Uf»l*gfl'
Office -of ar%
-.!»^'.an CC
PCP Immunoassay
Technologies
nnovative Technology
Evaluation Report
atesi i'.•Bs[,ii\L!«"oi-aJN^'AfSS
I *-„*!> . i '•
.» - " '•" """*' *.'*¥: :
-------
United States Environmental Monitoring Report No.
Environmental Protection Systems Laboratory Date 1995
Agency P.O. Box 93478
Las Vegas NV 89193-3478
PCP Immunoassay
Technologies
Innovative Technology
Evaluation Report
-------
EPA/No.
Date 1995
PCP Immunoassay Technologies
Innovative Technology Evaluation Report
Environmental Monitoring Systems Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, Nevada 89193
-------
Notice
This report was prepared for the U.S. Environmental Protection Agency's (EPA) Environmental Monitoring Systems
Laboratory, Las Vegas (EMSL-LV) by PRC Environmental Management, Inc. (PRC), in partial fulfillment of Contract
No. 68-CO-0047, Work Assignment No. 0-40. The opinions, findings, and conclusions expressed herein are that of
the contractor and not necessarily those of the EPA or other cooperating agencies. Mention of company or product
names is not to be construed as an endorsement by the agency.
111
-------
Abstract
This report describes the demonstration and evaluation of three immunoassay field screening technologies designed
to determine pentachlorophenol (PCP) contamination in soil and water. The three immunoassay technologies were
(1) the Penta RISc Test System developed by EnSys, Inc., (2) the Penta RaPID Assay developed by Ohmicron
Corporation, and (3) the EnviroGard Pentachlorophenol (PCP) Test Kit developed by Millipore Corporation. The
technologies were demonstrated in Morrisville, North Carolina, in August 1993, by PRC Environmental Management,
Inc. (PRC), under contract to the Environmental Protection Agency's (EPA) Environmental Monitoring Systems
Laboratory-Las Vegas (EMSL-LV).
The principal objective of the demonstration was to evaluate each technology for accuracy and precision at detecting
high and low levels of PCP in soil and water samples by comparing their results to those attained by a confirmatory
laboratory using standard EPA analytical methods. Each technology also was qualitatively evaluated for the length
of time required for analysis, ease of use, portability, and operating cost. Accuracy was also assessed through analysis
of performance evaluation (PE) samples, and precision was further assessed by comparing the results obtained on
duplicate samples. A secondary objective of the demonstration was to evaluate the specificity of each technology. The
evaluation of specificity was performed by examining any problems due to naturally occurring matrix effects,
site-specific matrix effects, and chemical cross reactivity. Information on specificity was gathered from each
developer, from the analysis of demonstration samples, and from a specificity study performed during the
demonstration.
IV
-------
Table of Contents
Section Page
Notice ii
Abstract iii
List of Abbreviations and Acronyms viii
Acknowledgments ix
1 Executive Summary 1
2 Introduction 4
EPA's Site Program and MMTP: An Overview 4
The Role of Monitoring and Measurement Technologies 4
Defining the Process 5
Components of a Demonstration 5
Rationale for this Demonstration 5
Demonstration Purpose, Goals, and Objectives 6
3 Predemonstration Activities 7
Identifying Developers 7
Selecting the Sites 7
Selecting the Confirmatory Laboratory and Analytical Methods 8
Training Technology Operators 8
Predemonstration Sampling and Analysis 8
4 Demonstration Design and Description 9
Demonstration Design 9
Implementation of the Demonstration Plan 9
Field Modifications to the Demonstration Plan 10
Data Collection 10
Statistical Analysis of Results 11
Intramethod Comparisons 11
Intermethod Comparisons 12
5 Confirmatory Analysis Results 14
Confirmatory Laboratory Procedures 14
Sample Holding Times 14
Sample Extraction 14
Reporting Limits and Initial and Continuing Calibrations 15
-------
Table of Contents (Continued)
Section
Sample Analysis 15
Quality Control Procedures 15
Data Reporting 16
Data Quality Assessment 17
Confirmatory Laboratory Costs and Turnaround Times 18
6 EnSys Inc.: Penta RISc Test System 20
Theory of Operation and Background Information 20
Operational Characteristics 22
Performance Factors 24
Detection Limits 24
Sample Throughput 24
Linear Range and Drift 25
Specificity Study 25
Intramethod Assessment 26
Comparison of Results to Confirmatory Laboratory Results 28
Accuracy 28
Soil Data Set 28
Water Data Set 32
Precision 33
7 Ohmicron Corporation: Penta RaPID Assay 34
Theory of Operation and Background Information 34
Operational Characteristics 34
Performance Factors 38
Specificity 39
Intramethod Assessment 40
Comparison of Results to Confirmatory Results 45
Soil Samples: Accuracy 48
Soil Samples: Precision 50
Water Samples: Accuracy 51
Water Samples: Precision 52
8 Millipore Corporation: EnviroGard PCP Test Kit 53
9 Applications Assessment 55
EnSys Inc.: Penta RISc Test System 55
Ohmicron Corporation: Penta RaPID Assay 56
Millipore Corporation: EnviroGard PCP Test Kit 56
10 References 58
VI
-------
List of Figures
7-1 Total Soil Data Set 48
7-2 Former Koppers Site Soil Samples 48
7-3 Winona Post Site Soil Samples 48
7-4 Total Water Data Set 51
7-5 Former Koppers Site Water Samples 51
7-6 Winona Post Site Water Samples 51
List of Exhibits
Exhibit Page
6-1 The processes for the EnSys PCP test kits 21
7-1 The process used by the Penta RaPID Assay 35
8-1 The processes for the Millipore technology 54
Vll
-------
List of Tables
Table Page
5-1 Soil Matrix Spike Sample Results for EPA Methods 8270A and 8151A 17
5-2 Water Matrix Spike Sample Results for EPA Methods 8270A and 515.1 17
6-1 Chemical Cross Reactivity as Reported by EnSys 22
6-2 Semiquantitative Penta RISc Test Data and Confirmatory Data for Soils 29
6-3 Semiquantitative Penta RISc Test Data and Confirmatory Data for Water 31
7-1 Penta RaPID ASSAY Inorganic Chemical Response as Reported
by Ohmicron 39
7-2 Penta RaPID ASSAY Compound Cross Reactivity as Reported
by Ohmicron 40
7-3 Penta RaPID ASSAY PCP Proficiency Sample Results: Training Results 41
7-4 Penta RaPID ASSAY PCP Proficiency Sample Results:
Demonstration Results 41
7-5 Penta RaPID ASSAY Soil Performance Evaluation Sample Results 42
7-6 Penta RaPID ASSAY Water Performance Evaluation Sample Results 42
7-7 Penta RaPID ASSAY Soil Matrix Spike Sample Results 42
7-8 Penta RaPID ASSAY Laboratory Duplicate Results: Soil Samples 43
7-9 Penta RaPID ASSAY Soil Field Duplicate Sample Results 44
7-10 Penta RaPID ASSAY Water Field Duplicate Sample Results 45
7-11 Summary of Demonstration Data: Former Koppers Site Soil Samples 46
7-12 Summary of Demonstration Data: Winona Post Soil Samples 47
7-13 Summary of Demonstration Data: Water Data 48
7-14 Summary of Regression and Residual Statistics: Ohmicron 49
Vlll
-------
List of Abbreviations and Acronyms
Beazer Beazer East, Inc.
CCA copper-chromium-arsenate
DCAA 2,4-dichlorophenylacetic acid
DQO data quality objective
BCD electron capture detector
ELISA enzyme-linked immunosorbent assay
EMSL-LV Environmental Monitoring Systems Laboratory-Las Vegas
EPA Environmental Protection Agency
ERA Environmental Research Associates
ESAT Environmental Services Assistance Team
FASP Field Analytical Screening Program
GC/MS gas chromatograph/mass spectrograph
IDW investigation-derived waste
ITER Innovative Technology Evaluation Report
Koppers Koppers Company
/xg/L micrograms per liter
/xg/kg micrograms per kilogram
MCL maximum contaminant level
mg/kg milligrams per kilogram
MMTP Monitoring and Measurement Technologies Program
ORD Office of Research and Development
OSWER Office of Solid Waste and Emergency Response
PCP pentachlorophenol
PE performance evaluation
ppb part per billion
ppm part per million
PRC PRC Environmental Management, Inc.
QADE Quality Assurance and Data Evaluation
QA/QC quality assurance/quality control
QAPjP quality assurance project plan
RB reagent blank
RCRA Resource Conservation and Recovery Act
RECAP Region 7 Environmental Collection and Analysis Program
RPD relative percent difference
RREL Risk Reduction Engineering Laboratory
SARA Superfund Amendments and Reauthorization Act of 1986
SITE Superfund Innovative Technology Evaluation
SMO Sample Management Office
SVOC Semivolatile Organic Compound
USI Unit Structures, Inc.
IX
-------
Acknowledgments
This demonstration and the subsequent preparation of this report required the services of numerous personnel from
the Environmental Protection Agency's Environmental Monitoring Systems Laboratory (Las Vegas, Nevada);
Environmental Protection Agency, Region 7 (Kansas City, Kansas); EnSys, Inc. (Research Triangle Park, Norfli
Carolina); Ohmicron Corporation (Newtown, Pennsylvania); Millipore, Inc. (Bedford, Massachusetts); Beazer East,
Inc. (Pittsburgh, Pennsylvania); Winona Post, Inc. (Winona, Missouri); and PRC Environmental Management, Inc.,
(Kansas City, Kansas; Cincinnati, Ohio; and Chicago, Illinois). The cooperation and efforts of these organizations
and personnel are gratefully acknowledged.
-------
Section 1
Executive Summary
This innovative technology evaluation report (ITER)
presents information on the demonstration and evaluation
of three immunoassay field screening technologies for
determining pentachlorophenol (PCP) contamination in
soil and water. These technologies were demonstrated in
Morrisville, North Carolina, in August 1993. The
demonstration was conducted by PRC Environmental
Management, Inc. (PRC), under contract to the
Environmental Protection Agency's (EPA)
Environmental Monitoring Systems Laboratory-Las
Vegas (EMSL-LV). The demonstration was developed
under the Monitoring and Measurements Technologies
Program (MMTP) of the Superfund Innovative
Technology Evaluation (SITE) Program.
The three immunoassay technologies evaluated
during this demonstration were (1) the Penta RISc Test
System developed by EnSys, Inc., (2) the Penta RaPID
Assay developed by Ohmicron Corporation, and (3) the
EnviroGard Pentachlorophenol (PCP) Test Kit developed
by Millipore Corporation. These technologies were
demonstrated in conjunction with the demonstration of
two other field screening technologies: the HNU-Hanby
Test Kit developed by HNU Systems and the Field
Analytical Screening Program (FASP) PCP Method
developed by EPA's Region 7 under the Superfund
Program. The demonstrations of these other two
technologies are presented in separate reports similar to
this one.
The first objective of this demonstration was to
evaluate each of the field screening technologies for
accuracy and precision in detecting high and low levels
of PCP in soil and water samples by comparing their
results to those attained by a confirmatory laboratory
using standard EPA analytical methods. These EPA-
-approved methods are used to provide legally defensible
analytical data for the purpose of monitoring or for the
enforcement of environmental regulations. Because
these EPA-approved methods are used by the regulatory
community, this demonstration also used these methods.
Though these methods may include inherent tendencies
which may bias data or may include procedures with
which developers disagree, they are the best methods for
providing legally defensible data as defined by the
regulatory community. To remove as much of these
inherent tendencies as possible, PRC used post-hoc
residual analysis to remove data outliers. Each
technology also was qualitatively evaluated for the length
of time required for analysis, ease of use, portability,
and operating cost.
The second objective of the demonstration was to
evaluate the specificity of each technology. The
evaluation of specificity was performed by examining
any problems due to naturally occurring matrix effects,
site-specific matrix effects, and chemical cross
reactivity. Information on specificity was gathered from
each developer, from the analysis of demonstration
samples, and from a specificity study performed during
the demonstration.
The site selected for demonstrating the technologies
was the former Koppers Company (Koppers) site in
Morrisville, North Carolina. One of the reasons this site
was selected was because a Risk Reduction Engineering
Laboratory (RREL) SITE demonstration was occurring
simultaneously, thus allowing EMSL-LV and RREL to
combine logistical and support efforts. Another reason
for selecting the former Koppers site was that historical
documentation indicated its PCP contamination ranged
from none detected to 3,200 parts per million (ppm) in
soil and from none detected to 1,490 parts per billion
(ppb) in groundwater. The PCP carrier used at this site
was a mixture of isopropyl ether and butane. Soil and
water samples also were collected from the Winona Post
site in Winona, Missouri. These samples were shipped
to the former Koppers site for inclusion as demonstration
samples. Winona Post samples were included to broaden
the scope of the demonstration by introducing a different
sample matrix and PCP carrier, diesel fuel, to the
evaluation of each technology. Findings for each of the
immunoassay field screening technologies demon-strated
are summarized below.
-------
EnSys, Inc.: Penta RISc Test System
The Penta RISc Test System is designed to quickly
provide semiquantitative results for PCP concentrations
in soil and water samples. In its standard format for soil
analysis, the semiquantitative ranges assessed are:
greater than 50 ppm, between 50 and 5 ppm, between 5
and 0.5 ppm, and less than 0.5 ppm. This technology's
ranges for water analysis are greater than 5,000 ppb,
between 5,000 and 500 ppb, between 500 and 5 ppb, and
less than 5 ppb. The developer will customize these
ranges to a user's needs. The immunoassay chemistry
produces compound-specific reactions to PCP allowing
its detection and quantitation. Polyclonal antibodies are
fixed to the inside wall of a test tube where they offer
binding sites for PCP. An enzyme conjugate containing
a PCP derivative is added to the test tube to compete
with PCP from samples for antibody binding sites.
Excess sample and enzyme conjugate are removed from
the test tube by washing, and reagents are added to the
test tube to react with the enzyme conjugate causing a
color formation. The amount of color formed by a
sample is then compared to the color formed by a PCP
standard taken through all of the immunoassay steps,
simultaneously. The comparison is made with the use of
a differential photometer.
The system is portable and can be operated
outdoors; however, temperature extremes and humidity
can affect its performance. It is easy to use even by
those with no immunoassay testing experience. The
highest number of demonstration samples analyzed in one
10-hour day was 40; the average number analyzed in one
10-hour day was 23. The detection limit reported by the
developer for soil samples is 0.5 ppm and that for water
samples is 5 ppb. The system can be affected by
naturally occurring matrix effects such as humic acids,
pH, or salinity. Site-specific matrix effects that can
affect the system include PCP carriers, such as
petroleum hydrocarbons or solvents, and other chemicals
used in conjunction with PCP, such as creosote, copper-
-chromium-arsenate (CCA), or herbicides. Chemicals
similar in structure to PCP can provide positive results.
The system was found to be most affected by tetra-
chlorophenols and trichlorophenols. A specificity study
performed during the demonstration showed that 2,3,4,6-
-tetra chlorophenol and 2,4,6-trichlorophenol would
provide a positive response when present at a
concentration of 10 ppm.
PRC evaluated field and laboratory duplicate
samples to determine the technology's precision. The
precision of the system for soil samples was found to be
79 percent. This is below the demonstration's criteria
for acceptable precision. The precision of the system for
water samples was found to be 100 percent. The
comparison of the accuracy of the soil analysis showed
that 73 percent of the time the technology was correct.
The technology gave false positive results 19 percent of
the time and gave false negative results 8 percent of the
time. All of the false negative results were for samples
containing less than 10 ppm. When examined for a PCP
carrier effect, the frequency of correct readings was
higher for the samples collected at the Winona Post site.
The frequency of false positive and false negative results
was similar between the two sites. The system produced
correct results 47 percent of the time for the water
analysis. It had a 42 percent false positive rate and an
11 percent false negative rate. When this data was
examined for a PCP carrier effect, no false negatives
were reported on samples from the former Koppers site;
in addition the frequency of correct and false positive
results was higher for the samples collected from this
site.
Overall, the technology was found not to be accurate
when compared to Level 3 data, but this technology can
produce Level 2 data. However, in some cases it
produced only Level 1 data. The technology is
conservative; however, using an absolute definition of
accuracy, it was accurate only 73 percent of the time. A
draft version of this ITER was distributed on March 4,
1994. EnSys submitted no comments on the draft ITER.
Ohmicron Corporation: Penta RaPID Assay
The Penta RaPID Assay is designed to quickly
provide quantitative results for PCP concentrations in
soil and water samples. It uses immunoassay chemistry
to produce compound-specific reactions to PCP allowing
its detection and quantitation. Polyclonal antibodies are
bound to paramagnetic particles and are introduced into
a test tube where they offer binding sites for PCP. An
enzyme conjugate containing a PCP derivative is added
to the test tube where it competes with PCP from
samples for antibody binding sites. A magnetic field is
applied to each test tube to hold the antibodies containing
the PCP and enzyme conjugate, while excess sample and
enzyme conjugate are removed from the test tube by
washing. Reagents are then added to the test tube where
they react with the enzyme conjugate causing a color
formation. A solution is added to each test tube to stop
color formation. The color formed by a sample is then
compared to the color formed by three PCP standards
taken through all of the immunoassay steps. The
comparison is made with the use of a spectrophotometer.
The technology is portable, but should be used
indoors because fluctuations and extremes in temperature
or humidity may affect its performance. Also, the
reagents require refrigeration, and the spectrophotometer
requires electricity. The technology was found to be
easy to operate by individuals with some prior analytical
laboratory experience, but can be used by individuals
-------
with no prior immunoassay testing experience. The
highest number of samples analyzed in one 10-hour day
during the demonstration was 64, and the average
number of samples analyzed in one 10-hour day was 21.
The detection limit reported by the developer for soil
samples is 0.1 ppm and for water samples is 0.06 ppb.
The technology may be affected by naturally occurring
matrix effects such as humic acids, pH, or salinity.
Site-specific matrix effects which can affect the
technology include PCP carriers, such as petroleum
hydrocarbons or solvents, other chemicals used in
conjunction with PCP, such as creosote, CCA, or
herbicides. Specific chemicals similar in structure to
PCP can provide positive results with the technology.
The technology was found to be most affected by tetra-
chlorophenols and trichlorophenols. A specificity study
performed during the demonstration showed that
2,3,4,6-tetrachlorophenol and 2,4,6-trichlorophenol
would provide a positive response when present at a
concentration of 10 ppm.
PRC found no significant difference between the
precision of the technology and that of the confirmatory
laboratory. This conclusion was the same for both soil
and water analysis. In addition, no PCP carrier effect on
precision was observed. Results from the technology did
not meet this demonstration's criteria for accuracy, as
compared to Level 3 data. However, many of the data
groupings produced were found to be linear, indicating
that the results can be corrected mathematically. If 10 to
20 percent of the soil samples are sent to a confirmatory
laboratory, then the results from the other 80 to
90 percent can be corrected. This need for mathematical
correction to improve accuracy and the comparability of
the technology's data with confirmatory data indicate that
this technology can produce Level 2 analytical data.
However, in some cases, this technology produced only
Level 1 data. PRC found that the technology's water
data was not significantly different from the confirmatory
laboratory's when samples from the former Koppers site
were analyzed. It indicated a significant difference
between the data sets when the Winona Post site samples
were analyzed. Therefore, the technology can produce
Level 3 data for the former Koppers site water samples
and Level 2 data for the Winona Post site water samples.
The developer submitted comments on a draft
version of this report on May 4, 1994. These comments,
which ranged from requests for clarification
to technical comments on data interpretation, are
available from EMSL-LV and PRC. In addition, the
developer informed PRC that it now provides its
customers with an "Environmental Users Guide," which
advises them how to use the technology's quantitative
data relative to action levels.
M MM pore Corporation:
EnviroGard PCP Test Kit
The EnviroGard PCP Test Kit is designed to quickly
provide semiquantitative or quantitative results for PCP
concentrations in soil and water samples. Polyclonal
antibodies are fixed to the inside wall of a test tube
where they offer binding sites for PCP. An enzyme
conjugate containing a PCP derivative is added to the test
tube to compete with PCP from samples for antibody
binding sites. Excess sample and enzyme conjugate is
removed from the test tube by washing, and reagents are
added to the test tube to react with the enzyme conjugate,
causing color formation. The amount of color formed by
a sample then is compared to the color formed by three
PCP standards taken through all of the immunoassay
steps. The results can be determined visually or a
solution can be added to the test tube to stop color
formation. A comparison of the sample and standards to
a blank water sample is made with the use of a
differential photometer. The differential photometer
readings can be used to provide semiquantitative results
or a standard curve can be prepared to allow for a
quantitative determination.
The developer was extremely concerned about the
results of this demonstration, particularly the results
showing the technology's tendency to produce false
negative results when concentrations of PCP were
greater than 1,000 ppm. The developer began an
investigation into the causes of this tendency and has
revised the technology as a result. The modifications
included the addition of a detergent wash instead of a
water wash. About one third of the samples analyzed
during the demonstration were then reanalyzed by the
developer. The developer has said the false positive and
false negative rates then both fell to 3 percent. The SITE
Program, though, has not demonstrated and evaluated the
technology as modified by the developer. All of the
modifications have been incorporated into the
technology's commercial product. The developer said in
its letter to EPA regarding this demonstration: "We
believe that, based on the knowledge gained from the
field demonstration, we have been able to make
significant improvements in our product. This
experience demonstrates the utility of the SITE program
and its goal of evaluating innovative technologies."
-------
Section 2
Introduction
This ITER summarizes the procedures used to
demonstrate three immunoassay field screening
technologies designed to detect PCP. The demonstration
was conducted under the EPA's SITE Program by PRC.
The three immunoassay technologies selected were: (1)
the Penta RISc Test System developed by EnSys, Inc.,
(2) the Penta RaPID Assay developed by Ohmicron
Corporation, and (3) the EnviroGard PCP Test Kit
developed by the Millipore Corporation. These three
immunoassay technologies were demonstrated in
conjunction with the demonstration of two other screen-
ing technologies: the HNU-Hanby Test Kit developed by
HNU Systems and the FASP PCP Method developed by
EPA Region 7 under the Superfund Program. The
results of the demonstration of these other technologies
are presented in separate reports similar to this one.
EPA's Site Program and MMTP:
An Overview
At the time of the Superfund Amendments and
Reauthorization Act of 1986 (SARA), it was well
recognized that the environmental cleanup problem
needed new and better methods. The SITE Program,
therefore, was created to fulfill a requirement of SARA
that the EPA address the potential of alternative or
innovative technologies. The EPA made this program a
joint effort between the Office of Solid Waste and
Emergency Response (OSWER) and the Office of
Research and Development (ORD). The SITE Program
includes four parts:
• The Demonstration Program (for remediation
technologies)
• The Emerging Technology Program
• The Monitoring and Measurement Technologies
Program (MMTP)
• The Technology Transfer Program
The largest part of the SITE Program is concerned
with treatment technologies and is administered by
ORD's RREL in Cincinnati, Ohio. The MMTP
component, though, is administered by EMSL-LV. The
MMTP is concerned with monitoring and measurement
technologies that identify, quantify, or monitor changes
in contaminants occurring at hazardous waste sites or
that are used to characterize a site.
The MMTP seeks to identify and demonstrate
innovative technologies that may provide less expensive,
better, faster, or safer means of completing this
monitoring or characterization. The managers of
hazardous waste sites are often reluctant to use any
method, other than conventional ones, to generate critical
data on the nature and extent of contamination. It is
generally understood that the courts recognize data
generated with conventional laboratory methods; still,
there is a tremendous need to generate data more cost
effectively. Therefore, the EPA must identify innovative
approaches, and through verifiable testing of the
technologies under the SITE Program, ensure that the
technologies are equivalent to or better than conventional
technologies.
The Role of Monitoring and Measurement
Technologies
Measurement and monitoring technologies are
needed to assess the degree of contamination; to
determine the effects of contamination on public health
and the environment; to supply data for selection of
appropriate remedial action; and to monitor the success
or failure of selected remedies. Thus, the MMTP is
concerned with evaluating screening technologies,
including remote sensing, monitoring, and analytical
technologies.
Candidate technologies may come from within the
federal government or from the private sector. Through
the program, developers are to rigorously evaluate the
performance of their technologies. Finally, by
distributing the results and recommendations of those
evaluations, the market for the technologies is enhanced.
-------
Defining the Process
The demonstration process begins by canvassing the
EPA's 10 regional offices (with input by OSWER and
ORD) to determine their needs. Concurrently, classes of
technologies are identified. An ideal match is made
when there is a clear need by EPA's regions and a
number of technologies that can address that need. The
demonstrations are designed to judge each technology
against existing standards and not "one against the
other."
