Академический Документы
Профессиональный Документы
Культура Документы
Wim Cofino
Steven Crum
Jacob de Boer
IVM
Institute for Environmental Studies
VU University Amsterdam
De Boelelaan 1087
1081 HV AMSTERDAM
The Netherlands
T +31-20-598 9555
F +31-20-598 9553
E info@ivm.vu.nl
Contents
Summary
Introduction
2.1
2.2
2.3
2.4
Study design
Material preparation
Methods used by participants
Data Assessment
9
9
9
11
Results
17
Discussion
19
4.1
4.2
Laboratory performance
Comparison with other ILSs
19
22
Conclusions
25
Acknowledgements
27
References
29
Appendices
31
Appendix 1
List of Participants
32
Appendix 2
33
Appendix 3
38
Appendix 4
Consistency of data
39
Appendix 5
40
Appendix 6
44
Summary
The first round of an international interlaboratory study (ILS) on chlorinated paraffins
(CPs) was organised by the Institute for Environmental Studies (IVM) in cooperation
with the proficiency testing scheme of QUASIMEME. This study was initiated because it
was generally agreed at a workshop on the analysis of CPs in Ostend, Belgium,
organised by QUASIMEME in March 2010, that there was a clear need for an
interlaboratory study, preferably designed in a step-wise way. The first objective of the
study was to assess the intercomparability of the data produced and secondly to
initiate improvements.
In total 15 laboratories participated in the present study on the analysis of CPs. The
participants were requested to quantify the total concentration of CPs and the
concentration of the three individual CPs 1,1,1,3,9,10-Hexachlorodecane (1,1,1,3,9,10HCD) , 1,1,1,3,11,12-Hexachlorododecane (1,1,1,3,11,12-HCDD), and
1,1,1,3,11,13,13,13-Octachlorotridecane (1,1,1,3,11,13,13,13-OCTD) in an iso-octane
solution of CPs. Two solutions containing CPs in known concentrations were provided
to the participants for calibration purposes. Participants used their in-house
quantification methods and techniques. The results were collected and statistically
evaluated using the Cofino statistics. Between lab coefficients of variation (CVs) and zscores were appointed to the laboratorys results as an expression of accuracy.
The majority of the laboratories (57-71%) obtained satisfactory z-scores for the
analysis of the three individual CPs. For individual CP analysis the model between lab
CVs varied between the compounds from 22 to 46%. For total CP a between lab CV of
56% was found. Since no other intercomparison studies were known for the analysis of
individual CPs, no comparison could be made with previous intercomparison studies.
Even though a variety of quantification techniques were used, the difference between
the median value and the target value of the total CP concentration has decreased
compared to a study of Tomy et al. (1999), in which all participants used the same
quantification technique.
The present study also showed that the overall performance of participants in the
analysis of total CPs has improved compared to the interlaboratory study of Pellizzato
et al. (2009) although the present study only dealt with a solution of CPs, while the
study of Pellizzato et al. (2009) dealt with an oil extract.
The overall experience in CP analysis in the present study varied between the
participants, but no significant difference could be calculated between the
performance of participants with different levels of experience.
Introduction
Chlorinated paraffins (CPs), also known as Polychlorinated alkanes (PCAs), are complex
mixtures of chlorinated n-alkanes with carbon chain lengths of 10 to 30 and a
chlorination degree between 30% and 70% by mass. Characterization of CPs is
performed by their alkane chain lengths. They are divided into three groups: shortchain (C10-13) (SCCP), medium-chain (C14-17) (MCCP) and long chain (C18-32) (LCCP)
CPs.
CPs are used in several industrial applications like flame-retardants in the rubber
industry, as high temperature lubricants and cutting fluids in the metalworking
industry and as additives in liquids, in paints and textile. The analysis of CPs is highly
complicated. There tens of thousands of congeners which make separation by gas
chromatography (GC) and even by GCxGC difficult. Alternative methods are scarce and
may suffer from false positive results. Therefore, the data reported on CPs up to the
present include very high uncertainties (100% or more).
Chlorinated paraffins are currently of utmost importance to the environmental analyst.
