience with scientific instrumentation.  Based on the experience during this  study, this expectation
 is true. Appropriate hands-on training with the IA  kit on known  samples is essential before beginning work  on
 unknown samples. Also,  two  problems occurred during sample  preparation which  might happen for other
 samples on this site.

 1) B44-02-102197 produced a large amount of foam when the acid was applied to the sample at the being of the
 extraction. During the study, the acid was added very slowly and the  bottle continuously tapped  on the bench top
 to break the foam bubbles. Even so, 2  of 4 aliquots lost about 1% of the sample due  to foaming out of the
 extraction bottle. This did  not appear to affect  the  assay results but  required considerable attention, persistence
 and time to avoid  significant sample loss. An experiment conducted after the assays were complete showed that
 adding the recommended amount of acid solution to a 5g sample aliquot resulted in foam overflowing the normal
 32 ml_ sample bottle and a 67 mL bottle. When a 140  ml_ bottle was used the foam filled 80% of the bottle before
 bursting and settling into the  bottom of the bottle. Thus, for samples similar to this one it  would be advisable to
 use a 140 mL bottle to avoid accidental sample loss, slow processing and increasing the skill requirements.

 2) The extract  filtration  step was much slower for the samples collected at the site than for the standard soils
 supplied by  BioNebraska. Sample B41-24-102297 in  particular was  very difficult to filter.  Several  samples
 required squeezing the sample bottle  with pliers  in  order to force  sufficient extract  filter. This would make
 reproducible results difficult using "eye dropper" volume measurements at this point in the process. Other types
 of IA kits have  more "user-friendly" filtration processes.  It may also be possible to adapt the current filtration
 process to improve its performance.

 Conclusion
 Method 4500 (IA)  produced semi quantitative results  which  matched the Method 7471  (CVAA) results for 15 of
 the 16 tests (94%) performed. There was one false positive where the immunoassay overestimated the mercury
 concentration. The false negative rate in  this four sample  evaluation study was 0%, while the false positive rate
 was 6%. This meets the requirements  of the NY DEC Alternative Testing Procedure. In addition, the average  IA
 absorbances from each quadruplicate  set of data correlated very well with the CVAA results. This  indicates that
 the IA kit may also be useful for quantitative analyses  in the future.

                Summary of Reporting Limits and Performance Results
Matrix
soil
Mercury
1 mg/kg
Accuracy
94%
False Negative
0%
False Positive
6%
Acknowledgments
The authors would like to thank Veronica Bortot from Quanterra - Pittsburgh for overall project coordination and
Craig Schweitzer from BioNebraska for IA product and technical support.
                                                   59

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
References
1. BioNebraska - BiMelyze® Soil Extraction and Mercury Assay Product Literature
2. US EPA SW-846 Method 4500 (proposed Update IVA)
3. DOE Method MB 100 Rev 1 (draft)
4. California EPA Evaluation Report Certificate No. 95-01-014
      UTILIZATION OF A FIELD METHOD FOR THE SEMIQUANTITATIVE DETECTION OF SILVER IN
                        ENVIRONMENTAL SAMPLES IN THE 0 - 50 ppb RANGE

                                              Dan Kroll
                               Hach, 100 Dayton Road, Ames, Iowa 50010

Abstract
Silver is commonly used in industry, and its bactericidal properties have also lead to its use for water disinfection
purposes. Excess concentrations of silver may damage human health  and be toxic to aquatic life. Current
methods of silver analysis, in the ppb range, require expensive equipment and careful technique. There is a need
for a  quick, easy  screening  method  for silver at  these levels.  The  procedure described employs the
bromopyrogallol red/1,10-phenanthroline method, described by Dagnall and West (1964), combined with a novel
concentration/detection method. At  a pH  of  7,  a ternary complex  is  formed with two 1,10-phenanthroline
molecules binding to each silver ion, and then two of these complexes bind to a bromopyrogallol red molecule.
This results in a blue precipitate. The colored  precipitate is caught on a 13 urn pore size filter, and the filter is
compared to a precalibrated  (0, 5, 10, 25, and  50 ppb) printed color chart for quantification. All the reagents are
combined in a single powder that contains both dyes,  a buffer, and a masking reagent. The system is easy to
use, fast, portable, and all reagents are stable for at least one year making the system ideal for field testing. This
method has been evaluated on a variety of tap water, pool & spa water, river water, and sewage effluent samples
that have been spiked with known amounts of silver. Some of the river water samples and the sewage effluent
required a sulfuric acid digestion, but all samples resulted in good recovery of the spikes and correlated  well with
numbers generated  using AA techniques. Various soil samples that were spiked with 25 ppb Ag resulted in good
recoveries  that corresponded to the appropriate  color spot on the chart. Use of this method as a screening
process may help to save time and money by cutting down on  the need to do more accurate analysis of all
samples.

Introduction
Silver is a common contaminant of industrial process and  wastewater. In private  industry, silver is used in
applications such as jewelry, coins,  dentalware, silverware,  solder,  electroplating,  photography,  and battery
production. In  low concentrations, silver's  antibiotic properties make it desirable for use as  a fungicide and for
drinking water disinfection purposes, and it has been gaining in popularity as a pool and spa biocide. However,
according to the World Health Organization (WHO), continuous  exposure to silver in drinking water (0.4 mg or
more) in humans causes arygaria,  an irreversible condition which  produces a bluish-gray discoloration  of  the
skin, hair, nails and eyes.4 Long term continuous exposure to silver has also been implicated in liver damage and
enzyme inactivation in humans."

Unpolluted  surface water levels of silver usually range between 0.1  4  ug/L. Drinking water levels range between
0 - 2 ug/L; average = 0.13 ug/L.3 The WHO has not as yet set limits for safe silver concentrations in drinking
water.5 The USEPA has adopted the Public Health Service (PHS) standard that silver in domestic water  not
exceed 50  ug/L.3 The USEPA-adopted PHS standard was set to protect aquatic life and human health. Canada
has adopted a similar 50 ug/L standard while  the EEC standard  is  10 ug/L.5 Silver is also on the list  of seven
priority pollutant metals that must be monitored in landfill leachate.7

There are many current methods to  measure silver in the ppb ranges.  These include atomic adsorption  by flame
or electrothermal techniques, inductively  coupled plasma,  or  colorimetry. Each of these methods requires
complicated and expensive apparatus, hazardous chemicals and/or a large  investment in time and equipment.
Each also  has its  drawbacks.  AA  is  accurate  at moderate concentrations, but displays sensitivity  to  ion
interference. ICP techniques have  higher  minimum detection limits  and are  sensitive to refractory elements

                                                  60

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


 Colorimetry loses sensitivity at these ranges and uses hazardous chemicals.6

 There is a need for a portable, rapid, easy, and safe screening  method for low levels of silver. The method
 described in this article meets all of these requirements and requires less time than conventional  methods. The
 user obtains an on-site reading that only takes minutes to obtain. The procedure features a  simple  visual
 comparison for  obtaining semiquanitative measurements for silver between  0-50 (jg/L. Single-reagent addition
 and quick results make this method ideal for field testing.

 Method and Chemistry
 This test makes use of the fact that under certain specified conditions, metals form brightly colored precipitates
 with certain dyes or mixtures of dyes. In the case of saver,  the complex formation is with  1-10 Phenanthroline
 and Bromopyrogallol Red. This chemistry is an  adaptation of a reaction described by Dagnall and  West.1  In this
 reaction each Ag ion  first reacts  with two 1-10 Phenanthroline molecules to form  a colorless complex. Two
 molecules of this  complex then react with a molecule of the Bromopyrogallol Red to form a blue precipitate.
 Dagnall and West made use of this system for  the spectrophotometric detection of silver with a lower detection
 limit of 20 ppb. The system that I am currently using follows this chemistry until the method of detection. I  make
 use of the fact that the blue complex is an insoluble precipitate in an aqueous sample.

 A reagent powder is prepared that contains a  Sodium Citrate - Citric Acid buffer system with a  pH of 7. The
 powder also contains Tetrasodium EDTA as a  masking reagent to remove interference from other metals. An
 excess of 1-10  Phenanthroline is  added  to make sure that there  is more than enough present to complex the
 silver along with any iron that  may be present. If excess 1-10 Phenanthroline is not added, it may all bind  to
 contaminant iron leaving none available for the silver. Finally, the reagent powder contains the Bromopyrogallol
 Red. The final  reagent powder is then packaged at a weight fill of 2.0 g in unit dose form. This is the amount
 needed to react with a  100 ml sample.

 After the initial chemical reaction is carried out, 100 ml_ of the reacted sample is filtered by being  forced with a
 syringe through a  nitrocellulose microporous filter.  (Schleicher & Schuell) The blue precipitate is trapped on the
 filter. The intensity and hue of the  filter is dependent in a quantitative manner upon the original concentration  of
 silver present in the sample. The colored filter is then  compared to a color matching chart with different shades
 of blue corresponding to different concentrations of silver. By manipulation of the dye concentrations, filter size,
 pore size, and sample volume, the levels to which the test can be made effective  can be altered. Using a 100 mL
 sample, 5 mm dia. filter size (a Gelman #4317 13mm plastic filter  holder is modified with washers to expose a 5
 nun surface area  on the filter), and  12  urn pore size, visual levels of detection down to 5 ppb Ag can  be
 achieved. The final color coding chart for silver has gradations
 of 0, 5, 10, 25, and 50 ppb Ag (See Fig. 1). This method has
 been validated on a number of different water matrices using
 NIST standard  spikes (SRM3151)  and compared to AA (Varian
 SPECTRAA 20-plus) for verification. This method  allows visual
 detection  of  silver in  ranges that are comparable to, and in
 some cases lower than, the levels  of detection that are possible
 with more costly and time consuming methods, such as AA.
                            Figure 1. Color Comparator Chart                   pp
Hach Rapid Silver™ Test Chart
      0 - 50 ppb Ag
Summary of Method
Silver ions at a pH of ~7 combine with two molecules of 1,10-Phenanthroline to form a colorless water soluble
complex. Two molecules of this complex then combine with a molecule of Bromopyrogallol Red to form a water
insoluble blue precipitate. (See Fig. 2) This blue precipitate is then captured on a nitrocellulose filter. The filter is
then compared to a color chart and matched to the appropriate spot for a semi-quantitative reading. The reagent
pillow contains appropriate  buffer, indicator, and EDTA as a masking reagent. A pretreatment of ascorbic acid is
necessary to remove CI2 in excess of 2 mg/L. Digestion is needed for some samples. Other samples such as
very concentrated soil digests that contain  large quantities of interfering metals may require the addition of extra
EDTA and or sodium citrate as chelating reagents.

Sampling and Storage
Collect samples in an acid-cleaned glass or plastic container. Adjust the pH to 2 or less with Nitric Acid (about 2


                                                   61

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
             rx^   .  ,«,*«,, —.AA.AA-
            OT        sllvar'            Jc>
            ^^      no, Ehenanthroline      HlT
         Brcmopyjcogallol                      ^5^
                                      Blu» Precipitate
mL per liter). Store preserved samples at
room temperature for up to six months.
Adjust the  pH to ~7 with Ammonium
Hydroxide before analysis.  Do not use  a
pH meter as  silver  contamination  from
the electrode  may  occur.  Phenol Red
may be used as an indicator.  Correct the
test results for volume additions.

Figure 2. Reaction Mechanism
Interferences
Interference studies were conducted by preparing a known silver solution (approximately 0.010 mg/L) and the
potential interfering ion. Positive interference was tested by running blanks of deionized water that contained the
potential interfering  ion. The ion was said to interfere when the resulting change threw off the color match by
more than 1/2 step on the chart. The following substances, at the stated levels,  show no interference on a 10 ppb
Ag standard test.  There should be  no problem with color  matching or bad  blanks at these levels. These tests
were run at the levels indicated, although greater concentrations may be tolerated.
Table of Interference Study Results

Aluminum Al
Antimony Sb
Bismuth Bi
Borate
Na2B4O7
Barium Ba
Chlorine CI2
Chloride Cl"
Chlorite
HOCI
Calcium Ca+2
Chromium Cr+3
Chromium Cr+6
Cadmium Cd+2
Cobalt Co
Copper Cu+2
Flouride F~
Gold Au
Iron Fe+2
Iron Fe+3
Lead Pb
Magnesium

10 ppm
10 ppm
10 ppm
1 g/Lor215
ppm as B
10 ppm
5 ppm
400 ppm
1 Ppm
1000 ppm
10 ppm
10 ppm
10 ppm
10 ppm
10 ppm
10 ppm
10 ppb
10 ppm
10 ppm
10 ppm
1000 ppm

none
none
none
none
(fades with time)
none
Dingy but
readable > OK
with ascorbic
acid
none > amounts
interfere
none
none
none
none
none
none
none
none
none > negative
interference
none
none (lite pink
tint) >with
addition extra
none
none

Manganese Mn
Mercury Hg
Molybdenum Mo
Nickel Ni
Nitrate NO3 as N
Nitrite NO2 as N
Ammonia
NH3+ as N
Palladium Pd
Phosphate PO4
asP
Potassium K
Platinum Pt
Sulfate SO4
Selenium Se
Silica SiO2
Thallium Tl
Tin Sn
Titanium Ti
Phenanthroline
Zinc Zn

8lSteftfeB*T\"j«^llfeB
10 ppm
10 ppm
10 ppm
10 ppm
250 ppm
100 ppm
1 000 ppm
10 ppm
1000 ppm
1 000 ppm
50 ppb
1 000 ppm
10 ppm
1000 ppm
1 ppm
10 ppm.
10 ppm
1 ppm
10 ppm


none
none
none
none
none > amounts
appear to fade color
none > amounts
appear to fade color
none
none
none
none
none > negative
interference
none
none
none
none
none
none
none
none

Turbid samples should be pre-filtered through a glass fiber filter. Oils and surfactants may interfere by preventing
                                                  62

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


 formation of the insoluble complex; however, appropriate digestion may eradicate some of these interferences.

 Materials that were not tested, that according to the literature do interfere, are Uranium (VI), Thorium (IV), and
 Niobium (V). They all form blue colored complexes with Bromopyrogallol Red. They can be masked at a 10-fold
 excess over silver by the addition of excess fluoride for Uranium and Thorium, and of hydrogen peroxide for
 Niobium.1

 Precision
 In a single laboratory, using a standard solution of 0.010 mg/L silver and three representative lots of reagents,  a
 single operator performing 50 tests per lot obtained no results that were not properly matched to the 10 ppb color
 dot within 1/2 step. In other words all results were within a range of 7.5 to  17.5 ppb. These results were compared
 to results obtained on a AA (Varian SPECTRAA 20-plus). The results for the AA on 30 repetitions using the same
 standard gave an average of 6  ppb with a standard deviation of 4.98 ppb. The recommended lower level of
 detection for the AA was 20 ppb.

 Assuming the worst case scenario where the standard deviation for the visual test method is 7.5 ppb a z test was
 performed to compare the mean  results from the two methods. This resulted in a calculated z value of 2.86. This
 value for the z statistic indicates  that the means are not the same with over 99% confidence. In this case, where
 the amounts of silver to be detected  are below the  recommended level of detection for the AA, the visual method
 out performs the AA.

 Performance on  Environmental Samples
 Samples from a hot tub utilizing bromine as a disinfectant, Ames, Iowa tap water and pool water from Carr pool,
 also in Ames, were spiked at a level of 10 ppb silver. Tests run in the field using the visual method resulted in
 100% recovery on these spiked samples. All samples matched the 10 ppb spot.

 Samples run on sewage effluent  from the Ames, Iowa sewer plant and surface water from the Skunk River near
 Ames resulted in  poor or no recovery. This is probably due to the binding of the silver ions by humic or fulvic
 acids present in the sample,  or  possibly  because of reduction of the  silver ions to silver metal upon addition.
 100% recovery on spiked 25 ppb silver samples was found after treatment of the samples with a simple sulfuric
 acid  - hydrogen  peroxide digestion procedure (Hach  Didesdahl™ Procedure)2.  Strongly  reducing samples,
 samples with high organic content, and samples which contain thiosulfate or cyanide should be digested before
 testing.

 Summary of Results
\ *f ?" §pyFce of Sample
i f' t
Tap Water
Pool Water
Spa Water
Sewage Efluent
River Water
Crowley Silt Loam
Coland Clay Loam
Clarion Loam
NIST SRM 271 1
Aeitow;1.'
10
10
10
25
25
30-32
25
25
25
1 tJbserve^ Cone, "^ IJ
A9.(fl8*)
10
10
10
25
25
35
25
25
25
Tests were run on digested samples of a Crowley Silt Loam soil obtained from the Louisiana State University
Extension Service. 0.5g of the sample was digested  using the Hach Digesdahl™  procedure. The digests were
then filtered through a glass fiber filter and diluted to a final volume of 1 liter. These tests resulted in no recovery.
After the addition of 1.5 g of extra disodium citrate  as  a  complexing  reagent to remove interfering ions, the
samples were found to give a blank of *8 ppb for the visual test and 4.66 with a standard deviation of 5.07 ppb
on the AA. 0.5g samples of the soil were then spiked with 25 ug of silver and the procedure was repeated. All of
30 samples measured by the visual test read  between the 25 and 50  ppb dots for an extrapolated average of «35
ppb. The same samples ran on the AA gave an average result of 36.66 ppb with a standard deviation of 4.34
ppb. Assuming the worst case scenario where the standard deviation for the visual test method is 12.5 ppb a z
test was performed to compare the mean results from the two methods. This  resulted in a small z  statistic of
-0.69. This value for the z statistic indicates that the means for this sample using the visual and AA methods are
                                                  63

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


statistically the same. The initial silver content of the soil explains the differences in the spiked and recovered
amounts. Tests performed on a Coland Clay Loam  and a Clarion Loam obtained in the Ames area gave similar
results, but with no detectable blanks.

Finally a 0.5407 g sample of Montana soil NIST SRM 2711 was digested using the Digesdahl™ procedure. This
amount of test soil results in a final solution, when diluted to 100 mL, with a reported assay value of 25 ppb
silver. All of the samples tested correlated well with the reported values. All samples tested matched the 25 ppb
dot using the visual method.

Stability
Reagents kept at 35°C for 1 year still function properly.

Summary
The method described was utilized to test various soil and water samples for the presence of Ag. All of the
matrices that were examined gave acceptable results. The results were comparable to those obtained with an
AA, and in the case of levels of 10 ppb and below, the visual  method was found to be superior to the  AA. This
method should be useful as a field method for determining trace amounts of silver in a variety of environmental
samples.

The use of this method should save time and money without compromising the accuracy of analysis.

References
1. Dagnall, R.M, and West, T.S. 1964. A selective and sensitive colour reaction for silver. Talanta,  1964, vol 11.
   pp. 1533-1541.
2. Digestion and Analysis of Wastewater, Solids, and Sludges.  1987, Hach Company. Loveland, Colorado.
3. Guidelines for Drinking Water Quality, vol. 1: Recommendations, World Health Organization, 1984.
4. Guidefines for Drinking Water Quality, vol. 2: Health Criteria and Supporting Info, 1984.
5. Hach Water Analysis Handbook, 2nd edition, 1992, Hach Company. Loveland, Colorado.
6. Standard Methods for Examination of Water and Wastewater,  18th edition, 1992.
7. 40 CM 7-1-96 edition. Section 261.24.
       DIAGNOSING ERRORS IN SPECIES ANALYSIS PROCEDURES USING SIDMS METHOD 6800

                             H.M. "Skip" Kingston, Dengwei Huo, Yusheng Lu
         Duquesne University, Department of Chemistry and Biochemistry Pittsburgh, PA 15282-1503
                                            412-396-5564

The determination of chemical  species in environmental samples is difficult when the species are easily altered
during the analysis process. A method has been developed to permit the species to undergo chemical  reactions
during analysis and correct for  these  changes.  Speciated Isotope  Dilution  Mass Spectrometry (SIDMS) is
especially suited for species that equilibrate in solution quickly and also degrade during extraction, oxidize or
reduce during analysis or are difficult to separate quantitatively using conventional  methods1 This method is a
new draft EPA RCRA method (Method 6800) that utilizes isotopically enriched speciated spikes combined with
isotope dilution to accurately determine and correct for species transformations that  occur in sample processing.
The errors in the measurement  are those that  are limited by the ability of the ratio measurement and  the
equilibrium of the species during spiking. This method was specifically developed  to address the problems of
accurately quantifying different  species in complicated matrices that also play a role in the stability of the species
during extraction, separation or manipulation. Additionally, it is a diagnostic tool for identifying both the  error and
bias inherent in specific method steps prior to and during sample analysis such as sampling process, storage,
sample preparation, and  chemical  modifications prior and during measurement. The  basic SIDMS method is
applicable to many species of  elements with  multiple isotopes and  extends to various oxidation states,
organometallics, and molecular forms of species. The method  is used  as a diagnostic tool and reference method
assisting  with error  identification  in other methods permitting  their development as more accurate and precise
species analysis tools.

                                                  64

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Validation data for Cr(VI) and Cr(lll) demonstrate the ability of SIDMS to examine other methods in a diagnostic
manner.  The extraction procedure Method 3060A and analysis Method 7196A are  used to demonstrate the
identification of specific chemical changes that take place in these methods. These changes can be corrected by
use of this very sensitive internal speciated tracer2.

The  objective of EPA Method 6800 is to provide a new reference method that is also  legally defensible as a
reference method for measurements that have  high degrees of uncertainty and error due to  highly reactive
species.

1. Kingston, H.M. Skip, Dengwei  Huo, Yusheng  Lu,  Stuart Chalk,  "Accuracy in species  analysis: Speciated
   Isotope  Dilution  Mass  Spectrometry (SIDMS)  exemplified  by the evaluation  of  chromium  species"
   Spectrachemica Acta B, (accepted) 1998.
2. Yusheng Lu,  Dengwei  Huo and H.M. Skip  Kingston. "Determination of Analytical Biases and  Chemical
   Mechanisms in the Analysis of Cr(VI) Using EPA Protocols" ES&T (submitted) 1998.
    THE USE OF 210Pb DATING AND DETAILED STRATIGRAPHY TO DETERMINE THE SIGNIFICANCE
             AND FATE OF CHROMIUM IN SEDIMENTS NEAR A HAZARDOUS WASTE SITE

                                           Richard Rediske
      The Robert B. Annis Water Resources Institute, Grand Valley State University, Allendale, Ml 49401
                                            616-895-3047
                                           Gary Fahnenstiel
      National Oceanic and Atmospheric Administration, Great Lakes Environmental Research Laboratory,
                     Lake Michigan Field Station, 1431  Beach St., Muskegon, Ml 49441
                                           Claire Schelske
             Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, FL
                                           Marc Tuchman
           USEPA Great Lakes National Program Office, 77 West Jackson Blvd., Chicago, IL 60604

ABSTRACT
An innovative investigation  using  210Pb dating  and  detailed stratigraphy was conducted  to  determine the
significance and fate of chromium in the sediments of White Lake (Michigan). Elevated sediment concentrations
of chromium, arsenic, and  mercury were found in the  vicinity  of the historical effluent discharge point of a
tannery. The  chromium levels found in the sediments were among the highest concentrations reported  in the
Great Lakes basin (20,000 mg/kg).  Since the direct discharge of effluent  from the tannery was discontinued in
1976, vertical depositional patterns may reflect changes in the flux of chromium into the system. Historical levels
of metals may be covered by less contaminated material  or resuspended by physical events.  Information on
sediment stability and deposition rates was critical to the development of remediation options for the site.

Traditional sampling  and analytical  methods  would  not  provide  information  on  sediment stability  and
accumulation patterns. Radiodating  using 210Pb,  a technique  commonly  used in  linmology,  was  employed to
determine the history of sediment deposition.  This technique was augmented with detailed stratigraphy analysis
to provide a current and historical record of chromium deposition in the  sediments. Two piston core samples
were collected in the tannery discharge area and sectioned in 2  cm intervals.  Total chromium was analyzed by
ICP. Radiometric measurements were  made using a low-background gamma counting system with a well-type
intrinsic germanium detector. Total 210Pb activity was obtained from the 46.5 kev photon peak, and 226Ra activity
was obtained  from the 609.2 kev peak  of 214Bi. The 661.7 kev photon  peak was  used to measure 137Cs activity.
The peak in 137Cs  activity was measured to evaluate its usefulness as  an independent time marker for the peak
period of fallout from nuclear weapons testing in 1962-63.

Chromium stratigraphy in the tannery  discharge area indicated that the top 15-20  cut  of sediment was less
contaminated (2,000-4,000  mg/kg)  than sediment located  at >30 cm  (>5,000 mg/kg).  Radionuclide results
suggested that this  surface sediment  layer was well  mixed,  however,  distinct from the deeper more highly


                                                  65

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


contaminated sediments. Presently this surface sediment  layer (15-20  cm) does not physically mix with the
deeper, more contaminated sediment. The surface layer was followed by a region  (30-80 cm) that contained
chromium levels in excess of 20,000 mg/kg. Since the direct discharge of tannery effluent to this area ceased in
1976, evidence of the deposition of sediment with less chromium contamination should have been apparent. The
lack of a decreasing gradient of chromium concentration in the near surface zone sediments suggested that the
processes of mixing and resuspension continue to be active. In addition, chromium transport to the 0-20 cm
sediment zone may also have occurred by other mechanisms  including surface runoff of contaminated soils and
possibly  groundwater  advection.  The lack of a  significant  137Cs horizon  in  the  sediments indicated that
groundwater was discharging  in this region;  however, the  linkage  with  chromium mobility  required  further
investigation.

INTRODUCTION
White  Lake is  a 2,571 acre,  drowned-rivermouth  lake located  on  the eastern shore of Lake  Michigan  in
Muskegon County. The  Lake is part of the White  River Watershed and discharges directly  to Lake Michigan
through a channel located on the western end. White Lake was designated an Area of Concern (AOC) in 1985 by
the International Joint Commission because of historical discharges of  heavy metals and organic chemicals.
Chromium, mercury, arsenic, and animal hides have been discharged into White Lake by Whitehall Leather. The
tannery began operating in Whitehall near the  turn of the century and used wood bark as the original tanning
agent.  In 1940, the tanning agent was changed to chromic sulfate, and a series of six waste treatment lagoons
were constructed near an area of the shoreline called Tannery  Bay. Effluent from these lagoons containing heavy
metals and leather byproducts was discharged directly into the bay.  In addition, dredged materials from the
lagoons and other process wastes were disposed of in landfill  areas adjacent to the shore. The direct discharge
of wastewater effluent  from the tannery  ceased in 1976. Previous investigations  have  indicated  extensive
contamination  of sediments  in White  Lake.  Elevated levels of chromium,  lead, arsenic, and mercury  were
detected in the northeastern section of the lake  in 1982 (WMSRDC 1982) and in 1994 (Bolattino and Fox 1995).
This area was the historical discharge point for tannery effluent from Whitehall  Leather. The chromium levels
found in the sediments of this area were some of the highest reported from any site in  the Great Lakes. Since the
direct discharge of effluent to Tannery Bay was discontinued  in  1976, vertical depositional patterns may reflect
changes  in the flux of chromium into the system. The stability  of the sediments in  this region was also unknown.
Without  more  information on  sediment stability and accumulation rates, it would  difficult  to determine the
residence time of contaminants within any specific region of the sediments. Whether historical levels of metals
are being covered by less  contaminated material or being resuspended by physical events are critical questions
that need to be answered before evaluating remediation options.

An innovative investigation  using 210Pb  dating and detailed stratigraphy was  conducted  to determine the
significance and  fate of  chromium in the sediments  of  White Lake.  Radiodating  using  210Pb provides a
continuous sequence of dates from a single core utilizing the natural decay of 210Pb. This technique has  been
widely used in limnology  and  has been  independently verified by comparisons with  other  techniques  (e.g.,
Robbins et al.  1978; Appleby et al. 1979; and Wolfe et al. 1994). 210Pb is a naturally  occurring radioisotope that
enters lakes through wet and dry deposition following the decay  of atmospheric 222Rn. Once in the take, 210Pb is
rapidly scavenged by particles  and settles to the bottom. The concentration of 210Pb  can then be analyzed at a
series  of  depths  in the cores  from the  surface to the depth where excess 210Pb  is  no longer  measurable,
approximately  5-8 half-lives or 150 years. From this 210Pb  profile, dates and sediment accumulation rates are
calculated using one of several mathematical  models,  such  as the  constant rate of supply  method. Using a
combination of 210Pb dating and detailed metal stratigraphy, critical information related to contaminant  profiles
and sediment stability was obtained. The technique described can be used at hazardous waste sites where the
evaluation of contaminated sediments is required for remediation.

EXPERIMENTAL
The sediment cores were  collected with a VibraCore device with core lengths ranging from  6  8 ft. The core
samples were then sectioned in three equal lengths for chemical analysis. Ponar samples were also collected at
these locations to provide an assessment of the near surface zone sediments. A piston corer (Fisher et al. 1992)
was used to obtain the samples for stratigraphy and  radiodating since the VibraCore causes some degree of
internal mixing  in the core tube. VibraCore samples were collected during  1994. The ponar and piston core
samples were collected  during 1996. Sampling  locations are shown in Figures I and 2. All samples were collected
using the USEPA R.V. Mudpuppy.

Piston  core samples were  extruded and cut into 2 cm intervals. Each interval was weighed and an aliquot was


                                                  66

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
removed for metals analysis. Sample preparation and analysis methods (EPA 1994) for metals are listed below:
       Arsenic
       Chromium
       Mercury
Graphite Furnace
Inductively Coupled Plasma
Cold vapor
                                                                  7060/3050
                                                                  6010/3050
                                                                  7471
All metals results were reported on a dry weight basis.
The preparation for radiometric analysis consisted of freeze drying the sediment and the grinding the sample to a
homogenous mixture. Sub samples were then packed and sealed with an epoxy resin in polypropylene tubes in
preparation for radiometric analysis.

Radiometric measurements were made  using low-background gamma counting systems with well-type intrinsic
germanium detectors (Schelske et al. 1994). To prepare samples for radiometric analysis, dry sediment from
each section  was packed to a nominal height of 30 min in a tared polypropylene tube (84 mm high x 14.5 mm
outside  diameter,  12 mm. inside diameter). Sample height was recorded and tubes were  weighed to  obtain
sample  mass. Samples in the tubes were sealed with a layer  of epoxy resin and  polyamine hardener, capped,
and stored before  counting to ensure equilibrium between 226Ra and 214Bi. Activities for each radionuclide were
calculated using empirically derived  factors of variation  in counting efficiency with sample mass and  height
(Schelske et  al. 1994). Total 210Pb  activity was obtained from the 46.5 kev photon peak, and 226Ra activity was
obtained from the 609.2 kev  peak of 214Bi  . 226Ra activity was assumed to represent supported 210Pb activity.
Excess  210Pb activity was determined from  the difference between total and supported 210Pb activity and then
corrected for decay from the coring date. The 661.7 kev  photon peak was used to measure 137Cs activity. The
peak in 137Cs activity was measured to evaluate its usefulness as an independent time marker for the peak period
of fallout from nuclear weapons testing in 1962-63.

Sediments were aged using  measurements of the activity of naturally  occurring radioisotopes in  sediment
samples. The method  is based  on determining the activity  of total 210Pb (22.3 yr half-life), a decay product of
226Ra (half-life 1,622 yr) in the 238U decay series. Total 210Pb  represents the sum of excess  210Pb and supported 2
10Pb activity in sediments. The ultimate source of excess 210Pb is the outgassing of chemically inert 222Rn (3.83
d half-life) from continents as 226Ra incorporated in soils and rocks decays. In  the atmosphere, 222Rn decays to
210Pb, which  is  deposited at the earth's surface with atmospheric washout as unsupported or excess 210Pb.
Supported 210Pb in lake sediments is produced by the decay of 226Ra that is deposited as one fraction of erosional
inputs. In the sediments, gaseous 222Rn produced from  226Ra is  trapped  and decays to 210Pb. By definition,
supported 210Pb is in secular equilibrium with  sedimentary  226Ra and is equal to total 210Pb activity at depths
where excess 210Pb activity is not measurable due to decay. Because the decay of excess 210Pb activity in
sediments provides  the basis for  estimating sediment ages,  it is necessary  to  make estimates of total and
supported   210Pb,   activities   so
excess 210Pb activity can be deter-
mined by  difference. Excess 210Pb
activity  was  calculated  either  by
subtracting 226Ra activity from total
210Pb activity at  each depth or by
subtracting  an  estimate  of  sup-
ported   210Pb  activity  based   on
measurements   of   total  210Pb
activity  at depths  where  excess
210Pb activity is negligible.

Cr ISM m(/kg
Aft 9.0 mg/kg
He O.TIm{/kg
                                                  Cr SIS n|/kf
                                                  AJ 8,6
                                                  Hi 0.39 mg/k>
                                              Cr  1650 mi/kg
                                              Al   1 Dug/lie
                                              Hg  1.04mj/ig
Figure    1.   Conentration    of
chromium, arsenic, and mercury in
ponar   samples   collected  from
Tannery Bay,  Whitehall, Michigan
(1996)

Sediment ages were calculated  using a CRS model (Appleby and Oldfield 1983). This model calculates ages
based on the assumption that the flux of excess 210Pb, to the lake was constant and therefore that variation in
210Pb activity from a  pattern of exponential decrease with depth depends on variation in rate of sedimentation.
                                                   67

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Errors in age and mass sedimentation  rate were propagated using  first-order  approximations and calculated
according to Binford (1990).
RESULTS AND DISCUSSION
The results of ponar and VibraCore samples are shown in Figures 1 and
               Depth    Cr
               0--24"  <360mg/kc
                     1610 mg/kg
               W-56"   88 me/kg
                             WL9417. WL9411
             Depth     Cr
             0"-15"  6700 mg/kg
             150-35"  2420 mg/kg
            35"-4T    30 rag/kg
                                                Cr
                                              8100 mg/kg
                                              7SIOmg/kg
                                               390 mg/kg
Depth Cr
fl"-l«" 9520 mg/kg
W-35" 2470 mg/kg
35"-49" 179 mg/kg


Depth Cr
>. (T-15- 5540 m(/kg
JS"-23" 14300 mg/kg
23"-40" 5620mg/ki
2 respectively. Ponar samples provide
an indication of the conditions present
in the  near surface  zone  of  the
sediments.  The  penetration  of  the
ponar is variable and can range from
0-15 cm.  depending on the condition
of the sediment. A comparison of the
results from  both  collection  methods
suggests  that the  highest degree of
the   chromium   contamination   is
present at depths below  the pene-
tration  of  the ponar. In  addition,  the
ponar results suggest that chromium
continues  to enter the sediments of
Tannery  Bay  since  concentrations
range  from  1,000  mg/kg  to 4,600
mg/kg near the surface.
Figure 2. Concentration of chromium
in   core   samples   collected  from
Tannery   Bay,  Whitehall,  Michigan
(1994)
The results of the stratigraphy analyses for total chromium are given in  Figures 3  and 4 for I-5M and I-7M
respectively. The I-5M core shows a relatively uniform region of chromium concentrations ranging from 2,500
mg/kg to 3,600 mg/kg between 0 and 26 cm. This region is followed by more concentrated strata that vary from
approximately 5,000 mg/kg to 23,000 mg/kg in the interval  from 26-84 cm. Chromium  in the remainder of the
core decreases after 84 cm. Since this station was located in  the discharge area of the waste treatment lagoons,


                                          Chromium (mg/kg)
                                 5000
                                               10000
                                                             15000
                                                                           20000
                                                                                          25000
              Figure 3. Results of chromium stratigraphy analysis on the core sample from I-5M

the variations in chromium concentrations observed reflect differences in effluent composition over time. Sudden
                                                   68

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
reductions in chromium levels also correspond to strata that contain animal hair and hide fragments. The I-7M
core follows a different depositional pattern. Concentrations of chromium gradually increase from approximately
2,000 mg/kg to 5,000 mg/kg over the  interval from  0-36  cm. Concentrations then rapidly rise and remain
elevated in the region from 38-128 cm. Chromium concentrations began to decrease after 128 cm. Higher levels
of chromium were found in the I-7M core than at I-5M. The highest level found at I-5M was 22,800 mg/kg while
several 2 cm strata at I-7M ranged from 34,000 mg/kg to 61,100 mg/kg.

                                           Chromium (mg/kg)
                          5000     10000    11000    JOOOO    35000    30000    31090    40000    46600
              Figure 4. Results of chromium stratigraphy analysis on the core sample from I-7M

The  radiochemistry data is summarized in Tables  1 and 2. The profiles of 210Pb activity for Station I-5M and
Station I-7M (Figure 5) provide other information about historical sedimentation at the Tannery Bay sites.  First,
210Pb activity generally decreases with depth. Second, several stratigraphic layers can be identified  based on
210Pb activity. Four layers are present in core I-5M: 0-15 cm, 15-30 cm, 30-50 cm, and 50-65 cm; and four layers
can also  be identified in I-7M:  0-20  cm, 20-35 cm, 35-45 cm, and 45-70 cm. The total 210Pb activity  in the top
layer in both cores was similar, ranging from  10-12 dpm/g, and the activity in the lowest layer was also similar in
both cores. Supported  210Pb is by far the largest fraction in the lowest  layer. Finally, because excess 210Pb is
generally not measurable in sediments with ages older than five  or six half lives, we can conclude that the ages
in sediments above the bottom layer with measurable levels of excess 210Pb activity are probably not older than
110 to 130 years.

Table 1. Results of radiochemistry analysis of the core sample from I-5M


Depth
(cm)
5
10
15
20
25
30
35
40
45
50
55
60
Total
Pb-210
Activity
(dpm/g)
11.951
11.552
10.70
8.55
8.81
7.455
4.931
2.641
2.696
2.889
1.243
0.44

Ra-226
Activity
(dpm/g)
1.735
1.592
1.73
1.276
1.015
1.423
2.967
1.753
1.398
1.224
0.703
0.757

Cs-137
Activity
(dpm/g)
1.565
1.437
1.614
1.449
1.598
2.414
3.769
3.142
2.468
1.139
0.437
0.024
Excess
Pb-210
Activity
(dpm/g)
10.358
10.103
9.1
7.38
7.911
6.123
1.996
0.904
1.32
1.689
0.549
-0.322


Age
(years)
3.965
10.534
17.992
29.222
46.450
68.841
81.028
90.010
104.612
149.286



Age
Error
(1s)
1.222
1.339
1.490
1.848
2.604
4.523
5.603
5.747
6.991
21.588




Date

1992.7
1986.2
1978.7
1967.5
1950.2
1927.9
1915.7
1906.7
1892.1
1847.4


Mass
Sedimentation
Rate
(mg/cm2/yr)
167.23
145.68
130.06
120.21
72.50
50.96
89.93
142.5
67.94
22.63



MSR
Error
ds)
8.36
7.68
8.22
7.18
5.86
5.95
22.61
61.92
22.25
9.26


                                                  69

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Combined plots of chromium stratigraphy and radiochemistry data are shown in Figure 6 for I-5M and I-7M. The
four regions in each core described earlier suggest distinct layers. Sediments that are well mixed would have
relatively uniform 210Pb activity as illustrated from the surface to the 10-20 cm zone. These results are significant
as the 210Pb  profile demonstrates a  mixed zone near the surface that is  isolated from  the  sediments below
approximately 20 cm. Levels of chromium in  excess of 20,000 mg/kg begin at 40  cm at I-7M and at 30 cm at
I-5M. Based on the 210Pb profile, a region of unmixed sediment lies between  the heavily contaminated strata and
the mixed sediment zone. The zones of greatest chromium contamination therefore appear to be isolated from
the surface sediments  that are  subject to  mixing. Sedimentation  rate data  for both  Table  1.  Results  of
radiochemistry analysis of the core sample  from  I-5M  stations suggest that  I-7M has  a greater rate (225
mg/cm2/yr) than I-5M (167 mg/cm2/yr). This observation  is supported by the chromium profile discussed above.
Total 2. Results of radiochemistry analysis of the core sample from I-7M.


Depth
(cm)
5
10
15
20
25
30
35
40
45
50
55
Total
Pb-210
Activity
(dpm/g)
12.510
12.700
1 1 .090
12.150
8.770
8.320
7.980
7.000
5.780
5.440
2.300

Ra-226
Activity
(dpm/g)
1.550
1.990
2.030
1.730
1.740
1.840
1.800
2.350
3.380
3.640
3.360

Cs-137
Activity
(dpm/g)
1.280
1.440
1.350
1.530
1.420
1.660
1.640
1.950
2.820
1.220
0.850
Excess
Pb-210
Activity
(dpm/g)
11.110
10.860
9.190
10.570
7.130
6.580
6.270
4.710
2.440
-1 .220
-1.080


Age
(years)
2.410
5.460
1 1 .620
21.900
31.670
44.500
65.110
101.820




Age
Error
(1s)
1.710
1.790
1.940
2.380
2.870
3.740
6.130
16.450





Date

1994.3
1991.2
1985.1
1974.8
1965.0
1952.2
1931.6
1894.9



Mass
Sedimentation
Rate
(mg/cm2/yr)
225.77
212.07
217.52
146.84
159.07
121.58
76.58
43.32




MSR
Error
ds)
14.85
16.31
19.10
9.77
16.91
15.67
12.89
13.97



    -2.5
      Activity (dpm/g)

  •2 0 2 4 6 8 10 12
                          u
                          **
                          v
                          a
                            fl-
                 I-5M
                           tCB-
132-

H4-I
                                        Excess "Pb
                                           I-7M
The peak input of fallout 137Cs in the late 1950s and early
1960s has been used to provide a time-dependent horizon
in cores. This approach was used to verify CRS dates in
Lake Erie cores  (Schelske  and Hodell 1995). Neither a
sharp peak nor a  large peak in 137Cs activity was found in
the Tannery Bay  cores. Therefore, this measurement was
not  useful in  establishing  the 137Cs  horizon. The  low
inventory of 137Cs activity in  both cores is in sharp contrast
to  the  high  inventory  of 210Pb  activity.  These  results
indicate that 137CS was deposited and not retained at these
sites for the following reasons:

• sediment   resuspension  focused  the  137Cs to  other
  locations
• the  137Cs  was diluted  by  the  introduction of large
  quantities of tannery wastes
• ionic 137Cs was  advected with pore waters from the core
  site

Figure 5. Activity versus depth of excess 210Pb and 137Cs
at stations I-5M and I-7M
The latter mechanism is prevalent at locations where groundwater is moving through deposited  sediments. It
seems unlikely that resuspension or dilution was a primary mechanism because of the large inventories of 210Pb
activity  at  both  sites.  The  most plausible  explanation  for the absence of  the  137Cs horizon  is  therefore,
groundwater advection.
                                                   70

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                                               Cr (rag/kg)

                                                           5000 10000 15000 20000 25000
                           Cr(mg/kB)

                       1200024000360004800060000
The influence of groundwater advection on chromium may also be a factor in its fate and transport. As discussed
previously, the absence of the 137Cs horizon suggested  that the movement of local groundwater through the
sediments was responsible for advective losses. Since the local groundwater is known to discharge in the near
shore area of Tannery Bay,  chromium may also be mobilized from the deeper layers and transported to the
surface. While the  solubility  of trivalent chromium is
generally limited due to the  precipitation  of insoluble
hydroxides, the formation of organic complexes  has
been  shown  to increase its solubility. Kaczynski  and
Kleber (1994), James and Bartlett (1983),  and  Hassan
and  Garrison (1996)  noticed  that  the solubility  of
trivalent chromium was increased  in  the  presence of
organic complexing agents. The latter authors noticed
an increase   in  solubility in  the  presence of cysteine
under  low  Eh  conditions.  The  low  Eh environment
present in the sediments of Tannery Bay, in addition to
the  organosulfur  compounds  produced  during  the
decomposition of animal hides and  hair, may produce
conditions that promote chromium complexation. It was
also  noted that a large amount of humic material was
released  from  the  Tannery  Bay  sediments  during
alkaline digestion. These materials may also serve as
complexing agents to increase chromium complexation.
The presence of a  complexed chromium fraction in the
sediment  pore  water  and  its potential  role in  the
advection of  chromium needs to be evaluated  as long
as groundwater continues to enter Tannery Bay.
 Figure 6. Chromium  concentrations and excess 210Pb
 versus depth at stations I-5M and I-7M
 2.5 5.0 7,5  10,0 12.5

Item "Pbfdpm/g)

    -c— Chromium

        Excess "Pb
                                                                                             10.0
                                                                                     ExcMcMPb(dpm/g)
                                                                                             Chromium

                                                                                             Excess ™Pb
 Since the direct discharge  of tannery effluent to Tannery Bay ceased in 1976,  evidence  of the deposition of
 sediments with  less chromium contamination should be evident. The lack of a decreasing gradient of chromium
 concentrations in the surface zone sediments (0-20 cm) may be explained  by several mechanisms:

    • continued surface runoff
    • groundwater advection
    • continual sediment mixing and resuspension in the 0-20 cm zone

 Since the levels of excess 210Pb in the surface zone sediments are normal and do not reflect excessive dilution
 with terrestrial soil,  surface runoff would only  be significant if small amounts of highly contaminated material
 were  continuously  eroding  into  Tannery  Bay.  As  discussed previously,  groundwater  advection may  be
 responsible for some migration of chromium from deeper sediment layers to the surface. It is, however, doubtful
 that this mechanism would be responsible for chromium levels in excess of 2,000 mg/kg. The most likely process
 that would produce the observed chromium levels is that of sediment  mixing and resuspension. The flocculent,
 fine-grained sediments in Tannery Bay may be mixed to a degree that prohibits  the formation of concentration
 gradients. The  continued mixing  of flocculent materials would result in  unstable,  resuspended sediment that
 could readily be exported into  White Lake by  currents and wave action. The  prevalence of the high levels of
 chromium in the White Lake samples collected down gradient from Tannery Bay would support the continued
 export of resuspended sediments (Rediske et al. 1998)

 The CRS model  (Appleby and  Oldfield 1983) was selected to calculate ages because of the absence of an
 exponential 210Pb  gradient (Table  1). Ages calculated  from the model placed 1892 at 45 cm and 1847 at 50 cm
 for core I-5M, and 1895 at 40 cm for core I-7M. These ages, however,  were much younger than expected based
 on other information available from the core. For example, the chromium  concentration exceeded 10,000 mg/kg
from 62-82 cm  in I-5M or in sediments older than 1847 according to the calculated 210Pb ages. The chromium
concentration exceeded 10,000  mg/kg at depths down to 84 cm in I-7M. By contrast only the upper 40 cm of this
core contained sufficient levels of excess 210Pb for dating. This lack of conformity  shows that the calculated ages
                                                  71

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


are not credible. Results from the CRS 210Pb age model that are not credible can be a product of the point
transformations that are used in the CRS model (Robbins and Herche 1993). Independent assessment of dating
is therefore required for the CRS model. For the piston cores, data for chromium and tannery waste by-products
provide independent time markers that are at variance  with the calculated 210Pb ages.  Since high levels of
chromium and tannery waste byproducts (hair and dye coloration) persist well below the calculated 1847 date,
the dating chronology must be rejected. Large inputs of waste materials could confound the chronological record
by diluting the natural sediments and by altering the physical-chemical environment.

Given  the insolubility of trivalent chromium  in  natural water  (Palmer and  Puls 1994), the predominant
mechanism driving the flux of this metal in White Lake appears to be sediment export. The hydrodynamics of
White  Lake support the  progressive transport of sediments in a westerly direction following the natural water
currents. Prevailing winds function to mix the near shore sediments and move the resuspended material out into
the main lake. The  0-20  cm zone of sediment mixing determined by  the 210Pb data reflects the action of the
prevailing winds and the wave induced resuspension. The  well mixed nature of the top 20 cm  zone also suggests
that these sediments are unstable and easily exported. In addition, the differences in stratigraphy between I-5M
and I-7M are consistent with wind induced wave action. Station I-5M has a greater exposure to the westerly wind
and has a lower calculated sedimentation rate (167.23 mg/cm2/yr) and a shallower interval of sediment above the
highly  contaminated zone. In contrast, station I-7M is more protected from wave action and exhibits a greater
calculated sedimentation  rate (225.77 mg/cm2/yr) and a  stratigraphy profile reflecting a greater depth of less
contaminated material.

The  discharge of tannery waste was located near the  shore of Tannery Bay  in the northeast corner.  The
EPA/MDEQ core samples indicate heavy sediment contamination with chromium in the near shore and  middle
areas of Tannery Bay. Stations near  the confluence with White Lake  have considerably less chromium in  the
sediments. This pattern  reflects  a discharge of  insoluble  chromium that was  rapidly incorporated into  the
sediments. Based on this information, the  historical and  current mechanism for chromium  transport in White
Lake is sediment export  from Tannery  Bay by the prevailing circulation pattern and wave action. Chromium
export  from Tannery Bay into White Lake proper will continue as long as the contaminated  sediments  are
influenced by hydrodynamic circulation patterns.

CONCLUSIONS
By  using a combination of combination of traditional  chemical analyses, radiometric  determinations,  and
stratigraphy, important information concerning the nature  and fate sediment contamination in the Tannery Bay
area of eastern White Lake was obtained. Chromium stratigraphy in indicated that the top 15-20 cm of sediment
were less contaminated  (2,000-4,000 mg/kg) than sediment located at >30 cm (>5,000 mg/kg). Radionuclide
results suggested that this surface sediment layer was well mixed, however, distinct from the deeper more highly
contaminated sediments. Presently this sediment layer (15-20 cm) does not physically mix with the deeper, more
contaminated sediment. The surface layer was followed by a region (30-80 cm) that contains  chromium levels in
excess of 20,000 mg/kg. Since the direct discharge of tannery effluent to this area ceased in 1976, evidence of
the deposition of sediment with less chromium  contamination  should have  been apparent. The  lack of a
decreasing gradient  of chromium concentration in the near surface zone sediments suggests that the processes
of mixing and resuspension continue to be active in Tannery Bay. In addition, the high inventories of 210Pb in the
0-20 cm zone show that surface  runoff from the waste piles has not  contributed significantly to the recent
sediment  record. The lack  of a  significant 137Cs horizon in  the sediments  indicates  that  groundwater is
discharging in this region; however, the linkage with chromium mobility requires further investigation. While
traditional chemical  analyses provide  important information  for determining the spatial extent of contamination,
additional techniques are  required to describe chemical fate  and transport. Stratigraphy and radiometric analysis
using 210Pb can provide critical information related to sediment stability, depositional patterns, and chemical flux
that is essential for the analysis of remediation alternatives.

REFERENCES
Appleby, P.G.,  G.F  Oldfield, R.  Thompson, P  Huttunen,  and K. Tolonen. 1979.  Pb-210  dating of annually
   laminated lake sediments from Finland. Nature 280:53-55.
Appleby,  P.G.  and  F   Oldfield.  1983.  The  assessment  of  210Pb data from  sites  with varying  sediment
   accumulation rates. Hydrobiologia 103: 29-35.
Binford, M.W. 1990. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake sediment cores. J.
   Paleolim. 3:253-267.
Bolattino, C. and R. Fox. 1995. White  Lake Area of Concern: 1994 sediment assessment. EPA Technical Report.


                                                  72

-------
                        WTO A '98 - 14th Annual Waste Testing & Quality Assurance Symposium


   Great Lakes National Program Office, Chicago.
EPA, 1994. Test Methods for Evaluating Solid Waste Physical/Chemical Methods. US. Environmental Protection
   Agency. SW-846, 3rd Edition.
Fisher, M.M., M. Brenner, and K.R. Reddy, 1992. A simple, inexpensive piston corer for collecting undisturbed
   sediment/water interface profiles. Journal of Paleominology 7:157-161.
Hassan, S.M. and A.W. Garrison. 1996. Distribution  of chromium species between soil and porewater. Chemical.
   Speciation and Bioavailability. 8:(3/4)85-103.
James,  B.R. and R.J. Bartlett.  1983.  Behavior of chromium in soils. V. Fate of organically complexed Cr(lll)
   added to soils. J. Environ. Qual. 12:169-172.
Kaczynski,  S.  E, and R.J. Kiebler. 1994. Hydrophobic Ci8 bound organic complexes of chromium and their
   impact on the geochemistry of chromium in natural waters. Environ. Sci. Technol. 28:799-804.
Palmer, C.D. and R.W. Puts. 1994. Natural Attenuation of Hexavalent Chromium in Groundwater and Soils. U.S.
   Environmental Protection Agency. EPA/540/5-94/505.
Robbins, J.A., D.N. Edgington, and A.L.W. Kemp. 1978. Comparative 210Pb, 137Cs, and pollen geochronologies of
   sediments from Lakes Ontario and Erie. Quat. Res. 10:256-278.
Robbins, J.A., and L.R. Herche. 1993. Models and uncertainty in 210Pb dating of sediments. Int. Ver.  Theor.
   Angew. Limnol. Verh 25:217-222.
Rediske, R.R., G. Fahnensteil,  C.  Schelske,  P  Meier. T.  Nalepa,  and  M.  Tuchman, 1998. Preliminary
   Investigation of the Extent and Effects of Sediment Contamination in White Lake near the Whitehall Leather
   Tannery. Final Report to U. S. EPA. Great Lakes National Program Office. Chicago, II.
Schelske, C.L., A. Peplow, M. Brenner, and C.N. Spencer. 1994. Low-background gamma counting: Applications
   for210Pb, dating of sediments. J. Paleolim. 10:115-128.
Schelske, C.L. and D. Hodell. 1995. Using carbon isotopes of bulk  sedimentary organic matter to reconstruct the
   history of nutrient loading and eutrophication in Lake Eric. Limnol. Oceanogr. 40:918-929.
Wolfe, B., H.J. Kling,  G.J. Brunskill, and P Wilkinson. 1994. Multiple dating of a freeze core from Lake 227, and
   experimental fertilized lake with varied sediments. Can. J. Fish.  Aquat. Sci. 51:2274-2285.

ACKNOWLEDGEMENTS
This work was supported by an Interagency Agreement (IAG) #DW13947766-01 between the  Environmental
Protection Agency Great Lakes National Program Office (GLNPO) and the National Oceanic and Atmospheric
Administration (NOAA)  Great Lakes Environmental  Research Laboratory (GLERL). Other funding was provided
by NOAA/GLERL and  Grand Valley State University  (GVSU) Water Resources Institute (WRI). The  authors
would also like to acknowledge the assistance of the following individuals:

NOAA/GLERL                 C. Beckman, S. Beattie, R. Stone, and G. Carter
GVSU                        A. Stiop, F Winkler, J. Oxford, D. Graeber, T. Hudson, and K. Onwuzulike
University of Florida            Dr. Jaye Cable
R/V Mudpuppy (USEPA)        J. Bonam
                AIR-FORCE WIDE BACKGROUND CONCENTRATIONS OF INORGANICS
                             OCCURRING IN GROUND WATER AND SOIL

                                           Philip M. Hunter
               Air Force Center for Environmental Excellence, Consultant Operations Division,
                              3207 North Road, Brooks AFB, Texas 78235

ABSTRACT
Background concentrations of naturally occurring inorganics are important to site characterization, establishing
cleanup levels, conducting risk assessments, designing and operating long-term monitoring programs, and the
like. The Air Force Center for Environmental Excellence (AFCEE) has streamlined the  process for determining
background  concentrations  by using computer algorithms  that interrogate  the Air  Force's  Environmental
Resources Program Information Management System (ERPIMS). Analysis of this database reveals that there are
a wealth of existing sampling locations that are known to be uncontaminated and available for use  in these
important calculations. Air-Force wide background concentrations for ground water and soil have been calculated

                                                 73

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


for inorganics  using all available sampling data from the ERPIMS  Database.  Insight can be gained by using
these concentrations as a representative baseline of numbers for ongoing and future investigations concerned
with monitoring and remediation of inorganic contamination.

INTRODUCTION
Analysis of the ERPIMS database reveals that most contamination across the Air Force is organic in nature and
is typically  associated with chlorinated solvents and fuels (i.e. BTEX and  related compounds). The presence of
key organics detected  in groundwater and soil samples is  a good indicator of both inorganic and organic
contamination.  Often Air  Force  resources  are expended  needlessly  to  perform a  separate background
investigation  at  individual  installations  using  newly-captured  data  rather  than  relying  on  existing data.
AFCEE-developed computer algorithms were used to automate the process of identifying background locations
using ERPIMS data. Over 10 years worth  of project data is available as an existing resource for background
determinations at  installations across the Air Force. This methodology reduces and some cases eliminates the
need  to perform  a  separate  background investigation  which can cost  hundreds of thousands of dollars to
accomplish. This  paper will  discuss AFCEE's  automated approach for  identifying  background  locations, the
statistical methodology used to calculate Air-Force wide background concentrations, and the nature of these
background concentrations for both ground water and soil.

ERPIMS DATABASE
ERPIMS (previously known as IRPIMS) stores some 12.5 million analytical sampling results from  196 Air Force
installations. Data from 40,000 distinct sampling locations (wells, borings, etc.) is captured by the system. The
ERPIMS hardware consists of a Digital Equipment Corporation (DEC) Alpha® 4100 computer that runs Oracle*
7.3 on a VMS operating system. The system has been operational since 1987 and is managed by AFCEE/MSC.

DETERMINATION OF BACKGROUND LOCATIONS
Even when investigations specifically target contamination by drilling wells and borings into areas known to be
hazardous waste sites,  the non-detect (ND) rates for organics are surprisingly high. For example, the  ND rates
for TCE, which is highly mobile and known to be the most ubiquitous constituent found on Air Force installations,
are on the  order of 65% in ground water. For most other organic constituents, the ND rates are approximately
90% for ground water. For soil, the ND rates for organics tend to be even higher. As a result, a wealth of existing
sampling locations  are known  to  be uncontaminated  and  available for use as locations for background
calculations. This  knowledge was used to construct a computer algorithm that identifies  background  locations
across the entire Air Force. The algorithm, which was written in Structured Query Language (SQL), searches out
all locations that have been sampled for both inorganics and organics. Sampling locations showing evidence (i.e.
detects) of organic  contamination are  then eliminated from the search and those remaining are retained for
further consideration as background locations. Both upgradient, downgradient, and sidegradient locations could
potentially be identified as background sampling locations. There were substantially more background locations
identified for soil  as opposed to ground water. On  average, at least 25  background well  locations and 50
background borehole locations  per Air  Force installation  have  been  identified  using these  procedures. As
indicated in the next section that follows, the magnitude  of distinct sample locations and  the  sample sizes
generated from these locations will more than adequately meet the  requirements for the statistical calculations
used to determine background levels.

DATA ANALYSIS AND CALCULATION OF BACKGROUND LEVELS
For calculation of background levels at individual installations and sites, AFCEE's  approach and statistical
methodology are similar to guidance published by EPA  (EPA 1989,  1992), ASTM  (1996), and Gibbons (1994).
Using this guidance,  normal 95% confidence  95% coverage upper tolerance limits are used.  Depending on the
distribution  of individual data sets and the percentage of detects, nonparametric tolerance limits are also typically
used. AFCEE, like  EPA and others, requires  at least  2 sampling  locations to minimally account for spatial
variability and  a total sample size of at least 8 (n = 8 for each constituent) to provide ample statistical power for
the background calculations. However, Air-Force wide inorganic data is complicated by multiple detection limits,
diverse hydrogeologic terranes, variability over 3-dimensional space,  a variety of types of hazardous waste sites,
multiple Air Force bases, different waste handling practices, and the like. All of these issues force  one ultimately
to discriminate background levels across more than one hydrostratigraphic unit or more than one soil horizon. As
a result and for the  purposes of this investigation,  the 95th  percentile (Prc95) of the data associated with each
analyte was used as the statistic of choice to  best  represent background. This  approach parallels  AFCEE's
guidance for individual  installations in that  the  95% upper tolerance limit focuses on the 95th percentile of the
data (i.e. a tolerance limit is similar to an  upper confidence limit on a specified percentile or coverage of the


                                                  74

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
data, in this case the 95th percentile). Calculation of the median and the 99th percentile of background levels for
analytes detected  in both ground water and soil were also calculated and can be found in Tables 1 and 2. These
tables also provide information that  qualifies the results of the background calculations including: sample size,
the number of distinct background well locations, the number of Air Force bases having background locations,
and the detection frequency. All statistical analysis was performed using SAS® statistical software.
                 Table 1. Air-Force Wide
Background Levels of Inorganics in Ground Water
    As of May 1998
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Cyanide
Fluoride
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Nitrate
Nitrite
Potassium
Selenium
Silver
Sodium
Strontium
Sulfate
Sulfide
Thallium
Vanadium
Zinc
Sample
Size
5656
5839
7259
6828
5891
812
7153
7892
5121
6412
1796
2132
6844
8916
6465
6017
3481
6471
1788
662
6561
6794
6812
6561
31
3175
176
5698
5378
6820
Wells
Sampled
2493
2843
2996
2750
2843
461
3202
3169
2538
2873
987
1351
2838
3552
2718
2808
1779
2915
1074
468
2750
2934
3165
2750
22
1794
122
2780
2443
2863
AF
Bases
86
97
107
98
98
26
110
112
83
102
49
60
93
118
92
105
58
104
61
33
91
105
108
91
2
77
14
97
84
104
Detection
Freq (%)
54
7
32
83
9
65
9
35
13
29
2
61
78
33
83
9
17
25
67
6
98
12
4
98
100
92
11
4
37
67
Median
(mg/L)
0.0735
ND
ND
0.07715
ND
0.042
ND
ND
ND
ND
ND
0.24
0.56
ND
0.0708
ND
ND
ND
0.8
ND
27.4
ND
ND
27.4
0.17
36.09
ND
ND
ND
0.02
Prc95
(mg/L)
44
0.014
0.044
0.6
0.002
1.46
0.0049
0.195
0.031
0.086
ND
2.263
54.4
0.047
2.840
0.00036
0.021
0.2
24.600
0.02
452
0.0081
ND
452
3670
430
0.14
ND
0.11
0.33
Prc99
(mg/L)
201
0.2
0.171
2.03
0.009
12
0.017
1.52
0.12
0.371
0.015
4.9
240
0.23
9.2
0.0017
0.147
0.83
67
1.1
4050
0.1
0.0155
4050
9070
2420
9.3
0.16
0.464
1.67
BACKGROUND LEVELS FOR GROUND WATER
The universe of distinct monitoring wells that were sampled simultaneously for both  inorganics and organics
across the Air Force is approximately 4000 wells as of this writing. The query  used to identify the background
data set resulted in the analysis of over 145,000 analytical records. Depending on the analyte,  the number of
background wells used in the analysis varied from 22 (strontium) to 3552 (lead) and sample sizes  varied from 31
(strontium) to 8916 (lead). Background data was captured from as many as 118 Air Force installations for lead
and as little as 2 installations for strontium. Potassium and sodium were detected 98% of the time while cyanide
was detected 2% of the time. Other analytes such as strontium and sulfate were detected over 90% of the time;
however, the detection frequency for strontium was represented by only 2 Air Force bases and 22 monitoring
wells. Some constituents were not typically detected at background  locations. The following analytes had median
concentrations that were  below the method  detection limit (MDL):  antimony, arsenic,  beryllium,  cadmium,
chromium, cobalt, copper, cyanide, lead, mercury, molybdenum, nickel, nitrite, selenium, silver, sulfide, thallium,
and vanadium. The 95th percentile of the data sets  for cyanide, silver, and thallium were also below MDL. This
indicates that they are rarely detected in ground water and was substantiated by detection frequencies that were
found to be in the neighborhood of 2% - 4%. Conversely, some inorganic constituents were detected frequently
and at levels that exceeded important environmental thresholds such as Maximum Contaminant  Levels (MCLs)
                                                  75

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
or Action Levels for drinking water. The following analytes had background levels (95th percentile) that exceeded
MCLs: antimony, chromium, and nitrate. The background level for lead exceeded the Action Level of 0.015 mg/L
set for drinking water measured at the tap. This may suggest that some regulatory limits are placed artificially
close to observed background levels.

                      Table 2. Air-Force Wide Background Levels of Inorganics in Soil
                                          As of May 1998
Analyte
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Cyanide
Fluoride
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Nitrate
Nitrite
Selenium
Silver
Sodium
Strontium
Sulfate
Sulfide
Thallium
Vanadium
Zinc
Sample
Size
13077
15051
17212
15290
14724
790
17464
17549
11815
15396
3220
1270
13719
20784
13495
15465
10584
15167
1400
107
16966
17600
12161
92
1416
204
15186
12342
16017
Wells
Sampled
4840
5683
6165
5765
5513
396
6738
6689
4359
5764
1299
224
4939
7523
4837
5492
3581
5677
273
30
6019
6598
4466
24
273
162
5580
4645
5996
AF
Bases
80
90
101
98
89
16
103
103
81
89
47
8
82
113
80
94
56
92
12
4
99
103
81
2
7
10
89
80
90
Detection
Freq (%)
99
8
66
98
64
63
20
93
60
83
5
79
99
76
99
8
8
68
47
50
8
7
64
100
96
12
6
97
98
Median
(mg/L)
6,510
ND
1.6
56.25
0.3
24.7
ND
9.1
3
7.9
ND
3.4
9180
5
187
ND
ND
6.1
ND
0.008
ND
ND
120
25.1
13
ND
ND
18.6
25.2
Prc95
(mg/L)
23700
5.5
13.8
332.64
1.1
108
2.56
51.8
15.2
53
0.155
9.9
33600
54
856
0.11
1.8
38.3
7.95
0.499
0.87
0.93
1300
111
200
24
0.352
66.6
111
Prc99
(mg/L)
84600
29.3
43.3
995
2.4
201
10
388
28.4
230
2.3
17
82900
340
2380
0.58
7.99
160
42.75
1.1
23.3
7.35
3260
8020
1340
99.7
19
142
540
BACKGROUND LEVELS FOR SOIL
The universe  of  distinct  boreholes  sampled  for both  inorganics and  organics  across the Air Force was
approximately  8100 boreholes. The query used to identify the background data set resulted in the analysis of
over 325,000 analytical records. Depending on the constituent, the number of background  boreholes used in the
analysis varied from 24 (strontium) to 7523 (lead) and sample sizes varied from 92 (strontium) to 20784 (lead).
Background data was captured from as many as 113 Air Force installations for lead and as little as 2 installations
for strontium. Since inorganics tend not to be particularly mobile in ground water, it is not surprising that they are
detected at higher frequencies in soil vis a vis ground water. The following constituents had detection frequencies
exceeding 95%: aluminum, barium, chromium, iron, manganese, strontium, sulfate, vanadium, and zinc. The
high detection  frequencies for strontium and sulfate,  however, are misleading since the number of Air Force
bases represented is 2 and 7, respectively. Strontium, in particular, had only 24 boreholes that were sampled and
identified as background  locations across  the Air Force.  Some analytes  were  not  commonly detected at
background  locations. The following  analytes  had  median  concentrations  that were below MDLs- antimony,
arsenic, cadmium,  cyanide, mercury,  molybdenum,  nitrate, selenium, silver, sulfide, and thallium. These same
constituents also had median concentrations below MDLs for  ground water. None of the 95th percentiles of any of
the data sets for soil fell  below  MDLs, unlike the situation found for ground water The following  analytes had
detection frequencies that were below 10%:  antimony, cyanide,  mercury, molybdenum, nitrate selenium  silver,
sulfide, and thallium. On rare occasion, inorganic constituents were detected at levels that exceeded important
                                                  76

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


environmental  thresholds (using residential  criteria) such as  Preliminary  Remediation Goals (PRGs), Human
Health  Screening  Levels (HHSLs), and Risk-Based Concentrations promulgated by various EPA regions. The
background level for both arsenic and beryllium exceeded the PRGs and HHSLs for EPA Region 9 and Region
6, respectively. Iron exceeded both the HHSLs and the PRGs for EPA Regions 6 and 9, respectively. As in the
case for ground water, these results also suggest that regulatory limits may be artificially placed too close to
observed background levels.

