US EPA SW-846 Method 8327
Multi-Laboratory Validation Study
Quality Control Summary Report
Introduction: This document summarizes quality control (QC) results from the multi-laboratory
validation of SW-846 Method 8327 Per- and PolyfluoroalkylSubstances (PFAS) Using External
Standard Calibration and Liquid Chromatography/Tandem Mass Spectrometry (LC/MS/MS) in
four water matrices using the dilution preparation method in Appendix B of Method 8327
(future Method 3512). Precision and bias were evaluated for 24 PFAS target chemicals and 19
associated isotopically-labeled surrogates in reagent water (RW), ground water, (GW), surface
water (SW) and waste water effluent (WW), and those results are discussed in the Statistical
Analysis Report for SW-846 Method 8327 Multi-Lab Validation Study, June 2019 (hereafter
referred to as the statistics report). Only the QC results associated with study sample analyses
are discussed in this report.
Validation Study Design: The validation of SW-846 Method 8327 (hereafter referred to as the
method) was performed in a multi-step process. In Phase I, laboratories performed and
submitted initial demonstration of capability (IDOC) information which included initial
calibration, lower limit of quantitation (LLOQ), and analyst demonstration of capability using
PFAS standards provided. These results are not presented in this document. Phase II of the
study was conducted in two steps, with blind replicate spiked and unspiked samples of each
matrix, shipped to six internal laboratories in 2017, and similarly prepared blind samples
shipped to seven external laboratories in 2018. Each laboratory was tasked with 1) following
the method and study instructions for sample preparation and analysis, 2) striving to meet the
recommended acceptance criteria for sample preparation and analysis in the method (refer to
Table 7) and study instructions, and 3) returning data to the study team in a prescribed format.
Of the 13 labs that received study samples, one of the external laboratories did not provide
sample results by the submittal deadline.
Aqueous samples (5 mL) were prepared by the EPA Region 5 Laboratory in 15 mL polypropylene
containers with screw-cap lids. Five replicates of each sample matrix (RW, GW, SW and WW)
were provided to laboratories unspiked (i.e., with no target compounds added) and at spiked
concentrations of 60 and 200 ng/L (nom.) in 5 mL water, for a total of 15 replicates per matrix,
with a total of 60 sample containers shipped to each laboratory. Trip blanks (2) were also
included with shipments of study samples to participating laboratories.
Laboratories were instructed to follow the sample preparation protocols embedded in the
method (Appendix B) by dilution (1:1) with a water-miscible organic solvent (methanol)
followed by manual filtration through a particle filter (0.2 |am) and addition of acetic acid (0.1%
by volume) prior to analysis.
1
-------
Laboratories were instructed to follow the sample analysis protocols embedded in the method
and additional information provided in the study instructions. See these documents for
detailed information. In the study instruments from several manufacturers were used to ensure
that the range is applicable across a variety of platforms.
To minimize variables, laboratories were provided with PFAS target and surrogate stock
standards and supplies for the study, including glass luer-lock syringes, filters, a liquid
chromatography column, and autosampler vials.
Data Evaluation: EPA staff evaluated the data from the 12 laboratories who did submit for
compliance with the study instructions and overall usability. Data from four laboratories were
excluded for not following the specified protocols (more detail about the basis used for
exclusion of each laboratory's data is provided in Appendix E of the statistics report. Data
verification and validation were performed by contract staff for completeness, correctness,
compliance, and analytical quality against criteria provided in the method and study
instructions. Statistical analyses were performed on the data from the remaining eight
laboratories and are presented in the statistics report.
A summary of laboratory performance for instrumental analysis and sample preparation Quality
Controls (QC) is presented in the following sections by QC type. Categories of instrumental QC
included initial calibration, continuing calibration verification, and reagent blanks. Categories of
QC that addressed both sample preparation and analysis included method blanks, lower limit of
quantitation (LLOQ) verification, laboratory control samples (LCS; spiked blank), and surrogates.
Table 1 provides a summary comparison of laboratory performance by QC category, including
frequencies at which the QC acceptance criteria were met. Acceptance criteria for all QC types
were met at a frequency of >90% by all laboratories except for the coefficient of determination
for initial calibration in data from labs 4 and 16 and for LCS recovery in data from lab 5, which
are discussed in more detail in the relevant sections below.
