proficiency study aqa 15-03 pfos/pfoa in soil … · · 2015-07-17proficiency study aqa 15-03...
TRANSCRIPT
June 2015
PROFICIENCY STUDY AQA 15-03 PFOS/PFOA IN SOIL AND WATER
i
ACKNOWLEDGMENTS
This study was conducted by the National Measurement Institute (NMI). Support funding was provided by the Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC CARE) and the Australian Government Department of Industry and Science. I would like to thank the management and staff of the participating laboratories for supporting the study. It is only through widespread participation that we can provide an effective service to laboratories.
The assistance of the following NMI staff members in the planning, conduct and reporting of the study is acknowledged.
Raluca Iavetz Geoff Morschel
Luminita Antin Gavin Stevenson
Jesuina De Araujo
I would also like to thank Syed Hasnain and Laura-Lee Innes from Environment Protection Authority Victoria for their input and advice into the planning of this study
Paul Armishaw Manager, Chemical Reference Values PO Box 138, North Ryde NSW 1670 Phone: 61-2-9449 0149 Fax: 61-2-9449 0123 [email protected]
ii
iii
TABLE OF CONTENTS
SUMMARY 1 1 INTRODUCTION 2
1.1 NMI Proficiency Testing Program 2 1.2 Study Background 2 1.3 Study Aims 2 1.4 Study Conduct 2
2 STUDY INFORMATION 3 2.1 Study Timetable 3 2.2 Participation 3 2.3 Test Material Preparation 3 2.4 Test Material Homogeneity and Stability Testing 3 2.5 Laboratory Code 3 2.6 Sample Storage, Dispatch and Receipt 3 2.7 Instructions to Participants 3 2.8 Interim Report 4
3 PARTICIPANT LABORATORY INFORMATION 5 3.1 Test Methods Reported by Participants 5 3.2 Basis of Participants’ Measurement Uncertainty Estimates 9 3.3 Participants’ Comments 10
4 PRESENTATION OF RESULTS AND STATISTICAL ANALYSIS 11 4.1 Results Summary 11 4.2 Assigned Value 11 4.3 Between-Laboratory Coefficient of Variation 11 4.4 Target Standard Deviation 11 4.5 z-Score 12 4.6 En-Score 12 4.7 Traceability and Measurement Uncertainty 12 4.8 Robust Average 12
5 TABLES AND FIGURES 13 6 DISCUSSION OF RESULTS 39
6.1 Assigned Value 39 6.2 Measurement Uncertainty Reported by Participants 39 6.3 z-Score 40 6.4 En-Score 40 6.5 Participants’ Methods 40
7 REFERENCES 42 APPENDIX 1 – PARTICIPATING LABORATORIES 43 APPENDIX 2 - SAMPLE PREPARATION, HOMOGENEITY TESTING AND STABILITY CHECK 44 APPENDIX 3 - ROBUST AVERAGE AND ASSOCIATED UNCERTAINTY 52 APPENDIX 4 - ACRONYMS AND ABBREVIATIONS 53
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 1
SUMMARY
AQA 15-03 was conducted in March 2015. Eleven laboratories registered to participate and ten submitted results. Four test samples were prepared in the NMI Sydney laboratory and consisted of:
Two soil samples, one from a contaminated site (S1) and one formulated to contain 10 µg/kg total and linear PFOS and PFOA (S2).
Two water samples, one from a contaminated site (S3) and one formulated to contain 4 µg/L total and linear PFOS and PFOA (S4).
Each participant received a set of four samples: two 10 g soil samples and two water samples, one 500 mL and one 100 mL. The samples were sufficiently homogeneous and stable for evaluation of participants’ performance. Of a possible 120 numeric results a total of 88 (73%) were submitted. The assigned values were the robust average of participants results.
The outcomes of the study were assessed against the aims as follows:
to compare the performances of participant laboratories and to assess their accuracy in the measurement of total and linear PFOS and PFOA in water and soil matrices;
Laboratory performance was assessed using both z-scores and En-scores. Of 78 z-scores, 62 (79%) were satisfactory with |z| 2. Of 78 En-scores, 58 (74%) were satisfactory with |En| 1. Laboratories 5 and 10 returned satisfactory z and En-scores for all analytes for which scores were calculated.
evaluate the laboratories’ methods; Participants used a variety of methods for extraction. No correlation between results and method was evident. The analytical detection method of choice was LC-MS/MS (ten participants) and LC-MS (one participant).
develop the practical application of traceability and measurement uncertainty and provide participants with information that will be useful in assessing their uncertainty estimates.
Seventy-nine of eighty-eight numeric results (90%) were reported with an associated estimate of expanded measurement uncertainty.
The magnitude of these expanded uncertainties was within the range 0.9% to 100% of the reported value. Ten were less than 10% relative, which the study coordinator believes is unrealistically small for a routine PFOS/PFOA measurement. Laboratory 9 did not report expanded measurement uncertainties.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 2
1 INTRODUCTION 1.1 NMI Proficiency Testing Program
The National Measurement Institute (NMI) is responsible for Australia’s national measurement infrastructure, providing a range of services including a chemical proficiency testing program. Proficiency testing (PT) is: ‘evaluation of participant performance against pre-established criteria by means of interlaboratory comparison.’1 NMI PT studies target chemical testing in areas of high public significance such as trade, environment, law enforcement and food safety. NMI offers studies in: pesticide residues in fruit and vegetables, soil and water; petroleum hydrocarbons in soil and water; metals in soil, water, food and pharmaceuticals; controlled drug assay; folic acid in flour; and allergens in food. 1.2 Study Background
Perfluorooctane sulfonate (PFOS) and Perfluorooctanoic acid (PFOA) are synthetic fluorinated chemicals used as a surface-active agent and in a variety of products such as firefighting foams, coating additives and cleaning products.
They are extremely persistent in the environment and resistant to typical environmental degradation processes. PFOS and PFOA were in 2010 added to the list of chemicals regulated under the international Stockholm Convention for Persistent Organic Pollutants, to which Australia is a signatory.
To ensure the quality and comparability of PFOS and PFOA measurements in Australia, the Cooperative Research Centre for Contamination Assessment and Remediation in the Environment (CRC CARE), under the Better Measurement program, instituted a project to facilitate improved PFOS/PFOA analysis through Proficiency Testing (PT) development and education; NMI and EPA Victoria are partners in this project. 1.3 Study Aims
The aims of the study were to:
compare the performances of participant laboratories and assess their accuracy in the measurement of total and linear PFOS and PFOA in water and soil matrices;
evaluate the laboratories’ test methods; and develop the practical application of traceability and measurement uncertainty and
provide participants with information that will be useful in assessing their uncertainty estimates.
