.: Slates
Environmental Protection
Agency
Office of
Toxic Substances
Washington DC 20460
EPA-560/5-85-026
August. 1985
Toxic Substances
>EPA
VERIFICATION OF
PCB SPILL CLEANUP
BY SAMPLING
AND ANALYSIS
10
Y 0
10
I I I I l I I I
I I I I I I I I I I I
10 8
0
10
Second Printing
-------
VERIFICATION OF PCB SPILL CLEANUP BY
SAMPLING AND ANALYSIS
By
Bruce A. Boomer
Mitchell D. Erickson
Stephen E, Swanson
Gary L. Kelso
MIDWEST RESEARCH INSTITUTE
and
David C. Cox
Bradley D. Schultz
WASHINGTON CONSULTING GROUP
INTERIM REPORT NO. 2
WORK ASSIGNMENT 37
EPA Contract No. 68-02-3938
MRI Project No. 8501-A(37)
and
EPA Contract No. 68-01-6721
WCG Subcontract to Battelle Columbus Laboratories
No. F4138(8149)435
Prepared for:
U.S. Environmental Protection Agency
Office of Toxic Substances
Exposure Evaluation Division (TS-798)
401 M Street, S.W.
Washington, DC 20460
Attn: Mr. Daniel T. Heggem, Work Assignment Manager
Dr, Joseph J. Breen, Project Officer
Richard A. Levy, Work Assignment Manager
Joseph S. Carrat Project Officer
Second Printing
-------
DISCLAIMER
This document has been reviewed and approved for publication by the
Office of Toxic Substances, Office of Pesticides and Toxic Substances, U.S.
Environmental Protection Agency. The use of trade names or commercial
products does not constitute Agency endorsement or recommendation for use.
-------
PREFACE
This Interim Report was prepared for the Environmental Protection
Agency under EPA Contract No, 68-02-3938, Work Assignment 37. The work as-
signment is being directed by Mitchell D. Erickson. This report was prepared
by Dr. Erickson, Bruce A. Boomer, Gary L. Kelso, and Steve E. Swanson of
Midwest Research Institute (MRI), The sampling design (Section IV,A) was
written by David C. Cox and Bradley D. Schultz of the Washington Consulting
Group, 1625 I Street, N.W., Washington, D.C. 20006, under subcontract to
Battelle Columbus Laboratories, Subcontract No, F4138(8149)435, EPA Contract
No. 68-01-6721 with the Design and Development Branch, Exposure Evaluation
Division. The EPA Task Managers, Daniel T. Heggem, Richard A. Levy and John H.
Smith, as well as Joseph J. Breen, Joseph S.Carra, and Martin P. Hal per, of
the Office of Toxic Substances, provided helpful guidance and technical in-
formation,
NOTE: The second printing of this report contains additional discussion of
sampling points located outside the original cleanup (contaminated)
area as discussed on pages 11-16.
MIDWEST RESEARCH INSTITUTE
John M. Hosenfeld
Section Head
Jo«Ti E, Going, Director
Chemical Sciences Department
January 13, 1986
-------
TABLE OF CONTENTS
Page
I. Introduction, , ..... 1
II. Summary ..... 1
III. Overview of PCS Spills and Cleanup Activities 3
A. Introduction to PCB Spills and Cleanup 3
1. Current Trends 3
2. Limitations of This Overview 3
B. Components of the Cleanup Process 4
1. Health and Safety 4
2. Reporting the Spill 4
3, Quick Response/Securing the Site. ...... 6
4, Determination of Materials Spilled/Cleanup
Plan. . 6
5. Cleanup Procedures 6
6. Proper Disposal of Removed PCB Materials. . . 7
7. Sampling and Analysis ............ 7
8. Remedial Action 8
9. Site Restoration 8
10, Records ........ ... 8
11. Miscellaneous Considerations 8
IV. Guidelines on Sampling and Analysis 9
A. Sampling Design 9
1. Proposed Sampling Design 9
2, Sample Size and Design Layout in the Field. . 16
3. Judgemental Sampling. ,,.,....,.,, 23
4. Compositing Strategy for Analysis of Samples. 23
5. Calculations of Average Number of Analyses,
and Error Probabilities 24
B. Sampling Techniques. ..... .... 40
1. Solids Sampling 40
2. Water Sampling 41
3. Surface Sampling 41
4. Vegetation Sampling ..... 42
C. Analytical Techniques 42
1. Gas Chromatography (GC) 42
2, Thin-Layer Chromatography (TLC) 49
3. Total Organic Halide Analyses 50
iii
-------
TABLE OF CONTENTS (concluded)
Page
D. Selection of Appropriate Methods . , , , 50
1. Criteria for Selection. , 50
2, Selection of Instrumental Techniques. , . , , 50
3. Selection of Methods 51
4. Implementation of Methods ..... 53
E. Quality Assurance. 54
1. Quality Assurance Plan. ........... 54
2, Quality Control 55
F. Documentation and Records 58
G, Reporting Results 59
V. References. 60
IV
-------
I. INTRODUCTION
The U.S. Environmental Protection Agency (EPA) under the authority
of the Toxic Substances Control Act (TSCA) Section 6(e) and 40 CFR Section
761.60(d), has determined that polychlorinated biphenyl (PCS) spills must be
controlled and cleaned up. The Office of Toxic Substances (OTS) has been re-
quested to provide written guidelines for cleaning up PCB spills, with par-
ticular emphasis on the sampling design and sampling and analysis methods to
be used for the cleanup of PCB spills.
This work assignment is divided into two phases. The reports of
Phase I are presented in Draft Interim Report No. 1, Revision No. 1, "Cleanup
of PCB Spills from Capacitors and Transformers," by Gary L. Kelso, Mitchell
D. Erickson, Bruce A, Boomer, Stephen E. Swanson, David C. Cox, and Bradley
D. Schultz, submitted to EPA on January 9, 1985. Phase I consists of a review
and technical evaluation of the available documentation on PCB spill cleanup,
contacts with EPA Regional Offices and industry experts, and preparation of
preliminary guidelines for the cleanup of PCB spills. The document was aimed
at providing guidance in all aspects of spill cleanup for those organizations
which do not already have working PCB spill cleanup programs.
Phase II, reported in this document, reviews the available sampling
and analysis methodology for assessing the extent of spill cleanup by EPA en-
forcement officials. This report includes some of the information from the
Phase I report, incorporates comments on the Phase I report and the general
issue which were received at a working conference on February 26-27, 1985,
and addresses the issue from the perspective of developing legally defensible
data for enforcement purposes.
This report, intended primarily for EPA enforcement personnel, out-
lines specific sampling and analysis methods to determine compliance with EPA
policy on the cleanup of PCB spills. The sampling and analysis methods can
be used to determine the residual levels of PCBs at a spill site following
the completion of cleanup activities. Although the methodologies outlined in
this document are applicable to PCB spills in general, specific incidents may
require special efforts beyond the scope of this report. Future changes in
EPA policy may affect some of the information presented in this document.
Following a summary of the report (Section II), Section III presents
an overview of PCB spills and cleanup activities. The guidelines on sampling
and analysis (Section IV) includes discussion of sampling design, sampling
techniques, analysis, and quality assurance.
II. SUMMARY
This report presents the results of Phase II of this work assign-
ment. Phase I consisted of a review and technical evaluation of the avail-
able documentation on PCB spill cleanup, contacts with EPA Regional Offices,
and preparation of preliminary guidelines for the cleanup of PCB spills.
-------
Phase II (this document) reviews the available sampling and analysis methodol-
ogy for assessing the extent of spill cleanup by EPA enforcement officials,
The report incorporates some of the information from the Phase I report and
general issues received at a working conference on PCB spills.
The EPA has set reporting requirements for PCB spills and views PCB
spills as improper disposal of PCBs. Cleanup activities have not been stan-
dardized since PCB spills are generally unique situations evaluated on a case-
by-case basis by both the PCB owner (or his contractor) and the responsible
EPA Regional Office. Components of the cleanup process may include protect-
ing the health and safety of workers; reporting the spill; quick response/
securing the site; determination of materials spilled; cleanup procedures;
proper disposal of removed PCB materials; and sampling and analysis. The
level of action required is dependent on the amount of spilled liquid, PCB
concentration, spill area and dispersion potential, and potential human expo-
sure.
A sampling design is proposed for use by EPA enforcement staff in
detecting residual PCB contamination above a designated limit after a spill
site has been cleaned. The proposed design involves sampling on a hexagonal
grid which is centered on the cleanup area and extends just beyond its bound-
aries. Guidance is provided for centering the design on the spill site, for
staking out the sampling locations, and for taking possible obstacles into
account. Additional samples can be collected at the discretion of the sam-
pling crew.
Compositing strategies, in which several samples are pooled and
analyzed together, are recommended for each of the three proposed designs.
Since an enforcement finding of noncompliance must be legally defensible, the
sampling design emphasizes the control of the false positive rate, the proba-
bility of concluding that PCBs are present above the allowable limit when, in
fact, they are not.
Sampling and analysis techniques are described for PCB-contaminated
solids (soil, sediment, etc,), water, oils, surface wipes, and vegetation, A
number of analytical methods are referenced; appropriate enforcement methods
were selected based on reliability. Since GC/ECD is highly reliable, widely
used, and is included in many standard methods, it is a primary recommended
method for most samples. Secondary methods may be useful for confirmatory
analyses or for special situations when the primary method is not applicable.
Quality assurance (QA) must be applied throughout the entire moni-
toring program. Quality control (QC) measures, including protocols, certifi-
cation and performance checks, procedural QC, sample QC, and sample custody
as appropriate, should be stipulated in a QA plan.
-------
III. OVERVIEW OF PCB SPILLS AND CLEANUP ACTIVITIES
A. Introductiorito PCB Spi 1Is and Cleanup
The EPA has established requirements for reporting PCB spills based
on the amount of material spilled and disposal requirements for the spilled
PCBs and materials contaminated by the spill. Under TSCA regulations [40 CFR
761.30(a)(l)(iii) and 40 CFR 761.60d], PCB spills are viewed as improper
disposal of PCBs. Although specific PCB cleanup requirements are not
established in the TSCA regulations, each regional administrator is given
authority by policy to enforce adequate clean-up of PCB spills to protect
human health and the environment.
1. Current Trends
Due to regional variations in PCB spill policy and the lack of a
national PCB cleanup policy, PCB cleanup activities have not been standardized.
Individual companies owning PCB equipment and contract cleanup companies have
developed their own procedures and policies for PCB cleanup activities keyed
to satisfying the requirements of the appropriate EPA Regional Office. In
addition, the EPA Regional Offices typically have provided suggestions for
companies unfamiliar with PCB cleanup.
PCB spills are generally viewed as unique situations to be evaluated
on a case-by-case basis by both the PCB owner {or his contractor) and the EPA
Regional Office. However, a general framework is often used to approach the
problem. Most cleanup activities involve quick response, removal or cleaning
of suspected contaminated material, and post-cleanup sampling to document
adequate cleanup. Major considerations involved in the cleanup process in-
clude minimizing environmental dispersion, minimizing any present or future
human exposure to PCBs, protecting the health and safety of the cleanup crew,
and properly disposing contaminated materials.
In general, the involvement of EPA Regional Offices is limited to
phone conversations often including a follow-up call to receive the analytical
results of the post-cleanup sampling. If the EPA representative is not satis-
fied with the reported data, additional documentation, sampling and analysis,
or cleanup (followed by further sampling and analysis) may be requested.
In cases of special concern (e.g., large spills), EPA Regional Of-
fices may work more closely with the PCB owner or contractor in planning the
cleanup, sampling and analysis activities, and on-site inspections.
2. Limitations of ThisOverview
The general discussion in this chapter refers to the procedures,
policy, and considerations that seem to be widely used at present by PCB
owners and spill cleanup contractors in meeting the requirements of the EPA
Regional Offices. The activities described do not involve EPA regulations or
policy except where indicated, since the EPA has not established requirements
on PCB cleanup procedures.
-------
Table 1 categorizes PCB spills into approximate levels of action
for PCB spill cleanup based on concern. Potential environmental problems in-
crease with increases in PCB concentrations, amount of spilled liquid, spill
area and dispersion potential, and potential human exposure. The three spill
types presented in Table 1 are based on very rough estimates. "Severity" in
one key item such as human exposure could raise a spill to a Type 3 (i. e.,
requiring special attention). On the other hand a spill of a large volume of
liquid may be considered a Type 2 spill due to a relatively low concentration
of PCBs, The three categories are only approximate and are intended to demon-
strate the flexibility needed in responding to PCB spills. EPA Regional Of-
fices should provide guidance on spill cleanup activities whenever questions
develop.
The situations described in this chapter are limited to recent PCB
spills of similar magnitude to the reported spills associated with PCB oil
transformers and capacitors (i.e., Type 2 in Table 1). Unusually severe spill
incidents (Type 3 in Table 1) involving large volumes of PCBs, a large spill
area, a high probability of significant human exposure, and/or severe en-
vironmental or transportation scenarios may require special considerations,
beyond the scope of this discussion.
All spills from regulated equipment are typically subject to the
detail of effort outlined in this chapter. Although cleanup of smaller spills
(Type 1 in Table 1) is required if the concentration of PCBs in the spilled
material is 50 ppm or greater, the spill and the cleanup activities normally
are not reported to EPA.
Future changes in EPA policy may invalidate some of the discussions
appearing in this chapter. For example, if EPA adopts any type of formal
categorization scheme for PCB spills, some of the assumptions made in this
chapter may become inappropriate.
B. Components of the Cleanup Process
1. Health and Safety
Protection of the health and safety of the clean-up crew during the
PCB cleanup operation is an important concern. References discussing health
and safety considerations relevant to some PCB spill incidents include NIOSH
Criteria for A Recommended Standard for Exposure to Po'lychlorinated Biphenyls
(PCBs) (1977c) and Health Hazards and Evaluation Report No. 80-85-745 (NIOSH
1980). The appropriate level of health and safety protection is dependent
upon the specifics of the spill.
2. Reporting the Spil1
If the regulatory limits are exceeded, the spill must be reported
to Federal, State, and local authorities as applicable. Under EPA regulations
[Fed. Reg. 50:13456-13475], spills over 10 Ib must be reported to The National
Response Center. The toll free phone number is (800) 424-8802.
-------
Table 1. Approximate Levels of Action for PCS Spill Cleanup Based on Concern
Type 1
Categories of increasing concern
Type 2
Type 3
Approximate gallons of
spilled liquid
Area of spil1 (sq ft)
PCB concentration in
spilled liquid
(ppm)
Types of spilled
1iquid
Exposure scenario
< 1
< 125
< 500
Mineral oil (or
variable)
Various
> 1
250 (avg.)
1 50
Variable
Various
> 5
> 1,000
Variable or high
Variable, Askarel
Special concern for high
exposure situations
Notes: • Type 1 spill is usually not reported.
• Type 2 spill is reported and discussed in this chapter.
* Type 3 spill is not discussed in this chapter and may require special
EPA assistance,
• ''Severity" in one key item may raise the spill to a higher risk category.
-------
3, Qu_ick Re s p o n s e /S ecu ring t he Site
Quick response is desirable to mitigate the dispersion of the
spilled material and to secure the site. Federal regulations require that
cleanup actions commence within 48 hr of discovery of a spill [40 CFR
761.30(a)(l) ("Mi)]. More rapid response is highly preferable,
A quick response allows removal or cleaning of the PCB-contaminated
material before it is dispersed by wind, rain, seepage, and other natural
causes or by humans or animals. In securing the site, the cleanup crew
determines the spill boundaries, prevents unauthorized access to the spill
site, and notifies all parties involved.