SUMMARY
Computer algorithms  developed by the Air Force were used to automate the process of identifying background
locations for inorganics  occurring in ground water and  soil.  These procedures identified large  numbers of
background locations and  a  more than  adequate  sample size which was  used to determine Air-Force wide
background levels for some 30  inorganic constituents. This baseline of numbers which was calculated in this
study provides insight on the  nature of background  variability across the Air Force and gives decision makers a
"feel" for representative  background levels.  The 95th  percentile statistic calculated from individual constituent
data sets is believed to  best  represent background  levels given  the inherent complexities associated with
analyzing these large and diverse data sets.  Potassium and sodium  were highly detected in ground water; while
aluminum, barium, chromium, iron, manganese, strontium, sulfate, vanadium, and zinc were frequently detected
in  background soil. Some constituents were not commonly detected at  background  locations  across the Air
Force.  The  following analytes were not typically found in ground water: antimony, arsenic, beryllium, cadmium,
chromium, cobalt, copper, cyanide, lead,  mercury, molybdenum, nickel, nitrite, selenium, silver, sulfide, thallium,
and vanadium. For background soil, the following analytes were not typically detected in soil: antimony, cyanide,
mercury, molybdenum,  nitrate, selenium, silver,  sulfide, and thallium. The results of this investigation suggest
that  some  regulatory limits may be  placed too close to observed  background levels for selected analytes.
Analytes that may fall into this category include antimony, chromium,  and  nitrate for ground water; and arsenic,
beryllium, and iron  for soil.  Background  levels  of these  constituents were  found to exceed   important
environmental  thresholds. This automated approach of performing background investigations using existing data
which has already been paid for, affords the Air Force many cost benefits. Use of this methodology can eliminate
the need to conduct a separate  background  investigation for individual sites or at the installation level and can
save hundreds of thousands of dollars that would otherwise be needlessly spent.

REFERENCES
American Society for Testing Materials (ASTM), 1996, Provisional Standard Guide for Developing Appropriate
   Statistical Approaches for Ground-Water Detection Monitoring Programs, Designation: PS 64 - 96.
EPA, Office of Solid Waste, 1989, Statistical Analysis of Ground-Water Monitoring Data at RCRA  Facilities,
   Interim Final Guidance.
EPA, Office of Solid Waste, 1992, Statistical Analysis of Ground-Water Monitoring Data at RCRA  Facilities,
   Addendum to Interim Final Guidance.
EPA Region 6, 1996,  Human Health Media-Specific Screening Levels.
EPA Region 9, 1996,  Preliminary Remediation Goals (PRGs).
EPA Region 3, 1997,  Risk-Based Concentrations.
Gibbons, Robert D., 1994, Statistical Methods for Groundwater Monitoring; John Wiley & Sons.
                                 NEW TOOLS FOR LIQUID SAMPLING
    'Evaluation and Comparison of the Performance of Liquid Sampling Devices In Stratified Liquids"

                                           James D. Hoover
                           Department of Civil and Environmental Engineering,
               Washington State University Tri-Cities, 100 Sprout Road, Richland, WA 99352
                                             509-372-6972
                                           Scott R. Somers
                Advanced Concepts & Design, Inc., 77 Symons Street, Richland, WA 99352
                                             509-943-1431
                                                  77

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


EXECUTIVE SUMMARY
Systematic differences in sampling performance were found to exist between the Drum Thief, COLIWASA, and
ACD liquid sampling devices in laboratory tests simulating  the containerized sampling of stratified liquids. The
study involved two phases of testing: (1) the collection of 250 samples with these devices using water and corn
oil at a ratio of 1:1 by 17 professional and 35 inexperienced sampling personnel, and (2) the collection of 216 test
samples at wateroil ratios ranging from 95:5 to 5:95  by a single user. Both small  volume (<250 ml) and large
volume (wlOOO ml) device models were evaluated. Statistically significant differences in the accuracy, precision,
spillage, and time of sampling were found to exist among these devices. The relative performance of the devices
were also found to coincide with the rating of the devices by 52 volunteer users.

Nearly  all differences in sampling performance were found to vary systematically with device type. Differences in
sampling accuracy between  devices range from 18%  to 175%, and sampling accuracy was found to be highly
dependent on  the liquid  ratio for some devices. Sampling  accuracy with  tile COLIWASA and  Drum  Thief
decreased sharply when the proportion of  either liquid was <30%, indicating that there are limits on the  liquid
proportions that  can  be  sampled with these  devices.  Sampling reproducibility among  users  also varied
systematically with device type, but when used by a single individual,  precision was >97% for all devices. The
amount of spillage  associated with sampling ranged from zero with some devices to as much as 10% of the
sample volume with others. Sampling time among devices differed by as much as 300%.

The performance of the small volume and  large volume models of the ACD devices exceeded that of the other
devices in essentially all  categories of evaluation. Nearly  all aspects of sampling performance appear  to be
dominated by differences  in  the inherent design and function of the devices, which cause some devices to also
be more user intensive than  others. Among the devices tested, sampling performance with the ACD devices was
found to  be  the  least  dependent on user factors,  including experience.  The ACD devices also received the
highest ratings in the user survey.

These  results have  important implications for the sampling of liquids that is routinely performed for the purpose
of characterizing their  composition and potential hazard. These  implications include health, safety, economic,
legal, and practical considerations for the  handling and disposition  of the sampled liquids. These results also
provide a quantitatively  basis  for  comparing the  performance  of liquid  sampling devices  used  in the
environmental industry and to provide a basis for improving sampling performance and sampler design.

PURPOSE AND BACKGROUND
The sampling of liquids is  an activity that is performed thousands of times daily by  numerous federal, state, and
local and agencies for the purpose of characterizing type, nature, and/or hazard of unknown liquid substances.
Environmental regulations also require  that essentially "all liquid waste materials be sampled and appropriately
characterized for the  purposes of appropriate  handling, transportation, treatment  and/or disposal. Although a
relatively small number of  sampling devices  are  used  to collect these samples, there  is little documented
information regarding the  performance of  these sampling devices or a quantitative  comparison to  serve as a
benchmark for assessing quality control or product improvement

These  studies were undertaken to assess the performance of five liquid sampling devices, The purpose of these
studies was to establish a  quantitative baseline for the evaluation and comparison of liquid samplers used  in the
acquisition of representative samples of stratified liquids. These studies were motivated by the need to establish
a quantitative basis for evaluating the quality of samples that are obtained with various sampling devices.

The sampling and analysis of waste materials, contaminated media, and other materials of unknown composition
are integral components of many environmental activities including regulatory compliance in accordance with the
Resource  Conservation and Recovery Act (RCRA; 1976) and the Comprehensive Environmental  Response
Compensation  and Liability Act (CERCLA; 1980). Sampling protocols have been established to ensure uniformity
in the generation on of sampling data (e.g.,  Environmental Protection Agency's (EPA's) Office of Solid Waste
and Emergency Response (OSWER (e.g., EPA, 1986). Although general guidelines and procedures for liquid
sampling exist (e.g.,  ASTM 1994; EPA  1991), there are no standards pertaining to  sample  quality  (e.g.,
accuracy), and essentially  no information on the quality of the samples that call be obtained with various types of
sampling devices.

EXPERIMENTAL DESIGN AND TEST METHODS
This  Study focused on the evaluation  of three types of  liquid sampling devices for the sampling of stratified


                                                 78

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


liquids from containers such as drums. The three types of liquid sampling devices evaluated were the  Drum
Thief, the Composite Liquid WAste SAmpler (COLIWASA),  and a new  product, the  Advanced  Concepts &
Design (ACD) liquid samplers. Further description of the sampling devices is  provided in Attachment 1. These
liquid sampling devices were chosen because they have all been  recognized by the standards organizations as
devices appropriate for liquid sampling (e.g., ASTM 1994,  1995, 1996, 1998;  EPA 1986).  The Drum Thief and
the COLIWASA have also been the most commonly used liquid sampling devices in the collection of composite
liquid samples. The Drum Thief and small volume models of the other two devices (COLIWASA-S  and ACD-S)
were evaluated for the collection of small volume liquid samples <250-ml. The COLIWASA-L and ACD-L models
were evaluated for the collection of larger volume samples (i.e., 1000 ml).

Conditions of stratified liquid sampling were chosen because most data quality issues in  containerized  liquid
sampling involve representative sampling of segregated liquids with different physical and chemical properties.
Two essentially immiscible liquids, distilled water (p=1.0) and  more viscous corn oil (p=0.9), were used in  these
tests to simulate simplified test conditions for stratified liquids. The evaluation involved two phases of study. In
the first phase of testing,  accuracy, precision, spillage,  and  sampling time were evaluated for samples collected   ^<
from simulated waste drums containing equal proportions of water and oil (ratio -1:1). Phase II tests  involved the   j]j
assessment of the accuracy and precision of the sampling devices for seven different  watenoil ratios ranging
from 95:5 to 5:95.                                                                                        (/)
                                                                                                         •
                                                                                                        D
Test Conditions                                                                                          ^
All samples were collected under controlled laboratory conditions simulating  the  sampling of stratified liquids   CC
from 55-gallon waste drums. In the Phase I tests, 17 professional and 35 inexperienced volunteers collected   ^
samples with each of the  five sampling devices from simulated waste drums under the supervision of laboratory   QJ
personnel. The simulated waste drums consisted of cut-away  55-gallon barrels fitted with 4" ID acrylic cylinders   j
34"  in length (5.6  liter capacity per cylinder) sealed on the bottom and mounted below the bung  hole. In the
Phase II tests, samples  were collected  from freestanding acrylic cylinders by a single individual.  Measured   ^
volumes of water and corn oil were placed into the cylinders to simulate stratified liquid conditions. A user survey   UJ
was also conducted in conjunction with the Phase I tests to  obtain all independent  assessment of the overall   JB
performance of  each device  from the  perspective  of  user  personnel. Details  regarding, the experimental    ?>
procedures of the  Phase I and II tests are described in Attachment 2.                                            96% accuracy across most  of the  range
of watenoil  ratios.  For tests performed with a watenoil  ratio of 1:1, under-sampling of one liquid resulted in
complementary oversampling, of the other liquid by an equivalent  amount.  Sampling biases for tests performed
at other ratios were not complementary (Figure 1).

Precision
Measurement reproducibility,  i.e.,  precision,  varied systematically with device  type among multiple users, but
varied little when  used by a single individual.  For the tests  performed with a watenoil ratio  of 1:1,  the ACD
devices produced  the greatest precision among both experienced  and inexperienced users. For the tests with a
watenoil ratio of  1:1,  measurement  uncertainty  (at the 95%  confidence interval)  with the  Drum  Thief and


                                                   79

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


COLIWASAs were about twice as large as that with the ACD devices (Figure 2). However, when all samples
were collected by a single user, the sampling precision (at the 95% confidence interval) for all devices was >97%
for all wateroil ratios. Thus, the level of precision produced by an individual user was significantly better that that
produced among multiple users. Sampling precision appears to be only slightly  influenced by user experience
and to be generally insensitive to differences in liquid ratios.

Spillage
The amount of spillage from sampling and the variability in the spillage  measurements differed systematically
with device type. The average spillage with the small volume devices ranged from essentially zero with the ACD
device (0.006 ml,  1a = 0.02 ml), to as much as 10% of sample volume (24 ml, 1c =13 ml) with the Drum Thief
(Figure 3). For the tests with large volume devices, the COLIWASA-L yielded the greatest spillage and variability
(up to 11 ml, 1a =6.7ml), and the ACD-L yielded the least spillage and variability  (0.02 ml, 1a =0.07 ml) (Figure
3). The variability in spillage was generally observed to increase with  the volume  spilled.  User experience
resulted in reduced spillage only with the Drum Thief (by about 30%).

Sampling Time
The time required  to collect samples with each of the devices varied systematically with the type of device used,
and was lower among experienced users. Average  sampling times for the devices ranged from 40 seconds to
120 seconds. The small volume samples obtained by inexperienced users were  collected fastest with the ACD
device.  The COLIWASA-S  required about 30% longer,  and  the  Drum Thief took nearly three times as long
(Figures 4). The same pattern of device performance was found for samples collected by experienced  users, but
with a 20%-30% reduction in sampling times with all  devices. The sampling times  with the larger volume devices
required only about 10-15 seconds longer. Sampling  with the ACD-L was consistently about 10 seconds (10-12%)
faster than with the COLIWASA-L.

User Ratings
The ratings of specific devices by experienced and inexperienced users  were nearly identical. The small and
large volume ACD devices received the highest ratings (4.2 and 4.3 out of a maximum  of 5.0: Figure 5). The two
COLIWASA devices received the second highest ratings; about 3.0 for the smaller device, and 2.7-3.2 for the
larger device. The Drum Thief was ranked lowest by both groups of users, receiving average scores of 1.3-1.5.
The average ratings and the 95% confidence intervals for these scores are  shown  in Figures 5.

Other Comparisons
It was hypothesized that some aspects of sampling performance may be related to factors such as the height of
the user and/or arm  length for devices requiring 42" tubes to be lifted  clear of the sampling vessel for the
collection of the sample. However, no obvious correlation was found between any of the sampling performance
parameters and  physical characteristics of the user including height, user stature, or sex.

SUMMARY AND INTERPRETATIONS
The relative performance of the five liquid sampling devices evaluated in these tests is summarized in Table 1.
As indicated in Table 1, the performance of the ACD-S device exceeded that of the Drum Thief and COLIWASA
in essentially all categories  of evaluation for small volume sampling devices, and the performance of the ACD-L
exceeded that of  the COLIWASA-L for large volume  devices. The  performance of the COLIWASA-S was
somewhat better than that of the Drum Thief for most test  parameters, but the accuracy obtainable with these
two devices appears to vary from user to user. The relative performance of all five devices is also consistent with
the ranking of device performance based on the survey of 17 professional and 35 inexperienced users. The two
ACD  devices received the highest ratings,  followed by the two COLIWASA  devices, with  the  Drum  Thief
receiving the lowest user ratings.

The systematic differences  in sampling performance among the devices can largely be attributed to their design
and function. Design factors that affect performance  include the relative size of the opening through which liquids
enter  the sampler, closing  mechanisms,  and the tendency for sample leakage. These factors also  appear to
control the extent to which device performance is affected by the user.

Much of the sampling error  with the COLIWASA and Drum Thief is related to the  relatively small diameter
opening on the  sampling tube, which causes disproportionate amounts  of the  liquids to enter the  tube. The
magnitude of this error appears to depend on the opening diameter, the rate of insertion of the sampling tube into
the liquids, and  on the proportions and viscosities of the liquids sampled. Although the guidelines for use of the


                                                 80

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


COLIWASA and Drum Thief recommend a slow insertion, the rate of insertion necessary to minimize this error is
never known in blind sampling. In practice, disproportionate amounts of water and oil were obtained with these
devices by all users. These relationships are discussed further in manuscripts being prepared for publication.

Another important manifestation of this source of bias is the dependency of sampling accuracy on the wateroil
ratio with the Drum Thief and COLIWASA because this error is magnified as the proportion of a liquid becomes
smaller. Although sampling accuracy for all devices was >80%  when the proportions of both liquids were greater
than 10%, accuracy with the COLIWASA and Drum Thief decreased substantially when the fraction of the liquid
was less than 30% (Figure 1a,b). At extreme wateroil ratios (e.g., 95:5) the accuracy with these two devices was
as low as 10%. These results indicate that there are limits on the proportions of a liquid that can be sampled with
these devices. Sampling accuracy with the ACD devices, however, was >96% for most liquid ratios and >85% for
extreme liquid proportions. Although there may also be sampling limits with the ACD devices for other testing
conditions, these limits appeal to be much smaller than those for Drum Thief and COLIWASA. Liquid viscosity is
expected to affect the accuracy and  limits of sampling with  all these  devices at some  point.  However, it is
indicated from the results of this study that  sampling accuracy with the ACD devices should be significantly
greater than with the Drum Thief or COLIWASA for most sampling conditions.

These design and function factors also affect sampling accuracy with the Drum Thief and COLIWASA because
they cause performance of these devices to vary from user to user. It is possible for individual users to obtain
highly reproducible results with the COLIWASA and Drum Thief, whether or not they are  accurate.  But it is
indicated from the  results of this study that the uncertainty in  sampling accuracy among users is two to three
times higher with the COLIWASA and Drum Thief than with the ACD devices. The Drum Thief appears to be the
most prone to user errors. The high sampling accuracy and precision obtained with the ACD devices among all
users indicates that sampling quality with these devices is essentially independent of user factors, including user
experience.

Design-related performance factors unique to the Drum Thief include the leakage of liquid from tire Drum Thief
and the small tube volume necessitating multiple strokes for the collection of samples. Spillage and the potential
for sampling error related to spillage is greatest with the Drum Thief because significant leakage of liquid from
this device during transfer is inherent to its' use. Although  leakage is negligible  with the COLIWASA and ACD
devices, spillage resulting from liquid on the outside  of the  sampling tube occurs with both the COLIWASA and
Drum Thief. This Source of spillage occurs because  the 42-inch tubes must be removed from the vessel  being
sampled, and lifted above eye-level to transfer the liquid to  a sample container. With the Drum Thief, both types
of spillage are compounded because multiple strokes are required to obtain a sufficient sample volume all forms
of spillage are essentially eliminated with  the ACID devices because they are designed for transfer of the liquid
directly to a sampling container without removal of  the tube from the vessel (e.g.,  drum). These factors also
affect the  time required to collect a sample. Samples were systematically collected more  rapidly with the ACD
devices than with the COLIWASA or Drum Thief because direct transfer of samples requires less time than
removal of the tube for transfer. Sampling times with the Drum Thief are also up to three times longer because
multiple strokes are required  to obtain samples.

Although statistically significant differences were found in the performance of liquid samplers, these differences
are specific to these test conditions. Based on the evaluation of these results, similar patterns of performance are
expected for most  other sampling  conditions. The magnitude  of the differences  in  device performance  is
expected to vary systematically with changes in physical  properties of liquids such  as viscosity  and  density.
However, these relationships, and the  effects of other factors such as suspended solids oil device performance
are not presently known, and are the subject of further investigations.

There are  many potentially important implications for the results obtained in this study. Although  the practical
significance  of the differences  in device performance must be  assessed oil a case-by-case basis, these results
have a number of broad implications for liquid sampling. The implications for significant differences in sampling
accuracy are important because the  objective of  sampling is to determine the  types and amounts of liquids
present so they can be properly (and  legally) handled and dispositioned. Representative sampling has safety,
economic,  legal,  and  practical implications  because  the accuracy  of sampling  determines,  correctly  or
incorrectly, the type and level of hazards associated with the liquids. Errors in sampling accuracy have direct and
indirect economic implications involving billings and associated costs that are based on the amounts of specific
chemicals  that must handled or otherwise dispositioned. The level of sampling accuracy required also depends
on the inherent hazard of the materials which is usually  not known prior to sampling. Thus, sampling should  be


                                                  81

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


carried out in a manner capable of providing the highest practicable accuracy. The indication that the proportion
limits on sampling with the Drum Thief and  COLIWASA  are relatively high (5%-30%) is especially notable
because this implies that liquid fractions smaller than 30% are subject to significant sampling errors with these
devices. Sampling precision requirements may also vary from case to case, but it is important that the quality of
sampling be comparable and reproducible from  user to user. Spillage has  direct implications  for worker safety
and  for the costs of handling  spillage and  any  subsequent contamination  in accordance with regulatory
requirements Sampling time differences of up to 300% can have cost-benefit implications, but  only if the quality
of sampling is not compromised. Sampling time and ease of use can also be important when sampling must be
performed in personal protection equipment  (PPE) or under adverse (e.g., weather)  conditions. Long-term
implications  of  sampling  performance also include the  integrated  cost  savings  associated  with  improved
sampling efficiency and minimizing the negative implications of poor quality sampling.

CONCLUSIONS & RECOMMENDATIONS
Systematic differences in sampling performance were  found between the  Drum Thief, COLIWASA,  and ACD
liquid sampling devices in tests simulating the  sampling of stratified liquids. Accuracy and precision were found
to be highest, and spillage and sampling time least, with the small volume (< 250 ml) and large volume (1000 ml)
ACD  devices. Sampling accuracy with the COLIWASA and Drum  Thief appear to depend on  the relative
proportions of the liquids, viscosities,  and on the users themselves. The ACD devices were found to be largely
independent of  user factors,  including user experience for the test conditions evaluated,  and  to  exhibit high
sampling accuracy for proportions ranging from 5% to  95%  of each liquid. The  Drum Thief and  COLIWASA
appear to have  limitations on proportion of a liquid that can be reliably sampled between 5%-30%.  The overall
performance of  the devices based on a survey of 17 professional samplers and 35 inexperienced  nolunteer
samplers indicate that both groups of  users rated the ACD devices highest followed by the COLIWASA, and the
Drum Thief.

These results have potentially important implications for the  handling and  disposition of potentially hazardous
liquids, and immediate implications for sampling efficiency and the reduction of risk to sampling  personnel from
spillage. The results provide a baseline for quantitatively comparing the performance of liquid sampling devices,
for matching sampling needs with product  performance,  and also provides a basis for the improvement of
sampling performance and sampler design.

Although these tests are specific to the test conditions  used, these results provide insight regarding the factors
that  affect liquid sampling  under these and other conditions. Tests with liquids over wider range  of densities,
viscosities, and  liquid conditions including suspended solids, and numbers of liquids, will be required to establish
functional relationships for the performance of these devices over the range of sampling conditions encountered
in the field. Based on the results of this study, the ACD devices appear to  provide significantly better sampling
performance and data quality than the COLIWASA  and Drum Thief for most  conditions of  stratified  liquid
sampling.

REFERENCES
American Society for Testing and Materials, 1994, Standard Practice for  Sampling With a Composite Liquid
   Waste Sampler (COLIWASA), ASTM D5495-94, 2pp.
American Society for Testing and Materials, 1995, Standard Practice for Sampling Single or Multilayered Liquids,
   With or Without Solids,  in Drums or Similar Containers, ASTM D5743-95, 5pp.
American Society  for Testing and  Materials,  1996,  Standard  Guide for Sampling of Drums   and Similar
   Containers by Field Personnel, ASTM D6063-96,18pp.
American Society for Testing and Materials, 1998, Standard Practice for Sampling Single or Multilayered Liquids,
   With or Without Solids,  in Drums or Similar Containers, ASTM D5743-98, 7pp.
Coin prehensive Environmental Response,  Compensation, and Liability Act (Superfund), 42 DSC 9601 et seq.,
   1980.
EPA, 1986,  Compendium of  ERT  Waste  Sampling  Procedures,  EPA/540/P-91/008,  U.S.  Environmental
   Protection Agency, Washington, D.C.
EPA, 1991, Test Methods for the Evaluation of Solid Waste Physical/Chemical Methods, SW-846 3rd ed., U.S.
   Environmental Protection  Agency, Washington, D.C.
The Resource Conservation and Recovery Act of 1976, 42 USC 6901 et seq., 1976.
                                                 82

-------
                          WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                   -

                e>   lit
                DJ>

               "5.   «
                £
               
-------
                          WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                                          GHucufmrienceil llsorn

                                                          • Experienced Uier»

                                                          OSiiigte Uner
                                 COUWASA-S
 Figure 2. Accuracy and precision test results.
 Accuracy and precision for liquid samples collected with five sampling devices by experienced inexperienced,
 and individual users. Sampling accuracy is indicated in terms of the average percent of over-sampling and/or
 under-sampling  (bias) for water at a wateroil ratio of 1:1. Zero over-sampling corresponds to  100%  accuracy.
 Bias levels  for  water mirror  bias  levels  for  oil  (i.e., 10%  over-sampling of  water corresponds to 10%
 under-sampling  of oil). The vertical lines at the top of each bar represent tile precision of the measurements at
 the 95% confidence interval (i.e., 95% confidence that the mean values lies within the range  of these error bars).
                                                                0 Inexperienced I'serx

                                                                • Experienced ( ncrs
                      M'O-S
                                             Drum I Kief
                                                                       ACU-I.
                                                                                 OOLIH ASA-!.
Figure 3. Spillage Test Results.
Amount of spillage resulting from the collection of liquid samples with five sampling devices  The height of the
columns denote the average volume of liquid spilled (ml) by experienced and inexperienced in the process of
™KSLTP  < ffromhsimulated  waste drums. Vertical  lines at the  top  of  each  bar represent  the 95%
confidence intervals for the measurements.
                                                   84

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
            S
            a
                                                                 E3 Inexperienced Users

                                                                 • Experienced Users
                      ,U:D-S
                                COLIWASA-S
                                             Drum Thtef
                                                                      ACD-I.
Figure 4. Sampling Time Test Results.
Time required for the collection of a sample with each of five sampling devices. The height  of the columns
denote the average time (in seconds) required for experienced and inexperienced users to collect a liquid sample
from a simulated waste drum and transfer the liquid to a sampling container. Vertical lines at the top of each bar
represent the 95% confidence intervals for these measurements.
                                               £3 Inexperienced Users

                                               • Experienced Users
                      A( TKS
                                 COUWASA-S
                                              Drum 1 hief
                                                                         ACD-L
                                                                                   fOI.IVVASA-l.
Figure 5. User Rating Results
Results of a  survey of 17 professional and 35 inexperienced users regarding the performance of five liquid
sampling devices. After collecting samples with each device, users scored each device on a scale of 1 to 5, with
5 = best, and 1 = worst. The heights of the columns denote the average user score. Vertical lines at the top of
each bar represent 95% confidence intervals for the mean values.
                                                    85

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


EXPERIENCED
USERS (n= 17)
Accuracy
Precision
Spillage
Sampling Time
User Ratings
INEXPERIENCED
USERS (n=35)
Accuracy
Precision
Spillage
Sampling Time
User Ratings
SINGLE USER
(n=80)
Accuracy
Fraction = 1 1
FiMion IO«
-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
jar.
ACD Samplers
The third type of sampling device tested was a product developed and manufactured by Advanced Concepts &
Design (AC&D), Inc. The ACD devices are essentially have an open tube design (0.87" and 1.5" ID) in which fluid
in the sample tube is manually drawn by pulling a small cone-shaped plastic plunger upward from the bottom of
the tube using an attached rod or cord extending the length of the collection tube. A sample container Gar)
adapter is attached at the upper end of the tube to receive the sample. Samples are obtained with this device by
lowering the collection tube into the liquid and transferring the liquid from the tube to the sampling container Gar)
by pulling the plunger up through the tube with the rod or cord, essentially pumping the  liquid from the tube into
the aligned sample jar without removing the collection tube from the sampled vessel (e.g., drum).

The COLIWASA and the ACD devices both have small volume (-200-250 ml) and large volume (-1000  ml)
models. All three devices are available in glass and high density polyethylene (HOPE) models.

ATTACHMENT 2
Laboratory Procedures
Phase / Testing
Samples were  collected with each  of  the five sampling  devices from  simulated waste  drums under  the
supervision of laboratory personnel. The simulated waste drums consisted of cut-away 55-gallon barrels fitted
with 4" ID acrylic cylinders 34" in length (5.6 liter capacity per cylinder) sealed on the bottom and mounted below
the bung hole. Sampling was performed by two groups of volunteer. A total of 250 samples were collected by 17
experienced (professional) samplers, and 35 inexperienced participants with no prior sampling experience. A
description of the study and procedures for the use of each device was provided to each participant  (herein
referred to as users) prior to testing. Each user was tasked with drawing a sample from each of five simulated
waste drums using one of the five liquid sampling devices at each drum station, and transferring the samples to
clean sampling containers Gars) placed on top of the drum.

Data on the sample volume, liquid ratios, sampling time, and spillage were obtained for each sample collected.
Pre-weighed drip trays and absorbent material (e.g., paper towels) were placed on the top of the barrels to
capture spillage associated with the transfer of liquid from the drum to the sampling jars. At the conclusion of
each  sampling  series,  laboratory personnel measured  the  wateroil volume  ratio and total volume  in each
sampling jar in  graduated cylinders. The drip trays were then re-weighed to quantitatively assess the spillage
associated with each device. The sequence in which the devices were used was rotated randomly between users
to minimize bias due to the sequence in which the devices were used. All devices and testing materials were
thoroughly cleaned  between sampling series,  liquid columns refilled and recalibrated,  and pre-measured spill
trays replaced.

All  participants completed a user survey at the conclusion  of their sampling series. In  the survey, users rated
each device in terms of overall performance, and commented on device features, attributes, and shortcomings.
Users also provided  personal information on experience and physical  characteristics such as height,  sex, and
age. Measurements were also made of wrist girth and forearm  length to assess correlation between  sampling
performance and physical characteristics of the users.

Phase II Tests
Sampling procedures similar to those used in the Phase I tests were employed in the Phase II tests. However, all
samples were collected from freestanding acrylic cylinders by  a  single  individual.  Multiple  samples were
collected with each sampling device for wateroil ratios ranging from 95:5 to 5:95. A total of 216 samples were
collected in Phase II testing. The following are  the wateroil ratios, and the  number of  measurements made at
each ratio.

Numbers of samples collected in Phase II testing at each wateroil ratio
Preliminary test, were performed with glass and plastic models of the COLIWASA and Drum Thief to assess
performance among these models of the same brand. The samples were collected with each of the seven device
models  at a wateroil ratio of 50:50, including ten single-stroke and double-stroke samples with both of the
COLIWASA-S models. The best performing models of the Drum Thief and COLIWASA-S were then used in tests
at other wateroil  ratios. The  glass model of the Drum Thief was selected based on  slightly better  sampling
accuracy than the high-density polyethylene (HOPE)  model. The COLIWASA-S with the (HOPE) plunger head


                                                  87

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
was selected for use in the remaining tests because its' sampling accuracy was slightly better than that with the
model with a borosilicate plunger head. The COLIWASA-L and ACD devices tested were all composed of HOPE.
Single stroke measurements were also made for the purpose  of independently evaluating the accuracy and
uncertainty of samples obtained from multiple stroke composites.
RATIO
water/oil
95/5
90/10
70/30
50/150
30/70
10/90
5/95
COLI WASA-Small
- Glass head
- Plastic head
Single
stroke
3
3
3
10
3
3
3
Double
stroke
3
3
3
10
3
3
3
Drum Thief
- Glass model
- Plastic model
Single
stroke
3
3
3
10
3
3
3
Five
stroke
3
3
3
10
3
3
3
COLIWASA
Large
Single
stroke
3
3
3
10
3
3
3
ACD
Small
Single
stroke
3
3
3
10
3
3
3
ACD
Large
Single
stroke
3
3
3
10
3
3
3
The following steps were routinely followed in the conduction of Phase I and Phase II testing:
    Each sampler was cleaned before use.
    The plunger head was consistently lifted to a height of 4-inches prior to inserting tile sampling tube into the
    liquid-filled acrylic cylinder.
    A constant insertion rate of approximately 0.5 inches per second was used with all devices.
    Sampling devices were closed and sample transfer was initiated immediately after the head of the sampling
    tube reached the bottom of the acrylic cylinder.
    The Drum Thief and COLIWASA samplers were removed from the acrylic cylinder by pulling them  up
    through a rag to wipe out excess fluid on the  outside surface.
    Samples  obtained with the  Drum  Thief and COLIWASA devices  were transferred  directly to  graduated
    cylinders.
    Samples  obtained with the  ACD-S and  ACD-L  devices were first collected  in 250ml and  1000ml jars
    respectively, and then  transferred to graduated cylinders. The jars were allowed to drain into the  graduated
    cylinders for up to 60 seconds so that transfer loss was negligible.
    Sufficient amount of time (up to 20 minutes) was allowed for the sample in the graduated cylinder to be
    segregate into  layers of water and oil prior to recording the total volume and water and oil volumes  of.
    Ambient room temperature  was recorded  at the beginning of each series of tests. An  ambient room
    temperature from 19-21°C was maintained for all tests.
    The density  of oil in each simulated waste cylinder was determined on daily  basis at the beginning of each
    experiment.  Distilled water was used in all  tests.
            ANALYSIS OF CHEMICAL WARFARE AGENT DECONTAMINATION BRINES FOR
               LEWISITE DEGRADATION PRODUCTS USING GAS CHROMATOGRAPHY
                               WITH ATOMIC EMISSION DETECTION

                        Kevin M. Morrissey. Theresa R. Connell, and Jeffrey Mays
                EAI Corporation, 1308 Continental Drive, Suite J, Abingdon, Maryland 21009
                               Tel: (410) 612-7332/Fax: (410) 612-7317
                                   E-mail: kmmorris@eaicorp.com
                                          H. DuPont Durst
 US Army, CBDCOM, Edgewcod Research, Development, and Engineering Center, Edgewood, Maryland 21009
                               Tel: (410) 671-5270/Fax: (410) 671-2081
                               E-mail: hddurst@cbdcom.apgea.army.mil
                                                88

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Bulk chemical warfare agent (CWA) storage containers which contain decontaminated CWA waste are stored at
several locations throughout the United States. Most of these containers have been in storage since the mid
1970's, and little analytical work has been  performed to characterize the contents. This work reports on the
efforts  to  determine whether these ton containers  may have  contained  Lewisite (L,  2-chlorovinylarsonous
dichloride, CAS No. 541-25-3), an arsenic containing CWA. This work summarizes the efforts at analyzing 250
ton containers stored at a single location.

In the presence  of water, Lewisite  quickly  hydrolyzes  to  2-chlorovinylarsenous  acid (CVAA, CAS  No.
159939-86-3), and this hydrolysis product is not directly amenable to gas chromatographic analysis. The CVAA
was derivatized with 1,3-propanedithiol (PDT, CAS No. 109-80-8), and analyzed by gas chromatography/atomic
emission  detection  (GC/AED). Quantitation was accomplished  using response  in the arsenic channel, with
supporting data collected  in the sulfur and carbon channels. Spike recovery experiments were  performed at 7
different levels, and data will be reported. In addition, total  arsenic was determined  by  ICP/MS,  and supporting
techniques of GC/MSD, LC/MS and CE were used  to confirm the  presence of CVAA and other L hydrolysis
products.

Detectable levels of total As were observed in 25 of  the ton containers by ICP/MS.  The total As values ranged
from just over the detection limit of 1 ppm, to well over 7800 ppm. Detectable levels of CVAA were observed in
17 of the ton containers by GC/AED. The CVAA levels ranged from just  over the detection limit of 0.008 ppm to
2.4 ppm. In  addition to the CVAA, additional organo-arsenic compounds were detected in several of the ton
container samples. These additional organo-arsenicals may be  indicative of the presence of other As containing
CWA, interaction of CVAA with sulfur mustard (HD, CAS No. 505-60-2) hydrolysis products, or As containing
industrial waste. Correlations will be made between the presence of CVAA and other CWA hydrolysis products.
                                                  89

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                            90

-------
  WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
     ERA'S
ENVIRONMENTAL
  MONITORING
   RESEARCH
   PROGRAM
       91

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                             92

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                          INTRODUCTION, SESSION SCOPE AND PURPOSE

                                            William Stelz
              US EPA, Office of Research and Development, 401 M St., SW, Washington, DC

                                     NO ABSTRACT AVAILABLE
                 BIOAVAILABILITY AND RISK ASSESSMENT OF COMPLEX MIXTURES

                 K.C. Donnelly. W.R. Reeves, T.J. McDonald, L.-Y. He and R.L. Autenrieth
                 Departments of Veterinary Anatomy & Public Health and Civil Engineering,
                               Texas A&M University, College Station, TX

There is an urgent need to develop an accurate method to assess the risk associated with contaminated soils
and complex mixtures. Perhaps more importantly, this method should provide a  means of defining acceptable
residue levels to allow a more cost effective approach to site remediation. This research program is developing a
methodology  which can be used to estimate bioavailability. Two soils have been prepared for evaluating the
bioavailability extraction method. One soil is a Weswood  silt loam amended with model chemicals, including
chrysene,  pyrene, phenanthrene and anthracene; while the second soil is a Weswood silt loam amended with
10% (wt/wt) wood preserving waste  (WPW). The soil was spiked with either the model chemicals or the complex
mixture and  samples collected  immediately after spiking, as well as after 60 and 360 days of incubation (note
day 360 samples will be collected in the fall of 1998). Soil was extracted with pH 7 water or a 1:1 methanohwater
mixture. Other solutions to be tested will include a gastric solution (pH 2) and an intestinal solution (pH 7) and a
3:1 methanol:water mixture. Extractions  are performed by shaking 40  g of soil with a 200 mL volume of
extracting  solution for 2, 3, or 5 hours (depending on the extractant) at 37°C. Recoveries were determined using
GC-FID. In addition, these extractions will  be compared to results from desorption kinetics studies. Results from
the digestive  fluid extractions indicate that the stomach to intestinal fluid conversion (Gl) extracted only 5.4%
and 0.11% of that recovered by standard methods for chrysene and pyrene, respectively. Using these numbers
as an estimate, the hypothetical excess lifetime cancer risk for the hexane:acetone extraction would be  7.1E-5,
while  the  estimate for the Gl fluid  was 1.3E-7. The desorption study reveals two  compartments: one slowly
desorbing  and solubility limited, and one limited by desorption/diffusion, which increases in size as the  soil ages.
An animal study is  planned  for this summer  using the soil  from  this study  as  a means of  evaluating these
methods in a rodent model. This research is supported by USEPA Grant No.  R825408.
               FIELD DETERMINATION OF ORGANICS FROM SOIL AND SLUDGE USING
                SUB-CRITICAL WATER EXTRACTION COUPLED WITH SPME AND SPE

               Steven B. Hawthorne. Carol B. Grabanski, Arnaud J.M. Lagadec, Martin Krappe,
                                  Cedric L. Moniot, and David J. Miller
                  Energy and Environmental Research Center, University of North Dakota,
                                 Grand Forks, North Dakota, USA 58202

We have demonstrated that subcritical water (hot water  maintained as a liquid by a few bar  pressure) is an
excellent solvent to quantitatively extract polar and non-polar organics from soils and sludges. Subcritical water
extractions  can be highly selective; polar organics extract at lower temperatures (e.g.,  phenols  and amines
extract at 50 to 100 °C), and non-polar organics extract at high temperatures (e.g., 200 to 250  °C). By heating
water under low pressure, solubilities of polar  organics increase dramatically, and even non-polar organics such
as PAHs can increase solubilities by > 106-fold.
                                                 93

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


For water samples, both SPE (solid phase extraction) and SPME (solid phase microextraction) can be used to
extract and  concentrate organics in  the  field for subsequent analysis  (e.g.,  field-portable  GC), but are not
applicable to extracting organic pollutants from solid samples. If organic pollutants on soils and sludges could be
efficiently transferred to water, both SPE and SPME could be  very useful for field determinations  of organic
pollutants from solids. The  primary purpose of the proposed investigations is to couple  SPE and SPME with
subcritical water  extraction  of soils  and  sludges to allow  field-portable water  methods  to  be applied to
contaminated solids.

We have coupled subcritical water extraction with SPME  using very simple, inexpensive,  and field-portable
equipment. The method uses a static extraction (no pump), no flow restrictor, and no organic solvent. Soil, water,
and internal standards are placed in an extraction cell and heated for 15-60 minutes. The cell is then cooled and
the water extracted using a SPME fiber followed by direct desorption in a GC injection port. Although the method
involves  multiple  partitioning steps (water/soil, and water/SPME), quantitative results  can be obtained using
proper internal standards, e.g., deuterated PAHs are  added to calibrate for PAH determinations. Methods have
been  developed  for  PCBs,  PAHs,  and  aromatic  amines  which  give  good  quantitative comparisons to
conventional (Soxhlet) extraction. Typical sample preparation time is < 1  hour, and detection limits of < ppb are
obtained.

In contrast to the multiple partitioning steps involved in the coupled subcritical water/SPME method, coupling
subcritical water with  SPE  discs (e.g., "Empore" discs) should allow quantitative  extraction and collection of
organic analytes.  For  example, when a static  extraction cell contains the soil, water, and an SDB disc, PAHs
extract from  the soil into the water during the 250 °C heating step, but then are efficiently collected (ca. 90 %) on
the sorbent disc as the extraction cell is cooled  to room temperature. The PAHs are then eluted from the disc in a
few ml  of solvent, and the extracts analyzed by conventional GC methods. Similar approaches  are being
developed for PCBs.  In addition,  the use of subcritical water to  aid in derivatization reactions for the SPME or
SPE collection and analysis of more polar solutes (e.g., acid  herbicides, natural pyrethrins) will  be presented.
                   A FIELD PORTABLE CAPILLARY LIQUID/ION CHROMATOGRAPH

                      T. Scott Kephart, C. Bradley Boring and Purnendu K. Dasgupta
         Department of Chemistry and Biochemistry, Texas Tech University, Lubbock, TX 79409-1061
                                        James N. Alexander, IV
    Rohm and Haas Company, 727 Norristown Road, P.O. Box 904, Spring House, Pennsylvania 19477-0904

INTRODUCTION
The need for on-site, real time characterization of environmentally  hazardous sites has led to  a  considerable
interest in the development of self-contained field-portable instrumentation. Presently, two factors limit the use of
field portable instruments for environmental analysis. First, most portable instruments do not compare favorably
to  laboratory-based  instruments with  respect  to reliability and performance.  Second,  the  availability  of
stand-alone field portable equipment is limited primarily to chemical analyzers or sensors which measure a single
physicochemical property such as pH, temperature, or UV/VIS absorbance,  although  more sophisticated
instruments such as X-ray fluorescence  analyzers,  mass spectrometers, and Fourier transform IR systems have
recently been developed1. Bringing samples collected in the fields back to the laboratory for analysis results in a
time lag that can compromise sample integrity as well as delay any needed response prompted by the analytical
result.

Analysis  in the field often requires the separation of multiple analyte species before detection and quantitation.
For  the  environmental  analyst,  liquid  chromatography (LC),  including ion chromatography  (1C), and gas
chromatography (GC) remain the primary  techniques of choice. Although field portable GC systems  have been
commercially available for some time, field portable LC systems are virtually non-existent. If used in the field, LC
systems typically have to  be located in a mobile laboratory, making them at best only moderately portable.

The practice of capillary LC has undergone extensive development since its introduction twenty years ago2. The
primary advantages of moving from conventional size  columns (> 4 mm i.d.) into the capillary domain include


                                                  94

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


higher efficiencies, higher mass sensitivities, low eluent consumption, and a very small sample requirement. LC
hardware components  other than the absorbance detector have been miniaturized with the development of
capillary chromatography. Fully automatable injection valves are available down to 20 nl_; small, compact, and
inexpensive syringe pumps with very low power requirements have been developed.

Recently,  we  described  a  capillary  1C  with suppressed conductometric detection  for anion analysis3. The
miniaturization of LC hardware components in combination with the  excellent  performance of this bench-top
capillary 1C, led us to investigate the feasibility of a field portable capillary IC/LC system.

In this paper, we present a fully computer controlled, stand-alone field portable capillary 1C. Further, we describe
a dual syringe pump capillary LC system that is in the process of being prepared for field use. The construction
and performance of each  of these capillary systems is reported.

EXPERIMENTAL
The layout of  the capillary  1C  and capillary HPLC systems are shown in Figure 1a and 1b, respectively. The
pumps used were fully computer  controlled 48 000 step, motor driven syringe-type dispensers (Model 50300,
Kloehn Inc., Reno, NV) equipped with syringes of appropriate size. A 500 uL glass syringe was used for capillary
1C; stainless steel syringes (constructed in-house) were used for capillary HPLC because operating pressures are
well above 1000 psi.  A stainless steel  block was  machined in-house  with appropriately  sized  ports  to
accommodate the appropriate syringe head, a low leak dual ball and seat inlet check valve (P/N 44541, Dionex
Corp., Sunnyvale, CA), and a liquid output port, A small volume eluent reservoir was affixed  next to the syringe
dispenser and was connected to the check  valve with PEEK tubing to avoid  CO2 intrusion.  A pressure sensor
(Model SP70-A3000, Senso-Metrics, Simi Valley, CA) was connected to the  liquid output port of the stainless
steel  block using  0.25  min  i.d.  PEEK tubing. The column  backpressure was  continuously monitored to insure
proper system performance.

Capillary 1C System
Water pumped by the  syringe was passed through a mixed bed ion exchange  resin  to remove any impurities
leached from glass and metal  parts  in  the upstream  components.  A  previously described  microscale
electrodialytic sodium hydroxide generator3 (EDG) was used for eluent production. A mildly pressurized reservoir
of 25 mM sodium  hydroxide was used as the donor solution for the EDG. A 10 cm long polystyrene capillary, -80
urn i.d., 250 urn o.d., was placed at the exit of the EDG to remove the H2 gas in the eluent stream by permeation
through this tube. The  polystyrene capillary was able to perform gas removal at pressures > 900 psi at NaOH
concentrations > 40 mM.

A hollow fiber suppressor, which has been previously described,3 was deployed prior to the  detector. A H2SO4
regenerant reservoir, mildly  pressurized  (<  1 psi), was connected to  the suppressor using  Tygon tubing. The
suppressor was able to  suppress NaOH concentrations ranging from 0.5 to 40 mM to a background of < 2 uS/cm
with eluent flow rates of 1.5-2.2 uL/min. A conductivity cell was connected at the suppressor exit. A bipolar pulse
conductivity detection system4 was used for the 1C system. A laptop computer in combination  with an executable
program, written in C, provided a user interface for the data acquisition system.5

Capillary HPLC System
The dual syringe  pump capillary HPLC system is shown in Figure 1b. The syringe pumps  were coupled to a
mixing chamber,  having an internal volume of 2 uL, for isocratic or gradient  eluent production in the capillary
HPLC system. To produce a constant flow  rate while operating in the gradient  mode, this setup required a
custom control program to be written. An executable program written in Microsoft Visual Basic® provided a user
interface for instrument control. A  typical gradient program utilizes a six step gradient, although more steps can
easily be added. The program calculates the appropriate delay between steps for  each pump,  given the total flow
rate, in terms of the percentage of  flow necessary from each pump. The program  then creates a command string
with the appropriate delays for each pump and  downloads this into the resident  memory of the pump hardware
via  an   RS-232  serial  communication  port.   A   Linear   model   UVIS   200  absorbance  detector
(Spectra-Physics/Thermoseparation systems), designed for on-column detection with capillaries, was used for
HPLC detection. A smaller detector will be developed in the future.

An  electrically actuated injection  valve  equipped with 20-100 nL internal sample  loops (Valco  Instruments,
Houston, TX) was  used  for sample  introduction.
                                                  95

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Analytical columns, -50 cm long, 180 um i.d. fused silica capillaries (Polymicro Technologies), were packed
in-house. The 1C columns were packed with the same packing as commercially available Dionex AS-11 columns.
The reverse phase HPLC columns were  packed with 5 urn PRP-1  and 5 urn HSC-18 particles, respectively. A frit
was made at the exit end of the column by placing several short pieces of glass wool into 0.3 mm i.d. PTFE
tubing and push fitting the end of the column and a 75 um i.d., 365 um o.d. fused silica capillary on either side of
the glass wool. No frit was needed at the front of the column. This configuration allowed the front of the column
to be easily trimmed when compression of packing material at the head of a column led to a void over a period
of time.

For 1C experiments using sample preconcentration, the preconcentration  column consisted of an -1.5 cm long
piece of 250 um i.d. fused silica capillary packed with AS-11  packing.  The exit frit was constructed by first
pushing a small  piece of glass fiber filter (Whatman type GF/A,  Maidstone, England) into this capillary -1 cm
from the end of the packing. A 50 um i.d.,  150 um o.d. fused silica capillary was then pushed inside the larger
capillary against  the glass fiber filter and epoxied into place. The  total length of the preconcentrator column was
-8 cm. The electrically actuated  sample  injector  was equipped  with  a  six port  valve (Cheminert  Model
C3-1006-EH, Valco Instruments  Co. Inc.),  having internal dead volumes of 200 nl_  between each port, to
accommodate the preconcentrator column.

A 24 Vdc power supply (Lambda Electronics, Melville, NY) was used to power the pumps and pressure sensor. A
24Vdc-10Vdc converter  was built in-house  to  provide the pressure  sensor with 10  Vdc operating voltage. The
bipolar pulse conductivity detector (capillary 1C system only) used  a 5 Vdc power supply for operation.

Atmospheric Sampling
The capillary ion chromatograph was interfaced to a  miniaturized  parallel  plate diffusion denuder (PPDD) to
monitor ambient levels of sulfur dioxide.  The PPDD construction is  shown in  Figure 2.  The  PPDD was
constructed from two Plexiglas plates, each measuring 2x17 cm.  The active area of the PPDD (0.6 x 10 cm)
was prepared by thermally pressing silica gel particles (120 mesh or smaller) onto the Plexiglas plates. The two
plates were separated by 1.5 mm thick Teflon coated Plexiglas spacers, 0.7 x 17 cm, which completely cover the
untreated edges  of the plates. Holes were machined in the top and bottom of the silica coating to provide a liquid
input and output, respectively. Stainless steel tubing (23 gauge) was push fit into these holes and epoxied to the
plates to provide rigid liquid input/output ports. The two plates were  clamped together along their edges. Tubing
for the air inlet and air outlet was fixed by epoxy adhesive at the  bottom  and top of the PPDD, respectively.
Hydrogen peroxide (0.5 mM,  30  uL/min flow  rate) was used as the denuder liquid. The denuder effluent was
loaded onto a preconcentration column at a  flow rate of 18 uL/min for 10 minutes for analysis.  The PPDD
displayed -100 % collection efficiency up to an air sampling rate  of 0.5 standard liters per minute (SLPM). Data
presented here was obtained using this sampling rate.

RESULTS AND  DISCUSSION
Portable Capillary 1C
System Performance
The day to day reproducibility of the  portable 1C is shown in Figure 3 for repeated injections of fluoride,  chloride,
sulfate, and phthalate. The chromatograms were obtained  under isocratic conditions using an -20 mM NaOH
eluent at a flow rate of 1.5 uL/min. The  relative standard deviation of retention times ranged from 0.1% to 0.7%
within one day and 0.3% to 0.8% day-to-day. Peak efficiencies for chloride,  sulfate, and phthalate were 27 133,
21 018, and  15 422 plates/m, respectively.

Response linearity  was studied  under the same chromatographic conditions as above.  A  sample solution
containing chloride, sulfate, and phthalate  over a  concentration  range of 10-200 uM was used; fluoride eluted
near an impurity and therefore was not used  for evaluating response linearity. Linear  r2 values for peak area
response vs. injected concentration for chloride, sulfate,  and  phthalate  were  0.9959, 0.9988,  and  0.9974,
respectively. Above a concentration of 200 uM, peak broadening resulted from column overloading.

The three constituent mixture was also evaluated in terms of attainable limits of detection (LOD) under the same
isocratic conditions. The intrinsic electronic noise of the bipolar pulse detector electronics was 0.3-0.4 nS. The
noise increased to 2-3 nS/cm during chromatography, regardless of NaOH concentration or flow rate employed.
Based  on the performance at an injection concentration near the baseline and the peak-to-peak noise level,  the
S/N = 3 LOD for the three anions are as follows (LOD in uM indicated  in parenthesis): chloride (0.03), sulfate
(0.12), and  phthalate (0.25). This performance is comparable to conventional size 1C  systems. However, an


                                                   96

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


increase of >2 orders of magnitude in terms of mass sensitivity is realized in the present system compared to the
performance of a bench-top 1C using conventional size columns.

Gradient Chromatography
Incorporation  of the EDG on the high pressure side of the pump also allows gradient Chromatography to be
performed easily. The lag time, or time required for the produced NaOH to reach the head of the column from
the EDG, was measured to be 1.5 minutes using an eluent flow rate of 1.5 uL/min. Therefore, only a short time is
needed  for  a specific programmed  NaOH  concentration to  reach  the head  of the  column. A gradient
chromatogram of a sample containing 15 anions is shown in  Figure 4. A  linear gradient of ~2 mM to 38 mM
NaOH from 5 min to 17 min was used for the separation. This corresponded to a current requirement of 5-95 uA
using a water flow rate of 1.5 uL/min. The resulting separation was excellent. Peak efficiencies ranged from 10
648 plates/m for acetate to 240 152 plates/m  for chromate, with an average of 80 000 plates/m being observed
for the separation.

Miniaturized PPDD Coupled Capillary Ion Chromatography System
An air  sampling rate  of 0.5  SLPM was chosen for  evaluation of the PPDD-capillary 1C system due to the
collection efficiency of the PPDD being -100% at this sampling rate. The response linearity was studied over an
SO2 concentration range of 23 pptv to 1944  pptv  (at an SO2 concentration > 2000 pptv, sample peak height
reached the maximum value  permitted with the conductivity detection system). The response linearity over this
concentration range was excellent. A log-log  plot of  peak height vs.  SO2 concentration  resulted in a linear r2
value of 0.998. The reproducibility of the data over this concentration range was < 3.2% RSD  for  each point
sampled. Figure 5 shows a chromatogram resulting from the sampling  of clean air and 80 pptv SO2. These data
lead to a computed limit of detection of 1.6 pptv SO2.

Ambient Air Studies
The ambient concentration of SO2 was studied in Lubbock,  TX over a period of 48 h. The system operated over
this time period without any user intervention. The results are shown in  Figure 6. These results correlate well with
the ambient SO2 levels at this location.

Capillary HPLC System
Isocratic Elution
Experiments evaluating retention time reproducibility were performed on the PRP-1 column injecting samples of
biological interest. A solution  of 100 mM ammonium formate (pH 4.25) was contained in pump A and the same
solution containing 10% acetonitrile was contained in pump B. Figure 7 shows system reproducibility for isocratic
elution of 8 sample components with a 50:50 A and B mix.

The average RSD in retention times for the 8 component mixture was 0.825%. Using only a single syringe pump
and employing the same eluent conditions, the average RSD  in retention times was 0.921%. The fact that the
dual pump system actually has a lower average RSD  indicates that the main source of retention irreproducibility
is not from the pumping system but comes from other components.

Gradient Elution
The gradient capabilities of the system were  examined by separating a series of benzene derivatives on the
HSC-18 column. Pump 1 contained a  mixture  of  acetonitrile and  water (50:50); pump 2 contained only
acetonitrile. Figure 8 shows a sample  chromatogram that also  indicates the gradient profile. Figure 9 shows the
dual pump gradient reproducibility. The average RSD in retention times under gradient conditions was 0.545%.
This corresponds to a variation of 2.05 (±.88) seconds for 10 peaks eluting in under 8.5 minutes.

System performance  in terms of peak  efficiency for  the gradient HPLC system was also evaluated.  The
maximum peak efficiency was observed for ethylbenzene, which had 320 000 theoretical plates per meter. The
average peak efficiency for the 10 components was 220 000 theoretical plates  per meter. This correlates to an
average of 17 000 plates per minute.

ACKNOWLEDGMENT
This research was supported  by the U. S.  Environmental Protection Agency through Grant R82-5344-01-0. The
manuscript has not, however,  been reviewed by the agency  and no endorsements should be inferred.
                                                 97

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
REFERENCES
1.   Newman, A. R. Anal. Chem 1991,  63, 641A-644A.
2.   Ishii,  D.;  Asai,   K.;  Hibi,  K., Jonokuchi,  T.;  Nagaya,  M.  J.
    Chromatogr. 1977,144, 157-168.
3.   Sjogren, A.; Boring, C.B.; Dasgupta,  P.K.; Alexander, J.N. Anal.
    Chem 1997, 69, 1385-1391.
4.   Kar, S.; Dasgupta, P.  K.; Liu, H.; Hwang,  H. Anal. Chem 1994, 66,
    2537-2543.
5.   Boring, C.B.; Dasgupta, P.K. and  Sjogren, A. J. Chromatogr. 804,
    45-54(1998).
Figure   1a.  Schematic  layout   of   portable  1C  system.   Figure
designations: SP, syringe pump; AP, air pressure pump; PS, pressure
sensor;  ITC,  ion trap column; EDG, electrodialytic sodium hydroxide
generator; PC, polystyrene capillary; I, motorized injector; C, capillary
column; SU, chemical suppressor; D, detector; EB, electronics box.
                            Top
                                             hlnga
                                               Figure 1b. Schematic layout of capillary HPLC system.
    Pump A      Pump B
            2    3
                             Computer
12.0 —
 8.0 —
                                  Day 2 Run 3+10 uS/cm
                                  Day 2 Run 2 + 8 uS/cm
                                                                                   PP-
                                        -TF
                                        VPS
                         Side View
                                                   Figure  2. Wet parallel plate diffusion  denuder. Figure
                                  Day 2 Run 1+5 us/cm   designations: AO,  air outlet;  LI,  liquid inlets; PS, Teflon
                                                   coated  Plexiglas  spacer;  SC, silica coating;  LO, liquid
                                                   outlet; PP, Plexiglas plates; Al, air inlet; TF, Teflon film
                                  Day 1 Run 3 + 4pS/cm
                                  Day 1 Run 2 + 2 uS/cm
                                  Day 1 Run 1
Figure  3.  Day-to-day system reproducibility;  repeated
100nL injections with  ~20  mM NaOH eluent.  Injected
concentration was 20 uM for each ion. Peak identities: 1,
fluoride; 2, chloride; 3,  sulfate; 4, phthalate.
                                                   98

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
   3.0 -
                                13
                                    14
                                        15
                                                      Figure    4.    Background    subtracted    gradient
                                                      chromatogram. A linear gradient from 2 mM NaOH to
                                                      38 mM  NaOH  from 5 to  17 minutes was used. Peak
                                                      identities: 1, acetate; 2, formate; 3, methanesulfonate;
                                                      4,  monochloroacetate; 5,  bromate;  6,  chloride;  7,
                                                      nitrite;  8,  trifluoroacetate;  9,  dichloroacetate;   10,
                                                      bromide;  11,  nitrate; 12,  chlorate;  13,  sulfate;  14,
                                                      phthalate; 15, chromate. All ions were 50 uM except
                                                      dichloroacetate which was 60 uM.
                     10             20
                        Time (min)
Figure 5. Chromatograms  resulting from
sampling  blank  air  (lower trace) and  80
pptv SO2 (upper trace). Peak identities:  1,
chloride; 2, carbonate; 3, sulfate.
o
o
                                                                                             80 pptv sulfur
                                                                                            dioxide -f 1 uS/cm
                                                                                             Blank
                                                               4.0                B.O
                                                                      Time (min)
                                                                                                12.0
                                                    99

-------
     1000 —i
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
     eoo
                                                         Figure 6. Ambient air levels of SO2 in Lubbock, TX
                                                         for a 48-hr period beginning in the afternoon of April
                                                         28, 1998.
         12
          1     I     I

24         12        24
       Diurnal Time (hrs)
                                                          zsoo
Figure 7. Sample of isocratic system reproducibility.
RSD in retention time is < 1%. All samples  are  500
\M. Peak identities from left to right: cytosine, uracil,
adenine, uridine, thymidine,  adenosine, xanthosine,
and inosine.
                                                           -500
•e
o
    0.20 —,
    0.15 —
    0.10 —
    0.05 —
    0.00
                                 u
                                                   ,— 100.00
                                                   — 80.00
                                                   — 80.00
                                                   — 40.00
                                                                              6789

                                                                                Time (min)
                                                                                                     12  13
                                                     20.00
                                           Figure 8. Sample chromatogram with gradient
                                           profile using an eluent flow rate of 5.0 uL/min.
                                           All samples are  0.5  ml / 100  ml solution in
                                           acetonitrile. Peak identities from  left to right:
                                           phenol,   benzaldehyde,   benzonitrile,  nitro-
                                           benzene,  benzene,  bromobenzene,  toluene,
                                           ethylbenzene,  propylbenzene,  and  t-butyl-
                                           benzene.
       0.00
                 2.00
                          4.00        6.00
                          Timp fminl
                          8.00
                                                    100

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                   0.1 AU
                                                        Figure  9.  Dual  pump gradient  reproducibility.
                                                        Average RSD in peak retention time is 0.545%.
                                                        Peak  identities  from  left  to   right:  phenol,
                                                        benzaldehyde,     benzonitrile,    nitrobenzene,
                                                        benzene, bromobenzene, toluene, ethylbenzene,
                                                        propylbenzene, and t-butylbenzene.
                     Time (min)
         RAPID DETERMINATION OF ORGANIC CONTAMINANTS IN WATER BY SOLID PHASE
                        MICROEXTRACTION AND INFRARED SPECTROSCOPY

        David C. Tilotta. Danese C. Stahl, Sheila A. Merschman, Daniel L Heglund, and Said H. Lubbad
      Department of Chemistry, P.O. Box 9024, The University of North Dakota, Grand Forks, ND 59202

The  objectives  of this research  project are to: identify suitable  solid phase  films for determining  organic
contaminants in water by SPME/IR, determine which  organic contaminants are  amenable to the SPME/IR
method, and adapt the basic methodology to field use.

Solid Phases
Of the 15 solid phases examined to date, three polymers have been found to be useful, for SPME/IR: Parafilm
M™  (a wax-impregnated polymer/rubber composite), poly(dimethylsiloxane) (PDMS, an important solid phase
material of the SPME syringe technology) and Teflon PFA™ (a perfluoroalkoxy teflon polymer).

Analyte Classes
To date, three classes of compounds have been examined for their suitability as analytes for SPME/IR using the
three aforementioned films. Table  1 shows formal  equilibration times, linear dynamic ranges, detection limits,
and precision data (expressed  as percent relative standard deviation) for these classes for the appropriate solid
phase film(s). Multiple entries in this table for a given analyte imply that more than one film is useful Conversely,
the  absence of an entry for a given analyte/film combination indicates that that film is  not suitable for the
analysis.

Volatile organic compounds (VOCs) examined include the BTEX compounds (benzene, toluene, ethylbenzene,
xylenes), and  halocarbons such as carbon tetrachloride,  chlorobenzene,  chloroform, and p-chlorotoluene.
Parafilm M™ and PDMS both have  been useful  for the analytical determination of these compounds. SPME/IR
analyses  using  Parafilm,  have  demonstrated  the  ability of  SPME/IR in distinguishing four  of  the  six
alkylbenzenes (benzene, o-xylene, m-mylene, p-xylene)  in petroleum industry wastewater samples. Quantitation
by simple univariate calibration based on absorbance band heights have provided  good agreement with purge
and trap GC/MS  standard  methods.  Analytical  determinations of ethylbenzene  and toluene  are, however,
complicated by the spectral overlap of other components in gasoline.
                                                 101

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Gasoline fuels include the  more volatile  organics such  as  the short chain hydrocarbons (e.g., 
-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


produced or processed for human use, are released into the environment. Upon reaching the environment, the
contaminants can be transported  in a variety of ways until they reach a depositional end point, possibly in soils,
waters or in biological  tissues. Depending on the reactivity and toxicity of the contaminants, there is a multitude
of possible  implications  for both  the  environment and  human  populations.  It is necessary,  therefore,  to
understand the nature of the contaminants, where they came from, and  how they reached their depositional
location. With this understanding, cleaning contaminated  sites and  preventing further contamination become
more manageable endeavors.

There are many  compound classes that  have been contributors of contaminants into  the environment. Of
growing interest in recent decades are polycyclic aromatic hydrocarbons (PAH), organochlorine (OC) pesticides
and  polychlorinated biphenyls  (PCBs). PAH are of interest  because  they can  be indicative of  a variety  of
contaminant sources,  such as petroleum  spills or combustion processes and are  known to be carcinogenic
(LaFlemme  and Hiles, 1978 and Hites et a/.,  1980).  Although  some organisms,  such  as fish,  are  able  to
metabolize PAH,  other organisms such  as mollusks and crustaceans are unable to do so  (Law and  Biscaya,
1994; Hickey et a/.,  1995; Roper  et a/.,  1997). Thus, PAH tend to accumulate in the tissue of these organisms.
Pesticides, such as DDT, and PCBs are of interest because they not only can reside in soil and water reservoirs,
but due to their lipophilic nature,  they can accumulate in tissues and are known to be toxic (Killops and Killops,
1993 and Hickey et a/., 1995). The storage of such contaminants in tissue results in heightened concentrations in
organisms residing at higher trophic levels, thus enhancing the risk to many predatory species.

In many earlier attempts at  deciphering the sources of contaminants such as PAH and PCBs, the approach has
been to look at absolute concentration levels and  relate  a concentration gradient to a point source. Another
approach was to determine  the concentrations of compound classes relative to one another and compare these
relationships to possible source relationships. One difficulty with the first approach is that  although measurable
amounts of  a contaminant may  be located near a source known to  produce such compounds,  there is no
absolute proof that  the contaminant came from that source. Furthermore, owing to the off-site acquisition  of
pesticides,  there  is rarely  a point source nearby which  can serve  as a  possible  answer to  contaminant
apportioning scenarios. A problem with the second approach is that chemical or biological activities such as
evaporation, water washing or biodegradation  could alter the concentration of one compound relative to the
others. This chemical  or biological alteration could  change the original "signature" of the  compound class that
could lead to  an  incorrect  assignment of  a source (O'Malley et a/.,  1994).  In the past, traditional  analytical
methods using gas chromatography (GC) and gas chromatography/mass spectrometry (GC/MS) have been used
to characterize contamination  sites. As the  above arguments show however, these techniques may yield
ambiguous results (Mansuy  et a/., 1997).

A Different Technique
In recent  years another approach has developed  which  is based  on the stable isotopic composition  of the
compounds  of interest. Carbon,  for instance,  has  two  naturally  occurring stable isotopes, 13C and 12C.  It is
reasonable to assume that the  ratio of the  amounts of these two isotopes is unique for each compound derived
from a different source. If two chemical companies were to  produce the chemically identical PCBs,  for instance,
it is  probable  that the identical  compounds will have  isotope ratios characteristic of different feedstocks  or
different manufacturing processes.  Upon entering the environment, therefore, a contaminant should be  able to
be linked to a source by its isotopic composition.

The technique for compound specific isotopic analysis (CSIA) involves coupling a gas chromatograph (GC) to a
combustion furnace, which is then attached to an isotope ratio mass spectrometer (GC/C/IRMS). The effluent of
the gas  chromatograph is introduced into  a microcombustion/CO2 purification interface.  Within the GC, fused
silica columns have been found to be most effective because carrier gas flows are low (a few ml/min.  of He),
resolving power is excellent (30 to 60 meter column lengths are routine), and  column bleed  is minimal (bonded
phase columns). The column effluent is combusted to CO2  in  the interface in the presence of CuO  and water is
removed by a cryogenic trap. Ion  current intensities of masses 44, 45 and 46, which represent the major isotopic
forms of CO2, are recorded simultaneously using a high speed on-line acquisition system. A schematic of the
system is shown below (Figure 1).