1.4 Study Conduct
The conduct of NMI proficiency tests is described in the NMI Chemical Proficiency Testing Study Protocol.2 The statistical methods used are described in the NMI Chemical Proficiency Statistical Manual.3 These documents have been prepared with reference to ISO 170431 and The International Harmonized Protocol for Proficiency Testing of (Chemical) Analytical Laboratories.4 PFOS and PFOA in environmental samples are outside the scope of NMI’s accreditation as a proficiency testing provider, nevertheless, the study was conducted according to NMI’s ISO 17043 compliant procedures.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 3
2 STUDY INFORMATION 2.1 Study Timetable
The timetable of the study was: Invitation issued 27 February 2015 Samples dispatched 02 March 2015 Results due 10 April 2015 Interim report issued 17 April 2015
2.2 Participation
One hundred and twenty two Australian and international laboratories were invited to participate. Eleven laboratories participated (see Appendix 1) and ten submitted results by the due date. 2.3 Test Material Preparation
Four test samples were prepared in February 2015. Care was taken to avoid any Teflon contamination during sample preparation.
Two soil samples each 10 g, one from a contaminated site and the other one formulated to contain 10 µg/kg of total and linear PFOS and PFOA.
Two water samples, Sample S3 was 500 mL of water from a contaminated site and Sample S4 was 100 mL of water formulated to contain 4 µg/L of total and linear PFOS and PFOA.
2.4 Test Material Homogeneity and Stability Testing
The preparation of the samples and their testing for homogeneity is described in Appendix 2. Sample preparation has been demonstrated to yield sufficiently homogeneous samples.
A stability study of the test samples at room temperature was also carried out over the period of two months (starting from 04/03/2015 when the test samples were prepared). The stability results are given in Appendix 2. Samples were demonstrated to be sufficiently stable. 2.5 Laboratory Code
All laboratories that agreed to participate were assigned a confidential code number. 2.6 Sample Storage, Dispatch and Receipt
The test samples were refrigerated at 4ºC prior to dispatch. Participants were sent two 10 g jar of soil for each of Sample S1 and Sample S2, one 500 mL bottle of Sample S3 and one 100 mL bottle of Sample S4. The samples were packed in a foam box with a cooler brick and sent by courier on 02 March 2015.
The following items were packaged with the samples:
a covering letter which included a description of the test samples and instructions for participants; and
a faxback form for participants to confirm the receipt and condition of the samples. An Excel spreadsheet for the electronic reporting of results was e-mailed to participants. 2.7 Instructions to Participants
Participants were instructed as follows: Quantitatively analyse the samples using your normal test method.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 4
Report results in units of µg/kg on as received basis for soil samples Report results in units of µg/L for water samples For each analyte in each sample report a single result expressed as if reporting
to a client (ie correct for recovery or not, according to your standard procedure). This figure will be used in all statistical analysis in the study report.
For each analyte in each sample report the associated expanded measurement uncertainty (eg 0.50 0.02 µg/kg). Report any analyte not tested as NT. No limit of reporting has been set for this study. Report results as you would to a client, applying the limit of reporting of the method used for analysis. Report the basis of your uncertainty estimates (eg uncertainty budget, repeatability precision, long term result variability). Return the completed results sheet by e-mail
[email protected]), Please return completed result sheet by 03 April 2015. Late results cannot be included in the study report. 2.8 Interim Report
An interim report was emailed to participants on 17 April 2015.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 5
3 PARTICIPANT LABORATORY INFORMATION 3.1 Test Methods Reported by Participants
Table 1 Test methods – Samples S1 and S2 soil
Lab Code
Sample weight
(g) Extraction solvent(s): Clean-up Equipment Column type Column
dimensions
Extra column for
blank separation
1 2 MTBE/TBAS ion pairing reagent Low resolution LC-MS/MS C18 10cm x 4.6mm x
2.6µm
2 10 Acetonitrile Low resolution LC-MS/MS Pursuit 5PFP 150x4.6
3 1 AcCN/MilliQ Water Low resolution LC-MS/MS C18 5mm*2.1mm, 1.7 µm
4 4 Methanol/EtAOc/H3P3O4 LC-MS C18 with guard column
4.6mmX100mmX3.5 µm
5 5, 10 Methanol/KOH Low resolution LC-MS/MS C18 yes
7 2 Acetone/Hexane Low resolution LC-MS/MS
8 2 Methanol/NaOH graphitised carbon Low resolution LC-MS/MS C18 100mm x 2.1mm x
1.8um
9 1.0428, 1.0697 Methanol SPE,HLB
200mg/6mL Low resolution LC-MS/MS C18 10 cmx2.0 mmx3 um,110Å
10 0.25/1 MeOH/Ammonium hydroxide (99/1)
Envi-Carb SPE Low resolution LC-MS/MS C18 Gemini 50 x 2mm yes
11 2.5 AcCN/MeOH/NaOH Filtration Low resolution LC-MS/MS EC- C18 7.5mm x 3mmx 2.7 µm
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 6
Table 2 Test methods – Samples S1 and S2 soil (cont’d)
Lab code Internal std Recovery std Recovery
correction If yes, which
method? Method used Blank corrected
1 13C4-PFOS/13C4-PFOA
ISO COMMITTEE DRAFT ISO/CD 25101. Reference number ISO/TC
147 / SC 2N 0914. Date 2006-11-08. no
2 13C PFOS/13C PFOA PFOA,PFOS yes 13C Int Std In house no
3 EPA537 no
4 PFOA/PFOS In-House no
5 13C PFOS/13C PFOA 18O PFOA yes 13C Int Std EPA537 no
7 13C PFOS/13C PFOA PFOA,PFOS no
8 13C PFOS/13C PFOA M8PFOA yes 13C Int Std in-house no
9 13C PFOS/13C PFOA yes 13C Int Std no
10 13C PFOS/13C PFOA 13C PFOS/13C PFOA yes 13C Int Std no
11 4-Octylbenzenesulfonic acid sodium salt
no
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 7
Table 3 Test methods – Samples S3 and S4 water
Lab Code
Sample volume (mL) Pretreatment Extraction
technique: Extraction
solvent Clean Up: Equipment Column Type
Column dimensions
Extra column for blank separation (yes/no)
1 1 Acidification Direct injection Low resolution LC-
MS/MS C18 10cm x 4.6mm x 2.6µm
2 100 SPE Methanol SPE Low resolution LC-MS/MS
Pursuit 5PFP 150x4.6
3 1 Acidification Dilution
and filtration
Low resolution LC-MS/MS C18 5mm*2.1mm,
1.7u
4 1 LC-MS C18 with
guard column
4.6mmX100mmX3.5 µm
5 S3: 250
Acidification SPE Methanol SPE Low resolution LC-MS/MS C18 yes
S4: 50
7 S3: 522
SPE Methanol 0.32% Ammonia Low resolution LC-
MS/MS S4: 101
8 100 SPE Methanol Low resolution LC-MS/MS C18 100mm x
2.1mm x 1.8um
9 S3: 100
SPE Methanol SPE,HLB 200mg/6mL
Low resolution LC-MS/MS C18 10 cmx2.0
mmx3 um,110Å S4: 5
10 S3: 200
SPE 0.2 % ammonia/MeOH
Strata X-AW SPE
Low resolution LC-MS/MS
C18 Gemini 50 x 2 mm yes
S4: 10
11 0.01 Direct
injection Low resolution LC-MS/MS EC-C18 7.5cm x 3mm x
2.7µm
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 8
Table 4 Test methods – Samples S3 and S4 water (cont’d)
Lab code Internal std Recovery std Recovery correction If yes, which method? If yes, which method? Blank corrected
(yes/no)
1 13C4-PFOS/13C4-PFOA
ISO COMMITTEE DRAFT ISO/CD 25101. Reference
number ISO/TC 147 / SC 2N 0914. Date 2006-11-08.