The methods used to secure the site will vary on a case-by-case
basis, depending on the specific circumstances. The extent of the spill is
usually determined by visual inspection with the addition of a buffer area
that may include PCBs finely dispersed from splattering. Evaluating the ex-
tent of the spill involves considerable judgment, including consideration
of the cause of the spill, weather conditions, and specifics of the site.
Field analysis kits may aid the crew in determining the extent of
the spill in some instances. The field kits, when used properly, can serve
as a screening tool. The need for quick response has limited the usefulness
of the more accurate field analytical techniques such as field gas chroma-
tography. Practical problems associated with availability of the equipment
and trained staff, set-up time, and cost have limited the use of such tech-
niques at this time.
4. Determination of Materials Spi1 led/Cleanup Plan
After securing the site, the response crew will either (a) immedi-
ately proceed with the cleanup operation, or (b) identify the materials
spilled and formulate an appropriate cleanup plan. A suitable cleanup plan
can be developed by identifying the type of PCB material (i.e., mineral oil,
PCB oil, Askarel) and considering such factors as the volume spilled, area
of the spill, and site characteristics.
Based on reasoning similar to Table 1, the crew leader can determine
the necessary level of effort in accordance with the policy of the PCB owner
and the EPA Regional Office. He can determine if additional guidance is
needed, plan the sampling and analysis, and make other decisions related to
the level of effort and procedures needed.
5. Cleanup Procedures
The cleanup procedure may include, but may not necessarily be limited
to, the following activities:
Removal or repair of failed/damaged PCB equipment,
Physical removal of contaminated vegetation;
-------
Physical removal of contaminated soils, liquids, etc.,
Decontamination or physical removal (as appropriate) of con-
taminated surfaces, and
Decontamination or removal of all equipment potentially con-
taminated during the cleanup procedures.
Encapsulation may be employed only with EPA approval,
The specific procedures used in a cleanup are selected by the PCB
owner or the cleanup contractor. Key considerations include removal of PCBs
from the site to achieve the standards required by the EPA region, company,
or other applicable control authority; avoidance of unintentional cross con-
tamination or dispersion of PCBs from workers' shoes, contaminated equipment,
spilled cleaning solvents, rags, and other sources; and protection of workers'
health.
The cleanup crew shall make every possible effort to keep the spilled
PCBs out of sewers and waterways. If this has already occurred, the crew needs
to contact the local authorities. Water is never used for cleaning equipment
or the spill site.
A simple PCB spill cleanup may involve the removal of the leaking
equipment, removal of contaminated sod and soil by shovel, cleaning pavement
with an absorbant material and solvents, and decontamination or disposal of
the workers' equipment (shovels, shoes, gloves, rags, plastic sheets, etc.).
More complicated situations may include decontamination of cars, fences,
buildings, trees and shrubs, electrical equipment, or water (in pools or
bodies of water).
In some cases, adequate decontamination of surfaces (pavements,
walls, etc.) may not be possible. An alternate to physical removal of the
surface material is encapsulation of the contaminated area under a coating
impervious to PCBs. (EPA approval would be required.)
6. Proper Oi sposa1of Removed PCB Materials
All PCB-contaminated materials removed from the spill site, must be
shipped and disposed in accordance with relevant Federal, State, and local
regulations. TSCA Regulations [40 CFR 761.60] outline the requirements for
the disposal of PCBs, PCB articles, and PCB containers in an incinerator,
high efficiency boiler, chemical waste landfill, or an approved alternative
method. Facility requirements for incineration and chemical waste landfills
are presented in 40 CFR 761.70 and 40 CFR 761.75, respectively. Applicable
Department of Transportation regulations are listed in 49 CFR 172.101.
7. Sampling and Analysis
Although sampling and analysis will be discussed in detail in Chap-
ter IV, this discussion gives an overview of applicable considerations and
current practice. Sampling and analysis may not always be needed (especially
for the spills described as Type 1 in Table 1), but enforcement authorities
or property owners may ask for proof that the spill site has been adequately
7
-------
decontaminated. This can be accomplished by taking a number of samples repre-
sentative of the area contaminated by the spill. Samples should represent
the full extent of the spill, both horizontal and vertical, as well as the
types of materials in the spill area (soil, surfaces, water, etc.).
Sampling design and technique as well as sample handling and preser-
vation should incorporate acceptable procedures for each matrix to be sampled
and concern for the adequacy and accuracy for the samples in the final analysis.
Analysis of the samples for PCB content should be performed by
trained personnel using acceptable procedures with due consideration of qual-
ity assurance and quality control.
Further discussion of sampling and analysis (applicable to EPA en-
forcement activities) appears in Chapter IV.
8. Remedial Action
If the analysis results indicate the cleanup was not in compliance
with designated cleanup levels, additional cleanup is needed. Additional
sampling can pinpoint the location of remaining contaminated areas if the
original sampling plan was not designed to identify contaminated sub-areas
within the spill site. If additional cleanup is needed, the cleanup crew will
continue as before, removing more material or cleaning surfaces more thoroughly.
Remedial action will be followed by additional sampling and analysis to ver-
ify the adequacy of the cleanup.
9. Site Restoration
This is not addressed under TSCA and is a matter to be settled be-
tween the company responsible for the PCB spill and the property owner.
10. Records
Although there are no TSCA requirements for records of PCB cleanup
activities except for documentation of PCBs stored or transported for disposal
[40 CFR 761.80(a)], the PCB owner should keep records of the spill cleanup
in case of future questions or concern. Relevant information may include
dates, a description of the activities, records of shipment and disposal of
PCB-contaminated materials, and a report of collected samples and results of
analysis.
11. Hi seel1aneous Consi derations
a. Expeditious and effective action are desired throughout the
cleanup process to minimize the concern of the public, especially residents
near the site or individuals with a special interest in the site. Likewise,
speed and effectiveness in the cleanup may prevent any future concern or action
related to the PCB spil1.
b. Education and training of the spill response crews and re-
sponsible staff members is a constant concern. The employees need sufficient
training to make proper judgements and to know when additional assistance or
guidance is needed.
-------
IV. GUIDELINES ON SAMPLING AND ANALYSIS
Reliable analytical measurements of environmental samples are an
essential ingredient of sound decisions for safeguarding public health and
improving the quality of the environment. Effective enforcement monitoring
should follow the general operational model for conducting analytical mea-
surements of environmental samples, including: planning, quality assurance/
quality control, verification and validation, precision and accuracy, sam-
pling, measurements, documentation, and reporting. Although many options are
available when analyzing environmental samples, differing degrees of reli-
ability, dictated by the objectives, time, and resources available, influence
the protocol chosen for enforcement monitoring. The following section out-
lines the factors critically influencing the outcome and reliability of en-
forcement monitoring of PCB spill cleanup.
A. Samp11 ing Design
This section presents a sampling scheme, for use by EPA enforce-
ment staff, for detecting residual PCB contamination above a limit designated
by EPA-OPTS after the site has been cleaned up. Two types of error traceable
to sampling and analysis are possible. The first is fal_se_ pos i tive, i.e.,
concluding that PCBs are present at levels above the allowable limit when, in
fact, they are not. The false positive rate for the present situation should
be low, because an enforcement finding of noncomplnance must be legally de-
fensible; that is, a violator must not be able to claim that the sampling re-
sults could easily have been obtained by chance alone. Moreover, all sampling
designs used must be documented or referenced.
The second type of error possible is a false negative, i.e. , failure
to detect the presence of PCB levels above the allowable limit. The false
negative rate will depend on the size of the contaminated area and on the
level of contamination, for large areas contaminated at levels well above
the allowable limit, the false negative rate must, of course, be low to en-
sure that the site is brought into compliance. The false negative rate can
increase as the area or level of contamination decrease.
1. Propo5ed S amp1i n g Design
In practice, the contaminated area from a spill will be irregular
in shape. In order to standardize sample design and layout in the field, and
to protect, against underestimation of the spill area by the cleanup crew, sam-
pling within a circular area surrounding the contaminated area is proposed.
Guidance on choosing the center and radius of the circle, as well as the number
of sample points to be used is provided in Section 2 below.
The detection problem was modeled as follows: try to detect a
circular area of uniform residual contamination whose center is randomly
placed within the sampling circle. Figure 1 illustrates the model. The
figure depicts a sampling circle of 10 ft centered on a utility pole (site of
the spill). After cleanup, a residually contaminated circle remains. How-
ever, in choosing locations at which to sample, the sampler has no knowledge
of either the location of the circle or the level of contamination. This
-------
Randomly Located
Area of Residual
Contamination
Sampling Circle
Figure 1. Randomly located area of residual contamination
within the sampling circle.
10
-------
lack of knowledge was modeled by treating the sampling locations as fixed and
the center of the contaminated circle as a randomly located point in the circle
of radius 10 ft. The implicit assumption that residual contamination is equally
likely to be present anywhere within the sampling area is reasonable, at least
as a first approximation (Lingle 1985). This is because more effort is likely
to have been expended in cleaning up the areas which were obviously highly
contaminated.
Two general types of design are possible for this detection problem:
grid designs and random designs. Random designs have two disadvantages com-
pared to grid designs for this application. First, random designs are more
difficult to implement in the field, since the sampling crew must be trained
to generate random locations onsite, and since the resulting pattern is ir-
regular. Second, grid designs are more efficient for this type of problem
than random designs. A grid design is certain to detect a sufficiently large
contaminated area while some random designs are not. For example, the sug-
gested design with a sample size of 19 has a 100% chance to detect a contam-
inated area of radius 2.8 ft within a sampling circle of radius 10 ft. By
contrast, a design based on a simple random sample of 19 points has only a
79% chance of detecting such an area.
Therefore, a grid design is proposed. A hexagonal grid based on
equilateral triangles has two advantages for this problem. First, such a grid
minimizes the circular area certain to be detected (among all grids with the
same number of points covering the same area). Second, some previous experi-
ence (Mason 1982; Matern 1960) suggests that the hexagonal grid performs well
for certain soil sampling problems. The hexagonal grid may, at first sight,
appear to be complicated to lay out in the field. Guidance is provided in
Section 2 below and shows that the hexagonal grid is quite practical in the
field and is not significantly more difficult to deploy than other types of
grid.
The smallest hexagonal grid has 7 points, the next 19 points, the
third 37 points as shown in Figures 2 through 4. In general, the grid has
3n2 + 3n + 1 points. To completely specify a hexagonal grid, the distance
between adjacent points, s, must be determined. The distance s was chosen to
minimize, as far as possible, the size of the residual contaminated circle
which is certain to be sampled, Values of s so chosen, together with number
of sampling points and radius of smallest circle certain to be sampled, are
shown in Table 2. For example, the grid spacing for a circle of radius 20 ft
for the 7-point design is s = (0.87)(20) = 17,4 ft. For a given size circle,
the more points on the grid, the smaller the residual contamination area which
can be detected with a given probability.
For cases in which the configuration of the contaminated area is
very different from a circle (e.g., an extremely elongated ellipse), the sam-
pling circle may be a poor approximation to the contaminated area, and a
moderate-to-large percentage of the sampling points may fall outside the con-
taminated area. If the sampler is certain that there is no contamination
outside the cleanup area, then it is permissible to disregard those sampling
points falling outside the cleanup area.
11
-------
Table 2. Parameters of Hexagonal Sampling Designs for a
Sampling Circle of Radius r Feet
No. of Distance between adjacent Radius of smallest circle
points points, s (ft) certain to be sampled
7 0.87r 0,5r
19 0.48r 0.28r
37 0.3r 0.19r
12
-------
4
3
2
Y 0 - a
D D
0
X
The outer boundary of the contaminated area
is assumed to be 4 feet from the center (C)
of the spill site .
Figure 2. Location of sampling points in
a 7-point grid.
13
-------
10
8
6
4- D D D
2
Y 0 -D a mC a D
nan
a
IQ| i i i t i i i i JL ,=l=Jt. i ; i i i i i i i
10 8 64 202468 10
X
The outer boundary of the contaminated area is assumed to be
10 feet from the center (C) of the spill site.
Figure 3. Location of sampling points in a 19-point grid.
14
-------
20 r
16
12
8
4
Y 0
4
8
12
16
20
D D a D
a o a D a
a D anna
-a a D
a a a D a D
a a
a a
' I i I i i t ...t I i l ) i i 1 l i I
20 16 12 8 4 04 8 12 16 20
X
The outer boundary of the contaminated area is assumed to be
20 feet from the center (C) of the spill site.
Figure 4. Location of sampling points in a 37-point grid.
15
-------
It is still good practice to collect samples from these outlying points even
if they are not ever analyzed because the cost of returning to the site to
perform sampling activity again is much greater than the cost of incremental
sampling performed while still onsite. If sampling points outside the con-
taminated area are ignored, and if it is a certainty that there is no con-
tamination outside this area, the absolute detection capability of the sam-
pling scheme is unaffected. For example, the chance of detecting a 5 sq ft
area of contamination within the restricted sampling area is the same as it
would be if the contaminated area comprised the entire sampling circle.
The first three hexagonal designs are shown in Figures 2 to 4, for
a sampling circle radius of r = 10 ft. the choice of sample size depends on
the cost of analyzing each sample and the reliability of detection desired
for various residually contaminated areas. Subsection 2 below provides some
suggested sample sizes for different spill areas, based on the distribution
of spill areas provided by the Utility Solid Waste Activities Group (USWAG
1984; tingle 1985).
2. Sample Size and Design Layout jn t.he_F_j_e_1_d
a. SampleSize
The distribution of cleanup areas for PCB capacitor spill
sites, based on data collected by USWAG (1984; Lingle 1985) is shown in Table
3. The smallest spill recorded in the USWAG database is 5 ft2, the largest
1,700 ft2. The median cleanup area is 100 ft, the mean 249 ft2; the wide dis-
crepancy between the mean and the median reflects the presence of a small per-
centage of relatively large spills in the database.
Recommended sample sizes are given in Table 4. Several con-
siderations were involved in arriving at these recommendations. First, the
maximum number of samples recommended for the largest spills is 37, in recog-
nition of practical constraints on the number of samples that can be taken.
Even so, it is important to note that not all samples collected will need to
be analyzed. The calculations in Section b below show that, even for the 37
sample case, no more than 8 analyses will usually be required to reach a de-
cision. Since the cost of chemical analyses is a substantial component of
sampling and analysis costs, even the 37-sample case should not, therefore,
be prohibitively expensive. Second, the typical spill will require 19 sam-
ples. Small spills, with sampling radius no greater than 4 ft, will have 7
samples, while the largest spills, with sampling radius 11.3 ft and up, will
require 37 samples. It should be noted that only capacitor spills are repre-
sented in Table 3. Transformer spills, however, would be expected to be
generally smaller than capacitor spills because energetic releases are less
likely from transformers. Thus, one would expect the s m a11e r sample sizes to
be relatively more likely for transformer spills than capacitor spills.
16
-------
Table 3. Distribution of PCB Capacitor SpiV
Cleanup Areas Based on 80 Cases
Cleanup area (ft*)
< 50
51-100
101-200
201-300
301-400
401-700
701-1,300
> 1,300
Percent of cases
32.5
18.8
15.0
12.5
3.8
7.5
8.8
1.3
Source: tingle 1985.
Table 4. Recommended Sample Sizes
Sampling area
(ft2)
£ 50
51-400
> 400
Radius of sampling
circle (ft)
5 4
4-11.3
> 11.3
Percent of PCB
capacitor spills
32.5
50.0
17.5
Sample size
7
19
37
17
-------
The final consideration in recommending sample sizes was to
achieve roughly comparable detection capability for different size spills.
The radius of the smallest contaminated circle certain to be sampled at least
once by the sampling scheme is used for comparative purposes (see Table 2).
Table 5 presents some calculations of this quantity. The absolute detection
capability of the sampling scheme is seen to be relatively constant for dif-
ferent spill sizes. This means that a given area of residual contamination
is about as likely to be detected in any sized spill.