The ratio of the ion current intensities of mass 45 to mass 44 is a measure of the ratio of 13C/12C and is compared
to the reference  value  of  13CI/12C. The reference  and sample  ion current intensities  are measured in  an
alternating manner,  which ensures a reliable comparison between the reference and sample isotope ratios. A
computer interfaced  to the instrument is used to calculate in the "per mil" (%o) notation. Isotope compositions are


                                                 103

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
reported as 5 (delta) values, using the following formula:
                                              R = 13C/12C
                                    5 13C = (Rsample/Rreference - 1) X 1000
     Gas Chromatograph


   Injector     Detector
      Heart split
      valve
Reference Gas-
Injector
                                                                             Figure    1.     Schematic
                                                                             showing the basic structure
                                                                             of  a  GC/C/IRMS  system
                                                                             (adapted from Freedman et
                                                                             a/., 1988).
Isotopic compositions are measured as ratios owing to the normally low absolute abundances of each chemical
species of interest. Samples with more 13C than the reference compound will have positive 5 13C values, and are
said to be enriched in 13C. Samples with less 13C than the  reference compound are said to be  depleted. It should
be noted that currently, while there are many elements with multiple stable isotopes, the most  commonly used in
analyses are carbon, nitrogen, oxygen,  sulfur  and hydrogen.  Like carbon, each of the other elements are
analyzed in a similar way, with deviations  occurring in the combustion and subsequent gas purification processes
occurring just before mass  abundances are measured. In every case,  a sample with more of the heavy stable
isotope is said to be enriched for that element under investigation. For instance, with respect to nitrogen, using
new modifications of the combustion system, it is now possible to perform CSIA on nitrogenous compounds
(Macko et a/., 1997). In this case,  the heavy isotope is 15N and the mass spectrometer measures the ion current
intensities of masses 29 and 28, representingthe two most common forms of molecular nitrogen.
With  a  state-of-the-art  instrument,  reproducibility  is
measurements typically better than 0.5 %o (Tables 1-3).
       quite  good,  with  standard  deviations  on  replicate
Table 1. Average 8 13C values (relative to PDB  reference) with  associated standard deviations  for five PAH
standards.
Compound
Naphthalene
Phenanthrene
Anthracene
Chrysene
2,3,6 - Trimethylnapthalene
Averaae 8 13C value
-26.2
-24.4
-23.7
-23.9
-22.6
Standard Deviation (+/-)
0.4
0.45
0.44
0.58
0.20
Table 2. Average 8 13C values (relative to PDB reference) with associated standard deviations for four alkane
standards.
Compound
Decane
Undecare
Dodecane
Tridecane
Averaqe 8 13C value
-29.0
-27.1
-33.3
-32.6
Standard Deviation
0.30
0.30
0.30
0.40
Also, with a state-of-the-art instrument, performing CSIA with a GC/C/IRMS system has many advantages over
other techniques for determining contaminant identities and sources. Because a GC is coupled to the mass
spectrometer, there is the capability of resolving complex mixtures. Also, with such a system, there is very high
                                                   104

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


sensitivity, which can  be a  very important consideration when dealing with trace contaminants. With  a
GC/C/IRMS system, detection is possible in the sub nM region.

Table 3. Average 6 15N values (relative to atmospheric N2 reference) with associated standard deviations for four
nitrogen-containing compounds.
Compound
Pyrazine
Tripropylamine
Quinoxaline
Nicotine
Averaae 8 15N value
0.7
9.5
-0.6
-1.7
Standard Deviation
0.12
0.55
0.16
0.41
The Approach
In applying CSIA to the problem of tracing environmental contaminants, measures must be taken to determine
whether the techniques are viable  under natural conditions. As  was mentioned above, using GC/IRMS to
measure isotopic compositions for the purpose of tracing contaminants seems to be an exciting alternative to
traditional methods. The first steps taken toward application of the technique must be field tests, however. This
can be stated in a series of objectives:

1)  Establish extraction, isolation, and purification techniques for the contaminants of interest. Of particular
   importance is obtaining high purity samples that have not been altered in isotopic composition.

2)  Determine the stable isotopic composition of the compounds of interest obtained from primary sources, such
   as manufacturers, petroleum suppliers and chemical storage locations.

3)  Determine the stable isotopic composition of the contaminants at locations where they are introduced into
   the environment. Possible locations include combustion areas (automobile exhaust and industrial exhaust),
   industrial emission, agricultural runoff, urban runoff and sewage effluent.

Once the first three objectives are  fulfilled,  the  next step is to apply them  to well-characterized sites where
pollution sources and times of occurrence are well known.  In this initial phase, carbon isotopes have been the
focus, and within this framework, early efforts have concentrated on the most  common PAH, OC pesticides and
PCBs (Table 4). The reasons for first analyzing the most common contaminants are two-fold. First, in trying to
determine the validity and  applicability of this technique, it  is undesirable to be sample-limited. From a sample
extraction and purification standpoint, it is necessary to deal with sufficient sample quantities. Once the method
proves successful,  samples of lower of concentration can be addressed. The second reason for focusing on the
most common contaminants is time  constraints. This approach is fairly time consuming, and thus only a limited
number of samples can be analyzed during the study. Choosing the most abundant  contaminants for analysis
should allow for the greatest chance of success.

Methods
In this study, the matrices of interest are soil and biological tissue.  Typically, the first step is the extraction
process. Soil samples are Soxhlet extracted with  methylene chloride  and then separated  based on compound
class with an alumina/silica column. Aliphatic hydrocarbons are eluted from the column with  50 ml of pentane.
Aromatic hydrocarbons, PCBs and OC pesticides are eluted with  150 ml of  pentane/methylene chloride  (1:1).
The second fraction is further separated on a  silica column. PCBs are eluted with an additional 90 ml of pentane.
High performance liquid chromatography (HPLC)  using a cyano/amino bonded phase  column  as per Killops and
Readman (1985) is used to purify the PCBs from fraction 2.  Pentane is used as the eluting solvent. Purity is then
checked with gas chromatography/electron capture detector (GC/ECD) and GC/MS systems.  Fraction two from
the silica column is separated into several subfractions, with the aromatic compounds being separated based on
the number of  double  bonds they contain (Killops and  Readman, 1985). A pentane and methylene chloride
gradient is used for the elution. It should be  noted that  others have shown that aromatic species can also be
separated with molecular sieves based on the arrangement of any alkyl substitutions present (Ellis et a/.,  1992;
1994). If need be, this is another viable alternative.

For the tissue samples, 20-30 grams of  wet tissue  are mixed with approximately 50  g of anhydrous  sodium
sulfate.  The extraction is  performed with three  100 ml methylene chloride  aliquots while macerating  with a
homogenizer. The  combined extract is then separated  as described above except there is  not an aliphatic
hydrocarbon component to contend with.  Before analyzing by GC/C/IRMS,  if the samples still  are not  pure


                                                  105

-------
                        MTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
enough, thin layer chromatography (TLC) can be utilized for further purification.
Table 4. List of contaminants that are initially being
investigated.

Contaminant
PAH
Pesticides
PCBs
                                                  Upon final separation  and purification, the analytes of
                                                  interest are taken up in an appropriate solvent (pentane
                                                  and methylene chloride, for example) to a concentration
                                                  in the 5-100 ng/ul range, depending  on the sensitivity of
                                                  the instrument for a particular compound class. Samples
                                                  are  then  analyzed  for  isotopic  composition  by  the
                                                  GC/C/IRMS system as described above.

                                                  Other Considerations
                                                  The stable isotopic composition  of a contaminant in the
                                                  environment  is the  end-result  of a  complex  chain  of
                                                  events. Naturally occurring  pollutants such as  spilled
                                                  crude  oil  are  most  easily traced  in  the  environment
                                                  because the starting  material is usually readily available
                                                  for analysis. Furthermore,  intrinsic tracers in spilled crude
                                                  oil would  also be  directly reflected in environmental
                                                  samples  since few  complicating   processes  would
                                                  intervene. Many synthetic materials  are  produced from
                                                  petrochemical   precursors   through   manufacturing
                                                  processes  such   as  distillation,   catalytic  cracking,
                                                  chlorination and polymerization to name a few. A change
                                                  in   isotopic  composition  might  occur   during   the
                                                  manufacturing process  if catalysis or high temperature is
                                                  involved.  Next the product is  applied  for  its  intended
                                                  purpose  which  could   include  combustion  (gasoline),
                                                  lubrication (lube oils), pesticide application (DDT), or use
                                                  as  a  transformer  oil (PCBs)  for example.  These
                                                  applications may cause an additional shift in  isotopic and
                                                  molecular composition. During  the  application, either
	       intentionally   or   indirectly,   some   portion   of   the
                                                  contaminant is released to the environment (such as soot
from combustion or the disposal of waste materials). Once released to the environment, the contaminant is then
subjected to  redistribution throughout various matrices including air, water,  sediments, and biological tissue
depending on its chemical properties and stability. The partitioning of the chemical among various phases might
be  accompanied by  a  shift  in  isotopic  composition  as  well  as  chemical  transformation.  Environmental
transformations are brought about by physical, chemical and microbiological processes. Each process defines an
independent set of isotopic and compositional changes. The complex history of a pollutant suggests that a
combination of compositional and stable isotopes can be linked to a specific series of events and processes
(Figure  2). Chemicals  of  identical structure may  have different isotopic composition if they have witnessed
different histories from  manufacture to environmental deposition. By analyzing contaminant  samples at different
stages of transport it should be possible to understand the dominant processes acting on them and to elucidate
the complex chain of events which led to the isotopic composition of the contaminants at their environmental end
point.

The issue of complex mixtures
As  was mentioned above, compound  specific isotopic analysis  using GC/C/IRMS has the potential to handle
complex mixtures.  It is inevitable, however, that during some analyses, there  are two or more chemical species
present in a mixture that are similar chemically and therefore coelute. Perhaps in some of the separated fractions
there could be additional compounds present other than  the desired analytes of interest that are similar to the
desired  compounds. If  there is only one other compound present, temperature programs for the  GC could be
altered, columns could be changed,  either in length or composition, or additional separation  steps  could be
added to alleviate this problem. When there is a high amount of background,  however, owing to more than one
additional chemical  species being present, the problem  is more difficult to handle. The coelution will lead to
erroneous isotopic results because what was thought to be the composition of  one compound is actually the
combined isotopic composition of two or more compounds. This issue is complicated further by the fact that
isotope fractionation occurs during elution, with portions of the peak varying dramatically in isotopic composition
Naphthalenes
Fluorenes
Phenanthrenes/Anthracenes
Dibenzothiophenes
Fluoranthenes/Pyrenes

Benzanthracenes
Benzofluoroanthenes
Benzopyrenes, Dibenzanthracenes

Perylene

Aldrin, Heptachlor, Endrin, Mirex
DDT, Dieldrin, Transnonachlor

Dichlorobiphenyls
Trichlorobiphenyls
Tetrachlorobiphenyls
Pentachlorobiphenyls
Hexachlorobiphenyls
Heptachlorobiphenyls
Octachlorobiphenyls
Nenachlorobiphenyls
                                                  106

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
(Figure 3). Although the isotope values can be averaged
across a peak, with coelution, the modifying effect of the
additional species can also vary across the peak  if the
analyte of  interest  and the  additional  species do not
elute at exactly the  same time. Some researchers have
used internal standards in their analyses to monitor the
degree to which unresolved complex mixtures (UCM)  or
other coelution problems affect the results of the isotopic
analyses. By knowing how much the  isotopic values  of
compounds   of  known  composition   have  shifted,
corrections can be made for the unknowns (Mansuy  et
a/.,  1997).  Other researchers  have  relied  on certain
separation techniques to try  to  keep the problem  to a
minimum (Ellis et a/.,  1994). Also,  within some software
packages,  monitoring  of  the background  is  possible,
along  with  the subsequent  subtraction  of  unwanted
contributions to a peak. Based on the above discussion,
it is obvious that this issue will require careful monitoring
in order to insure results are as accurate as possible.

Initial Applications
Extraction,  purification, and  analysis  on some of the
analytes of interest  has already been performed (Table
5). PAH mixtures from two sediment sites as well as a
sample  of  creosote  were   analyzed  by  GC/C/IRMS
yielding variable carbon isotope values. This variability
could  indicate different sources of the  PAH influence
into the various reservoirs.
Figure 2. Schematic representing possible pathways for
contaminants that enter the environment.
             Contaminant
                Source
          (Original Compounds)
                                    -manuf«clurins -
A S13C.
      AS'3C,T
Possible Transformation
(Biodegradation, other chemical
 reactions?)
           Continued Transport
                                                              Deposition in Environment
                                    A S
                                                            Soil
               Water
   Tissue
Atmosphere
Table 5. 8 13C values (relative to PDB reference) for PAH mixtures derived from three different sources.
Compound
Naphthalene
2-Methylnaphthalene
1-Methylnaphthalene
Biphenyl
2,6-Dimethylnaphthalne
Acenaphthalene
Fluorene
Phenanthrene
1 -Methylphenanthrene
Fluoranthene
Pyrene
Chrysene
Oreaon Inlet Sediment
-27.1
-27.8
-28.2
-28
-28.8
-27.4
-27.7
-25.6
-25.8
-24.6
-23.0
-23.0
Creosote
-23.5
-23.1
-21.2
-21.4


-18.4
-24.3

-25.2
-25.2

Casco Bay Sediment

-27.9
-29.5
-26.4








                                                   107

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                                            nf XC. / TRMS
               t.UOB-?
             o 1.120B-2-
               1.100B-2:
               1.080B-2:
                                                                                    > I
                      0    200    400   £00
                l.OOE-31
                8. COB-9-
                6.00E-9-
                4.00E -9-
                2.00E-*
                O.OOE+I)
Figure 3. Plot of the carbon isotope signal of three hydrocarbon species. The upper trace is the ratio of ion
current intensities  of  masses  45 and 44. The lower trace  is the ion current  intensity of  mass 44. Note the
changing isotope ratio across the peak.

Stable Isotopic Analysis as a Versatile Tool
In recent years, more attention has been focused on stable isotopic analysis,  and in recent years, on compound
specific isotopic analysis. While traditional molecular approaches are still in active  use today, stable isotope
analysis has gained popularity as a complement to other techniques or a stand-alone technique when other
approaches are  ineffective or too time consuming. Furthermore, as more attention is paid to stable isotopic
analysis, the versatility of the technique has become more evident. As was mentioned in the above discussion,
stable isotope analysis can be used as an effective tracer or method to apportion sources of compounds to a site.
GC/C/IRMS  has been used to  distinguish oils based  on the isotopic composition  of alkane  and isoprenoid
constituents present in the oil  (Bjoroy et a/.,  1991). This technique has been applied to oil spills as well;  even
under conditions where some of the oils have been weathered (Mansuy et a/.,  1997). With respect to the analytes
                                                   108

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


of interest in this study, CSIA has been used to not only trace the transport of PAH in aerosols (Ballentine et a/.,
1995) but to distinguish between biomass burning and fossil fuel burning sources of PAH (Ballentine et a/., 1996).
Also, sources of PAH  have been identified along with the relative amount of influence of each in marine soils.
(O'Malley  el a/.,  1994). With respect to  PCBs, it was  discovered that individual compounds tend to be more
depleted in 13C the more chlorine atoms there are present in the structure (Jarman et a/., 1998). Stable isotopic
analyses have also been used in groundwater studies, both as a source apportioning tool (Kelley et a/., 1997) and
also as a  means of tracing compounds during groundwater flow (Dempster et a/., 1997). Perhaps one  of the most
unique characteristics of stable isotope analysis is that is has been applied to many different scientific areas. Two
areas that are utilizing this approach more and more are biology and ecology. Characteristics of food  webs, from
trophic structure to influences on the diet on individual organisms have found application in stable isotope
analysis (Fang et a/., 1993 and Jarman et a/.,  1996). As has been shown, beyond the realm of just contaminant
studies, compound specific stable isotope analyses has  a wide range of applicability. This should  prove very
beneficial  as the call for more interdisciplinary science continues.

Conclusions
With the current heightened awareness  of the need to  understand and control contamination sites and sources,
the demand for powerful and accurate analytical tools to answer these questions is very high. Compound specific
isotopic analysis has  been suggested as a  new tool  to be used to answer these questions. Upon purifying
contaminant fractions, GC/C/IRMS  can  be  used  to  determine the  isotopic composition  of the individual
components in these fractions. These stable isotope compositions  can compared to the isotopic compositions of
different  source  materials and contaminants entering the  environment through various pathways. With this
information, the analyst can begin decipher together the history of the contaminants, linking the source and the
process of introduction to the contaminants of interest. Whereas  first only being applied to well characterized
sites, upon finding that the techniques are valid, the methodology can be applied to sites of unknown composition
and influence. Compound specific isotopic analysis could  then be used to complement other analytical methods
or as a stand-alone technique where the other methods  prove to be ineffective.

Works Cited
Ballentine, D.C., Macko,  S.A., Turekian, V.C., Gilhooly,  W.P. and  B. Martincigh. 1995. Transport  of biomass
   burning products through compound  specific isotope  analysis,  Selected Papers from the 17th International
   Meeting on Organic Geochemistry, 644-646.
Ballentine, D.C., Macko,  S.A., Turekian, V.C., Gilhooly, W.P  and B. Martincigh. 1996. Compound  specific
   isotope analysis of  fatty acids and polycyclic aromatic  hydrocarbons in aerosols: implications for biomass
   burning, Organic Geochemistry, 25, 97-104.
Bjoroy, M., Hall, K., Gillyon, P  and J. Jumeau. 1991. Carbon isotope variations in n-alkanes and isoprenoids of
   whole oils, Chemical Geology,  93, 13-20.
Dempster, H.S., Sherwood Lollar, B. and S. Feenstra. 1997.  Tracing organic contaminants in groundwater: a new
   methodology  using compound specific  isotopic  analysis,  EnvironmentalScience   and Technology,  31,
   3193-3197.
Ellis,  L,  Alexander, R.,  and  R.I. Kagi. 1994. Separation  of  petroleum hydrocarbons using dealuminated
   mordenite  molecular  sieve-ll.  Alkylnapthalenes  and  alkylyphenanthrenes,  Organic Geochemistry,  21,
   849-855.
Ellis,  L.,   Kagi,  R.I., and R.  Alexander. 1992. Separation  of  petroleum hydrocarbons using dealuminated
   mordenite molecular sieve. I. Monoaromatic hydrocarbons, Organic Geochemistry, 18, 587-593.
Fang, J., Abrajano, T.A., Comet, P.A., Brooks, J.M.,  Sassen, R. and I.R. MacDonald.  1993.  Gulf of Mexico
   hydrocarbon seep  communities XI.  Carbon isotopic  fractionation  during fatty acid  biosynthesis  of seep
   organisms and its implications for chemosynthetic processes, Chemical Geology, 109, 271-279.
Freedman, P.A., Gillyon,  E.G.P.  and E.J.  Jumeau.  1988. Design  and application of  a new instrument for
   GC-isotope ratio MS, American Laboratory, June, 114-119.
Mickey, C.W., Roper, D.S.,  Holland,  P.T. and T.M. Trower.  1995.  Accumulation of organic contaminants  in two
   sediment-dwelling shellfish  with contrasting feeding modes: deposit- (Macomona liliana)  and filter-feeding-
   (Austrovenus stutchburyi), Archives of Environmental Contamination and Toxicology, 29, 221-231.
Hites, R.A., LaFlamme, R.E., Windsor, Jr., J.G., Farrington, J.W., and W.G. Denser. 1980. Polycyclic aromatic
   hydrocarbons in  an anoxic sediment core from  the Pettaquamscutt River (Rhode  Island, USA), Geochim.
   Cosmochim. Acta, 44,  873-878.
Jarman, W.A., Hobson, K.A., Sydeman,  W.J., Bacon, C.E., and E.B. Mclaren. 1996. Influence of trophic position
  and feeding  location on contaminant levels in the  Gulf of Farallones food web revealed by stable isotope


                                                  109

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


   analysis, Environmental Science and Technology, 30, 654-660.
Jarman, W.M, Hilkert, A., Bacon, J.W.,  Ballschmoter, K. and R.W. Risebrough. 1998. Compound specific carbon
   isotopic analysis of Aroclors, Clophens, Kaneclors, and Phenoclors, Environmental Science and Technology,
   32, 833-836.
Kelley, C.A.,  Hammer, B.T. and R.B. Coffin.  1997.  Concentrations and stable isotope values of BTEX in
   gasoline-contaminated groundwater, Environmental Science and Technology, 31, 2469-2472.
Killops, S.D. and V.J. Killops. 1993.  An introduction to organic geochemistry,  Longman Scientific and Technical,
   265 pp.
Killops, S.D. and J.W.  Readman. 1985. HPLC fractionation and GC-MS determination of aromatic hydrocarbons
   from oils and sediments, Organic Geochemistry, 8, 247-257.
LaFlamme, R.E. and R.A. Hites. 1978. The global distribution of polycyclic aromatic hydrocarbons in recent
   sediments, Geochim. Cosmochim. Acta, 42, 289-303.
Law, R.J. and J.L. Biscaya. 1994. Polycyclic aromatic hydrocarbons (PAH)-problems and processes in sampling,
   analysis and interpretation, Marine Pollution Bulletin, 29, 235-241.
Macko, S.A., Uhle, M.E., Engel, M.H. and V. Andrusevich. 1997. Stable nitrogen isotope analysis of amino acid
   enantiomers by gas chromatography/combustion/isotope ratio mass spectrometry, Analytical Chemistry, 69,
   926-929.
Mansuy, L, Philp, R.P. and J. Allen. 1997. Source identification of oil spills based on the isotopic composition of
   individual components in weathered  oil samples, Environmental Science and Technology, 31, 3417-3425.
O'Malley, V.P.,  Abrajano, Jr., T.A.,  and J. Hellou. 1994. Determination of the 13C/12C ratios of individual  PAH
   from environmental samples: can PAH sources be apportioned?, Organic Geochemistry, 21, 809-822.
Roper, J.M., Cherry, D.S., Simmers, J.W. and H.E. Tatem. 1997. Bioaccumulation of PAHs in the Zebra mussel
   at Times Beach, Buffalo, New York, Environmental Monitoring and Assessment, 46, 267-277.
            RECENT DEVELOPMENTS IN IMMUNOBIOSENSORS & RELATED TECHNIQUES
                      FOR THE DETECTION OF ENVIRONMENTAL POLLUTANTS

                                  M. Masila, H. Xu, E. Lee, O.A. Sadik*
                   Department of Chemistry, State University of New York at Binghamton,
                              P.O. Box 6016, Binghamton, New York 13902
                                         Fax: (607)777-4478
                                    E-mail: osadik@binghamton.edu

ABSTRACT
This paper describes inimunobiosensors and other related multianalyte detection  methods of identification and
quantitation of various environmental compounds and metabolites. It presents the synthesis and characterization
of polymers, biological  conjugates and metabolites  for the detection  of  a range  of toxic chemicals, including
chlorinated phenols, s-triazine  herbicides, polychlorinated biphenyls,  and heavy  metals.  A new multi-analyle
detection technique  utilizing polypyrrole derivatives was developed for the detection of chlorinated phenols and
other organics. The detection of polychlorinated biphenyls (PCBs), volatile and semivolatile, halogenated organic
compounds of environmental interest was conducted using the new polymer sensors. A new promising approach
for heavy metal detection is described that utilizes o-hydroxypyridylazo metal-protein conjugates which may also
be used in the development of nonradioactive  immunosensor labels  for  environmental compliance  monitoring
and clinical applications.

INTRODUCTION
Very few analytical  methods for environmental monitoring that are fast,  low-cost and continuous are currently
available.  The monitoring of residue or contamination in soil, water and air can be classified into two main
categories. These are:  (i) screening  or diagnostic techniques in which only  a yes-or-no (qualitative) answer is
required, and (ii) semi-quantitative or quantitative techniques in which  the detection of unwanted chemicals, and
the testing  of whether or not the residues of the contaminants are within permissible levels are required. It is
possible for the former  methods to generate false positive or negative results if the sensitivities are  insufficient
for the detection of the threshold levels.
                                                 110

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


The  successful coupling of suitable transducers (e.g.  electrochemical,  optical  or mass) with biomolecular
systems is of great importance in the search for novel sensing technologies that are inexpensive, highly selective
and  capable  of  generating  early warning signals in the presence  of toxic environmental chemicals. The
emergence of on-site chemical and  immunochemical biosensors for environmental pollutant monitoring has
tremendous potentials due to their small size, low costs and the case of analytical signal generation in real time.
These sensors represent a step forward over the conventional laboratory analytical methods.

In recent  years,  biosensors; that  can  detect  a range  of  analytes  in  environmental  samples  have  been
developed12 Basically, a sensor consists of a chemically selective layer, a transducer, and a signal processor. If
the selective layer utilizes a biological or biochemical species, then it can be classified as a biosensor. Thus, an
immunosensor is a subset of biosensor since it comprises either an antibody or an antigen. Each sensor has a
number of desirable characteristics depending on its applications. Essentially, a practical biosensor  for the
monitoring of environmental pollutants must be  specific, reversible, able to provide fast response time, and
capable of direct detection of an immunoreaction  with minimal sequential addition of immunoreagents. Also, the
sensor should be capable of continuous flow measurements and capable of determining multiple analytes in
complex samples with little or no  need for sample preparation steps. Finally, the sensor must be able to process
signals, or capable of being integrated  into other devices that can exercise real-time feedback as required for
pollution monitoring or surveillance studies. Although, a number of pollutant measurement techniques have been
reported, only few possess these specific requirements.

One of the major objectives of our research is to develop field-portable sensors that meet or exceed the above
sensor requirements for use in  the assessment of toxic chemical residues in various environmental media. This
paper discusses sensors developed in our laboratory for the identification and quantitation of environmental
contaminants.

MATERIALS  AND METHODS
Instrumentation
The  following  instruments were used to conduct the experiments described in this paper: A Hewlett-Packard
Diode-array UVA/is spectrophotometer was used  for the characterization of all protein conjugates. ELX 800 UV
Plate Reader (from  Bio-Tek Instruments) was used for all of the  enzyme-linked immunosorbent assay (ELISA)
experiments.  EG&G PAR potentiostat/galvanostat Model 263A  and EG&G 270 software were employed for the
electrochemical experiments with silver/silver chloride reference electrode, platinum wire counter electrode and
gold (A =  0.2 cm2) as working electrode. Quartz crystal microbalance (QCM) measurements were carried out
using EG&G  quartz crystal analyzer (Model  QCA917). A 9MHz EG&G At-cut quartz crystals was sandwiched
between two gold electrodes (A = 0.186 cm2). AromaScanner Model A32S (from AromaScan, Inc.,  NH) was used
for the multiarray electronic nose  experiments.

Sensor Preparation and Characterization
Immobilization on Quartz: The Au-coated quartz  crystal was initially pretreated by cycling the potential between
1.4 and 0.0V for a minimum of 15 minutes in  0.2M perchloric acid. The cell was then rinsed with copious  amount
of water, and one surface of  the  crystal  was soaked  in a  0.02 M cystamine solution. The Au surface was
thoroughly  rinsed with  water  to remove  any physically adsorbed cystamine before  being soaked  in 3mM
cyanazine hapten solution containing 0.01 M HEPES (N-[2-Hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid)
buffer solution (pH = 7.3) using 10mM 1-ethyl-3-(3-dimethylamino-propyl) cabodimide EDO coupling reagent.

Electrochemical  Immobilization: The Au electrode was pretreated as described above before being modified with
the  cyanazine hapten  using  EDC  as  the coupling  reagent.  The   modified  electrode was  used   in the
electrochemical analysis, first without soaking in an antibody solution.  Later the electrode was incubated in an
anti-cyanazine antibody solution  at 35°C using a thermostated water-bath. All cyclic voltammetry experiments
were conducted at the same temperature. The other electrochemical immobilization procedures were as  recently
reported34

Polymer Synthesis:  Various pyrrole derivatives  were polymerized  by  electrochemical  oxidation to  enable  the
conducting polymer films to be used for conductivity, electrochemical, and mass measurements. Some selective
electrodes for phenols, PCBs and s-triazines were prepared by the electropolymerization of pyrrole onto platinum
electrodes in the presence of tetrabutyl ammonium perchlorate. The selectivities were comparable to a range of
structurally   similar   organic   compounds,   including  2,3,5,6-tetra   chloroanisole,   2,3,4-trichloroanisole,
2-chloroanisole, 2,4,5-trichlorophenol, simazine, cyanazine, and substituted benzenes.


                                                  111

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Antibody Production: The design and preparation of analyte analogs and immunogens are essential steps in the
development of low molecular weight immunosensors. Triazine analogs were obtained from Dr. J.R. Fleeker's
laboratory. These were prepared from active esters of the carboxylic acid analog of the triazine haptens using
N-hydroxysuccinamide5  The triazines were coupled to a high  molecular weight carrier, bovine serum albumin
(BSA), or keyhole limpet hemocyanin (KLH) which were used for the production of the antibodies. The antibodies
were purified by gel filtration and protein-A immunoaffinity columns, and were subsequently characterized using
ELISA and nuclear magnetic resonance (NMR) techniques. By using these antibodies, sensors for s-triazine were
developed  based  on the  antibody inhibition of the  current generated by  the  ferricyanide  mediator on
antigen-immobilized gold electrodes.

Pesticide Immunosensors
Several pesticides and herbicides are routinely used to improve crop harvesting and  pest-control. Due to the
growing concern about health effects, several investigations have been conducted  in order to understand how
pesticides and herbicides degrade in the environment67  Current methods of monitoring pesticides include liquid
chromatography  and gas chromatography with mass spectrometry. The high costs and labor involved in the
chromatographic methods have led to the  search for low-cost alternatives capable of providing rapid analysis. In
this paper, we report on  the development  of  immunosensors for atrazine,  cyanazine,  simazine and  their
metabolites. The sensor chemistries are shown in Scheme 1  below:
                                            NH-C(CH3)3
                    x-^NH2 HEPES buffer, °-°' M. pH=7.4 * EDC


                Au                                       Au
                                                           s-VN-C
!TpI
                Scheme 1. The assembly of a cyanazine hapten monolayer on Au electrode.

The  hapten  monolayer electrode sensor assembly was used in  the detection  of cyanazine in a flow injection
analysis mode. The interaction of the electrode with different antibody concentrations resulted in the formation of
an antibody-antigen (Ab-Ag) complex which insulated the electrode towards  the [Fe(CN)6]47Fe(CN)6]3"  redox
probe, hence resulting in no charge transfer. The extent of the insulation depended on the antibody concentration
and the time of exposure to the antibody solution. The decrease  in the amperometric response of the antigenic
monolayer to corresponding antibody solution  for a fixed time  produced a quantitative  measurement of the
antibody concentration (Figure 1). Typical voltammetric responses obtained for the cyanazine hapten monolayer
electrode to  different antibody concentrations are shown in Figure 2. The lowest detection  limit achieved for the
cyanazine sensor was 4.0 ug/ml with a response time of a few minutes and a less-than 2% cross-reactivity to
atrazine, simazine and other metabolites.

Multianalyte  Sensors
The  presence of halogenated organic compounds  in the environment has posed  a great concern due to their
persistent  toxicity and the ability to bioaccumulate.  Of all the 19 known chlorinated phenols, the most important
congeners include the 2,4-Dinitrophenol (2,4-D), 2,4,5-trichlorophenol, (2,4,5-TCP) and pentachlorophenol. While
these compounds  can be  determined using mass spectrometry  and  gas chromatographic  techniques, the
structural similarities of substituted phenols and their derivatives enable the development of a rapid, multianalyte
method rather than for one or two  analytes. The combination  of gas  sensor arrays  and pattern recognition
techniques has  resulted in a fast and objective method for the simultaneous measurement of a  wide range of
volative and  semi-volative organics.

A  32-array  conducting polymer sensor was used for the rapid  measurement  of  volatile  and semivolatile
halogenated organic compounds of environmental  interest. The  mathematical  expressions for the  microscopic
polymer network  model  was described in  a recent  article4. A classical,  nonparametric, and unsupervised
technique  of cluster analysis was used  to  discriminate  between  the  polychlorinated organic  phenol  vapor


                                                  112

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


response vectors in a 2-dimensional space, and to identify clusters, or groups, to which unknown vectors were
likely to belong. Consequently, the characteristic pattern for each sample was generated. The pattern was used
to generate the database employed in the determination of the Euclidean distances between two given patterns
and the  normalized sensor response.  Also,  this was used to develop the  2-dimensional  mapping  from a
multi-dimensional space to quantify the distinctions of the samples.

The resulting sensor arrays were found to recognize small molecules on the basis of their chemical structures
which were related to the nature of the  chemical class, the type and the position of the  functional groups.  Each
sensor responded in varying degrees to chlorinated organic molecules with standard deviation of less than  0.05.
The time averages  for the sensor response databases, datamaps, response patterns, and  the intensity  profiles
were  obtained for  different phenols.  Tables  2 and  3 showed the representative  databases obtained  for
2,4,6-trichlorophenol and 2-CP created from the raw sample data files by selecting sample data between 60 and
120 sec. The limit  of detection obtained  for 2,4,6-trichlorophenol  and 2-chlorophenols using the conducting
polymer sensor array were 0.1 and 0.25 ng/mL respectively. This results demonstrated the viability of conducting
polymer  sensor  arrays  for the identification  and quantitation of chlorinated  organic  phenols  based  on the
differences in their  Euclidean  distances.  The qualitative  differences as defined by  the Euclidean distance
measurements were most clearly visible when the nature and the type of the functional groups were considered.

Direct Electrochemical Sensors for Polychlorinated Biphenyls (PCBs)
A direct electrochemical immunosensor has been developed for the determination of PCBs in water.  The assay
was  based on the  measurement of the current  due to the specific binding between  PCB  and  anti-PCB
antibody-immobilized conducting polymer matrix. The linear dynamic  range of  the immunosensor was between
0.3-100 ng/mL with a correlation coefficient of 0.997 for Aroclor 1242. A typical flow injection analysis signal
obtained for Aroclor 1254 is shown in Figure  3. Well defined responses were  recorded  for all  aroclors. The
method detection  limits for Aroclors  1242, 1248,  1254  and 1016  were  3.3, 1.56,  0.39,  and  1.66 ng/mL
respectively, and a signal-to-noise (S/N) ratio of 3. The immunosensor exhibited high selectivity for PCBs in the
presence  of potential  interference  such  as  chlorinated anisoles,  benzenes and  phenols.  The   highest
cross-reactivity measured for chlorinated  phenolic compounds relative to Aroclor 1248  was less than 3%. The
recoveries  of spiked Aroclors  1242  and 1254 from industrial  effluent water, rolling  mill and  seafood  plant
pretreated water at  0.5 and 50  ng/mL ranged from 103-106%. The  effect of ionic compounds on the detection
indicated that no significant change in  immunosensor  signal was observed within the uncertainty of the assay
procedure.  The detection method can be used  for continuous monitoring of effluent such as waste streams and
ground water.

Rational Design of Immunosensors: Sensors for Heavy Metals
In order to  increase  the  sensitivity of  immunosensing  methods, a rational design of sensors using
o-hydroxypyridylazo  compounds was explored. The two most  important of these compounds employed  were
1-(2-pyridylazo)-2-naphthol  (PAN) and  4-2-pyridylazo resorcinol (PAR). Both PAR and PAN have been  used
extensively for the analysis of metals, and they  posses lots of useful spectroscopic and luminescence properties.

The use of  2-pyridylazo compounds as precursors for the preparation of protein conjugate by coupling the ligand
to BSA, KLH, and ovalbumin was considered.  It was anticipated that  using these conjugates would lead to the
development  of  new biosensing  chemistries  and transduction principles.  Ultimately,  any  protein conjugate
developed  may  become  useful  in  developing novel non-radioactive  molecular labels for  immunoassay,
molecular labeling  and environmental compliance monitoring applications. The metal-chelate conjugates  were
tested to determine if the system was simpler and rapid for the identification and quantitation of lead and  other
heavy metals.

Finally,  the  PAR-lead-BSA,   PAR-lead-Ovalbumin,  and  PAR-lead-alkaline  phosphatase enzymes  were
successfully designed   and synthesized8  These conjugates  were  characterized  using  UV/Vis, intra-red
spectroscopy, NMR, and electrochemical techniques. Figure 3 shows the absorption spectrum obtained for the
coupling  of PAR (510 nm) and BSA (280  mn) conjugate.  A  preliminary test of the PAR-conjugates and the
detection of lead and mercury were conducted  using optical, differential pulse voltammetry and anodic stripping
voltammetry  techniques.  The  binding strategies employed  include  a  sandwich configuration  using the
synthesized PAR-protein conjugates.

Pb2+ binding was monitored by recording the change in the cathodic reduction of the ion and  the absorbance of
the lead-PAR chelate. The  binding affinity was controlled by an electro-optical technique  which influenced the


                                                  113

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
PAR-chelate's  geometry. However, it is our understanding that no  published literature exists on  protein that
utilizes  PAR metal-ion  binding  and the  chromophoric  PAR-protein  conjugates.  Using  the electro-optical
techniques  it  became possible to quantitatively determine mercury  and lead at approximately 3x1 Q-7 and
1x10*M respectively.  Thus, a good knowledge of the selective interaction of the conjugates with  biological
macromolecules may result in new applications of metal-ion immunoselective adsorbents.

CONCLUSIONS                                                                  .  ,  ,
We have developed various sensors for the detection of pesticides, PCBs and heavy metals for environmental
monitoring Other sensors developed and  their analytical characteristics are summanzed in Table  3. The new
metal-chelate protein conjugates reported in this manuscript could lead to the development of new immunoassay
formats and instrumentation capable of providing rapid and selective environmental monitoring. Details  of the
design synthesis and characterization of the conjugates and their analytical applications for metal detection are
being compiled for journal  publication. This  work demonstrates that  new and  promising applications  of the
chemical and immunobiosensors and the emerging immunoassay labels will continue to make immunochemical
methods more valuable to environmental monitoring.

ACKNOWLEDGMENTS                                                                ^^ AJ
The  authors would like to acknowledge the following: Office of Research & Development,  NCERQ Advance
Monitoring Program of the United States Environmental Protection Agency for funding. AromaScan Inc., NH, for
loaning their AS  32A Electronic Nose instrument used in the  evaluation of the  chlorinated organic sensors
developed in this project, and Dr. James R. Fleeker of North Dakota State University for the s-triazine analogs
used in the pesticide sensors.

REFERENCES
1.  Sadik O.A., Van Emon J.M., Designing Immunosensors for Environmental Monitoring, ChemTech, Vol. 27
    No. 6, pp. 39 - 46, June 1997.
2.  Sadik, O.A., Van Emon J.M., Biosensors & Bioelectronics, Vol 11 (8), AR1, 1996.
3.  Bender Sharin, Omowunmi  Sadik, "Direct Electrochemical Immunosensor for Polychlorinated  Biphenyls
    (PCBs)," Environmental Science & Technology, Vol. 32, No.6, pp. 788-797, 1998.
4.  M. Masila, A. Sargent, F Van, O.A. Sadik, "Pattern Recognition Studies of Chlorinated Organic Compounds
    Using Polymer Sensor Arrays," Electroanalysis, Vol 10, No.4, pp.  1-9, 1998.
 5.  Lawruk T.S., Hottenstein  C.S., Fleeker J.R., Rubio  F.M., Herzog D.P., ACS  Symposium Series No. 630,
    Herbicide Metabolites in Surface Water and  Groundwater,  (M.T.,  Meyer and  E.M., Thurman  - Editors) pp
    43-52,1996.
 6.   Roy-Keith Smith, Handbook of Environmental Analysis, 2nd Edition, Genium Publishing Corp., USA, 1995.
 7.   Barnett D., Laing D.G., Skopec, S., Sadik O.A., Wallace G.G., Analytical Letters, 27(13), 2417, 1994.
 8.   Hongwu Xu, E. Lee, S. Bender, O.A. Sadik, "Immunosensors Based on Metal-Chelates for Monitoring  Heavy
     Metals," PittCon 98, New Orleans, Louisiana, March 1-5, 1998, paper #1244.
 9.   Sadik O.A., Wallace G.G., Anal. Chim. Acta,  279, 209, 1993
 10.  Sadik O.A., John M. J., Wallace G.G., Barnett D., Clarke C., Laing  D.,  Analyst,  119, 1997,  1994.
 11.  Anita  Sargent, Omowunmi  Sadik, "Pulsed  Electrochemical Technique for  Monitoring Antibody-Antigen
     Reactions at Interfaces," Trends in Analytical Chemistry, 1998 (In Press).
 12.  Sadik O.A., Wallace G.G., Electroanalysis, 5, 555, 1993.
                                                  Predicted Y
                 10      20      30

                   Ab cone. (uL/mL)
40
Figure  1. Cathodic  peak current  versus
antibody   concentration   at   constant
incubation time of 5 minutes.
                                                  114

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1. Representative database obtained for 2,4,6-Trichlorophenol at conducting polymer sensor arrays
Sensor
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Intensity
-2.560
-2.450
-2.710
-2.700
-2.540
-2.460
-2.730
-2.600
-2.490
-2.700
-2.320
-2.450
-2.310
-2.720
-2.310
-2.260
-2.150
-2.500
-2.720
-2.790
-2.570
-2.170
-2.300
-2.120
-2.070
-2.440
-2.520
-2.350
-2.430
-2.700
-2.810
-2.570
Pattern
-3.219
-3.079
-3.407
-3.392
-3.198
-3.096
-3.428
-3.269
-3.132
-3.393
-2.920
-3.080
-2.905
-3.426
-2.902
-2.839
-2.709
-3.145
-3.424
-3.508
-3.238
-2.728
-2.892
-2.659
-2.602
-3.068
-8.175
-2.957
-3.053
-3.398
-3.530
-3.229
SD
0.038
0.003
0.021
0.024
0.037
0.033
0.009
0.013
0.005
0.004
0.014
0.008
0.017
0.007
0.041
0.036
0.070
0.013
0.030
0.006
0.056
0.110
0.109
0.194
0.010
0.012
0.028
0.009
0.039
0.037
0.029
0.054
SD = Standard deviation, experimental conditions.
30 min.
  relative humidity 50%, temperature 25°C, equilibration time
                                                             Figure 2.  Voltammetric responses obtained
                                                             for 1.1 mM K4Fe(CN)6 with cyanazine hapten
                                                             monolayer    electrode    using   different
                                                             concentrations of anti-cyanazine antibody at
                                                             5-minute incubation time interval, (a) Blank
                                                             (phosphate buffer), (b) 5ML/ml; (c) 15 |jl_/ml;
                                                             (d) 30gL/ml; (e) 40ul_/ml.
     -400      -200        0       200
                   E (mV) vs. Ag/AgCI
400
600
                                                  115

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 2. Representative database obtained for 2-Chlorophenol at conducting polymer sensor arrays
Sensor
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Intensity
-0.21
-0.08
-0.15
-0.15
-0.17
-0.18
-0.08
-0.12
0.11
-0.05
-0.08
-0.09
0
-0.09
-0.17
-0.18
0.32
0.46
0.02
0.36
-0.24
0.6
0.73
0.81
-0.04
-0.03
-0.11
-0.01
-0.17
-0.11
-0.09
-0.24
Pattern
-3.63
1.26
-2.58
-2.59
-2.94
-3.1
-1.56
-2.06
-1.66
-0.97
-1.36
-1.56
0.1
-1.62
-3.09
-3.16
4.77
6.05
0.43
5.44
-4.31
8.45
10.09
11.67
-0.59
-0.64
-1.96
-0.15
-3
-2.07
-1.65
-4.22
SD
1.11
0.51
0.85
0.8
0.86
0.99
0.58
0.59
0.66
0.62
0.37
0.5
0:090
0.55
1.03
1.04
1.03
4.58
0.1
1.51
1.49
3.39
4.38
3.26
0.16
0.42
0.61
0.14
0.9
0.72
0.47
1.45
SD = Standard deviation, experimental conditions: relative humidity 50%, temperature 25°C, equilibration time
30 min.
Table 3. Immunobiosensors & Chemical Sensors Developed
Compound
Chlorinated
Phenols
PCBs
Cyanazine
Heavy metals
Cyanide
Anions
Atrazine
HSA
P-Cresol

P-Athau
* MDL = Method
MDL
0.25 ug/ml

0.05 ng/ml
4.0 ug/ml

low ppt
3X10-8M
4.0 ng/ml
3X10'7M
0.5 mg/l

0.01 mg/ml
Detection Limit. The MDL
Detection TechniquelRemarks
AR

FIA mode
A[Ab]
absorbance measurement
Fluorescence
Current
A[Ab]
Regenerate
Dynamic range 3 orgers of magnitude,
reusable
Regenerable
was computed using MDL = t(n-ri^=o99i *
value, S = standard deviation of the replicate analyses.
b Currently under
investigation.

Ref.
2

1-3
b
b
b
12
b
9
7

10
S, where t = the students t


                                                   116

-------
                      WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                            Figure 3. UVA/is  Spectrum  recorded for the coupling of
                                            BSA (labeled A at 280 nm) and PAR (labeled B at 510 nm)
                                            conjugate.
                       MULTIPLEXED DIODE LASER GAS SENSOR SYSTEM
                     FOR IN-SITU MULTI-SPECIES EMISIONS MEASUREMENTS

                                           R. Hanson

                                   NO ABSTRACT AVAILABLE
                     OVERVIEW/FUTURE OF NCERQA RESEARCH PROGRAM

                                          Bala Krishnan
              US EPA, Office of Research and Development, 401 M St, SW, Washington, DC

                                   WO ABSTRACT AVAILABLE
              ADVANCED ANALYTICAL METHODS FOR THE DIRECT QUANTIFICATION
           AND CHARACTERIZATION OF AMBIENT METAL SPECIES IN NATURAL WATERS

                                  Janet G. Hering and Joon H. Min
                  California Institute of Technology, Environmental Engineering Science,
                        1200 E. California Blvd. (138-78), Pasadena, CA 91125

Neither the biogeochemical  cycling of trace metals nor their ecotoxicological effects can be fully understood
without careful  consideration of  metal speciation. Investigations of metal speciation  in natural waters have
demonstrated that,  for many metals (particularly Cu, Zn, and Fe), the  predominant ambient metal species are
nonreactive  in chemical and/or biological assays. However, the methods commonly used  to  monitor metal
speciation cannot provide definitive information on the nature of ambient metal species.

Electrospray-mass spectrometry (ESMS) offers a powerful tool for the investigation of ambient  metal species.
Unlike  more common  GC-MS  techniques,  ESMS  can  be applied  to  non-volatilizable  species  and  the
comparatively gentle ionization in ESMS allows compounds to  be analyzed with minimal fragmentation thus
                                              117

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


preserving the molecular ion signature.

Preliminary studies with model compounds  have been performed to investigate the  application  of ESMS for
analysis  of  metal-organic  complexes. Work  to  date with  the  strong  organic complexing  agent  EDTA
(ethylenediaminetetraacetic acid)  has demonstrated that both uncomplexed EDTA and metal-EDTA complexes
can be detected in  the positive  ion mode as protonated  species with a  single positive charge. Despite the
extreme conditions imposed by the electrospray interface, protonation (and, in the case of uncomplexed EDTA,
formation of the Na+ adduct) appears to  be the only perturbation of the  initial distribution of EDTA species.
Molecular ions were detected for EDTA and its complexes with Cu,  Pb, Cd,  Al, and Fe(lll). In the case of the
Pb-EDTA complex, the  isotopic signature of Pb was also observed. Based on this work, the application of ESMS
for the detection of ambient metal species  in natural waters appears promising but is limited by the sensitivity of
the  technique and by  variable  response  to different compounds.  Detection limits for EDTA and  its metal
complexes are  approximately micromolar though better sensitivity has been reported for other metal-organic
complexes.

Work is in progress to characterize the  ESMS response to organic ligands of varying structures (and their metal
complexes) and to improve the sensitivity of the technique. The instrument currently in  use is a Hewlett Packard
Series 1100 LC/MSD; adjustable instrumental parameters include drying gas flow rate and temperature, capillary
and fragmentor voltages, gain and nebulizer pressure. The effects of the sample matrix (e.g., pH and methanol
concentration) on sensitivity will also be tested for selected model compounds. Model compounds have been
selected  to  include  a  range  structural  characteristics  including:  type  of heteroatom(s) and  complexing
functionalities, metal-ligand stoichiometry,  and ligand  charge and hydrophobicity. Preliminary, screening studies
will be performed to examine ESMS response to reference humic and fulvic  acids.
                                  RADICAL BALANCE IN URBAN AIR

                                          Robert J. O'Brien
                                        Chemistry Department
                                           Linda A. George
                                     Center for Science Education
                                          Thomas M. Hard
           Chemistry Department, Portland State University, PO Box 751, Portland, Oregon 97207

Atmospheric free  radicals hydroxyl and hydroperoxyl (OH and HO2,  collectively  HOX) are the catalysts which
cause secondary or photochemical air pollution. Chemical mechanisms for oxidant and acid formation, on which
expensive air pollution control strategies are  based, must accurately predict these radical concentrations. We
used the PAGE technique to carry out the first  simultaneous, in-situ, measurements of these two radicals in
highly polluted air at downwind sites in the Los Angeles airshed.

To  compare  the  measured OH  and  HO2 concentrations  with  photochemical  models, a complete  suite of
simultaneous ancillary measurements was necessary, and was obtained during each measurement campaign.
The suite included speciated hydrocarbons, carbonyl  compounds,  carbon monoxide,  nitric oxide,  nitrogen
dioxide, ozone,  and meteorological  parameters. With  this  suite as input, we tested the ability of a  lumped
chemical mechanism to accurately predict the  measured OH and HO2 radical  concentrations.

Due to the short photochemical  lifetime of HO*  (less than  1  minute), this test of radical balance in urban air
depends  directly and quantitatively on the measured parents and  reaction  partners of the  radicals, and  only
indirectly on the upstream history of the sample.

Results of the measurements,  and of the radical balance  tests, will  be presented, with acknowledgments to the
organizations and  scientists who provided assistance.
                                                 118

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


   ENVIRONMENTAL APPLICATIONS OF NOVEL INSTRUMENTATION FOR MEASUREMENT OF LEAD
        ISOTOPE RATIOS IN ATMOSPHERIC POLLUTION SOURCE APPORTIONMENT RESULTS

                                               Keeler

                                     NO ABSTRACT AVAILABLE
                     REMOTE SAMPLING PROBE WITH FAST GC/MS ANALYSIS:
                  SUBSURFACE DETECTION OF ENVIRONMENTAL CONTAMINANTS

                                          Albert Robbat. Jr.
                  Center for Field Analytical Studies & Technology, Chemistry Department,
                            Tufts University, Medford, Massachusetts 02155

ABSTRACT
This paper describes the results of an in situ sampling probe that is capable of thermally extracting volatile and
semivolatile organics  bound to soil from depths of 30 ft. The organic vapor is swept to the surface by an inert
carrier gas and trapped (volatiles) or condensed (semivolatiles) in appropriate  sampling tubes. The organics are
subsequently thermally desorbed into  a gas chromatograph/mass spectrometer and analyzed in under 5 minutes.

INTRODUCTION
The EPA estimates that  the cost for hazardous waste site cleanups will exceed $300 billion over the next 10
years,1 with the cost  for Superfund alone exceeding $26 billion since  1980,  The  following questions can be
posed: Do inadequate site investigations and, therefore, a lack of understanding with respect to the chemical and
physical dynamics  affecting the  cleanup contribute to these costs? Can field-based analytical instrumentation
and methods give on-site project engineers the kind of data needed that will lead to faster, better, and cheaper
cleanups?

Toward this end, research is leading to  the development technology and methods that can produce quantitative
analysis  of  environmental contaminants  in  minutes  by thermal  desorption  gas  cnromatography/mass
spectrometry (TDGC/MS).  The analysis is based  on a ballistically heated thermal desorber to achieve  large
volume sample  introduction and mass spectral  data analysis algorithms  that can "look through" complex matrix
signals to identify and quantify target  compounds.2 The TDGC/MS when  used  in a dynamic workplan framework
can provide data fast enough  to influence the  on-site decision  making  process.3 We have shown that on-site
chemical analyses employing dynamic workplans can reduce  the time and cost of hazardous waste site
investigations.4

Cone Penetrometer (CP) systems can  collect  samples at much faster rates than  can  traditional drilling rigs.
Figure 1  depicts the sampling probe used to collect subsurface soil and water samples. Typically, 5 cm o.d. pipes
are threaded together and pushed underground by truck weights of up to 40 tons. The challenge therefore, is to
design 1) a  flexible heated, 300 °C, transfer line that can  be woven through each pipe section  and 2)  a
programmable thermal extraction sample collection probe that can heat the soil to at least  350 °C. These target
temperatures  are based an past studies aimed at developing direct measuring  thermal desorption gas
chromatography  (TDGC)  sample introduction  system.56789  The  design  of a  thermal  extraction  cone
penetrometer (TECP) system  for subsurface sampling of soil-bound organics from depths of up to 25 m  is
described in this paper as well  as a new data analysis system that provides unique compound identification and
quantification capabilities under fast TDGC/MS conditions.

MATERIALS
Heated Transfer Line
The materials used to fabricate the heated transfer line  include: deactivated  fused silica  lined stainless steel
tubing 30 in x 1  mm,  i.d. 0.76 mm Silcosteel® (Restec Corp.,  Bellefonte, PA);  Nextel  312 thermal  insulation
sleeving  and Viton shrinkable  tubing (Omega,  Stamford, CT), heat  shrinkable Teflon tubing (Patriot Plastics,


                                                119

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Woburn,  MA);  aluminum  foil  with  silicon  adhesive  backing
(COMCO), polyimid moisture insulation tape (Newark Electronics,
Chicago,  IL);  thermal  insulated fiber  glass  cloth tape  (Fisher
Scientific, Pittsburgh, PA). The heated  transfer line is heated by
connecting  high  temperature  power  lead  wires   (Newark
Electronics) to both ends of the Silcosteer* tube. Temperature was
measured using Thermocouples C01-K and C02-K (Omega).

Heated Probe
The probe was made from a 1  m x 4.5 cm, 2.5 cm i.d., threaded
steel pipe. A 20 mm i.d. hole was cut in the pipe at one-third the
distance  from  the bottom.  The heat was supplied by inserting an
aluminum casing into the pipe,  which  contained a 10 cm x 1.5 cm
heating  cartridge   L4A712,  240V/1000W.  The  same  model
Thermocouples used to measure the  transfer line  temperature
were used in the Probe.
Figure  1. Cone penetrometer and  thermal  extraction  sampling
probe.

Equipment
The Silcosteel® was heated by passing current through the tube using an electrical isolation step-up transformer
(Grainger,  Haverhill,  MA)  with  power and  temperature controllers model  DCIP-50245-FOO and  model
988A-10FD-AARG, respectively (Watlow, St. Loius, MO). A Hewlett Packard (Palo Alto, CA) model 5972 mass
spectrometer was  ruggedized for the field and used in combination  with a  Tufts University (Medford, MA)
designed thermal  desorption  gas chromatograph. All GC/MS  total  ion  current chromatograms (TIC) were
acquired by  HP's data acquisition system. A new mass spectrometry data  analysis software developed at Tufts
was used to identify and quantify polycyclic  aromatic hydrocarbons (PAHs).  The Ion Fingerprint Detection™
software is available from Ion Signature Technology (Cambridge,  MA).
                                       UEfr


                                       l.OEfr


                                       11.8
Figure 2. GC/MS analysis of a soil
sample  collected  from Hanscom  Air
Force    Base    (Bedford,    MA);
compounds  found: 1) 1,1,1-trichloro-
ethene,   2)  methylene  chloride,  3)
1,1 -dichloroethene,  4)   1,4-difluoro-
benzene  (surrogate),  5)  toluene-da
(internal  standard),  6) ethylbenzene,
7) m/p-xylene, 8) o-xylene, 9) styrene,
10)   4-bromofluorobenzene  (surro-
gate).
Total Ion Chromatograra
  AAJ
                                                     Reconstructed Ion Chromatogram
RESULTS
Two key breakthrough technologies  have been  developed that meet the EPA data quality  measurement
objectives and EPA Soil Screening Level quantitation levels under fast GC/MS  conditions. First is the mass
spectral data analysis software. The software extracts between three and ten characteristic fragment ions for
each targeted organic and then, based on a patented set of algorithms, compound  identity and concentration are
determined. Algorithmic details can be found elsewhere.2 Although all MS systems can extract ions, they cannot
handle the amount of extracted ion information and determine compound presence using current statistical or
                                                 120

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
library matching routines when high levels of interferents are present in the sample. For example, the software
can provide compound identification in complex environmental samples without the need for extensive sample
cleanup. The second technology breakthrough is the thermal desorber. Unlike other commercially available units,
the TD can be ballistically heated from subambient temperatures to 320 °C in 8-sec. The TD uses a standard
Tenax tube for trapping VOCs and an empty glass sleeve for semivolatiles that have been swept from depth to
the surface by the carrier gas. Direct desorption of organics from solid materials or an organic extract into the GC
is made by placing known quantities into an empty glass sleeve.

Figure 2 illustrates a representative total ion current chromatogram (TIC) and  a reconstructed ion current (RIG)
chromatogram where the data analysis software was able  to "see through" the  matrix. Note that the internal
standard (peak #5) and target compounds  1,1,1-trichloroethene and  1,1-dichloroethene (peaks 1 and  3) are
buried within the  matrix and that their corresponding signals are 102 to 104 times smaller than the matrix signal.
Typically, analysts would dilute this sample prior to analysis  based on visual inspection of the petroleum present
in the sample. This practice will result in the  loss of low level target compounds such as the chlorinated solvents
found in this sample.

Figure 3 shows the TIC and RIC chromatograms produced from a 10-min TDGC/MS analysis
of a  standard mixture  of polychlorinated biphenyls (PCBs, Aroclor 1248), polycyclic aromatic
hydrocarbons (PAHs), chlorinated pesticides, and engine oil  (25% v/v). A total  of 1,000-ng
PCBs was thermally desorbed into a 15-m  GC  column along with 19 chlorinated pesticides
(20-ng/compound),  16 PAHs (40-ng/compound), and pyrene-dio (50-ng) added as an internal
standard. An expanded view of the RIC chromatogram between 6.7 min and 6.9 min is shown
here. Note that  there are  six compounds  that elute within this time  domain. Compound
identification was made  based on a set of  algorithms that extracted 3-6  fragment ions per
compound from  the TIC and computed their match against standard reference spectra in
seconds. Each compound's RIC signal, based upon preselected quantitation ions, are then
used to  produce the RIC chromatogram  and to quantify  compound concentration. The
algorithms and results will be presented documenting measurement accuracy, precision, and
sensitivity.

Figure 4 depicts  the schematic of the heated transfer line and the electronic circuitry used to control  the power
and  temperature. The figure shows the various layers  including moisture, electrical,  and thermal  insulation
sleeves as well as the  fused silica coated stainless steel tube Silcosteel® The goal was to heat the transfer line to
300 °C and achieve a 15 cm bend radius so that the pipes in the truck could  be  stacked efficiently. This feature is
important  since cone penetrometer systems can reach subsurface depths  of  up to 60  m when the geology is
                                                       amenable. Based  on the design  shown in the figure
                                                       a  10 cm  bend  radius was  obtained,  with the
                                                       Silcosteel®  temperature   programmable   from
                                                       ambient soil temperatures up to 300 °C. To date, a
                                                       30 m transfer  line has been made, which can  be
                                                       woven through ten  1 m pipe sections to achieve
                                                       subsurface depths of  10  m.  The  transfer  line is
                                                       heated  by  passing direct current  through the
                                                       stainless steel tube.  Imbedded  in the transfer line
                                                       are the electrical  and thermal couple wires needed
                                                       to carry current to the probe head and to monitor
                                                       both the transfer line and probe temperatures.
                                                       Figure  3.  10-min  TDGC/MS analysis  of a  soil
                                                       sample  fortified with a standard  mixture of PCBs,
                                                       PAHs, pesticides and  gasoline/engine oil (1:3  by
                                                       vol).
                    Reconstructed Ion Current
Total Ion Current
                                                  121

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                                                     collection gas
Figure 4. TECP system, which includes 1)
heat  shrinkable  sleeve,  2)  fiber  glass
insulation, 3)  high temperature electrical
wires, 4)  Viton shrinkable  tubing,  5) heat
shrinkable Teflon tubing,  6)  Nextel 312
thermal insulation sleeving, 7) deactivated
fused silica lined stainless steel tubing 30
m x 1  mm, i.d.  0.76 mm Silcosteel® 8)
aluminum   foil   with  silicon  adhesive
backing, 9) polyimid  moisture  insulation
tape,  10)  thermal insulated  fiber  glass
cloth   tape,   11)   polyimid  moisture
insulation  tape,  12)  thermocouple  wire,
13)  isolation  step-up  transformer, 14)
temperature  and  power controllers, 15)
transfer line thermocouple, and 16) high
temperature electrical wires.
                               Table 1. Material Balance of Closed Chamber
                               (Rem = 5700; V0 = 32; wg = 0.6 m/s; dry sand)

Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluorarthene
Benzo(a)pyrene
Dibenz(a,h)anthracene
lndeno(1 ,2,3-ed)pyrene
Benzo (g,h,i)perylene
250 °C
% Rec
5
61
74
72
76
75
79
86
56
50
41
41
65
NA
NA
NA
% Soil
ND
ND
ND
ND
5
6
5
9
42
40
53
53
31
NA
NA
NA
% Ads
ND
ND
3
2
2
3
3
2
2
2
6
6
4
NA
NA
NA
300 °C
% Rec
15
63
78
75
83
78
80
83
67
61
55
55
65
37
37
23
% Soil
ND
ND
ND
3
5
6
4
6
30
31
40
40
23
59
59
67
% Ads
ND
ND
ND
0.6
2
2
3
1
1
1
4
4
3
2
2
4
ND - not detected; NA - not analyzed

Little degradation of the silica lining  is observed as long as air is purged from the system. Air is flushed from the
transfer line by nitrogen prior to transfer line heating. After the Silcosteel® has been conditioned the gas valve is
switched to re-direct nitrogen into the carrier gas  line. At this point, a vacuum pump is turned on to collect the
soil/organic vapor at the collection window and to transport the organics to the surface through the Silcosteel*
tube.  The valve can then  be  re-positioned to cleanse the transfer line tube when high  levels of  sample are
collected. This step is intended to eliminate sample carry over from one sample location to the next. Work is in
progress to automate the TECP system to control the probe and transfer line temperatures as well as the carrier,
flush, and collection gas lines.

Material balance experiments were conducted to determine the  thermal extraction efficiency for PAHs at 250°C
and 300°C, see Table 1. At optimum conditions, i.e., soil temperature, carrier gas flow  rate, collection volume,
and under closed cell conditions, greater than 55% recovery was obtained. Closed cell conditions represent the
maximum amount that can be extracted, as opposed to the TECP, since  none of the organic vapor produced is
                                                   122

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


lost to the environment and all soil is uniformly heated. The organic vapor is efficiently flushed from the cell into
Tenax for VOCs or the cold trap for semi-VOCs. Note that less than 4% of the  collected organics remained in the
transfer line after collection. Research is in progress to chemically modify the surface to minimize (eliminate)
organic absorption. Nonetheless, back-flushing for 5-min reduced the percent adsorbed to non detectable levels.

                            Table 2. Comparison of TECP vs. Closed System
                      T^pe = 450 °C; Tsoi, = 280 °C; Rem = 6000; wg = 1.2 m/s; Ud = 15

Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Chrysene
Benzo(a)anthracene
Benzo(b&k)fluoranthene
Benzo(a)pyrene
lndeno(1 ,2,3-ed)pyrene
Dibenz(a,b) anthracene
Benzo(g,h,i)perylene
TECP % Recovery
50-ppm
20
34
41
49
37
37
65
64
27
49
29
25
26
26
17
25-ppm
37
36
39
36
32
32
53
53
37
37
27
23
19
19
16
1 5-ppm
26
36
33
38
35
34
48
45
31
36
29
36
20
19
16
5-ppm
37
33
30
37
34
34
36
68
41
41
23
24
23
23
16
Ave Rec
30
35
36
40
35
34
50
47
34
40
27
27
22
22
10
% RSD
23
5
18
17
7
7
34
15
16
12
12
26
14
14
4
Closed
Chamber
% Recovery
18
62
77
73
80
77
80
85
57
62
48
65
65
62
20
%
Diff
68
44
53
46
57
56
37
32
40
36
44
58
31
32
19
 Note: % Difference is between Average TECP Recovery and Closed Chamber Recovery

 The TECP and closed cell data comparison study is shown in Table 2. Recall that the closed system represents
 maximum recovery of analyte while the TECP is what one should  expect to achieve in the field. The TECP
 measurement  precision  determined  over an  order  of  magnitude  for PAHs  is excellent. The  results are
 remarkable and are as good or better than what is achievable through soil/solvent extraction. The accuracy for
 the TECP approximates one-half that of the closed (ideal) system. Results will be presented illustrating collection
 efficiency as a function of soil type, probe depth and moisture content.

                               Table 3. TECP Field Study, Berlin Vermont
                    TDiDe = 450 °C; Tsoji = 280 °C; Ttmne = 280 °C; Rem = 6000; wfl = 1.6 m/s

Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Dibenz(a,h)anthracene
lndeno(1 ,2,3-cd)pyrene
Benzo(g,h,i)perylene
TECP Measured PAH/Soil Concentration
Boring 1
97-cm
2
9
2
7
11
11
44
29
45
45
5
6
12
2
3
6
142-cm
2
6
2
3
5
5
3
4
11
11
0.4
0.3
0.6
ND
ND
ND
Boring 2
1 30-cm
4
14
3
10
5
7
6
6
6
6
0.5
0.5
ND
ND
ND
ND
155-cm
ND
ND
2
4
2
2
2
5
3
3
ND
ND
ND
ND
ND
ND
1 90-cm
ND
5
3
3
4
4
3
4
2
2
ND
ND
ND
ND
ND
ND
The TECP and heated transfer line was tested in the field employing ARA's CPT in Berlin, Vermont. The location
was a Vermont state central maintenance facility known to contain petroleum hydrocarbon contamination. The
                                                  123

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


TECP was tested for mechanical ruggedness and its ability to collect subsurface bound organics. The vertical
profile of two borings are shown in Table 3. PAHs were found at depths of 97-cm and 142-cm at boring one and
130-cm. The hot organic vapor was cooled by dry ice and collected in  an empty glass sleeve attached to the
transfer line. Total collection and analysis time  required at each depth  location was approximately 15-min per
sample inclusive of sample collection, transfer line back flushing, and analysis time. Unfortunately, comparison
measurements between the TECP measured concentrations and actual soils collected at depth and analyzed by
standard GC/MS  methods were not possible since ARA was  unable to re-enter the hole with a soil  sample
collection probe without breaking the CP rod.  The results produced to date are promising suggesting that direct
on-line chemical measurements of subsurface contaminants may be possible within a couple years.