Initial Calibration:
The method and study instructions specified analysis of 9 initial calibration standards over a
concentration range of 5-200 ng/L, representing 10 -400ng/L in the samples before dilution.
Initial calibration acceptance criteria included minimum acceptance limits for coefficient of
determination (r2); % error, i.e., recalculated concentrations <+50% for the lowest standard and
<+30% of the higher standards; and a signal-to-noise (s/n) >3 for the calibration standard the
LLOQ. (Note: no specification was provided for calculating s/n). For bias measurements, higher
importance was placed on meeting % error acceptance followed by r2 (for the internal
validation study, initial calibrations were specified to be linear only with a minimum r2 >0.98,
while for the external validation study linear or quadratic regressions were permitted, with a
minimum r2 >0.99).
2
-------
Participating labs were able to identify and calibrate most target analytes and surrogates in
standards in the concentration range of 5-200 ng/L (nom.; Table 2). The upper limit of
calibration linearity for quantitative analysis was not evaluated as part of the scope of this
validation study. Laboratory-reported LLOQs were in the 10-20 ng/L concentration range for
most target analytes, with higher ranges (up to 40-80 ng/L). % Error/r2 and target analyte
recoveries in LLOQ verification samples did not always support the LLOQs reported by some
laboratories.
Laboratories met % error criteria across 6 or more calibration standards at a higher frequency
than the minimum r2 criteria (whether assessed using a minimum r2of 0.98 or 0.99 for linear
regressions). Laboratories did not meet r2 criteria and/or did not meet % error criteria at the
lower calibration levels for certain target analytes and their associated surrogates, including the
long-chain carboxylic acids (PFTreA, PFTriA, PFDoA, PFUnDA, PFDA), the short-chain carboxylic
acids with no or low-abundance qualifier transitions (PFBA, PFPeA, PFHxA), the telomer
sulfonates (particularly 8:2 FTS and 6:2 FTS), and the sulfonamidoacetic acids (N-MeFOSAA and
N-EtFOSAA). 6:2 FTS met the initial calibration % error criteria across a minimum of six initial
calibration standards at the lowest frequency of any target analyte or surrogate (77%), likely
due to background contamination reported by multiple laboratories.
Calibration options that could have reduced calibration-related measurement bias (i.e., reduced
% error and/or increased r2) are within the scope of the method, but not all laboratories applied
these options to meet the ICAL acceptance criteria. Note: Recommendations were added to the
post validation method to calibrate target analytes with lower signal, more variable
performance, or background at higher concentrations relative to other targets, and the
chemicals above were provided as examples. Initial calibration acceptance criteria (r2>0.99 for
linear or quadratic regressions, <±50% error at LLOQ and <±30% error for higher concentration
standards) are retained in the post validation method as these are standard criteria from
Method 8000D. 6:2 FTS and associated surrogate are also listed as especially problematic.
Continuing Calibration Verification (CCV):
The method and study instructions specified analysis of CCV standards at a concentration near
the middle of the calibration range (80 ng/L, nom., with some labs using 60 or 100 ng/L). Phase
II study instructions provided to internal laboratories specified analysis of CCV standards at the
end of the analysis sequence, while instructions for external laboratories specified analysis of
CCVs at the beginning (if ICAL standards were not analyzed), after every 10 samples, and at the
end of the analysis sequence. The CCV acceptance criterion is ±30% for calculated
concentrations of target analytes and surrogates.
CCV criteria were met at high frequency across laboratories (Table 3), with all target analytes
and surrogates except 6:2 FTS and M2-6:2 FTS meeting acceptance criteria in >95% of CCVs.
Laboratories that had problems meeting CCV criteria for 6:2 FTS also had problems meeting the
initial calibration acceptance criteria due to background contamination.
3
-------
Blanks:
The method and study instructions specified the preparation of one reagent blank (RB) and the
preparation of two method blanks (MBs) with each batch of 20 or fewer samples to monitor for
contamination introduced by reagents and materials during analysis and sample preparation.
The acceptance criterion for RBs and MBs was for target analyte concentrations to be <1/2 the
LLOQ. Note: The requirement for two MBs per batch was reduced to one in the post validation
draft method. A caution was also added that more than one blank may be needed to evaluate
for commonly observed laboratory contaminants (e.g., 6:2 FTS) or very low levels (i.e., at or near
the LLOQ) are of interest.