no
2 13C PFOS/13C PFOA PFOA,PFOS yes 13C Int Std no
3 EPA537 no
4 PFOA/PFOS no
5 13C PFOS/13C PFOA 18O PFOA yes 13C Int Std EPA537 no
7 13C PFOS/13C PFOA PFOA,PFOS no
8 13C PFOS/13C PFOA M8PFOA yes 13C Int Std no
9 13C PFOS/13C PFOA yes 13C Int Std no
10 13C PFOS/13C PFOA 13C PFOS/13C PFOA yes 13C Int Std no
11 4-Octylbenzenesulfonic acid sodium salt no
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 9
3.2 Basis of Participants’ Measurement Uncertainty Estimates
Table 5 Basis of Participants’ Uncertainty Estimate
Lab Code
Approach to Estimating MU Information Sources for MU Estimation Guide Document for Estimating MU
Precisiona Method Biasa
1 Top Down - precision and estimates of the method and laboratory bias
Control samples Duplicate analysis Instrument calibration
Laboratory bias from PT studies Recoveries of sample spike
NATA Technical Note 33
2 Method validation data Control samples Duplicate analysis
Instrument calibration Recoveries of sample spike Standard Purity
NMI Uncertainty Course
3 Professional judgment Control samples Instrument calibration Recoveries of sample spike
NATA Technical Note 33
4 Standard deviation of replicate analyses multiplied by 2 or 3
Duplicate analysis Recoveries of sample spike NEPM 1999, Schedule B(3)
5 Top Down - precision and estimates of the method and laboratory bias
Control samples
Recoveries of sample spike
ASTM E2554-13 Standard Practice for Estimating and Monitoring the Uncertainty of Test Results of a Test Method Using Control Chart Techniques
7 Standard deviation of replicate analyses multiplied by 2 or 3
Control samples Instrument calibration
Laboratory bias from PT studies Recoveries of sample spike
NATA Technical Note 33
8 Standard deviation of replicate analyses multiplied by 2 or 3
Duplicate analysis Standard Purity
9 Instrument calibration
10 Top Down - precision and estimates of the method and laboratory bias
Duplicate analysis Instrument calibration
Instrument calibration Recoveries of sample spike
Measurement uncertainty was calculated taking into account the relative standard deviation of duplicate analysis of the samples, the inter days standard deviation and the bias from recovery spikes.
11 Top Down - precision and estimates of the method and laboratory bias
Duplicate analysis Instrument calibration
Instrument calibration Recoveries of sample spike Standard purity
NATA Technical Note 33
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 10
3.3 Participants’ Comments
The study co-ordinator welcomes comments or suggestions from participants about this study or possible future studies. Participants’ comments are reproduced in Table 6.
Table 6 Participants’ Comments
Lab Code
Sample Participants’ Comments
1 All Linear PFOS was quantified using PFOS multipoint calibration standards containing both linear and branched PFOS.
4 All LORs for soil samples: PFOS 1 µg/kg and PFOA 1.5 µg/kg LORs for water samples: PFOS 0.5 µg/L and PFOA 1 µg/L
5 All 6:2 FTS, PFHxA and PFBA to be included for future studies.
9 S3 and S4 MDL : PFOA=5.40 ng/L;Total PFOS*=6.55 ng/L;Linear PFOS*=6.55 ng/L
10 All Calculation of total PFOS based on linear PFOS
11 All Results are not corrected for recovery.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 11
4 PRESENTATION OF RESULTS AND STATISTICAL ANALYSIS 4.1 Results Summary
Participant results are listed in Tables 6 to 12 with the summary statistics robust average, mean, median, maximum, minimum, robust standard deviation (SDrob) and robust coefficient of variation (CVrob). Bar charts of results and performance scores are presented in Figures 2 to 8.
An example chart with interpretation guide is shown in Figure 1.
Figure 1 Guide to Presentation of Results
4.2 Assigned Value
The assigned value is defined1 as: ‘value attributed to a particular property of a proficiency test item.’ In this study property is the mass fraction of analyte. Assigned values were the robust average of participants’ results; the expanded uncertainties were estimated from the associated robust standard deviations. 4.3 Between-Laboratory Coefficient of Variation
The between laboratory coefficient of variation is a measure of the between laboratory variation that in the judgement of the study coordinator would be expected from participants given the analyte concentration. It is important to note this is a performance measure set by the study coordinator; it is not the coefficient of variation of participant results. 4.4 Target Standard Deviation
The target standard deviation (σ) is the product of the assigned value () and the between-laboratory coefficient of variation (CV). This value is used for calculation of participant z-score.
σ = * CV Equation 1
Assigned value and associated expanded uncertainty (coverage factor is k= 2).
Uncertainties reported by participants.
Distribution of results around the assigned value as kernel density estimate (illustrates participant consensus).
