United States Office of Research and EPA/600/R-98/113
Environmental Protection Development August 1998
Agency Washington, D.C. 20460
vvEPA Environmental Technology
Verification Report
Immunoassay Kit
Strategic Diagnostics Inc.
EnviroGard PCB Test Kit
4^^^^^^^^^^ d^^^^^^^^^^^^^t
-------
EPA/600/R-98/113
August 1998
Environmental Technology
Verification Report
Immunoassay Kit
Strategic Diagnostics Inc.
EnviroGard PCB Test Kit
By
Amy B. Dindal
Charles K. Bayne, Ph.D.
Roger A. Jenkins, Ph.D.
Oak Ridge National Laboratory
Oak Ridge Tennessee 37831-6120
Stephen Billets, Ph.D.
Eric N. Koglin
U.S. Environmental Protection Agency
Environmental Sciences Division
National Exposure Research Laboratory
Las Vegas, Nevada 89193-3478
This demonstration was conducted in cooperation with
U.S. Department of Energy
David Bottrell, Project Officer
Cloverleaf Building, 19901 Germantown Road
Germantown, Maryland 20874EPA Report No.
August 1998
Superfund Innovative Technology
Evaluation Program
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development
(ORD), and the U.S. Department of Energy's Environmental Management (EM) Program, funded and
managed, through Interagency Agreement No. DW89937854 with Oak Ridge National Laboratory, the
verification effort described herein. This report has been peer and administratively reviewed and has been
approved for publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use of a specific product.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
^^^ Office of Research and Development
* *m T" Washington, D.C. 20460
"".
«-- ..^_._.J \
• ssszj
ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
VERIFICATION STATEMENT
TECHNOLOGY TYPE: POLYCHLORINATED BIPHENYL (PCB) FIELD ANALYTICAL
TECHNIQUES
APPLICATION: MEASUREMENT OF PCBs IN SOILS AND SOLVENT EXTRACTS
TECHNOLOGY NAME: EnviroGard PCB TEST KIT
COMPANY: STRATEGIC DIAGNOSTICS INC.
ADDRESS: 111 PENCADER DRIVE
NEWARK, DE 19702-3322
PHONE: (302) 456-6789
The U.S. Environmental Protection Agency (EPA) has created a program to facilitate the deployment of innovative
technologies through performance verification and information dissemination. The goal of the Environmental Technology
Verification (ETV) Program is to further environmental protection by substantially accelerating the acceptance and use
of improved and more cost-effective technologies. The ETV Program is intended to assist and inform those involved in
the design, distribution, permitting, and purchase of environmental technologies. This document summarizes the results
of a demonstration of the Strategic Diagnostics Inc. (SDI) EnviroGard PCB test kit.
PROGRAM OPERATION
EPA, in partnership with recognized testing organizations, objectively and systematically evaluates the performance of
innovative technologies. Together, with the full participation of the technology developer, they develop plans, conduct
tests, collect and analyze data, and report findings. The evaluations are conducted according to a rigorous demonstration
plan and established protocols for quality assurance. EPA's National Exposure Research Laboratory, which conducts
demonstrations of field characterization and monitoring technologies, with the support of the U.S. Department of
Energy's (DOE) Environmental Management (EM) program, selected Oak Ridge National Laboratory (ORNL) as the
testing organization for the performance verification of polychlorinated biphenyls (PCBs) field analytical techniques.
DEMONSTRATION DESCRIPTION
In July 1997, the performance of six PCB field analytical techniques was determined under field conditions. Each
technology was independently evaluated by comparing field analysis results with those obtained using approved reference
methods. Performance evaluation (PE) samples were also used to assess independently the accuracy and comparability
of each technology.
The demonstration was designed to detect and measure PCBs in soil and solvent extracts. The demonstration was
conducted at ORNL in Oak Ridge, Tennessee, from July 22 through July 29, 1997. The study was conducted under two
environmental conditions. The first site was outdoors, with naturally fluctuating temperatures and relative humidity
conditions. The second site was inside a controlled environmental chamber, with generally cooler temperatures and lower
relative humidities. Multiple soil types, collected from sites in Ohio, Kentucky, and Tennessee, were analyzed in this
study. Solutions of PCBs were also analyzed to simulate extracted surface wipe samples. The results of the soil and
extract analyses conducted under field conditions by the technology were compared with results from analyses of
homogenous replicate samples conducted by conventional EPA SW-846 methodology in an approved reference
laboratory. Details of the demonstration, including a data summary and discussion of results, may be found in the report
EPA-VS-SCM-16 The accompanying notice is an integral part of this verification statement August 1998
iii
-------
entitled Environmental Technology Verification Report: Immunoassay Kit, Strategic Diagnostics Inc., EnviroGard
PCB Test Kit,, EPA/600/R-98/113.
TECHNOLOGY DESCRIPTION
The EnviroGard PCB test kit is a competitive binding enzyme immunoassay that performs rapid, interval testing for
PCBs in soils and solutions at specified action levels of 1, 5, 10, and 50 parts per million (ppm). These results are
reported in intervals (e.g., < 1 ppm, 1 to 5 ppm, etc.) rather than in specific quantities (e.g., 6.7 ppm). The test kit is
standardized using Aroclor 1248, but it can also detect Aroclors 1016, 1242, 1254, and 1260. The following items are
needed to run a test: the EnviroGard PCB test kit, the Soil Extraction Bottle kit, the equipment contained in the Soil Field
Lab, methanol, and water. The test procedure entails collecting a 5-g soil sample and extracting the PCBs from it using
methanol. To initiate the PCB test, PCB-enzyme conjugate is added to the antibody-coated test tubes. The soil extract
sample is then added to the test tube. After a 15-min incubation period, the tubes are rinsed and a color developing
solution is added. Color development is inversely related to the PCB concentration (e.g. the darker the color, the less
analyte PCB is present in the sample). PCBs are detected using a photometer that measures the absorbance of each tube.
VERIFICATION OF PERFORMANCE
The following performance characteristics of the EnviroGard PCB test kit were observed:
Throughput: Throughput was 18 samples/hour under outdoor conditions and 9 to 10 samples/hour under chamber
conditions. This rate included sample preparation and analysis.
Ease of Use: Three operators analyzed samples during the demonstration, but the technology can be run by a single
trained operator. Minimal training (2 to 4 h) is required to operate the EnviroGard kit, provided the user has a
fundamental understanding of basic chemical and field analytical techniques.
Completeness: The EnviroGard kit generated results for all 232 PCB samples for a completeness of 100%.
Blank results: All of the blank soil samples were reported as the lowest reporting interval, which included zero; therefore,
the percentage of false positive results was 0%. One false positive result (13%) was reported for the extract samples. The
EnviroGard kit reported no false negative results.
Precision: The overall precision—based on the percentage of combined sample sets where all four replicates were
reported as the same interval—was 38% for the PE soils, 47% for the environmental soils, and 67% for the extracts.
Accuracy: Accuracy was assessed using PE soil and extract samples. Accuracy, defined as the percentage of EnviroGard
results that agreed with the accepted concentration, was 51% for PE soils and 58% for extracts. In general, the percentage
of samples that was biased high was much greater (47% for PE soils and 38% for extracts) than the percentage biased
low (1% for PE soils and 4% for extracts).
Comparability: Comparability, like accuracy, was defined as the percentage of results that agreed with, was above (i.e.,
biased high), or was below (i.e., biased low) the reference laboratory result. The percentage of samples that agreed with
the reference laboratory results was 53% for all soils (PE and environmental) and 63% for extracts. The percentage of
samples that was biased high was again much greater (45% for soils and 38% for extracts) than the percentage that was
biased low (2% for soils and 0% for extracts).
Regulatory Decision-making: One objective of this demonstration was to assess the technology's ability to perform at
regulatory decision-making levels for PCBs, specifically 50 ppm for soils and 100 (jg/100cm2 for surface wipes. For PE
and environmental soil samples in the range of 40 to 60 ppm, 39% of the EnviroGard results agreed with the reference
laboratory. In contrast, 59% were biased high and 2% were biased low. For extract samples representing surface wipe
sample concentrations of 100 fjg/100 cm2 and 1000 fjg/100 cm 2 (assuming a 100cm wipe sample), 63% of the
EPA-VS-SCM-16 The accompanying notice is an integral part of this verification statement August 1998
iv
-------
EnviroGard results agreed with the extract spike concentration. In comparison, the percentage of extract samples that
was biased high was 38%, and the percentage of samples that was biased low was 4%.
Data quality levels: The performance of the EnviroGard PCB test kit was characterized as biased and imprecise about
50% of the time, because nearly half of the data were biased relative to the accepted concentration values (in terms of
accuracy) and had replicate results that were not reported as the same interval (in terms of precision). It should be noted
that there was an increased likelihood of results being biased high as a result of the conservatism that the manufacturer
has incorporated into the calculation of results.
The results of the demonstration show that the EnviroGard PCB test kit can provide useful, cost-effective data for
environmental problem-solving and decision-making. Undoubtedly, it will be employed in a variety of applications,
ranging from serving as a complement to data generated in a fixed analytical laboratory to generating data that will stand
alone in the decision-making process. As with any technology selection, the user must determine if this technology is
appropriate for the application and the project data quality objectives. For more information on this and other verified
technologies, visit the ETV web site at http://www.epa.gov/etv.
Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Office of Research and Development
NOTICE: EPA 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, under circumstances other than those tested, operate at the levels verified. The end user
is solely responsible for complying with any and all applicable Federal, State, and Local requirements.
EPA-VS-SCM-16 The accompanying notice is an integral part of this verification statement August 1998
V
-------
VI
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the nation's natural
resources. The National Exposure Research Laboratory (NERL) is EPA's center for the investigation of
technical and management approaches for identifying and quantifying risks to human health and the
environment. NERL's research goals are to (1) develop and evaluate technologies for the characterization and
monitoring of air, soil, and water; (2) support regulatory and policy decisions; and (3) provide the science
support needed to ensure effective implementation of environmental regulations and strategies.
EPA created the Environmental Technology Verification (ETV) Program to facilitate the deployment of
innovative technologies through performance verification and information dissemination. The goal of the ETV
Program is to further environmental protection by substantially accelerating the acceptance and use of
improved and cost-effective technologies. The ETV Program is intended to assist and inform those involved
in the design, distribution, permitting, and purchase of environmental technologies. This program is
administered by NERL's Environmental Sciences Division in Las Vegas, Nevada.
The U.S. Department of Energy's (DOE's) Environmental Management (EM) program has entered into active
partnership with EPA, providing cooperative technical management and funding support. DOE EM realizes
that its goals for rapid and cost-effective cleanup hinge on the deployment of innovative environmental
characterization and monitoring technologies. To this end, DOE EM shares the goals and objectives of the
ETV.
Candidate technologies for these programs originate from the private sector and must be commercially ready.
Through the ETV Program, developers are given the opportunity to conduct rigorous demonstrations of their
technologies under realistic field conditions. By completing the evaluation and distributing the results, EPA
establishes a baseline for acceptance and use of these technologies.
Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Office of Research and Development
vn
-------
-------
Abstract
In July 1997, the U.S. Environmental Protection Agency (EPA) conducted a demonstration of poly chlorinated
biphenyl (PCB) field analytical techniques. The purpose of this demonstration was to evaluate field analytical
technologies capable of detecting and quantifying PCBs in soils and solvent extracts. The fundamental
objectives of this demonstration were (1) to obtain technology performance information using environmental
and quality control samples, (2) to determine how comparable the developer field analytical results were with
conventional reference laboratory results, and (3) to report on the logistical operation of the technology. The
demonstration design was subjected to extensive review and comment by EPA's National Exposure Research
Laboratory (NERL) Environmental Sciences Division in Las Vegas, Nevada; Oak Ridge National Laboratory
(ORNL); EPA Regional Offices; the U.S. Department of Energy (DOE); and the technology developers.
The demonstration study was conducted at ORNL under two sets of environmental conditions. The first site
was outdoors, with naturally variable temperature and relative humidity conditions typical of eastern Tennessee
in the summer. A second site was located inside a controlled environmental chamber having lower, and
relatively stable, temperature and relative humidity conditions. The test samples analyzed during this
demonstration were performance evaluation (PE) soil, environmental soil, and extract samples. Actual
environmental soil samples, collected from sites in Ohio, Kentucky, and Tennessee, were analyzed and ranged
in concentration from 0.1 to 700 parts per million (ppm). Extract samples were used to simulate surface wipe
samples, and were evaluated at concentrations ranging from 0 to 100 (jg/mL. The reference laboratory method
used to evaluate the comparability of data was EPA SW-846 Method 8081.
The field analytical technologies tested in this demonstration were the L2000 PCB/Chloride Analyzer (Dexsil
Corporation), the PCB Immunoassay Kit (Hach Company), the 4100 Vapor Detector (Electronic Sensor
Technology), and three immunoassay kits: D TECH, EnviroGard, and RaPID Assay System (Strategic
Diagnostics Inc.). The purpose of an Environmental Technology Verification Report (ETVR) is to document
the demonstration activities, present demonstration data, and verify the performance of the technology. This
ETVR presents information regarding the performance of SDFs EnviroGard PCB test kit. Separate ETVRs
have been published for the other technologies demonstrated.
The EnviroGard PCB test kit is an immunoassay kit used to determine PCB concentrations as interval threshold
values. The EnviroGard PCB test kit uses a competitive binding enzyme immunoassay to perform rapid,
interval testing for PCBs in soils and solutions at specified action levels of 1, 5, 10, and 50 ppm. The test kit
is standardized using Aroclor 1248, but it can also detect Aroclors 1016, 1242, 1254, and 1260. The presence
of PCBs is detected by a photometer based on a colored reaction in which the color development is inversely
proportional to the concentration of PCB in the sample (e.g., the darker the color, the less analyte PCB is
present in the sample). The EnviroGard kit provides no information on Aroclor identification.
IX
-------
The EnviroGard's quantitative results were based on the analysis of threshold standards with every batch of
12 samples. Because the EnviroGard kit was an interval technique, method detection limits are not applicable.
Precision, defined as the percentage of the sample sets where all four replicates were reported as the same
interval range, was 38% for PE soils, 47% for environmental soils, and 67% for extracts. Accuracy, defined
as the percentage of EnviroGard results that agreed with the accepted concentration, was 51% for PE soils and
58% for extracts. In general, the percentage of results that was biased high was much greater (47% for PE soils
and 38% for extracts) than the percentage of samples that was biased low (1% for PE soils and 4% for
extracts). Comparability was defined similarly to accuracy, but the EnviroGard result was compared with the
reference laboratory result rather than with the accepted concentration to determine comparability. For all soil
samples (PE and environmental), the percentage of EnviroGard results that agreed with the reference laboratory
results was 53%, and the percentage that was biased high was much greater than the percentage biased low.
The demonstration found that the EnviroGard kit was simple to operate in the field, requiring about an hour
for initial set-up and preparation for sample analysis. Once the kit was operational, the sample throughput of
the EnviroGard kit was 18 samples/hour under outdoor conditions and 9 to 10 samples/ hour under chamber
conditions. Three operators analyzed samples during the demonstration, but the technology can be run by a
single, trained operator. Minimal training (2 to 4 h) is required to operate the EnviroGard kit, provided the user
has a fundamental understanding of basic chemical and field analytical techniques. The overall performance
of the EnviroGard PCB test kit was characterized as biased and imprecise about 50% of the time; however,
the kit generated no false positive or false negative results for soil samples. It should be noted that there was
an increased likelihood that results would be biased high as a result of the conservatism that the manufacturer
has incorporated into the calculation of results.