RB and MB contamination overall was infrequent and generally limited to concentrations near
or below laboratory-reported LLOQs (Table 4). The only target analytes found in a MB or RB at
a concentration > 1/z the LLOQ at a frequency >5% were PFTreA, PFBS, and 6:2 FTS. 6:2 FTS was
reported in at least one MB at concentrations > 100 ng/L by two different laboratories, and it
was reported in an RB from a third laboratory at a concentration around 30 ng/L. The
maximum measured concentration of any target analyte other than 6:2 FTS in any blank was
<20 ng/L.
LLOQ Verification
The method and study instructions specified preparation and analysis of one or more LLOQ
verification samples with each batch of 20 or fewer samples. LLOQ verification samples were
recommended to be prepared at concentrations of 10 and/or 20 ng/L (nom.) in 5 mL water, but
some laboratories included LLOQ verification QC samples at 40 and/or 80 ng/L. The recovery
criterion for LLOQ verification samples is 50-150% of the expected (prepared) concentrations.
The frequency of target analytes meeting LLOQ verification acceptance criteria was higher at 20
ng/L than at 10 ng/L for all target analytes (Table 5). At a concentration of 20 ng/L, only a few
target analytes did not meet the LLOQ verification criteria at a frequency >90%, including
PFTriA, 8:2 FTS, 6:2 FTS, N-EtFOSAA, and N-MeFOSAA. Two laboratories did not meet LLOQ
verification acceptance criteria for 6:2 FTS in any batch. Note: Recommendations were added
to the post validation method to 1) either prepare LLOQ verification QC samples at multiple
concentrations, 2) to calibrate the instrument to below the reported LLOQ, or 3) consider the
concentration levels of interest for the project to determine if the LCS (at a higher concentration)
would meet requirements for the LLOQ verification.
LCS
The method and study instructions specified preparation and analysis of Laboratory Control
Sample/Laboratory Control Sample Duplicate (LCS/LCSD) QC samples with each batch of 20 or
fewer study samples. LCS samples were specified to be prepared at concentrations of 160 ng/L
(nominal) in 5 mL water. The acceptance criteria for LCS is 70-130% recovery and <30% for the
relative percent difference (RPD) between LCS and LCSD concentrations.
4
-------
LCS recovery and LCS/LCSD RPD (Table 6) acceptance criteria were met for all target analytes at
frequencies >90% except for PFTriA, PFBA, and 6:2 FTS. Other target analytes that had higher
frequencies of QC failures for ICAL, CCV, and LLOQ verification also met LCS recovery
acceptance criteria at lower frequency, including PFTreA, PFDoA, 8:2 FTS, N-EtFOSAA, and N-
MeFOSAA, but rates of QC failure were acceptable. Note: Statistically-derived acceptance
limits will be recommended in the post validation method for some targets, including PFTriA,
PFBA, and 6:2 FTS, as 70-130% default limits may be too narrow, provided the calculated limits
are not more restrictive than current limits.
Surrogates
The method and study instructions specified addition of 19 isotopically-labeled surrogates to
every field sample and QC sample prior to any further preparation steps at a concentration of
160 ng/L (nominal) in 5 mL water. The acceptance criterion for surrogates is 70-130% recovery.
Of the overall study sample surrogate recoveries (Table 7a), 36% had at least 1 of 19 surrogates
recovered outside of 70-130%, 13% had 2 or more of 19 out, <5% had 3 or more of 19 out, and
<2% had 4 or more of 19 out. In study samples (Table 7b), M2PFTreDA met the acceptance
criteria at the lowest frequency (around 90%), and M2-4:2FTS, d5-N-EtFOSAA, and d3-N-
MeFOSAA were the only other surrogates that met the recovery criteria at frequencies <95%.
Surrogates were recovered within the acceptance criteria at similar frequencies in laboratory-
prepared QC samples (method blanks, LLOQ verifications, and LCS), which suggested that the
sample matrix was not a strong determinant of surrogate performance in this study.
Note: Statistically-derived acceptance limits will be recommended in the post validation method
for some surrogates as 70-130% default limits may be too narrow, provided the calculated limits
are not more restrictive than current limits.