REFERENCES
1. Cleaning Up the Nation's Waste Sites: Markets and Technology Trends, Office of Solid Waste and Emergency
  Response, U.S. EPA, Washington, DC, 1993, EPA 542-R-92-012, 164 pp.
2. Y. Gankin, A. Gorshteyn, S. Smarason, and A. Robbat, Jr., Anal. Chem, 70:_1655-1663 (1998).
3. A. Robbat,  Jr., "Field Analytics, Dynamic Workplans" The Encyclopedia of Environmental Analysis and
  Remediation, ed. by R.A. Meyers, John Wiley & Sons, Inc. New York, NY., July 1998.
4. A. Robbat, Jr., "A Dynamic Site Investigation Adaptive Sampling and Analysis Program for Operable Unit 1 at
  Hanscom Air Force Base, Bedford, Massachusetts", U.S. Environmental Protection Agency, Region I, October
  1997; see http://clu-in.com/char1.htmtfregional.
5. A. Robbat, Jr., T-Y Liu, and B. Abraham, Anal. Chem.,  64:358-364 (1992).
6. A. Robbat, Jr., C. Liu, and T.-Y. Liu, J. Chromatography, 625: 277-288  (1992).
7. A. Robbat, Jr., C. Liu, and T.-Y. Liu, J., Anal. Chem., 64: 1477-1483 (1992).
8. K M. Abraham, T-Y Liu, and A. Robbat, Jr., Hazardous Waste & Management, 10: 461-473 (1993).
9. K. Jiao and A. Robbat, Jr., J. of AOAC International, 79: 131-142 (1996).
                    AN INTEGRATED NEAR INFRARED SPECTROSCOPY SENSOR
                            FOR IN-SITU ENVIRONMENTAL MONITORING

                                  Roland A. Levy and John F. Federici
                 New Jersey Institute of Technology, University Heights, Newark, NJ 07102
                                         Tel. (973)-596-3561

The monitoring of environmental organic contaminants currently involves off-site methods which prohibit optimal
usage. This study explores the possibility of combining the principles of interferometry with that of near infrared
evanescent  wave  absorption spectroscopy to  produce  a  novel  integrated sensor  technology  capable  of
monitoring and determining in-situ the concentration of numerous organic analyte species  simultaneously. This
novel sensor promises to be non-intrusive and to provide accurate, rapid, and cost effective data.  The overall
instrument is envisioned to be compact, portable, rugged,  and suitable for real time monitoring of organics. The
sensor consists of a symmetric, single mode Mach-Zehnder interferometer with one arm (sampling) that is either
exposed directly to the analyte or coated with a thin hydrophobic layer that  enhances the binding  of  pollutant
molecules onto its surface. A glass buffer  layer protects the  second arm (reference) from the influence of
pollutants. Light  is coupled into the  waveguide and split between the sampling and reference arms using a
Y-splitter configuration. Changes in the refractive index caused by the presence of organic contaminants result in
a measurable phase difference between the sampling and  reference arm.  Selectivity of the sensor is achievable
by utilizing  evanescent wave  absorption spectroscopy  in the near  infrared,  a  technique  which measures
wavelength dependent refractive index changes. The waveguide structures used in this study are fabricated on
10 cm silicon wafers. V-grooves are  first formed in  the silicon substrate to hold the fibers which couple to the
ends of the Mach-Zehnder interferometer. A 10 urn thick SiO2 film is synthesized by low pressure chemical vapor
deposition (LPCVD) to  act  as cladding material for the waveguide and  prevent light from coupling  with the
underlying silicon. A 4 urn thick phosphorus-doped (7.5 wt% P) LPCVD SiO2 film is then deposited to act as core
material for the waveguide. This layer  is patterned using standard lithographic exposure and  plasma etching
techniques and subjected to a 1050 °C anneal to cause viscous flow and  round off the edges. This  rounding-off
procedure is necessary to minimize coupling losses between fiber and waveguide. The refractive index of the
doped glass  is near 1.48, thus, producing with the underlying SiO2 (n=1.46) substrate a single mode waveguide


                                                 124

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


device. Deposition of a 0.5 |jm thick LPCVD undoped SiO2 buffer layer over the entire wafer and a subsequent
lithographic step results in selective removal  of that layer over the sampling arm of the  interferometer. This
configuration  allows for exposure of the  sampling arm (uncoated  or coated) to  various contaminants in the
environment which cause a change in  the effective refractive index of that arm. The arm coated with the SiO2
buffer layer sees a  constant refractive index  of n=1.46. A lithographic mask is used to produce the required
patterns.  Five micron wide waveguides form  the  two interferometer paths using a splitting angle of 2°. The
sampling and reference arms are at a  fixed separation of 50 urn and variable lengths (4, 6, 8, and 10 mm). A
detailed analysis of these processing steps together with the principles of operation  of the sensor will be
discussed.
     A NITRIC OXIDE AND AMMONIA SENSOR ARRAY FOR FOSSIL FUEL COMBUSTION CONTROL

                                B.T. Marquis, J.F Vetelino and T.D. Kenny
       Laboratory for Surface Science & Tech. and Department of Electrical and Computer Engineering,
                                University of Maine, Orono, Maine 04469

ABSTRACT
In fossil fuel combustion processes, nitric oxide (NO) emissions are minimized by a selective catalytic reduction
(SCR) technique where  ammonia  (NH3)  is  injected  into the flue gas  stream to react with  NO  to form
environmentally safe gases such as nitrogen and water vapor.  Unfortunately, this process is usually incomplete,
resulting in either NO emissions or excess NH3 (NH3 slip). Therefore, a critical need  exists for an in situ sensor
array to continuously monitor the NO and NH3 levels at the output of the SCR system near the stack to provide
real time control  of the NH3 injection and hence minimize the NO emissions to the environment. Chemiresistive
sensor technology is being used  to develop a small,  portable, sensitive and selective sensor array  that has
potential to continuously measure NO and NH3 emissions. The sensor array utilizes a tungsten trioxide  (WO3)
film as the sensing element to simultaneously identify NO and NH3 concentrations present in the fossil fuel gas
exhaust. Several film parameters such as thickness, dopant (gold vs. ruthenium), doping  method (post-sputter
vs. co-sputter), doping amount, deposition temperature, annealment procedure and operating temperature were
varied to determine their effect  on the film's  sensing properties. As a  result, a 500A  WO3:16A Au  film
post-sputtered at 200°C demonstrated high sensitivity and selectivity to NH3 whereas a 1000A WO3 undoped film
sputtered at 200°C exhibited greater sensitivity to NO.

INTRODUCTION
The detection and measurement of flue gases are critical not only for achieving real time process control of new
clean combustion systems, but also to minimize  their emissions of dangerous air pollutants. Among the most
dangerous of these air pollutants are nitric oxide (NO) and nitrogen dioxide (NO2), collectively referred to as NOX.
Currently about one half of all NOX emissions into the environment are due to power plants and industrial boilers.
NOX  gas which is the precursor to nitric and nitrous acid, causes acid rain and photochemical smog  and is a
critical factor in the formation of ozone in the troposphere. Ground level ozone is a severe irritant, responsible for
the choking,  coughing  and  burning  eyes associated  with smog.  Ozone  often  damages lungs,  aggravates
respiratory disorders, increases susceptibility  to  respiratory infections and is particularly harmful to  children.
Elevated ozone levels can also inhibit plant growth and cause widespread damage to trees and crops. Therefore,
exceeding critical NOX levels poses immediate health and environmental problems.

In fossil fuel combustion NOX is formed by high temperature chemical processes from both nitrogen present in
the fuel  and oxidation of nitrogen in air. Typically, the NOX emissions consist of 90-95% NO with the remainder
being N2O and NO2.1

Several  methods have been  examined as potential control  systems for  the  reduction  of  NO emissions in
combustion  processes  including combustion  control and  flue gas treatment  techniques  such as  selective
noncatalytic reduction (SNCR) and selective catalytic reduction (SCR)23. SCR technology achieves the highest
overall control efficiency  of 60-80% NO reduction. In this process,  NH3 is  uniformly injected into the flue gas
stream that passes through a catalyst  bed to enhance the kinetics of the NO/NH3 reaction. This can be  further
improved with  air staging to provide a longer residence time allowing more  NO  to react with the NH3 ,  thus


                                                 125

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


increasing NO reduction and decreasing NH3 emissions (commonly called NH3 slip). Although there are several
chemical reactions that may take place in the catalyst bed, the most common reactions are as follows,

                 4 NH3 + 4 NO + O2 -> 4 N2 + 6 H2O,                          (1)
                 4 NH3 + 6 NO -> 5 N2 + 6 H2O,                               (2)
                 4 NH3 + 3 02 -» 2 N2 + 6 H2O                                (3)
                 and
                 4 NH3 + 5O2 -- NO + 6 H2O.                                  (4)

Reaction (1) is the predominant reaction1 and reduces NO into harmless nitrogen and water vapor. The reaction
given in equation (3) neutralizes some of the excess NH3 that did not fully react with the NO in equations (1) and
(2) and  hence decreases the NH3  slip. An unwanted side reaction given by equation (4) oxidizes  NH3 to form
water vapor and NO. This harmful oxidation  process however usually takes place at elevated temperatures, and
can be minimized by temperature control. If  a judicious choice of NH3 injection levels is made, NO emissions in
SCR systems, as well as NH3 slip into the atmosphere can be reduced significantly.

In order to control the  precise levels of NH3  injection, a critical need exists for an NO/NH3 sensor system in the
flue gas exhaust prior  to the stack emission. The output from the sensor would be fed back to the SCR system
via a  real time control  system to adjust NH3  injection levels  thereby maximizing NO reduction with minimal NH3
slip  into the atmosphere.  Current  NO  and  NH3 measurement  techniques, such  as  chromatography,
chemiluminescence and  infrared  adsorption  are  very  expensive  and too  bulky  for  in  situ  operation.
Chemiresistive semiconducting metal oxide  (SMO) films offer the technology to develop a small,  inexpensive
and reliable in situ sensor array for the simultaneous identification of NO and NH3 present in a flue gas exhaust.

The first report on Chemiresistive SMO films for gas sensing appeared in 19624  Since  that time, considerable
efforts have been made5 to study SMO films for the detection of a variety of gases. Investigators have examined
SMO films such as SnO2 615, TiO215'16, indium tin oxide (ITO) 17, ZnO 1S1718 and WO319'22 for detecting NO, as
well as ZnO 2328, MoOs/TiO2 29 and WO3 3034 for detecting NH3. WO3 films can  operate at elevated temperatures
for long periods of time and selectively detect NO and NH3 in the presence of interferent gases such as H2, CO,
CO2,  CH4 and various other hydrocarbons. These  films  are  also electrically and structurally stable  at elevated
temperatures.

The electrical conductivity of thin WO3 films doped with metals such as gold (Au) and ruthenium (Ru), change
upon  exposure to NO and  NH3.  The  sensitivity  of WO3  to NO  and NH3  depends  significantly upon film
parameters such as thickness, dopant type, doping method, deposition temperature, annealing  procedure and
operating temperature. Furthermore,  the functional relationship between sensitivity and each of these parameters
is different for both gases.

THEORY
The WO3 film conductivity changes  as a function of NO and NH3 gas concentrations. These films exhibit very
fast response and recovery  times and, after initial film conditioning,  show no appreciable aging  effects after
repeated gas exposures. The basic chemical sensing mechanism involves the dissociative chemisorption of the
target molecules and the formation of transitory concentrations of chemisorbed atoms on the WO3 film surface.
The  rate of dissociation of the target molecules on the surface  can be  greatly enhanced by the addition  of
catalytic metals such as gold or ruthenium. Although the overall chemistry of the possible interactions of NO and
NH3 with WO3 is complex and not well defined,  certain primary interactions dominate.

In the case of NH3, which acts as  a  reducing agent (an oxygen scavenger), the carrier concentration in the film
rises as a result of a decrease in adsorbed surface oxygen, as follows,

                             2 NH3 + 3O'(ad)^ N2 + 3 H2O + 3 e-                     (5)

This rise in carrier concentration within the film  is then manifested as  a decrease in resistivity.

Unlike NH3, NO acts as an oxidizing agent (oxygen donor) at the temperatures of interest. Thus, reactions with
NO result in an increase in chemisorbed oxygen in the film, decreasing the free carrier concentration, as follows,

                             2 NO + 2 e- -* N2 + 2 O (ad).                           (6)


                                                  126

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
This decrease in carrier concentration causes the film resistivity to rise. This behavior is completely opposite to
that observed with NH3.

EXPERIMENTAL SETUP
The  WO3 sensing films  were deposited on alumina and sapphire substrates  by reactively sputtering a pure
tungsten target in an 80/20 argon-oxygen atmosphere using an RF magnetron  sputtering system. Alumina and
sapphire substrates were used because they are good electrical insulators and thermodynamically stable. A DC
magnetron gun was used to precisely dope  the WO3 film with gold or ruthenium to selectively catalyze the
reaction with NO or NH3. After the films were sputtered, they were subjected to an annealing process which
transformed the as-deposited amorphous film into a polycrystalline film. This results in a more stable film that is
chemically and conductiometrically inert to moisture and many potential interferent gases. Microheaters which
controlled the temperature of the sensing film were fabricated by depositing a thin chrome serpentine structure
on alumina.
The chemiresistive sensing element is placed on the surface of the serpentine microheater with silicone heat sink
                                               grease. The heater and a thermocouple are used to control
                                               the film's  operating temperature. The entire  apparatus is
                                               suspended on two wires attached to  the sensor  package to
                                               provide thermal  isolation from other  nearby  sensors and
                                               electronics. Electrical connections to both the sensing film
                                               and microheater are made with 100um aluminum bond wire
                                               to the  sensor  package. The chemiresistive  sensing  element
                                               and microbeater is shown in Figure 1.
Thermocouple

Chemiresistive
Sensing Element
  Serpentine
  Microheater
                                               Figure 1. Chemiresistive sensing element and microheater.
 In order to determine the electrical conductivity of a large number of WO3 films exposed to a wide range of gas
 concentrations, a system capable of simultaneously testing and controlling up to eight sensors was designed and
 built.  This system  improved testing  efficiency and  insured that all films were subjected to the same gas
 environment during each test. A block diagram of this system is shown in Figure 2.

 The sensing elements shown in Figure 1 reside inside the sealed teflon gas chamber. Two-point conductivity
 measurements of each sensing film are performed with an electrometer and read into the  computer via a HPIB
                                                                 interface. The computer outputs an analog
                                                                 voltage   to  the  microheater  that  is
                                                                 determined by feeding the  measured film
                                                                                             proportional
                                                                                             temperature
   Gas Exhaust

Analog Interface
Digital Interface
i AV >
f
 Tefloii Gaa Chamber
 with Eight Separately
 Addressable Sensors
 and Heaters
                 Electrometer ^—        ^	
                 Multiplexing;  ^J Electrometer' HPFB Interface
                        '  r\
                    Gas Delivery
                   i System
                   Interface Electronics
                                             Computer Control/
                                             Data Acquisition
                                                                temperature   through    a
                                                                integral  differential  (PID)
                                                                control algorithm. The support electronics
                                                                include  power  stages  to   supply   the
                                                                necessary  current to all the  microheaters
                                                                and  amplifiers that linearize  the  ther-
                                                                mocouples'  temperature  measurements
                                                                into  a 10mV/°C  analog signal  for  the
                                                                computer's temperature control algorithm.
Figure 2. Block diagram of the
experimental testing system.
The delivery of the  NO and NH3 gases to the sensor is achieved with the system shown in Figure 3. The
apparatus consists of a  bottle of the dilution gas (simulated flue gas, in this case), bottles of each test  gas (NO
and NH3), an electronic three-way solenoid valve (Y-valve) and a system  of mass flow controllers and plumbing.
The entire system is computer controlled and is capable of delivering precise concentrations of NO and NH3 in
an atmosphere of controlled humidity for precise lengths of time. It is fully programmable and capable of running
                                                   127

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
multiple tests and collecting data without user input or intervention.

RESULTS
In order to engineer the film to respond to NH3 and  NO in a fashion unique to each gas,  a large number of
experiments were performed to determine the film parameters that would result in responses to NH3 and NO that
were clearly different. In particular, relationships between film thickness, gold doping, annealing environment,
operating   temperature   and   electrical

                                                 ill
conductivity when the film  is exposed to
NH3 and NO gas were determined. A  few
of these experiments are discussed below
in order to demonstrate the feasibility of
using  WO3 films  in  a  sensor  array to
selectively detect NH3 and NO.
          Figure 3. Gas Delivery System
Simulated
Flue Gas
(Diluent)
                          Exhaust
                                        iP'
                                       -.•*:.--\-	
                                       r
                                   	[•*..  J-	-IX)
                                                                    Gas
                                                                    Chamber
                                                        Bubble
~l
 Exhaust
     15
                                                    SOppm NO
                                         40ppmNO
                              SOppmNO
                     20ppm NO
          10ppm NO
                 30
                            60          90
                             Time (minutes)
                                                 120
                                                            150
                      A  100Q.A WO3  undoped  film  was
                      sputtered  at  500°C  on a  sapphire
                      substrate and  annealed in  dry air at
                      300°C for 10  hours and exposed to
                      concentrations  of  NO  ranging from
                      0-50ppm at a temperature of 300°C in
                      an environment of dry CO2. As can be
                      seen in  Figure  4,  the film  demon-
                      strated  a very fast linear response to
                      0-50ppm NO. It also recovered to the
                      baseline resistance when the gas was
                      taken away which allows the sensor to
                      measure absolute NO concentrations
                      in real time.

                      Figure  4. Response of a 1000A W03
                      film to  0-50 ppm  NO at 300°C in dry
                      CO2
Likewise, Figure 5 shows the same 1000A WO3 film responding to a SOppm pulse of NH3 in 50% humid air. It is
important to note that the response to  NO  is manifested as an increase in resistivity, while  NH3 results in an
increase in conductivity (i.e. a decrease in  resistivity), as predicted by equations (5) and (6)  above. The film's
change in resistivity is 6 times greater to 50ppm  NO than to SOppm NH3. In fact, the film's response to only
10ppm  NO was still greater than  the response to SOppm NH3. The typical NH3 and NO concentration ranges
found in the fossil fuel combustion exhaust are 0-5ppm and 5-75ppm,  respectively. Therefore, this film  could
selectively measure up to 10-50ppm NO in the presence of up to SOppm NH3.

A 500A WO3:16A Au film post-sputtered at  200°C  on  alumina and annealed in dry air at 350°C  for 5 hours was
exposed to 3 pulses of 10ppm NH3 and 3 pulses of SOppm NO in dry air at 350 °C. Figure 6 shows the film's high
sensitivity and reproducibility to 10ppm NH3 and relatively low sensitivity to SOppm NO. Again,  the response to
NH3 is  manifested  as an  decrease in resistivity, while NO  results  in  an  increase  in  resistivity.  The film
demonstrates adequate resolution to potentially measure NH3 concentrations between 0-10ppm  in the presence
of up to 5-75ppm of NO.

SUMMARY
As a result of the tests described above, it can be  deduced that the film parameters,  such  as thickness, doping,
annealing  procedure, and operating temperature significantly effect the film's sensitivity and  selectivity to NH3
and NO gas concentrations. Both, the 1000A WO3 undoped film operated at 300°C and the 500A WO3:16AAn
post-sputtered film operated at 350°C show  potential as  a two sensor array capable of simultaneously identifying
concentrations  of NO and NH3.  Eventually, a  neural  network  could be  trained  and  integrated with  the
                                                  128

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
chemiresistive sensor array to improve
real time process control and correlate
the  sensors' responses to the NO and
NH3 concentrations present in  a flue
gas exhaust.  In future  work,  this
technology  could  be   extended  to
include other  combustion gases such
as SOX and H2S
Figure 5. Response of 1000A WO3
film to SOppm NH3 at 300°C  in humid
air.
                    16
                 ?  13
                 0  10
                                     S   7
                                        30
                                                  60
                                                      90
                                                                   Time (minutes)
     15


     12


     9


     6


     3


     0
                              60


                              50 -


                              40 |


                              30 |

                              20 o
                                 ID
                                 O
                              10
             60
120
180    240

Time (minutes)
300
360
420
                                             ACKNOWLEDGMENTS
                                             The   authors  wish  to  thank  the
                                             Environmental Protection Agency for
                                             their  financial  support  (Grant  No.
                                             GAD#R826164).  Also,   T.D.  Kenny
                                             was   funded  through  the   National
                                             Science Foundation Research Experi-
                                             ence  for  Undergraduates   program
                                             (Grant No. EEC9531378).
Figure  6.  Response  of  a  500A
WO3:16A Au film to (3)  10ppm pulses
of NH3 and (3) SOppm pulses of NO at
350°C in humid air,  respectively.
REFERENCES
1.  J.H.A. Kiel, A. Edelaar, W. Prims and  W. Van  Swaajj, "Selective  Catalytic Reduction of  Nitric Oxide by
    Ammonia", Applied Catalysis B: Environmental 1, pp. 41-60 (1992).
2.  J.A. Eddinger, "Status of EPA Regulatory Development Program for Revised NO* New Source Performance
    Standards for Utility and Nonutility Units-Performance and Costs of Control Options", Joint Symposium on
    Stationary Combustion NOX Control 6 (1995).
3.  K.J. Fewel and J.H. Conroy, "Design Guidelines for NH3 Injection Grids Optimize SCR NOX Removal",  pp.
    56-64(1993).
4.  T.  Seiyama,  A. Kato, K. Fujiishi and  M. Nagatani, "A New Detector for Gaseous Components Using
    Semiconductive Thin Films", Anal. Chem. 34,  1502 (1962).
5.  See,  for example, Proceedings of the  First, Second, Third, Fourth  and Fifth International Meetings on
    Chemical  Sensors, 1983,  1986, 1990, 1992 and 1994 and references contained therein. Editors Elsevier
    (1983, 1992).  Unpublished (1986, 1990, 1994).
6.  K. Tanaka, S. Morimoto,  S. Sonoda, S. Matsuura,  K. Moriya and M.  Egashira,  "Combustion Monitoring
    Sensor Using  Tin Dioxide Semiconductor", Sensor and Actuators B 3, 247 (1991).
7.  S. Sberveglieri, G. Faglia, S. Groppelli and P. Nelli, "Methods for the Preparation of NO, NO2 and H2 Sensors
    Based on  Tine Oxide Thin Films Grown by Means of RF Magnetron Sputtering Techniques", Sensors and
    Actuators 68,79(1992).
8.  G. Williams and G.S.V. Coles, "NOX Response of Tin Dioxide Based Gas Sensors", Sensors and Actuators B
    15-16.349(1993).
9.  F.J. Gutierrez, L. Ares, J.I. Robla, M.C. Horillo, I. Sayago, J.M. Getino  and J.A.  de Agapito, NOX Tin Dioxide
    Sensor Activities as a  Function of Doped Materials and Temperature", Sensors and Actuators B 15-16, 354
                                                 129

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


   (1993).
10.  G.B. Barbi and J.S. Blanco, "Structure of Tin Oxide  Layers  and Operating Temperature as Factors
    Determining the Sensitivity to NOX", Sensors and Actuators B 15-16, 372 (1993).
11. C. DiNatale, A. D'Amico, F.A.M. Davide,  G. Faglia, P. Nelli and G. Sberveglieri, "Performance Evalation of
    an SnO2-based Sensor Array for the Quantitative Measurement of Mixtures of H2S and NO2", Sensors and
    Actuators 820,217(1994).
12. G. Wiegleband J. Heitbaum, "Semiconductor Gas Sensor for Detecting NO and CO Traces in Ambient Air of
    Road Traffic", Sensors and Actuators B17, 93 (1994).
13.  I. Sayago, J.  Gutierrez, L. Ares, J.I. Robla, M.C. Horillo, J. Getino, J. Rino and J.A. Agapito, "The Effect of
    Additives in Tin Oxide on the Sensitivity and Selectivity to NOX and CO", Sensors and Actuators B 26-27.19
    (1995).
14. Ibid., "Long-term Reliability of Sensors for Detection of Nitrogen Dioxides", Sensors and Actuators B 26-27,
    56(1995).
15.  K. Satake, A. Kobayashi, T. Inoue, T. Nakahara and T. Takeuchi, "NOX Sensors for Exhaust Monitoring",
    Proc. Third Int. Mtg. on Chem. Sensors, Sept. 24-26, 1990, Cleveland Ohio, pp. 334-337.
16. J.  Huusko, V. Lantto and H. Torvela, "TiO2 Thick Film Gas Sensors and Their Suitability for NOX Monitoring",
    Sensors and Actuators B 15-16. 245 (1993).
17.  G. Sberveglieri, P. Benussi, G. Coccoli, S. Groppelli and P Nelli, "Reactively Sputtered Indium Tin Oxide
    Polycrystalline Thin Films as NO and NO2 Gas Sensors", Thin Solid Films 186, 349 (1990).
18. S. Matsushima, D.  Ikeda, K. Kobayashi and G. Okada "NO2 Gas Sensing Properties of Ga-Doped ZnO Thin
    Films", Proc.  Fourth Int. Mtg. on Chem. Sensors, Sept. 13-17, 1992, Tokoyo, Japan, pp. 704-705.
19.  M. Akiyama,  J. Tamaki, N.  Miura and N. Yamazoe, "Tungsten Oxide-Based Semiconductor Sensor Highly
    Sensitive to NO and NO2", Chem. Letters, 1611 (1991).
20. M. Akiyarna, Z. Zhang, J. Tamaki, T. Harada, N. Miura and N. Yamazoe, "Tungsten Oxide-Based Sensor for
    Detection of Nitrogen Oxides in Combustion Exhaust", Sensors and Actuators B 13-14, 619 (1993).
21. J.  Tamaki, Z. Zhang, K. Fujimori, M. Akiyama, T. Harada, N. Miura and N. Yamazoe, "Grain-Size Effects in
    Tungsten Oxide-Based Sensor for Nitrogen Oxides", J. Electrochem. Soc. 141. 2207 (1994).
22. G. Sberveglieri, L. Depero, S.  Gropelli and P  Nelli,  "W03  Sputtered Tin Films for NOx Monitoring", Sensors
    and Actuators B 26-27. 89 (1995).
23. H. Nanto, H. Sokooshi, and T.  Usuda, "Smell Sensor Using Aluminum-Doped Zinc Oxide Thin Film Prepared
    by Sputtering Technique", Sensors and Actuators B10_,_pp. 79-83 (1993).
24.  Hidehito Nanto, Tadatugu Minami, and Shinso Takata,  "Ammonia Gas  Sensor Using Sputtered Zinc Oxide
    Thin Film", Proceedings of the 5th Sensor Symposium, pp. 191-194 (1985).
25.  Hidehito Nanto, Tadatugu Minami, and Shinso  Takata, "Zinc-Oxide Thin-Film Ammonia Gas  Sensors with
    High Sensititivity and Excellent Selectivity", Journal  of Applied Physics 60 (2). pp. 482-484 (1986).
26.  Arya, D'Amico, and Verona,  "Study of  Sputtered ZnO-Pd Thin Films  as  Solid State H2 and NH3 Gas
    Sensors", Thin Solid Films 157, pp. 169-174 (1988).
27.  H.  Nanto, H.  Sokooshi, T.  Kawai,  and T. Usuda, "Zinc-Oxide  Thin-Film Trimethylamine Sensor With  High
    Sensitivity and Excellent Selectivity", Journal of Materials Science  Letters H, pp. 235-237 (1992).
28.  G. Sberveglieri, S. Groppelli,  and P.  Nelli, "A Novel Method for the Preparation of ZnO-ln Thin Films for
    Selective NH3 Detection", 5th International Meeting on Chemical Sensors, pp. 748-751 (1994).
29. A.R. Raju, C.N.R.,  (MoO3/TiO2 and Bi2MoO6 as Ammonia  Sensors", Sensors and Actuators B, 21, pp. 23-26
    (1994).
30. A. Bryant, K.-Poirer, D. Lee, and J.F. Vetelino, "Gas Detection Using Surface Acoustic Wave  Delay Lines",
    Sensors and Actuators 4_, pp. 105-111  (1983).
31. Tomoki  Maekawa,  Jun Tamaki, Norio Miura, and Noboru Yamazoe,  "Gold-Loaded Tungsten  Oxide Sensor
    for Detection  of Ammonia in Air", Chemistry Letters, pp.  639-642 (1992).
32. Tomoki  Maekawa, Jun Tamaki, Norio Miura, and Noboru  Yamazoe, "Promoting Effects of Noble Metals on
    the Detection of Ammonia  By Semiconducting Gas Sensor", New Aspects of Spillover Effects in Catalysis,
    pp. 421-424(1993).
33.  G. Sberveglieri, L.  Depero, S.  Groppelli, and  P Nelli, "WO3 Sputtered Thin Films for NOX Monitoring",
    Sensors and Actuators B 26-27. pp.89-92  (1995).
34.  H. Meixner, J. Gerblinger, U. Lampe, and M.  Fleischer, "Thin-Film Gas Sensors Based on Semiconducting
    Metal Oxides", Sensors and Actuators B 23, pp. 119-125
    (1995).
                                                 130

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


           RELEVANCE OF ENANTIOMERIC SEPARATIONS IN ENVIRONMENTAL SCIENCE

                                        Daniel W. Armstrong
                     Department of Chemistry, University of Missouri-Rolla, Rolla, MO

A significant number of all organic chemicals that are released into the environment are racemic mixtures. Most
environmental regulations and  scientific environmental studies treat racemic mixtures as though they were
single,  pure compounds. This can lead to incorrect lexicological, distribution, degradation and other data. A
series of new enantioselective analytical techniques have been developed that allow the facile separation and
quantitation of chiral compounds of environmental importance.1 Table 1 shows a typical example of the disparate
biological activities of paclobutrazol stereoisomers.2

               Table 1. Biological activity of paclobutrazol diastereoisomers and enantiomers.
                            Compound               Fungicidal activity        Plant growth regulatory
  c—c-cw—CM—CK-<'  *>—a                            (cereal mildews and rust)  activity (apple seedlings)
     c     i       ~       2RS.3RS                High                    High
                            2R.3R (+)                High                    Low
                            2S,3S(-)	Low	High
Note: The 2R,3R(+) enantiomer has a high fungicidal but a low plant growth regulatory activity. For the 2S.3S (-)
enantiomer the reverse situation holds true. Separation of the enantiomers implies separation of the desired and
the undesired action

We have examined the enantioselective biodegradation of chlorinated pesticides and the herbicides dichlorprop
and mecoprop. While it is known that the herbicidally active enantiomer is the (+) enantiomer for both herbicides,
and fate of each enantiomer in broadleaf weeds and grass has not previously been reported.  Both dichlorprop
and mecoprop are sold as racemic mixtures and are among the most commonly used herbicides for the control
of broadleaf  weeds in grass. This presentation compares the biodegradation of each enantiomer of dichlorprop
and  mecoprop  in several  types  of  broadleaf weeds and common grasses. The results indicate that  one
enantiomer is degraded faster in weeds, while both enantiomers degrade at equal rates in grass.

References
1. D.A. Armstrong, G.L. Reid, M.L. Hilton, C.-D. Chang, Environ. Pollution 79 (1993) 51-58.
2. E. J. Ariens, in Chiral Separations by HPLC, Ed. A. M. Kistulovic, John Wiley & Sons, New York, 1989, Ch. 2,
   pp. 69-80.
                     DEVELOPMENT OF AEROSOL MASS SPECTROMETER FOR
                 REAL TIME ANALYSIS OF PAH BOUND TO SUBMICRON PARTICLES

                                J.T. Jayne, D.R. Worsnop and C.E. Kolb
                      Aerodyne Research, Inc. 45 Manning Road Billerica,  MA 01821
                                 Tel: 978-663-9500, Fax: 978-663-4918
                                       X. Zhang and K.A. Smith
     Massachusetts Institute of Technology, Department of Chemical Engineering, Cambridge, MA 01239
                                 Tel: 617-253-1973, Fax: 617-253-2701

Polycyclic aromatic hydrocarbon  (PAH) are mutagenic pollutants formed as  by-products  of combustion.
Measurements of the distribution of PAH species with different aerosol particles of different sizes are critical for a
complete understanding of the environmental fate of the human exposure to PAH. We present here an aerosol
mass spectrometer (AMS) designed to simultaneously  measure particle size, particle number density and
size-resolved particle composition for volatile and semi-volatile compounds. The instrument combines a unique
aerodynamic sampling  inlet which focuses the particles into a narrow beam and efficiently transports them from


                                                 131

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


atmospheric pressure into a vacuum chamber where particle mass/size is determined by time-of-flight (TOP)
measurement. Size-resolved single particles are flash vaporized on a heated filament and ionized by resonance
enhanced  multi-photon ionization (REMPI) using an excimer laser (248 nm). The PAH ions are detected by a
molecular  TOP mass spectrometer. The ionization process is highly selective for aromatic PAH.

The AMS described  here provides size-resolved single particle composition detection down to 50 nm. Molecular
mass spectra for a number of PAH aerosols  have been  measured. The  sampling efficiency and the size
resolution  of the instrument have been  investigated  using  a differential mobility  analyzer (DMA) and
condensation nucleus counter. Size  resolution defined as ADp/Dp = 4 (where Dp = particle diameter) has been
measured  for tenth  micron size  particles with sampling efficiencies approaching unity, and found to be  about
25%.
          REAL-TIME TRACE DETECTION OF ELEMENTAL MERCURY AND ITS COMPOUNDS

                                           Robert B. Barat
                    New Jersey Institute of Technology, Dep't. of Chemical Engineering,
                        Chemistry, and Environmental Science Newark, NJ 07102

Introduction
Emission of elemental mercury  [Hg] vapor and  volatile mercury  compounds [e.g.  HgCI2,  Hg(CH3)2] from
combustion and other processes  is an important environmental issue  [Von  Burg and Greenwood, 1991]. The
present research  addresses the  need to develop real time  stack monitoring of emissions of Hg and its
compounds. Such continuous emission monitor (CEM) technology would enable identification of peak emission
events and the possibility of corrective action. Realistic emission inventory  will enable regulatory agencies to
better assess health risks associated with future siting of emission sources, such as waste incinerators.

To be  useful to a wide range of applications, the CEM should  be capable of detection in the range of 1-5000
ug/m3, with an ultimate sensitivity limit on the order of 0.1  ug/m3 (ca. 10 pptv). Current best technology for Hg
detection uses cold vapor trap resonant atomic fluorescence [Tekran, 1998].  However, this technology provides
no information on mercury compounds.

In this work, detection of mercury compounds is  based on  photo-fragment fluorescence (PFF) excited by deep
ultraviolet (UV) light.  The fluorescence spectra can facilitate identification of the original  mercury compounds.
Photo-fragment fluorescence has been successfully applied for the gas-phase analysis of HgCI2, Hg(CH3)CI, and
Hgl2 (Barat and Poulos, 1998).

The detection for Hg uses Doppler-shifted resonant  atomic fluorescence excited by a UV laser or a low pressure
mercury lamp. The Doppler-shifted fluorescence will be separated from the unshifted background signal by use
of an optically dense  Hg vapor filter precisely matched to the spectral linewidth of the source. An alternative to
the vapor filter is precise use of time  gating  to distinguish the nearly instantaneous background scattering from
the relatively long fluorescence signal.

The experimental program involves three stages.  In the first, a static cell (no flow) containing  mercury (elemental
or compound)  vapor is probed with a deep UV laser to generate spectra. An atmospheric pressure flow cell is
used in the second stage. The third stage utilizes the expanding jet.

Compounded Mercury
In general, mercury compounds absorb light  strongly below 250 nm [Gowenlock and Trotman, 1955], and these
absorption bands are generally dissociative. In PFF,  a photolyzing UV photon dissociates the target molecule
into fragments, some of which are imparted with excess energy. The energy might then be lost by fluorescence:

                                hv
                           A-B-->A + B*                                (1a)
                            B* — > B + hv'                                (1b)

                                                 132

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


where A and B can be atoms or polyatomic species. The fragment identities and distributions, as revealed in the
fluorescence spectrum, can, in principle,  provide information on the parent species, in a manner analogous to
mass spectrometry and other fragmentation spectroscopies.

For example, Figure 2 shows the PFF spectrum from 193 nm excimer laser excitation of Hg(CH3)CI vapor in a
static cell (Figure 1). Two  features are evident: atomic  Hg  emission lines at 546 and 579 nm; and a  broad
continuum assigned to the B->X system of HgCI** excited state (Mandl and Parks, 1978; Whitehurst and King,
1987). Likely photochemical processes resulting in these observations are:

                                    hv
                        Hg(CH3)CI	> HgCI** + CH3                        (2)
                            HgCI**	> HgCI + hv1                           (2a)
                                  	> Hg + Cl                             (2b)

Steps (2a) and  (2b) are clearly implicated, but the origin of the Hg atomic emission  lines is open to question.
Focussing the laser beam  increases the emission from  Hg, and generates the blue emission at 431 nm from
CH*.

The lowest concentration of mercury compound [Hg(CH3)CI] measured in the static cell was approximately 50
pg/m3. Using reasonable improvements in optics and electronics, it is estimated that the limit-of-detection can be
lowered by at least a factor  of 500.

The concentration of the target compound is related to the fluorescence intensity from a hot fragment. Excitation
laser energy of about 2 mj/pulse  at a repetition rate of 10 Hz at 222 nm was applied to HgBr2 vapor (in Argon) in
an atmospheric pressure flow cell  (Figure 3). The PFF was monitored using a photomultiplier tube and narrow
interference filters centered at 254 nm - a strong Hg* emission line - (see Figure 4).  Excellent linearity was
obtained over a wide concentration range. Similar results were obtained using a compact monochromator + CCD
system.

The supersonic jet spectroscopy  (to be discussed in the next section) is expected to further improve sensitivity of
PFF detection.  Spectra will be  sharpened,  leading to better discrimination of fragment vibrations!  structure.
Quenching by O2 will be reduced due to the low pressure.

Elemental Mercury
Atomic Fluorescence Spectroscopy (AFS) is a highly sensitive spectroscopic marker for elemental Hg detection.
Current AFS  instruments (e.g. Tekran, 1998) use a cold vapor trap for collection + concentration of the air
sample,  purging  (to remove O2), desorption,  excitation  with  an  Hg  vapor lamp (253.7 mn), and then
measurement of the resonant fluorescence  (at the same wavelength). Sensitivity is limited by the elastically
scattered light from the  exciting source.

The technique under study, shown in Figure 5, will expand the Hg-contaminated air stream across a supersonic
nozzle into a high vacuum  chamber. Light at 253.4 rim will be directed across the jet. Atomic Hg fluorescence
will be Doppler-shifted  by between 1 and 3 GHz due to the jet motion. Total  collected  light, comprised of the
shifted fluorescence and stray elastic scattering, will be passed out of the vacuum chamber and through (Figure
6) an optically dense, sharp cut-off Hg vapor filter (centered at the excitation wavelength) to reduce  elastic
scattering while transmitting the fluorescence signal (Miles, 1991; Finkelstein, et al., 1994).

The expansion reduces the  collision rate of Hg* with O2, so collisional quenching and collisional broadening are
both substantially  reduced. The low temperature associated with this expansion further reduces the fluorescence
linewidth, enhancing the performance of the Hg vapor filter.

In the absence (or in conjunction with) the atomic filter, time gating of the collected signal offers an alternative
means to extract the  fluorescence signal  from the  scattering  background. It  has been observed  in the
atmospheric pressure flow cell experiments with HgBr2 that scattering is essentially instantaneous, lasting for the
duration of the laser pulse (i.e. 10 nanoseconds). The relatively long lifetime for Hg* (117 nanoseconds at 253.7
nm - Dodd et. al., 1970) results in the  fluorescence signal appearing as a long-tailed shoulder on the scattering
signal. Prudent placement of the  signal detection gate, as well as subtraction of the background signal measured
in the absence of Hg vapor,  should allow for extraction of the desired fluorescence signal.


                                                  133

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Quality Assurance / Performance Assessment
For comparison of time-averaged test concentrations, the mercurycontaining sample stream will be diverted to a
reference method, such as optical absorption. Testing of the research technology will consist of obtaining data on
five performance measures:

Relative Accuracy: the absolute mean difference between the metals concentration determined by the monitor
and that determined by the reference method,  plus a 2.5 percent uncertainty confidence coefficient based on a
test series.

Calibration Drift: the difference in the monitor output reading from the established reference value after a stated
period  of operation. The  reference value is established by  a  calibration standard which has a  concentration
nominally 80 percent or greater of the full scale reading capability of the monitor.

Zero Drift: calibration drift when the reference value is zero.

Response Time: the time interval between the start of a step change in the concentration of the monitored gas
stream and the time when the output signal reaches 95 percent of the final value.

Detection Limit: three  times the standard deviation of nine repeated measurements of a low-level (near blank)
sample.

Acknowledgement
The  author thanks Dr. Arthur T. Poulos, President of Poulos Technical Services,  Inc., for  his significant
contributions to this research.  The author would also like to thank the  U.S.  Environmental  Protection Agency -
National Center for Environmental Research and Quality Assurance for its financial support of this work (Grant #
R-825380-01-0), and grant project officer Mr. William Stelz for his administrative guidance.

References
Barat,   R.B.  and  Poulos,  A.T.,  "Detection of  Mercury  Compounds  in  the  Gas   Phase  by   Laser
   Photo-Fragmentation/Emission Spectroscopy," Applied Spectroscopy (accepted -1998).
Dodd, J. N., Sandle, W. J., and Williams, O.M., J. Phys. B: Atom. Molec. Phys. 3, 256 (1970).
Finkelstein, N., Gambogi, J., Lempert, W.R., Miles, R.B., Rine, G.A., Finch,  A. and Schwarz, R.A., Proceedings
   of 32nd Aerospace  Sciences Meeting, Jaunary, 1994,  Reno, NV, Paper #A1AA 94-0492.
Gowenlock, B. G.  and  Trotman, J., J. Chem. Society, pt. 2,  1454 (1955).
Mandl, A. and J.H. Parks, Appl. Phys. Let. 33 (6), 498 (1978).
Miles, R.B., "Absorption Line Filter Window and Method for Velocity Measurements by Light Scattering," U.S.
   Patent #4,988,190 (1991).
Tekran, Inc., 1-132 Railside Road, Toronto, Canada.
Von  Burg, R. and Greenwood, M.R., in Metals and
   Their Compounds in the Environment,  ed. by E.
   Merian, VCH, Weinheim (1991).
Whitehurst, C.  and T. A. King,  J. Phys. D: Appl.
   Phys. 20, 1577(1987).

                                                                                             to
                                                                                          manifold
Figure 1. Research Apparatus #1: Static Cell for
Atomic and Photo fragment Fluorescence
                                                 134

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Figure 2b. PFFS of Hg(CH3)CI
193 nm excitation - focused
                                                                 Figure 2a. PFFS of Hg(CH3)CI
                                                                 193 nm excitation - not focused
                                           sen
                               1.2 -

                               I. I -

                                1 -

                               0 S -

                               08 -

                               07-

                               06-

                               01-

                               0 ^ -

                               03-

                               02-

                               0 I -

                                0
     Ar, Air,
      orN2
     Frits
   Heated
    Tube
     Solid
    Samples
 NOT TO SCALE
                   Metal
                  compound
                   vapor
  w
-*•
Mono.
                MPC
                                   W
                   Heated
                   Optical
                   Chamber
                            Mono-
                           chromator
                -—i  r—o
            —i—        I   vent
            ^"    U
                                        Trap
                                                                   Figure 3.  Research Apparatus #2: Flow
                                                                   Cell for PFF
                            Beam Stop
                                     M « mirror
                                      L-lens
                                 MPC • multi-pass cell
                                                    135

-------
                          WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
    4 -
  1  3
  w
  u.
  t
     2 -
     1 -
                                                         Figure 4.  Photofragment Fluorescence (PFF) Signal
                                                         vs. HgBr2 Concentration
                 2000       4000

                  HgBr2 Concentration (
                                       6000
                    8000
Figure 5. Apparatus #3: Atomic Fluorescence and
Compound PFF in Vacuum
      NARROW
     LINEWIDTH
       LIGHT
    (on resonance)
dilution
nr °" \












HO
Saturates



Diffusion
pimp

nozzle
\
Hg
Compound .
Saturator
pr«oture
valve

III



VB.jm
Stop
"•--|- - ~"
1




r
i
i
i
to



r


r*w
P^1
mchMkal
1 — 1 vacuum
cold PUTOP

Vacuum
Chamber
j !__.] windows
1 L±b
UV source t 1
vent 1 1
cold
trap



Claaer) £-
Mhe.m
-W — i
^f — 1
n •
Hg Atomic Fitter
(Princeton)
lor$ pass
ptpcul fitter
CCDor

to data
acquisition

L.J
j
PMT


    BACKGROUND
    SCATTERING
    FROM WALLS
   AND WINDOWS
     ATOMIC
   FLUORESCENCE
1  DOPPLER-SHIFTED
                        TRANSMISSION PROFILE
                         OF ATOMIC Hg FILTER
                                                 Figure 6. Doppler-shifted Resonant Fluorescence
                 FLUORESCENCE
                SEEN BY DETECTOR
                                     FREQUENCY
                                                       136

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium



                      ORTHOGONAL BACKGROUND SUPPRESSION TECHNIQUE
                          FOR EPA'S FIELD INFRARED DATA PROCESSING

                                              Blaterwick

                                     NO ABSTRACT AVAILABLE
                     DEVELOPMENT OF A CONTINUOUS MONITORING SYSTEM
                               FOR PM10 AND COMPONENTS OF PM25

                              Morton Lippmann. Judy Q. Xiong and Wei Li
              New York University Medical  Center, Nelson Institute of Environmental Medicine,
                              57 Old Forge Road, Tuxedo, New York 10987

ABSTRACT
While particulate matter with  aerodynamic diameters below 10 and 2.5 urn (PMio and PM25) correlate with
excess mortality and morbidity, there is evidence for still closer epidemiological associations with sulfate ion, and
experimental exposure-response studies suggest that the hydrogen ion and ultrafine (PM015) concentrations may
be important risk factors. Also, there are measurement artifacts in current methods used to measure ambient
PMio and PM25, including negative artifacts because of losses of sampled semivolatile components (ammonium
nitrate and some organics) and positive  artifacts due to particle-bound water. In order to study such issues, we
are developing a semi-continuous monitoring system for PMio, PM25,  semivolatiles (organic compounds and
NH4NO3), particle bound water, and other PM2s constituents that may be causal factors. PM« is aerodynamically
sorted into three size-fractions:  1) coarse  (PMi0-PM25); 2) accumulation mode (PM2.s-PM0is); and 3) ultrafine
(PMo.-is). The mass concentration  of  each fraction is measured in  terms of the linear  relation  between
accumulated mass and pressure drop on  polycarbonate pore filters. The PMois mass, being highly correlated
with the ultrafine number concentration, provides a good index of the total number concentration in ambient air.
For the accumulation mode (PM25-PMois), which contains nearly all of the semivolatiles and particle-bound water
by mass, aliquots of the aerosol  stream flow  into system components that continuously monitor sulfur  (by flame
photometry), ammonium  and nitrate (by chemiluminescence following catalytic transformations to NO), organics
(by thermal-optical  analysis) and particlebound water (by  electrolytic hygrometer after vacuum evaporation of
sampled particles). The concentration of I-T can be calculated (by ion balance using the monitoring data on NO3,
NHV, and SO4=).

OBJECTIVES
Background
Particulate  matter  (PM) is an  ambient  air criteria pollutant that does not,  as listed,  have  any specific
compositional definition. When initially defined (in  1971) as total suspended particulate matter (TSP), it had  no
specific particle size distinction either. The TSP inlet cut-size was determined by  the inlet aspiration efficiency,
whose upper 50% cut-size  (20-50 gm) varied with ambient wind speed and direction. The 1987 PMio National
Ambient Air Quality Standard (NAAQS) revision was defined in terms of the mass concentration of PM aspirated
by a  non-directional and  wind  speed insensitive inlet with a 50% sampling  efficiency at  -10 urn aerodynamic
diameter. The PMio cut approximates that  of the normal human upper respiratory tract during oral inhalation, so
that the sampled particles represent those  that can penetrate to the thoracic airways (tracheobronchial tree and
more distal gas-exchange airways).  Particles depositing  along  the  tracheobronchial  tree  (lung conductive
airways) can exacerbate asthma and cause bronchitis and  bronchial  cancer, while those depositing in the
gas-exchange airways can cause  lung fibrosis, emphysema and  peripheral lung  cancers.  Particles that are
retained in the lung airways can cause  persistent local irritation and/or dissolve  and be translocated to more
distant organs via the bloodstream.12

Neither the  PMio nor the TSP NAAQS made any distinction as to the chemical composition of the particles
sampled. In terms of composition, there is  a  relatively clear distinction between coarse and fine particles in the
ambient air, as illustrated in Figure 1.
                                                 137

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
The  coarse  mode  of the ambient aerosol is generally composed of soil and  soil-like dust and ash particles
generated and/or dispersed by mechanical forces. It is typically dominated by basic mineral oxides in particles
larger than ~1.0 urn. By contrast, the fine mode particles are generally derived from gas-phase precursors, which
form in the atmosphere as ultrafme (nuclei mode) particles after chemical transformations. The transient nuclei
mode particles rapidly coagulate and coalesce to form larger accumulation mode particles in the light-scattering
                                                         range (0.1 to 1 pm), are persistent in the air for
                                                         many days, and contain almost all of the sulfate,
                                                         ammonium and hydrogen ion, as well as most of
                                                         the  nitrate   ion  and   carbonaceous  particles.
                                                         Size-selective ambient  air  samplers have been
                                                         developed to collect the coarse  mode  and fine
                                                         mode  particles  separately,  typically   with  a
                                                         relatively sharp size-cut at 2.5 urn in aerodynamic
                                                         diameter (PM25).  When using a 2.5 urn  cut, the
                                                         fine mode includes  essentially all of the  sulfates
                                                         and organics, but also includes  the  lower tail of
                                                         the coarse mode.
          0.01
           0.1
                        DP (urn)
Nuclei Mode Accumulation Mode
        »< 	^^_
                                         10
                                      Coarse Mode
100
         Fine Particles
                              -x-
                                    Coarse Particles
        TSP, Total Suspended Paniculate Matter
            PMz.5
                                       	>•
                                  Coarse Fraction
                                                         Figure  1.  Measured  mass  size  distribution
                                                   	   showing  particles in nuclei and  accumulation
                                                         modes  of  fine   particles.  Also   shown  are
                                                         transformation  and  growth  mechanisms  (e.g.,
                                                         nucleation, condensation, and coagulation).

From some epidemiological studies, there is suggestive evidence that  the excess  mortality and morbidity
associated with elevated PM was more closely correlated with PM25 than with PMio.3 The fine particle mass in
the eastern  U.S. is  dominated by  sulfates (SCv), and  the  epidemiological evidence  suggests that excess
mortality and morbidity are as well or better correlated with ambient SO4= than with PM25.4 The  SO4= is, in turn,
highly correlated with aerosol  acidity ((-T), a  more likely causal factor than SOC on the basis of controlled
exposure studies in humans and animals.4

At this time, the actual causal factor(s) for the excess  mortality  and morbidity  are  not clearly established.
However, in  1997, the EPA adopted both 24 hr and annual average PM25 standards on the basis  that PM25 is the
best currently available surrogate index for the health effects associated with ambient air PM.5 At the same time,
EPA retained the PMio NAAQS, with some relaxation  in stringency, because  of residual concerns about the
health risks  from the coarser particles that deposit on lung conductive airways, and may cause or exacerbate
asthma, bronchitis, or bronchial cancer.5

Currently available monitoring methods,  which are based on filter sampling of ambient  PMio or PM25, with
subsequent  determination of sampled  mass,  have significant  limitations for  the accurate determination  of
ambient PM mass concentrations. Some  of the ambient PM is semi-volatile, especially ammonium nitrate and
some of the organic  constituents. Sampled PM mass can be lost, especially when the temperature is elevated
during  the sampling  and/or prior to analysis.  Nitrate (NO3-)  can also be lost from  the sampling filter after
collection of acidic sulfates as the H+ combines with NO3 to form nitric acid vapor (HNO3) that is carried off by air
passing through the filter. There can be positive artifacts as well, most notably due to water of hydration at high
ambient humidities. Their associated water may contribute to the measured fine particle mass without adding to
the health risk.  In view of these  considerations, there is a need to  be able to determine the concentrations of
several key species within the accumulation mode particles that may be either candidate causal factors or likely
sources of sampling artifacts.  The  most important of  these  aerosol  components are SO4=,  NO3, NH4+,  H+,
semi-volatile organics (SVOC) and water vapor.

There is  also some concern about the  health  effects of ultrafine particles (those  less than 0.1 um diameter).
Some animal studies indicate that extremely small mass concentrations of ultrafines 25 ug/m3) can cause excess
mortality  and pathological changes after brief exposures67 Evidence  for an  important role  for ultrafines is
reported  by  Peters et al.,8 who found somewhat closer associations between  reduced  pulmonary function in
nonsmoking  adult asthmatics with number concentration than with PM25 mass concentration. Finally, it remains
                                                  138

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
possible that the effects of PM25 may really be due to its total mass concentration and not to any of its specific
chemical constituents. Thus, it is important to accurately determine the overall mass concentration.

Design Objectives
These  considerations lead to our design  objectives for the continuous monitoring of PMio and  PM25. The
concentrations of components to be measured separately are: 1) coarse mode particle mass (i.e., PMi0-PM25); 2)
fine mode sulfate (SO4=); 3) fine mode nitrate (NO3~); 4) fine mode ammonium (NH4+); 5) fine mode organic
carbon (OC) and elemental carbon (EC); 6) fine mode water (H2O); 7) fine particle  mass (PM25); and 8) ultrafine
mode mass (PM015).

The sum of mass components 1 and 7 equals PMio mass.  I-T, a PM parameter of interest  can also be derived
from the monitoring data. As shown in Figure 2, the milliequivalent sum of the accumulation mode anions (SCv
and NO3~)  is equal to the sum of the cations (NH4+ + H+).  Thus,  we can estimate H+ concentration by net ion
difference.  For  ultrafines,  the  parameter of  interest may  be  number  concentration  rather than  mass
concentration. However, as indicated in Figure 3, the mass and number concentrations for particles below 0.1
urn are highly  correlated.  Thus, the ultrafine number concentration can  be  reliably estimated from  the
corresponding measured mass concentration. Alternatively,  the  number concentration can be  directly measured
using a condensation nucleus counter (CMC).
1-300.
•E
^f 200.
z
•«:
a) Albany. NY ''
(n=140)
f
„ ttf 1:1 line
*2''
JJV* f=0.94
r8 	
300 .
200 .

100 -
0 J
b) Buffalo, NY ,**
(n=178) *'
*
*
4
* <»)*'
j|/P *
•J^ r=x0.96
                                                                    400
                                                                    300.
                                                                    200.
                                                                    100
                                                                         c) White Plains, NY
                                                                            (r*=164)
                                                                                0   '
             *     •
       • jr*
                                                                                       -0.94
                       100   200   300   400    0    100   200   300  400   0

                                               Measured H + (nmol/m3)
                                                                            100   200   300   400
 Figure 2. Comparison of estimated H+ (via ion difference) to directly measured H+ (via pH) for data collected on
 Teflon filters at 3 NYDEC sites (June-August, 1988 and 1989). Unpublished data from Dr. G.D. Thurston, NYU.
160,000
140.000
1W.OOO
100,000
80.000
60,000
40*00
20,000
n
•
0 '
•
'•
m • •
• Qj? "•" *"
• • n
3 • a D n
\ °a$S?f ff 0
U 	 ,_ 	 , 	 . 	 . 	 .
                              2.00    4.00     fi.00

                             Volume i.1nm (|«n3cm-3)
  SO     100     ISO

Volum* < 2.5 iim (jim3cm-3)
                                                                                     200
 Figure 3. Relationships between particle number and volume concentrations: Lett panel for ultrafine particles
 (smaller than 0.1 urn); right panel for fine particles (smaller than 2.5 M). From: USEPA (1996).

 Design Concept
 The overall design concept is illustrated schematically in Figure 4.

 The PM10 inlet limits access to those particles that can penetrate into the human thorax. This is followed by a
 virtual impactorwith a 2.5 urn cut-size. The coarse particle mode is directed onto a spot on a polycarbonate pore
 ucleporeTM) filter tape using the filter resistance method developed by Koutrakis et al.9 After a suitable sampling
 interval and determination of particle mass collected, the tape spot is mechanically advanced for sample storage
                                                  139

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


and presentation of a fresh filter surface for the next sampling interval. The  fine particle fraction (suspended in
29 Ipm of the inlet flow)  is carried  into a second virtual  impactor with a  0.15  urn  cut-size (lowest practical
cut-size). The smaller (ultrafine) particles are collected on a filter spot on a sequential  filter sampler for periodic
mass concentration analyses in a manner similar to that  used to  measure the coarse particle fraction. The
extension of the method of Koutrakis et al.9 to ultrafine particles is described in detail later in this paper. The
accumulation mode particles (0.15 to 2.5 urn), suspended in 2.9 Ipm of the inlet air, are directed into: 1) a stream
of 1.8 Ipm leading to the aerosol water detector; and 2) a stream of 1.1  Ipm leading to the inlets of the continuous
detectors for accumulation mode aerosol components of primary interest, i.e., SCV, NO3-, NH4+, OC, and EC. A
third filter tape sampler draws off 0.2 Ipin of the fine particle (PM25) stream into  a filter tape sampler for the
determination of the overall mass concentration of PM2.5.
       Virtual Impactor
        (djO = 15 Mm)
                                                                    2.9 Ip
                                                                 Sulfate
                                                                Analyzer
                                                                                 1

                                             Nitrate and
                                             \mmonium
                                             Analyzer
                                   Organlcs
                                   Analyzer
                      Particle-bound
                      Water Mass
                       Analyzer
             Direct Readout of
             (PM |0-PM2.9)
             Mass Concentration
Direct Readout
of PM 2.5 Mass
Concentration
Direct Readout
of PM0 u Mass
Concentration
Direct Analysisof PMj 5-PM 0.15 Components
       Figure 4. Overview of the system for continuous measurement of PM 10 and components of PM25.

With the exception of the aerosol water and  paniculate mass detectors,  each of the continuous monitors uses
well established detection methods available in widely used commercial instruments.  Similarly, the PMio inlet
and 2.5 urn  virtual  impactor  are widely  used and commercially available.  A  0.15 urn  virtual  impactor was
designed and tested  by Sioutos et al.10

Monitoring System Elements Based on Commercial Available Equipment
1) Flame-Photometric Detector (FPD) for Aerosol Sulfate: The Meloy Model 285 FPD  can be used to measure
total concentration  of sulfur in the aerosol by using  PbO diffusion denuders at the  inlet to remove ambient
vapors, such as SO2, H2S, and mercaptans. The sulfur in the aerosol is, with rare  exceptions, due to its presence
in sulfates (H2SO4, NI-UHSCX,, and (NH4)2SO4). Thus, the sulfate ion mass concentration is  essentially equivalent
to three times the measured sulfur concentration. This application of the FPD has been  described by Cobourn et
al.11 and Allen et al.12 The instrument detection limit is 1  ppb (4 ug/m3 SCv) with a sampling flowrate of 180
ml/min. Since the sample is preconcentrated  10 times by  means of a 0.15 gm virtual impactor before analysis,
the detection limit is  ~0.4 ug/m3 for accumulation mode particulate SO4 in ambient air.

2) Thermal-Optical Technique for Measurement of Aerosol Organic Carbon (OC) and  Elemental Carbon (ECli
For measurement of aerosol organic and elemental carbon,  we plan to adopt an in situ aerosol carbon analysis
method developed  by Turpin  et al.13 The  method combines the sampling function of a  two-port parallel filter
sampling technique with the analytical function of a thermal-optical carbon analyzer,14'15  and was employed in the
Carbonaceous Species Method Comparison Study (CSMCS) in Glendora, California, in the summer of 1986, for
side by side  measurement of  sub-2.5 gm aerosol carbon  concentrations with  other conventional sampling and
analysis methods.16 The detection limit of the method was reported to be as low as 0.2 ug carbon with a precision
of about 3%.
                                                   140

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
3)    PMio Inlet and PM?5 Virtual Impactor: We use 16.7 liter/min (Ipm) inlets and 2.5 um virtual impactors from
Series 241 Graseby-Andersen PMio Manual  Dichotomous Samplers. They separate the 16.7 Ipm of inlet flow
into: 1) a 1.7 Ipm stream containing the PMio coarse-mode fraction along with 10% of the fine fraction; and 2) a
15 Ipm stream containing 90% of the fine PM fraction.

Monitoring System Elements Under Development and/or Undergoing Evaluation Detail in this Research
1)  Particulate Mass Concentration Monitors:  We are using the newly developed method of Koutrakis et al.,9 in
which the mass accumulated  on a  polycarbonate  pore (Nuclepore™) filter can be  shown to  be directly
proportional to the pressure drop across that  filter. The basis for the method is that the particles are collected at
the entries to and within the pores of the Nuclepore™ filter by interception or Brownian diffusion,  rather than on
the surfaces between the  pores by impaction. The flow through the pores is restricted  by the presence of the
collected particles in proportion to their volume.

For the coarse particles (PMio-PM25), a Nuclepore™ filter with 10 urn pores is being used, while for the fine mass
2  urn pores are used. In  order to measure the mass  concentration of  ultrafine particles, we use a modified
system that uses a filter with 0.2  um  pores. The  0.2 urn pore filter
has a relatively high  baseline flow resistance, and very small mass                   1
increments due to collected ultrafines will  markedly increase the               h -J-
resistance, providing  a  very  sensitive  measure  of the  mass
concentration of ultrafine particles.
Design and Validation of the Ultrafine Mass Monitors
The ultrafine, mass monitor  (CPMM-U)  consists  of  two parallel
channels, four capillary pore filters (N1  ... N4), and two HEPA filters
(Figure 5).  In addition, needle valves and flow meters are used to
control and  monitor the flow rates in each channel. Two sensitive
pressure transducers (T1 and T2) are used to measure the change in
pressure drop at two locations along each  channel as shown in
Figure 5. The measured pressure drops can be related to the mass
loading of the first capillary pore filter of the measurement channel.
The left channel, which has a capillary pore filter exposed to ambient
particles, is the measurement channel. The right channel, with both
capilliary  pore  filters  behind  the  HEPA  filter,  is the  reference
channel.
 Figure 5. Schematic layout of the basic elements of the system for
 the measurement of the mass concentration of the ultrafine function
 of ambient  air particulate matter. The  increased  flow  resistance
 across the N1  Nuclepore filter is proportional  to the accumulated
 mass and number of ultrafine particles.
 The pressure drop across a capillary pore filter is affected by relative humidity, temperature, flow rate, and the
 static pressure at the entrance of the filter. The effect of any one of these factors can exceed the change  in
 pressure drop due to particle loading. While N2 serves as a reference to eliminate the fluctuations in relative
 humidity, temperature and flowrate, a one-channel design has serious limitations. First, since the range of the
 change in pressure drop across N1 is very small (<5%) in comparison to the overall pressure drop of the filter, it
 is difficult to accurately measure the change in pressure drop due to particle loading. Also, it cannot be assumed
 that N2 operates  at the  same conditions as  N1.  This requirement is  approximately satisfied for humidity,
 temperature  and  flow rate. However, the pressure at the  entrance of  N2 may be different if N1  causes a
 significant pressure drop,  as in the case of using a  capillary pore filter with a small pore size. A two-channel
 design greatly improves on these limitations.

 To  measure the mass of ultrafine particles, a very small  pore size is  needed  in order to  match the size  of
 particles of interest and optimize the sensitivity of the instrument. This results in a very high baseline pressure
 drop The typical pressure drop in the Koutrakis9 CPMM design using  a 2 um pore size was -12 inches of water,
 while the pressure drop in CPMM-U using a 0.2 um pore size is -70 inches. Two problems arise: 1) the difficulty
                                                   141

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


of balancing  both  pressure transducers as the sampling time increases (use of a pressure transducer with a
broad measurement range limits the sensitivity); 2) the difficulty of maintaining a leak-free system increases (a
small leak in the system will cause errors in mass measurement).

Only those particles depositing at the entrances to, or inside of the pores of, a filter can contribute to the increase
in pressure drop Therefore, the flow rate was selected to minimize the impaction of particles on the surface of
the filter and to maximize the diffusion and interception of  particles inside of the pores. Other factors important
in the selection include: 1) ensuring that the change in pressure drop across N1  can be measured  in a reasonably
period of time for  the typical ambient particle concentration; 2) achieving a pressure drop that  is not too high
(causing operational problems). A capillary pore filter with 0.2 urn  pore  size and a face velocity of 5 cm/sec was
selected.18

Koutrakis  et  al9 have shown  that,  for the CPMM, the  increase  in  pressure drop of  N1 can be calculated by
2T1-T2 One of the assumptions is that the  pressure drop of N2 is the  same as that of N1. This condition  is
satisfied when using capillary pore filter of  larger pore size (2.0  urn). However, this condition is not satisfied when
0 2 urn capillary pore  filters are used, because the pressure drop across a capillary pore filter depends upon the
flow rate and pressure at the filter face. Because of the pressure drop across N1 (-70 in. of water), the pressure
drop  across  N2 is only about 85% of that across N1.  Generally,  the pressure drop across N1, AP, can be

expressed as,                               Ar_ H-K-n


                                                  _ APft/2
                                                a - APW1

where T1  and T2 are  the pressure differentials recorded  by pressure transducer 1 and 2. APNi and APN2 are the
pressure drop across  N1 and N2 at a given flow rate, respectively. For a face velocity of 5 cm/sec, the a has a
value of 0.85.

As aerosol enters  the CPMM-U, it is dried by passing through  a diffusion dryer. The flow then splits.  On the left
path, particles deposit on N1  , causing a  increase in flow restriction  across N1. Any particle  that penetrates N1
will be removed completely by the HEPA filter that is located further down the line. Therefore, the flow restriction
of N2 will not change due to particle loading during sampling. On the right path, particles are immediately
removed  by the HEPA filter. The flow restrictions of both the first  and second capillary filters will  not change due
to particle loading during sampling because: 1) the capacity  of  the HEPA  filter is much larger than that of a
capillary filter; and 2) the pressure drop across the HEPA filter is less than one hundredth of that across a
capillary pore  filter under our experimental  conditions. Thus, the increase  in  flow restriction of HEPA filter is
negligible in  comparison to that of the capillary pore filter.

                                                As particles deposit  on N1,  the balance of the system is
                                                self-adjusted  to  accommodate the change in flow restriction
                                                of N1. Thus, the pressure drop across the N1 increases and
                                                the  T1 reading  increases.  Assuming that the flow rates of
                                                 both lines  are not significantly changed  during the period of
                                                sampling,  the T2 reading also increases. For the calibration
                                                 of the  ultrafine  particle  mass  monitor  (CPMM-U)  we
                                                 employed   a method   using  an   Ultrafine Condensation
                                                 Particle Counter and monodisperse particles to calibrate the
                                                 CPMM-U.  The   system  used is  described  in  a separate
                                                 paper.19 The  results of the calibration tests for various sizes
                                                 of monodisperse ultrafine particles are shown in Figure 6.
1
i
"3
a
z
1
o
s
Q.

Q
p
1
£
.g

8)

/
6

5

4

3


2

1


n
. i i | < i i | . i . | i . | . i i | i i • l i i y^
7 ]12 nm Q-'
y-0.136+0.0492«, R. 0.992 f.- J
.'*
j,-'
^.O 95.0 nm "
L ..- ' -|
: .••"" :
— ft" —
: f.-''o 753 nm :
; -t5 66.8 nm '.
51.3 nm ..•' 66'8 "m "
• °.-'' "
_.•" _:
O" 42.0 nm
: P 35.3 nm
"0 31.7 nm, ..,...,... | ... | 	 -
            20    40    60    80   100
                 Mass Concentration (ug/m3)
                                       120
                                            140
Figure  6. The results  of the  calibration tests for various
sizes of monodisperse ultrafine particles.
 Since the pressure drop across the capilliary pore filters in the CPMM-U  is 5-7 times higher than that in the
 CPMM, the filter tape advancement system for CPMM was considered unsuitable for the CPMM-U. Koutrakis et
 al.20 have shown that a  small  leak in the  CPMM can cause a significant error in the measurement of mass
                                                    142

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
concentration. To avoid leakage problems we use a multi-head filter holder that can provide a tight seal for each
filter. Instead of a filter advancement system, we  use a sequential  sampling approach. Figure 7 shows the
systematic design  of the multi-head filter holder. The multi-head filter holder is machined from  an aluminum
block. Recessed filter holders (the number depends on the optimization of the CPMM-U) are machined into the
block. Each filter holder includes a stainless steel screen backup and  a gasket. The solenoid valves, which are
controlled by a  computer,  select one sampling channel at a  time. The capilliary pore filters are loaded into the
filter holder block in the laboratory, and the multi-head filter holder is replaced as a whole in the field. Since the
linear range of the instrument extends to a mass concentration up to 20 ug/m3, one filter may  be  used for long
period sampling if the concentration of particles is relatively  low. For example, if the mass concentration of the
                                                         ultrafine  fraction is 1  ug/m3, the linear response
                                                         range  will  not  be exceeded until  20 one-hour
                                                         consecutive  samplings  are  made on one  filter.
                                                         Therefore, a multi-head holder containing  14 filter
                                                         holders  should  be  sufficient  for  a  one  week
                                                         sampling under  almost all ambient  conditions in
                                         Filter holder block  the U.S.
                 Out
                                     Solenoid valve
                                                         Figure 7. Systematic design of the multi-head filter
                                                         holder.
                    In
2) Particle-Bound Water Detection System: The basic problem that must  be overcome is the extremely high
background water  vapor concentration in the  air compared  to  the water  bound  by particles at normal
environmental  conditions.  Due to the  rapid  equilibrium  between  water vapor and  particle  surface  water
(milliseconds time-scale), there is no  conventional method for separating  the  particle-bound water and  its
coexisting  vapor without disturbing the phase  equilibrium. The innovative concepts of our method are: 1) to
collect particles over a relatively short time period (compared with the time  scale of environmental variation);
and 2) to collect the particles without disturbing the water equilibrium between the particle and gas phases (by
maintaining the sampling system at the condition of the ambient environment); and 3) to minimize the sample
cell volume in which the associated air remains. The particle-bound water can then be readily detected above the
background in air by means  of a highly sensitive moisture detector,  such as  a P2O5-Pt electrolytic hygrometer.
The electrolytic hygrometer  was  chosen as a  water detector because: 1)  it has the  lowest detection limits
currently available; 2) it is relatively inexpensive; and 3) it is convenient to operate.