The demonstration is designed to provide for
detailed quality assurance and quality control (QA/QC)
to insure that a potential user can evaluate the accuracy,
precision, representativeness, completeness, and
comparability of data derived from the innovative
technology. In addition, a description of the necessary
steps and activities associated with operating the
innovative technology is prepared. Cost data, critical to
any environmental activity, are generated during the
demonstration and allow a potential user to make
economic comparisons. Finally, information on practical
matters such as operator training requirements, detection
levels, and ease of operation are reported. Thus, the
demonstration report and other informational materials
produced by MMTP provide a real-world comparison of
that technology to traditional technologies. With cost
and performance data, as well as "how to" information,
users can determine whether a new technology better
meets their needs.
Components of a Demonstration
Once a decision has been made to demonstrate
technologies to meet a particular EPA need, the MMTP
performs a number of activities. First, the MMTP
identifies potential participants and determines whether
they are interested in participating. Each developer is
advised of the general nature of the demonstration and is
provided with information common to all MMTP
demonstrations. Information is sought from each
developer about its technology to insure that the
technology meets the parameters of the demonstration.
Then, after evaluation of the information, respondents
are told whether they have been accepted into the
demonstration or not. While participants are being
identified, potential sites also are identified, and basic
site information is obtained.
The next component, probably the most important,
is the development of plans that describe how the
demonstration will be conducted. A major part of the
EPA's responsibility is the development of a
demonstration plan, quality assurance project plan
(QAPjP), and a health and safety plan. While the EPA
pays for and has the primary responsibility for these
plans, each is developed with input from all of the
demonstration's participants. The plans define how
activities will be conducted and how the technologies will
be evaluated. The MMTP also provides each developer
with site information and often predemonstration samples
so the developer can maximize the field performance of
its innovative technology. Generally, the developers
train EPA-designated personnel to operate their
technologies so that performance is not based on the
special expertise of the developers. This also insures
that potential users have valid information on training
requirements and the types of operators who typically use
a technology successfully.
The field demonstration itself is the shortest part of
the process. During the field demonstration, data is
obtained on cost, technical effectiveness (compared to
standard methods), and limiting factors. In addition,
standardized field methods are developed and daily logs
of activities and observations (including photos or
videotape) are produced. The EPA is also responsible
for the comparative, conventional method analytical costs
and the disposal of any wastes generated by the field
demonstration.
The final component of an MMTP demonstration
consists of reporting the results and insuring distribution
of demonstration information. The primary product of
the demonstration is an ITER, like this one, which is
peer-reviewed and distributed as part of the technology
transfer responsibility of the MMTP. The ITER fully
documents the procedures used during the field
demonstration, QA/QC results, the field demonstration's
results, and its conclusions. A separate QA/QC data
package also is made available for those interested in
evaluating the demonstration in greater depth. Two-page
"Technical Briefs" are prepared to summarize the
demonstration results and to insure rapid and wide
distribution of the information.
Each developer is responsible for providing the
equipment or technology product to be demonstrated, its
own mobilization costs, and the training of EPA-
-designated operators. The MMTP does not provide any
funds to developers for costs associated with preparation
of equipment for demonstration or for development, and
it does not cover the costs developers incur to
demonstrate their products.
Rationale for this Demonstration
PCP is a regulated chemical, is included in the EPA
Extremely Hazardous Substances List, and is reported in
the EPA Toxic Substances Control Act. Recently, PCP
regulations under the Resource Conservation and
Recovery Act (RCRA) have been created specifically for
wood treatment facilities. PCP is included as a target
-------
compound of many EPA-approved analytical methods Level 1:
including: EPA 500 Series Methods 515.1 and 525,
EPA 600 Series Methods 604 and 625, and EPA
SW-846 Manual Methods 8040, 8151, 8250, and 8270.
All of these methods use solvent extraction and gas
chromatography. Detection and quantitation are
performed with flame ionization, electron capture, or
mass spectrometer detectors. Analyzing samples for
PCP using these methods is typically costly and time Level 2:
consuming. EMSL-LV identified the need for effective,
accurate, low-cost screening technologies that could
provide near real-time analytical data for PCP to
Superfund and RCRA decisionmakers.
Demonstration Purpose, Goals,
and Objectives
Each of the three immunoassay technologies was
evaluated on its accuracy and precision in detecting high
and low levels of PCP in environmental samples, and the
effects, if any, of both PCP carrier and natural matrix
interferences on the technologies. The accuracy and Level 3:
precision of each technology were statistically compared
to the accuracy and precision of a conventional
confirmatory laboratory using EPA-approved analytical
methods. These comparisons also were used to
determine the highest data quality level that each
technology could attain in field applications. For the
purpose of this demonstration, the three primary data
quality levels are defined as follows (EPA 1990):
This data is not necessarily compound-
specific. Technologies that generate Level 1
data provide only an indication of
contamination. Generally, the use of these
technologies requires sample documentation,
instrument calibration, and performance
checks of equipment.
This data is compound-specific. To provide
an accuracy check, verification analysis for
at least 10 percent of the samples by an
EPA-approved method is necessary. The
method's analytical error is quantified. Use
of QC procedures such as sample
documentation, chain-of-custody
procedures, sample holding time criteria,
initial and continuing instrument calibration,
method blank analysis, rinsate blank
analysis, and trip blank analysis is
recommended.
This data is considered formal or
confirmatory analysis. Analytical error is
quantified (precision, accuracy, coefficient
of variation) and monitored. The following
QC procedures are used: sample
documentation, chains of custody, sample
holding time criteria, initial and continuing
instrument calibration, rinsate blank
analysis, trip blank analysis, and PE
samples. Detection limits are determined
and monitored.
Each technology also was qualitatively evaluated for
specificity, the length of time required for its analysis,
ease of use, portability, and operating cost.
-------
Section 3
Predemonstration Activities
Several activities were conducted by EMSL-LV,
PRC, and other demonstration participants before the
demonstration began. These activities included
identifying developers, selecting the demonstration sites,
selecting the confirmatory laboratory and analytical
methods, conducting predemonstration sampling, and
training technology operators. Predemonstration samp-
ling and analysis are normally used to allow developers
to refine their technologies and revise their operating
instructions, if necessary, prior to the demonstration.
Identifying Developers
EMSL-LV supplied PRC with the names of the
immunoassay developers to be included in the
demonstration, which were EnSys, Inc.; Ohmicron
Corporation; and Millipore Corporation, and asked that
PRC search for other technologies that could be
included. The EPA Superfund Program's FASP PCP
Method and the HNU-Hanby test kit produced by HNU
Systems were included after PRC searched for other
methods. These two methods are discussed in separate
ITERs.
Selecting the Sites
To evaluate the field screening technologies under
field conditions, hazardous waste sites suitable for the
demonstration were needed. The following criteria were
used to select the appropriate sites:
• The technologies needed to be demonstrated at
sites with a wide range of PCP contamination.
• PCP concentrations at the sites had to be well
characterized and documented.
• The sites had to be accessible for conducting
demonstration activities without interfering with
any other activities being conducted on the site.
• Because various carriers have been used with
PCP and because those carriers may influence
the technologies, it was determined that the sites
used should offer two different carriers.
The former Koppers wood treatment site was
selected as one of the two sites for this demonstration
based on these criteria. This site also was selected
because EPA's RREL was planning a SITE demon-
stration of the ETG Environmental, Inc., Base-Catalyzed
Decomposition technology there and choosing the former
Koppers site would allow logistical and support efforts
between RREL and EMSL-LV could be combined. The
second demonstration site selected was the Winona Post
site wood treatment facility. The Winona Post site is
contaminated with PCP in a diesel fuel carrier solvent.
The former Koppers site is contaminated with PCP in
butane and isopropyl ether carrier solvents.
The former Koppers site is located in Morrisville,
North Carolina, at the intersection of Highway 54 and
Koppers Road. The site is currently owned by two
companies: Beazer East, Inc. (Beazer), and Unit
Structures, Inc. (USI). The portion of the site owned by
Beazer is inactive. The portion of the site owned by USI
is currently used as a wood laminating facility. The site
occupies about 52 acres and includes the wood
laminating building, an office, and several warehouses.
Surrounding land use is a mixture of commercial, light
industrial, and rural residential. During investigations at
the former Koppers site, samples from the following
media were collected: soil, groundwater, surface water,
sediment, and fish. This sampling detected PCP from
not detected to 3,200 ppm in soils and from not detected
to 1,490 ppb in water.
The Winona Post site is located in Winona,
Missouri, on Old Highway 60 West. It has operated as
a saw milling and wood preserving facility since at least
the early 1950s. The saw milling, wood preserving, and
-------
storage areas of the facility cover about 4 acres. The
remaining portion of the 40-acre facility is wooded and
largely undeveloped. The main features of the
facility include a sawmill, office, treatment building,
debarker, storage building, and pond. Currently, the
company uses a solution of 5 percent PCP in diesel fuel.
The solution is stored in the 20,000-gallon aboveground
storage tank located adjacent to the treatment building.
In the past, the Winona Post Company mixed its own
solution from concentrated PCP. Prior to the mid-1950s,
the Winona Post Company treated wood with cresol. Six
samples were collected at the Winona Post site in 1992
and revealed PCP concentrations ranging from 886 to
24,000 ppm in the soil and sediment samples and from
10 to 528 ppm in the surface water samples.
PCP is an organic chemical with an empirical
formula of C6C15OH and a molecular weight of 266
grams per mole. PCP is an organic acid with a pKa of
4.7. PCP has a melting point of 191 °C and a boiling
point of 310 °C. The specific gravity of PCP is 1.978
grams per cubic centimeter. PCP is described as
slightly soluble in water, with 8 milligrams able to
dissolve into 100 milliliters of water. The octanol-water
partition coefficient of PCP is 6,400, which indicates that
PCP is tightly bound to the soil matrix when it is released
into the environment. PCP is used as a wood
preservative, an insecticide, a preharvest defoliant, a
slimicide, and a defoaming agent. The largest user of
PCP is the wood treating industry. For treating wood,
PCP is usually diluted to a 5 percent solution with
solvents such as mineral spirits, kerosene, diesel fuel, or
fuel oil. PCP also has been applied to wood with
methylene chloride and liquified petroleum gas, such as
butane. It has been manufactured under numerous trade
names.
Selecting the Confirmatory Laboratory and
Analytical Methods
Before this demonstration, the EPA Region 7
Laboratory arranged for all soil samples to be analyzed
under the Region 7 Environmental Collection and
Analysis Program (RECAP) Contract and all water
samples to be analyzed under its Environmental Services
Assistance Team (ESAT) Contract. SW-846 protocols
for Level 3 data were to be used to analyze soil and
water samples during this demonstration. All samples
were to be extracted by EPA Method 3540A and
analyzed by EPA Method 8270A. Any soil samples in
which PCP were not detected using Method 8270A were
to be reanalyzed by Method 8151A calibrated to PCP.
Use of Method 8151A would allow PCP to be detected
at concentrations closer to the reported lower detection
limits of the immunoassay technologies being
demonstrated. Any groundwater samples in which PCP
was not detected using Method 8270A were to be
reanalyzed using Method 515.1. This method delivers
detection levels for PCP in groundwater at or below the
detection limits of the immunoassay technologies being
evaluated. All of these analytical methods are well
established and approved by EPA. The QA procedures,
reporting requirements, and data quality objectives
(DQO) of these methods are consistent with the goals of
the SITE Program.
Training Technology Operators
During the demonstration, the technologies were
operated by PRC operators. Before the demonstration,
these individuals were trained on how to use the
technologies. This training involved a review of
operating instructions provided by the developers and
formal training by the developers. Training was
equivalent to that recommended by the developers for
users of their technologies on actual site characterization
projects.
Predemonstration Sampling
and Analysis
In July 1993, PRC prepared a predemonstration
sampling plan (PRC 1993a), and on July 12, 1993, PRC
collected predemonstration soil samples from areas at the
former Koppers site previously identified as containing
high, medium, low, and not detected concentrations of
PCP. These samples were split into replicates. One
replicate of each sample was submitted to each of the
developers. These samples were not analyzed by a
confirmatory laboratory because the contract for the
confirmatory analyses were not yet finalized. The pre-
demonstration sampling was limited to the former
Koppers site. The Winona Post site was not added to the
demonstration until after the predemonstration sampling
had already occurred. All developers agreed to
participate in the demonstration without having access to
predemonstration samples from the Winona Post site.
The unanimous finding from the developers regarding the
predemonstration samples was that they did not exhibit
their expected PCP concentrations.
-------
Section 4
Demonstration Design and Description
This section describes the organization of the
demonstration, presents an overview of the
demonstration's design, and details all deviations from
the developer- and EPA-approved demonstration plan.
Among the key portions of the demonstration plan
presented here are the types of data collected and the
statistical methods used to determine the accuracy and
precision of the technologies. A detailed description of
the demonstration is presented in the demonstration plan
andQAPjP(PRC 1993b).
Demonstration Design
The primary objective of the demonstration was to
evaluate field portable analytical technologies for their
effectiveness at detecting PCP in soil and water when
operated in field conditions. This objective included
defining the precision, accuracy, cost, and range of
usefulness for each technology. A secondary objective
was to define the DQOs that each technology could be
used to address. The evaluation was designed so that the
results from the technologies could be compared to those
of a confirmatory laboratory that analyzed each sample
using standard EPA-approved methods. The design
limited, as much as possible, those elements of sample
collection and analysis that would interfere with direct
comparison of the results. These elements included
heterogeneity of the samples and interference from other
chemicals or other controllable sources.
The design also insured that the data was collected
in a normal field environment. To do this, each
technology was operated by an operator who worked in
a trailer located at the former Koppers site. The
operators were trained by representatives from each
developer and were able to call the developers with
questions when necessary. The operators, though,
obtained all results on their own and reported the results
once they believed the results were accurate and precise.
Standard QC samples were analyzed with each
batch of environmental samples. Numerous laboratory
and field duplicate samples were analyzed to insure a
proper measure of precision. The technologies were
tested for common interferants. Qualitative measures,
such as portability and ease of operation, were noted by
the operators.
Overall, the demonstration was executed as planned
in the demonstration plan and QAPjP. The final version
of that plan was approved by all participants before the
demonstration began. Below is a discussion of selected
elements of that plan and a full discussion of deviations
from it.
Implementation
of the Demonstration Plan
For the demonstration, 98 soil samples, 14 soil
sample field duplicates, 10 water samples, and six water
sample field duplicates were collected. Each soil sample
was thoroughly homogenized and then split into replicate
samples. One replicate from each water and soil sample
was submitted to the confirmatory laboratory; three
replicates were analyzed in trailers at the former Koppers
site using the immunoassay technologies being evaluated.
In addition to these samples, two soil PE samples and
three water PE samples were analyzed by each
technology.
The final demonstration plan called for the collection
of 90 soil samples with the following distribution:
(1) 40 samples containing 0 to 100 ppm PCP, (2) 25
samples containing 100 to 1,000 ppm PCP, and (3) 25
samples containing greater than 1,000 ppm PCP. During
this demonstration, 98 soil samples were collected. The
actual distribution of these samples, when the
demonstration was complete, was as follows:
(1) 60 samples contained 0 to 100 ppm PCP, (2) 16
samples contained 100 to 1,000 ppm PCP, and (3)
22 samples contained greater than 1,000 ppm PCP. This
skewing of the sample set to the 0 to 100 ppm range
should not affect the usability of this report since the
majority of EPA PCP soil action levels occur in the 20 to
100 ppm range.
Of the samples collected for the demonstration, 53
-------
soil samples, 9 soil field duplicates, 5 water samples, and
5 water field duplicate samples were collected at the
former Koppers site. The soil samples were collected
from areas known to exhibit a wide range of PCP
concentrations. The areas sampled ranged in PCP
concentration from not detected to 3,220 ppm. Most of
the samples were collected from areas characterized
during the remedial investigation. The water samples
were collected from five existing groundwater
monitoring wells located at the former Koppers site. The
PCP concentrations in these wells were well documented
from past sampling. The PCP concentrations sampled
ranged from not detected to almost 1,500 ppb.
Of the samples collected for the demonstration, 45
soil samples, 5 soil field duplicates, 5 water samples, and
1 water duplicate sample were collected at the Winona
Post site. The soil samples were collected in areas
believed to be contaminated with high (greater than
1,000 ppm), medium (100 to 999 ppm), and low
concentrations (less than 99 ppm) of PCP. The
identification of these areas was based on past sampling
data and visual signs of waste disposal. The water
samples were collected from surface water located on or
near the site. All of the Winona Post samples were
collected, packaged, and shipped to the former Koppers
site using the methods in the demonstration plan.
Field Modifications
to the Demonstration Plan
Two field modifications were made to the approved
demonstration plan. First, fluorescein was not added to
the soil samples prior to homogenization, as specified in
the demonstration plan. The nature of the soil samples
at both the former Koppers site and the Winona Post site
allowed easy and thorough homogenization. The
saturated stiff clay matrix for which the fluorescein
additions were designed was not encountered at the
former Koppers site, and thus, for consistency, this
technique was eliminated at both sites. PRC believes
that the elimination of the fluorescein from the
homogenization process was offset by the long
homogenization times used during this demonstration.
To further examine this position, PRC conducted a
side-by-side comparison of homogenization with and
without fluorescein. Samples from the former Koppers
site were used for this comparison. Due to the dry
nature of the soil, the soil had to be hydrated with water
to allow visible distribution of the fluorescein. The
addition of the water and fluorescein caused a two-unit
increase in the soil sample pH. This alteration of the
sample chemistry coupled with the reactive nature of
PCP invalidated the fluorescein homogenization approach
for environmental applications. PRC used an
EPA-approved homogenization method and applied it to
each sample for between 10 and 15 minutes. This
method involved vigorous kneading of the sample in a
clear plastic bag.
The second modification to the approved
demonstration plan involved the sampling of the water
matrix. The change was made because the EPA
Region 7 project sponsor altered the design of the
demonstration with regard to the evaluation of the water
assays. The EPA project sponsor required that the
number of water samples collected and analyzed for the
demonstration be reduced to a total of 5 or 10. The
approved demonstration plan called for the collection of
50 groundwater samples. To maximize the usefulness of
the reduced number of water samples, PRC and
EMSL-LV agreed to combine data from both sites if they
could be shown statistically to come from the same
distribution, thereby increasing the sample set size.
Also, the EPA Region 7 project sponsor agreed to allow
the following: (1) the Region 7 Laboratory would split
and analyze samples from the Winona Post site in
sample-plus duplicate (split) pairs; (2) the excess water
from the original Winona Post site water sample that was
duplicated would be used for laboratory QA/QC; and
(3) only five monitoring wells would be sampled at the
former Koppers site, and each sample would be
duplicated. This would have resulted in five paired
(sample plus its duplicate) samples from each site. This
in turn would have provided five samples from each site
for an accuracy assessment, and five paired samples
from each site for a precision assessment. Although
these are minimal sample sizes, it was felt that this
design would provide the most useful data given the
reduction in analytical resources that the EPA Region 7
sponsor required. However, when PRC delivered the
soil samples and the water samples from the former
Koppers site, it learned that the Winona Post site water
samples had been extracted and analyzed as they were
delivered, as five environmental samples with one
duplicate. This failure to follow the modified
experimental design resulted in only four single sample
results and one duplicate result for precision analysis for
this data set. The former Koppers site data set consisted
of five duplicate pairs.
Data Collection
The technology operators prepared a subjective
evaluation of how difficult each technology was to use.
Other qualitative measures included portability,
ruggedness, instrument reliability, and health and safety
considerations. Information on these qualitative factors
was collected both by the operator of each technology
and by the project's lead chemist.
Accuracy and precision were statistically evaluated
during this demonstration. To evaluate accuracy and
precision, all samples collected for the demonstration
10
-------
were split between the technologies and the confirmatory
laboratory for analysis. The results from the
confirmatory laboratory, for the purposes of this
demonstration, were considered the actual concentration
of PCP in each sample. The cost of using each
technology also was assessed. Cost, for the purposes of
this demonstration, included expendable supplies,
nonexpendable equipment, labor, and investigation-
derived waste (IDW) disposal. These costs were tracked
during the demonstration.
Statistical Analysis of Results
For each technology, two data sets were created:
one for soil samples and the other for water samples. In
addition, each water and soil data set was composed of
two subsets, one for the samples taken from the former
Koppers site and one for those collected at the Winona
Post site. This grouping was intended to assess potential
PCP carrier effects. A third data separation involved
grouping the site-specific data sets into results greater
than 100 ppm PCP and less than 100 ppm PCP. This
grouping was intended to assess potential concentration
effects on the data analysis.
These data sets were prepared for the statistical
analysis following the methods detailed in the approved
demonstration plan. When comparing duplicate samples
or when comparing the results of a technology to those
from the confirmatory laboratory, sample pairs that
contained a nondetect were removed from the data sets.
While other statistical methods can be used when
nondetects are encountered, PRC felt that the variance
introduced by eliminating these data pairs would be less
than, or no more than equal to, the variance produced by
giving these results an arbitrary value.
Intramethod Comparisons
Sample results from each technology were compared
to their duplicate sample results and to other QA/QC
sample results. These comparisons are called
intramethod comparisons. Intramethod accuracy was
measured by assessing each technology's performance in
analyzing PE samples. If the method produced a result
considered accurate by the manufacturer that produced
the PE samples, the technology was considered to have
acceptable intramethod accuracy for this demonstration.
Intramethod precision was assessed through the statistical
analysis of relative percent differences (RPD). First, the
RPDs of the results for each sample pair, in which both
the sample and its duplicate were found to contain PCP,
were determined. The RPDs then were compared to
upper and lower control limits. When using conventional
technologies, such data is often available from analysis
of samples collected during previous investigations.
Because the technologies being demonstrated were
themselves being assessed, the control limits used were
calculated from data provided during this investigation.
To determine these control limits, the standard deviation
of the RPDs was calculated for each technology. This
standard deviation was then multiplied by two and added
to its respective mean RPDs. This established the upper
control limit for the technology. Because an RPD of
zero would mean that the duplicate samples matched
their respective samples perfectly, zero was used as the
lower control limit. This resulted in a large range of
acceptable values. Because duplicate analyses seldom
match perfectly, even for established technologies, all
samples that fell within the control limits were
considered acceptable. PRC determined that if at least
90 percent of the duplicate samples fell within these
control limits, the technology had acceptable intramethod
precision.
Intermethod Comparisons
The data sets from the technologies, also, were
statistically compared to the results from the con-
firmatory laboratory, and the precision of a technology
was statistically compared to the precision of the
confirmatory laboratory. These comparisons are called
intermethod comparisons. In both cases, the results from
the confirmatory laboratory were considered to be as
accurate and precise as analytically possible.
The statistical methods used to determine
intermethod accuracy were linear regression analysis, the
Wilcoxon Signed Ranks Test, and the Fisher's Test.
Linear regression was used for the technologies that were
capable of determining quantitative results. PRC further
prepared the data sets for the linear regression by
averaging the field duplicate results. This was done to
insure that samples were not unduly weighted in the
regression analysis. This further preparation of the data
sets was presented in the demonstration plan and agreed
to by the developers before the demonstration began.
PRC calculated linear regression by the method of
least squares. Calculating linear regression in this way
makes it possible to determine whether two sets of data
are reasonably related, and if so, how closely.
Calculating linear regression results in an equation that
can be visually expressed as a line. Three factors are
determined during calculations of linear regression.
These three factors are the y-intercept, the slope of the
line, and the correlation coefficient, also called an r2.
11
-------
All three of these factors had to have acceptable values
before a technology's accuracy was considered to meet
Level 3 data quality requirements.
The r2 expresses the mathematical relationship
between two data sets. If the r2 is one, then the two data
sets are directly related. Lower r2 values indicate less of
a relationship. Because of the heterogeneous nature of
environmental samples, r2 values between 0.85 and 1
were considered to meet data quality Level 3 accuracy
requirements; r2 values between 0.75 and 0.85 were
considered to meet data quality Level 2 accuracy
requirements; and r2 values below 0.75 were considered
not accurate, meeting, at best, Level 1 accuracy
requirements. The classification of data as Levels 2 or
1 was implied in the approved demonstration plan;
however, these specific criteria were not presented.
If the regression analysis resulted in an r2 between
0.85 and 1, then the regression line's y-intercept and
slope were examined to determine how closely the two
data sets matched. A slope of one and a y-intercept of
zero would mean that the results of the technology
matched those of the confirmatory laboratory perfectly.
Theoretically, the farther the slope and y-intercept differ
from these expected values, the less accurate the
technology. Still, a slope or y-intercept can differ
slightly from their expected values without that
difference being statistically significant. To determine
whether such differences were statistically significant,
PRC used the normal deviate test statistic. This test
statistic results in a value that is compared to a table.
The value at the 90 percent confidence level was used for
the comparison. To meet data quality Level 3
requirements, both the slope and y-intercept had to be
statistically the same as their ideal values.