Independent estimates of analyte concentration with associated uncertainties (coverage factor is 2). Md = Median (of participants’ results) R.A = Robust Average HV = NMI Homogeneity Value S = Spike (formulated concentration)
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 12
4.5 z-Score
For each participant result a z-score is calculated according to Equation 2 below:
)( Xz Equation 2
where: z is z-score is participant result is the study assigned value is the target standard deviation from equation 1
A z-score with absolute value (|z|): |z| 2 is satisfactory; 2 < |z| 3 is questionable; |z| > 3 is unsatisfactory.
4.6 En-Score
The En-score is complementary to the z-score in assessment of laboratory performance. En-score includes measurement uncertainty and is calculated according to Equation 3 below:
22
)(
X
nUUXE
Equation 3
where:
nE is En-score is a participant’s result is the assigned value U is the expanded uncertainty of the participant’s result
XU is the expanded uncertainty of the assigned value
An En-score with absolute value (|En|): |En| 1 is satisfactory; |En| > 1 is unsatisfactory.
4.7 Traceability and Measurement Uncertainty
Laboratories accredited to ISO/IEC Standard 17025:20055 must establish and demonstrate the traceability and measurement uncertainty associated with their test results.
Guidelines for quantifying uncertainty in analytical measurement are described in the Eurachem /CITAC Guide.6
4.8 Robust Average
The robust averages and associated expanded measurement uncertainties were calculated using the procedure described in ‘ISO13528:2005(E), Statistical methods for use in proficiency testing by interlaboratory comparisons’.7
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 13
5 TABLES AND FIGURES Table 7
Sample Details Sample No. S1 Matrix. Soil Analyte. Linear PFOS Units µg/kg Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 122 20.4 75 -1.35 -0.81 2 NT NT NT 3 220 50 115 1.59 0.73 4 286.8 13.2 107 3.59 2.23 5 183.8 46.0 NR 0.50 0.24 7 108 20 92 -1.77 -1.06 8 118 31 132 -1.47 -0.81 9 144 NR 116 -0.69 -0.44 10 190.8 20 86 0.71 0.43 11 250 150 NT 2.49 0.52 Statistics Assigned Value* 167 52 Spike Not Spiked Homogeneity Value 210 42 Robust Average 180 60 Median 184 70 Mean 180 N 9 Max. 286.8 Min. 108 Robust SD 59 Robust CV 35% *Robust average excluding laboratory 4
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 14
Figure 2
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 15
Table 8 Sample Details Sample No. S1 Matrix. Soil Analyte. PFOA Units µg/kg Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 5.53 2.04 104 -0.23 -0.10 2 7.5 2.3 100 1.47 0.61 3 10 5 105 3.62 0.80 4 5.3 0.2 65 -0.43 -0.31 5 5.5 1.65 NR -0.26 -0.13 7 3.2 0.8 123 -2.24 -1.45 8 4.7 1.1 100 -0.95 -0.57 9 8.63 NR 110 2.44 1.77 10 5.5 0.4 87 -0.26 -0.18 11 3.8 2.6 NT -1.72 -0.66 Statistics Assigned Value 5.8 1.6 Spike Not Spiked Homogeneity Value 5.8 1.2 Robust Average 5.8 1.6 Median 5.5 1.3 Mean 6.0 N 10 Max. 10 Min. 3.2 Robust SD 2.0 Robust CV 35%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 16
Figure 3
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 17
Table 9 Sample Details Sample No. S1 Matrix. Soil Analyte. Total PFOS Units µg/kg Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 191 31.6 75 -1.20 -0.83 2 301 90 100 1.00 0.45 3 NR NR NR 4 352.0 3.1 107 2.01 1.55 5 256.9 64.2 NR 0.12 0.06 7 190 48 92 -1.22 -0.75 8 163 42 156 -1.75 -1.14 9 261 NR 116 0.20 0.15 10 290.0 31 86 0.78 0.54 11 NT NT NT Statistics Assigned Value 251 65 Spike Not Spiked Homogeneity Value 334 67 Robust Average 251 65 Median 259 68 Mean 251 N 8 Max. 352 Min. 163 Robust SD 74 Robust CV 30%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 18
Figure 4
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 19
Table 10 Sample Details Sample No. S2 Matrix. Soil Analyte. Linear PFOS Units µg/kg Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 5.07 1.37 75 -2.05 -1.74 2 NT NT NT 3 10 5 NR 0.81 0.27 4 <LOR NT 130 5 9.3 2.3 NR 0.41 0.25 7 5.0 1.2 92 -2.09 -1.87 8 8.9 2.3 95 0.17 0.11 9 8.92 NR 116 0.19 0.21 10 9.9 1.1 89 0.76 0.70 11 18 11 NT 5.47 0.85 Statistics Assigned Value* 8.6 1.5 Spike 10.0 0.5 Homogeneity Value 9.8 2.0 Robust Average 9.1 1.3 Median 9.1 1.0 Mean 9.4 N 8 Max. 18 Min. 5 Robust SD 1.6 Robust CV 18% *Robust average excluding laboratory 11.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 20
Figure 5
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 21
Table 11 Sample Details Sample No. S2 Matrix. Soil Analyte. PFOA Units µg/kg Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 8.34 2.81 104 -0.31 -0.18 2 10 3.0 100 0.62 0.33 3 20 10 NR 6.24 1.10 4 8.6 0.3 93 -0.17 -0.21 5 9.8 2.45 NR 0.51 0.32 7 5.6 1.4 123 -1.85 -1.67 8 9.0 2.2 98 0.06 0.04 9 13.1 NR 110 2.36 3.00 10 10.0 0.7 86 0.62 0.70 11 5.4 3.5 NT -1.97 -0.93 Statistics Assigned Value* 8.9 1.4 Spike 10.0 0.5 Homogeneity Value 10.1 2.0 Robust Average 9.3 1.3 Median 9.4 1.0 Mean 10.0 N 10 Max. 20 Min. 5.4 Robust SD 1.7 Robust CV 19% * Robust average excluding laboratory 3.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 22
Figure 6
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 23
Table 12 Sample Details Sample No. S2 Matrix. Soil Analyte. Total PFOS Units µg/kg Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 5.07 1.37 75 -2.21 -2.69 2 9.7 2.9 100 0.33 0.20 3 NT NT NT 4 NT NT NT 5 9.3 2.3 NR 0.11 0.08 7 5.0 1.3 92 -2.25 -2.86 8 8.9 2.3 95 -0.11 -0.08 9 9.29 NR 116 0.10 0.32 10 9.9 1.1 89 0.44 0.64 11 NT NT NT Statistics Assigned Value 9.1 0.6 Spike 10.0 0.5 Homogeneity Value 9.8 2.0 Robust Average 9.1 0.6 Median 9.3 0.6 Mean 8.2 N 7 Max. 9.9 Min. 5 Robust SD 0.7 Robust CV 7.7%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 24