Assessment of Method Ruggedness or Robustness:
Three laboratories whose data was not included in this summary demonstrated issues related
to method robustness by not directly following study instructions for preparing standards or
samples. Two labs prepared spiking solutions in the solvent matrix used for calibration
standards and sample extracts (50:50 methanol-water with 0.1% acetic acid) instead of 95:5
acetonitrile-water and stored these solutions in glass containers, and study data submitted by
both of these labs exhibited variable performance of the longer chain acids, likely due to loss
from solution. Note: Cautions were included in the post validation draft method regarding the
minimum organic solvent content of higher concentration solutions of target compounds and/or
surrogates and avoiding storage of calibration standards and sample extracts in glass
containers to prevent loss of longer chain PFASfrom solution. Another laboratory subsampled
from aqueous sample containers prior to addition of organic solvent or surrogates, which also
resulted in apparent loss of some target analytes, particularly the longer chain carboxylic acids,
N-MeFOSAA and N-EtFOSAA. Note: More cautions were added to the post validation draft
method to help users understand the target analyte specific biases resulting from sub-sampling.
5
-------
Conclusions:
The majority of analytes met acceptance criteria for precision, bias and method DQIs. The
determinative method (LC/MS/MS) used is highly selective because multiple reaction
monitoring (MS) is used. Interferences are most likely to be seen as suppression or
enhancement to ionization rather than a false signal, and these types of matrix effects are
monitored with isotopically-labeled surrogates added to every sample.
The principal issues with this analysis are retaining the analytes in solution, background
contamination and instrument sensitivity. The 50% aqueous solvent composition of the
analytical samples and standards imposes an upper limit on the concentration of the CIO - C14
acids and FOSAAs that will be stable in solution. That limit was not determined in this study.
A general caution was also added to the method for problems observed with multiple QC
failures for 6:2 FTS in the validation study, including the enhancement of M262FTS when high
concentrations of 6:2FTS are present.
Every data set submitted by the laboratory participants was reviewed carefully and provided
valuable insight regarding both flexibilities that could be incorporated in the reference method
and additional cautions that might be needed in the method regarding potential sources of
measurement bias. The SW-846 methods team is grateful for the effort all participating
laboratories invested in this validation study.
EPA will post draft methods for public comment, as soon as possible, through the recently
approved SW-846 streamlined process.
6
-------
Table 1 Summary of study performance across QC types, by laboratory
Initial calibration (ICAL)1
Continuing
Reagent
Method
LLOQ Verification
LCS
Surrogates in
Calibration
Blank (RB)
Blank (MB)
Study Samples
Verification
(CCV)
% of target
% of target analytes
% of target
RB % of
MB % of