                                                        A schematic diagram  of our semi-realtime analyzer
                                                        for  particle-bound  water  in  accumulation  mode
                                                        aerosol  is illustrated in Figure 8. The basis for the
                                                        technique is the  accretion of PM25-PM015 particles
                                                        by means  of a filter over a  preset period  of time.
                                                        Two identical sample cells, each one consisting of a
                                                        13 mm Teflon membrane filter (2 urn pore size) and
                                                        a small  enclosure, are  connected  in  series.  The
                                                        Teflon membrane  filter was selected due  to its
                                                        excellent particle collection efficiency,  low moisture
                                                        uptake, and low  trace background. The upstream
                                                        sample  cell  collects  and   preconcentrates  the
                                                        particles in the  sampled air from the 0.15 pm virtual
                                                        impactor,  while  the  second cell analyzes  the
                                                        backgrounds from the air and filter.
Aerosol
                                       Pressure Sensor

                                             Vacuum
   A 13mm Teflon filter
   (2 Jim pare size) with
   Teflon support is
   mounted between two
   Teflon washers
                               £   Vacuum Valve

                              Ijjl  Three-way Solenoid Valve

                                   Two-way Solenoid Valve

                                   Filler
Figure  8.  Schematic  diagram  of  the  system  for
continuous measurement of  particle-bound water
within PM2s
A pair of parallel sampling lines  are  used for alternating the sampling  and analysis  processes.  They  are
                                                   143

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
controlled by 2 three-way solenoid valves, the sample inlet valve (S1) and the exhaust valve (S2). When S1 and
S2 switch to line A, the line A starts sampling and the line B starts analysis of the samples collected in the prior
sampling period. At the end of the sampling period, both S1  and S2 switch, and the functions are reversed.
During Line A sampling, solenoid valve SA opens. When switched to analysis, SA closes, and the two sample
cells will  be separated and analyzed in order. The air background goes first and the aerosol sample next. All the
solenoid valves are controlled by an interfacing computer with a preset program.

The  water detector consists  of a 1000 cm3 Vapor Dilution  Cell  (VDC),  an Electrolytic Hygrometer,  a flow
controller, a  pressure sensor, a vacuum line, and a helium gas line. The procedures for analysis of the water
contents  collected by each sample cell are: 1) evacuate the VDC; 2) close the vacuum line and open the VDC to
the cell to be analyzed; 3) extract the  water bound in the particles and absorbed on the filter and draw it into the
VDC along with the air remaining in the cell by vacuum; 4) fill the system with helium gas to a pressure of about
900 torr  (slightly higher than atmosphere);  5) open the system to the hygrometer,  and allow a small flow (10
cm3/min) to pass through the electrolytic sensor; and 6) start one measurement.  The output signal is recorded
and  processed  by an interfacing data acquisition system. The particle-bound water is determined from the
difference between the measurements of the two cells. The filter of the aerosol sample cell is changed after each
run.

A P2O5-Pt electrode is used to sense water vapor concentration in the system. It consists of a quartz tube wound
with  two  platinum (Pt) wires. The winding is coated with a phosphorus pentoxide  (P2O5) film, which has a high
affinity for H2O. A voltage is applied to the  platinum winding so that the water molecules adsorbed on the P2O5
film are electrolyzed to H2 and O2 and  a current flow is generated as described below:
        Cathode reaction:
        Anode reaction:
                              2 H2O+2e- -> 2 OH + H2
                              2 OH -+ H2O + 1/2 O2 + 2 e
Since every H2O molecule electrolyzed  produces two electrons (based on Faraday's Law of Electrolysis) the
current is directly proportional to the concentration of the H2O molecules in the gas stream. This correlation is
insensitive to gas pressure and mass flowrate. The sensor is commercially available for sampling moisture in gas
streams and at normal atmospheric pressure. The working range is 0-1000 ppm with an accuracy of 2%. The
lowest reported detection limit is 10 ppb.21 The sample cell volume is 2 cm3 (See Figure 8). For a flowrate of 1.8
Ipm and a sampling time of 60  minutes, the concentration of particle-bound water is elevated by a factor of
5.4x10s in relation to its carrier air stream (including a preconcentration factor of 10 provided by the 0.15 virtual
impactor). Therefore, the particle-bound  water is measured despite the associated vapor in  the air  stream. A
detection limit of the system for measuring particle-bound water is ~5 pg/m3 of total particle mass concentration
with a water composition of 15% at RH  above 40%, and  5% at RH below 40%,  in a sampling period  of 60
minutes.
3) Tungstic Acid  Technique - Chemiluminescent NO* Detector (TAT-CLD) System for Measurements  of
Particulate Nitrate (NCv) and Ammonium (NH/): A system for continuous measurement of particulate NO3 and
NH4+ is shown in Figure 9. The system combines a two-channel Chemiluminescent NOx Analyzer (Monitor Labs,
Model  8840) with the Tungstic Acid Technique (TAT)  developed by Braman et  al.22 The  TAT was used  by
Braman et al.22 for preconcentration and determination of gaseous and particulate nitrate  and  ammonia, based
                                                                   on  the  principle  that  nitrate  and
                                              Exhaust                ammonia  are quantitatively chemi-
                                                                   sorbed at  the  tungstic acid-coated
                                             er"1                    surfaces  at room  temperature, and
                                                                   thermally desorbed at  high tempera-
                                                                   ture (350°C).
He
Tungstic
Acid
Denuder
Tungstic
Acid
Converter
Makeup
Air
i
fc 1-
Molybde
Convefl
                     Remove particles
                         350 °C
                     NO3Xp)—*~ NOjIg)
                          3iO°C
                            » NH3(g)
                               T
                              Makeup
                               Air
                                       Exhaust
                                                           Figure   9.  Schematic  diagram  of
                                                           tungstic  acid denuder and converter-
                                                           chemiluminescent NOX detector sys-
                                                           tem for  continuous  measurement of
                                                           nitrate and ammonium concentrations
                                                           in accumulation mode particles.
                                                  144

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


The advantage of the TAT is that there is essentially no interference of aseous NO* in either the NO3- or NH4+
analyses. The weakness of the Braman et al.22 TAT is the limited particle collection efficiency of the TAF. For a
34 cm long three-section TAF, the particle penetration  was as high as 22% at a sampling flowrate of 1 IpM.22
Large packed tubes can be used to improve efficiency, but this limits the sampling flowrate attainable. In  our
system, the sample air is preconcentrated 10 times prior to entering the analysis system. The sampling flowrate
of the TAT-CLD may be as low as 0.1 - 0.2 Ipm without  increasing the detection limit of the original TAT. Thus a
much higher particle collection efficiency is achieved. A two-channel CLD is used for parallel detection of NO3
and NH4+ in our system, which simplifies the processes of sample separation. It also minimizes the  sampling
artifacts reported by Braman et al.22

The instrument detection limit of CLD is 2 ppb with a sampling flowrate of 250 ml/min for each channel. The
detection limits of TAT-CLD system for ambient accumulation mode particulate  NO3  and NH4+ are below  0.1
gg/m3 for a 30-minute sampling period.

4) Data Acquisition and System  Control:  To automate the detection system and the data acquisition, we use a
computer interfaced data acquisition  and instrument control  system. As  shown in  Figure 10, it consists of  an
IBM-PC computer, an Analog-to-Digital Converter Board (ADC), a four-channel Signal Conditioner,  a Digital
Interface  Board,  and a  Multichannel Relay Board (MRB). The Signal Conditioner converts the signals from
sensors to standard signals (0 - ± 5V, or 0 - ± 10V), which can be accepted by ADC. The signals received from
ADC  are  recorded  and  stored  in   the
computer  at specified intervals. To  control                                   cp  cp  cp cin  senors
the solenoid valves, a  MRB  and  a Digital                              ~  "
Interface is used.
Figure 10. Schematic of the data acquisition
and instrument control system.
                                          L
_L
1 Plug-In ADC
_ \ Board
Cable

Signal Conditioner
H Power
supply
Plug-In Digital
Interface
Cable

Multichannel
Relav Board
DISCUSSION
An  ambient PM monitoring package  (prototype monitor)  has been designed  to  be capable of continuous
operation at our laboratories in New York City and in Tuxedo, NY, a location 50 miles north-northwest (NNW) of
New York City. It will provide  records of concentration data as a fimction of time for PMio and  eight of its
components, including coarse mode particles (PMio-PM25), fine particles (PM25),  ultrafine particles (PM015) and
the constituents of the accumulation mode aerosol (PM25-PMois) of primary interest,  including SCv,  NO3~, NH4+,
l-f, H2O, OC, and EC. Using three different pore sizes for the three mass fractions helps insure that  interception
will be the dominant collection mechanism and that each fraction will be collected within the optimum range for
linear response. Using this monitoring system, researchers will be  able to  measure the mass concentrations of
PMio and PM25 in near real-time, and without sampling artifacts due to sample volatilization and particle-bound
water. They will also  be able to develop data bases for the concentrations of specific components of PM25 that
may be causal factors for the PM-associated health effects of concern (e.g., H+, SCv , OC, EC, and ultrafines),
thereby providing opportunities for more definitive epidemiological  studies.  Such studies could provide the basis
for future NAAQS for specific PM components and thereby more rationale  design and implementation of source
controls of PM  and/or PM  precursors.

ACKNOWLEDGMENTS
This manuscript is  a  condensed version of a  paper that  has  been submitted  for publication in  Applied
Occupational and Environmental Hygiene, and premission to reproduce parts of  it has been graciously granted
by its Editor-in-Chief.

This research was supported by Grant R825305  from the U.S. Environmental Protection Agency, and is part of a
Center Program supported by Grant ES 00260 from the National Institute of Environmental Health Sciences.

The consultation and support of Dr. Petros Koutrakis and his colleagues for our application of the filter resistance
method for PM mass concentration measurement developed  at the  Harvard School of Public Health is also
gratefully acknowledged.

LITERATURE  CITED
1     Lippmarm, M.; Yeates, D.B.; Albert, R.E.:  Deposition, Retention and Clearance of Inhaled Particles. Brit. J.
      Ind. Med. 37:337-362(1980).
                                                 145

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


2.    Lippmann, M.: Size-Selective Health Hazard Sampling. In: Air Sampling Instruments, 8th Ed., pp. 81-119.
      S.V. Hering, Ed. Am. Conference of Governmental Industrial Hygienists, Cincinnati, OH (1995).
3.    U.S. Environmental Protection Agency: Air Quality Criteria for Particulate  Matter.  EPA/600/P-95/001F.
      USEPA, Washington, DC (1996).
4.    Lippmann, M.; Thurston, G.D.: (1996). Sulfate Concentrations as an Indicator  of Ambient  Particulate
      Matter Air Pollution for Health Risk Evaluations. J. Exposure Anal. Environ. Epidemiol. 6:123-146 (1996).
5.    U.S. Environmental Protection Agency: National Ambient Air Quality Standards for Particulate Matter.
      Federal Register 62:38762-38896 (July 1997).
6.    Oberdorster,  G.; Gelein, R.M.;  Ferin,  J.;  Weiss, B.: Association of Particulate Air Pollution  and Acute
      Mortality. Inhal. Toxicol.  7:111-124 (1995).
7.    Chen,  L.C.; Wu, C.Y.; Qu, Q.S.; Schlesinger,  R.B.: Number Concentration and  Mass Concentration as
      Determinants of Biological Response to Inhaled Irritant Particles. Inhal. Toxicol. 7:557-588 (1995).
8.    Peters, A.; Wichmann, E.; Tuch, T.; et  al.:  Respiratory Effects are Associated with the Number of Ultrafine
      Particles. Am. J. Respir. Grit. Care Med. (In press).
9.    Sioutos, C.;  Koutrakis,  P.; and Olson, B.A.  Development and Evaluation of  a Low Cutpoint Virtual
      Impactor. Aerosol Sci. Technol. 21:223-235 (1994).
10.   Koutrakis, P.; Wang, P.Y.; Sioutas, C.;  Wolfson, J.M.: U.S. Patent No. 5571945 (1995).
11.   Cobourn, W.G.; Husar,  R.B.; Husar, J.D.: Continuous In Situ Monitoring of Ambient  Particulate Sulfur
      Using Flame  Photometry and Thermal Analysis. Atmos. Environ. 12:89-98 (1978).
12.   Allen, G.A.; Turner, W.A.; Wolfson, J.M.; Spengler, J.D.: Description of a Continuous Sulfuric Xcid/Sulfate
      Monitor. In:  Proceedings of the National  Symposium  on Recent Advances  in Pollutant Monitoring of
      Ambient Air and Stationary Sources, Raleigh, NC, pp. 140-151. USEPA, EPA-600/9-84-019 (1984).
13.   Turpin, B.; Gary, R.; Huntzicker, J.: An In  Situ Time Resolved Analyzer for Aerosol Organic and Element
      Carbon. Aerosol Sci. Technol. 12:161-171 (1990).
14.   Johnson, R.L.; Shah,  J.J.; Gary, R.A.; Huntzicker, J.J.: An Automated Thermal  Optical Method for the
      Analysis of  Carbonaceous Aerosol.  In:  Atmospheric Aerosol:  Source/Air Quality Relationships, pp.
      223-233. E.S. Macias and P.K. Hopke, Eds.  American Chemical Society, Washington, DC, ACS Symp.
      Ser. No. 167(1981).
15.   Huntzicker, J.J.; Johnson, R.L.;  Shah, J.J.; Gary, R.A.:  Analysis of Organic and Elemental Carbon in
      Ambient Aerosols by a Thermal-Optical Method. In: Particulate Carbon:  Atmospheric Life Cycle, pp. 79-88.
      G.T. Wolff and R.L. Klimisch,  Eds. Plenum, New York (1982).
16.   Hering, S.V.; Appel, B.R.; Cheng, W.; et a].: Comparison of Sampling Methods for Carbonaceous Aerosol
      in Ambient Air. Aerosol Sci. Technol. 12:200-213 (1990).
17.   Marple, V.A.; Rubow,  K.L.; Turner, W.; Spengler, J.D.: Low  Flow Rate Sharp Cut Impactors for Indoor Air
      Sampling: Design and Calibration. JAPCA 37:1303-1307 (1987).
18.   Spurny K.R.; Lodge, J.P.; Frank, E.R.; Sheesley, D.C.: Aerosol Filtration by Means of  Nuclepore Filters:
      Structural and Filtration Properties. Environ. Sci. Technol.  3(5):453-464  (1969).
19.   Li, W.; Xiong, J.Q.; Lippmann, M. The Development of a Continuous Particle Mass Monitor for Ultrafine
      Ambient Particles (In preparation).
20.   Koutrakis, P.; Wang, P.Y.; Wolfson, J.M.: Private Communication (1996).
21.   McAndrew, J.J.:  Moisture Analysis in Process Gas Streams. Solid State Technol. 35:55-60 (1992).
22.   Braman, R.S.;  Shelley,  T.; McClenny, W.A.:  Tungstic Acid for Preconcentration and Determination of
      Gaseous and Particulate Ammonia and Nitric Acid in Ambient Air. Anal. Chem. 54:358-364 (1982).
                 A REAL-TIME SAMPLER RAMS, FOR THE DETERMINATION OF PM25,
                                INCLUDING SEMI-VOLATILE SPECIES

                                F Obeidi, E. Patterson and D.J. Eatough
         Department of Chemistry and Biochemistry, Brigham Young University, Provo UT 84602 USA

The RAMS, Figure 1, is a real-time ambient monitor for the determination of fine paniculate mass, including the
volatile components (Eatough 1998). The RAMS has a particle concentrator, followed by  diffusion denuders to
remove gas phase compounds which can be  absorbed  by charcoal, a Nafion dryer to remove water, and a
"sandwich filter" containing a Teflon coated filter to collect particles and a charcoal impregnated filter to retain

                                                 146

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


volatile components which can be lost from the particles during sample collection. Semi-volatile fine particulate
material retained on the "sandwich filter include ammonium nitrate and semi-volatile organic compounds. The
"sandwich filter" is located at the tip of the tapered oscillating element of a TEOM monitor and mass retained on
the "sandwich filter" is measured as a function of time.

The results obtained with the RAMS have been validated by comparison with results obtained from diffusion
denuder integrated samples, to determine the mass of fine  particulate  material retained on a filter and the
semi-volatile organic material and ammonium nitrate  lost  torn the filter during sampling. This has included
comparisons with sampling  periods for the denuder samplers  as short as 1 hour. Results obtained with  RAMS
and denuder samplers for samples collected  in Riverside CA  in the summer and Bakersfield, CA in  the winter
show that semi-volatile fine particulate species are accurately monitored with the RAMS.

Research is currently underway to validate the measurement of volatile  constituents of fine particles with the
RAMS in chamber experiments  using  well characterized particles of ammonium  sulfate, ammonium nitrate,
glycerol and a carboxylic acid.

ACKNOWLEDGMENTS
The research  reported here was supported by the U.S.  Environmental Protection Agency STAR grant R825367
to Brigham Young University.   Some support to the program has also  been provided  by the Electric  Power
Research Institute, and Rupprecht and Patashnick, Inc.

REFERENCE
Eatough  D.J.,  Obeidi F., Pang Y., Ding Y.,  Eatough  N.L.  and Wilson W.E. (1998) "Integrated and real-time
diffusion denuder samplers for PM25 based on BOSS, PC and TEOM technology," Atmos. Environ., submitted.
                             DYNAMIC NUCLEAR POLARIZATION (DNP):
                 A NEW DETECTOR FOR ANALYSIS OF ENVIRONMENTAL TOXICANTS

                                            Harry C. Dorn
      Department of Chemistry, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061

The  objective of  this study  is the development  of  a new  analytical instrument designed for environmental
monitoring  applications. Specifically, this will consist of  dynamic nuclear polarization detection  with either
direct-coupling  continuous-flow  supercritical  fluid chromatography  (SFC/DNP)  or  recycled-flow  13C  (DNP)
analysis of toxicant mixtures. The  DNP  detector is a variant of the  well known nuclear magnetic resonance
(NMR) phenomena. A salient feature of NMR is the  chemical shift parameter which  provides a very sensitive
probe of the local  electronic environment about a given atom in a molecule. Thus, the DNP detector could have
wide ranging applications  for specific monitoring of various organic toxicants mixtures (e.g.,  chlorocarbons,
organophosphates, pesticides, petroleum  pollutants,  etc.).  A major limitation of NMR for  most environmental
monitoring  applications  has  been  sensitivity  constraints.  The  DNP  approach helps alleviate the  sensitivity
limitation of NMR by transfer of polarization from an electron spin to the nuclear spin of interest (1H, 13C, 31P,
etc.). The corresponding DNP signal enhancements are proportional to the electron-to-nuclear magnetogyric
ratio (YB/YH) which is on the order of 103-104 for most nuclides.

In this presentation, LC/DNP and SFC/DNP results for chlorocarbon mixtures will be presented, In addition, the
results for continuous monitoring of a  mixture of  benzene and several  chlorocarbons with a recycled-flow 13C
DNP instrument will also be presented. Finally, progress towards development of a routine SFC/DNP instrument
will be reported.
                                                 147

-------
        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


PARTITIONING TRACERS FOR IN-SITU DETECTION AND MEASUREMENT
           OF NONAQUEOUS LIQUIDS IN POROUS MEDIA

                            Brusseau

                   ABSTRACT NOT AVAILABLE
                              148

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
  QUALITY
ASSURANCE
      149

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                            150

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


              CD-R ARCHIVE AND CATALOG OF HEWLETT-PACKARD LOW RESOLUTION
                        FORMAT DATA FROM NETWORKED GC/MS SYSTEMS

                                 Robert G. Briggs and Herman Valente
                      The Wadsworth Center, New York State Department of Health,
                       Empire State Plaza, Box 509, Albany, New York 12201-0509
                                         Phone (519)473-4871
                                     robert. briggs@wadsworth. org

ABSTRACT
(1) Taped GC/MS data are not efficiently accessible;  (2)  Cataloging of archived data with descriptive information
requires retyping information entered by the operator at the time of data acquisition. We have addressed these
two problems, in our laboratory by archiving all  of our data, generated on various  Hewlett-Packard GC/MS
systems, on CD-R (Compact Disk, Recordable) media. Data are initially sent to a Mylex Level 5 RAID array; then
a CD is prepared using the Pinnacle Computing RCD-202 and associated RCD-PC software. A Microsoft Visual
Basic program developed in-house (Arch-CD) extracts sample information entered at analysis time from each
data file selected for archive and generates a Microsoft  Access database (.MDB) file. This MDB file is included
with the data files on the CD and also  merged with an integrated database. The database along with the CDs on
a changer provide instant access to five years of GC/MS data.

INTRODUCTION
Our laboratory applies EPA CLP (Contract Laboratory Protocol) volatilcs and semivolatiles methodology to water
and soil samples or their extracts.
•   We have  four GC/MS systems dedicated to this work, three Hewlett Packard MSD systems and one Varian
    Saturn Ion Trap system.
•   Submission rate for samples of all types averages a modest 750 per year. With daily calibration and other
    QC, which proceeds irrespective of sample submission, this translates to  about 3000 GC/MS runs per year.
    The average input rate  belies the fact that  submission is episodic (figure 1) and that samples for a given
    instrument are received at a rate  up to 30 per day, rather than the more prosaic one per day. Turnaround
    requirements cannot be adjusted to sample submission rate so we must be ready to handle the maximum.
•   Prior to 1993, we were limited in disk space, even with a network of PCs,  and data was archived on tape. In
    a major project where non-target analytes are  of concern, it is desirable to reevaluate previously processed
    samples when a new class of substances is detected. The  report time is inordinately extended because of the
    need to move data in and out of storage. Taped GC/MS data are not efficiently viewed or accessed.
•   Once data are sent to long-term  archive, key information should be put in a  database to facilitate future
    access. A catalog of archived data should contain, at minimum, the descriptive information entered by the
    operator at the time of data acquisition, but retyping this information into a database is tedious and should be
    unnecessary.

SUMMARY
We have addressed these two problems of short-term data access and long-term accessibility  in our laboratory.
For five years we have been archiving all of our data, generated on various Hewlett-Packard GC/MS systems, on
CD-R (Compact Disk, Recordable).
•   Data are initially sent to a 1.2 Gigabyte partition on a Mylex Level 5 RAID array.
•   Then they are mastered using RCD-PC software for the Pinnacle Computing RCD-202 onto a  0.8 Gigabyte
    partition on the RAID array.
•   CD-R disks are prepared and transferred to SCSI chained Pioneer 604X/624X six disk changers.
•   CDs beyond the  twelve accessible disks are kept in 6-disk cartridges  for insertion in  the  changer  and
    reasonably rapid access.
•  A  Microsoft Visual Basic program developed  in-house (Arch-CD) extracts sample  information entered at
   analysis time from each data  file  selected for archive and generates a Microsoft Access  database (.MDB)
   file. This is included with the data files on the CD.
•  A master database is built up from individual MDB format files.
•  The success of this system is illustrated for analysis  of volatile organics in water (figure 2).  Although the
   percentage  of total  time from sample receipt to the analytical report dedicated to data processing  has
   remained a constant 80% - 90%, the total time  necessary for analysis and  report has fallen significantly since


                                                 151

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


    implementation of this system, from 40  days (or more)  to 10 -20 days.  During  this time, the rate  of
    submission for samples in this category actually increased (figure 1).
•   The long report time events of 1995 (figure 2)  were due to data glut on our ion trap system, one of the two
    systems used for water volatiles. This low-level ion-trap system was only recently connected to the network,
    has limited  hard disk storage and has depended on tape backup.  We have avoided recurrence  of this
    problem by networking the instrument and archiving its data with those from the HP  systems. We have not
    yet developed  information-extraction methodology to incorporate these  Saturn data  into the  database.
    Hewlett Packard publishes the structure of data files for users whereas Varian does not.

CONCLUSION
Accumulated sample information from a five-year-period is now accessible for HP systems and the data files are
readily retrieved for use.
               THE USE OF ACCEPTABLE KNOWLEDGE FOR THE CHARACTERIZATION
               OF TRANSURANIC WASTE IN THE DEPARTMENT OF ENERGY COMPLEX

                                           R. Vann Bynum
       Science Applications International Corporation, 2109 Air Park Road S.E., Albuquerque, NM 87109
                                 R. Butch Stroud and R. Denny Brown
       Department of Energy, Carlsbad Area Office, 4021 National Parks Highway, Carlsbad, NM 88220
                                           Laurie D. Sparks
                   Waste Isolation Division, Westinghouse Electric Co., a Division of CBS,
                                  P.O. Box  2078, Carlsbad, NM 88221

ABSTRACT
The Resource Conservation and Recovery Act (RCRA) regulations codified in 40 CFR Parts 260 through 265,
268, and 270, authorize the use  of acceptable  knowledge as a method which can be  used in  appropriate
circumstances by waste generators, or treatment, storage,  or disposal facilities to make hazardous waste
determinations. Acceptable knowledge, as an alternative  to sampling and analysis, can be used to meet all or
part of waste characterization requirements under  RCRA.

One example of the use of acceptable knowledge within the U. S. Department of Energy (DOE) complex is the
waste characterization requirements for the  Waste Isolation Pilot Plant (WIPP), a deep geologic repository for
the disposal of transuranic (TRU) waste. The WIPP will be disposing of TRU mixed and non-mixed waste from
various generators within the DOE  complex  in accordance with the provisions of the DOE Carlsbad  Area Office
(CAO) Quality Assurance Program  Plan, the EPA Final Certification Decision, an  anticipated RCRA  Permit, and
the WIPP Waste Acceptance Criteria. Part B of the WIPP RCRA Permit application contains a Waste Analysis
Plan which  describes the measures that will  be  taken by the DOE/CAO to assure that mixed wastes received at
the WIPP repository are characterized appropriately. To satisfy the characterization requirements, acceptable
knowledge  is confirmed by radiography, drum headspace gas  sampling and analysis,  and  solidified waste
sampling and  analysis. Acceptable knowledge is primarily used in  TRU  waste characterization activities to
delineate TRU waste streams, to  determine if TRU debris wastes  exhibit  a toxicity characteristic  (40 CFR
261.24), and to determine if TRU wastes are listed (40 CFR 261.31). The  physical form and the increased health
and safety risks associated with obtaining a  representative sample of TRU debris wastes,  clearly justify the  use
of acceptable knowledge in making  hazardous waste determinations.

DOE complex waste generators apply knowledge of their waste based on the materials and processes used to
generate the waste. Acceptable knowledge  includes information regarding the physical form of the waste, the
base materials composing the waste, the nature of the  radioactivity present, and the process(es) generating the
waste.

The DOE/CAO audits the TRU waste generators to grant TRU waste certification authority to the generators. The
DOE/CAO conducts audits at least annually thereafter to verify ongoing compliance with approved plans  and
procedures  including those for waste characterization  using acceptable knowledge.  The  information  resulting

                                                152

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


from DOE/CAO audits is used by the DOE/CAO to help sites across the DOE complex prevent or solve problems
associated with the compilation and use  of acceptable knowledge for  TRU  waste  characterization. Past and
current issues with respect to the compilation, use, and defensibility of acceptable  knowledge for TRU waste
characterization along with the solutions to those problems and their prevention will be addressed.

INTRODUCTION
Waste Isolation Pilot Plant (WIPP)
The WIPP is a deep geologic repository designed to permanently dispose of transuranic (TRU) radioactive waste
generated from the research,  development, and  production of nuclear weapons in an evaporite salt  formation.
The mission of the WIPP Project, as authorized by the U.S. Congress in  1979  (Public Law 96-164), is  to provide
a research and development  facility to demonstrate the safe disposal of TRU waste generated as a result of
United States defense activities. WIPP is located in Eddy County in southeastern New Mexico about 26  miles
east of Carlsbad,  New Mexico, in an area known  as Los Medanos - a relatively flat, sparsely inhabited plateau
with little surface water. WIPP encompasses a  16-square mile area under the jurisdiction of the U.S. Department
of Energy (DOE) pursuant to the Land Withdrawal  Act.1 The site boundary was established to ensure that at least
1 mile of intact salt  exists laterally between  the  waste disposal area and  the accessible  environment and to
ensure that no permanent residences will be established in close proximity to the facility.

The DOE's objective  is to operate and  maintain the WIPP free of both chemical and  radiological contamination.
Therefore, all waste sampling  and analyses will be conducted by the DOE generator/storage sites in accordance
with the requirements  of the WIPP Waste  Analysis  Plan  (WAP,  Chapter  C  of  the RCRA Part  B Permit
Application)2 and  the WIPP Compliance Certification  Application (Chapter 4)3 as allowed by 20 New Mexico
Administrative Code  (NMAC)  4.1, Subpart V,  Paragraph 264.13, and consistent with joint U.S. Environmental
Protection Agency (EPA) and U.S. Nuclear Regulatory Commission guidance.

Transuranic Waste
The transuranic (TRU)  elements have  atomic  numbers greater  than that of uranium (92). Examples  of
transuranic elements include neptunium (Np), plutonium (Pu), americium (Am), curium (Cm), etc. Each element
typically  has several isotopes and  is produced  during  nuclear reactions.  These man-made elements  are
radioactive and provide the key components for building nuclear weapons.

Transuranic waste consists of clothing, tools, rags,  and other items contaminated  with trace  amounts  of
radioactive TRU elements - mostly plutonium. TRU  waste  is defined1 as "waste  containing more than 100
nanocuries of alpha-emitting  transuranic isotopes, per gram  of waste, with  half-lives  greater than  20 years,
except for a)  high-level radioactive waste, b) waste that the Secretary has determined,  with concurrence of the
Administrator,  does not need  the degree of isolation  required by the disposal regulations;  or c) waste that the
Nuclear Regulatory Commission has approved for disposal on a case-by-case basis  in accordance with Part 61
of Title 10, Code of Federal Regulations."

TRU-rmixed waste is waste that is commingled with hazardous materials, such as lead or organic solvents. It is
regulated by  both  the Atomic Energy  Act and RCRA (as defined  in 20 NMAC 4.1, Subpart VIII,  Paragraph
268.35(d)  and in  the  Federal Facility Compliance Act.4  TRU-mixed waste  has  physical  and  radiological
characteristics similar to TRU waste. The majority of  TRU-mixed waste contains relatively small quantities of
spent halogenated solvents,  which  were  used   in cleaning  and degreasing of equipment, glassware, and
components. Based on sampling of gases within TRU waste drums, the most common volatile organic  hazardous
constituents are methylene chloride, carbon  tetrachloride, and 1,1,1-trichloroethane.5  TRU-mixed waste also
contains various RCRA-regulated metals.  These metals are usually associated with solid materials, such as lead
shielding.  Lead, chromium, and  cadmium are the most prevalent hazardous  metals in TRU-mixed waste.
TRU-mixed waste  constitutes approximately 60 percent of the DOE's TRU waste.6 The hazardous components of
TRU-mixed waste  to be  managed at  the  WIPP  are designated in  Part A of the  RCRA Permit  application.2
Henceforth, the term TRU waste used in this paper includes both TRU and TRU-mixed waste.

Transuranic Waste Generator/Storage Sites
TRU waste has accumulated over the past 50 years as a result of weapons development and  production at U.S.
defense facilities.  Since 1970, DOE has segregated TRU waste from other radioactive  waste and stored it in a
manner that allows it to  be retrieved.  TRU waste to be emplaced at  WIPP has resulted primarily from the
following: (1)  nuclear weapons development and  manufacturing,  (2) plutonium recovery, (3) defense research
and  development, (4)  environmental restoration, (5)  decontamination  and  decommissioning,  (6) waste


                                                 153

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
management, and (7) testing at facilities that are under DOE contract.

Table 1 has been reproduced from the National TRU Waste Management Plan.7 It lists the volumes of TRU
waste currently in storage and  the volumes of TRU waste projected to be generated  by ongoing and new
missions during the  life  of  the WIPP   Estimates  from  environmental restoration, decontamination and
decommissioning and future missions are also  included. The nuclear weapons complex  consists of ten major
facilities, including those at large reservations listed in the upper half of Table 1.

             Table 1. TRU Waste Storage Locations and Pre-treatment Volumes (cubic meters)7
Contact-Handled
TRU Waste
Site
Argonne National Lab.-East
Hanford Reservation
Idaho National Engineering and
Environmental Lab.
Lawrence Livermore National Lab.
Los Alamos National Lab.
Mound Plant
Nevada Test Site
Oak Ridge National Lab.
Rocky Flats Environmental
Technology Site
Savannah River Site
Small-Quantity Sites
Ames Laboratory
ARCO Medical Products
Company
Babcock & M/ilcox-NES
Battelle Columbus Laboratories
Bettis Atomic Power Laboratory
Energy Technology Engineering
Center
General Electric-Vallecitos
Nuclear Center
Knolls Atomic Power Lab.
Lawrence Berkeley Lab.
Missouri University Research
Reactor
Paducah Gaseous Diffusion Plant
Sandia National Laboratories
U.S. Army Material Command
Total Waste Volumes***
Location
Argonne, IL
Richland, WA
Idaho Falls, ID

Livermore, CA
Los Alamos, NM
Miamisburg, OH
Nevada
Oak Ridge, TN
Golden, CO

Aiken, SC

Ames, IA
West Chester, PA

Lynchberg, VA
Columbus, OH
West Mifflin, PA
Santa Susana, CA

Pleasanton, CA

Niskayuna, NY
Berkeley, CA
Columbia, MO

Paducah, KY
Albuquerque, NM
Rock Island, IL

Remote-Handled
TRU Waste
Stored* Projected Stored* Projected
through through
2033** 2033**
94
16,127
64,575

297
8,255
241
618
917
1,505

11,725

0
<1

20
0
0
7

6

0
<1
<1

2
7
2.5
104,400
109
7,305
15,009

835
8,544
6
19
180
6,988

17,811

<1
<1

0
0
114
0

3

0
4
1

0
44
0
56,972
0
200
86

0
101
0
0
1,268
1,268

1

0
0

0
0
0
0

8

<1
0
0

0
1
0
1,666
0
1,592
53

0
128
0
0
100
100

21

0
0

0
369
2
1

5

5
0
0

0
3
0
2,268
   Volumes Prior to treatment and repackaging.
** Projected volumes include estimates from environmental restoration, decontamination and decommissioning,
    and future  Departmental missions,  for example,  the  disposition of weapons-useable plutonium at the
    Savannah River Site. Estimates will change based upon future compliance actions under environmental law.
*** Totals reflect rounding of numbers.

Continued temporary storage of TRU waste at these and other sites across the nation poses potential problems.
For example, some of the metal drums used to store TRU waste are showing signs of corrosion, and the contents
of these drums  eventually will have to be repackaged. Not only would additional storage facilities be needed at
the generator/storage sites, but also  additional  worker exposures to penetrating radiation would occur due to
                                                 154

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


repackaging and inspection of waste containers. New treatment capacity would also be needed because much of
the TRU waste is subject to RCRA Land Disposal Restrictions and cannot be placed in or on the land unless it is
treated to satisfy  those  restrictions.  Sound environmental practice  requires that TRU waste be  permanently
isolated to prevent human exposure for many generations to come.

TRANSURANIC WASTE CHARACTERIZATION
The process of  identifying and classifying the chemical, physical, and radiological constituents of each drum of
waste is  a critical aspect of waste characterization. TRU waste  characterization  is  a subset of the waste
certification process and is based on sampling and analysis combined with acceptable knowledge of each waste
stream in accordance with the requirements of the TRU Waste Characterization Quality Assurance Program
Plan8 (QAPP). The QAPP utilizes a performance-based approach to allow individual sites to have the flexibility
to employ analytical and examination methods that  meet the quality assurance objectives specified in the Waste
Analysis Plan (WAP)2 and implemented by the requirements of the QAPP

Retrievably stored1 TRU waste will be characterized  by the generator/storage sites as the waste  is retrieved.
Newly generated"  TRU waste will be characterized as it is generated.  Waste characterization  requirements for
retrievably stored  and newly generated wastes are  slightly  different  and  are discussed in the WAP Waste
characterization activities  at the generator/storage  sites include  the following, although  not all  of these
techniques will be  used on each container:

•  Acceptable  Knowledge:  Compilation of documented characterization  and/or process knowledge into an
   auditable  record
•  Headspace-gas  sampling and analysis: Used  to determine volatile  organic  compound (VOC) content  of
   gases in the void volume of the containers
•  Sampling and analysis of homogeneous solid waste forms: Used to determine concentrations of hazardous
   waste constituents and toxicity characteristic contaminants of waste in containers
•  Radiography: an x-ray technique used to determine physical contents of containers
•  Visual examination: Used to verify radiography results
•  Radioassay: Used to identify isotopic inventory and associated activity

The  origins of  these requirements in the WAP  are traceable to applicable  regulatory requirements and
commitments made to regulatory authorities such as the NMED and the EPA.

NMED Basis for Waste Characterization
The Waste Analysis Plan (WAP), Chapter  C of the RCRA Part B Permit Application,2 describes the measures
that will be taken to assure that TRU waste received at the WIPP facility is within the scope of the RCRA permit
as established in and with unit-specific requirements of Title 20 of the New Mexico Administrative Code, Chapter
4, Part 1, Subpart V,  Paragraph 264.13 (20 NMAC  4.1).  The  WAP  establishes  waste characterization
requirements  that are referenced in the Waste Acceptance Criteria (WAC),9  the QAPP, and the TRU-waste
generator/storage  sites Quality Assurance Project Plans (QAPjPs). It includes descriptions of waste parameters,
rationale, and characterization methods; waste sampling and analysis strategies;  waste shipment screening and
verification processes; and a quality assurance (QA)/quality control (QC) program.

The WIPP underground disposal  unit is classified as a "miscellaneous unit" subject to regulation under 40 CFR
Part 264, Subpart X. Permit  applications for miscellaneous units must describe the wastes to be managed and
assess the potential environmental impacts associated with the proposed waste management activities. A  listing
of the  EPA  Hazardous Waste  Numbers  that may be  associated  with the waste to be  emplaced  in the
miscellaneous unit is contained in Part A of the WIPP RCRA Permit Application.2 This listing was determined by
a survey  of the  generator/storage sites' TRU waste inventories and  includes such RCRA-regulated constituents
as:
•  toxic   characteristic  contaminants listed  in  20  NMAC  4.1,  Subpart II,  Paragraph  261.24,  Table  1
   (corresponding to 40 CFR 261, Subpart C, Paragraph 261.24) except for pesticides,
1 Retrievably stored waste is defined as waste generated after 1970 and before implementation of the QAPP
  characterization requirements.
"  Newly generated  waste is defined  as waste  generated after  implementation  of  QAPP  characterization
  requirements.
                                                 155

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
•   F-listed solvents (F001, F002, F003, F004, F005, F006, F007, and F009) found in 20 NMAC 4.1, Subpart I,
    Paragraph 262.31 (corresponding to 40 CFR 261, Subpart D, Paragraph 261.3 1), and
•   hazardous constituents included in 20 NMAC 4.1, Subpart H, Paragraph 261,  Appendix VII (corresponding to
    40 CFR 261, Appendix VIII).

Waste is characterized  on a waste stream basis. A waste stream is defined as waste material generated from a
single process or from an activity that is similar in  material, physical form, and hazardous constituents. Wastes
are initially categorized  into three broad Summary Category Groups that are related to the final physical form of
the wastes.  These groups  include homogeneous solids  (Summary Category  S3000),  soil/gravel (Summary
Category S4000), and debris wastes (Summary Category S5000). Waste streams are grouped by Waste Matrix
Code Groups related to the physical and chemical properties of the waste. Generator/storage sites must use the
characterization techniques described in the WAP to assign appropriate Waste  Matrix Code  Groups for WIPP
disposal. The Waste Matrix Code Groups  are solidified inorganics,  solidified  organics,  salt waste, soils,
lead/cadmium metal, inorganic nonmetal waste,  combustible  waste,  graphite, filters,  heterogeneous debris
waste, and uncategorized metal.

A statistically selected portion of waste containers from waste streams of homogeneous solids and soil/gravel will
be sampled and analyzed for total volatile organic carbons (VOCs), semi-volatile organic compounds (SVOCs),
and  metals. TRU waste classified as debris wastes will be characterized  based on acceptable  knowledge.
Acceptable knowledge refers to applying knowledge of the hazardous characteristic of the waste in light of the
materials or processes  used to generate the waste. The use of acceptable knowledge is outlined in a guidance
manual10 wherein the EPA has specifically referred to the characterization of  radioactive mixed waste as a
situation where the use of acceptable knowledge is appropriate.

Since  waste  containers will  not be  opened  to  perform confirmatory sampling at the WIPP  site, waste
characterization  data produced  at the  site is reviewed at three levels to ensure it meets Program needs and
objectives. At the data generation level, data are reviewed, validated,  and verified. Data packages are submitted
to the project level for validation and verification. The third and final  level is the CAO level at which data from
the project level are verified.

Data review determines  if the raw data  have been properly collected and  ensures raw data are properly reduced.
Data  validation  confirms that  the data reported satisfy the requirements defined by the user (e.g.,  quality
assurance objectives and data quality objectives) and is accompanied by signature release. Validation at each
level  ensures that certain aspects of characterization and quality assurance have  been  met. Data verification
authenticates that data are  in fact that which is claimed.

EPA Basis for Waste Characterization
An estimate of each generator/storage site TRU  waste  inventory was  compiled  in the  TRU Waste Baseline
Inventory  Report,11 modeled by the performance  assessment,3 and evaluated  using  a  sensitivity analysis to
determine the impact of each waste component on the long term performance of the repository. The results of
the sensitivity analysis identified several parameters that must be monitored and  tracked to assure the validity of
assumptions  used in the   performance  assessment.  These  parameters are  listed  in  an  Appendix  to  the
Compliance  Certification Application (CCA)3  entitled "Waste  Component Limits" (WCL). Being that Appendix
WCL identifies the radionuclide content of the waste  as  one  of the components to be monitored  and tracked,
radiological characterization of the waste by radioassay is of concern to the EPA.

ACCEPTABLE KNOWLEDGE
The use of acceptable knowledge is discussed  in the  EPA document, "Waste Analysis: EPA  Guidance  Manual
for Facilities that Generate, Treat, Store and Dispose  of Hazardous Wastes."10 This document points out that
there  are situations where it may be appropriate to apply acceptable knowledge to characterize hazardous waste.
One of these, which is applicable to the type of waste that will be accepted for disposal in the WIPP, is that the
physical nature of the waste does not lend itself to the acquisition of a  representative sample. Additionally, the
RCRA regulations codified in 40 CFR Parts  260 through 265,  268,  and 270, authorize  the  use of acceptable
knowledge as a method which  can  be used in  appropriate  circumstances by waste generators, or treatment,
storage, or disposal facilities to make hazardous waste determinations. Acceptable knowledge, as an alternative
to sampling and analysis, can be used to meet all or part of waste characterization requirements under RCRA.
                                                  156

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Acceptable knowledge refers to applying knowledge of the  waste based on the materials or processes used to
generate the waste. Acceptable knowledge  includes information regarding the physical form of the waste, the
base materials composing the waste,  the nature of the  radioactivity present, and the  process generating the
waste. Acceptable knowledge is used to assign matrix parameter categories and EPA hazardous waste numbers
to waste streams and to determine the waste material parameters and  radionuclides present in waste streams.
The  collection and  use of acceptable knowledge information  applies to both  retrievably stored and newly
generated waste streams.

To satisfy the characterization requirements of the WAP, acceptable knowledge is confirmed by  radiography,
drum headspace gas sampling and analysis, and solidified waste sampling and analysis. Acceptable knowledge
is primarily used in TRU waste characterization activities to delineate TRU waste streams, to determine if TRU
debris wastes exhibit a toxicity characteristic (40 CFR 261.24), and to  determine if TRU wastes are listed (40
CFR 261.31).  The  physical form  and  the increased  health  and  safety risks associated with  obtaining a
representative  sample  of TRU debris wastes, clearly justify  the  use of acceptable knowledge in making
hazardous waste determinations.

DOE complex waste generators apply knowledge of their waste  based  on the materials and processes used to
generate the waste. A generator site can establish the characterization  of a waste stream by demonstrating an
understanding of the materials which are introduced into the process, and the process(es) which those materials
undergo. Understanding the process(es) which a  material may undergo is  very important, particularly  with
respect  to toxicity  characteristic wastes,  because some chemical processes may  result  in a change  in
concentration of the RCRA constituents. Assignment of F-listed wastes  also depends on knowledge of the
process that produces the waste. A change in concentration of the RCRA constituents  during a process could
result in a new waste stream containing constituents with concentrations that are above the regulatory threshold.
In the case of certain compounds  which may be  either  listed or toxicity characteristic, the use of acceptable
knowledge is the only viable route to determine whether a substance was utilized for its solvent properties or not.
If a  constituent was used for  its solvent properties, the waste stream would be assigned an F  code. If the
constituent was not  utilized for its solvent properties, it would be evaluated for the assignment of  a D code, if
appropriate.

The  Land Withdrawal Act1 included a number of requirements and  restrictions on the wastes  that can be
disposed of at the WIPP  Among these requirements and restrictions are that the waste be generated by atomic
energy  defense  activities and that it is neither  high-level  waste nor spent nuclear fuel.  DOE can verify
compliance with these conditions  only  through  the use of historical  information about  the processes that
generated a particular waste stream. This historical information is a component of the acceptable knowledge
record that is assembled and assessed by the waste generators.

Consistency among  sites in using acceptable knowledge information to characterize TRU waste involves a three
phase process: 1) compiling  the minimum acceptable knowledge  documentation in  an auditable record; 2)
confirming acceptable  knowledge  information using radiography, and headspace gas and solidified waste
sampling and analysis; and 3) auditing acceptable knowledge records. The consistent presentation of acceptable
knowledge among sites in  auditable records will allow WIPP personnel to verify the completeness of acceptable
knowledge and determine that the accuracy of acceptable knowledge  has been documented for TRU waste
characterization during the audit process.

AUDITING ACCEPTABLE KNOWLEDGE RECORDS
The DOE/CAO audits the TRU waste generators to grant TRU waste certification authority to the generators. The
DOE/CAO conducts audits at least annually thereafter to verify ongoing compliance with approved plans and
procedures including those for waste  characterization using  acceptable knowledge. The information resulting
from DOE/CAO audits is used by the DOE/CAO to help sites across the DOE complex prevent or solve problems
associated with the  compilation and use of acceptable  knowledge for TRU waste characterization.  Past and
current issues with respect to the compilation, use, and defensibility of acceptable knowledge for TRU waste
characterization along with the solutions to those problems and their prevention are addressed in this section.

The Audit Process
CAO conducts  an initial audit of  each  site to evaluate waste stream and  program  documentation prior to
certifying the site for shipment of TRU waste to the WIPP facility. The initial audit establishes a baseline that will
be reassessed annually. The audits are used to ensure the consistent compilation, application, and  interpretation


                                                  157

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


of acceptable knowledge information  throughout the DOE complex  and to evaluate the completeness and
defensibility of site-specific acceptable knowledge documentation related to hazardous waste determinations.

Acceptable knowledge audit checklists  typically include, but are not limited to, the following elements for review
during the audit:

•   Documentation of the process used to compile, evaluate, and record acceptable knowledge is available and
    implemented

•   Personnel training and qualifications are documented

•   All of the  required acceptable knowledge documentation has been compiled in an auditable record

•   Procedures exist for:
    •  assigning EPA hazardous waste numbers to waste streams
    •  assigning a matrix parameter category to a waste stream
    •  determining waste material parameters present in a waste stream
    •  determining the radionuclides present in a waste stream
    •  resolving inconsistencies in acceptable knowledge documentation
    •  confirming   acceptable knowledge  information  through:  a)  radiography or  visual  examination,  b)
       headspace gas sampling and analysis, and c) solidified waste sampling

•   Results of other audits of the TRU waste characterization programs at the site are  available in site records

Auditors assess all documents associated with the evaluation of the acceptable knowledge documentation for at
least one debris waste  stream and one solidified  waste  stream during the  audit.  For these waste streams,
auditors review all procedures and associated processes developed by the site for: documenting  the process of
compiling acceptable knowledge documentation; correlating information to specific waste inventories; assigning
EPA hazardous waste numbers; assigning matrix parameter categories; determining waste material parameters;
determining  the radionuclides;  and   identifying, resolving,  and  documenting  discrepancies  in  acceptable
knowledge records. The  adequacy of  acceptable knowledge procedures and processes is assessed and any
discrepancies in procedures are documented in the audit report.

Auditors  review the acceptable knowledge documentation for selected waste streams for logic,  completeness,
and defensibility. The criteria used by auditors to evaluate the logic and defensibility of the acceptable knowledge
documentation include completeness  and traceability of the information, clarity of  presentation,  degree  of
compliance with  the requirements of the QAPP and WAP with  regard to acceptable knowledge confirmation
data, nonconformance  procedures, and  oversight  procedures. Auditors evaluate  compliance with  written site
procedures for developing the acceptable knowledge record. A  completeness review is done to evaluate the
availability of the minimum required TRU waste management and TRU waste stream information. Records are
reviewed  to  assess the  correlations to  specific waste streams  and to assess the  basis for  making  waste
determinations. Auditors verify that sites  include all  required information and conservatively assign  all potential
EPA hazardous waste numbers indicated  by the acceptable knowledge  records. All deficiencies found  in the
acceptable knowledge documentation are included in the audit report.

Auditors verify and document that sites use management controls and follow written  procedures to make waste
determinations for newly generated and  retrievably  stored wastes. Auditors review procedures used by sites to
confirm acceptable  knowledge information using radiography or visual examination, headspace gas sampling
and analysis, and  solidified waste sampling and analysis. Procedures to  document  changes  in acceptable
knowledge documentation, EPA hazardous waste number assignments, matrix parameter category assignments,
waste material parameter determinations,  and  radionuclide determinations to specific waste streams are also
evaluated.

After the audit is complete, CAO prepares a final audit report that includes all observations and findings identified
during the audit.  Sites are required  to respond to all audit findings and identify corrective actions. If acceptable
knowledge procedures  do not exist, the minimum required  information is not  available,  or  findings of


                                                 158

-------
                        WTQA '98 - J 4th Annual Waste Testing & Quality Assurance Symposium


noncompliance associated with waste determinations are identified, CAO will not grant the site characterization
and certification authority for the subject waste. Waste stream characterization and certification authority may be
revoked or suspended if findings during subsequent annual audits indicate a lack of compliance with approved
acceptable knowledge procedures.

Acceptable Knowledge Issues
To date, the Department of Energy, Carlsbad Area Office has certified three generator/storage sites: Los Alamos
National Laboratory (September 1997), Rocky Flats  Environmental Technology Site (March 1998), and Idaho
National Engineering and Environmental Laboratory  (April 1998). These sites have undergone initial and final
audits of their certification  and acceptable knowledge processes and have successfully met all  WAP criteria for
characterizing their TRU waste via acceptable knowledge. Issues that the audit team encountered during audits
of the adequacy, implementation, and effectiveness of the sites' acceptable knowledge procedures/processes are
summarized below.

Adequacy
Procedural inadequacy  was a common  problem encountered  by the  audit team.  The WIPP  requires the
"proceduralization" of every aspect of the acceptable knowledge process, from compilation and evaluation of the
acceptable knowledge documentation, to reconciliation of discrepancies in the documentation,  to waste stream
designation and assignment of EPA hazardous waste codes. The following  procedural  inadequacies were noted
by the audit team:

•   The responsible party(ies) for some tasks identified in the procedures was not specified
•   A list of the documents generated as a result of implementing the procedure was not specified
•   Proper approvals for procedures were not obtained prior to implementation
•   Approvals for procedures were not adequately documented
•   Various requirements were omitted from procedures, such as a  specification to look for prohibited waste
    items
•   Necessary or required procedure steps, or forms, or references  to  other related  procedures were not
    specified

Implementation and Effectiveness
If the acceptable knowledge process procedures are adequate and complete, the satisfactory implementation of
these procedures results in  accurate  and defensible characterization of the waste by the generator, i.e.  an
effective process.  When procedures are not implemented correctly, the result is an incorrect determination, or an
inability to characterize the waste using acceptable knowledge, or an  ineffective process.  The audit team noted
several  deficiencies with respect to the  implementation and effectiveness of  acceptable  knowledge process
procedures:

•   Various processes, such as that for amending acceptable knowledge records, were not  established
•   Some procedures did not reflect current practice
•   Some procedures were not being property implemented
•   The acceptable knowledge documentation was not traceable to source documents the acceptable knowledge
    documentation was not compiled into an auditable package or record
•   Some sites  that subcontracted the acceptable knowledge process work did not  include requirements for
    qualification/training of personnel or a requirement to prepare  and  train the site personnel in the statement of
    work for the  subcontractor
•   The required analytical data and/or acceptable knowledge documentation was not available or insufficient to
    support the conclusions or EPA hazardous waste code assignments
•   Retrieval of records (source documents, etc.) was difficult, time consuming, or impossible
•   Documentation of training for various acceptable knowledge processes (review of acceptable knowledge
    documentation, etc.) was non-existent or insufficient
•   EPA hazardous waste codes were removed without sufficient documentation that a discrepancy process was
    followed
•  Acceptable  knowledge documentation contains  statements that  are not corroborated by  the source
   documentation


                                                  159

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


•   The process for the assignment  of EPA  hazardous  waste codes was  not  consistent with QAPP/WAP
    requirements
•   The acceptable knowledge summary report did  not list all potential constituents as shown in  supporting
    documentation
•   Documentation to justify that the waste was generated as a result of atomic energy defense activities was
    insufficient
•   Acceptable knowledge documentation for radionuclide  distribution was not adequate to support  radioassay
    data to confirm the radionuclide inventory and obtain total activity in TRU waste
•   The EPA hazardous waste code for a constituent identified in the acceptable knowledge documentation as
    being present in the waste was not assigned
•   Inconsistencies or discrepancies between acceptable knowledge documentation and/or analytical data were
    not reported (packaging configuration, volume of waste, etc.)

CONCLUSIONS
Acceptable knowledge  provides a reasonable and appropriate means to characterize nuclear waste for disposal.
Acceptable knowledge  is the only route by which certain characterization decisions (e.g. the origin of the waste
being from defense activities, the use of solvents) may be  made. The DOE and the waste generator sites have
developed and implemented an adequate system for the collection,  assessment,  and utilization of  acceptable
knowledge for characterizing waste to be disposed of at the Waste Isolation Pilot Plant.

ACKNOWLEDGEMENTS
Portions of this work were conducted under contract to Sandia  National Laboratories.  Sandia is a multiprogram
laboratory operated by Sandia Corporation,  a  Lockheed Martin Company for  the United States Department of
Energy Contract DE-AC04-94-AL85000. This work was conducted under the  Sandia  WIPP Quality  Assurance
Program which is equivalent to NQA-1, NQA-2 (Part 2.7), and NQA-3 Standards.

REFERENCES
1. Waste Isolation Pilot Plant Land Withdrawal Act, Public Law 102-579,  1992.
2. "Resource conservation and Recovery Act Part B Permit Application for the Waste Isolation Pilot Plant," U. S.
   Department of Energy (DOE), 1997, DOE/WIPP 91-005, Revision 6.4.
3. Waste Isolation Pilot  Plant Compliance  Certification Application, 40  CFR Part 191/194, DOE/CAO-2056,
   Carlsbad, New Mexico, U. S. Department of Energy.
4. Federal Facility Compliance Act,  Public Law 102- 386, Title 1,'3021(d).
5. "Waste Isolation Pilot Plant No-Migration Variance Petition," DOE/WIPP-89-003, Revision 1, Carlsbad, New
   Mexico, U. S. Department of Energy, 1990.
6.  "Waste  Isolation  Pilot Plant   Disposal  Phase  Draft  Supplemental  Environmental Impact Statement,"
   DOE/EIS-0026-S-2, Carlsbad, New Mexico, U. S. Department of Energy, 1996.
7. "National  TRU Waste  Management Plan. DOE  Carlsbad  Area  Office: Leadership  in Safe and  Efficient
   Cleanup of Transuranic Waste," DOE/NTP-96-1204, Revision 1, Carlsbad, New Mexico, U. S. Department of
   Energy, 1998.
8. "Transuranic Waste  Characterization Quality Assurance  Program Plan," CAO-94-1010, Revision 0, Carlsbad,
   New Mexico, U. S. Department of Energy, 1995.
9. "Waste Acceptance  Criteria for the Waste Isolation Pilot Plant," WIPP-DOE-069, Revision 5, Change Notice
   #2, Carlsbad, New Mexico, U. S. Department of Energy,  1991.
10.  "Waste Analysis at Facilities that  Generate, Treat, Store,  and Dispose of Hazardous Waste; A Guidance
    Manual," EPA-530-R-94-024, Washington D.C., Office of Solid Waste and  Emergency Response, U.S. EPA,
    1994.
11.  "Waste  Isolation  Pilot  Plant Transuranic Waste Baseline  Inventory Report," CAO-94-1005, Revision 3,
    Carlsbad, New Mexico, U. S. Department of Energy, 1995.
                                                160

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


        PERFORMANCE EVALUATION SOIL SAMPLES FOR VOLATILE ORGANIC COMPOUNDS
                        UTILIZING SOLVENT ENCAPSULATION TECHNOLOGY

                                            James Dahlgran
            U.S. Department of Energy, Idaho National Engineering  and Environmental Laboratory,
                            850 Energy Drive, NIS 4149, Idaho Falls, ID 83401
                                               Curt Thies
                   Thies Technology, 3720 Hampton, Suite 207, St.Louis, MO 63109-1438

Abstract
A mixture of volatile organic compounds (VOCs) was encapsulated and mixed with soil to produce a product
suitable for  use as a double blind source of VOCs in soil performance evaluation sample. Two independent
laboratories  analyzed the standard encapsulated VOC/soil mixture for  benzene, toluene, ethylbenzene and
xylene by using U.S. EPA SW-846 Method 5030 in conjunction with SW-846 Method 8020. One laboratory
received the sample as a single blind standard while the other laboratory received  the sample as a double blind
standard. The percent relative standard deviation (%RSD) for triplicate analyses  ranged  from 2.4 to 7.7. The
lowest %RSD was for meta/para-xylene (2.4%) from the sample analyzed as a double blind sample. Analytical
results from these pilot studies indicate that it  is possible to prepare standard soil samples contaminated with
known amounts of VOCs, which unlike current market technology, will enable soil samples to be submitted to
environmental analytical laboratories as a truly blind sample.

Introduction
For  most  environmental  analytical  procedures, demonstrating  proficiency  of  an analytical  method  is
accomplished utilizing known spiked samples, blanks, surrogate spikes and appropriate performance evaluation
standards. For analysis of volatile organic compounds  (VOCs) in soils, the performance evaluation standards are
primarily methanolic solutions that contain target analytes and are spiked into a soil sample immediately prior to
analysis.

Demand for precise performance evaluation samples for VOCs in  soil matrices has stagnated due to lack of
sample preparation technology. Current technology for preparing volatile performance evaluation samples  utilize
solvent spiking procedures1 or vapor fortification methodologies developed by J. Hewitt1 of the U.S. Army Cold2
Regions Research  and  Engineering Laboratory.  Private sector companies typically provide  analysts a  dilute
solution of analytes in a solvent, usually methanol.  These methanol  solutions are introduced either into the
analytical  technique (purge and  trap  or headspace  analyses) or  placed onto sand3 (used  to  simulate soil)
immediately prior to instrumental analysis. These practices do not adequately replicate  soil  sample  handling
procedures  in the  analytical laboratory. Liquid standards also inappropriately provide the  analysts with an
opportunity to analyze the spiking solution.

Although vapor-fortified soils provide a means of examining spiked soils that are analogous to soils isolated from
the environment, such standards are difficult to disseminate to analytical laboratories as a double blind quality
control standard. An ideal VOCs in soil standard should provide analyte concentrations across the concentration
range of 5 ug/kg  to 100  ug/g. Vapor fortification methodology can  be customized to decrease concentrations
below 100 ug/g, but it is less amenable to water soluble analytes such as acetone or 2-butanone and some target
analytes are lost due to the  varying absorbtivity of vastly differing  soil  matrixes. Early attempts  to  spike,
homogenize and transfer soil performance standards were unsatisfactory.4

To create a true volatiles in soil performance evaluation standard:
1. Volatile  target analytes of interest must be unknown  to the analyst.
2. Target analytes must be provided to the analyst in the soil matrix.
3. Volatile  components must be protected from potential soil biological activity.
4. The standard must be stable over an extended period of time.
5. The target analytes must provide the laboratory with a wide range  of VOC concentrations that are  accurate
   and reproducible independent of analytical methodology applied.

In order to meet these specifications, a novel method of spiking native soil samples is required. The objective of
this work is to demonstrate  that microencapsulated VOCs provide a novel method of spiking soils and offer an
improved  means  of assessing precision and accuracy of VOC  analyses of contaminated soils  reported by
different analytical laboratories.


                                                  161

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Experimental
Microcapsules loaded with a VOC mixture (1:1:1  (V/V/V) toluene, p-xylene and ethylbenzene) were formed by
complex coacervation5. Figure 1  is a diagram of a typical complex coacervation encapsulation protocol. The
shell of the microcapsules formed is primarily gelatin cross-linked with glutaraldehyde.

Standard contaminated soil samples were produced by dispersing a known weight of microcapsules loaded with
the VOC mixture into a known weight of soil. Dry subsurface soil taken  from the Idaho National Engineering and
Environmental Laboratory  (INEEL) was sieved (60-80 mesh) and  mixed with microcapsules to form a uniform
soil/microcapsule mixture.  Figure 2 shows the steps envisioned for the preparation of a standard performance
evaluation soil sample that unitizes microencapsulated  VOCs.

VOCs  in soil standards were characterized by gas chromatography. A Hewlett Packard 5880 Series  II flame
ionization gas chromatograph was used.  For toluene, ethylbenzene and p-xylene mixture, the column was a 30
meter X 0.53 mm J&W Scientific DB-624 capillary column. For gasoline, the column was a 30 meter X 0.32 mm
J&W Scientific DB-1 fused silica capillary column.

In order to analyze the VOC content of capsules and  soil containing capsules, a known weight of sample was
equilibrated in methanol at room temperature for a finite period. The samples were gently swirled to assure
complete wetting of the soil with the solvent. The methanol solution obtained was assayed directly (in the case of
gasoline  in   soil)   or  diluted   with   acetone  and   subsequently  assayed   (in  the   case   of  1:1:1
toluene/ethylbenzene/p-xylene). Equilibration time was  30 min. for all samples.

Results and  Discussion
In order to have a stable standard soil sample contaminated with VOCs, it is necessary to be able to prepare in a
reproducible  manner VOC-loaded  microcapsules that  are stable for prolonged  periods when mixed with a soil
sample. For a microcapsule  loaded with  VOCs to be stable, the capsule shell or coating  must have essentially
zero permeability to the VOCs encapsulated. The shell must also be susceptible to water and/or methanol, since
the analytical methodology  used to characterize  VOC  contaminated  soil  involves  these  solvents.  Dry
microcapsules shells formed by complex  coacervation are able to retain VOCs for prolonged periods as long as
they are not subject to high humidity storage conditions. In the presence of water or water/methanol mixtures, the
predominately gelatin shells become permeable thereby enabling release of the VOCs.

The first objective of the study was to  demonstrate that a mixture of pure solvents could be encapsulated and
retained. The solvent mixture encapsulated was a 1:1:1  (V/V/V) mixture of toluene,  ethylbenzene and p-xylene.
Three batches of capsules were made in  order to examine lot-to-lot reproducibility. The capsules produced were
clean,  uniform free-flowing powders suitable for incorporation into a soil sample. Triplicate analyses of 2 gram
samples of each  capsule  batch established that the  capsules were  95.8 weight % solvent (relative percent
standard deviation;  1.0%). When  5.08 grams of one capsule sample, fraction passing 60 mesh screen,  was
combined  with  100.2  grams clean  INEEL  subsurface  soil,  the   contaminated soil  produced  contained
approximately 56.5 milligrams (estimated by calculation) of toluene/ethylbenzene/p-xylene mixture per gram of
soil.

In order to examine homogeneity and accuracy this  standard contaminated soil sample, three one-gram and
three five-gram samples of contaminated soil were assayed. Table 1  summarizes the replicate analytical  data
obtained. Overall variability at one standard deviation was about 5% for all components at either the 1.0 gram or
5.0 gram  aliquot size.  This  single analysis indicates  that the concept of preparing an accurate performance
standard by utilizing microcapsules is realistic.  It demonstrated adequate  homogeneity in the sample at various
sample sizes as well as relative agreement with the amount of material expected (estimated 56.5 mg/gram vs.
found 48.6 mg/gram).

A second pilot involved the encapsulation of gasoline and preparation of soil sample contaminated with gasoline.
Four batches of capsules loaded with gasoline were prepared. The mean gasoline loading of these four samples
was 73.2% (standard deviation 6.2). Three ~2 gram  samples  of gasoline-loaded  capsules  were subjected in
triplicate to analysis and found to 3274 ug/gram toluene (std. dev. = 248, %RSD = 7.6), 4347  ug/g ethylbenzene
(std. dev. = 296, %RSD = 6.8) and 14922 ug /g p-xylene  (std. dev. 931, %RSD = 6.2).

Ten grams of gasoline-loaded capsules were combined with 193.0 grams of soil to produce a pilot performance
evaluation  sample.   Three samples of  this  contaminated soil  sample  placed  into individual 40  milliliter,


                                                  162

-------
                          WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposia
precleaned sampling vials each containing approximately 20 grams, were shipped to an
laboratory for analysis for benzene, toluene, ethylbenzene  and xylenes (BTEX)  Each
triplicate utilizing U.S.  EPA SW-846 Method 5030 in  conjunction  with SW846 Method
results of these analyses.
                                                                                       environmental analytical
                                                                                       sample was analyzed in
                                                                                       8020. Table 2 contains
Another sample was sent as
a double blind sample to a
second analytical laboratory
for analysis. Table 3) sum-
marizes  results  of  these
analyses. Other than  a  dis-
agreement  concerning  the
quantitation   of  benzene,
results  obtained  indepen-
dently  by  the two labora-
tories agree well.
Figure  1.  Coacervation of
Solvents for Addition to Soil
Sample
                                           core particle dispersed in solution of
                                           polymer by agitation
                                           Coacervation visible as droplets of colloid-rich
                                           phase induced by one or more agents and Deposition
                                           of coacervation droplets on surface of core
                                           Mergence of Coacervation droplets to form the coating
 KEY:
o
  -    Coacervation
 °o   droplets
 O
 o
                                                                                              Coating
                                                                                              Hardened coating
                                           Shrinkage and crosslinking of the coating to rigidize it as necessary.