If the r2 was between 0.75 and 1, and one or both of
the other two regression parameters were not equal to
their ideal, the technology was considered inaccurate but
producing Level 2 quality data. Results in this case
could be mathematically corrected if 10 to 20 percent of
the samples were sent to a confirmatory laboratory.
Analysis of a percentage of the samples by a
confirmatory laboratory would provide a basis for
determining a correction factor. Only in cases where the
r2, the y-intercept, and the slope were all found to be
acceptable did PRC determine that the technology was
accurate, meeting Level 3 data quality requirements.
Data placed in the Level 1 category had r2 values
less than 0.75, the data was not statistically similar to the
confirmatory data, based on parametric testing, or the
results did not meet the manufacturer's performance
specifications.
A second statistical method used to assess the
intermethod accuracy of the data from each technology
was the Wilcoxon Signed Ranks Test. This test is a
nonparametric method for comparing matched pairs of
data. It can be used to evaluate whether two sets of data
are significantly different. The test requires no
assumption regarding the population distribution of the
two sets of data being evaluated other than that the
distributions will occur identically. In other words,
when one data point deviates, its respective point in the
other set of data will deviate similarly. Because the only
deviation expected during the demonstration was a
difference in the concentrations reported by each
technology, the two sets of data were expected to deviate
in the same way. The calculation performed in the
Wilcoxon Signed Ranks Test uses the number of samples
analyzed and a ranking of the difference between the
result obtained from a technology and the corresponding
result from the confirmatory laboratory. The rankings
can be compared to predetermined values on a standard
Wilcoxon distribution table, which indicates whether,
overall, the two methods have produced similar results.
Two of the field screening technologies, those
produced by EnSys and by Millipore, produced semi-
quantitative results. These technologies produce results
that indicate only whether a sample contains PCP above
or below a predetermined range. Semiquantitative
technologies, by definition, cannot produce Level 3 data.
Linear regression analysis and the Wilcoxon Signed
Ranks Test cannot be used to compare semiquantitative
results. Instead, PRC used a 2 by 2 contingency table and
a Fisher's Test. The Fisher's Test determines whether
both data sets are correlated. When used in a two-tailed
manner, as it was in this case, its formula is usually
conservative. Therefore, use of a modified Chi-square
formula is recommended (Pearson and Hartley 1976).
The Fisher's Test statistics were compared to the 90
percent significance level obtained from a standard
Chi-square distribution table. This comparison indicated
whether, overall, there was a correlation between the
results of the two methods. If a correlation existed, the
technology was considered accurate, capable of
producing Level 2 data.
Finally, the precision of each quantitative technology
was statistically compared to the precision of the
confirmatory laboratory using Dunnett's Test. This test
was used to assess whether the precision of the
technology and that of the confirmatory laboratory were
statistically equivalent. First, the mean RPD for all
samples and their respective duplicates analyzed by the
confirmatory laboratory was determined. The RPD of
each duplicate pair analyzed by each of the technologies
was then statistically compared to this mean. It should
be noted that a Dunnett's result showing the precisions
are not similar does not mean that the precision of the
12
-------
technology was not acceptable, only that it was different of the Burnett's results was provided by the Wilcoxon
from the precision of the confirmatory laboratory. In Signed Ranks Test.
particular, Dunnett's Test has no way of determining
whether or not any difference between the two data sets Overall, for this demonstration, the determination of
actually resulted because a technology's data was more significance for inferential statistics was set at 90
precise than the confirmatory laboratory's. Verification percent. However, regression analysis was considered
to show a significant relationship if the r2 was greater
than 0.85 for Level 3 data and between 0.75 and 0.85 for
Level 2 data.
13
-------
Section 5
Confirmatory Analysis Results
All samples collected during this demonstration were
submitted to the EPA Region 7 Laboratory for
confirmatory analysis. The water samples were analyzed
by the EPA Region 7 Laboratory under the ESAT
Contract, and the soil samples were analyzed under the
RECAP Contract. The EPA Region 7 Laboratory
assigned sample numbers to each sample submitted for
analysis. In this manner, analyst bias was eliminated
during analysis of the samples. The result for each
sample is presented within Sections 6 and 7.
Confirmatory Laboratory Procedures
EPA Region 7 Laboratory Quality Assurance and
Data Evaluation (QADE) Branch personnel conducted a
Level II data review on the results provided by the
confirmatory laboratory. A Level II data review does
not include an evaluation of the raw data or a check of
calculated sample values. A review of the raw data and
a check of the calculations was performed by QC
personnel from the confirmatory laboratory before the
data package was submitted to the EPA Region 7
Laboratory QADE Branch. PRC was not able to review
the raw data generated. However, PRC did review the
laboratory case narratives and the EPA Region 7
Laboratory QADE Branch comments generated by the
Level II data review.
The following sections discuss specific procedures
used to identify and quantitate semivolatile organic
compounds (SVOC), and specifically PCP, using the
following methods: SW-846 Method 8270A (soil and
water), SW-846 Method 8151A (soil), and EPA Method
515.1 (water).
Sample Holding Times
All of the analytical methods used for confirmatory
analysis require that all sample extractions be completed
within 7 days from the time a sample was collected. Due
to the stability of PCP, EPA's ORD Methods Validation
Section extended these holding time requirements by 4
days for this demonstration. The analysis of the sample
extracts must be completed within 40 days of sample
receipt. The holding time requirements for the
demonstration's samples were met.
Sample Extraction
The method used for the extraction of soil samples
prior to analysis by EPA Method 8270A was EPA
Method 3550. This method involves sonication
extraction of the soil using methylene chloride. The
confirmatory laboratory used both the low concentration
extraction method and the high concentration extraction
method discussed in the method. To determine the
appropriate extraction method to use for the analysis of
individual soil samples, the confirmatory laboratory
screened each sample using the screening techniques
recommended in EPA Method 8270A. EPA Method
3510A was used for the extraction of water samples and
involves a separatory extraction of the water with
methylene chloride. To ensure that phenolic compounds,
such as PCP, will be adequately extracted from the water
samples, two extractions of each water sample were
performed. The pH of the water was adjusted to greater
than 12 and extracted, then the pH of the water sample
was adjusted to below 2 and extracted. The two sample
extracts were then combined for sample analysis.
The low-level detection analytical methods for PCP
included different procedures for sample extraction. The
method used for the soil samples, EPA Method 8151A,
involved an acidification of the soil sample, followed by
an ultrasonic extraction with methylene chloride. This
extraction is similar to the EPA Method 3550 sonication
extraction. The soil sample extract was then taken
through an acid-base partition. The acid-base partition
was used to remove potentially interfering compounds
from the sample extract. The sample extract was then
concentrated and taken through a diazomethane deri-
vatization. This procedure replaces the hydrogen atom
of the phenolic hydroxide group with a methyl anion.
This derivatization removes the polarity associated with
PCP and enables improved chromatographic behavior.
PCP standards used for sample identification and
14
-------
quantitation were taken through the same derivatization
steps as samples to allow a direct comparison of
concentration. That is, no correction factor needs to be
used for the molecular weight of the derivatization
product.
The low-level detection analytical method used for
water samples, EPA Method 515.1, involved a
separatory extraction of the water sample with methylene
chloride. A pH adjustment of the water samples was
performed similar to the pH adjustment used for the
water samples extracted with EPA Method 3510A. In
EPA Method 515.1, the solvent extract from the basic
extraction is discarded because it contains no PCP. This
step also removes potential interferences. The water
sample extract is then concentrated and derivatized in the
same manner as the soil sample extracts. Again, this
derivatization removes the polarity associated with PCP
and provides improved chromatographic behavior. PCP
standards used for sample identification and quantitation
were taken through the same derivatization steps as the
samples to allow a direct comparison of concentration.
Reporting Limits and Initial
and Continuing Calibrations
The reporting limit for soil samples analyzed by
EPA Method 8270A was 0.330 ppm. The reporting limit
for soil samples analyzed by EPA Method 8151A was
0.076 ppb. The reporting limit for water samples
analyzed by EPA Method 8270A was 50 ppb. The
reporting limit for water samples analyzed by EPA
Method 515.1 was 0.076 ppb. Method-required initial
and continuing calibration procedures were appropriately
conducted, and all method-required criteria for these
calibrations were met.
Sample Analysis
The confirmatory laboratory performed sample
analysis by first analyzing samples using EPA Method
8270A. Based upon the screening results, the samples
were extracted with either the low concentration method
or the high concentration method. Samples which did
not provide a positive response for PCP with EPA
Method 8270A were analyzed by one of two low-level
detection methods, EPA Method 8151A for soil samples
and EPA Method 515.1 for water samples.
EPA Method 8270A uses gas chromatography for
compound separation and a mass spectrometer for
identification and quantification of PCP. EPA Method
8151A uses gas chromatography from compound
separation and an electron capture detector (BCD) for
identification and quantification of PCP. The
chromatographic column used for EPA Methods 8270A
and 8151A can vary and appropriate columns can be
found in the methods.
For EPA Method 8270A, compound identification
was required to meet two criteria: (1) the sample
component relative retention time was to fall within +
0.06 relative retention time units of the standard
component, and (2) the mass spectrum of the sample
compound was to correspond with the standard
compound mass spectrum.
Soil and water samples, which were found to contain
no PCP during the EPA Method 8270A analysis, were
analyzed using the EPA Method 8151A and 515.1,
respectively. PCP identification was made if a sample
peak eluted within the retention time window established
during the initial calibration.
Quality Control Procedures
Method blanks are used to monitor the presence of
laboratory-induced contamination. The EPA Region 7
Sample Management Office (SMO) provided blank soil
and blank water samples for use as method blank samples
during the analysis of demonstration samples. An
acceptable method blank must not provide a positive
response for the target compounds above the reported
detection limit. Method blank samples were stored,
extracted, and analyzed in exactly the same manner as
the demonstration samples. Results for all method blank
samples extracted and analyzed along with the
demonstration samples were found to be acceptable.
Internal standards were used for the analysis of
demonstration samples by EPA Method 8270A. Internal
standards were added to all standards, blanks, samples,
and QC samples prior to injection into the GC/MS
system. The internal standards were used to provide
response factors for each of the target compounds.
During the analysis of soil samples, seven samples
exhibited internal standard responses which were outside
of the QC limits of 50 to 150 percent recovery. All of
the affected samples provided internal standard responses
which were less than 50 percent. The soil samples
affected were samples 038, 060, 062, 068, 090, 091,
095. Of these samples, three - 090, 091, and 095 -
were found to contain no detectable levels of PCP, and
no corrective action was taken. Instead, they were
reanalyzed using EPA Method 8151A. The remaining
samples were reanalyzed to verify that the internal
standard response was below 50 percent recovery. The
reanalysis showed that internal standard response was
below 50 percent recovery. No corrective action was
taken by the laboratory, which attributed the low
recovery to matrix effects inherent to the samples. In the
Region 7 Laboratory QADE Branch review of the data,
the same conclusion was reached.
15
-------
Surrogate standards were used to evaluate the
efficiency of the extraction and analysis processes and to
evaluate matrix effects. Surrogate standards used for
EPA Method 8270A include deuterated standards, which
provide a different mass spectrum when compared to the
nondeuterated compound. Surrogate standard recoveries
for the soil samples all fell within the acceptance ranges.
The data review performed by the Region 7 QADE
Branch indicated that surrogate recoveries for some of
the water samples were outside of the acceptance ranges,
but no information indicated which samples or how many
samples fell outside surrogate recovery acceptance
ranges. Corrective action was not taken because the
acceptance ranges listed in the method are for advisory
purposes only. The surrogate standard used for EPA
Method 8151A and EPA Method 515.1 was 2,4-dichlor-
ophenylacetic acid (DCAA). The acceptance range for
DC A A was determined by the RECAP and Region 7
Laboratory through a statistical analysis of 30 or more
standard surrogate recoveries. The mean and standard
deviation were then calculated, and the acceptance range
was determined by applying a + 3 standard deviations
around the mean. All samples analyzed with EPA
Method 8151A and EPA Method 515.1 provided
surrogate recoveries which fell within the laboratory-
generated control limits.
Matrix spike samples were aliquots of original
samples into which a known concentration of the target
compounds was added. The EPA Region 7
Environmental Services Division specified which
samples were to be used as confirmatory laboratory
matrix spike samples. The specified soil samples were
samples 036, 048, 053, 073, 087, and 098, all analyzed
using EPA Method 8270A, and sample 089, analyzed
using EPA Method 8151A. The specified water samples
were samples 101 and 111, analyzed using EPA Method
515.1, and sample 104, analyzed using EPA Method
8270A. The soil matrix spike samples analyzed using
EPA Method 8270A were spiked with all of the target
compounds reported by the method. Water sample
matrix spike samples analyzed using EPA Method 8270A
were spiked with nine of the target compounds reported
by the method. Matrix spike samples analyzed with EPA
Methods 8151A and 515.1 were only spiked with PCP.
Soil matrix spike data for PCP is shown in Table 5-1;
water matrix spike data for PCP is shown in Table 5-2.
It should be noted that the confirmatory laboratory does
not reprepare or reanalyze matrix spike samples.
The soil sample matrix spike recoveries were greatly
influenced by the high concentrations of PCP present in
the original sample relative to the amount spiked. Only
one sample, 098, resulted in recoveries for both the
matrix spike and matrix spike duplicate sample which
could be considered acceptable. A clear evaluation of
the effects of matrix on PCP recovery is not possible due
to the high concentrations of PCP in the original sample
and the comparatively low levels of PCP added to the
matrix spike samples. The water sample matrix spike
sample analyzed using EPA Method 8270A resulted in
high recoveries. These recoveries are on the high end of
the QC acceptance criteria for PCP recoveries listed in
EPA Method 8270A (14 to 176 percent recovery).
However, the agreement between the matrix spike and
matrix spike duplicate was excellent as determined by the
RPD of the matrix spike recoveries. The water matrix
spike samples analyzed using EPA Method 515.1 were
affected by the concentration of PCP in the original
sample. Although the matrix spike recoveries for sample
101 were found to be acceptable, the recoveries of PCP
spiked into the sample were affected by the much larger
concentration of PCP in the original sample. Sample 111
also was affected by the concentration of PCP in the
original sample. The results of both the matrix spike
sample and the matrix spike duplicate sample were less
than the result for the original sample. This may indicate
a heterogeneity problem with the sample. The low levels
of PCP added to this sample were not enough to obtain
an accurate indication of matrix spike recovery.
Blank spike samples were prepared by the EPA
Region 7 Laboratory SMO for the water sample analysis
performed by EPA Methods 8270A and 515.1. These
samples are used to evaluate the accuracy of the
laboratory. The blank spike samples were stored,
extracted, and analyzed in the same manner as all other
samples. The percent recoveries of the blank spike
samples fell within the 14 to 176 percent QC acceptance
criteria listed in EPA Method 8270A and the 67.6 to
192.4 percent acceptance criteria listed in EPA Method
515.1. The accuracy of the analysis of water samples
using EPA Methods 8270A and 515.1 was found to be
acceptable, based on the blank spike sample results.
Data Reporting
The data report PRC received from the EPA Region
7 Laboratory included a standard EPA Region 7 Analysis
Request Report. Results were reported on a dry-weight
basis, as required in the methods. PRC obtained data on
the loss-on-drying determination for each of the samples.
The loss-on-drying values were used to convert the
16
-------
TABLE 5-1.
Sample
No.
036
048
053
073
087
089
098
TABLE 5-2.
Sample
No.
101
104
111
SOIL MATRIX SPIKE SAMPLE RESULTS FOR EPA METHODS 8270A AND 8151 A
Amount
Found in
Original
Sample
(ppm)
40.0
30,000
2.30
86.0
46.0
0.247
0.70
Amount
Added to
Matrix Spike
Sample
Duplicate
(ppm)
1.90
46.0
1.50
11.0
1.40
0.098
0.41
Amount
Found in
Matrix Spike
Sample
(ppm)
66.0
22,000
12.0
130
57.0
0.315
0.82
WATER MATRIX SPIKE SAMPLE RESULTS
Amount
Found in
Original
Sample
(ppb)
4.14
50.0 U
1.85
Amount
Added to
Matrix
Spike
Sample and
Duplicate
(ppb)
0.446
200
0.398
Amount
Found in
Matrix Spike
Sample
(ppb)
4.46
348
1.55
Percent
Recovery
(%)
1,370
0
647
400
786
69
29
FOR EPA
Percent
Recovery
(%)
72
174
0
Amount
Found in
Matrix
Spike
Duplicate
Sample
(ppm)
51.0
24,000
5.20
93.0
64.0
0.241
0.98
METHODS 8270A
Amount
Found in
Matrix
Spike
Duplicate
Sample
(ppb)
4.24
353
1.64
Duplicate's
Percent
Recovery
(%)
579
0
193
64
1,285
0
68
AND 515.1
Duplicate's
Percent
Recovery
(%)
22
177
0
Relative
Percent
Difference
(%)
81
0
108
145
48
200
80
Relative
Percent
Difference
(%)
106
2
0
Note:
U Not detected above detection limit.
confirmatory laboratory data from a dry-weight basis to
a wet-weight basis.
Results were reported by the confirmatory
laboratory in micrograms per kilogram (/xg/kg) for soil
samples and micrograms per liter (/xg/L) for water
samples. Soil sample results were converted to
milligrams per kilogram (microgram/kg) so they could
be compared to the results from the technologies, all of
which reported results for soil samples in microgram/kg.
The results from the technologies for water samples were
reported in //g/L, so no conversion of the confirmatory
laboratory data was needed.
Data Quality Assessment
Accuracy refers to the difference between the
sample result and the true concentration of compound in
the sample. Bias, a measure of the departure from
complete accuracy, can be caused by such processes as
loss of compound during the extraction process,
interferences, and systematic contamination or carryover
of a compound from one sample to the next. Accuracy
for the confirmatory laboratory was assessed through the
use of PE samples. Four of the PE samples used for this
demonstration were purchased from Environmental
Research Associates (ERA). Two of these PE samples
were soil and two were water. These samples contained
a known quantity of PCP. ERA supplied data sheets for
each PE sample which included the true concentration
and an acceptance range for the sample. The acceptance
range was based on the 95 percent confidence interval
taken from data generated by ERA and EPA
interlaboratory studies. A third water PE sample was
prepared by the PRC lead chemist to widen the range
covered by the PE samples.
The PE samples contained different concentrations
of PCP. These samples were extracted and analyzed in
the same manner as the other water and soil samples.
The confirmatory laboratory did not know which samples
were PE samples or the certified values and acceptance
ranges. The true value concentration of soil PE sample
099 (the low-level sample) was 7.44 ppm with an
acceptance range of 1.1 to 13 ppm. The result reported
by the confirmatory laboratory for this sample was 4.02
ppm, which was within the acceptance range. The
percent recovery of this sample by the confirmatory
laboratory was 54 percent. The true concentration of soil
PE sample 100 (the high-level sample) was 101 ppm with
an acceptance range of 15 to 177 ppm. The result
reported for this sample by the confirmatory laboratory
was 52.4 ppm, which was within the acceptance range.
The percent recovery of this sample by the confirmatory
laboratory was 52 percent.
17
-------
The true value concentration of water PE sample
106 (the low-level sample) was 68.4 ppb with an
acceptance range of 10 to 120 ppb. The result reported
by the confirmatory laboratory for this sample was 10.3
ppb, which was within the acceptance range. The
percent recovery of this sample by the confirmatory
laboratory was 15 percent. The true concentration of
water PE sample 107 (the high-level sample) was 2,510
ppb with an acceptance range of 377 to 4,420 ppb. The
result reported for this sample by the confirmatory
laboratory was 2,050 ppb of PCP, which was within the
acceptance range. The percent recovery of this sample
by the confirmatory laboratory was 82 percent. The true
value concentration of water PE sample 113 (the PE
sample prepared by PRC) was 7.50 ppb. No acceptance
range was statistically determined for this PE sample.
Instead, PRC established a 30 to 170 percent window of
acceptable values around the true value result of the
low-level PE sample. This window is consistent with
both the acceptance ranges of the PE samples prepared
by ERA and the QC Acceptance Criteria for PCP
recovery stated in EPA Method 8270A. The
confirmatory laboratory result for this PE sample was
within the acceptance range. Based on the results for all
of the PE samples, the accuracy of the confirmatory
laboratory was acceptable.
Precision refers to the degree of mutual agreement
between individual measurements and provides an
estimate of random error. Precision for the confirmatory
laboratory results was determined through the use of
field duplicate samples. Normally laboratory duplicates
are used for this. However, no laboratory duplicates
were analyzed by the confirmatory laboratory. Field
duplicates are two samples collected together, but
delivered to the laboratory with separate sample
numbers. Typically, field duplicate samples are used to
measure both sampling and analysis error. PRC
established control limits for field duplicate RPDs.
These control limits are similar to those used to
determine matrix spike recovery acceptance control
limits. To establish the control limits, all sample pairs
that did not produce two positive results were removed
from the data set. Then the RPD for each pair was
calculated, and the mean RPD and standard deviation
were determined. The lower control limit was set at
zero because this would mean that the results from a
duplicate and its sample matched perfectly. The upper
control limit was set by multiplying the standard
deviation by two and adding it to the mean RPD. The
RPD of each sample pair was then compared to these
control limits. Each sample pair RPD was expected to
fall within the control limits.
Fourteen soil field duplicate samples were collected
and analyzed by the confirmatory laboratory during this
demonstration. Field duplicate samples represented 17
percent of all soil samples collected and analyzed. The
original results ranged from 0.10 to 26,100 ppm. The
duplicate sample results ranged from 0.09 to
30,260 ppm. RPD values for the soil field duplicate
pairs ranged from 1 to 168 RPD. The mean RPD value
of the soil field duplicate pairs was 33 percent, with a
standard deviation of 47 percent. For the soil field
duplicate pairs, the control limits were found to be 0 to
128 RPD. Thirteen of the fourteen, or 93 percent, of the
field duplicate sample pairs fell within this range.
Six water field duplicate samples were collected and
analyzed by the confirmatory laboratory during this
demonstration. Field duplicate samples represented 32
percent of all water samples collected and analyzed. The
original results ranged from 0.175 to 1,810 ppb. The
field duplicate sample results ranged from 0.63 to
2,020 ppb. RPD values for the water field duplicate
pairs ranged from 0 to 113 RPD. The mean RPD value
of the water field duplicate pairs was 30 percent with a
standard deviation of 41 percent. For the water field
duplicate pairs, the control limits were found to be 0 to
112 RPD. Five of the six, or 83 percent of the field
duplicate sample pairs, fell within this range.
Overall, this data shows excellent agreement
between the samples and their respective field duplicates,
indicating a high degree of precision by the confirmatory
laboratory. The mean RPD also indicated that the
method used to homogenize the samples before splitting
them for analysis was highly effective.
Completeness refers to the amount of data collected
from a measurement process compared to the amount
that was expected to be obtained (Stanley and Verner
1983). The completeness objective for this project was
95 percent. This demonstration resulted in the analysis
of 98 soil samples, 14 soil sample duplicates, 2 soil PE
samples, 10 water samples, 6 water sample duplicates,
and 3 water PE samples. Results were obtained for all
of these samples. Completeness for the confirmatory
laboratory was 100 percent.
Confirmatory Laboratory Costs
and Turnaround Times
The cost for performing PCP analysis by
EPA-approved analytical methods varies from laboratory
to laboratory. The cost of analysis depends upon the
18
-------
number of samples submitted for analysis, the matrix, for samples submitted for analysis with EPA-approved
and the level of QC performed. The following costs are analytical methods range from 14 to 30 days. The
given as general guidelines. EPA Method 8270A turnaround time also depends upon the number of
analysis costs range from $250 to $400 per sample. EPA samples submitted for analysis, the matrix, and the level
Method 8151A analysis costs range from $150 to $250 of QC performed. Faster turnaround times may be
per sample. EPA Method 515.1 analysis costs range available for an additional cost.
from $125 to $200 per sample. Turnaround times
19
-------
Section 6
EnSyslnc.: Penta RISc Test System
This section provides information on the Penta RISc
Test System including background information,
operational characteristics, performance factors, a data
quality assessment, and a comparison of its results with
those of the confirmatory laboratory.
Theory of Operation and
Background Information
The system is designed to provide quick,
semiquantitative results for PCP concentrations in soil
and water samples (see Exhibit 6-1). The system is
composed of a soil test kit and a water test kit. Soil and
water samples must be analyzed separately and compared
to different PCP calibrators. The system uses the
principles of enzyme-linked immunosorbent assays
(ELISA). ELISA-based analytical technologies use
antibodies to provide compound-specific reaction,
detection, and quantitation. These antibodies are
produced by injecting animals, usually rabbits, with
either a measured amount of PCP or a PCP analog. The
PCP or analog also are known as immunogens. The
animal's immune system produces an antibody specific
to the PCP or analog in much the same manner that the
human immune system produces antibodies to fight
infection by common viral and bacterial pathogens.