% of target analytes in LLOQ verification
% of target
% within
analytes and
and surrogates that
analytes and
target
target
recovered within 50-150%
analytes that
70-130%
surrogates
met % ICAL error
surrogates
analytes
analytes
met 70-130%
recovery
that met r2
criteria across at
within ±30%
cones
cones
10 ng/L
20 ng/L
40 ng/L
80 ng/L
recovery
criteria
least 6 standard
% drift
-------
Table 2. Initial calibration (ICAL) summary of performance by eight laboratories
Target or
Surrogate
Range of
LLOQs
reported by
laboratories
in ng/L1
% of initial calibrations that met % error acceptance criteria across
the stated ranges across all laboratories (n=22)2
% of ICALs that met minimum r2 and
minimum number of calibration
points, all laboratories (n=22)2
5-200
ng/L
10 - 200
ng/L
20 - 200
ng/L3
40 - 200
ng/L
% that met
acceptance criteria for
range of 6 or more
calibration standards4
Using r2>0.98 for
linear or 0.99 for
quadratic
regressions
Using r2>0.99 for
linear and
quadratic
regressions
PFTreA
10-20
68
23
5
-
95
82
73
PFTriA
10-20
64
27
9
-
100
77
73
PFDoA
10-40
68
27
5
-
100
91
82
PFUnA
10-20
82
14
5
-
100
91
86
PFDA
10-80
82
14
5
-
100
77
73
PFNA
10
95
-
5
-
100
96
77
PFOA
10-20
95
-
5
-
100
100
100
PFHpA
10-20
91
5
5
-
100
100
100
PFHxA
10-40
77
9
14
-
100
91
86
PFPeA
10-40
77
9
14
-
100
96
95
PFBA
10-40
73
18
5
-
95
100
95
PFDS
10-20
77
9
9
5
100
96
82
PFNS
10-20
86
9
5
-
100
96
86
PFOS
10-20
77
5
9
9
100
96
77
PFHpS
10
95
-
5
-
100
96
95
PFHxS
10
91
5
5
-
100
100
95
PFPeS
10-20
91
5
5
-
100
100
100
PFBS
10-20
91
5
5
-
100
100
95
PFOSA
10
95
-
5
-
100
100
100
FtS 8:2
10-20
73
14
9
5
100
96
77
FtS 6:2
10-20
73
-
5
-
77
86
77
FtS 4:2
10-40
91
-
9
-
100
100
95
NEtFOSAA
10-40
59
23
14
5
100
82
73
NMeFOSAA
10-40
55
18
18
9
100
77
64
M2PFTeDA
NA
77
5
5
-
86
82
68
MPFDoA
NA
95
-
5
-
100
96
91
M7PFUdA
NA
95
-
5
-
100
91
82
M6PFDA
NA
91
-
5
5
100
96
91
M9PFNA
NA
91
5
5
-
100
96
91
M8PFOA
NA
95
-
5
-
100
96
95
M4PFHpA
NA
95
-
5
-
100
100
91
M5PFHxA
NA
95
-
5
-
100
100
95
M5PFPeA
NA
95
-
5
-
100
100
100
MPFBA
NA
86
5
5
-
95
96
95
M8PFOS
NA
95
-
5
-
100
100
95
M3PFHxS
NA
95
-
5
-
100
100
100
M3PFBS
NA
95
-
5
-
100
100
100
M8FOSA-I
NA
82
14
5
-
100
100
86
M2-8:2FTS
NA
64
14
9
9
95
96
73
M2-6:2FTS
NA
82
9
-
-
91
96
68
M2-4:2FTS
NA
86
-
9
-
95
86
82
d5-N-EtFOSAA
NA
64
18
14
5
100
82
68
d3-N-MeFOSAA
NA
50
18
18
14
100
82
73
8
-------
^ne laboratory did not report LLOQs by target analyte; LLOQs were determined during validation based on meeting
acceptance criteria
2AII laboratories reported three initial calibrations except Laboratory 10, which only reported a single initial calibration,
using continuing calibration verification standards to demonstrate the initial calibration was valid. Note that initial
calibrations from which calibration points in the middle of the calibration range were removed to meet initial calibration
% error or r2 criteria were re-evaluated by EPA during validation, and some initial calibrations were counted as
unacceptable
3One laboratory excluded ICAL standards at 5 and 10 ng/L concentrations in one of the reported initial calibrations for all
target analytes and surrogates