They are being produced in high volumes, they are under discussion in the United
Nations Environmental Programme (UNEP) for possible definition as a persistent
organic pollutant (POP), and they are a mandatory determinand in the European Water
Framework Directive. However, their analysis is subject to very large variation, as the
CP patterns are extremely complex. In March 2010, QUASIMEME organized a workshop
on the analysis of CPs in Ostend, Belgium. A number of critical issues in the analysis of
these CPs were discussed. It was generally agreed that there was a clear need for an
interlaboratory study, preferably designed in a step-wise way and accompanied by
expert comments.
The present study focuses on the analysis of CPs in a solution of undisclosed
concentration, to assess the possibility of producing precise and accurate data for a
range of CPs in a mixture.
In total 15 laboratories participated in the present ILS, of which 11 submitted data. The
ILS study focussed on the CPs listed in table 1.
Table 1
Full name
Abbreviation
1,1,1,3,9,10-Hexachlorodecane
1,1,1,3,11,12-Hexachlorododecane
1,1,1,3,11,13,13,13-Octachlorotridecane
1,1,1,3,9,10-HCD
1,1,1,3,11,12-HCDD
1,1,1,3,11,13,13,13-OCTD
Chemical
formula
C10H16Cl6
C12H20Cl6
C13H20Cl8
Chlorination
degree (% Cl)
61.0
56.4
61.7
This study was carried out by the Institute for Environmental Studies (IVM) in
collaboration with QUASIMEME (www.QUASIMEME.org).
2.1
Study design
The aim of the sample preparation was to select individual CPs as representative for
CPs in the environment. Unfortunately, only commercial standards with chlorine atoms
at the end carbon atoms of the CP backbone were available. Due to this lack of
standards, three congeners with CPs on the ends of the backbone were selected.
Three Ampoules were prepared and sent to the participants with the aim of quantifying
in triplicate the three individual CPs, mentioned in table 1, and the total amount of CPs
in ampoule C by using ampoule A and B (see chapter 2.2).
The first invitation for participation in the study was sent out in November 2010 and
the samples were distributed in August 2011.
2.2
Material preparation
Ampoule A
Ampoule A contained an iso-octane solution of three individual CPs of a known identity
and in a known concentration (see Table 2).
Table 2
Content of Ampoule A
Abbreviation
Concentration (ng/mL)
1,1,1,3,9,10-HCD
1366
1,1,1,3,11,12-HCDD
1389
1,1,1,3,11,13,13,13-OCTD
1009
Ampoule B
Ampoule B contained an iso-octane solution of a technical mixture of SCCPs (C10-C13,
51,5% Cl) in a disclosed concentration (53 g/mL).
Ampoule C
The iso-octane solution in ampoule C consisted of a technical mixture of SCCPs (C10C13, 55.5% Cl) spiked with the three individual CPs, which are mentioned in table 1.
Details on the concentrations of the CPs can be found in Table 3 and in Appendix 2.
2.3
10
varied. Almost half of the participants (46%) used chemical ionization (CI), while 31%
used electron impact (EI) and 23% used electron capture chemical ionization (ECNI) (see
Figure 2). Whenever EI was used the polarity used was positive, while the polarity used
with CI and ECNI was negative.
Ionization
Analysis method
GCxGC-ECD
8%
ECNI
23%
GC-MS/MS
23%
CI
46%
GC-MS
69%
EI
31%
Figure 1
Figure 2
For the GC separation often a DB-5MS column was used (46%), but also other columns
were utilised, such as Rtx-5MS and Rtx-1614 (see Figure 3). Splitless injection was used
(69%) as well as on-column injections (23%) and cooled injection with solvent venting
(8%).
DB-1MS xColumn
innowax
8%
Rtx-200
8%
Rtx-1614
8%
DB-5MS
46%
Restek
15%
Rtx-5MS
15%
Figure 3
2.3.1
Quantification
For the quantification of the concentration of total CPs in ampoule C participants were
requested to use ampoule B. Since the mixture in ampoule B contained CPs with 51.5%
Cl, and the mixture in ampoule C contained 55.5% Cl some participants reported that
an in-house mixture was a better match to quantify CPs in ampoule C. Participant CPP4
and CPP8 handed in two data sets. The first data set, marked with m1, was obtained by
quantification of total CP with ampoule B and the second data set, marked with m2,
11
was obtained with quantification with in-house mixtures. In-house mixtures were also
used by participants CPP3 for the quantification of total CPs in ampoule C.
