EPA Region 10 SOP For the Validation of
of Method 1668 Toxic, Dioxin-like,
PCB Data

EPA Region 10
Environmental Services Division
12 00 Sixth Avenue
Seattle, WA 98101

Revision 1.0
12/8/95

APPROVAL:

Quality Assurance Manager:

Barry Towns	Date:

Region 10 Chemists:

Peggy Knight	Date:

Robert Melton

Date:


-------
TABLE OF CONTENTS

I.0	HOLDING TIME AND PRESERVATION OF SAMPLES 		1

1.1	Objective	1

1.2	Criteria	1

1.3	Action	2

2.0 GC/MS PERFORMANCE CHECK 		2

2.1	Objective	2

2.2	Criteria	3

2.3	Action	3

3.0 INITIAL CALIBRATION 		3

3.1	Objective	3

3.2	Criteria	3

3.3	Action	4

4.0 CALIBRATION VERIFICATION MEASUREMENTS 		4

4.1	Objective	4

4.2	Criteria	4

4.3	Action	4

5.0 SYSTEM PERFORMANCE 		5

5.1	Objective	5

5.2	Criteria	5

5.3	Action	5

6.0 METHOD BLANKS 		6

6.1	Objective	6

6.2	Criteria	6

6.3	Action	6

7.0 RECOVERY OF SPIKED C-13 LABELED PCB CONGENERS 		6

7.1	Objective	6

7.2	Criteria	6

7.3	Action	7

8.0 RECOVERY OF C-13 LABELED INTERNAL STANDARDS 		7

8.1	Objective	7

8.2	Criteria	7

8.3	Action	7

9.0 PROJECT AND REGIONAL QUALITY ASSURANCE SAMPLES 		8

9.1	Objective	8

9.2	Criteria	8

9.3	Action	8

10.0 COMPOUND IDENTIFICATION 		8

10.1	Objective	8

10.2	Criteria	8

10.3	Action	9

II.0	LABORATORY CONTACTS 		9


-------
Revision: 1.0
Date: 12/8/95
Page: 3 of 12

12.0 OVERALL ASSESSMENT OF THE QUALITY OF THE DATA	10

12.1	Objective	10

12.2	Criteria	10

12.3	Action	10

DATA QUALIFIER DEFINITIONS 		11


-------
Revision: 1.0
Date: 12/8/95
Page: 1 of 12

EPA Region 10 SOP For the Validation of
of Method 1668 Toxic, Dioxin-like,

PCB Data

The Quality Assurance Unit of EPA Region 10 has developed the
following guidelines which should be used to access the quality
of toxic, dioxin-like PCB data from samples originating from
Region 10 sampling sites. This SOP is based upon the data
validation principles specified in National Functional Guidelines
For Organic Data Review. December, 1990, and the quality control
(QC) requirements of EPA Method 1668, Draft Revision, 10/4/95.
The validator of toxic, dioxin-like PCB data should obtain a copy
of the site-specific Quality Assurance Project Plan (QAPP) and
use the Data Quality Objectives and QA requirements of the QAPP
to assess the data. This SOP requires that the following
criteria be evaluated when determining the quality of toxic,
dioxin-like PCB data:

1.0	HOLDING TIME AND PRESERVATION OF SAMPLES

1.1	Objective. To determine the validity of the measurement
results based upon EPA requirements for preservation and holding
time of the samples from day of collection to day of extraction.
EPA also has holding time requirements for extracts which is the
time from extraction of the samples to injection of the sample
extracts.

1.2	Criteria. Holding time and preservation requirements for
the measurement of PCBs as Aroclors in water samples under the
CWA (40CFR Part 136), SDWA, and RCRA have been promulgated and
codified under 40CFR. These regulations require that water
samples be preserved by cooling to 4°C using a holding time of 7
days from day of collection to day of extraction of the sample.
In addition, the maximum holding time of extracts is 40 days from
day of extraction to day of injection of the extract.