Figure 7
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 25
Table 13 Sample Details Sample No. S3 Matrix. Water Analyte. Linear PFOS Units µg/L Participant Results Lab Code Result Uncertainty Recovery 1 < 0.02 0.004 90 2 NT NT NT 3 <0.5 NR 96 4 <LOR NR 71 5 0.001 0.001 NR 7 <0.005 0.001 96 8 0.0030 0.0004 38 9 NR NR 94.0 10 0.0031 3.3E-04 77 11 <1 0.47 NT Statistics Assigned Value Not Set Spike Not Spiked Homogeneity Value 0.0035 0.0007 Robust Average 0.0030 0.0003 Median 0.0030 0.0004 Mean 0.0024 N 3 Max. 0.0031 Min. 0.001 Robust SD 0.0002 Robust CV 6.7%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 26
Figure 8
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 27
Table 14 Sample Details Sample No. S3 Matrix. Water Analyte. PFOA Units µg/L Participant Results Lab Code Result Uncertainty Recovery 1 < 0.02 0.005 91 2 <0.02 NR 99 3 <0.5 NR 101 4 <LOR NR 93 5 0.004 0.002 NR 7 <0.005 0.001 99 8 0.0036 0.0010 80 9 NR NR 111 10 0.0037 1.9E-04 81 11 <1 0.47 NT Statistics Assigned Value Not Set Spike Not Spiked Homogeneity Value 0.0039 0.0008 Robust Average 0.0037 0.0003 Median 0.0037 0.0004 Mean 0.0038 N 3 Max. 0.004 Min. 0.0036 Robust SD 0.0002 Robust CV 5.4%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 28
Figure 9
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 29
Table 15 Sample Details Sample No. S3 Matrix. Water Analyte. Total PFOS Units µg/L Participant Results Lab Code Result Uncertainty Recovery 1 < 0.02 0.004 90 2 <0.02 NR 121 3 NT NT NT 4 NT NT NT 5 0.002 0.001 NR 7 0.006 0.002 96 8 0.0055 0.0008 38 9 NR NR 94.0 10 0.0055 5.9E-04 77 11 NT NT NT Statistics Assigned Value Not Set Spike Not Spiked Homogeneity Value 0.0058 0.0012 Robust Average 0.0055 0.0006 Median 0.0055 0.0006 Mean 0.0048 N 4 Max. 0.006 Min. 0.002 Robust SD 0.0005 Robust CV 9.1%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 30
Figure 10
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 31
Table 16 Sample Details Sample No. S4 Matrix. Water Analyte. Linear PFOS Units µg/L Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 3.01 0.519 90 -1.14 -0.73 2 NT NT NT 3 4.3 1 NR 0.51 0.27 4 5.4 0.1 104 1.92 1.36 5 2.8 0.42 NR -1.41 -0.93 7 3.1 0.78 96 -1.03 -0.59 8 4.3 0.65 77 0.51 0.31 9 2.34 NR 94.0 -2.00 -1.42 10 3.9 0.41 94 0.00 0.00 11 5.7 1.6 NT 2.31 0.93 Statistics Assigned Value 3.9 1.1 Spike 4.0 0.2 Homogeneity Value 3.9 0.8 Robust Average 3.9 1.1 Median 3.9 1.0 Mean 3.9 N 9 Max. 5.7 Min. 2.34 Robust SD 1.3 Robust CV 33%
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 32
Figure 11
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 33
Table 17 Sample Details Sample No. S4 Matrix. Water Analyte. PFOA Units µg/L Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 3.92 0.666 91 0.30 0.32 2 3.7 1.1 99 0.00 0.00 3 3.4 1 NR -0.41 -0.29 4 8.7 0.1 102 6.76 22.36 5 3.5 0.88 NR -0.27 -0.22 7 3.8 1.0 99 0.14 0.10 8 3.7 1.0 91 0.00 0.00 9 4.98 NR 111 1.73 6.40 10 3.9 0.2 98 0.27 0.71 11 3.6 1.1 NT -0.14 -0.09 Statistics Assigned Value* 3.7 0.2 Spike 4.0 0.2 Homogeneity Value 4.0 0.8 Robust Average 3.8 0.2 Median 3.8 0.2 Mean 4.3 N 10 Max. 8.7 Min. 3.4 Robust SD 0.3 Robust CV 6.8% * Robust average excluding laboratory 4.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 34
Figure 12
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 35
Table 18 Sample Details Sample No. S4 Matrix. Water Analyte. Total PFOS Units µg/L Participant Results Lab Code Result Uncertainty Recovery z-Score En-Score 1 3.01 0.519 90 -0.30 -0.20 2 0.94 0.28 121 -3.53 -2.67 3 NT NT NT 4 NT NT NT 5 2.8 0.42 NR -0.63 -0.44 7 3.1 0.78 96 -0.16 -0.09 8 4.3 0.65 77 1.72 1.07 9 2.38 NR 94.0 -1.28 -1.03 10 3.9 0.41 94 1.09 0.78 11 NT NT NT Statistics Assigned Value* 3.2 0.8 Spike 4.0 0.2 Homogeneity Value 3.9 0.8 Robust Average 3.0 1.0 Median 3.0 0.9 Mean 2.9 N 7 Max. 4.3 Min. 0.94 Robust SD 0.7 Robust CV 23% * Robust average excluding laboratory 2.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 36
Figure 13
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 37
Figure 14 z-Score Dispersal by Laboratory
Figure 15 z-Score Dispersal by Analyte
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 38
Figure 16 En-Score Dispersal by Laboratory
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 39
6 DISCUSSION OF RESULTS 6.1 Assigned Value
The robust average of participants’ results was used as the assigned value for Samples S1, S2 and S4. The robust averages and associated expanded uncertainties, were calculated using the procedure described in ‘ISO13528:2005(E), Statistical methods for use in proficiency testing by interlaboratory comparisons’.7 The calculation of the expanded uncertainty for the robust average of PFOA in Sample S2 is presented in Appendix 3. Results less than 50% and greater than 150% of the robust average were removed before calculation of the assigned value.3,4 Assigned values for spiked Samples S2 and S4 were within the range 80– 98% of the spiked concentration for that analyte (Table 19). No assigned values were calculated for analytes in Sample S3 because only three results were reported. Traceability: The consensus of participants’ results is not traceable to any external reference, so although expressed in SI units, metrological traceability has not been established.
Table 19 Comparison of Assigned Value and Spiked Concentration.
