»EPA
           United States       Office of Research and    EPA/540/R-05/003
           Environmental Protection   Development       March 2005
           Agency         Washington, DC 20460
           Innovative Technology
           Verification Report
           Technologies for Monitoring
           and Measurement of Dioxin
           and Dioxin-like Compounds
           in Soil and Sediment

           Abraxis LLC
           Coplanar PCB ELISA Kit

-------

-------
                                    EPA/540/R-05/003
                                    March 2005
    Innovative Technology
       Verification Report

Abraxis LLC Coplanar PCB ELISA Kit
                   Prepared by

                    Battelle
                  505 King Avenue
                Columbus, Ohio 43201

                Contract No. 68-C-00-185
                  Stephen Billets
              Environmental Sciences Division
            National Exposure Research Laboratory
             Office of Research and Development
             U.S. Environmental Protection Agency
                 Las Vegas, NV 89119

-------
                                               Notice
This document was prepared for the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology
Evaluation Program under Contract No. 68-C-00-185. The document has met the EPA's requirements for peer and
administrative review and has been approved for publication. Mention of corporation names, trade names, or
commercial products does not constitute endorsement or recommendation for use.

-------
                                               Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the nation's natural
resources. Under the mandate of national environmental laws, the Agency strives to formulate and implement actions
leading to a compatible balance between human activities and the ability of natural systems to support and nurture
life. To meet this mandate, the EPA's Office of Research and Development (ORD) provides data and scientific
support that can be used to solve environmental problems, build the scientific knowledge base needed to manage
ecological resources wisely, understand how pollutants affect public health, and prevent or reduce environmental
risks.

The National Exposure Research Laboratory is the Agency's center for investigation of technical and management
approaches for identifying and quantifying risks to human health and the environment. Goals of the Laboratory's
research program are to (1) develop and evaluate methods and technologies for characterizing and monitoring air, soil,
and water; (2) support regulatory and policy decisions; and (3) provide the scientific support needed to ensure
effective implementation of environmental regulations and strategies.

The EPA's Superfund Innovative Technology Evaluation (SITE) Program evaluates technologies designed for
characterization and remediation of contaminated Superfund and Resource Conservation and Recovery Act (RCRA)
sites. The SITE Program was created to provide reliable cost and performance  data in order to speed the acceptance
and use of innovative remediation, characterization, and monitoring technologies by the regulatory and user
community.

Effective monitoring and measurement technologies are needed to assess the degree of contamination at a site,
provide data that can be used to determine the risk to public health or the environment, and monitor the success or
failure of a remediation process. One component of the EPA SITE Program, the Monitoring and Measurement
Technology (MMT) Program, demonstrates and evaluates innovative technologies to meet these needs.

Candidate technologies can originate within the federal government or the private sector. Through the SITE Program,
developers are given the opportunity to conduct a rigorous demonstration of their technologies under actual field
conditions. By completing the demonstration  and distributing the results, the Agency establishes a baseline for
acceptance and use of these technologies. The MMT Program is managed by the ORD's Environmental Sciences
Division in Las Vegas, Nevada.

                                                   Gary Foley, Ph.D.
                                                   Director
                                                   National Exposure Research Laboratory
                                                   Office of Research and Development
                                                    in

-------
                                                Abstract
A demonstration of technologies for determining the presence of dioxin and dioxin-like compounds in soil and
sediment was conducted under the U.S. Environmental Protection Agency's (EPA's) Superfund Innovative
Technology Evaluation Program in Saginaw, Michigan, at Green Point Environmental Learning Center from April 26
to May 5, 2004. This innovative technology verification report describes the objectives and the results of that
demonstration, and serves to verify the performance and cost of the Abraxis LLC Coplanar PCB Enzyme-Linked
Immunosorbent Assay (ELISA) Kit. Four other technologies were evaluated as part of this demonstration, and
separate reports have been prepared for each technology. The objectives of the demonstration included evaluating the
technology's accuracy, precision, sensitivity, sample throughput, tendency for matrix effects, and cost. The test also
included an assessment of how well the technology's results compared to those generated by established laboratory
methods using high-resolution mass spectrometry (HRMS). The demonstration objectives were accomplished by
evaluating the results generated by the technology from 209 soil, sediment, and extract samples. The test samples
included performance evaluation (PE) samples (i.e., contaminant concentrations were certified or the samples were
spiked with known contaminants) and environmental samples collected from 10 different sampling locations.

The Abraxis LLC Coplanar PCB ELISA Kit is an immunoassay technology that reports the total toxicity equivalents
(TEQ) of coplanar polychlorinated biphenyls (PCBs) in the sample. As part of the performance evaluation, the
technology results were compared to TEQ results generated by a reference laboratory, AXYS Analytical Services,
using EPA Method 1668A. It should be noted  that this technology may not directly correlate to HRMS TEQPCB in all
cases because it is known that the congener responses and cross-reactivities to the kit are not identical to the World
Heath Organization toxicity equivalency factors that are used to convert congener HRMS concentration values to
TEQPCB.  The  effect of cross-reactivities may contribute to this technology reporting results, which are biased high or
low compared to HRMS TEQPCB results. Therefore, the Abraxis kit should not be viewed as producing an equivalent
measurement  value to HRMS TEQPCB but as a screening value to approximate HRMS TEQPCB concentration. It has
been suggested that correlation between the Abraxis TEQPCB results and HRMS TEQPCB results could be improved by
first characterizing a site and calibrating the Abraxis results to HRMS results. Subsequent analysis using the Abraxis
kit for samples obtained from this site may then show better correlation with the HRMS  TEQPCB result. This approach
was not evaluated during this demonstration.

The Abraxis kit reported data higher and lower than the certified PE values. Abraxis generally reported data that were
higher than the reference laboratory TEQPCB values, with the exception of ultra-high level PCB samples [> 10,000
picogram/gram (pg/g) TEQ] where Abraxis reported values lower than the reference method. The technology's
estimated MDL was 6 to 31 pg/g TEQPCB; the developer's reporting limit was 6.25 pg/g TEQPCB. No  statistically
significant matrix effects on precision were observed by sample type  (performance evaluation vs. environmental vs.
extract), matrix type (soil vs. sediment vs. extract), or polynuclear aromatic hydrocarbon (PAH) concentration. One
result (5% of total) from replicate sample sets that were analyzed in the laboratory and in the field showed a
significant statistical difference, but the one sample was a PE sample that was spiked with only PAHs and no PCBs.
The kit had a false positive rate of 35% and a false negative rate of 7% around 6.25 pg/g TEQPCB (the reporting limits
of the technology). Abraxis reported significantly fewer false positives (8%) and false negatives (3%) around 50 pg/g
TEQPCB. This  evaluation indicates that the Abraxis kit could be an effective screening tool for screening sample
concentrations above and below 50 pg/g TEQPCB, particularly considering that the cost ($22,668 vs. $184,449) and the
time to analyze the 209 demonstration samples were significantly less than those of the reference laboratory.
                                                    IV

-------
                                              Contents
Chapter                                                                                           Page

Notice	ii
Foreword  	 iii
Abstract 	iv
Abbreviations, Acronyms, and Symbols	ix
Acknowledgments	xii

1   Introduction	 1
    1.1   Description of the SITE MMT Program  	 1
    1.2   Scope of This Demonstration	 3
         1.2.1   Organization of Demonstration	 4
         1.2.2   Sample Descriptions and Experimental Design	 5
         1.2.3   Overview of Field Demonstration	 5

2   Description of Abraxis Coplanar PCB ELISA Kit	 6
    2.1   Company History  	 6
    2.2   Product History	 6
    2.3   Technology Description 	 6
    2.4   Developer Contact Information  	 8

3   Demonstration and Environmental Site Descriptions  	 9
    3.1   Demonstration Site Description and Selection Process	 9
    3.2   Description of Sampling Locations  	 10
         3.2.1   Soil Sampling Locations 	 10
         3.2.2   Sediment Sampling Sites	 12

4   Demonstration Approach 	 14
    4.1   Demonstration Objectives  	 14
    4.2   Toxicity Equivalents	 14
    4.3   Overview of Demonstration Samples	 16
         4.3.1   PE Samples  	 16
         4.3.2   Environmental Samples	 19
         4.3.3   Extracts  	 22
    4.4   Sample Handling	 22
    4.5   Pre-Demonstration Study	 24
    4.6   Execution of Field Demonstration  	 24
    4.7   Assessment of Primary and  Secondary Objectives  	 24
         4.7.1   Primary Objective PI: Accuracy	 25
         4.7.2   Primary Objective P2: Precision	 25
         4.7.3   Primary Objective P3: Comparability	 25
         4.7.4   Primary Objective P4: Estimated Method Detection Limit 	 26
         4.7.5   Primary Objective P5: False Positive/False Negative Results  	 26
         4.7.6   Primary Objective P6: Matrix Effects	 26

-------
                                        Contents (continued)

                                                                                                   Page

         4.7.7   Primary Objective P7: Technology Costs	  27
         4.7.8   Secondary Objective SI: Skill Level of Operator	  27
         4.7.9   Secondary Objective S2: Health and Safety Aspects 	  27
         4.7.10  Secondary Objective S3: Portability	  27
         4.7.11  Secondary Objective S4: Sample Throughput 	  27

5   Confirmatory Process	  28
    5.1   Traditional Methods for Measurement of Dioxin and Dioxin-Like
         Compounds in Soil and Sediment  	  28
         5.1.1   High-Resolution Mass Spectrometry  	  28
         5.1.2   Low-Resolution Mass Spectrometry	  28
         5.1.3   PCB Methods	  28
         5.1.4   Reference Method Selection	  29
    5.2   Characterization of Environmental Samples 	  29
         5.2.1   Dioxins and Furans  	  29
         5.2.2   PCBs  	  30
         5.2.3   PAHs  	  30
    5.3   Reference Laboratory Selection  	  30
    5.4   Reference Laboratory Sample Preparation and Analytical Methods	  31
         5.4.1   Dioxin/Furan Analysis	  31
         5.4.2   PCB Analysis	  31
         5.4.3   TEQ Calculations 	  31

6   Assessment of Reference Method Data Quality 	  33
    6.1   QA Audits	  33
    6.2   QC Results  	  34
         6.2.1   Holding Times and Storage Conditions  	  34
         6.2.2   Chain of Custody	  34
         6.2.3   Standard Concentrations	  34
         6.2.4   Initial and Continuing Calibration   	  34
         6.2.5   Column Performance and Instrument Resolution  	  35
         6.2.6   Method Blanks	  35
         6.2.7   Internal Standard Recovery	  35
         6.2.8   Laboratory Control Spikes	  35
         6.2.9   Sample Batch Duplicates  	  35
    6.3   Evaluation of Primary Objective PI: Accuracy	  35
    6.4   Evaluation of Primary Objective P2: Precision  	  36
    6.5   Comparability to Characterization Data	  37
    6.6   Performance Summary	  37

7   Performance of Abraxis Coplanar PCB ELISA Kit	  40
    7.1   Evaluation of Coplanar PCB ELISA Kit Performance 	  40
         7.1.1   Evaluation of Primary Objective PI: Accuracy 	  40
         7.1.2   Evaluation of Primary Objective P2: Precision 	  41
         7.1.3   Evaluation of Primary Objective P3: Comparability 	  41
         7.1.4   Evaluation of Primary Objective P4: Estimated Method Detection Limit	  44
         7.1.5   Evaluation of Primary Objective P5: False Positive/False Negative Results	  44
         7.1.6   Evaluation of Primary Objective P6: Matrix Effects 	  45
         7.1.7   Evaluation of Primary Objective P7: Technology Costs  	  45
                                                   VI

-------
                                       Contents (continued)

                                                                                                  Page

    7.2   Observer Report: Evaluation of Secondary Objectives 	 45
         7.2.1   Evaluation of Secondary Objective SI: Skill Level of Operator  	 47
         7.2.2   Evaluation of Secondary Objective S2: Health and Safety Aspects	 48
         7.2.3   Evaluation of Secondary Objective S3: Portability 	 49
         7.2.4   Evaluation of Secondary Objective S4: Throughput  	 49
         7.2.5   Miscellaneous Observer Notes 	 49

8   Economic Analysis	 51
    8.1   Issues and Assumptions  	 51
         8.1.1   Capital Equipment Cost 	 51
         8.1.2   Cost of Supplies	 51
         8.1.3   Support Equipment Cost	 51
         8.1.4   Labor Cost	 52
         8.1.5   Investigation-Derived Waste Disposal Cost	 52
         8.1.6   Costs Not Included  	 52
    8.2   Coplanar PCB ELISA Kit Costs	 53
         8.2.1.   Capital Equipment Cost 	 53
         8.2.2   Cost of Supplies	 53
         8.2.3   Support Equipment Cost	 53
         8.2.4   Labor Cost	 53
         8.2.5   Investigation-Derived Waste Disposal Cost	 54
         8.2.6   Summary of Coplanar PCB ELISA Kit Costs	 54
    8.3   Reference Method Costs	 54
    8.4   Comparison of Economic Analysis Results	 55

9   Technology Performance Summary	 57

10  References	 60

Appendix A SITE Monitoring and Measurement Technology Program Verification Statement	A-l
Appendix B  Supplemental Information Supplied by the Developer	 B-l
Appendix C  Reference Laboratory Method Blank and Duplicate Results Summary	 C-l
Appendix D Summary of Developer and Reference Laboratory Data	D-l
                                                  vn

-------
                                       Contents (continued)

                                                                                                   Page

                                                Figures

1-1    Representative dioxin, furan, and polychlorinated biphenyl structure	 3
2-1    Abraxis Coplanar PCB ELISA Kit  	 6
2-2    Microplate reader used by Abraxis during demonstration  	 7
2-3    Abraxis processing samples during the field demonstration  	 8
6-1    Comparison of reference laboratory and characterization D/F data for environmental samples	  39


                                                 Tables

2-1    Cross-Reactivities for the Abraxis Coplanar PCB ELISA Kit 	 8
3-1    Summary of Environmental Sampling Locations 	  11
4-1    World Health Organization Toxicity Equivalency Factor Values	  15
4-2    Distribution of Samples for the Evaluation of Performance Objectives	  16
4-3    Number and Type  of Samples Analyzed in the Demonstration  	  17
4-4    Summary of Performance Evaluation Samples	  17
4-5    Characterization and Homogenization Analysis Results for Environmental Samples	  21
4-6    Distribution of Extract Samples	  23
5-1    Calibration Range of HRMS Dioxin/Furan Method	  28
5-2    Calibration Range of LRMS Dioxin/Furan Method 	  28
6-1    Objective PI Accuracy - Percent Recovery	  36
6-2    Evaluation of Interferences 	  36
6-3a   Objective P2 Precision - Relative Standard Deviation	  38
6-3b   Objective P2 Precision - Relative Standard Deviation (By Sample Type)	  39
6-4    Reference Method Performance Summary - Primary Objectives	  39
7-1    Objective PI Accuracy - Percent Recovery	  40
7-2a   Objective P2 Precision - Relative Standard Deviation	  42
7-2b   Objective P2 Precision - Relative Standard Deviation (By Sample Type)	  43
7-3    Objective P3 - Comparability Summary Statistics of RPD	  43
7-4    Objective P3 - Comparability Using Interval Assessment  	  43
7-5    Objective P3 - Comparability for Blank Samples	  43
7-6    Objective P4 - Estimated Method Detection Limit	  44
7-7    Objective P5 - False Positive/False Negative Results	  45
7-8    Objective P6 - Matrix Effects Using Descriptive Statistics and ANOVA Results
       Comparing In-Field to Laboratory Analysis  	  46
7-9    Objective P6 - Matrix Effects Using RSD as a Description of Precision by Matrix Type	  47
7-10   Objective P6 - Matrix Effects Using RSD as a Description of Precision by PAH Concentration Levels
       (Environmental Samples Only) 	  47
7-11   Objective P6 - Matrix Effects Using PE Materials	  47
8-1    Cost Summary 	  55
8-2    Reference Method Cost Summary	  56
9-1    Abraxis Coplanar PCB ELISA Kit Performance Summary - Primary Objectives	  58
9-2    Abraxis Coplanar PCB ELISA Kit Performance Summary - Secondary Objectives	  59
                                                  Vlll

-------
                          Abbreviations, Acronyms, and Symbols
Ah
ANOVA
ATSDR
CIL
CoA
COC
CRM
DER
D/F
DNR
D/QAPP
ELC
ELISA
EMDL
EMPC
EPA
ERA
g
GC
HPLC/GPC
HRGC
HRMS
HRP
i.d.
IDW
ITVR
kg
L
LDD
aryl hydrocarbon
analysis of variance
Agency for Toxic Substances and Disease Registry
Cambridge Isotope Laboratories
Certificate of Analysis
chain of custody
certified reference material
data evaluation report
dioxin/furan
Department of Natural Resources
demonstration and quality assurance project plan
Environmental Learning Center
enzyme-linked immunosorbent assay
estimated method detection limit
estimated maximum possible concentration
Environmental Protection Agency
Environmental Resource Associates
gram
gas chromatography
high-performance liquid chromatography/gel permeation chromatography
high-resolution capillary gas chromatography
high-resolution mass spectrometry
horseradish peroxidase
internal diameter
investigation-derived waste
innovative technology verification report
kilogram
liter
least detectable dose
                                                IX

-------
                   Abbreviations, Acronyms, and Symbols (Continued)
LRMS
m
MDEQ
MDL
mg
mL
mm
MMT
MS
NERL
ng
nm
NIST
NOAA
ORD
PAH
PCB
PCDD/F
PCP
PE
Pg
ppm
ppb
ppt
psi
QA/QC
RM
RPD
RSD
SDL
low-resolution mass spectrometry
microliter
micrometer
meter
Michigan Department of Environmental Quality
method detection limit
milligram
milliliter
millimeter
Monitoring and Measurement Technology
mass spectrometry
National Exposure Research Laboratory
nanogram
nanometer
National Institute for Standards and Technology
National Oceanic and Atmospheric Administration
Office of Research and Development
polynuclear aromatic hydrocarbons
polychlorinated biphenyl
polychlorinated dibenzo-p-dioxin/dibenzofuran
pentachlorophenol
performance evaluation
picogram
parts per million; microgram/g; |ig/g
parts per billion; nanogram/g; ng/g
parts per trillion; picogram/g; pg/g
pound per square inch
quality assurance/quality control
reference material
relative percent difference
relative standard deviation
sample-specific detection limit

-------
SIM
SITE
SOP
SRM
TCDD
TEF
TEQ
TEQD/F
TOC
total TEQ

WHO
Abbreviations, Acronyms, and Symbols (Continued)

  selected ion monitoring
  Superfund Innovative Technology Evaluation
  standard operating procedure
  Standard Reference Material®
  tetrachlorodibenzo-/>-dioxin
  toxicity equivalency factor
  toxicity equivalent
  total toxicity equivalents of dioxins/furans
  total toxicity equivalents of World Health Organization polychlorinated biphenyls
  total organic carbon
  total toxicity equivalents including the sum of the dioxin/furan and World Health
  Organization polychlorinated biphenyls
  World Health Organization
                                                 XI

-------
                                         Acknowledgments

This report was prepared for the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology
Evaluation (SITE) Program under the direction and coordination of Stephen Billets of the EPA's National Exposure
Research Laboratory (NERL)—Environmental Sciences Division in Las Vegas, Nevada. George Brilis and
Brian Schumacher of the EPA NERL reviewed and commented on the report. The EPA NERL thanks Michael Jury,
Al Taylor, and Sue Kaelber-Matlock of the Michigan Department of Environmental Quality (MDEQ) and
Becky Goche and Doug Spencer of the U.S. Fish and Wildlife Service for their support in conducting the field
demonstration. We appreciate the support of the Dioxin SITE Demonstration Panel for their technical input to the
demonstration/quality assurance project plan. In particular, we recognize Andy Beliveau, Nardina Turner,
Greg Rudloff, Allen Debus, Craig Smith, David Williams, Dwain Winters, Jon Josephs, Bob Mouringhan,
Terry Smith, and Joe Ferrario of the U.S. EPA. Thanks also go to EPA Region 2, EPA Region 3, EPA Region 4, EPA
Region 5, EPA Region 7, and the MDEQ for collecting and supplying environmental samples for inclusion in the
demonstration. Andy Beliveau, Allen Debus, and Nardina Turner served as EPA reviewers of this report. Michael
Jury (MDEQ), Sue Kaelber-Matlock (MDEQ), Jim Sanborn (California-EPA), and Jeffrey Archer (U.S. Food and
Drug Administration) served as additional reviewers of this report.  Computer Sciences Corporation provided a
technical editing review of the report. This report was prepared for the EPA by Battelle. Special acknowledgment is
given to Amy  Dindal, who was the Battelle Project Manager, and to Josh Finegold, Nicole Iroz-Elardo, Mark Misita,
Tim Pivetz, Mary Schrock, Rachel Sell, Bea Weaver, and Zack Willenberg for their contributions to the preparation
of this report.
                                                  Xll

-------
                                              Chapter 1
                                             Introduction
The U.S. Environmental Protection Agency (EPA),
Office of Research and Development (ORD), National
Exposure Research Laboratory (NERL) contracted with
Battelle (Columbus, Ohio) to conduct a demonstration of
monitoring and measurement technologies for dioxin
and dioxin-like compounds in soil and sediment. A field
demonstration was conducted as part of the EPA
Superfund Innovative Technology Evaluation (SITE)
Monitoring and Measurement Technology (MMT)
Program. The purpose of this demonstration was to
obtain reliable performance and cost data on the
technologies to provide (1) potential users with a better
understanding of the technologies' performance and
operating costs under well-defined field conditions and
(2) the technology developers with documented results
that will help promote the acceptance and use of their
technologies.

This innovative technology verification report (ITVR)
describes the SITE MMT Program and the scope of this
demonstration (Chapter 1); the Abraxis LLC Coplanar
PCB Enzyme-Linked  Immunosorbent Assay (ELISA) kit
(Chapter 2);  the demonstration site and the sampling
locations (Chapter 3); the demonstration approach
(Chapter 4);  the confirmatory process (Chapter 5); the
assessment of reference method data quality (Chapter 6);
the performance of the technology (Chapter 7); the
economic analysis for the technology and reference
method (Chapter 8); the demonstration results in
summary form (Chapter 9); and the references used to
prepare this report (Chapter 10). Appendix A contains a
verification statement; Appendix B contains
supplemental information provided by the developer;
Appendix C  is a summary of method blank and batch
duplicate data by the reference laboratory; and
Appendix D  contains  a one-to-one matching of the
developer and reference laboratory data.
1.1    Description of the SITE MMT Program
Performance verification of innovative environmental
technologies is an integral part of the regulatory and
research mission of the EPA. The SITE Program was
established by the EPA Office of Solid Waste and
Emergency Response and  ORD under the Superfund
Amendments and Reauthorization Act of 1986. The
overall goal of the Program is to conduct performance
verification studies and to  promote the acceptance of
innovative technologies that may be used to achieve
long-term protection of human health and the
environment. The program is designed to meet three
primary objectives: (1) identify and remove obstacles to
the development and commercial use of innovative
technologies, (2) demonstrate promising technologies
and gather reliable performance and cost information to
support site characterization and  remediation activities,
and (3) develop procedures and policies that encourage
use of innovative technologies at Superfund sites as well
as at other waste sites or commercial facilities. The SITE
Program includes the following elements:

    MMT Program—Evaluates technologies that
    sample, detect, monitor, or measure hazardous and
    toxic substances. These technologies are expected to
    provide better, faster, or more cost-effective methods
    for producing real-time data during site characteriza-
    tion and  remediation efforts than conventional
    laboratory technologies.

•   Remediation Technology Program—Conducts
    demonstrations of innovative treatment technologies
    to provide reliable performance, cost, and
    applicability data for site cleanups.

    Technology Transfer Program—Provides and
    disseminates technical information in the form of
    updates,  brochures, and other publications that
    promote  the SITE Program and participating

-------
    technologies. It also supports the technologies by
    offering technical assistance, training, and
    workshops.

The MMT Program's technology verification process is
designed to conduct demonstrations that will generate
high-quality data so that potential users have reliable
information  regarding the technology performance and
cost. Four steps are inherent in the process: (1) needs
identification and technology selection, (2) demon-
stration planning and implementation,  (3) report
preparation, and (4) information distribution. The first
step of the technology verification process begins with
identifying technology needs of the EPA and regulated
community.  The EPA Regional offices, the U.S.
Department  of Energy, the U.S. Department of Defense,
industry, and state  environmental regulatory agencies are
asked to identify technology needs for  sampling,
measurement, and  monitoring of environmental media.
Once a need is identified, a search is conducted to
identify suitable technologies that will  address the need.
The technology search and identification process
consists of examining industry and trade publications,
attending related conferences, and exploring leads from
technology developers and industry experts. Selection of
technologies for field testing includes evaluation of the
candidate technologies based on several criteria. A
suitable technology for field testing

    is designed for use in the field or in a mobile
    laboratory,

•   is applicable to a variety of environmentally
    contaminated sites,

•   has potential for solving problems  that current
    methods cannot satisfactorily address,

•   has estimated costs that are lower than those of
    conventional methods,

    is likely to achieve equivalent or better results than
    current methods in areas such as data quality and
    turnaround time,

•   uses techniques that are easier or safer than current
    methods, and

    is commercially available.

Once candidate technologies are identified, developers
are asked to  participate in a developer conference. This
conference gives the developers an opportunity to
describe their technologies' performance and to learn
about the MMT Program.

The second step of the technology verification process is
to plan and implement a demonstration that will generate
representative, high-quality data to assist potential users
in selecting a technology. Demonstration planning
activities include a pre-demonstration sampling and
analysis investigation that assesses existing conditions at
the proposed demonstration site or sites. The objectives
of the pre-demonstration investigation are to (1) confirm
available information on applicable physical, chemical,
and biological characteristics of contaminated  media at
the sites to justify selection of site areas for the demon-
stration; (2) provide the technology developers with an
opportunity to evaluate the areas, analyze representative
samples, and identify logistical requirements; (3) assess
the overall logistical and quality assurance requirements
for conducting the demonstration; and (4) select and
provide the reference laboratory with an opportunity to
identify any matrix-specific analytical problems
associated with the contaminated media and to propose
appropriate solutions. Information generated through the
pre-demonstration investigation is used to develop the
final demonstration  design and to confirm the nature and
source of samples that will be used in the demonstration.

Demonstration planning activities also include
preparation of a demonstration plan that describes the
procedures to verify the performance and cost  of each
technology. The demonstration plan incorporates
information generated during the pre-demonstration
investigation as well as input from technology
developers, demonstration site representatives, and
technical peer reviewers. The demonstration plan also
incorporates the quality assurance (QA)/quality control
(QC) elements needed to produce data of sufficient
quality to document the performance and cost of each
technology.

During the demonstration, each technology is evaluated
independently and, when possible and appropriate, is
compared to a reference technology. The performance
and cost of one technology are not compared to those of
another technology evaluated in the demonstration.
Rather, demonstration data are used to evaluate the
individual performance, cost, advantages, limitations,
and field applicability of each technology.

-------
As part of the third step of the technology verification
process, the EPA publishes a verification statement
(Appendix A) and a detailed evaluation of each
technology in an ITVR. To ensure its quality, the ITVR
is published only after comments from the technology
developer and external peer reviewers are satisfactorily
addressed. All demonstration data used to evaluate each
technology are summarized in a data evaluation report
(DER) that constitutes a complete record of the
demonstration. The DER includes audit reports, observer
reports, completed data validation checklists, certificates
of analysis, and the data packages (i.e., raw data) from
the reference laboratory. The DER is not published as an
EPA document, but a copy may be obtained from the
EPA project manager.

The fourth step  of the verification process is to distribute
demonstration information. To benefit technology
developers and potential technology users, the EPA
makes presentations, publishes and  distributes  fact
sheets, newsletters, bulletins, and ITVRs through direct
mailings and on the Internet. Information on the SITE
Program is available on the EPA ORD Web site
(http://www.epa.gov/ORD/SITE). Additionally, a
Visitor's Day, which is held in conjunction with the
demonstration, allows the developers to showcase their
technologies and gives potential users the opportunity to
have a firsthand look at the technologies in operation.

1.2     Scope of This Demonstration
Polychlorinated dibenzo-/>-dioxins and polychlorinated
dibenzofurans, commonly referred to collectively as
"dioxins," are of significant concern in site remediation
projects and human health assessments because they are
highly toxic. Dioxins and furans are halogenated
aromatic hydrocarbons and are similar in structure as
shown in Figure 1-1. They have similar chemical and
physical properties. Chlorinated dioxins and  furans are
technically referred to as polychlorinated dibenzo-/?-
dioxins (PCDD) and polychlorinated dibenzofurans
(PCDF). For the purposes of this document, they will be
referred to simply as "dioxins," "PCDD/F," or "D/F."
Dioxins and furans are not intentionally produced in
most chemical processes. However, they can be
synthesized directly and are commonly generated as  by-
products of various combustion and chemical
processes.(1) They are colorless crystals or solids with
high melting points, very low water solubility, high fat
solubility, and low volatility. Dioxins and furans are
extremely stable under most environmental conditions,
making them persistent once released in the
environment. Because they are fat soluble, they also tend
to bioaccumulate.

There are 75 individual chlorinated dioxins and 135
individual chlorinated furans.  Each individual dioxin and
furan is referred to as a congener. The properties of each
congener vary according to the number of chlorine
atoms present and the position where the chlorines  are
attached. The congeners with chlorines attached at a
minimum in the 2,3,7, and 8  positions are considered
most toxic. A total of seven dioxin and 10 furan
congeners contain chlorines in the 2, 3, 7, 8 positions
and, of these, 2,3,7,8-tetrachlorodibenzo-/?-dioxin
(2,3,7,8-TCDD) is one of the most toxic and  serves as
the marker compound for this  class.

Certain polychlorinated biphenyls (PCBs) have
structural and conformational  similarities to dioxin
compounds (Figure 1-1) and are therefore expected to
exhibit toxicological similarities to dioxins as well.
Currently only 12 of the total 209 PCB congeners are
thought to have "dioxin-like" toxicity. These 12 are
     Cl
Cl
      2,3,7,8-Tetrachlorodibenzo-p-dioxin
     Cl
Cl
        2,3,7,8-Tetrachlorodibenzofuran
          Cl
           Cl                       Cl
        3,3',4,4',5,5'-Hexachlorobiphenyl

     Figure 1-1. Representative dioxin, furan,
     and polychlorinated biphenyl structure.