Table 5, Detection Capability of the Recommended Sampling Schemes
Sampl ing area
(ft?)
50
150
400
875
Radius
(ft)
4,0
6.9
11.3
16.7
Sample
size
1
19
19
37
Radius of smallest circle to
be sampled (ft)
2.0
1.9
3.2
3.2
b. Design Laygut i n the Fie 1d
Figure 5 presents a typical illustration of design layout in
the field. The first step is to determine the boundaries of the original
cleanup area (from records of the cleanup). Next, find the center and radius
of the sampling circle which is to be drawn surrounding the cleanup area.
The following approach is recommended:
(a) Draw the longest dimension, Ll5 of the spill area.
(b) Determine the midpoint, P, of Lj.
(c) Draw a second dimension, U, through P perpendicular to
LI-
(d) The midpoint, C, of L2 is the required center.
(e) The distance from C to the extremes of LL is the required
radius, r.
Figure 5 shows an example of the procedure; Figure 6 demonstrates how the center
is determined for several spill shapes. Even if the center determined is
slightly off, the sampling design will not be adversely affected.
-------
Original cleanup area
(b) Locating the center of the
sampling circle
(c) Centering the hexagonal grid
(d) Staking out the grid points
Figure 5
-------
Once the sampling radius,
can be selected based on Table 4,
r, has been found, the sample size
Example: Suppose r = 5 ft. From Table 4, a sample size of 19
should be used.
Having selected the sample size, the grid spacing can be calculated from
Table 2.
Example (continued):
the grid spacing is s
For a 19-point design with radius r
= 0.48r = (0.48)(b)"= 2.4 ft.
= 5.
The
The procedure for laying out a 19 point design is as follows.
first sampling location is the center C of the sampling circle, as shown
in Figure 5. Next, draw a diameter through C and stake out locations 2
through 5 on it as shown; adjacent locations are a distance s apart. The
orientation of the diameter (for example east-west) used is not important; it
may be chosen at random or for the convenience of the samplers. The next 4
locations, Nos. 6-9, are laid out parallel to the first row, again a distance
s apart. The only difficulty is in locating the starting point, No. 6, for
this row. To accomplish this the sampler needs two pieces of rope (or sur-
veyor's chain, or equivalent measuring device) of length s. Attach one piece
of rope to the stake at each location 4 and 5. Draw the ropes taut horizontally
until they touch at location 6. Once the second row is laid out, the third
and final row of 3 locations in the top half of the design is found similarly,
starting with number 10. In the same way, the bottom half of the design is
staked out. The 7-point or 37-point designs are laid out in an analogous
fashion.
Once the sampling locations are staked out the actual samples
can be collected. In the example in Figure 5, three of the sampling locations
fall outside the original cleanup area. Samples should be taken at these
points, to detect contamination beyond the original cleanup boundaries. This
verifies that the original spill boundaries were accurately assessed. However,
if the sampler is certain that there is no contamination outside the origina
cleanup area, then it is permissible to disregard those sampling points fall-
ing outside the cleanup area. It is still good practice, however, to collect
such samples even if they are not ever analyzed because the cost of returning
to the site to sample again is much greater than the cost of incremental sam-
pling performed while still onsite. As indicated on page 16, ignoring the
sampling points outside the original cleanup area does not affect the absolute
detection capability of the sample scheme.
In practice, various obstacles may be encountered in laying
out the sampling grid. Many "obstacles" can be handled by taking a different
type of sample, e.g., if a fire hydrant is located at a point in a samp ing
grid otherwise consisting of soil samples, then a wipe sample should be taken
at the hydrant, rather than taking a sample of nearby soil. The obstacle most
likely to be encountered is a vertical surface such as a wall. To determine
the sampling location on such a surface, draw taut the ropes (chains) of
length s attached to two nearby stakes and find the point on the vertical
surface where their common ends touch. See Figure 7 for an illustration of
20
-------
Figure 6. Locating the center and sampling circle radius of an
Irregularly shaped spill area.
21
-------
Figure 7. Location of a sampling point on a vertical surface.
22
-------
the procedure. If more samples from the vertical surface are called for, the
same principle may be applied, always using the last two points located Lo
find the next one.
3. Judge me n t a I S a mp1i ng
The inspector or sampling crew may use best judgement to collect
samples wherever residual PCB contamination is suspected. These samples are
in addition to those collected from the sampling grid. Examples of extra
sampling points include suspicious stains outside the designated spill area,
cracks or crevices, and any other area where the inspector suspects inade-
quate cleanup.
4. Compositing Strategy for Analysis of Samples
Once the samples have been collected at a site, the goal of the
analysis effort is to determine whether at least one sample has a PCB concen-
tration above the allowable limit. This sampling plan assumes the entire spil"
area will be recleaned if a single sample contaminated above the limit is
found. Thus, it is not important to determine precisely which samples are
contaminated or even exactly how many. This means that the cost of analysis
can be substantially reduced by employing compositing strategies, in which
groups of samples are thoroughly mixed and evaluated in a single analysis.
If the PCB level in the composite is sufficiently high, one can conclude that,
a contaminated sample is present; if the level is j_ow enough, all individual
samples are clean. For intermediate levels, the samples from which the com-
posite was constructed must be analyzed individually to make a determination.
Thus, the number of analyses needed is greatly reduced in the presence of
very high levels of contamination in a few samples or in the presence of very
low levels in most samples.
For purposes of this discussion, assume that the maximum allowable
PCB concentration in a single soil sample is 10 ppm. The calculations can
easily be adapted for a different level or for different types of samples.
Based on review of the available precision and accuracy data (Erickson 1985),
method performance of 80% accuracy and 30% relative standard deviation should
be attainable for soil concentrations above 1 ppm.
To protect against false positive findings due to analytical €;rror,
the measured PCB level in a single, sample must exceed some cutoff greater than
10 ppm for a finding of contamination. Assume that a 0.5% false positive rate
for a single sample is desired. As will be shown later, this single sample
false positive rate controls the overal1 false positive rate of the sampling
schemes to acceptable levels. Then, using standard statistical techniques,
the cutoff level for a single sample is
(0.8)(10) + (2.576X0.3X0.8X10) = 14.2 ppm,
where 0.8(80%) represents the accuracy of the analytical method, 10 ppm is
the allowable limit for a single sample, 2.576 is a coefficient from the stan-
dard normal distribution, and 0.3(30%) is the relative standard deviation of
the analytical method. Thus, if the measured level in a single sample is
23
-------
14.2 ppm or greater, one can be 99.5% sure that the true level is 10 ppm or
greater.
Now suppose that a composite of, say, 7 samples is analyzed. The
true PCB level in the composite (assuming perfect mixing) is simply the aver-
age of the 7 levels of the individual samples. Let X ppm be the measured PCB
level in the composite. If X S (14.2/7) = 2.0, then all 7 individual samples
are rated clean. If X -> 14.2, then at least one individual sample must be
above the 10 ppm limit. If 2.0 < X S 14.2, no conclusion is possible based
on analysis of the composite and the 7 samples must be analyzed individually
to reach a decision. These results may be generalized to a composite of any
arbitrary number of samples, subject to the limitations noted below.
The applicability of compositing is potentially limited by the size
of the individual specimens and by the performance of the analytical method
at low PCB levels. First, the individual specimens must be large enough so
that the composite can be formed while leaving enough material for individual
analyses if needed. For verification of PCB spill cleanup, adequacy of speci-
men sizes should not be a problem. Ihe second limiting factor is the analyt-
ical method. Down to about 1 ppm, the performance of the stipulated analytical
methods should not degrade markedly. Therefore, since the assumed permissible
level is 10 ppm, no more than about 10 specimens should be composited at a
time.
In compositing specimens, the location of the sampling points to be
grouped should be taken into account. If a substantial residual area of con-
tamination is present, then contaminated samples will be found close together.
Thus, contiguous specimens should be composited, if feasible, in order to
maximize the potential reduction in the number of analyses produced by the
compositing strategy. Rather than describe a (very complicated) algorithm
for choosing specimens to composite, we have graphically indicated some possi-
ble compositing strategies in Figures 8 Through 11. Based on the error proba-
bility calculations presented in Section 4 below, we recommend the compositing
strategies indicated in Table 6. The recommended strategy for the 7-point
design requires no explanation. The strategies for the 19- and 37-point cases
are shown in Figures 9 and 11, respectively. The strategies shown in Figures
8 and 10 are used in Section 5 for comparison purposes. For details on the
reduction in number of analyses expected to result (as compared to individual
analyses), see the next Section, 5.
b. Calculations of Average Number of Analyses, and Error Probabil-
ities
Estimates of expected number of analyses and probabilities of fa 1 se
pos i ti ves (incorrectly deciding the site is contaminated above the limitVi
and false negatives (failure to detect residual contamination) were obtained
for various scenarios. The calculations were performed by Monte Carlo simula-
tion using 5,000 trials for each combination of sample size, compositing
strategy, level, and extent of residual contamination. The computations were
based on the following assumptions:
24
-------
A 2 GROUP COMPOSITING PLAN FOR 7 SAMPLE POINTS
Figure 8
A 2 GROUP COMPOSITING PLAN FOR 19 SAMPLE POINTS
Figure 9
25
-------
A 6 GROUP COMPOSITING PLAN FOR 19 SAMPLE POINTS
V «o to <">.*> Q> a
Figure 10, Location of sample points in a 19 sample point plan,
with detail of a 2 group compositing design.
26
-------
A 4 GROUP COMPOSITING PLAN FOR 37 SAMPLE POINTS
Figure 11. Location of sample points In 37 sample point plan,
with detail of a 4 group compositing design.
27
-------
Table 6, Recommended Compositing Strategies
No. of samples collected Compositing strategy
7 One group of 7
19 One group of 10, one of 9
37 Three groups of 9, one of 10
a. Only soil samples are involved. In practice other types
of samples will often be obtained and analyzed. Although the results of this
section are not directly applicable to such cases, they do indicate in gen-
eral terms the type of accuracy obtainable and the potential cost savings from
compositing,
b. If the true PCB level in a sample is C, then the measured
value is a normally distributed random variable with mean O.SC and standard
deviation (0.3)(0.8C) = 0.24C. Thus, it is assumed that the analytical method
is 80% accurate, with 30% relative standard deviation.
c. The maximum allowable level in a single sample is 10 ppm.
However, the measured level for a single sample must exceed 14.2 ppm for a
finding of noncompliance. As previously discussed, this corresponds to a
single-sample false positive rate of 0,5%.
d. The residual contamination present is modeled as a randomly
placed circle of variable radius and contamination level. The PCB level is
assumed to be uniform within the randomly-placed circle and zero outside it.
e. Analysis of samples is terminated as soon as a positive
result is obtained on a single analysis. If a composite does not give a de-
finitive result (positive or negative), the individual specimens from which
the composite was formed are analyzed in seguence before any other composite.
f. The compositing strategies used are shown in Figures 8 and
11.
28
-------
The results of the computations are shown in Tables 7 through 20.
Tables 7 through 12 show the performance of the compositing strategies recom-
mended in Section 3. For each strategy, there is a pair of tables. The first
table shows the probability of reporting a violation of a 10 ppm cleanup stan-
dard, for different levels of residual contamination and percent of cleanup
area contaminated. When the contamination level is 10 ppm or less, the number
in the table is the probability of a false positive, i.e., a false finding of
noncompliance. These probabilities are all very low, as they should be. When
the level is above 10 ppm, the number in the table is the probability that a
violation will be detected by the sampling design. For levels close to 10
ppm, and for small percentages of cleanup area residually contaminated, the
detection probability is low. When the level is high and the percent of area
contaminated is large, however, detection probability approaches 100%. For
small areas with high contamination, detection capability is modest. This is
because there is only a small chance that the contaminated area will be sam-
pled. Similarly, detection capability is also modest for large areas contam-
inated near the 10 ppm limit. The reason for this is that, even though a
number of contaminated samples will be found in such cases, the analytical
method is not likely to give positive identification of levels near the 10
ppm cutoff. This is the price paid for reducing the single-sample false pos-
itive rate to 0.5%.
The second table for each compositing strategy shows the expected
(average) number of analyses needed to reach a decision, for a fixed percent
of area contaminated, the smallest number of analyses is needed if the level
of contamination is very high or very low. For intermediate levels, more
analyses are needed. The largest number of analyses are required with a
large area contaminated at close to 10 ppm. In such a situation, the levels
of the composite(s) will mostly lie in the intermediate range for which no
conclusion is possible based on analysis of the composite. Thus, individual
analyses will almost always be required, so that the advantage of compositing
is lost.
Tables 13 through 20 compare the recommended compositing strategies
for the 7-point and 19-point designs to alternative compositing strategies
for these designs, for 4 different contaminated percentages (1%, 9%, 25%, and
49%). The comparison is based on the expected number of analyses required.
Overall detection capabilities are comparable for the different strategies.
The tables show that the recommended strategies are best, except for larger
areas contaminated close to the 10 ppm level.
29
-------
Table 7. Probability of Declaring a Violation of a 10 gpm
Cleanup Standard, for the 7 Point, 1 Composite Design
Level of residual
PCB contamination
(ppm)
Compliant 8
10
Noncompliant 11
12
13
14
15
16
18
20
25
50
75
100
150
200
300
500
Percent of cleanup area with residual PCB
1
< 0.
< 0.
< 0.
< 0.
0.
0.
0.
0.
0.
0.
0,
0.
0.
0,
0.
0.
0,
0,
001
001
001
001
001
003
006
009
019
030
048
070
071
068
070
073
069
070
4
< 0.
< 0.
< 0.
0.
0.
0.
0.
0,
0.
0.
0.
0.
0,
0,
0.
0.
0.
0.
001
001
001
001
005
010
016
029
074
110
186
245
245
255
246
254
257
242
9
< 0.
< 0.
< 0.
0.
0.
0,
0.
0,
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
001
001
001
001
005
019
039
064
137
199
342
487
496
499
481
489
494
492
16
< 0.
< 0.
< 0.
0.
0.
0.
0,
0.
0.
0.
0.
0.
0.
0,
0.
0.
0,
0.
001
001
001
002
009
028
065
102
218
335
554
767
787
800
796
806
792
811
contamination
25
< 0.
0,
0.
0,
0,
0.
0.
0.
0.
0,
0,
0.
0,
0.
0.
> 0.
> 0.
> 0.
,001
,002
,009
,017
,045
.085
,134
202
,344
,479
,736
,977
,992
995
998
999
999
999
49
< 0,001
0.007
0.032
0.092
0.184
0.298
0.396
0.517
0.655
0.787
0.905
0.989
0.995
0.997
0.999
> 0.999
> 0.999
> 0.999
Seven samples analyzed first as a composite, then individually if necessary
to reach a decision.
30
-------
Table 8. Expected Number of Analyses to Decide Compliance or
Violation, for a 10 ppm Cleanup Standard, for the
7-Point, 1-Composite Design
Level of residual
PCB contamination
(ppm)
Compliant
Noncompl iant
4
6
8
10
11
12
13
14
15
16
18
20
25
50
75
100
150
200
300
500
Percent
1
1.00
1,00
1.00
1.00
1.01
1.04
1.04
1.10
1.13
1.15
1.19
1.24
1.26
1.28
1.28
1.21
1.09
1.03
1.01
1.00
of cleanup
4
1.00
1.00
1.00
1.01
1.04
1.08
1.18
1.32
1.45
1.52
1.69
1.85
1.98
1.96
1.94
1.79
1.28
1.11
1.01
1.00
area with
9
1.00
1.00
1.00
1.02
1.05
1.17
1.40
1.63
1.85
2.03
2.41
2.57
2.85
2.93
2.93
2.53
1.52
1.15
1.04
1.01
residual
16
1.00
1.00
1.00
1.03
1.11
1.32
1.59
2.02
2.35
2.67
3.18
3.59
3.84
3.99
3.98
3.45
1.86
1.34
1.09
1.02
PCB contami
25
1.00
1.06
1.44
1.75
2.01
2.21
2.56
2.86
3.22
3.50
3.95
4.19
4.47
4.45
4.23
3.54
1.89
1.33
1.06
1.02
nation
49
1. 11
2.31
3.96
4.96
5.31
5.39
5.35
5.18
4.90
4.71
4.36
4.04
3.61
2.96
2.26
1.87
1.30
1.13
1.03
1.01
Seven samples analyzed first as a composite, then individually if necessary
to reach a decision.