-------
Table of Contents
Notice ii
Verification Statement iii
Foreword vii
Abstract ix
List of Figures xv
List of Tables xvii
List of Abbreviations and Acronyms xix
Acknowledgments xxii
Section 1 Introduction 1
Technology Verification Process 2
Needs Identification and Technology Selection 2
Demonstration Planning and Implementation 3
Report Preparation 3
Information Distribution 3
Demonstration Purpose 4
Section 2 Technology Description 5
Objective 5
Principle 5
Method Overview 5
Materials 5
Procedure 6
Extraction 6
Assay 6
Interpreting Results 7
Possible Interfering Compounds 8
Section 3 Site Description and Demonstration Design 9
Objective 9
xi
-------
Demonstration Site Description 9
Site Name and Location 9
Site History 9
Site Characteristics 10
Experimental Design 10
Environmental Conditions during Demonstration 12
Sample Descriptions 13
Performance Evaluation Materials 13
Environmental Soil Samples 13
Extract Samples 14
Sampling Plan 14
Sample Collection 14
Sample Preparation, Labeling, and Distribution 14
Predemonstration Study 16
Predemonstration Sample Preparation 16
Predemonstration Results 17
Deviations from the Demonstration Plan 17
Section 4 Reference Laboratory Analytical Results and Evaluation 19
Objective and Approach 19
Reference Laboratory Selection 19
Reference Laboratory Method 20
Calibration 20
Sample Quantification 20
Sample Receipt, Handling, and Holding Times 21
Quality Control Results 21
Objective 21
Continuing Calibration Verification Standard Results 21
Instrument and Method Blank Results 22
Surrogate Spike Results 22
Laboratory Control Sample Results 22
Matrix Spike Results 23
Conclusions of the Quality Control Results 23
Data Review and Validation 23
Objective 23
Corrected Results 24
Suspect Results 24
Data Assessment 25
Objective 25
Precision 25
Performance Evaluation Samples 25
Environmental Soil Samples 26
Extract Samples 28
Accuracy 28
xii
-------
Performance Evaluation Soil Samples 29
Extract Samples 30
Representativeness 30
Completeness 30
Comparability 31
Summary of Observations 31
Section 5 Technology Performance and Evaluation 33
Objective and Approach 33
Interval Reporting 33
Data Assessment 33
Objective 33
Precision 34
Performance Evaluation Samples 34
Environmental Soil Samples 35
Extract Samples 37
Precision Summary 37
Accuracy 38
Performance Evaluation Soil Samples 38
Extract Samples 39
False Positive/False Negative results 40
Representativeness 40
Completeness 41
Comparability 41
Summary of PARCC Parameters 42
Regulatory Decision-Making Applicability 43
Additional Performance Factors 43
Sample Throughput 43
Cost Assessment 43
EnviroGard Costs 44
Reference Laboratory Costs 45
Cost Assessment Summary 46
General Observations 46
Performance Summary 47
Section 6 Technology Update and Representative Applications 49
Objective 49
Technology Update 49
Reconfiguration of Soil Extraction (Sample Preparation) Products 49
Instrument Consolidation 49
Representative Applications 49
Data Quality Objective Example 49
Section 7 References 51
xiii
-------
Appendix A Description of Environmental Soil Samples 53
Appendix B Characterization of Environmental Soil Samples 57
Appendix C Temperature and Relative Humidity Conditions 61
Appendix D PCB Technology Demonstration Sample Data 67
Appendix E Data Quality Objective Example 77
Disclaimer 79
Background and Problem Statement 79
DQO Goals 79
Use of Performance Information to Implement the Decision Rule 80
Determining the Number of Samples 80
Alternative FP Parameter 83
xiv
-------
List of Figures
3-1. Schematic map of ORNL, indicating the demonstration area 11
C-l. Summary of temperature conditions for outdoor site 64
C-2. Summary of relative humidity conditions for outdoor site 64
C-3. Summary of temperature conditions for chamber site 65
C-4. Summary of relative humidity conditions for chamber site 65
xv
-------
-------
List of Tables
2-1. Test tube labels 7
2-2. Interpretation of photometer readings 8
2-3. Converting to Aroclor-specific concentrations 8
3-1. Summary of experimental design by sample type 12
3-2. Summary of EnviroGard predemonstration results 17
4-1. Suspect measurements within the reference laboratory data 24
4-2. Precision of the reference laboratory for PE soil samples 26
4-3. Precision of the reference laboratory for environmental soil samples 27
4-4. Precision of the reference laboratory for extract samples 28
4-5. Accuracy of the reference laboratory for PE soil samples 29
4-6. Accuracy of the reference laboratory for extract samples 30
4-7. Summary of the reference laboratory performance 31
5-1. EnviroGard PCB test kit reporting intervals 34
5-2. Classification of precision results 34
5-3. Precision of the EnviroGard PCB test kit for PE soil samples 35
5-4. Precision of the EnviroGard PCB test kit for environmental soil samples 36
5-5. Precision of the EnviroGard PCB test kit for extract samples 37
5-6. Overall precision of the EnviroGard PCB test kit for all sample types 37
5-7. EnviroGard test kit accuracy data for PE soil samples 38
5-8. Evaluation of agreement between EnviroGard's PE sample results and the certified PE values
as a measure of accuracy 39
5-9. Accuracy of the EnviroGard test kit for extract samples 40
5-10. Evaluation of agreement between EnviroGard's extract results and the spike concentration
as a measure of accuracy 40
5-11. Evaluation of agreement between EnviroGard's soil results and the reference laboratory's
results as a measure of comparability 41
5-12. Comparison of the EnviroGard results with the reference laboratory's suspect measurements ... 42
5-13. EnviroGard PCB test kit performance for precision, accuracy, and comparability 42
5-14. Estimated analytical costs for PCB soil samples 44
5-15. Performance summary for the EnviroGard PCB test kit 48
A-l. Summary of soil sample descriptions 55
B-l. Summary of environmental soil characterization 59
C-l. Average temperature and relative humidity conditions during testing periods 63
D-l EnviroGard PCB technology demonstration soil sample data 69
D-2 EnviroGard PCB test kit technology demonstration extract sample data 74
D-3. Corrected reference laboratory data 75
xvn
-------
-------
List of Abbreviations and Acronyms
AL action level
ANOVA analysis of variance
ASTM American Society for Testing and Materials
BHC benzenehexachloride
C concentration at which the false positive error rate is specified
CASD Chemical and Analytical Sciences Division (ORNL)
CCV continuing calibration verification standard
CSCT Consortium for Site Characterization Technology
DCB decachlorobiphenyl
DOE U. S. Department of Energy
DQO data quality objective
ELISA enzyme-linked immunosorbent assay
EM Environmental Management (DOE)
EPA U. S. Environmental Protection Agency
ERA Environmental Resource Associates
EST Electronic Sensor Technology
ETTP East Tennessee Technology Park
ETV Environmental Technology Verification Program
ETVR Environmental Technology Verification Report
EvTEC Environmental Technology Evaluation Center
fn false negative result
FN false negative decision error rate
xix
-------
fp false positive result
FP false positive decision error rate
HEPA high-efficiency participate air
ID identifier
LCS laboratory control sample
LMER Lockheed Martin Energy Research
LMES Lockheed Martin Energy Systems
LV Las Vegas
MDL method detection limit
MS matrix spike
MSB matrix spike duplicate
n number of samples
NERL National Exposure Research Laboratory (EPA)
NCEPI National Center for Environmental Publications and Information
NRC Nuclear Regulatory Commission
ORD Office of Research and Development (EPA)
ORNL Oak Ridge National Laboratory
ORO Oak Ridge Operations (DOE)
PARCC precision, accuracy, representativeness, completeness, comparability
PCB polychlorinated biphenyl
PE performance evaluation
ppb parts per billion
ppm parts per million; equivalent units: mg/kg for soils and (jg/mL for extracts
Pr probability
QA quality assurance
QC quality control
R2 coefficient of determination
xx
-------
RDL reporting detection limit
RH relative humidity
RFD request for disposal
RPD relative percent difference
RSD relative standard deviation (percent)
RT regulatory threshold
S2 variance for the measurement
SARA Superfund Amendments and Reauthorization Act of 1986
SD standard deviation
SDI Strategic Diagnostics Inc.
SITE Superfund Innovative Technology Evaluation
SMO sample management office
SOP standard operating procedure
SSM synthetic soil matrix
TCMX tetrachloro-m-xylene
TSCA Toxic Substance Control Act
Z!_P the (l-p)th percentile for the standard normal distribution
%D percent difference
xxi
-------
-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the demonstration,
analyze the data, and prepare this report. In particular, we recognize the technical expertise of Mitchell
Erickson (Environmental Measurements Laboratory), Viorica Lopez-Avila (Midwest Research Institute), and
Robert F. O'Brien (Pacific Northwest National Laboratory), who were peer reviewers of this report. For
internal peer review, we thank Stacy Barshick (ORNL); for technical support during the demonstration, Todd
Skeen and Ralph Ilgner (ORNL); for site safety and health support, Kim Thomas, Marilyn Hanner, and Fred
Smith (ORNL); for administrative support, Betty Maestas and Linda Plemmons (ORNL); for sample collection
support, Wade Hollinger, Charlotte Schaefer, and Arlin Yeager (LMES), and Mike Rudacille and W. T.
Wright (EET Corporation); for preliminary soil characterization support, Frank Gardner, John Zutman, and
Bob Schlosser (ORNL, Grand Junction, Colo.); for sample management support, Angie McGee, Suzanne
Johnson, and Mary Lane Moore (LMES); for providing performance evaluations samples, Michael Wilson
(EPA's Office of Solid Waste and Emergency Response's Analytical Operations and Data Quality Center);
and for technical guidance and project management of the demonstration, David Garden, Marty Atkins, and
Regina Chung (DOE's Oak Ridge Operations Office), David Bottrell (DOE, Headquarters), Deana Crumbling
(EPA's Technology Innovation Office), and Stephen Billets, Gary Robertson, and Eric Koglin (EPA's National
Exposure Research Laboratory, Las Vegas, Nevada). The authors also acknowledge the participation of
Strategic Diagnostics Inc., in particular, Craig Kostyshyn, Tim Lawruk, Chris Jones, and Penny Kosinski, who
performed the analyses during the demonstration.
For more information on the PCB Field Analytical Technology Demonstration, contact
Eric N. Koglin
Project Technical Leader
Environmental Protection Agency
Characterization and Research Division
National Exposure Research Laboratory
P. O. Box 93478
Las Vegas, Nevada 89193-3478
(702) 798-2432
For more information on the EnviroGard PCB test kit, contact
Tim Lawruk
Strategic Diagnostics Inc.
Ill Pencader Drive
Newark, DE 19702-3322
(302) 456-6789
xxni
-------
Section 1
Introduction
The performance evaluation of innovative and alternative environmental technologies is an integral part of the
U.S. Environmental Protection Agency's (EPA's) mission. Early efforts focused on evaluating technologies
that supported the implementation of the Clean Air and Clean Water Acts. In 1987, the Agency began to
evaluate the cost and performance of remediation and monitoring technologies under the Superfund Innovative
Technology Evaluation (SITE) program. This was in response to the mandate in the Superfund Amendments
and Reauthorization Act (SARA) of 1986. In 1990, the U.S. Technology Policy was announced. This policy
placed a renewed emphasis on "making the best use of technology in achieving the national goals of improved
quality of life for all Americans, continued economic growth, and national security." In the spirit of the
Technology Policy, the Agency began to direct a portion of its resources toward the promotion, recognition,
acceptance, and use of U.S.-developed innovative environmental technologies both domestically and abroad.
The Environmental Technology Verification (ETV) Program was created by the Agency to facilitate the
deployment of innovative technologies through performance verification and information dissemination. The
goal of the ETV Program is to further environmental protection by substantially accelerating the acceptance
and use of improved and cost-effective technologies. The ETV Program is intended to assist and inform those
involved in the design, distribution, permitting, and purchase of environmental technologies. The ETV Program
capitalizes upon and applies the lessons that were learned in the implementation of the SITE Program to the
verification of twelve categories of environmental technology: Drinking Water Systems, Pollution
Prevention/Waste Treatment, Pollution Prevention/ Innovative Coatings and Coatings Equipment, Indoor Air
Products, Air Pollution Control, Advanced Monitoring Systems, EvTEC (an independent, private-sector
approach), Wet Weather Flow Technologies, Pollution Prevention/Metal Finishing, Source Water Protection
Technologies, Site Characterization and Monitoring Technology [also referred to as the Consortium for Site
Characterization Technology (CSCT)], and Climate Change Technologies. The performance verification
contained in this report was based on the data collected during a demonstration of polychlorinated biphenyl
(PCB) field analytical technologies. The demonstration was administered by CSCT.
For each pilot, EPA utilizes the expertise of partner "verification organizations" to design efficient procedures
for conducting performance tests of environmental technologies. To date, EPA has partnered with federal
laboratories and state, university, and private sector entities. Verification organizations oversee and report
verification activities based on testing and quality assurance protocols developed with input from all major
stakeholder/customer groups associated with the technology area.
-------
In July 1997, CSCT, in cooperation with the U.S. Department of Energy's (DOE's) Environmental
Management (EM) Program, conducted a demonstration to verify the performance of six field analytical
technologies for PCBs: the L2000 PCB/Chloride Analyzer (Dexsil Corporation), the PCB Immunoassay Kit
(Hach Company), the 4100 Vapor Detector (Electronic Sensor Technology), and three immunoassay kits from
Strategic Diagnostics Inc. (SDI): D TECH, EnviroGard, and RaPID Assay System. This environmental
technology verification report (ETVR) presents the results of the demonstration study for one PCB field
analytical technology, SDFs EnviroGard PCB test kid. Separate ETVRs have been published for the other five
technologies.
Technology Verification Process
The technology verification process is intended to serve as a template for conducting technology demonstrations
that will generate high-quality data that EPA can use to verify technology performance. Four key steps are
inherent in the process:
• Needs identification and technology selection
• Demonstration planning and implementation
• Report preparation
• Information distribution
Needs Identification and Technology Selection
The first aspect of the technology verification process is to determine technology needs of EPA and the
regulated community. EPA, DOE, the U.S. Department of Defense, industry, and state agencies are asked to
identify technology needs and interest in a technology. Once a technology need is established, a search is
conducted to identify suitable technologies that will address this need. The technology search and identification
process consists of reviewing responses to Commerce Business Daily announcements, searches of industry and
trade publications, attendance at related conferences, and leads from technology developers. Characterization
and monitoring technologies are evaluated against the following criteria:
• meets user needs;
• may be used in the field or in a mobile laboratory;
• is applicable to a variety of environmentally impacted sites;
• has high potential for resolving problems for which current methods are unsatisfactory;
• is cost competitive with current methods;
• performs better than current methods in areas such as data quality, sample preparation, or analytical
turnaround time;
-------
• uses techniques that are easier and safer than current methods;
• is a commercially available, field-ready technology.
Demonstration Planning and Implementation
After a technology has been selected, EPA, the verification organization, and the developer agree to the
responsibilities for conducting the demonstration and evaluating the technology. The following tasks are
undertaken at this time:
• identifying demonstration sites that will provide the appropriate physical or chemical environment,
including contaminated media;
• identifying and defining the roles of demonstration participants, observers, and reviewers;
• determining logistical and support requirements (for example, field equipment, power and water
sources, mobile laboratory, communications network);
• arranging analytical and sampling support;
• preparing and implementing a demonstration plan that addresses the experimental design, sampling
design, quality assurance/quality control (QA/QC), health and safety considerations, scheduling of
field and laboratory operations, data analysis procedures, and reporting requirements.
Report Preparation
Innovative technologies are evaluated independently and, when possible, against conventional technologies. The
field technologies are operated by the developers in the presence of independent technology observers. The
technology observers are provided by EPA or a third-party group. Demonstration data are used to evaluate the
capabilities, limitations, and field applications of each technology. Following the demonstration, all raw and
reduced data used to evaluate each technology are compiled into a technology evaluation report, which is
mandated by EPA as a record of the demonstration. A data summary and detailed evaluation of each
technology are published in an ETVR.
Information Distribution
The goal of the information distribution strategy is to ensure that ETVRs are readily available to interested
parties through traditional data distribution pathways, such as printed documents. Documents are also available
on the World Wide Web through the ETV Web site (http://www.epa.gov/etv) and through a Web site supported
by the EPA Office of Solid Waste and Emergency Response's Technology Innovation Office (http://CLU-
in.com).
-------
Demonstration Purpose
The purpose of this demonstration was to obtain performance information for PCB field analytical
technologies, to compare the results with conventional fixed-laboratory results, and to provide supplemental
information (e.g., cost, sample throughput, and training requirements) regarding the operation of the
technology. The demonstration was conducted under two climatic conditions. One set of activities was
conducted outdoors, with naturally fluctuating temperatures and relative humidity conditions. A second set was
conducted in a controlled environmental facility, with lower, relatively stable temperatures and relative
humidities. Multiple soil types, collected from sites in Ohio, Kentucky, and Tennessee, were used in this study.
PCB soil concentrations ranged from approximately 0.1 to 700 parts per million (ppm). Developers also
analyzed 24 solutions of known PCB concentration that were used to simulate extracted wipe samples. The
extract samples ranged in concentration from 0 to 100 (jg/mL.
-------
Section 2
Technology Description
Objective
The objective of this section is to describe the technology being demonstrated, including the operating principles
underlying the technology and the overall approach to its use. The information provided here is excerpted from
that provided by the developer. Performance characteristics described in this section are specified by the
developer; they may or may not be substantiated by the data presented in Section 5.
Principle
The EnviroGard PCB test kit assay system for PCB analysis applies the principles of enzyme-linked
immunosorbent assay (ELISA) to the determination of PCBs. In such an assay, an enzyme has been chemically
linked to a PCB molecule or PCB analog to create a labeled PCB reagent. The labeled PCB reagent (called a
conjugate) is mixed with an extract of native sample containing the PCB contaminant. A portion of the mixture
is applied to a surface to which an antibody specific for PCB has been affixed. The native PCB and
PCB-enzyme conjugate compete for a limited number of antibody sites. After a period of time, the solution is
washed away, and what remains is either PCB-antibody complexes or enzyme- PCB-antibody complexes
attached to the test surface. The proportion of the two complexes on the test surface is determined by the
amount of native PCB in the original sample. The enzyme present on the test surface is used to catalyze a color
change reaction in a solution added to the test surface. Because the amount of enzyme present is inversely
proportional to the concentration of native PCB contaminant, the amount of color development is inversely
proportional to the concentration of PCB contaminant.
The EnviroGard PCB test kit is a system that performs rapid, semi-quantitative testing for PCBs in soil and
in solution at specified action levels of 1, 5, 10, and 50 ppm. The kit is standardized using Aroclor 1248, but
this test can also detect Aroclors 1016, 1242, 1254, and 1260. The test screens for PCBs with 95% confidence
of no false negatives at the action levels.
Method Overview
The following items are needed to run a test: the EnviroGard PCB test kit, the Soil Extraction Bottle kit, the
equipment contained in the Soil Field Lab, methanol, and water. The test procedure entails collecting a 5-g soil
sample and extracting the PCBs from it using methanol. To initiate the PCB test, PCB-enzyme conjugate is
added to the antibody-coated test tubes. The soil extract is then added to the test tube. After a 15-min
incubation period, the tubes are rinsed and a color developing solution is added. A photometer is used to
measure the absorbance of each tube. Color development is inversely related to the PCB concentration (i.e.,
a darker color indicates less PCB than calibrator material).
Materials
The following is a list of the materials needed to perform an assay:
-------
• EnviroGard PCB test kit (including 20 antibody-coated test tubes, four calibrators, and one negative
control standard; store at 4to 8°C; do not freeze; check expiration date). Note: The stop solution in
the kit is hydrochloric acid. Avoid contact with skin or eyes.
• EnviroGard Soil Extraction Bottle kit (including weighing boats, wooden spatulas, soil extraction
bottles containing three mixing beads, 20-cc syringe, syringe coupler, filter units, filter caps, 4-mL
glass storage vials, stoppers, and blank labels). Note: This kit contains enough material for 14 soil
samples.
• EnviroGard Soil Field Lab (including photometer, charger, Repeater™ pipettes and tips, positive
displacement pipette and tips, timer, wash bottle, test tube rack). Note: Make sure the portable
balance works and has fresh batteries.
• Portable balance
• Methanol
• Tap or distilled water
• Marker with water-resistant ink
• Lab coat, gloves, and goggles
• Paper towels
Procedure
Extraction
Weigh out 5 g of soil using the portable balance, weigh boat, and wooden spatula. Uncap the soil extraction
bottle and label it appropriately. Fold the weigh boat into the mouth of the bottle and pour in the soil sample.
Attach the 50-mL tip to the Repeater pipette and set the dial to 5. (This is equivalent to 5 mL.) Add 5 mL of
methanol to each bottle. (Note: It may be necessary to add an additional 5 mL ofmethanol if the sample soaks
up all of the methanol, leaving little or no excess liquid to decant.) Screw the cap onto the extraction bottle
and tighten. Agitate the bottle for 2 min. Remove and discard the extraction bottle cap. Tightly screw a filter
cap on the bottle. Attach a filter unit to the filter cap. Draw air into the syringe by pulling the plunger to the
20 mL mark. Twist the syringe firmly onto the open end of the filter unit. Push down the plunger. This creates
enough pressure in the soil extraction bottle to drive the soil extract through the filter. Hold the filter unit and
twist off the syringe coupler to remove the syringe assembly. Immediately invert the pressurized extract bottle
and insert the filter outlet into the mouth of the glass storage vial. Hold the vial steady (or place it in a rack)
until the necessary amount of soil extract (a minimum of 25 ^L) has been filtered. Cap the glass storage vial.
Assay
Remove up to 20 of the antibody-coated test tubes from the kit and label them according to Table 2-1.
-------
Table 2-1. Test tube labels
Tube label
NC
1 ppm
5 ppm
10 ppm
50 ppm
SI
S2
Indicates...
negative control
1 ppm PCB calibrator"
5 ppm PCB calibrator"
10 ppm PCB calibrator"
50 ppm PCB calibrator"
Sample 1
Sample 2
a The calibrators have actual concentrations of 0.5, 3, 5, and 22 ppm
of Aroclor 1248, respectively.
Using the Repeater pipette labeled "CON," dispense 500 i\L of enzyme conjugate into each tube. Using the
positive displacement pipette, dispense 25 yL of the negative control, calibrators, and sample(s) into the
appropriate test tubes. Shake the test tube rack thoroughly for 5 s to mix the contents. Then set the timer for
15 min and let the tubes incubate undisturbed.
After the incubation period, empty the test tube contents into a sink or suitable container. Fill the tubes with
water. Then empty them and shake out the remaining drops. Repeat the wash three times. Then invert the tubes
and tap them on paper towels to remove excess water. Using the Repeater pipette labeled "SUB," dispense 500
(iL of the substrate into each tube. Let the tubes incubate for 5 min. During this time, you should see the tubes
turn varying shades of blue, depending on the PCB concentration. (Note: If a blue color does not develop in
the "NC" test tube within 5 min after adding the substrate, this test is invalid and must be repeated.)