                                                      Raf: Deasy. Patrick B . Microencapsulation and Related Drug Processes Marcel Dekker, Inc
   Mix core material (solvents) to appropriate
   ratios
                           \
 The analytical data obtained to date demonstrates using microcapsules to  prepare soil performance evaluation
 standards is attainable. The performance evaluation soil standard has the advantage of being as close to a real
 world soil sample as technologically feasible and has demonstrated the ability to measure the quality associated
 with the entire analytical methodology. This technology is capable of producing a performance evaluation sample
 containing volatile target organic analytes within well-defined concentration ranges. The encapsulated  process
                                                                                        can provide  variable
                                                                                        concentration  ranges
                                                                                        for  numerous  target
                                                                                        volatile analytes.  The
                                                                                        resulting  performance
                                                                                        evaluation sample will
                                                                                        be  amenable to many
                                                                                        of the current US  EPA
                                                                                        methodologies for the
                                                                                        analysis  of  environ-
                                                                                        mentally    significant
                                                                                        volatile   organics  in
                                                                                        soils  including  analy-
                                                                                        sis   carried  out   by
                                                                                        thermal     desorption
                                                                                        techniques.
                      Emulsify to form small droplets
                                 Form coating or interfacial polymerization
                                           Mieroeapsule suspended in polymer forming phase
                                                                  i
                                                         Filer, wish and recover microcapsulea
                                                                         T
                                                Mix capsules with diluvnl soil In appropriate
                                                ratio to generate required standard
                                                                                        Figure 2. Typical
                                                                                        Steps of Standard
                                                                                        Preparation Process

Current efforts are focused on  reducing the concentration of target analytes to the 5 - 200  ppb range  and to
examine product storage stability. Once through this pilot examination, the carrier solvent will be spiked with
other  analytes  including; acetone,  carbon  tetrachloride,  chloroform,  methylene  chloride,  trichloroethane,
tetrachloroethylene. The author has submitted soil  samples that contain microencapsulated VOCs to the Mixed
Analyte Performance Evaluation  Program, operated by the U.S. Department  of Energy, for  round robin study
utilizing contracted environmental analytical laboratories.
                                                     163

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table  1. Data from  Simulated  Contaminated Soil  Sample that Contains Microcapsules Loaded with a 1:1:1
(V/V/V) Toluene/ Ethylbenzene/p-Xylene Mixture, (units are milligrams per grain of soil)
 Sample Weight (g.)
                     Toluene
            Ethylbenzene
               p-Xylene
 1.0
 5.0
Average
Std. Dev.
 %RSD

Average
Std. Dev.
 %RSD
  11.96
   0.57
    4.8

  12.00
   0.62
    5.1
   22.21
    1.05
     4.7

   22.78
    1.16
     5.1
        22.41
          1.08
          4.8

        23.02
          1.17
          5.1
Table 2. Gasoline in Soil Pilot Performance Evaluation Standard - Single Lab Analysis, (units are milligrams per
gram of soil)
                      Benzene
             Toluene
Ethylbenzene
o-Xylene
m & p-Xylene
Sample 1
Sample 2
Sample 3
Average
Std. Dev.
%RSD
444
406
407
367
402
416
387
384
406
402.1
20.5
5.1
143
134
130
131
130
133
131
124
131
131.9
4.7
3.6
187
206
168
192
198
205
197
191
200
193.8
10.9
5.6
341
306
303
267
286
280
298
261
287
292.1
22.5
7.7
898
846
841
792
793
828
814
766
836
823.8
36.3
4.4
Table 3. Gasoline in Soil Pilot Performance Standard Sent as Double Blind to Second Lab for Analysis, (units
are milligrams per gram of soil)
                      Benzene
             Toluene
Ethylbenzene
o-Xylene
m & p-Xylene
Sample 1



Average
Std. Dev.
%RSD
Sample 1
Confirmation


Average
Std. Dev.
%RSD
2.5 U
2.5U
2.5 U
2.5 U



NA
NA
NA
NA



120
120
99
100
109.8
10.2
9.3
120
120
120
120
120
0

150
150
120
110
130
16.7
12.9
200
210
220
210
210
7.1
3.4
270
270
230
250
255
16.6
6.5
350
360
380
350
360
12.2
3.4
750
760
610
680
700
60.4
8.6
900
940
960
920
930
22.4
2.4
Literature Cited
1. Conversations with RTC, ERA and APG.
2. Hewitt, A.D.; Clarence, L.G., Environ. Sci. Technol. 1995, 29, 769-773.
3. Environmental Research Associates Product Catalogue, 1997, 16.
4. Maskarinec, M.P.; Johnson, L.H.; Bayne, C.K., J Assoc. Off. Anal. Chem. 1989, 72, 823-827.
5. Deasy, Patrick B., "Microencapsulation and Related Drug Processes", Marcel Dekker, Inc., 1984.
                                                  164

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


       U.S. EPA AND U.S. A.F. INTERAGENCY AGREEMENT FOR FIELD ANALYTICAL SERVICES

                                         Raymond A. Flores
       US EPA Region 6, Houston Community College, 10625 Fallstone Road, Houston TX 77099-4303
                                       281-983-2139 ext. 139
                              FLORES.RAYMOND@EPAMAIL.EPA.GOV
                                           George H. Lee
          Analytical Services Division, AL/OEA,  Armstrong Laboratory, Brooks AFB, TX 78235-5114
                                           210-536-6166
                                GLEE@GUARDIAN.BROOKS.AF.MIL

The interagency agreement (IAG) between the  parties above will provide for more  timely analyses of EPA
samples while at the same time making more efficient use of the federal government's  resources. At the time of
the writing of this abstract the IAG has just reached implementation phase and consequently  data is limited. We
are confident that with the data generated in the next four months, we will be able to show that time and money
are saved by the use of this IAG. As the Project officers for the EPA-AF IAG, we will use the data from each field
project in making the cost savings comparison. We will discuss project startup and implementation. The onsite
analytical capabilities of the Air Force, both field screening  techniques and field confirmation methods, will be
examined. USAF's Armstrong  Laboratory located at Brooks Air Force Base, San  Antonio Texas, will provide
onsite analyses for USEPA  Region 6's Superfund Division.  The elements of the Superfund Program  that can
request work are the Brownfields, Removal, Emergency Response, Remedial, and Site  Assessment Teams. The
Air Force Responders will be onsite in a time frame dependent upon the nature of the site activity.  Most site
activities are  non-emergency. Activities which are not time critical  such as Brownfields  and Remedial for
example, will  allow sufficient time for a significant amount of preplanning for onsite activities. Other  activities
which are more time critical  require immediate action. Air Force responds to those  activities in which there is a
threat or possible threat of risk to human health in less than 24 hours dependent on site location. In  the near
future, AF is projected to have a mobile laboratory which can  be driven to the site immediately upon notice.

Onsite field screening  and confirmation  techniques have been projected to be a  more cost effective way to
perform portions of the chemical analysis evaluation of a site as early as 1988. Air Force's trained personnel and
specialized equipment can be used for  EPA projects in  EPA Region 6  and of course for Air Force's needs
worldwide. Air Force benefits in  a  number of ways such as becoming familiarized with  EPA's SOPs for site
activities. EPA benefits by having  access to a cost-effective service with unique capabilities and world wide
experience. This IAG provides for technology transfer and more efficient resource utilization.

SUMMARY
This  agreement between Federal Agencies pursuing  common goals allows for  greater efficiency  and  cost
effectiveness for all and provides an opportunity for technology transfer. We recommend the use of interagency
agreements in similar cases. An IAG can be a commensal relationship with mutual advantages.
       SEVERAL ORGANIC PARAMETERS ON UNDERLYING HAZARDOUS CONSTITUENTS LIST
               CAN NOT BE MEASURED AT THE UNIVERSAL TREATMENT STANDARDS

                                         Howard C. Johnson
                                Sample Management Risk Technologies
                                         Clifford S. Watkins
     Sample Management Risk Technologies, Idaho National Engineering and Environmental Laboratory,
                Sample Management Office, Lockheed Martin Idaho Technologies Company,
                             P.O. Box 1625-3960, Idaho Falls, Idaho 83415

ABSTRACT
The Idaho National Engineering and Environmental Laboratory (INEEL) has several permitted treatment, storage
and disposal facilities. The  INEEL Sample Management Office  (SMO), operated  by  Lockheed Martin Idaho


                                                165

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Technologies Company (LMITCO),  conducts all analysis subcontracting  activities for Department of Energy
Environmental Management programs at the INEEL. In this  role, the INEEL SMO has had the opportunity to
subcontract the analyses of various wastes (including ash from an interim status incinerator) requesting a target
analyte list equivalent to the constituents listed in 40 Code of Federal Regulations (CFR) § 268.48. Per 40 CFR §
268.40,  these analyses are  required to ensure  that treated  wastes do not contain underlying  hazardous
constituents (UHC) at concentrations greater than the universal treatment standards (UTS)  prior to land disposal.
The  language in  40 CFR § 268.40 (d) (3)  states that,  "The treatment  or disposal facility may demonstrate
compliance with organic constituents if good-faith analytical  efforts achieve detection limits for the regulated
organic  constituents that do  not exceed the  treatment  standards specified in  this section by an order of
magnitude."

The  INEEL SMO has conducted this "good-faith effort" by negotiating with several commercial laboratories to
identify the lowest possible quantitation and detection limits that can be achieved for the organic LTHC analytes.
The results of this negotiating effort has been the discovery that no single laboratory (currently under subcontract
with  the INEEL SMO)  can  achieve a  detection level that is  within an order of magnitude of the UTS for all
organic  parameters on a clean sample matrix  (e.g., sand). This indicates that for a typical waste sample, the
chances of the order of magnitude requirement not being met for many more than  just the "problem analytes" is
likely. This does not mean that there is no laboratory that can achieve the order of magnitude requirements for
all organic UHCs on a clean  sample matrix.  The negotiations held  to  date indicate that it is likely that no
laboratory can achieve the order of magnitude requirements for a difficult sample  matrix (e.g., an incinerator
ash). The authors suggest  that the regulation needs to be revised to address the disparity between what is
achievable in the laboratory and the regulatory levels required  by the UTS.

INTRODUCTION
The  INEEL SMO conducts all analysis subcontracting activities for the Department of Energy Environmental
Management programs at the  INEEL.  Contracted  analyses are primarily  in the following  analytical disciplines;
radiological, inorganic,  organic and physical properties testing. Within each discipline numerous analytical tests
are requested on a  large variety of sample matrices. Analytical  test requirements  range from  field screening or
processing information to data  required  to  satisfy  U.S.  Environmental  Protection Agency (EPA)  and  Idaho
Division of Environmental Quality (ID-DEQ)  requirements. The INEEL SMO subcontracts analytical work on a
variety of wastes (including ash  from an interim status incinerator). Analytical requests have included a target
analyte  list equivalent to the constituents listed in 40 CFR § 268.48. Per  40 CFR  § 268.40, these analyses are
required to ensure that treated wastes do not contain UHCs at concentrations greater than the UTS prior to land
disposal. The language in 40 CFR § 268.40 (d) states:

        "Notwithstanding the  prohibitions specified  in paragraph (a)  of this section,  treatment  and  disposal
        facilities may demonstrate (and certify pursuant to 40 CFR § 268.7(b)(5)) compliance with the treatment
        standards  for  organic constituents  specified  by a  footnote  in  the table "Treatment Standards for
        Hazardous Wastes" in this section, provided the following conditions are satisfied:
        (1) The treatment standards for the  organic constituents were established  based on  incineration in units
           operated in accordance with the technical requirements of 40 CFR part 264, subpart 0, or based on
           combustion in fuel substitution units operating in accordance with applicable technical requirements;
        (2) The treatment or disposal facility has used the methods referenced in paragraph (d)(1) of this section
           to treat the organic constituents; and
        (3) The treatment or disposal facility may demonstrate compliance with organic constituents if good-faith
           analytical efforts achieve detection limits for the  regulated organic constituents that do not exceed
           the treatment standards specified in this section by an order of magnitude."

The  INEEL SMO has  sought this "good-faith effort"  by  negotiating  with  several  commercial laboratories to
identify  the lowest  possible quantitation or  detection limits that can be  achieved for the organic UHCs. The
primary  emphasis has been on the nonwastewater standard. At this time, the wastewater standard for the organic
UHCs has not been performed by all of the  INEEL SMO contracted laboratories. The  results of this negotiating
effort has been the discovery that no single laboratory (currently under subcontract with  the  INEEL SMO) can
achieve the detection level of the UTS for all organic parameters of the UHC list on a clean sample matrix (e.g.,
sand). This indicates that for  a typical waste sample,  the chances of the order of magnitude requirement not
being met for many more than just the "problem analytes" is likely. This does not mean that there is no laboratory
that can achieve the order of magnitude requirements for all organic UHCs on a clean sample matrix. The INEEL
SMO continues to seek laboratory capability to analyze for the UHCs to the UTS  in a cost-effective manner.


                                                  166

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Additionally, the  regulated community should be aware that the "order-of-magnitude" provision might not be
allowed  on  incinerator ash from  an interim  status incinerator. In the September 19  1994  final  rule  the
"order-of-magnitude" provision is only applicable where incinerated wastes were treated in permitted (Part 264
Subpart O) units.

Within the capabilities of the current INEEL SMO subcontracted laboratories, no two laboratories analyze the
UHCs to the UTS the  same way. In the guidelines of USEPA SW-846 a variety of approved methods can be
used to analyze for the various constituents. The SW-846 methods being used to analyze for the  complete UHC
list by the laboratories currently contracted through the INEEL SMO are:

•   Method 1311 Toxicity Characteristic Leaching Procedure
•   Method 8015A Nonhalogenated Volatile Organics by Gas Chromatography
•   Method 8015M (modified) Nonhalogenated Volatile Organics
•   Method 8080A Organochlorine Pesticides And Polychlorinated Biphenyls By Gas Chromatography
•   Method  8081 Organochlorine Pesticides, Halowaxes And  PCBs As Aroclors  By Gas Chromatography:
    Capillary Column Technique
•   Method 8081A Organochlorine Pesticides (PCBs) By Gas Chromatography
•   Method 8082 Polychlorinated Biphenyls By Gas Chromatography
•   Method 8140 Organophosphorus Pesticides
•   Method 8141 Organophosphorus Compounds By Gas Chromatography: Capillary Column Technique
•   Method 8150 Chlorinated Herbicides By Gas Chromatography
•   Method  8151 Chlorinated Herbicides By GC Using Methylation Or Pentafluorobenzylation  Derivatization:
    Capillary Column Technique
•   Method  8260A Volatile  Organic  Compounds By Gas  Chromatography  Mass  Spectrometry (GC/MS):
    Capillary Column Technique
•   Method  8270B Semivolatile Organic Compounds By Gas Chromatography/Mass Spectrometry (GC/MS):
    Capillary Column Technique
•   Method 8310 Polynuclear Aromatic Hydrocarbons
    Method 8316 Acrylamide, Acrylonitrile And Acrolein By High Performance Liquid Chromatography (HPLC)

The UHC  list is extensive and there are many "problem analytes" to  analyze. The methodology is left to the
capability of the laboratory and the laboratory's discretion. It would not  be practical or reasonable for the INEEL
SMO  to specify to a laboratory a required method for every constituent. Under the fixed  contracts, obsolete
methods are still  listed but current methodology  is used whenever possible, incorporating  the  June  13, 1997
promulgated methods  from the Third Edition of the  EPA-approved test  methods  manual  "Test Methods for
Evaluating Solid Waste, Physical/Chemical Methods."

To procure UHC organic analyses,  it becomes necessary to negotiate with each contracted laboratory concerning
achievable detection and quantitation limits for the entire UHC list. The negotiations held to date indicate that it is
likely that no laboratory can achieve the order of magnitude requirements  for a difficult sample matrix (e.g., an
incinerator  ash).  When  negotiating  with  commercial laboratories that  are  under fixed  price subcontracts
government contractors  like LMITCO have difficulty authorizing additional financial support, for the "research
project"  samples. Analytical difficulties arise when dealing with complex matrices such as an incinerator fly ash.
For example,  alternate  solvent system  extractions may  be  more  effective for the semivolatile organic
compounds. Laboratories currently contracted through the INEEL SMO  routinely extract semivolatiles using
acetone/hexane or acetone/methylene chloride systems. They have  found  fewer fly ash matrix interferences
when they substituted  a  methylene chloride only extraction system. The alternate solvent system also requires
additional calibration with a resulting burden on the laboratory under a fixed price contract.

The organic UHC list is extensive  and standards are hard to locate for the entire list. For example, one of the
contracted laboratories had difficulty initially procuring a standard for 4,4'-Methylene-bis-(o-chloroaniline) CAS#
101-14-4. No laboratory contracted through the INEEL SMO  has determined quantitation  limits or calibrated for
the  40 additional analytes that will become "underlying hazardous  constituents" August 26, 1998. Commercial
laboratories contracted through the INEEL SMO have expressed concerns  about the effort and expense that will
be required to obtain standards, conduct MDL determinations, etc. in order to analyze for the  added constituents.
The 40 additional constituents are identified  in Table 1 UTS.
                                                 167

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
SUMMARY
The  current INEEL SMO contracted laboratories operate within the  highly competitive  commercial laboratory
industry. With laboratory capacity available to DOE (i.e.  possessing a NRC  license to receive radioactively
contaminated samples) limited nationwide, extensive "non-routine" analytical requests on complex matrices are
not desired by the laboratories.  Obtaining the  full suite  of UHC analyses can be difficult.  The  inability to
determine if a treated waste has achieved the required treatment standard concentrations can place programs in
a position subject to violation of regulatory requirements. The authors' suggest that the  regulation needs to be
revised to address the disparity between what is achievable in the laboratory and the regulatory levels required
by the UTS.

The authors' poster delineates the achievable detection limits for organic UHC parameters at several commercial
laboratories. The intent of the poster  is to bring this  issue to the attention  of the regulators and regulated
community. Another benefit is to discuss analytical approaches used by other laboratories that may achieve
greater sensitivities for the difficult organic analytes on the UHC list.

Table  1  UTS  identifies the organic hazardous constituents, along  with the  nonwastewater and wastewater
treatment standard levels. 40 CFR § 268.48 also states: "For determining compliance with treatment standards
for underlying hazardous constituents as defined in § 268.2(i), these treatment standards  may not be exceeded.
Compliance with these treatment standards is measured by an analysis of grab  samples,  unless otherwise noted
in  the following Table 1  UTS".  Included in the table are the  negotiated quantitation limits for the nonwastewater
standard from three  INEEL SMO contracted laboratories (A, B, and  C). These limits can be compared to the
listed regulatory nonwastewater standard concentration. Presently the most difficult ash  samples are requiring
five to ten fold dilutions -in addition to alternate solvent system  extractions- which further limit the ability of the
laboratory to achieve quantitation limits at (or below) the UTS for all of the UHCs. The  negotiated quantitation
limits are the  laboratories  "ideal" quantitation limits  on a  clean matrix  (e.g.,  sand) and do not represent the
current achievable limits  on  difficult  sample  matrices (e.g.,  incinerator  fly ash).  When the laboratories
quantitation limit is at the treatment standard concentration,  the  INEEL SMO requires the laboratory to have the
instrument detection  limit for that analyte less than 0.33 times the treatment standard. This limit is also specified
in  the task order statement of work between the laboratory and the INEEL SMO.

                              Table 1. UTS - Universal Treatment Standards
                                     (Note: NA means not applicable.)
Regulated
constituent/common
name
Organic Constituents
A2213(6)
Acenaphthylene
Acenapthene
Acetone
Acetonitrile
Acetophenone
2-Acetylaminofluorene
Acrolein
Acrylamide
Acrylonitrile
Aldicarb sulfuric (6)
Aldrin
4-Aminobiphenyl
Aniline
Anthracene
Aramite
alpha-BHC
beta-BHC
delta-BHC
gamma-BHC
Barban (6)
Bendiocarb (6)
Bendiocarb phenol (6)
Benomyl (6)
Benzene
CAS(1)
Number
30558-43-1
208-96-8
83-32-9
67-64-1
75-05-8
96-86-2
53-96-3
107-02-8
79-06-1
107-13-1
1646-88-4
309-00-2
92-67-1
62-53-3
120-12-7
140-57-8
319-84-6
319-85-7
319-86-8
58-89-9
101-27-9
22781-23-3
22961-82-6
17804-35-2
71-43-2
Wastewater
standard
Concentration
in mg/l (2)
0.042
0.059
0.059
0.28
5.6
0.010
0.059
0.29
19
0.24
0.056
0.021
0.13
0.81
0.059
0.36
0.00014
0.00014
0.023
0.001
0.056
0.056
0.056
0.056
0.14
Nonwastewater
standard
Concentration in
mg/kg (3)
unless noted as
"mg/l TCLP"
1.4
3.4
3.4
160
38
9.7
140
NA
23
84
0.28
0.066
NA
14
3.4
NA
0.066
0.066
0.066
0.066
1.4
1.4
1.4
1.4
10
Negotiated
Nonwastewater
Quantitation Limit
INEEL SMO
contracted
Laboratory A
undetermined
0.33
0.33
0.01
0.02
0.33
0.33
0.01
50
0.01
undetermined
0.005
0.33
0.33
0.33
0.66
0.005
0.005
0.005
0.005
undetermined
undetermined
undetermined
undetermined
0005
Negotiated
Nonwastewater
Quantitation Limit
INEEL SMO
contracted
Laboratory B
undetermined
0.67
0.67
0.01
0.05
0.67
0.67
NA
23
0.05
undetermined
0.066
NA
0.67
0.67
NA
0.002
0.004
0.006
0.002
undetermined
undetermined
undetermined
undetermined
0.005
Negotiated
Nonwastewater
Quantitation Limit
INEEL SMO
contracted
Laboratory C
undetermined
3.4
3.4
160
38
9.7
140
NA
23
84
undetermined
0.066
NA
14
3.4
NA
0.066
0.066
0.066
0.066
undetermined
undetermined
undetermined
undetermined
10
                                                  168

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Benz(a)anthracene
Benzal chloride
Benzo(b)fluoranthene
(difficult to distinguish
from benzo(k)
fluoranthene
Benzo(k)fluoranthene
(difficult to distinguish
from benzo(b)
fluoranthene)
Benzo(gji,i)perylene
Benzo(a)pyrene
Bromodichloromethane
Bramomethane/Methyl
bromide
4-Bromophenyl phenyl
ether
n-Butyl alcohol
Butylate (6)
Butyl benzyl phthalate
2-sec-Butyl-4,6-
dinitrophenol/Dinoseb
Carbaryl (6)
Carbenzadim (6)
Carbofuran (6)
Carbodfuran phenojJ6}
Carbon disulfide
Carbon tetrachloride
Carbosulfan (6)
Chlordane (alpha and
gamma isomers)
p-Chloroaniline
Chlorobenzene
Chlorobenzilate
2-Cloro-1 ,3-butadiene
Chlorodibromomethane
Chloroethane
bis(2-chloroethoxy)
methane
bis(2-Chloroethyl)ether
Chloroform
bis(2-Chloroisopropyl)
ether
p-Chloro-m-cresol
2-Chloroethyl vinyl ether
Chloromethane/Methyl
chloride
2-Chloronaphthalene
2-Chlorophenol
3-Chloropropylene
Chrysene
o-Cresol
m-Cresol (difficult to
distinguish from
p-cresol)
p-Cresol (difficult to
distinguish from
m-cresol)
m-Cumenyl
methylcarbamate (6)
Cyclohexanone
o,p'-DDD
p,p'-DDD
o,p'-DDE
p,p'-DDE
o,p'-DDT
p,p'-DDT
Dibenz(a,h)anthracene
Dibenz(a.e)pyrene
56-55-3
98-87-3
205-99-2
207-08-9
191-24-2
50-32-8
75-27-4
74-83-9
101-55-3
71-36-3
2008-41-5
85-68-7
88-85-7
63-25-2
10605-21-7
1563-66-2
1563-38-8
75-15-0
56-23-5
55285-14-9
57-74-9
106-47-8
108-90-7
510-15-6
126-99-8
124-48-1
75-00-3
111-91-1
111-44-4
67-66-3
39638-32-9
59-50-7
110-75-8
74-87-3
91-58-7
95-57-8
107-05-1
218-01-9
95-48-7
108-39-4
106-44-5
64-00-6
108-94-1
53-19-0
72-54-8
3424-82-6
72-55-9
789-02-6
50-29-3
53-70-3
192-65-4
0.059 I
0.055
0.11
0.11
0.0055
0.061
0.35
0.11
0.055
5.6
0.042
0.017
0.066
0.006
0.056
0.006
0.056
3.8
0.057
0.028
0.0033
0.46
0.057
0.10
0.057
0.057
0.27
0.036
0.033
0.046
0.055
0.018
0.062
0.19
0.055
0.044
0.036
0.059
0.11
0.77
0.77
0.056
0.36
0.023
0.023
0.031
0.031
0.0039
0.0039
0.055
0.061
3.4
6.0
6.8
6.8
1.8
3.4
15
15
15
2.6
1.4
28
2.5
014
1.4
014
1.4
4.8 mg/l TCLP
6.0
1.4
0.26
16
6.0
NA
28
15
6.0
7.2
6.0
6.0
7.2
14
NA
30
5.6
5.7
30
3.4
5.6
5.6
5.6
1.4
0.75 mg/l TCLP
0.087
0.087
0.087
0.087
0.087
0.087
8.2
NA
0.33
1.65
0.33
0.33
0.33
0.33
0.005
0.01
0.33
0.5
undetermined
0.33
1.65
undetermined
undetermined
undetermined
undetermined
.005 mg/l TCLP
0.005
undetermined
0.005
0.33
0.005
0.33
0.01
0.005
0.01
0.33
0.33
0.005
0.33
0.33
0.01
0.01
33
0.33
0.02
0.33
0.33
Reported as
p-cresol
0.33
undetermined
0.1 mg/l TCLP
0.01
0.01
0.01
0.01
0.01
0.01
0.33
0.01
0.67
0.05
0.67
0.67
0.67
0.67
0.05
0.05
0.67
2.6
undetermined
67
1.3
undetermined
undetermined
undetermined
undetermined
4.8 mg/l TCLP
0.005
undetermined
0.009
0.67
0.005
NA
0.05
0.005
0.05
0.67
0.67
0.005
0.67
0.67
NA
0.05
0.67
0.67
0.05
0.67
0.67
Reported as
p-cresol
0.67
undetermined
0.75 mg/l TCLP
0,087
0.087
0.087
0.087
0.087
0.087
0.67
NA
3.4
6.0
68
6.8
1.8
3.4
15
15
15
2.6
undetermined
28
2.5
undetermined
undetermined
undetermined
undetermined
4.8 mg/ TCLP
6.0
undetermined
0.26
16
6.0
NA
0.28
15
6.0
7.2
6.0
6.0
7.2
14
NA
30
5.6
5.7
30
3.4
5.6
Reported as
p-cresol
5.6
undetermined
0.75 mg/l TCLP
0.087
0.087
0.087
0.087
0.087
0.087
8.2
NA
                             169

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
1 ,2-Dibromo-3-
chloropropane
1 .2-Dibromoethane/
Ethylene dibromide
Dibromomethane
m-Dichlorobenzene
o-Dichlorobenzene
p-Dichlorobenzene
Dichlorodinuoromethane
1 , 1 -Dichloroethane
1 ,2-Dichloroethane
1 , 1 -Dichloroethylene
trans-1 ,2-
Dichloroethylene
2,4-Dichlorophenol
2,6-Dichlorop henol
2,4-Dichlo phenoxyacetic
acid/2,4-D
1 ,2-Dichloropropane
cis-1 ,3-
Dichloropropylene
trans-1 ,3-
Dichloropropylene
Dieldrin
Diethylene glycol,
dicarbamate (6)
Diethyl phthalate
p-Dimethylamino-
azobenzene
2,4-Dimethyl phenol
Dimethyl phthalate
Dimetilan (6)
Di-n-butyl phthalate
1 ,4-Dinitrobenzene
4,6-Dinitro-o-cresol
2.4-Dinitrophenol
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Di-n-octyl phthalate
Di-n-propylnitrosamine
1 ,4-Dioxane
Diphenylamine (difficult
to distinguish from
diphenylnitrosamine)
Diphenylnitrosamine
(difficult to distinguish
from diphenylamine)
1 ,2-Diphenylhydrazine
Disulfoton
Dithiocarbamates (total)
(6)
Endosulfan I
Endosulfan II
Endosulfan sulfate
Endrin
Endrin aldehyde
EPIC (6)
Ethyl acetate
Ethyl benzene
Ethyl cyanide/
Propanenitrile
Ethyl ether
96-12-8
106-93-4
74-95-3
541-73-1
95-50-1
106-46-7
75-71-8
75-34-3
107-06-2
75-35-4
156-60-5
120-83-2
87-65-0
94-75-7
78-87-5
10061-01-5
10061-02-6
60-57-1
5952-26-1
84-66-2
60-11-7
105-67-9
131-11-3
644-64-4
84-74-2
100-25-4
534-52-1
51-28-5
121-14-2
606-20-2
117-84-0
621-64-7
123-91-1
122-39-4
86-30-6
122-66-7
298-04-4
137-30-4
959-98-8
33213-46-9
1031-07-8
72-20-8
7421-93-4
759-94-4
141-78-6
100-41-4
107-12-0
60-29-7
0.11
0.028
0.11
0.036
0.088
0.090
0.23
0.059
0.21
0.025
0.054
0.044
0.044
0.72
0.85
0.036
0.036
0.017
.0.056
0.20
0.130
0.036
0.047
0.056
0.057
0.32
0.28
0.12
0.32
0.55
0.017
0.40
12.0
0.92
0.92
0.087
0.017
0.028
0.023
0.029
0.029
0.0028
0.025
0.042
0.34
0.057
0.24
0.12
15
15
15
6.0
6.0
6.0
7.2
6.0
6.0
6.0
30
14
14
10
18
18
18
0.13
1.4
28
NA
14
28
1.4
28
2.3
160
160
140
28
28
14
170
13
13
NA
62
28
0.066
0.13
0.13
0.13
0.13
1.4
33
10
360
160
0.02
0.02
0.01
0.33
0.33
0.33
0.01
0.005
0.005
0.005
0.005
0.33
0.33
10
0.005
0.005
0.005
0.01
undetermined
0.33
0.33
0.33
0.33
undetermined
0.33
1.65
1.65
1.65
0.33
0.33
033
0.33
0.33
0.33
0.33 cannot be
separated from
Dipheylamine
0.33
0.02
undetermined
0.005
0.01
0.01
0.01
0.01
undetermined
0.01
0.005
0.25
0.01
0.05
0.05
0.05
0.67
0.67
0.67
0.05
0.005
0.005
0.005
0.005
0.67
0.67
10
0.005
0.005
0.005
0.01
undetermined
0.67
NA
0.67
0.67
undetermined
0.67
1.6
3.3
3.3
0.67
0.67
0.67
0.67
0.5
13
1 3 reported as
Dipheylamine
NAfor
nonwastewater,
reported as
decomposition
product
azobenzene for
wastewater
0.03
undetermined
0.066
0.13
0.13
0.13
0.13
undetermined
33
0.005
360
160
15
15
15
6.0
6.0
6.0
7.2
6.0
6.0
6.0
30
14
14
10
18
18
18
0.13
undetermined
28
NA
14
28
undetermined
28
2.3
160
160
140
28
28
14
0.1
13
1 3 reported as
Dipheylamine
NA
6.2
undetermined
0.066
0.13
0.13
0.13
0.13
undetermined
33
10
360
160
                             170

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
bis(2-Ethytthexyl)
phthalate
Ethyl methacrylate late
Ethylene oxide '
Famphur
Fluoranthene
Fluorene
Formetanate
hydrochloride (6)
Formparanate (6)
Heptachlor
Heptachlor epoxide
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclo-
pentadiene
HxCDDs (All Hexa-
chlorodibenro-p-
dioxins)
HxCDFs (All Hexa-
chlorodibenzo-furans)
Hexachloroethane
Hexachloropropylene
Indeno (1 ,2,3-c,d)
pyrene
lodomethane
Isobutyl alcohol
Isodrin
Isolan (6)
Isosafrole
Kepone
Methacrylnitrile
Methanol
Methapyrilene
Methiocarb (6)
Methomyl (6)
Methoxychlor
3-Methylcholanthrene
4,4-Methylene
bls(2-chloroaniline)
Methylene chloride
Methyl ethyl ketone
Methyl isobutyl ketone
Methyl methacrylate
Methyl methansulfonate
Methyl parathion
Metolcarb (6)
Mexacarbate (6)
Molinate (6)
Naphthalene
2-Naphthylamine
o-Nitroaniline
p-Nitroaniline
Nitrobenzene
5-Nitro-o-toluidine
o-Nitrophenol
p-Nitrophenol
N-Nitrosodiethylamine
N-Nitrosodimethylamine
N-Nitroso-di-n-
butylamine
N-Nitrosomorpholine
N-Nitrosopiperidine
N-Nitrosomethyl-
ethylamine
N-Nitrosopyrrolidine
117-81-7
97-63-2
75-21-8
52-85-7
206-44-0
86-73-7
23422-53-9
17702-57-7
76-44-8
1024-57-3
118-74-1
87-68-3
77-47-4
NA
NA
67-72-1
1888-71-7
193-39-5
74-88-4
79-83-1
465-73-6
119-38-0
120-58-1
143-50-0
126-98-7
67-56 -1
91-80-5
2032-65-7
16752-77-5
72-43-5
56-49-5
101-14-4
75-09-2
78-93-3
108-10-1
80-62-6
66-27-3
298-00-4
1129-41-5
315-18-4
2212-67-1
91-20-3
91-59-8
88-74-4
100-01-6
98-95-3
99-55-8
88-75-5
100-02-7
55-18-5
62-75-9
924-16-3
59-89-2
100-75-4
10595-95-6
930-55-2
0.28
0.14
0.12
0.017
0.068
0.059
0.056
0.056
0.0012
0.016
0.055
0.055
0.057
0.000063
0.000063
0.055
0.035
0.0055
0.19
5.6
0.021
0.056
0.081
0.001
0.24
5.6
0.081
0.056
0.028
0.25
0.0055
0.50
0.089
0.28
0.14
0.14
0.018
0.014
0.056
0.056
0.042
0.059
0.52
0.27
0.028
0.068
0.32
0.028
0.12
0.40
0.40
0.40
0.40
0.013
0.40
0.013
28
160
NA
15
3.4
3.4
1.4
1.4
0.066
0.066
10
5.6
2.4
0.001
0.001
30
30
3.4
65
170
0.066
1.4
2.6
0.13
84
0.75 mg/l TCLP
1.5
1.4
0.14
0.18
15
30
30
36
33
160
NA
4.6
1.4
1.4
1.4
5.6
NA
14
28
14
28
13
29
28
2.3
17
2.3
35
2.3
35
0.33
0.33
10
0.1
0.33
0.33
undetermined
undetermined
0.005
0.005
0.33
0.33
0.33
0.0005
0.0005
0.33
0.33
0.33
0.01
0.1
0.1
undetermined
0.33
0.05
0.02
5 mg/l TCLP
0.33
undetermined
undetermined
0.05
0.33
1.65
0.005
0.01
0.01
0.33
0.33
0.02
undetermined
undetermined
undetermined
0.33
0.33
1.65
1.65
0.33
0.33
33
1.65
0.33
0.33
0.33
0.33
1.65 _j
0.33
0.33
0.67
0.05
NA
1.3
0.67
0.67
undetermined
undetermined
0.002
0.056
0.67
0.67
0.67
0.001
0.001
0.67
0.67
0.67
0.05
5
0.66
undetermined
0.67
1.3
0.05
Analyzed as total
10mg/kg,
equivalent to 0.5
mg/l TCLP
0.67
undetermined
undetermined
0.12
0.67
30
0.005
0.01
0.01
0.05
NA
0.06
undetermined
undetermined
undetermined
0.67
NA
3.3
3.3
0.67
0.67
0.67
3.3
0.67
67
0.67
0.67
1.3
0.67
0.67
28
160
NA
15
3.4
3.4
undetermined
undetermined
0066
0.066
10
5.6
2.4
0.001
0.001
30
30
3.4
65
170
0.066
undetermined
2.6
0.13
84
0.75 mg/l TCLP
1.5
undetermined
undetermined
0.18
15
30
30
36
33
160
NA
4.6
undetermined
undetermined
undetermined
5.6
NA
14
28
14
28
13
29
28
2.3
17
2.3
undetermined
2.3
undetermined
                             171

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Oxamyl (6)
Parathion
Total PCBs (sum of all
PCB isomers, or all
Aroclors)
Pebulate (6)
Pentachlorobenzene
PeCDDs (All
Pentachlorodibenzo-
p-dioxins)
PeCDFS (All Penta-
chlorodibenzo-furans)
Pentachloroethane
Pentachloronitrobenzene
Pentachlorophenol
Phenacetin
Phenanthrene
Phenol
o-Phenylenediamine (6)
Phorate
Phthalic acid
Phthalic anhydride
Physostigmine (6)
Physostigmine salicylate
(6)
Promecarb (6)
Pronamide
Propham (6)
Propoxur (6)
Prosulfocarb (6)
Pyrene
Pyridine
Safrole
Silvex/2,4.5-TP
1 ,2,4,5-Tetrachloro-
benzene
TCDDs (All
Tetrachlorodibenzo-p-
dioxins)
TCDFS (All Tetra-
chlorodibenzofurans)
1,1,1,2-
Tetrachloroethane
1,1,2,2-
Tetrachloroethane
Tetrachloroethylene
2,3,4,6-
Tetrachlorophenol
Thiodicarb (6)
Thiophanate-methyl (6)
Tirpate (6)
Toluene
Toxaphene
Triallate (6)
Tribromomethane/
Bromoform
1 ,2,4-Trichlorobenzene
1,1,1-Trichloroethane
1 ,1 ,2-Trichloroethane
Trichloroethylene
Trichloromonofluoro-
methane
2,4,5-Trichlorophenol)
2,4,6-Trichlorophenol
2,4,5-Trichlorophen-
oxyacetic acid/2, 4,5-T
23135-22-0
56-38-2
1336-36-3
1114-71-2
608-93-5
NA
NA
76-01-7
82-68-8
87-86-5
62-44-2
85-01-9
108-95-2
95-54-5
298-02-2
100-21-0
85-44-9
57-47-6
57-64-7
2631-37-0
23950-58-5
122-42-9
114-26-1
52888-80-9
129-00-0
110-86-1
94-59-7
93-72-1
95-94-3
NA
NA
630-20-6
79-34-5 1
127-18-4
58-90-2
59669-26-0
23564-05-8
26419-73-8
108-88-3
8001-35-2
2303-17-5
75-25-2
120-82-1
71-55-6
79-00-5
79-01-6
75-69-4
95-95-4
88-06-2
93-76-5
0.056
0.014
0.10
0.042
0.055
0.000063
0.000035
0.055
0.055
0.089
0.081
0.059
0.039
0.056
0.021
0.055
0.055
0.056
0.056
0.056
0.093
0.056
0.056
0.042
0.067
0.014
0.081
0.72
0.055
0.000063
0.000063
0.057
0.057
0.056
0.030
0.019
0.056
0.056
0.080
0.0095
0.042
0.63
0.055
0.054
0.054
0.054
0.020
0.18
0.035
0.72
0.28
4.6
10
1.4
10
0.001
0.001
6.0
4.8
7.4
16
5.6
6.2
5.6
4.6
28
28
1.4
1.4
1.4
1.5
1.4
1.4
1.4
8.2
16
22
7.9
14
0.001
0.001
6.0
6.0
6.0
7.4
1.4
1.4
0.28
10
2.6
1.4
15
19
6.0
6.0
6.0
30
7.4
7.4
7.9
undetermined
0.02
0.8
undetermined
0.33
0.0005
0.0005
0.33
1.65
1.65
0.33
0.33
0.33
undetermined
0.02
3.3
0.66
undetermined
undetermined
undetermined
0.33
undetermined
undetermined
undetermined
0.33
0.33
0.33
7.9
0.33
0.0005
0.0005
0.01
0.005
0.005
0.33
undetermined
undetermined
undetermined
0.005
0.5
undetermined
0.005
0.33
0.005
0.005
0.005
0.005
1.65
0.33
7.9
undetermined
0.03
10
undetermined
0.67
0.001
0.001
0.01
0.67
3.3
0.67
0.67
0.67
undetermined
0.02
28 reported as
Phthalic
anhydride
28
undetermined
undetermined
undetermined
0.67
undetermined
undetermined
undetermined
0.67
0.67
0.67
7.9
0.67
0.001
0.001
0.05
0.005
0.005
0.67
undetermined
undetermined
undetermined
0.005
0.16
undetermined
0.005
0.67
0.005
0.005
0.005
0.005
3.3
0.67
7.9
undetermined
4.6
10
undetermined
10
0.001
0.001
6.0
4.8
7.4
16
5.6
6.2
undetermined
4.6
undetermined
undetermined
undetermined
undetermined
undetermined
1.5
undetermined
undetermined
undetermined
8.2
16
22
7.9
14
0.001
0.001
6.0
6.0
6.0
7.4
undetermined
undetermined
undetermined
10
2.6
undetermined
15
19
6.0
6.0
6.0
30
7.4
7.4
7.9
                             172

-------
                          WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
1 ,2,3-Trichloropropane
1,1,2-Trichloro-1,2,2-
trifluoroethane
Triethylamine (6)
trls-(2,3-Dibromopropyl)
phosphate
Vernolate (6)
Vinyl chloride
Xylenes-mixed isomers
(sum of o-, on, and
Pj*ylene
96-18-1
76-13-1
10-144-8
126-72-7
1929-77-7
l~~ 75-01-4
133020-7
0,85
0.057
0.081
0.11
0.042
0.27
0.32
30
30
1.5
0.10
1.4
6.0
30
0.01
0.01
undetermined
3.3
undetermined
0.01
0.005
0.05
0.05
undetermined
0.33
undetermined
0.05
0.005
30
30
undetermined
undetermined
undetermined
6.0
30
Footnotes to Universal Treatment Standards Table:
(1) CAS means Chemical Abstract Services. When the waste code and/or regulated constituents we described as a combination of a chemical with
   it's salts and/or esters; the CAS number is given for the parent compound only.
(2) Concentration standards for wastewaters are expressed in mg/l and are based on analysis of composite samples.
(3) Except for Metals (EP or TCLP) and Cyanides (Total and Amemable) the nonwastewater treatment standards expressed as a concentration
   were established, in part, based upon incineration in units operated in accordance with the technical requirements of 40 CFR part 264, subpart
   O, or 40 CFR part 265, subpart O, or based upon combustion in  fuel substitution units operating in accordance with applicable technical
   requirements. A  facility may comply with these treatment standards according to provisions in See 268.40(d). All concentration standards for
   nonwastewaters  are based on analysis of grab samples.
(4) Both Cyanides (Total) and Cyanides (Amenable) for nonwastewaters are to be antlyzed using Method 9010 or 9012, found in "Test Methods for
   Evaluating Solid  Waste, Physical/Chemical Methods". EPA Publication SW-846, as incorporated by reference in 40 CFR 260.11, with a sample
   size of 10 grams and a distillation time of one hour and 15 minutes.
(5) These constituents am not "underlying hazardous constituents" in characteristic news, according to the definition at See. 268.2(i).
(6) Between August 26,1997, and August 26, 1998, these constituents are not "underlying hazardous constituents" as defined at Sec. 268.2(i).

REFERENCES
Protection of the Environment, 40 Code of Federal Regulations.
Test Methods for Evaluating Solid Waste, Physical/Chemical Methods SW-846 Third Edition, United States
   Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, D.C.,
   December, 1996.
RCRA Regulations and Keyword Index, Elsevier Science Inc., New York, New York, 1997 Edition.
                     IGNITABILITY PERFORMANCE EVALUATION STUDY ARE YOUR
                        WASTE STREAMS BEING CORRECTLY CHARACTERIZED?

                                        Lester J. Dupes. Rock J. Vita I e
        Environmental Standards,  Inc., 1140 Valley Forge Road, Valley Forge, Pennsylvania 19482-0911
                                              Debra J. Caillouet
                           Stationary Environmental & Energy, Chrysler Corporation,
                           800 Chrysler Drive E, Auburn Hills, Michigan 48326-2757

ABSTRACT
Thirteen commercial laboratories  were evaluated  for consideration of providing  waste stream characterization
support. An initial single-blind performance evaluation (PE) study was performed to determine the accuracy of
the laboratory-reported results when compared to known values.

The initial PE study was conducted  for constituents typically analyzed for  hazardous waste characterization.
These constituents included Toxicity Characteristic Leaching Procedure (TCLP)  volatiles, TCLP semivolatiles,
TCLP metals,  and ignitability. A review of the ignitability tests performed by SW-846 Method 1010 and  1020
exhibited a wide range  of both  positive and negative bias  from the known flashpoint temperature of the pure
compound used as the  PE sample. Results were compared to reproducibility criteria generated from American
Society of Testing Materials (ASTM) Method D93-96.

Because of the erratic  ignitability results observed for many of the participating laboratories, the laboratories
were requested to perform a self-audit inspection of their current  SOP and actual laboratory procedures. These
self-audit results were used to ensure inter-laboratory consistency and compliance to SW-846  Method 1010 and
Method 1020.
                                                     173

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


A second round of PE samples were subsequently submitted to each of the thirteen participating laboratories for
analysis. These results showed some improvement for several of the laboratories, whereas other laboratories
again reported results with  a  significant  bias. The results of this study exhibited  a wide range of ignitabilities
which would have represented incorrect waste characterization  if the PE samples  were actual waste  stream
samples. This paper will focus on discussion of the PE sample results, the laboratory self-audit results,  and
method compliance issues.

INTRODUCTION
A comprehensive laboratory evaluation process was performed  to identify commercial laboratories to provide
waste characterization testing services. This evaluation included  an initial technical survey of approximately 40
laboratories. The survey results were  used  to  further  reduce  the  potential candidates  to 13 laboratories.
Laboratory  audits specifically focused on waste characterization analyses and performance evaluation  (PE)
samples for specific parameters of interest  for the characterization of complex  wastes was used to  further
evaluate the  13  short-listed laboratories.  The parameters   included ignitability,  TCLP volatiles,  TCLP
semivolatiles, and TCLP metals. An initial  set  of single-blind  PE samples were prepared  as whole volume
samples by a reputable commercial PE provider.

A single-blind PE study sample is defined as a "test" sample in  which the laboratory is aware that the sample
submitted is a PE sample but does not have knowledge relative to the analytes or true concentrations contained
in the PE. A single-blind sample permits the data user to better understand a laboratory's analytical accuracy and
by inference, draw conclusions on the accuracy of actual waste sample results.

The  matrix of the ignitability  PE sample was a pure compound. The TCLP volatile sample was an aqueous
matrix, and the TCLP metals sample was a soil matrix. The PE samples for all 13 laboratories were prepared
from the same lot number to  further reduce variance and permit a direct and meaningful comparison between
laboratories. Whole-volume samples, which  require no further dilution or preparation by the laboratory prior to
analysis, were chosen to further reduce variables caused by different technicians or chemists. The PE samples
were  prepared and  shipped on ice  under Chain-of-Custody  procedures to  each  laboratory directly from the
provider.

In addition, an actual complex industrial waste stream sample was also used in this PE study. This waste  stream
was  selected to  provide the  laboratories with a  representative  "real-world"  sample that the laboratories may
receive as an actual  sample for waste testing. In  particular, a paint purge solvent was selected for testing. This
paint purge  solvent is relatively homogeneous and has previously been characterized as hazardous due to a
flashpoint temperature below the regulatory limit and volatile concentrations above the regulatory limit.

To collect this sample, the PE provider furnished  bottles to the industrial facility for packaging this solvent. This
sample was collected and shipped back to the PE provider for homogenization and repackaging for delivery to
the 13 laboratories.  This PE sample was shipped by ground carrier and was not refrigerated. Since it was not
certified, this sample was to be evaluated only for interlaboratory comparison purposes and was analyzed for
ignitability, TCLP volatiles and TCLP metals.

Performance by the individual laboratories was evaluated not only for the results proximity to the certified values
but also for communications, data packaging and reporting, method compliance, and timeliness of deliverables.
These latter issues were evaluated using  techniques and procedures addressed in a previous manuscript  (Dupes
and Rose, 1995). The issues are not discussed any further herein.

INITIAL PERFORMANCE EVALUATION SAMPLE RESULTS
In general,  the laboratories reported acceptable results for the TCLP volatile organic compounds and  metals,
relative to the program  requirements. However, ignitability results ranged from 114°F to 153°F, with only  four
laboratories  meeting acceptance criteria for this parameter. The ignitability PE  sample was certified  by the
provider of the pure compound to have a flashpoint of 140°F ± 5°F. This certified value represents the flashpoint
of the compound as determined by various manufacturers of this compound.  The temperature of the PE  sample
was specifically chosen to determine variance around 140°F This temperature is also defined as the regulatory
limit  for characterization of the waste as hazardous. Wastes that flash at a temperature of <140°F are considered
hazardous and wastes that flash at  >140°F are considered non-hazardous by the definition for ignitability (40
CFR Part 261).
                                                  174

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Ten of the laboratories reported the  temperatures higher than the regulatory  limit of 140°F. resulting in  a
classification as  non-hazardous.  Results for the three laboratories classifying the PE  sample as hazardous
reported results of 129°F, 114°F, and 120°F. Those temperatures are all well below the lower acceptance limits
of 135°F for the PE sample. Six other laboratories were slightly above the 145°F acceptance criteria. A summary
of the results is presented in Table 1.

The "real world" purge solvent sample submitted to the laboratories was consistently identified by all laboratories
as hazardous based on flashpoint. The reported flashpoints ranged from 66-76°F (several laboratories reported
the sample flashed at less than their reporting limit of room temperature).

The variance in the certified (pure compound) PE results may have been due to the use of thermometers which
have not been  adequately calibrated against  a National Institute of Standards and Technology (NIST) standard
thermometer (this is routinely noted in  a number of audits recently performed by the authors), the lack of
correction  for barometric pressure, or analytical variance of the methodologies used by the laboratories. Each of
the laboratories management or quality assurance staff were required to conduct an intensive self-audit review of
the PE  sample analysis, general analysis conditions, and  quality  control procedures to determine  potential
causes for the wide range of flashpoint temperatures obtained for the PE sample.

SELF-AUDIT QUESTIONS AND FINDINGS
Upon completion of the initial PE study,  a self-audit questionnaire was generated for completion by each of the
participating laboratories.  The audit questionnaire consisted of 26 questions obtained from a technical review of
SW-846 Methods 1010 and 1020A; ASTM Methods D93-80, D93-90,  D93-96, D3278-96, and E502-84. These
questions  included  verification of the method used and referenced for analysis of the initial PE study  sample,
sample storage conditions, instrument analysis conditions, analysis procedures,  thermometer calibrations, quality
control measures, acceptance criteria, and corrective actions.

The questionnaire was provided to  each  of the laboratories for completion  by laboratory management or quality
assurance personnel. In  addition  to  providing written responses to the  questions, a  copy of the laboratory
standard operating procedure was requested for review.

A review of the thirteen laboratory  responses and  SOPs were conducted to determine possible reasons for the
significant variation observed. A summary of issues identified in the  self-audit checklists by the authors  is
presented  below.

•  According to the original  request for analysis,  all laboratories were requested to  perform Method  1020A.
   Several laboratories did request a change to Method 1010, which was granted and documented. However,
   many  laboratories did not request this change. Whenever a method change is necessary, data users (viz.,
   the client) should be contacted prior to implementing the method change.

•  Several laboratories used compounds (i.e. acetic acid, mixed xylene, and kerosene) that were not stated  in
   the method for quality control purposes.  The laboratories were directed to use p-xylene conforming to the
   ASTM specifications,  as required by  Method 1010. The acceptance criteria should be 81±2°F as required by
   ASTM D93-90 or 81±1°F as required  by ASTM D93-96.

•  Several laboratories did not store the ignitability PE sample at 4±2°C. Although it could be argued that a pure
   single-component compound used for the PE would be unaffected by volatile constituent loss, actual waste
   samples are typically complex mixtures, and lighter fractions can be lost due to volatilization. All laboratories
   were requested to store future samples scheduled for ignitability at 4±2°C to reduce the  possibility of loss  of
   lighter fraction volatile constituents.

•  Five laboratories had  not  calibrated or could not demonstrate that the  thermometer  used  for ignitability
   determination had been calibrated with a NIST-standard thermometer within the last year. One laboratory
   noted a 2-5°F difference upon calibration of three different thermometers. In addition,  a laboratory also noted
   that after the recalibration a mercury gap in one thermometer was  noted and that the gap increased as the
   temperature increased. All laboratories were directed to recalibrate  the thermometers on an annual basis. A
   multi-point  calibration should be performed to determine accuracy at  several temperatures, and correction
   factors must be taken into account when measuring flashpoint temperatures.
                                                  175

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


•   Six laboratories did not correct  for barometric pressure, as required by the ASTM D93 methods. Several of
    the laboratories noted that the correction would be very small (< 0.2°F); however, one laboratory noted the
    flashpoint temperature would increase by 0.9°F All  laboratories were directed  to correct for the ambient
    barometric pressure of the laboratory, as required by ASTM methodologies referenced in Method 1010 and
    Method 1020A.

•   Many of the  laboratories did  not report a duplicate result  for the ignitability analysis  by  Method  1010.
    Although not specifically stated in Method 1010 or ASTM D93, duplicate analyses of all samples should be
    performed to  comply with ASTM E502, which  indicates that the average  results for flashpoint should be
    reported. This will also provide a consistent approach with respect to Method 1020A, which requires duplicate
    analyses. Sufficient volume (>150 ml) should be collected or provided in order to conduct duplicate analyses
    by Method 1010.

•   All laboratories  indicated the Pensky Martens apparatus used for analysis  by  Method 1010  is located in a
    hood to reduce surrounding drafts and for health and safety considerations. However, several laboratories
    indicated  that the  hood was turned on during the actual testing, which can cause loss of volatiles through
    drafts. Several laboratories indicated that problems have occurred when hoods containing the apparatus are
    turned on during analysis. Laboratories were directed to limit the amount of draft around the apparatus.

•   Three laboratories did not meet p-xylene acceptance criteria of 81±2°F (ASTM  D9390) or 81±1°F (ASTM
    D93-96) prior to analysis of the PE sample. One laboratory  reported an acceptance range for p-xylene of
    80-91 °F. Two laboratories did  not report p-xylene results on the logbook  pages. These practices are  not
    acceptable. All p-xylene calibrations must meet acceptance criteria and be recorded prior to sample analysis.
    Several laboratories have  implemented the use of an additional laboratory control sample  (LCS) near the
    regulatory limit of 140°F  This is not a method requirement but is recognized  as a good laboratory practice.

•   One  laboratory was increasing  the rate of temperature of the sample by 2°F/minute ; prior to application of
    the test flame. The laboratories were directed to follow the rate of temperature increase, as specified  in the
    ASTM D93 Method as 9-11°F/minute.

•   One  laboratory  reported that the  sample was stirred during  the actual application of the test flame to the
    sample. Samples must not be stirred during the application of, the test flame to the vapor space, as required
    by the ASTM method.

•   Several laboratories indicated  the introduction  of the test flame into the vapor space is maintained from
    0.5-1.5 seconds. Analysts should  attempt to consistently introduce the test flame into the vapor space for 1
    second, as required by the ASTM  protocol referenced by Method 1010.

•   One  laboratory  reported the ignitability results to the nearest tenth degree.  One laboratory reported the raw
    results on a  stenopad, which  did  not  include,  the  analysis test name,  method  number,  analyst name,
    laboratory name, or QC results. The laboratories were directed to report result to the  nearest 1°F  Adequate
    documentation of quality control (QC) results, barometric pressure, analysis date,  etc.  must be maintained by
    all laboratories through the use  of formal logbooks.

•   Several laboratories had not reviewed their SOP within the last year to verify method  compliance and  actual
    laboratory procedures. All  laboratories were directed to review and  update laboratory SOPs on  an annual
    basis to reflect the actual procedures performed and to monitor for continuing method compliance.

SECOND ROUND IGNITABILITY PE STUDY
Upon completion  of the self-audit  review, laboratories were  notified of the findings and requested  corrective
actions prior to performing the second ignitability PE sample study. Three ignitability  PE samples were selected
for the second round study. The flashpoint of the PE samples was specifically designed by the authors to provide
a range  of flashpoints. The flashpoint-certified values of Flashpoint Sample  #1, Flashpoint Sample #2, and
Flashpoint Sample #3 were 140°F, 120°F, and 170°F, respectively. These three flashpoints provide a sample that
would be considered hazardous (120°F), a sample at the hazardous regulatory limit  (140°F) and a sample that
would be considered non-hazardous (170°F). The actual compounds used for the PE samples were chosen and
prepared by the same commercial  PE provider used for the  initial study. The expected  flashpoint temperature

                                                  176

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


represents the flashpoint for the compound as determined by various manufacturers of this compound. All PE
samples were pretested prior to use by the PE supplier by  Pensky Martens closed-cup procedures (US  EPA
SW-846 Method 1010) and were corrected for barometric pressure of 620 mm. All PE samples were contained in
glass bottles  with Teflon-lined caps and were shipped on ice under Chain-of-Custody to  the 13 participating
analytical laboratories. Sufficient volume (150 ml) was sent to each laboratory to permit duplicate analyses of the
PE sample.

ANALYTICAL RESULTS AND CRITERIA FOR EVALUATION
Reproducibility criteria were calculated based on ASTM D93-96, which provides precision and bias data for the
ASTM  method. Reproducibility in this ASTM method  is defined as "the difference between two single and
independent results obtained  by different operators working  in different laboratories on identical material that
would,  in the  long run, in the normal and correct operation of the test method, exceed the following values only
one case in 20" (ASTM D93-96).

       where:

              R=BX

              B = 0.078
              X = the reported result in °C
              R = reproducibility

Since the  ASTM method is unclear as to which temperature should  be used in calculating the reproducibility
criteria, criteria for  the three flashpoint PE samples were generated  by using the expected flashpoint (converted
to degrees Celsius) in the equation listed above. The criteria calculated should represent a range encompassing
the expected  value of the PE sample. The following criteria were calculated for the PE samples:

 Sample Identifier                         Expected Flashpoint               Reproducibilily Criteria
 Flashpoint Sample #1                           140°F                             ± 8°F
 Flashpoint Sample #2                           120°F                             ± 7°F
 Flashpoint Sample #3                           170°F                            ± 11 °F

According  to  the previous ASTM Method  revision D93-90, reproducibility criterion of ± 6°F should be used for
evaluation of the analytical  data. According to the ASTM  D93-90 method, this criteria  applies to  all liquid
samples with  flashpoint temperatures less than 220°F Evaluation of the data using the more stringent criteria in
D93-90 excluded only two more results when compared to the criteria calculated  by Method  D93-96. A summary
of the results  are presented in  Tables 2, 3 and 4.

All participating  laboratories  except  one analyzed the three  PE  samples  by SW-846  Method  1010.  One
laboratory  analyzed the samples by SW-846 Method 1020. Three provided results after the due date, missing the
specifically stated date as a requirement in the documentation accompanying the PE samples.

DISCUSSION OF SECOND ROUND IGNITABILITY PE STUDY RESULTS
Upon review of the PE sample results, the inter-laboratory accuracy  of the  results obtained from the laboratories
was observed to be variable.  Four laboratories reported results which, when compared to the known flashpoint,
were greater than the Method  D93-96 criteria for all analyses (except for the analysis of Flashpoint #1 sample by
one laboratory, which was within criteria). Reported results, when compared to  the known flashpoint, ranged from
10° to  70°F absolute difference from the  known value  of the PE samples. Results  reported by one laboratory
were biased very low (-27° to -70°F  below the known flashpoints  of the materials). Results reported by the
laboratories are unacceptable  and may indicate  a recurring problem  relating to the analyses being performed by
these laboratories.

Two laboratories reported two of the three results outside the Method D93-96 criteria. When compared to the
known  flashpoint, reported results ranged from  8° to 12°F absolute  difference from the known value of the PE
samples.

Five laboratories reported  one  of the three results  outside the  calculated Method  D93-96  criteria. When
compared  to the known flashpoint, reported results ranged from 7°  to 13°F absolute difference from the known


                                                 177

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


value of the PE samples.  One laboratory reported duplicate results, but did not report and average result for
each PE sample. Finally, two laboratories reported results for all three of the PE samples within the Method
D93-96 criteria.

In general, a majority of the laboratories reported a positive bias when compared to the known flashpoints of the
samples. Furthermore, several laboratories did  not follow the instructions provided after the self-audit.  Two
laboratories did not report the results to the nearest 1°F Instead, the laboratories reported results to one decimal
place. In addition, two laboratories did not analyze all samples in duplicate as requested. Finally, one laboratory
sent the PE samples from the facility being considered for program inclusion, where ignitability is apparently no
longer performed, to an affiliated network laboratory. This transfer occurred without notification to the authors.
This network facility was not previously audited for this program.

SUMMARY
Based upon a review of the initial ignitability study, self-audit review, and subsequent PE study,  the  use of
ignitability data for actual waste testing should be highly scrutinized. The regulatory-approved methods require a
significant amount of  operator experience in conducting tests that are consistent  and method  compliant.
Evaluating ignitability through  audits and frequent double-blind  PE studies should be implemented  to verify
acceptable  performance. Without continuous  review  of this method,  inaccurate ignitability results could be
causing  your waste to be disposed of improperly.

ACKNOWLEDGMENTS
Mr. Chuck Wibby and Environmental Resource Associates for providing technical assistance in identification and
evaluation of compounds for the ignitability performance evaluation study, providing  assistance  in collection,
homogenization, and repackaging of the Chrysler waste stream  sample and supply of all PE samples in this
study.

Mr. Brian Miller, of Chrysler Corporation for collection,  packaging, and shipment of the waste stream sample
used in this PE study.

REFERENCES
American Society for Testing and Materials. D93-96 Standard Test Methods for Flash-Point by Pensky-Martens
   Closed Cup Tester. Philadelphia, PA:ASTM, 1996.
American Society for Testing  and Materials.  E502: Standard Test Method for Selection and Use of ASTM
   Standards  for the  Determination of Flash Point of  Chemicals by  Closed  Cup Methods.  Philadelphia,
   PA: ASTM, 1994.
American Society for Testing and Materials. D93-90: Standard Test Methods for Flash- Point by Pensky-Martens
   Closed Cup Tester. Philadelphia, PA:ASTM, 1990.
American Society for Testing  and Materials.  D3278-81  Test Method  for Flash-Point of Liquids  by Setaflash
   Closed Tester. Philadelphia, PA:ASTM, 1981.
American Society for Testing and Materials. D93-80: Standard Test Methods for Flash-Point by Pensky-Martens
   Closed Cup Tester. Philadelphia, PA:ASTM, 1980.
Dupes,  L. J.  and G.  Rose. "Conducting  a Performance Evaluation Study -It's Not Just Analytical  Results."
   Eleventh Annual Waste Testing & Quality Assurance Symposium. Washington, DC, 1995.
United States Environmental Protection Agency. Code of Federal Regulations, EPA Regulations for Identifying
   Hazardous Waste, 40 CFR Part 261, Subpart C 261.21. Continuously Updated.
United States  Environmental Protection Agency. "Chapter Seven: Characteristics Introduction and  Regulatory
   Definitions." Test Methods for Evaluating Solid  Waste, Physical/Chemical Methods, SW-846, Third Edition,
   Update III, Office of Solid Waste, December 1996.
United States Environmental Protection Agency.  "Method 1020A Setaflash Closed-Cup Method for Determining
   Ignitability." Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846, Third Edition,
   Update I, Office of Solid Waste, July 1992.
United States Environmental  Protection Agency.  "Method  1010:  Pensky-Martens Closed-Cup Method for
   Determining Ignitability." Test Methods for Evaluating Solid Waste, Physical/Chemical Methods,  SW-846,
   Third Edition, Office of Solid Waste, September 1986.
                                                  178

-------
         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
          Table 1. Initial Ignitability Performance Evaluation Study
                     56789
                     Number of Participating LataoratortM
10
              12      13
             Acceptance Limits
Table 2. Ignitability Performance Evaluation Study PE Flashpoint Sample #1
                         56789
                          Number of Participating Laboratory*
                                                              10
               11      12

               Acceptance Units
                                     179

-------
                      WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
             Table 3. Ignitability Performance Evaluation Study PE Flashpoint Sample #2
140
100
                                    56789
                                      Number of Participating Laboratories
                                                                         10
11
                                                                                       12
             Table 4. Ignitability Performance Evaluation Study PE Flashpoint Sample #3
120
                                    56789
                                      Number of Participating Laboratories
                                                                         10
11      12

Acceptance (Junta
                                                                                               13
                                                 180

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                   TECHNIQUES FOR IMPROVING THE ACCURACY OF CALIBRATION
                               IN THE ENVIRONMENTAL LABORATORY

                                           Dennis A Edaerlev
        Quanterra Environmental Services, 880 Riverside Parkway, West Sacramento, California 95605

Abstract
Consistent  and reliable procedures  for generating calibrations are essential to accurate laboratory  results.
Unfortunately the  interpretation  of acceptable  practice  is often based on misunderstanding or derived from
practices commonly utilized in non-environmental methods, and therefore does not provide a reliable means for
maintaining  data  quality.  This paper  presents a demonstration that some common practices  used  in the
calculation and evaluation of calibration factors, including the use of unweighted  regression and the associated
correlation coefficient, are inappropriate for environmental analysis due to high relative errors which result at the
low end of the curve. Alternate criteria for evaluation of calibration curves are  proposed based on the Relative
Standard Error (%RSE). Statistical derivations and examples are presented to demonstrate how this  approach
provides an improved measure for the evaluation of calibration data based on weighted regression. Other related
considerations for assessing acceptability of calibration data are also  presented.

Introduction
Any analytical measurement must employ reference elements to ensure traceability to relevant basic quantities.
The quality of a calibration depends on the uncertainty of the reference, the appropriateness of the reference and
how well the calculation procedures match the requirements of the analysis. A majority of the methods employed
in  the  environmental laboratory are based on a relative calibration  where standards of known  content and
concentration  are  analyzed  by a suitable  detector.  The  responses  of samples analyzed  under  the same
conditions are then used to  calculate concentrations by numerical  interpolation to a response curve from the
calibration standards.

The only criterion for initial calibration  in SW-846 method 8000A1 reads,  "If %RSD is less than 20 an average
calibration factor can be  used otherwise data should be fitted to a curve." There are many examples of well
defined and reproducible  calibrations which either due to nonlinearity,  or a non-zero intersection of the axis will
not meet this 20% criterion for acceptability. Recognizing  this, update  III to SW-846, Method 8000B2 provides
additional direction on use mid evaluation of least squares regression, adding criteria for higher order curves.
However,  several  critical issues are  not  sufficiently considered and overall the  current guidance remains
incomplete with regard to error weighitng and evaluation of acceptability.

Linear Calibration
The most commonly adopted option for handling calibration data, which is linear but  does not meet the criteria
for averaging, involves the calculation of coefficients for a linear equation of the  form:
                                        y = Ax + B
Equation 1
4.0 -,
35 -
§ 3'° '
X 2.5
3fc
720
w
§ 1.5 •
a
0.5 -
0.0 -
c
Average
%RSD = 15.0 ^ 9
^
.' Linear fit
L ,•' ^ = 0.928
S
y
S-*
12345
Concentration (x)
                                          Concentration is defined here as the independent variable (x) and
                                          response as the dependent variable (y) in compliance with method
                                          8000B. A least squares regression  is employed and the value of
                                          the correlation coefficient (r) or the coefficient of determination (r2)
                                          is  evaluated as  a measure of  acceptability.  A value  of  1.00
                                          represents a perfect correlation. Generally in practice, a value of r2
                                          greater than 0.990 is considered satisfactory.