-------
PCBs with four or more chlorines with just one or no
substitution in the ortho position, and which assume a
flat configuration with rings in the same plane. These
"dioxin-like" PCBs are often refered to as non-ortho and
mono-ortho substituted coplanar PCBs.

Conventional analytical methods for determining
concentrations of dioxin and dioxin-like compounds are
time-consuming and costly. For example, EPA standard
methods require solvent extraction of the sample,
processing the extract through multiple cleanup
columns, and analyzing the cleaned fraction by gas
chromatography (GC)/high-resolution mass
spectrometry (HRMS). The use of a simple, rapid, cost-
effective analytical method would allow field personnel
to quickly assess the extent of contamination at a site
and could be used to direct or monitor remediation or
risk assessment activities. This data could be used to
provide immediate feedback on potential health risks
associated with the site and permit the development of a
more focused and cost-effective sampling strategy. At
this time, more affordable and quicker analytical
techniques will not replace HRMS. However, before
adopting an alternative to traditional laboratory-based
methods, a thorough assessment of how commercially
available technologies compare to conventional
laboratory-based analytical methods using certified,
spiked, and environmental samples is warranted. A
summary of the demonstration activities to evaluate
measurement technologies for dioxin and dioxin-like
compounds in soil and sediment is provided below. The
experimental design and demonstration approach are
described in greater detail in Chapter 4 and in the
Demonstration and Quality Assurance Project Plan
(D/QAPP).(2)

1.2.1   Organization of Demonstration
The key organizations and personnel involved in the
demonstration, including the roles and responsibilities of
each, are fully described in the D/QAPP.(2) EPA/NERL
had overall responsibility for this project. The EPA
reviewed and concurred with all project deliverables
including the D/QAPP and the ITVRs, provided
oversight during the demonstration, and participated in
the Visitor's Day. Battelle served as the verification
testing organization for EPA/NERL. Battelle's
responsibilities included developing and implementing
all elements of the D/QAPP; scheduling and
coordinating the activities of all demonstration
participants; coordinating the collection of environ-
mental samples; serving as the characterization
laboratory by performing the homogenization of the
environmental samples and confirming the efficacy of
the homogenization and approximate sample
concentrations; conducting the demonstration by
implementing the D/QAPP; summarizing, evaluating,
interpreting, and documenting demonstration data for
inclusion in this report; and preparing draft and final
versions of each developer's ITVR. The developers were
five companies who submitted technologies for
evaluation during this demonstration. The responsibili-
ties of the developers included providing input to,
reviewing, and concurring with the D/QAPP; providing
personnel and supplies as needed for the demonstration;
operating their technologies during the demonstration;
and reviewing and commenting on their technologies'
ITVRs. AXYS Analytical Services, Ltd. was selected to
serve as the reference analytical laboratory.  AXYS
analyzed each demonstration sample by EPA Method
1613B(3) and EPA Method 1668A(4) according to the
statement of work provided in the D/QAPP. The
Michigan Department of Environmental Quality
(MDEQ)  hosted the demonstration, coordinated the
activities  of and participated in Visitor's Day, and
collected  and provided some of the environmental
samples that were used in the demonstration. The Dioxin
SITE Demonstration Panel served as technical advisors
and observers of the demonstration activities. Panel
membership, which is outlined in the D/QAPP, included
representation from EPA Regions 1, 2, 3, 4, 5, 7, and 9;
EPA Program Offices; the MDEQ; and the U.S. Fish and
Wildlife Services. Members of the panel participated in
five conference calls with the EPA, Battelle, AXYS, and
the developers. The panel contributed to the
experimental design and D/QAPP development; logistics
for the demonstration, including site selection, sample
collection, reference laboratory selection, and data
analysis; and technology evaluation procedures. As an
example of the significant impact the panel had on the
demonstration, it was the EPA members of the panel
who suggested expanding the scope of the project from
focusing exclusively on dioxins and furans, to also
include PCBs and the generation of characterization data
for polynuclear aromatic hydrocarbons (PAHs).

-------
1.2.2  Sample Descriptions and Experimental
       Design
Soil and sediment samples with a variety of
distinguishing characteristics such as high levels of
PCBs and PAHs were analyzed by each participant.
Samples were collected from a variety of dioxin-
contaminated soil and sediment sampling locations
around the country. Samples were identified and
supplied through EPA Regions 2, 3, 4,  5, and 7 and the
MDEQ. The samples were homogenized and
characterized by the characterization laboratory prior to
use in the demonstration to ensure a variety of
homogeneous, environmentally derived samples with
concentrations over a large dynamic range (<50 to
> 10,000 picogram/gram [pg/g]) were included. The
environmental samples comprised 128  of the 209
samples included in the demonstration  (61%).
Performance evaluation (PE) samples were obtained
from five commercial sources. PE samples consisted of
known quantities of dioxin and dioxin-like compounds.
Fifty-eight of the  209 demonstration samples (28%)
were PE samples. A suite of solvent extracts was
included in the demonstration to minimize the impact of
sample homogenization and to provide a uniform matrix
for evaluation.  A total of 23 extracts (11% of the total
number of samples) was included in the demonstration.
The demonstration samples are described in greater
detail in Section 4.3.
1.2.3  Overview of Field Demonstration
All technology developers participated in a pre-
demonstration study where a representative subset of the
demonstration samples was analyzed. The pre-
demonstration results indicated that the Abraxis
technology was suitable for participation in the
demonstration. The demonstration of technologies for
the measurement of dioxin and dioxin-like compounds
was conducted at the Green Point Environmental
Learning Center in Saginaw, Michigan, from April 26 to
May 5, 2004. Five technologies, including immuno-
assay test kits and aryl hydrocarbon (Ah) receptor
binding technologies, participated in the demonstration.
The operating procedures for the participating tech-
nologies are described in the D/QAPP.

The technologies were operated by the developers.
Because the sample throughput of the technologies
varied widely, it was at the discretion of the developers
how many of the 209 demonstration samples were
analyzed in the field. Results from the demonstration
samples,  in comparison with results generated by AXYS
using standard analytical methods, were used to evaluate
the analytical performance of the technologies, including
the parameters of accuracy, precision, and comparability.
Observations from the field demonstration were used to
assess sample throughput, ease of use, health and safety
aspects, and the field portability of each technology. The
performance evaluation of the Abraxis LLC Coplanar
PCB ELISA Kit is presented in this ITVR. Separate
ITVRs have been published for the other four
participating technologies.

-------
                                              Chapter 2
                       Description of Abraxis Coplanar PCB ELISA Kit
This technology description is based on information
provided by Abraxis LLC and only editorial changes
were made to ensure document consistency. Actual cost
and performance data, as reported and observed during
the demonstration, will be provided later in this
document. The Abraxis Coplanar PCB ELISA Kit
(Figure 2-1) applies the principle of enzyme
immunoassays for the qualitative or semiquantitative
analysis of coplanar PCBs in a variety of sample
extracts. Extracts from  soil, sediment, fish tissue, and
other matrices can be exchanged to methanol for ELISA
analysis. Water samples can be diluted 1:1 in methanol
and analyzed directly in the assay.
Figure 2-1. Abraxis Coplanar PCB ELISA Kit.

2.1    Company History
Abraxis LLC is a biotechnology company that develops,
manufactures, and markets products for the
environmental and food testing markets.  The company
was founded in 1998 and has its headquarters in
Warminster, Pennsylvania.  The company's primary
product lines are antibody-based testing kits (ELISA) for
detecting pesticides, algal toxins, endocrine disrupters,
surfactants, antibiotics, and industrial chemicals.

2.2    Product History
For several years, Abraxis has been marketing an ELISA
kit for PCB (Aroclor) analysis. The development of this
Coplanar PCB ELISA kit was a natural progression for
Abraxis, which sensed from its customers the need for
such a product.

2.3    Technology Description
The Abraxis Coplanar PCB ELISA kit can screen
samples according to their PCB toxic equivalency (TEQ)
concentration. The specificity of the test is
predominantly for those congeners with high toxicity
equivalency factor (TEF) values (i.e., congeners 126 and
169). Samples extracted with organic solvents that are
incompatible with ELISA can be evaporated and
redissolved in methanol. For a quick screen of soil and
sediment samples, the samples can be extracted in 20%
acetone in hexane, cleaned up with concentrated sulfuric
acid, evaporated, diluted 1:10 in the provided diluent,
and run directly in the assay.

A solution containing a primary antibody (rabbit) that
reacts with coplanar PCBs is added to a microplate
containing a secondary antibody that captures the
primary antibody. Calibrators (congener 126) and
samples are added and allowed to incubate, followed by
the addition of a coplanar PCB-horseradish peroxidase
(HRP) enzyme conjugate. Any coplanar PCBs that may
be in the sample compete with the coplanar PCB
enzyme-labeled  conjugate for a finite number of
antibody binding sites. At the end of the incubation
period, the unbound conjugate is removed, and the plate
is washed. A substrate/chromogen solution is then added
and enzymatically converted from a colorless to a blue
solution by the captured coplanar PCB-HRP conjugate
on the plate. The reaction is then terminated by
acidification. The coplanar PCB concentration is
determined by measuring the absorbance [at 450
nanometer (nm)] of the sample solution using a
microplate reader (see Figure 2-2) and comparing it to
the absorbance of the calibrators. The amount of color
             Information was provided by the developer and does not necessarily reflect the opinion of the EPA.
                                                    6

-------
 Figure 2-2.  Microplate reader used by Abraxis during
 the field demonstration.
produced is inversely proportional to the amount of
coplanar PCBs present in the sample.

The final value measured by ELISA is the sum of the
various congeners responses. This value approximates
TEQPCB because the immunoassay kit cross-reaction
profile, shown in Table 2-1, for coplanar PCBs
approximates TEF values. The cross-reactivity of the
Abraxis coplanar PCB assay for various congeners and
Aroclors can be expressed as the least detectable dose
(LDD) which is estimated at 90% B (mean absorbance
obtained with the standard)/Bo (mean absorbance value
for the zero standard), or as the dose required for the
50% absorbance inhibition (50% B/Bo). The following
compounds demonstrated no reactivity in the Abraxis
coplanar PCB assay at concentrations up to 1,000 ppb:
aldicarb, aldicarb sulfoxide, aldicarb sulfone, alachlor,
atrazine, benomyl, butachlor, butylate, captan, carbaryl,
carbendazim, carbofuran, 2,4-D, 1,3-dichloropro-pene,
dinoseb, 4-chloro-2-methylphenoxy)acetic acid,
metolachlor, metribuzin, pentachloro-phenol, picloram,
propachlor, terbufos, thiabendazole, and thiophanate-
methyl. Accuracy among samples may vary because of
the variability of congener composition. To help
maximize accuracy, the variability of congener
composition in the target sample should be known.

The primary use of the Abraxis Coplanar PCB ELISA
Kit is to screen samples that have low coplanar PCB
concentrations. The sensitivity of the test is claimed by
Abraxis to be 4 parts per trillion (ppt) in water samples
and 6.25 pg/g in soil or sediment samples. These values
are related to the original sample concentration by using
the appropriate dilution and volume factors. Detection
levels depend on how much sample is evaporated and
the volume of solvent used to resuspend the sample.
Matrix detection limits will vary according to the matrix
being analyzed, sample size, and dilution factor. Up to
100 samples per day can be analyzed using the
procedure described.

The Abraxis Coplanar PCB ELISA Kit consists of

1.   Microtiter Plate coated with Goat-Antirabbit
    Antibody
    96-well test kit: 8X12 strips

2.   Coplanar PCB Antibody Solution
    Rabbit  anti-coplanar PCB solution in a colored
    buffered saline solution with preservative and
    stabilizers.
    96-well test kit: one 6-milliliter (mL) vial

3.   Coplanar PCB Standards (Congener 126)
    Seven concentrations (0, 25, 50, 100, 250, 500, and
    1,000 ppt) in 50% methanol.
    96-well test kit: one 1-mLvial

4.   Coplanar PCB-HRP Enzyme Conjugate
    Coplanar PCB labeled with horseradish peroxidase
    diluted in colored buffered solution with
    preservative and stabilizers.
    96-well test kit: one 6-mL vial

5.   Diluent/Zero Standard
    50% methanol in distilled water (v/v) without any
    detectable PCB.
    96-well test kit: one 30-mL vial

6.   Color Solution
    A solution of hydrogen peroxide and
    3,3',5,5'-tetramethylbenzidine in an organic base.
    96-well test kit: one 16-mL vial

7.   Stopping Solution
    A solution of diluted acid.
    96-well test kit: one 6-mL vial
             Information was provided by the developer and does not necessarily reflect the opinion of the EPA.

-------
Table 2-1.  Cross-Reactivities for the Abraxis Coplanar PCB ELISA Kit
Compound
PCB Congener 77
PCB Congener 81
PCB Congener 105
PCB Congener 114
PCB Congener 118
PCB Congener 123
PCB Congener 126
PCB Congener 156
PCB Congener 157
PCB Congener 167
PCB Congener 169
PCB Congener 170
PCB Congener 180
PCB Congener 189
Aroclor 1016
Aroclor 1056
Aroclor 1221
Aroclor 1242
Aroclor 1248
Aroclor 1254
Aroclor 1262
Aroclor 1268
Biphenyl
Least Detectable Dose
(LDD) (pg/g)
300
700
3000
5000
26,000
2,200
14
3,300
10,000
3,000
4
50% absorbance inhibition
(50%B/Bo) (pg/g)
5,100
10,000
400,000
115,000
240,000
270,000
270
500,00
140,000
54,000
90
TEF
0.0001
0.0001
0.0001
0.0005
0.0001
0.0001
0.1
0.0005
0.0005
0.00001
0.01
NR
NR
700
540,000
90,000
120,000
7,000
70,000
100,000
90,000
120,000
9,000
4,000,000
3,200,000
4,200,000
480,000
440,000
1500,000
10,000,000
40,000,000
0.0001
NA
NA
NA
NA
NA
NA
NA
NA
NR
NR = nonreactive up to 1,000,000 pg/g.
NA = not applicable.

8.  Washing Buffer 5X Concentrate
    Buffer salts with detergent and preservatives.
    96-test kit: one 100-mLvial
     Phone:(215)357-3911
     Email: frubio@abraxiskits.com
     Web site: www.abraxiskits.com.
This is the method that Abraxis implemented during the
field demonstration. A photo of the technology in
operation during the demonstration is presented in
Figure 2-3. Abraxis provided supplemental information
about the performance of its technology during the
demonstration and it is presented in Appendix B.

2.4  Developer Contact Information
Additional information about this technology can be
obtained by contacting:
        Fernando  Rubio
        Abraxis LLC
        54 Steamwhistle Drive
        Warminster, Pennsylvania 18974
Figure 2-3. Abraxis processing samples during the
field demonstration
             Information was provided by the developer and does not necessarily reflect the opinion of the EPA.

-------
                                              Chapter 3
                     Demonstration and Environmental Site Descriptions
This chapter describes the demonstration site, the
sampling locations, and why each was selected.

3.1    Demonstration Site Description and
       Selection Process
This section describes the site selected for hosting the
demonstration, along with the selection rationale and
criteria. Several candidate host sites were considered.
The candidate sites were required to meet certain
selection criteria, including necessary approvals,
support, and access to the demonstration site; enough
space and power to host the technology developers, the
technical support team, and other participants; and
various levels of dioxin-contaminated soil and/or
sediment that could be analyzed as part of the
demonstration. Historically, these demonstrations are
conducted at sites known to be contaminated with the
analytes of interest. The visibility afforded the sites is a
valuable way of keeping the local community informed
of new technologies and to help promote the EPA's
commitment to promote and advance science and
communication.

After review of the information available, the site
selected for the demonstration was the Green Point
Environmental Learning Center (ELC) site, located
within the city of Saginaw, Michigan. The Saginaw
city-owned, 76-acre Green Point ELC, formerly known
as the Green Point Nature Center, is managed by the
Shiawassee National Wildlife Refuge. The Green Point
ELC is situated within the Tittabawassee River flood
plain. The MDEQ found higher than normal levels of
dioxins in soil and sediment samples taken from the
flood plain of the Tittabawassee River. The flood plain is
not heavily laden with PCBs; however, low levels of
PCBs have been detected in some areas. Soil samples
taken from areas outside the flood plain were at typical
background levels. The source of the contamination was
speculated to be attributed to legacy contamination from
chemical manufacturing.

To summarize, Green Point ELC was selected as the
demonstration site based on the following criteria:

• Access and Cooperation of the State and Local
  Community—Representatives from the MDEQ, EPA
  Region 5, and the local U.S. Fish and Wildlife
  Services supported the demonstration by providing site
  access for the demonstration, logistical support for the
  demonstration, and supported a Visitor's Day during
  the demonstration.

• Space Requirements and Feasibility—The demon-
  stration took place in the parking lot adjacent to the
  Green Point ELC, not directly on an area of
  contamination. The  site had electrical power and
  adequate space to house the trailers and mobile labs
  that were used for the demonstration. Furthermore, the
  site was close to an international airport. The weather
  in Michigan at the time of the demonstration was
  unpredictable; however, all participants were provided
  heated containment (a mobile laboratory or
  construction trailer).

• Site Diversity—The area encompassing the Green
  Point site had different levels and types of dioxin
  contamination in both the soil and sediment that were
  used to evaluate the performance of the technologies.

The demonstration was conducted at the Green Point
ELC over a 10-day period from April 26 to May 5, 2004.
All technologies were operated inside trailers equipped
with fume hoods  or inside mobile laboratories. As such,
the ambient weather conditions during the demonstration
had little impact on the operation of the technologies,

-------
since all of the work spaces were climate-controlled with
heat and air conditioning. The outdoor weather
conditions were generally cool and rainy, but the
developers kept their working environment at
comfortable temperatures (16 to 18°C).  The low
temperature over the 10-day demonstration period was
2°C, the high temperature was 26°C, and the average
temperature was 9°C. Precipitation fell on eight of the
10 days, usually in the form of rain, but occasionally as
sleet or snow flurries, depending on the temperature. The
largest amount of precipitation on a given demonstration
day was 0.50 inches.

3.2    Description of Sampling Locations
This section provides an overview of the 10 sampling
sites and methods of selection. Table 3-1 summarizes
each of the locations, what type of sample (soil or
sediment) was provided, the number of samples
submitted from  each location, and the number of
samples included in the demonstration from each
location.  Samples were collected from multiple
sampling sites so that a wide variety of matrix conditions
could be used to evaluate the performance of the
technologies in addressing monitoring needs at a diverse
range of Superfund sites.

Samples consisted of either soil or sediment and are
described below based on this distinction. It should be
noted that it was not an objective of the  demonstration to
accurately characterize the concentration of dioxins,
furans, and PCBs from a specific sampling site. It was,
however, an objective to ensure comparability between
technology samples and the reference laboratory
samples. This was accomplished by homogenizing each
matrix, such that all sub-samples of a given matrix had
consistent contaminant concentrations. As a result,
homogenized samples were not necessarily repre-
sentative of original concentrations at the site.

3.2.1   Soil Sampling Locations
This section provides descriptions of each of the soil
sampling locations, including how the sites became
contaminated and approximate dioxin concentrations, as
well as the type  and concentrations of other major
constituents, where known [such as PCBs,
pentachlorophenol (PCP), and PAHs]. This information
was provided by the site owners/sample providers (e.g.,
the EPA, EPA contractors, and the MDEQ).

3.2.1.1 Warren County, North Carolina
Five areas of the Warren County PCB Landfill in North
Carolina, a site with both PCB and dioxin
contamination, were sampled. Dioxin concentrations in
the landfill soils range approximately from 475 to
700 pg/g, and PCB concentrations are greater than
100 parts per million (ppm). The Warren County PCB
Landfill contains soil that was contaminated by the
illegal spraying of waste transformer oil containing
PCBs from over 210 miles of highway shoulders. Over
30,000 gallons of contaminated oil were disposed of in
14 North Carolina counties. The landfill is located on a
142-acre tract of land. The EPA permitted the landfill
under the Toxic Substances Control Act. Between
September and November 1982, approximately
40,000 cubic yards (equivalent to 60,000 tons) of PCB-
contaminated soil were removed and hauled to the newly
constructed landfill located in Warren County, North
Carolina. The landfill is equipped with both polyvinyl
chloride and clay caps and liners. It also has a dual
leachate collection system. The material in the landfill is
solely from the contaminated roadsides. The landfill was
never operated as a commercial facility. The remedial
action was funded by the EPA and the State of North
Carolina. The site was deleted from the National
Priorities List on March 7, 1986.

3.2.1.2 Tittabawassee River Flood Plain
The MDEQ sampled the Tittabawassee River flood plain
soils from three sites in the flood plain. The source of the
contamination was speculated to be attributed to legacy
contamination from chemical manufacturing. Two
samples were collected from two locations at Imerman
Park in Saginaw Township. The first sample was taken
near the boat launch, and the second sample was taken in
a grassy area near the  river bank. Previous analysis from
these areas of this park indicated a range of PCDD/F
concentrations  from 600 to 2,500 pg/g. Total PCBs from
these previous measurements were in the low ppt range.
Two samples were collected from two locations at
                                                   10

-------
Table 3-1.  Summary of Environmental Sampling Locations
Sample Type
Soil
Sediment
Sampling Location
Warren County, North Carolina
Tittabawassee River, Michigan
Midland, Michigan
Winona Post, Missouri
Solutia, West Virginia
Newark Bay, New Jersey
Raritan Bay, New Jersey
Tittabawassee River, Michigan
Saginaw River, Michigan
Brunswick, Georgia
Total
Number of Samples
Submitted for Consideration
5
6
6
6
6
6
6
6
6
5
58
Included in Demonstration
3
o
J
4
3
3
4
3
o
3
o
3
3
32
Freeland Festival Park in Freeland, MI. The first sample
was taken above the river bank, and the second sample
was taken near a brushy forested area within the park
complex. Previous PCDD/F concentrations were from
300 to 3,400 pg/g, and total PCBs were in the low ppt
range. The final two samples were collected from
Department of Natural Resources (DNR)-owned
property in Saginaw, which was formerly a farming area
located almost at the end of the Tittabawassee River
where it meets the Shiawassee River to form the
Saginaw River. Previous PCDD/F concentrations ranged
from 450 to 1,150 pg/g. Total PCBs were not previously
analyzed, but concentrations were expected to be less
than 1 ppm. The DNR property is approximately a
10-minute walk from where the demonstration was
conducted at the Green Point ELC.

3.2.1.3 Midland, Michigan
Soil samples were collected by the MDEQ from various
locations in Midland, Michigan. The soil type and nature
of dioxin contamination are different in the Midland
residential area than it is on the Tittabawassee River
flood plain, but it is from the same suspected source
(legacy contamination from chemical manufacturing).
Samples were collected in various locations around
Midland. Estimated TEQ concentrations ranged from 10
pg/g to 1,000 pg/g.

3.2.1.4 Winona Post
The Winona Post site in Winona, Missouri, was a
Superfund cleanup of a wood treatment facility.
Contaminants at the site included PCP, dioxin, diesel
fuel, and PAHs. Over a period of at least 40 years, these
contaminants were deposited into an on-site drainage
ditch and sinkhole. Areas of contaminant deposition
(approximately 8,500 cubic yards of soils/sludge) were
excavated in late 200I/early 2002.  This material was
placed into an approximate 2!/2-acre treatment cell
located on facility property.  During 2002/2003, material
at the treatment cell was treated through addition of
amendments (high-ammonia fertilizer and manure) and
tilling. Final concentrations achieved in the treatment
cell averaged 26 milligrams  (mg)/kg for PCP from 8,000
to 10,000 for pg/g dioxin equivalents.  Samples obtained
for this study from this site were obtained from the
treatment cell after these concentrations had been
achieved.

3.2.1.5  Solutia
The chemical production facility at the Solutia site in
Nitro, West Virginia, is located along the eastern bank of
the Kanawha River, in Putnam County, West Virginia.
The site has been used for chemical production since the
early 1910s. The initial production facility was
developed by the U.S. government for the production of
military munitions during the World War I era between
1918 and 1921. The facility was then purchased by a
small private chemical company, which began manu-
facturing chloride, phosphate, and phenol compounds at
the site. A major chemical manufacturer purchased the
facility in 1929 from Rubber Services Company. The
company continued to expand operations and accelerated
its growth in the 1940s. A variety of raw materials has
been used at the facility over the years, including
                                                    11

-------
inorganic compounds, organic solvents, and other
organic compounds, including Agent Orange.
Agent Orange is a mixture of chemicals containing
equal amounts of two herbicides: 2,4-D
(2,4 dichlorophenoxyacetic acid) and 2,4,5-T (2,4,5
trichlorophenoxyacetic acid). Manufacture of this
chemical herbicide began at the site in  1948 and ceased
in 1969. The source of the dioxin contamination in the
site soils was associated with the manufacture of 2,4,5-T,
where dioxins are an unintentional by-product. The site
has a dioxin profile from ppt to low ppb range. No PCBs
or PAHs were identified in the soil.

3.2.2  Sediment Sampling Sites
This  section provides descriptions of each of the
sediment sites that includes how the sites became
contaminated and approximate dioxin concentrations, as
well  as the type and concentrations of other major
constituents (such as PCBs, PCP, and PAHs). This infor-
mation was provided from site owners/samples providers
(e.g., the EPA, EPA contractors, and the MDEQ).

3.2.2.1 New York/New Jersey Harbors
Dredged materials from the New York  and New Jersey
harbors were provided as samples for the demonstration.
The U.S. Army Corps of Engineers, New York District,
and EPA Region 2 are responsible for managing dredged
materials from the New York and New Jersey harbors.
Dioxin levels affect the disposal options for dredged
material. Dredged materials are naturally occurring
bottom sediments, but some in this area have been
contaminated with dioxins and other compounds by
municipal or industrial wastes or by runoff from
terrestrial sources such as urban areas or agricultural
lands.

3.2.2.1.1 Newark Bay
Surrounded by manufacturing industries, Newark Bay is
a highly contaminated area with numerous sources
(sewage treatment plants, National Pollutant Discharge
Elimination System discharges, and nonpoint sources).
This  bay is downstream from a dioxin Superfund site
that contains some of the highest dioxin concentrations
in the United States and also is downstream from a
mercury Superfund site. The dioxin concentration in the
area  sampled for this demonstration was approximately
450 pg/g. Average PCB concentrations ranged from 300
to 740 ppb. Fine-grained sediments make up 50% to
90% of the dredged material. Average total organic
carbon (TOC) was about 4%.

3.2.2.1.2 RaritanBay
Surrounded by industry and residential discharges,
Raritan Bay has dioxin contamination in the area, but it
is not to the degree of Newark Bay. No major Superfund
sites are located in the vicinity. Dioxin concentration
should be significantly less than in Newark Bay. PCB
concentrations are around 250 ppb. The fine-grained
sediment and TOC values were similar to percentages in
Newark Bay.

3.2.2.2  Tittabawassee River
The first Tittabawassee River location was
approximately %-mile upstream of the Bob Caldwell
Boat launch in Midland, Michigan. The sediments are
dark gray, fine sand with some silt. The estimated TEQ
concentration was 260  pg/g; however, concentrations as
high as 2,100 pg/g TEQ have been found in this area.
The second site was on the Tittabawassee River
approximately 100 yards downstream from old Smith's
Crossing Bridge in Midland, Michigan. The sediment
was brown and sandy with organic material. The
estimated TEQ concentration was 870 pg/g; but, again,
concentrations as high  as 2,100 pg/g TEQ are possible in
the area. The third site  was on Tittabawassee River at the
Emerson Park Golfside Boat Launch. The sediment was
gray black silty sand, with many leaves and high organic
matter. The estimated TEQ concentration was < 5 pg/g.
The fourth site was on the Tittabawassee River adjacent
to Imerman Park in Saginaw County across from the
fishing dock. The sediment was sand with some silt. The
estimated TEQ concentration was between 100 and
2,000 pg/g TEQ. The fifth site was on the Tittabawassee
River approximately 1 mile downstream of Center Road
Boat Launch in Saginaw Township. The sediment
consisted of sand and gravel with some shells and not
much organic matter. The estimated TEQ concentration
was between 100 and 1,000 pg/g TEQ. The sixth site
also was on the Tittabawassee River across from the
Center Road Boat Launch. The sediment was fine sand
with high organic matter. The estimated TEQ
concentration was 1,000 pg/g TEQ.  The source of the
contamination was speculated to be attributed to legacy
contamination from chemical manufacturing.
                                                   12

-------
3.2.2.3 Saginaw River
Saginaw River samples were collected at six locations.
The first sampling location was in the Saginaw River
just downstream of Green Point Island. Samples were
collected near the middle of the river in about 21 feet of
water. The sample was granular with some organic
material. The estimated TEQ concentration was 100 ppt.
Another Saginaw River sample was taken upstream of
Genesee Bridge on the right side of the river. The sample
was a brown fine sand from about 15 feet of water. The
estimated TEQ concentration was 100  ppt. The third
location was in the Saginaw River downstream of the
Saginaw wastewater treatment plant in about eight feet
of water. The sample was gray silty clay with an
unknown TEQ concentration. The fourth location was in
the Saginaw River in about eight feet of water. The
sample was a black sandy material. The estimated TEQ
concentration for this location was unknown. The fifth
location was downstream of a petroleum pipeline
crossing upstream of the Detroit and Mackinaw railroad
bridge crossing. This location was selected because of its
proximity to a former PCB dredging location. The
sediment sample consisted of dark black silt with some
sand. The estimated TEQ concentration was unknown,
but PCB concentrations are expected to be high. The
sixth and final sampling location was near the mouth of
the Saginaw River in about five feet of water. The
sediment was a mix of fine black silt and layers of sand
and shells. The estimated TEQ concentration for this
location was also unknown.
3.2.2.4 Brunswick Wood Preserving Site
The Brunswick Wood Preserving Superfund site is
located in Glynn County, Georgia, north of the city of
Brunswick. The site was originally located in the city of
Brunswick, but moved to its present location around
1958. The site is approximately 84 acres and is about
two-thirds of a mile long. Burnett Creek, a tidally
influenced stream, is located at the western corner of the
site. At several points, most, if not all, of the drainage
from the site flows into Burnett Creek. The site was first
operated by American Creosote Company, which
constructed the facility sometime between 1958 and
1960. The site was acquired by Escambia Treating
Company in 1969 from Georgia Creosoting Company
and the Brunswick Creosoting Company. In  1985, a
corporate reorganization resulted in the purchase of the
facility by the Brunswick Wood Preserving Company,
which operated the site until it closed in early 1991.
Each of the three major wood-treating operations was
carried out at the facility: PCP, creosote, and
chromium-copper-arsenic (CCA). The site was listed on
the EPA's National Priorities List on April 1, 1997.