After 5 min, add 500 \\L of the stop solution to each tube using the Repeater pipette labeled "STOP". The tube
contents should turn yellow. Use caution when handling the stop solution because it is an acid.
Add 1 mL of the stop solution or wash water to a new empty test tube . (This is the photometer "blank' tube.)
Insert the blank tube into the left well of the photometer. Dry the outside of each labeled tube with a clean paper
towel and place each tube (one by one) into the right well of the photometer. Record the absorbance (optical
density) reading of each tube. The tube contents must be read with a photometer within 30 min after adding
the stop solution. Dispose of the tube in an appropriate waste container.
Interpreting Results
Interpret the photometer readings using Table 2-2.
The EnviroGard PCB test kit is calibrated to Aroclor 1248. If it is known that other Aroclors are present in
the sample, Table 2-3 should be used for result interpretation.
-------
Table 2-2. Interpretation of photometer readings
A sample with an absorbance reading...
> (the reading of the 1 ppm PCB calibrator)
< (the reading of the 1 ppm PCB calibrator) and
> (the reading of the 5 ppm PCB calibrator)
< (the reading of the 5 ppm PCB calibrator) and
> (the reading of the 10 ppm PCB calibrator)
< (the reading of the 10 ppm PCB calibrator) and
> (the reading of the 50 ppm PCB calibrator)
< (the reading of the 50 ppm PCB calibrator)
...contains...
< 1.0 ppm PCB
> 1 ppm PCB and
< 5 ppm PCB
> 5 ppm PCB and
< 10 ppm PCB
> 10 ppm PCB and
< 50 ppm PCB
> 50 ppm PCB
...and is reported as...
[0,1)
[1,5]
(5, 10]
(10,50]
(50, co)
Table 2-3. Converting to Aroclor-speciflc concentrations
Aroclor 1248
1 ppm
5 ppm
10 ppm
50 ppm
Aroclor 1260
2 ppm
10 ppm
20 ppm
100 ppm
Aroclor 1242
2 ppm
10 ppm
20 ppm
100 ppm
Aroclor 1254
1.1 ppm
5.5 ppm
11.1 ppm
55 ppm
Possible Interfering Compounds
The following substances were tested and found to have less than 0.5% weight-to-weight of the
immunoreactivity of Aroclor 1248: 1,2-dichlorobenzene, 1,3-dichlorobenzene, 1,4-dichlorobenzene, 1,2,4-
trichlorobenzene, 2,4-dichlorophenol, 2,5-dichlorophenol, 2,4,5-trichlorophenol, 2,4,6-trichlorophenol,
biphenyl, and pentachlorophenol.
-------
Section 3
Site Description and Demonstration Design
Objective
This section describes the demonstration site, the experimental design for the verification test, and the sampling
plan (sample types analyzed and the collection and preparation strategies). Included in this section are the
results from the predemonstration study and a description of the deviations made from the original
demonstration design.
Demonstration Site Description
Site Name and Location
The demonstration of PCB field analytical technologies was conducted at Oak Ridge National Laboratory
(ORNL) in Oak Ridge, Tennessee. PCB-contaminated soils from three DOE sites (Oak Ridge; Paducah,
Kentucky; and Piketon, Ohio) were used in this demonstration. The soil samples used in this study were
brought to the demonstration testing location for evaluation of the field analytical technologies.
Site History
Oak Ridge is located in the Tennessee River Valley, 25 miles northwest of Knoxville. Three DOE facilities are
located in Oak Ridge: ORNL, the Oak Ridge Y-12 Plant, and East Tennessee Technology Park (ETTP).
Chemical processing and warhead component production have occurred at the Y-12 Plant, and ETTP is a
former gaseous diffusion uranium enrichment plant. At both facilities, industrial processing associated with
nuclear weapons production has resulted in the production of millions of kilograms of PCB-contaminated soils.
Two other DOE facilities—the Paducah plant in Paducah, Kentucky, and the Portsmouth plant in Piketon,
Ohio—are also gaseous diffusion facilities with a history of PCB contamination. During the remediation of the
PCB-contaminated areas at the three DOE sites, soils were excavated from the ground where the PCB
contamination occurred, packaged in containers ranging in size from 55-gal to 110-gal drums, and stored as
PCB waste. Samples from these repositories—referred to as "Oak Ridge," "Portsmouth," and "Paducah"
samples in this report—were used in this demonstration.
In Oak Ridge, excavation activities occurred between 1991 and 1995. The Oak Ridge samples comprised PCB-
contaminated soils from both Y-12 and ETTP. Five different sources of PCB contamination resulted in soil
excavations from various dikes, drainage ditches, and catch basins. Some of the soils are EPA-listed hazardous
waste due to the presence of other contaminants (e.g., diesel fuels).
A population of more than 5000 drums containing PCB-contaminated soils was generated from 1986 to 1987
during the remediation of the East Drainage Ditch at the Portsmouth Gaseous Diffusion Plant. The ditch was
reported to have three primary sources of potential contamination: (1) treated effluent from a radioactive liquid
treatment facility, (2) runoff from a biodegradation plot where waste oil and sludge were disposed of, and (3)
storm sewer discharges. In addition, waste oil was reportedly used for weed control in the ditch. Aside from
-------
PCB contamination, no other major hazardous contaminants were detected in these soils. Therefore, no EPA
hazardous waste codes are assigned to this waste.
Twenty-nine drums of PCB-contaminated soils from the Paducah plant were generated as part of a spill
cleanup activity at an organic waste storage area (C-746-R). The waste is considered a listed hazardous waste
for spent solvents (EPA hazardous waste code F001) because it is known to contain trichloroethylene. Other
volatile organic compounds, such as xylene, dichlorobenzene, and cresol, were also detected in the preliminary
analyses of some of the Paducah samples.
Site Characteristics
PCB-contaminated environmental soil samples from Oak Ridge, Portsmouth, and Paducah were collected from
waste containers at storage repositories at ETTP and Paducah. Many of the soils contained interfering
compounds such as oils, fuels, and other chlorinated compounds (e.g., trichloroethylene). Specific sample
descriptions of the environmental soils used in this demonstration are given in Appendix A. In addition, each
sample was characterized in terms of its soil composition, pH, and total organic carbon content. Those results
are summarized in Appendix B.
Field demonstration activities occurred at two sites at ORNL: a natural outdoor environment (the outdoor site)
and inside a controlled environmental atmosphere chamber (the chamber site). Figure 3-1 shows a schematic
map of a section of ORNL indicating the demonstration area where the outdoor field activities occurred.
Generally, the average summer temperature in eastern Tennessee is 75.6° F, with July and August temperatures
averaging 79.1 °F and 76.8 °F, respectively. Average temperatures during the testing periods ranged from 79
to 85 °F, as shown in Appendix C. Studies were also conducted inside a controlled environmental atmosphere
chamber, hereafter referred to as the "chamber," located in Building 5507 at ORNL. Demonstration studies
inside the chamber were used to evaluate performance under environmental conditions that were markedly
different from the ambient outdoor conditions at the time of the test. Average temperatures in the chamber
during the testing periods ranged from 55 to 70°F. The controlled experimental atmosphere facility consists
of a room-size walk-in chamber 10 ft wide and 12 ft long with air processing equipment to control temperature
and humidity. The chamber is equipped with an environmental control system, including reverse osmosis water
purification that supplies the chamber humidity control system. High efficiency particulate air (HEPA) and
activated charcoal filters are installed for recirculation and building exhaust filtration.
Experimental Design
The analytical challenge with PCB analysis is to quantify a complex mixture that may or may not resemble the
original commercial product (i.e., Aroclor) because of environmental aging, and to report the result as a single
number [1]. The primary objective of the verification test was to compare the performance of the field
technology with laboratory-based measurements. Often, verification tests involve a direct one-to-one
comparison of results from field-acquired samples. However, because sample heterogeneity can preclude
replicate field or laboratory comparison, accuracy and precision data must often be derived from the analysis
of QC and performance evaluation (PE) samples. In this study, replicates of all three sample types (QC, PE,
and environmental soil) were analyzed. The ability to use environmental soils in the verification test was made
possible because the samples, collected from drums
10
-------
Building
**»
Figure 3-1. Schematic map of ORNL, indicating the demonstration area.
containing PCB-contaminated soils, could be thoroughly homogenized and characterized prior to the
demonstration. This facet of the design, allowing additional precision data to be obtained on actual field-
acquired samples, provided an added performance factor in the verification test.
Another objective of this demonstration was to evaluate the field technology's capability to support regulatory
compliance decisions. For field methods to be used in these decisions, the technology must be capable of
informing the user, with known precision and accuracy, that soil concentrations are greater than or less than
50 ppm, and that wipe samples are greater than or less than 100 (jg/100 cm2 [2]. The samples selected for
analysis in the demonstration study were chosen with this objective in mind.
The experimental design is summarized in Table 3-1. This design was approved by all participants prior to the
start of the demonstration study. In total, the developers analyzed 208 soil samples, 104 each at both locations
(outdoors and chamber). The 104 soil samples comprised 68 environmental samples (17 unique environmental
samples prepared in quadruplicate) ranging in PCB concentration from 0.1 to 700 ppm and 36 PE soils (9
unique PE samples in quadruplicate) ranging in PCB concentration from 0 to 50 ppm. To determine the impact
of different environmental conditions on the technology's performance, each batch of 104 samples contained
5 sets of quadruplicate soil samples from DOE's Paducah site. These were analyzed under both sets of
environmental conditions (i.e., outdoor and chamber conditions). For the developers participating in the extract
sample portion (i.e., simulated wipe samples) of the demonstration, 12 extracts, ranging in concentration from
0 to 100 (jg/mL, were analyzed in each location (chamber and outdoors). All samples were analyzed without
prior knowledge of sample type or concentration and were analyzed in a randomized order that was unique for
each developer.
11
-------
Table 3-1. Summary of experimental design by sample type
Concentration
range
Sample ID "
Outdoor site
Chamber site
Total #
samples
analyzed
PE materials
0
2.0 ppm
2.0 ppm
5.0 ppm
10.9 ppm
20.0 ppm
49.8 ppm
50.0 ppm
50.0 ppm
126
118
124
120
122
119
125
121
123
226
218
224
220
222
219
225
221
223
8
8
8
8
8
8
8
8
8
Environmental soils
0.1-2.0 ppm
2. 1-20.0 ppm
20. 1-50.0 ppm
50. 1-700.0 ppm
0
10 |ig/mL
100 |ig/mL
Grand total
101,107,108,109,113,114
102,103,104,115
111,116
105,106,110,112,117
Extracts
129*71 32 c
127/130
128/131
116
201,202,206
203,207,212,213
204,208,209,214,215
205,210,211,216,217
229/232
227/230
228/231
116
36
32
28
40
8
8
8
232 d
" Each sample ID was analyzed in quadruplicate.
6 Extract prepared in iso-octane for Dexsil and the reference laboratory.
0 Extract prepared in methanol for Electronic Sensor Technology, Strategic Diagnostics Inc., and the
reference laboratory.
d All samples were analyzed in random order.
12
-------
Environmental Conditions during Demonstration
As mentioned earlier, field activities were conducted both outdoors under natural environmental conditions and
indoors in a controlled environmental atmosphere chamber to evaluate the effect of environmental conditions
on technology performance. The weather outside was relatively uncomfortable during the July demonstration,
with highs approaching 100°F and 90% relative humidity (RH). Daily average temperatures were around 85 °F
with 70% RH. While outside, the developers set up canopies to provide shade and protection from frequent late
afternoon thundershowers.
In the indoor chamber tests, conditions were initially set to 55 °F and 25% RH. An independent check of the
conditions inside the chamber revealed that the temperature was closer to 68 °F with a 38% RH on the first day
of testing. A maintenance crew was called in to address the inconsistencies between the set and actual
conditions. By the middle of the third day of testing, the chamber was operating properly at 55 °F and 50% RH.
Appendix C contains a summary of the environmental conditions (temperature and relative humidity) during
the demonstration. The SDI team analyzed samples using the EnviroGard test kit outdoors July 26 and 27 and
in the chamber on July 23 and 24.
Sample Descriptions
PCBs (C^HK^QJ are a class of compounds that are chlorine-substituted linked benzene rings. There are 209
possible PCB compounds (also known as congeners). PCBs were commercially produced as complex mixtures
beginning in 1929 for use in transformers, capacitors, paints, pesticides, and inks [1]. Monsanto Corporation
marketed products that were mixtures of 20 to 60 PCB congeners under the trade name Aroclor. Aroclor
mixtures are identified by a number (e.g., Aroclor 1260) that represents the mixture's chlorine composition as
a percentage (e.g., 60%).
Performance Evaluation Materials
Samples of Tennessee reference soil [3] served as the blanks. Pre-prepared certified PE samples were obtained
from Environmental Resource Associates (ERA) of Arvada, Colorado, and the Analytical Operations and Data
Quality Center of EPA's Office of Solid Waste and Emergency Response. The soils purchased from ERA had
been prepared using ERA's semivolatile blank soil matrix. This matrix was a topsoil that had been dried,
sieved, and homogenized. Particle size was approximately 60 mesh. The soil was approximately 40% clay. The
samples acquired from EPA's Analytical Operations and Data Quality Center had been prepared using
contaminated soils from various sites around the country in the following manner: The original soils had been
homogenized and diluted with a synthetic soil matrix (SSM). The SSM had a known matrix of 6% gravel, 31%
sand, and 43% silt/clay; the remaining 20% was topsoil. The dilution of the original soils was performed by
mixing known amounts of contaminated soil with the SSM in a blender for no less than 12 h. The samples were
also spiked with target pesticides (a, P, A, and 6-BHC, methoxychlor, and endrin ketone) to introduce some
compounds that were likely to be present in an actual environmental soil. The hydrocarbon background from
the original sample and the spiked pesticides produced a challenging matrix. The PE soils required no additional
preparation by ORNL and were split for the developer and reference laboratory analyses as received.
Environmental Soil Samples
As noted in the site description, PCB-contaminated environmental soil samples from Oak Ridge, Portsmouth,
and Paducah were used in this demonstration. The soils were contaminated with PCBs as the result of spills
13
-------
and industrial processing activities at the various DOE facilities. Originally, the contaminated soils were
excavated from dikes, drainage ditches, catch basins, and organic waste storage areas. The excavated soils were
then packaged into waste containers and stored at the repositories in ETTP and Paducah in anticipation of
disposal by incineration. The environmental soil samples used in this study were collected from these waste
containers. Many of the soils contained interfering compounds such as oils, fuels, and other chlorinated
compounds, while some contained multiple Aroclors. For more information on sampling locations and sample
characteristics (soil composition, pH, and total organic carbon content), refer to Appendices A and B,
respectively.
Extract Samples
Traditionally, the amount of PCBs on a contaminated surface is determined by wiping the surface with a cotton
pad saturated with hexane. The pad is then taken to the laboratory, extracted with additional hexane, and
analyzed by gas chromatography. Unlike soil samples, which can be more readily
homogenized and divided, equivalent wipe samples (i.e., contaminated surfaces or post-wipe pads) were not
easily obtainable. Therefore, interference-free solutions of PCBs were analyzed to simulate an extracted surface
wipe pad. Extract sample analyses provided evaluation data that relied primarily on the technology's
performance rather than on elements critical to the entire method (i.e., sample collection and preparation).
Because different developers required the extract samples prepared in different solvents (e.g., methanol and iso-
octane), the reference laboratory analyzed sets of extracts in both solvents. SDI analyzed extracts prepared in
methanol. A total of 12 extracts were analyzed per site; these consisted of four replicates each of a blank and
two concentration levels (10 and 100 (jg/mL).
Sampling Plan
Sample Collection
Environmental soil samples were collected from April 17 through May 7, 1997. Portsmouth and Oak Ridge
Reservation soils were collected from either storage boxes or 55-gal drums stored at ETTP. Briefly, the
following procedure was used to collect the soil samples. Approximately 30 Ib of soil were collected from the
top of the drum or B-25 box using a scoop and placed in a plastic bag. The soil was sifted to remove rocks and
other large debris, then poured into a plastic-lined 5-gal container. All samples were subjected to radiological
screening and were determined to be nonradioactive. Similarly, soil samples were collected from 55-gal drums
stored at Paducah and shipped to ORNL in lined 5-gal containers.
Sample Preparation, Labeling, and Distribution
Aliquots of several of the environmental soils were analyzed and determined to be heterogeneous in PCB
concentration. Because this is unsatisfactory for accurately comparing the performance of the field technology
with the laboratory-based method, the environmental soils had to be homogenized prior to sample distribution.
Each Portsmouth and Oak Ridge environmental soil sample was homogenized by first placing approximately
1500 g of soil in a glass Pyrex dish. The dish was then placed in a large oven set at 35 °C, with the exhaust and
blower fans turned on to circulate the air. After drying overnight, the soil was pulverized using a conventional
blender and sieved using a 9-mesh screen (2 mm particle size). Last, the soil was thoroughly mixed using a
spatula. A comparison of dried and undried soils showed that a minimal amount of PCBs (< 20%) was lost as
the result of sample drying, making this procedure suitable for use in the preparation of the soil samples. The
Paducah samples, because of their sandy characteristics, required only the sieving and mixing preparation
steps. Extract sample preparation involved making solutions of PCBs in methanol and iso-octane at two
14
-------
concentration levels (10 and 100 pg/mL). Multiple aliquots of each sample were analyzed using the analytical
procedure described below to confirm the homogeneity of the samples with respect to PCB concentration.
To provide the developers with soils contaminated at higher concentrations of PCBs, some of the environmental
soils (those labeled with an "S" in Appendix B) were spiked with additional PCBs. Spiked soils samples were
prepared after the soil was first dried in a 35 °C oven overnight. The dry soil was ground using a conventional
blender and sieved through a 9-mesh screen (2 mm particle size). Approximately 1500 g of the sieved soil were
spiked with a diethyl ether solution of PCBs at the desired concentration. The fortified soil was agitated using
a mechanical shaker and then allowed to air-dry in a laboratory hood overnight. A minimum of four aliquots
were analyzed using the analytical procedure described below to confirm the homogeneity of the soil with
regard to the PCB concentration.
The environmental soils were characterized at ORNL prior to the demonstration study. The procedure used to
confirm the homogeneity of the soil samples entailed the extraction of 3 to 5 g of soil in a mixture of solvents
(1 mL water, 4 mL methanol, and 5 mL hexane). After the soil/solvent mixture was agitated by a mechanical
shaker, the hexane layer was removed and an aliquot was diluted for analysis. The hexane extract was analyzed
on a Hewlett Packard 6890 gas chromatograph equipped with an electron capture detector and autosampler.
The method used was a slightly modified version of EPA's SW-846 dual-column Method 8081 [4].
After analysis confirming homogeneity, the samples were split into jars for distribution. Each 4-oz sample jar
contained approximately 20 g of soil. Four replicate splits of each soil sample were prepared for each
developer. The samples were randomized in two fashions. First, the order in which the filled jars were
distributed was randomized so that the same developer did not always receive the first jar filled for a given
sample set. Second, the order of analysis was randomized so that each developer analyzed the same set of
samples, but in a different order. The extract samples were split into 10-mL aliquots and placed into 2-oz jars.