The holding time and preservation requirements of toxic, dioxin-
like PCB isomers in non-water matrixes have not been promulgated
by EPA. Therefore, the data validator should use the holding
time specified in the EPA approved site-specific Quality
Assurance Project Plan (QAPP).

Method 1668, Draft Revision, 10/4/95 recommends different
preservation and holding times for PCB congeners. Consult
Section 8.0 of Method 1668 for preservation and holding time
recommendations.


-------
Revision: 1.0
Date: 12/8/95
Page: 2 of 12

Section 8.2 states that aqueous samples
should be tested for chlorine residual. If
chlorine is present, 80mg of sodium
thiosulfate should be added per liter of
water. Adjust pH to 2-3 with sulfuric acid.

Store samples in dark at 0 to 4°C. Method
1668 recommends a holding time of less that
one year.

Section 8.3 states that solid, semi-solid,
oily, and mixed phase samples should be
stored in wide mouth bottle at <4°C. Section
8.3 states that solid, semi-solid, oily, and
mixed phase samples should be stored in the
laboratory at < -10°C. Method 1668
recommends a holding time of less that one
year.

Section 8.4 states that fish and tissue
samples should be wrapped in aluminum foil,
cooled to <4°C, and shipped to lab. Upon
receipt at the lab, tissue samples should be
stored in the dark at < -10°C. Method 1668
has recommended a holding time of one year
for tissue samples which are frozen at < -
10°C. Once frozen tissue samples are thawed,
tissue samples must be extracted within 24
hours.

Extracts should be analyzed within 40 days of extraction.

1.3 Action. If 40CFR Part 136 and the QAPP for the samples do
not specify a holding time, then the holding time which is
recommended by applicable EPA method -- Method 16 6 8 should be
used. Whenever samples or extracts are analyzed after holding
time expiration date, the results should be considered to be
minimum concentrations and must be qualified with a "J3".

Samples which are not preserved correctly should be qualified
with a "J" flag.

2.0	GC/MS PERFORMANCE CHECK

2.1	Objective. Gas chromatograph/mass spectrometer (GC/MS)
instrument performance checks stated in Method 1668 Section 10.0
are performed to ensure mass resolution, identification, and
calibration. Conformance is determined using standard materials,
therefore, these criteria should be met in all circumstances.


-------
Revision: 1.0
Date: 12/8/95
Page: 3 of 12

2.2	Criteria. For the PFK molecular leak, the resolution must
be greater than or equal to 10,000. The deviation between the
exact mass and the theoretical mass for each of the three to five
ions monitored must be less than 5 ppm. If the mass spectrometer
is adjusted the resolution must be tested again and the
resolution documented.

The mass spectrometer shall be operated in a mass-drift
correction mode using PFK to provide lock-masses. Each lock-mass
shall be monitored and shall meet the QC requirements of Section
7.1 of Method 1668.

Ion abundance ratios. All labeled and unlabeled PCB congeners in
the CS1 standard shall be within the QC limits described in
Section 10.2 and in Table 9 for their respective ion abundance
ratios.

The HRGC/HRMS must meet the minimum levels in 1668 Table 2. All
labeled and unlabeled analytes in the CS1 calibration standard
must have signal to noise ratios greater than or equal to 10.0.
(see Method 1668/Section 10.2)

The absolute retention time of PCB 169 shall exceed 20.0 minutes
on the SPM-Octyl column, and the retention time of PCB 157 shall
exceed 25.0 minutes on the DB-1 column, (see Method 1668/Section
10.2.4)

The compound pairs in the window defining mixtures shall be
determined, (see Method 1668/Section 10.3)

The isomer specificity requirements stated in Method 1668 Section
10.4 shall be met.

2.3	Action. Failure to meet either the resolution or the
retention window criteria invalidates all calibration or sample
data collected during the 12 hour time window.