Sediment samples from the Brunswick Wood Preserving
site in Brunswick, Georgia, were collected from six
locations on the site, including areas thought to have
lower (< 300 pg/g TEQ) and higher (> 10,000 pg/g
TEQ) dioxin/furan concentrations. Due to the processes
that occurred on this site, the samples also contain
varying levels of PAHs and PCP, but they were not
expected to contain PCBs.
                                                    13

-------
                                               Chapter 4
                                     Demonstration Approach
This chapter discusses the demonstration objectives,
sample collection, sample homogenization, and
demonstration design.

4.1    Demonstration Objectives
The primary goal of the SITE MMT Program is to
develop reliable performance and cost data on innovative,
commercial-ready technologies. A SITE demonstration
must provide detailed and reliable performance and cost
data so that technology users have adequate information
to make sound decisions regarding comparability to
conventional methods. The demonstration had both
primary and  secondary objectives. Primary objectives
were critical to the technology evaluation and required
the use of quantitative results to draw conclusions
regarding a technology's performance. Secondary
objectives pertained to information that is useful to know
about the technology but did not require the use of
quantitative results to draw conclusions regarding a
technology's performance.

The primary objectives for the demonstration of the
participating technologies were as follows:

P1.    Determine the accuracy.
P2.    Determine the precision.
P3.    Determine the comparability of the technology to
       EPA standard methods.
P4.    Determine the method detection limit (MDL).
P5.    Determine the frequency of false positive and
       false negative results.
P6.    Evaluate the impact of matrix effects on
       technology performance.
P7.    Estimate costs associated with the operation of
       the technology.

The secondary objectives for the demonstration of the
participating technologies were as follows:
S1.    Assess the skills and training required to
       properly operate the technology.
S2.    Document health and safety aspects associated
       with the technology.
S3.    Evaluate the portability of the technology.
S4.    Determine the sample throughput.

Application of these objectives to the demonstration was
addressed based on input from the Dioxin SITE
Demonstration Panel members,(2) general user
expectations of field measurement technologies, the time
available to complete the demonstration, technology
capabilities that the developers participating in the
demonstration intend to highlight, and the historical
experimental components of former SITE Program
demonstrations to maintain consistency.

Note that this demonstration does not assess all
parameters that can affect performance of the
technologies in comparison to the reference methods
(i.e., not all compounds have been characterized in the
test samples, calibration of technologies results to
HRMS results on site-by-site basis was not evaluated,
etc.). However, the demonstration as outlined below was
agreed upon by the Dioxin SITE Demonstration Panel
members to provide a reasonable evaluation of the
technologies.

4.2    Toxicity Equivalents
For risk assessment purposes, estimates of the toxicity of
samples that contain a mixture of dioxin, furan, and PCB
congeners are often expressed as TEQs. TEQs are
calculated by multiplying the concentration of each
congener with a TEF, according to the equation:
                  TEQ = CC  *TEF

where Cc is the concentration of the congener. The TEF
(see Table 4-1) provides an equivalency factor for each
                                                    14

-------
Table 4-1.  World Health Organization Toxicity Equivalency Factor Values
Compound'3'
PCDDs
2,3,7,8-TCDD
1,2,3,7,8-PeCDD

1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD

1,2,3,4,6,7,8-HpCDD
555555 f
OCDD
Dioxin-like PCBs
Coplanar
3,3',4,4'-TCB (PCB 77)
3,4,4',5-TCB (PCB 81)
3,3',4,4',5-PeCB (PCB 126)
3,3',4,4',5,5'-HxCB (PCB 169)




WHO TEF

1
1

0.1
0.1
0.1

0.01

0.0001


0.0001
0.0001
0.1
0.01




Compound
PCDFs
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
555555 f
1,2,3,4,7,8,9-HpCDF
OCDF

mono-ortho
2,3,3',4,4'-PeCB (PCB 105)
2,3,4,4',5-PeCB (PCB 114)
2,3',4,4',5-PeCB (PCB 118)
2,3,4,4',5-PeCB (PCB 123)
2,3,3',4,4',5-HxCB (PCB 156)
2,3,3',4,4',5-HxCB (PCB 157)
2,3',4,4',5,5'-HxCB (PCB 167)
2,3,3',4,4'5,5'-HpCB (PCB 189)
WHO TEF

0.1
0.05
0.5
0.1
0.1
0.1
0.1
0.01
0.01
0.0001


0.0001
0.0005
0.0001
0.0001
0.0005
0.0005
0.00001
0.0001
    T = Tetra, Pe = Penta, Hx = Hexa, Hp = Hepta, O = Octa, CDD = chlorinated dibenzo-/>-dioxin, CDF = chlorinated
    dibenzofuran, CB = chlorinated biphenyl
congener's toxicity relative to the toxicity of 2,3,7,8-
TCDD. The TEFs used in this demonstration were
determined by the World Health Organization (WHO)
for mammalian species.(5) The total TEQ from dioxin
and furans (TEQD/F) in a sample is calculated by adding
up all of the TEQ values from the individual dioxin and
furan congeners. The total TEQ contribution from PCBs
(referred to as TEQPCB) is calculated by summing up the
individual PCB TEQ values. The total TEQ in a sample
is the sum of the TEQD/F and TEQPCB values. TEQ
concentrations for soils and sediments are typically
reported in pg/g, which is equivalent to ppt.

Concentrations of dioxins, furans, and PCBs,
represented as total TEQ concentration, provide a
quantitative estimate of toxicity for all congeners
expressed as if the mixture were a TEQ mass of 2,3,7,8-
TCDD only. While the TEQ concept provides a way to
estimate potential health or ecological effects, the
limitations of this approach should be understood. The
WHO report noted that the TEF indicates an order of
magnitude estimate of the toxicity of a compound
relative to 2,3,7,8-TCDD.(5) Therefore, the accuracy of
the TEF factors could be affected by differences in
species, in the functional responses elicited by the
compounds, and in additive and nonadditive effects
when the congeners are present in complex mixtures.
The WHO report(5) concluded, however, that it is
unlikely that a significant error would be observed due
to these differences. The larger impact to the TEF
concept is the presence of Ah receptor binding
compounds, such as PAHs (including naphthalenes,
anthracenes, and fluorenes) and brominated and
chloro/bromo-substituted analogues of PCDD/Fs that
have not been assigned TEF values but which may
contribute to the total TEQ. This potentially can result
in an underestimation of TEQs in environmental samples
using the TEF approach.(5)

This demonstration was designed with these limitations
of the TEQ concept in mind.  The samples chosen
contained a variety of combinations of dioxins, furans,
and PCBs and at a wide range of concentration levels.
Some samples were high in analytes with better
understood TEFs,  while others were high in analytes
                                                    15

-------
with TEFs that have more uncertainty. Some were high
in other Ah-receptor binding compounds such as PAHs,
while still others were free of these possible TEQ
contributing compounds. The purpose was to evaluate
each of the technologies under a variety of conditions
and assess the comparability of the TEQD/F and TEQPCB
values determined by the reference laboratory.

4.3    Overview of Demonstration Samples
The goal of the demonstration was to perform a detailed
evaluation of the overall performance of each
technology for use in the field or mobile environment.
The demonstration objectives were centered around
providing performance data that support action levels for
dioxin at contaminated sites. The Centers for Disease
Control's Agency for Toxic Substances and Disease
Registry (ATSDR) has established a decision framework
for sites that are contaminated with dioxin and dioxin-
like compounds.(6) If samples are determined to have
dioxin TEQ levels between 50  and 1,000 pg/g, the site
should be further evaluated; action is recommended for
levels above 1,000 pg/g (i.e., 1 ppb) TEQ. A mix of PE
samples,  environmentally contaminated ("real-world")
samples,  and extracts were evaluated that bracket the
ATSDR guidance levels. Table 4-2 lists the primary and
secondary performance objectives for this demonstration
and which sample types were used in each evaluation.
The PE samples were used primarily to determine the
accuracy of the technology and consisted of purchased
soil and sediment standard reference materials with
certified concentrations of known contaminants and
newly prepared spiked samples. The PE samples also
were used to evaluate precision, comparability, MDL,
false positive/negative results, and matrix effects.
Environmentally contaminated samples were collected
from dioxin-contaminated sites around the country and
were used to evaluate the precision, comparability, false
positive/negative results, and matrix effects. Extracts,
prepared in toluene, which was the solvent used by the
reference laboratory, were used to evaluate precision,
MDL, and matrix effects. All samples were used to
evaluate qualitative performance objectives such as
technology  cost, the required skill level of the operator,
health and safety aspects, portability, and sample
throughput. Table 4-3 shows the number of each sample
type included in the experimental design. The following
sections describe each sample type in greater detail.

4.3.1  PE Samples
PE standard reference materials are available through
Cambridge  Isotope  Laboratories (CIL)  (Andover,
Massachusetts), LGC Promochem (United Kingdom),
Wellington  Laboratories (U.S. distributor TerraChem,
Shawnee Mission, Kansas), the National Institute of
Standards and Technology (NIST) (Gaithersburg,
Maryland),  and Environmental Resource Associates
(ERA, Arvada, Colorado). All of these sources were
utilized to obtain PE samples for use in this
demonstration, and  Table 4-4 summarizes the PE
samples that were included. PE samples consisted of
Table 4-2. Distribution of Samples for the Evaluation of Performance Objectives
Performance Objective
P 1 : Accuracy
P2: Precision
P3: Comparability
P4: MDL
P5: False positive/negative results
P6: Matrix effects
P7: Cost
SI : Skill level of operator
S2: Health and safety
S3: Portability
S4: Sample throughput
Sample Type Used in Evaluation
PE
PE, environmental, extracts
PE, environmental, extracts
PE, extracts
PE, environmental, extracts
PE, environmental, extracts
PE, environmental, extracts
PE, environmental, extracts
PE, environmental, extracts
PE, environmental, extracts
PE, environmental, extracts
                                                    16

-------
Table 4-3. Number and Type of Samples Analyzed in the Demonstration
Sample Type
PE
Environmental
Extracts
Total number of samples per technology
No. of Samples
58
128
23
209
Table 4-4. Summary of Performance Evaluation Samples
Sample
Type
ID
PE#1
PE#2
PE#3
PE#4
PE#5
PE#6
PE#7
PE#8
PE#9
PE#10
PE#11
PE#12
Source
CIL
LGC
Promochem
Wellington
CIL
NIST
ERA
ERA
ERA
ERA
ERA
ERA
ERA
PE Type
Certified
Certified
Certified
Certified
Certified
Spiked
Spiked
Spiked
Spiked
Spiked
Spiked
Organic,
Semivolatile,
Blank Soil
Product
No.
RM5183
CRM 529
WMS-01
RM5184
SRM 1944
custom
custom
custom
custom
custom
custom
056
(lot 56011)
Certified Concentration
TEQD/F
(Pg/g)
3.9
6,583
62
171
251
11
33
NS
NS
NS
11
0.046
TEQPCB
(Pg/g)
5.0
424C
10.5
941
41°
NSf
NS
NS
11
1,121
3,760C
0.01
PAH
(mg/kg)
0.18
NAd
NA
27
2.4e
0.33
<0.33
61g
<0.33
<0.33
<0.33
<0.33
Total Number ofPE samples
Correlation
to Environ.
Sample
Type D)a
6
5
6
2,8,9
3,4
10
10
5,7
1
1
1
not
applicable

No. of
Replicates
Per
Sample
7b
4
7b
4
4
4
4
4
4
4
4
8
58
  Environmental Sample IDs are provided in Table 4-5.
  Seven replicates were analyzed for MDL evaluation.
  Little or no certified PCB data was available; mean of reference laboratory measurements was used.
  NA = no data available.
  Approximate concentration of 2-methyl naphthalene, acenaphthene, and fluorene, which were the only PAHs that were included in the
  analysis.
  NS = not spiked.
  Each of the 18 target PAHs was spiked at levels that ranged from 1 to 10 mg/kg. (See Section 5.2.3 for the list of 18 PAHs.)
three types of samples: (1) reference materials (RMs) or
certified samples, which included soil and/or sediment
samples with certified concentrations of dioxin, furan,
and/or PCBs; (2) spiked samples, which included a
certified dioxin, furan, PCB, and PAH-clean matrix
spiked with known levels of dioxin and/or other
contaminants; and (3) blank samples that were certified
to have levels of dioxins, furans, WHO PCBs, and PAHs
that were nondetectable or were considerably lower than
the detection capabilities of developer technologies.  The
PE samples were selected based on availability and on
the correlation of the PE composition as it related to the
environmental samples that were chosen for the
demonstration (e.g., the PE sample had a similar
congener pattern to one or more of the environmental
sites). Table 4-4 indicates a correlation between the
composition of the PE sample and the samples from the
environmental sites, where applicable.  The certified
samples only required transfer from the original jar to
the demonstration sample jar. The spiked samples were
shipped to the characterization laboratory in bulk
quantities so each had to  be aliquoted in 50-g quantities.
Additional details about each source of PE sample are
provided further in this section.
                                                      17

-------
4.3.1.1 Cambridge Isotopes Laboratories
Two RMs were obtained from CIL for use in this
demonstration. RM 5183 is a soil sample that was
collected from a location in Texas with the intended
purpose of serving as an uncontaminated soil for use as a
spiking material. The soil was sieved to achieve uniform
particle size and homogenized to within 5% using a
disodium fluorescein indicator.  Samples were then
sterilized three times for 2 hours at 121°C and 15 pounds
per square inch (psi). Analytical results indicated that the
soil had low levels of D/F and PCBs.

RM 5184 is a heavily-contaminated soil sample with
relatively high levels of D/F and PCBs.  According to
the Certificate of Analysis (CoA), approximately
75 kilograms (kg) of contaminated sediment were
obtained from an EPA Superfund site in Massachusetts
that was known to contain considerable contamination
from PCBs and other chemical pollutants. The sediment
was sieved to achieve uniform particle size and
homogenized to within 5% using a disodium fluorescein
indicator. Samples were then sterilized three times for
2 hours at 121 °C and  15 psi.

RM 5183 and RM 5184 are newly available SRMs from
CIL. For both RM 5183 and RM 5184, certified
analytical values are provided for the D/F and the
12 WHO PCB congeners. The samples were included in
an international interlaboratory study conducted by CIL
and Cerilliant Corporation. More than 20 laboratories
participated in analysis of the D/Fs; up to 20 laboratories
participated in the analysis of the PCBs. Participating
laboratories used a variety of sample preparation and
analytical techniques.

4.3.1.2 LGCPromochem
Certified reference material (CRM) 529 was obtained
from LGC Promochem. The following description is
taken from the reference material report that
accompanied CRM 529.  The soil for CRM 529 was
collected in Europe from a site where chloro-organic and
other compounds had been in large-scale production for
several decades, but where production had ceased more
than five years before sampling. The site had been
contaminated during long-term production of
trichlorophenoxyacetic acid. An area of sandy soil was
excavated to a depth of several meters. Several hundred
kgs of this mixed soil were air-dried at about 15°C for 3
months. After removal of stones and other foreign matter
by sieving, the remaining material was sterilized in air at
120°C for 2 hours, thoroughly mixed, and ground in an
Alpine air jet mill to a particle size of < 63 micrometers
(|im). The material was homogenized once more in a
Turbula mixer and packaged in 50-g quantities. The final
mean moisture content at the time of bottling was found
to be 1.5%. According to the CoA, certified values are
provided for five dioxin congeners, seven furan
congeners, three chlorobenzene compounds, and three
chlorophenol compounds. No PCBs were reported with
certified values on the CoA, so the mean concentration
determined by the reference laboratory was used as the
certified value.

4.3.1.3 Wellington
PE sample WMS-01 was obtained from TerraChem, the
U.S. distributor for Wellington, an Ontario-based
company. As described in the CoA, WMS-01 is a
homogeneous lake sediment that was naturally
contaminated (and not fortified). The crude, untreated
sediment used to prepare WMS-01 was collected from
Lake Ontario. The sediment obtained was subsequently
air-dried; crushed to break up agglomerates; air-dried
again; and then sieved, milled, and re-sieved (100%
< 75 |im). The sediment was then  subsampled into 25-g
aliquots. The demonstration samples for only the
Wellington PE samples were 25 g rather than 50 g based
on the package size available from Wellington.  Certified
values for the 17 D/F congeners and the 12 WHO PCB
congeners are provided on the CoA.

4.3.1.4 National Institute for Standards and
       Technology
Standard Reference Material® (SRM) 1944 was
purchased through NIST. As described in the CoA,
SRM 1944 is a mixture of marine  sediment collected
from six sites in the vicinity of New York Bay and
Newark Bay in October 1994. Site selection was based
on contaminant levels measured in previous samples
from these sites and was intended  to provide relatively
high concentrations for a variety of chemical classes of
contaminants. The sediment was collected using an
epoxy-coated modified Van Veen-type grab sampler
designed to sample the sediment to a depth of
10 centimeters (cm).  A total of approximately 2,100 kg
of wet sediment was collected from the six sites. The
sediment was freeze-dried,  sieved (nominally 6.1 to
250 |im), homogenized in a cone blender, radiation
sterilized, then packaged in 50-g quantities.  Certified
                                                    18

-------
values are provided on the CoA for the 17 D/F
congeners, 30 PCB congeners, 24 PAHs, four
chlorinated pesticides, 36 metals, and total organic
carbon. Since only three WHO PCBs were reported out
of the 30 PCB congeners, the mean concentration of the
reference laboratory measurements was used as the
certified value so that the TEQPCB concentration would
not be underestimated when compared to the developer
technologies.

4.3.1.5 Environmental Resource Associates
ERA synthesized PE samples for this demonstration.
ERA spiked  blank, uncontaminated soil to pre-
determined levels of D/Fs, PCBs, and/or PAHs. Spiked
PE samples were prepared to include additional
concentration ranges and compositions that were not
covered with the commercially available certified
materials. The organic semivolatile soil blank (ERA
Product #056, Lot 56011) is atopsoil that was obtained
from a nursery and processed according to ERA
specifications by a geochemical laboratory. The particle
size distribution of the soil was -20/+60 mesh. The soil
was processed and blended with a sandy loam soil to
create a blank soil with the following make-up: 4.1%
clay, 4.5% silt, 91.2% sand, and 0.2% organic material.
Initially, ERA was required to certify that the blank soil
matrix to be  used as the blank and for the preparation of
the spiked PE samples was "clean" relative to the list of
required target analytes. This was accomplished through
a combination of ERA-conducted analyses (PAHs,
pesticides, semivolatile organic compounds, Aroclors
which are trade mixtures of PCB congeners) and
subcontracted analytical verification (D/F and PCBs).
The subcontracted analyses were performed by Alta
Analytical Perspectives, LLC, in Wilmington, North
Carolina. The Alta Analytical Certificate  of Results and
the ERA Certification sheets for the organic semivolatile
soil blank indicated that trace levels of the octa-dioxins
and several WHO PCB congeners were detected, but the
total TEQ (combined  D/F and PCBs) was less than
0.06 pg/g. The level of PAHs, pesticides, Aroclors, and
semivolatile  organic compounds in the soil was
determined to be < 0.33 pg/g. The TEQ level was
considerably below the detection capabilities of the
participating technologies, so the organic semivolatile
soil blank was considered adequately clean for use in
this demonstration.
The manufacturing techniques that ERA used to prepare
the PE samples for this demonstration were consistent
with those used for typical semivolatile soil products by
ERA. These techniques have been validated through
hundreds of round robin performance test studies over
ERA's more than 25 years in business. The D/F stock
solutions used in the manufacture of these PE samples
were purchases from CIL. The PCB and PAH stock
solutions were purchased from ChemService. For each
PE sample, a spiking concentrate was prepared by
combining appropriate weight/volume aliquots of stock
materials required for that PE sample. Typically,
additional solvent was added to this concentrate to yield
sufficient volume of solution, appropriate for the mass of
soil to be spiked. Based on a soil mass of 1,600 g, the
volume of spike concentrate was approximately 10 to
30 mL. For each PE sample, the blank soil matrix was
weighed into a 2-liter (L) wide mouth glass jar, the spike
concentrate was distributed onto the soil, and the soil
was allowed to air-dry for 30 to 60 minutes. The PE
samples were then capped and mixed in a rotary tumbler
for 30 minutes. Each PE sample was certified as the
concentration of target analytes present in the blank
matrix, plus the amount added during manufacture,
based on volumetric and gravimetric measurements.
CoAs were provided by ERA for all six ERA-provided
PE samples. The certified values provided by ERA were
different from the commercially available certified
samples since the data were not based on analytically
derived results. Further confirmation of the
concentrations was conducted by the reference
laboratory.

4.3.2  Environmental Samples
Handling of the environmental samples is described in
this section. Note that once the environmental samples
were collected, they were dried and homogenized as best
as possible to eliminate variability introduced by sample
homogeneity. As such, the effect of moisture on the
sample analysis was not investigated.

4.3.2.1 Environmental Sample Collection
Samples were collected by the EPA, an EPA contractor,
or the MDEQ and shipped to the characterization
laboratory. When determining whether a soil or sediment
site had appropriate dioxin contamination, a guideline
concentration range of < 50 pg/g to 5,000 pg/g was used.
                                                    19

-------
Once necessary approvals and sampling locations had
been secured, sample containers were shipped to site
personnel. Each site providing samples received one-
gallon containers (Environmental Sampling Supply,
Oakland, California, Part number 3785-1051, wide-
mouth, 128-ounce high-density polyethylene round
packer) for collecting five or six samples.

Instructions for sample collection, as well as how the
containers were to be labeled and returned, were
included in a cover letter with the sample containers that
were shipped to each site. Personnel collecting the
samples were instructed to label two containers
containing the same sample as "1 of 2" and "2 of 2" and
to attach a description or label to each container with a
description of the sample, including where the sample
was collected and the estimated concentrations of dioxin
and any other anticipated contamination [e.g., PCBs,
PAHs, PCP]. Final instructions to sample providers
indicated that collected samples were to be shipped back
to the characterization laboratory using the provided
coolers. Federal Express labels that included an account
number and the shipping address were enclosed in each
shipment.

Sample providers also were asked to provide any
information about the possible source of contamination
or any historical data and other information, such as
descriptions  of the sites, for inclusion in the demon-
stration and quality assurance project plan (D/QAPP).(2)

4.3.2.2 Homogenization of Environmental Samples
If the material had very high moisture content, the jar
contents were allowed to settle, and the water was
poured off. Extremely wet material was poured through
fine mesh nylon material to remove water. After water
removal, the material was transferred to a Pyrex™ pan
and mixed. After thorough mixing, an aliquot was stored
in a pre-cleaned jar as a sample of "unhomogenized"
material and was frozen.1 The remaining bulk sample
was mixed and folded bottom to top three times. This
material was split equally among multiple pans. In each
  Ideally, the samples would have been stored at 4° ± 2°C;
  but, due to the large volume of buckets and jars that needed
  to be stored, the most adequate available storage at the
  characterization laboratory was a walk-in freezer that was at
  approximately minus 20°C.
pan, the material was spread out to cover the entire
bottom of the pan to an equal depth of approximately
0.5 inches. The pans were placed in an oven at 35 °C and
held there until the samples were visibly dry. This
process took from 24 to 72 hours, depending on the
sample moisture. The trays were removed from the oven
and allowed to rise to room temperature by sitting in a
fume hood for approximately 2 hours. Approximately
500 g of material were put in a blender and blended for
2 minutes. The blender sides were scraped with a spatula
and the sample blended for a second 2-minute period.
The sample was sieved [USA Standard testing, No. 10,
2.00-millimeter (mm) opening] and the fine material
placed in a tray. Rocks and particles that were retained
on the sieve were placed in a pan. This process was
repeated until all of the sediment or soil were blended
and sieved. The blended and sieved sediment or soil in
the tray was mixed well, and four aliquots of 100- to
300-g each were put into clean jars (short, wide-mouth
4-ounce, Environmental Sampling Supply, Oakland,
California, Part number 0125-0055) to be used for the
characterization analyses.  The remaining sediment or
soil was placed in  a clean jar, and the particles that were
retained on the sieve were disposed of. The jars of
homogenized sediment and soil were stored frozen
(approximately -20°C), unless the samples were being
used over a period of several days, at which time they
were temporarily stored at room temperature.

4.3.2.3  Selection  of Environmental Samples
Once homogenized, the environmental samples were
characterized for dioxin/furans (EPA Method 1613B(3)),
PCBs, low-resolution mass spectrometry (LRMS)
modified EPA Method 1668A(4), and 18 target PAHs
[National Oceanic and Atmospheric Administration
(NOAA) method(7)] to establish the basic composition  of
the samples. (Characterization analyses are described in
Chapter 5.) Because the soil and sediment samples were
dried and homogenized, they were indistinguishable. As
such, the soil and sediment samples were jointly referred
to as "environmental" samples, with no distinction made
between soil or sediment other than during the matrix
effects evaluations, as described in Section 4.7.6.
Environmental samples were selected for inclusion in the
demonstration based on the preliminary characterization
data. The number and type of samples from each
sampling location  included in the demonstration are
presented in Table 4-5.
                                                    20

-------
Table 4-5.  Characterization and Homogenization Analysis Results for Environmental Samples
Sample
Type ID
Env Site #1
Env Site #2
Env Site #3
Env Site #4
Env Site #5
Env Site #6
Env Site #7
Env Site #8
Env Site #9
Env Site
#10
Environmental
Site Location
Warren County,
North Carolina
Tittabawassee
River, Michigan
Newark Bay,
New Jersey
Raritan Bay, New
Jersey
Winona Post,
Missouri
Tittabawassee
River, Michigan
Brunswick,
Georgia
Saginaw River,
Michigan
Midland,
Michigan
Solutia, West
Virginia
Soil or
Sediment
soil
soil
sediment
sediment
soil
sediment
sediment
sediment
soil
soil
Sample
No.
1
2
3
1
2
3
1
2
3
4
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
4
1
2
3
Average Total
TEQD/F
Concentration
(Pg/g)
274
5,065
11,789
42
435
808
16
62
45
32
12
14
13
3,831
11,071
11,739
1
55
16
69
65
14,500
921
1,083
204
239
184
149
25
48
1,833
3,257
RSD (%)
11
7
3
23b
5
10
26b
14
26b
6
2
3
7
1
2
1
23b
7
26b
8
1
2
9
28C
34C
5
5
7
10
10
19
11
Average RSD for all environmental samples used in demonstration
Total number of environmental samples
No. of Replicates
Per Sample
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
Correlation
with PE
Sample Type
roa
9, 10, 11
4
5
5
2,8
1,3
8
4
4
6,7
77%
72S
1 PE Sample IDs are provided in Table 4-4.
b RSD values up to 30% were allowed for samples where the characterization analyses determined concentration to be < 50 pg/g total TEQD/F.
c RSD value slightly exceeded the homogeneity criteria, but samples were included in the demonstration because they were samples of interest.
                                                          21

-------
Four aliquots of the homogenized material and one
aliquot of unhomogenized material were analyzed. Two
criteria had to be met for the environmental sample to be
considered for inclusion in the demonstration. The first
criterion was that the relative standard deviation (RSD)
of the total D/F TEQ values from the four aliquots had to
be less than 20% for samples with total TEQ values
> 50 pg/g; RSD values up to 30% were considered
acceptable if the concentration was < 50 pg/g. The
second criterion was that no single RSD for an
individual congener could be greater than 30%. If both
of these criteria were met, the sample met the homogeni-
zation criteria and was considered for inclusion in the
demonstration. If either of these criteria was not met.
options for the sample included (a) discarding it and not
considering it for use in the demonstration,
(b) reanalyzing it to determine if the data outside the
homogenization criteria were due to analytical issues, or
(c) rehomogenizing and reanalyzing it. Of these options,
(a) and (b) were utilized, but (c) was not because an
adequate number of environmental samples were
selected using criteria (a) and (b). The average D/F
concentration and RSDs for the homogenization
analyses of environmental samples are shown in
Table 4-5. The composition of two particular Saginaw
River samples was of interest for inclusion in the
demonstration because of their concentration and unique
congener pattern, but the homogenization criteria were
slightly exceeded (i.e., 28% and 34% RSD for Saginaw
River Sample #2 and Saginaw River Sample #3,
respectively). Since multiple replicates of every sample
were analyzed, those samples were included in the study
because of their unique nature but are flagged as slightly
exceeding the homogenization criteria. A similar
correlation of environmental samples to PE samples,
similar to that presented in Table 4-4, is presented in
Table 4-5.

 4.3.3 Extracts
A summary of the extract samples is provided in
Table 4-6. The purpose of the extract samples was to
evaluate detection and measurement performance
independent of the sample extraction method. As shown
in Table 4-6, two environmental samples (both
sediments) were extracted using Soxhlet extraction with
toluene. These extractions were performed by AXYS
Analytical Services consistent with the procedures to
extract the demonstration samples for reference
analyses.(2) The environmental sample extracts repre-
sented a 10-g sediment sample extraction and were
reported in pg/mL, which was calculated by the
following equation:
            (Pg/gsamples)x(lOgaliquot)
    pg/mL =
             (300 mL extraction volume)
                                      ,
where DF = dilution factor.
Total extract volume per 10-g aliquot was 300 mL, but
the sample extracts were concentrated and provided to
the developers as 10-mL extracts, so a 30x dilution
factor is included. The extracts were not processed
through any cleanup steps, but they were derived from
sediment samples that also were included in the suite of
environmental samples. All environmental sample
extractions were prepared in the same solvent (toluene).
The extract samples also included three toluene-spiked
solutions that were not extractions of actual environ-
mental samples. Because adequate homogenization at
trace quantities was difficult to achieve, one set of
extract samples was spiked at low levels (approximately
0.5 pg/mL of 2,3,7,8-TCDD) and used as part of the
MDL evaluation.

4.4     Sample Handling
In preparation for the demonstration, the bulk
homogenized samples were split into jars for distribu-
tion. Each 4-ounce, amber, wide-mouth glass sample jar
(Environmental Sampling Supply, Oakland, California,
Part number 0125-0055) contained approximately 50 g
of sample. Seven sets of samples were prepared for five
developers, the reference laboratory, and one archived
set. A minimum of four replicate splits of each sample
was prepared for each participant, for a total of at least
28 aliquots prepared for each sample. The purchased  PE
samples (i.e., standard reference materials and spiked
materials) were transferred from their original packaging
to the jars to be used in the demonstration for the
environmental samples, making the environmental and
PE samples visually indistinguishable.