The extracts were stored in the refrigerator (at <4°C) until released to the developers. Each sample jar had
three labels: (1) developer order number, (2) sample identifier number, and (3) a PCB warning label. The
developer order number corresponded to the order in which the developer was required to analyze the samples
(e.g., SDI 1001 through SDI 1116). The sample identifier number was in the format of "xxxyzz," where "xxx"
was the three-digit sample ID (e.g., 101) listed in Table 3-1, "y" was the replicate (e.g., 1 to 4), and "zz" was
the aliquot order of each replicate (e.g., 01 to 11). For example, sample identifier 101101 corresponded to
sample ID "101" (an Oak Ridge soil from RFD 40022, drum 02), "1" corresponded to the first replicate from
that sample, and "01" corresponded to the first jar filled in that series.
Once the samples were prepared, they were stored at a central sample distribution center. During the
demonstration study, developers were sent to the distribution center to pick up their samples. Samples were
distributed sequentially in batches of 12 to ensure that samples were analyzed in the order specified.
Completion of chain-of-custody forms and scanning of bar code labels documented sample transfer activities.
Some of the developers received information regarding the samples prior to analysis. SDI received information
pertaining to which Aroclors were in the samples and, if multiple Aroclors were present, in what ratio. This
was provided at the request of SDI to simulate the type of information that would be available during actual
field testing. The developers returned the unused portions of the samples with the analytical results to the
distribution center when testing was completed. The sample bar codes were scanned upon return to document
sample throughput time.
15
-------
Three complete sets of extra samples, called archive samples, were available for distribution in case the
integrity of a sample was compromised. Very few (<5) archive samples were utilized over the course of the
demonstration.
Predemonstration Study
Ideally, environmental soil samples are sent to the developers prior to the demonstration study to allow them
the opportunity to analyze representative samples in advance of the verification test. This gives developers the
opportunity to refine and calibrate their technologies and revise their operating procedures on the basis of the
predemonstration study results. The predemonstration study results can also be used as an indication that the
selected technologies are of the appropriate level of maturity to participate in the demonstration study.
According to ORNL regulations, however, one of two conditions must exist in order to ship environmental soils
that were once classified as mixed hazardous waste. First, the recipient—in this case, the developer's
facilities—must have proper Nuclear Regulatory Commission (NRC) licensing to receive and analyze
radiological materials. Second, the soils must be certified as entirely free of radioactivity, beyond the no-rad
certification issued from radiological screening tests based on ORNL standards. Because none of the developers
had proper NRC licensing and proving that the soils were entirely free of radioactivity was prohibitive, spiked
samples of Tennessee reference soil were used for the predemonstration study. The developers had an
opportunity to evaluate the Tennessee reference soils spiked with PCBs at concentrations similar to what would
be used in the demonstration study. The developers also analyzed two performance evaluation samples and one
solvent extract. The reference laboratory analyzed the same set of samples, which included two extracts
samples, prepared in the two solvents (methanol and iso-octane) requested by the developers.
Predemonstration Sample Preparation
Two soil samples were prepared by ORNL using Tennessee reference soil [3]. The soil was a Captina silt loam
from Roane County, Tennessee, that was slightly acidic (pH ~5) and low in organic carbons (-1.5%). The soil
composition was 7.7% sand, 29.8% clay, and 62.5% silt. To prepare a spiked sample, the soil was first ground
using either a mortar and pestle or a conventional blender. The soil was then sieved through a 16-mesh screen
(1 mm particle size). Approximately 500 g of the sieved soil was spiked with a diethyl ether solution of PCBs
at the desired concentration. The soil was agitated using a mechanical shaker, then allowed to air-dry overnight
in a laboratory hood. A minimum of five aliquots were analyzed by gas chromatography using electron capture
detection. The PCB concentration of the spiked samples was determined to be homogeneous. The remaining
two soil samples used in the predemonstration study were PE materials acquired from ERA and EPA (see the
section "Performance Evaluation Materials" above). In addition, a solvent extract was prepared by ORNL to
simulate an extracted surface wipe sample. The extracts were prepared in two different solvents (iso-octane
and methanol) to accommodate developer requests.
Predemonstration Results
The predemonstration samples were sent to the developers and the reference laboratory on June 2, 1997.
Predemonstration results were received by June 26, 1997. Table 3-2 summarizes the test kit's results for the
predemonstration samples. Results indicated that SDI's EnviroGard PCB test kit was ready for field evaluation.
16
-------
Table 3-2. Summary of EnviroGard predemonstration results
Sample description
2 ppm of Aroclor 1260
100ppm(total)of
Aroclors 1254 and 1260
1 1 ppm of Aroclor 1260
50 ppm of Aroclor 1254
5 ppm of Aroclor 1242
Matrix
Soil
Soil
Soil
Soil
Extract
Source
ORNL
ORNL
EPA
ERA
ORNL
EnviroGard"
Result
(ppm)
[2, 8.2)6
(85,00)
[8.2, 14.4)
Duplicate
result (ppm)
[2, 8.2)
85, oo )
[8.2, 14.4)
(52, »r
(1.0,5.0)
(1.0,5.0)
Reference laboratory
Result
(ppm)
2.2
78.0
11.0
Duplicate result
(ppm)
2.3
89.0
9.5
37.0"
4.7
4.9
" Results were Aroclor-adjusted (see Section 2 for more details).
6 The notation [2, 8.2) indicates that the sample concentration was greater than or equal to 2 and less than 8.2. See
Sections 2 and 5 for more information on interval reporting.
c Replicate was not analyzed because of lack of adequate sample for second analyses.
Deviations from the Demonstration Plan
A few deviations from the demonstration plan occurred. In Appendix B of the technology demonstration plan
[5], the reference laboratory's procedure states that no more than 10 samples will be analyzed with each
analytical batch (excluding blanks, standards, QC samples, and dilutions). The analytical batch is also stated
as 10 samples in the Quality Assurance Project Plan of the demonstration plan. The reference laboratory
actually analyzed 20 samples per analytical batch. Because a 20-sample batch is recommended in SW-846
Method 8081, this deviation was deemed acceptable.
Table 5 of the demonstration plan [5] delineates the environmental soils according to concentration. The
classification was based on a preliminary analysis of the soils at ORNL. Table 3-1 of this report arranges the
concentrations as characterized by the reference laboratory. The reference laboratory determined that five
sample sets (sample IDs 102, 105, 110, 111, and 210) were in the next highest concentration range, differing
from what was originally outlined in the demonstration plan. Also, the highest concentration determined by the
reference laboratory was 700 ppm, while the preliminary analysis at ORNL found the highest concentration
to be 500 ppm.
During the demonstration study, the SDI team did not note any deviations from the procedure described in the
technology demonstration plan [5] for the EnviroGard test kit.
17
-------
-------
Section 4
Reference Laboratory Analytical Results and Evaluation
Objective and Approach
The purpose of this section is to present the evaluation of the PCB data generated by the reference laboratory.
Evaluation of the results from the analysis of PE, environmental soil, and extract samples was based on
precision, accuracy, representativeness, completeness, and comparability (PARCC) parameters [6]. This
section describes how the analytical data generated by the reference laboratory were used to establish a baseline
performance for PCB analysis.
Reference Laboratory Selection
The Oak Ridge Sample Management Office (SMO) has been tasked by DOE Oak Ridge Operations (DOE-
ORO) with maintaining a list of qualified laboratories to provide analytical services. The technology
demonstration plan [5] contains the SMO's standard operating procedures (SOPs) for identifying, qualifying,
and selecting analytical laboratories. Laboratories are qualified as acceptable analytical service providers for
the SMO by meeting specific requirements. These requirements include providing pertinent documentation
(such as QA and chemical hygiene plans), acceptance of the documents by the SMO, and satisfactory
performance on an on-site prequalification audit of laboratory operations. All laboratory qualifications are
approved by a laboratory selection board, composed of the SMO operations manager and appointees from all
prime contractors that conduct business with the SMO.
All of the qualified laboratories were invited to bid on the demonstration study sample analysis. The lowest-cost
bidder was LAS Laboratories, in Las Vegas, Nevada. A readiness review conducted by ORNL and the SMO
confirmed the selection of LAS as the reference laboratory. Acceptance of the reference laboratory was
finalized by satisfactory performance in the predemonstration study (see Table 3-2). The SMO contracted LAS
to provide full data packages for the demonstration study sample analyses within 30 days of sample shipment.
The SMO conducts on-site audits of LAS annually as part of the laboratory qualification program. At the time
of selection, the most recent audit of LAS had occurred in February 1997. Results from this audit indicated
that LAS was proficient in several areas, including program management, quality management, and training
programs. No findings regarding PCB analytical procedure implementation were noted. A second on-site audit
of LAS occurred August 11-12, 1997, during the analysis of the demonstration study samples. This
surveillance focused specifically on the procedures that were currently in use for the analysis of the
demonstration samples. The audit, jointly conducted by the SMO, DOE-ORO, and EPA-Las Vegas (LV),
verified that LAS was procedurally compliant. The audit team noted that LAS had excellent adherence to the
analytical protocols and that the staff were knowledgeable of the requirements of the method. No findings
impacting data quality were noted in the audit report.
19
-------
Reference Laboratory Method
The reference laboratory's analytical method, also presented in the technology demonstration plan [5], followed
the guidelines established in EPA SW-846 Method 8081 [4]. According to LAS's SOP, PCBs were extracted
from 30-g samples of soil by sonication in hexane. Each extract was then concentrated to a final volume that
was further subjected to a sulfuric acid cleanup to remove potential interferences. The analytes were identified
and quantified using a gas chromatograph equipped with dual electron-capture detectors. Each extract was
analyzed on two different chromatographic columns with slightly different separation characteristics (primary
column: RTX-1701, 30 m x 0.53 mm ID x 0.5 (jm; confirmatory column: RTX-5, 30 m x 0.53 mm ID x 0.5
(jm). PCBs were identified when peak patterns from a sample extract matched the patterns of standards for
both columns. PCBs were quantified based on the initial calibration of the primary column.
Calibration
Method 8081 states that, because Aroclors 1016 and 1260 include many of the peaks represented in the other
five Aroclor mixtures, it is only necessary to analyze two multilevel standards for these Aroclors to demonstrate
the linearity of the detector response for PCBs. However, per LAS SOPs, five-point (0.1 to 4 ppm) initial
calibration curves were generated for Aroclors 1016, 1248, 1254, and 1260 and the surrogate compounds
[decachlorobiphenyl (DCB) and tetrachloro-m-xylene (TCMX)]. Single mid-level standards were analyzed for
the other Aroclors (1221, 1232, and 1242) to aid in pattern recognition. All of the multi-point calibration data,
fitted to quadratic models, met the QC requirement of having a coefficient of determination (R2) of 0.99 or
better over the calibration range specified. The detection limits for soil samples were 0.033 ppm (ng/g) for all
Aroclors except Aroclor 1221, which was 0.067 ppm. For extract samples, the detection limits were 0.010 ppm
((ig/mL) for all Aroclors except Aroclor 1221, which was 0.020 ppm. Reporting detection limits were
calculated based on the above detection limits, the actual sample weight, and the dilution factor.
Sample Quantification
For sample quantification, Aroclors were identified by comparing the samples' peak patterns and retention
times with those of the respective standards. Peak height ratios, peak shapes, sample weathering, and general
similarity in detector response were also considered in the identification. Aroclor quantifications were
performed by selecting three to five representative peaks, confirming that the peaks were within the established
retention time windows, integrating the selected peaks, quantifying the peaks based on the calibrations, and
averaging the results to obtain a single concentration value for the multicomponent Aroclor. If mixtures of
Aroclors were suspected to be present, the sample was typically quantified in terms of the most representative
Aroclor pattern. If the identification of multiple Aroclors was definitive, total PCBs in the sample were
calculated by summing the concentrations of both Aroclors. Aroclor concentrations were quantified within the
concentration range of the calibration curve. If PCBs were detected and the concentrations were outside of the
calibration range, the sample was diluted and reanalyzed until the concentration was within the calibration
range. If no PCBs were detected, the result was reported as a non-detect (i.e., "< reporting detection limit").
Sample Receipt, Handling, and Holding Times
The reference laboratory was scheduled to analyze a total of 256 PCB samples (208 soil samples, 24 iso-octane
extract samples, and 24 methanol extract samples). Of these same samples, the developer was scheduled to
analyze a total of 232 PCB samples (208 soil samples and 24 extract samples in solvent of choice). The
samples were shipped to LAS at the start of the technology demonstration activities (July 22). Shipment was
coordinated through the SMO. Completion of chain-of-custody forms documented sample transfer. The
20
-------
amples were shipped on ice in coolers to maintain <6° mperatures during shipment. Samples were shipped
with custody seals to ensure sample integrity and to prevent tampering during transport.
on receipt of the samples, the reference laboratory checked the receipt temperature and conditions of th
sample containers, assigned each sample a unique number, an
All samples were received at the proper temperature and in good condition. Demonstration samples wer
divided into 11 analytical batches (with no
order specified by ORNL to ensure that the analysis of sample types was ran
supplied by the reference laboratory to indicate method performance, were performed with each analytica
batch of soils.
Prior lysis, samples were stored in refrigerators kept at 4 to 6° e
r ct samples and to extract the soil samples within 14 days
o soils were extracted, the reference laboratory had an additional 40 days to
a r any of the demonstration samples. The
f es was received at ORNL in 72 days, on October 1, 1997.
The contractual obligation was 30 days.
he remainder of this section is devoted to summarizing the data generated by the reference laboratory and to
Quality Control Results
Objective
he purpose of this section is to provide an assessment of the data generated by the reference laboratory's QC
rocedures. The QC samples included continuing calibration verification standards (CCVs), instrument blanks,
ethod blanks, surrogate spikes, laboratory control samples (LCSs), and matrix spike (MS)/duplicate matrix
pike (MSB) samples. Each control type is described in more detail in the following text and in the technology
on plan [5]. Because extraction of these liquid samples was not required, calibration chec
standards and instrument blanks were the only control samples implemented for the extract samples. Th
reference laboratory's implementation of QC procedures was consistent with SW-846 guidance.
Continuing Calibration Verification Standard Results
CCV is a single calibration standard of known concentration, usually at the midpoint of the calibration range.
s standard is evaluated as an unknown and is quantified against the initial calibration. The calculate
concentration is then compared with th
calibration is still valid. CCVs
acceptance was
QC requirement was
for
course of sample analysis.
21
-------
Instrument and Method Blank Results
Instrument blanks (hexane) were analyzed prior to each CCV. The QC requirement was that instrument blanks
must contain less than the reporting detection limit for any analyte. All instrument blanks were acceptable.
A method blank is an analyte-free soil matrix sample that is taken through the extraction process to verify that
there are no laboratory sources of contamination. One method blank was analyzed for each analytical batch.
The QC requirement was that method blanks must contain less than the reporting detection limit for any
Aroclor. No PCBs were detected in any of the eleven method blanks that were analyzed. These results
demonstrated that the reference laboratory was capable of maintaining sample integrity, and that it did not
introduce PCB contamination to the samples during preparation.
Surrogate Spike Results
A surrogate is a compound that is chemically similar to the analyte group but is not expected to be present in
the environmental sample. A surrogate is added to test the extraction and analysis methods to verify the ability
to isolate, identify, and quantify a compound similar to the analyte(s) of interest without interfering with the
determination. Two different surrogate compounds, DCB and TCMX, were used to bracket the retention time
window anticipated in the Aroclor chromatograms. All soil samples, including QC samples, were spiked with
surrogates at 0.030 ppm prior to extraction. Surrogate recoveries were deemed to be within QC requirements
if the measured concentration fell within the QC acceptance limits that were established by past method
performance. (For LAS this was 39 to 117% for DCB, and 66 to 128% for TCMX). The results were
calculated using the following equation:
measured amount lnno/
percent recovery = x 100%
actual amount (4-1)
In all undiluted samples, both of the surrogates had percentage recoveries that were inside the acceptance limits.
Surrogate recoveries in diluted samples were uninformative because the spike concentration (0.030 ppm, as
specified by the method) was diluted below the instrument detection limits. The surrogate recovery results for
undiluted samples indicated that there were no unusual matrix interferences or batch-processing errors for these
samples.
Laboratory Control Sample Results
An LCS is an aliquot of a clean soil that is spiked with known quantities of target analytes. The LCS is spiked
with the same analytes and at the same concentrations as the MS. (MSs are described in the next section.) If
the results of the MS analyses are questionable (i.e., indicating a potential matrix effect), the LCS results are
used to verify that the laboratory can perform the analysis in a clean, representative matrix.
Aroclors 1016 and 1260 were spiked into the clean soil matrix at approximately 0.300 ppm, according to the
reference laboratory's SOP. The QC requirements (defined as percent recovery) for the LCS analyses were
performance-based acceptance limits that ranged from 50 to 158%. In all but 1 of the 11 LCSs analyzed, both
Aroclor percent recoveries fell within the acceptance limits. Satisfactory recoveries for LCS verified that the
reference laboratory performed the analyses properly in a clean matrix.
22
-------
Matrix Spike Results
In contrast to an LCS, a MS sample is an actual environmental soil sample into which target analytes are
spiked at known concentrations. MS samples are used to assess the efficiency of the extraction and analytical
methods for real samples. This is accomplished by determining the amount of spiked analyte that is
quantitatively recovered from the environmental soil. An MSD sample is spiked and analyzed to provide a
measure of method precision. Ideally, to evaluate the MS/MSD results, the environmental soil is analyzed
unspiked so that the background concentrations of the analyte in the sample are considered in the recovery
calculation.
For the demonstration study samples, one MS and MSD pair was analyzed with each analytical batch. The MS
samples were spiked under the same conditions and QC requirements as the LCS (50 to 158% acceptance
limits), so that MS/MSD and LCS results could be readily compared. The QC requirement for MS and MSD
samples was a relative percent difference (RPD) of less than 30% between the MS/MSD pair. RPD is defined
as
MS recovery - MSD recovery inn0/
average recovery (4-2)
A total of 11 MS/MSD pairs were analyzed. Because the MS/MSD spiking technique was not always properly
applied (e.g., a sample which contained 100 ppm of Aroclor 1254 was spiked ineffectively with 0.300 ppm of
Aroclor 1260), many of the MS/MSD results were uninformative. For the samples that were spiked
appropriately, all MS/MSD QC criteria were met.
Conclusions of the Quality Control Results
The reference laboratory results met performance acceptance requirements for all of the samples where proper
QC procedures were implemented. Acceptable performance on QC samples indicated that the reference
laboratory was capable of performing analyses properly.
Data Review and Validation
Objective
The purpose of validating the reference laboratory data was to ensure usability for the purposes of comparison
with the demonstration technologies. The data generated by the reference laboratory were used as a baseline
to assess the performance of the technologies for PCB analysis. The reference laboratory data were
independently validated by ORNL and SMO personnel, who conducted a thorough quality check and reviewed
all sample data for technical completeness and correctness.