4Sum of percentages in each of the calibration range columns; If an ICAL did not meet the % error criteria across 6
consecutive ICAL standards, this column is <100% (note that some values may not sum to exactly 100% due to rounding)
9
-------
Table 3. Continuing calibration verification (80 ng/L, nom.) performance summary
for eight laboratories (n=42)
target analyte
or surrogate
average %
drift1
Standard
Deviation % drift
% of CCVs that met % drift
criteria (<±30%, or 70-130%
of expected concentration)
PFTreA
-0.4
13.8
98
PFTriA
0.2
11.7
98
PFDoA
3.2
11.0
98
PFUnA
3.0
11.1
98
PFDA
0.4
11.0
98
PFNA
-1.8
9.9
100
PFOA
-0.5
7.7
100
PFHpA
-1.8
9.9
100
PFHxA
-1.5
10.6
95
PFPeA
-0.1
5.9
100
PFBA
-1.4
10.3
100
PFDS
-1.1
7.2
100
PFNS
0.3
6.3
100
PFOS
-0.1
7.3
100
PFHpS
1.2
5.4
100
PFHxS
1.2
6.8
100
PFPeS
0.4
6.4
100
PFBS
-0.2
12.6
98
PFOSA
0.3
4.3
100
FtS 8:2
1.5
9.6
100
FtS 6:2
27.5
121
86
FtS 4:2
0.1
8.2
100
NEtFOSAA
1.4
6.0
100
NMeFOSAA
-1.3
7.0
100
M2PFTeDA
-2.3
11.0
100
-------
target analyte
or surrogate
average %
drift1
Standard
Deviation % drift
% of CCVs that met % drift
criteria (<±30%, or 70-130%
of expected concentration)
MPFDoA
-1.4
9.7
100
M7PFUdA
1.8
10.6
100
M6PFDA
-0.2
7.9
100
M9PFNA
0.4
8.5
100
M8PFOA
-3.0
9.2
100
M4PFHpA
-0.3
10.0
100
M5PFHxA
-2.1
10.9
98
M5PFPeA
0.3
5.0
100
MPFBA
-0.9
8.9
100
M8PFOS
1.9
6.3
100
M3PFHxS
0.8
6.3
100
M3PFBS
-1.5
10.2
98
M8FOSA-I
1.7
4.7
100
M2-8:2FTS
1.2
8.7
100
M2-6:2FTS
8.0
26.2
90
M2-4:2FTS
1.6
9.5
98
d5-N-EtFOSAA
1.7
9.2
100
d3-N-
MeFOSAA
-1.3
9.9
98
-------
Table 4. Method blank and reagent blank performance summary for eight laboratories
Target
Analyte
Method blank (n=49)1
Reagent blank (n=54)
Maximum
reported
concentration
(ng/L)
% with measured
concentration <50%
of laboratory-
reported LLOQ(ng/L)
Maximum
reported
concentration
(ng/L)
% with measured
concentration <50%
of laboratory-
reported LLOQ(ng/L)
PFTreA
14.3
90
11.9
93
PFTriA
11.1
96
8.2
96
PFDoA
10.4
98
5.0
98
PFUnA
8.0
98
5.1
98
PFDA
9.0
98
9.8
100
PFNA
00
00
98
2.4
100
PFOA
6.8
98
6.3
98
PFHpA
3.4
100
1.7
100
PFHxA
9.7
98
0.6
100
PFPeA
9.2
100
4.6
100
PFBA
8.0
100
7.5
96
PFDS
4.3
100
4.6
100
PFNS
2.7
100
2.3
100
PFOS
9.9
98
1.8
100
PFHpS
3.8
100
3.0
100
PFHxS
7.3
98
12.3
98
PFPeS
1.8
100
2.2
100
PFBS
19.0
92
17.0
96
PFOSA
3.1
100
3.2
100
FtS 8:2
2.0
100
1.8
100
FtS 6:2
116.2
86
29.6
89
FtS 4:2
2.2
100
1.5
100
NEtFOSAA
8.0
98
5.0
98
NMeFOSAA
9.4
98
3.6
100
1 Laboratory 4 submitted data for spiked blanks instead of method blanks in batch 2; this data was excluded from the method blanks statistical
summary
12
-------
Table 5. LLOQ verification performance summary for eight laboratories
Target
Analyte
10 ng/L (nom.) in 5 mL water
(n=21 across 6 labs)
20 ng/L (nom.) in 5 mL water
(n=18 across 6 labs)
40 ng/L (nom.) in 5 mL water
(n=5 across 2 labs)
Average
% recovery
Standard
deviation
of %
recovery
% that met
50-150%
recovery
Average
% recovery
Standard
deviation
of %
recovery
% that met
50-150%
recovery
Average
% recovery
Standard
deviation
of %
recovery
% that met
50-150%
recovery
PFTreA
112
25.1
81
110
30.0
94
112
18.2
100
PFTriA
126
53.0
71
118
32.4
83
128
36.6
80
PFDoA
103
29.9
81
109
20.5
94
94.6
11.0
100
PFUnA
108
27.4
81
99.9
16.0
100
88.7
10.1
100
PFDA
96.8
26.3
86
103
19.5
94
100