Booster injections of the immunogen are continued until
a maximum antibody-binding response is attained. The
antibodies are then collected and separated from the
animal's blood for use in the manufacture of the
immunoassay kit. The developer covalently binds the
antibodies to the inside walls of test tubes for use in its
test system.
In addition to the antibodies, ELISA-based
technologies generally use an enzyme conjugate in the
analysis step of the immunoassay test. The enzyme
conjugate is formed by covalently binding a PCP analog
to a horseradish peroxidase enzyme. This enzyme
conjugate competes with PCP in an environmental
sample for antibody binding sites on the walls of the test
tube. The enzyme conjugate provides the means for
identification and quantitation of PCP. ELISA-based
technologies use chromogenic reagents that react
specifically with the enzyme conjugate. These results
allow identification and quantitation of PCP.
In this technology, a mixture of two substrates reacts
with the enzyme conjugate to produce a brilliant blue
color. Because an exact number of antibody binding
sites are available on each test tube and an exact number
of enzyme conjugate molecules are introduced into each
sample, the only variable which exists is the number of
PCP molecules present in the environmental sample.
The PCP in the sample will compete with the enzyme
conjugate molecules for antibody binding sites, thus
reducing the amount of blue color formed by the reaction
of the substrate with the enzyme conjugate molecules. It
is the amount of blue color formed by the sample that is
used for identification and quantitation of PCP in the
environmental sample.
Semiquantitative interpretation is performed by
comparing differential photometer readings of the sample
to differential photometer readings of a PCP standard
taken through the same immunoassay steps required for
the samples. If the differential photometer reading of a
sample is zero, or a negative value, the sample contains
close to or more than the detection level of interest. If
the differential photometer reading of a sample is
positive value, the sample contains less than the detection
level of interest. The use of three sample dilutions
allows for a closer approximation of PCP concentrations
in samples.
The system is designed to detect PCP, but other
compounds may respond to the system. This is referred
to as cross reactivity and is measured by determining at
what concentration a specific compound will yield a
positive result using the system. Another phenomenon
that may affect immunoassay tests is masking cross
reactivity, which is caused by cross reacting compounds
inhibiting or obscuring the reaction of the target
compound with the antibodies. The developer has
evaluated a number of compounds and has provided
20
-------
Exhibit 6-1. The processes for the EnSys PCP test kits.
information on cross reactivity. Table 6-1 presents this
positive value, the sample contains less than the detection
level of interest. The use of three sample dilutions
allows for a closer approximation of PCP concentrations
in samples.
The system is designed to detect PCP, but other
compounds may respond to the system. This is referred
to as cross reactivity and is measured by determining at
what concentration a specific compound will yield a
positive result using the system. Another phenomenon
that may affect immunoassay tests is masking cross
reactivity, which is caused by cross reacting compounds
inhibiting or obscuring the reaction of the target
compound with the antibodies. The developer has
evaluated a number of compounds and has provided
information on cross reactivity. Table 6-1 presents this
a information. As part of this demonstration, a
21
-------
specificity study was conducted that concerned the test
system's reaction to diesel fuel, which is a common
carrier for PCP, and to chemicals with molecular
structures similar to PCP. Data from the specificity
study is presented later in this report.
TABLE 6-1. CHEMICAL CROSS REACTIVITY
AS REPORTED BY ENSYS
Compound
PCP
Phenol
4-Chlorophenol
2,4-Dichlorophenol
2,6-Dichlorophenol
3,5-Dichlorophenol
2,3,4-Trichlorophenol
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
Tetrachlorophenol
Tetrachlorohydroquinone
Pentachlorobenzene
Aroclor 1254
Concentration
Required for
a Positive
Interpretation
by Soil Test
System
(ppm)
0.5
> 1 ,000
> 1 ,000
> 1 ,000
700
NA
400
100
16
1.2
500
> 1 ,000
> 1 ,000
Concentration
Required for
a Positive
Interpretation
by Water Test
System
(ppb)
5
>600
>800
> 1 ,000
600
> 1 ,000
600
500
100
7
> 1 ,500
> 1 ,400
> 100
Operational Characteristics
Instrumentation required for the system is shipped in
a small suitcase-sized, plastic container. This container
has a handle and is very easy to carry. The reagents and
supplies required for the system are shipped in large
plastic boxes. The containers and boxes needed for
analysis of 200 samples would fit in the trunk of a large
car or in the back seat of a smaller car for transportation
to close proximity sites. The containers and boxes can
easily be shipped by commercial carriers to sites farther
away. During this demonstration, all containers and
boxes were shipped to the site directly from the
developer.
The system is composed of two test kits: the Penta
RISc Soil Test System and the Penta RISc Test System,
used for water samples. An accessory kit, containing
instrumentation, also is required for the analysis of soil
and water samples. The Penta RISc Soil Test System
includes the following: (1) four sample extraction jars
with screw caps, each with stainless-steel ball bearings
and premeasured extraction solvent, (2) four weigh
boats, (3) four wooden spatulas, (4) four bulb transfer
pipettes, (5) four filtration devices composed of four
filtration barrels and four filtration plungers, (6) four
mechanical pipette pistons and four mechanical pipette
capillaries, (7) four sample dilution vials labeled 0.5
ppm, (8) four sample dilution vials labeled 5.0 ppm, (9)
four sample dilution vials labeled 50.0 ppm, (10) twelve
buffer tubes containing premeasured diluent, (11) eight
tubes containing standard, (12) twenty antibody-coated
test tubes in an aluminized pouch, (13) one bottle of
wash buffer solution, (14) one bottle of Substrate A, (15)
one bottle of Substrate B, (16) one bottle of stop
solution, and (17) one instruction booklet.
The Penta RISc Test System (for water) includes:
(1) four filtration devices composed of four filtration
barrels and four filtration plungers, (2) four bulb transfer
pipettes, (3) one capillary pipette p lunger, (4) four
capillary pipettes, (5) twelve sample dilution tubes
containing premeasured diluent, (6) eight tubes
containing standard, (7) twelve antibody-coated test
tubes, (8) one bottle of enzyme conjugate solution, (9)
one bottle of wash buffer solution, (10) one bottle of
Substrate A, (11) one bottle of Substrate B, (12) one
bottle of stop solution, and (13) one instruction booklet.
The accessory kit includes the following: (1) a
Pocket Pro Series electronic balance, (2) an Artel DP
differential photometer, (3) a Model P-250 Poppette
Micropipettor, adjustable from 5 to 250 microliters, and
(4) a Westbend electronic timer.
Other equipment that is helpful when using the
system, which is not supplied by the developer includes
protective gloves, laboratory tissue, a permanent
marking pen, paper towels, and liquid and solid waste
containers.
For this demonstration, the system was operated in
a 28-foot trailer located at the former Koppers site.
Electricity was supplied to the trailer for the use of air
conditioning and to provide lighting. A refrigerator was
required to store the enzyme conjugate solution used with
the water test kit. The enzyme conjugate supplied with
the soil test kit is in a pelletized form bonded within a
sucrose tablet and is more stable than the liquid form of
the enzyme conjugate. For this reason, the developer
states that the soil test kit does not require refrigeration.
The developer recommends storage of the soil test kit at
room temperature and recommends that these
components should not be exposed to temperatures below
0 °C or above 37 °C. The developer recommends
storage of the water test kit components at 4 to 8 °C and
that reagents should not be stored at or below 0 °C or
above 37 °C. Electricity to operate the differential
photometer during this demonstration was supplied by
the photometer's rechargeable battery. This battery
requires 8 to 24 hours of charging and, when fully
charged, can perform up to 500 tests without being
recharged. The battery was recharged nightly by the
electricity in the trailer. Electricity required is a
110-volt circuit. A 3-foot by 2-foot hood was used by
the operator for performing soil sample extractions. A
4-foot table was used to perform the assay steps, sample
22
-------
dilutions, sample analysis using the differential
photometer, and logbook entries and other data
documentation.
The operator chosen for analyzing samples using the
technology was Mr. Frank Douglas. Mr. Douglas is an
employee of PRC. He earned a Bachelor of Science
degree in journalism with a minor in science, a Master
of Arts degree in English. While at PRC, Mr. Douglas
has worked as an editor in PRC's QC system. Mr.
Douglas also has worked as a technician on a similar
SITE demonstration performed by PRC. In this
capacity, he helped explain how immunoassay systems
work and helped interpret the statistics that resulted from
the test.
Mr. Douglas's training in the use of the test system
included a review of the information provided by the
developer before the start of the demonstration. Mr.
Douglas also received approximately 6 hours of training
at the start of the demonstration by Ms. Rhonda Mudd of
EnSys. This training included step-by-step procedures
for extracting, preparing, and analyzing soil and water
samples using the system, and instructions on use of the
pipettes and instrumentation. Interpretation of sample
results was discussed as well as QC requirements. In
addition, Mr. Douglas analyzed soil and water samples
using the system under the supervision of Ms. Mudd.
Mr. Douglas noted that after he had completed the
analysis of both soil and water samples, Ms. Mudd felt
confident that he was ready to properly operate the
system.
The developer states that this product is intended for
use by environmental professionals and requires a
minimum of training. Mr. Douglas had no prior
experience performing immunoassay testing. Mr.
Douglas noted that he found the system very easy to
operate. Mr. Douglas noted that the developer has
dramatically increased the ease of operation for the
system by limiting the amount of pipetting and measuring
necessary. He also noted that the developer has
premeasured the extraction liquid, dilution liquid, buffer
liquid, and standards. The developer also color codes
the reagents and numbers the reagents used in the water
test kit.
The technology is promoted for use as a portable
field system. However, it does require special care and
handling in the field to avoid damage. The differential
photometer requires care during use, shipping, and
transportation to avoid breaking or damaging internal
components. No mechanical or electronic problems
were experienced with this equipment during the course
of the demonstration.
Instrument reliability for the system was to be
evaluated by monitoring the standard calibration
responses. QC criteria for the calibration responses
included the analysis of a duplicate standard with each
standard prepared. The standard which exhibited the
higher optical density reading was to be used for sample
comparison, and the optical density readings between a
standard and its duplicate were not to vary by more than
0.35 optical density units. For the soil system, 46 pairs
of standards were prepared. Only three times did the
optical density readings between a standard and its
duplicate vary by more than 0.35 optical density units.
In these cases, all samples analyzed with these pairs were
reanalyzed unless their optical density readings were
greater than the optical density of the high-level
standard. The developer recommended to PRC that
samples which gave optical density readings that were
greater than the high-level standard did not have to be
reanalyzed because, even when reanalyzed with another
acceptable calibration, these samples would still give the
same result. For the water system, 37 standard pairs
were prepared. None of these pairs produced optical
density readings which varied by more than 0.35 optical
density units.
Two times during the analysis of soil samples the
operator of the system noted that a sample turned bright
orange upon the addition of the stop solution and
produced an extremely high optical density reading. The
operator contacted the developer with this problem.
EnSys personnel commented that they had seen this
problem before and that the sample needed to be
reanalyzed. Ms. Mudd referred to this phenomena as a
"high flyer" and speculated that it may be attributed to a
higher than normal amount of enzyme conjugate in the
pellet. These two samples were reanalyzed. The
operator of the system also noted that two samples
provided inconsistent results. In both cases, the
technology reported results less than 5 mg/kg but greater
than 50 mg/kg. These samples were reanalyzed, and
acceptable results were obtained from the reanalysis.
Immunoassays are enzymatic reactions and can be
effected by changes in ambient temperatures. The
technology seemed to be affected by the temperature in
the trailer during this demonstration. The operator of the
system noted several times during the course of the
demonstration that analysis problems occurred most
frequently in the afternoon. Although PRC did not
record temperatures in the trailer during the
demonstration, the technology operators noticed a
temperature increase in the trailer during the afternoon.
One operator noted that he believed the temperature in
the trailer increased 5 to 10 °F in the afternoon. This
was noticed even while the trailer was air conditioned.
Although this problem seemed to affect standards and
QC criteria, the problems associated with the
temperature increase were found to have an insignificant
23
-------
effect on the results obtained from the system.
However, PRC believes that use of the test system is
improved when used in a temperature-controlled
environment. This will help minimize problems
associated with extremes of temperature or humidity
which may affect the immunoassay chemistry.
The test system contains various chemicals in
quantities ranging from parts per million levels to 500
milliliters. The more dangerous chemicals include
methanol, a flammable solvent and poison, and sulfuric
acid, a strong acid. Other chemicals include parts per
million levels of PCP. Gloves, safety glasses, and
protective laboratory coats are recommended when using
the system and were used during this demonstration.
Both the soil and water test kits contain enough
reagents and supplies to analyze four samples and can be
purchased from the developer for $225 each. The
developer states that either of these test kits can be
modified to provide the user with the specific detection
level of interest for no additional charge. The developer
offers accessory items required for the soil and water
systems. The differential photometer is required for both
systems and is available for $935. The balance is
available for $100, and the mechanical pipet is available
for $219. All of these items are required for the soil
system. A timer also is sold by EnSys for $29.95.
EnSys offers an accessory pack that includes the
differential photometer, the mechanical pipette, the
balance, and the timer. This accessory pack costs
$1,250. The accessory pack can be rented from EnSys
on a daily or weekly rate. The daily rental rate is $150
and the weekly rate is $400. According to EnSys, the
shelf-life of the reagents used for its technology is 4
months for the water test kit and 3 months for the soil
test kit. An expiration date and lot number is printed on
each batch of reagents produced.
Logistical costs will vary depending on the scope of
the project. As discussed earlier, the system can be
operated outdoors or indoors, with electricity or without.
A refrigerator or cooler is required for storage of
reagents used with the water system. The best results
will be attained when the system is used indoors in a
temperature-controlled environment. Logistical costs
may include trailer rental, electrical hookup fees,
refrigerator rental or purchase, and electrical or gas
usage. Waste disposal is another operating cost of the
system. During this demonstration, about 200 samples
were analyzed using the test system. The waste
generated by these analyses filled half a 55-gallon drum.
The appropriate way to dispose of this waste is through
an approved incinerator facility. The cost for disposal of
one drum of this waste is estimated at $1,000.
Performance Factors
The following paragraphs describe performance
factors, including detection limits, sample throughput,
linear range, and drift. Specificity, another performance
factor, is discussed separately because of its complexity.
Detection Limits
The detection limit of the system for water samples
is 0.005 ppm. This is based on the analysis of
200 microliters of a water sample and 5 drops of an
unspecified amount and concentration of a PCP standard.
If the water sample's optical density is greater than that
of the PCP standard, then the sample contains less than
0.005 ppm. The detection limit of the system for water
samples is greater than the 1.0 ppb maximum
contaminant level (MCL) for PCP.
The detection limit of the system for soil samples is
0.5 ppm. This is based on the analysis of 100 microliters
of a soil extract (10 grams to 10 milliliters) and an
unspecified amount and concentration of a PCP standard.
If the soil sample's optical density is greater than that of
the PCP standard, then the sample contains less than 0.5
ppm.
Sample Throughput
Sample throughput was determined by evaluating
both the time required to extract and analyze one sample
and the number of samples analyzed in 1 work day.
According to the developer, about 20 minutes is required
to analyze a water or soil sample. When run in batches
of four samples, the developer states that analysis of all
four samples can be completed in less than 30 minutes.
EnSys estimates that 50 to 75 samples can be processed
in 1 day. The analysis of samples during this
demonstration was completed in 8 days. The average
work day was 10 hours in duration. The number of
samples analyzed during the demonstration was 184.
This number included field duplicate samples, specificity
samples, and QC samples. The 184 samples were
analyzed in 80 hours or slightly more than two samples
per hour. The operator was able to average a sample
every 27 minutes, slightly longer than the 20 minutes
claimed by the developer. The average number of
samples analyzed in a 10-hour work day was 23. This
was less than the 50 to 75 samples the developer
estimates can be processed per day. The largest number
of samples analyzed in one 10-hour day was 40. The
operator noted that the maximum number of soil samples
that could be extracted and analyzed in one 10-hour day
was 40. This equals one sample every 15 minutes,
which is less than the 20 minutes claimed by the
developer. The operator also noted that 33 water
samples were completed in one 10-hour work day, and
that three dilutions of the water samples were prepared,
24
-------
rather than the one dilution usually required. When
analyzing water samples with one dilution of the water
sample, a much higher sample throughput can be
expected. The operator noted that sample analysis time
did not include the time required for sample handling,
data documentation, difficult extractions, or the
preparation of QC samples. The time required by the
operator to perform these tasks prevented him from
completing analysis of 50 to 75 samples per day, as
estimated by the developer.
Linear Range and Drift
This technology uses one level of standard for both
water and soil analysis. There is no linear range
established for the standards. Samples are serially
diluted, 1 to 10, to obtain different levels of detection.
Drift normally is a measurement of an instrument's
variability in quantitating a known amount of a standard.
This technology eliminates the variability associated with
drift by requiring that a new standard be analyzed with
each set of samples analyzed. A common method for
evaluating drift for immunoassays is to compare the
responses of standards to a negative control standard. A
negative control standard is not required for sample
analysis when using the system, and this method of drift
evaluation could not be performed during this
demonstration.
Specificity Study
Specificity refers to a technology's ability to identify
and quantitate a particular contaminant when in the
presence of other chemicals that could act as interferants.
For this technology, interferants might be natural
chemicals present in a matrix, carriers or other chemicals
used to introduce PCP during wood treatment processes,
or other contaminants that might be present at such
facilities. To assess the specificity of this technology,
PRC studied its chemistry, reviewed its developer's
literature, and conducted a specificity study.
Organic sample matrix effects for this technology
are primarily caused by humic acids. Humic acids exist
in most soils and are found in highest concentrations in
topsoil. Because they are slightly polar, humic acids can
leech from soil and, therefore, are found in water
samples as well. High concentrations of humic acids in
a sample can inhibit the extraction of PCP through
absorption. Inorganic sample matrix effects are
primarily caused by pH, salinity, and inorganic chemical
composition. The pH of water samples may affect the
immunoassay chemistry of the system. The pH of water
samples should be measured and, if extremes are noted,
the sample may need to be neutralized. Salinity of water
may sometimes affect immunoassay results. The
developer provides no information about how salinity
may affect its system. Some inorganic chemicals may
have an effect on the system's performance by inhibiting
the binding of PCP or the enzyme conjugate to the
antibody binding sites. The developer furnished no
information concerning the effects of inorganic
chemicals.
The operation at the former Koppers site used an
isopropyl ether butane carrier for PCP application. The
operation at the Winona Post site used a diesel fuel-based
carrier for PCP application. One objective of this
demonstration was to evaluate the technology's ability to
quantitate PCP concentrations in samples contaminated
with these carriers. Both solvents are highly volatile and
will not have long residence times in surficial soils. The
developer reports that diesel fuel in concentrations as
high as 10 percent will have no effect on soil results and
that concentrations as high as 10 ppm will have no effect
on water results. In addition to these PCP carriers, other
wood-preserving agents are used in conjunction with
PCP. The developer has determined that its system is
not affected by concentrations of 1,000 ppm creosote in
soil or 1,000 ppb creosote in water, nor by
concentrations of 1,000 ppm CCA in soil or 10,000 ppb
CCA in water. Another site-specific matrix effect is the
presence of other chemicals in the samples. Historical
data revealed the presence of other phenols, including
chlorophenols, at the sites. It also revealed the presence
of dioxins and furans, particularly the octa-isomers, at
the sites, but the effects of these chemicals on the
technology are believed to be insignificant compared to
those of the chlorophenols.
Immunoassays are sometimes depicted as a
lock-and-key type of chemical interaction. The lock
system is the antibody, and it is designed to interact with
the key, PCP or the enzyme conjugate. However, the
antibody also will interact with other chemicals. This is
referred to as cross reactivity and can be evaluated by
determining the concentration of a specific chemical
which will provide a positive result when analyzed using
the technology. The developer has experimentally
determined the cross reactivity for a number of
compounds which have similar chemical properties to
PCP. This cross reactivity data was presented earlier on
Table 6-1. Although the table shows that the system is
very specific to PCP, chemicals other than PCP can
provide a positive result. Chlorinated phenols, especially
the tetrachlorophenols and the trichlorophenols, have the
highest cross reactivities. The system will provide false
positive results in samples containing other chlorophenols
in high concentrations. Many other industrial and
natural chemicals have not been evaluated to determine
cross reactivity, and for this reason approved
methodologies should be used to confirm positive results.
The specificity study was conducted to determine if
25
-------
this technology would show cross reactivity to several
chemicals other than PCP. The specificity study
involved spiking clean sand with known concentrations
of chlorophenols and diesel fuel. It also involved spiking
blank water with a mixture of PCP in diesel fuel. The
specificity study's samples were prepared by the lead
chemist and distributed to the technology operator along
with the demonstration samples. The operator knew that
the samples were for the specificity study but did not
know what chemical each sample was spiked with, nor
the concentration.
The soil specificity samples were prepared by
weighing 10 grams of clean sand into a soil extraction
bottle and spiking it with microliter amounts of chemical
standards. One soil specificity sample, SS-05, was an
unspiked sand sample to ensure that the sand would not
provide a positive response. Soil specificity samples
SS-14 through SS-17 were spiked with 100 ppm of diesel
fuel, and soil specificity samples SS-06 through SS-09
were spiked with 10 ppm of 2,4-dichlorophenol. None
of these samples produced a positive result.
Soil specificity samples SS-01 through SS-04 were
spiked with 10 ppm of 2,3,4,6-tetrachlorophenol. These
samples produced greater than 5 ppm, but less than 50
ppm results. The developer states that 1.2 ppm of
2,3,4,6-tetrachlorophenol in a soil sample is required to
produce a positive result, and the specificity study
supports this claim. Soil specificity samples SS-10
through SS-13 were spiked with 10 ppm of
2,4,6-trichlorophenol. These samples all provided a
greater than 0.5 ppm, but less than 5 ppm result. The
developer states that 16 ppm of 2,4,6-trichlorophenol in
a soil sample is required to produce a positive result with
the system, but the study showed that the system will
respond to concentrations of 10 ppm of
2,4,6-trichlorophenol.
The water specificity samples were SS-18 through
SS-21 and were spiked with 125 ppm of diesel fuel and
50 ppb of PCP. This was done to evaluate the effects of
diesel fuel on the recovery of PCP in water samples. The
results for these samples were the expected results,
greater than 5 ppb, but less than 500 ppb.
Intramethod Assessment
Intramethod measures of the technology's
performance included its results on reagent blanks, the
completeness of its results, its intramethod accuracy, and
its intramethod precision. Reagent blank samples were
prepared by taking reagents through all extraction,
cleanup, and reaction steps of the analysis. An
acceptable reagent blank sample must not provide a
positive result for PCP. Six soil and one water reagent
blank samples were analyzed during the demonstration,
and none of these reagent blank samples provided a
positive response. For this demonstration, completeness
refers to the proportion of valid, acceptable data
generated. Results were obtained for all of the samples;
therefore, completeness was 100 percent.
Intramethod accuracy was assessed by using PE
samples and matrix spike samples. Five PE samples
were analyzed during the demonstration, two for the soil
matrix, and three for the water matrix. Both of the soil
samples and two of the water samples were purchased
from ERA; the other water sample was produced by
PRC. These samples were extracted and analyzed in the
same way as all other samples. The operator did not
know the samples were PE samples, nor did the operator
know the true concentrations or the acceptance ranges.
The true value concentration of soil PE sample 099
was 7.44 ppm with an acceptance range of 1.1 to 13
ppm. The result reported by the technology for this
sample was greater than 5 ppm, but less than 50 ppm.
The true value concentration of soil PE sample 100 was
101 ppm, with an acceptance range of 15 to 177 ppm.
The system indicated the concentration was greater than
50 ppm. These results agree with the true values. The
true value concentration of water PE sample 106 was
68.4 ppb, with an acceptance range of 10 to 120 ppb.
The technology indicated the concentration was greater
than 5 ppb, but less than 500 ppb. The true value
concentration of water PE sample 113, the PE sample
prepared by PRC, was 7.50 ppb, with an acceptance
range of 2.25 to 12.8 ppb. The technology indicated that
the concentration was greater than 5 ppb, but less than
500 ppb. In both of these cases, the system's results
agreed with the true results. However, the true value
concentration of water PE sample 107 was 2,510 ppb,
with an acceptance range of 377 to 4,420 ppb. The
technology indicated that the concentration was greater
than 5,000 ppb, which was above the upper acceptance
limit. The accuracy of the PE sample results, therefore,
was 100 percent for the soil results and 67 percent for
the water results. Overall, the accuracy as measured by
the PE samples was 80 percent. The technology could
only have been more accurate if all results fell within the
PE sample ranges. Still, the demonstration's criteria
stated that more than 90 percent of the samples had to
have acceptable results, and the technology, therefore,
was not considered accurate. Based on this data the
performance of the technology on analyzing PE samples
was unacceptable.