Corrected Results
Approximately 8% of the results provided by the reference laboratory (20 of 256) were found to have
correctable errors. So as not to bias the assessment of the technology's performance, errors in the reference
laboratory data were corrected. These changes were made conservatively, based on the guidelines provided in
the SW-846 Method 8081 for interpreting and calculating Aroclor results. The errors (see Appendix D, Table
D-3) were categorized as transcription errors, calculation errors, and interpretation errors. The corrections
23
-------
listed in Table D-3 were made in the final data set that was used for comparison with the demonstration
technologies.
Suspect Results
Normally, one would not know if a single sample result was "suspect" unless (1) the sample was a PE sample,
where the concentration is known or (2) a result was reported and flagged as suspect for some obvious reason
(e.g., no quantitative result was determined). The experimental design implemented in this demonstration study
provided an additional indication of the abnormality of data through the inspection of the replicate results from
a homogenous soil sample set (i.e., four replicates were analyzed for each sample ID).
Data sets were considered suspect if the standard deviation (SD) of the four replicates was greater than 30 ppm
and the percent relative standard deviation (RSD) was greater than 50%. Five data sets (sample IDs 106, 205,
216,217, 225) contained measurements that were considered suspect using this criteria, and the suspect data
are summarized in Table 4-1. A number of procedural errors may have caused the suspect measurements (e.g.,
spiking heterogeneity, extraction efficiencies, dilution, etc.). In the following subsections for precision and
accuracy, the data were evaluated with and without these suspect values to represent the best and worst case
scenarios.
Table 4-1. Suspect measurements within the reference laboratory data
Criteria
SD > 30 ppm
and
RSD > 50%
Qualitative result
Sample ID
106
205
216
217
225
110
112
PCB concentration (ppm)
Replicate results
(ppm)
255.9,269.9,317.6
457.0,483.3,538.7
47.0, 54.3, 64.0
542.8, 549.8, 886.7
32.1,36.5,56.4
< reporting detection
limits
Suspect result(s)
(ppm)
649.6
3,305.0
151.6
1,913.3
146.0
< 66, < 98, < 99, < 490
< 66, < 130, < 200,< 200
Data usability
Performed data analysis with
and without this value
Used as special case for
comparison with developer
results
Samples that did not fall into the above criteria, but were also considered suspect, were non-blank samples that
could not be quantified and were reported as "< the reporting detection limit." This was the case for
environmental soil sample IDs 110 and 112. It is believed that the reference laboratory had trouble quantifying
these soil samples because of the abundance of chemical interferences. These samples were diluted by orders
of magnitude to reduce interferences, thereby diluting the PCB concentrations to levels that were lower than
the instrument detection limits. With each dilution, the reporting detection limits values were adjusted for
sample weight and dilution, which accounts for the higher reporting detection limits (up to 490 ppm). It is
24
-------
believed that these samples should have been subjected to additional pre-analytical cleanup to remove these
interferences before quantification was attempted. Sample IDs 110 and 112 were collected from the same
cleanup site (see Appendix B), so it is not surprising that similar difficulties were encountered with both sample
sets. Because the results for sample IDs 110 and 112 were not quantitative, these data were compared with the
technology data only on a special case basis.
Data Assessment
Objective
The purpose of this section is to provide an evaluation of the performance of the reference laboratory results
through statistical analysis of the data. The reference laboratory analyzed 72 PE, 136 environmental soil, and
48 extract samples. All reference laboratory analyses were performed under the same environmental
conditions. Therefore, site differentiation was not a factor in data assessment for the reference laboratory. For
comparison with the technology data, however, the reference laboratory data are delineated into "outdoor site"
and "chamber site" in the following subsections. For consistency with the technology review, results from both
sites were also combined to determine the reference laboratory's overall performance for precision and
accuracy. This performance assessment was based on the raw data compiled in Appendix D. All statistical tests
were performed at a 5% significance level.
Precision
The term "precision" describes the reproducibility of measurements under a given set of conditions. The SD
of four replicate PCB measurements was used to quantify the precision for each sample ID. SD is an absolute
measurement of precision, regardless of the PCB concentration. To express the reproducibility relative to the
average PCB concentration, RSD is used to quantify precision, according to the following equation:
RSD = Standard Deviation x WQ%
Average Concentration
Performance Evaluation Samples
The PE samples were homogenous soils containing certified concentrations of PCBs. Results for these samples
represent the best estimate of precision for soil samples analyzed in the demonstration study. Table 4-2
summarizes the precision of the reference laboratory for the analysis of PE samples. One suspect measurement
(sample ID 225, 146.0 ppm) was reported for the PE soil samples. The RSDs for the combined data ranged
from 9 to 33% when the suspect measurement was excluded, and from 9 to 79% including the suspect
measurement. The overall precision, determined by the mean RSD for all PE
25
-------
Table 4-2. Precision of the reference laboratory for PE soil samples
Outdoor Site
Sample
ID
126"
118
124
120
122
119
125
121
123
Average
concentration
(ppm)
0
1.6
1.7
5.0
11.1
20.1
37.9
54.6
60.1
SD
(ppm)
n/a
0.6
0.2
1.0
0.9
3.4
6.9
3.4
4.6
RSD
(%)
n/a
39
13
20
8
17
18
6
8
Chamber Site
Sample
ID
226
218
224
220
222
219
225
221
223
Average
concentration
(ppm)
0
2.6
1.7
5.8
12.8
23.3
41.7"
44.9
55.8
SD
(ppm)
n/a
0.2
0.5
1.8
0.3
6.1
12.9"
11.3
7.7
RSD
(%)
n/a
6
29
31
3
26
31"
25
14
Combined Sites
Average
concentration
(ppm)
0
2.1
1.7
5.4
11.9
21.7
39.5C
49.8
58.0
SD
(ppm)
n/a
0.7
0.4
1.4
1.1
4.9
9.2 c
9.3
6.3
RSD
(%)
n/a
33
21
26
9
23
23 c
19
11
" All PCB concentrations were reported as non-detects.
6 Results excluding the suspect value (results including the suspect value: mean = 67.8 ppm,
c Results excluding the suspect value (results including the suspect value: mean =52.8 ppm,
SD = 53.2 ppm, and RSD =79%).
SD = 38.6 ppm, and RSD = 73%).
samples, was 21% for the worst case (including the suspect result) and 18% for the best case (excluding the
suspect result).
Environmental Soil Samples
The precision of the reference laboratory for the analysis of environmental soil samples is reported in Table
4-3. In this table, results including suspect measurements are presented in parentheses. Average concentrations
were reported by the reference laboratory as ranging from 0.5 to 1,196 ppm with RSDs that ranged from 7 to
118% when the suspect results were included. Excluding the suspect results, the highest average concentration
decreased to 660 ppm, and the largest RSD decreased to 71%. Because the majority of the samples fell below
125 ppm, precision was also assessed by partitioning the results into two ranges: low concentrations (< 125
ppm) and high concentrations (> 125 ppm). For the low concentrations, the average RSD was 23% excluding
the suspect value and 26% including the suspect value. These average RSDs were only slightly larger than the
RSDs for the PE soil samples of comparable concentration (18% for best case and 21% for worst case). Five
soil sample sets (sample IDs: 106, 117, 205, 211 and 217) were in the high-concentration category. The
average precision for high concentrations was 56% for the worst case and 19% for the best case. The precision
estimates for the low and high concentration ranges were comparable when the suspect values were excluded.
This indicated that the reference laboratory's precision for the environmental soils was consistent
(approximately 21% RSD) and was comparable to the PE soil samples when the suspect values were excluded.
The Paducah soils (indicated as bold sample IDs in Table 4-3) were analyzed by the technologies under both
outdoor and chamber conditions to provide a measure of the effect that two different environmental
26
-------
Table 4-3. Precision of the reference laboratory for environmental soil samples
Outdoor site
Sample
ID
101
102
103
104
105
106
107
108
109
110
111
112
113 c
114
115
116
117
Average
concentration
(ppm)
0.5
2.0
2.3
9.4
59.4
281.0(373.2)"
1.3
1.8
2.0
n/a6
38.7
n/a
1.1
1.3
14.8
41.3
383.9
Standard
deviation
(ppm)
0.1
0.3
0.6
4.0
16.5
32.4(186.2)
0.3
0.1
0.4
n/a
4.3
n/a
0.6
0.3
1.8
5.9
55.2
RSD
(%)
16
16
27
43
28
12 (50)
20
8
20
n/a
11
n/a
55
20
12
14
14
Chamber site
Sample
ID
206
207
208
209
210
211
212
213
214
215
216
217
201
202
203
204
205
Average
concentration
(ppm)
1.9
18.8
30.5
40.2
88.6
404.5
3.2
8.1
25.2
26.7
55.1 (79.2)
659.8(973.2)
0.9
1.4
13.9
44.3
493.0(1196.0)
Standard
deviation
(ppm)
0.9
3.5
7.9
28.5
25.6
121.8
1.6
1.6
3.7
3.2
8.5 (48.7)
196.6(647.0)
0.2
0.2
1.7
2.9
41.7(1406.4)
RSD
(%)
49
19
26
71
29
30
50
20
15
12
15 (62)
30 (66)
24
12
12
7
8(118)
" Data in parentheses include suspect values.
6 n/a indicates that qualitative results only were reported for this sample.
c Bold sample IDs were matching Paducah sample pairs (i.e., 113/201, 114/202, 115/203, 116/204, 117/205).
conditions had on the technology's performance. Although this was not an issue for the reference laboratory
(because all the samples were analyzed under laboratory conditions), the reference laboratory's results were
delineated into the different site categories for comparison with the technologies. Sample IDs 113 and 201, 114
and 202, 115 and 203, 116 and 204, and 117 and 205 each represent a set of eight replicate samples of the
same Paducah soil. The RSDs for four of the five Paducah pairs (excluding the suspect value for sample ID
205) ranged from 11 to 17%. The result from one pair (sample IDs 113 and 201) had an RSD of 42%, but the
reported average concentration was near the reporting limits.
27
-------
Extract Samples
The extract samples, which were used to simulate surface wipe samples, were the simplest of all the
demonstration samples to analyze because they required no extraction and were interference-free. Three types
of extract samples were analyzed: solvent blanks, spikes of Aroclor 1242 at 10 (jg/mL, and spikes of Aroclor
1254 at 100 (jg/mL. Identical extract samples were prepared in two solvents (iso-octane and methanol) to
accommodate the developer's request. The reference laboratory analyzed both solvent sets. A student's t-test
[7, 8] was used to compare the reference laboratory's average PCB concentrations for the two different solvents
and showed that no significant differences were observed at either concentration. Therefore, the reference
laboratory results for the two extract solvents were combined. Additionally, all blank samples were quantified
as non-detects by the reference laboratory.
Table 4-4 summarizes the reference laboratory results for the extract samples by site. RSDs for the four
replicates for each sample ID ranged from 3 to 24%. For the combined data set (16 replicate measurements),
the average RSD at the lO-pg/mL level was 19%, while the average RSD at the 100-(jg/mL level was 8%. For
the entire extract data set, an estimate of overall precision was 14%. The overall precision for the extract
samples was comparable to the best-case precision for environmental soil samples (21%) and PE soil samples
(18%).
Table 4-4. Precision of the reference laboratory for extract samples
Outdoor site
Sample
ID
129"
132"
127
130
128
131
Average
cone
(ug/mL)
0
0
10.9
12.1
67.4
63.8
SD
(ug/mL)
n/a
n/a
0.4
2.9
2.3
5.0
RSD
(%)
n/a
n/a
4
24
3
8
Chamber site
Sample
ID
229
232
227
230
228
231
Average
cone
(ug/mL)
0
0
9.6
8.9
65.2
57.7
SD
(ug/mL)
n/a
n/a
0.8
1.4
5.1
3.1
RSD
(%)
n/a
n/a
8
16
8
5
Combined sites
Average
cone
(ug/mL)
0
10.4
63.5
SD
(ug/mL)
n/a
1.9
5.2
RSD
(%)
n/a
19
8
" All PCB concentrations reported as non-detects by the laboratory.
Accuracy
Accuracy represents the closeness of the reference laboratory's measured PCB concentrations to the accepted
values. Accuracy was examined by comparing the measured PCB concentrations (for PE soil and extract
samples) with the certified PE values and known spiked extract concentrations. Percent recovery was used to
quantify the accuracy of the results. The optimum percent recovery value is 100%. Percent recovery values
greater than 100% indicate results that are biased high, and values less than 100% indicate results that are
biased low.
28
-------
Performance Evaluation Soil Samples
The reference laboratory's performance for the PE samples is summarized in Table 4-5. Included in this table
are the performance acceptance ranges and the certified PCB concentration values. The acceptance ranges,
based on the analytical verification data, are guidelines established by the provider of the PE materials to gauge
acceptable analytical results. As shown in Table 4-5, all of the average concentrations were within the
acceptance ranges, with the exception of sample ID 218. The average result of sample ID 225 was outside of
the acceptance range only when the suspect result was included. All of the replicate measurements in sample
ID 225 were biased slightly high. Average percent recoveries for the PE samples (excluding suspect values)
ranged from 76 to 130%. Overall accuracy was estimated as the average recovery for all PE samples. The
overall percent recovery was 105% as a worst case when the suspect value was included. Excluding the suspect
value as a best case slightly lowered the overall percent recovery to 101%. A regression analysis [9] indicated
that the reference laboratory's results overall were unbiased estimates of the PE sample concentrations.
Table 4-5. Accuracy of the reference laboratory for PE soil samples
Certified
concentration
(ppm)
(acceptance
range, ppm)
0"
(n/a)
2.0
(0.7-2.2)
2.0
(0.9-2.5)
5.0
(2.1-6.2)
10.9
(4.0-12.8)
20.0
(11.4-32.4)
49.8
(23.0-60.8)
50.0
(19.7-63.0)
50.0
(11.9-75.9)
Outdoor site
Sample
ID
126
118
124
120
122
119
125
121
123
Average
cone
(ppm)
0
1.6
1.7
5.0
11.1
20.1
37.9
54.6
60.1
Recovery
(%)
n/a
79
85
99
102
100
76
109
120
Chamber site
Sample
ID
226
218
224
220
222
219
225
221
223
Average
cone
(ppm)
0
2.6
1.7
5.8
12.8
23.3
41.7"
44.9
55.8
Recovery
(%)
n/a
130
85
117
117
116
84"
90
112
Combined sites
Average
cone
(ppm)
0
2.1
1.7
5.4
11.9
21.7
39. 5 c
49.8
58.0
Recovery
(%)
n/a
105
85
108
109
109
79 c
100
116
"All PCB concentrations reported as non-detects by the laboratory.
6 Results excluding the suspect value (results including the suspect value: average = 67.8 ppm and recovery = 136%).
'Results excluding the suspect value (results including the suspect value: average = 52.8 ppm and Recovery = 106%).
Extract Samples
29
-------
Percent recovery results for extract samples are summarized in Table 4-6 for the reference laboratory. The
average percent recoveries for extract samples ranged from 58 to 121%. In terms of concentration levels, the
average recovery at the lO-pg/mL level (for both solvents) was 104%, compared with 64% at the 100-(jg/mL
level. The reference laboratory classified all 16 samples spiked at 10 (jg/mL as Aroclor 1016; however, these
samples were actually spiked with Aroclor 1242. Despite this misclassification, the results did not appear to
be biased. In contrast, the samples spiked at 100 (ig/mL were correctly classified as Aroclor 1254 but were
all biased low. Although these results suggested that Aroclor classification had little effect on the quantification
of the extract samples, there was an obvious, consistent error introduced into the analysis of the 100-(jg/mL
samples to cause the low bias. For the entire extract data set, the overall percent recovery was 84%.
Table 4-6. Accuracy of the reference laboratory for extract samples
Spike
concentration
(ug/mL)
0"
0"
10
10
100
100
Outdoor site
Sample
ID
129
132
127
130
128
131
Avg
cone
(ug/mL)
0
0
10.9
12.1
67.4
63.8
Recovery
(%)
n/a
n/a
109
121
67
64
Chamber site
Sample
ID
229
232
227
230
228
231
Avg
cone
(ug/mL)
0
0
9.6
8.9
65.2
57.7
Recovery
(%)
n/a
n/a
96
89
65
58
Combined sites
Avg
cone
(ug/mL)
10.4
63.5
Recovery
(%)
n/a
104
64
a All PCB concentrations reported as non-detects by the laboratory.
Representativeness
Representativeness expresses the degree to which sample data accurately and precisely represent the capability
of the method. Representativeness of the method was assessed based on the data generated for clean-QC
samples (i.e., method blanks and laboratory control samples) and PE samples. Based on the data assessment
(discussed in detail in various parts of this section), it was determined that the representativeness of the
reference laboratory data was acceptable. In addition, acceptable performance on laboratory audits
substantiated that the data set was representative of the capabilities of the method. In all cases, the performance
of the reference laboratory met all requirements for both audits and QC analyses.
Completeness
Completeness is defined as the percentage of measurements that are judged to be usable (i.e., the result was
not rejected). Usable results were obtained for 248 of the 256 samples submitted for analysis by the reference
laboratory. Eight results (for sample IDs 110 and 112) were deemed incomplete and therefore not valid because
the measurements were not quantitative. To calculate completeness, the total number of complete results were
divided by the total number of samples submitted for analysis, and then multiplied by 100 to express as a
percentage. The completeness of the reference laboratory was 97%, where a completeness of 95% or better is
typically considered acceptable.
30
-------
Comparability
Comparability refers to the confidence with which one data set can be compared with another. The
demonstration study was designed to have a one-to-one, sample-by-sample comparison of the PCB results
obtained by the reference laboratory and the PCB results obtained by the technology being evaluated. Based
on thorough examination of the data and acceptable results on the PE samples, it was concluded that the
reference laboratory's SOPs for extraction and analysis, and the data generated using these procedures, were
of acceptable quality for comparison with the field technology results. Additional information on comparability
was available because the experimental design incorporated randomized analysis of blind, replicate samples.
Evaluation of the replicate data implicated some of the individual data points as suspect (see Table D-2). The
reference laboratory's suspect data were compared with the technology data on a special-case basis, and
exceptions were noted.
Summary of Observations
Table 4-7 provides a summary of the performance of the reference laboratory for the analysis of all sample
types used in the technology demonstration study. As shown in Table 4-7, the precision of the PE soils was
comparable to the environmental soils. A weighted average, based on the number of samples, gave a best-case
precision of 21% and a worst-case precision of 28% for all the soil data (PE and environmental). The extract
samples had a smaller overall RSD of 14%. Evaluation of overall accuracy was based on samples with certified
or known spiked concentrations (i.e., PE and extract samples). The overall
Table 4-7. Summary of the reference laboratory performance
Sample matrix
Blank
Environmental soil with
interferences
Soil
Best case
(excluding suspect data)
Soil
Worst case
(including suspect data)
Extract
Sample type
Soil
Extract
Sample ID 110
Sample ID 112
PE
Environmental
< 125 ppm
> 125 ppm
overall
PE
Environmental
< 125 ppm
> 125 ppm
overall
10 ppm
100 ppm
overall
Number of samples
8
16
4
4
63
107
17
187
64
108
20
192
16
16
32
Precision
(average % RSD)
n/a"
n/a"
18
23
19
21
21
26
56
28
19
8
14
Accuracy
(average % recovery)
All samples were
reported as non-detects.
All samples were
reported as non-detects.