3.0	INITIAL CALIBRATION

3.1	Objective. Compliance requirements for satisfactory
instrument calibration are established to ensure that the
instrument is capable of producing acceptable qualitative and
quantitative data for PCBs. Initial calibration demonstrates
that the instrument is capable of producing a linear calibration
curve.

3.2	Criteria. There shall be an initial calibration curve
consisting of five points for each analyte. The initial
calibration curve shall be determined less than 30 days from the


-------
Revision: 1.0
Date: 12/8/95
Page: 4 of 12

time the first samples of a Sample Delivery Group (SDG) are
measured by the lab. The lab shall use the same calibration
standards with the same lot number, for all internal standards,
and labeled standards used in measuring the initial calibration
curve, verification standards, field samples, and method blanks
on both the primary GC column and on the secondary confirmation
GC column. If an analyte is calculated by the isotope dilution
method, an averaged response factor may be used if the RSD is
less than 20% For analytes calculated by the internal standard
method, an averaged response factor may be used if the RSD is
less than 35%. Otherwise, for either calculation method, the
complete curve must be used (see Method 1668/Sections 10.5 and
10.6) .

3.3 Action. If the Initial Calibration Curve is older that 30
days, or if internal standards or labeled standards used in
measuring of the initial calibration curve, verification
standards, field samples, and method blanks on both the primary
GC column and on the secondary confirmation GC column or not from
the same lot number, then all measurement data should be
qualified with a "J" qualifier.

If the RSD exceeds 20% for those analytes analyzed by isotope
dilution or 35% for those analytes analyzed by the internal
standard method, qualify positive results with "J", and non-
detected analytes using professional judgement. At the
reviewer's discretion, a more in-depth review may be conducted to
minimize data qualification by examining the entire curve and the
quantitation method used.

4.0	CALIBRATION VERIFICATION MEASUREMENTS

4.1	Objective. Compliance requirements for satisfactory
instrument calibration are established to ensure that the
instrument remains capable of producing acceptable qualitative
and quantitative data each day that samples are measured.

4.2	Criteria. Native and labeled PCB congeners in the
calibration verification standard (CS3) and in the Ongoing
Precision and Recovery Standard (OPR) shall meet the acceptance
criteria which are specified in Method 1668, Section 15.0.

4.3	Action. The reviewer should use professional judgement to
determine if it is necessary to qualify the data. The following
are guidelines:

If the %D for an analyte is outside the acceptance window,
qualify positive results "J" and non-detected "UJ" for that


-------
Revision: 1.0
Date: 12/8/95
Page: 5 of 12

analyte. If the ion abundance criteria are not met results
qualify all results for that analyte "R".

5.0	SYSTEM PERFORMANCE

5.1	Objective. The performance of the method by the Laboratory
is examined by determination of the Laboratory's ability to
perform the method (Initial Precision and Recovery (IPR) study)
and to demonstrate the Laboratory's continuing ability to perform
the analysis. See Section 9.0 of Method 1668, Draft Revision,
10/4/95 for initial and ongoing QA and QC requirements.

As part of measuring system performance, Method 1668 require that
samples and standards be measured within require QC limits. QC
criteria such as required relative retention times of labeled and
native isomers, theoretical ion abundance ratios, recovery limits
for OPR and VER standards, and recovery limits for spiked labeled
target compounds must be met in order to demonstrate that the
measurement system is within the specified control limits of
Method 1668. In addition, all samples will be spiked with the
labeled compound spiking solution described in Section 7.10.3.

5.2	Criteria. Initial precision and accuracy (IPR). All
cleanup steps used in processing samples shall be included in the
IPR study. All analytes shall be within the IPR limits in Table

6 of Method 1668 (use Table 6a if only PCBs 77, 126, and 169 are
determined). There will be one PAR sample for each sample set
analyzed. The recovery of labeled spiked isomers in samples
shall be within the QC limits specified in Table 7 (use Table 7a
if only PCBs 77, 126, and 169 are measured).