Matrix spike and matrix spike duplicate samples also
were used to assess intramethod accuracy. Matrix spike
samples are aliquots of an original sample into which a
known concentration of PCP is added. Original samples
to which the spike solution was added were required to
contain much less PCP than the amount added. These
26
-------
samples were then extracted and analyzed. Each also
was duplicated. Six soil samples and one water sample
were used for matrix spike samples. Therefore, there
were 12 matrix spike results for soil samples and two for
water samples. Soil samples were spiked with 2.00 ppm
of PCP. An accurate result for a soil matrix spike
sample would be greater than 0.5 ppm, but less than 5
ppm. Of the 12 soil matrix spike samples, 11 provided
accurate results. The matrix spike duplicate of sample
003 was found to contain greater than 5 ppm, but less
than 50 ppm, overestimating the amount of the PCP
spiked into the sample. Overall, the accuracy of the
results for the soil matrix spike samples was found to be
92 percent, and the system's performance on the matrix
spike analysis was assessed as acceptable. Water matrix
spike samples were spiked with 50 ppb of PCP. An
accurate result for a water matrix spike sample would be
greater than 5 ppb, but less than 500 ppb. Both of the
water matrix spike samples provided results of greater
than 5 ppb, but less than 500 ppb. The accuracy of the
water matrix spike samples, therefore, was found to be
100 percent.
Precision is evaluated by determining the number of
duplicate sample results which agreed with the original
sample results. Precision was assessed by comparing the
results obtained on duplicate samples. Three types of
precision data were generated: data from laboratory
duplicate samples, data from field duplicate samples, and
data from matrix spike duplicate samples. Usually these
duplicate samples are used to determine matrix vari-
ability and the effects of using several operators. To use
the duplicates to measure the method's precision, PRC
both controlled for matrix variability by thoroughly
homogenizing the samples and controlled for operator
effects by using only one operator
Laboratory duplicate samples are two analyses
performed on a single sample submitted for analysis.
Laboratory duplicate samples were analyzed with each
set of 20 samples submitted for analysis. Laboratory
duplicate samples were analyzed after the original sample
results were obtained. Only original samples with
positive results were used for laboratory duplicate
analysis. Six soil and one water laboratory duplicate
samples were performed during the demonstration. Of
the six soil laboratory duplicate samples analyzed, five
were found to agree with the original results. Sample
045 had a result of greater than 5 ppm, but less than 50
ppm the first time it was analyzed and a result of greater
than 50 the second time it was analyzed. Only one water
laboratory duplicate sample was analyzed during the
demonstration. The result for both the original sample
and its duplicate were found to be greater than 5 ppb, but
less than 500 ppb.
Field duplicates are two samples collected together,
but delivered to the laboratory with separate sample
numbers. Fourteen soil field duplicate samples were
collected and analyzed during this demonstration. Field
duplicate samples represented 17 percent of all soil
samples collected and analyzed. Of the 14 soil field
duplicate samples analyzed, 11 were found to agree with
the original results, and 3 were found not to agree with
the original results. Ten water field duplicate samples
were collected and analyzed during the demonstration.
Field duplicate samples represent 43 percent of all water
samples collected and analyzed during the demonstration.
All 10 water field duplicates were found to agree with
the original results.
For this demonstration, precision was considered
acceptable if 90 percent of the duplicate pairs provided
the same result. Overall, 20 soil duplicate pairs were
analyzed during this demonstration, and 16 pairs
provided matching results. The precision of the
technology during this demonstration was found to be 80
percent. This is below the 90 percent criteria established
as acceptable precision. Eleven water duplicate pairs
were analyzed during this demonstration, and all 11
provided matching results. The precision of the water
system during this demonstration was found to be 100
percent.
Six soil matrix spike duplicate samples and one
water matrix spike duplicate sample were analyzed and
their results were compared to the results of their
respective matrix spike samples. Precision was
evaluated by determining the number of matrix spike
duplicate pairs which provided the same result. Five of
the six matrix spike duplicate results agreed with their
respective matrix spike samples. The precision of the
soil matrix spike samples was determined to be 83
percent, below the 90 percent criteria. It should be noted
that, for the soil matrix spike duplicates to meet the
precision criteria, all six of the duplicate results would
have had to agree with the matrix spike results. Only
one water matrix spike duplicate pair was analyzed
during this demonstration. The matrix spike and matrix
spike duplicate results were found to agree, and the
precision of the water system, therefore, was found to be
acceptable.
Comparison of Results
to Confirmatory Laboratory Results
The following paragraphs detail the accuracy and
precision of the data from analyses using the Penta RISc
Test System when compared to that of the confirmatory
laboratory. The results from the confirmatory labora-
tory are considered accurate, and its precision is
considered acceptable. The results for soil and water
sample analysis are summarized in Tables 6-2 and 6-3.
The results from the soil and water sample analyses will
27
-------
be discussed separately. Within each of these two
sample matrices the data will be examined as a whole
and according to the site from which the samples were
collected.
Accuracy
To assess the accuracy of this technology, PRC
compared its data to the data from the confirmatory
laboratory by using a 2' by 2' contingency table and a
Fisher's Test, as approximated by a corrected Chi-square
statistic. This statistic was used to test the null
hypothesis that the frequency of correct assays is
independent of the analytical method used to produce
them. In other words, this statistic can be used to
determine whether two test methods produce statistically
similar results. This test loses most of its statistical
power if any expected frequencies fall below 5. A 2' by
2' contingency table was set up for the overall data set,
the site-specific data sets, and for each range evaluated
by the technology. The contingency tables were set up
to compare the number of correct and incorrect results
from a test kit relative to the confirmatory laboratory's
data. A Fisher's Test, at a 90 percent confidence level,
was then used to determine whether a relationship existed
between the two sets of results. Accuracy in this section
is defined as relating to the number of correct results.
While false positives may not impact the intended
application of this technology, they are not correct
results.
Soil Data Set
This data set consisted of 114 matched pairs of data.
It included two PE samples, 67 samples from the former
Koppers site and 45 samples from the Winona Post site.
The Penta RISc Test System places analytical data into
four categories: (1) below 0.5 ppm, (2) between 0.5 and
5 ppm, (3) between 5 and 50 ppm, and (4) greater than
50 ppm.
Examination of the entire data set revealed that the
soil test kit gave 83 correct results and 31 incorrect
results. The Fisher's Test showed that results from the
two analytical methods were statistically different. The
lack of correlation indicates that for the entire data set
the technology's results were not accurate. These
findings did not change when the entire data set was
divided by where the samples had been collected. The
technology produced 45 correct results and 22 incorrect
results on samples from the former Koppers site, and 37
correct results and 8 incorrect results on those from the
Winona Post site.
The confirmatory laboratory found that 14 samples
had PCP concentrations of less than 0.5 ppm. For this
range, the technology produced 12 samples in this range
that had concentrations above 0.5 ppm. The Fisher's
Test showed that results from the two sets of data were
not statistically similar. The lack of correlation indicates
that within this range, the technology's results were not
accurate. These findings did not change when the entire
data set was divided by where the samples were from.
The confirmatory laboratory placed nine samples from
the former Koppers site in this range; the technology
placed 13 there, nine of them correct and four incorrect.
The confirmatory laboratory placed three samples from
the Winona Post site in this range; the technology placed
eight there, three correct and five incorrect. Due to the
small data sets in this range for both the Koppers and
Winona Post site samples, this statistical test has little
power and at best may indicate a trend of greater
accuracy for detecting PCP in an isopropyl ether and
butane carrier.
The confirmatory laboratory placed 30 samples into
the range between 0.5 and 5 ppm. The technology
produced 12 correct results and 23 incorrect results for
the entire data set. Eighteen times it placed samples that
should have been in this range into a different range and
five times it placed samples into this range that should
have been elsewhere. The Fisher's Test showed correct
results. Two times it incorrectly placed samples that
should have been in this range into a different range, and
it incorrectly placed seven that results from the two sets
of data were not statistically similar. The lack of
correlation indicates that within this range, for all data,
the technology's results were not accurate. These
findings did not change when the entire data set was
divided into former Koppers site data and Winona Post
site data. The technology produced 11 correct results
and 15 incorrect results for the former Koppers site data.
The confirmatory laboratory had 24 results in this range.
The confirmatory laboratory placed six samples from the
Winona Post site into this range; the technology
produced one correct result and seven incorrect results.
28
-------
TABLE 6-2. SEMIQUANTITATIVE PENTA RISc TEST DATA AND CONFIRMATORY DATA FOR SOILS'
Sample
No.
001
001 D
002
003
004
005
006
007
008
009
010
011
011D
012
013
014
015
016
017
018
019
020
020D
021
022
023
024
025
026
027
Penta
RISc Test
System
(ppm)
>0.5<5
>5<50
<0.5
<0.5
>5<50
>50
>5<50
>5<50
>0.5<5
>0.5<5
>50
>50
>50
<0.5
>50
>50
>50
>50
>50
>5<50
>5<50
<0.5
<0.5
>50
>0.5<5
>5<50
>0.5<5
>50
<0.5
>5<50
Confirmatory
Laboratory
(ppm)
4.20
4.18
1.64
0.13
2.04
3.70
1.89
2.66
0.66
3.52
435.0
106.0
112.0
0.056
32.80
99.60
1,190
273.0
1,335
2.13
6.89
0.10
0.09
5,320
1.85
1.86
1.57
593.0
0.42
11.30
Technology
Accuracy
Correct
FP
FN
Correct
FP
FP
FP
FP
Correct
Correct
Correct
Correct
Correct
Correct
FP
Correct
Correct
Correct
Correct
FP
Correct
Correct
Correct
Correct
Correct
FP
Correct
Correct
Correct
Correct
Sample
No.
028
029
030
030D
031
032
033
034
035
036
037
038
039
040
040D
041
042
043
044
045
046
047
048
048D
049
050
050D
051
052
053
Penta
RISc Test
System
(ppm)
<0.5
>0.5<5
>50
>50
<0.5
<0.5
<0.5
>0.5<5
>50
>50
>0.5<5
>50
>0.5<5
>50
>50
>5<50
>5<50
>50
>50
>5<50
>0.5<5
>50
>50
>50
>50
>5<50
>0.5<5
<0.5
>50
>0.5<5
Confirmatory
Laboratory
(ppm)
0.45
1.06
28.60
29.00
1.43
0.62
0.40
0.31
145.0
36.80
1.19
77.00
3.32
400.0
34.40
6.44
4.09
655.0
6,956
22.10
0.95
13,920
26,100
30,260
255.0
2.16
1.25
0.43
28.20
2.23
Technology
Accuracy
Correct
Correct
FP
FP
FN
FN
Correct
FP
Correct
FP
Correct
Correct
Correct
Correct
FP
Correct
FP
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
FP
Correct
Correct
FP
Correct
TABLE 6-2. Continued
29
-------
Penta RISc
Test
Sample System
No. (ppm)
054 <0.5
055 >50
055D >50
056 >50
057 >5<50
058 >5<50
058D >0.5<5
059 >50
059D >50
060 >50
061 >50
062 >50
063 >50
064 >50
065 >50
066 >50
067 >50
068 >5<50
069 >50
070 >50
071 >50
072 >50
073 >50
073D >50
074 >50
074D >50
075 >50
Notes:
Confirmatory
Laboratory
(ppm)
0.47
3,135
3,003
9.90
8.74
3.53
9.13
9,600
102.6
1,008
2,744
138.0
1,610
1,978
1,577
57.80
110.0
47.70
798.0
2,888
289.0
336.0
74.80
78.20
836.0
1,520
3,692
Technology
Accuracy
Correct
Correct
Correct
FP
Correct
FP
FN
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Sample
No.
076
077
078
079
080
081
082
083
084
085
086
086D
087
087D
088
089
090
091
092
093
094
095
096
097
098
099
100
Penta RISc
Test
System
(ppm)
>50
>50
>50
>50
>50
>50
>50
>50
>50
>50
>0.5<5
>0.5<5
>50
>50
>0.5<5
>0.5<5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
>50
>5<50
<0.5
>5<50
>50
Confirmator
y
Laboratory
(ppm)
4,590
2,040
1,720
792.0
2,550
125.0
2,400
270.0
1,140
57.70
6.59
6.88
34.00
51.80
2.58
0.21
0.55
0.28
0.57
0.19
1.02
0.088
59.80
14.60
0.57
4.02
52.40
Technology
Accuracy
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
Correct
FN
FN
FP
Correct
Correct
FP
FN
Correct
FN
Correct
FN
Correct
Correct
Correct
FN
FP
Correct
a Samples 1 through 58D were collected from the former Koppers site; Samples 59 through 98 were
collected from the Winona Post site. Samples 99 and 100 were PE samples.
FP False positive.
FN False negative.
J Reported amount is below detection limit or not valid by approved QC procedures.
ND PCP was not detected above the detection limit.
NA Either the technoloc
y, the confirmatc
ry laboratory, or both, did not
detect PCP.
30
-------
TABLE 6-3. SEMIQUANTITATIVE PENTA RISc DATA AND CONFIRMATORY DATA FOR WATER8
Sample No.
PENTA RISc Test
System
(ppm)
Confirmatory Laboratory
(Ppm)
Technology Accuracy
101
102
103
104
105
105D
106
107
108
108D
109
109D
110
110D
111
111D
112
112D
113
>0.005<0.5
>5
>5
>5
>0.005<0.5
>0.005<0.5
>0.005<0.5
>5
>0.005<0.5
>0.005<0.5
>0.005
>0.005
>0.005<0.5
>0.005<0.5
>0.005
>0.005
>5
>5
>0.005<0.5
0.004
15.900
13.500
0.012
0.849
0.640
0.010
2.050
0.002
0.002
0.000
0.001
0.018
0.018
<0.001
<0.001
1.810
2.020
0.002
False Positive
Correct
Correct
False Positive
False Negative
False Negative
Correct
False Negative
False Positive
False Positive
False Positive
False Positive
Correct
Correct
False Positive
False Positive
False Positive
False Positive
False Positive
Note:
Samples 101 through 105D were collected from the Winona Post site; samples 108 through 112D were collected
from the former Koppers site. Samples 106, 107 and 113 were PE samples.
Due to the small data set in this range for the Winona
Post site samples, this statistical test has little power and
at best may indicate a trend of greater accuracy for PCP
in an isopropyl ether and butane carrier.
The confirmatory laboratory placed 18 samples into
the range between 5 and 50 ppm. The technology
produced seven correct results and 21 incorrect results.
Eleven times it placed samples that should have been in
this range into other ranges, and 10 times it placed
samples into this range that should have been elsewhere.
The Fisher's Test showed that results from the two sets
of data were not statistically similar. The lack of
correlation indicates that within this range, for all data,
the technology's results were not accurate. These
findings did not change when the entire data set was
divided into the former Koppers site data and the Winona
Post site data. The technology produced five correct
results, and 16 incorrect results for the former Koppers
site data. When compared to the 13 results the
confirmatory laboratory had in this range, the former
Koppers site results were statistically different from the
confirmatory laboratory's data. The technology
produced two correct results and three incorrect results
on samples from the Winona Post site. The confirmatory
laboratory had five results in this range. Due to the
small data set in the range for the Winona Post site
samples, this statistical test has little power and at best
may indicate a trend of no carrier effect.
The confirmatory laboratory reported that 52
samples had concentrations of PCP greater than 50 ppm.
The technology produced 52 correct results and nine
incorrect results. All of the incorrect results were due to
samples incorrectly being placed into this range. The
Fisher's Test showed that results from the two sets of
data were not statistically similar. The lack of
correlation indicates that within this range, for all data,
the technology's results were not accurate. These
findings changed when the entire data set was divided
into the former Koppers site data and the Winona Post
site data. The confirmatory laboratory placed 20
31
-------
samples from the former Koppers site into this range.
The technology produced 20 correct results and 8
incorrect results for this site. The Fisher's test showed
that the two sets of data from the former Koppers site
were statistically different. Therefore, for the former
Koppers site samples, the technology's results were not
accurate. The confirmatory laboratory placed 31
samples collected from the Winona Post site into this
range. The technology produced 31 correct results and
one incorrect result. When these results were compared,
no statistically significant difference was seen between
the two data sets. Therefore, for the Winona Post site
samples in this range, the results were accurate.
Overall, 83 of 114 times the technology was correct.
This is 73 percent of the time. Of the other 31 times, the
technology gave 21 false positive results and nine false
negative results. This equates to an 18 percent false
positive rate and a 9 percent false negative rate. All of
the false negative results were produced when samples
containing less than 10 ppm of PCP were analyzed.
When the former Koppers site data is examined
alone relative to the confirmatory laboratory's results, 45
of 67 times the technology was correct. This is 68
percent of the time. Of the other 22 times, the
technology gave 18 false positive results and four false
negative results. This equates to a 26 percent false
positive rate and a 6 percent false negative rate. When
the Winona Post site data was examined alone relative to
the confirmatory laboratory's results, 37 of 45 times the
technology was correct. This is 82 percent of the time.
Of the other eight times, the technology gave two false
positive results and six false negative results. This
equates to a 4 percent false positive rate and a 13 percent
false negative rate.
Overall, for the soil matrix the technology is
conservative. It was not always accurate, relative to the
confirmatory laboratory. The semiquantitative nature of
the technology does not allow it to be placed in a Level
3 data quality category. Based on the developer's QA
requirements and performance specifications for this
technology's use, it can produce Level 2 data. However,
the technology never met its developer's specifications
for percentages of correct, false positive, and false
negative results, and it exhibited a lack of correlation to
the confirmatory laboratory. The developer's
performance criteria are: (1) less than 12 percent false
positives, (2) less than 1 percent false negative results,
and (3) greater than 88 percent correct results. The
failure to meet its developer's specifications and its lack
of correlation to the confirmatory data places this
technology in a Level 1 data quality category. Because
the false negatives all occurred in samples containing less
than 10 ppm PCP, this technology may be applied to
sample characterization for applications that require
accuracy at concentrations greater than 10 ppm PCP. In
these cases, false negatives would most likely occur
below target levels and other errors (false positives)
would result in a conservative interpretation of data.
Therefore, the technology could be used to assist in some
activities.
The technology had a slightly higher percentage of
correct readings when used on the samples from the
Winona Post site. The former Koppers site samples
produced a higher percentage of false positives, while the
Winona Post samples provided a higher percentage of
false negatives.
Water Data Set
This data set consisted of 19 matched pairs of data.
This data set included three PE samples, five duplicate
sample pairs from the former Koppers site, and five
samples and one duplicate sample pair from the Winona
Post site. Due to the small data set size, PRC considered
duplicate pairs as two individual samples for this
assessment. Therefore, the data set as a whole consisted
of 19 data points while the data set for the former
Koppers site samples contained only 10 samples, and the
data set for the Winona Post site samples contained only
six samples. The data sets for the individual sites, when
divided into the technology's analysis ranges, are too
small to give any quantitative statistical analysis
meaningful power. Because of this, PRC only evaluated
the entire data set. Significance within the individual
concentration ranges was not evaluated though trends
will be identified.
The results from this technology place analytical
data into four categories: (1) below 0.005 ppm,
(2) between 0.005 and 0.5 ppm, (3) between 0.5 and 5
ppm, and (4) greater than 5 ppm. Examination of the
entire data set revealed that the water test kit gave nine
correct results and 10 incorrect results. The Fisher's
Test showed that results from the two sets of data were
statistically different. The lack of correlation indicates
that for the entire data set the technology's results were
not accurate. These findings did not change when the
entire data set was divided by the site where the samples
were collected. The technology produced six correct
results and four incorrect results for the former Koppers
site data and two correct results and four incorrect results
for the Winona Post site data. The other three samples
were the PE samples.
In a comparison of the technology's sample results
for the entire data set to the confirmatory laboratory's
results, 9 of 19 times the technology was correct. This
is 47 percent of the time. Of the other 10 times, the
technology gave eight false positive results and two false
negative results. This equates to a 42 percent false
32
-------
positive rate and an 11 percent false negative rate. When
the data from water samples collected at the former
Koppers site is examined alone, 6 of 10 times the
technology was correct. This is 60 percent of the time.
Of the other four times, the technology gave four false
positive results and no false negative results. This
equates to a 40 percent false positive rate and a 0 percent
false negative rate. When the data from samples
collected at the Winona Post site was examined alone,
two of six times the technology was correct. This is 33
percent of the time. Of the other four times, the
technology gave two false positive results and two false
negative results. This equates to a 33 percent false
positive rate and a 33 percent false negative rate.
Overall, for the water matrix the test system is
conservative. It is not always accurate, relative to the
confirmatory laboratory. The lack of correlation to the
confirmatory laboratory indicates that this
technologydoes not produce Level 3 data. Based on the
developer's QA requirements for the technology's use,
this technology can produce Level 2 data. However, the
technology never met its developer's specifications for
percentages of correct, false positive, and false negative
results. These criteria are: (1) less than 12 percent false
positive results, (2) less than 1 percent false negative
results, and (3) greater than 88 percent correct results.
Exceeding the developer's false positive and false
negative rates could indicate that this technology is not
appropriate for site characterization activities unless a
high percentage of samples are analyzed by a
confirmatory laboratory. This technology only exceeded
the developer's specifications for false negative
frequency for the Winona Post site samples. This
indicates that, if this technology is used to assist removal
actions at a PCP site where diesel fuel is the PCP
carrier, all negative results should be submitted for
confirmatory analysis. The failure of the test system to
meet its manufacturer's accuracy criteria and its lack of
correlation to confirmatory data places this technology
into a Level 1 data category.
The technology produced almost twice as high a
percentage of correct results for the former Koppers site
samples relative to the samples from the Winona Post
site. This indicates a trend for greater accuracy for
detecting PCP in water when isopropyl ether and butane
are the PCP carrier solvents. The two data sets had
similar false positive results while the former Koppers
site samples produced no false negatives and the Winona
Post site samples produced 33 percent false negatives.
Precision
The precision of the technology when compared to
that of the confirmatory laboratory had seven out of 14
soil duplicate pairs agree with the confirmatory
laboratory's corresponding duplicate pairs. This equates
to a 50 percent precision. This is an unacceptable
precision. A similar comparison for the water duplicates
showed that two out of six duplicate pairs matched. This
equates to a 33 percent precision for the water sample
field duplicates. This is an unacceptable precision.
33
-------
Section 7
Ohmicron Corporation: Penta RaPID
Assay
This section provides information on the Penta
RaPID Assay including background information,
operational characteristics, performance factors, a data
quality assessment, and a comparison of its results with
those of the confirmatory laboratory.
Theory of Operation
and Background Information
Ohmicron has developed the assay kit to determine
PCP concentrations in water, soil, crops, and food. The
water and soil applications of the assay kit were
evaluated during this demonstration (See Exhibit 7-1).
The assay kit uses the principles of ELISA to determine
PCP concentrations in water and soil samples. A review
of the principles of ELISA is presented in Section 6.
The differences in ELISA technology between this kit
and the one discussed in Section 6 are presented below.
The developer covalently binds its antibodies to
magnetic particles for use. The developer claims there
are three advantages to the use of magnetic particles over
covalently binding the antibodies to the walls of a test
tube. The first is that the magnetic particles are easier to
coat and manufacture than the test tubes. The second is
that less of the antibody is required for coating the
magnetic particles than for the test tubes. The third
advantage is that the magnetic particles provide more
surface area creating more opportunities for binding.
In addition to the antibodies, ELISA-based
technologies use an enzyme conjugate in the analysis step
of the immunoassay test. The enzyme conjugate is
formed by covalently binding a PCP analog to a
horseradish peroxidase enzyme. In the case of the
RaPID assay, this enzyme conjugate competes with PCP
in an environmental sample for antibody binding sites on
the magnetic particles. The enzyme conjugate provides
the means for identification and quantitation of PCP.
ELISA-based technologies use chromogenic reagents that
react specifically with the enzyme conjugate to perform
identification and quantitation of PCP.
In the case of the assay kit, the enzyme substrate,
hydrogen peroxide, and the chromogen, 3,3',5,5'-tetra-
methylbenzidine react with the enzyme conjugate to
produce a brilliant blue color. Because an exact number
of antibody binding sites are used with each sample and
an exact number of enzyme conjugate molecules are
introduced into each sample, the only variable which
exists is the number of PCP molecules present in the
environmental sample. The PCP in the sample will
compete with the enzyme conjugate molecules for
antibody binding sites, thus reducing the amount of blue
color formed by the reaction. It is the amount of blue
color formed by the sample that is used for identification
and quantitation of PCP.