Sample Analyte Spiked Concentration
Assigned Value
Assigned/ Spike (%)
S2 Linear PFOS 10 8.6 86
S2 Total PFOS 10 9.1 91
S2 PFOA 10 8.9 89
S4 Linear PFOS 4 3.9 98
S4 Total PFOS 4 3.2 80
S4 PFOA 4 3.7 93
6.2 Measurement Uncertainty Reported by Participants
Participants were asked to report an estimate of the expanded uncertainty associated with their results and the basis of this uncertainty estimate (Table 4).
It is a requirement of the ISO Standard 17025 that laboratories have procedures to estimate the uncertainty of chemical measurements and to report this uncertainty in specific circumstances, including: ‘when the client’s instruction so requires.’ Seventy-nine of eighty-eight results (90%) were reported with an associated estimate of expanded measurement uncertainty. Laboratory 9 did not report expanded measurement uncertainties.
The magnitude of the reported expanded uncertainties was within the range 0.9% to 100% of the reported value. Ten were less than 10% relative, which the study coordinator believes is unrealistically small for a routine PFOS/PFOA measurement. Results returning a satisfactory z-score but an unsatisfactory En-score may have underestimated the uncertainty. Some participants attached an estimate of the expanded measurement uncertainty to a result reported as less than their limit of reporting.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 40
In some cases the results were reported with an inappropriate number of significant figures. The recommended format is to write uncertainty to no more than two significant figures and then to write the result with the corresponding number of decimal places (for example instead of 9.8 ± 2.45 mg/kg better report 9.8 ± 2.5 mg/kg)6. 6.3 z-Score
A target standard deviation equivalent to 20% coefficient of variation (CV) was used to calculate z-scores. The between laboratory coefficient of variation predicted by the modified Horwitz equation8 and the between laboratories CV are presented for comparison in Table 20.
Table 20 Target standard deviation, between laboratories CV and modified Horwitz values
Sample Analyte Assigned
value
Unit
Target SD
(as CV, %)
Modified Horwitz CV
(%)
Between laboratories’
CV (%)
S1 Linear PFOS 167 µg/kg 20 21 35
S1 Total PFOS 251 µg/kg 20 20 30
S1 PFOA 5.8 µg/kg 20 22 35
S2 Linear PFOS 8.6 µg/kg 20 22 18
S2 Total PFOS 9.1 µg/kg 20 22 7.7
S2 PFOA 8.9 µg/kg 20 22 19
S4 Linear PFOS 3.9 µg/L 20 22 33
S4 Total PFOS 3.2 µg/L 20 22 23
S4 PFOA 3.7 µg/L 20 22 6.8
The dispersal of participants’ z-scores is graphically presented in Figure 14 and by analyte in Figure 15. Of seventy-eight results for which z-scores were calculated, 62 (79%) returned a satisfactory z-score of |z| 2. Laboratories 5, 8 and 10 returned satisfactory z-scores for all analytes for which z-scores were calculated. 6.4 En-Score
Where a laboratory did not report an uncertainty estimate an uncertainty of zero (0) was used to calculate the En-score.
Of seventy-eight En-scores, 58 (74%) were satisfactory with |En| 1. Laboratories 5 and 10 returned satisfactory En-score for all analytes for which En-scores were calculated.
The dispersal of participants’ En-scores is presented in Figure 16. 6.5 Participants’ Methods
Participants were requested to analyse the samples using their normal test method and to report a single result as they would normally report to a client. Results reported in this way reflect the true variability of results reported to laboratory clients. The method descriptions provided by participants are presented in Tables 1 to 4.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 41
Soil samples Extraction solvent: ion pairing (1), methanol (1), alkaline methanol (3), acetonitrile (3) and acetone/hexane (1). Clean-up: graphitised carbon (1) and SPE (2).
Water samples Pre-treatment: Acidification (3) Extraction technique: Direct injection (2), Dilution and filtration (1), SPE (6) and not specified (1). Extraction solvents: methanol (4), ammonia/methanol (2) and not specified (3). Two laboratories used extra column to suppress instrument background contamination due to Teflon tubing for blank separation.
For both matrices the analytical detection method of choice was LC-MS (1) or LC-MS/MS (9). Every laboratory used a different extraction and sample preparation method. Due to the limited amount of data and the variety of analytical methods used no significant trend with extraction and sample preparation was identified.
Five laboratories reported being accredited to ISO 17025.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 42
7 REFERENCES
[1] ISO/IEC 17043:2010, Conformity assessment – General requirements for proficiency testing, ISO, Geneva.
[2] NMI Chemical Proficiency Testing Study Protocol. http://www.measurement.gov.au → products and services → chemical proficiency testing → details of our program → Study Protocol [3] NMI Chemical Proficiency Testing Statistical Manual. http://www.measurement.gov.au
→ products and services → chemical proficiency testing → details of our program → Statistical Manual
[4] Thompson, M. Ellison S. L. R and Wood, R., The international harmonized protocol for proficiency testing of (chemical) analytical laboratories, Pure Appl. Chem. 78, 145-196, 2006. [5] ISO/IEC 17025:2005, General requirements for the competence of testing and
calibration laboratories, ISO, Geneva. [6] Eurachem/CITAC Guide Quantifying Uncertainty in Analytical Measurement, Third
edition, (2012), http://www.eurachem.org/images/stories/Guides/pdf/QUAM2012-P1.pdf (Accessed July 2014)
[7] ISO13528:2005(E), Statistical methods for use in proficiency testing by interlaboratory comparisons, ISO, Geneva.
[8] Thompson, M., Recent trends in inter-laboratory precision at ppb and sub-ppb concentrations in relation to fitness for purpose criteria in proficiency testing, Analyst, 125, 385-386, 2000. [9] ISO Guide 35, Reference materials – General and statistical principles for certification,
ISO, Geneva, Switzerland, 2006.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 43
APPENDIX 1 – PARTICIPATING LABORATORIES
ACS Laboratories (Australia), VIC Advanced Analytical Australia, NSW
Analytical Reference Laboratory (WA) Pty Ltd, WA AsureQuality Limited, NEW ZEALAND
Australian Laboratory Services Sydney Environmental Laboratory, NSW
CERAR University of South Australia, SA
Eurofins mgt, Brisbane Laboratory, QLD National Institute of Environmental Analysis Environmental Analysis Laboratory, TAIWAN (R.O.C)
National Research Centre for Environmental Toxicology ENTOX, Queensland University, QLD
Queensland Health Forensic and Scientific Services, QLD
SGS Leeder Consulting, VIC
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 44
APPENDIX 2 - SAMPLE PREPARATION, HOMOGENEITY TESTING AND STABILITY CHECK A2.1 Sample Preparation
Sample S1: Soil was obtained from a contaminated site, placed in a Pyrex dish and allowed to dry in fume cupboard. Dried soil was then ground using Retsch Cross Beater Mill. Uncontaminated dried soil was sieved through 212 µm sieve. 255 g of contaminated soil and 250 g of uncontaminated soil were mixed for 24 h using a drum-hoop mixer. The mixture was divided into 10 g portions using a rotary sample divider and packed into labelled 50 mL polypropylene tubes.