                                          The practical difficulty encountered in this approach is displayed in
                                          Figure 1. For the example data set, the line based on an average
                                          calculation  (shown as a  dotted  line)  easily  meets  the  20%
                                          acceptance criteria for %RSD, yet r2, does not meet the criteria for
                                          acceptability with a value of 0.928.

                                          Figure 1
                                                   181

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
To understand this inconsistancy we must examine more closely the statistical difference between an average
calculation and the regression line. The average calibration factor is determined according to equation 2.
                                          C = -n £
Equation 2
                       Where:
                          C = Average Calibration factor
                         C, = Calibration factor for calibration level i (y/x,)
                         n = number of calibration levels

This and all following equations may also be adapted to internal standard methods by substituting the relative
response calculated as in Equation 3 for the measured response (y/).
                                         ./Relative _ ,/
                                         Yi      -yi
                                                        v/s
 Equation 3
                                       y, = response of target analyte
                                   x/s = concentration of internal standard
                                     y/s = response of internal standard
Graphically the average  and  associated  error limits based on  ±1
standard deviation are shown in Figure 2. This type of normalized plot
makes  visual  examination of the  data  more straight-forward3  as
values at the low end of the calibration are shown at the same relative
scaling as those toward the high end.

The average represents the value which will minimize the variance in
the calibration factors (s2^ for all  calibration points4 For reference the
calculation of variance is shown in Equation 4.

                                                         Figure 2
1.1 1
15 0.85
"• 0.9
0 0.85
"§ 0,6
& 0.7S
0 0.7
0.65
0.0
• ' -k +1 Std Dev

	 — - -1 Std Dev - — — — ^
"•«
0 1.00 2.00 3.00 4.00 500
Concentration (x)
                                          2 _
                                         SC -
Equation 4
If y, is defined as the expected response for calibration level i from the relationship y, = x,C, by substitution the
variance can also be expressed as:
                                         c2 -
                                         Sc-
                                                 n-1
Equation 5
Using the method of least squares we can determine a mathematical relationship between the dependent and
independent variables which minimizes the residual variance.  By definition the residual variance represents the
variability due to experimental  error5 and does not include that contribution to variance which  is attributable to
differences in the independent variable. The residual variance of yon x is defined as:
                                                  n-1
 Equation 6
A comparison of equations 5 and 6 demonstrates that the average value is the same as a coefficient derived
from least squares regression if a weighting of 1/x2 is applied. The average calibration factor gives a result which
is identical to that produced by the method of least squares using (1/concentration2) weighting and intercept
                                                   182

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
forced through zero.
Alternatively a generalized equation for calculating the coefficient which fits the simple relationship y = Ax and
minimizes the residual variance without weighting is determined by substituting Ax for y in Equation 6 and setting
the derivative with respect to A equal to zero. This gives for the calibration factor:
                                                                               Equation 7
For the  example data set, calibration coefficients  and  residual variances are  compared in table  1.  The
coefficients  determined from these two approaches are quite  different  and it  is obvious that an  average
calibration coefficient does not minimize the residua! variance of y on x.
                                 Table 1
        Calculation type    CF
                          s2
       Average
       Linear fit
0.916
0.784
1.335
0.768
Transforming the vertical axis of Figure 2 to a non-normalized  form
gives the plot shown in Figure 3. A comparable plot for the unweighted
regression  line using the same set of example data  is presented in
Figure 4.  Both figures  represent graphically, with the dashed lines,
effective error weighting based on ±1 standard deviation.
                                                                                   Average
                                                        12345
                                                          Concentration (x)
               Linear Fit
               1234
                Concentration (x)
                                                                                     Figure 3
                                        Figure  3 demonstrates that 1/x2 weighting gives a relationship that
                                        emphasizes precision at the  low end  of  the  calibration  range.  In
                                        environmental analyses frequently the objective  is to ensure that
                                        target analytes do  not  exceed  defined regulatory or action limits,
                                        hence reducing quantitation error at low concentrations is  especially
                                        important.
               Figure 4
Deriving the best form for  a  calibration curve must include  consideration of the weighting factors which  are
appropriate to the requirements of the analysis. As was shown empirically for the single factor calibrations and
can be  proven for the general case6, the value of the coefficients is sensitive to the weighting. Method 8000B
requires that at least three replicates at a minimum of 5 concentration levels are used to derive weighting factors
which are defined as the inverse of the standard deviation squared for each concentration. In practice this is an
especially burdensome requirement, both in the amount of data required and  the computational difficulty which
results when using individually derived weighting values for  each calibration  level. Faced with the options in
method 8000B, most  laboratories  will  probably choose  curve fitting without any consideration of weighting
regardless of the potential negative impact on data quality.

For many methods in the environmental laboratory errors in measurement are proportional to the magnitude of
the parameter measured. Likewise it is common to consider percentage errors relative to the concentration as we
have demonstrated in  the case of the average calibration calculations. Rather than  perform multiple replicates
each time a calibration is run, the laboratory  should make an initial determination of the relationship between
concentration  and the standard deviation throughout the calibration  range,  Where  a direct proportionality is
found,  all subsequent  calibrations should be based on the 1/concentration2 weighting which  is consistent with
average calibration curves.  Conversely, if standard  deviations are of equal absolute magnitude throughout the
concentration  range,  calibrations should  be  unweighted  and  average calibration factors should  not  be used.
                                                   183

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
When the weighting is expressed as a function of the concentration most data systems are capable of perforating
the calculation.

Percent Relative Standard Error
The magnitude of the residual variance with error weighting applied will provide a measure of the experimental
error for the derived curve. Figure 5 shows the  relationship between variance and calibration coefficient for a
fitting equation of the form: y = Ax.

The value of  relative standard deviation defined  according  to Equation  8 has widespread  acceptance as a
measure of the error associated with an average calibration.
                 1     ¥*-c')'
 %RSD = 1OOx^rxV ^  -,
                 c  v    ""1
                                                                                 Equation 8
     \
     Weighting = 1/xs
                          Weighting = 1
    0.7
         0.75
              0.8   O.B5   0.6

                Calibration (actor
Because standard deviation is equal to the square root of variance, it
can be  shown that %RSD is also equal to the square  root of the
weighted residual variance of y on  x, calculated  as a percentage,
using a derivation similar to that for Equation 5.

Likewise the coefficient of determination (r2) is normally, recognized
as an indication of error associated with regression curves.
Figure 5
                                                                               Equation 9
                             (    \2
Since for any set of data the I,(y-9)  term is not dependent on the form of the calibration curve function or the
coefficients, the value of (1-r2) will be directly  proportional to the unweighted residual variance as defined  in
Equation 6.

It becomes apparent that for an  average calibration, the coefficient of determination calculated will not provide
an optimum measure of error as  r2 applies only to an unweighted least squares determination. The %RSD on the
other hand is  only meaningful as a measure of error when  applied to an average calibration. A generalized
indicator of error,  Percent Relative  Standard Error (%RSE), which can be applied to any form  of weighted
regression  function is derived  similarly to  %RSD making  adjustment for the degrees  of  freedom  in the
relationship by replacing  the n   1 factor with n  - p where p is an integer equal to the number of coefficients as
defined in Equation 10.

%RSE is equivalent to %RSD  when calculated for the average. In Figure 6 the least squares line derived using
(1/concentration2) weighting is  compared to the average line and the unweighted least squares line for the same
data set used  in the previous figures. In this example the elimination of the  zero point is responsible for an
increase in the %RSE for the weighted line (p = 2).

                                                                             Equation 10
                      Where:
                              y, = Actual response of calibration level
                              y, = Calculated response from curve
                              p = number of terms in the fitting equation
                                     (average = 1, linear = 2, quadratic 3)
                              n = number of calibration points
                                                   184

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
             1234
               Concentration (x)
                                       The improved low-end accuracy for this weighted line compared to the
                                       unweighted  regression  line  is shown  in  Table 2.  Concentrations
                                       measured close to the lowest calibration point would be reported with
                                       values  approaching 100%  lower by  using an unweighted  linear
                                       calibration as compared to the weighted curve.

                                                                     Table 2

                                                                   % Error (calculated - True)
Concentration
0.2
0.5
1.0
2.5
5.0
Weiahted
1.7
1.2
2.7
16.7
28.7
Unweiahted
109
29.9
2.8
25.1
8.9
             Figure 6

Only recently, in the  update to SW-846 method 8000B and  also in a draft of guidance for development or
modification of water  methods7 has the EPA recognized the importance of considering weighting in calibration
calculations. These documents do  not, however provide options for evaluating weighted regression fits. The
coefficient of determination (COD) used in method 8000B for evaluating polynomial curves is "weighted" only to
adjust for the degrees of freedom  in the fitting equation and does not provide a measure of error which is suitable
for regression curves derived with error weighting.

Non-Linear Calibration
While it is reasonable to use the simplest mathematical  relationship, which gives acceptable accuracy, it should
not always be  assumed that an average or linear curve  is preferred. Some detector systems commonly used in
the environmental laboratory are  inherently non-linear. As an example the
electron  capture  detector (ECD) commonly utilized for its sensitivity to
chlorinated  pesticides and herbicides can for most  analytes  provide
calibrations  which are linear over a 20X concentration  range. It  is often
difficult to optimize detector conditions for multiple analytes  with widely
varying electron  affinities  and even under  ideal conditions  many  ECD
detectors show linearity within 20%  RSD only over a narrow concentration
range.  The calibration data plotted  in Figure 7 was curved for enhanced
accuracy over a 32X range by applying a quadratic fit. While the %RSD for
the average was not  acceptable at 44.9% the quadrati  curve  gives a
%RSE of only  3.7%.
                                                             Figure 7
MCPP (ECDI
           Awrage
 16     32
  Concentration (x)
In lieu of higher order curve fitting the approach often defaults to reducing the calibration range, which normally
requires re-analysis of some or all calibration points as well as increasing the number of sample dilutions, both of
which will add  to the cost of the  analysis without necessarily providing significant improvement in  quality. A
simple recalculation of the curve  parameters based on a quadratic fit allows the full concentration range of the
calibration data to be utilized with improved accuracy. The use of second or higher order curves should not be
applied simply to  achieve minor improvements in the %RSE that will allow an otherwise unacceptable curve to
be  used. Justification  for  higher  order calculations should  be  based  primarily on an  understanding  of the
performance characteristics of the detector or the method. The use of higher order curve fitting should also not
be substituted for proper instrument maintenance, nor should  an effort be made to extend calibrations beyond
the detector saturation level.

Evaluation of Calibrations
Whenever possible, data evaluation should be performed against rigid criteria that will prevent any tendency for
analyst bias to  affect reported results and allow automated data validation processes to be implemented. In a
production-oriented laboratory setting, visual examination of every calibration curve is not routinely performed.
                                                   185

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Rather curves are evaluated against calculated criteria only. Modern data systems allow regression parameters
to be calculated with very little analyst effort. Although the %RSE calculation  is not commonly provided it can
usually be implemented either in a user function or by exporting the results to a spreadsheet.
              Toluene (PID)
            10    20    30    40
               Concentration (x)
    14 •
                 0.4    0.6
              Concentration (x)
The behavior of the calibration curve near the reporting limit should
also  be  a  primary  consideration  for  environmental  methods.
According to the proposed update  to SW-846 method 8000B, data
should be considered unreliable for instrument response less than 3
times the y-intercept from the curve if the value is positive, or less
than the concentration calculated from zero response if the y-intercept
is negative. The use of weighted curve  fitting will greatly reduce the
possibility that the intercept exceeds these limits. The unweighted line
in Figure 8a appears to provide a very good fit to the data with r2 =
0.999. The expanded view near the low end of the calibration shown
in Figure  8b clearly demonstrates the improved fit for the weighted
line.

Figure 8a

Recommendations for the minimum number of data points needed to
prepare a calibration  vary  from one to fifteen depending  on the
method and the order of the fitting equation. Normally five points are
considered  adequate  for linear  curves.  Some  problems  may  be
corrected   by   elimination  of  data  points  from  the  calibration
calculation.  This should only done  if they are either the highest or
lowest concentration and the number of points remaining meets the
minimum  requirements of the  method.  The analyst should  also be
aware that removing the low point might adversely affect the reporting
limit, as quantitation must not be performed outside the concentration
range of the calibration.
                                       Figure 8b

When  second order curves are evaluated, the acceptability should
include an evaluation of the curve  inflection points. In Figure 9, the
quadratic curve (solid line) provides  a  significant  improvement  in
%RSE over the weighted least squares line (dashed). However, the
y-value of the curve inflection for this line is below the response of the
highest data point,  thus quantitation near the upper  end of the
calibration range could give erroneously high results.
                                                        Figure 9
                                                                             Freon113{ELCD)
                                12-
                                10 -.
                              3
                              5
                                 64
                                                                        2 -
                                            5        10
                                          Concentration (x)
                                                                                                     15
As a rule, all points in the calibration curve should demonstrate a consistent relationship between concentration
and response (response increases with  increasing concentration). According to SW-846 method 8000B "... the
curve must be continuous, continuously differentiable and monotonic over the calibration range."

Summary
The correlation coefficient as a criterion for evaluating regression curves does not apply to weighted regressions
that are necessary for accurate low concentration reporting of environmental data. The correlation coefficient is
also not consistent with the %RSD criterion used to evaluate average curves, often leading to calibration data
that is not acceptable for a linear regression based on the correlation coefficient but does meet %RSD criterion
of an average calibration.

The Percent Relative Standard Error (%RSE) provides an improved criteria for evaluation of calibration curves in
environmental laboratory methods. The advantages of this approach are as follows:
                                                   186

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


•  All curve types can be calculated with the same (1/concentration2) weighting applied, This places emphasis
   on relative error with improved accuracy at concentrations near the reporting limit.

•  The %RSE criterion is consistent for evaluation of all curve fitting types. Interpretation of acceptability for
   calibration curves is simplified with the same criterion applied to all curve types and no conflicting criteria.

•  The most  appropriate curve  fitting  model can  be applied to each set of calibration data with  evaluation
   criteria between different curve types directly comparable.

The evaluation of all calibration data should include,  as a minimum, the following checks:
   •   %RSE < maximum limit
   •   concentration level of low standard < reporting  limit
   •   low point intersection values < reporting limit
   •   Number of calibration levels meets method requirements
   •   All points must be monotonically increasing
   •   Also for second or higher  order curves inflection points should  not be within the calibration range

With the widespread availability of powerful computer  hardware and software in the laboratory, it is unnecessary
to sacrifice data quality for the sake of simplification. Analysts should be familiar with the  productivity and quality
benefits of least squares curve-fitting algorithms. Clients and  regulators  must understand the importance of
weighted curve fitting and the need for complete and consistant evaluation criteria, which will provide high quality
results for the lowest cost.

References
1. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods,  SW846, 3rd Edition, Final  Update 1,
   July 1992, Section 8000A
2. Test  Methods  for Evaluating Solid  Waste, Physical/Chemical Methods, SW846, 3rd Edition, Revision 2,
   December 1996, Section 8000B
3. R. Cassidy and M. Janoski, LC-GC 10(9), 692  (1992)
4. Allen L. Edwards, An Introduction to  linear Regression and Correlation, W.H. Freeman and Company, 1976
5. Maurice G. Kendall and William R. Buckland, A Dictionary of Statistical Turns, Hafner Publishing Co., 1972
6. J.C. Miller and J. N,  Miller,  Statistics for Analytical  Chemistry, John Wiley, 1984
7. Guide to  Method Flexibility  and Approval of  EPA  Water Methods,  U.S.  EPA, Office of Water, Draft:
   December 1996.
8. Grant T. Wernimont, Use of  Statistics to Develop and Evaluate Analytical Methods, Association  of Official
   Analytical Chemists, 1985
9. G. Bulmer, Principles of Statistics, Dover Publications, 1979
                   QUALITY CONTROL PROTOCOL FOR ANALYSIS OF DRUGS AND
                         EXPLOSIVES USING ION MOBILITY SPECTROMETRY

                               Juliana Homstead and Edward J. Poziomek
           Old Dominion University, Department of Chemistry and Biochemistry, Norfolk, VA 23529
                                             (757)440-4005
                                      email: epoziome@ODU.EDU

Ion mobility spectrometry  (IMS) is widely used  as a  screening tool  in the detection  of contraband drugs and
explosives. It is also gaining popularity as a semiquantitative instrument used in the research of these and other
compounds, including  hazardous environmental  contaminants. We  have successfully used  IMS in studies
involving cocaine and TNT, and have developed  a quality control protocol to assess and ensure the quality of our
data.  This protocol employs  cocaine  hydrochloride as  the  reference  standard  in  the  positive  mode and
trinitrotoluene  (TNT) as the reference standard  in the negative mode. A five point calibration  curve was


                                                  187

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


generated for each of the reference compounds in order to determine a concentration level suitable for quality
control (QC) check solutions. We have established peak amplitudes and reduced mobility constant (Ko) for the
QC check solutions that must be met each day before proceeding with analyses. Any deviation from these
criteria requires assessment of the  problem and appropriate  corrective action. We have found this procedure
helpful in maintaining data quality and in providing an early indication of potential problems.
                 REFERENCE MATERIAL FOR THE ANALYSIS OF METALS IN SLUDGE

 Stuart J. Nagourney. Nicholas J. Tummillo, Jr., John Birri, Kenneth Peist, Bruce MacDonald and Jean S. Kane

INTRODUCTION
New Jersey Department of Environmental Protection (NJDEP) regulations limit the levels of discharged toxic
substances in sludge effluents by sewage treatment plants. Once maximum contaminant levels are established,
they become part of the facility's operating permit. When chemical analyses indicate that these permit levels are
exceeded, the NJDEP has statutory authority to assess significant monetary penalties.

Sludge samples vary widely in their physical and chemical composition,  ranging from liquids with low dissolved
solids content  and  small quantities  of  organics and  metals, to multi-phase samples and cakes  with solids
contents  often  greater than 50 % and concentrations of metals and other constituents at percent levels. Since
permit limits for metals  are based  on the amount leached during mineral acid digestion, these acid  extractable
concentrations, rather than total  concentrations, are the values of interest. With several options for methods of
sample preparation  and measurement being  proposed by State  and Federal agencies, comparability of data
among methods is a critical issue.  Reference materials of similar matrix and composition to the various sludge
matrices, and having known metal levels with defined uncertainties, are essential to insure that an accurate
assessment of the data being generated is made,

To  expand  the  utility  of existing  environmental SRMs,  NIST  has  initiated efforts  provide  data on the
acid-extractable levels of metals in selected new solid sample environmental SRMs. A recent  paper by these
authors described a collaborative project among the NJDEP, USEPA, Region II Technical Support Branch and
NIST to develop sludge reference materials from  domestic and industrial sources having reference values for
their acid-extractable metals content. The study  was successful for the domestic  material, resulting in the
derivation of reference  values for  EPA acid-extractable methods as  part of SRM 2781, Domestic Sludge. The
addendum to SRM 2781 contains leachable mass fractions for 14 metals and compares the leach recoveries to
total metal  concentrations  obtained separately. This  paper discusses the  analysis of the industrial sludge
material,  NIST SRM 2782.

EXPERIMENTAL
Reference Samples
The  candidate industrial sludge reference material  SRM  2782,  was  prepared from  more than  100 kg of
electroplating waste supplied by AT&T Bell Laboratories,  Murray Hill, NJ.  This material was shipped to NIST,
wherein  it  was freeze-dried,  processed,  radiation sterilized  and homogenized by  contractors  according to
procedures used to prepare United States Geological Survey and NIST geological reference materials. Samples
were aliquotted and shipped to  the  participating laboratories, the NJDEP Bureau  of Radiation and  Inorganic
Analytical Services (BRIAS) and  USEPA, Region II, Technical Support Branch, for chemical analysis.

Sample Preparation
The  NJDEP used open  vessel hot plate digestion (NJDEP Method 100)  for all of their acid digestions. The
USEPA used open vessel hot plate digestion (USEPA  Method 3050) for the industrial sludge and Method 3050
and closed vessel microwave digestion  techniques (USEPA Method 3051) to prepare the domestic sludge for
measurement. EPA prepared two sets of samples using microwave digestion, one without and one with HCI.

Instrumentation
The  NJDEP used  a  Perkin-Elmer (PE) Model 5000 atomic absorption spectrometer (AAS) for its flame atomic
absorption (FAAS) metal measurements. A similar unit, equipped with a PE Model 500 furnace and a PE AS-50

                                                 188

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


autosampler, was used for the graphite furnace atomic absorption (GFAAS) measurements. The NJDEP also
employed  a Thermo-Jarrell  Ash Model 25  sequential  inductively  coupled plasma  emission  spectrometer
(ICPOES) for some of its metal determinations. The USEPA used a CEM Model MDS-2000 microwave digestion
system for some sample preparations and a Thermo Jarrell-Ash Model 61 Simultaneous ICPOES and a PE 5100
GFAAS, equipped with a model AS-600 furnace and  AS-60 autosampler for  its  metal determinations. The
USEPA performed the metals measurements by either simultaneous ICPOES and GFAAS, using one technique
per metal. The NJDEP employed more than one technique to measure most of the metals.

Sample Processing
Each  laboratory  analyzed its own digests by  one or more instrumental methods. Once completed, the two
laboratories exchanged extracts and conducted  another series of measurements using the same techniques.

RESULTS AND DISCUSSION
Developing "reference" values for the composition of real environmental samples requires inclusion of random
errors of measurement and any systematic bias which might be inherent in the method(s) employed to estimate
the true value. Issues such as the method of measurement, the  complexity and composition of the sample matrix
and the concentration and heterogeneity of the analytes all contribute to the overall uncertainty. While relative
uncertainties of -20% to + 20% are expected when  EPA methods such as FAAS, GFAAS and ICPOES are
applied to  nonaqueous media,  reference values based  upon these methods are still useful as confirmation  of
results of environmental monitoring such as the analysis of sludge effluents from treatment plants.

Current Results for Industrial Sludge
The initial  set of analyses performed on NIST SRM 2782 yielded poor agreement between the two laboratories.
For the most recent series of tests, the precision of the 6 aliquots of NIST SRM 2782 ranged from 0.2 % to 2.0 %
for all  elements  measured by  both participating laboratories.  There  is no discernable change in the level  of
precision of the replicate measurements obtained  in the earlier and later  studies Table 1 compares the results
obtained by both laboratories in the  initial ('94)  and most recent ('96) studies. There is no discernable pattern in
the NJDEP data  from 1994 to 1996; the more recent series of USEPA results are generally higher than those
obtained previously. Differences between the grand means of  all  results obtained for each element by the two
laboratories for the more recent set of measurements were within 6% for 6 elements and 10%  for 12 of the 17
metals studied; 2 others were within  12 %. Cd, Cr and K were the only elements outside this range,

Analysis of Variance (ANOVA) can be utilized to evaluate whether the individual element means obtained by the
two laboratories  for each element,  generated  by different methods  of sample preparation  and analysis, are
statistically significant. As shown in  Table 2, for the 14 metals where agreement  between the grand means was
within 12 %. ANOVA analyses, as shown by the F value (ratio of variance between data sets to the variance
within a data set) and P value (measure of the probability the actual sample set  fell within hypothetical
frequencies for infinitely large data populations) showed  that the means obtained by the two laboratories agreed
within specified tolerances for 11 of the 14 metals; the exceptions being Ba, Cr and Mn. This serves as further
corroboration to the extent of  the agreement between the results.

The USEPA obtained good results for 14 metals where extractions were obtained  by both Method 3050 (hot plate
digestion) and Method 3051 (microwave digestion), with  analyses performed by ICP-OES. Agreement is within ±
6 % for all metals studied but Na and K; with a range  in  bias from - 4.8 % for Al to +5.2 % for Zn. These results
are shown in Table 3.  The values  for the alkali metals Na and  K are approximately  30-40 % higher for the
microwave technique. When  the t-test for means assuming unequal variance is applied,  for 6  metals (Ba, Ca,
Cu, Mn, Ph and Zn), the differences between the calculated means obtained by each method were statistically
significant. However, these detected differences are relatively small, especially when compared  to the control
limits  assigned to acid-extractable analytes measured in non-aqueous  media. All standard deviations are ± 2
sigma.

Comparison of Leachable Concentrations for Domestic Sludge and Industrial Sludge
The addendum  to the  Certificate of Analysis  for NIST SRM 2781,  domestic sludge  compares the % leach
recoveries of selected metals with the  values obtained for their total  mass fractions. For  elements where
leachable metal results were  obtained from both the USEPA and NJDEP laboratories, a ratio of leachable metal
concentration to  total metal concentration (obtained by NIST) was calculated. For the domestic sludge these
ratios were between 0.82 and 0.96 for 9 out of  11 elements; only Al (0.50) and Cr (0.71) were below this range.
Preliminary values supplied  by NIST for the  industrial sludge material shows  a different  pattern;  for the  14


                                                 189

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
elements for which such data is available, only 6 have ratios of leachable to total that are greater than 0.80. Na,
K, Mg, Al have ratios below 0.20, the ratio for V is 0.21, Ba and Ni are 0.60 and 0.62 respectively, while Ca is
0.72. Only certain transition elements (Cu, Ph, Zn, Mn,  Fe and Co) have leachable/total ratios for the industrial
sludge that are greater than 0.80. The results are summarized in Table 4. Note that for the industrial sludge all of
the ratios for the  alkali and alkaline earth metals as well  as Al are 0.72  or below. While the reasons for the
specific analyte-matrix interactions are uncertain,  it is apparent that the more complex nature of the industrial
sludge plays a key role in how much metal is leached by the regulatorily-approved digestion methods, and that
the amount  that  leaches  varies significantly  by  element and group. This issue must  be considered when
evaluations  of leachable  metal  concentrations are  made  for environmental  assessment   and  policy
considerations.
                                Table 1. Industrial Sludge - Grand Means
     Element
       Al
       Ba
       Ca
       Cd
       Co
       Cr
       Cu
       Fe
        K
       Mg
       Mn
       Na
       Ni
       Pb
        V
       Zn
     Table 2. Industrial Sludge - ANOVA
                 Analyses
Mean DEP'94
1502 ±43
150 ±1
4687 ± 470
4.0 ±0.1
70.0 ± 6
79.8 ± 1
2485 ± 146
255020± 17000
116±1
508±37
274 ±13
2430±18
125±2
581 ±18
23.2± 1
1266 ±31
Mean EPA'96
1587 ±107
161 ±11
4950 ± 144
11. 4 ±3.0
51 .6 ±7
76.7 ± 4
2459 ± 37
256600± 2300
121 ±2
498 ±69
274 ±7
2431 ±20
95.5 ±6
558±9
17.5 ±2
1181 ±59
Mean DEP'94
1380 ±50
132 ±4
4320 ±141
15.4 ±0.8
54.4 ± 2
55 3 ±2
2270 ± 53
232000±5600
58.9 ±6
441 ±17
224 ±7
2000 ±123
908±2
51 9 ±30
20.6 ±1
1170±21
Mean DEP'96
1528 ±46
144 ±4
4685 ±151
2.3 ±0.1
56.0 ± 3
58.3 ±1
2420 ± 42
253000±14000
78.6 ±21
470±15
247±7
2573 ±307
96.1 ±4
553 ±22
15.9±1
11 58 ±71
      Al
      Ba
      Ca
      Co
      Cr
      Cu
      Fe
      Mg
      Mn
      Na
      Ni
      Pb
      V
      Zn
 F
0.82
33.9
1.73
0.25
3.67
0.51
4.09
2.25
69.6
0.06
4,20
3.52
2.32
0.07
 P
0.38
0.00
0.22
0.62
0.00
0.49
0.06
0.14
0.00
0.81
0.09
0.09
0.14
0.79
                                                                   DEP/EPA '96
                                                                       1.038
                                                                       1.118
                                                                       1.056

                                                                       0.921
                                                                       1014

                                                                       1.059
                                                                       1.109
                                                                       0.945
                                                                       0.994
                                                                       1.010
                                                                       1.100
                                                                       1.019
CONCLUSIONS AND RECOMMENDATIONS
Reference  values  have   been  derived  for  the  leachable
concentration of  14 metals  in  an industrial  sludge  standard
reference material (NIST SRM 2782). Both open vessel hot-plate
acid digestion (NJDEP  Method 100 or USEPA  Method 3050) or
microwave  digestion (USEPA  Method  3051)  are  appropriate
methods for  sample preparation of these  materials and  the
leachable  concentrations may  be  measured by  either  FAAS,
sequential or simultaneous ICPOES. This information supports
the application  of performance-based  methodology  since there
are several combinations of sample preparation  and measure-
ment systems  that can   achieve  similar  results. Following
complete statistical review by NIST, sludge  Standard Reference
Materials from  both domestic and industrial sources-will  be
commercially  available.  Reference  values  for  their  acid
extractable metals content  and associated  uncertainties  will  be
provided to support the  quality assurance of  sludge  metal
measurements.  Analyses of these SRMs should be  required as
part of  a  POTW's  compliance data submitted  to regulatory
agencies  Such   as the   USEPA  and  NJDEP-  It  is  also
recommended that these materials  become part  of any future
laboratory certification program for sludge effluents.
                                                 190

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


                             Table 3. Industrial Sludge - Method Comparison

     Element               Method 3050                       Method 100               3050/100
                      Mean             SD             Mean            SD
       Al             1520             18             1540              8             0.987
       Ba              148              2              144              1             1.027
       Ca             4870             58             4720             34            1.032
       Co             58.5              0.9             57.0             0.9            1.026
       Cr             57.9              0.6             58.2             2.4            0.994
       Cu             2500             37             2420             45            1.033
       Fe            260000           4500           236000           1100           1.102
       K             72.4             11.8            80.1             1250           0.904
       Mg              470              5              490              7             0.959
       Mn              256              3              246              3             1.041
       Na             2300             39             2470             60            0.931
       Ni             99.6              3.7             95.1             1.1            1.047
       Pb              545              6              524             13            1.040
       Zn             1210             15             1210             11             1.000

                   Table 4. Domestic Sludge -Recoveries and Comparisons w/Certificate

                    Element         Domestic Sludge           Industrial Sludge
                                 Ratio: Leachable / Total     Ratio: Leachable / Total
                      Ag                  0.88
                      Al                  0.50                      0.11
                      Ba                    -                        0.60
                      Cd                  0.86
                      Ca                  0.93                      0.72
                      Co                    -                        0.82
                      Cr                  0.71
                      CU                  0.96                      0.94
                      Fe                  0.87                      0 95
                       K                    -                        0.02
                      Pb                  0.91                       0.97
                      Mg                  0.82                      0.18
                      Mn                  0.82                      0.86
                      Na                    -                        0.19
                      Ni                  0.90                      0.62
                      V                                            0.21
                      Zn                  0.88                      0.93

                - Only NJDEP  or USEPA leachable results were available,  mean  values
                  between  the two laboratories was not computed.
                               PERSPECTIVES ON DIOXIN ANALYSIS

                                           Stevie Wilding
                  USEPA, Region III, Office of Analytical Services and Quality Assurance,
                               839 Bestgate Road, Annapolis, MD 21401
                               Phone: 410-573-2733, FAX: 410-573-2771

Dioxin analysis  is very complex.  Many  people do not understand how dioxin analysis is  performed within a
laboratory or how to validate the  data when they receive it. Through my technical assistance to internal and
external customers concerning dioxin analysis, it has become clear they do not understand and are afraid to

                                                191

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


understand the methodology of dioxin analysis. Prior to validating PCDD/PCDF data, it is imperative to have
complete knowledge of the method used. The purpose of this paper is to try and simplify dioxin analysis. To
break the analysis into understandable pieces  and to compare it to other organic analyses. The basis for the
Dioxin/Furan analysis is Selected  Ion Monitoring (SIM) Gas Chromatography/Mass Spectrometry (GC/MS). The
objective of this paper is to  explain  SIM  GC/MS  and isotope dilution  for  the average nonchemist.  The
advantages and disadvantages of using isotope dilution and  SIM over the "routine" organic methods will be
discussed.
                  INTERPRETATION OF GROUND WATER CHEMICAL QUALITY DATA

                                            G.M. Zemanskv
              Compass Environmental, Inc., 3000 W. 19th Court, Lawrence, Kansas 66047-2300

ABSTRACT
Ground water is sampled to assess its quality for a variety of purposes. Whatever the purpose,  it can only be
achieved  if results are  representative of actual site conditions and are interpreted in  the  context of those
conditions.

Substantial costs are incurred to obtain and analyze samples. Field costs for drilling, installing,  and sampling
monitoring wells and laboratory costs for analyzing samples are not trivial. The utility of such  expenditures can
be jeopardized by the manner in which reported results are interpreted as well as by problems in how samples
were obtained and analyzed. Considerable  attention has been given to standardizing procedures for sampling
and analyzing ground water. Although following such standard procedures is important and provides a necessary
foundation for understanding  results,  it neither guarantees that reported  results  will be representative nor
necessarily have any real relationship to actual site conditions. Comprehensive data analysis and evaluation by a
knowledgeable professional should be the final quality assurance step, it may indeed help to find errors in field or
laboratory work that went  otherwise unnoticed, and  provides the best chance for real  understanding  of the
meaning of reported results.

The focus of this paper is  on the interpretative part of the process.  Although formal interpretation necessarily
comes late in  a project, when data  have been generated and the report is being written, it will be most useful if
relevant elements can be integrated into the project from the beginning. When  this is done, it increases the
likelihood  of achieving project  objectives as well as understanding the data. To facilitate interpretation, the
following steps should be included:

1. Collection, analysis, and evaluation of background data on regional and site-specific geology, hydrology, and
   potential anthropogenic factors  that could influence  ground water quality and collection of background
   information on the environmental chemistry of the analytes of concern.
2. Planning and carrying out of field activities using accepted standard procedures capable of  producing data of
   known quality.
3.  Selection  of a laboratory to analyze ground water samples based  on careful  evaluation of laboratory
   qualifications.
4.  The use of appropriate quality control/quality assurance  (QC/QA)  checks  of field  and laboratory work
   (including field blank, duplicate, and performance evaluation samples).
5. Comprehensive interpretation of reported  analytical data by a knowledgeable professional. The  analytical data
   must be accompanied by appropriate QC/QA data, be cross-checked using standard water quality checks and
   relationships where possible, and be correlated with information  on regional  and site-specific geology and
   hydrology, environmental chemistry, and  potential anthropogenic influences.

Application of this sequence of steps and  their importance in interpretation of ground water quality data are
discussed  in this paper. The discussion includes several  illustrative case examples.

INTRODUCTION
There  area number of reasons we might want to know about groundwater  chemical quality.  For example, we

                                                  192

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


might be interested in using an aquifer as a source of water today or we may want to ensure that current
practices haven't caused contamination so that ambient ground water quality remains legally acceptable whether
or not it is used in the future. At first glance, you'd think that obtaining and interpreting ground water quality data
would be a fairly straight forward exercise. Simply follow the usual scientific approach: (1) take a sample; (2)
analyze it; and (3)  compare analytical results to a set of criteria. However, obtaining reliable data and properly
interpreting them turns out to be more complicated than that; even in the relatively simple case of surface water.
When it comes to ground water, everything seems more complex.

To begin with, obviously, with ground water you can't just go down to the stream and get a sample. Because
ground water is underground, you need to do more preparation in advance of sampling. This involves obtaining
background subsurface information and an access point.  Such information will be helpful, both in planning how to
obtain  a sample as well as in interpreting  analytical results when they become  available. Whether you use a
standard monitoring well, direct push technology, or something else, getting access to  sample isn't always easy
and can influence the quality of the sample obtained. There are also a number of  potentially confounding factors
with regard to the next step of the  process, laboratory analysis. Sample quality can change between the time a
sample is obtained and the time it is analyzed and,  even if it doesn't, the overall reliability of laboratory results is
not the sure thing many people assume it is. Finally, selection of appropriate criteria to  compare data to may not
always be straight-forward.  Relevant criteria may either not exist or be incomplete.

Comprehensive data interpretation  by a knowledgeable professional should be the final  quality assurance step of
any project involving ground water quality data. It may indeed help to find errors in field or laboratory work that
went otherwise unnoticed and provides the best chance for real understanding of the meaning of reported results.
Proper project planning should prepare for this final step by obtaining  relevant information early on and including
relevant data collection into field segments of the project. The following steps must be integrated into and carried
out throughout the project to facilitate final interpretation:

1. Collection, analysis, and evaluation of background data on regional and site-specific geology, hydrology, and
   potential anthropogenic  factors that  could influence ground water quality  and  collection  of background
   information on the  environmental chemistry of the analytes of concern.
2. Planning and carrying out of field activities using accepted standard procedures capable of producing data of
   known quality.
3. Selection  of  a  laboratory  to analyze ground  water samples based on careful  evaluation of laboratory
   qualifications.
4. The use of appropriate QC/QA checks (including  field blank, duplicate, and performance evaluation samples).
5. Comprehensive interpretation of reported analytical data by a knowledgeable professional.

The analytical data must be accompanied  by appropriate QC/QA data, be cross-checked using standard water
quality checks and relationships where possible, and be  correlated with information  on  regional and site-specific
geology and hydrology, environmental chemistry, and potential anthropogenic influences.

BACKGROUND INFORMATION
Background information serves several functions: (1) it facilitates obtaining access to sample groundwater; (2) it
provides guidance regarding selection of appropriate sampling methods and analytical variables; (3) it places
ground water quality data into context; and (4) it provides a quality assurance check. The following background
information  is necessary  to fulfill these functions:  (1) regional and site-specific geology; (2) regional and
site-specific hydrology; (3) information on the environmental chemistry of variables of concern to the project; and
(4) broad  information on potential  anthropogenic influences including site conditions and possible contaminant
sources. The latter includes not only those conditions which may have impacted ground water, but those which
could have influenced sample quality as a result of installation and testing of monitoring wells or otherwise during
the sampling  process. For example, shallow contamination can  be carried down into a deeper aquifer during field
work and  airborne chemicals  may cause trace contamination of ground water samples  if they enter an open
borehole, monitoring well, or sample being placed into a container.

FIELD PROCEDURES
-Access Points
The nature of subsurface conditions will  influence the type of access point that  is possible. Sample quality will
also be impacted by the type of access  point selected. Commonly used ground water  access points  include: (1)
monitoring wells; (2) wells or piezometers installed for other purposes; and (3) direct push technology.


                                                   193

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


The nature of the access point has, in particular, a relationship to the level of total suspended solids (TSS) likely
to be present in samples. TSS levels would be expected to be relatively high in samples obtained using direct
push technology and low in samples obtained from a water supply well. Assuming proper design and installation
(including development), TSS levels in samples obtained from monitoring wells will normally be low  except
where fine-grained materials are screened. However, development is something that is often neglected or treated
in a pro-forma manner when monitoring wells  are installed. Additional development, as opposed to routine
purging, may also be required when there are long periods  between  sampling events. This is  illustrated in Table
1, showing results for inorganic variables in unfiltered samples from a well that was properly  developed prior to
initial sampling but was not redeveloped when sampled again after two years.

Sampling Methods
Available sampling methods are often constrained by the type of access point utilized. Since they will have a
direct bearing on sample quality, the sampling methods used  must  be taken  into account in  planning sampling
events and in  interpretation of the data obtained. This category includes consideration of both field equipment
and procedures. For example, regulatory agencies are more frequently requiring analysis of unfiltered samples.
This may introduce substantial variation into the process, particularly for inorganic variables.

Sample Handling and Preservation
The order of sampling, type  of container, and sample preservation  method  utilized  can  affect the  quality of
samples  analyzed in the laboratory. As discussed further below, U.S. Environmental Protection Agency (USEPA)
specified sample preservation methods and maximum allowed holding times do  not always ensure sample
quality will not  change between the field and laboratory analysis.

Field Analysis/Observations
Reliable  sample  preservation methods  do not  exist  for  some water quality variables.  In  other cases, field
measurements carried  out for other purposes (i.e., well purging) are routinely available and preferable (e.g., to
laboratory analysis for the  same  variable) or field observations can  be made that provide  useful information
otherwise lost if not recorded  at the time. USEPA requires field analysis (by specifying immediate analysis) for
only five  variables: (1) chlorine residual; (2) pH; (3) dissolved oxygen (by probe); (4) sulfite; and (5) temperature.

The validity of  some  USEPA maximum  holding times and  preservation combinations is  questionable. For
example,  although USEPA recognizes that pH  and dissolved oxygen  levels may  change substantially if not
analyzed in the field and therefore requires immediate analysis without holding for them, substantial changes in
alkalinity may  occur for the same reasons but  14 days holding time  is allowed  for this variable.  Because it is
allowed and more convenient, samples are frequently analyzed in the laboratory for alkalinity instead of the field.
Similarly, USEPA allows maximum holding times of 48 hours and seven days for color and total suspended
solids (TSS), respectively.  USEPA also allows seven and 14 days holding time for unacidified  and acidified
volatile aromatic compounds,  respectively. Both color and TSS may change substantially in samples over these
allowed time frames and research has shown both substantial loss of volatile aromatic compounds in less than
seven days in  unacidified samples and that substantially greater holding times than  14 days are appropriate for
many compounds when samples are acidified.1

When sampling ground water, field analysis should routinely occur for the purge variables conductivity, pH, and
temperature, using appropriate equipment operated, calibrated, and maintained in accordance with manufacturer
recommendations. The following field observations related to sample quality may  also be made: (1) color; (2)
odor;  and (3) turbidity.  Although this information  is essentially "free" for the taking, it is often not recorded. In
particular, the  latter observation can provide at least a qualitative indication of the presence of TSS and their
possible  effect on inorganic variables. The difference that filtering  makes in  reported concentrations of many
inorganic  variables when unfiltered  samples  contain high  solids  concentrations is substantial and  readily
demonstrated with split samples.

LABORATORY ANALYSIS
Too often, the analytical laboratory for a project  is selected on the basis of cost only. Such cost savings may
prove to have been very expensively purchased if quality isn't also delivered. Laboratory qualifications should be
carefully  researched in advance of selection. This can be  done by  such things as  reviewing historic results on
performance evaluation samples, auditing the facility, submitting performance evaluation samples, and checking
references. Checks should continue throughout the project with the appropriate use of blank,  duplicate, and
performance evaluation samples.


                                                  194

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Unfortunately,  no laboratory  is perfect. Even the better  laboratories often  get  some  wrong answers when
analyzing USEPA performance evaluation samples. These are in a relatively clean matrix, without the kinds of
interferences that can complicate real world samples, and are being analyzed under test conditions. A laboratory
would be foolish not to make its  best effort on these samples. You  can  expect  somewhat lower quality with
regard to samples being routinely analyzed for the average client. Routine  QC/QA and data validation practices
are not the complete  answer to  data quality.  In  particular,  the latter  usually don't  include  various  data
cross-checks or take into account what is  known about site conditions. In the worst case, laboratory reported
analytical results can be more artifacts of the sampling and analysis process than representative of ambient
ground water quality.

There are a  number of factors which can influence laboratory results. But even when analytical laboratories are
performing well, it  should be recognized that:  (1) at best, they can only analyze samples  in  the condition
received; and  (2) standard limits  of precision  and accuracy allow considerable variation. As noted above, a
number of factors may result in samples  reaching the laboratory which  are  less than  representative  of site
conditions.  Even  when laboratory QC/QA requirements are met, the level  of  allowed  variation  limits the
conclusions which can be based on laboratory data. USEPA Superfund acceptance criteria for percent recovery
of matrix spikes (MS) and laboratory control standards (LCS) are shown in Table 2.23

Data reported by laboratories should be carefully reviewed or "validated" before being  utilized. USEPA national
functional guidelines for data review under Superfund are an example of typical validation guidelines. These
specify  a number  of checks intended  to assure that correct procedures were followed (e.g., holding times,
calibration, blanks, matrix spikes and spike duplicates, and laboratory control standards). Although this type of
review is useful, it is important to  realize its limitations. The standard  conclusion one consulting company uses
after a successful validation process contains the statement that "the data collected during this investigation are
valid as qualified for use in representing Site conditions and for use in risk assessments."4 Since  the validation
process referred to doesn't take relevant site information or  available data cross-checks into account,  such a
statement goes too far.  Neither is it necessarily correct.

CASE EXAMPLES OF  INTERPRETING GROUND WATER DATA
Organic Contamination  at a Superfund Site
Trichloroethylene (TCE) groundwater contamination was recognized at a Superfund site.  The presence of other
halocarbons associated with TCE  degradation, notably cis-1,2-dichloroethylene (c-1,2-DCE) and vinyl chloride,
was also recognized. However, the potentially  responsible party (PRP) claimed the TCE originated elsewhere
and was part of a wider regional problem involving a number of sources and contaminants.  Site work performed
by the PRP in 1996 identified  a variety of other compounds in samples from relatively deep wells drilled using air
rotary equipment. The compounds reported in various samples fell within the following categories:

1. Volatile organic compounds (VOCs) -
   a. Halocarbons including TCE and TCE degradation products.
   b. Petroleum hydrocarbons, including aromatic compounds (e.g., xylenes).
   c. Trihalomethanes,  particularly chloroform and bromodichloromethane.

2. Semivolatile organic  compounds (SVOCs) -
   a. Petroleum hydrocarbons (benzoic acid and naphthalene).
   b. Phenol and other  phenolic compounds.
   c. Phthalates, particularly bis(2-ethylhexyl)phthalate,

Review of these data indicated that other interpretations were much  more likely.  Among the  reasons the data
were suspect were the following:

1. With the exception of halocarbons, most of the compounds involved had not been reported in  samples from
   shallow wells previously drilled  using hollow-stem augers. Contamination in such cases generally moves from
   surface or near-surface sources downward. Furthermore, the concentrations involved were generally low and
   appeared to be randomly distribution rather than in a pattern suggesting  any relationship to possible sources.

2. Air rotary drilling is a  possible source of petroleum hydrocarbon and phenolic compound contamination.

3. Some of the boreholes had been left open to the atmosphere for substantial periods of time  (i.e., on the order


                                                  195

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


   of months) after drilling before monitoring wells were installed in them. The ones open the longest were also
   adjacent to an  Interstate Highway. Petroleum hydrocarbons have been identified in vehicle emissions and
   ambient  urban  air.5 Research has  also indicated the  potential for petroleum hydrocarbons  and phenolic
   compounds in urban air to be transported into ground water.67

4. PRP  consultants  had  not  decontaminated  at least  some downhole equipment (particularly water level
   indicators) when used between boreholes or between monitoring wells. This may cause cross-contamination.

5. THM contamination is common in chlorinated tap water from surface sources, but unusual in ambient ground
   water. Thousands of gallons  of chlorinated tap water from a surface source known to contain substantial
   concentrations of trihalomethane compounds had been introduced into most of the boreholes as a part of the
   testing program during field  work.

6. Phthalates are recognized by USEPA as a common SVOC laboratory contaminant.8

7. Background regional information strongly indicated the PRP was the source of halocarbon contamination and
   that there was no other more widespread regional problem involving halocarbons or other compounds.

Resampling in 1997 provided further confirmation of this interpretation.  Data for the 19 wells sampled in both
years are shown in Table 3  (parts a and b). With the exception of a reduced number of low concentration hits for
phthalate  compounds  (which  are recognized  as common laboratory  contaminants) and,  in  one case,  a
trihalomethane compound, only halocarbons were  reported  in  1997  This indicates  that the various  other
compounds reported in 1996 samples were a  transitory impact of drilling, testing, sampling, and analysis rather
than regional  ground water contamination. This transitory impact dissipated over time as  a result  of natural
mechanisms such as flushing by ambient ground  water flow and biodegradation. The change in  apparent
distribution of halocarbon contamination between 1996 and 1997 apparent from these results may also indicate
that  shallower contamination was  carried downward by intrusive work  performed  in  1996 and that this  also
produced a transitory impact on ground water samples.

Ground/Surface Water Relationship at a RCRA Site
Ground water monitoring at landfills  and other  Resource Conservation and Recovery Act (RCRA) facilities is
oriented towards ensuring that a release of contaminants,  if it occurs, will be detected.  In general,  detection
monitoring  requires that data from wells downgradient of the facility and subject to impact in event of  a release
be statistically compared to data from  upgradient background wells. If the comparison results in a statistically
significant difference with a downgradient increase,  a release is  assumed. However, other regional  conditions
must be considered if such comparisons are to be useful.

In this case, there  is a network of monitoring wells installed up and downgradient of a land treatment  unit (LTU)
located at a refinery adjacent to a major river. Statistical  analysis showed that concentrations of some variables
were elevated in downgradient wells  and that the elevations were statistically significant.  Did this mean that a
release had occurred?

A linkage between the river and the adjacent alluvial ground water aquifer would be expected based on general
principles. This was confirmed by analysis of two lines of evidence: (1) correlation of  ground water  elevations
with  river flow (a surrogate for stage); and (2) statistical  analysis of  water quality data.  Time series plots  of
ground water elevations versus river flow showed an evident visual correlation, which  was confirmed by linear
regression analysis. The correlation was best  for those wells closest to the river (correlation  coefficient of 0.80)
and  decreased with distance from the  river. A comparison of water quality data is supportive. Data  indicating
central tendency for upgradient monitoring wells and the river are presented for six variables in Table 4. For the
three major ions in Table 4, including chloride, river concentrations far exceed those in upgradient ground water.
In these cases, statistical analysis shows that concentrations in samples from downgradient monitoring wells are
significantly higher than  upgradient  ground water.  The reverse  is true for the three elements  listed. Their
concentrations are higher in upgradient ground water than the river and statistical analysis shows that their levels
in downgradient monitoring wells  are significantly lower than upgradient.

Inorganic Water Quality at a Superfund Site
Consultants for a  PRP  at a Superfund  site involving a limestone aquifer in the midwest US developed several
theories  regarding the nature  of  the  aquifer  involved  based on  their interpretation of  data  for  inorganic


                                                  196

-------
                         WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


constituents.  Data for  alkalinity, aluminum, calcium,  iron,  silica, sodium, and sulfate were  prominent in their
interpretation. However, these consultants did not consider relevant site-specific information  and admitted they
were unaware of the degree of variation that can occur in laboratory analysis of ground water  samples that meet
acceptance criteria (see Table 2).

A potentially  important piece of site-specific information not considered by the PRP's consultants was that most
of the borings involved had been drilled considerably deeper than the wells later installed in them. For the wells
being installed in  such deeper borings, boreholes were first partially filled by pouring in sodium bentonite chips.
This occurred for  about two-thirds of the relatively deep wells drilled. On the average, approximately one-third of
the borehole  was filled (i.e., 71 of 221  feet). This process undoubtedly resulted in the introduction of chemicals
from the hydrating chips into the water (both as dissolved and suspended solids) as they fell through the water
column. Of the analytes relevant to this site, sodium bentonite chips are typically composed of silica and oxides
of aluminum, iron, calcium, sodium, magnesium, and potassium  (in order of concentration). They also contain a
small level of water soluble nitrate. Given the chemistry of silica and calcium and the likelihood that calcium in a
limestone aquifer would be expected  to already be near  saturation,  concentrations of these  variables would
probably not  be greatly affected by this. However, concentrations of aluminum, iron, sodium, magnesium,  and
potassium could be and this appears to have been the case. The  potential for this was increased by the fact that,
although consultants for the PRP purged three well volumes immediately prior to sampling,  they did  not develop
the monitoring wells after installation.

To  evaluate  the  effect of filling boreholes with bentonite on inorganic ground  water  quality, monitoring wells
sampled both during 1996 (shortly after installation) and 1997 (nearly a year since last sampled) were divided
into two groups: (1) bentonite filled (BF); and (2) unfilled (UF). Median  data for major cations, major anions,  and
several other variables grouped into these two categories for both 1996 and 1997 sampling events are  presented
in Table 5. The 1996 data clearly indicate impact  from bentonite filling  for most of the variables listed except
calcium and silica, BF:UF ratios for sodium, aluminum, and iron indicate nearly an order of magnitude  or greater
level of enrichment for those variables as a result of bentonite filling. This is also  evident in the STIFF diagrams
of median grouped data in Figure 1. The STIFF diagram for unfilled wells (Figure 1a) is typical of what would be
expected for  a limestone aquifer.9 There are several relatively minor differences between the  STIFF diagram for
bentonite filled wells (Figure 1b) and  unfilled wells (Figure la), but by  far the most notable difference is the
sodium "bulge" to the  lower left of the diagram. The  impact of the bentonite appears to have been  transitory.
With the possible exception of nitrate, the enrichment appears to  have  been flushed away due to ambient ground
water flow by the time wells were resampled  nearly a  year later.  STIFF diagrams for both sets of wells (UF and
BF) when resampled in 1997 were similar to each other and the one for unfilled wells in 1996 (Figure  1a).

The PRP's consultants pointed to two aspects of the data to support their interpretation that the aquifer involved
was an "open" system (i.e., rapidly recharged from the surface throughout  the  aquifer):  (1) abnormally high
concentrations of aluminum and  iron in samples from some relatively deep wells (exceeding solubility limits); and
(2)  lack of any apparent ground water evolution (change in quality along a flow line) between wells at higher
elevations and those at lower elevations (a distance  of roughly one  mile). As discussed above and shown in
Table 5, aluminum and iron enrichment appears to have been related  to filling boreboles with sodium bentonite
chips.  Not knowing  about this circumstance  and  since the overlying clay soils involved would  be normally
expected to be rich  in silica as well  as aluminum  and iron, the  consultants interpreting inorganic water quality
data for the PRP  felt compelled to provide another explanation why silica concentrations were not also enriched
when aluminum and iron were. Their explanation was an assertion, without any data, that the  soils were lateritic.
Lateritic soils develop in hot, wet tropical climates subject to heavy rainfall when the intense chemical weathering
that occurs under those conditions removes both soluble materials and much of the silica. Lateritic  soils are not
characteristic of the temperate climate midwest US.10 A much more likely explanation for the relative magnitudes
of aluminum, iron, and silica has to do with the manner in which samples were taken, preserved, and prepared
for  analysis.  Samples to be analyzed for aluminum  and iron  are acidified in the field  and digested in  the
laboratory. Since the samples involved were unfiltered, this meant that  some  particulate aluminum and iron
would be included in the measurement. In contrast, samples to be analyzed for silica are not acidified in the field,
but  are filtered prior to  analysis.

With respect  to ground water "evolution," this phenomena has generally been documented on a  regional rather
than a site scale.  This  may be in part because very substantial "evolution" must occur to be detectable against a
background of variation sampling and  analysis introduced  variation as well as environmental variation. For an
extreme  example, using the acceptance criteria of 75 to 125 percent recovery for matrix spike samples  (see


                                                   197

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Table 2), a sample having a true value of 100 mg/L of calcium could be reported to have either 75 or 125 mg/L.
Although the higher of these two numbers is 67 percent greater than the low number, either one would be within
acceptance criteria. With enough sampling  events and statistical analysis of data, it might be possible to detect
an "evolutionary" change within acceptance criteria boundaries (e.g., a change from 50 to 65 mg/L along the flow
path); however, the possibility of seeing this change with a single set of samples appears slim when so much
variation is acceptable.

SUMMARY AND CONCLUSIONS
Comprehensive data interpretation by a knowledgeable professional should be the final quality assurance step of
any project involving ground water quality data. It may indeed help to find errors in field or laboratory work that
went otherwise unnoticed and provides the best chance for real understanding of the meaning of reported results.
Proper project planning should prepare for this final step by obtaining relevant information early on and including
relevant data  collection into field segments of the project. The following steps must be integrated into and carried
out throughout the project to facilitate final interpretation:

1  Collection, analysis, and evaluation of background data on regional and site-specific geology, hydrology, and
   potential anthropogenic factors that  could  influence ground water quality  and collection  of background
   information on the environmental chemistry of the analytes of concern.
2. Planning and carrying out of field activities using accepted standard procedures capable of producing data of
   known quality.
3. Selection  of  a  laboratory to analyze ground water  samples based  on  careful evaluation of  laboratory
   qualifications.
4. The use of appropriate QC/QA checks (including field blank, duplicate, and performance evaluation samples).
5. Comprehensive interpretation of reported analytical data by a knowledgeable professional.

The analytical data must be accompanied by appropriate QC/QA data, be  cross-checked  using standard water
quality checks and relationships where possible, and be correlated with information on regional and site-specific
geology and hydrology, environmental chemistry, and potential anthropogenic influences.

References Cited
1. USEPA  holding  times cited  in this paragraph are from Table II of Part  136, Title 40, U.S. Code of Federal
   Regulations (40 CFR 136).
2. Office of Emergency and Remedial Response. 1994. USEPA contract laboratory program national functional
   guidelines for inorganic data review. Publication No. EPA 540/R-94/013,  USEPA, Washington, DC, pp. 21-24
   and 27-29.
3. USEPA Superfund contract laboratory program (CLP) statement of work  (SOW) Form III VOA-1  and Form III
   SV-1.
4. Burns &  McDonnell  Waste  Consultants,  Inc.  1996.  Quality control  evaluation  report 96-072-4  dated
   September. Kansas City, MO, p. 10-1.
5. Schleyer, R.  1991.  Influence of the deposition of anthropogenic organic  substances from the atmosphere on
   ground-water quality.  IN:  Ground Water: Proceedings of the International Symposium, G.P, Lennon, ed.,
   American Society of Civil Engineers, pp. 293-298.
6. Harley,  R.A., et al. 1992. Respeciation  of  organic gas emissions and the detection of excess unburned
   gasoline in the atmosphere. ES&T, Vol. 26, No. 12, pp. 2395-2408.
7. Pankow, J.F., et al. 1997.  The urban atmosphere as a non-point source  for the transport of MTBE and other
   volatile organic compounds (VOCs) to shallow groundwater. ES&T, Vol. 31, No. 10, pp. 2821-2828.
8. Office of Emergency and Remedial Response. 1994. USEPA contract laboratory program national functional
   guidelines for organic data review. Publication No. EPA 540/R-94/012, USEPA, Washington, DC, pp. 62-65.
9. Hounslow,  A.W. 1995. Water Quality  Data: Analysis and Interpretation,  CRC Lewis Publishers,  Boca  Raton,
   FL, pp. 86-87.
10. Tarbuck, E.J. and  F.K. Lutgens.  1992.  The Earth: An Introduction to Physical Geology, 4th Ed.,  Macmillan
    Publishing Co., New York,  pp. 130-131.
                                                  198

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 1. Effect of Monitoring Well Development1
$jJiJJ!$il?:'-'': .;:.:. ,. •• , ,-.'.L>:^. '^d^;^
Calcium
Chloride
Iron
Magnesium
Manganese
Potassium
Sodium
Appearance
>i:-'ii!5v^ '^^p^HpUP^'''^" "'A ;;::^''v'i
92.
43.
0.08
2.4
0.069
ND
5.8
Clear
^m^w^mmm^1 : 	 -j^;
104.
58.3
5.53
2.84
0.13
533
6.38
Turbid
1. Concentrations in mg/L. ND means non-detect at 1 mg/L.
2. New well developed and purged prior to sampling.
3. Well unused for two years. Three well volumes purged prior to sampling.

                             Table 2. USEPA Superfund Acceptance Criteria
'^IWl^^^lilt-'fBett^^1.'" '" '•-.; ;7:'ixS
Elements (60 10)
VOCs (8240):
Benzene
Trichloroethylene
SVOCs (8270):
Pyrene
Phenol
:.S';x"-v- i^J^^ptw^ f?" ?;, vi-:
75-125

76-127
71-120

26-127
12-110
:. " wi*! •''•'r&WM*' "t"j; ^'~Kr}*i&ys&-~FfKf,; ;.." .;: J-. ',:,.«'• rf'i
>: Vs'l.atriK, ?'-' .r,ta>*&i:S.i -'.. v' '-TV — •..: .' :•;.•.=•
80-120

-
-

-
-
1.  Identification of variable and USEPA analytical method with  example volatile and semivolatile organic
  compounds (VOCs and SVOCs, respectively).
2.  Matrix spike (MS) and laboratory control sample (LCS)  acceptance criteria in percent recovery.  Inorganic
  criteria from USEPA national functional guidelines.  Organic criteria from CLP SOW Forms III VOC-1 and
  SV-1.
Table 3a. Reported Organic Compounds
- •" • :*
°0#ilfi^yf$£ i- ti'j-^ -» .'.' '-•
VOCS:
Halocarbon
Petroleum Hydrocarbon
Trihalomethane
SVOCS:
Petroleum Hydrocarbon
Phenolic
Phthalate
Highest Cortcentrafil^H * ^ , r
V 1998 s »;: e--,

236.
29.6
26.3

44.
175.
72.3
/lVr '-; aff>97 ' sr i«;.

172.
ND
5.08

ND
ND
43.2
1. Highest reported concentration of any compound in category out of 19 wells sampled in both years.
Table 3b. Reported Organic Compounds
:;f«apfiVft;|^:iA, - „ ^
'jft^RP5*1^'^''1 " "
VOCS:
Halocarbon
Petroleum Hydrocarbon
Trihalomethane
SVOCS:
Petroleum Hydrocarbon
Phenolic
Phthalate
NuBi^fDfWfeflsFfeporteain1 :^ ys Aj3r
•*-'-4,r "\ ( 1-99(1 L, ~~ t ^ "':'' :?

6.
2.
12.

3.
7.
12.
1 ' ' • '~ f' • - - '"• -^'SSir9rM;£q. " -" ' ,"i - i -^"' i" -~~' ^fe
.f • ' L? - • . v •"' '' .""-^ F* "f^f*-f'ry3 •• - J' i •!•.•:- -|j~.-,!+^:^-.rir|jJi

3.
0.
1.

0.
0.
6.
1. Number of wells out of 19 sampled in both years in which any compound in the indicated category was
reported at any concentration.
                                                 199

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                               Table 4. Ground/Surface Water Relationship

Major Ions:
Chloride
Sodium
Sulfate
Elements:
Arsenic
Barium
Iron


6.96
6.76
323

6.85
725.
13.2
Sa^^^SilJaHiiKk'.-s3A~*-^ifai!i

300.
340.
150.

2.
130.
0.01
1. Site data. Mean for upgradient monitoring wells (1994-1997).
2. USGSdata. Median for river at nearby gaging station (1981-1995),
                                   Table 5. Grouped Inorganic Data
 1996 Data:
   BF
 62.2
4.75
1.04
35.55
   UF
 59.4
2.97
  .5
 0.5
    BF:UF Ratio
 1.04
 1.60
2.06
 8.46
 1997 Data:
   BF
58.66
4.47
 .55
 3.06
    UF
58.48
4.02
  .5
 2.76
    BF:UF Ratio
 1.00
 1.11
                     1.11
 1996 Data:
    BF
 204.
7.31
 12.8
 33.5
    UF
 181.
 5.82
6.69
 16.6
    BF:UF Ratio
 1.13
 1.26
 1.91
 2.02
 1997 Data:
    BF
 177.
 3.93
7.62
 10.2
    UF
 174.
4.65
4.52
 9.5
    BF:UF Ratio
 1.02
0.845
 1.69
 1.07

 1996 Data:
    BF
 1.21
 0.94
 8.72
 305.
    UF
 0.10
 0.05
 9.68
 196.
    BF:UF Ratio
 12.1
 18.8
0.901
 1.56
 1997 Data:
    BF
 0.31
 0.10
 9.2
 231.
    UF
 0.29
 0.06
 3.8
 226.
   BF:UF Ratio
 1.06
 1.67
 2.42
 1.02
1.  All concentrations are median values for grouped data in units of mg/L.
                                                  200

-------
                          WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
             Milliequivalents per liter
    54321012345
  Ca

  Mg

N»+K

  ft
HCOj+COj


S04


Cl

NO3
Figure 1. Grouped Median Data STIFF Diagrams -1996


a. Wells in unfilled boreholes (UF).
              Milliequivalents per liter
    54321012345
  Ca

  Mg

 N»+K

   Fe
HCOj+COj


SO4


Cl


NO,
                                                       b. Wells in bentonite filled boreholes (BF).
                                                      201

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                             202

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
  GENERAL
         203

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                            204

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


 SAMPLE INTRODUCTION TECHNIQUES FOR FAST GC ANALYSIS OF ORGANOCHLORINE PESTICIDES

                        C. Eric Boswell. Michael S. Clark, and Rattana M. Woodard
               National Air and Radiation Environmental Laboratory, 540 South Morris Avenue,
                                   Montgomery, Alabama 36115-2601

ABSTRACT
Advances in  GC instrumentation  in  recent years have  allowed the analysis of organochlorine pesticides to
become more efficient. Pressure programable injectors, for example, have allowed analysts to reduce run times
as well as control injector phenomena such  as vapor cloud expansion and thermal degradation. GC analysts in
environmental laboratories have many options available to them in configuring instrumentation, especially in the
techniques they choose for sample  introduction.  When  attempting to maximize  the  benefits  of  modem GC
instrumentation, analysts must  be prepared to adapt their working knowledge  of classical GC techniques to
newer ones. At the National Air and Radiation Environmental Laboratory (NAREL) we learned that several of our
ideas about classical splitless injections had to be modified when we optimized SW846 Method 8081 for speed,
Endrin/DDT degradation, and chromatographic resolution.

In an attempt  to reduce the  run time  for SW846 Method 8081, we configured a  GC  with  dual pressure
programmable  injectors, two  dissimilar 0.53 mm  capillary columns, and  dual ECDs.  While the biggest time
savings resulted from  a relatively high  initial oven temperature, having  the ability to pressure program the
injectors also helped to reduce run times.  Programming the injectors  at  a high initial pressure also reduced
Endrin/DDT degradation by insuring a short injector residence time. A high initial oven temperature and high
initial injector pressure presented  a challenge since high pressure applications  typically depend on low initial
oven temperatures to achieve solvent focusing. In this analysis solvent focusing  is not  practical because the
boiling point difference between the solvent, hexane, and the earliest elating compound is too large. We found
that  although we wanted to  use techniques that seemed to preclude each other we could empirically match our
starting  oven temperature with a relatively high  initial injector pressure and have a degree of cold trapping to
achieve both short GC run times, low  Endrin/DDT degradation, and  good  chromatographic resolution. This
deviation from classical splitless injection used both the principles of solvent focusing and high pressure injection
to achieve the desired goal. The result was a GC method for twenty-two of the organochlorine pesticides on the
Method  8081  list with a run time of less than 20 minutes, very low Endrin/DDT degradation, good resolution on
difficult analyte pairs, and all the benefits of splitless injections.

INTRODUCTION
When attempting to maximize the benefits of modern GC instrumentation,  analysts must balance their desire to
improve the  efficiency of an analysis with  certain pragmatic concerns:  (1)  ease of maintenance, (2) use of
standard consumables, and  (3) ruggedness of new procedures. At the NAREL we wanted a GC method  for
twenty-two of the organochlorine pesticides on the Method 8081 list with a run time of less than 20 minutes, very
low Endrin/DDT degradation, good resolution on difficult analyte  pairs, and  all the benefits of splitless injections.
We also wanted the resulting chromatographic system to  be one that was easy to use by any reasonably trained
analyst. Most of the modifications  employed  in  our design  were in the  sample introduction  part of the
chromatographic analysis.