101
n/a6
n/a6
101
105
n/a6
n/a6
105
104
64
84
* Because the results were reported as non-detects, precision assessment is not applicable.
6 Accuracy assessment calculated for samples of known concentration only.
31
-------
accuracy, based on percent recovery, for the PE samples was 105% for the worst case (which included the
suspect value) and 101% for the best case (which excluded the suspect value). These results indicated that the
reference laboratory measured values were unbiased estimates of the certified PE concentrations (for samples
that contained <50 ppm of PCBs). Accuracy for the extract samples at 10 ppm was also unbiased, with an
average percent recovery of 104%. However, the accuracy for the extract samples at 100 ppm was biased low,
with an average recovery of 64%. Overall, the average percent recovery for all extract samples was 84%. The
reference laboratory correctly reported all blank samples as non-detects but had difficulty with two soil sample
IDs (110 and 112) that contained chemical interferences. In general, the reference laboratory's completeness
would be reduced, at the expense of an improvement in precision and accuracy, if the suspect measurements
were excluded from the data analysis. Based on this analysis, it was concluded that the reference laboratory
results were acceptable for comparison with the developer's technology.
32
-------
Section 5
Technology Performance and Evaluation
Objective and Approach
The purpose of this section is to present the evaluation of the data generated by the EnviroGard PCB test kit.
The technology's precision and accuracy performance are presented for the data generated in the demonstration
study. In addition, an evaluation of comparability, through a one-to-one comparison with the reference
laboratory data, is presented. An evaluation of other aspects of the technology (such as cost, sample
throughput, hazardous waste generation, and logistical operation) is also presented in this section.
Interval Reporting
The EnviroGard results were reported as concentration ranges that were designated as intervals incorporating
parentheses/bracket notation. The parentheses indicated that the end-points of the concentration range were
excluded, while brackets indicated that the end-points were included. As shown in Table 5-1, the interval (5,
10] indicates that the PCB concentration range is greater than 5 and less than or equal to 10.
As discussed briefly in Section 2 of this report, this technology cannot distinguish between different Aroclors.
The test kit has been calibrated to respond in a one-to-one ratio with Aroclor 1248. If site history or
information indicates that a different Aroclor (e.g., 1260, 1242, or 1254) is present in the samples, a conversion
can be applied (see Table 2-3 for more information). The Aroclor-specific reporting intervals for the
EnviroGard results are listed in Table 5-1. For the purposes of the demonstration, SDI was provided
information about the type of Aroclor present in the samples. Dilution of samples during analysis to optimize
method performance altered some of the standard intervals shown in Table 5-1 for select samples.
Data Assessment
Objective
The purpose of the data assessment section is to present the evaluation of the performance of SDFs EnviroGard
PCB test kit through a statistical analysis of the data. PARCC parameters were used to evaluate the test kit's
ability to measure PCBs in PE soil, environmental soil, and extract samples. The developer analyzed splits of
replicate samples that were also analyzed by the reference laboratory (72 PE soil samples, 136 environmental
soil samples, and 24 extract samples). See Section 4 for a more detailed
analysis of the reference laboratory's results. Replicate samples were analyzed by the developer at two different
sites (under outdoor conditions and inside an environmentally controlled chamber) to evaluate the effect of
environmental conditions on the test kit's performance; see Section 3 for further details on the different sites.
Evaluation of the measurements made at each site indicated that there were no
significant differences between the two data sets. Because environmental conditions did not appear to affect
the results significantly, data from both sites were also combined for each parameter (precision and
Table 5-1. EnviroGard PCB test kit reporting intervals
33
-------
Default mode
Interval
[0,1)
[1,5]
(5, 10]
(10,50]
(50, oo)
Aroclor 1248
concentration range
0 50
Conversion to specific Aroclor
Interval
[0,2)
[2, 10]
(10,20]
(20, 100]
(100,»)
Aroclor 1242/1260
concentration range
0 100
Interval
[0,1.1)
[1.1,5.5]
(5.5,11.1]
(11.1,55]
(55,oo)
Aroclor 1254
concentration range
0< PCB ppm < 1.1
1.1 < PCB ppm < 5.5
5.5< PCB ppm < 11.1
ll.KPCBppm< 55
PCB ppm > 55
accuracy) to determine the test kit's overall performance. All statistical tests were performed at the 5%
significance level. Appendix D contains the raw data that were used to assess the performance of the
EnviroGard test kit.
Precision
Precision is the reproducibility of measurements under a given set of conditions. The frequency with which the
same interval was reported within a set of replicates was used to quantify precision. Examples of how the
precision was classified are presented in Table 5-2. Reporting a higher number of replicates in the same interval
for a given replicate set indicates higher precision. In other words, reporting all four replicate results as the
same interval indicates the highest possible precision.
Performance Evaluation Samples
Table 5-3 summarizes the precision information for the EnviroGard test kit's analysis of the PE samples. The
EnviroGard test kit reported all four replicates as the same interval (i.e., high precision) for three PE sample
sets under each set of environmental conditions (i.e., outdoor and chamber conditions). Operating under the
outdoor conditions, seven of eight replicate sets were classified as having medium to high precision. None of
the replicate sets was reported with the lowest precision (i.e., none). Under the chamber conditions, medium
to high precision was achieved for five of eight replicate sets, with the remaining three replicate sets classified
as having low precision. A more detailed analysis of the data showed that the replicates having medium to low
precision classifications were never more than one interval away from the most frequently reported interval.
Table 5-2. Classification of precision results
If the replicate results are...
[0,1), [1,5], (5, 10], (10, 50]
[0,1), [1,5], [1,5], (5, 10]
[0, 1), [1, 5], [1, 5], [1, 5]
[1, 5], [1, 5], [1, 5], [1, 5]
...and the number reported
in identical intervals are...
0
2
3
4
...then the precision
classification is...
none
low
medium
high
34
-------
Table 5-3. Precision of the EnviroGard PCS test kit for PE soil samples
Certified
PE
Cone.
(ppm)
0
2.0
2.0
5.0
10.9
20.0
49.8
50.0
50.0
Outdoor site
Sample
ID
126"
118
124
120
122
119
125
121
123
# in each precision
classification
precision
Number of replicates reported in
identical intervals
0"
0
2
X
1
3
X
X
X
X
4
4
X
X
X
X
3
Chamber site
Sample
ID
226"
218
224
220
222
219
225
221
223
precision
Number of replicates reported in
identical intervals
0"
0
2
X
X
X
3
3
X
X
2
4
X
X
X
X
3
" Indicates that all four replicates
6 Blank data were not included in
were reported as different intervals.
the determination of the overall precision.
Environmental Soil Samples
The EnviroGard results for the replicate environmental soil sample measurements are presented in Table 5-4.
Under the outdoor conditions, 6 of 17 replicate sets achieved the highest precision classification (i.e., the same
interval was reported for all 4 replicates). Under the chamber conditions, 10 of 17 sample sets were classified
as high precision. Of the sample sets for which precision was classified
as medium to low, only Sample ID 115 had one replicate result that differed by more than one interval range.
Because the majority of measurements fell below 125 ppm, precision was also assessed by partitioning the
results into two ranges: low (reference laboratory values < 125 ppm) and high concentrations (reference
laboratory values > 125 ppm). See Section 4 for the delineation of which Sample IDs were in the low and high
categories. For the low concentrations, 40% of the sample sets were reported with all four replicates in the
same interval (i.e., highest possible precision). For the high concentration category, 100% of the sample sets
(five of five) were reported with the highest possible precision.
Table 5-4. Precision of the EnviroGard PCB test kit for environmental soil samples
35
-------
Outdoor site
Sample
ID
101
102
103
104
105
106
107
108
109
110
111
112
113"
114
115
116
117
# in each
precision
classification
precision
Number of replicates reported in
identical intervals
0"
0
2
X
X
X
X
X
5
3
X
X
X
X
X
X
6
4
X
X
X
X
X
X
6
Chamber site
Sample
ID
206
207
208
209
210
211
212
213
214
215
216
217
201
202
203
204
205
precision
Number of replicates reported in
identical intervals
0"
0
2
X
1
3
X
X
X
X
X
X
6
4
X
X
X
X
X
X
X
X
X
X
10
" Indicates that all four replicates were reported as different intervals.
" Bold sample IDs were matching Paducah sample pairs (i.e., 113/201, 114/202, 115/203, 116/204, 117/205).
The Paducah soils (indicated by bold Sample IDs in Table 5-4) were analyzed at both sites to provide an
assessment of EnviroGard's performance under different environmental conditions. For these samples, the data
generated under both environmental conditions were also combined to provide an overall assessment of
precision. Sample IDs 113 and 201, 114 and 202, 115 and 203, 116 and 204, and 117 and 205 represented
replicate Paducah soil sample sets, where the 100 series were samples analyzed under the outdoor conditions,
and the 200 series were samples analyzed inside the chamber. Additional statistical analysis was used to
compare the effect of the two environmental conditions on the measurements. Results from this analysis showed
that there were no significant differences in the data generated at each site. This indicated that these different
environmental conditions did not impact the performance of the test kit.
36
-------
Extract Samples
The EnviroGard results for the replicate extract measurements are presented in Table 5-5. All three sample sets
analyzed under the outdoor conditions were reported with the highest possible precision (i.e., all four replicates
within the same interval). One sample set (the blank) analyzed under the chamber conditions achieved the
highest precision, and the remaining two sample sets were reported with medium precision (i.e., three replicates
were reported in the same interval).
Precision Summary
A summary of the EnviroGard test kit's overall precision is presented by sample type (PE, environmental soil,
and extract samples) in Table 5-6. For PE and environmental soil samples, 38% and 47% of the samples,
respectively, achieved the highest possible precision (i.e., all four sample replicates were reported as the same
interval). For the extract samples, 50% of the samples achieved the highest precision.
Table 5-5. Precision of the EnviroGard PCB test kit for extract samples
Outdoor site
Sample ID
130"
131
132
# in each
precision
classification
precision
Number of replicates reported in
identical intervals
0"
0
2
0
3
0
4
X
X
X
3
Chamber site
Sample
ID
230"
231
232
precision
Number of replicates reported in
identical intervals
0"
0
2
0
3
X
X
2
4
X
1
" Indicates that all four replicates were reported as different intervals.
6 Blank data were not included in the determination of the overall precision
Table 5-6. Overall precision of the EnviroGard PCB test kit for all sample types
Environmental
site
Outdoor site
Chamber site
Combined sites
Percentage of samples classified in each precision category
PE samples
none
0
0
0
low
13
38
25
med
50
25
38
hidh
38
38
38
Environmental soil samples
none
0
0
0
low
29
6
18
med
35
35
35
hidh
35
59
47
Extract samples
none
0
0
0
low
0
0
0
med
0
100
50
hidh
100
0
50
37
-------
Accuracy
Accuracy represents the closeness of the EnviroGard test kit's measured PCB concentrations to the certified
values. Because the EnviroGard test kit produced interval results, accuracy was evaluated in terms of the
percentage of samples that agreed with, were above (i.e., biased high), and were below (i.e., biased low) the
certified value.
Performance Evaluation Soil Samples
Table 5-7 contains a comparison between the EnviroGard interval result and the corresponding certified PE
value. The interval(s) listed under a particular column indicates how many of the four replicates were
reported as that interval. For example, for Sample ID 120, two replicates were reported as (11.1, 55], and two
were reported as (5.5, 11.1]. For Sample ID 119, three are reported as (10, 50], and one is reported as (50, °°).
Note that performance acceptance ranges for the PE results, which are the guidelines established by the
provider of the PE materials to gauge acceptable analytical results, are also presented in Table 5-7 for
information. These ranges were not used to evaluate the EnviroGard results because the acceptance ranges
overlap several EnviroGard reporting intervals.
Table 5-7. EnviroGard test kit accuracy data for PE soil samples
Certified
cone.
(ppm)
(acceptance
range, ppm)
0
(n/a)
2.0
(0.7-2.2)
2.0
(0.9-2.5)
5.0
(2.1-6.2)
10.9
(4.0-12.8)
20.0
(11.4-32.4)
49.8
(23.0-60.8)
50.0
(19.7-63.0)
50.0
(11.9-75.9)
Outdoor site
Sample
ID
126
118
124
120
122
119
125
121
123
# of replicates reported at each interval
1
(5, 10]
(50, co)
(20, 100]
(20, 100]
2
(5.5,11.1]
(11.1,55]
3
[1,5]
(10,50]
(55,°o)
(100,»)
4
[0,1)
[2, 10]
(20, 100]
(55,oo)
Chamber site
Sample
ID
226
218
224
220
222
219
225
221
223
# of replicates reported at each interval
1
(5, 10]
[1.1,5.5]
(5, 10]
(5.5,11.1]
(11.1,55]
[2, 10]
(20, 100]
(20, 100]
(100,oo)
2
(10,20]
(55,oo)
(20, 100]
(100,oo)
3
[1,5)
4
[0,1)
[2, 10]
(10,50]
(55,»)
38
-------
The data in Table 5-7 were used to derive the accuracy results that are presented in Table 5-8. Accuracy was
based on a comparison of the certified PE value with the interval reported by the EnviroGard test kit. If the
interval encompassed the certified PE value, the EnviroGard result "agreed" with the certified value. If the
EnviroGard result was above the certified value, the result was classified as "biased high." If the EnviroGard
result was below the certified value, the result was classified as "biased low." For example, for Sample ID 222,
the certified value was 10.9 ppm (for Aroclor 1260). The comparison would be classified as "agreed" for the
EnviroGard interval result (10, 20], as "biased high" for the interval result (20, 100], or as "biased low" for
the interval result [2, 10]. Separate comparisons were made for the two environmental conditions to determine
if ambient temperature and humidity had an effect on the technology performance. Statistical analysis showed
that there was no significant difference between the results obtained by the test kit under the two different
environmental conditions evaluated in this demonstration. Therefore, all PE sample results were combined to
determine the overall percentage of agreement between EnviroGard results and the certified PE value. The
overall percentage of agreement was 51%. A single EnviroGard sample result was biased low. A total of 47%
of the test kit's results were biased high, and in approximately 80% of these samples, the difference between
the certified PE value and the lower end of the reporting interval was less than 10 ppm. This indicated that
many of the results that were biased high were reported in the next highest reporting interval. For example, for
PE samples where the certified PE value was 50 ppm, the EnviroGard test kit would generally report the
sample concentration as (55, °°). The large number of sample results that were biased high relative to the
certified PE value relates to the conservatism that the manufacturer intentionally incorporates into the
EnviroGard's calculation of reported results.
Table 5-8. Evaluation of agreement between EnviroGard's PE sample results and the certified PE values as a measure of
accuracy
Environmental
site
Outdoor site
Chamber site
Combined sites
Relative to certified values for performance evaluation samples
Biased Agree Biased
low high
0%
3%
1%
44%
58%
51%
56%
39%
47%
Number of samples
36
36
72
Extract Samples
Table 5-9 contains a comparison between the EnviroGard interval result and the corresponding spike
concentration for the extract samples. The EnviroGard test kit's percentage of agreement with the spike
concentration of the extract samples is summarized in Table 5-10. Statistical analysis showed that
environmental conditions had no significant effect upon the performance of the EnviroGard test kit.
Therefore, the data sets generated under the outdoor and chamber conditions were combined. Overall, 14 of
24 extract samples (58%) agreed with the spike concentration. Approximately 38% were biased high, where
most of the high bias was associated with the 10 ppm extract samples, all of which were reported
39
-------
Table 5-9. Accuracy of the EnviroGard test kit for extract samples
Spike cone.
(ug/mL)
0
10
100
Outdoor site
Sample
ID
132
130
131
# of replicates reported at each interval
1
2
3
4
[0,1)
(20, 100]
(555«>)
Chamber site
Sample
ID
232
230
231
# of replicates reported at each interval
1
[1,5]
(11.1,55]
2
3
[0,1)
(55,»)
4
(20, 100]
Table 5-10. Evaluation of agreement between EnviroGard's extract results and the spike concentration as a measure of
accuracy
Environmental
site
Outdoor site
Chamber site
Combined sites
Relative to spike concentration for extract xamples
Biased Agree Biased
low high
0%
8%
4%
67%
50%
58%
33%
42%
38%
Number of samples
12
12
24
as (20, 100]. Under the outdoor conditions, no measurements were biased low; under the chamber conditions,
one measurement was biased low relative to the spike concentration.
False Positive/False Negative Results
A false positive (fp) result [10] is one in which the technology detects PCBs in the sample when there actually
are none. A false negative (fh) result [10] is one in which the technology indicates that there are no PCBs
present in the sample, when there actually are. Both fp and fh results are influenced by the method detection
limit of the technology. All of the eight blank soil samples were reported as the lowest reporting interval, which
included zero, so the fp result was 0%. Of the 192 non-blank soil samples analyzed, EnviroGard reported seven
in the lowest reporting interval (e.g., 0 to 2 ppm). All of the corresponding reference laboratory results fell into
EnviroGard's reporting interval (e.g., 1.5 ppm). Therefore, the fn result for the soil samples was 0%. For the
eight extract samples, the EnviroGard test kit reported one blank as [1, 5]. Therefore, the fp result was 13%.
All other extract samples were reported as non-blanks, so the fn result was 0%.
Representativeness
Representativeness expresses the degree to which the sample data accurately and precisely represent the
capability of the technology. The performance data were accepted as being representative of the technology
because the EnviroGard test kit was capable of analyzing diverse sample types (PE samples, simulated wipe
extract samples, and actual field environmental samples) under multiple environmental conditions. When this
technology is used, quality control samples should be analyzed to assess the performance of the EnviroGard
PCB test kit under the testing conditions.
40
-------
Completeness
Completeness is defined as the percentage of measurements that are judged to be useable (i.e., the result was
not rejected). Valid results were obtained by the technology for all 232 samples. Therefore, completeness was
100%.
Comparability
Comparability refers to the confidence with which one data set can be compared to another. A one-to-one
sample comparison of the EnviroGard results and the reference laboratory results was performed for all soil
samples. Accuracy was evaluated in terms of the percentage of samples that agreed with, were above (i.e.,
biased high), and were below (i.e., biased low) the certified value. For comparability, the EnviroGard results
were compared with the results generated by the reference laboratory, including both environmental soils and
PE samples. Sample IDs 110 and 112 were excluded because the reference laboratory did not generate
quantitative results for these samples. The results are summarized in Table 5-11. The percentage of EnviroGard
results that agreed with the reference laboratory results was 53%. Approximately 45% were biased high, and
approximately 2% were biased low relative to the results reported by the reference laboratory.
Table 5-11. Evaluation of agreement between EnviroGard's soil results and the reference laboratory's results as a measure
of comparability
Environmental
site
Outdoor site
Chamber site
Combined sites
Relative to reference laboratory results for soil samples
Biased Agree Biased
low high
2%
2%
2%
52%
54%
53%
46%
44%
45%
Number of samples
96
104
200
For the extract samples, the comparison of the EnviroGard test kit's result with the reference laboratory result
was similar to the comparison with the spike concentrations (shown in Table 5-10). There was 63% agreement
between the laboratory method and the field technology, and for the rest of the samples (37%), the EnviroGard
results were biased high
The soil data not included in previous comparability evaluations (because the replicate data for the reference
laboratory were considered suspect) are shown in Table 5-12. Refer to Section 4, in particular Table 4-1, for
more information on the reference laboratory's suspect measurements. The reference laboratory's suspect data
were compared with the EnviroGard's matching results. For Sample IDs 110 and 112, the reference laboratory
obtained qualitative results only. The EnviroGard test kit also had some difficulty with Sample ID 110,
producing results in two different intervals, in contrast to Sample ID 112, where all four replicates were
reported as the same interval. For the other five suspect values for the reference laboratory data, the
EnviroGard test kit generated results that agreed with the replicate means of the reference laboratory. These
comparisons demonstrated that the EnviroGard test kit did not have difficulty with most of the samples that
were troublesome for the reference laboratory.