Sample S2: 500.12 g of dried uncontaminated soil sieved through 212 µm sieve was placed in a stainless steel pot. A slurry was produced by adding acetone. The slurry was spiked with solutions of PFOA and linear PFOS prepared in methanol. The slurry was stirred using an IKA stirrer, left to evaporate overnight and purged with nitrogen for further drying. The soil was then mixed, dispensed into 10 g portions using a spatula and packed into labelled 50 mL polypropylene tubes. Sample S3: Contaminated water was placed into a stainless steel pot, mixed with an IKA stirrer and dispensed into labelled 500 mL plastic bottles using a Masterflex peristaltic pump. Sample S4: 5000.1 g of Milli-Q water was spiked with standard solutions of PFOA and linear PFOS prepared in methanol. The spiked water was stirred using an IKA stirrer and dispensed into labelled 100 mL plastic bottles using a Masterflex peristaltic pump.
All the samples were stored in a cool room at 40C prior to dispatch to participants.
A2.2 Sample Analysis and Homogeneity Testing
Homogeneity testing was based on that described in the International Protocol.4 Eight bottles were selected at random from each set of samples and analytes measured. The measurement were made under repeatability conditions in random order.
Soil samples S1 and S2 were prepared in duplicate by accurately weighing 1 g of sample and spiking with 50 µL of labelled surrogate standard in methanol. The samples were extracted by overnight tumbling in alkaline methanol (0.01 N potassium hydroxide), then centrifuged and a portion was filtered and evaporated. The filtered extract was reconstituted to 600 µL in mobile phase and spiked with 10 µL labelled recovery standard in methanol. Water samples S3 and S4 were prepared in duplicate by accurately weighing 100 g (S3) and 10 g (S4) then spiking with 20 µL (S3), and 50 µL (S4) of labelled surrogate standard in methanol. Portions of each sample were used to determine density to allow reporting by mass/volume. The samples were pre-treated with 1N acetic acid then extracted by solid phase extraction (Waters Oasis HLB, 6 mL, 500 mg, 60 µm particle size) under vacuum and eluted using methanol. After evaporation under nitrogen, the extract was reconstituted to 600 µL in mobile phase and spiked with 10 µL labelled recovery standard in methanol. All chemicals are analytical reagents or LCMS grade solvents. Instrument analysis was performed using an Agilent 1100 High Performance Liquid Chromatography (HPLC) coupled with an ABSciex 4000 Qtrap Mass spectrometer, operating in multiple reaction monitoring mode. 5 µL of extract was injected onto a Waters XBridge BEH column (1 mm x 150 mm x 3.5 µm, 130 Å) with a mobile phase gradient consisting of water:methanol (2 mM ammonium acetate). Two mass transitions were monitored for each target analyte and labelled surrogate, and abundance ratios checked. The instrument mass accuracy is calibrated annually during preventative maintenance, and the six point calibration curve
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 45
established three monthly. A continuing calibration standard is run with each batch of samples (±15% acceptance). A solvent batch blank is extracted and analysed with each batch, and sample results must be at least three times the level of any analyte detected in the batch blank to be reported. Quantification is based on the use of the labelled surrogates using relative retention factors from the multipoint calibration, and is corrected for surrogate recoveries. The analysis is based on USEPA 537 and used calibration, surrogate and recovery standards supplied by Wellington Laboratories, Canada.
Chromatographic separation of the main four PFOS isomers was performed by optimising column and mobile phase conditions. Total PFOS was determined by integrating all peaks that eluted from the LC column within the retention time window established during analysis of a technical PFOS mixture standard (Wellington Laboratories). Linear PFOS was determined by integrating the peak that co eluted with the linear labelled surrogate PFOS. Both total and linear PFOS were quantitated using linear PFOS calibration standards.
Results of the homogeneity testing are presented in tables 21-30. For each sample, the mean of the 16 measurements were used as the NMI homogeneity value. All samples were found to be sufficiently homogeneous for use in this PT study.