The samples were randomized in two ways. First, the
order in which the filled jars were distributed was
randomized. All jars had two labels. The label on the  top
of the jar was the analysis order and contained sample
numbers 1 through 209. A second label placed on the
side of the jar contained a coded identifier including a
                                                    22

-------
Table 4-6.  Distribution of Extract Samples
Sample Type ID
Extract #1
Extract #2
Extract #3
Extract #4
Extract #5
Sample ID
Environmental #6,
Sample #2
Environmental #7,
Sample #1
Spike #la
Spike #2a
Spike #T
Sample Description
Soxhlet extraction in toluene; no
cleanup
Soxhlet extraction in toluene; no
cleanup
0.5 pg/mL
2,3,7,8-TCDD
100 pg/mL 2,3,7,8-TCDD
1,000 pg/mL each WHO PCB
(TEQ-11)
10,000 pg/mL each WHO PCB
(TEQ~1,000)C
Total number of extracts
No. of Replicates per Sample
4
4
7b
4
4
23
1  Prepared in toluene.
b  Seven replicates were analyzed for MDL evaluation.
c  This extract was spiked with PCBs only but a low-level (approximately 0.3 pg/mL) 2,3,7,8-TCDD contamination was confirmed by the
  reference laboratory.
series of 10 numbers coded to include the site, replicate,
developer, and matrix.

All samples believed to have at least one D/F or PCB
congener greater than 10,000 pg/g were marked with an
asterisk for safety purposes.  This was consistent for both
the developer and reference  laboratory samples. The
developer was given the option of knowing which
environmental site the samples came from and whether
the sample was a soil  or sediment.  Abraxis elected not
to have any of this information. As  described in the
D/QAPP, AXYS was informed of which environmental
site that the samples came from so it could use congener
profiles and dilution schemes determined during the
pre-demonstration phase as a guide along with the
concentration range data that was provided in the
D/QAPP. This information was supplied to the reference
laboratory with the samples, along with which samples
contained high (i.e., a sample with at least one congener
with concentration > 120,000 pg/g) or ultrahigh (i.e., a
sample with at least one congener with concentration
> 1,200,000 pg/g) PCB levels. Using this information,
AXYS regrouped the  samples in  batches so that, to the
extent possible, samples from the same site would be
analyzed within the same analytical batch. Because an
analytical laboratory might know at least what site
samples came from, and because it is reasonable from an
analytical standpoint to group samples that might require
similar dilution schemes and which have similar
congener patterns in an analytical batch, this approach
was an acceptable deviation from the original intention
of having the samples run completely blind by the
reference laboratory completely blind and in the
prescribed analytical order.

Abraxis analyzed the samples in the order received. The
extracts were the first 23 samples in the analysis order.
The randomization was generated so that an equal split
of the sample replicates were analyzed in the field and in
the laboratory. For example, when four replicates of a
particular sample were included in the suite of
demonstration samples, two replicates were analyzed
among the first 116 samples that were analyzed in the
field by Abraxis and two replicates were among the
second set of 93 samples that were analyzed in the
Abraxis laboratories. The environmental samples were
stored at room temperature until homogenized. After
homogenization and prior to distribution during the
demonstration, the samples were stored in a walk-in
freezer (approximately -20°C) at the characterization
laboratory. At the demonstration site, the samples were
stored at ambient temperature. After the demonstration
analyses were completed, the samples were stored at the
characterization laboratory in the walk-in freezer until
the conclusion of the project.
                                                    23

-------
4.5    Pre-Demonstration Study
Prior to the demonstration, pre-demonstration samples
were sent to Abraxis for evaluation in its laboratory. The
pre-demonstration study comprised 15 samples,
including PE samples, environmental samples, and
extracts. The samples selected for the pre-demonstration
study covered a wide range of concentrations and
included a representative of each environmental site
analyzed during the demonstration.

The pre-demonstration study was conducted in two
phases. In Phase 1, Abraxis was sent six soil/sediment
samples with the corresponding D/F, PCB, and PAH
characterization data to perform a self-evaluation of the
coplanar PCB ELISA kit performance. In Phase 2, seven
additional soil/sediment samples and two extracts were
sent to Abraxis for blind evaluation. AXYS analyzed all
15 pre-demonstration samples blindly. The Abraxis
pre-demonstration results were paired with the AXYS
results and returned to Abraxis so they could use the
HRMS pre-demonstration sample data to refine the
performance of the coplanar PCB ELISA kit prior to
participating in the field demonstration. Results for the
pre-demonstration study can be found in the data
evaluation report, which can be obtained by contacting
the EPA program manager for this demonstration. The
results confirmed that Abraxis was a viable candidate to
continue in the demonstration process.

4.6    Execution of Field  Demonstration
Abraxis arrived on-site on Sunday, April 25, and spent a
few hours setting up its mobile laboratory. The
demonstration officially commenced on Monday,
April 26 after 1.5 hours of safety and logistical training.
During this meeting, the health and safety plan was
reviewed to ensure all participants understood the safety
requirements for the demonstration.  Logistics, such as
how samples would be distributed and results reported,
were also reviewed during this meeting. After the safety
and site-specific training meeting and prior to samples
being received by the developers, each trailer and mobile
laboratory was surface wipe sampled on the floor to the
entrance of the developer work area to establish the
background level of D/F and PCB contamination. The
wipe sampling procedure was followed as described in
the D/QAPP. Following demobilization by the
developers, all of the trailers and mobile laboratories
were cleaned and surface-wipe-sampled. Analysis of the
pre- and post-deployment wipe samples indicated that all
trailers and mobile laboratories met the acceptable
clearance criteria that were outlined in the D/QAPP.
Only one fume hood had to be recleaned and resampled
before receiving final clearance.

Ideally, all 209 demonstration samples would have been
analyzed on-site, but sample throughput of some of the
technologies participating in the demonstration would
require three weeks or more in the field to analyze 209
samples. Consequently, it was decided, as reported in
the D/QAPP, that the number of samples to be analyzed
in the field by each developer would be determined at
the discretion of the developer.

Abraxis received its first batch of samples by
midmorning on April 26. Abraxis completed the field
sample results in three working days (on April 28). It
should be noted that the morning of April 28 was
dedicated to a Visitor's Day, so minimal work on sample
analyses was performed.  Abraxis analyzed the 23
extracts and exactly half of the soil/sediment samples
(116) during the field demonstration. The remaining 93
samples were completed by Abraxis in its laboratories.
These samples were shipped to Abraxis on April 28 and
received at Abraxis on April 29. The remaining
93 samples analyzed in the Abraxis laboratories were
reported on June 2.  Abraxis estimated that it spent one
week completing the analyses in its laboratory. Once the
complete data set was submitted, Abraxis was offered
the opportunity to reanalyze any samples before
reporting final results. Abraxis reanalyzed all of the field
samples but elected to keep all of the results that were
reported in the field.

4.7    Assessment of Primary and Secondary
       Objectives
The purpose of this section is  to describe how the
primary and secondary objectives are assessed, as
presented in Chapters 6 and 7.

Abraxis reported its results in pg/g TEQPCB. The Abraxis
results were compared to the certified values and
reference laboratory results for pg/g TEQPCB.
                                                   24

-------
4.7.1  Primary Objective PI: Accuracy
The determination of accuracy was based on agreement
with certified or spiked levels of PE samples. PE
samples containing concentrations from across the
analytical range of interest were analyzed. Percent
recovery values relative to the certified or spiked
concentrations were calculated. The PE samples were
analyzed by the laboratory reference method for
confirmation of certified and spiked values.

To evaluate accuracy, the mean replicate results from the
field technology measurement were compared to the
certified or spiked value of the PE samples to calculate
percent recovery. The equation used was:
                R= C/CRxlOO%

where  C is the mean concentration value calculated
from the technology replicate measurements (in pg/g
TEQPCB) and CR is the certified value (in pg/g TEQPCB).
Nondetects and values reported as "> (value)" were not
included in the accuracy assessment. Mean
concentration values were determined when at least three
of four replicates were reported as actual values [i.e.,
were not reported as, "< (value)" or "> (value)"]. The
mean, median, minimum, and maximum R values are
reported as an assessment of overall accuracy. An ideal
Rvalue would be 100%.

4.7.2  Primary Objective P2: Precision
To evaluate precision, all samples (including PE,
environmental, and extract samples) were analyzed in at
least quadruplicate. Seven replicates of three different
samples were analyzed to evaluate MDLs.

Precision was evaluated at both low and high
concentration levels and across different matrices. The
statistic used to evaluate precision was RSD. The
equation used to calculate standard deviation (SD)
between replicate measurements was:
             SD =
        RSD =
               SD
               C
x 100%
where SD is the standard deviation and  C  is the mean
measurement. Both values are reported in pg/g TEQPCB.

The equation used to calculate RSD between replicate
measurements was:
RSD, reported in percent, was calculated if detectable
concentrations were reported for at least three replicates.
The mean, median, minimum, and maximum RSD
values are reported as an assessment of overall precision.

Low RSD values (< 20%) indicated high precision. For a
given set of replicate samples, the RSD of results was
compared with that of the laboratory reference method's
results to determine whether the reference method is
more precise than the technology or vice versa for a
particular sample set. The mean RSD for all  samples was
calculated to determine an overall precision estimate.

4.7.3  Primary Objective P3: Comparability
Data comparability was maximized by using the
homogenization procedures and applying criteria for
acceptable results prior to a sample being included in the
demonstration. (See Section 4.3.2.3 for additional
information.)

Technology results reported by Abraxis were compared
to the corresponding reference laboratory results by
calculating  a relative percent difference (RPD). The
equation for RPD is as follows:
                   K-MD)
        RPD =
                average (MR,MD)
                   100%
where MR is the reference laboratory measurement (in
pg/g TEQPCB) and MD is the developer measurement (in
pg/g TEQPCB). Nondetects were not included in this
evaluation.  For the PE samples, RPD calculations were
only performed for those samples that contained PCBs.
For example, PE sample #6 was only spiked with
2,3,7,8-TCDD. Consequently, RPD calculations were
not performed.

The absolute value of the difference between the
reference and developer measurements in the equation
above was not taken so that the RPD would indicate
whether the technology measurements were greater than
the reference laboratory measurements (negative RPD
values) or less than the reference laboratory
measurements (positive RPD). Because negative values
for RPD could be obtained with this approach, the
median RPD of all individual RPDs was  calculated
                                                   25

-------
rather than the average RPD in calculation of
comparability between the Abraxis results and reference
laboratory measurements. The median, minimum, and
maximum RPD values were reported as an assessment of
overall comparability. RPD values between positive and
negative 25% indicated good agreement between the two
measurements.

As another measure of comparability, the reference data
were grouped into four TEQ concentration ranges. The
ranges were_< 50 pg/g, 50 to 500 pg/g, 500 to
5,000 pg/g, and > 5,000 pg/g.  The intervals were
determined by the Demonstration Panel and were based
on current guidance for cleanup levels. The percentage
of developer results that agreed with those ranges of
values was reported.

The accuracy of reporting blank samples was assessed.
The blanks included eight replicate samples that
contained levels of PCBs that were below the reporting
limits of the developer technology but contained levels
that could be detected by the reference methods (see
Table 4-4). If the reference laboratory result was in the
nondetect interval reported by the developer technology
reporting limit, this result was considered accurately
reported by the developer. The accuracy of the blank
samples was reported in terms of % agreement. Ideal %
agreement values would be 100%.

4.7.4  Primary Objective P4: Estimated Method
       Detection Limit
The method detection limit (MDL) calculation procedure
described in the demonstration plan was 40 CFR
Part 136, Appendix B, Revision 1.11.  This procedure is
based on an assumption that the replicates are
homogeneous enough to allow proper measurement of
the analytical precision and that the concentration is in
the appropriate range for evaluation of the technology's
sensitivity. For this evaluation, Abraxis analyzed seven
aliquots each of a low-level PE soil, PE sediment, and a
toluene-spiked extract. MDL-designated samples are
indicated in Tables 4-4 and 4-6.  The developer reported
nondetect values for some of the replicates, so
provisions had to be made for the treatment of
nondetects. As such, the results from these samples were
used to calculate an estimated MDL (EMDL) for the
technology.
A Student's t-value and the standard deviation of seven
replicates were used to calculate the EMDL in pg/g TEQ
is shown in the following equation:
            EMDL= t
                       (n-l,l-co=0.99)
(SD)
where  t(n_l}_^=OM) = Student's t-value appropriate for a
99 percent confidence level and a standard deviation
estimate with n-1 degrees of freedom. Nondetect values
were assigned the reported value (i.e., "< 1" was
assigned as value of 1), half of the reported value (i.e.,
"< 1" was assigned 0.5), or zero. The various treatments
of nondetect values were performed to see the impact
that reduced statistical power (i.e., lower degrees of
freedom) had on the EMDL calculation. The lower the
EMDL value, the more  sensitive the technology is at
detecting contamination.

4.7.5  Primary Objective P5: False Positive/False
       Negative Results
The tendency for the Abraxis kit to return false positive
results (e.g., results reported above a specified level for
the field technology but below a specified level by the
reference laboratory) was evaluated. The frequency of
false positive results was reported as a fraction of results
available for false positive analysis. Similarly, the
frequency of false negatives results was examined. For
this purpose, the results were evaluated for samples
reported as having concentrations above and below
6.25 pg/g TEQPCB and above and below 50 pg/g TEQPCB.
As such, the samples that were reported as < 6.25 (or 50)
pg/g TEQPCB by the reference laboratory but > 6.25 (or
50) pg/g TEQPCB by Abraxis were considered false
positive. Conversely, those samples that were reported as
< 6.25  (or 50) pg/g TEQPCB by Abraxis, but reported as
> 6.25  (or 50) pg/g TEQPCB by the reference laboratory,
were considered false negatives. In the case of
semiquantitative results (reported as < or >), if the
laboratory result was within the interval reported by the
developer, it was not considered a false positive or false
negative result.  Ideal false positive and negative
percentages would be equal to zero.

4.7.6  Primary Objective P6: Matrix Effects
The likelihood of matrix-dependent effects on
performance was investigated by grouping the data by
matrix type (i.e., soil, sediment, extract), sample type
(i.e., PE, environmental, and extract), varying levels of
                                                    26

-------
PAH, environmental site and known interferences.
Precision (RSD) data were summarized by soil,
sediment, and extract (matrix type); environmental, PE,
and extract (sample type); and PAH concentration.
Analysis of variance (ANOVA) tests were performed to
determine whether there was a dependence on matrix
type or sample type. Only the environmental samples
were included in the matrix effect assessment based on
PAH concentration, because only the environmental
samples were analyzed for PAHs during the
characterization analysis (Section 5.2.3).  Some PAH
data were available for the PE samples, but data were not
available for all of the same analytes that were
determined during the characterization analysis. The
environmental samples were segregated into four ranges
of total PAH concentrations: < 1,000 nanogram/g (ng/g),
1,000 to 10,000 ng/g, 10,000 to 100,000 ng/g, and >
100,000 ng/g. The precision (RSD) data were
summarized for samples within these PAH
concentrations. ANOVA tests were used to determine if
the summary values for RSD were statistically different,
indicating performance  dependent upon PAH
concentration. For the environmental site evaluation, the
comparability (RPD) values  from each of the 10
environmental sites were compared to see whether the
developer results were comparable to the reference
laboratory for a particular site. For known interferences,
the developer's reported results for PE samples were
summarized for samples where the PE samples did not
contain the target analyte (e.g., did the  developer report
PCB detections for a sample only spiked with D/Fs).

This objective also evaluated whether performance was
affected by measurement location (i.e., in-field versus
laboratory measurements), although this is not a
traditional matrix effect. To evaluate the effect of
measurement location, ANOVA tests were performed
for sample results  within a replicate set that were
generated both in the laboratory and in the field. For
these analyses, p-values < 0.05 indicated statistically
different results between the laboratory and field
measurements and therefore a significant effect of the
measurement location. The percentage of replicate sets
having p-values < 0.05 was reported.

4.7.7  Primary  Objective P7: Technology Costs
The full cost of each technology was documented and
compared to typical and actual costs  for standard HRMS
PCB analytical results. Cost inputs included equipment,
consumable materials, mobilization and demobilization,
and labor. The evaluation of this objective is described
in Chapter 8, Economic Analysis.

4.7.8  Secondary Objective SI: Skill Level of
       Operator
Based on observations during the field demonstration,
the type of background and training required to properly
operate the  coplanar PCB ELISA kit was assessed and
documented. The skill required of an operator was also
evaluated. The evaluation of this secondary objective
also included user-friendliness of the technology.

4.7.9  Secondary Objective S2: Health and Safety
       Aspects
Health and  safety issues, as well as the amount and type
of hazardous and nonhazardous waste generated, were
evaluated based on observer notes during the field
demonstration. This also included an assessment of the
personal protective equipment required to operate the
technology.

4.7.10 Secondary Objective S3: Portability
Observers documented whether the coplanar PCB
ELISA kit could be readily transported to the field and
how easy it was to operate in the field. This included an
assessment of what infrastructure requirements were
provided to Abraxis (e.g., a mobile laboratory), and an
assessment of whether the infrastructure was adequate
(or more than adequate) for the technology's operation.
Limitations of operating the technology in the field are
also discussed.

4.7.11 Secondary Objective 84: Sample
       Throughput
Sample throughput was measured based on the observer
notes, which focused  on the time-limiting steps of the
procedures, as well as the documentation of sample
custody. The number  of hours Abraxis worked in the
field was documented using attendance log sheets where
Abraxis recorded the time they arrived and departed
from the demonstration site. Time was removed for
training and Visitor's  Day activities. The number of
operators involved in  the sample analyses also was
noted. Throughput of the developer technology was
compared to that of the reference laboratory.
                                                   27

-------
                                             Chapter 5
                                       Confirmatory Process
This chapter describes the characterization analyses and
the process for selecting the reference methods and the
reference laboratory.

5.1    Traditional Methods for Measurement of
       Dioxin and Dioxin-Like Compounds in
       Soil and Sediment
Traditional methods for analysis of dioxin and dioxin-
like compounds involve extensive sample preparation
and analysis using expensive instrumentation resulting in
very accurate and high-quality, but costly, information.
The ability to use traditional methods for high-volume
sampling programs or screening of a contaminated site
often is limited by budgetary constraints. The cost of
these analyses can range approximately from $500 to
$1,100 per sample per method, depending on the method
selected, the level of QA/QC incorporated into the
analyses, and the reporting requirements.

5.1.1  High-Resolution Mass Spectrometry
EPA Method 1613B(3) and SW846 Method 8290(8)are
both appropriate for low and trace-level analysis of
dioxins and furans in a variety of matrices. They involve
matrix-specific extraction, analyte-specific cleanup, and
high-resolution capillary GC (HRGC)THRMS analysis.
The main differences between the two methods are that
EPA Method 1613B has an expanded calibration range
and requires use of additional 13C12-labeled internal
standards resulting in more accurate identifications and
quantitations. The calibration  ranges for the HRMS
methods based on atypical 10-g sample  and
20-microliter (\\L) final sample volume are presented in
Table 5-1.
Table 5-1.  Calibration Range of HRMS
           Dioxin/Furan Method
Compound
Terra
Compounds
Penta-Hepta
Compounds
Octa
Compounds
EPA Method
1613B
1-400 pg/g
5-2,000 pg/g
10-4,000 pg/g
SW846 Method
8290
2-400 pg/g
5-1,000 pg/g
10-2,000 pg/g
5.1.2  Low-Resolution Mass Spectrometry
SW846 Method 8280 is appropriate for determining
dioxins and furans in samples with relatively high
concentrations, such as still bottoms, fuel oils, sludges,
fly ash, and contaminated soils and waters. This method
involves matrix specific extraction, analyte-specific
cleanup, and HRGC/LRMS analysis. The calibration
ranges in Table 5-2 are based on a typical 10-g sample
size and 100-(iL final volume.

Table 5-2.  Calibration Range of LRMS
           Dioxin/Furan Method
Compound
Tetra-Penta Compounds
Hexa-Hepta Compounds
Octa Compounds
SW846 Method 8280
1,000-20,000 pg/g
2,500-50,000 pg/g
5,000-100,000 pg/g
                                                      5.1.3   PCB Methods
                                                      There are more options for analysis of dioxin-like
                                                      compounds such as PCBs. EPA Method 1668A(4) is for
                                                      low- and trace-level analysis of PCBs. It involves
                                                      matrix-specific extraction, analyte-specific cleanup, and
                                                   28

-------
HRGC/HRMS analysis. This method provides very
accurate determination of the WHO-designated dioxin-
like PCBs and can be used to determine all 209 PCB
congeners. Not all PCBs are determined individually
with this method because some are determined as sets of
coeluting congeners. The calibration range for PCBs
based on a typical 10-g sample and 20-(iL final sample
volume is from 0.4 to 4,000 pg/g. PCBs also can be
determined as specific congeners by GC/LRMS or as
Aroclors1 by GC/electron capture detection.

5.1.4  Reference Method Selection
Three EPA analytical methods for the quantification of
dioxins and furans were available: Method 1613B,
Method 8290, and Method 8280. Method 8280 is a
LRMS method that does not have adequate sensitivity
(i.e., the detection limits reported by the developers are
less than that of the LRMS method). Methods 1613B
and 8290 are HRMS methods with lower detection
limits. Method 1613B includes more labeled internal
standards than Method 8290, which affords more
accurate congener quantification. Therefore, it was
determined that Method 1613B best met the needs of the
demonstration, and it was selected as the dioxin/furan
reference method. Reference data of equal quality
needed to be generated to determine the PCB
contribution to the TEQ, since risk assessment is often
based on TEQ values that are not class-specific. As such,
the complementary HRMS method for PCB TEQ
determinations, Method 1668A,(4) was selected as the
reference method for PCBs. Total TEQD/F concentrations
were generated by Method 1613B, and total TEQPCB
concentrations were generated by Method 1668A. These
data were summed to derive a total TEQ value for each
sample.

5.2    Characterization of Environmental
       Samples
All of the homogenized environmental samples were
analyzed by the Battelle characterization laboratory to
determine which would be included in the
demonstration. The environmental samples were
characterized for the 17 D/Fs by Method 1613B, the
12 WHO PCBs by LRMS-modified Method  1668A, and
18 target PAHs by the NOAA Status and Trends
GC/Mass Spectrometry (MS) method.(7)
5.2.1  Dioxins and Furans
Four aliquots of homogenized material and one
unhomogenized (i.e., "as received") aliquot were
prepared and analyzed for seventeen 2,3,7,8-substituted
dioxins and furans following procedures in EPA Method
1613B. The homogenized and unhomogenized aliquots
were each approximately 200 g. Depending on the
anticipated levels of dioxins from preliminary
information received from each sampling location,
approximately 1 to 10 g of material were taken for
analysis from each aliquot, spiked with 13C12-labeled
internal standards, and extracted with methylene chloride
using accelerated solvent extraction techniques. One
method blank and one laboratory control spike were
processed with the batch of material from each site. The
sample extracts were processed through various cleanup
techniques, which included gel permeation chroma-
tography or acid/base washes, as well as acid/base silica
and carbon cleanup columns. As warranted, based on
sample compositions, some samples were put through
additional acid silica cleanup prior to the carbon column
cleanup. Extracts were spiked with 13C12-labeled
recovery standards and concentrated to a final volume of
20 to 50 (iL.  Dilution and reanalysis of the extracts were
performed if high levels of a particular congener were
observed in the initial analysis; however, extracts were
not rigorously evaluated to ensure that all peaks were
below the peak area of the highest calibration standard.

Each extract was analyzed by high-resolution gas
chromatography/HRMS in the selected ion monitoring
(SIM) mode  at a resolution of 10,000 or greater. A DB-5
column was used for analysis of the seventeen 2,3,7,8-
PCDD/F congeners. The instrument was calibrated for
PCDD/F at levels specified  in Method 1613B with one
additional calibration standard at concentrations
equivalent to one-half the level of Method 1613B's
lowest calibration point. Using a DBS column, 2,3,7,8-
TCDF is not separated from other non-2,3,7,8-TCDF
isomers. However, since the primary objective was to
determine adequacy of homogenization and not
congener quantification, it was determined that sufficient
information on precision could be obtained with the DBS
analysis of 2,3,7,8-TCDF and no second column
confirmation of 2,3,7,8-TCDF was performed. PCDD/F
data were reported as both concentration (pg/g dry) and
TEQs (pg TEQ/g dry).
                                                   29

-------
5.2.2  PCBs
One aliquot of material from each sampling location was
prepared and analyzed for the 12 WHO-designated
dioxin-like PCBs by GC/LRMS. The LRMS PCB
analysis method is based on key components of the PCB
congener analysis approach described in EPA Method
166 8A and the PCB homologue approach described in
EPA Method 680. Up to 30 g of sample were spiked
with surrogates and extracted with methylene chloride
using shaker table techniques. The mass of sample
extracted was determined based on information supplied
to the laboratory regarding possible contaminant concen-
trations. The extract was dried over anhydrous sodium
sulfate and concentrated. Extracts were processed
through alumina column cleanup, followed by high-
performance liquid chromatography/gel permeation
chromatography (HPLC/GPC). Additionally, sulfur was
removed using activated granular copper. The post-
HPLC extract was concentrated and fortified with
recovery internal standards. Extracts were concentrated
to a final volume between 500 (iL and 1 mL, depending
on the anticipated concentration of PCBs in the sample,
as reported by the sample providers. PCB congeners and
PCB homologues were separated via capillary gas
chromatography on a DB5-XLB column and identified
and quantified using electron ionization MS. This
method provides  specific procedures for the
identification and measurement of the selected PCBs in
SIM mode.

5.2.3  PAHs
One aliquot of material from each sampling location was
analyzed for PAHs. The 18 target PAHs included:
•  naphthalene
•  2-methylnaphthalene
•  2-chloronaphthalene
•  acenaphthylene
•  acenaphthene
•  fluorene
•  phenanthrene
•  anthracene
•  fluoranthene
•  pyrene
•  benzo(a)anthracene
•  chrysene
•  benzo(b)fluoranthene
•  benzo(k)fluoranthene
•  benzo(a)pyrene
• indeno(l,2,3-cd)pyrene
• dibenzo(a,h)anthracene
• benzo(g,h,i)perylene.

The method for the identification and quantification of
PAH in sediment and soil extracts by GC/MS was based
on the NOAA Status and Trends method(7) and,
therefore, certain criteria (i.e., initial calibrations and
daily verifications) are different from those defined in
traditional EPA methods 625 and 8270C. Up to 30 g of
sample were spiked with surrogates and extracted using
methylene chloride using shaker table techniques. The
mass of sample extracted was determined based on
information supplied to the characterization laboratory
regarding possible contaminant concentrations. The
extract was dried over anhydrous sodium sulfate and
concentrated. The extract was processed through an
alumina cleanup column followed by HPLC/GPC. The
post-HPLC extract was concentrated and fortified with
recovery internal standards. Extracts were concentrated
between 500 (iL and 1 mL, depending on the anticipated
concentration of PCBs in the sample, as reported by the
sample providers. PAHs were separated by capillary gas
chromatography on a DB-5, 60-m column and were
identified and quantified using electron impact MS.
Extracts were analyzed in the SIM mode to achieve the
lowest possible detection limits.

5.3    Reference Laboratory Selection
Based on a preliminary evaluation of performance and
credibility, 10 laboratories were contacted and were sent
a questionnaire geared toward understanding the
capabilities of the laboratories, their experience with
analyzing dioxin samples for EPA, and their ability to
meet the needs of this demonstration. Two laboratories
were selected for the next phase of the selection process
and were sent three blind audit samples. Each laboratory
went through a daylong audit that included a technical
systems audit and a quality systems audit. At each
laboratory, the  audit consisted of a short opening
conference; a full day of observation of laboratory
procedures, records, interviews  with laboratory staff; and
a brief closing meeting. Auditors submitted followup
questions to each laboratory to address gaps in the
observations.

Criteria for final selection were  based on the
observations of the auditors, the performance on the
                                                   30

-------
audit samples, and cost. From this process, it was
determined that AXYS Analytical Services (Sidney,
British Columbia, Canada) would best meet the needs of
this demonstration.

5.4    Reference Laboratory Sample
       Preparation and Analytical Methods
AXYS Analytical Services received all 209 samples on
April 27, 2004. To report final data, AXYS submitted
14 D/F and 14 PCB data packages from June 11 to
December 20, 2004.  The following sections briefly
describe the reference methods performed by AXYS.

5.4.1   Dioxin/Furan Analysis
All procedures were carried out according to protocols
as described in AXYS Summary Method Doc MSU-018
Rev 2 18-Mar-2004 [AXYS detailed Standard Operating
Procedure (SOP) MLA-017 Rev 9 May-2004], which is
based on EPA Method 1613B. AXYS modifications to
the method are summarized in the D/QAPP.(2)  Briefly,
samples were spiked with a suite  of isotopically labeled
surrogate standards prior to extraction, solvent extracted,
and cleaned up through a series of chromatographic
columns that included silica, Florisil, carbon/Celite, and
alumina columns. The extract was concentrated and
spiked with an isotopically labeled recovery (internal)
standard. Analysis was performed using an HRMS
coupled to an FiRGC equipped with a DB-5 capillary
chromatography column [60 meters (m), 0.25-mm
internal diameter (i.d.), 0.1-|im film thickness]. A
second column, DB-225 (30 m, 0.25-mm i.d., 0.15-|im
film thickness), was used for confirmation of
2,3,7,8-TCDF identification.  Samples that were known
to contain extremely high levels of PCDD/F were
extracted without the addition of the surrogate standard,
split, then spiked with the isotopically labeled surrogate
standard prior to cleanup.  This approach allowed
extraction of the method-specified 10-g  sample volume,
and subsequent sufficient dilution that high level
analytes were brought within the  instrument calibrated
linear range. While this approach induces some
uncertainty because the actual recovery of analytes from
the extraction process is unknown, it was decided by the
demonstration panel that in general analyte recovery
through the extraction procedures are known to be quite
good and that the uncertainty introduced by this
approach would be less than the uncertainty introduced
by other approaches such as extracting a significantly
smaller sample size.