QC limits such as required relative retention times of labeled
and native isomers, theoretical ion abundance ratios, recovery
limits for OPR and VER standards, and recovery limits for spiked
labeled target compounds must be within control limits of Method
1668 .

5.3	Action. Results for analytes which do not meet either IPR
or PAR requirements should be qualified with either "J" or "UJ".
If an analyte is not recovered for an PAR sample, results must be
qualified with an "R" for that analyte. Failure to meet QC
limits of the method may result in measurement values which are
qualified with a "J" or "UJ". In specific cases where major QC
limits are exceeded, the data validator may determine that the
measurement system is out of control, which would require that
all measurement results for a sample be qualified with a "J",
"UJ", or "R" flag.


-------
Revision: 1.0
Date: 12/8/95
Page: 6 of 12

6.0	METHOD BLANKS

6.1	Objective. To determine the existence and magnitude of
contamination of samples resulting from laboratory activities.
The criteria for evaluation of blanks will apply to any blank
associated with the samples, including any method blanks,
instrument blanks, field equipment blanks, transfer blanks, trip
blanks, or solvent blanks.

6.2	Criteria.

1.	The criteria for the frequency of extraction and
analysis of method blanks as stated in Section 9.5 of Method
1668 shall be followed and demonstrated in the documented
data. The maximum amount of toxic, dioxin-like PCB isomer
contamination in method blanks is stated in Table 2 of
Method 1668.

2.	A method blank must be measured on each GC/MS system
which is used to measure a group of samples. This
requirement includes measuring method blanks for PCBs 156
and 157 on the secondary GC confirmation column (DB-1) if
PCBs 156 and 157 are detected on the primary GC column, SPB-
Octyl (see GC confirmation requirements in Method 1668,
Section 16.5).

6.3	Action. If the maximum contamination requirements of
specific PCB congeners stated in Table 2 of Method 1668 are not
met, then all isomers in all samples associated with a method
blanks shall be qualified with a "Jl" flag. If the frequency of
measuring method blanks is not met by the laboratory in the data
submitted, then the results of all samples which do not meet the
frequency of extraction and measurement of method blanks shall be
qualified with a "R" flag. Any measurement of PCB congeners in a
sample that is also measured in any associated blank, is
qualified with a "U" flag if the sample concentration is less
than 5 times the blank concentration.

7.0	RECOVERY OF SPIKED C-13 LABELED PCB CONGENERS

7.1	Objective. Labeled PCB congeners are added to each sample
and method blank prior to extraction. The role of these C-13
labeled spiked compounds is to be an internal standard for the
quantitation of native toxic, dioxin-like PCB isomers, and to
serve as surrogates for the assessment of method performance in
the sample matrix.

7.2	Criteria. The recovery of each C-13 labeled toxic, dioxin-
like PCB isomer using Method 1668 must be within recovery limits


-------
Revision: 1.0
Date: 12/8/95
Page: 7 of 12

specified in Table 7 (see Table 7a if PCBs 77, 126, and 169,
only, are measured).

7.3 Action. If any of the labeled percent recoveries are
outside the guideline windows for individual analytes listed in
Table 7 (see Table 7a if PCBs 77, 126, and 169, only, are
measured), the individual isomer for that sample will be
qualified with a "J" flag. For non-detected toxic, dioxin-like
PCB compounds whose percent recoveries are outside the guideline
windows for individual analytes, these will be qualified with a
"UJ" flag.

8.0	RECOVERY OF C-13 LABELED INTERNAL STANDARDS

8.1	Objective. The purpose of adding four labeled internal
standards (see Method 1668, Section 7.12) prior to injecting
sample extracts and standards into the GC/MS is to determine the
recovery efficiency of the extraction and cleanup procedures, to
determine if the GC/MS sensitivity and response are stable during
every analytical run, and to determine if the same amount of
extract was injected into the GC/MS.

8.2	Criteria. The sum of the area counts of two masses for each
of the two cleanup standards for samples, blanks, and standards
must not vary by more than a factor of four (-25% to +200%) from
the sum of the associated average areas from the five initial
calibration standards.