The samples are quantified using the RPA-1 RaPID
analyzer. PCP results are directly reported in parts per
billion for water samples and parts per million for soil
samples. The RPA-1 RaPID analyzer is an internally
calibrated spectrophotometer, operating at a frequency of
450 nanometers. It is equipped with an electronic
integrator that has been programmed to perform the
three-point standard calibration required for the assay
kit. The integrator also has been programmed to flag
any calibration QC criteria that are not acceptable.
The assay kit is designed to detect PCP, but other
compounds may respond to it as well. This is referred to
as cross reactivity, and it is measured by determining the
concentration of a compound needed to yield a positive
result. The developer has evaluated a number of
compounds and has provided information on cross
reactivities. The developer suggests that all positive
results should be confirmed through the use of an
independent, nonimmunological method.
Operational Characteristics
34
-------
Exhibit 7-1. The process used by the PENTA RaPID ASSAY
The Penta RaPID Assay is composed of three kits:
the PCP RaPID Assay Kit, used for both soil and water
samples; the Soil Collection Kit, used for soil samples;
and the PCP Extraction Kit, used for soil samples.
Instrumentation required is shipped in a suitcase-size,
plastic container. These containers have small handles
and are easy to carry. The reagents and supplies re-
quired are shipped in large cardboard boxes. The
containers and boxes needed for analysis of 200 samples
would fit in the trunk of a large car or in the back seat of
a smaller car. The containers and boxes also can easily
be shipped by commercial carriers. During this demon-
stration, all containers and boxes were shipped to the
site directly from the developer.
The PCP RaPID Assay Kit can be purchased in
either a 30-sample size or a 100-sample size. The
following are contained in the assay kit: (1) one bottle of
PCP antibody coupled paramagnetic particles (20-millil-
iter bottle in 30-sample kit, 65-milliliter bottle in 100-
sample kit), (2) one bottle of PCP enzyme conjugate
(10-milliliter bottle in 30-sample kit, 35-milliliter bottle
in 100-sample kit), (3) three vials of PCP standards: 0.1,
2.0, and 10 ppb (2 milliliters each), (4) one vial of a
PCP control standard: 1.0 ppb (2-milliliter), (5) one
bottle of diluent/zero standard (10-milliliter bottle in the
30-sample kit, 35-milliliter bottle in the 100-sample kit),
(6) one bottle of peroxide solution (10-milliliter bottle in
the 30-sample kit, 35-milliliter bottle in the 100-sample
kit), (7) one bottle of chromogen solution (10-milliliter
bottle in the 30-sample kit, 35-milliliter bottle in the 100-
sample kit), (8) one bottle of stop solution (20-milliliter
bottle in the 30-sample kit, one 25-milliliter and one 40-
milliliter bottle in 100-sample kit), (9) one bottle of
35
-------
washing buffer (70-milliliter bottle in the 30-sample kit,
250-milliliter bottle in the 100-sample kit), (10)
polystyrene test tubes (1 box of 36 in the 30-sample kit,
3 boxes of 36 in the 100-sample kit), (11) one copy of
instructions, and (12) two sheets of graph paper. The
Soil Col-lection Kit contains the following: (1)
twenty-one soil collection devices with detachable
plungers and screw caps, (2) twenty-one screw caps with
filters, (3) twenty-one extract collection vials, (4) one
Styrofoam tube holder, (5) one base piece for the soil
collection devices, (6) one copy of the instructions for
use of the kit, and (7) forty-five chain-of-custody labels.
The PCP Extraction Kit contains the following: (1)
twenty bottles of PCP extract solution (20-milliliter
each), (2) twenty bottles of PCP extract diluent
(25-milliliter each), (3) one 50-microliter, disposable
precision pipet, (4) twenty disposable pipet tips, and (5)
thirty chain-of-custody labels.
In addition to the above-mentioned kits, other
equipment is required for the assays and can be obtained
separately from the developer. These items include an
RPA-I RaPID Analyzer, a magnetic separation unit, a
vortex mixer, a digital balance, a repeating pipet capable
of delivering 250, 500, and 1,000 microliters, a 200-
microliter fixed volume pipet, and a digital timer. Other
equipment which is helpful when using the assays and
which is not supplied by the developer includes
protective gloves, twenty-milliliter vials with screw caps
for diluting samples, a 10-microliter fixed volume pipet,
a pen, and both liquid and solid waste containers.
During the demonstration, the technology was
operated in a 28-foot trailer. Electricity was supplied to
the trailer for air conditioning and to provide lighting.
Electricity also was required to operate the RPA-I RaPID
Analyzer during this demonstration. The electricity
required was a 110-volt circuit. A refrigerator was
required to store analytical reagents for the assay. The
developer recommends storing reagents at 2 to 8 ° C and
has stated that reagents should not be stored at or below
0 °C. A 3-foot by 2-foot hood was used by the operator
for soil sample extractions. A 4-foot table was used to
perform the assay steps, sample dilutions, sample
analysis using the RPA-I RaPID Analyzer, logbook
entries, and other data documentation.
The operator of the assay kit was Mr. Nathan
Meyer, an employee of PRC with a bachelor of arts
degree in biology and a master's of science in
environmental science. While at PRC, Mr. Meyer has
conducted preliminary site assessments and investigations
at hazardous waste sites in EPA Region 7. He also has
assisted in RCRA compliance evaluation inspections and
in other RCRA enforcement oversight activities. Mr.
Meyer's training in the use of the assay kit included
viewing a videotape produced by the developer which
explained the equipment and provided step-by-step
instructions, and reviewing literature provided by the
developer. Mr. Meyer also received approximately 8
hours of training at the start of the demonstration from
Dr. Scott Jourdon of Ohmicron. This included training
in the procedures for extracting, preparing, and
analyzing soil and water samples using the assay kit and
instructions for use of the pipets and instrumentation.
QC procedures and requirements also were discussed.
Mr. Meyer then analyzed both soil and water samples
using the kit while under the supervision of Mr. Jourdon.
After analyzing these samples, Mr. Meyer noted that he
felt comfortable with his ability to properly analyze
samples.
The assay kit is designed to be operated by persons
with some understanding of laboratory procedures. Mr.
Meyer had worked in a laboratory and was familiar with
laboratory procedures. Mr. Meyer noted that he found
the assay kit and its components easy to use, but he also
noted that certain steps were tedious and required
concentration and consistent technique. The techniques
required can be acquired within one week's use of the
assay kit. As shown during this demonstration, a person
with no experience using the assay kit can produce
results. However, the best results can be expected from
operators familiar with the kit and have used it before.
The assay kit is promoted for use as a portable field
instrument, but does require special care and handling in
the field to avoid damage. The RPA-I RaPID Analyzer
is a spectrophotometer and electronic spectrophotometric
integrator and requires care during shipping and
transportation to avoid breaking its internal components.
It also requires protection from the elements, such as
moisture, sunlight, and extremes of temperature. The
balance and pipets are precise measuring instruments and
need to be handled with care. No mechanical or
electronic problems were experienced during the
demonstration.
Instrument reliability was evaluated by monitoring
specific calibration and QC checks. Calibration QC
criteria for the assay kit include the following: (1)
meeting or exceeding a correlation coefficient of 0.990,
and (2) obtaining a percent coefficient of variation for
calibration standard replicates of less than or equal to 10
percent. The assay kit also includes a PCP control
standard. This standard is analyzed with each
calibration. The true value of the control standard is 1.0
ppb and the assay kit must determine a value within the
range of 0.7 to 1.3 ppb or the calibration is not
acceptable. If any of these criteria are not met, sample
results cannot be considered valid.
During the demonstration 18 batches of samples
were analyzed. This required 18 calibrations. Fourteen
36
-------
of the 18 calibrations met the criteria mentioned above.
The first calibration was unacceptable due to a 33.8
percent coefficient of variation for the 2.0 ppb calibrator.
The third calibration was unacceptable because the
control sample fell outside the acceptable range. The
fourth calibration was unacceptable due to a 35.6 percent
coefficient of variation for the 2.0 ppb calibrator. The
tenth calibration was also unacceptable because both
control samples were outside the acceptance range.
After three of the first four calibrations were found to be
unacceptable, the operator contacted the developer for
technical assistance. Ohmicron officials suggested
studying the response values for calibrators that exceeded
the 10 percent coefficient of variation criteria, then
determining which of the two calibrators most closely
matched the response value of previous calibrators which
met this criterion. The calibration would then be
repeated using the calibrator which most closely matched
previous acceptable response values. This calibration
would then be analyzed twice. It was believed that this
would bring the percent coefficient of variation to an
acceptable level. The developer also suggested analyzing
two aliquots of the control calibrator: one after the
acceptable calibration, but before sample analysis, and
one at the end of sample analysis. One of the standards
must fall within the acceptance range to consider the
results valid.
All of the samples analyzed during an unacceptable
calibration were reanalyzed during an acceptable
calibration, with the exception of sample 026. This
sample was analyzed and reported during the
unacceptable third calibration. The analysis of this
sample was never repeated during an acceptable
calibration. The result for sample 026 was not detected
above the soil quantitation limit. This was discovered
during a technical review of the data. The technical
review advised reporting the result, rather than excluding
it from data comparison. The reason for this decision
was that the QC criterion which was not met, less than
or equal to 10 percent coefficient of variation, should not
affect a result of this sort.
Overall, PRC found the instrumentation required for
the assay kit to be reliable. Immunoassays are enzymatic
reactions and can be effected by changes in ambient
temperatures. The assay kit may have been affected by
the temperature in the trailer during this demonstration.
The operator of the assay kit noted several times during
the course of sample analysis that QA/QC criteria were
most frequently violated in the afternoon. Although PRC
did not record temperatures in the trailer during the
demonstration, the technology operators noticed a
temperature increase in the trailer during the afternoon.
The operator of this technology noted that he believed the
temperature in the trailer increased 5 to 10 °F in the
afternoon, even while the trailer was air conditioned.
This temperature fluctuation seemed to affect the
standards and QC criteria. The developer stated that
increased temperature can increase PCP binding in the
substrate, raising reported PCP concentrations.
Therefore, temperature fluctuation could cause
calibrators and controls to fail QC criteria set at cooler
morning times. Operation of this technology in a
temperature-controlled environment could eliminate this
problem.
The assay kit contains chemicals in quantities
ranging from a few milliliter to 500 milliliters. The
more dangerous chemicals include methanol, sodium
hydroxide, and sulfuric acid. The kit should not be used
near ignition points or open flames and care should be
taken to avoid dermal, respiratory, and oral contact with
methanol. Also, sodium hydroxide, a strong base, and
sulfuric acid, a strong acid, should never be mixed due
to the strong reaction that will occur. Other chemicals,
including PCP in standards, are used in much smaller
amounts.
Costs associated with using the assay kit include the
costs of the analyzer and reagents, logistical
requirements, the operator, and waste disposal. The
developer offers two different sizes of kits, a 30-tube kit
and a 100-tube kit. These assay kits contain all of the
necessary reagents needed to perform water sample
analysis and are also required for soil sample analysis.
Prices for the 30-tube kit are $200 per kit for the
purchase of one to three and $180 per kit for the
purchase of four or more. Prices for the 100-tube kit are
$450 per kit for the purchase of one to five kits, $425 per
kit for the purchase of six to 10, and $400 per kit for the
purchase of 11 or more.
Soil samples require the purchase of the above-
mentioned kit, as well as the Soil Collection Kit and the
PCP Extraction Kit. The Soil Collection Kit contains the
plasticware and filtration devices required for soil
extraction, and the PCP Extraction Kit contains the
reagents and pipets needed for soil extraction. Both of
these kits contain enough supplies to perform 20 soil
sample extractions. The price of the Soil Collection Kit
is $100, and the price of the PCP Extraction Kit is $127.
Ohmicron also offers additional 100-milliliter bottles of
PCP assay and extract diluent for $10. These diluents
are needed to perform the immunoassay and to perform
dilution of soil sample extracts.
Some of the equipment needed to analyze samples is
sold separately. The repeating pipet is offered for $340,
and pipet tips are $105 for a pack of 100. The 100-
microliter, fixed-volume pipet is offered for $162, and
pipet tips are available for $57 for 10 racks of 96 pipet
tips. The developer also offers the 200-microliter,
fixed-volume pipet and the 250-microliter, fixed-volume
37
-------
pipet for $162 per pipet. An alternative to the purchase
of these pipets is the purchase of a trivolume pipet which
can be adjusted to 100, 200, and 250 microliters,
offered for $260. Pipet tips for the 200- and 250-
microliter fixed-volume pipets and for the trivolume
pipet are available for $65 for 10 racks of 100 pipet tips.
The developer offers two different magnetic
separation racks for PCP analysis. A 60-position, bench-
top separator is available for $405, and a fiveposition
field kit separator is available for $130. The 60-position,
bench-top separator was used for sample analysis during
this demonstration. Other miscellaneous items needed to
perform sample analysis are a vortex mixer which costs
$225, a digital timer which costs approximately $28, and
500 test tubes which cost approximately $16.50. The
RPA-I RaPID Analyzer is sold by the developer and
costs $3,985. Additional rolls of paper for the analyzer
are $6 per roll.
According to the developer, the shelf-life of the
reagents used for the assay kit is 1 year from the date of
manufacture. An expiration date is printed on each lot of
reagents produced. If kits expire within 90 days of the
date of purchase, the developer will replace any unused
reagents at no charge.
Logistical requirements include an electrical supply
capable of operating the instrumentation, a refrigerator,
and temperature-control equipment. Costs associated
with these requirements may include trailer rental,
electrical hookup fees, refrigerator rental or purchase,
and electrical or gas usage.
Operator costs will vary depending on the technical
knowledge of the operator. As discussed earlier, the kit
can be used by individuals with some understanding of
laboratory techniques and a minimal amount of technical
training, thereby decreasing this cost. Waste disposal is
another operating cost. During this demonstration, about
200 samples were analyzed using the analyzer. The
waste generated by these analyses filled half a 55-gallon
drum. The cost for disposal of one drum of this waste is
estimated at $1,000.
Performance Factors
The following paragraphs describe the assay kit's
detection limits and sensitivities, throughput, linear
range, and drift. Specificity is discussed separately due
to its complexity.
PRC conducted three test runs in which water
samples were analyzed. In all three test runs, the
absorbance of the low calibrator divided by the
absorbance of the 0.0 ppm calibrator (B/Bo ratios) were
less than the 0.90 level recommended by Ohmicron.
Because the B/Bo ratio criteria was met, PRC used the
minimum detectable concentration of 0.06 ppb that was
specified by the developer as the limit of quantitation for
water samples analyzed during this demonstration. This
is below the 1.0 ppb MCL for PCP.
Soil sample extracts required a 1 to 1,000 dilution
when analyzed with the assay kit. When this dilution
factor was multiplied by the minimum detectable
concentration, and the units were corrected, the
theoretical soil detection limit was 0.06 ppm. PRC
conducted 12 test runs in which soil samples were
analyzed. Five of the 12 runs provided B/Bo ratios that
were greater than the 0.90 level recommended by
Ohmicron, with the highest ratio being 0.95. Because
the 0.90 B/Bo ratio was not met on all test runs, PRC did
not use 0.06 ppm as the minimum detectable
concentration for soils. Developer product literature
indicates that the quantitation limit for soil samples is
0.10 ppm. PRC used 0.10 ppm for the soil quantitation
limit.
The developer reports that about 60 minutes is
required to analyze a sample using its assay kit. Up to
26 samples can be prepared at the same time to reduce
the average analysis time per sample. The developer
claims that a proficient user of the assay kit should be
able to analyze 100 to 150 samples in an 8-hour day.
During the demonstration, the 193 samples were
analyzed in 90 hours; this equals 2 samples per hour.
The operator was able to analyze samples in 60 minutes,
as reported by the developer. The average number of
samples analyzed in a 10-hour work day was 21. This
was less than the 100 to 150 samples reported by the
developer. The largest number of samples analyzed in
one 10-hour day was 64. The operator noted that sample
analysis time did not include the time required for sample
handling, data documentation, difficult extractions, or
the preparation of QC samples. Samples which exceeded
the linearity range required dilution. Seventy-nine
samples required at least one dilution, and several of
these samples required more than one dilution. The time
required by the operator to perform these tasks prevented
him from completing analysis of 100 to 150 samples per
day.
The linear range of the assay kit is from 0.10 to 10.0
ppb for water samples, and from 0.10 to 10.0 ppm for
soil samples. The ranges are established by a three-level
calibration using 0.10, 2.00, and 10.0 ppb calibrators.
The linear range for the soil samples is defined as the
concentration of the calibration standards (low- and high-
level calibrators) multiplied by the appropriate factors
introduced by the soil extraction. Linearity of the
calibration is evaluated using the standard curve of the
absorbance versus concentration for each calibration
standard level. The developer defines acceptable
38
-------
linearity as a correlation coefficient of 0.990 or greater.
Samples which exceed the linear range require dilution
to determine the concentration of PCP in the sample.
Water samples are diluted by using less of the sample for
the assay procedure. This procedure can be employed
down to volumes of 10 microliters. When the analysis of
10 microliters exceeds the linear range, the water sample
is diluted with deionized water. Soil samples are diluted
by performing 1- to 100-fold dilutions of the sample
extract. These dilutions are continued until sample
results are within the linear range. Several times during
the demonstration, a sample which was found to be
above the linear range was diluted and reanalyzed. If the
result for the diluted sample was still above the upper
linear range, the sample extract was diluted further and
reanalyzed.
Drift normally is a measurement of an instrument's
variability in quantitating a known amount of a standard.
The assay kit eliminates the variability associated with
drift by requiring a new calibration with each batch of
samples analyzed. The absorbance values from the
standards analyzed were found to drift during the 18
calibrations performed. For example, the absorbance
values obtained from the 0.10 ppb calibrator ranged from
0.792 to 1.656. This is a significant range and gives
justification to the requirement of performing a new
three-level calibration for each set of samples analyzed.
Specificity
Specificity refers to a technology's ability to identify
and quantitate a particular contaminant in the presence of
chemicals that could act as interferants. For this
technology, interferants might be natural chemicals
present in a matrix, carriers or other chemicals used to
introduce PCP during wood treatment, or other
contaminants that might be present at such facilities. To
assess the specificity, PRC studied the assay's chemistry,
reviewed its developer's literature, and conducted a
specificity study.
Organic sample matrix effects are primarily caused
by humic acids found in highest concentrations in topsoil.
Humic acids can leach from soil and, therefore, can be
found in water samples, as well. High concentrations of
humic acids in a sample can cause a loss of recovery of
PCP through absorption. The assay kit uses a methanol
and sodium hydroxide buffer solution for sample
extraction and assay diluent to eliminate the absorption
of PCP by humic acids. The developer claims that
humic acids up to 10 ppm have no effect on the assay kit.
Inorganic sample matrix effects are primarily caused by
pH, salinity, and inorganic chemical composition. The
immunoassay portion of the analysis must be performed
under basic conditions. The developer recommends that
samples with a low pH be neutralized with a 6 Normal
sodium hydroxide solution before performing the assay.
The salinity of water may sometimes affect immunoassay
results. The developer reports that the assay kit is not
affected by solutions containing up to 0.65 Molar sodium
chloride.
Some inorganic chemicals may have an effect on the
performance of the assay kit. Inorganic compounds may
inhibit the binding of PCP or the enzyme conjugate to the
antibody binding sites. The developer has evaluated a
number of inorganic compounds which may affect the
performance of the kit. These chemicals and their tested
concentrations, as reported by the developer, are shown
in Table 7-1.
One objective of this demonstration was to evaluate
the technology's ability to quantitate PCP concentrations
in samples contaminated with diesel fuel and isopropyl
ether, which were used as carriers to introduce the PCP
into wood at the two wood treatment facilities.
Ohmicron has evaluated some solvents similar to the
isopropyl ether for their effect on the assay and found
that the minimum concentration needed to affect the
39
-------
TABLE 7-1. INORGANIC CHEMICAL RESPONSE AS
REPORTED BY OH MICRON
Concentration3 Tested
Compound (ppb)
Calcium 250,000
Copper 250,000
Iron 50,000
Manganese 250,000
Magnesium 250,000
Mercury II 250,000
Nickel 250,000
Nitrate 250,000
Phosphate 250,000
Sodium chloride 650,000
Sulfate 10,000,000
Sulfite 250,000
Thiosulfate 250,000
Zinc 250,000
Note:
a No response was found at these concentrations.
results is 5 percent acetone, 2 percent acetonitrile, and
10 percent methanol. Because isopropyl ether has a
similar chemical structure to these solvents, the
maximum concentration before an effect is produced
should be similar. Percent levels of isopropyl ether or
butane were not expected in the samples from the former
Koppers site, due to their volatility and relatively short
residence time in surficial soils. It is possible that
percent levels of diesel fuel were encountered in the
samples from the Winona Post site, but the developer
reports that diesel fuel concentrations up to 10 percent
will not affect the kit. In addition to these PCP carriers,
other wood preserving agents are often used in
conjunction with PCP. The developer has evaluated two
of these wood preserving agents and determined that the
kit is not affected by concentrations of 100 ppm creosote
or 1,000 ppm CCA.
Another site-specific matrix effect is the presence of
other chemicals present in the samples. Historical data
for the two demonstration sites revealed the presence of
other phenols, including chlorophenols, as well as
dioxins and furans, particularly the octa-isomers. The
effects of the nonchlorinated phenols, dioxins, and furans
were not evaluated during this demonstration, nor was
any information regarding their effects provided by the
developer. However, the effects of these chemicals on
the assay kit are believed to be insignificant compared to
those of the chlorophenols, due to their lesser chemical
similarity to PCP.
As shown in Table 7-2, chemicals that have very
similar chemical structure to PCP (trichlorophenols and
tetrachlorophenols) show a high degree of cross
40
-------
TABLE 7-2. COMPOUND CROSS REACTIVITY AS
REPORTED BY OH MICRON
Concentration
Needed to Provide
a Positive Result
Compound (ppb)
Pentachlorophenol
2,3,5,6-Tetrachlorophenol
2,3,4,6,-Tetrachlorophenol
2,3,5-Trichlorophenol
2,3,6-Trichlorophenol
Tetrachlorohydroquinone
2,4,6-Trichlorophenol
2,4,5-Trichlorophenol
2,3,4-Trichlorophenol
2,5-Dichlorophenol
2,6-Dichlorophenol
2,3-Dichlorophenol
2,4-Dichlorophenol
3,5-Dichlorophenol
Hexachlorobenzene
Hexachlorocyclohexane
0.06
0.21
0.91
1.52
2.44
8.70
15.1
21.5
53.2
62.9
286
611
887
1,670
1,560
5,790
reactivity; other chemicals, including the dichlorophenols
are less chemically similar to PCP, and have a very low
cross reactivity, less than 0.0001 of the response of PCP.
Ohmicron also has tested the following chemicals at
a level of 10,000 ppb and found they did not give a
positive result using the assay: alachlor, aldicarb,
benomyl, butachlor, butylate, captan, carbaryl, carben-
dazim, carbofuran, 4-chlorophenol, chlorthalonil, 2-4-D,
3,4-dichlorophenol, 1,3-dichloropropane, dinoseb,
matelaxyl, metalochlor, metribuzen, pentachlorobenzene,
pentachloronitrobenzene, picloram, propachlor, terbos,
triclopyr, thiobendazole, and thiphenate-methyl. Many
of these chemicals contain an aromatic ring group similar
to PCP; the antibody lock appears to be geared for the
hydroxyl group of the PCP key.
The specificity study involved spiking clean sand
with known concentrations of chlorophenols and diesel
fuel. The soil and water specificity samples were
prepared by the lead chemist and given to the technology
operator along with the demonstration samples. The soil
specificity samples were prepared by weighing 10 grams
of clean sand, placing it into a soil extraction device, and
spiking it with microliter amounts of the chemical
standards. One soil specificity sample, SS-17, was an
unspiked sand sample. It did not give a positive result.
Soil specificity samples SS-01 through SS-04 were
spiked with 50 ppm of diesel fuel. These samples did
not give a positive result. Soil specificity samples SS-09
through SS-12 were spiked with 5 ppm of 2,4-dichlor-
ophenol. These samples did not give a positive result.
Soil specificity samples SS-13 through SS-16 were
spiked with 5 ppm of 2,4,6-trichlorophenol. Two of
these samples did not give positive results. Two of the
samples, though, did give positive results. The result of
samples SS-13 and SS-15 were 0.11 ppm and 0.06 ppm,
respectively. The result for SS-15 was below the soil
quantitation limit, and it would be reported as not
detected during a field sampling event. Still, the
developer states that the cross reactivity of
2,4,6-trichlorophenol is 15.1 ppm. The results of the
specificity study may indicate that the actual cross
reactivity o this compound is lower than this level. Soil
specificity samples SS-05 through SS-08 were spiked
with 5 ppm of 2,3,4,6-tetrachlorophenol. All of these
samples produced a positive result. Sample results
ranged from 0.32 ppm to 0.53 ppm. The mean result
from the four samples was 0.41 ppm. The developer
claims that the cross reactivity of 2,3,4,6-tetrachloro-
phenol is 0.91 ppm. The specificity test results show
that this claim is accurate.