41
-------
Table 5-12. Comparison of the EnviroGard results with the reference laboratory's suspect measurements
Sample ID
110
112
106
205
216
217
225
Reference laboratory
Suspect measurement
(ppm)
-------
Regulatory Decision-Making Applicability
One objective of this demonstration was to assess the technology's ability to perform at regulatory decision-
making levels for PCBs, specifically 50 ppm for soils and 100 (jg/100cm2 for surface wipes. To assess this,
the EnviroGard test kit's performance for PE and environmental soil samples ranging in concentration from
40 to 60 ppm (as determined by the paired reference laboratory analyses) can be used. For this concentration
range, the test kit's results agreed with the reference laboratory's results 39% of the time. Results were biased
high 59% of the time, and 2% of the results were biased low. No fns were observed for this concentration
range. As discussed in the "Accuracy" section, the absence of fns and the presence of a large percentage of
results that were biased high were most likely a result of the degree of conservatism that the manufacturer has
incorporated into the calculation of results.
Assuming a 10-mL extract volume, extract samples (at 10 and 100 (jg/mL) represented surface wipe sample
concentrations of 100 (jg/100cm2 and 1000 (jg/100cm2. For simulated wipe extract samples, the percentage
of the EnviroGard's measurements that agreed with the reference laboratory results was 63%. Approximately
38% of the results were biased high and none were biased low. No fn results were observed for the extract
samples.
Additional Performance Factors
Sample Throughput
Sample throughput is representative of the estimated amount of time required to extract the PCBs, to perform
appropriate reactions, and to analyze the sample. Operating under the outdoor conditions, the SDI team's
sample throughput rate was 18 samples/hour. Working in the chamber, the rate was lower, around 9 to 10
samples/hour. This increased sample throughput under the outdoor conditions may be attributed to the analysis
order: because SDI analyzed samples under the chamber conditions first, they may have gained valuable
experience that was applied during the analysis of the outdoor samples. Alternatively, SDI may have had more
difficulty with the sample matrices that were analyzed only under the outdoor conditions.
Cost Assessment
The purpose of this economic analysis is to provide an estimation of the range of costs for an analysis of PCB-
contaminated soil samples using the EnviroGard PCB test kit and a conventional analytical reference laboratory
method. The analysis was based on the results and experience gained from this demonstration, costs provided
by SDI, and representative costs provided by the reference analytical laboratories that offered to analyze these
samples. To account for the variability in cost data and assumptions, the economic analysis was presented as
a list of cost elements and a range of costs for sample analysis by the EnviroGard test kit and by the reference
laboratory.
Several factors affected the cost of analysis. Where possible, these factors were addressed so that decision-
makers can independently complete a site-specific economic analysis to suit their needs. The following
categories are considered in the estimate:
sample shipment costs,
labor costs,
43
-------
equipment costs,
waste disposal costs.
Each of these cost factors is defined and discussed and serves as the basis for the estimated cost ranges
presented in Table 5-14. This analysis assumed that the individuals performing the analyses were fully trained
to operate the technology. SDI recommends that new users attend a training session that SDI offers on the use
of the EnviroGard test kit. Costs for sample acquisition and pre-analytical sample preparation, which are tasks
common to both methods, were not included here.
Table 5-14. Estimated analytical costs for PCB soil samples
EnviroGard PCB test kit
Strategic Diagnostics Inc.
Sample throughput rate: 18 samples/hour (outdoors)
9-10 samples/hour (chamber)
Cost category Cost ($)
Sample shipment 0
Labor
Mobilization/demobilization 250^00
Travel 15-1 ,000 per analyst
Per diem 0-150 per day per analyst
Rate 30-75 per hour per analyst
Equipment
Mobilization/demobilization 0-150
Kit rental fee 450 per week
Kit purchase price 2,495
Training < 935
Reagents/supplies 26 per sample
Waste disposal 75-1,060
EPA SW-846 Method 8080/8081/8082
Reference laboratory
Typical turn-around time: 14-30 days
Cost category Cost (S)
Sample shipment
Labor 100-200
Overnight shipping charges 50-150
Labor
Mobilization/demobilization included"
Travel included
Per diem included
Rate 44-239 per
sample
Equipment
Mobilization/demobilization included
Rental/purchase of system included
Reagents/supplies included
Waste disposal included
""Included" indicates that the cost is included in the labor rate.
EnviroGard Costs
Because the samples were analyzed on site, no sample shipment charges were associated with the cost of
operating the EnviroGard test kit. Labor costs included mobilization/demobilization, travel, per diem, and on-
site labor.
• Labor mobilization/demobilization: This cost element included the time for one person to prepare for
and travel to each site. The estimate ranged from 5 to 8 h, at a rate of $50 per hour.
44
-------
• Travel: This element was the cost for the analyst(s) to travel to the site. If the analyst is located near
the site, the cost of commuting to the site (estimated to be 50 miles at $0.30 per mile) would be minimal
($15). The estimated cost of an analyst traveling to the site for this demonstration ($1000) included the
cost of airline travel and rental car fees.
• Per diem: This cost element included food, lodging, and incidental expenses and was estimated ranging
from zero (for a local site) to $150 per day per analyst.
• Rate: The cost of the on-site labor was estimated at a rate of $30 to $75 per hour, depending on the
required expertise level of the analyst. This cost element included the labor involved with the entire
analytical process comprising sample preparation, sample management, analysis, and reporting.
Equipment costs included mobilization/demobilization, rental fees or purchase of equipment, and the reagents
and other consumable supplies necessary to complete the analysis.
• Equipment mobilization/demobilization: This included the cost of shipping the equipment to the test site.
If the site is local, the cost would be zero. For this demonstration, the cost of shipping equipment and
supplies was estimated at $150.
• Rental/purchase: The fee to rent the EnviroGard test kit at the time of the demonstration was $450 per
week. At the time of the demonstration, the cost of purchasing the equipment was $2495. The purchase
price included a photometer and accessories.
• Reagents/supplies: These items are consumable and are purchased on a per-sample basis. At the time
of the demonstration, the cost of the reagents and supplies needed to prepare and analyze PCB soil
samples using the EnviroGard was $26 per sample. This cost included the sample preparation supplies,
assay supplies, and consumable reagents.
Waste disposal costs are estimated based on the 1997 regulations for disposal of PCB-contaminated waste. The
EnviroGard test kit generated approximately 20 Ib of vials containing soils and liquid solvents (classified as
solid PCB waste suitable for disposal by incineration) and approximately 20 Ib of other solid PCB waste (i.e.,
used and unused soil, gloves, paper towels, ampules, etc.). The cost of disposing of PCB solid waste by
incineration at a commercial facility was estimated at $1.50 per pound. The cost for solid PCB waste disposal
at El IP was estimated at $18 per pound. The test kit also generated approximately 19 Ib of liquid waste. The
cost for liquid PCB waste disposal at a commercial facility was estimated at $0.25 per pound, while the cost
at ETTP was estimated at $11 per pound.
Reference Laboratory Costs
Sample shipment costs to the reference laboratory included the overnight shipping charges, as well as labor
charges associated with the various organizations involved in the shipping process.
• Labor: This cost element included all of the tasks associated with the shipment of the samples to the
reference laboratory. Tasks included packing the shipping coolers, completing the chain-of-custody
documentation, and completing the shipping forms. Because the samples contained PCBs, the coolers
45
-------
were inspected by qualified personnel to ensure compliance with the U.S. Department of
Transportation's shipping regulations for PCBs. The estimate to complete this task ranged from 2 to
4 h at $50 per hour.
• Overnight shipping: The overnight express shipping service cost was estimated to be $50 for one 50-lb
cooler of samples.
The labor bids from commercial analytical reference laboratories that offered to perform the PCB analysis for
this demonstration ranged from $44 to $239 per sample. The bid was dependent on many factors, including
the perceived difficulty of the sample matrix, the current workload of the laboratory, and the competitiveness
of the market. In this case, the wide range of bids may also be related to the cost of PCB waste disposal in a
particular laboratory's state. LAS Laboratories was awarded the contract to complete the analysis as the lowest
qualified bidder ($44 per sample). This rate was a fully-loaded analytical cost, including equipment, labor,
waste disposal, and report preparation.
Cost Assessment Summary
An overall cost estimate for the EnviroGard test kit versus the reference laboratory was not made because of
the extent of variation in the different cost factors, as outlined in Table 5-14. The overall costs for the
application of each technology will be based on the number of samples requiring analysis, the sample type, and
the site location and characteristics. Decision-making factors, such as turn-around time for results, must also
be weighed against the cost estimate to determine the value of the field technology versus the reference
laboratory.
General Observations
The following are general observations regarding the field operation and performance of the EnviroGard test
kit:
• The system was light, easily transportable, and rugged. It took about an hour for the SDI team to
prepare to analyze samples on the first day of testing. While working at the outdoor site, the SDI team
completely disassembled their work station, bringing everything inside, at the close of each day. It took
the SDI team less than an hour each morning to prepare for sample analyses.
• Three operators were used for the demonstration because of the number of samples and working
conditions, but the technology can be operated by a single person. With three SDI technologies
(D TECH, EnviroGard, and RaPID Assay) being demonstrated, SDI elected to work as a team to
complete the analyses for each technology (as opposed to three SDI people working with three different
technologies).
• Operators generally require 2 to 4 h of training. They should have a basic knowledge of field analytical
techniques.
• The SDI team calibrated the photometer often, analyzing 1,5, 10, and 50 ppm standards and a negative
control with every batch of 12 samples. This was done to account for changing environmental
conditions (i.e., temperature and humidity).
46
-------
• Data processing and interpretation was minimal. The results were quantified relative to the four
calibration standards and reported in terms of intervals using the sample information provided (Aroclor
type and ratio).
• New start and stop solutions were used with every 12 samples. All reagents were allowed to come to
room temperature before use. Although it is recommended that all of the reagents in the test kit be stored
under refrigerated conditions, the SDI team noted that the reagents can be stored at ambient conditions
for several hours.
• The measurement system (photometer) was battery-operated.
• The EnviroGard test kit generated approximately 20 Ib of vials containing soils and liquid solvents
(classified as solid PCB waste suitable for disposal by incineration) and approximately 20 Ib of other
solid PCB waste (i.e., used and unused soil, gloves, paper towels, ampules, etc.). The test kit also
generated approximately 19 Ib of liquid waste (aqueous with trace methanol).
Performance Summary
A summary of the performance characteristics of SDFs EnviroGard PCB test kit, presented previously in this
chapter, is shown in Table 5-15. The performance of the EnviroGard test kit was characterized as biased,
because nearly half (49%) of the EnviroGard results disagreed with the certified PE values, and imprecise,
because over half (62%) of the PE replicate results were not reported as the same interval. The test kit had no
fp or fh results for the soil samples. For extract samples, the test kit had one fp result and no fh results. It
should also be noted that there was an increased likelihood that results would be biased high as a result of the
conservatism that the manufacturer has incorporated into the calculation of results.
47
-------
Table 5-15. Performance summary for the EnviroGard PCB test kit
Feature/parameter
Blank results
Precision
Accuracy
False positive results
False negative results
Comparison with reference laboratory
results
Regulatory decision-making applicability
Sample throughput
Power requirements
Operator requirements
Cost
Hazardous waste generation
Performance summary
Soils: Correctly reported all 8 samples as [0,1) ppm
Extracts: 7 samples reported correctly as [0,1) ppm; 1 sample reported as
[1,5]
Percentage of combined sample sets where all four replicates were
reported as the same interval
PE soils: 38%
Environmental soils: 47%
Extracts: 50%
PE soils Extracts
agreed =51% agreed = 58%
biased high = 47% biased high = 38%
biased low = 1% biased low = 1%
Blank soils: 0% (0 of 2 samples)
Blank extracts: 13% (1 of 8 samples)
PE and environmental soils: 0% (0 of 192 samples)
Spiked extracts: 0% (0 of 16 samples)
PE and environmental soils Extracts
agreed =53% agreed = 63%
biased high = 45% biased high = 38%
biased low = 2% biased low = 0%
PE and environmental soils Extracts
(40 to 60 ppm) (100 ug/100cm2, 1000 ug/100cm2)
agreed = 39% agreed = 63%
biased high =59% biased high =38%
biased low = 2% biased low = 0%
9-10 samples/hour (chamber)
1 8 samples/hour (outdoors)
Battery-operated photometer
Basic knowledge of chemical techniques; 1-A hours technology-specific
training
Incremental: $26 per sample
Instrumental: $2,495 (purchase); $450 (weekly rental)
Approximately 20 Ib of solid/liquid (classified as solid PCB waste suitable
for disposal by incineration)
Approximately 20 Ib of solid (used gloves, pipettes, paper towels, etc.)
Approximately 1 9 Ib of liquid waste (aqueous with trace methanol)
48
-------
Section 6
Technology Update and Representative Applications
Objective
The purpose of this section is to allow the developer to provide information regarding new developments with
its technology since the demonstration activities. In addition, the developer has provided a list of representative
applications in which its technology has been or is currently being used.
Technology Update
Reconfiguration of Soil Extraction (Sample Preparation) Products
SDI is in the process of commercializing a common extraction kit for three of four remediation immunoassay
test kit product lines. The affected product lines include the EnviroGard, EnSys (not demonstrated here), and
RaPID Assay test kit systems. The new "Universal Extraction Kit" will be used with assay kits of these three
product lines, with extraction solvents or dilution reagents specifically formulated to match individual kits
available as kit component options where required. The new test kit configuration will provide increased user
convenience and simplify the product specification and ordering process without affecting test kit analytical
performance. Commercialization of the new Universal Extraction Kit was initiated in April 1998. The new kits
are not for use with the D TECH product line, which will continue to use the existing SDI Soil Extraction Pac
products.
Instrument Consolidation
Associated with the incorporation of several independently developed product lines into SDFs product
offerings, some consolidation of equipment and instrumentation is anticipated in the near future. This will
consist primarily of reducing the number of pipette types and photometers used to perform the assays. While
pipette types and procedures for pipetting reagents and reading and interpreting assay results may change
slightly, no effect on assay performance will result.
Representative Applications
In a 1997 report entitled Field Analytical and Site Characterization Technologies: Summary of Applications
[11], the use of SDI immunoassay kits is documented at more than 30 remediation sites under state or federal
oversight. Contact information is provided for many of the immunoassay kit users at these sites. The summary
report can be obtained from the National Center for Environmental Publications and Information (NCEPI).
Hard copies of the report can be ordered, free of charge, by telephone, (513) 891-6561; by fax (513) 891-6685;
or through the NCEPI homepage on the Internet at http://www.epa.gov /ncepihom/. The summary report is
available for viewing or downloading as a .pdf file from the CLU-IN Internet website. The Internet address is
http://clu-in. com/pubichar. htm.
49
-------
Data Quality Objective Example
This application of SDFs EnviroGard PCB immunoassay kit is based on data quality objective (DQO) methods
for project planning advocated by the American Society for Testing and Materials [12, 13] and EPA [14].
ORNL derived a DQO example from the performance results in Section 5. This example, which is presented
in Appendix E, illustrates the use of the performance data for SDFs EnviroGard from the ETV demonstration
in the DQO process to select the number of samples to characterize the decision rule's fp and fn error rates.
50
-------
Section 7
References
[1] Erickson, M. D. Analytical Chemistry ofPCBs, 2nd ed., CRC Press/Lewis Publishers, Boca Raton,
Fla., 1997.
[2] "Polychlorinated Biphenyls (PCBs) Manufacturing, Processing, Distribution in Commerce, and Use
Prohibitions," Code of Federal Regulations, 40 CFR, pt. 761, rev. 7, December 1994.
[3] Maskarinec, M.P., et al. Stability of Volatile Organics in Environmental Soil Samples, ORNL/TM-
12128, Oak Ridge National Laboratory, Oak Ridge, Tenn., November 1992.
[4] U.S. Environmental Protection Agency. "Method 8081: Organochlorine Pesticides and PCBs as
Aroclors by Gas Chromatography: Capillary Column Technique," in Test Methods for Evaluating
Solid Waste: Physical/Chemical Methods (SW-846), 3d ed., Final Update II, Office of Solid Waste
and Emergency Response, Washington, D.C., September 1994.
[5] Oak Ridge National Laboratory. Technology Demonstration Plan: Evaluation of Polychlorinated
Biphenyl (PCB) Field Analytical Techniques, Chemical and Analytical Sciences Division, Oak
Ridge National Laboratory, Oak Ridge, Tenn., July 1997.
[6] U.S. Environmental Protection Agency. Data Quality Objectives for Remedial Response Activities,
EPA 540/G-87/003, EPA, Washington, D.C., March 1987.
[7] Sachs, Lothar. Applied Statistics: A Handbook of Techniques, 2nd ed., Springer-Verlag, New York,
1984.
[8] Snedecor, G. W., and William G. Cochran. Statistical Methods, Iowa State University Press, Ames,
Iowa, 1967.
[9] Draper, N. R., and H. Smith. Applied Regression Analysis, 2nd ed., John Wiley & Sons, New York,
1981.
[ 10] Berger, Walter, Harry McCarty, and Roy-Keith Smith. Environmental Laboratory Data Evaluation,
Genium Publishing Corp., Schenectady, NY., 1996.
[11] U.S. Environmental Protection Agency. Field Analytical and Site Characterization Technologies:
Summary of Applications, EPA-542-R-97-011, Office of Solid Waste and Emergency Response,
Washington, D.C., November 1997
51
-------
[12] American Society for Testing and Materials (ASTM). Standard Practice for Generation of
Environmental Data Related to Waste Management Activities Quality Assurance and Quality
Control Planning and Implementation, D5283-92, 1997.
[13] American Society for Testing and Materials (ASTM), Standard Practice for Generation of
Environmental Data Related to Waste Management Activities Development of Data Quality
Objectives, D5 792-95, 1997.
[14] U.S. Environmental Protection Agency. Guidance for Data Quality Assessment.. EPA QA/G-9;
EPA/600/R-96/084, EPA, Washington, D.C., July 1996.
52
-------
Appendix A
Description of Environmental Soil Samples
53
-------
54
-------
Table A-l. Summary of soil sample descriptions
Location
Oak Ridge
Oak Ridge
Oak Ridge
Oak Ridge
Oak Ridge
Paducah
Portsmouth
Tennessee
Reference Soil
Request for
Disposal
(RFD)#
40022
40267
24375
43275
134555
97002
7515
n/a
Drum#
02
01
02
03
04
01
02
03
01
02
03
01
02
03
04
858
1069
1096
1898
2143
2528
3281
538
940
4096
n/a
Description
Soil from spill cleanup at the Y-12 Plant in Oak Ridge, Tennessee.
This soil is PCB-contaminated soil excavated in 1992.