31
-------
Table 9. Probability of Declaring a Violation of a 10 pom Cleanup
Standard, for the 19 Point, 2 Composite Design
Level of residual
PCB contamination
(ppm)
Compl iant
Noncompl iant
8
10
11
12
13
14
15
16
18
20
25
50
75
100
150
200
300
500
Percent
1
< 0.001 <
< 0.001 <
< 0.001 <
0.001
0.003
0,005
0.012
0.025
0.046
0.077
0.125
0.161
0.171
0.168
0.166
0.175
0.168
0.180
of cleanup
4
0.001 <
0.001
0.001
0.002
0.007
0.021
0.052
0.083
0.167
0.263
0.461
0.631
0.651
0.642
0.657
0.648
0.654
0.661
area with
9
0.001 <
0.002
0.007
0.029
0.062
0.114
0.178
0.264
0.421
0.556
0.784
0.978
0.993
0.994
0.998
0.999
0.999 >
0.999 >
residual
16
0.001
0.007
0.034
0.084
0.179
0.304
0.407
0.518
0.698
0.812
0.923
0.992
0.997
0.999
0.999
0.999
0.999
0.999
PCB contami
25
< 0.001 <
0.015
0.058
0.153
0.304
0.455
0.606
0.744
0.883
0.945
0.990
0,999 >
> 0.999 >
> 0.999 >
> 0.999 >
> 0.999 >
> 0.999 >
> 0.999 >
nation
49
0.001
0.028
0.017
0.281
0.497
0.693
0.832
0.908
0.978
0.993
0.999
0.999
0.999
0.999
0.999
0.999
0.999
0.999
Nineteen samples analyzed first as two composites, then individually if
necessary to reach a decision.
32
-------
Table 10. Expected Number of Analyses to Decide Compliance or
Violation, for a 10 ppm Cleanup Standard, for the
19-Point, 2-Composite Design3
Level of residual
PCB contamination
(ppm)
Compliant 4
6
8
10
Noncompliant 11
12
13
14
15
16
18
20
25
50
75
100
150
200
300
500
Percent of cleanup area with residual PCB contamination
1
2.00
2.00
2.00
2.01
2.03
2.10
2.21
2.25
2.37
2.49
2.60
2.68
2.82
2.80
2.80
2.77
2.53
2.21
1.99
1.92
4
2.00
2.00
2.00
2.03
2.14
2.32
2.74
3.02
3.40
3.84
4.36
4.65
5.02
5.03
5.05
4.95
3.94
2.67
1.89
1.69
9
2.00
2.00
3.01
3.72
4.07
4.57
4.84
5.16
5.50
5.89
6.11
6.26
6.20
5.96
5.69
5.37
3.99
2.61
1.70
1.48
16
2.18
3.79
6.15
7.46
7.90
8.08
7.94
7.90
7.65
7.30
6.57
6.18
5.45
4.70
3.68
3.46
2.59
1.91
1.50
1.39
25
3.30
6.70
9.20
10.55
10.74
10.67
9.95
9.31
8.42
7.59
6.29
5.48
4.57
3.48
2.63
2.26
1.80
1.55
1.34
1.30
49
7.49
11.22
13.18
14.02
13.81
12.78
11.00
9.27
7.80
6.63
5.02
4.25
3.36
2.28
1.84
1.69
1.46
1.33
1.19
1.16
Nineteen samples analyzed first as two composites, then individually if
necessary to reach a decision.
33
-------
Table 11. Probability of Declaring a Violation of a 10 ppm Cleanup
Standard, for the 37 Point, 4 Composite Design
Level of residual
PCB contamination
(ppm)
Compliant 8
10
Noncompliant 11
12
13
14
15
16
18
20
25
50
75
100
150
200
300
500
Percent of cleanup area with residual PCB
1
< 0.
< 0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
001
001
001
001
005
012
023
039
091
147
249
340
343
353
339
357
344
348
4
< 0.
0.
0.
0.
0.
0,
0,
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
001
002
008
024
053
094
159
242
390
542
771
976
991
993
997
996
997
999
9
< 0.
0.
0.
0.
0.
0.
0.
0,
0.
0.
0.
0.
0.
0.
> 0.
> 0.
> 0.
> 0.
001
010
041
103
224
360
501
621
785
884
958
997
999
999
999
999
999
999
16
< 0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
> 0.
> 0.
> 0.
> 0.
> 0.
001
022
084
217
388
575
740
831
940
981
995
999
999
999
999
999
999
999
contamination
25
< 0.
0,
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
> 0.
> 0,
> 0.
> 0.
> 0.
> 0.
001
031
124
305
536
726
859
936
985
996
999
999
999
999
999
999
999
999
49
< 0,001
0.060
0,225
0.488
0.751
0.908
0.950
0.991
> 0.999
> 0.999
> 0.999
> 0.999
> 0.999
> 0.999
> 0.999
> 0.999
> 0.999
> 0.999
Thirty-seven samples analyzed first as four composites, then individually if
necessary to reach a decision.
34
-------
Table 12. Expected Number of Analyses to Decide Compliance or
Violation, for a 10 ppm Cleanup Standard, for the
37-Point, 4-Composite Design
Level of residual
PCB contamination
(ppm)
Compl iant
Noncompliant
4
6
8
10
11
12
13
14
15
16
18
20
25
50
75
100
150
200
300
500
Percent
1
4,00
4.00
4.00
4.02
4.07
4.18
4.35
4.57
4.73
4.90
5.09
5.26
5.34
5.27
5.23
5.22
4.55
3.95
3.59
3.49
of cleanup
4
4.01
4.15
4,77
5.36
5.69
5.97
6.28
6.78
7.04
7.33
7.59
7.74
7.55
7.14
6.84
6.43
4.89
3.57
2.67
2.48
area with
9
4.41
6.66
9.01
10.56
10.87
10.94
10,56
10.21
9.60
9.08
8.02
7.28
6.53
5.39
4,31
3.73
3.02
2.53
2.28
2.22
residual
16
6.72
10.22
12.76
14.29
14.29
13.74
12.74
11.21
9.71
8.77
7.05
6.26
5.28
3.78
3.04
2.64
2.37
2.15
2.04
1.99
PCB contain
25
9.85
13.48
15.98
17.18
16.93
15.68
13.44
11.13
9.33
7.83
6.16
5.30
4.37
3.06
2.55
2.32
2.07
1.90
1.81
1.79
i nation
49
15.69
19.36
22.08
23.04
21.28
17.84
13.54
10.10
7.78
6.12
4.71
3.96
3.08
2.16
1.90
1.73
1.57
1.52
1.44
1.44
Thirty-seven samples analyzed first as four composites, then individually if
necessary to reach a decision.
35
-------
Table 13. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 7-Point Design, When an Area 1%
of the Size of the Cleanup Site Remains Contaminated
Level of res
contaminati
Compl iant
Noncompl iant
idual PCB
on (ppm)
4
8
10
12
14
16
20
25
50
100
200
500
1 Composite
1.00
1.00
1.00
1.04
1.10
1.15
1.24
1.26
1.28
1.21
1.03
1.00
2 Composites
2.00
2.00
2.00
2.02
2.05
2.07
2.10
2.11
2.09
1.98
1.96
1.96
Individual ly
7.00
7.00
7.00
6.98
6.96
6.92
6.88
6.84
6.80
6.78
6.80
6.81
Table 14. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 7-Point Design, When an Area 9%
of the Size of the Cleanup Site Remains Contaminated
Level of
contami
residual PCB
nation (ppm)
Compliant 4
Noncompl
8
10
iant 12
14
16
20
25
50
100
200
500
1 Composite
1.00
1.00
1.02
1.17
1.63
2.03
2.57
2.85
2.93
2.53
1.15
1.01
2 Composites
2.00
2.00
2.01
2.09
2.32
2.50
2.77
2.79
2.60
1.85
1.72
1.17
Individual ly
7.00
7.00
6.99
6.91
6.69
6.49
6.06
5.65
5.45
5.46
5.45
5.45
36
-------
Table 15, Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 7-Point Design, When an Area 25%
of the Size of the Cleanup Site Remains Contaminated
Level of res
contaminati
Compl iant
Noncompl iant
idual PCB
on (ppm)
4
8
10
12
14
16
20
25
50
100
200
500
1 Composite
1.00
1.44
1.71
2.21
2.86
3.50
4.19
4.47
4.45
3.54
1.33
1.02
2 Composites
2.00
2.13
2.24
2.44
2.84
3.23
3.54
3.56
2.97
1.61
1.38
1.37
Individual ly
7.00
7.00
6.98
6.81
6.29
5.64
4.68
4.12
3.58
3.51
3.50
3.50
Table 16. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 7-Point Design, When an Area 49%
of the Size of the Cleanup Site Remains Contaminated
Level of res
contaminati
Compl iant
Noncompl iant
idual PCB
on (ppm)
4
8
10
12
14
16
20
25
50
100
200
500
1 Composite
1.11
3.96
4.96
5.39
5.18
4.71
4.04
3.61
2.96
1.87
1.13
1.01
2 Composites
2.02
2.99
3.50
3.81
3.94
3.86
3.49
3.03
2.22
1.36
1.23
1.20
Individual ly
7.00
7.00
6.96
6.61
5.79
4.82
3.53
2.87
2.40
2.40
2.39
2.39
37
-------
Table 17. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 19-Point Design, When an Area 1%
of the Size of the Cleanup Site Remains Contaminated
Level of
contami
residual PCB
nation (ppm)
Compliant 4
Noncompl
8
10
iant 12
14
16
20
25
50
100
200
500
2 Composites
2,00
2,00
2.01
2.10
2.25
2.49
2.68
2.82
2.80
2.77
2.21
1.92
6 Composites
6.00
6.00
6.00
6.03
6.07
6.11
6.07
6.01
5.80
5.56
5.53
5.57
Individual ly
19.00
19.00
19.00
18.93
18.74
18.46
18.06
17.75
17.49
17.46
17.46
17,46
Table 18. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 19-Point Design, When an Area 9%
of the Size of the Cleanup Site Remains Contaminated
Level of residual PCB
contamination (ppm)
Compl iant
Noncompl iant
4
8
10
12
14
16
20
25
50
100
200
500
2 Composites
2.00
3.01
3.72
4.57
5.16
5.89
6.26
6.20
5.96
5.37
2.61
1.48
6 Composites
6.00
6.19
6.32
6.54
6.74
6.83
6.33
5.74
4.45
3.34
3.17
3.17
Individual ly
19.00
19.00
18.96
18.40
16.90
14.86
11.89
10.22
8.94
8.64
8.63
8.62
38
-------
Table 19. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 19-Point Design, When an Area 25%
of the Size of the Cleanup Site Remains Contaminated
Level of residual PCB
contamination (ppm)
Compl iant
Noncompliant
4
8
10
12
14
16
20
25
50
100
200
500
2 Composites
3.30
9.20
10,55
10,67
9.31
7.59
5.48
4.57
3.48
2.26
1.55
1.30
6 Composites
6.07
7.73
8.44
8,47
7.67
6.57
5.09
4.24
3.22
2.51
2.41
2.43
Individually
19.00
19.00
18.83
17.31
13.72
10.58
6.25
4.35
3.34
3.29
3.26
3.23
Table 20. Comparison of Expected Number of Analyses for Different
Compositing Strategies for the 19-Point Design, When an Area 49%
of the Size of the Cleanup Site Remains Contaminated
Level of
contami
residual PCB
nation (ppm)
Compliant 4
Noncompl
8
10
iant 12
14
16
20
25
50
100
200
500
2 Composites
7.49
13.18
14. 02
12.78
9.27
6.63
4.25
3,36
2.28
1.69
1.33
1,16
6 Composites
6.28
9.85
10.84
10.10
7.78
5.87
3.92
3.23
2.46
1.85
1.79
1.78
Individual ly
19.00
19.00
18.73
16.15
11.34
7.14
3.74
2.61
2.10
2.06
2.04
2.02
39
-------
The major conclusions that can be drawn from these results are as
follows. First, the proposed cutoff on the measured PCB level for a finding
of noncompliance for a single sample, 14.2 ppm, is successful in controlling
the overall false positive rate of the sampling scheme. For example, when an
area half the size of the entire site remains contaminated just at the allow-
able limit of 10 ppm, the false positive rate is 1% for the 7-point design,
3% for the 19-point design, and 6% for the 37-point design. Note, that the
overall false-positive rate is highest for contamination just at the allow-
able limit. Second, the detection capabilities of the design appear satis-
factory, bearing in mind the difficulty of detecting randomly-located contam-
ination by any sampling scheme without exhaustive sampling. As an example,
the proposed 19-point design can detect 50 ppm contamination present in 9% of
the cleanup area with 98% probability. Similarly, the 19-point design can
detect 20 ppm contamination present in 25% of the area with 95% probability.
Third, the proposed compositing strategies are quite effective in reducing
the number of analyses needed to reach a decision in all cases except those
involving large areas contaminated near the cutoff of 10 ppm. For example,
for contaminated levels of 25 ppm or greater, the expected number of analyses
to reach a decision never exceeds 5 for the 7-point design, or 7 for the 19-
point design, or 8 for the 37-point design. Larger number of analyses are
needed in cases of contamination close to the allowable limit of 10 ppm, up
to 23 for the 37-point design when 49% of the area is contaminated at 10 ppm.
B. Sampling Techniques
The types of media to be sampled will include soil, water, vegeta-
tion and solid surfaces (concrete, asphalt, wood, etc.). General sampling
methods are described below. Additional sampling guidance documents are avail-
able (Mason 1982, USWAG 1984).
1. Sol ids Sampling
When soil, sand, or sediment samples are to be taken, a surface
scrape samples should be collected. Using a 10 cm x 10 cm (100 cm2) template
to mark the area to be sampled, the surface should be scraped to a depth of
1 cm with a stainless steel trowel or similar implement. This should yield
at least 100 g soil. If more sample is required, expand the area but do not
sample deeper. Use a disposable template or thoroughly clean the template
between samples to prevent contamination of subsequent samples. The sample
should be scraped directly into a precleaned glass bottle. If it is free-
flowing, the sample should be thoroughly homogenized by tumbling. If not,
successive subdivision in a stainless steel bowl should be used to create a
representative subsaniple.
In some cases, such as sod, scrape samples may not be appropriate,
For these cases, core samples, not more than 5 cm deep, should be taken using
a soil coring device. These core samples should be wel1-homogenized in a
stainless steel bowl by successive subdivision. A portion of each sample
should then be removed, weighed and analyzed.
Samples should be stored in the dark at 4°C in precleaned glass
bottles. If samples are to be analyzed quickly, the storage requirements may
be relaxed as long as sample integrity is maintained. Before collection of
40
-------
verification samples, this equipment must be used to generate a field blank
as described in Section IV.E.