The water specificity samples, SS-18 through SS-21,
were spiked with 125 ppm of diesel fuel and 50 ppb of
PCP. This was done to evaluate the effects of diesel fuel
on the recovery of PCP in water samples. The water
specificity samples ranged in concentration from 42 to 70
ppb, with a mean result of 57.8 ppb. Recoveries for
PCP ranged from 84 to 140 percent, with a mean
recovery of 116 percent. If these values are compared
with the matrix spike values, the mean recovery of the
water specificity samples is equivalent to the water
matrix spike recoveries obtained. Based on this
comparison, diesel fuel at 125 ppm does not interfere
with the recovery of PCP at a concentration of 50 ppb.
Intramethod Assessment
Intramethod measures of the technology's
performance included its results on reagent blanks, the
completeness of its results, its intramethod accuracy, and
its intramethod precision. Reagent blank samples were
prepared by taking reagents through all extraction,
cleanup, and reaction steps of the analysis. Soil reagent
blank samples were extraction solvent added to the
extraction tubes, while the water reagent blank sample
was deionized water added to the assay tube. An
acceptable reagent blank sample must not contain PCP
above the method quantitation limit. Eleven reagent
blanks were analyzed during the demonstration. Ten of
the eleven reagent blanks were found to contain no PCP
above the method quantitation limit and were determined
to be acceptable.
41
-------
TABLE 7-3. PENTA RaPID ASSAY PGP PROFICIENCY SAMPLE RESULTS: Training Results
Sample
No.
A
B
C
Result
(ppb)
0.63
1.87
4.22
Mean
Result
(ppb)
0.64
2.63
5.12
Percent
Recovery
(%)
98
71
82
± 2 Standard Deviation
Range
(ppb)
0.40 to 0.87
1.87 to 3.40
3. 33 to 6. 90
Did Result Fall Within
Acceptance Range
(Yes or No)
Yes
Yes
Yes
TABLE 7-4. PENTA RaPID ASSAY POP PROFICIENCY SAMPLE RESULTS: Demonstration Results
Sample
No.
A
B
C
Result
(ppb)
0.31
1.05
3.94
Mean
Result
(ppb)
0.64
2.63
5.12
Percent
Recovery
(%)
48
40
77
± 2 Standard
Deviation Range
(ppb)
0.40 to 0.87
1.87 to 3.40
3.33 to 6.90
Did Result Fall Within
Acceptance Range
(Yes or No)
No
No
Yes
One reagent blank sample (RB-6) was found to
contain 0.13 ppm of PCP. This did not meet the reagent
blank requirements. Corrective action required for
unacceptable reagent blanks included a detailed data
review and, if necessary, the reanalyzing of the reagent
blank. If the reagent blank was still unacceptable,
corrective action called for the reanalysis of the reagent
blank and all samples associated with it. RB-6 was
analyzed during test run six, which included 20 soil
samples, one laboratory duplicate sample, and a matrix
spike and matrix spike duplicate sample. PRC closely
reviewed the data from test run 6 to determine whether
the unacceptable reagent blank sample affected sample
results. Two samples were reported as "ND" by the
RPA-I RaPID Analyzer. This indicates that the sample
contained less than 0.10 ppm, as reported by the
Analyzer. Since these samples contained less than the
unacceptable reagent blank, PRC feels confident in
reporting these samples as not detected above 0.10 ppm.
The other samples analyzed during test run six were
found to contain PCP. All these sample results, though,
were more than five times the amount of PCP found in
RB-6. PRC feels confident the reported values for all
samples analyzed during test run six were not affected by
the unacceptable reagent blank, RB-6. Therefore, no
corrective action was taken.
For this demonstration, completeness refers to the
proportion of valid, acceptable data generated.
Completeness for the samples analyzed by the assay was
100 percent, well above the objective of 90 percent.
Intramethod accuracy was assessed for the
technology through the use of proficiency samples, PE
samples, and matrix spike and matrix spike duplicate
samples.
Proficiency samples were provided by the developer to
verify and document operator accuracy. The proficiency
samples were analyzed two times, once during the
training period under the supervision of the developer
representative and once during the demonstration.
The developer has analyzed each of the proficiency
samples using its technology and has determined a mean
value and the 90 percent confidence interval (+2
standard deviation range) for each of the proficiency
samples. As Table 7-3 shows the operator was able to
produce results for all three of the proficiency samples
which were within this range. These results show that
the operator was proficient in the use of the assay kit
prior to the start of the demonstration. Table 7-4 shows
the results for the proficiency samples analyzed during
the demonstration. Two of the three sample results were
outside of the + 2 standard deviation range
recommended by the developer. In both samples, the
results were lower than expected. Recoveries of these
two samples compared to the mean value of the
proficiency sample were 48 and 40 percent. Overall, 67
percent of the proficiency samples analyzed were within
the + 2 standard deviation range deemed by the
developer to be acceptable. The results of these
proficiency samples may indicate that field personnel
recently trained in the use of the assay kit may not be
able to continuously produce results which are as
accurate as those produced by more experienced
personnel.
Five PE samples were analyzed during the
demonstration, two for soil and three for water. Both of
the soil PE samples and two of the water PE samples
were purchased from ERA; the other water PE sample
was produced by PRC. These samples were extracted
42
-------
TABLE 7-5. PENTA RaPID ASSAY SOIL PERFORMANCE EVALUATION SAMPLE RESULTS
Sample No.
099
100
Result
(ppm)
5.01
38
True Result
(ppm)
7.44
101
Percent
Recovery
(%)
67
38
Acceptance
Range
(ppm)
1.1 to 13
15to177
Did Result Fall Within
Acceptance Range
(Yes or No)
Yes
Yes
TABLE 7-6. PENTA RaPID ASSAY WATER PERFORMANCE EVALUATION SAMPLE RESULTS
Sample
No.
106
107
113
Result
(ppb)
62
1,100
7.62
True Result
(ppb)
68.4
2,510
7.50
Percent
Recovery
(%)
91
44
102
Acceptance
Range
(ppb)
10 to 120
377 to 4,420
2.25 to 12.8
Did Result Fall Within
Acceptance Range
(Yes or No)
Yes
Yes
Yes
TABLE 7-7. PENTA RaPID ASSAY SOIL MATRIX SPIKE SAMPLE RESULTS
Sample
No.
020
029
039
051
086
090
095
Note:
ND
Amount
Found
In Original
Sample
and
Duplicate
Sample
(ppm)
0.13
ND
ND
ND
1.86
ND
ND
Amount
Added
To Matrix
Spike
Sample
(ppm)
5.00
5.00
5.00
5.00
5.00
5.00
5.00
Amount
Found
In Matrix
Spike
Sample
(ppm)
3.60
4.57
5.50
5.20
8.46
5.87
6.31
None detected above soil quantitation limit of 0.
Sample
Percent
Recovery
(%)
69
91
110
104
132
117
126
10 ppm
Amount
Found
In Matrix
Spike
Duplicate
Sample
(ppm)
4.21
4.65
6.23
4.74
8.68
4.95
4.94
Duplicate
Percent
Recovery
(%)
82
93
125
95
136
99
99
Relative
Percent
Difference
(%)
17
2
13
5
3
17
24
and analyzed in the same way as the other samples. The
operator did not know that the samples were PE samples,
nor did the operator know the true concentration and
acceptance range. The results for the PE samples are in
Tables 7-5 and 7-6 All values reported for the PE
samples were within acceptance ranges. Accuracy of the
samples analyzed was found to be 100 percent for both
sample matrices.
Matrix spike samples are aliquots of original sample
into which a known concentration of PCP is added.
These samples are then extracted and analyzed in the
same way as original samples. Seven soil samples and
one water sample were used for matrix spike samples.
They also were duplicated. Therefore, 14 spiked soil
samples and two spiked water samples are used in this
assessment. The average recovery of the soil matrix
spike samples and duplicates was 106 percent or 5.3
ppm. This is very close to the 5 ppm actually added to
the samples. The standard deviation of the matrix spike
samples was 19 percent or 0.95 ppm. Control limits for
soil matrix spike recovery can be established following
guidelines outlined in the SW-846 Manual Method 8000.
Control limits are defined as + 2 standard deviations
from the mean. For the soil matrix spike samples
analyzed during the demonstration, the calculated control
limits ranged from 68 to 144 percent recovery. All soil
matrix spike samples analyzed fell within these control
limits. The soil matrix spike results are shown on Table
7-7. The recoveries for the water matrix spike sample
43
-------
TABLE 7-8. PENTA RaPID ASSAY LABORATORY DUPLICATE RESULTS: Soil Samples
Sample
No.
004
018
019
038
057
058
088
099
Original Sample Result
(ppm)
0.65
2.77
3.90
4.61
7.72
2.92
1.51
5.01
Laboratory Duplicate Sample
Result
(ppm)
0.75
2.02
3.78
2.24
3.09
3.01
1.19
2.86
Relative Percent
Difference
(%)
14
31
3
69
86
3
24
55
and its duplicate were 116 and 115 percent, respectively.
Their relative percent difference was 1 percent. Because
only two recoveries were produced for water matrix
spike samples, control limits similar to those developed
with the soil matrix spike samples were not established.
Intramethod precision for the kit was assessed by
comparing the results obtained on duplicate samples.
Three types of precision data were generated: data from
laboratory duplicate samples, data from field duplicate
samples, and data from matrix spike duplicate samples.
Usually these duplicate samples are used to determine
matrix variability and the effects of using several
operators. To use the duplicates to measure the method's
precision, PRC both controlled for matrix variability by
thoroughly homogenizing the samples and controlled for
operator effects by using only one operator for the entire
demonstration. Results for soil laboratory duplicate
samples are provided in Table 7-8. Soil laboratory
duplicate samples were analyzed after the original sample
results were obtained. Only samples with positive results
were used for laboratory duplicate analysis. Eight soil
laboratory duplicate samples were analyzed during the
demonstration. The original results obtained for these
samples ranged from 0.65 to 7.72 ppm. When the
analysis was duplicated, the results ranged from 0.75 to
3.78 ppm. RPD values for the soil laboratory duplicate
samples ranged from 3 to 86 percent. The mean RPD
value of the soil laboratory duplicate samples was 36
percent, with a standard deviation of 31 percent.
Water laboratory duplicate samples were analyzed
after the original sample results were obtained. Only
samples with positive results were used for laboratory
duplicate analysis. Two water laboratory duplicate
samples were analyzed. The original results obtained for
these samples were 4.34 and 1.42 ppb. When the
analysis was duplicated, the results were 4.34 and 1.43
ppb. RPD values for the water laboratory duplicate
samples were 0 and 1 percent.
Fourteen field duplicate soil samples were analyzed.
In one, the kit did not detect PCP. This sample was
eliminated from the statistical analysis. The remaining
results obtained for the soil field duplicate samples
ranged from 1.17 to 11,800 ppm. The field duplicate
sample results ranged from 1.46 to 11,700 ppm. RPD
values for the soil field duplicate samples ranged from 1
to 118 percent. The mean RPD value these samples was
28 percent, with a standard deviation of 31 percent. The
results are shown on Table 7-9. Ten water field
duplicate samples were analyzed. Field duplicate
samples represent 43 percent of all water samples
analyzed during the demonstration. The original results
obtained for the water samples ranged from 0.71 to
122,000 ppb. The field duplicates results were from
0.94 to 95,600 ppb. RPD values for the water field
duplicate samples ranged from 1 to 34 percent. The
mean RPD value of the water field duplicate samples was
20 percent, with a standard deviation of 9 percent.
These results are shown on Table 7-10.
PRC used the laboratory and field duplicates
together to evaluate the technology's precision. To do
this, PRC established control limits like those sometimes
used to evaluate laboratory duplicates. These control
limits were then used to determine whether the
difference between a result from a duplicate and the
result from its respective sample was reasonable. To
establish the control limits, all sample pairs that did not
produce two positive results were removed from the data
population. Then, the RPD for each pair was calculated
and the mean RPD and population standard deviation
were determined. The lower control limit was set at
zero because this would mean that the results from a
duplicate and its sample matched perfectly. The upper
44
-------
TABLE 7-9. PENTA RaPID ASSAY SOIL FIELD DUPLICATE SAMPLE RESULTS
Sample
No.
001
011
020
030
040
048
050
055
058
059
073
074
086
087
Original Sample Result
(ppm)
1.61
55
0.13
8.3
19
11,800
1.17
748
2.92
1,800
123
649
1.86
29
Field Duplicate Sample Result
(ppm)
2.53
64
ND
8.6
10
11,700
1.46
670
2.16
2,200
125
690
7.21
20.2
Relative Percent
Difference
(%)
44
15
NA
3
62
1
22
11
30
20
2
6
118
36
Notes:
ND None detected above soil quantitation limit of 0.10 ppm.
NA Not applicable due to ND result.
control limit was set by multiplying the standard
deviation by two and adding it to the mean RPD. The
RPD of each sample pair was then compared to these
control limits. If greater than 90 percent fell within the
control limits, the technology's precision was considered
adequate. If fewer than 90 percent of them fell within
this range, the data was reviewed, and if no explanation
could be found, the technology's precision was
considered inadequate.
The assay kit for soils had 21 duplicates in which
both a sample and its duplicate had positive results. The
data from these 21 pairs had a mean RPD of 31 percent
and a standard deviation of 31. The control limits were,
therefore, set at 0 and 93 percent. All but one of the 21
RPDs fell within the control limits. The sample pair that
was outside the control limits had results of 1.86 and
7.21 ppm, respectively. That sample pair had an RPD of
118 percent. Still, 95.2 percent of the sample pairs had
RPDs within the control limits. Based on this, the
precision of the assay kit for soil samples was found to
be acceptable.
The assay kit for water had 12 duplicate pairs in
which both a sample and its duplicate had positive
results. The data from these pairs had a mean RPD of
17 percent and a standard deviation of 11. The control
limits were, therefore, set at 0 and 39 percent. All 12
RPDs fell within these limits, and the precision for the
assay kit for water samples was found to be acceptable.
Matrix spike duplicate samples were used to further
evaluate precision. Seven soil matrix spike duplicate
samples and one water matrix spike duplicate sample
were analyzed and their results were compared to the
results of their respective matrix spike samples.
Precision of the matrix spike duplicate samples was
evaluated through the RPD of the matrix spike result and
the matrix spike duplicate result. RPD values for the
seven pairs of matrix spike soil samples ranged from 2 to
24 percent. The mean RPD value from these seven pairs
was 12 percent, and the standard deviation was 8
percent. If an upper control limit of two times the
standard deviation is used, the upper control limit for
RPD determined for soil samples analyzed during the
demonstration was 28 percent. All RPD values for the
soil matrix spike duplicate samples were below the upper
control limit. The test kit's precision, therefore, was
found to be acceptable. Precision of the water matrix
spike duplicate sample was evaluated by examining the
45
-------
TABLE 7-10. PENTA RaPID ASSAY WATER FIELD DUPLICATE SAMPLE RESULTS
Sample
No.
101
102
103
104
105
108
109
110
111
112
Original Sample Result
(PPP)
9.00
53,000
122,000
2,300
31
4.34
1.42
30
0.71
1,500
Field Duplicate Sample
Result
(ppb)
8.87
59,100
95,600
1,800
40
5.01
2.00
24
0.94
1,800
Relative Percent
Difference
(%)
1
11
24
24
25
14
34
22
28
18
RPD between the matrix spike result and the matrix
spike duplicate result. The RPD value for the matrix
spike water samples was 1 percent.
Comparison of Results
to Confirmatory Results
The quantitative results of the assay kit were
compared to those of the confirmatory laboratory
using the statistical methods detailed in Section 4 (see
Tables 7-11, 7-12, and 7-13) . The purpose of this
statistical data evaluation is to assess whether the
technology meets Level 3 criteria for accuracy and
precision. If the technology cannot, but the technology's
data can be mathematically corrected to become
accurate, it can be placed into a Level 2 data quality
category.
The Wilcoxon Signed Ranks Test was used to
supplement the findings of the regression analysis. This
nonparametric test was used to test the hypothesis that
the technology's data was not significantly different from
the confirmatory laboratory's data. In cases where the
regression analysis criteria for accuracy were not met,
but the Wilcoxon Signed Ranks Test indicated that there
was no significant difference between the two data sets,
the technology's data was placed in the Level 2 category.
The contradiction between the regression and inferential
statistic indicates that the distribution of one or both of
duplicate had positive results. The data from these 12
pairs had a mean RPD of 17 percent and a standard
deviation of 11. The control limits were, therefore, set
at 0 and 39 percent. All 12 RPDs fell within the control
the data sets violated a fundamental assumption of
regression analysis, normal data distribution. In these
cases, the regression analysis was discarded, and
evaluation was based on the inferential statistic. Data not
meeting any of the above criteria was placed into a Level
1 data quality category, unless it was unable to detect
PCP when PCP was in a sample. Identifying the
presence of a compound is a criterion of Level 1 data.
The Wilcoxon probability and the parameters for the
regression analysis are presented in Table 7-14.
This technology was assessed in both soil and water.
For each of these matrices two different sites were
sampled. One site was contaminated with PCP in an
isopropyl ether and butane carrier solvent and the other
site was contaminated with PCP in a diesel fuel carrier
solvent.
The statistical evaluation of this technology was
carried out in a tiered approach. For the soil matrix, if
the initial analysis of the entire data set (tier 1) showed
that the technology's data was statistically different from
the confirmatory laboratory's data, the data set was
divided by the site producing the samples. This
comparison was used to conduct a preliminary
assessment of any carrier effect on the technology's data
(tier 2). A third tier of analysis was conducted within
each site-specific data set. This assessment involved
splitting the data sets into two subsets again, one having
samples the confirmatory laboratory found had
concentrations less than 100 ppm, the other those having
concentrations greater than 100 ppm. This assessment
was used to address any potential concentration effects
on the technology's performance. The 100 ppm level
was selected as a general action level for PCP in soil.
EPA action levels for this contaminant are generally site
specific; however, they typically occur in a range
between 20 and 100 ppm.
46
-------
TABLE 7-11. SUMMARY OF DEMONSTRATION DATA: FORMER KOPPERS SITE SOIL SAMPLES
Sample
No.
001
001 D
002
003
004
005
006
007
008
009
010
011
011D
012
013
014
015
016
017
018
019
020
020D
021
022
023
024
025
026
027
028
029
030
030D
Notes:
a
b
J
u
Penta
RaPID Assay
(O.IOppmf
1.61
2.53
0.81
<0.10U
0.65
4.82
0.66
3.83
0.32
0.45
289
55.0
64.0
<0.10U
17.0
46.0
135
122
43.2
2.77
3.90
0.13
<0.10U
400
19.8
9.70
0.52
86.0
<0.10U
3.80
0.28
<0.10U
8.30
8.60
Confirmatory
Laboratory
(ppm)
4.42
4.18
1.64
0.1 3b
2.04
3.70
1.89
2.66
0.66
3.52
435.0
106.0
112.0
0.056b
32.80
99.60
1,190
273.0
1,335
2.13
6.89
0.10
0.09b
5,320
1.85
1.86
1.57
593
0.42
11.3
0.45
1.06
28.6
29.0
Sample
No.
031
032
033
034
035
036
037
038
039
040
040D
041
042
043
044
045
046
047
048
048D
049
050
050D
051
052
053
054
055
055D
056
057
058
058D
Detection limits are presented in parentheses.
Sample was analyzed by Method 81 51 A; all other samples were
Reported amount is below detection limit or not valid by approved
PCP was not detected.
Penta
RaPID Assay
(O.IOppmf
0.70
0.12
0.15
0.23
49.8
28.9
1.54
4.61
<0.10U
19.0
10.0
12.0
5.39
585
4,000
13.0
1.14
32,000
11,800
11,700
105
1.17
1.46
<0.10U
24.00
1.49
<0.10U
748
670
668
7.72
2.92
2.16
analyzed by Method
QC procedures.
Confirmatory
Laboratory
(ppm)
1.43
0.62b
0.40
0.31b
145
36.8
1.19
77.0
3.32
400.0
34.40
6.44
4.09
655.0
6,956
22.10
0.95
13,920
26,100
30,260
255.0
2.16
1.25
0.43
28.20
2.23
0.47
3,135
3,003
9.90
8.74
3.53
9.13
8270A.
47
-------
TABLE 7-12. SUMMARY OF DEMONSTRATION DATA: WINONA POST SOIL SAMPLES
Sample
No.
059
059D
060
061
062
063
064
065
066
067
068
069
070
071
072
073
073D
074
074D
075
076
077
078
079
Notes:
a
b
J
u
Penta
RaPID Assay
(0.10 ppmf
1,800
2,200
877
1,840
32.0
462
1,030
604
96.0
52.0
19.0
406
766
236
522
123
125
649
690
1,800
1,800
1,900
1,250
505
Confirmatory
Laboratory
(ppm)
9,600
10,260
1,008
2,744
138.0
1,610
1,978
1,577
57.80
110.0
47.70
798.0
2,888
289.0
336.0
74.80
78.20
836.0
1,520
3,692
4,590
2,040
1,720
792.0
Sample
No.
080
081
082
083
084
085
086
086D
087
087D
088
089
090
091
092
093
094
095
096
097
098
099
100
Detection limits are presented in parentheses.
Sample was analyzed by Method 81 51 A; all other samples were
Reported amount is below detection limit or not valid by approved
PCP was not detected.
Penta
RaPID Assay
(O.IOppmf
1,700
138
1,000
107
565
67.0
1.86
7.21
29.0
20.2
1.51
0.16
<0.10U
0.27
0.31
0.36
0.27
<0.10U
22.0
10.0
1.25
5.01
38.00
analyzed by Method
QC procedures.
Confirmatory
Laboratory
(ppm)
2,550
125.0
2,400
270.0
1,140
57.70
6.59
6.88
34.00
51.80
2.58
0.21b
0.55b
0.28b
0.57b
0.1 9b
1.02b
0.088b
59.80
14.60
0.57
4.02
52.40
8270A.
48
-------
TABLE 7-13. SUMMARY OF DEMONSTRATION DATA:
Water Data3
Penta Confirmatory
Sample RaPID Assay Laboratory
No. (0.00006 ppm)b (ppm)
101 0.00900 0.004140°
102 53.0 15.90
103 122 13.50
104 2.30 0.01230
105 0.031 0.8490
105D 0.040 0.6400
106 0.062 0.01 030C
107 1.10 2.050C
108 0.00434 0.001850°
108D 0.00501 0.002210°
109 0.00142 0.0001750°
109D 0.00200 0.0006300°
110 0.0300 0.01810°
110D 0.0240 0.01810°
111 0.000710 0.0003480°
111D 0.000940 0.0003200°
112 1.50 1.810
112D 1.80 2.020
113 0.00762 0.002270°
Notes:
a Samples 101 through 105D were collected from
the Winona Post site. Samples 108 through
112D were from the former Koppers site.
Samples 106, 107, and 113 were PE samples.
b Detection limit.
° Samples analyzed by Method 515.1; all other
samples were analyzed by Method 8270A.
J Reported amount is below detection
limit or not valid by approved QC
procedures.
U PCP was not detected.
Precision for this technology was assessed through
the use of the Dunnett's Test, and a Wilcoxon Signed
Ranks Test. The Dunnett's Test was used to see if the
technology's precision was different from the
confirmatory laboratory's precision, and the Wilcoxon
Signed Ranks Test was used to verify the results of the
Dunnett's Test.
Soil Samoles: Accuracv
FIGURE 7-1. TOTAL SOIL DATA SET
FIGURE 7-2. FORMER KOPPERS SITE SOIL
SAMPLES
FIGURE 7-3. WINONA POST SITE SOIL SAMPLES
The initial linear regression analysis of the entire data
set was based on results from 90 samples. The other
samples did not contain PCPs above the detection limit of
0.10 ppm. The r2 for this regression was 0.47, indicating
that a relationship may exist between the data sets. A
residual analysis of the data, though, identified samples
21, 44, 47, 48, 59, and 77 as outliers. PRC removed
these six points and recalculated the linear regression.
When the regression was recalculated on the 84 remaining
sample results, it defined an r2 of 0.81, indicating that a
relationship exists between the two data
sets. The Wilcoxon Signed Ranks Test was used to verify
these results. It indicated that the assay kit's data was
significantly different from that of the confirmatory
laboratory. These results indicate that this technology
is not accurate, but its results can be mathematically
corrected to estimate corresponding confirmatory data.