Soil from the Elza Gate area, a DOE Formerly Utilized Sites Remedial
Action Program site in Oak Ridge, Tennessee. This soil is PCB-
contaminated soil that was excavated in 1992.
Catch-basin sediment from the K-71 1 area (old Powerhouse Area) at
the DOE East Tennessee Technology Park (formerly known as Oak
Ridge Gaseous Diffusion Plant) in Oak Ridge, Tennessee. This soil is
PCB-contaminated storm drain sediment that was excavated in 1991.
Soil from the K-25 Building area at the DOE East Tennessee
Technology Park (formerly known as Oak Ridge Gaseous Diffusion
Plant) in Oak Ridge, Tennessee. This soil is PCB-contaminated soil
that was excavated in 1993.
Soil from the K-707 area at the DOE East Tennessee Technology Park
(formerly known as Oak Ridge Gaseous Diffusion Plant) in Oak Ridge,
Tennessee. This soil is PCB-contaminated soil from a dike spillage that
was excavated in 1995.
Soil from the DOE Paducah Gaseous Diffusion Plant in Kentucky. This
soil is PCB-contaminated soil from a spill cleanup at the C-746-R
(Organic Waste Storage Area) that was excavated in 1989.
Soil from the DOE Portsmouth Gaseous Diffusion Plant in Ohio. This
soil is PCB-contaminated soil from a probable PCB oil spill into the
East Drainage Ditch that was excavated in 1986.
Captina silt loam from Roane County, Tennessee; used as a blank in
this study (i.e., not contaminated with PCBs)
55
-------
56
-------
Appendix B
Characterization of Environmental Soil Samples
57
-------
-------
Table B-l. Summary of environmental soil characterization
Location
Oak Ridge
Paducah
Portsmouth
Sample
ID
101
102
103
104
105
106
107
108
109
110
111
112
126, 226
113,201
114,202
115,203
116,204
117,205
206
207
208
209
210
216
211
217
212
213
214
215
RFD
drum # a
40022-02
40267-03
40267-01
40267-04
40267-01 S*
24375-03
24375-01
40267-02
24375-02
43275-01
134555-03S*
43275-02
non-PCB soil
97002-04
97002-01
97002-03
97002-02
97002-02S *
7515-4096
7515-1898
7515-1096
7515-2143
7515-0940
7515-0538
7515-0538S*
7515-0538S*
7515-2528
7515-3281
7515-0858
7515-1069
Composition
% gravel
0
0.5
0.2
0.6
0.5
0.5
2.5
0.4
0.3
0
0.5
0.1
0
0
0.2
0.1
0.4
0
0.2
0.4
0
0.3
0.5
0.5
0.5
0
1.3
% sand
91.8
99.3
96.7
98.2
94.8
87.8
92.5
94.2
93.1
89.2
88.1
91.4
85.6
92.4
87.6
83.6
93.7
87.1
78.0
74.4
74.3
73.0
73.3
70.4
72.6
65.8
75.0
% silt + clay
8.2
0.2
3.1
1.2
4.7
11.7
5.0
5.4
6.6
10.8
11.4
8.5
14.4
7.6
12.2
16.3
5.8
12.9
21.8
25.2
25.7
26.7
26.3
29.1
26.8
34.2
23.7
Total organic
carbon
(mg/kg)
5384
13170
13503
15723
14533
19643
1196
9007
1116
14250
10422
38907
9249
1296
6097
3649
4075
3465
3721
3856
10687
7345
1328
5231
5862
6776
4875
PH
7.12
7.30
7.21
7.07
7.28
7.36
7.26
7.30
7.48
7.57
7.41
7.66
7.33
7.71
7.64
7.59
7.43
7.72
7.66
7.77
7.71
7.78
7.78
7.92
7.67
7.85
7.56
* Request for disposal drum number (see Table A-l).
6 "S" indicates that the environmental soil was spiked with additional PCBs.
59
-------
-------
Appendix C
Temperature and Relative Humidity Conditions
61
-------
62
-------
Table C-l. Average temperature and relative humidity conditions during testing periods
Date
7/22/97
7/23/97
7/24/97
7/25/97
7/26/97
7/27/97
7/28/97
7/29/97
Outdoor site
Average
temperature
(°F)
85
85
85
80
85
80
79
b
Average
relative humidity
(%)
62
70
67
70
55
75
88
b
Chamber site
Average
temperature
(°F)
70 "
60 "
58
56
57
55
57
55
Average
relative humidity
(%)
38 "
58 fl
66
54
51
49
52
50
a The chamber was not operating properly on this day. See discussion in Section 3.
* No developers were working outdoors on this day.
63
-------
120
100 --
n
7/22/97 7/23/97 7/24/97 7/25/97
7/26/97
7/27/97
7/28/97
Figure C-l. Summary of temperature conditions for outdoor site.
120
100 --
7/22/97
7/23/97
7/24/97
7/25/97
7/26/97
7/27/97
7/28/97
Figure C-2. Summary of relative humidity conditions for the outdoor site.
64
-------
80
70 --
40 --
20 --
10 --
0 -I
TT
7/22/97 7/23/97 7/24/97 7/25/97 7/26/97 7/27/97 7/28/97 7/29/97
Figure C-3. Summary of temperature conditions for chamber site.
7/22/97 7/23/97 7/24/97 7/25/97 7/26/97 7/27/97 7/28/97 7/29/97
Figure C-4. Summary of relative humidity conditions for chamber site.
65
-------
-------
Appendix D
EnviroGard PCB Test Kit
PCB Technology Demonstration Sample Data
67
-------
Legend for Appendix D Tables
Table Heading
Obs
Sample ID
Rep
EnviroGard Result
Ref lab result
Reference Aroclor
Type
Order
Definition
Observation
Sample identification
101 to 126 = outdoor site soil samples
127 to 130 = outdoor site extract samples
201 to 226 = chamber site soil samples
227 to 230 = chamber site extract samples
Replicate of sample ID (1 through 4)
EnviroGard' s measured PCB concentration (ppm)
LAS reference laboratory measured PCB concentration (ppm)
Values with "<" are samples that the reference
laboratory reported as "< reporting detection limit"
Aroclor(s) identified by the reference laboratory
Sample = environmental soil
1242, 1248, 1254, 1260 = Aroclor in PE samples
Blank = non-PCB-contaminated sample
Order of sample analysis by SDI
(started with 2001-21 16, then 1001-1 1 16)
68
-------
Table D-l. EnviroGard PCB technology demonstration soil sample data
Obs
Sample
ID
Rep
EnviroGard
Result
(ppm)
(5.5,11.1]
(5.5,11.1]
[1.1,5.5]
(5.5,11.1]
(11.1,55]
(11.1,55]
(11.1,55]
(5.5,11.1]
Ref Lab
Result
(ppm)
Reference
Aroclor
Type
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Order
[1.1,5.5]
(5.5,11.1]
(5.5,11.1]
[1.1,5.5]
[1.1,5.5]
(5.5,11.1]
[1.1,5.5]
(5.5,11.1]
(5.5,11.1]
[1.1,5.5]
[1.1,5.5]
(5.5,11.1]
(55,
-------
Obs
Sample
ID
Rep
EnviroGard
Result
(ppm)
Ref Lab
Result
(ppm)
Reference
Aroclor
Non-Detect
Non-Detect
Non-Detect
Non-Detect
Type
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Order
(11.1,55]
(5.5,11.1]
(5.5,11.1]
(11.1,55]
70
-------
Obs
Sample
ID
Rep
EnviroGard
Result
(ppm)
Ref Lab
Result
(ppm)
Reference
Aroclor
Type
Order
Non-Detect
Non-Detect
Non-Detect
Non-Detect
Blank
Blank
Blank
Blank
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
71
-------
Obs
Sample
ID
Rep
EnviroGard
Result
(ppm)
Ref Lab
Result
(ppm)
Reference
Aroclor
Type
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Sample
Order
72
-------
Obs
Sample
ID
Rep
EnviroGard
Result
(ppm)
Ref Lab
Result
(ppm)
Reference
Aroclor
Type
Order
1.2
1.4
Non-Detect
Non-Detect
Non-Detect
Non-Detect
Blank
Blank
Blank
Blank
73
-------
Table D-2. EnviroGard PCB test kit technology demonstration extract sample data
Obs
Sample
ID
Rep
EnviroGard
Result
(ppm)
Ref Lab
Result
(ppm)
Reference
Aroclor
Type
Spikea
(ppm)
10
10
10
10
100
100
100
100
Order
1113
1110
1109
1114
.100]
Non-Detect
Non-Detect
Non-Detect
Non-Detect
Blank
Blank
Blank
Blank
0
0
0
10
10
10
10
Non-Detect
Non-Detect
Non-Detect
Non-Detect
Blank
Blank
Blank
Blank
100
100
100
100
0
0
0
0
"Nominal spike concentration of the extract sample prepared by ORNL.
74
-------
Table D-3. Corrected reference laboratory data
Error
Transcription
Calculation
Interpretation
Sample ID
106
130
205
207
210
118
119
209
214
219
101 a
101 fl
107
109
113*
113*
119
127
201
219
Reported Result
(ppm)
<490
5.6
32,000
180
160
3.6
4.3
2.3
43.0
29.0
<0.7
<0.7
<1.3
18.0
<0.9
<1.0
18.0
7.2
< 1.0
21.0
Corrected Result
(ppm)
255.9
10.3
3,305.0
17.8
123.2
2.1
17.4
37.9
26.0
22.4
0.5
0.6
1.2
1.5
0.6
0.7
21.2
10.9
0.6
26.0
a Two of four measurements in Sample ID 101 were corrected.
Two of four measurements in Sample ID 113 were corrected.
75
-------
-------
Appendix E
Data Quality Objective Example
77
-------
-------
Disclaimer
The following hypothetical example serves to demonstrate how the information provided in this report may be
used in the data quality objectives (DQO) process. This example serves to illustrate the application of
quantitative DQOs to a decision process, but it cannot attempt to provide a thorough education in this topic.
Please refer to other educational or technical resources for further details. Additionally, since the focus of this
report is on the analytical technology, this example makes the simplifying assumption that the contents of these
drums will be homogeneous. In the real world, however, this assumption is seldom valid, and matrix
heterogeneity constitutes a source of considerable uncertainty which must be adequately evaluated if the overall
certainty of a site decision is to be quantified.
Background and Problem Statement
An industrial company discovered a land area contaminated with PCBs from an unknown source. The
contaminated soil was excavated into waste drums. Preliminary characterizaton determined that a number of
PCB drums had to be incinerated to reduce or eliminate the PCB contamination. The incinerated soil was
placed in drums for disposal in a landfill. However, a final check of each drum was required to verify for the
regulator that the appropriate level of cleanup had been achieved. The regulator required that no drum have
more than 2 ppm of PCB. The company's DQO team was considering the use of the EnviroGard PCB test kit
to measure the PCB concentration in each drum. Because the type of Aroclor was unknown, all measurements
would be reported as Aroclor 1254. The plan was to randomly select soil samples collected from each drum
and test with the kit to determine if the concentration was in one of the three intervals: [0,1.1), [1.1, 5.5], or
(5.5,<=°). Recall that this notation describes the concentration ranges 0 ppm < PCB < 1.1 ppm, 1.1 ppm < PCB
< 5.5 ppm, and PCB > 5.5 ppm, as used in Section 5. The DQO team decided that a drum would be
reprocessed by incineration if any of the EnviroGard results indicated a concentration in the intervals [1.1, 5.5],
or (5.5, °°). In agreement with the regulator, the DQO team determined that a decision rule for disposal would
be based on the number of samples with PCB concentrations in the intervals [1.1, 5.5], or (5.5, <=°).
General Decision Rule
If all of the PCB sample results show concentrations in [0, 1.1), then send the soil drum to the landfill.
If any of the PCB sample results are different from [0, 1.1) then reprocess the soil drum by
incineration.
DQO Goals
EPA's Guidance for Data Quality Assessment [14] states in Sect. 1.2: "The true condition that occurs with
the more severe decision error...should be defined as the null hypothesis." The DQO Team decided that the
more severe decision error would be for a drum to be erroneously sent to a landfill if the drum's PCB
concentration actually exceeded the 2 ppm limit. Therefore, the null hypothesis is constructed to assume that
a drum's true PCB concentration exceeds the 2 ppm limit, and as a "hot" drum, it would be sent to the
incinerator. Drums would be sent to the landfill only if the null hypothesis is rejected and it is concluded that
the "true" average PCB concentration is less than 2 ppm.
79
-------
With the null hypothesis defined in this way, a false positive (fp) decision is made when it is concluded that a
drum contains less than 2 ppm PCBs (i.e., the null hypothesis is rejected), when actually the drum is "hot" (i.e.,
the null hypothesis is true). The team required that the error rate for sending a "hot" drum to the landfill (i.e.,
the fp error rate for the decision) could not be more than 5%. Therefore, a sufficient number of samples must
be taken from each drum that the fp decision error rate (FP) is 0.05 (or less) if the true drum concentration is
2 ppm. This scenario represents a 5% chance of sending a drum containing 2 ppm or more of PCBs to the
landfill. The EnviroGard interval boundary of 1.1 ppm can be used as a conservative estimate of the 2 ppm
criterion.
The DQO team did not want to send an excessive number of drums to the incinerator if the average PCB
concentration was less than 2 ppm because of the expense. In this situation, a false negative (fh) decision is
made when it is concluded that a drum is "hot" (i.e., the null hypothesis is not rejected), when in actuality, the
drum contains soil with less than 2 ppm PCBs (i.e., the null hypothesis is actually false). After considering the
guidelines presented in Section 1.1 of EPA's Guidance for Data Quality Assessment [14], the DQO team
recommended that the fh decision error rate (FN) be 0.10 if the true drum concentration was less than 1.1 ppm.
That is, there would be a 10% probability of sending a drum to the incinerator (denoted as Pr[Take Drum to
Incinerator] if the true PCB concentration for a drum was less than 1.1 ppm.
Permissible FP and FN Error Rates and Critical Decision Point
FP: Pr[Take Drum to Landfill] < 0.05 when true PCB concentration > 1.1 ppm
FN: Pr[Reprocess Drum in Incinerator] < 0.10 when true PCB concentration < 1.1 ppm
Use of Technology Performance Information to Implement the Decision Rule
Technology performance information is used to evaluate whether a particular analytical technology can produce
data of sufficient quality to support the site decision. Because the DQO team is considering the use of the
EnviroGard PCB kit, the performance of this technology [as reported in this Environmental Technology
Verification (ETV) report] was used to assess its applicability to this project. The question arises: How many
samples are needed from a single drum to permit a statistically valid decision at the specified certainty? Recall
that the simplifying assumption was made that the PCB distribution throughout the soil within a single drum
is homogeneous and thus matrix heterogeneity will not contribute to overall variability. The only variability,
then, to be considered in this example is the variability in performance of the EnviroGard kit's analytical
method, which is determined by precision and accuracy studies.
Determining the Number of Samples
The number of samples needed to satisfy the FP and FN requirements depends on the misclassification error
rates of the EnviroGard PCB kit. Two types of misclassifications have to be considered:
1. Underestimating the PCB concentration—classifying a sample concentration in [0, 1.1) when the true PCB
concentration is greater or equal to 1.1 ppm
80
-------
2. Overestimating the PCB concentration—classifying a sample concentration in [1.1, 5.5] or (5.5, °°) when
the PCB concentration is less than 1.1 ppm.
The ETV demonstration results on performance evaluation soil samples and on environmental soil samples
indicated the error rates for the two types of misclassifications to be
PU = Pr[Underestimating the PCB concentration ] = 0.017
P0 = Pr [Overestimating the PCB concentration ] = 0.400
The probability distribution of classifying the number of soil samples in different concentration intervals
follows a binomial probability distribution [7]. This probability distribution and the requirements for FP and
FN can be used to determine the number of samples to meet the DQO goals. The FP for the decision rule is
related to PU by
FP = Pr[ All EnviroGard results < 1.1 ppm for PCB > 1.1 ppm ] = (PVY (E-l)
The FP error rate decreases as the sample size increases. The sample size is solved as
Log(FP)
— ^ - -
Log(Pu)
Log(FP)
n = — ^ - -
where
n = number of samples from a drum to be measured
FP = false positive decision error rate (e.g., FP = 0.05)
PU = probability of underestimating the PCB concentration
Zog(0.05) = -1.301 = Q ?4
Zog(0.017) -1.770
The sample size was rounded up to the next integer, an operation that will decrease the FP for the decision rule.
The DQO team would have to analyze one sample from each drum to meet the decision rule's fp requirement.
The FN for the decision rule is related to P0 by
FN = Pr[ Some of EnviroGard results > 1.1 ppm for PCB < 1.1 ppm ] = 1 - (1 - Po)n (E-3)
81
-------
The probability of an fn decision (FN = sending a drum for reprocessing) actually increases with increasing
sample size because the chance that the kit will overestimate a result increases with continued testing. The
sample size required to meet the FN requirement is
= Log( I - FN)
Log(l - P0) (E~4)
where
n = number of samples from a drum to be measured
FN = false negative decision error rate (e.g., FN = 0.10)
P0 = probability of overestimating a PCB concentration
n = Log(l - 0.10) = -0.046
Log( 1 - 0.400 ) -0.222
The sample size must be rounded up to n = 1 (fractions of a sample analysis are not possible). When n = 1,
and the above equation is solved for FN, it is found that the DQO team cannot meet its goal of 10% FN and
would have to accept an FN of 40%. This situation occurs because of the 40% overestimation error rate of the
kit. If a decision about a drum is based on a single sample, and that sample has a 40% chance of being
overestimated, there is consequently a 40% chance that the drum will be unnecessarily sent for reprocessing
through the incinerator (which was the definition of FN). Although this amount of conservatism may be
desirable in some situations, in others it may not be. The only way to reduce the FN in this kind of scenario is
to use an analytical technology with a lower overestimation error rate.
The DQO team in our example decided that the sampling procedure would be to randomly select one soil
sample from each drum and test the sample with the EnviroGard PCB kit. The DQO team would send the drum
to the landfill if the EnviroGard result was less than 1.1 ppm, and send the drum to be reprocessed by
incineration if the EnviroGard result was greater than 1.1 ppm. To meet the FP requirement of 5%, the DQO
team would have to accept an FN of 40%.
Decision Rule for 5% FP and 40% FN
If one randomly selected soil sample has a PCB test result reported as the interval [0, 1.1) then send the
soil drum to the landfill.
If one randomly selected soil sample has a PCB test result different from [0, 1.1) then reprocess the soil
drum by incineration.
82
-------
Alternative FP Parameter
The following statement describes how changing the FP requirement from 5% to 0.1 % would affect the decision
rule. Using FP = 0.001, the calculated sample sizes would be n = 1.70, which is rounded up to 2. The FN
would be 64%. The higher FN occurs because each of the two samples has a 40% chance of being
overestimated, and if only one is overestimated, the drum is sent for reprocessing. The decision rule for the
lower FP requirement would be
Decision Rule for FP = 0.1% and FN = 64%
If two randomly selected soil samples have their PCB test results reported as [0, 1.1) then send the soil
drum to the landfill.
If either of the two randomly selected soil samples has a PCB test result different from [0, 1.1) then
reprocess the soil drum by incineration.
83
------- |