2. Water Sampling
a. Surface Sampl1ng
If PCBs dissolved in a hydrocarbon oil were spilled, they will
most likely be dispersed on the surface. Therefore, a surface water collec-
tion technique should be used. Surface water samples should be collected by
grab techniques. Where appropriate, the precleaned glass sample bottle may
be dipped directly into the body of water at the designated sample collection
point. A sample is collected from the water surface by gently lowering a
precleaned sample bottle horizontally into the water until water begins to
run into it. The bottle is then slowly turned upright keeping the lip just
under the surface so that the entire sample is collected from the surface.
b. Subsurface Sampling
If the PCBs were in an Askarel or other heavier-than-water
matrix, the PCBs will sink. In these cases water near the bottom should be
collected. To collect subsurface water, the bottle should be lowered to the
specified depth with the cap on. The cap is then removed, the bottle allowed
to fill, and the bottle brought to the surface.
c. Other Sampling Approaches
When the above approaches are not feasible, other dippers,
tubes, siphons, pumps, etc., may be used to transfer the water to the sample
bottle. The sampling system should be of stainless steel, Teflon, or other
inert, impervious, and noncontann'nating material. Before collection of sam-
ples, this equipment must be used to generate a field blank as described in
Section IV.E.
d. Sample Preservation
The bottle is then lifted out of the water, capped with a PTFE~
or foil-lined lid, identified with a sample number, and stored at approximately
4°C (USEPA 1984a) until analysis to retard bacterial growth. If samples are
to be analyzed quickly, the storage requirements may be relaxed as long as
sample integrity is maintained.
3. S u r face Samp ling
a. Wipe Samples
If the surface to be sampled is smooth and impervious (e.g.,
rain gutters, aluminum house siding), a wipe sample should indicate whether
the cleanup has sufficiently removed the PCBs. These surfaces should be sam-
pled by first applying an appropriate solvent (e.g., hexane) to a piece of
11 cm filter paper (e.g., Whatman 40 ashless, Whatman "50" smear tabs, or
equivalent) or gauze pad. This moistened filter paper or gauze pad is held
with a pair of stainless steel forceps and used to thoroughly swab a 100-cm2
area as measured by a sampling template.
41
-------
Care must be taken to assure proper use of a sampling template,
Different templates may be used for the variously shaped areas which must be
sampled. A 100 cm2 area may be a 10 cm x 10 cm square, a rectangle (e.g.,
1 cm x 100 cm or 5 cm x 20 cm), or any other shape. The use of a template
assists the sampler in the collection of a 100 cm2 sample and in the selec-
tion of representative sampling sites. When a template is used it must be
thoroughly cleaned between samples to prevent contamination of subsequent
samples by the template.
The wipe samples should be stored in precleaned glass jars at
4°C. Before collection of verification samples, the selected filter paper or
gauze pad and solvent should be used to generate a field blank as described
in Section IV.E.
b. Sampling Porous Surfaces
Wipe sampling is inappropriate for surfaces which are porous
and would absorb PCBs. These include wood and asphalt. Where possible, a
discrete object (e.g., a paving brick) may be removed. Otherwise, chisels,
drills, saws, etc., may be used to remove a sufficient sample for analysis.
Samples less than 1 cm deep on the surface most likely to be contaminated with
PCBs should be collected.
4. Vegetation Sampling
The sample design or visual inspection may indicate that samples of
vegetation (such as leaves, bushes, and flowers) are required. In this case,
samples may be taken with pruning shears, a saw, or other suitable tool and
placed in a precleaned glass bottle.
C. Analytical Techniques
A number of analytical techniques have been used for analysis of
PCBs in the types of samples which may be associated with PCB spills. Some
of the candidate analytical methods are listed in Table 21. The analysis
method(s) most appropriate for a given spill will depend upon a number of
factors. These include sensitivity required, precision and accuracy required,
potential interferents, ultimate use of the data, experience of the analyst,
availability of laboratory equipment, and number of samples to be analyzed.
As shown in Table 21, many analytical methods are available. The
general analytical techniques are discussed and then compared below.
1. Gas Chromatography (GC)
As can be seen in Table 21, analysis of PCBs by gas Chromatography
is frequently the method of choice. PCBs are chromatographed using either
packed or capillary columns and may be detected using either specific detec-
tors or mass spectrometry. A comprehensive method for analysis of PCBs in
transformer fluid and waste oils was developed by Bellar and Lichtenberg
(1982). This method describes six different cleanup techniques, recommends
three GC detectors, and suggests procedures for GC calibration and for mea-
surement of precision and accuracy. This method also discusses several cal-
culation methods.
42
-------
lab IP ?i
Procedure? of An*i 1 v^ i s for
Procedure
designation
D3531-80
60S
62$
30th
EPA (by-
products)
ANSI
Monsanto
UK-DOE
03304-74
EPA (homo log)
EPA 625-S
Matrix
Water
Water
Water
Water
Water-
Water
Water
Water
Ai r
Water
Soil,
sediment
Sol ids and
1 i quids
Sludge
Extraction
Hexane/CH;O2
CHZCI2
CH2C!2
Hexane/
CHZC12
(85/15)
Several
Hexane
Hexane
Hexane
DI
Hexane
lf20/Clt3CN
Several
CH2C12
Cleanup
(Mori si I)
o,/L
WehL McCall
Area 0,0-1-0. 15 (iq/L
Area 30-36 pg/L
Siintmeri areas NS
or Wfibb-McCal 1
!nd, peaks NS
Single peak or ? ppm
simtmpd peaks
individual or f pph
total peat
heights
NS 106 nq/L
Total area NS
Ind. ppnks NS
Area NS
QC
No
Yes
Yes
VPS
Ves
Yes
No
No
Yes
Yes
Ves
Reference
ASIH, 1381,1
FPA, 198<5a:
Looylioltom and
LichtenhPrn, 198?
EPA, 1934IJ.
Lontjtioll'im anil
Lichtenberg,
198?
EPA, 1178
Erickson et al. .
198?, 1933d;
FPA, 19fl4c
ANSI, 1971
Moeio, 1976
UK-DOE, 1979;
Oevenish and
llarl inq-Bonen,
1980
AS IK, 19fl!b
Ericksort el al - ,
Hai te and
Lopez-Avi la,
1984
-------
lable n (Continued)
Procedure
designation
CPA
(Ha locarbon)
Priority
Pollutant
BUM
8080
8250
a??o
EPA (spills)
EPA
Monsanto
ANSI
fPA (by-
products)
EPA (ambient
air)
Matrix
5 1 tidqe
Sludqp
Sludge
Solid waste
Sol id waste
Solid waste
Unspeei f ted
Soil and
Sediment
Serf) me Fit
Sediment,
soil
Air collected
on f lerisi 1 or
XflD-2
Air near hai-
ardous waste
Extraction
Kexane/
C.B2CI2/
acetone
(83/15/2)
CH2C12
(base/
neutral
and acid
fractions!
CHjClj,
(3 fractions)
CHSC1?
ClljClj
CHClj
Itexane/
acetone
Acetone/
Hexanp
CH3CN
CHjC.N
Hexane
Hexane/
ether
C 1 psniip
GPC
S removal
GPC
GPC
Si Hca gel
(Florisll)
None
None
(CH3CN)
(FlorisH)
{Silica gel)
(Mercury)
riorisi !
Silica qe 1
(S removal)
Saponi fi cat ion
HjSO,
Alumina
Saponif icalion
HjSO^
A 1 um i oa
(HZSO,)
(Florisi 1)
Al utn i n-3
Determination C
method ;
PGC/ECR
PGC/tlMS
HRGC/E1HS
or PGC/EIMS
PGC/ECO
PGC/EIHS
CGC/EIHS
PGC/ECD
PGC/ECD
PGC/ECD
PGC/ECD
HRGC/f IMS
PGC/ECO
Jual itative
issessneM
Yes
Ye;
¥P^
No
No
No
No
No
Nn
No
Yes
No
Quant Station
method LOD
Peak area or NS
peak height
HS MS
NS NS
Area 1 pg/g
MS 1 pg/g
US I pg/g
Total area or NS
Webb-McCall
Computer NS
Individual or ? ppb
total peak
heights
Single peak or 2 ppro
summed peaks
Intl. peaks NS
Iota! ares or 10-50 nq/m5
peak height
qc
Yes
YPS
Yet
Ye;
Yes
Yes
No
Yes
No
Yes
VPS
No
Reference
Rodriguez
et al. . 1980
EPA, n?9c
Batlinqer, 1578
EPft, 1982e
EPA, l?8Ze
EPA, 198?p
Beard and
Schaum, 1976
EPft, 198?d
Moein, 1976
ANSI, 1974
Erickson et al. ,
1982, 1983d;
f.rickson, 19B4b
lewis, 1982
sites col-
lected on PUF
-------
Table ?J (Cnntinuert)
Procedure
designation
ETA (stack)
EPA
EPA (incin-
e f it to r 5 )
ANSI
NIOSH
(PiCAH 244)
-fa
3~
lion
H?S(S4 PGC/ECD
{Florid 1) Pat/ECO No
Florisi! Pf.C/ECD No
Saponi f icalion TIC No
DorK-H
CH3CN PGC/ECO Yes
f loristl
Si 1 iea acirt
Qtpanl i tat ion
method 1 On
A red 10 nq
Area/homnlog 01 no/inj
Single peak KS
Single peak ? ppb
Peak height or 0.01 mg/m3
area from stan-
dard curve nr
Webb-HcCsl!
Peak height or O.fll mq/m'
area from stan-
dard rurve
Total area, peak 0. J-? ng/m1
height or Webb-
HcCall
(Perchlorinalion)
NS MS
H5 NS
Spsni quant. 10 ppm
Intl. ppaks 50 ppb
QC
discussed Reference
No Maile ami
Baladi , 1977;
Rear"d ar^d 5ch3um,
1179
No levins et a I .
1979
Yes Beard and
Schaum, 1978
Yes ANSI, 1.974
Ho NIOSH, 1977s
No NIOSH, I977b,c
No Harris et al . ,
1981
No Watts, 1980
Yes Watts, 1980
No Walls, 1980
Yes Watts, 1980
Sherma, 1981
-------
Table 71 (Conlinnpd)
Procedure
des i Qfiat i on
AOAC (?<»)
Japan
RAM
AOAC (29)
0330.1-74
43, D4059-83
CTi
EPA (oil)
EPA (by-
products )
rjCHA
DOW
EPA { isomer
groups)
Matrix
Fnod
Food
Food
Paper and
paperboard
Capaci tor
AsVarets
Minpra & r>i 1
T rans f oriBer
fluids sr
waste oils
Products or
was Les
3 p i qmen t
types
Chlorinated
Unspef. if ted
Extrac t i nrt
CH;5CN/Pet .
e. ther
Pet. ether/
CH3CN
Pel. pi hpr/
CHatN
Saponi tics-
It on
D.b
Oi lute with
hexane or
isooctane
D!
Several
A. Hexane/
HjSO,
B, CHjCl,
D!
Not addressed
Cleanup1"
Cel He
Sapooi f i cat ton
Silica gel
Sapon i f i cat i an
(FlorisiT)
Silicic ac iri
f Saponi f icalion)
(Oxidation)
(Flnrisit)
FToristT HqO/
Cel He
Sapnn \ f i cat. ion
None
F lori ^ i 1 si urry
(H?504 )
(F lori si 1 column)
(Ftorisil)
(Alumina)
(Silica gel)
(GPC), (CH3CN)
Scvfral
Flnriii!
None
Not addressed
Determination Qualitative Quant i tat ion
method a^^^ssment method
PfiC/ECU No Int.il area 01
\ nrl peaks
PGC/FCD YPF. SiiJimed areas
pprchtorinat inn
PCC/ECO No Area
(PGC/HltD)
(NP-TLC)
(RP-TLC)
PGC/FCD No Total area or
Ind. peaks
SCOT HRGC/FID No Total area
PGC/tCO Yes Ind, peaks or
(PGC/HECD) Wsbh-HcCall
PGC/HECD No Total area or
or /ECO or Webb-HcCall
/El MS
(HRGC)
HRfiC/EIMS Yes I«L peaks
PGC/ECD No 10 isomers
PGC/EIM") Yes Intal peak
he in.M /homo locf
HflGC/EIMS Yes Ind. peaks
QC
IOD discussed
NSa No
NS No
NS No
NSa No
2.8 * in"8 mol/t No
50 ppm No
1 mg/kg Yes
NS Yes
•>• 1 ppm/homolog Ye s
N.S YPS
NS YPS
Reference
AOAC, 1980,1
(aoabe, 197fi
FDA, 1977
AOAC. 1930b
*5TM, 19SOa
ASTM, ltfl.1
EPA, 1981
Bel lar and
Lichtenberg,
1961
Erickson et aL,
1982, 19-fl.W;
Frickson. 19P4a
RCMA, 1.982
(low, 1961
fl'A, 1984d
Source: M. D" frfckton, The AnaJyt7c^rjhem7sl.ry of PCfls , IfuHerworfFs, finstmi.'MA," I9R5, in
a Mo specific detai ts.
b Direct injection or dilute and inject.
c lechnirjues in parenth^ses are doscrihed as optional in thf? prorncliire.
d Dr PfiC with mif rocouf ontetric or elecirtilyt ir cnodocli v i. ty.
-------
a. Gas Chrpmatograph/E1ectron Capture Detection
Packed column gas chromatography with electron capture detec-
tion (GC/ECD) is generally the method of choice for analysis of spill site
samples, transformer oils, and other similar matrices which must be analyzed
for PCB content prior to disposal (Copland and Gohmann 1982). GC/ECD is very
sensitive, highly selective against hydrocarbon background, and relatively
inexpensive to operate. The technique is most appropriate when the PCB resi-
due resembles an Aroclor® (Aroclor® is a registered trademark of Monsanto
Company; the trademark designation is not used throughout this report) stan-
dard and other halogenated compounds do not interfere.
While it is considered a selective detector, ECD also detects
non-PCB compounds such as halogenated pesticides, polychlorinated naphthal-
enes, chloroaromatics, phthalate and adipate esters, and other compounds.
These compounds may be differentiated from PCBs only by chromatographic re-
tention time. Elemental sulfur can interfere with PCB analysis in sediment
and other samples which have been subjected to anaerobic degradation condi-
tions. There are also common interferences which do not give discrete peaks.
An example of a nonspecific interference is mineral oil (ASTM 1983). Mineral
oil, a complex mixture of hydrocarbons, can cause a general suppression of
ECD response. Mineral oils from transformers often contain PCBs as a result
of cross-contamination of transformer oils.
A major disadvantage of ECD is the range of response factors
which different PCB congeners exhibit. Zitko et al. (1971) and Hattori et al.
(1981) published response factors ranges of about 540 and 9000, respectively.
Boe and Egaas (1979), Onsuka et al. (1983) and Singer et al. (1983) have also
published ECD response factors. The range of response factors seriously in-
hibits reliable quantHat ion of individual PCB congeners or norr Aroclor PCBs
unless the composition of the sample and standard are the same.
When PCBs are analyzed by packed column gas chromatography,
the PCBs are usually quantitated by total areas or individual peaks. In the
total areas method, the areas of all peaks in a retention window are summed
and this total compared with the corresponding response of an Aroclor stan-
dard. With the individual peak quantitation method, response factors are
calculated for each peak in the packed column chromatogram. The most prom-
inent individual peak quantitation method was originated by Webb and McCall
(1973). These results may be reported as an Aroclor concentration or as
total PCB. Packed column GC techniques are generally useful for quantitation
of samples which resemble pure Aroclors but are prone to errors from inter-
fering compounds or from PCB mixtures that do not resemble pure Aroclors
(Albro 1979). For this reason analysts have been using capillary gas chro-
matography for the analysis of PCBs. Capillary gas chromatography offers the
analyst the ability to separate most of the individual PCB isomers. Bush
et al. (1982) has proposed a method of obtaining "total PCB" values by inte-
gration of all PCB peaks, using response factors generated from an Aroclor
mixture. Zell and BalIschmiter (1980) have developed a simplified approach
where only a selected few "diagnostic peaks" are quantitated. In a similar
approach Tuinstra et al. (1983) have quantitated six specific, diagnostic
congeners which appear to be useful for regulatory cutoff analyses.