8.3	Action. The reviewer should use professional judgement to
determine if it is necessary to qualify the data. The following
are guidelines:

1.	If the sum of the two quantitation area counts of each
internal standard in samples or blanks are outside a -25% to
+200% window which is determined by averaging the sum of the
area counts present in the five initial calibration
standards, then positive measurement results for native
compounds should be qualified with a "J".

2.	If the sum of the two quantitation area counts is
greater that 200%, then non-detected compounds should not be
qualified.

3.	If the sum of the two area counts is less than 25%, then
non-detect compounds should be qualified with a "UJ".

4.	If the sum of the area counts is less than 10%, then
non-detect target compounds should be qualified with a "R".


-------
Revision: 1.0
Date: 12/8/95
Page: 8 of 12

9.0	PROJECT AND REGIONAL QUALITY ASSURANCE SAMPLES

9.1	Objective: The data validator should consider the data of
samples which are identified as field duplicates, transfer
blanks, trip blanks, blind spikes, blind blanks, and performance
evaluation (PE) samples.

9.2	Criteria. If QA samples are included among the field
samples for measurement by the laboratory, then the data
validator should refer to the applicable QAPP for any QA
requirements regarding QA samples. Results from the measurement
of project and regional QA samples will reflect upon the ability
of the laboratory to report analytical results of known and
documented quality which meet the PARCC requirements of the QAPP.

9.3	Action. The data validator should recommend action in
accordance with Regional specifications, QAPP specifications, or
criteria for acceptable PE sample results. Poor performance by
the laboratory on blind PE samples may indicate that the
laboratory analytical system is out of control, or that the
amount of toxic, dioxin-like PCB isomers reported by the
laboratory is an estimated quantity. The data validator should
use her/his professional judgement to assess if "J" or "R"
qualifiers should be placed upon the data due to the measurement
of QA or PE samples.

10.0	COMPOUND IDENTIFICATION

10.1	Objective. The qualitative criteria for target compound
identification have been established by EPA Method 1668 to
minimize the number of erroneous identifications. An erroneous
identification can be either a false-positive (reporting a target
compound when it is not present in the sample), or false-negative
(not reporting a compound that is present in the sample). The
addition of single or double blind PE samples among field samples
provides ancillary data to support the laboratory's ability to
meet QAPP objectives.

10.2	Criteria. EPA Method 1668 specifies certain requirements
and guidelines for the positive identification of certain toxic,
dioxin-like PCB isomers such as PCBs 156 and 157 (see Section
16.5). The most frequently encountered interfering compounds to
the measurement of toxic, dioxin-like PCB isomers are chlorinated
substances such as other PCB congeners, Polychlorinated dioxins
and furans (PCDDs/PCDFs), methoxy biphenyls, hydroxydiphenyl
ethers, benzylphenyl ethers, polynuclear aromatics, and
pesticides that may be found at concentrations several orders of
magnitude higher than the analytes of interest. Method 1668
requires that if certain PCB congeners such as PCB 156 and 157


-------
Revision: 1.0
Date: 12/8/95
Page: 9 of 12

are measured on the primary GC column, SPB-Octyl, that PCBs 156
and 157 must be confirmed using second dissimilar GC column (DB-
1) before specific identifications can be made.

In this part of the SOP for the validation of toxic, dioxin-like
PCB data, the qualitative identification criteria specified in
Method 1668, Section 16.0 must be met for a GC peak to be
identified as a PCB congener:

1.	The signals for the two exact m/z's listed in Table 8
must be present, and must maximize within plus or minus 2
seconds of one another (see 1668/Section 16.1).

2.	The signal-to-noise ratio (S/N) of each of the two exact
m/z's must be greater than or equal to 2.5 for a sample
extract, and greater than or equal to 10 for a calibration
standard (see 1668/Section 16.2).