17.6
100
PFNA
100
26.1
91
99.7
14.1
100
108
9.3
100
PFOA
101
26.8
86
99.5
16.3
100
98.1
6.6
100
PFHpA
93.2
15.7
100
99.5
13.8
100
97.9
6.4
100
PFHxA
99.1
41.6
86
94.9
17.6
100
98.1
19.5
100
PFPeA
103
36.2
86
99.1
13.2
100
102
6.9
100
PFBA
89.1
27.4
86
95.2
20.6
94
94.3
7.9
100
PFDS
105
24.6
81
100
24.2
100
105
24.2
100
PFNS
102
28.5
95
106
18.0
100
112
15.3
100
PFOS
112
23.0
86
106
16.8
100
114
8.2
100
PFHpS
89.1
41.8
91
105
14.5
100
100
9.8
100
PFHxS
99.0
17.7
100
99.3
12.7
100
103
11.4
100
PFPeS
95.8
12.4
100
99.7
12.3
100
100
8.0
100
PFBS
93.1
17.2
95
91.6
12.5
100
108
24.7
100
PFOSA
101
15.0
100
99.7
8.5
100
111
14.7
100
FtS 8:2
112
36.5
67
129
57.9
72
120
32.5
100
FtS 6:2
1470
5540
57
125
152
50
85.6
21.0
100
FtS 4:2
102
16.4
91
96.2
14.3
100
101
14.2
100
NEtFOSAA
122
33.6
71
111
18.9
78
106
25.3
80
NMeFOSAA
109
52.8
71
104
34.1
83
100
25.9
100
13
-------
Table 6. LCS performance for eight laboratories (160 ng/L nom. expected concentration in 5 mL
water)
Target
Analyte
LCS % Recovery, All Labs (n=48)
Relative % Difference (RPD) between
concentration in LCS and LCSD, All labs (n=24)
Average %
recovery
Standard
deviation of
% recovery
% that met 70-
130% Recovery
Average
RPD
Standard
Deviation RPD
% that met RPD
(<±30%)
PFTreA
103
18.9
90
7.1
4.9
100
PFTriA
107
22.7
83
8.9
8.0
96
PFDoA
104
16.7
92
10.3
9.7
96
PFUnA
101
12.1
100
9.0
8.7
100
PFDA
102
11.5
98
8.9
7.4
96
PFNA
103
12.3
96
6.9
7.2
100
PFOA
101
12.1
98
6.8
6.6
96
PFHpA
96.4
8.7
100
5.4
5.3
100
PFHxA
95.8
10.5
100
8.1
7.0
100
PFPeA
94.1
10.1
100
4.1
3.3
100
PFBA
91.5
15.1
88
4.4
4.4
100
PFDS
100
10.2
100
5.5
4.8
100
PFNS
105
12.6
100
6.9
6.3
100
PFOS
99.9
8.9
100
5.1
4.9
100
PFHpS
101
9.1
100
5.2
5.2
100
PFHxS
97.9
8.1
100
4.5
4.7
100
PFPeS
98.0
7.2
100
5.5
5.2
100
PFBS
93.2
9.6
100
3.3
5.5
100
PFOSA
98.7
8.2
100
3.6
2.7
100
FtS 8:2
104
15.0
94
8.2
6.9
100
FtS 6:2
91.1
33.0
65
10.2
8.4
100
FtS 4:2
98.0
12.0
96
00
00
8.9
96
NEtFOSAA
102
15.6
94
9.0
8.6
100
NMeFOSAA
102
15.2
94
9.2
7.7
96
14
-------
Table 7a. Overall surrogate performance in study samples, by laboratory
Laboratory #
2 4s 5 6s 10" 11 12 16
All
# of study samples with reported
results
60
60
60
60
59
60
60
60
479
# of samples with one or more
surrogates outside 70-130% recovery
34
25
42
9
14
14
5
29
172
# of samples with two or more
surrogates outside 70-130% recovery
12
6
18
6
3
3
0
12
60
# of samples with three or more
surrogates outside 70-130% recovery
4
1
5
5
0
1
0
6
22
# of samples with four or more
surrogates outside 70-130% recovery
1
1
1
1
0
0
0
3
7
# of surrogates reported across all
samples
1140
1140
1140
1140
1121
1140
1140
1140
9101
# of surrogates that met 70-130%
recovery across all samples
1089
1060
1074
1104
1104
1122
1135
1090
8778
% of surrogates that met 70-130%
recovery across all samples
95.5
93
94.2
96.8
98.5
98.4
99.6
95.6
96.4
'Reported surrogate recoveries were near 200% in a study sample, suggesting it was double-spiked with
surrogates. Surrogate recoveries from these samples are included in the summary table above.
"Results for one study sample were rejected due to lack of identified surrogates or target analytes;
Surrogate data from this sample was excluded from these summary statistics.