47
-------
b. GC/Ha 11 E1 ectrolytic__Conductj_yiJLy Detector
Electrolytic conductivity detectors have also been used with
packed column gas chromatography to selectively detect PCBs (Webb and McCal1
1973, Sawyer 1978). The Hall electrolytic conductivity detector (HECD) mea-
sures the change in conductivity of a solution containing HC1 or HBr which is
formed by pyrolysis of halogenated organic GC effluents. The HECD exhibits
10r'-10G selectivity for halogenated compounds over other compounds. It also
gives a linear response over at least a 103 range. HECD and ECD were com-
pared for their use in detecting PCBs in waste oil, hydraulic fluid, capacitor
fluid, and transformer oil (Soncnik et al. 1984). They found both detectors
acceptable, but noted that the HECD gave higher results with less precision
than the ECD. The method detection limits ranged from 3-12 ppm for HECD and
2-4 ppm for ECD. Greater than 100% recovery of spikes analyzed by HECD indi-
cated a nonspecific response to non-PCB components, since extraneous peaks
were not observed. Another comparison of HECD and ECD for the analysis of
PCBs in oils at the 30-500 ppm levels found that the type of detector made no
significant difference in the results (Levine et al. 1983). The authors noted
that they had expected higher accuracy from the more specific HECD. They
postulated that the cleanup procedures (Florisil, alumina, and sulfuric acid)
all had effectively removed the non-PCB species which would have caused
interferences in the ECD and reduced its accuracy.
c, Gi C /M a s s 5 pe c t r om e t ry
Highly specific identification of PCBs is performed by GC with
mass spectrometric (GC/MS) detection. High resolution gas chromatography is
generally used with mass spectrometry, so individual PCB isomers may be
separated and identified. A GC/MS produces a chromatogram consisting of data
points at about 1 second intervals, which are actually full mass spectra. The
data are stored by a computer and may be retrieved in a variety of ways. The
data file contains information on the amount of compound (signal intensity),
molecular weight (parent ion), and chemical composition (fragmentation pat-
terns and isotopic clusters).
GC/MS is particularly suited to detection of PCBs because of
its intense molecular ion and the characteristic chlorine cluster. Chlorine
has two naturally occurring isotopes, 35C1 and 37C1, which occur in a ratio
of 100:33. Thus, a molecule with one chlorine atom will have a parent ion,
M, and an M+2 peak at 33% relative intensity. With two chlorine atoms, M+2
has an intensity of 66% and M+4, 11%.
Because of its expense, complexity of data, and lack of sensi-
tivity, GC/MS has not been used as extensively as other GC methods (particu-
larly GC/ECD), despite its inherently higher information content. As the
above factors have been improved, GC/MS has become much more popular for
analysis of PCBs, and will probably continue to increase in importance. Sev-
eral factors including the introduction of routine instruments without costly
accessories, decreasing data system costs, and mass-marketing, have combined
to keep the costs of GC/MS down while prices of other instruments have risen
steadily. With larger data systems and more versatile and "user-friendly"
48
-------
software, the large amount of data is more easily handled. However, data re-
duction of a GC/MS chromatogram still requires substantially more time than
for a GC/ECD chromatogram. In addition, the sensitivity of GC/MS has im-
proved.
d. Field- Portable Gas Chjrgrnatography Instrutnentat i on
Gas chromatography may be used for analysis of samples in the
field. Gas chromatography is a well-established laboratory technique, and
portable instruments with electron capture detectors are available (Spittler
1983, Colby et al. 1983, Picker and Colby 1984). A field-portable GC/ECD
was used to obtain rapid measurements of PCBs in sediment and soil (Spittler
1983). The sample preparation consisted of a single solvent extraction. The
PCBs were eluted from the GC within 9 min. In a 6-h period, 40 soils and
10 QC samples were analyzed, with concentrations ranging from 0.2 to 24,000
ppm. The use of field analysis permits real-time decisions in a cleanup op-
eration and reduces the need for either return visits to a site.
Mobile mass spectrometers are also available. An atmospheric
pressure chemical ionization mass spectrometer, marketed by SCIEX, has been
mounted in a van and used for J_n jnjtu analyses of soil and clay (Lovett et al.
1983). The instrument has apparently been used for field determination of
PCBs in a variety of emergency response situations, including hazardous waste
site cleanups. Other, more conventional mass spectrometers, should also be
amenable to use in the field.
2. Thin-Layer Chromatography(TLC)
Thin-layer chromatography is a well-established analytical tech-
nique which has been used for the determination of PCBs for many years.
Since the publication of a TLC method for PCBs by Mulhern (Mulhern 1968,
Mulhern et al. 1971), several researchers have used TLC to measure PCBs in
various matrices. Methods have been reported by Willis and Addison (1972)
for the analysis of Aroclor mixtures, by Piechalak (1984) for the analysis of
soils, and by Stahr (1984) for the analysis of PCB containing oils. Even with
a densitometer to measure the intensity of the spots, TLC is not generally
considered quantitative. Order-of-magnitude estimates of the concentration
are certainly obtainable, but the precision and accuracy probably do not
approach that of the gas chromatographic methods.
A spill site sample extract will probably need to be cleaned up
before TLC analysis. Levine et al. (1983) have published a comparison of
various cleanup procedures. Stahr (1984) has compared the Levine sulfuric
acid cleanup to a SepPak§ C!8 cleanup method.
Different TLC techniques have been used to improve the sensitivity
and selectivity of the method. Several researchers have reported that the
use of reverse-phase TLC (C!8-bonded phase) achieves a better separation
of PCBs from interferences (DeVos and Peet 1971, DeVos 1972, Stalling and
Huckins 1973, Brinkman et al. 1976). Koch (1979) has reported an order of
magnitude improvement in the PCB limit of detection through use of circular
49
-------
TLC. The two most common methods of visualization are fluorescence (Kan et al.
1973, Ueta et al. 1974) and reaction with AgN03 followed by UV irradiation
(DeVos and Peet 1971, DeVos 1972, Kawabata 1974, Stahr 1984).
No direct comparison of the performance of TLC with other techniques
for analysis of samples from spill sites has been made. Two studies (Bush et
al. 1975', Collins et al. 1972) compared TLC and GC/ECD. In both studies, the
PCB values obtained were comparable. However., the study by Bush et al. indi-
cated that the TLC results were generally lower than GC/ECD.
3. Total Organic Halide Analyses
Total organic halide analysis can be used to estimate PCB concen-
trations for guiding field work, but is not appropriate for verification or
enforcement analyses. A total organic halide analysis indicates the presence
of chlorine and sometimes the other halogens. Many of the techniques also
detect inorganic chlorides such as sodium chloride. The reduction of organo-
chlorine to free chloride ion with metallic sodium can be used for PCB analy-
sis. The free chloride ions can be then detected colorimetrically (Chlor-N-
Oil§) or by a chloride ion-specific electrode (McGraw-Edison). The perfor-
mance of these kits has not been tested with any matrix other than mineral
oil. X-ray fluorescence (XRF) has also been studied as a PCB screening tech-
nique (McQuade 1982, Schwalb and Marquez 1982).
D. Selection of Appropriate Methods
1. Criteria for Selection
The primary criterion for an enforcement method is that the data be
highly reliable (i.e., they are legally defensible). This does not necessarily
imply that the most exotic, state-of-the-art methods be employed; rather that
the methods have a sound scientific basis and validation data to support their
use. Many other criteria also enter into selection of a method, including
accuracy, precision, reproducibility, comparability, consistency across ma-
trices, availability, and cost.
For PCB spills, it is assumed that the spills will be relatively
fresh and therefore that PCB mixtures will generally resemble those in com-
mercial products (i.e., Aroclor©). It is further assumed that, for most of
the matrices likely to be encountered, the levels of interferences will be
relatively low.
2. Selection of InstrumentalTechniques
Based upon the above criteria and assumptions, either GC/ECD or
GC/MS should provide suitable data. Since GC/ECD is included in more stan-
dard methods and since the technique is more widely used, it appears to be
the technique of choice. The primary methods recommended below are all based
on GC/ECD instrumental analysis. Some of the secondary and confirmatory tech-
niques are based on GC/EIMS.
50
-------
3. Selection of Methods
Ideally, a standard method would be available for each matrix likely
to be encountered in a PCB spill. The matrices of concern include solids (soil
sand, sediment, bricks, asphalt, wood, etc.), water, oil, surface wipes, and
vegetation. The methods for these matrices are summarized in Table 22 and
discussed in detail below. A primary recommended method is given and should
be used in most spill instances. The secondary method may be useful for con-
firmatory analyses, or where the situation (e.g., high level of interferences)
indicates that the primary method is not applicable. The methods used must
be documented or referenced.
a. Solids (Soil, Sand, Sediment, Bricks, Asphalt, Wood, Etc.)
EPA Method 8080 from SW-846 (USEPA 1982e) is the primary recom-
mended method. The secondary methods, Method 8250 and Method 8270, are GC/MS
analogs. Method 8080 entails an acetone/hexane (1:1) extraction, a Florisil
column chromatographic cleanup, and a GC/ECD instrumental determination. A
total area quantitation versus Aroclor standards is specified. No qualitative
criteria are supplied. A detection limit of 1 ug/g is prescribed. No valida-
tion data are available.
Bulk samples (bricks, asphalt, wood, etc.) should be readily
extractable using a Soxhlet extractor according to EPA Method 8080 (USEPA
1982e). The sample must be crushed and subsampled to ensure proper solvent
contact.
b. Water
EPA Method 608 (USEPA 1984e) is recommended as the primary
method. This is one of the "priority pollutant" methods and involves extrac-
tion of water samples with dichloromethane. An optional Florisil column
chromatographic cleanup and also an optional sulfur removal are given. Sam-
ples are analyzed by GC/ECD and quantitated against the total area of Aroclor
standards. No qualitative criteria are given. This method has been exten-
sively validated and complex recovery and precision equations are given in
the method for seven Aroclor mixtures. The average recovery is about 86% and
average overall precision about ± 26%. The average recovery and precision
for the more common Aroclors (1242, 1254, and 1260) are about 78% and ± 26%,
respectively. Detection limits are not given in the current version (USEPA
1984a), although they were listed as between 0.04 and 0.15 M9/L for the seven
Aroclor mixtures listed as priority pollutants in the method validation study
(Millar et al. 1984).
c. Oils
Spilled oil samples should be analyzed according to an EPA
method (Bellar and Lichtenberg 1981). The method is written for transformer
fluids and waste oils, but should also be applicable to other similar oils
such as capacitor fluids. In this method, samples are diluted by an appro-
priate factor (e.g., 1:1000). Six optional cleanup techniques are given.
51
-------
Table 22. Summary of Recommended Analytical Methods
ro
Matrix
Solids
Water
Oil
Surface
wipes
Primary method
Designation
8080
608
"oil"
Hexane extrac-
tion/608
(GC/ECD)
Reference
USEPA 1982e
USEPA 1984a
USEPA 1981a;
Bellar and
Lichtenberg,
1981
None
Designation
8250, 8270
625
"oil"
Hexane extrac-
tion/625
Secondary method
GC detector
MS
MS
MS
MS
Reference
USEPA 1982e
USEPA 1984b
USEPA 1981a;
Bellar and
Lichtenberg,
1981
None
Vegetation
AOAC (29)
AOAC 1980a
None
None
None
-------
JUI
The sample may be analyzed by GC/ECD as the primary method. Secondary instru-
mental choices, also presented in the method, are GC/HECD, GC/MS, and capil-
lary GC/MS. PCBs are quantitated by either total areas or the Webb-McCall
(1973) method. No qualitative criteria are given. QC criteria are given. A
detection limit of 1 mg/kg is stated, although it is highly dependent on the
amount of dilution required. An interlaboratory validation study (Sonchik
and Ronan 1984) indicated 81 to 126% recoveries for different PCB mixtures,
with an average of 97% for Aroclors 1242, 1254, and 1260, as measured by ECD.
The overall method precision ranged from ± 11 to ± 55%, with an average of
± 12% for Aroclors 1242, 1254, and 1260. The method validation statistics
were presented in more detail as regression equations.
d. Surface Wipes
No standard method is available for analysis of PCBs collected
on surface wipes. However, since this matrix should be relatively clean and
easily extractable, a simple hexane extraction should be sufficient. Samples
should be analyzed according to EPA Method 608 (USEPA 1984a), except for
Section 10,1 through 10.3. In lieu of these sections, the sample should be
extracted three times with 25 to 50 ml of hexane. The sample can be extracted
by shaking for at least 1 min per extraction in the wide-mouthed jar used for
sample storage. Note that the rinses should be with hexane so that solvent
exchange from methylene chloride to hexane (Section 10.7) is not necessary.
e. Vegetation
The AOAC (1980a) procedure for food is recommended for analysis
of vegetation (leaves, vegetables, etc.). This method involves extraction of
a macerated sample with acetonitrile. The acetonitrile is diluted with water
and the PCBs extracted into petroleum ether. The concentrated extract is
cleaned up by Florisil column chromatography by elution with a mixture of ethyl
ether and petroleum ether. The sample is analyzed by GC/ECD with quantitation
by total areas or individual peak heights as compared to Aroclor standards.
No qualitative criteria are given. Validation studies with chicken fat and
fish (Sawyer 1973) are not relevant to the types of matrices to be encountered
in PCB spills.
4. Implementation of Methods
Each laboratory is responsible for generating reliable data. The
first step is preparation of an in-house protocol. This detailed "cookbook"
is based on methods cited above, but specifies which options must be followed
and provides more detail in the conduct of the techniques. It is essential
that a written protocol be prepared for auditing purposes.
53
-------
Each laboratory is responsible for generating validation data to
demonstrate the performance of the method in the laboratory. This can be
done before processing of samples; however, it is often impractical. Valida-
tion of method performance (replicates, spikes, QC samples, etc.) while ana-
lyzing field samples is acceptable.
Changes in the above methods are acceptable, provided the changes
are documented and also provided that they do not affect performance. Some
minor changes (e.g., substitution of hexane for petroleum ether) do not
generally require validation. More significant changes (e.g., substitution
of a HECD for ECD) will require documentation of equivalent performance.
E. Duality Assurance
Quality assurance must be applied throughout the entire monitoring
program including the sample planning and collection phase, the laboratory
analysis phase, and the data processing and interpretation phase.
Each participating EPA or EPA contract laboratory must develop a
quality assurance plan (QAP) according to EPA guidelines (USEPA 1980). Ad-
ditional guidance is also available (USEPA 1983). The quality assurance plan
must be submitted to the regional QA officer or other appropriate QA official
for approval prior to analysis of samples.
1. Quality AssurancePlan
The elements of a QAP (U.S. EPA, 1980) include:
Title page
Table of contents
Project description
Project organization and responsibility
QA objectives for measurement data in terms of precision, ac-
curacy, completeness, representativeness, and comparability
Sampling procedures
Sample tracking and traceability
Calibration procedures and frequency
Analytical procedures
Data reduction, validation and reporting
Internal quality control checks
Performance and system audits
Preventive maintenance
Specific routine procedures used to assess data precision,
accuracy and completeness
Corrective action
Quality assurance reports to management
54
-------
2.
Each laboratory that uses this method must operate a formal quality
control (QC) program. The minimum requirements of this program consist of an
initial and continuing demonstration of acceptable laboratory performance by
the analysis of check samples, spiked blanks, and field blanks. The labora-
tory must maintain performance records which define the quality of data that
are generated.