Table 21 Sample S1, Linear PFOS homogeneity testing results
Bottle Fill Number
Linear PFOS µg/kg Replicate 1 Replicate 2
2 215 221
12 217 205
13 214 217
17 210 221
22 208 209
24 209 207
29 203 205
31 209 196
Mean 210
CV 3.2%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.35 0.68 Pass
San/ 0.13 0.5 Pass
s2sam 16.6 358.3 Pass
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 46
Table 22 Sample S1, Total PFOS homogeneity testing results
Bottle Fill Number
Total PFOS µg/kg Replicate 1 Replicate 2
2 344 348
12 350 325
13 339 343
17 326 355
22 332 331
24 333 333
29 320 327
31 333 311
Mean 334
CV 3.5%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.41 0.68 Pass
San/ 0.17 0.5 Pass
s2sam 9.054 967.65 Pass
Table 23 Sample S1, PFOA homogeneity testing results
Bottle Fill Number
PFOA µg/kg Replicate 1 Replicate 2
2 5.9 5.7
12 6.1 5.8
13 5.9 6.1
17 5.7 5.8
22 5.7 5.9
24 5.5 6.1
29 5.7 5.7
31 6.1 5.5
Mean 5.8
CV 3.4%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.38 0.68 Pass
San/ 0.21 0.5 Pass
s2sam 0.000 0.32 Pass
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 47
Table 24 Sample S2, Total/Linear PFOS homogeneity testing results
Bottle Fill Number
Total/Linear PFOS µg/kg Replicate 1 Replicate 2
3 9.9 9.7
12 10.5 9.5
15 9.6 9.2
19 9.5 10.0
21 9.8 9.6
24 10.4 9.5
32 9.3 10.2
35 10.5 9.9
Mean 9.8
CV 4.2%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.29 0.68 Pass
San/ 0.24 0.5 Pass
s2sam 0.000 0.97 Pass
Table 25 Sample S2, PFOA homogeneity testing results
Bottle Fill Number
PFOA µg/kg Replicate 1 Replicate 2
3 9.2 10.7
12 10.3 10.3
15 10.5 9.4
19 9.7 10.0
21 10.1 9.8
24 10.6 10.3
32 8.8 10.5
35 10.7 10.9
Mean 10.1
CV 5.8%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.43 0.68 Pass
San/ 0.32 0.5 Pass
s2sam 0.000 1.26 Pass
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 48
Table 26 Sample S3, Linear PFOS homogeneity testing results
Bottle Fill Number
PFOA µg/L Replicate 1 Replicate 2
11 0.00422 0.00426
12 0.00325 0.00378
24 0.00307 0.00329
25 0.00380 0.00376
33 0.00316 0.00305
34 0.00368 0.00301
40 0.00323 0.00296
11 0.00422 0.00426
Mean 0.0035
CV 13%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.52 0.73 Pass
San/ 0.36 0.5 Pass
s2sam 0.000 0.00 Pass
Table 27 Sample S3, Total PFOS homogeneity testing results
Bottle Fill Number
Total/Linear PFOS µg/L Replicate 1 Replicate 2
11 0.00640 0.00676
12 0.00557 0.00572
24 0.00558 0.00542
25 0.00623 0.00612
33 0.00612 0.00553
34 0.00620 0.00516
40 0.00565 0.00512
11 0.00640 0.00676
Mean 0.0058
CV 8.3%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.57 0.73 Pass
San/ 0.32 0.5 Pass
s2sam 0.000 0.00 Pass
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 49
Table 28 Sample S3, PFOA homogeneity testing results
Bottle Fill Number
PFOA µg/L Replicate 1 Replicate 2
11 0.00375 0.00457
12 0.00414 0.00393
24 0.00419 0.00382
25 0.00383 0.00357
26 0.00397 0.00373
33 0.00394 0.00382
34 0.00387 0.00371
40 0.00341 0.00412
Mean 0.0039
CV 6.9%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.44 0.68 Pass
San/ 0.40 0.5 Pass
s2sam 0.000 0.00 Pass
Table 29 Sample S4, Total/Linear PFOS homogeneity testing results
Bottle Fill Number
PFOA µg/L Replicate 1 Replicate 2
2 3.9 4.0
5 4.2 3.8
10 4.1 3.5
12 4.1 3.6
22 3.8 3.7
28 4.1 3.8
32 3.8 4.1
37 4.0 3.4
Mean 3.9
CV 5.9%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.27 0.68 Pass
San/ 0.35 0.5 Pass
s2sam 0.000 0.20 Pass
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 50
Table 30 Sample S4, PFOA homogeneity testing results
Bottle Fill Number
PFOA µg/L Replicate 1 Replicate 2
2 4.0 3.9
5 4.2 4.0
10 4.0 3.9
12 4.1 4.1
22 4.0 4.2
28 4.2 4.1
32 3.9 4.1
37 4.0 3.6
Mean 4.0
CV 3.5%
Thompson and Fearn Homogeneity Tests Test Value Critical Result Cochran 0.56 0.68 Pass
San/ 0.17 0.5 Pass
s2sam 0.001 0.14 Pass
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 51
y = -0.1286x + 4.1559R² = 0.1505
0123456
Tota
l PFO
S (u
g/L)
Day 0 Day 36 Day 78
A2.3 Stability Testing
The stability of total and linear PFOS and PFOA in soil and water samples was assessed using a classical stability testing approach. The procedure was based on that described in ISO Guide 359. The storage condition used for stability testing was room temperature. The test samples were prepared in March 2015 and analysed over a period of two months (78 days).
Figure 17 Stability PFOA/PFOS in Sample S4
A linear approximation model was used to describe the changes in analyte concentration during the stability study. The concentration was plotted versus time. Day 0 was the date for the first measurement (4 March 2015). Three sets of data were obtained for each analyte in each sample. Each sample was analysed in duplicate. No significant change was observed in analytes concentration in Samples S1 to S4 over the entire period of testing (see example Figure 17 for Sample S4).
y = 0.0405x + 4.1289R² = 0.0876
0
1
2
3
4
5
6
PFO
A (u
g/L)
Day 78Day 36Day 0
y = -0.1205x + 4.1498R² = 0.1404
0
1
2
3
4
5
6
Line
ar P
FOS
(ug/
L)
Day 0 Day 36 Day 78
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 52
APPENDIX 3 - ROBUST AVERAGE AND ASSOCIATED UNCERTAINTY
The robust average was calculated using the procedure described in ‘ISO13258:2005(E), Statistical methods for use in proficiency testing by interlaboratory comparisons – Annex C’7 the uncertainty was estimated as:
urob av = 1.25*Srob av / p Equation 4
where: urob av robust average standard uncertainty Srob av robust average standard deviation p number of results
The expanded uncertainty (Urob av) is the standard uncertainty multiplied by a coverage factor of 2 at approximately 95% confidence level. The robust average of all ten results was calculated (Table 22).
Table 31 Robust average for all PFOA results in Sample S2
Lab. Code PFOA (µg/kg)
1 8.34 2 10 3 20 4 8.6 5 9.8 7 5.6 8 9.0 9 13.1 10 10.0 11 5.4
Robust average 9.3
Result from laboratory 3 was removed as this is outside of a range ± 50% of the robust average. The robust average and associated uncertainty of the remaining nine results were then used as the assigned value for this analyte (Table 23).
Table 32 Uncertainty estimate for PFOA in Sample S2 after removing outlier No. results (p) 9 Robust Average 8.94 µg/kg Srob av 1.70 µg/kg urob av 0.71 µg/kg k 2 Urob av 1.42 µg/kg
The assigned value for PFOA in Sample S2 is 8.9 1.4 µg/kg.
AQA 15-03 PFOS/PFOA IN SOIL AND WATER 53
APPENDIX 4 - ACRONYMS AND ABBREVIATIONS
CV Coefficient of Variation
EPA Environment Protection Authority ISO International Standards Organisation
LC Liquid Chromatography Max Maximum value in a set of results
Md Median Min Minimum value in a set of results
MS Mass Spectrometry NMI National Measurement Institute (of Australia)
NR Not Reported NT Not Tested
PFOA Perfluorooctanoic acid PFOS Perfluorooctane sulfonate
PT Proficiency Test Robust CV Robust Coefficient of Variation
Robust SD Robust Standard Deviation S Spiked or formulated concentration of a PT sample
SPE Solid Phase Extraction Target SD Target Standard Deviation
Target standard deviation