We configured a GC with dual pressure programmable injectors, two dissimilar 0.53 mm capillary columns, and
dual ECDs. Although most of the  work was in choosing  the appropriate sample introduction techniques,  there
were other parameters to optimize. For example, we chose a relatively high initial oven temperature to reduce
the cycle time of the analysis We  also adjusted the detector makeup gas flow to achieve acceptable sensitivity.
We wanted to take advantage of injector pressure programming which has become standard equipment on most
GCs. We chose splitless injections over direct injections because we believe the splitless technique to be more
rugged.  In our dual injector system, we kept the injector pressures similar during injection and programmed them
differently after the split valve opened  to achieve chromatographic resolution.  We chose a single taper inlet
sleeves  because they are easily obtainable and easy  to maintain.  We chose a  compromise initial  oven
temperature that was hot enough to keep the GC cycle time short and low enough to prevent the earliest eluting
compounds from tailing too much. This is different from  classical splitless injection technique where the initial
oven temperature is kept low until  most of the sample can slowly migrate from the inlet sleeve to the head of the
column. This deviation from classical splitless injection used both the principles of solvent focusing and high
pressure injection to achieve the desired goal. The result was a GC method for twenty-two of the organochlorine
pesticides on the Method 8081  list with a run time of less than 20 minutes, very low Endrin/DDT  degradation,


                                                 205

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
good resolution on difficult analyte pairs, and all the benefits of splitless injections.

EXPERIMENTAL
A mix containing twenty pesticides, two surrogate standards, and three internal standards (Ultra Scientific) was
prepared from the  dilution of certified standard  mixes with pesticide grade hexane (Burdick & Jackson). Our
experimental results are  best discussed in terms of  injection technique, pressure programming,  and column
selection.

We decided to use splitless injections  rather than direct injections to both simplify routine maintenance and
increase system ruggedness. Using  a  4 mm ID  single taper  inlet liner eased  routine column  maintenance
because the insertion distance of the column into the liner was the only critical step. A 1.0 ul_ volume of hexane
extract will expand to approximately 300 uL of vapor  at 6.5 psig.  This is easily contained within the 900 uL of
available volume in a 4mm ID liner. We found that a pressure increase after injection provided two benefits: (1)
the vapor cloud moved  onto the analytical column in a  smaller plug and (2) the compounds were pushed through
the column quicker. We chose 12.4 psig as  the injector pressure which resulted in a column linear velocity of 102
cm/sec. We used  120°C as our initial oven temperature. These represent a compromise between classical
solvent focusing and high pressure injection. Moving the sample vapor cloud onto the column quickly has several
benefits. Along with keeping early  eluting peaks  from  tailing,  high  pressure injections dramatically reduce
Endrin/DDT degradation. Combined  Endrin/DDT degradation  rarely exceeds  5.0%  with  this configuration.
Optimization of the opening and closing times for the split vent are still crucial for quantitative accuracy.

We chose analytical columns based on chromatographic separation  and ruggedness. A 5% diphenyl 95%
dimethyl polysiloxane column such as Restek's Rtx-5, provided both the good chromatographic separation and a
high temperature limit.  As a  confirmation column we chose a hybrid column developed by Tammy Carey1 that
met our chromatographic needs. The GC  oven  temperature program  is given in Table 1 along with the other
chromatographic conditions. The temperature  program  was optimized to  separate  critical  pairs such  as
Endosulfan I/alpha  Chlordane and Dieldrin/DDE on the Rtx-5 column and Heptachlor epoxide/gamma-Chlordane
and Endosulfan II/DDT on the Hybrid column. Chromatograms of the twenty-two organochlorine pesticides are
shown in Figure 1 and calibration  curve summaries are given in Tables 2 and 3.

                                  Table 1. Chromatographic Conditions
GC Parameter
Carrier Gas
Injector
Pressure Program
Temperature Program
Detector
Value
Helium
Splitless, 1uL, Purge Delay 0.75 min.
Inlet Temperature 250°C
Initial Linear Velocity: 102 cm/sec. (£
Rtx-5
12.4 psi Hold 0.50 min.
to 5.8 psi @ 99.0 psi/min. Hold 4.00
to 14.5 psi @ 0.75 psi/min.
to 19.0 psi @ 1.5 psi/min.
Hybrid
12.4 psi Hold 0.50 min.
to 7.5 psi @ 99.0 psi/min. Hold 6.00
to 14.0 psi @ 0.75 psi/min.
to 20.0 psi @ 1 .5 psi/min.
g 120°C
min.
min.
120°C Hold 0.5 min.
to150°C@30°C/min.
to 200°C @ 4°C/min.
to 285°C @ 20°C/min. Hold 1.0 min.
N2 Makeup @ 50 mL/min.
Anode Purge @ 5mL/min.
ECD@310°C
SUMMARY
At the NAREL we wanted a GC method for twenty-two of the organochlorine pesticides on the Method 8081 list
with a run time of less than 20 minutes, very low Endrin/DDT degradation, good resolution on difficult analyte
                                                  206

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


pairs, and all the benefits of splitless injections. We also wanted the resulting chromatographic system to be one
that was easy to use by any trained analyst. Although most of the work was in choosing the appropriate sample
introduction techniques, there were other parameters to optimize. We deviated from classical splitless injection
technique slightly to accommodate the use of pressure programmable injectors. This deviation from classical
splitless injection used both the principles of solvent  focusing and high pressure injection to achieve the desired
goal. The result was a GC method for twenty-two of the organochlorine pesticides on the Method 8081 list with a
run time of less than 20 minutes, very low Endrin/DDT degradation, good resolution on difficult analyte pairs and
all the benefits of splitless injections.

REFERENCES
1. Carey, Tamela; Chlorinated Pesticide Analysis Using Hybrid Columns (unpublished), CH2M Hill, Montgomery,
   AL

                       Table 2. Calibration Curve Summary of 5% Diphenyl Column

INITIAL CALIBRATION REPORT
 INSTRUMENT:
 CALIBRATION DATE: 03/09/97
GC#5/FRONT (5%DP)
 METHOD:
 SEQUENCE:
 DATA FILES:
SW846/ 8081
E030997A
010F010-014F0101
COMPONENT
Pentachlorobenzene/IS#1
Tetrachloro-m-xylene
alpha-BHC
beta-BHC
Lindane
delta-BHC
Heptachlor
Aldrin
Heptachlor Epoxide
gamma-Chlordane
Endosulfan I
alpha-Chlordane
Dieldrin
DDE
o-p'-DDD/IS#2
Endrin
Endosulfan II
ODD
Endrin Aldehyde
Endosulfan Sulfate
DDT
Endrin Ketone
Methoxyclor
Pentabromobiphenyl/IS#3
Decachlorobiphenyl
5ppb
1.00
0.65
0.51
0.37
0.50
0.46
0.70
1.16
1.47
1.65
1.35
1.69
1.35
1.44
1.00
1.46
1.40
1.12
1.16
1.33
1.22
1.55
0.45
1.00
1.55
10ppb
1.00
0.66
0.54
0.37
0.53
0.47
0.67
1.28
1.46
1.65
1.37
1.71
1.29
1.35
1.00
1.40
1.48
1.05
1.17
1.22
1.12
1.51
0.46
1.00
1.53
25ppb
1.00
0.72
0.63
0.39
0.60
0.54
0.68
1.37
1.46
1.70
1.42
1.78
1.31
1.47
1.00
1.37
1.53
1.02
1.18
1.17
1.15
1.51
0.45
1.00
1.44
50ppb
1.00
0.74
0.75
0.40
0.69
0.65
0.67
1.40
1.43
1.70
1.41
1.77
1.32
1.59
1.00
1.37
1.50
1.07
1.10
1.17
1.19
1.50
0.45
1.00
1.35
100ppb
1.00
0.73
0.80
0.39
0.74
0.69
0.66
1.40
1.35
1.61
1.34
1.67
1.25
1.55
1.00
1.28
1.38
1.01
1.01
1.13
1.17
1.43
0.43
1.00
1.34
SD
0.00
0.04
0.11
0.01
0.09
0.09
0.01
0.09
0.04
0.03
0.03
0.04
0.03
0.08
0.00
0.06
0.06
0.04
0.07
0.07
0.04
0.04
0.01
0.00
0.09
MEAN
RRF
1.00
0.70
0.65
0.38
0.61
0.56
0.68
r 1.32
1.43
1.66
1.38
1.73
1.31
1.48
1.00
1.37
1.46
1.05
1.12
1.20
1.17
1.50
0.45
1.00
1.44
%RSD
NA
5.13
17.61
3.40
14.88
16.77
1.99
7.05
2.94
2.09
2.29
2.60
2.65
5.62
NA
4.26
4.01
3.60
5.79
5.93
3.05
2.56
2.36
NA
6.16
P/F
NA













NA








NA

COMMENTS:
                                                207

-------
                     WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


                       Table 3. Calibration Curve Summary of Hybrid Column

INITIAL CALIBRATION REPORT
INSTRUMENT:
GC#5/REAR (HYBD)
CALIBRATION DATE: 03/09/97
METHOD:
SEQUENCE:
DATA FILES:
SW846/ 8081
E030997A
060R010-064R0101
COMPONENT
Pentachlorobenzene/IS#1
Tetrachloro-m-xylene
alpha-BHC
Lindane
beta-BHC
Heptachlor
delta-BHC
Aldrin
Heptachlor Epoxide
gamma-Chlordane
alpha-Chlordane
DDE
Endosulfan I
o-p'-DDD/IS#2
Dieldrin
Endrin
ODD
Endosulfan II
DDT
Pentabromobiphenyl/IS#3
Endrin Aldehyde
Methyoxyclor
Endosulfan Sulfate
Endrin Ketone
DCB
5ppb
1.00
0.76
0.65
0.63
0.46
0.95
0.54
0.64
1.10
1.29
1.30
1.29
1.35
1.00
1.22
1.15
0.91
1.06
1.11
1.00
1.09
0.61
1.40
1.80
1.73
10ppb
1.00
0.79
0.65
0.64
0.39
0.94
0.55
0.63
1.10
1.44
1.48
1.28
1.25
1.00
1.15
1.12
0.85
1.03
1.03
1.00
1.04
0.58
1.36
1.73
1.80
25ppb
1.00
0.89
0.77
0.75
0.42
0.93
0.63
0.68
1.12
1.47
1.46
1.29
1.25
r 1.00
1.19
1.16
0.87
1.08
1.04
1.00
1.01
0.55
1.41
1.70
1.82
SOppb
1.00
0.80
0.82
0.79
0.42
0.90
0.70
0.70
1.07
1.43
1.39
1.28
1.20
1.00
1.22
1.12
0.92
1.07
1.04
1.00
0.93
0.51
1.23
1.63
1.69
100ppb
1.00
0.81
0.94
0.89
0.47
0.87
0.81
0.78
1.12
1.50
1.40
1.38
1.25
1.00
1.23
1.15
0.92
1.05
1.06
1.00
0.89
0.48
1.22
1.55
1.54
SD
0.00
0.04
0.11
0.10
0.03
0.03
0.10
0.05
0.02
0.07
0.06
0.04
0.05
0.00
0.03
0.02
0.03
0.02
0.03
0.00
0.07
0.05
0.08
0.06
0.10
MEAN
RRF
1.00
0.81
0.77
0.74
0.43
0.92
0.65
0.69
1.10
1.42
1.40
1.30
1.26
1.00
1.20
1.14
0.90
1.06
1.06
1.00
0.99
0.55
1.32
1.63
1.71
%RSD
NA
5.47
14.20
13.15
6.62
3.35
15.67
7.89
1.88
5.17
4.49
2.86
3.99
NA
2.46
1.48
3.06
1.59
2.55
NA
7.18
8.51
6.25
3.77
5.84
P/F
NA












NA





NA





COMMENTS:
                                            208

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
    .O«BS-
  4.Oe4-
              — C  OCF-estStci
                             1O
                                                         Figure 1. Chromatograms of twenty-two
                                                         Organochlorine Pesticides on a dueal injector
                                                         GC system.
              — C OCI=estSt
-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


analytical chemistry is gaining importance because of their persistence and toxicity. Furthermore, many countries
have now passed legislation to ensure that pesticides are used safely and responsibly in order to protect the
public from unsafe levels of pesticide residues.  Due to the  high polarity, low volatility  and thermolability  of
carbamates, GC  and GC/MS methods are not amenable to this class of compounds. HPLC with UV detection
has been the standard  analytical method for  analysis and  quantitation  of polar compounds.  Benomyl,  a
carbamate  and a widely used systemic fungicide for  disease control  in crops, poses additional analytical
challenges to the environmental chemist due to its extreme instability in the environment. Benomyl decomposes
to three stable  major  degradation  products:   2-aminobenzimidazole  (2-AB),  methyl  1-H-benzimidazol-2-
ylcarbamate (MBC), and 3-butyl-2,4-dioxo-s-triazino [1,2-a] benzimidazole (STB).

EPA Method 631  for the determination of benomyl and MBC in industrial  and municipal wastewater restricts the
detection and quantitation to MBC only since benomyl is hydrolyzed to MBC in this method. This isocratic UV
method has a Method Detection Limit (MDL) of 8.7 ppb for MBC.

The UV method  adapted in this report monitors the three major degradation products of benomyl by gradient
elution. The UV analysis has a method detection limit of 2 ppb for 2-AB, MBC, and STB.

METHOD
Materials
2-AB was obtained from Aldrich Chemical Company (Milwaukee, Wl). MBC was obtained from Chem Services
(West Chester, PA) and STB was obtained from E.I.  du  Pont  Nemours & Co. (Wilmington, DE).  Solvents used
for HPLC and extraction were HPLC grade acetonitrile, methanol and water from JT Baker (Philipsburg, NJ). The
formic acid used in the HPLC solvents,  as well as the injection solvent,  was from Sigma (St. Louis, MO). The
standard bore (4.6 x 250 mm) HPLC column was a C18 with a 5-um particle size by Whatman (Clifton, NJ). The
narrow bore column (2 x 50 mm) was a Monitor C18 with a 3-um particle size by Column  Engineering (Ontario,
CA). A Waters Oasis (Milford, MA) 3-mL, 60-mg SPE cartridge was used for the solid  phase extraction.

Preparation. Benlate Extraction Procedure
A Benlate mix spike solution was  comprised of a  mixture of 2-AB and STB in methanol at a concentration of 2
ppm. 200 mL of  water sample was poured into a 400-mL beaker.  HPLC grade water was used for blanks and
laboratory fortified blanks. The pH of all  samples was checked and recorded using a pH meter. For any spiked
samples  (laboratory fortified blanks and matrix spikes),  200 uL of Benlate mix spike solution was added and
mixed well. Two  Normal NaOH was added to adjust the  pH of all the samples to 10. The solid phase extraction
(SPE) cartridge was conditioned by aspirating 5  mL of methanol through the cartridge followed by 10 mL of
HPLC grade water without allowing the cartridge  to dry. The sample  was filtered  through the conditioned SPE
cartridge under vacuum. The sample  beaker was then rinsed with 5 mL of HPLC grade water adjusted to pH 10.
The rinse was passed through the  SPE cartridge and vacuum dried for 3 min. Elution with 5 mL of methanol and
collection of the eluate into a 15 mL centrifuge test tube was completed. The eluate was dried down with nitrogen
in a water bath at 40°C to a volume of 1 mL. One mL of 0.2% formic acid was added and mixed well.

HPLC Conditions
A Shimadzu VP series  HPLC (Columbia, MD) was used with the gradients, flow rates  and columns listed in Table
1.0 and Table 1.1. The flow from the  standard bore column was split down from 1  mL/min to approximately 200
ul_/min into  the  Turbolonspray source  at ambient temperature.  For the  narrow  bore  column no split was
necessary and the entire effluent was sent into a Turbolonspray source set to 200°C.

Table 1.0. HPLC Gradient
Standard Bore Column
4.6 x 250 mm
Flow rate: 1 mL/min
Whatman C18 5 urn




Time (min)
0
10
20
20.1
30
Solvent A
0.1% Formic Acid
70
30
15
70
70
Solvent B
Methanol + 0.1% Formic Acid
30
70
85
30
30
                                                 210

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                              Table 1.1. HPLC Gradient
Narrow Bore Column
2 x 50 mm
Flow rate: 0.2 mL/min
Column Engineering
Monitor C1 8 3 pm




Time(min)
0
0.85
2.25
4
6
7
7.1
10
0.1% Formic Acid
100
50
45
45
30
30
100
100
Acetonitrile + 0.1%
0
50
55
55
70
70
0
0
Formic Acid








Mass Spectrometer
A PE Sciex API 150 (Concord, Ontario, Canada) single quadrupole mass spectrometer was operated in positive
ion mode with a Turbolonspray source. The source temperature was ambient for the standard bore HPLC column
and was set to 200°C for the narrow bore column. The compounds were monitored using selected ion monitoring
(SIM) with the following m/z values:
                           m/z-134.1
The ion optics path was optimized for each compound separately in the separations solvent. The values for the
orifice, ring and QO are listed in  Table  2.0. The mass spectrometer was operated at unit resolution with  an
ionspray voltage of 4600 V.

                                      Table 2.0. Ion Path Voltages
Compound
2-AB
MBC
ST13
OR
45
20
30
RNG
160
120
155
QO
-4
-3
-4
RESULTS
Initially the primary emphasis was on transferring an already validated UV HPLC method to a single quadrupole
MS method altering as few parameters as possible. Only two changes to the original procedure were necessary
for the method transfer. Formic acid was added to the mobile  phase  of the HPLC solvents, and the injection
solvent was changed to contain a final concentration of 0.1% formic acid in 50% methanol. All other parameters
remained the same. As a  consequence of the change in pH, the elution order of the compounds changed but
chromatographic resolution was maintained and all compounds were baseline resolved.  The eluant from the
4.6-mm column was initially split to 50 pL/min into  the Turbolonspray source at room temperature. Occasionally
at a high percentage of organic, the 50-uL/min split into the mass spectrometer would lose flow resulting in a loss
of signal. This was corrected by decreasing the split ratio sending approximately 200 uL/min into the ion source
at the expense of sensitivity. Nevertheless, the limit of quantitation for the standards (0.01 ppm) was an order of
magnitude better than that  of the UV analysis of standards (0.1 ppm). (See Figure 1.0 and Figure  1.1.)

The Method Detection Limit (MDL) was calculated with the following equation:

          MDLcalculated = t x s x Vext/Vsampie
          Vext = 2 mL            t = Student Coefficient
          Vsampie = 200mL        s = Standard Deviation

Table 3 0  contains the solid phase extraction validation data and Table 4.0  contains the  MDL calculations for
both UV and LC/MS analyses. While the calculated MDL for the UV and LC/MS methods are similar, the actual
                                                  211

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
MDL achieved by the two methods is an order of magnitude different.

                                  Table 3.0. SPE Validation for Water
Recoveries
UV 4.6x250 mm
Compound Spjke Level
2-AB 5 ug/L
MBC 5 ug/L
STB 5 ug/L
MDL-1
ug/L
25
5
54
%
50
100
108
MDL-2
ug/L
2.5
53
5.5
%
50
106
110
MDL-3
ug/L
2.7
54
5.8
%
54
108
112
MDL-4
ug/L
2.7
4.6
4.7
%
54
92
94
MDL-5
ug/L
32
56
5.7
%
64
112
114
MDL-6
gg/L
3.3
5.3
5.5
%
66
106
110
MDL-7
yg/L
3.1
5.2
53
%
62
104
106
%Mean
57.1
104
108
SD
6.7
6.4
6.6
%RSD
12
6
6
LC/MS 4.6. 250 mm
2-AB 2 ug/L
MBC 2 ug/L
STB 2 ug/L
1 74
1.98
1 57
87
99
79
1.86
2.12
1 64
93
106
82
1 76
2.04
1.54
88
102
77
1.9
2.2
1.58
95
110
79
1.86
2.18
1.54
93
109
77
1.94
2.26
1.62
97
113
81
1.64
218
1 66
82
109
83
90.7
107
79.7
3.9
5.2
2
4
5
3
                                   Table 4.0. MDL
Method Detection Limit
UV 4.6 x 250mm
Compound
2-AB
MBC
STB
Student
Coefficient
3.143
3.143
3.143
n
7
7
7
Std. Deviation
6.7
6.4
6.6
MDL
(calculated)
0.21 ppb
0.20 ppb
0.21 ppb
MDL (actual)
2 ppb
2 ppb
2 ppb
LC/MS 4.6. 250 mm
2-AB
MBC
STB
3.143
3.143
3.143
7
7
7
3.9
5.2
2
0. 12 ppb
0.16 ppb
0.063 ppb
0.2 ppb
0.2 ppb
0.2 ppb
In order to improve the setup and ease of use the LC/MS method was altered further. A 2 X 50-mm HPLC
column replaced the larger diameter  column and the entire eluant flow was sent directly into the Turbolonspray
source. The source was operated at 200°C. This source temperature gave the best overall response for the three
compounds of interest while not significantly adding to the background signal of the final method. The initial
gradient was based on the same column volumes and used the same buffer system of the larger column.  The
result was a fine separation of 2-AB and MBC but STB did not elute as a sharp peak. With the addition of heat
the methanol  solvent system generated an additional  problem. The background signal increased dramatically
making detection of the low standards difficult. All efforts to reduce the background with methanol as the organic
solvent failed. Methanol was replaced with acetonitrile, resulting in much lower background counts and  a much
improved STB peak shape. The gradient was changed to accommodate the  stronger solvent. In fact, the peak
shapes improved for all three analytes. 2-AB and  MBC each elute in roughly 20 uL and are baseline-resolved.
STB elutes in approximately 100 uL. Because of this difference  in  elution  volumes  a period MS  method
generated better S/N for each compound. This reduced elution volume from  the narrow bore column improved
the limit of quantitation (LOQ) even  further. While the reduction in column  dimension should only result  in a
four-fold gain in sensitivity, the above changes resulted in a 50-fold gain since only 10 uL were injected onto the
2-mm column versus  100 uL injected onto the 4.6-mm column. The  small volume of  the 2-mm  column as
compared to the standard bore column also reduced the analysis time from 30 minutes to 10 minutes.

Figures 1.0, 1.1  and 1.2 are views of  the low standards for each method. Each has a S/N greater than 10:1. The
LOQ for the standards is 0.1  ppm, 0.01 ppm, and  0.002 ppm for the UV, 4.6, and 2-mm methods, respectively.
The resultant calibration curves are shown in Figure 2.0. A linear 1/X weighting was used for the LC/MS analyses
    	  while  no  weighting  was used  in  the UV
                                                            analysis.  Table   5.0  lists   the   percent
                                                            deviation  of the mean from the theoretical
                                                            (%DMT) and the  %RSD  obtained over five
                                                            days for the various levels using the 2-mm
                                                            LC/MS method. Table 5.0 demonstrates the
                                                            high  degree of accuracy and consistency
                                                            required for successful quantitative results.
', net:
                          £ - H
                                                             Figure 1.0. UV Low Standard (0.1  ppm;
                                                             100-uL injection volume).
                                                 212

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                               STB
        2-AB
V  Mt
1   \
           MBC
   #;,•'«#%:; U< k^'i^^vAy,
    /W\T
                                            Figure 1.1. 4.6 x 250 mrn Low Standard (0.01 ppm; 100-uL
                                            injection volume).
Figure 1.2. 2 x 50 mm Low Standard (0.002 ppm; 10-uL
injection volume).
Table 5.0. %DMT Statistics
                                                         2-AB
                                                          MBC
                                                                         STB
Five-Day Percent Deviation from Theoretical LC/MS 2 x 50 mm method

Level

1000ppb
200 ppb
1 00 ppb
20 ppb
10 ppb
2 ppb
1ppb
2-AB
%DMT
2-AB
96.19
111.15
110.49
106.65
104.38
94.05
78.97
% RSD
2-AB
1.12
0.94
2.35
1.95
4.04
2.51
13.26
MBC
%DMT
MBC
97.8
107.69
106.79
102.2
99.57
98.91
95.67
% RSD
MBC
0.44
1.87
1.98
2.81
2.38
4.67
9.29
STB
%DMT
STB
99.45
105.99
105.38
98.76
91.64
108.2
114.25
% RSD
STB
0.05
1.38
2.06
5.81
5.67
11.72
16.74
While the 0.001-ppm standard is detected with the 2-mm method, the %RSD for the  accuracy of the three
compounds was deemed to be too high for the LOQ. The MDL-7 for the 2 x 50-mm LC/MS method needs to be
completed, but based on the previous results for the standards the MDL should be approximately 0.02 ppb.

CONCLUSIONS
A method was developed to quantitate the major degradation products of benomyl in a simple, single quadrupole
LC/MS assay. An existing UV HPLC method was slightly modified for analysis by a PE Sciex. API  150 mass
spectrometer. An immediate order  of magnitude  gain in sensitivity  over the UV MDL resulted. This  LC/MS
method obviates  the additional step of hydrolysis described in EPA Method 631, allowing for direct quantitation
of the major degradation products of benomyl. Currently the benomyl LC/MS method can be used with extraction
protocols for water. Additional optimization of the HPLC and MS method produced further gains in sensitivity and
also simplified the system setup. The narrow bore method requires the validation process to be completed,  but
this should not prove difficult based on the clean matrix provided by the SPE protocol and the consistency of the
standards analyzed. The increased  specificity afforded by MS also allowed for a reduction in HPLC and data
analysis time. Future work will result in the development of extraction protocols for waste and sediment. Also,
implementation of an internal standard should improve the results obtained by this method.
                                                 213

-------
                       WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                         Figure 2.0. Calibration Curves for2-AB, MBC, and STB.
            METHODS OPTIMIZATION OF MICROWAVE ASSISTED SOLVENT EXTRACTION
                      TECHNIQUES FOR VARIOUS REGULATORY COMPOUNDS

                                      Greg LeBlanc, Mike Miller
                          CEM Corp., PO Box 200, Matthews, NC 28106-0200
                                           (704)821 7015

We have seen a period of rapid growth in the separation sciences, with gas and liquid chromatography as well as
the hyphenated techniques becoming commonplace. There is a need to streamline the extraction techniques
preceding these chromatographic analyses. Time and solvent usage  have become critical factors. In response to
this need several new sample preparation technologies have emerged. As a result analysts are faced with the
challenge of adapting current methods or developing entirely new ones depending on the extraction technology
they use.

Microwave Assisted Solvent Extraction is a new technique that works with existing solvent regimes with minimal
modifications and significant solvent reduction. It also has the benefit of reducing the extraction time. This study
will review the methods  optimization process when using a  closed vessel microwave assisted solvent extraction
                                                214

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


technique. It  will  focus  on extraction temperature,  sample water content  and ratios for  solvent mixtures.
Recovery data will be presented for pesticides and fungicides from reference soils and vegetable samples.
       NEW DEVELOPMENTS IN CLOSED VESSEL MICROWAVE DIGESTION TECHNOLOGY FOR
              PREPARATION OF DIFFICULT ORGANIC SAMPLES FOR AA/ICP ANALYSIS

                               Greg LeBlanc. Bobbie Haire and Bob Fidler
                        CEM Corporation, PO Box 200, Matthews, NC 28106-0200
                                            (704)821-7015

Microwave digestion techniques  are well accepted as a means of preparing samples prior to AA and ICP
analysis. Pressurization of the  sample vessel accelerates the dissolution  process by permitting  acids to attain
higher temperatures than under  ambient conditions.  However, it has been  historically difficult to  control the
microwave digestion process under ultra-high  temperature and  pressure conditions.  As well, throughput  of
difficult high-pressure samples using conventional microwave  instrument technology  has been limited, thus
inhibiting its use for preparation of some difficult organic and inorganic samples. This paper will demonstrate new
microwave  instrument technology  for  optimization  of  ultra-high  pressure and  temperature  conditions  for
microwave preparation of difficult organic and inorganic samples for AA and ICP analysis. Data will be provided
to demonstrate the digestion performance under significantly elevated pressure and temperature conditions  as
well as the impact of such conditions on the microwave preparation instrumentation.
                      EVALUATION OF ICP-OES AND ICP-MS FOR ANALYSIS OF
                         THE FULL TCLP INORGANIC TARGET ANALYTE LIST

                                    Marc Paustian and Zoe Grosser
              The Perkin-Elmer Corporation, 50 Danbury Road, MS-219, Wilton, CT 06897-0219

 Abstract
 Toxicity Characteristic testing is performed to assess the potential of a material to leach hazardous constituents
 after disposal in a landfill. The Toxicity Characteristic Leaching Procedure (TCLP) extract solution is analyzed for
 31  organic  components  and 8 metals and if the regulated limits  are  exceeded  (Table 1) the material is
 considered hazardous and must be treated appropriately.

 Currently, ICP-OES and ICP-MS have the technical capability of determining the full  suite of TCLP elements in
 one run. EPA  ICP-OES method 601 OB includes Hg in the target analyte list. EPA ICP-MS draft method 6020A
 now includes Se and Hg in the target analyte list. This provides an opportunity to demonstrate the utility of TCLP
 analysis using  one sample preparation and one technique for analysis.

 Compliance with the generally accepted requirements of the RCRA program will be demonstrated for the full
 suite of TCLP analytes including mercury. ICP-OES and ICP-MS will be compared for the analysis.

 The results will be validated with the analysis of the required Quality Control for performance-based methods and
 will include detection limits, spike recoveries, and duplicate sample analytical data.

 Experimental
 Raw TCLP extracts were provided  by Trinity River Authority in Dallas Texas. They were prepared by digestion
 with EPA method 3015 using the Multiwave Microwave Digestion System (Perkin-Elmer Corp.). To evaluate the
 recovery vs. hotplate digestion, for mercury,  the samples were also spiked with mercury and digested on a
 hotplate using method  3020. Gold  was added  to some samples to evaluate the effect on retention of mercury


                                                  215

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
during the digestion process and resulted in a concentration of 2 mg/L in the final solution.
The samples were analyzed using ICP-OES and ICP-MS, for comparison. The instrumental conditions are shown
in Tables 2 and 3.

                                      Table 1.TCLP Metal Limits
Element
As
Ba
Cd
Cr (Total)
Pb
Hg
Se
MCL(mg/L)
5.0
100.0
1.0
5.0
5.0
0.2
1.0
                               Table 2. ICP-OES Instrumental Conditions
                                          Optima 3000 DV
Parameter
RF Power
Nebulizer Flow
Auxiliary Flow
Plasma Flow
Sample Pump Flow
Plasma Viewing
Processing Mode
Auto integration
Read Delay
Rinse
Replicates
Background Correction
Settings
1 500 watts
0.4 L/min
0.5 L/min
15.0 L/min
1.5 ml_/min
Axial
Area
5 sec min-20 sec max
45 sec
45 sec
2
one or two points
                                Table 3. ICP-MS Instrumental Conditions
                                             ELAN 6000
Parameter
RF Power
Nebulizer Flow
Auxiliary Flow
Plasma Flow
Sample Pump Flow
Sample/Skimmer Cones
Scanning Mode
Lens
Read Delay
Rinse
Replicates
Detector Mode
Settings
1500 watts
0.94 L/min
1.0 L/min
15.0 L/min
1.5 mL/min
Nickel
Peak Hopping
Lens Scan Enabled
10 sec
35 sec
3
Dual Mode
Results and Discussion
The instruments were qualified by measuring  method detection limits using  the  procedures specified in  the
Federal Register,  part 1361  The  method detection limits are shown in Table 4 and compared with the MCL
(shown here in ug/L for easier comparison) for the element. It is generally accepted that the MDL should be ten
times  less than the MCL to  ensure sufficient confidence  at the decision-making point. The detection limits for
both the ICP-OES and ICP-MS are well below the limits set by the criteria specified.
                                                216

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Table 4. Method Detection Limits
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
MCL (ug/L)
5000
5000
100,000
1000
5000
200
5000
1000
MDL (ug/L)
ICP-OES
6
2
0.3
0.1
0.6
8
2
3
MDL (|jg/L)
ICP-MS
0.007
0.09
0.01
0.03
0.05
0.34
0.006
0.18
The results were measured on the samples digested with the microwave procedure for the full suite of elements.
The results are shown in Table 5 and compared with results reported by the supplier. The digested samples
compared well with the results provided. The ICP-OES and ICP-MS results compared well with each other, The
results were all well below the maximum contaminant levels for all elements.
Table 5. Comparison of Sample Results
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
ICP-OES
Sample 1
(Mg/L)
8
8
253
2
15
<8
7
5
ICP-MS
Sample 1
(ug/L)
0.1
5
214
0.4
12
<0.3
4
4
TR-1
(pg/L)
<2
<5
232
<8
11
<0.5
4
<5
ICP-OES
Sample 2
(ug/L)
6
6
260
1
6
<8
5
2
ICP-MS
Sample 2
(ug/L)
0.05
3.4
218
0.08
7
<0.3
2
2
TR-2
(ug/L)
<2
<2
278
<8
5
<0.5
<4
<5
Predigestion spike recoveries were evaluated for samples digested with the microwave procedure. The samples
were spiked with 2.5 mg/L of As, Cr, Pb, Ag, 0. 5 mg/L of Cd, Se, 0.1 mg/L Hg, and 50 mg/L of Ba. The samples
labeled "Spike 2" also had gold added (2 mg/L). The spike recoveries are shown in Table 6 for ICP-OES analysis
and Table 7 for ICP-MS analysis.

                                   Table 6. ICP-OES Spike Recoveries
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
Sample 1
Spike 1
%Rec
33
104
102
106
100
60
99
104
Sample 1
Spike 2
%Rec
2
107
102
108
101
62
102
107
Sample 2
Spike 1
%Rec
30
107
102
107
100
62
100
105
Sample 2
Spike 2
%Rec
2
108
101
108
101
63
100
106
Blank
Spike 1
%Rec
63
99
101
105
99
54
102
96
Blank
Spike 2
%Rec
2
100
101
106
99
60
103
96
The spike recoveries were excellent. The silver recoveries diminished with time, as is typical with a nitric-only
digestion. Since the ICP-MS analysis was performed first and the ICP-OES analysis was performed after several
days had passed the effect is more evident in the  ICP-OES results. The addition of small amounts of chloride
with the gold solution  in the second spike accelerated the precipitation of silver. The mercury spike recoveries
were  less than expected and  may be due  to  insufficient cooling  of the vessels before the solution was
transferred. The solutions were not stored in Teflon® bottles,  providing another possible source of loss. Since the
addition of gold  improved the recoveries  of  mercury only  slightly in the microwave digestion, better overall
performance may be obtained if the gold is left out. The effect of gold addition on the retention of mercury in a
traditional hotplate digestion was investigated. The  recoveries compared to spikes digested with microwave are
                                                  217

-------
shown in Table 8.
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                               Table 7. ICP-MS Spike Recoveries
Element
Ag
As
Ba
Cd
Cr
Hg
Pb
Se
Sample 1
Spike 1
92
104
115
101
96
62
109
103
Sample 1
Spike 2
6
100
120
103
96
53
107
108
Sample 2
Spike 1
97
102
116
103
96
92
109
98
Sample 2
Spike 2
4
104
118
103
99
53
106
107
Blank
Spike 1
106
93
117
104
100
55
109
103
Blank
Spike 2
7
102
116
103
101
54
106
113
                            Table 8. Effect of Gold on Mercury Spike Recovery

Sample 1
Spike
(% Rec)
Sample 1
Spike Dup.
(% Rec)
Sample 1
Spike+Gold
(% Rec)
Sample 1
Spike Dup. +Gold
(% Rec)
Hot Plate Digestion
ICP-OES
ICP-MS
52.1
44.4
32.8
37.0
56.9
51.2
61.5
47.5
Microwave Digestion
ICP-OES
ICP-MS
60
62
62
62
62
53
63
53
The addition of gold  only slightly  improves the recoveries of mercury  using the microwave digestion.  The
hotplate digestion results show more dramatic recovery improvements. In addition, the duplicates are much more
consistent in the hotplate digestion when gold is added before the digestion process.

Summary
Much progress has made in increasing the  scope of inorganic methods to include all the analytes of interest.
Analyses can be performed more economically when only one sample preparation and one  analysis technique
are needed to fully evaluate a sample.

This work has demonstrated that  mercury can be successfully determined  in a digested TCLP extract matrix, as
a part of the full suite of elements  traditionally measured. Both ICP-OES and ICP-MS show excellent method
detection  limits for the  elements  measured and good predigestion spike  recoveries  with the  exception of
mercury. Further work is  needed  to define the mechanism for mercury loss in the  digestion  and transfer
procedure used before the instrumental analysis.

Method 601 OB for ICP-OES and  proposed 6020A for ICP-MS contain the full  list of analytes  for TCLP analysis.
This work validates the utility of these methods for this type of matrix. Microwave sample  preparation procedures
are especially  useful when mercury is included in the analyte list, coupled  with  careful  sample transfer and
storage.

Acknowledgement
The authors would like to thank Bill Cyrus of the Trinity River Authority, Dallas, Texas for providing the TCLP
extracts.

References
1. US Code of Federal Regulations 40, Ch. 1, Pt 136, Appendix B.
                                                 218

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


                                     CLEAN METALS SAMPLING

                                          Roger E. Stewart II
       Virginia Department of Environmental Quality, Office of Water Quality Assessment and Planning,
                              629 East Main Street, Richmond, Virginia 23219

ABSTRACT
Scientific evidence1 of aquatic organism toxicity to low concentrations of dissolved metals has led to the need for
accurate and precise measurements of these compounds in a variety of aqueous matrices. A practical method2
will be presented for the collection and analysis of freshwater and wastewater samples for trace metals. A hands
on demonstration of clean techniques will be presented to show the simplicity and ease of collecting scientifically
defensible  samples for compliance monitoring purposes. Field sampling equipment specifications, collection
procedures, the importance  of  blanks, and preservation of samples, will be discussed as topics  important to
designing  a procedure for collecting  water samples with  metals concentrations in the  range where toxicity is
significant.

INTRODUCTION
To  be protective  of  aquatic organisms  and  human  health  the  US  Environmental  Protection  Agency has
established water quality standards3, which specify  safe concentration levels in background waters, for toxic
metal  species. Regulatory agencies must evaluate the  concentrations  of  analytes in receiving waters and
discharges to determine "reasonable potential" for impact to  an  ecosystem. Reliable methods for measuring
trace concentrations of target compounds are essential to developing permit limits on discharges.

As  regulatory levels  for metals  have decreased,  sometimes several orders  of magnitude,  field collection
protocols  which are practical and applicable  to a wide variety  of site conditions and  personnel  must be
developed.

The US Geological Survey (USGS) and the US Environmental Protection Agency (USEPA) have made progress
in the  last two years towards developing  widely applicable field and  laboratory  procedures. The USGS has
focused on internal procedures4 designed to  be used by USGS personnel. As a regulatory agency the USEPA
has developed procedures56'789101112  which  can be used by monitoring groups,  public agencies, and  the
regulated community.

The freshwaters appropriate for collection  by this method include all surface waters and groundwaters with a
specific conductivity of approximately 1000 umhos/cm or less. Appropriate wastewaters include treated effluents
(conductivity < 1000 umhos/cm).  Saltwaters, brackish waters,  highly turbid wastewaters, i.e.  landfill leachates,
are not appropriate as they require special laboratory preparation and analysis.

The protocols contained in this Standard Operating Procedure are  applicable to the compounds listed in Table 1
Target Analytes, Analytical Test Methods, and Detection  Limits on page 8.

Additionally this SOP is suitable for  freshwater and treated final  effluents with concentrations of toxic metals
below  approximately  200 ug/L.  The  200 ug/L threshold  should be applied cautiously  as this is only a
generalization of the effect of contamination. For example, because of well documented contamination problems
with copper and zinc, if a final effluent has historically had copper or zinc reported in the 0.2 mg/L range use of
this protocol may reveal that the actual concentrations are significantly lower. However if the historical numbers
for  cadmium, arsenic,  or mercury have been in the 0.2 mg/L range use of  this protocol  may not affect these
concentrations. Another factor to  consider when  determining the  applicability of this SOP is that typically the
reporting level of historic data is much higher than the lowest  concentrations when compared to Water Quality
Standards.

For concentrations above approximately 200 ug/L,  existing 40  CFR 136  procedures are adequate and contain
the  necessary Quality Controls  (including the requirement to collect blanks) to make reliable measurements in
the  mg/L range. The United  States EPA Region III has  prepared extensive guidance for existing and new data
which falls into this range.

METHODOLOGY
Improving on a combination of the more salient features from the USGS, USEPA, and  others413 the Virginia


                                                  219

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Department of Environmental Quality  (VADEQ) began a pilot  project  designed to 1) determine if  current
technology could be used to accurately measure metal concentrations  in water below  1.0 ug/L using clean
techniques, 2) develop protocols for wastewater treatment plant effluent collection and analysis which could be
incorporated  into guidance  and, 3)  develop freshwater  protocols for an ambient water quality monitoring
program.

The main design considerations were to develop a system to 1) collect field samples free of contamination and,
2) minimize the level of effort required by field technicians.

When measuring at trace concentrations small amounts of background contamination and interference  can be
significant. The most likely sources of contamination are:

1. improper sample handling  techniques,
2. improperly cleaned sampling equipment and sample containers, and
3. atmospheric dust and debris.

These ideas led to the design of a closed loop sample container which could be filled onsite, transported to the
laboratory,  prepared, and digested all in the same container. Figure 1 Loop Sample Bottle shows the container
used to collect the samples.

Figure 2 Sample Collection  Scheme  illustrates the closed loop sample collection  system. Sample collection is
accomplished by pumping water from the sampling zone through a teflon  tube, peristaltic tubing, and a capsule
filter, directly into the loop sample container.

The entire process consists of the following  steps:
1. Sample containers and tubing kits are cleaned and certified in a controlled laboratory environment.
2. The containers and kits are packaged into coolers  and shipped to the field.
3.  Blank ultra pure water from a blank sample container is processed through the tubing kit to condition  and
    clean the capsule filter.
4.  Ultra pure water from the sample container is pumped through the tubing assembly into the empty blank
    container to produce the  field blank.
5. Sample is collected into the empty sample container using the tubing assembly.
6. Blanks and samples are shipped to the laboratory.
7. All samples and blanks are acidified, digested, prepared in the original container.
8. Samples are analyzed accordingly.

The advantage of this  design is that field technicians  receive the sample containers and equipment ready to use.
Other than filter conditioning1415 there are no cleaning steps required and no need to preserve the samples other
than icing. The sample containers include the water to be used for field blank collection with the blank collection
exactly duplicating the site conditions of the sample collection.  All the plasticware is inexpensive and  can be
discarded after single use.

DISCUSSION
Ambient Sample Collection Protocol
1.  Locate an area where sample processing  will occur. This should be an area free of falling debris and swirling
    dust, flat, smooth,  and protected from the wind. The tailgate of a vehicle or the back of a Suburban are good
    locations.
2.  Locate the equipment box and  coolers containing the sample containers and kits in the area where  sample
    processing will occur.
3.  Cover the work area with  a large piece of plastic film. Set  out the pump and connect the  battery. Switch
    pump on for a quick burst to check that it is working. Dial the pump speed to 5.
4.  Remove a tubing  kit, two sample containers, and  a bridge bottle from the cooler and place on the plastic
    near the pump.
5.  Remove a pack of sample gloves from  the storage container and place on the plastic.
6.  Remove the plastic sample caddie from the storage box and place it on the sample processing area near the
    pump. Secure the two sample bottles in the caddie.
                                                  220

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


Bridge Bottle Filling
1 . Locate the sample weights for connection to the BRIDGE BOTTLE.
2.  Locate  the polypropylene sampling rope spool, cut a sufficient length of rope to allow for deployment, and
    place on plastic.
3.  Don one or two pairs of vinyl gloves using clean precautions.
4.  Tie one end of the sampling rope to the five pound weight leaving approximately a V long end for connection
    to the BRIDGE BOTTLE.
5.  Untie or open by tearing the top of the outer plastic bag  containing the BRIDGE BOTTLE.
6.  Reach into the outer bag and untie or tear the inner bag near the handle connection. Check the configuration
    of the tubing to ensure that proper filling will occur. While the bottle is still in the inner bag it is acceptable to
    remove the top fitting to check the inner sipper tube. Adjust all fittings appropriately.
7.  When the fittings have been properly secured and adjusted remove the BRIDGE BOTTLE from the inner bag
    and lay on the plastic film. Tie the weighted end of the  rope onto the handle of the bottle leaving about 6" of
    line between the bottle and the weight.
8.  Proceed to the sampling location with  the BRIDGE BOTTLE apparatus. If appropriate carry several extra
    pairs of gloves to the site to facilitate bridge bottle handling.
9.  When deploying from  bridges with moderate to low stream velocities collect the sample upstream of the
    bridge by lowering the  assembly  into the water. Ensure  that the assembly does not contact any structures or
    other objects as it is lowered into  the water.
10. Once in the water the weight will partially submerge the BRIDGE BOTTLE which will begin to fill. Check to
    insure the air release tube is above the water level and not obstructed. When the bottle is first submerged a
    good indication it is filling properly is a small slug of water may be expelled from the air vent tube. The bottle
    will fill quickly if it has been properly adjusted.
11. Problems with filling from bridges can occur when stream velocities are high. Sampling on the  downstream
    side of bridges is acceptable to avoid the risk of losing  the assembly due to the current sweeping it  under a
    bridge or other obstruction. When stream velocities are high an additional 7.5 pounds of weight will aid in
    sample collection. The added weight will cause the container to sink lower when partially filled which may
    submerge the vent tube. The vent tube can be extended past the bottom of the bottle to prevent filling with
    water when the weight is heavy or the water is rough.
12. Other problems with filling can occur when the inlet tube is clogged, the vent tube contains a slug of water or
    other obstruction, the vent  tube is below the surface of the water, the weight is not positioned close enough
    to the bottle, or the vent tube or inlet tube has become disconnected from the bottle.
13. When the BRIDGE  BOTTLE is  approximately 1/3 to  1/2 full retrieve the bottle and return  to the sample
    processing area. Ensure that the  assembly does not contact any structures or other objects as  it is retrieved.
14. When deploying while wading or from a small craft the BRIDGE BOTTLE can be submerged by hand without
    the weights.
15. When the water level at the sample site is very shallow  it may be difficult to submerge the BRIDGE BOTTLE
    deep enough to begin siphoning.  The  alternative is to use the effluent sample configuration where the stream
    sample is pumped directly into  the  loop sample container. Sampling in  this manner requires the pump
    assembly to be transported to the site. This is best accomplished by attaching the  pump  assembly to a
    backpack.
16. Once the BRIDGE BOTTLE has been brought back to the sample processing area set it next to the pump
    and remove the weight. With the inlet and vent tubing properly configured the BRIDGE BOTTLE can remain
    on the plastic outside of a bag without any danger of atmospheric contamination.

Ambient Dissolved Grab Blank
1 .  Refer to Figure 2 Sample  Collection Scheme for the  schematic of the field  sampling equipment  used to
    process blanks and samples.
2.  Determine which tech will be clean hands and which will be dirty hands.
3.  Dirty hands and clean hands don  one  or two pairs of vinyl gloves. Dirty hands opens the sample  bottles outer
    plastic bag, clean hands opens the inner plastic bag.
4.  Dirty hands opens the grab kit's outer plastic  bag, clean hands opens the inner plastic bag and removes the
    tubing assembly.
5.  Clean hands disconnects one side of the sample loop on a sample container. Clean hands connects the end
    of the tubing kit opposite the filter to the opened sample container. Remember the sample container is full of
    clean water from the lab.
6.  Dirty hands connects the peristaltic tubing at approximately the mid-point of the length to the field pump,
    clean hands inverts the sample container, and dirty hands switches on the pump.


                                                  221

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


7.  Process the entire contents, 1000 mis, of the sample container through the tubing and filter apparatus at a
    flow rate of 500 mls/min. (pump setting of 5). At the beginning of the sample processing orient the filter
    cartridge with the flow arrow pointing up. This will insure proper wetting of the filter. After the last noticeable
    air bubbles have been expelled from the filter it can be oriented in any direction during the rinsing. When the
    sample container is nearing empty orient the capsule filter with the arrow down. Once the sample container
    is empty continue processing until the free water in the tubing and filter has been expelled.
8.  Dirty hands turns off the pump. This step is a rinse of the filter which cleans and conditions the media. The
    rinse can be pumped directly to waste.
9.  Use the same flow and orientation scheme throughout all sample and blank processing.

Field Blank
1 .  Clean hands removes the pump tubing from the empty bottle and  connects the capsule filter to the empty
    container via the sample loop  tubing. Remove one side of the loop fitting from a second sample container
    and connect the free  end of the tubing to the container. Remember that this second bottle is full of clean
    water from the lab. Invert the container, dirty hands switches on  the pump. Process  the  entire contents,
    approximately 1000 mis, of the sample container through the tubing and filter apparatus into the first sample
    container.
2.  Dirty hands switches  off the pump when the filter has been emptied. Clean hands  disconnects the outlet
    tubing from the blank  sample container and  immediately reconnects the loop tubing, seals the inner bag and
    places inside  the  outer  bag.  Dirty hands  identifies this  as a blank  by recording  the  ULTRA #  as  the
    CONTAINER ID, seals the outer bags, and places on ice in a sample cooler.
3.  Clean hands immediately (immediately means the sooner the switch is made the less  likely contamination
    can adhere to the end of an open tube, immediately means less than one minute) disconnects the tubing
    from the sample container and then connects the free end containing the filter, which was just removed from
    the blank sample container, to the empty sample container.  Clean hands then  disconnects the vent tubing
    from the BRIDGE BOTTLE and connects the pump tubing in place of the vent tubing.
4.  The field blank collected  in this manner is a comprehensive blank because  it is collected in the same
    equipment as the sample and  it is processed  like the sample through  all  steps of the  protocol. This is  the
    most important check  of contamination in the protocol.

Ambient Dissolved Grab
1.  In  a field notebook dirty hands records the sample container identification number, date, time. Clean hands
    secures the sample container in the sample caddie.
2.  While keeping the BRIDGE BOTTLE near level dirty hands holds the bottle at the midsection or lower and
    shakes the vessel to ensure mixing. Dirty hands switches on the pump and clean hands collects a full sample
    container while following  the  filter orientation scheme. It  is acceptable  to fill the sample container to
    overflowing, however  avoid filtering more than  1000 mis through the filter.
3   Once the sample container  is  full, dirty hands switches off the pump. Clean hands  disconnects the outlet
    tubing from the sample container and immediately reconnects the loop tubing on the sample container, seals
    the inner bag and places inside the outer bag. Dirty hands identifies the sample by recording the ULTRA # as
    the CONTAINER ID, seals the outer bags, and  places on ice in a sample cooler.
4.  The clean protocol is complete at this step and field parameters can now be taken from the remaining water
    in the BRIDGE BOTTLE. Suggested field parameters include pH, Conductivity, Temperature, and Dissolved
    Oxygen. Additional laboratory  samples for the solid series  and  total organic carbon should  be prepared;
    group code SOLIDS, catalog number 190-243,  and group code NME10. catalog number  190-25.
5.  When finished with the  BRIDGE BOTTLE seal the apparatus  in a plastic  bag  and return to the cooler for
    recycling back to DCLS. DCLS will  reclean, certify, and repackage for reuse.
6.  Rinse the rope and weights with ambient water to remove any visible dirt, place inside a plastic bag, and
    store in the storage container. Rope may be reused several times if rinsed frequently.

Effluent Sample Collection Protocol
Equipment Setup
1    Locate an area near the final effluent sampling location where sample processing will occur. This should be
    an area free of falling debris and swirling dust, flat, smooth, and protected from the wind. The tailgate of a
    vehicle or the back of  a Suburban are good locations.
2.  Locate the equipment box and coolers containing the sample containers and kits in the area where sample
    processing will occur.
3.  Cover the work area with a large pick of plastic film. Set out the pump and connect the battery. Switch pump


                                                 222

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium


    on for a quick burst to check that it is working. Dial the pump speed to 5.
4.  Remove a tubing kit and two sample containers from the cooler and place on the plastic near the pump.
5.  Remove a pack of sample gloves from the storage container and place  on the plastic.  Refer to, for the
    schematic of the field sampling equipment  used to collect treatment plant grab samples.
6.  Remove the plastic sample caddie from the storage box and place it on the  sample processing area near the
    pump.
7.  Locate the sample wand used for positioning the teflon sample tubing in the  effluent.

Effluent Dissolved Grab Blank
1 .  Refer to Figure 2 Sample Collection  Scheme for the  schematic of the field sampling equipment  used to
    process blanks and samples.
2.  Determine which tech will be clean hands and which will be dirty hands.
3.  Dirty hands and clean hands don one or two pairs of vinyl gloves. Dirty hands opens the sample bottles outer
    plastic bag, clean hands opens the inner plastic bag.
4.  Dirty hands opens the grab kit's outer plastic bag, clean hands opens the inner plastic bag  and removes the
    tubing assembly.
5.  Clean hands disconnects one side of the sample loop on a sample container.  Clean hands connects the
    teflon end of the tubing  kit  opposite the filter to the opened sample container. Remember the sample
    container is full  of clean water from the lab.
6.  Dirty hands connects the  peristaltic tubing to the field pump on the end of the  pump tubing  closest to the
    connection with the teflon tubing. This allows for slack on the  filter end of the  pump tubing. Clean hands
    inverts the sample container, and dirty hands switches on the pump.
7.  Process the entire  contents, 1000 mis, of  the sample container through the tubing and filter apparatus at a
    flow rate of 500 mls/min. (pump setting of 5). At the  beginning of the sample processing orient the filter
    cartridge with the flow arrow pointing up. This will insure proper wetting of the filter. After the last noticeable
    air bubbles have been expelled from the filter it can be oriented in any direction during the rinsing. When the
    sample container is nearing empty orient the capsule filter with the arrow down.  Once the  sample container
    is empty continue processing until the  free  water in the tubing and filter has  been expelled.
8.  Dirty hands turns off the pump. This step  is a rinse of the filter which cleans and conditions the  media. The
    rinse can be pumped directly to waste.
9.  Use the same flow and  orientation scheme throughout all sample and blank processing.
10. Clean hands removes the teflon tubing from the empty bottle and connects the capsule filter to the empty
    container via the sample loop tubing. Remove one side of the loop fitting  from  a second  sample container
    and connect the teflon  end of the tubing to the container. Remember that  this second bottle is full  of clean
    water from the  lab. Clean hands inverts the container, and dirty hands switches on the pump. Process the
    entire contents, approximately 1000 mis, of the sample container through the tubing and filter apparatus into
    the first sample container.
11. Dirty hands switches off the pump when  the filter has  been emptied as  evidenced by air bubbles  in the
    system. Clean hands disconnects the outlet tubing from the sample container and  immediately reconnects
    the loop tubing on the  sample container,  seals the inner bag and places  inside the outer bag.  Dirty  hands
    identifies this as a blank, seals the outer bags, and places on ice in a sample cooler.
12. This is  a  comprehensive blank because  it is collected in the same  equipment as the  sample and  it is
    processed  like the sample through all  steps  of the protocol.  This is the most important check  of
    contamination in the protocol.

Effluent Dissolved Grab
1 .  In a field notebook dirty hands records the sample container identification  number, date, time. Clean hands
    secures the sample container in the sample caddie.
2.  Clean hands connects the filter to the empty sample container.  Clean hands presents a section of the Teflon
    tubing just past the inlet to dirty hands who then attaches the tubing to the sample wand.
3.  The entire assembly: sample caddie  containing the empty sample container, sample tubing, pump/battery,
    and sample wand are transported to the effluent sampling location.
4.  Dirty hands places the sample wand sample collection zone taking precaution not to touch the tip of the
    sampling tube on any items. Once the sample tube is located in the effluent take precaution not to let the tip
    contact anything but water.
5.  Dirty hands switches on the pump and clean hands collects a full sample  container while  following the filter
    orientation scheme. It is acceptable to fill  the sample container to overflowing, however avoid filtering more
    than 1000 mis through the filter.


                                                  223

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
6.   Once the sample container is full dirty hands switches off the pump. Clean  hands disconnects the outlet
    tubing from the sample container and immediately reconnects the loop tubing on the sample container, seals
    the inner bag and places inside the  outer bag. Dirty hands identifies the sample, seals the outer bags,  and
    places on ice in a sample cooler.
7.   The clean protocol is complete at this step and field parameters can now be taken directly from the effluent.
    Suggested field parameters include pH, Conductivity, Temperature,  and Dissolved Oxygen. Additional
    laboratory samples  for the  solid series and  total  organic carbon should be  prepared; group code SOLIDS,
    catalog number 190-243, and group code NME10, catalog number 190-25.

CONCLUSIONS
Using these clean procedures  to collect more than two hundred samples the  data  indicate that collection of
contaminant free samples at trace concentrations is possible. State regulatory agencies should be encouraged to
move forward  in  promulgating  guidance for trace element  wastewater characterizations and  ambient water
quality monitoring. The  careful  application of the techniques developed by the U.S. Geological Survey and the
U.S. Environmental Protection Agency can produce high quality data, satisfactory for regulatory decision making.
By  adopting the system of field procedures proposed here,  investigations of water  quality can  be performed
simply and cost effectively.

ACKNOWLEDGMENTS
This work was prepared under contract with the U.S. Environmental  Protection Agency. The author would like to
thank the following people for their assistance in conducting  this work: Jim Anderson, Lisa  McMillan, and Ron
Gregory.

                  Table 1. Target Analytes, Analytical Test Methods, and Detection Limits
Parameter CAS Number
,;,.-,,.., ,,-'.-..:•' • . ,; ^;:,, •,
Aluminium
Antimony
Arsenic
Cadmium
Calcium
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Selenium
Silver
Zinc
7429-90-6
7440-36-0
7440-38-2
7440-43-9
7440-70-2
7440-47-3
7440-50-8
7439-89-6
7439-92-1
7439-95-4
7439-96-5
7439-97-6
7440-02-0
7782-49-2
7440-22-4
7440-66-6
•*!<*- Method Detection Limits, ug/L t ! t; ;
ICPMS USN '* ICPMS MSN ICPM$ : ICP A€S CVAA '
TR • .' ,, .v.:. •:..,!•* USN •„{& ', . ;;*;• ,,,;i!'

0.05
0.07
0.06

0.02
0.02

0.17

0.02

0.04

0.19
0.26
0.04
0.03
0.03
0.04

0.04
0.06

0.03

0.03

0.02
0.06
0.03
0.03
0.37
0.33
0.24
1.72


0.87

0.28

1.32
0.12
0.39
0.77
0.15
2.18


5.37
2.37
0.08
2.27
4.98
2.3

0
0.58

1.71


1395











0.07




ICPMS inductively coupled plasma mass spectrometry sample introduction by ultrasonic nebulization
USN . , - , " i ^ ]' ' • • "• - , ;:; • -
ICPMS inductively coupled plasma mass spectrometry sample introduction by ultrasonic nebulization
USNTR total recoverable
ICPMS inductively coupled plasma rrtas^ spectrometry i
ICP A6S inductively coupled plasma atorriic emission spectrometey sample introduction #y ultrasonic ! ,
USN nebulization
CVAA cold vapor atomic absorption spectrorrietry
REFERENCES
1.  Quality Criteria of Water 1986. United States Environmental Protection Agency, EPA 440/5-86-001, May 1,
   1986.
2.  Standard Operating Procedure for the  Collection  of Freshwaters and Wastewater for the Determination of
                                                 224

-------
                        WTQA '98 - 14th Annual Waste Testing & Qualify Assurance Symposium
   Trace Elements, Virginia Department of Environmental, 15 May 1997.
3. Federal Register, 60848, Vol. 57, No. 246, Tuesday, December 22, 1992.
4.  U.S. Geological Survey  Protocol for the  Collection  and Processing of Surface-Water Samples for  the
   Subsequent Determination of Inorganic Constituents in Filtered Water, Open-File Report 94-539.
5. Method 1631: Mercury in Water by Oxidation Purge and Trap, Cold Vapor Atomic Fluorescence Spectrometry,
   EPA 821-R-95-027, April 1995 DRAFT.
6. Method 1632:  Determination of Inorganic Arsenic in Water by Hydride Generation Flame Atomic Absorption.
   EPA 821-R-95-028, April 1995 DRAFT.
7.  Method  1636:  Determination of Hexavalent Chromium  by Ion Chromatography,  EPA  821-R-95-029, April
   1995.
8.  Method  1637: Determination  of Trace  Elements  in Ambient Waters by Chelation  Preconcentration with
   Graphite Furnace Atomic Absorption, EPA 821-R-95-030, April 1995.
9.  Method  1638:  Determination of Trace Elements  in Ambient Waters  by Inductively Coupled Plasma-Mass
   Spectrometry,  EPA 821 -R-95-031, April 1995.
10. Method 1639: Determination  of Trace Elements in Ambient Waters by Stabilized Temperature Graphite
   Furnace Atomic Absorption, EPA 821-R-95-032, April 1995.
11. Method1640:  Determination of Trace Elements in Ambient  Waters by On-line Chelation Preconcentration
   and Inductively Coupled  Plasma-Mass Spectrometry, EPA 821-R-95-033, April 1995.
12. Method 1669: Sampling Ambient Water for  Trace  Metals at  EPA Water Quality Criteria Levels,  U.S.
   Environmental Protection Agency, 821-R-5-034, April 1995.
13. Benolt, Gaboury, Clean Technique Measurement of Pb, Ag, and Cd  in Freshwater: A Redefinition of Metal
   Pollution, Environ. Sci. Technol., Vol. 28, No. 11, 1994.
14. Horowitz,  Arthur J., et.al., Problems Associated with  Using Filtration to Define  Dissolved Trace Element
   Concentrations in Natural Water Samples, Environ. Sci. Technol., Volume 30, Number 3. Pages 954-963,
   1996.
15.  Horowitz, A.J.  et.  al.,  The  Effect  of Membrane  Filtration  Artifacts  on Dissolved Trace Element
   Concentrations, Wat. Res. Vol. 26, No. 6, pp. 753-763,  1992.

                                                                                     6" X 0.19" ID
                                                                                     size 25 C-Rex
            Figure 1. Loop Sample Bottle
                                            1 liter plastic
                                            wide mouth

                                            high-density
                                            polyethylene

                                            leakproof
    1.5" X Q.I 9" ID
    size 25 C-Flex
1.5"X1/4"ODFEP
                                                                      polypropylene screw cap
                                                                      with 2X19/64" holes
                              Peristaltic Pump
                                                      capsule filter
                                                      •L
                        -r
                    4'CFLEX25tubing
                • 2" white PVC
    •f
  1Ltwo port
polyethylene bottle
                                                                     Figure 2. Sample Collection
                                                                     Scheme
                                                  225

-------
WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
                            226

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Author

Alexander IV, J.N.
Armstrong, D.W.
Autenrieth, R.L.

Bakaltcheva, I.B.
Barat, R.B.
Belanger, J.M.R.
Birri, J.
Blaterwick
Bonde, S.E.
Boring, C.B.
Boswell, C.E.
Boylan, H.M.
Briggs, R.G.
Broske,A.
Brown, R.D.
Bruce, M.
Bruce, M.L.
Brusseau
Bynum, R.V.

Caillouet, D.J.
Calcavecchio, R.
Carlson, R.E.
Charles, P.T.
Chin, C.
Clark, M.S.
Connell, T.R.
Cortese, N.A.

Dahlgran, J.
Dasgupta,  P.K.
Donnelly, K.C.
Dorn, H.C.
Drake, E.N.
Dupes, L.
Dupes, L.J.
Durst, H.D.

Eatough, D.J.
Edgerley, D.A.
Eilcone, C.
Ellickson, M.L.

Fahnenstiel, G.
Fan, T.S.
Fan, T.S.
Federici, J.F.
Felix, A.
Fidler, B.
Flores, R.A.
Fordham, O.
AUTHOR INDEX
aper
o.
1
3
9

0
5

9
5
4
1
2
0
1

2

1
0
2
6
5
2


2
7
3

3
1
9
9
5
8
6
7

8
7
3


4
1
3
1
5
5
4
9
Page
No.
94
131
93

26
132
11
188
137
28
94
205
53
151
4
152
17
55
148
152
173
29
27
26
11
205
88
102

161
94
93
147
29
43
173
88

146
181
209
4

65
27
28
124
29
215
165
53

Author
Forman, R.

Gauger, P.R.
George, L.A.
Gere, D.R.
Grabanski, C.B.
Grosser, Z.

Haire, B.
Hard, T.M.
Hanson, R.
Harrison, R.O.
Hawthorne, S.B.
He, L.-Y.
Heglund, D.L.
Hering, J.G.
Hoberecht, H.
Homstead, J.
Hunter, P.M.
Huo, D.
Jayne, J.T.
Johnson, H.C.

Kane, J.S.
Kaphart, T.S.
Keeler
Kennicutt II, M.C.
Kenny, T.D.
Kinghorn, R.
Kingston, H.M.
Kingston, H.M.
Kolb, C.E.
Krappe, M.
Krishnan, B.
Kroll, D.
Kustarbeck, A.W.

Lagadec, A.J.M.
LeBlanc, G.
LeBlanc, G.
Lee, E.
Lee, G.H.
LeMoine, E.A.
Lesnik, B.
Lesnik, B.
Lesnik, B.
Levy, R.A.
Li, W.
Lippmann, M.
Lu, Y.
Lubbad, S.H.
Paper
No.

18
Page
No.

43
9
38
5
30
66
65
38
35
12
30
29
32
37
16
58
25
23
44
55
59
31
39
33
42
5
20
23
44
30
36
22
9
30
64
65
34
54
16
1
2
7
41
47
47
23
32
26
118
4
93
215
215
118
117
27
93
93
101
117
35
187
73
64
131
165
188
94
119
102
125
4
53
64
131
93
117
60
26
93
214
215
110
165
35
3
3
11
124
137
137
64
101
                                                   227

-------
                        WTQA '98 - 14th Annual Waste Testing & Quality Assurance Symposium
Author

MacDonald, B.
Macko, S.A.
Marquis,  B.T.
Masila, M.
Mays, J.
McDonald,  T.J.
Merschman, S.A.
Miller,  D.J.
Miller,  LM.
Miller,  M.
Min, J.H.
Moniot, C.L.
Morrissey, K.M.

Nagourney, S.J.

O'Brien,  R.J.
Obeidi, F.

Pare, J.R.J.
Patterson, C.
Patterson, E.
Paustian, M.
Peist, K.
Poziomek,  E.J.

Quin, Y

Reddy, S.M.
Rediske, R.
Reeves,  W.R.
Richards, K.L.
Richards, K.L.
Richter, R.C.
Risden, R.M.
Robbat Jr, A.

Sakik,  O.A.
Schelske, C.
Shriver-Lake, L.C.
Singhvi,  R.
Skoczenski, B.A.
Skoczenski, B.A.
Smith, K.A.
Smith, R-K.
Solsky, J.F.
Somers,  S.R.
Sparks, L.D.
Stahl, D.C.
Stelz, W.
Stewart II, R.E.
Stroud, R.B.
Stutz, M.H.
Paper
No.

59
33
42
34
27
29
32
30
21
64
37
30
27

59

38
48
7
9
48
66
59
58

33
63
24
29
8
21
20
8
40
34
24
10
7
11
13
44
17
3
26
52
32
28
67
52
6
Page
No.

188
102
125
110
88
93
101
93
55
214
117
93
88

188

118
146
11
26
146
215
188
187

102
209
65
93
17
55
53
17
119
110
65
26
11
27
28
131
40
3
77
152
101
93
219
152
5
Author

Thies, C.
Tillotta, D.C.
Tuchman, M.
Tummillo Jr, N.J.
Turle, R.
Turpin, R.

Valente, H.
van Bergen, S.
VanGaalen, G.C.
Vetelino, J.F
Vitale, R.J.
Vitale, R.J.

Watkins, C.S.
Weisberg, S.
Wilding, S.
Woodard, R.M.
Worsnop, D.R.
Xu, H.
Xuan, T.
Yanik, P.J.

Zemansky, G.M.
Zhang, X.
Ziong, J.Q.

























Paper
No.
53
32
24
59
7
7
51
10
15
42
18
58
55
4
60
62
44
34
63
33
61
44
47
Page
No.
161
101
65
188
11
11
151
26
29
125
43
173
165
4
191
205
131
110
209
102
192
131
137
                                                   228

-------