The exact, quality control measures will depend on the laboratory,
type and number of samples, and client requirements. The QC measures should
be stipulated in the QA Plan. The QC measures discussed below are given for
example only. Laboratories must decide on which of the measures below, or
additional measures, will be required for each situation.
a. Protocol s
Virtually all of the available PCB methods contain numerous
options and general instructions. Effective implementation by a laboratory
requires the preparation of a detailed analysis protocol which may be followed
unambiguously in the laboratory. This document should contain working instruc-
tions for all steps of the analysis. This document also forms the basis for
conducting an audit,
b. Certification and Performance Checks
Prior to the analysis of samples, the laboratory must define
its routine performance. At a minimum, this must include demonstration of
acceptable response factor precision with at least three replicate analyses
of a calibration solution; and analysis of a blind QC check sample (e.g., the
response factor calibration solution at unknown concentration submitted by an
independent QA officer). Acceptable criteria for the precision and the ac-
curacy of the QC check sample analysis must be presented in the QA plan.
Ongoing performance checks should include periodic repetition
of the initial demonstration or more elaborate measures. More elaborate mea-
sures may include control charts and analysis of QA check samples containing
unknown PCBs, and possibly with matrix interferences.
c. Procedural QC
The various steps of the analytical procedure should have qual-
ity control measures. These include, but are not limited to, the following:
I ns t r ume n t a 1 P e r f q rma nc e : Instrumental performance cri-
teria and a system for routinely monitoring the performance should be set out
in the QA Plan. Corrective action for when performance does not meet the
criteria should also be stipulated.
55
-------
Qualitative Identification: Any questionable results
should be confirmed by a second analytical method, A least 10% of the
identifications, as well as any questionable results, should be confirmed
a second analyst.
by
(juant itati on: At least 10% of all calculations must be
checked. The results should be manually checked after any changes in computer
quantisation routines.
d.
Each sample and each sample set must have QC measures applied
to it to establish the data quality for each analysis result. The following
should be considered when preparing the QA plan:
Field _B1anks: Field blanks are analyzed to demonstrate
that the sample collection equipment has not been contaminated. A field blank
may be generated by using the sampling equipment to collect a blank sample
(e.g., using the water sampling equipment to sample laboratory reagent grade
water) or by extracting the sampling equipment (e.g., extracting a sheet, of
filter paper from the lot used to collect wipe samples or rinsing the soil
sampling apparatus into the sample jar). A field blank must be collected and
analyzed for each type of sample collected.
L a bo r a t o ry R e a g e n t B_j_an_ ks: These blanks are generated in
the laboratory and are analyzed to assess contamination of glassware, reagents,
etc., in the laboratory. Generally, a reagent blank is processed through the
entire analysis process. Although in special circumstances, additional reagent
blanks may be generated which are processed through only part of the procedure
to isolate sources of contamination. At least one laboratory reagent blank
must be generated and analyzed for each type of sample analyzed.
Check Samples:
of PCBs in the sample matrix. They
demonstrate the method performance.
the analyst.
These samples contain known concentrations
are analyzed along with field samples to
The PCB concentrations may be known to
B 1 i n d C h eg k Samples:
check samples discussed above, except the
the analyst.
These samples are
PCB concentration
the same as the
is not known to
Replicate Samples: One sample from each batch of 20 or
fewer will be analyzed in triplicate. The sample is divided into three rep-
licate subsamples and all these subsamples carried through the analytical pro-
cedure, blind to the analyst. The results of these analyses must be compara-
ble within the limits required for spiked samples.
Spj_k^o^_SajiipJ_es_: The sensitivity and reproducibi I ity must
be demonstrated for any method used to report verification data. This can be
done by analyzing spiked blanks near the required detection limit. To demon-
strate the ability of the method to reproducibly detect the spiked sample,
one or more spiked samples should be analyzed in at least triplicate for each
group of 20 or fewer samples within each sample type collected. Samples will
56
-------
be spiked with a PCB mixture similar to that spilled (e.g., Aroclor 1260).
Example concentrations are:
Matrix Spi ke Level
Soil, etc, 10 M9/9 (10 ppro)
Water 100 ug/L (100 ppb})
Wipes 100 ug/wipe (100 ug/100 cm2)
Quantitative techniques must detect the spike level within ±30% for all spiked
samples.
e. SampleCustody
As part of the Quality Assurance Plan, the chain-of-custody
protocol must be described. A chain-of-custody provides defensible proof of
the sample and data integrity. The less rigorous sample traceability docu-
mentation merely provides a record of when operations were performed and by
whom. Sample traceability is not acceptable for enforcement activities.
Chain-of-custody is required for analyses which may result in
legal proceedings and where the data may be subject to legal scrutiny.
Chain-of-custody provides conclusive written proof that samples are taken,
transferred, prepared, and analyzed in an unbroken line as a means to maintain
sample integrity. A sample is in custody if:
It is in the possession of an authorized individual;
It is in the field of vision of an authorized
individual;
It is in a designated secure area; or
It has been placed in a locked container by an
authorized individual.
A typical chain-of-custody protocol contains the following elements:
1. Unique sample identification numbers.
2. Records of sample container preparation and integrity
prior to sampling.
3. Records of the sample collection such as:
Specific location of sampling.
57
-------
Date of collection.
Exact time of collection.
Type of sample taken (e.g., air, water, soil).
Initialing each entry.
Entering pertinent information on chain-of-
custody record.
Maintaining the samples in one's possession or
under lock and key.
Transporting or shipping the samples to the
analysis laboratory.
Filling out the chain-of-custody records.
The chain-of-custody records must accompany the
samples,
4. Unbroken custody during shipping. Complete shipping
records must be retained; samples must be shipped in
locked or sealed (evidence tape) containers.
5. Laboratory chain-of-custody procedures consist of:
Receiving the samples.
Checking each sample for tampering.
Checking each sample against the chain-of-custody
records.
Checking each sample and noting its condition.
Assigning a sample custodian who will be responsible
for maintaining chain-of-custody.
Maintaining the sign-offs for every transfer of each
sample on the chain-of-custody record.
Ensuring that a 11 manipulations of the sample are
duly recorded in a laboratory notebook along with
sample number and date. These manipulations will
be verified by the program manager or a designee.
F. Documentation and Records
Each laboratory is responsible for maintaining complete records of
the analysis. A detailed documentation plan should be prepared as part of
58
-------
the QAP. Laboratory notebooks should be used for handwritten records. Digi-
tal or other GC/MS data must be archived on magnetic tape, disk, or a similar
device. Hard copy printouts may also be kept if desired. Hard copy analog
data from strip chart recorders must be archived. QA records should also be
retained.
The documentation must completely describe how the analysis was
performed. Any variances from a standard protocol must be noted and fully
described. Where a procedure lists options (e.g., sample cleanup), the op-
tion used and specifics (solvent volumes, digestion times, etc.) must be
stated.
The remaining samples and extracts should be archived for at ~east
2 months or until the analysis report is approved by the client organization
(whichever is longer) and then disposed unless other arrangements are made.
The magnetic disks or tapes, hard copy chromatograms, hard copy spectra, quan-
titation reports, work sheets, etc., must be archived for at least 3 years.
All calculations used to determine final concentrations must be documented.
An example of each type of calculation should be submitted with each verifi-
cation spot.
G. Reporting Results
Results of analysis will normally be reported as follows:
HatHx Reporting Units
Soil, etc. ug PCB/g of sample (ppm)
Water mg PCB/L of sample (ppm)
Surfaces (wipes) ug PCB/wipe (jjg PCB/100 cm2)
In some cases, the results are to be reported by homolog. In this
case, 11 values are reported per sample: one each for the 10 homo logs and
one for the total. Some TSCA analyses require reporting the results in terms
of resolvable gas chromatographic peak (U.S. EPA, 1982c, 1984e). In these
cases, the number of results reported equals the number of peaks observed on
the chromatogram. These analyses are generally associated with a regulatory
cutoff (e.g., 2 ug/g per resolvable chromatographic peak (U.S. EPA, 1982c,
1984). In these cases it may be sufficient, depending on the client organi-
zation's request, to report only those peaks which are above the regulatory
cutoff.
Even if an Aroclor is used as the quantitation standard, the re-
sults are never to be reported as "u.g Aroclor®/g sample." TSCA regulates all
PCBs, not merely a specific commercial mixture.
59
-------
V. REFERENCES
Albro PW. 1979. Problems in analytic methodology:
tion, and cleanup. Ann NY Acad Sci 320:19-27,
American National Standards Institute, Inc. 1974.
guidelines for handling and disposal of capicator-
karels containing polychlorinated biphenyls. ANSI
sample handling, extrac-
American national standard
and transformer-grade as-
C107.1-1974. New York, NY.
American Society for Testing and Materials. 1980. Standard method for rapid
gas chromatographic estimation of high boiling homologues of chlorinated bi-
phenyls for capacitor askarels. ANSI/ASTM 0 3303-74 (Reapproved 1979). In:
Annual book of ASfM standards, Part 40. Philadelphia, Pennsylvania, pp.
870-876,
American Society for Testing
chlorinated biphenyls (PCBs)
of ASTM standards, Part 31.
and Materials. 1981a. Standard method for poly-
in water. ANSI/ASTM D 3534-80. In: Annual book
Philadelphia, Pennsylvania, pp. 816-833.
American Society for Testing and Materials. 1981b, Standard method for
analysis of environmental materials for polychlorinated biphenyls. ANSI/ASTM
D 3304-77. In: Annual book of ASTM standards, Part 31, Philadelphia,
Pennsylvania, pp. 877-885.
American Society for Testing and Materials.
ysis of polychlorinated biphenyls in mineral
tography. ANSI/ASTM 0 4059-83. In: Annual
Philadelphia, Pennsylvania, pp. 542-550.
1983. Standard method for anal-
insulating oils by gas chroma-
book of ASTM standards, Part 40.
Association of Official Analytical Chemists. 1980a. General method for
organochlorine and organophosphorus pesticides, Method 29.001. Official Meth-
ods ofAnalysis of the Association of Official Analytical Chemists, W. Horwitz
Ed. (13th ed. , Washington, DC), pp. 466-474.
Association of Official Analytical Chemists. 1980b. PCB in paper and paper-
board, Method 29.029. Official Methods of Analysis of the Association of
Official Analytical Chemists. W. Horwitz, Ed. (13th ed., Washington, DC),
p. 475-476.
Ballinger DG. 1978 (December 11).
municipal wastewater and sludges,
Cincinnati, Ohio.
Test procedures for priority organics in
U.S. Environmental Protection Agency,
Beard JH III, Scnaum J. 1978 (February 10). Sampling methods and analytica
procedures manual for PCB disposal: Interim Report, Revision 0. Office of
Solid Waste, U.S. Environmental Protection Agency, Washington, DC.
Bellar TA, Lichtenberg JJ. 1982. The determination of polychlorinated bi-
phenyls in transformer fluid and waste oils. Prepared for U.S. Environmenta
Protection Agency, EPA-600/ 4-81-045.
60
-------
i
Boe B, Egaas E. 1979. Qualitative and quantitative analyses of polychlori-
nated biphenyls by gas-liquid chrotnatography. J Chromatogr 180:127-132.
Brinkman UATh, De Kok A, De Vries G, Reymer HGM, 1976. High-speed liquid
and thin-layer chromatography of polychlorinated biphenyls. J Chromatogr
128:101-110.
Bush B, Baker F, Dell'Acqua R, Houck CL, Lo F-C. 1975. Analytical response
of polychlorinated biphenyl homologues and isomers in thin-layer and gas chro-
matography. J Chromatogr 109:287-295.
Bush B, Connor S, Snow J. 1982. Glass capillary gas chromatography for sen-
sitive, accurate polychlorinated biphenyl analysis. J Assoc Off Anal Chem
65(3):555-566.
Colby BN, Burns EA, Lagus PL. 1983. The S-Cubed PCBA 101, an automated field
analyzer for RGBs. Abstract No. 731, 1983 Pittsburgh Conference and Exposi-
tion on Analytical Chemistry and Applied Spectroscopy.
Collins GB, Holmes DC, Jackson FJ. 1972. The estimation of polychloro-
biphenyls. J Chromatogr 71:443-449.
Copland GB, Gohmann CS. 1982. Improved method for polychlorinated biphenyl
determination in complex matrices. Environ Sci Technol 16:121-124.
De Vos RH. 1972. Analytical techniques in relation to the contamination of
the fauna. TNQ-nieuws 27:615-622.
De Vos RH, Peet EW. 1971. Thin-layer chromatography of polychlorinated bi-
phenyls. Bull Environ Contam Toxicol 6(2):164-170.
Devenish I, HarIing-Bowen L. 1980. The examination and estimation of the
performance characteristics of a standard method for organo-chlorine in-
secticides and PCB. In: Hydrocarbons and Halogenated Hydrocarbons in the
Aquatic Environment, B. K. Afghan and D. Mackay, Eds. (New York; Plenum
Press), pp. 231-253.
Oow Chemical Company. 1981 (July 1). Determination of chlorinated biphenyls
in the presence of chlorinated benzenes. Midland MI.
Dry Color Manufacturers Association. 1981, An analytical procedure For the
determination of polychlorinated biphenyls in dry phthalocyanine blue, phthalo-
cyanine green, and diarylide yellow pigments, Arlington, VA.
Erickson MD, Stanley JS, Turman K, Radolovich G, Bauer K, Onstot J, Rose D,
Wicknam M. 1982. Analytical methods for by-product PCBs--preliminary vali-
dation and interim methods. Interim Report No, 4, Office of Toxic Substances,
U.S. Environmental Protection Agency, Washington, 0. C., EPA-560/5-82-006,
NTIS No, PB83 127 696, 243 pp.
61
-------
Erickson MD, Stanley JS, Radolovich G, Blair RB. 1983 (August 15). Analyt-
ical method: the analysis of by-product chlorinated biphenyls in commerical
products and product wastes. Revision 1, Prepared by Midwest Research Insti-
tute for Office of Toxic Substances, U.S. Environmental Protection Agency,
Washington, DC, under Subcontract No. A-3044(8149)-271, Work Assignment No. 17
to Battelle, Washington, DC.
Erickson MD 1984a. Analytical method: The analysis of by-product chlo-
rinated biphenyls in commercial products and product wastes, revision 2,
U.S. Environmental Protection Agency, Office of Toxic Substances, Washington,
DC, EPA 560/5-85-010.
Erickson MD. 1984b, Analytical method: The analysis of by-product chlo-
rinated biphenyls in water, revision 2. U.S. Environmental Protection
Agency, Office of Toxic Substances, Washington, DC, EPA 560/5-85-012.
Erickson
MA.
MD, 1985. The analytical chemistry of PCBs. Butterworths, Boston,
Erickson MO, Stanley JS, Turman JK, and Radolovich G. 1985a. Analytical
method: The analysis of chlorinated biphenyls in liquids and solids. U.S.
Environmental Protection Agency, Office of Toxic Substances, Washington,
DC, EPA-560/5-85-Q23.
Fisher DJ, Rouse TO, Lynn TR.
former oil "CLOR-N-OIL Kit."
and Komai RY, Eds. Report No.
Research Institute.
Food and Drug Administration.
1977.
1984. Field determination of PCB in trans-
In: Proceedings: 1983 PCB Seminar, Addis G,
EPRI-EL-3581, Palo Alto, CA: Electric Power
Pesticide Analytical Manual. Vol. 1, August 1,
Haile CL, Baladi E. 1977. Methods for determining the total polychlorinated
biphenyl emissions from incineration and capacitor and transformer filling
plants. U. S. Environmental Protection Agency, EPA-600/4-77-048, NTIS No.
PB-276 745/7G1.
Haile CL, Lopez-Avi'la V. 1984. Development of analytical test procedures for
the measurement of organic priority pollutants—project summary. U.S. Envi-
ronmental Protection Agency, Environmental Monitoring and Support Laboratory,
Cincinnati, Ohio, EPA-600/S4-84-001; (Full Report available as NTIS No.
