»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 ------- |