15
-------
Table 7b. Comparison of surrogate recovery in study samples and clean matrix QC
samples (Method blank, LLOQ verification, LCS) across all laboratories and matrices
Surrogate
Surrogates in Study Samples, All Labs (n=477)1
Surrogates in Laboratory QC Samples, All Labs
(n=143)
Average %
Recovery
Standard
Deviation of
% recovery
% that met 70-
130% recovery
limits
Average %
Recovery
Standard
Deviation of
% recovery
% that met 70-
130% recovery
limits
M2PFTeDA
96.8
18.8
90
101
19.5
90
MPFDoA
101
14.7
96
101
13.3
97
M7PFUdA
103
11.6
99
101
10.3
99
M6PFDA
104
12.1
98
102
10.5
99
M9PFNA
102
11.6
99
99.7
9.3
100
M8PFOA
101
9.5
100
99.7
9.1
100
M4PFHpA
98.9
10.9
98
97.1
9.9
99
M5PFHxA
97.4
11.8
98
95.3
10.6
98
M5PFPeA
98.7
7.5
100
95.7
8.0
100
MPFBA
95.6
10.9
98
93.7
11.5
94
M8PFOS
104
11.2
99
102
9.3
100
M3PFHxS
102
8.0
100
99.1
7.5
100
M3PFBS
96.9
12.0
98
94.5
9.1
98
M8FOSAI
101
8.9
99
99.8
00
00
99
M282FTS
106
13.9
96
101
14.2
97
M262FTS
100
15.4
98
100
17.3
95
M242FTS
97.8
19.4
92
97.4
18.4
92
d5NEtFOSAA
104
16.0
91
106
14.1
95
d3NMeFOSAA
102
16.1
92
103
14.2
95
Surrogate data for three samples were removed prior to calculating these statistics (one sample had no
qualitatively identifiable surrogates or target analytes, and the other two samples had surrogate recoveries around
200% and appear to have been double spiked).
16
-------
Appendix A. Target analyte abbreviations used for the validation study
Analyte
CAS RN
Abbreviation
PFAS sulfonic acids
Perfluoro-l-butanesulfonic acid 375-73-5
Perfluoro-l-pentanesulfonic acid 2706-91-4
Perfluoro-l-hexanesulfonic acid 355-46-4
Perfluoro-l-heptanesulfonic acid 375-92-8
Perfluoro-l-octanesulfonic acid 1763-23-1
Perfluoro-l-nonanesulfonic acid 68259-12-1
Perfluoro-l-decanesulfonic acid 335-77-3
1H, 1H, 2H, 2H-perfluorohexane sulfonic acid 757124-72-4
1H, 1H, 2H, 2H-perfluorooctane sulfonic acid 27619-97-2
1H, 1H, 2H, 2H-perfluorodecane sulfonic acid 39108-34-4
PFAS carboxylic acids
Perfluorobutanoic acid 375-22-4
Perfluoropentanoic acid 2706-90-3
Perfluorohexanoic acid 307-24-4
Perfluoroheptanoic acid 375-85-9
Perfluorooctanoic acid 335-67-1
Perfluorononanoic acid 375-95-1
Perfluorodecanoic acid 335-76-2
Perfluoroundecanoic acid 2058-94-8
Perfluorododecanoic acid 307-55-1
Perfluorotridecanoic acid 72629-94-8
Perfluorotetradecanoic acid 376-06-7
PFAS sulfonamides and sulfonamidoacetic acids
N-ethylperfluoro-l-octanesulfonamidoacetic acid 2991-50-6
N-methylperfluoro-l-octanesulfonamidoacetic acid 2355-31-9
Perfluoro-l-octanesulfonamide (FOSA) 754-91-6
PFBS
PFPeS
PFHxS
PFHpS
PFOS
PFNS
PFDS
4:2 FTS
6:2 FTS
8:2 FTS
PFBA
PFPeA
PFHxA
PFHpA
PFOA
PFNA
PFDA
PFUdA (PFUnA)*
PFDoA
PFTrDA (PFTriA)*
PFTeDA (PFTreA)*
N-EtFOSAA
N-MeFOSAA
PFOSA
*Two abbreviations were used during the study, both are given here
17
------- |