5.4.2  PCB Analysis
The method was carried out in accordance with the
protocols described in AXYS Summary Method Doc
MSU-020 Rev 3 24-Mar-2004 (AXYS detailed SOP
MLA-010 Rev 5 Sep-2003), which is based on EPA
Method 1668A, with changes through August 20, 2003.
AXYS modifications to the method are summarized in
the D/QAPP. Briefly, samples were spiked with
isotopically labeled surrogate standards, solvent
extracted, and cleaned up on a series of chromatographic
columns that included silica, Florisil, alumina,  and
carbon/Celite columns. The final extract was spiked with
isotopically labeled recovery (internal) standards prior to
instrumental analysis. The extract was analyzed by
FiRMS coupled to an HRGC equipped with a DB-1
chromatography column (30 m, 0.25-mm i.d., 0.25-|im
film thickness).  Because only the WHO-designated
dioxin-like PCBs were being analyzed for this program
and in order to better eliminate interferences, all samples
were analyzed using the DB-1 column, which is an
optional confirmatory column in Method 1668A rather
than the standard SPB Octyl column. Samples that were
known to contain extremely high levels of PCBs were
extracted without the addition of the surrogate  standard,
split, then spiked with the isotopically labeled surrogate
standard prior to cleanup.  This approach allowed
extraction of the method-specified 10-g sample volume,
and subsequent sufficient dilution that high level
analytes were brought within the instrument calibrated
linear range. While this approach induces some
uncertainty because the actual recovery of analytes from
the extraction process is unknown, it was decided by the
demonstration panel that in general analyte recovery
through the extraction procedures are known to be quite
good and that the uncertainty introduced by this
approach would be less than the uncertainty introduced
by other approaches such as extracting a significantly
smaller sample size.

5.4.3  TEQ Calculations
For the reference laboratory data, D/F and PCB congener
concentrations were converted to TEQ and subsequently
summed to determine total TEQ, using the TEFs
established by WHO in 1998 (see Table 4-l).(5)
                                                   31

-------
Detection limits were reported as sample-specific
detection limits (SDLs). SDLs were determined from
2.5 times the noise in the chromatogram for D/F and 3.0
times the noise for PCBs, converted to an area, and then
converted to a concentration using the same calculation
procedure as for detected peaks. Any value that met all
quantification criteria (> SDL and isotope ratio) were
reported as a concentration. A "J" flag was applied to
any reported value between the SDL and the lowest level
calibration. The concentration of any detected congener
that did not meet all quantification criteria (such as
isotope ratio or peak shape) was reported but given a
"K" flag to indicate estimated maximum possible
concentration (EMPC).(8) TEQs were  reported in two
ways to cover the range of possible TEQ values:

(1)    All nondetect and EMPC values were assigned a
       zero concentration in the TEQ calculation.

(2)    Nondetects were assigned a concentration of
       one-half the  SDL.  EMPCs were assigned a
       value equal to the EMPC.
In both cases, any total TEQ value that had 10%
contribution or more from J-flagged or K-flagged data
was flagged as J or K (or both) as appropriate.

TEQs were calculated both ways for all samples.  For
TEQD/F, 63% of the samples had the same TEQ value
based on the two different calculation methods, and the
average RPD was 8% (median = 0%). For TEQPCB, 65%
of the samples had the same TEQ value based on the two
different calculation methods, and the average RPD was
9% (median = 0%). Because overall there were little
differences between the two calculation methods, as
presented in Appendix D, TEQ values calculated by
option # 1 were used in comparison with the developer
technologies. On a case-by-case basis, developer results
were compared to TEQs calculated by option #2 above,
but no significant differences in comparability results
were observed so no additional data analysis results
using these TEQ values were presented.
                                                   32

-------
                                              Chapter 6
                         Assessment of Reference Method Data Quality
Ensuring reference method data quality is of paramount
importance to accurately assessing and evaluating each
of the innovative technologies. To ensure that the
reference method has generated accurate, defensible
data, a quality systems/technical audit of the reference
laboratory was performed during analysis of
demonstration samples after the first batch of
demonstration sample analyses was complete.  The
quality systems/technical audit evaluated
implementation of the demonstration plan. In addition, a
full data package was prepared by the reference
laboratory for each sample batch for both dioxin and
dioxin-like PCB analyses. Each data package was
reviewed by both a QA specialist and technical
personnel with expertise in the reference methods for
agreement with  the reference method as described in the
demonstration plan.  Any issues identified during the
quality systems/technical audit and the data package
reviews were addressed by the reference laboratory prior
to acceptance of the data. In this section, the reference
laboratory performance on the QC parameters is
evaluated. In addition, the reference data were
statistically evaluated for the demonstration primary
objectives of accuracy and precision.

6.1    QA Audits
A quality systems/technical audit was conducted at the
reference laboratory, AXYS Analytical Services, Ltd.,
by Battelle auditors on May 26, 2004, during the
analysis of demonstration samples. The purpose of the
audit was to verify AXYS compliance with its internal
quality system and the D/QAPP.(2) The scope
specifically included a review of dioxin and PCB
congener sample processing, analysis, and data
reduction; sample receipt, handling, and tracking;
supporting laboratory systems; and followup to
observations and findings identified during the
independent laboratory assessment conducted by Battelle
on February 11, 2004, prior to contract award.
Checklists were prepared to guide the audit, which
consisted of a review of laboratory records and
documents, staff interviews, and direct observation.

The AXYS quality system is documented in a
comprehensive QA/QC manual and detailed SOPs.  No
major problems or issues were noted during the audit.
Two findings were identified, one related to a backlog of
unfiled custody records and the  other related to the need
for performance criteria for the DB-1 column used for
the analysis of PCB congeners by HRMS. Both issues
were addressed satisfactorily by AXYS after the audit.
One laboratory practice that required procedural
modification was identified:  the laboratory did not
subject all QC samples to the most rigorous cleanup
procedures that might be required for individual samples
within a batch. The AXYS management team agreed that
this procedure was incorrect. As corrective action, the
QA manager provided written instructions regarding
cleanup of the quality control samples to the staff, and
the laboratory manager conducted follow up discussion
with the staff. Other isolated issues noted by the auditors
did not reflect systemic problems and were typical of
analytical laboratories (e.g., occasional documentation
lapses or an untrackable balance weight).

The audit confirmed that the laboratory procedures
conformed to the SOPs and D/QAPP and that the quality
system was implemented effectively.  Samples were
processed and analyzed according to the  laboratory
SOPs and D/QAPP using the Soxhlet Dean Stark
extraction method. No substantial deviations were
noted.  The audit verified the traceability of samples
within the laboratory, as well as the traceability of
standards, reagents, and solvents used in preparation,
and that the purity and reliability of the latter materials
were demonstrated through documented  quality checks.
In addition, the audit confirmed that analytical
                                                    33

-------
instruments and equipment were maintained and
calibrated according to manufacturers' specifications and
laboratory SOPs. Analytical staff members were
knowledgeable in their areas of expertise. QC samples
were processed and analyzed with each batch of
authentic samples as specified by the D/QAPP. QA/QC
procedures were implemented effectively, and corrective
action was taken to address specific QC failures. Data
verification, reporting, and validation procedures were
found to be rigorous and sufficient to ensure the
accuracy of the reported data. The auditors concluded
that AXYS is in compliance with the D/QAPP and its
SOPs, and that the data generated at the laboratory are of
sufficient and known quality to be used as a reference
method for this project.

In addition, each data package was reviewed by both a
QA specialist and technical personnel with expertise in
the reference methods for agreement with the reference
method as described in the demonstration plan.
Checklists were prepared to guide the data package
review. This review included an evaluation of data
package documentation such as chain-of-custody (COC)
and record completeness, adherence to method
prescribed holding times and storage conditions,
standard spiking concentrations,  initial and continuing
calibrations meeting established criteria, GC column
performance, HRMS instrument resolution, method
blanks, lab control spikes (ongoing precision and
recovery samples), sample duplicates, internal standard
recovery, transcription of raw data into the final data
spreadsheets, calculation of TEQs, and data flag
accuracy. Any issues identified during the data package
reviews were addressed by the reference  laboratory prior
to acceptance of the data. All of the audit reports and
responses are included in the DER.

6.2    QC Results
Each data package was reviewed for agreement with the
reference method as  described in the demonstration plan.
This section summarizes the evaluation of the reference
method quality control data.

6.2.1   Holding Times and Storage Conditions
All demonstration samples were stored frozen (< -10°C)
upon receipt and were analyzed within the method
holding time of one year.
6.2.2  Chain of Custody
All sample identifications were tracked from sample
login to preparation of record sheets, to instrument
analysis sheets, to the final  report summary sheets and
found to be consistent throughout.  One COC with an
incomplete signature and one discrepancy in date of
receipt between the COC and sample login were
identified during the Battelle audit and were corrected
before the data packages with these affected items were
accepted as final.

6.2.3  Standard Concentrations
The concentration of all calibration and spiking
standards was verified.

6.2.4  Initial and Continuing Calibration
All initial calibrations met the criteria for response factor
RSD and minimal signal-to-noise ratio requirements for
the lowest calibration point.

Continuing calibrations were performed at the beginning
and end of every 12-hour analysis period with one minor
exception for dioxin/furan sample batch WG13551,
which contained five samples from Environmental Site
#1 (North Carolina) and 12 samples from Environmental
Site #5 (Winona Post).  On  one analysis day, a high-
level sample analyzed just prior to the ending calibration
verification caused the verification to fail. In this
instance, the verification was repeated just outside of the
12-hour period.  The repeat calibration verification met
the acceptance criteria and was considered to show
acceptable instrument performance in the preceding
analytical period; therefore, the data were accepted.

Continuing calibration results were within the criteria
stated in Table 9-2 (D/F) and Table 9-4 (PCB) of the
D/QAPP, with one exception.  For PCB sample batch
WG12108, which contained nine samples from
Environmental Site #3 (Newark Bay) and  12 samples
from Environmental Site #4 (Raritan Bay), isotopically
labeled PCB 169 was above the acceptable range during
one calibration verification  on May 15, 2004. The
acceptance range included in the  D/QAPP is tighter than
the acceptance range in  Method 1668A Table 6.
Because the result for labeled PCB 169 was within the
Method 166 8A acceptance  limits, the data were
accepted.
                                                    34

-------
The minimum signal-to-noise criteria for analytes in the
calibration verification solution were met in all
instances.

6.2.5  Column Performance and Instrument
       Resolution
Column performance was checked at the beginning of
each 12-hour analytical period and met method criteria.

Instrument resolution was documented at the beginning
and end of each 12-hour period with one exception. In
PCB sample batch WG13554, which contained five
performance evaluation samples and 15 extract samples,
on one analysis day (September 17, 2004), the ending
resolution documentation was conducted at 12 hours and
54 minutes.  However, as this resolution documentation
met all criteria, it was considered representative of
acceptable instrument performance during the analytical
period, and the data were accepted.

6.2.6  Method Blanks
Method blanks were analyzed with each sample batch to
verify that laboratory procedures did not introduce
significant contamination. A summary of the method
blank data is presented in Appendix C. There were many
instances for both D/F and PCB data where analyte
concentrations in the method blank exceeded the target
criteria in the D/QAPP. Samples from this
demonstration, which had very high D/F and PCB
concentrations, contributed to the difficulty in achieving
method blank criteria in spite of steps the reference
laboratory took to minimize contamination (such as
proofing the glassware before use in each analytical
batch). In many instances, the concentrations of D/F and
PCBs in the  samples exceeded 20 times the
concentrations in the blanks. For all instances, the
sample results were unaffected because the method
blank TEQ concentration was compared to the sample
TEQ concentrations to ensure that background
contamination did not significantly impact sample
results.

6.2.7  Internal Standard Recovery
Internal standard recoveries were generally within the
D/QAPP criteria. D/QAPP criteria were tighter than the
standard EPA method criteria; in instances where
internal standard recoveries were outside of the D/QAPP
criteria, but within the standard EPA method criteria,
results were accepted. In several instances, the dioxin
cleanup standard recoveries were affected by
interferences. As the cleanup standard is not used for
quantification of native analytes, these data were
accepted. Any samples affected by internal standard
recoveries outside of the D/QAPP and outside of the
EPA method criteria were evaluated for possible impact
on total TEQ and for comparability with replicates
processed during the program before being accepted.

6.2.8  Laboratory Control Spikes
One laboratory control spike (ongoing precision and
recovery sample), which consisted of native analytes
spiked into a reference matrix (sand), was processed
with each analytical batch to assess accuracy.  Recovery
of spiked analytes was within the D/QAPP criteria in
Table 9-2 for all analytes in all laboratory control spike
samples.

6.2.9  Sample Batch Duplicates
A summary of the duplicate data is presented in
Appendix C. One sample was prepared in duplicate in
most sample  batches; four batches were reported without
a duplicate. Three of 14 dioxin sample batches and 5 of
14 PCB sample batches did not meet criteria of <20%
RPD between duplicates. Data where duplicates did not
meet D/QAPP criteria were evaluated on an individual
basis.

6.3    Evaluation of Primary Objective PI:
       Accuracy
Accuracy was assessed through the analysis of PE
samples consisting of certified standard reference
materials, certified spikes, and certified blanks. A
summary of reference  method percent recovery (R)
values is presented in Table 6-1. The Rvalues are
presented for TEQPCB, TEQD/F, and total TEQ. The
minimum, maximum, mean, and median R values are
presented for each set of TEQ results. The reference
method values were in best agreement with the certified
values for the TEQPCB results, with a mean R value of
96%. The mean R values for TEQD/F and total TEQ were
125% and 94%, respectively. The mean and median R
values for the TEQPCB and total TEQ  were identical. The
mean and median R values for TEQD/F were not similar
and were largely influenced by the TEQD/F recovery for
ERA Aroclor of 324%. The ERA Aroclor-certified
TEQD/F values were based on TCDD  and TCDF only,
                                                   35

-------
Table 6-1.  Objective PI Accuracy - Percent Recovery
PE Sample
ID
1
2
3
4
5
6
7
8
9
10
11
12
PE Sample
Description
Cambridge 5 183
LCG CRM-529
Wellington WMS-01
Cambridge 5 184
NIST 1944
ERA TCDD 10
ERA TCDD 30
ERA PAH
ERA PCB 100
ERAPCB 10000
ERA Aroclor
ERA Blank
All Performance Evaluation Samples
% Recovery
TEQPCB
81
100
93
120
102
NA
NA
NA
96
95
82
NA
NUMBER
MESf
MAX
MEDIAN
MEAN
8
81
120
96
96
TEQD/F
111
106
106
106
91
79
77
NA
NA
NA
324
NA
NUMBER
MIN
MAX
MEDIAN
MEAN
8
77
324
106
125
Total TEQ
94
106
105
118
93
79
77
NA
95
95
83
NA
NUMBER
MIN
MAX
MEDIAN
MEAN
10
77
118
94
94
 NA = not applicable; insufficient data were reported to determine R or the sample was not spiked with those analytes.
whereas the reference method TEQD/F values were based
on contributions from all 2,3,7,8-substituted D/F
analytes. The Rvalues presented in Table 6-3 indicate
that the reference method reported data that were on
average between 94 and 125% of the certified values of
the PE samples.

The effect of known interferences on reference method
TEQs is listed in Table 6-2. D/F and PCB TEQs were
not affected by PAH as evidenced through the analysis
of ERA PAH standard reference material. D/F and PCB
TEQs were not affected by each other as evidenced by
spikes that contained only one set of analytes having
negligible influence on the TEQ of the other analyte set.
Table 6-2.  Evaluation of Interferences
6.4    Evaluation of Primary Objective P2:
       Precision
The 209 samples included in the demonstration
consisted of replicates of 49 discrete samples. There
were four replicates of each sample except for PE
sample Cambridge 5183 (7 replicates), ERA blank
reference material (8 replicates), Wellington WMS-01
standard reference material (7 replicates), and 0.5 pg/mL
2,3,7,8-TCDD extract (7 replicates). Reference method
data were obtained for all 209 samples; however, TEQD/F
and total TEQ data for samples Ref 197 (ERA PCB  100)
and Ref 202 (LCG CRM-529) were omitted as outliers
as it appeared that these two samples were switched
during preparation after observing results of the
replicates and evaluating the congener profiles of these
two samples.
PE Material with Known Interference
ERA PAH
ERA PCB 100
ERA PCB 10,000
ERA TCDD 10
ERA TCDD 30
Mean TEQ (pg/g)
0.195 (D/F + PCB)
0.073 (D/F)
0.220 (D/F)
0.025 (PCB)
0.036 (PCB)
                                                   36

-------
A summary of the reference method replicate RSD
values is presented in Tables 6-3a and 6-3b. The RSD
values are presented for TEQPCB, TEQD/F, and total TEQ
in Table 6-3a, and a summary by sample type is
presented in Table 6-3b, along with the minimum
R value, the maximum R value, and the mean R value
for each set of TEQ results and sample types. In terms of
sample type, the reference method had the most precise
data for the environmental sample TEQD/F results, with a
mean RSD value of 12%. This was followed closely by
environmental sample TEQPCB and total TEQ results,
which both had mean RSDs of 13%.  In terms of TEQ
values, the reference method had the most precise data
for the total TEQ values, with a mean overall RSD of
13%.  Overall RSD values ranged from l%to 119%.
Precision was significantly worse for certified blanks
and blank samples (e.g., samples that contained spikes of
only one analyte set and were blank for the other
analytes) as might be expected due to the very low levels
detected in these samples.

 6.5   Comparability to Characterization Data
To assess comparability, reference laboratory D/F data
for environmental samples were plotted against the
characterization data that was generated by Battelle prior
to the demonstration. Characterization data were
obtained as part of the process to verify homogenization
of candidate soil and sediment samples as described in
Chapter 5 and reported in Table 4-5. It should be noted
that second column confirmations of 2,3,7,8-TCDF
results were not performed during characterization
analyses; therefore, characterization TEQs are biased
high for samples where a large concentrations of non-
2,3,7,8-TCDF coeluted with 2,3,7,8-TCDF on the DB-5
column. Characterization samples also were not
rigorously evaluated to ensure that high concentration
extracts were diluted sufficiently so that all peak areas
were less than the peak areas of the highest calibration
standard. In spite of these differences between reference
and characterization analyses, the results had fairly high
correlation (R2 = 0.9899) as demonstrated in Figure 6-1.
6.6    Performance Summary
This section provides a performance summary of the
reference method by summarizing the evaluation of the
applicable primary objectives of this demonstration
(accuracy, precision, and cost) in Table 6-4. A total of
209 samples was analyzed for seventeen
2,3,7,8-substituted D/F and 12 PCBs over an eight-
month time frame  (April 27 to December 20, 2004).
Valid results were obtained for all 209 PCB analyses,
while 207 valid results were obtained for D/F. The
TEQD/F and total TEQ results for samples Ref 197 (ERA
PCB 100) and Ref 202 (LCG CRM-529) were omitted
as outliers because it appeared that these two  samples
were switched during preparation after observing results
of the replicates and evaluating the congener profiles of
these two samples. The demonstration sample set
provided particular challenges to the reference
laboratory in that there was a considerable range of
sample concentrations for D/F and PCB. This caused
some difficulty in  striving for low MDLs in the presence
of high-level samples. The range of concentrations in the
demonstration sample set also required the laboratory to
modify standard procedures, which contributed to
increased cost and turnaround time delay. For example,
an automated sample  cleanup system could not be used
due to carryover from high-level samples; instead, more
labor-intensive manual cleanup procedures were used;
glassware required extra cleaning and proofing before
being reused; cleanup columns sometimes became
overloaded from interferences and high-level  samples,
causing low recoveries so that samples had to be
re-extracted or cleanup fractions had to be analyzed for
the lost analytes; and method blanks often showed trace
levels of contamination, triggering the repeat of low-
level samples.

Because the reference method was not to be altered
significantly for this demonstration, the reference
laboratory was limited in its ability to adapt the trace-
level analysis to higher level samples. In spite of these
challenges, the quality of the data generated met the
project goals. The main effect of the difficulties
associated with these  samples was on schedule and cost.
                                                    37

-------
Table 6-3a. Objective P2 Precision - Relative Standard Deviation
Sample Type
Environmental
Extract
Performance
Evaluation
Sample ID
Brunswick #1
Brunswick #2
Brunswick #3
Midland #1
Midland #2
Midland #3
Midland #4
NCPCBSite#l
NC PCB Site #2
NC PCB Site #3
Newark Bay #1
Newark Bay #2
Newark Bay #3
Newark Bay #4
RaritanBay#l
Raritan Bay #2
Raritan Bay #3
Saginaw River #1
Saginaw River #2
Saginaw River #3
Solutia#l
Solutia #2
Solutia #3
Titta. River Soil* 1
Titta. River Soil #2
Titta. River Soil #3
Titta. River Sed #1
Titta. River Sed #2
Titta. River Sed #3
WinonaPost#l
Winona Post #2
Winona Post #3
Envir Extract #1
Envir Extract #2
Spike #1
Spike #2
Spike #3
Cambridge 5183
Cambridge 5184
ERA Aroclor
ERA Blank
ERA PAH
ERA PCB 100
ERA PCB 10000
ERA TCDD 10
ERA TCDD 30
LCG CRM-529
NIST 1944
Wellington WMS-01
RSD for TEQPCB
(%)
8
o
J
5
4
10
4
77
21
21
25
7
2
6
1
6
3
o
J
8
7
60
36
4
11
7
9
12
19
14
13
13
4
9
71
83
119
1
4
7
3
44
62
83
4
7
60
39
14
4
5
RSD for TEQD/F
(%)
6
16
8
9
6
6
9
15
2
12
28
22
6
12
5
2
5
25
19
19
13
7
5
6
10
26
27
37
9
2
9
4
50
2
6
5
13
19
4
6
65
27
65 a
91
5
6
T
9
3
RSD for Total TEQ (%)
6
16
8
9
6
6
10
20
21
24
25
20
6
11
4
1
4
23
18
19
13
7
5
5
10
26
26
37
8
2
9
4
50
2
9
3
4
9
2
43
61
30
3
7
5
6
1
7
3
"Does not include sample excluded due to sample preparation error.
                                                       38

-------
Table 6-3b. Objective P2 Precision - Relative Standard Deviation (By Sample Type)
Sample Type
Environmental
Extract
PE
Overall
RSD for TEQprR (%)
N
32
5
12
49
MIN
1
1
3
1
MAX
77
119
83
119
MED
8
71
11
8
MEAN
13
56
28
21
RSD for TEQn;F (%)
N
32
5
12
49
MIN
2
2
2
2
MAX
37
50
91
91
MED
9
6
7
9
MEAN
12
15
25
16
RSD for Total TEQ (%)
N
32
5
12
49
MIN
1
2
1
1
MAX
37
50
61
61
MED
10
4
7
8
MEAN
13
14
15
13
Table 6-4. Reference Method Performance Summary - Primary Objectives
Objective
PI: Accuracy
P2: Precision
P7: Cost
Performance
Statistic
Number of data points
Median Recovery (%)
Mean Recovery (%)
Number of data points
Median RSD (%)
Mean RSD (%)
TEQPCB
8
96
96
49
8
21
TEQD/F
8
106
125
49
9
16
Total TEQ
10
94
94
49
8
13
209 samples were analyzed for 17 D/F and 12 PCBs. Total cost was $398,029.
D/F cost was $213,580 ($1,022 per sample) and PCB cost was $184,449 ($883 per sample)
        16000
                          5000        10000        15000
                        Characterization Data (TEQ D/F pg/g)
20000
Figure 6-1. Comparison of reference laboratory and characterization D/F data
for environmental samples.
                                                39

-------
                                             Chapter 7
                      Performance of Abraxis Coplanar PCB ELISA Kit
7.1    Evaluation of Coplanar PCB ELISA Kit
       Performance
It should be noted that this technology may not directly
correlate to HRMS TEQPCB in all cases because it is
known that the congener responses and cross-reactivities
to the kit are not identical to the World Heath
Organization toxicity equivalency factors that are used
to convert congener HRMS concentration values to
TEQPCB. The effect of cross-reactivities may contribute
to this technology reporting results that are biased high
or low compared to HRMS TEQPCB results.  Therefore,
the Abraxis kit should not be viewed as producing an
equivalent measurement value to HRMS TEQPCB but as a
screening value to approximate HRMS TEQPCB
concentration. It has been suggested that correlation
between the Abraxis TEQPCB results and HRMS TEQPCB
results could be improved by first characterizing a site
and calibrating the Abraxis results to HRMS results.
Subsequent analysis using the Abraxis kit for samples
obtained from this site may then show better correlation
with the HRMS TEQPCB result. This approach was not
evaluated during this demonstration.

The following sections describe the performance of the
Coplanar PCB ELISA Kit, according to the primary
objectives for this demonstration. The developer and
reference laboratory data are presented in Appendix D.
The statistical methods used to evaluate the primary
objectives are described in Section 4.7. Detailed data
evaluation records can be found in the DER.

7.1.1   Evaluation of Primary Objective PI:
       Accuracy
A summary of the Abraxis Coplanar PCB ELISA Kit
percent recovery (R) values is presented in Table 7-1.
The description of how R values were calculated is pre-
sented in Section 4.7.1. The Rvalues were calculated by
comparing Abraxis kit values to the certified TEQPCB
values. The minimum R value, the maximum R value,
the median R value, and the mean R value were 16%,
204%, 79%, and 92%, respectively. As presented in
Table 7-1, there were only six R values that could be
calculated from the 12 PE samples because many of the
PE samples were reported as nondetects by Abraxis.
Reporting these samples as nondetects in some cases
was accurate. For example, ERA TCDD 10, ERA TCDD
30, ERA PAH, and Cambridge 5183 contained only
spiked D/Fs or PAHs, or it had TEQ PCB concentrations
below Abraxis's 6.25 pg/g reporting limits. These
samples should not have been detections for PCBs by
the Abraxis kit. The lack of reported data for ERA PCB
100 (TEQPCB =11 pg/g) indicates that this sample
Table 7-1. Objective PI Accuracy: Percent Recovery
PE Sample
Number
1
2
3
4
5
6
7
8
9
0
1
2
PE Sample
Description
Cambridge 5 183*
LCG CRM-529
Wellington
WMS-01
Cambridge 5184
NIST 1944
ERA TCDD 10*
ERA TCDD 30*
ERA PAH*
ERA PCB 100
ERA PCB 10000
ERA Aroclor
ERA Blank*
All Performance Evaluation
Samples
% Recovery
TEQprR
NA
138
204
16
86
NA
NA
NA
NA
73
32
NA
NUMBER
MLN
MAX
MEDIAN
MEAN
6
16
204
79
92
NA = not applicable (insufficient results to calculate R
  values).
* Indicates sample did not contain PCBs or the levels of
  PCBs were below the Abraxis reporting limits.
                                                  40

-------
contained PCB concentrations that were below the
capabilities of the Abraxis kit.

7.1.2  Evaluation of Primary Objective P2:
       Precision
A summary of the Abraxis Coplanar PCB ELISA Kit
RSD values is presented in Tables 7-2a and 7-2b.  The
description of how RSD values were calculated is
presented in Section 4.7.2. Low RSD values (< 20 %)
indicate high precision. The RSD values are presented in
Table 7-2a for each sample where Abraxis reported
values for three or more of the replicate samples. A
summary by sample type is presented in Table 7-2b,
along with the minimum R value, the maximum R value,
and the mean R values. In terms of sample type, the
Abraxis kit values had the most precise data for the
extract results, with a mean RSD value of 23%. Overall
RSD values ranged from 4% to 172%, with a mean RSD
of 52% and a median RSD of 45%.

7.1.3  Evaluation of Primary Objective P3:
        Comparability
The description of the statistical analyses used in the
comparability evaluations are described in Section 4.7.3.
The comparability of the Abraxis and reference
laboratory data was assessed by calculating RPD values
for TEQPCB, as presented in Table 7-3. The summary
statistics presented in Table 7-3 provide an overall
assessment of the RPD values that is reported by sample
type.  The Abraxis values agreed best with the reference
laboratory PCB measurements for PE samples, with a
median RPD value of 20%. The median, minimum, and
maximum RPD values for all samples were -129%,
-200%, and 200%, respectively. RPD values that are
between positive and negative 25% indicate good
agreement between the developer and reference
laboratory data.  Abraxis reported 12 of 114 RPD  results
within positive and negative 25%. This evaluation
indicates that the Abraxis PCB results were generally
higher than the reference laboratory (as evidenced by
three of the four median values that were negative).

Comparability was also assessed using the interval
approach discussed in Section 4.7.3.  The agreement
when sorting the developer and reference laboratory
TEQPCB results into four intervals (< 50 pg/g, 50 to
500 pg/g, 500 to 5,000 pg/g, and >5,000 pg/g) are
described in Table 7-4. The agreement between the
developer and reference laboratory TEQPCB data was
83%. Interval reporting addresses the question whether a
value reported by the technology would result in the
same decision of what to do next with the sample if it
was analyzed by the reference method. This interval
assessment table indicates that 17% of the time, the
Abraxis result would have indicated a different interval
(and therefore a different decision to be made about the
sample) than if it was analyzed by the reference
laboratory, based on the concentrations chosen for the
intervals.