PB 84-129 048).
Harris RW, Grainger CF, Mitchell WJ. 1981, Validation of a method for mea-
suring polychlorinated biphenyls in natural gas pipelines. EPA 600/4-81-048;
NTIS No. PB82-207556.
Hattori Y, Kuge Y, Nakamoto M. 1981. The correlation betv/een the electron-
capture detector response and the chemical structure for polychlorinated bi-
phenyls. Bull Cnem Soc Jpn 54(9):2807-2810; Chem Abstr 96:34427s (1981).
62
-------
Kan I, Kamata K, Ueta T, Yamazoe R, Totani T. 1973, Fluorescence reactions
of organohalogen compounds. I. Fluorometry of polychlorinated biphenyls
(PCB) with diphenylamine on thin-layer chromatograms. Tokyo Torit.su Eisei
Kenkyusho Kenky Nempo 24:137-145; Chem Abst 80:115771w (1974).
Kawabata J. 1974. Simple method for the determination of PCBs [polychlorin-
ated biphenyls] by a combination of thin-layer chromatography and UV absorp-
tion. Kogai To Taisaku 10(10):1112-1115; Chem Abst 83:201652b (1975).
Koch R. 1979. Circular thin-layer chromatography as a rapid method for a
qualitative detection of organochlorine compounds. Acta Hydrochim Hydrobiol
7(3):355-356; Chem Abst 91:101574z (1979).
Levine SP, Homsher MT, Sullivan JA. 1983. Comparison of methods of analysis
of polychlorinated biphenyls in oils. J Chromatogr 257:255-266.
Levins PL, Rechsteiner CE, Stauffer JL. 1979. Measurement of PCB emissions
from combustion sources. U.S. Environmental Protection Agency, Report No.
EPA-600/7-79-047.
Lewis RG. 1982 (March). Procedures for sampling and analysis of polychlori-
nated biphenyls in the vicinities of hazardous waste disposal sites. U.S.
Environmental Protection Agency, Research Triangle Park, NC, 14 pp.
Lingle JW. Wisconsin Electric Power Company, P.O. Box 2046, Milwaukee, WI
53201. May 24, 1985. Personal communication.
Longbottom JE, Lichtenberg JJ, Eds. 1982 (July). Methods for organic chem-
ical analysis of municipal and industrial wastewater. U.S. Environmental
Protection Agency, Report No. EPA-600/4~82~057.
Lovett AM, Nacson S, Hijazi NH, Chan R. 1983. Real time ambient air mea-
surements for toxic chemical. In: Proceedings: a specialty conference on:
measurement and monitoring of non-criteria (toxic) contaminants in air,
Frederick ER, Ed., The Air Pollution Control Association, Pittsburgh, PA,
113-125 pp.
Mason BJ. 1982 (October). Preparation of soil sampling protocol: techniques
and strategies. ETHURA, McLean, VA, under subcontract to Environmental Re-
search Center, University of Nevada, for U.S. Environmental Protection Agency,
Las Vegas.
Matern B. 1960. Spacial variation. Medd. fr. Stateus Skogsforsknings
Institut. 49:1-144.
McQuade JM. 1982. PCB analysis by X-ray fluorescence. In: Proceedings:
1981 PCB Seminar, Addis G, Harks J, Eds., Report No. EPRI-EL-2572, Palo Alto,
CA: Electric Power Research Institute, pp. 2-9.
Millar JD, Thomas RE, Schattenberg HJ. 1984 (June). EPA Method Study 18,
Method 608--organochlorine pesticides and PCB'a. Quality Assurance Branch,
Environmental Monitoring and Support Laboratory, U.S. Environmental Protection
Agency, Cincinnati, Ohio. Report No. EPA-600/4-84-061, NT1S No. PB84 211358,
197 pages.
63
-------
Moein GJ. 1976. Study of the distribution and fate of po'lychlori nated bi-
phenyls and benzenes after spill of transformer fluid. Report No. EPA
904/9-76-014, NT1S No. PB288484.
Mulhern BM. 1968. An improved method for the separation and removal of
organochlorine insecticides from thin-layer plates. J Chromatogr 34:556-558.
Mulhern BM, Cromartie E, Reichel WL, Belisle A. 1971. Semiquantitative de-
termination of polychlorinated biphenyls in tissue samples by thin layer
chromatography. J Assoc Offie Anal Chem 54(3):548-550.
National Institute for Occupational Safety and Health. 1977a (April). NIOSH
Manual of A n a 1y tic a1 Methods, Second Edition, Part I, NIOSH Monitoring Methods,
Vol. 1, "Polychlorinated Biphenyls (PCB) in Air, Analytical Method P&CAH 244,"
U.S. Department of Health, Education, and Welfare, Cincinnati, Ohio.
National Institute for Occupational Safety and Health. 1977b (April). NIOSH
Manual of Analytical Methods, Second Edition, Part I, NIOSH Monitoring Methods,
Vol. 1, "Polychlorinated Biphenyls (PCB) in Air, Analytical Method P&CAM 253,"
U.S. Department of Health, Education, and Welfare, Cincinnati, Ohio.
NIOSH. 1977c (September). National Institute for Occupational Safety and
Health. Criteria for a recommended standard.... occupational exposure to poly-
chlorinated biphenyls (PCBs). U.S. Department of Health, Education, and
Welfare (Public Health Service, Center for Disease Control, and National
Institute for Occupational Safety and Health), DHEW (NIOSH) Publication No.
7-225, 224 pp.
NIOSH. 1980 (September). National Institute for Occupational Safety and
Health, U.S. Department of Health and Human Services. Health Hazard Evalua-
tion Report No. 80-85-745. Oakland, CA: Pacific Gas and Electric Company.
Onsuka FI, Kominar RJ, Terry KA. 1983. Identification and determination of
Polychlorinated biphenyls by high-resolution gas chromatography. J Chromatogr
279:111-118.
Picker JE, Colby BN. 1984. Field determination of Aroclors using an auto-
mated electron capture detector gas chromatograph. In: Proceedings: 1983
PCB Seminar, Addis G, Komai RY, Eds., Report No. EPRI-EL-3581. Palo Alto,
CA: Electric Power Research Institute.
Piechalak B. 1984. The semiquantitative detection of polychlorinated bi-
phenyls (PCBs) in contaminated soils by thin-layer chromatography. In:
Proceedings: 1983 PCB Seminar, Addis G, Komai RY, Eds., Report No. EPRI-EL-
3581. Palo Alto, CA; Electric Power Research Institute.
Rodriguez CF, McMahon- WA, Thomas RE. 1980 (March). Method development for
determination of polychlorinated hydrocarbons in municipal sludge. Final
Report, Contract No. 68-03-2606, Environmental Protection Agency, EPA-6QO/
2-80-029; NTIS No. PB 82-234 071.
64
-------
Sawyer ID. 1973. Collaborative study of the recovery and gas chromatographic
quantitation of polychlorinated biphenyls in chicken fat and polychlorinated
biphenyl-DDT combinations in fish. J Assoc Offic Anal Chem 56(4):1015-1023.
Sawyer LD. 1978. Quantitation of polychlorinated bipheny'l residues by elec-
tron capture gas-liquid chromatography: reference material characterization
and preliminary study. J Assoc Offic Anal Chem 61(2):272-281.
Schwalb AL, Marquez A. 1982. Salt River Project's experience with the Horiba
Sulfur/Chlorine-in-Oi1 Analyzer. In: Proceedings: 1981 PCB Seminar, Addis
G, Marks J, Eds., Report No. EPR1-EL-2572. Palo Alto, CA: Electric Power
Research Institute, pp. 2-23.
Sherma, J. Manual of Analytical QuaJJtyControl for Pesticides and Related
Compounds in Human and Environmental Samples, EPA-600/2-81-Q59; NTIS No.
PB81-222721 (April 1981).
Singer E, Jarv T, Sage M. 1983. Survey of polychlorinated biphenyls in am-
bient air across the province of Ontario. Chapter 19 in Physical Behavior
of PCBs in the Great Lakes, Mackay 0, Paterson S, Eisenreich SJ, Simmons MS,
Eds. Ann Arbor, MI: Ann Arbor Science Publishers, Inc., pp 367-383.
Sonchik S, Madeleine D, Macek P, Longbottom J, 1984. Evaluation of sample
preparation techniques for the analysis of PCBs in oil. J Chromatogr Sci
22:265-271.
Spittler TM. 1983. Field measurement of PCB's in soil and sediment using a
portable gas chromatograph. Natl Conf Manage Uncontrolled Hazard Waste Sites
105-107; Chem Abst 100:220890p (1984).
Stalling DL, Huckins JN. 1973. Reverse phase thin layer chromatography of
some Aroclors, halowaxes, and pesticides. J Assoc Offic Anal Chem 56(2):
367-372.
Stahr HM. 1984. Analysis of PCBs by thin layer chromatography. J Liq Chrom
7(7):1393-1402.
Tahiliani VH. 1984. CLOR-N-QIL field test program. In: Proceedings: 1983
PCB Seminar, Addis G, Komai RY, Eds., Report No". EPRI-EL-3581. Palo Alto, CA:
Electric Power Research Institute.
Tanabe H. 1976. PCB microanalysis. In PCB Poisoning and Pollution, K. Higuchi
Ed. (Tokyo: Kodansha, Ltd; New York: Academic Press), pp, 127-145.
Tuinstra LGMTh, Driessen JJM, Keukens HJ, Van Munsteren TJ, Roos AH, Traag
WA. 1983. Quantitative determination of specified chlorobiphenyls in fish
with capillary gas chromatography and its use for monitoring and tolerance
purposes. Intern J Environ Anal Chem 14:147-157.
65
-------
Ueta T, Kamata K, Kan T, Kazama M, Totani T. 1974. Fluorescence reactions
for organic halogen compounds. II. I_n situ fluorometry of polychlorinated
biphenyls and their isomers on thin-layer chrcmatograms using diphenylamine.
Tokyo Toritsu Eisei Kenkyusho Kenkyu Nempo 25:111-118; Chem Abst 83:21991c
(1975).
United Kingdom Department of the Environment, 1979. Organochlorine Insecti-
cides and Polychlorinated Biphenyls in Waters 1978; Tentative Method. Methods
for the Examination of Waters and Associated Materials. Organochlorjne Insectic.
Polychlorinated Biphenyjs Waters 28 pp.
USEPA. 1978 (September). U.S. Environmental Protection Agency. Methods for
benzidine, chlorinated organic compounds, pentachlorophenol and pesticides in
water and wastewater. Interim Report, Environmental Monitoring and Support
Laboratory, Cincinnati, OH.
USEPA. 1979a (December 3). U.S. Environmental Protection Agency. Organo-
chlorine pesticides and RGBs—Method 608. 44 FR 69501-69509^
USEPA. 1979b (December 3). U.S. Environmental Protection Agency, Base/
neutrals, acids, and pesticides--Method 625. 44 FR 69540-69552.
USEPA. 1979c (September). U.S. Environmental Protection Agency. Analytical
protocol for screening publicly owned treatment works (POTW) sludges for
organic priority pollutants. Environmental Monitoring and Support Laboratory,
Cincinnati, OH.
USEPA. 1980. U.S. Environmental Protection Agency. Guidelines and specifi-
cations for preparing quality assurance project, plans. Office of Monitoring
Systems and Quality Assurance, QAMS-005/80.
USEPA. 1981a (February). U.S. Environmental Protection Agency. The analysis
of polychlorinated biphenyls in transformer fluid and waste oils. Office of
Research and Development, Environmental Monitoring and Support Laboratory,
Cinci nnati, OH.
USEPA. 1981b. U.S. Environmental Protection Agency. PCB disposal by thermal
destruction. Solid Waste Branch, Air and Hazardous Materials Division, Region
6, Dallas, TX, EPA-200/9-81-001; NTIS No. PB82 241 860, 606 pp.
USEPA. 1981c (March). U.S. Environmental Protection Agency. TSCA Inspection
Manual.
USEPA. 1982a (November 4). U.S. Environmental Protection Agency. Analysis
of pesticides, phthalates, and polychlorinated biphenyls in soils and bottom
sediments. HWI Sample Management Office, Alexandria, VA, unpublished method,
12 pp.
USEPA. 1982b (July). U.S. Environmental Protection Agency. Test methods
for evaluating solid waste, physical/chemical methods, SW-846, 2nd ed. Office
of Solid Waste and Emergency Response, Washington, DC.
66
-------
USEPA, 1982c (October 21). 40 CFR Part 761, Polychlorinated Biphenyls (PCBs);
Manufacturing, Processing, Distribution in Commerce, and Use Prohibitions;
Use in Closed and Controlled Waste Manufacturing Processes.
47:46980-46986.
USEPA. 1982d (November 4). Analysis of Pesticides,
chlorinated Biphenyls in Soils and Bottom Sediments.
Office, Alexandria, VA, unpublished method, 12 pp.
Phthalates, and Poly-
HWI Sample Management
USEPA. 1982e (July). Test Methods for Evaluating Solid Waste-Physical/Chem-
ical Methods, SW-846, 2nd Edition. Office of Solid Waste and Emergency
Response, Washington, DC.
USEPA. 1983. U.S. Environmental Protection Agency. Quality assurance pro-
gram plan for the Office of Toxic Substances, Office of Pesticides and Toxic
Substances, Washington, D.C.
USEPA. 1984a (October 20). Organochlorine Pesticides and PCBs--Method 608.
Fed. Reg. 49(209):89-104.
USEPA. 1984b (October 26). Base/Neutrals, Acids, and Pesticides—Method 625.
Fed. Reg. 49(209):153-174.
USEPA. 1984c (October 11). 40 CFR Part 761, Polychlorinated Biphenyls (PCBs);
Manufacture, Processing, Distribution in Commerce and Use Prohibitions; Use
in Electrical Transformers. Fed. Reg. 49:39966-39989.
USEPA. 1984d (October). Mass Spectrometric Identification and Measurement
of Polychlorinated Biphenyls as Isomer Groups. Draft Report by Physical and
Chemical Methods Branch, Office of Research and Development, Cincinnati, OH.
USEPA. 1984e (July 10). 40 CFR Part 761, Polychlorinated Biphenyls (PCBs);
Manufacturing, Processing, Distribution in Commerce and Use Prohibitions;
Response to individual and Class Petitions for Exemptions, Exclusions, and
Use Authorization, Final Rule. Fed. Reg. 49:28154-28209.
USWAG. 1984 (October 15).
posed spill cleanup policy
tection Agency.
The Utility Solid Waste Activities Group. Pro-
and supporting studies. U.S. Environmental Pro-
Watts RR (Ed.). 1980 (June). Analysis of Pesticide Residues in Human and
Environmental Samples, A Compilation of Methods Selected for Use in Pesti-
cide Monitoring Programs, U.S. Environmental Protection Agency, Research
Triangle Park/NC, EPA-600/8-80-Q38.
Webb RG, McCall AC.
gas chromatography.
1973. Quantitative PCB standards for electron capture
J Chromatogr Sci 11:366-373.
Willis OE, Addison RF. 1972. Identification and estimation of the major
components of a commercial polychlorinated biphenyl mixture, Aroclor 1221.
J Fish Res Board Can 29(5):592-595.
67
-------
Zell M, Ballschmiter K. I960. Baseline study of the global pollution. Ill,
Trace analysis of polychlorjnated biphenyls (PCB) by ECD glass capillary gas
chromatography in environmental samples of different trophic levels.
Fresenius1 Z Anal Chem 304:337-349,
Zitko V, Hutzinger 0, Safe S. 1971. Retention times and electron-capture
detector responses of some individual chlorobiphenyls. Bull Environ Corttam
Toxicol 6(2):160-163,
68
------- |