Based on these results, 10 to 20 percent of the samples
analyzed by this method need confirmation analysis so
that the regression parameters can be defined. This
places this technology, for the combined soil data set,
into the Level 2 data quality category. Figure 7-1
shows the relationship between the assay kit's data and
the confirmatory laboratory's data. All data points are
49
-------
TABLE 7-14. SUMMARY OF REGRESSION AND RESIDUAL STATISTICS: OHMICRON
All Data
All Data
All Data
<100
>100
ppm
ppm
Koppers-AII Data
Koppers
Koppers
<100
>100
ppm
ppm
Winona-AII Data
Winona
Winona
Notes:
<100
>100
ppm
ppm
N
84
46
33
48
32
13
33
15
20
r2
.81
.76
.68
.65
.61
.60
.90
.69
.75
N Number of data points.
r2 Coefficient of determination adjusted
Y-int. Y-axis intercept of the regression line
Standard Error Standard error of the estimate.
Y-int
28
.55
71
-3.4
1.5
-110
34
-3.8
98
for variance.
Slope
.43
.71
.43
.37
.57
.39
.51
1.2
.46
Standard
Error
197
6.1
350
350
4.7
680
160
22
280
Wilcoxon Probabilitv
Significant
Significant
Significant
Significant
Significant
Significant
Significant
Difference
Difference
Difference
Difference
Difference
Difference
Difference
No Significant
Difference
Significant
Difference
shown. This figure has been divided at 100 ppm to show
the response of the kit relative to a common action level.
The second tier of the evaluation involved the
separation of the data by site. This evaluation was
conducted to assess potential carrier effects on a
technology's performance. Figures 6-2 and 6-3 illustrate
these assessments. The initial analysis on the data
produced by samples from the former Koppers site was
based on results from 50 samples. The r2 for this
regression was 0.48, indicating that a relationship may
exist between the data sets. A residual analysis of the
data identified samples 47 and 48 as outliers. PRC
removed these two points as outliers and recalculated the
linear regression. When the regression was recalculated
on the 48 remaining sample results, it defined an r2 factor
of 0.65, indicating that some relationship may exist
between the two data sets. The Wilcoxon Signed Ranks
Test verified these results. It indicated that the kit's data
was significantly different from that of the confirmatory
laboratory. This data indicates that the technology is not
accurate, and places it, for the former Koppers site soil
data set, into the Level 1 data quality category.
The initial linear regression analysis on the data
produced by samples from the Winona Post site was
based on results from 38 samples. The r2 for this
regression was 0.66, indicating that a relationship may
exist between the data sets. A residual analysis of the
data identified samples 59, 61, 70, 76, and 77 as
outliers. PRC removed these 5 points and recalculated
the linear regression. When the regression was
recalculated on the 33 remaining results, it defined an f
factor of 0.90, indicating that this technology meets the
first criteria for Level 3 data classification. The slope
(0.51) and the y-intercept (34.4) are not both statistically
similar to their expected values for Level 3 classification.
The Wilcoxon Signed Ranks Test was used to verify
these results. It indicated that the kit's data was
significantly different from that of the confirmatory
laboratory. This indicates that the kit's data must be
mathematically corrected to become accurate. This
indicates that this technology is not accurate but its
results can be mathematically corrected to estimate
corresponding confirmatory data. Generally, 10 to 20
percent of the samples analyzed by this technology
would need to be submitted for confirmatory analysis to
define the regression parameters necessary for the
predictive model. This factor places this technology, for
the Winona Post soil data set, into the Level 2 data
quality category. Based on these results for the data set
as a whole and divided by site, there appears to be a
carrier effect. The kit showed a stronger correlation to
confirmatory data for samples with PCP in a diesel
carrier.
Finally, PRC assessed the data by concentration
range. The confirmatory laboratory found that 52
samples had concentrations of less than 100 ppm. The
initial linear regression on these samples, defined an r2 of
0.00, indicating that no relationship exists between the
two data sets. A residual analysis of the data, though,
identified samples 14, 38, 56, 66, 73, and 96 as outliers.
PRC removed these six points and recalculated the linear
regression. The r2 improved to 0.76, indicating that a
relationship exists between the two data sets. The
Wilcoxon Signed Ranks Test verified these results. It
indicated that the assay kit's data was significantly
different from that of the confirmatory laboratory. This
indicates that the kit's data must be mathematically
50
-------
corrected to simulate confirmatory data. Based on these
results, 10 to 20 percent of samples analyzed by this
method need confirmation analysis to calculate the
regression parameters for the predictive model. This
factor places this technology into the Level 2 data quality
category for samples with concentrations less than 100
ppm.
The data set of all results greater than 100 ppm PCP
contained 38 samples. The initial linear regression
defined an r2 of 0.42, indicating that a relationship may
exist between the two data sets. A residual analysis of
the data, identified samples 21, 44, 47, 48, and 59 as
outliers. PRC removed these five points and recalculated
the linear regression. When the regression was
recalculated, it defined an r2 of 0.68, below the 0.75
necessary for Level 2 classification. This indicates that
too little of a relationship exists between the two data sets
to produce a useable predictive model. The Wilcoxon
Signed Ranks Test indicated, at a 90 percent confidence
level, that the kit's data was significantly from that of the
confirmatory laboratory. This data indicates that there
is a concentration effect for these samples. For samples
with concentrations of PCP below 100 ppm the
technology produced Level 2 data; for those with
concentrations above 100 ppm, the technology produced
Level 1 data.
Of the samples from the former Koppers site, the
confirmatory laboratory found that 35 had concentrations
of less than 100 ppm. The initial linear regression on
these samples defined an r2 of 0.00. A residual analysis
of the data, though, identified samples 14, 38, and 56 as
outliers. PRC removed these three points as outliers and
recalculated the linear regression. An r2 of 0.61 was
then defined, indicating that a weak relationship exists
between the two data sets. The Wilcoxon Signed Ranks
Test, at a 90 percent confidence level, verified that the
test kit's data was significantly different from that of the
confirmatory laboratory. This factor places this
technology, for the Koppers samples with less than 100
ppm of PCP, into the Level 1 data quality category. Of
the samples from the former Koppers site, different the
confirmatory laboratory found that 15 had results greater
than 100 ppm. The initial linear regression on these
samples defined an r2 of 0.39, indicating that a weak
relationship exists between the two data sets. A residual
analysis of the data identified samples 47 and 48 as
outliers. PRC removed these two points, recalculated
the linear regression, and defined a new r2 of 0.60,
indicating little improvement in the relationship between
the two data sets. The Wilcoxon Signed Ranks Test
verified that the data sets were significantly different.
This factor places this technology, for the former
Koppers site samples with more than 100 ppm PCP, into
the Level 1 data quality category. This data indicates
that there is no concentration effect for the samples from
the former Koppers site. At both concentration ranges,
the technology produced Level 1 data.
The results from the Winona Post site samples were
grouped in the same way. The data set for results of less
than 100 ppm contained 15 samples, defined an r2 factor
of 0.69, and identified no outliers. However, the
Wilcoxon Signed Ranks Test, at a 90 percent confidence
level, indicated that the assay kit's data was not
significantly different from that of the confirmatory
laboratory. The contradiction between the two tests
indicates that one or more of the data sets are not
distributed normally, violating a fundamental assumption
of regression analysis. This makes the regression results
suspect. Therefore, based on the Wilcoxon Signed
Ranks Test, the technology's data can be placed into a
Level 2 category. The confirmatory laboratory found
that 22 samples from the Winona Post site had results
greater than 100 ppm. The initial linear regression on
these results defined an r2 factor of 0.71, indicating that
a relationship may exist between the two data sets.
Residual analysis of the data identified samples 59 and 76
as outliers. PRC removed these two points as outliers
and recalculated the linear regression. The new
regression defined an r2 of 0.75, just within the Level 2
cutoff criteria. This indicates that a relationship exists
between the two data sets, allowing a useable predictive
model to be produced. The Wilcoxon Signed Ranks Test
was used to verify these results. It indicated that the test
kit's data was significantly different from that of the
confirmatory laboratory. These results indicate that this
technology is not accurate, but its results can be
mathematically corrected to estimate corresponding
confirmatory data. Therefore, 10 to 20 percent of the
samples analyzed by this method need confirmation
analysis. This factor places this technology, for the
combined soil data set for Winona Post samples with
PCP concentrations above 100 ppm, into the Level 2 data
quality category. Based on this concentration-related
data analysis, concentration levels appeared not to affect
the technology on these samples. Level 2 data were
produced on samples in both ranges.
So/7 Samples: Precision
Intermethod precision was assessed by a comparison
of the assay kit's results on duplicate samples to similar
results from the confirmatory laboratory. Three
assessments were done; one on all duplicate pairs, and
one on the samples from each site. Due to the small
sample size for duplicates when grouped by site, those
evaluations have less statistical power than the evaluation
of the combined data set. The evaluation of precision on
a site-specific basis, though, reveals precision trends and
possibly identifies potential PCP carrier influences on
precision. When the Dunnett's Test compared the RPDs
between all of the field duplicates, it found that the
51
-------
precisions were similar. When precision was examined
relative to the site from which the duplicates were
collected, the Dunnett's Test again showed that the
precisions were similar. This indicates that the
technology's precision is not different from the
confirmatory laboratory's, for all data groupings. The
Wilcoxon Rank Sum Test confirmed this.
Water Samples: Accuracy
This section presents an assessment of accuracy by
tier. Tier 1 involves the examination of the data set as a
whole. Tier 2 involves the examination of data sorted by
site and is intended to give an indication of possible PCP
carrier effects on the technology's performance.
The initial linear regression analysis was based on
results from 19 samples. The r2 for this regression was
0.75, indicating that a relationship exists between the
data sets. Figure 7-4 shows this relationship relative to
the MCL of 1 ppb. A post-hoc analysis of residuals
identified samples 102 and 103 as outliers. The PCP
concentrations detected in these samples by the assay test
were both more than 10 times greater than their
corresponding confirmatory results. These two points
were removed and the regression was run again. The
second regression analysis produced an r2 of 0.33, a
slope of 0.60, and a y-intercept of 0.14. The reduction
in the r2 value out of the acceptable range for Level 2
data, shows that these two points were significantly
biasing regression analysis. However, the Wilcoxon
Signed Ranks Test did not verify these results. Both the
original data set and the data set with the outliers
removed indicated, at a 90 percent confidence level, that
the assay kit's data was not significantly different from
that of the confirmatory laboratory. The contradiction
between the regression analysis and the Wilcoxon Signed
Ranks test indicates that one or both of the data sets do
not meet the assumption of normality essential for
accurate regression analysis. Therefore, the regression
data was considered suspect. Based on the Wilcoxon
Signed Ranks Test, this technology, for the combined
water data set, falls into the Level 2 data quality
category.
Ten samples from the former Koppers site were
analyzed. The initial linear regression analysis on these
samples resulted in an r2 of 0.99, indicating that a strong
relationship exists between the data sets (Figure 7-5).
This r2 value meets the first condition for a technology to
be classified as capable of producing Level 3 analytical
data. However, an examination of the slope (0.86) and
y-intercept (0.003) showed that both of these parameters
are not statistically equivalent to their expected values for
a Level 3 analytical method. The Wilcoxon Signed
Ranks Test indicated that the assay kit's data was not
FIGURE 7-4. TOTAL WATER DATA SET
FIGURE 7-5. WINONA POST SITE WATER
SAMPLES
FIGURE 7-6. WINONA POST SITE WATER SAMPLES
significantly different from that of the confirmatory
laboratory. All of this indicates that the technology is
not accurate, but it does produce Level 2 data. The data
produced from this technology can be corrected if 10 to
20 percent of the samples are submitted for confirmatory
analysis.
Six samples were from the Winona Post site. The
initial linear regression analysis on these samples resulted
in an r2 of 0.65, indicating that a relationship may exist
between the data sets (Figure 7-6). However, this data
set contains the two points identified as significant
outliers when the combined data set was evaluated.
These points also were identified as outliers for this
subset. Removal of these points reduced the r2 to 0.00
52
-------
indicating that no relationship exists. This reduction in
r2 shows that these two data points strongly biased the
initial regression analysis. The r2 based on the smaller
data set confirms that this data does not meet either Level
2 or Level 3 criteria. This confirmed the regression
analysis and indicated that the assay kit's data was not
accurate for the Winona Post site water sample analysis.
This data, therefore, does not meet either Level 2 or
Level 3 criteria. This data falls into the Level 1
category.
The water data indicates that a carrier effect may
exist. The accuracy of this technology was greater for
the samples where isopropyl ether and butane was the
PCP carrier solvent.
Water Samples: Precision
When the Dunnett's Test compared the RPDs
between the technology's entire field duplicate data set
and the corresponding confirmatory laboratory data set,
the data sets were found to be statistically similar. This
indicates that the technology's precision is not different
from the confirmatory laboratory's. The Wilcoxon
Signed Ranks Test confirmed this data. The data set was
not reassessed by site because only one duplicate pair
was collected from the Winona Post site.
53
-------
Section 8
Millipore Corporation: EnviroGard PCP Test Kit
The EnviroGard PCP Test Kit is designed to quickly
provide semiquantitative or quantitative results for PCP
concentrations in soil and water samples. In its standard
configuration for soil analysis, the kit detects PCP
concentrations in the following ranges: greater than 5
ppm, between 5 and 0.25 ppm, between 0.25 and 0.025
ppm, and less than 0.025 ppm. For water analysis, the
kit detects PCP in the following ranges: greater than 100
ppb, between 100 and 20 ppb, between 20 and 5 ppb,
and less than 5 ppb. The developer will customize these
ranges to meet specific needs.
The kit uses the principles of ELISA to determine
PCP concentrations in water and soil samples. A review
of the principles of ELISA is presented in Section 5.
Unlike the EnSys technology, though, the colors
produced by samples tested by the kit are compared to
three PCP standards taken through all of the
immunoassay steps (Exhibit 8-1). The results can be
determined visually or by using a differential
photometer. For quantitative comparisons, the sample
and standards are compared to a blank water sample
using the differential photometer. The differential
photometer readings can be then be graphed as a
standard curve, allowing for the quantitative
determination. Overall, the kit is portable and can be
operated outdoors, but temperature extremes and
humidity can affect their performance. The reagents
used for sample analysis require refrigeration. The
differential photometer requires electricity but can be
operated using a rechargeable battery. The kit was found
to be easy to operate even by individuals with no prior
PCP immunoassay testing experience.
The highest number of demonstration samples
analyzed in one 10-hour day was 43. The average
number of demonstration samples analyzed in one
10-hour day was 20. Sample throughput was higher
when using the kit to obtain semiquantitative results.
Quantitative results require dilutions and reanalyses of
samples which decreases sample throughput. The
detection limit reported by the developer for soil samples
is 0.025 ppm and for water samples is 5 ppb.
The intramethod and intermethod statistical
comparisons for this technology revealed that it produced
a high percentage of false negative results and that they
had difficulty reproducing their results. A draft version
of these results was issued on March 4, 1994. The
developer submitted comments on the draft version on
May 17, 1994. In its comments, Millipore stated:
"Upon reviewing the data we were extremely concerned
with the lack of agreement with the reference method
particularly in regard to the number of false negative
results. We immediately began an investigation into the
source of these discrepancies." The developer noted that
the concentrations encountered during the demonstration,
especially those above 1,000 ppm, "significantly
exceeded concentrations we used in fortified samples
used to develop the immunoassay when there is more
emphasis on assay sensitivity." The developer then
fortified samples to these concentrations in the laboratory
and "noted two phenomena: there was a high incidence
of false negative results and there was excessive variation
between duplicate determinations." The developer
theorized that this was due to hydrophobic binding of the
enzyme conjugate. As a solution, the developer
proposed replacing a deionized water wash with a wash
that uses a detergent and adding an additional dilution
step to the technology's protocols. The developer claims
these changes greatly improved the technology's
laboratory results. The developer then used the new
protocol on about 33 extracts left over from the
demonstration. "Using the revised protocol, the false
negative rate was reduced to 3 percent while the false
positive rate remained at 3 percent." The developer
stated that the new protocol has now been included in its
commercial product. It added, "We believe that, based
on the knowledge gained from the field demonstration,
we have been able to make significant improvements in
our product. This experience demonstrates the utility of
the SITE program and its goal of evaluating innovative
technologies." It should be noted that the SITE
Program has not yet reevaluated the EnviroGard PCP
Test Kit since the change in protocol.
54
-------
Exhibit 8-1. The processes for the Millipore technology.
55
-------
Section 8
Applications Assessment
The principal advantage of the ELISA technology is
that it is very specific to PCP. This specificity reduces
the chances of determining that a sample contains PCP,
when in fact it does not. This specificity also greatly
reduces the chances of determining that a sample
contains no PCP, when it in fact does. ELISA
technologies are generally inexpensive when compared
to formal laboratory analysis using EPA-approved
methods for PCP. They are generally simple to operate
even for individuals with no prior immunoassay testing
experience. These technologies are portable and can be
operated outdoors under certain conditions. By batching
samples ELISA methods can have high sample
throughput, and they are capable of quickly providing
sample results.
ELISA systems are most applicable to sites where
PCP is a known contaminant and where large
concentrations of other chemicals are not present in the
samples. Generally, the larger the number of samples to
be collected, the greater the advantage of using these
systems. The use of these systems at these sites can
often decrease the cost of an investigation by decreasing
the number of samples requiring confirmatory laboratory
analysis and by enabling more work to be completed
during the sampling visit. Using these systems can allow
work to continue without having to wait for confirmatory
laboratory results. These systems also can be used by
laboratories as screening tools for PCP. Results can be
used to determine appropriate sample extraction
techniques, as well as to determine dilutions which may
be required for sample analysis. These results also can
be used to determine the appropriate analytical method to
be used for sample analysis.
Both semiquantitative and quantitative ELISA
systems are available. The semiquantitative systems
provide sample results greater than or less than a specific
detection level, while the quantitative systems provide
estimates of actual contaminant concentrations. The
detection levels for the systems can be customized to
meet site-specific action levels. For quantitative
systems, more than one detection level can be used to
obtain a closer estimate of PCP concentrations in
samples. A potential limitation of ELISA systems is that
their results may not always agree with results from the
analysis of the same sample by EPA-approved
methodologies. These systems' results, when compared
to confirmatory laboratory results, included both false
positive results, results which overestimate the
concentration of PCP in the sample, and false negative
results, results which underestimate the concentration of
PCP in the sample. Both false positive and false
negative results have important implications on
investigative and remedial activities. Another general
limitation of ELISA systems is that some organic
compounds, particularly tetrachlorophenols and tri-
chlorophenols, will provide a positive response when
present in a sample at part per million levels.
These systems also can be affected by other
chemicals found naturally in environmental samples,
such as humic acids, and by chemicals which are
associated with PCP treatment of wood products, such as
petroleum hydrocarbon products and solvents. These
chemicals must generally be present in percent level
concentrations before these systems are affected.
Logistical limitations of these systems generally include
that the enzyme conjugate used for soil and water sample
analysis requires refrigeration. The ELISA technologies
are generally temperature sensitive and should not be
operated in temperature extremes. Temperatures less
than 32 °F will cause certain reagents to freeze, and
temperatures greater than 97 °F may cause inaccurate
results. High humidity also may affect the immunoassay
chemistry.
EnSys Inc.:
Penta RISc Test System
This system is designed to provide semiquantitative
screening results for PCP in water and soil samples.
This system's soil test kit does not require refrigeration.
This is a specific advantage of this technology. During
the demonstration, this system appeared to be highly
temperature sensitive. When ambient temperatures
approached 90 °F, the system's calibration failed at a
56
-------
high frequency. Therefore, during the summer, the use
of this technology in an air-conditioned environment is
advised.
For the soil sample analysis, this system produced its
highest percentage of correct readings for samples
contaminated with PCP in an isopropyl ether and butane
carrier solvent. However, this carrier also produced a
higher percentage of false negatives with this carrier
relative to the diesel carrier. For the water sample
analysis, this system produced a higher percentage of
correct readings and a lower percentage (0 percent) of
false negative results for samples contaminated with PCP
in an isopropyl ether and butane carrier relative to a
diesel carrier.
Overall, the results of the demonstration indicate
that the system does not meet its developer's
performance specifications and its data is statistically
different from corresponding confirmatory data. Based
on this, the system produces Level 1 screening data.
This classification is based on the technology's failure to
meet its developer's performance specifications. This
demonstration indicated that for soils, the system tended
to give false negative results when used on samples with
PCP concentrations below 10 ppm. All of the remaining
inaccuracies were false positive results. Based on these
results, if target action levels are greater than 10 ppm
PCP, and when false positive results are acceptable, the
system can be used to guide field work and sampling
efforts. These activities include: determining the vertical
and horizontal extent of PCP contamination in soil,
tracking PCP groundwater contamination plumes, and
determining PCP contamination in surface waters.
Another use of the system is to monitor the effectiveness
of remediation techniques employed to reduce or
eliminate PCP contamination. The Penta RISc Test
System can be used to determine whether PCP
concentrations in soil or water samples exceed
site-specific action limits. In such applications, the
results would be considered conservative based on the
potential frequency of false positive results. False
positive results will cause remediation efforts to be
performed in areas that do not require them. This is
considered a conservative error compared to false
negative results which cause no remediation to take place
in areas where it is needed. Field investigators must
realize that the system is designed as a screening tool to
assist in evaluating PCP contamination. It is an approved
method for determining whether PCP concentrations are
above or below a site-specific action limit. Limited
confirmatory laboratory analysis, using EPA-approved
methods, will provide QC and check the performance of
the technology.
Ohmicron Corporation:
Penta RaPID Assay
The PCP RaPID Assay is designed to provide
quantitative screening results for PCP in water and soil
samples. Applications for the assay include both
laboratory and field uses. During the demonstration, this
system appeared to be temperature sensitive. Therefore,
during summer, the use of the assay in an air-conditioned
environment is advised.
This demonstration indicated that this technology's
soil results may be affected by the type of carrier used.
When the carrier solvent was diesel fuel, this technology
produced Level 2 data. This means that its data can be
used to guide field work and sampling efforts. These
activities include determining the vertical and horizontal
extent of PCP contamination in soil, tracking PCP
groundwater contamination plumes, and determining
PCP contamination in surface waters. Another use of the
system is to monitor the effectiveness of remediation
techniques employed to reduce or eliminate PCP
contamination. The PCP RaPID Assay can be used to
determine whether PCP concentrations in soil or water
samples exceed site-specific action limits. In such
applications, the results would be considered
conservative based on the potential frequency of false
positive results, unless the data was corrected through
confirmatory analysis. False positive results will cause
remediation efforts to be performed in areas that do not
require them. This is considered a conservative error
compared to false negative results which cause no
remediation to take place in areas where it is needed.
Field investigators must realize that the system is
designed as a screening tool. It is a field method for
determining PCP concentrations. Limited confirmatory
laboratory analysis, using EPA-approved methods, will
allow data correction, provide QC, and check the
performance of the technology.
For samples where isopropyl ether and butane were
the PCP carrier solvent, the assay produced Level 1 data.
In this application, the technology only could be used to
detect the presence or absence of PCP. Therefore, its
field uses would be limited to initial site survey activities
only. The results and carrier effects identified through
the soil analysis were reversed for the water analysis.
M MM pore Corporation:
EnviroGard PCP Test Kit
This test kit can be used on both soil and water
samples. It was developed as a semiquantitative field
test, but was assessed during this demonstration for its ability to produce both semiquantitative and quantitative
57
-------
results. During the demonstration, this test was found to analytical protocol, including adding a wash step that
produce false negative results and poor precision when uses a detergent. The developer claims that the modified
high concentrations of PCP were found in a sample. As EnviroGard PCP Test Kit has only a 3 percent false
a result, the kit's developer has changed part of the negative rate and a 3 percent false positive step, but the
SITE Program has not yet reevaluated the kit since the
modification.
58
-------
Section 9
References
Pearson, E. S. and H. O. Hartley. 1976. Biometrika Tables for Statisticians. Charles Griffin and Company, Ltd.
Third edition with corrections.
PRC Environmental Management, Inc. (PRC). 1993a. "Predemonstration Sampling Plan for the Evaluation of
Pentachlorophenol Field Screening Technologies." EPA Contract 68-CO-0047.
-. 1993b. "Final Demonstration Plan for the Evaluation of Pentachlorophenol Field Screening Technologies."
EPA Contract No. 68-CO-0047.
Stanley, T. W. and S. S. Verner. 1983. "Interim Guidelines and Specifications for Preparing Quality Assurance
Project Plans." U.S. Environmental Protection Agency, Washington D.C. EPA/600/4-83/004.
U.S. Environmental Protection Agency. 1990. "Quality Assurance/Quality Control Guidance for Removal
Activities." EPA/540/G-90/004.
59
------- |