The ERA blank samples contained levels  of PCBs that
were below the reporting limits of the developer
technologies (see Table 4-4 certified value: 0.01 pg/g
TEQPCB). The Abraxis reported concentrations were
compared with the reference laboratory reported data for
these samples in Table 7-5. Abraxis reported two of the
eight TEQPCB values as detections (6.30 and 16 pg/g) and
six results were reported as nondetects (< 6.25 or
< 6.30). As such, 75% agreed with the reference
laboratory results.  It should be noted that the reference
laboratory data presented in Table 7-5 were calculated
with nondetect values assigned a zero concentration.
When applying the TEQ calculation method of assigning
nondetects with a concentration of one-half the SDL, the
reference data increased, but the  conclusions regarding
agreement with the developer data remain the same.
                                                    41

-------
Table 7-2a.  Objective P2 Precision - Relative Standard Deviation
Sample Type
Environmental
Extracts
Performance Evaluation
Sample ID
Brunswick #1
Brunswick #2 *
Brunswick #3
Midland #1 *
Midland #2 *
Midland #3 *
Midland #4 *
NC PCB Site #1
NC PCB Site #2
NC PCB Site #3
Newark Bay #1 *
Newark Bay #2 *
Newark Bay #3 *
Newark Bay #4 *
RaritanBay#l *
Raritan Bay #2 *
RaritanBay#3 *
Saginaw River # 1
Saginaw River #2
Saginaw River #3 *
Solutia #1 *
Solutia #2
Solutia #3
Titta. River Soil #1
Titta. River Soil #2 *
Titta. River Soil #3 *
Titta. River Sed #1 *
Titta. River Sed #2 *
Titta. River Sed #3 *
WinonaPost#l *
WinonaPost#2*
WinonaPost#3*
Envir. Extract # 1 *
Envir. Extract #2 *
Spike #1 *
Spike #2
Spike #3
Cambridge 5 183 *
Cambridge 5184
ERA Aroclor
ERA Blank *
ERA PAH*
ERA PCB 100
ERA PCB 10000
ERA TCDD 10 *
ERA TCDD 30 *
LCG CRM-529
NIST 1944
Wellington WMS-01
Relative Standard Deviation (%)"
22
NA
43
NA
NA
NA
NA
NA
NA
NA
95
15
43
NA
NA
NA
45
67
35
NA
NA
NA
NA
NA
48
42
NA
NA
NA
62
58
76
9
14
NA
44
25
NA
33
172
NA
46
NA
67
NA
NA
107
75
4
NA = not applicable (i.e., one or more of the replicates were reported as a nondetect value).
* Indicates sample did not contain PCBs or the levels of PCBs were below the Abraxis reporting limits.
a Three or four replicate results were used to calculate the RSD values.
                                                        42

-------
Table 7-2b. Objective P2 Precision - Relative Standard Deviation (By Sample Type)
Sample
Type
Environ.
Extract
PE
Overall
Relative Standard Deviation (%)
N
13
4
7
24
MIN
15
9
4
4
MAX
95
44
172
172
MEAN
50
23
72
52
MEDIAN
45
20
67
45
Table 7-3.  Objective P3 - Comparability Summary Statistics of RPD
Sample Type
Environmental
Extract
PE
Overall
TEQpr« RPD (%)
N
74
16
24
114
MIN
-200
-196
-119
-200
MAX
200
112
191
200
MEDIAN
-155
-98
20
-129
Table 7-4.  Objective P3 - Comparability Using Interval Assessment
Agreement
Number Agree
% Agree
Number Disagree
% Disagree
TEQprB
174
83
35
17
Table 7-5.  Objective P3 - Comparability for Blank Samples
Rep
1
2
3
4
5
6
7
8
% agree
TEQPCB
Abraxis
(Pg/g)
<6.30
6.30
16
<6.25
<6.25
<6.30
<6.25
<6.25
Ref Lab"
(Pg/g)
J0.0243b
0.00385
0.00277
J0.042
J0.0229
J0.0191
J0.0325
J0.0225
Agree?
Yes
No
No
Yes
Yes
Yes
Yes
Yes
75% (6 of 8)
1 All nondetect and EMPC values were assigned a zero concentration for the reference laboratory TEQ calculation.
3 J flag was applied to any reported value between the SDL and the lowest level calibration.
                                                     43

-------
7.1.4  Evaluation of Primary Objective P4:
       Estimated Method Detection Limit
It should be noted that these calculations did not strictly
follow the definition presented in the Code of Federal
Regulations (i.e., t-value with  6 degrees of freedom).
Since detections were not reported for all seven replicate
samples, the degrees of freedom and statistical power of
the analysis were reduced accordingly. The only
approach that led to the use of the definitional
calculation with 6 degrees of freedom required special
treatment of the nondetect values (i.e., assigning values
that were one-half or equal to the nondetect value).
However, these calculations are provided as estimated
method detection limits (EMDL) to give the reader a
sense of the detection capabilities of the technology.

The EMDL of the Abraxis Coplanar PCB ELISA Kit
was determined by assessing the values that Abraxis
reported for two PE samples: Wellington WMS-01 and
Cambridge 5183. Extract Spike #1 was also included in
the demonstration with seven replicates, but this sample
was not appropriate for use in  the EMDL calculation for
the Abraxis kit because it was  spiked with only
0.5 pg/mL of 2,3,7,8-TCDD. As shown in Table 7-6
because some of the results for the samples were
nondetects, the TEQPCB MDL was calculated in three
ways: by setting nondetect values to zero, by setting
nondetect values to half of the reporting limit value, and
by setting nondetect values to the reporting limit value
itself. For Cambridge 5183 samples, Abraxis reported
only two samples as actual values (as opposed to
nondetects), so an MDL could not be calculated for
those samples using the approach of setting nondetect
values to zero. The MDLs calculated using the
Wellington WMS-01 replicates with the nondetects set
to zero resulted in an MDL calculation with 2 degrees of
freedom, yielding an MDL of 5.7 pg/g; this value is an
outlier due to the few samples with detected values. The
MDLs from the remaining Cambridge 5183 and
Wellington WMS-01 samples ranged from 12 to 31 pg/g
TEQ. The detection limit reported by Abraxis in the
demonstration plan was 6.25 pg/g TEQPCB.

7.1.5  Evaluation of Primary Objective P5: False
       Positive/False Negative Results
The summary of false positive/false negative results is
presented in Table 7-7. Abraxis reported many more
false positive results (74) than false negative  results (14),
relative to the reference laboratory's reporting of
samples above and below 6.25 pg/g TEQPCB.  This
analysis indicated that the Abraxis kit had more of an
issue with correctly reporting positive results than it did
with reporting negatives around the reporting limit of
6.25 pg/g TEQPCB. Abraxis reported significantly fewer
false positives (8%) and false negatives (3%) around 50
pg/g TEQPCB. Given the calculated EMDLs presented in
Section 7.1.4 and the false positive rate of 35% at
6.25 pg/g TEQPCB, this evaluation indicates that the
Abraxis kit could be an effective screening tool for
sample concentrations above and below 50 pg/g TEQPCB.
Table 7-6.  Objective P4 - Estimated Method Detection Limit
Statistic
Degrees of
Freedom
Standard Deviation
(pg/g TEQPCB)
EMDL
(pg/gTEQPCB)
Wellington WMS-01
Nondetect
values set to
zero
2
0.8
5.7
Nondetect
values set to 1A
value
6
9.8
31
Nondetect values
set to reported
value
6
8.1
25
Cambridge 5183
Nondetect
values set
to zero
NA
Nondetect
values set to 1A
value
6
5.0
16
Nondetect
values set to
reported value
6
3.8
12
                                                     44

-------
Table 7-7.  Objective P5 - False Positive/False
Negative Results
Rate
False Positive
False Negative
TEQPCB
Around 6.25 pg/g
35%
(74 out of 209)
7%
(14 out of 209)
Around 50 pg/g
8%
(16 out of 209)
3%
(6 out of 209)
7.1.6  Evaluation of Primary Objective P6: Matrix
       Effects
Six types of potential matrix effects were investigated:
(1) measurement location (field vs. laboratory), (2) matrix
type (soil vs. sediment vs. extract), (3) sample type (PE
vs. environmental vs. extract), (4) PAH concentration,
(5) environmental site, and (6) known interferences. A
summary of the matrix effects is provided in the bullets
below, followed by a detailed discussion:

•  Measurement location: 1 sample set statistically
  different
•  Matrix type: none
•  Sample type: none
•  PAH concentration: none
•  Environmental site: none
•  Known interferences: slight

An equal number of sample replicates were analyzed by
Abraxis  during the field demonstration and in its
laboratories. A one-way ANOVA was performed on
samples that had at least one detected replicate analyzed
in the field and in the laboratory to determine if
performance was affected by the samples being analyzed
in the field. A p-value less than 0.05 in Table 7-8
indicates that the mean of samples analyzed in the field
was significantly different from the mean of those
analyzed in the laboratory. Only one sample set (5%
overall)  showed statistically significant location effects,
and this  was a PE sample that was spiked with PAHs only
and should not have been a detection for TEQPCB.  In
Table 7-9, precision summary values are presented by
matrix type. A one-way ANOVA model  was used to test
the effect of soil vs. sediment vs. extract on RSD. These
tests showed no significant effect on RSD for TEQPCB. In
Table 7-10, precision summary values are presented by
PAH concentrations for environmental samples only. A
one-way ANOVA model was used to test the effect of
PAH concentration on RSD. These tests showed no effect
for TEQPCB. The summary of RSD values segregated by
sample type is presented in Table 7-2b.  A one-way
ANOVA model was used to test the effect of sample type
on RSD. These tests showed no significant effect on RSD
for TEQPCB. The mean RSD for extracts (23%) was less
than environmental (50%) and PE (72%) samples, but the
number of data points for evaluation (4) made this
difference not significant. Based on the comparability
(RPD) results, Abraxis's results were not more or less
comparable for one particular environmental site,
suggesting that matrix effects were not dependent on
environmental site.

The effect of known interferences was also assessed by
evaluating the results of PE materials that contained one
type of contaminant (D/F or PAHs) but not PCBs.
Table 7-11 summarizes the TEQPCB values reported by
Abraxis in the PE samples that did not contain PCBs,
along with the percent recovery values (from Table 7-1).
For the ERA PAH sample, Abraxis reported a mean
TEQPCB value of 12.1 pg/g. Only one of the four
replicates from each of the D/F-only spiked PE samples
was reported as detections for PCBs by Abraxis.

7.1.7  Evaluation of Primary Objective P7:
       Technology Costs
Evaluation of this objective is fully described in
Chapter 8, Economic Analysis.

7.2    Observer Report: Evaluation of Secondary
       Objectives
The Abraxis Coplanar PCB ELISA Kit is a screening
method specifically for coplanar PCBs.  This test kit has
high specificity for PCBs with higher TEFs such as PCB
126 and PCB 169. During the field demonstration, 2 g of
each sample were extracted in an acetone/hexane mix and
then underwent an oxidation cleanup using concentrated
sulfuric acid. Samples were evaporated to dryness,
redissolved using methanol, and then diluted 50/50 with
deionized water. The sample was added to a well of a
pretreated microtiter plate along with an antibody specific
for the coplanar PCBs to begin the competitive ELISA.
After a 30-minute incubation period, a coplanar PCB
ligand labeled with an enzyme was added and the plate
was allowed to incubate for 90  minutes. The samples
were washed and a substrate was added that caused a
                                                    45

-------
Table 7-8.  Objective P6 - Matrix Effects Using Descriptive Statistics and ANOVA Results Comparing In-Field to
Laboratory Analysis
Sample Type
Environmental
PE
Sample
Brunswick #1
Brunswick #3
Newark Bay #1
Newark Bay #2
Newark Bay #3
Raritan Bay #3
Saginaw River #1
Saginaw River #2
Titta. River Soil #2
Titta. River Soil #3
WinonaPost#l
Winona Post #2
Winona Post #3
Cambridge 5 184
ERA Aroclor
ERA PAH
ERA PCB 10000
LCG CRM-529
NIST 1944
Wellington WMS -01
Location
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
Field
Lab
TEQPrn
N
2
2
2
2
2
2
2
1
2
2
2
1
2
2
2
2
2
2
1
2
1
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
1
2
1
2
Mean(SD)
(P2/2)
18.8 (6.0)
19.0 (4.0)
368.8 (79.5)
1,82.8(33.2)
37.8(38.5)
16.1 (0.7)
13.8 (2.5)
16.2
51.9(18.5)
28.9 (4.8)
19.0 (5.7)
9.0
67.5 (35.4)
34.3 (33.2)
35.7(15.1)
26.6 (6.2)
10.4 (5.8)
19.2(6.3)
10.5
11.1(6.5)
21.0
100.8 (6.0)
140.0(106.1)
92.4(14.1)
83.8 (76.0)
47.7 (19.6)
176.3 (51.3)
126.7 (47.2)
2,170.0 (2941.6)
207.4 (24.2)
8.9(0.1)
18.5
190.0
1125.0(70.7)
632.3 (873.6)
541.4 (648.6)
65.0
20.7(10.1)
22.0
21.2(0.9)
p-Value Comparing Field
to Laboratory
0.9724
0.0928
0.5103
0.5672
0.2313
0.3857
0.4353
0.5145
0.2851
0.9521
0.0586
0.5932
0.5822
0.4200
0.4450
0.0115a
0.0588
0.9168
0.1734
0.5883
"Bold indicates in-field measurement statistically different from the laboratory measurement at the p < 0.05 significance level.
                                                         46

-------
Table 7-9. Objective P6 - Matrix Effects Using RSD as a Description of Precision by Matrix Type
Matrix Type
Soil
Sediment
Extract
Overall
RSD for TEQPrB (%)
N
10
10
4
24
MIN
33
4
9
4
MAX
172
95
44
172
MED
71
44
23
52
MEAN
60
43
20
45
Table 7-10. Objective P6 - Matrix Effects Using RSD as a Description of Precision by PAH Concentration Levels
(Environmental Samples Only)
PAH Concentration Level
(ng/g)
100,000+
10,000-99,999
1,000-9,999
0-999
Overall
(Environmental Samples Only)
RSD for TEQprR (%)
N
1
4
6
2
13
MIN
43
22
15
42
15
MAX
43
76
95
48
95
MED
43
55
50
45
50
MEAN
43
60
44
45
45
Table 7-11. Objective P6 - Matrix Effects Using PE Materials
PE Sample
ERA PAH
ERA TCDD 10
ERA TCDD 30
% Recovery for Spiked
Analytes
NAa
NA
NA
Mean TEQPCB (pg/g) Reported by Abraxis
when PCBs were not spiked in the PE Sample
12.1
16.2b
8.8b
 NA = not applicable; percent recovery value could not be calculated.
 Three replicates were reported as nondetects.
color reaction (inversely proportional to amount of PCB).
The color reaction was stopped and stabilized after 20 to
30 minutes. The plate was then read on a plate reader and
the data compiled.

Some of the samples included in the demonstration test
were received as toluene extracts. These extracts went
directly to the oxidation step and were handled like the
other samples from that point on.

All steps of both extraction and ELISA analysis were
observed during the demonstration. The observed method
and the method described in the demonstration plan  were
similar, with one exception. In the observed method, an
extra step was added when pipetting the sample into the
plate, where the samples and antibody were first pipetted
into an uncoated plate and then transferred by
multichannel  pipette to the coated plate.

7.2.7  Evaluation of Secondary Objective SI: Skill
       Level of Operator
In the field demonstration, Fernando Rubio performed all
assays with the Abraxis kit.  Mr. Rubio has an M.S.
degree  in biochemistry and is one of the kit's designers.
Dr. Gary Hinshaw assisted with sample weighing and
                                                     47

-------
data analysis. Dr. Hinshaw has a Ph.D. in environmental
science.

The developer suggests users have some experience with
the 96-well ELISA, as well as training specifically in use
of this kit. The developer will not sell this technology to
users without a background in chemistry or biology.
Upon observation, it would be beneficial to have a
background in both wet chemistry and some assay work
before this kit is used. Technical ability is more important
than education level,  since good analytical and lab skills
are important when using this kit.

The instructions included with the kit are generally
detailed and explain most of the extraction and assay
steps. The steps described are clearly defined, affording
easy understanding of each step in the process. The
instructions could provide some additional clarifications.
For example, after the samples are extracted in the
acetone/hexane mix, no specific final volume is given for
the level during the evaporation step. The instructions
also don't describe lengths of mixing times for the
samples once they are in the wells or storage conditions
for samples and extracts. The instructions also make the
assumption that the user has some experience with ELISA
and the equipment and techniques that are listed in the
instructions. These assumptions could make the
instructions somewhat less useful to  a complete novice.
There are some techniques, such as how to dry the plates,
that don't come across well in the instructions, but they
are easily understood if the user has experience with
assays and typical laboratory operations.

The kit is a standard 96-well format and requires only a
constant room temperature for all activities. The
extraction method and assay are both easy to use and
should present minimal problems in the field for a user
with some experience with the kit or, at a minimum, with
familiarity with immunoassay techniques. The difficulty
lies in the fact that the weights and volumes used have to
be measured accurately. The weight must be recorded
accurately since the final calculations are based on the
weight. The volumes  of sample and reagent added to the
assay are also critical steps, since adding incorrect
amounts or having large variations of volume can cause
problems with the calibration curve due to inaccurate
results for standards,  as well as variability between
duplicate sample results.
7.2.2   Evaluation of Secondary Objective S2:
        Health and Safety Aspects
The majority of the waste from Abraxis came from the
sample extraction. The bulk of the remaining waste was
concentrated sulfuric acid used in the cleanup step. The
amount of sulfuric acid waste generated is 4 mL of acid
per sample per oxidation step. The final amount of waste
depends on the "cleanliness" of the samples and is
variable. The waste  generated by the assay itself is
relatively low in volume, being composed mainly of the
washes from the wells and the sample/enzyme conjugate
mix. The amount of waste will increase with the disposal
of the remaining extract. Solid waste will also be
generated with the soil sample itself, glass tubes, plastic
jars used for extraction, plates and pipette tips, as well as
paper towels and other miscellaneous lab supplies.

A complete inventory of the waste generated was
performed after the demonstration for the processing of
116 samples by Abraxis and the following was recorded.
None of the containers was verified as full. Note that this
summary does not include the samples that were analyzed
in the Abraxis laboratories:

(1)  One 5-gallon container labeled "high concentration"
    containing solid waste such as wooden tongue
    depressors (used as a disposable  scoop for sample
    aliquoting), weighing paper, and gloves.

(2)  One 5-gallon container labeled "low concentration"
    containing solid waste such as personal protective
    equipment, weighing papers, and wooden tongue
    depressors.

(3)  One 5-gallon container filled with hundreds of vials
    containing water and methanol and vials in plastic
    jugs.

(4)  One 5-gallon container with a 1-gallon plastic jug
    containing rinse water waste.

(5)  One 5-gallon container filled with several hundred
    screw cap vials. Vials contained several milliliters of
    spent  sulfuric acid.

(6)  One 5-gallon container with sulfuric acid
    contaminated cardboard and plastic gloves.
                                                     48

-------
(7) One 5-gallon container with 200 mL concentrated
    sulfuric acid and 18 vials containing spent sulfuric
    acid.

(8) One used broken glass container.

The reader should be advised that, although no difficulties
were encountered during this project, difficulties could
arise with disposal of dioxin-contaminated waste.

7.2.3 Evaluation of Secondary Objective S3:
     Portability
A mobile laboratory was used during the demonstration.
The need for power, nitrogen tanks, fume hood, plate
reader, and a sink preclude this technology from being
used in the field with anything less than a trailer, but a
fully equipped mobile laboratory was more infrastructure
than was needed for this technology.  It took
approximately half of a day to set up the trailer for this
demonstration. The setup mainly involved ensuring that
all of the equipment was in place and that the nitrogen
flow was hooked up correctly. Some reagents needed to
be kept cold; however, this was accomplished by keeping
the reagents in a cooler on ice, eliminating the need for a
refrigerator.

While the kit is not difficult to use, there are a few areas
that could cause problems when using this technology in
the field. Users should consider these factors in their
planning.  The first is the plate reader that is necessary to
read the assay results. During the demonstration, a new
plate reader was being used by Abraxis, and there was
quite a bit of difficulty in connecting the reader to the
computer. It was finally decided that the  data should be
printed from the reader and not collected by using the
computer connection. A malfunctioning reader will
prevent any samples from being run, so this could be a
major problem in working in the field if the plate reader
does not work properly. The second area where problems
could develop in the field is the need for nitrogen in the
sample extraction process. The requirement of having
tanks of nitrogen could be a limiting factor when taking
the technology to the field since nearly two nitrogen tanks
(2,000 psi gauge each) were consumed over the course  of
three days of testing. The final area to consider when
using this technology in the field is having enough
supplies. The type of sample being extracted, the level of
contamination, and other factors could affect how many
times a sample would need to undergo both the oxidation
cleanup and how many times a sample would have to be
diluted.  The additional cleanups and dilutions would
increase the number of kits and glassware needed to
complete all samples.

Differences in reported results due to measurement
location (in field vs. laboratory) are described in
Section  7.1.6.

7.2.4 Evaluation of Secondary Objective S4:
     Throughput
During the demonstration, 116 samples were processed
by Abraxis. This was accomplished in about three full
working days, with one person doing the majority of the
extraction on the first two days and a second person doing
the data workup on the second and third day. Sample
throughput was approximately 40 samples per day during
the field demonstration.

According to the developer, one batch of 14 to
16 samples would take approximately half of a working
day for one person to process. This is dependent upon
how clean the samples are and the experience of the user.
A novice with the kit would take slightly longer per
batch, as would samples that require additional cleanup.
The first results from a small batch with an experienced
user would be available approximately five hours after
sample processing began. The most time-consuming steps
of the extraction and analysis of the samples are the
necessary incubation periods. The samples are extracted
for 1 hour, then incubated for a total of 2!/2 hours before
being analyzed. Based on observations during the
demonstration, throughput might more realistically be
slightly lower than the developer's assessment with a 14
to 16 sample batch  size taking a full day to complete all
of the steps from extraction to final analysis. The number
of samples that could be analyzed by one kit is highly
variable. The kit includes one coated 96-well plate, but
factors such as the number of sample replicates and QC
decisions can have  an impact on the number of samples
that could be processed using the one plate.

7.2.5 Miscellaneous Observer Notes
Abraxis is a U.S.-based company and can provide
training on site, as well as extensive phone support for the
technology.
                                                     49

-------
As described in Chapter 2, the Abraxis ELISA kit comes
with one 96-well microtiter plate, the coplanar PCB
antibody solution, seven coplanar PCB standard solutions
(0, 25, 50, 100, 250, 500, 1,000 pg/mL), the coplanar
PCB-HRP enzyme conjugate, the diluent/zero standard,
color solution, stopping solution and the concentrate
washing buffer. The user must supply micropipettes (i.e.,
Eppendorf™), a plate reader capable of reading at 450-
nm, distilled or deionized water, reagent grade methanol,
transfer pipettes, disposable glass tubes with Teflon™
caps, and Parafilm™. The kit itself is an off-the-shelf
product, and availability is only limited by the amount of
stock being kept by the company.

The extraction method used by Abraxis during the
demonstration is not the only extraction option available.
Users purchasing the kit can use their own extraction
method. However, in this case, all supplies for alternate
extraction methods need to be provided by the user.

A standard curve, negative control, and positive control
are required by the kit instructions, but the use of spikes,
blanks, and any additional QC are determined by the user.
For the demonstration, the developer analyzed all samples
in triplicate, including spikes and blanks. Results were
reported based on the mean value of the triplicates. The
developer also recommends that results requiring some
type of regulatory action be checked using HRMS. The
kit is not meant to give an exact correlation to HRMS but
serve as a screen for low and high values.
                                                      50

-------
                                               Chapter 8
                                          Economic Analysis
During the demonstration, the coplanar PCB ELISA kit
and the reference laboratory analytical methods were
each used to perform more than 200 analyses of dioxin-
contaminated samples, including samples with a variety
of distinguishing characteristics such as high levels of
polychlorinated biphenyls and PAHs. The purpose of the
economic analysis was to estimate the total cost of
generating results by using the coplanar PCB ELISA kit
and then comparing this cost to that for the reference
method. This cost estimate also is provided so that
potential users can understand the costs involved with
using this technology.

This chapter provides information on the issues and
assumptions involved in the economic analysis
(Section 8.1), discusses the costs associated with using
the coplanar PCB ELISA kit (Section 8.2), discusses the
costs associated with using the reference method
(Section 8.3), and presents a comparison of the
economic analysis results for the coplanar PCB ELISA
kit and the reference method (Section 8.4).

8.1    Issues and Assumptions
Several factors affect sample measurement costs.
Wherever possible in this chapter, these factors are
identified in such a way that decision-makers can
independently complete a project-specific economic
analysis. The following five cost categories were
included in the economic analysis for the demonstration:
capital equipment, supplies, support equipment, labor,
and investigation-derived waste (IDW) disposal. The
issues and assumptions associated with these categories
and the costs not included in the analysis are briefly
discussed below. The issues and assumptions discussed
below only apply to the coplanar PCB ELISA kit unless
otherwise stated.
8.1.1  Capital Equipment Cost
The capital equipment cost was the cost associated with
the purchase of the coplanar PCB ELISA kit. Compo-
nents of the coplanar PCB ELISA kit are presented in
detail in Chapters 2 and 7. Abraxis offers a purchase
option for potential coplanar PCB ELISA kit users. The
purchase price information was obtained from a standard
price list provided by Abraxis.

8.1.2  Cost of Supplies
The cost of supplies  was estimated based on the supplies
required to analyze all demonstration samples using the
coplanar PCB ELISA kit that were not included in the
capital equipment cost category. Examples of such
supplies include filters, cleanup columns, gas cylinders,
solvents,  and distilled water. The supplies that Abraxis
used during the demonstration fall into two general
categories: consumable (or expendable) and reusable.
Examples of expendable supplies utilized by Abraxis
during the demonstration include hexane, acetone,
distilled water, sulfuric acid, nitrogen cylinders, glass
disposable pipettes, pipette tips, and glass disposable
extraction tubes. Examples of reusable supplies include a
microplate reader, digital balance, vortex mixer, and
water bath. It should be noted that this type of equipment
may or may not be already owned by a potential
coplanar PCB ELISA kit user; however, for this
economic analysis, an assumption was made that the user
does not possess these items.

The purchase price of these supplies was either obtained
from a standard price list provided by Abraxis, or it was
estimated based on price quotes from independent
sources.

8.1.3  Support Equipment Cost
This section details the equipment used at the
demonstration  such as the mobile laboratory, fume hood,
and laptop computer required by the technology. Costs
                                                    51

-------
for these items will be reported per actual costs for the
demonstration.

8.1.4  Labor Cost
The labor cost was estimated based on the time required
for work space setup, sample preparation, sample
analysis, and reporting. For the demonstration,
developers reported results by submitting a chain-of-
custody (COC)/results form. The measurement of the
time required for Abraxis to complete the 116 sample
analyses during the demonstration (50 labor-hours) was
estimated by the sign-in log sheets that recorded the time
that the Abraxis operators were on-site. Time was
removed for site-specific training activities and Visitor's
Day. Time estimates were rounded to the nearest hour.

During the demonstration, the skill level required for the
operators to complete analyses and report results was
evaluated. As stated in Section 7.4.1, based on the field
observations, a field technician with laboratory
experience with extraction and sample handling
protocols was considered to be qualified to use the
coplanar PCB ELISA kit. A high school graduate is
needed to perform sample extractions; however, a
technician with a college degree is preferred for
performing sample analysis. This information was
corroborated by Abraxis.

Education levels of the actual field operators included a
master's degree for the primary operator and a Ph.D.
degree for the secondary operator. For the economic
analysis, costs were estimated using both actual and
projected necessary skill levels for operators.

8.1.5  Investigation-Derived Waste Disposal Cost
During the demonstration, Abraxis was provided with
5-gallon containers for collecting wastes generated
during the demonstration. Sample by-products such as
used samples, aqueous and solvent-based effluents
generated from analytical processes, used glassware, and
personal protective equipment were disposed of in the
containers. The total cost to dispose of these wastes
generated during the demonstration is included in the
economic analysis. Items such as coffee cups, food
waste, and office waste were  disposed of in regular
public refuse containers and were not included as IDW
and are not discussed in this economic analysis.
8.1.6   Costs Not Included
Items whose costs were not included in the economic
analysis are identified below along with a rationale for
the exclusion of each.

Electricity.  During the demonstration, some of the
capital equipment was operated using AC power. The
costs associated with providing the power supply were
not included in the economic analysis as it is difficult to
estimate the electricity used solely by the Abraxis
technology. The total cost for electricity usage over the
10-day demonstration was $288. With seven mobile
labs/trailers and miscellaneous equipment being operated
continuously during the course of the demonstration, the
cost of Abraxis electricity usage would be no more than
$41. There was significantly more cost (approximately
$13,000) to install an electrical board and additional
power at the demonstration site, but this was a function
of the demonstration site and not the responsibility of the
individual developers, so this cost was not included in
the economic analysis.

Oversight of Demonstration Activities. Atypical user
of the coplanar PCB ELISA kit would not be required to
pay for customer oversight of sample analysis. The EPA,
the MDEQ, and Battelle representatives were present
during the field demonstration, but costs for oversight
were not included in the economic analysis because
these activities were project-specific. For these same
reasons, cost for auditing activities (i.e., technical
systems audits at the reference laboratory and during the
field demonstration) were also not included.

Travel and Per Diem for Operators. Operators may be
available locally. Because the availability of operators is
primarily a function of the location of the project site,
travel and per diem costs for operators were not included
in the economic analysis.

Sample Collection and Management. Costs for sample
collection and management activities, including sample
homogenization and labeling, were not included in the
economic analysis because these activities were project-
specific and were not dependent upon the selected
reference method or developer technology. Additionally,
sample shipping, COC activities, preservation of
samples, and distribution of samples were specific
requirements of this project that applied to all developer
                                                     52

-------
technologies and may vary from site to site. None of
these costs were included in the economic analysis.

Shipping. Costs for (1) shipping equipment and supplies
to the demonstration site and (2) sample coolers to the
reference laboratory were not included in the economic
analysis because such costs vary depending on the
shipping distance and the service used (for example, a
courier or overnight shipping versus economy shipping).

Items Costing Less Than $10. The cost of inexpensive
items was not included in the economic analysis when
the estimated cost was less than $10. Items where it is
estimated that the cost was less than $10 included:
- Distilled water
- Personal protective equipment
- Waste containers
- Lab stools

8.2     Coplanar PCB ELISA Kit Costs
This section presents information on the individual costs
of capital equipment, supplies, support equipment, labor,
and IDW disposal for the coplanar PCB ELISA kit as
well as a summary of these costs. Additionally,
Table 8-1 summarizes  the coplanar PCB ELISA kit
costs. As described in Section 4.6, Abraxis analyzed 116
samples during the  field demonstration and 93 samples
in its laboratory (total 209 demonstration samples). It is
important to note that costs estimated in this section are
based on actual costs to analyze the  116 samples during
the field demonstration.  Cost estimates for analyzing the
entire set of 209 demonstration samples were then
determined based on the  field demonstration costs.

8.2.1    Capital Equipment Cost
The capital equipment cost was the cost associated with
the purchase of the  technology in order to perform
sample preparation and analysis. The coplanar PCB
ELISA kit can be purchased from Abraxis for $ 1,000.
One kit contains enough  supplies for 100 samples to be
analyzed. Because the  kit is consumable, Abraxis does
not rent the coplanar PCB ELISA kit. During the field
demonstration, Abraxis utilized two coplanar PCB
ELISA kits  for approximately three days to analyze 116
samples.
8.2.2  Cost of Supplies
The supplies that Abraxis used during the demonstration
fall into two general categories: expendable or reusable.
Table 8-1 lists all the expendable and reusable supplies
that Abraxis used during the demonstration and their
corresponding costs. The cost of each item was rounded
to the nearest $ 1. Expendable supplies are ones that are
consumed during the preparation or analysis. Reusable
costs are items that must be used during the analysis but
ones that can be repeatedly reused. The estimated life of
reusable supplies could not be assessed during this
economic analysis.