3.	The ratio of the integrated ion currents (EICPs) of both
the exact m/z's monitored must be within the limits which
are listed in Table 9 of the method (see 1668/Section 16.3).

4.	The relative retention time (RRT) of the peaks
representing a unlabeled PCB congener must be within 5% of
the limits listed in Table 2 (see Method 1668, Section
16.4) .

5.	The measurement of PCBs 156 and 157 on the primary SPD-
Octyl GC column must be confirmed by analysis on a
confirmatory column such as DB-1. All QC requirements of
the method must be met on both the primary and secondary GC
columns (see 1668/Section 16.5).

10.3 Action. The validator of the data must use his/her
professional judgement in evaluating the data using the above
identification criteria. Generally, if all of the above criteria
for the identification of toxic, dioxin-like PCB isomers are not
met, then each reported positive measurement of a PCB congener
should be considered a non-detect, and therefore flagged with a
"R" flag. The "R" flag in this case is based upon the fact that
the presence of the isomer in the sample can not be corroborated
by the laboratory data.

11.0 LABORATORY CONTACTS

Provide and attached to the validation memo a copy of all
telephone logs and correspondence with the laboratory concerning
the quality of data submitted by the laboratory.


-------
Revision: 1.0
Date: 12/8/95
Page: 10 of 12

12.0	OVERALL ASSESSMENT OF THE QUALITY OF THE DATA

12.1	Objective. The overall assessment of a data package is a
brief narrative in which the data reviewer expresses concerns and
comments of the quality of the data. The overall assessment of
the data should be made after the data validator considers the
DQOs and other QA requirements of the site-specific QAPP. It
should be noted that the data reviewer does not determine or
report the useability of the data. This determination is made by
the Site Manager and by the other users of the data.

12.2	Criteria. The criteria for overall assessment is the QA
and DQO criteria of the QAPP and the criteria listed above in
this data validation SOP.

12.3	Action. Use professional judgement to determine if there
is a need to further qualify the data. Write a brief narrative
to give the user an indication of any analytical limitations of
the data. Note if there are any inconsistencies observed between
the raw data and the laboratory reported sample results.


-------
Revision: 1.0
Date: 12/8/95
Page: 11 of 12

DATA QUALIFIER DEFINITIONS

U - The analyte was analyzed for, but was not detected above
the sample quantitation limit. The associated numerical value
indicates the approximate concentration necessary to detect the
analyte in this sample.

If a decision requires quantitation of the analyte below the
associated numerical level, reanalysis or alternative
analytical methods should be considered.

J - The analyte was analyzed for and was positively identified,
but the associated numerical value may not be consistent with the
amount actually present in the environmental sample.

A subscript may be appended to the "J" that indicates which
of the following quality control criteria were not met:

J1 Blank Contamination: indicates possible high bias
and/or false positives.

J2 Calibration range exceeded: indicates possible low
bias.

J3 Holding times not met: indicates low bias for most
analytes.

J4 Other QC parameter outside control limits: bias
not readily determined.

J5 Other QC parameter outside control limits. The
reported results appear to be biased high. The actual
value of target compound in the sample may be lower
than the value reported by the laboratory.

J6 Other QC parameter outside control limits. The
reported results appear to be biased low. The actual
value of target compound in the sample may be higher
than the value reported by the laboratory.

R - The sample results are rejected due to serious deficiencies
in the ability to analyze the sample and meet critical quality
control criteria. The presence or absence of the analyte cannot
be verified.

Resampling and reanalysis are necessary to confirm or
deny the presence of the analyte.


-------
Revision: 1.0
Date: 12/8/95
Page: 12 of 12

UJ - The analyte was analyzed for and was not detected above the
reported quantitation limit. However, the reported quantitation
limit is approximate and may or may not represent the actual
limit of quantitation necessary to accurately and precisely
measure the analyte in this sample.

If a decision requires quantitation of the analyte
close to the associated numerical level, reanalysis or
alternative analytical methods should be considered.

12


-------