The total cost of the supplies employed by Abraxis
during the field demonstration was $7,008, and the total
supplies cost for all 209 samples was $7,377. Supplies
have to be purchased from a retail vendor of laboratory
supplies. Reusable items listed in Table 8-1 can be
substituted for other models  that operate under the same
specifications, thereby modifying the cost of supplies to
the potential kit user.

8.2.3  Support Equipment Cost
Abraxis analyzed demonstration samples in a 24-foot
mobile lab equipped with a fume hood.  The rental cost
for the mobile lab for use during sample extraction and
sample analysis was $2,750. The minimum rental rate
for the mobile lab was  one month. Abraxis only used the
mobile laboratory for three days. Since weekly or daily
rental rates for the mobile lab were not an option, the
entire cost is reported.  As determined by the observers, a
construction trailer with a fume hood would have been
sufficient for operation of this technology in the field.
Use of a construction trailer with fume hood would have
been more cost efficient, lowering the support equipment
cost by at least $1,000.

A laptop computer is necessary for the efficient
operation of this technology. This is a one-time purchase
that is reusable.

8.2.4  Labor Cost
As described in Section 8.1.4, 50 labor-hours were spent
in the field to analyze 116 samples. An hourly rate of
$32.10 was used for a research scientist performing
sample extractions and sample analysis, and a
multiplication factor of 2.5 was applied to labor costs in
order to account for overhead costs.(9) Based on this
                                                     53

-------
hourly rate and multiplication factor, a labor rate of
$4,013 was determined for the analysis of the 116
samples during the field demonstration. It was estimated
that the labor cost for the total 209 samples was $7,223.

Based on observation, it is anticipated that lower-cost
field technicians, with proper training and skill levels,
could have analyzed the samples in a similar amount of
time. As such, the labor rate for the analysis of 116
samples during the field demonstration could have been
as low as $2,537 (hourly rate of $20.30 with 2.5
multiplication factor for 50 labor-hours), and $4,568 for
all 209 demonstration samples.

8.2.5  Investigation-Derived Waste Disposal Cost
As discussed in Chapter 7, Abraxis was provided with
5-gallon containers for collecting wastes generated
during the demonstration. Chapter 7 discusses the type
and amount of waste generated by the technology during
the field demonstration in more detail.

During the demonstration, Abraxis analyzed
116 samples. The total cost to dispose of the waste
generated for these samples was $399. The cost to
dispose of waste for  all 209 samples is estimated at
$718.

8.2.6  Summary of Coplanar PCB ELISA Kit
       Costs
The total cost for performing coplanar PCB analyses
using the coplanar PCB ELISA kit purchase option was
$22,668. The PCB analyses were performed for 58 soil
and sediment PE samples, 128 soil and sediment
environmental samples, and 23 extracts. When Abraxis
performed multiple dilutions or reanalyses for a sample,
these were not included in the number of samples
analyzed.

The total cost of $22,668 for analyzing the
demonstration samples under the coplanar PCB ELISA
kit included $3,600 for capital equipment; $7,377 for
supplies; $3,750 for  support equipment; $7,223 for
labor; and $718 for IDW disposal. Of these five costs,
the largest cost was for the supplies (33% of the total
cost).

8.3    Reference Method Costs
This section presents the costs associated with the
reference method used to analyze the 209 demonstration
samples for dioxin-like PCBs. Typical costs of these
analyses can range from $800 to $1,100 per sample,
depending on the method selected, the level of quality
assurance/quality control incorporated into the analyses,
and reporting requirements. The reference laboratory
utilized EPA Method 1668A for all soil and sediment
samples for comparison with the coplanar PCB ELISA
kit. The reference method costs  were calculated using
cost information from the reference laboratory invoices.

To allow an accurate comparison of the coplanar PCB
ELISA kit and reference method costs, the reference
method costs were estimated for the same number and
type of samples as was analyzed by Abraxis. For
example, although the reference laboratory analyzed soil
and sediment samples for dioxin/furans, the associated
sample analytical costs were not included in the
reference method costs because Abraxis did not analyze
samples for dioxin/furans during the demonstration.

Table 8-2 summarizes the projected and actual reference
method costs. At the start of the demonstration, the
reference laboratory's projected cost per sample was
$885 for PCB analysis. This cost covered the  preparation
and analysis of the demonstration samples, required
method QC samples, electronic  data deliverable, and the
data package for each. The actual cost for the  209
demonstration analyses ($ 184,449) was higher than the
projected ($158,465) due to reanalysis, re-extractions,
dilutions and additional cleanups that were above the
30% repeats allowable by the original quote. The
turnaround time by the reference laboratory for reporting
all 209 samples was approximately eight months (171
business days). The quoted turnaround time was three
months.
                                                     54

-------
Table 8-1. Cost Summary
Quantity Used
Item

Capital equipment
Purchase of Coplanar PCB ELISA kit
Supplies
Expendableb
Hexane (4-liter bottle)
Acetone (4-liter bottle)
Sulfuric acid (10-liter carboy)
Nitrogen cylinder
Cylinder regulator
Pasteur Pipets (package of 250; 2 mL
size)
Pipette tips (10, 200, and 1,000 |_iL)
Glass tubes (16x125 mm; case of
1,000)
Reusable
Microplate reader
Digital balance
Vortex mixer
Water bath
Support Equipment
Mobile laboratory
Laptop computer
Labor
Operator
IDW Disposal*
Total Cost
During Field
Demo
(116

2


1
1
1
3
2

1
1

1

1
1
1
1

1
1

samples)

kits


unit
unit
unit
unit
unit

unit
unit

unit

unit
unit
unit
unit

unit
unit

50 labor hours
1

unit

Unit Cost ($)


1,000


98
104
155
31
182

10
50

84

4,500
300
150
1,000

2,750
1,000

80C
399

Itemized
116
samples

2,000


98
104
155
93
364

10
150

84

4500
30,0
150
1,000

2,750
1,000

4,013
399
$17,170
Cost3 ($)
209
samples

3600


176
187
279
167
364

20
150

84

4,500
300
150
1,000

2,750
1,000

7,223
718
$22,668
          a Itemized costs were rounded to the nearest $1.
          b All reagents are at a minimum American Chemical Society grade.
          0 Labor rate for field technicians to operate technology rather than research scientists was $50.75 an hour,
           $2,537 for 116 samples and $4,568 for 209 samples.
          d Further discussion about waste generated during demonstration can be found in Chapter 7.
8.4    Comparison of Economic Analysis Results
The total costs for the coplanar PCB ELISA kit ($22,668)
and the reference method ($184,449) are listed in
Tables 8-1 and 8-2, respectively. The total cost for the
coplanar PCB ELISA kit was $ 161,781 less than that for
the reference method. It should be noted that Abraxis
analyzed 116 samples in three days on-site during the
demonstration and completed the remaining 116 samples
in its laboratory after the demonstration. Abraxis reported
that they completed the 116 analyses in its laboratory in
one week. For comparison, the reference laboratory took
eight months to report all 209 samples.
Use of the kit in the field will likely produce additional
cost savings because the results will be available within a
few hours of sample collection; therefore, critical
decisions regarding sampling and analysis can be made in
the field, resulting in a more complete data set. Additional
possible advantages to using field technologies include
reduction of multiple crew and equipment mobilization-
demobilization cycles to a single cycle, dramatically
increased spatial resolution mapping for higher statistical
confidence, leading to reduced insurance costs and
reduced disposal costs, and compression of total project
time to reduce administrative overhead. However, these
                                                     55

-------
Table 8-2. Reference Method Cost Summary
Analyses Performed
WHO PCBs EPA Method 1668A,
GC/HRMS
1668 Optional Carbon Column
DB1
Total Cost
Number of
Samples
Analyzed
23 extracts
186
soil/sediment
40
209 samples
Cost per sample
Quotation ($)
685
735
150

Itemized Cost ($)
Quotation3
15,755
136,710
6,000
$158,465
Actual
$184,449
  Price includes up to 30% of samples requiring additional work of some kind (dilutions or extra cleanup). Greater than
  that would require additional work with further charges associated to them ($150 to $180 per sample per procedure).
savings cannot be accurately estimated and thus were
not included in the economic analysis. Project-specific
costs associated with the use of the technology, such as
the cost for HRMS confirmation analyses and training
costs to be proficient in the use of the technology, were
also not accounted for in this analysis.
The Abraxis Coplanar PCB Kit is a screening method
that only reports TEQPCB, unlike the reference method
that reports concentrations for individual congeners.
Although the coplanar PCB kit analytical results did not
have the same level of detail as the reference method
analytical results (or comparable QA/QC data), the
coplanar PCB kit provided coplanar PCB analytical
results on-site at significant cost and time savings
compared to the reference laboratory.
                                                      56

-------
                                              Chapter 9
                                Technology Performance Summary
The purpose of this chapter is to provide a performance
summary of the Abraxis Coplanar PCB ELISA Kit by
summarizing the evaluation of the primary and secondary
objectives of this demonstration in Tables 9-1 and 9-2,
respectively. Detailed information about these
evaluations, including a complete evaluation of the
reference laboratory data, can be found in previous
sections of this report.

The data generated and evaluated during this
demonstration showed that the Abraxis kit in many cases
did not directly correlate with HRMS TEQPCB values, but
that the kits could be an effective tool as a screen for
sample concentrations above and below 50 pg/g TEQPCB,
particularly considering that both the cost ($22,668 vs.
$184,449) and the time (< two weeks vs. eight months) to
analyze the 209 demonstration samples were significantly
less than those of the reference laboratory. Because the
Abraxis kit is not expected to directly correlate to HRMS
TEQPCB in all cases, the technology should not be viewed
as producing an equivalent measurement value to HRMS
TEQ, but as a screening value to approximate HRMS
TEQPCB concentration.  It has been suggested that
correlation between the Abraxis TEQPCB results and
HRMS TEQPCB results could be improved by first
characterizing a site and calibrating the Abraxis results to
HRMS results. Subsequent analysis using the Abraxis kit
for samples obtained from this site may then show better
correlation with the HRMS TEQ PCB result. This approach
was not evaluated during this demonstration.
                                                    57

-------
Table 9-1. Abraxis Coplanar PCB ELISA Kit Performance Summary - Primary Objectives
Objective
PI: Accuracy
P2: Precision
P3: Comparability
P4: Estimated Method
Detection Limit
P5: False Positive/False
Negative Rate
P6: Matrix Effects
P7: Cost
Statistic
Number of data points
Median Recovery (%)
Mean Recovery (%)
Number of data points
Median RSD (%)
MeanRSD (%)
Number of data points
Median RPD (%)
Interval agreement (%)
Blank agreement (%)
EMDL(pg/gTEQPCB)
False positive rate around 6.25 pg/g TEQ (%)
False positive rate around 50 pg/g TEQ (%)
False negative rate around 6.25 pg/g TEQ (%)
False negative rate around 50 pg/g TEQ (%)
Performance
6
79
92
24
45
52
114
-129
83
75
6-31
35
8
7
3
• Measurement location: 1 sample set statistically different
• Matrix type: none
• Sample type: none
• PAH concentration: none
• Environmental site: none
• Known interferences: slight
116 samples during field demonstration: $17,170
If all 209 samples were analyzed during field demonstration: $22,668
                                               58

-------
Table 9-2.  Abraxis Coplanar PCB ELISA Kit Performance Summary - Secondary Objectives
           Objective
                                     Performance
 SI: Skill level of Operator
By observation of the kit in operation, it was determined that it would be beneficial for the user
of this test kit to have a background in wet chemistry and assay work. The developer prefers
that the user have a degree in chemistry or biology, but it was determined that technical ability
is more important than education level or background since good analytical and lab skills are
important when using this kit.
 S2: Health and Safety Aspects
The majority of the waste by this technology was generated during the sample extraction.
Solvents involved included hexane, acetone, and methanol. The bulk of the remaining waste
was concentrated sulfuric acid that was generated during the cleanup step. The volume of
sulfuric acid waste generated by this technology is dependent on the cleanliness of the samples
and can be variable. A fume hood is necessary for the operation of this technology.
 S3 Portability
This technology is readily deployable in a field or mobile environment. For the demonstration,
a mobile laboratory was used. The need for power, nitrogen tanks, fume hood, plate reader, and
a sink precluded this technology from being operated in the field with anything less than a
trailer, but a fully-equipped mobile laboratory was more infrastructure than was needed for this
technology.
 S4: Sample Throughput
During the field demonstration, 116 samples were processed by Abraxis, equating to a sample
throughput rate of 40 samples per day. This was accomplished in about three full working
days (50 labor-hours), with one person doing the majority of the extraction on the first two
days and a second person performing the data workup on the second and third days while the
sample processing was being completed by the first operator. Abraxis reported that the
remaining 93 sample analyses were completed in its laboratory in one week.	
                                                           59

-------
                                             Chapter 10
                                             References
1.  EPA. 2001. Database of Sources of Environmental
   Release of Dioxin-like Compounds in the United
   States, EPA/600/C-01/012, March.

2.  EPA. 2004 "Technologies for the Monitoring and
   Measurement of Dioxin and Dioxin-like Compounds
   in Soil and Sediment," Demonstration and Quality
   Assurance Project Plan, U.S. EPA/600/R-04/036,
   April.

3.  EPA Method 1613B. 1994. Dioxins, Tetra-thru Octa-
   (CDDs) and Furans (CDFs), EPA/821/B-94-005, 40
   Code of Federal Regulations Part 136, Appendix A,
   October.
   equivalency factors (TEFs) for PCBs, PCDDs,
   PCDFs for humans and wildlife. Environmental
   Health Perspectives 106: 775-792.

6.  De Rosa, Christopher T., et al. 1997. Dioxin and
   dioxin-like compounds in soil, Part 1: ATSDR
   Interim Policy Guideline. Toxicology and Industrial
   Health, Vol. 13, No. 6, pp. 759-768.

7.  NOAA.  1998.  Sampling and analytical methods of
   the national status and trends program mussel watch
   project:  1993-1996 update. NOAA Technical
   Memorandum NOS ORCA 130. Silver Spring,
   Maryland.
4.  EPA Method 1668A. 1999. Chlorinated biphenyl
   congeners byHRGC/HRMS, EPA/821/R-00-002,
   December.

5.  van den Berg, M., Birnbaum, L., Bosveld, A. T. C.,
   Brunstrom, B., Cook, P., Feeley, M., Giesy, J. P.,
   Hanberg, A., Hasagawa, R., Kennedy, S. W., Kubiak,
   T., Larsen, J. C., van Leeuwen, F. X. R., Liem, A. K.
   D., Nolt, C., Peterson, R. E., Poellinger, L., Safe, S.,
   Schrenk, D., Tillitt, D., Tysklind, M., Younes, M.,
   Waern, F., and Zacharewski, T. 1998. Toxic
9.
EPA SW-846 Method 8290. 1994. Polychlorinated
dibenzodioxins (PCDDs) and polychlorinated
dibenzofurans (PCDFs) by high-resolution gas
chromatography/high-resolution mass spectrometry
(HRGC/HRMS), September.

U.S. Bureau of Labor Statistics, National
Compensation Survey. Accessed on 7/26/04.
Available at:
http://data.bls.gov/labjava/outside jsp?survey=nc
                                                   60

-------
                   Appendix A
SITE Monitoring and Measurement Technology Program
               Verification Statement

-------

-------
     UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                             Office of Research and Development
                                   Washington, DC 20460
                                             EPA
        SITE Monitoring and Measurement Technology Program
                              Verification Statement
        TECHNOLOGY TYPE:  Enzyme-Linked Immunosorbent Assay
        APPLICATION:
MEASUREMENT OF DIOXIN AND DIOXIN-LIKE
COMPOUNDS
        TECHNOLOGY NAME: Coplanar PCB ELISA Kit
        COMPANY:
        ADDRESS:

        PHONE:

        WEB SITE:
        E-MAIL:
Abraxis LLC
54 Steamwhistle Drive
Warminster, Pennsylvania 18974
(215) 357-3911

www.abraxiskits.com
frubio(S>abraxiskits.com
VERIFICATION PROGRAM DESCRIPTION

The U.S. Environmental Protection Agency (EPA) created the Superfund Innovative Technology Evaluation (SITE)
Monitoring and Measurement Technology (MMT) Program to facilitate deployment of innovative technologies through
performance verification and information dissemination. The goal of this program is to further environmental protection
by substantially accelerating the acceptance and use of improved and cost-effective technologies. The program assists
and informs those involved in designing, distributing, permitting, and purchasing environmental technologies. This
document summarizes results of a demonstration of the Abraxis LLC Coplanar Polychlorinated Biphenyl (PCB)
Enzyme-Linked Immunosorbent Assay (ELISA) Kit.

PROGRAM OPERATION

Under the SITE MMT Program, with the full participation of the technology developers, the EPA evaluates and
documents the performance of innovative technologies by developing demonstration plans, conducting field tests,
collecting and analyzing demonstration data, and preparing reports. The technologies are evaluated under rigorous
quality assurance protocols to produce well-documented data of known quality. The EPA's National Exposure Research
Laboratory, which demonstrates field sampling, monitoring, and measurement technologies, selected Battelle as the
verification organization to assist in field testing technologies for measuring dioxin and dioxin-like compounds in soil
and sediment.
                                            A-l

-------
DEMONSTRATION DESCRIPTION

The demonstration of technologies for the measurement of dioxin and dioxin-like compounds was conducted at the
Green Point Environmental Learning Center in Saginaw, Michigan, from April 26 to May 5, 2004. The primary
objectives for the demonstration were as follows:
P1.  Determine the accuracy.
P2.  Determine the precision.
P3.  Determine the comparability of the technology to EPA standard methods.
P4.  Determine the estimated method detection limit (EMDL).
P5.  Determine the frequency of false positive and false negative results.
P6.  Evaluate the impact of matrix effects on technology performance.
P7.  Estimate costs associated with the  operation of the technology.

The secondary objectives for the demonstration were as follows:
S1.  Assess the skills and training required to properly operate the technology.
S2.  Document health  and safety aspects associated with the technology.
S3.  Evaluate the portability of the technology.
S4.  Determine the sample throughput.

A total of 209 samples was analyzed by each technology, including a mix of performance evaluation (PE) samples,
environmentally contaminated samples, and extracts. Abraxis analyzed 116 sample during the field demonstration and
93 samples in its laboratory. The PE samples were used primarily to determine the accuracy of the technology and
consisted of purchased reference materials with certified concentrations. The PE samples also were used to evaluate
precision, comparability, EMDL, false positive/negative results, and matrix effects. Dioxin-contaminated samples from
Warren County, North  Carolina; the Saginaw River, Michigan; Tittabawassee River, Michigan; Midland, Michigan;
Winona Post, Missouri; Nitro, West Virginia; Newark Bay, New Jersey; Raritan Bay, New Jersey; and Brunswick,
Georgia were used to evaluate precision, comparability, false positive/negative results, and matrix effects. Extracts
prepared in toluene were used to evaluate  precision, EMDL, and matrix effects. All samples were used to evaluate
qualitative performance objectives such as technology cost, the required skill level of the operator, health and safety
aspects, portability, and sample throughput.  AXYS Analytical Services (Sidney, British Columbia) was contracted to
perform the reference analyses by high-resolution mass spectrometry (HRMS)  (EPA Method 1668A). The purpose of
the verification statement is to provide a summary of the demonstration and its results; detailed information is available
in Technologies for Monitoring and Measurement of Dioxin and Dioxin-like Compounds in Soil and Sediment—Abraxis
LLC CoplanarPCB ELISA Kit (EPA/540/R-05/003).

TECHNOLOGY DESCRIPTION

The technology description and operating  procedure below are based  on information provided by Abraxis LLC.

The Abraxis Coplanar PCB ELISA Kit screens samples according to their PCB toxicity equivalent (TEQ) concentration.
The specificity of the test is predominantly for those congeners with high toxicity equivalency factor (TEF) values.
Samples extracted with organic solvents that are incompatible with ELISA can be evaporated and re-dissolved in
methanol. For a quick screen of soil and sediment samples, the samples can be extracted in 20% acetone in hexane,
evaporated, diluted 1:10 in the provided diluent, and run directly in the assay.  A solution containing a primary antibody
(rabbit) that reacts with coplanar PCBs is added to a microplate containing a secondary antibody that captures the
primary antibody. Calibrators and samples are added and allowed to incubate, followed by the addition of a coplanar
PCB-horseradish peroxidase (HRP) enzyme conjugate. Any coplanar PCBs that may be in the sample compete with the
coplanar PCB enzyme-labeled conjugate for a finite number of antibody binding sites. At the end of the incubation
period, the unbound conjugate is removed, and the plate is washed. A substrate/chromogen solution is then added and
enzymatically converted from a colorless to a blue solution by the captured coplanar PCB-HRP conjugate on the plate.
The reaction is then terminated by acidification. The coplanar PCB concentration is determined by measuring the
absorbance (at 450 nanomaters) of the sample solution using a microplate reader and comparing it to the absorbance of
the calibrators. The amount of color produced is inversely proportional to the amount of coplanar PCBs present in the
sample. Results are reported as picogram/gram (pg/g) total PCB TEQ (TEQPCB).  The final value measured by ELISA is


                                                    A-2

-------
the sum of the various congener responses. This value approximates toxicity equivalent (TEQPCB) because the
immunoassay kit cross-reaction profile for coplanar PCBs approximates TEF values. The cross-reactivity of the Abraxis
coplanar PCB assay for various congeners and Aroclors can be expressed as the least detectable dose, which is
estimated at 90% B (mean absorbance obtained with the standard)/Bo (mean absorbance value for the zero standard), or
as the dose required for the 50% absorbance inhibition (50% B/Bo). The primary use of the Abraxis Coplanar PCB
ELISA Kit is to screen samples that have low coplanar PCB concentrations. The sensitivity of the test is claimed by
Abraxis to be 4 parts per trillion in water samples and 6.25 pg/g TEQPCB in soil or sediment samples. These values are
related to the original sample concentration by using the appropriate dilution and volume factors. Detection levels
depend on how much sample is evaporated and the volume of solvent used to resuspend the sample. Matrix detection
limits will vary according to the matrix being analyzed, sample size, and dilution factor. Up to 100 samples per day can
be analyzed using the procedure described.

VERIFICATION OF PERFORMANCE

The Abraxis kit is an immunoassay technology that reports total coplanar PCBs in a sample. It should be noted that this
technology may not directly correlate to HRMS TEQPCB in all cases because it is known that the congener responses and
cross-reactivities to the kit are not identical to the World Heath Organization TEFs that are used to convert congener
FIRMS concentration values to TEQPCB. Therefore, the Abraxis kit should not be viewed as producing an equivalent
measurement value to HRMS TEQPCB, but as a screening value to approximate HRMS TEQPCB concentration. It has
been suggested that correlation between the Abraxis TEQPCB results and HRMS TEQPCB results could be improved by
first characterizing a site and calibrating the Abraxis results to HRMS results. Subsequent analysis using the Abraxis kit
for samples obtained from this  site may then show better correlation with the HRMS TEQPCB result.  This approach was
not evaluated during this demonstration.

Accuracy: The determination of accuracy was based on the agreement of the Abraxis results with the certified or spiked
levels of the PE samples that were obtained from commercial sources. Accuracy was assessed by percent recovery (R),
which is the average of the replicate  results from the coplanar PCB ELISA kit divided by the certified or spiked value of
the PE sample, multiplied by 100%.  Ideal R values are near 100%. The overall R values were 92% (mean), 79%
(median),  16% (minimum), and 204% (maximum).

Precision: Replicates were incorporated for all samples (PE, environmental, and extracts) included in the 209 samples
analyzed in the demonstration.  Three samples had seven replicates in the experimental design, one sample had eight
replicates, and all other samples had four replicates. Precision was determined by calculating the standard deviation of
the replicates, dividing by the average concentration of the replicates, and multiplying by 100%.  Ideal relative standard
deviation (RSD) values are less than 20%. The overall RSD values were 52% (mean), 45% (median), 4% (minimum),
and 172% (maximum).

Comparability: The Abraxis results were compared to EPA Method 1668A results for TEQPCB. The results were
compared by determining the relative percent difference (RPD) by dividing the difference of the  two numbers by the
mean of the two numbers and multiplying by 100%.  Ideal RPD values are between positive and negative 25%.  The
overall RPD values were -129% (median), -200% (minimum), and 200% (maximum).  The Abraxis results were also
compared to the reference laboratory results using an interval approach to determine if the Abraxis results and the
reference laboratory results would place the samples in the same action-level interval, thereby resulting in the same
action-oriented decision. The developer and reference data were grouped into four TEQ concentration ranges. The
ranges were < 50 pg/g, 50 to 500 pg/g, 500 to 5,000 pg/g, and > 5,000 pg/g. The intervals were determined based on
current guidance for cleanup levels.  The percentage of developer results that agreed with reference laboratory results
forTEQPCBwas83%.

Estimated method detection limit:  EMDL was calculated generally according to the procedure  described in 40 CFR
Part 136, Appendix B, Revision 1.11. Lower EMDL values indicate better sensitivity. The calculated EMDLs ranged
from 6 to 31 pg/g TEQPCB, depending on whether nondetect values were assigned values of zero, one-half the reporting
limit value, or the reporting limit value itself. The detection limit reported by Abraxis in the demonstration plan was
6.25 pg/g TEQPCB.
                                                    A-3

-------
False positive/negative results: The tendencies of the Abraxis ELISA kit to return results that were above a specified
level when the reference laboratory result was below that level (i.e., false positive) and to report values that were below
the specified level when the reference laboratory reported a result that was greater than the specified level (i.e., false
negative) were determined.  Ideal false positive and false negative rates would be zero. The kit had a false positive rate
of 35% and a false negative rate of 7% around 6.25 pg/g TEQPCB (the reporting limits of the technology). Abraxis
reported significantly fewer false positives (8%) and false negatives (3%) around 50 pg/g TEQPCB. This evaluation
indicates that the Abraxis kit could be an effective tool for screening sample concentrations above and below 50 pg/g
Matrix effects: The likelihood of matrix-dependent effects on performance was investigated by evaluating results in a
variety of ways. The Abraxis TEQPCB results that were generated in the laboratory and in the field for replicate samples
were statistically different for only one sample (overall 5% of the total number of samples), which was the PE sample
that was spiked for only polynuclear aromatic hydrocarbons  (PAHs). No significant effect was observed for the
reproducibility of Abraxis results by matrix type (soil, sediment, and extract), sample type (PE vs. environmental vs.
extract), or by PAH concentration. PE samples spiked with dioxin/furans or PAHs were sometimes reported as
detections for PCBs that were not spiked in the sample. The Abraxis results were not more or less comparable to the
reference laboratory results based on environmental site.

Cost: The full cost of the technology was documented and compared to the cost of the reference  analyses. The total cost
for the coplanar PCB ELISA kit to analyze all 209 demonstration samples was $22,668. The total cost for the reference
laboratory to analyze all 209 demonstration samples by EPA Method 1668A was $184,449. The total cost for the
coplanar PCB ELISA kit was $161,781 less than for the reference method.

Skills and training required: By observation of the kit in operation, it was determined that it would be beneficial for
the user of this test kit to have a background in wet chemistry and assay work. The developer prefers that the user have a
degree in chemistry or biology, but it was determined that technical ability is more important than education level or
background since good analytical and lab skills are important when using this kit.

Health and safety aspects: The majority of the waste by this technology was generated during the  sample extraction.
Solvents involved included hexane, acetone, and methanol. The bulk of the remaining waste was concentrated sulfuric
acid which was generated during the cleanup step. The volume of sulfuric acid waste generated by this technology is
dependent on the cleanliness of the samples and can be variable. A fume hood is necessary for the operation of this
technology.

Portability: This technology is readily deployable in  a field or mobile environment. For the demonstration, a mobile
laboratory was used. The need for power, nitrogen tanks, fume hood, plate reader, and a sink precluded this technology
from being operated in the field with anything less than a trailer, but a fully equipped mobile laboratory was more
infrastructure than was needed for this technology.

Sample throughput: During the field demonstration, 116 samples were processed by Abraxis, equating to a sample
throughput rate of 40 samples per day. This was accomplished in about  three full working days (50 labor-hours), with
one person doing the majority of the extraction on the first two days and  a second person performing the data workup
on the second and third days when the sample processing  was being completed by the first operator. Abraxis completed
the remaining 93 samples in their laboratory and reported that it took them one week to complete those analyses. For
comparison, the reference laboratory took eight months to report all 209  samples.
       NOTICE: Verifications are based on an evaluation of technology performance under specific, predetermined
       criteria and the appropriate quality assurance procedures. EPA makes no expressed or implied warranties as to
       the performance of the technology and does not certify that a technology will always operate as verified. The
       end user is solely responsible for complying with any and all applicable federal, state, and local requirements.
                                                     A-4

-------
                                         Appendix B
                   Supplemental Information Supplied by the Developer


 The purpose of this section is for the developer to provide additional information about the technology.  This can
include updates/changes/modifications planned for the technology or which have occurred since the technology was
       tested. The developers can also use this section to comment and expand on the findings of the report.

-------

-------
                                         Abraxis Comments

General Comments

Overall we are pleased with the results obtained on this SITE demonstration. The results obtained indicate
that this technology can be used as an effective tool for the rapid and cost-effective screening of samples at
PCB concentrations at or below a 50 pg/g TEQ action level, as well as for action level intervals of 50-500,
500-5000, and >5000 pg/g.

The Abraxis Coplanar PCB Kit utilizes immunoassay (ELISA) technology. This technology utilizes
antibodies which exhibit different sensitivities and reactivities against the various coplanar PCB congeners.
Because the concentrations and type of specific coplanar PCB congeners can vary greatly between samples, it
is virtually impossible to directly correlate to HRMS TEQPCB Therefore, the Abraxis kit should not be viewed
as producing an equivalent measurement value to HMRS TEQPCB but only as a screening method to provide a
value which is approximate to HMRS TEQPCB concentrations, as it was originally intended.

Overall Comments on Demonstration Objective Conclusions

Sample Throughput: Approximately 1 week needed to analyze 209 samples using the Coplanar PCB kit
compared to approximately 8 months for the reference lab analysis.

Cost: Approximately $100 per sample compared to about $900 for lab analysis

False Positive/Negative:  8% false positive, 3% false negative at 50 pg/g TEQPCB

Precision: Relative standard deviation values were higher than what we would of liked to see. These results
might have been due to the very complex samples used during this demonstration and the quick extraction
procedure and sample clean-up we chose to use during the study. This same comment applies for accuracy
results.

We have requested the complete data analysis containing PAH, PCP, etc. concentrations obtained by the
reference lab. This data will help Abraxis understand the exact nature of the samples and to address/develop
improvements to the extraction/sample clean up method.
            Information was provided by the developer and does not necessarily reflect the opinion of the EPA.
                                                 B-l

-------

-------
                         Appendix C
Reference Laboratory Method Blank and Duplicate Results Summary

-------

-------
Table C-1. Summary of Method Blank Performance
Sample Batch
Number
D/F
WG12107
D/F
WG12148
D/F
WG12264
D/F
WG12534
D/F
WG12641
D/F
WG12737
D/F
WG12804
D/F
WG13547
Criteria
Met
Y
N
N
N
N
N
N
N
Method
Blank TEQa
(Pg/g)
0.000812
0.133
0.0437
0.610
0.0475
0.348
0.0153
0.0553
Sample TEQ Range3 (pg/g)
26.1-74.1 (Newark Bay)
9.93-13.3 (RaritanBay)
13. 5-50.4 (Newark Bay)
49.5-15,200 (Brunswick)
1.0-94.1 (Titta. River sediment)
0.237-6,900 (PE)
25.3-7, 100 (PE)
3 1-269 (Midland)
72.8 (Brunswick)
123 (Titta. River sediment)
0. 159-7,690 (PE)
25.7-192 (Midland)
35.2- 1,300 (Titta. River soil)
3.89-188 (PE)
57.5-3,000 (Nitro)
37.9 (North Carolina)
122 (Saginaw River)
26.4-222 (Midland)
Comments

Many samples had concentrations >20x
blank. Few that didn't were not
significantly affected on a total TEQ basis.
Most samples had concentrations >20x
blank. Low level Tittabawassee River
sediment samples L6749-2 (Ref 48b), -9
(Ref 130), -10 (Ref 183), and -12 (Ref 207)
were evaluated based on their replication
within the demonstration analyses and
comparison to characterization results and
considered unaffected by method blank
exceedances. Low level PE samples
L6760-1 (Ref 25), -3 (Ref 28), and -4
(Ref 29) were D/F blanks with resulting
TEQs sufficiently low enough to still be
distinguished as blank samples.
Sample concentrations > 20x blank.
All but PE sample Ref 177 (0. 159 TEQ)
had significantly higher total TEQ than
blank. Ref 177 was confirmed by running
in another batch and results, which agreed
within 18%. Additionally, Ref 177 was
compared to its replicates within the
program and considered acceptable.
Sample concentrations >20x blank.
A few analytes higher than criteria but no
significant contribution to total TEQ.
Several analytes exceeded criteria, but
blank total TEQ contribution to sample is
relatively small.
                                                 C-1

-------

Sample Batch
Number
D/F
WG13548





















D/F
WG13549

D/F
WG13551

D/F

WG13552
D/F
WG13984








D/F
WG14274









Criteria
Met











N












N


N


Y





N









N





Method
Blank TEQa
(Pg/g)











0.0114












0.0925


2.40


0 000969





0.0154









0.0434







Sample TEQ Range3 (pg/g)
99.6-99.7 (Saginaw River)
32.9-36.4 (North Carolina)
0.268-100 (Extracts)




















2, 160-3,080 (Nitro)
146-1,320 (Saginaw River)
788-8,410 (North Carolina)
1,100-10,800 (North Carolina)
7,160-11,300 (WinonaPost)

0.0386-9.28 (PE)

25.8 (Midland)
0.524-24.8 (PE)
10.4 (Raritan Bay)
53. 1-444 (Extracts)







2800 (Nitro)
35.5-8,320 (North Carolina)
0.0530-5.93 (PE)









Comments
Several analytes exceeded criteria. In
general, the blank contribution to total TEQ
was negligible and in those cases results
were accepted. Several low-level extract
samples were evaluated as follows: Extract
Spike #1 samples L6754-4 (Ref 4), -8
(Ref 8), -10 (Ref 10), -14 (Ref 14), -19
(Ref 19), -22 (Ref 22), and -23 (Ref 23)
were known TCDD spikes at 0.5 pg/mL.
Results were compared to the known spiked
TEQ and considered unaffected by blank
contribution to TEQ. Extract Spike #3
samples L6754-1 (Ref 1), -7 (Ref 7), -12
(Ref 12), and -15 (Ref 15) were PCB spikes
and not expected to contain D/F. These
spikes consistently contained a D/F TEQ of
~0.3. However, this came from a consistent
-0.3 pg/mL of TCDD detected in these
extracts that was confirmed as a low-level
TCDD contamination by AXYS. Since
TCDD was not present in the lab blank,
these results were accepted as unaffected by
any blank contribution to TEQ.
Many analytes exceeded limits, but the
blank contribution to total TEQ is small
relative to sample TEQs.
Many analytes exceeded limits, but the
blank contribution to total TEQ is small
relative to sample TEQs.



Blank contribution to total TEQ was
negligible except for PE samples L7 179-7
(Ref 94), -8 (Ref 96), -11 (Ref 108), -12
(Ref 109), -17 (Ref 132), and L7182-6
(Ref 150). All but L7179-8 were certified
blanks. L7 179-8 was a PAH spike with no
D/F TEQ expected. The TEQs of these
samples were considered sufficiently low
enough to still be distinguished as blank
samples and were accepted.
Sample TEQs were large enough to be
unaffected by the blank TEQ except for
four PE samples L7179-4 (Ref 85), -16
(Ref 124) and L7182-12 (Ref 169) and -14
(Ref 184). These PE samples were either
certified blanks or PCB spikes with no
expected D/F TEQ. Resulting TEQs for
these samples were considered low enough
to be distinguished as blank samples and
were accepted.
C-2

-------
Sample Batch
Number
PCB
WG12108
PCB
WG12147
PCB
WG12265
PCB
WG12457
PCB
WG12687
PCB
WG12834
PCB
WG12835
PCB
WG12836
PCB
WG13008
PCB
WG13256
PCB
WG13257
PCB
WG13258
PCB
WG13554
PCB
WG14109
Criteria
Met
N
Y
Y
N
N
N
N
N
N
Y
Y
Y
N
N
Method
Blank TEQa
(Pg/g)
0.000137
0.000
0.0000584
0.000208
0.0183
0.000405
0.000125
0.0499
0.0221
0.000102
0.000251
0.000301
0.0000900
0.000288
Sample TEQ Range3 (pg/g)
2.63-5. 19 (Newark Bay)
2.04-2.82 (Raritan Bay)
1.21-5.06 (Newark Bay)
0.104-0.330 (Brunswick)
0.132-0.369 (Brunswick)
0.034-0.649 (Titta. River
sediment)
0.00277-1,030 (PE)
4.20-1,020 (PE)
0.974-2.73 (Midland)
10.3-1, 180 (PE)
0.0157-62.4 (Saginaw River)
0.181-0.203 (Brunswick)
0.986-7.57 (Titta. River Soil)
0.822-2.06 (WinonaPost)
1,060-904,000 (North Carolina)
2.38-3. 15 (Midland)
1.03-8.37 (Titta. River soil)
4 1.0-1 140 (PE)
0.00385-0.051 (PE)
0.253-0.3 18 (Midland)
0.135-2.08 (Extracts)
3.53-9.62 (PE)
1.14-1.33 (Titta. River Soil)
0. 163-37.0 (Nitro)
29.8-73.6 (Saginaw River)
40.1-42.1(PE)
0.000103-1,080 (Extracts)
435-1,160 (PE)
0.388-0.452 (Nitro)
0.0467 (Saginaw River)
0.654-1.87 (WinonaPost)
0.00300-0.0420 (PE)
Comments
PCB 77 slightly high, but all samples >20x
blank levels.


PCB 77 slightly high. Did not report any
samples where PCB 77 was <10x blank.
No significant effect on total TEQ.
PCB 77 and 156 high, but all samples >
20x blank levels.
PCB 77 slightly high. Does not affect total
TEQ.
PCB 77 slightly high. Sample TEQs much
greater than blank TEQ.
PCBs 77, 123, 126, 156, 167, and 118 high,
but most samples significantly > 20x blank
levels.
PCBs 77 and 1 18 high, but all samples
>20x blank levels.



PCB 77 slightly high. Does not affect total
TEQ.
PCB 77 high. PE certified blanks Ref 85,
Ref 85 duplicate, and Ref 108 were the only
samples where PCB 77 was not >20x blank.
TEQs for these certified blanks were
considered low enough to be distinguished
as blank samples and were accepted.
1 All nondetect and EMPC values were assigned a zero concentration for the TEQ calculation.
b "Ref XX" is a reference laboratory sample ID number.
                                                               C-3

-------
Table C-2. Sample Batch Duplicate Summary
Sample Batch
Number
D/FWG12107
D/FWG12148
D/F WG12264
D/F WG12534
D/FWG12641
D/F WG12737
D/F WG12804
D/F WG13547
D/FWG13548
D/FWG13549
D/FWG13551
D/FWG13552
D/F WG13984
D/F WG14274
PCB WG12108
Criteria Met
N
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
N
N
Duplicate RPDa (%)
23
2.1
1.2
5.7
4.6
14
none
16
5.9
3.6
0.0
20
(onU=l/2DLbasisb)
3.4
54
22
Comments
L6744-5, Ref 100 Newark Bay
Because this was above the 20% criteria, an additional
aliquot of this sample was prepared. Results for the
additional aliquot were within 1 1% RPD from the original
results; therefore, this duplicate result was accepted.
L6744-9, Ref 122 Newark Bay
L6760-2,Ref27PE
L6760-14, Ref 55 PE
L6747-1, Ref 32 Midland
L6750-3, Ref 78 Tittabawassee River Soil
The duplicate processed with this batch was to be repeated
due to some analytes being <20x blank level. However, it
was reprocessed as a single sample and not a duplicate.
Samples in this set were accepted based on their agreement
with other replicates within the demonstration program.
L7163-l,Ref26Nitro
L6751-14, Ref 83 North Carolina
L6751-7, Ref 135 North Carolina
L6751-1, Ref 42 North Carolina
L7179-3, Ref 74 PE. Fails on a U=0 DL basis due to
presence of "K" flagged analytes in one replicate. When
compared on U-l/2 DL basis where "K" concentrations are
included in the TEQ calculation, the duplicate passed.
L7179-14,Ref 113PE
L7 179-16, Ref 124 PE
This was a PCB PE sample and contained only trace levels
of D/F. Replicate precision is affected because D/F content
is so low. This is not expected to indicate any problems with
precision within this sample set. Samples in this set were
accepted based on their agreement with other replicates
within the demonstration program.
L6744-2, Ref 49 Newark Bay
This result is only slightly above the acceptance criteria of
20%. The variability was influenced by 25% RPD for
PCB 126 (which has the highest TEF of the PCBs and,
therefore, a larger influence on total TEQ). The slight
exceedance in duplicate criteria was not considered to have
any significant impact on the data reported in this sample
batch. All samples in this set were also evaluated based on
their agreement with other replicates within the
demonstration program and deemed to be acceptable.
                                              C-4

-------
Sample Batch
Number
PCB WG12147
PCB WG12265
PCB WG12457
PCB WG12687
PCB WG12834
PCB WG12835
PCB WG12836
PCB WG13008
PCB WG13256
PCB WG13257
PCB WG13258
PCB WG13554
PCB WG14109
Criteria Met
N
Y
N
Y
Y
N
Y
Y
Y
Y
Y
Y
N
Duplicate RPDa (%)
none
2.5
none
43
4.2
none
2.6
5.1
1.7
(on U= 1/2 DL basis)
15
19
12
85
(on U= 1/2 DL basis)
Comments
L6748-9, Ref 129 Brunswick
The duplicate sample for this batch required reprocessing.
When reprocessed, it was not prepared in duplicate.
Samples in this set were accepted based on the RPD of site
replicates that were processed within the batch (RPDs
<10%).
L6760-5, Ref35PE
L6760-16,Ref62PE
This duplicate set was to be repeated due to low internal
standard recovery. When repeated, it was not prepared in
duplicate. Data for this set was accepted because all
samples in the set were PE samples. These PE samples met
accuracy criteria and reproducibility criteria to other
replicates of the same PE material processed within the
demonstration.
L6762-12, Ref 169 PE
L6750-8, Ref 164 Tittabawassee River Soil
Duplicate sample repeated in WG13258. Results reported
with that sample set. Three sets of sample replicates within
this batch were also compared and found to have <13.5%
RPD showing acceptable precision with this sample set.
L6751-6, Ref 126 North Carolina
L6750-6, Ref 121 Tittabawassee River Soil
L6761-3, Ref 74 PE. Fails on a U=0 DL basis due to
presence of "K" flagged analytes in one replicate. When
compared on U=l/2 DL basis where "K" concentrations are
included in the TEQ calculation, the duplicate passed.
L7187-5, Ref 92 Tittabawassee River Soil
L6743-2, Ref36Nitro
L6762-l,Ref202PE
L7 '179-4, PE. Fails based on both U=0 and U= 1/2 DL.
This was a blank PE sample and contained only trace levels
of PCBs. Replicate precision is affected because the PCB
content is so low. This is not expected to indicate any
problems with precision within this sample set. Samples in
this set were accepted based on their agreement with other
replicates within the demonstration program.
1 Nondetects were assigned a concentration of zero unless otherwise noted and are referred to as U=0 DL values.
bU=l/2 DL indicates that non-detects were assigned a concentration equal to one-half the SDL and EMPC concentrations were assigned a value
equal to the EMPC.
                                                               C-5

-------

-------
                  Appendix D
Summary of Developer and Reference Laboratory Data

-------

-------
Appendix D. Abraxis and Reference Laboratory One-to-One Matching
Sample Type
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Sample
Number
ABRAXIS 178
ABRAXIS 42
ABRAXIS 97
ABRAXIS 206
ABRAXIS 140
ABRAXIS 30
ABRAXIS 85
ABRAXIS 177
ABRAXIS 187
ABRAXIS 116
ABRAXIS 159
ABRAXIS 38
ABRAXIS 83
ABRAXIS 156
ABRAXIS 131
ABRAXIS 29
ABRAXIS 34
ABRAXIS 174
ABRAXIS 62
ABRAXIS 199
ABRAXIS 167
ABRAXIS 78
ABRAXIS 154
ABRAXIS 88
ABRAXIS 50
ABRAXIS 184
ABRAXIS 129
ABRAXIS 92
ABRAXIS 112
ABRAXIS 89
ABRAXIS 149
ABRAXIS 182
ABRAXIS 79
ABRAXIS 160
ABRAXIS 186
ABRAXIS 111
ABRAXIS 98
ABRAXIS 148
ABRAXIS 76
ABRAXIS 191
ABRAXIS 136
ABRAXIS 125
ABRAXIS 82
ABRAXIS 113
ABRAXIS 208
ABRAXIS 114
ABRAXIS 123
ABRAXIS 107
ABRAXIS 86
ABRAXIS 72
ABRAXIS 169
ABRAXIS 183
ABRAXIS 181
Measurement
Location
Laboratory
Field
Field
Laboratory
Laboratory
Field
Field
Laboratory
Laboratory
Field
Laboratory
Field
Field
Laboratory
Laboratory
Field
Field
Laboratory
Field
Laboratory
Laboratory
Field
Laboratory
Field
Field
Laboratory
Laboratory
Field
Laboratory
Laboratory
Laboratory
Laboratory
Field
Laboratory
Laboratory
Field
Field
Laboratory
Field
Laboratory
Laboratory
Laboratory
Field
Field
Laboratory
Field
Laboratory
Field
Field
Field
Laboratory
Laboratory
Laboratory
Sample Description
Brunswick #1
Brunswick #1
Brunswick #1
Brunswick #1
Brunswick #2
Brunswick #2
Brunswick #2
Brunswick #2
Brunswick #3
Brunswick #3
Brunswick #3
Brunswick #3
Midland #1
Midland #1
Midland #1
Midland #1
Midland #2
Midland #2
Midland #2
Midland #2
Midland #3
Midland #3
Midland #3
Midland #3
Midland #4
Midland #4
Midland #4
Midland #4
NC PCB Site #1
NC PCB Site #1
NC PCB Site #1
NC PCB Site #1
NC PCB Site #2
NC PCB Site #2
NC PCB Site #2
NC PCB Site #2
NC PCB Site #3
NC PCB Site #3
NC PCB Site #3
NC PCB Site #3
Newark Bay #1
Newark Bay #1
Newark Bay #1
Newark Bay #1
Newark Bay #2
Newark Bay #2
Newark Bay #2
Newark Bay #2
Newark Bay #3
Newark Bay #3
Newark Bay #3
Newark Bay #3
Newark Bay #4
Replicate
1
2
3
4
1
2
3
4
1
2
o
J
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
TEQPrR(pg/g)
Developer3
16.1
14.5
23
21.8
13.9
<6.3
<6.3
<6.25
206.3
425
159.3
312.5
<6.3
<6.25
<6.25
<6.3
11.3
<6.25
<6.3
14.8
13.3
<6.3
<6.25
<6.3
<6.3
<6.25
<6.25
<6.3
>2500
>2500
5050
4600
>2500
241.8
23775
>2500
>2500
>25000
>2500
>25000
16.6
15.6
65
10.5
<6.25
15.5
16.2
12
65
38.8
32.3
25.5
<6.25
Reference Laboratory1"
0.314
0.342
0.369
0.313
0.127
0.128
0.132
0.123
0.19
0.181
0.203
0.182
2.59
2.73
2.5
2.53
2.7
2.81
2.48
3.15
2.28
2.17
2.23
2.38
0.253
0.318
0.974
0.263
53000
65300
80500
85100
311000
305000
210000
361000
848000
618000
533000
904000
1.22
1.44
1.39
1.34
5.01
5.19
5.14
5.09
4.61
5.04
4.5
5.03
2.73
                                                D-l

-------
Sample Type
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Sample
Number
ABRAXIS 170
ABRAXIS31
ABRAXIS 91
ABRAXIS 57
ABRAXIS 51
ABRAXIS 145
ABRAXIS 119
ABRAXIS 39
ABRAXIS 153
ABRAXIS 87
ABRAXIS 144
ABRAXIS 27
ABRAXIS 205
ABRAXIS 75
ABRAXIS 128
ABRAXIS 162
ABRAXIS 188
ABRAXIS 55
ABRAXIS 24
ABRAXIS 68
ABRAXIS 193
ABRAXIS 138
ABRAXIS 73
ABRAXIS 59
ABRAXIS 185
ABRAXIS 64
ABRAXIS 146
ABRAXIS 81
ABRAXIS 155
ABRAXIS 122
ABRAXIS 71
ABRAXIS 32
ABRAXIS 152
ABRAXIS 157
ABRAXIS 58
ABRAXIS 74
ABRAXIS 164
ABRAXIS 130
ABRAXIS 60
ABRAXIS 198
ABRAXIS 43
ABRAXIS 104
ABRAXIS 141
ABRAXIS 28
ABRAXIS 166
ABRAXIS 46
ABRAXIS 176
ABRAXIS 41
ABRAXIS 209
ABRAXIS 172
ABRAXIS 44
ABRAXIS 158
ABRAXIS 127
ABRAXIS 54
ABRAXIS 37
Measurement
Location
Laboratory
Field
Field
Field
Field
Laboratory
Laboratory
Field
Laboratory
Field
Laboratory
Field
Laboratory
Field
Laboratory
Laboratory
Laboratory
Field
Field
Field
Laboratory
Laboratory
Field
Field
Laboratory
Field
Laboratory
Field
Laboratory
Laboratory
Field
Field
Laboratory
Laboratory
Field
Field
Laboratory
Laboratory
Field
Laboratory
Field
Field
Laboratory
Field
Laboratory
Field
Laboratory
Field
Laboratory
Laboratory
Field
Laboratory
Laboratory
Field
Field
Sample Description
Newark Bay #4
Newark Bay #4
Newark Bay #4
RaritanBay#l
RaritanBay#l
RaritanBay#l
RaritanBay#l
Raritan Bay #2
Raritan Bay #2
Raritan Bay #2
Raritan Bay #2
Raritan Bay #3
Raritan Bay #3
Raritan Bay #3
Raritan Bay #3
Saginaw River #1
Saginaw River #1
Saginaw River #1
Saginaw River #1
Saginaw River #2
Saginaw River #2
Saginaw River #2
Saginaw River #2
Saginaw River #3
Saginaw River #3
Saginaw River #3
Saginaw River #3
Solutia#l
Solutia#l
Solutia#l
Solutia#l
Solutia #2
Solutia #2
Solutia #2
Solutia #2
Solutia #3
Solutia #3
Solutia #3
Solutia #3
Titta. River Soil #1
Titta. River Soil #1
Titta. River Soil #1
Titta. River Soil #1
Titta. River Soil #2
Titta. River Soil #2
Titta. River Soil #2
Titta. River Soil #2
Titta. River Soil #3
Titta. River Soil #3
Titta. River Soil #3
Titta. River Soil #3
Titta. River Sed #1
Titta. River Sed #1
Titta. River Sed #1
Titta. River Sed #1
Replicate
2
3
4
1
2
3
4
1
2
o
J
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
TEQpr«(P2/2)
Developer3
<6.25
14.5
21.3
11.8
13.8
<6.25
<6.25
12.5
<6.25
21
<6.25
15
9
23
<6.25
57.8
10.8
42.5
92.5
25
31
22.2
46.3
16
<6.25
28.8
<6.25
26.3
<6.25
<6.25
<6.3
<6.3
18.4
<6.25
19
16
17.8
<6.25
<6.3
14.8
9
<6.3
<6.25
6.3
14.7
14.5
23.6
<6.3
15.7
6.5
10.5
<6.25
<6.25
10
9.5
Reference Laboratory1"
2.65
2.72
2.7
2.33
2.06
2.35
2.25
2.7
2.67
2.68
2.85
2.43
2.43
2.3
2.33
62.4
73.6
69.9
63.7
30.6
31
26.7
29.8
0.0202
0.0164
0.0467
0.0157
0.452
0.163
0.388
0.391
17.6
18.8
19.2
18.5
29.7
36.9
37
31.5
7.32
8.26
7.57
8.37
0.986
1.2
1.03
1.06
1.26
1.16
1.54
1.33
0.0527
0.034
0.0407
0.0403
D-2

-------
Sample Type
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Environmental
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Extract
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Sample
Number
ABRAXIS 69
ABRAXIS 53
ABRAXIS 120
ABRAXIS 143
ABRAXIS 49
ABRAXIS 163
ABRAXIS 202
ABRAXIS 80
ABRAXIS 103
ABRAXIS 150
ABRAXIS 93
ABRAXIS 126
ABRAXIS 142
ABRAXIS 61
ABRAXIS 48
ABRAXIS 137
ABRAXIS 207
ABRAXIS 63
ABRAXIS 105
ABRAXIS 147
ABRAXIS 3
ABRAXIS 17
ABRAXIS 10
ABRAXIS 19
ABRAXIS 15
ABRAXIS 5
ABRAXIS 22
ABRAXIS 13
ABRAXIS 9
ABRAXIS 6
ABRAXIS 11
ABRAXIS 4
ABRAXIS 21
ABRAXIS 18
ABRAXIS 12
ABRAXIS 7
ABRAXIS 23
ABRAXIS 1
ABRAXIS 20
ABRAXIS 14
ABRAXIS 8
ABRAXIS 16
ABRAXIS 2
ABRAXIS 196
ABRAXIS 95
ABRAXIS 70
ABRAXIS 26
ABRAXIS 133
ABRAXIS 173
ABRAXIS 118
ABRAXIS 151
ABRAXIS 192
ABRAXIS 102
ABRAXIS 100
ABRAXIS 200
Measurement
Location
Field
Field
Laboratory
Laboratory
Field
Laboratory
Laboratory
Field
Field
Laboratory
Field
Laboratory
Laboratory
Field
Field
Laboratory
Laboratory
Field
Field
Laboratory
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Field
Laboratory
Field
Field
Field
Laboratory
Laboratory
Laboratory
Laboratory
Laboratory
Field
Field
Laboratory
Sample Description
Titta. River Sed #2
Titta. River Sed #2
Titta. River Sed #2
Titta. River Sed #2
Titta. River Sed #3
Titta. River Sed #3
Titta. River Sed #3
Titta. River Sed #3
WinonaPost#l
WinonaPost#l
WinonaPost#l
WinonaPost#l
Winona Post #2
Winona Post #2
Winona Post #2
Winona Post #2
Winona Post #3
Winona Post #3
Winona Post #3
Winona Post #3
Envir. Extract #1
Envir. Extract #1
Envir. Extract #1
Envir. Extract #1
Envir. Extract #2
Envir. Extract #2
Envir. Extract #2
Envir. Extract #2
Spike #1
Spike #1
Spike #1
Spike #1
Spike #1
Spike #1
Spike #1
Spike #2
Spike #2
Spike #2
Spike #2
Spike #3
Spike #3
Spike #3
Spike #3
Cambridge 5 183
Cambridge 5 183
Cambridge 5 183
Cambridge 5 183
Cambridge 5 183
Cambridge 5 183
Cambridge 5 183
Cambridge 5 184
Cambridge 5 184
Cambridge 5 184
Cambridge 5 184
ERA Aroclor
Replicate
1
2
3
4
1
2
o
J
4
1
2
3
4
1
2
o
J
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
o
J
4
5
6
7
1
2
3
4
1
2
3
4
1
2
3
4
5
6
7
1
2
3
4
1
TEQpr«(P2/2)
Developer3
16
<6.3
<6.25
<6.25
<6.3
<6.25
<6.25
<6.3
21
96.5
>2500
105
102.3
65
215
82.4
33.8
30
137.5
61.5
9
9.5
8
10
12
12.5
16
12.5
<6.3
<6.3
<6.3
<6.3
<6.3
<6.3
<6.3
45
32
80
37.5
1125
1350
1850
1950
7.3
<6.3
16.5
<6.3
<6.25
<6.25
<6.25
160
93.3
212.5
140
190.3
Reference Laboratory1"
0.649
0.71
0.566
0.515
0.0719
0.0973
0.083
0.09
0.654
0.904
0.829
0.822
1.2
1.3
1.32
1.28
1.68
1.87
1.8
2.06
0.629
0.673
0.64
2.08
0.742
0.135
0.297
0.17
0.0638
0.00013
0.0001
0.0275
0.0562
0.00724
0.139
113
113
111
113
1060
1080
1060
990
3.81
4.33
4.2
4.24
4.25
3.86
3.53
1080
1120
1140
1160
1060
D-3

-------
Sample Type
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Performance
Sample
Number
ABRAXIS 110
ABRAXIS 139
ABRAXIS 65
ABRAXIS 36
ABRAXIS 33
ABRAXIS 77
ABRAXIS 179
ABRAXIS 161
ABRAXIS 56
ABRAXIS 189
ABRAXIS 121
ABRAXIS 135
ABRAXIS 40
ABRAXIS 190
ABRAXIS 106
ABRAXIS 171
ABRAXIS 52
ABRAXIS 124
ABRAXIS 96
ABRAXIS 204
ABRAXIS 66
ABRAXIS 90
ABRAXIS 175
ABRAXIS 203
ABRAXIS 67
ABRAXIS 132
ABRAXIS 109
ABRAXIS 101
ABRAXIS 201
ABRAXIS 45
ABRAXIS 194
ABRAXIS 35
ABRAXIS 165
ABRAXIS 108
ABRAXIS 168
ABRAXIS 134
ABRAXIS 117
ABRAXIS 84
ABRAXIS 115
ABRAXIS 94
ABRAXIS 47
ABRAXIS 195
ABRAXIS 99
ABRAXIS 180
ABRAXIS 197
ABRAXIS 25
Measurement
Location
Field
Laboratory
Field
Field
Field
Field
Laboratory
Laboratory
Field
Laboratory
Laboratory
Laboratory
Field
Laboratory
Field
Laboratory
Field
Laboratory
Field
Field
Field
Field
Laboratory
Laboratory
Field
Laboratory
Field
Field
Laboratory
Field
Laboratory
Field
Laboratory
Field
Laboratory
Laboratory
Laboratory
Field
Field
Field
Field
Laboratory
Field
Laboratory
Laboratory
Field
Sample Description
ERA Aroclor
ERA Aroclor
ERA Aroclor
ERA Blank
ERA Blank
ERA Blank
ERA Blank
ERA Blank
ERA Blank
ERA Blank
ERA Blank
ERA PAH
ERA PAH
ERA PAH
ERA PAH
ERA PCB 100
ERA PCB 100
ERA PCB 100
ERA PCB 100
ERA PCB 10000
ERA PCB 10000
ERA PCB 10000
ERA PCB 10000
ERA TCDD 10
ERA TCDD 10
ERA TCDD 10
ERA TCDD 10
ERA TCDD 30
ERA TCDD 30
ERA TCDD 30
ERA TCDD 30
LCG CRM-529
LCG CRM-529
LCG CRM-529
LCG CRM-529
NIST 1944
NIST 1944
NIST 1944
NIST 1944
Wellington WMS - 01
Wellington WMS - 01
Wellington WMS -01
Wellington WMS -01
Wellington WMS -01
Wellington WMS -01
Wellington WMS -01
Replicate
2
3
4
1
2
3
4
5
6
7
8
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
1
2
3
4
1
2
o
3
4
5
6
7
TEQpr«(P2/2)
Developer3
4250
224.5
90
<6.3
6.3
16
<6.25
<6.25
<6.3
<6.25
<6.25
18.5
9
<6.25
8.8
<6.25
<6.3
<6.25
25
1075
190
>1250
1175
<6.25
<6.3
16.2
<6.3
<6.3
<6.25
8.8
<6.25
14.5
82.8
1250
1000
13.5
27.8
<6.3
65
<6.3
<6.3
21.8
<6.3
<6.25
20.5
22
Reference Laboratory1"
3690
3790
3800
0.0243
0.00385
0.00277
0.042
0.0229
0.0191
0.0325
0.0225
0.0254
0.00429
0.00423
0.026
10.6
11.1
10.6
9.95
1030
1030
1180
1020
0.0147
0.0123
0.0299
0.045
0.0451
0.0153
0.0436
0.04
435
405
498
356
40.1
43.7
42.1
41
10.6
9.4
9.62
9.07
10.3
9.62
9.68
1 Data listed exactly
b Qualifier flags (e.g
as reported by the developer.
., J and K flags) included in the raw data have been removed for the purposes of statistical analysis.
                                                                  D-4

-------