The majority of the participants (64%) used in addition an internal standard, like 13C
Chlordane, 13C12-138-HxBDE, d10 anthracene, and 13C HCB, for the quantification of
CPs.
2.4
Data Assessment
The data assessment was carried out according to the principles employed in the data
assessment of the QUASIMEME laboratory performance studies. All data received from
the participants were entered into a database and assessed using a standard
procedure enabling direct comparison between participants. The approach to the
assessment is based on standard ISO 13528 (Thompson et al., 2006). Additions or
differences in the assessment from these standards are given or referred to in this
report. However, the assigned value and the laboratory assessment using z-scores is
calculated using the model developed by Cofino et al. (2000).
The standard, (ISO 13528), includes statistics for proficiency testing schemes, and
uses robust statistics as a basis for the assessment. However, it is generally
acknowledged that robust statistics cannot cope with more than 10% extreme values,
particularly with a skewed distribution. The Cofino model is able to routinely handle
these types of distribution and provide the best estimate of the consensus value,
which may be used as the assigned value.
The Cofino model is used routinely for QUASIMEME assessments. In this study the
Normal Distribution Approximation (NDA) has been employed (Cofino et al., 2000).
This approach includes all data in the evaluation and no subjective truncation or
trimming is made. This model has been further developed to include Left Censored
Values (LCV, which is the correct nomenclature for less than values) The development
of these models has been fully documented and published (Cofino et al., 2000, Cofino,
et al., 2005, Wells, et al., 2004) and accepted by the Dutch Accreditation Body. An
overview of the assessment with explanation and examples is given in the Assessment
Rules for the Evaluation of the QUASIMEME LP Studies Data (Wells and Scurfield, 2004).
The details of the Cofino Model were provided elsewhere (Wells et al., 2004, Wells and
Scurfield, 2004), but in summary the approach is as follows:
All data included in the assessment;
No data trimmed or down weighted;
Assigned values (AV) based on Cofino NDA model;
All LCV are also included, provided certain criteria are met (see chapter 2.4.2).
2.4.1 Plots
The performance of the laboratories in this study is illustrated in the z-score
histograms in Appendix 5, and an example of the plots is shown in Figure 4. Where
the assigned value for a determinand is indicative, the values are plotted as their
original reported concentrations. The rules for confirming whether the consensus
value should be an assigned value or an indicative value are given in the Assessment
Rules for the Evaluation of the QUASIMEME LP Studies Data (Wells and Scurfield, 2004)
with appropriate examples.
Normally, four plots are given for each determinand (see Figure 4). The upper left plot
provides an impression of the probability density function (PDF) model for all data
(black) and for the first mode (PMF1) model of the data (blue dotted). Superimposed on
12
these PDFs is a histogram of the individual measurements (grey bars). This plot shows
the distribution of the data as a whole, and of the data in the main mode (PMF1) model
on which the assigned value is based.
Figure 4. Examples of the graphical output of the Cofino Model statistics for
1,1,1,3,11,13,13,13-OCTD in ampoule C.
The Kilt Plot (Overlap Matrix) (upper right plot) provides an overview of the degree of
overlap of each pair of data. It gives a clear indication of the degree of homogeneity of
the data. As a key, the white areas indicate maximum overlap of the PDFs and,
therefore, highest agreement (an overlap of one implies that the two laboratories of
the pair report exactly the same results), while the black area show the pairs in poor
agreement.
The lower left plot is a ranked overview of all data with an error bar of 2 SD. The
numerical values are given in blue and the LCVs are given in red.
The ranked z-score plot (lower right) is based on the mean of the data, which is
normally also the assigned value. However, if there is any adjustment required to the
assigned value as a result of the assessment, e.g. use of the nominal concentration or
a trimmed value, then the final z-score given in the z-score histograms will reflect
these changes.
13
In some instances it is not possible to set an AV, and an indicative value is given. No
assessment of laboratory performance is given where an indicative value is set. An
overview of the assessment, with explanation, decision flowcharts and examples, is
given in the paper Assessment Rules for the evaluation of the QUASIMEME Laboratory
Performance Studies Data, available on the QUASIMEME website
(www.quasimeme.org). A summary of the categories is given below:
Category 1
For data with the number of numerical observations 7
An assigned value is based on the mean when 33% of values have a z-score of |Z| < 2.
Where < 33% of the data has |Z| < 2 the value is indicative. i.e. at least 33% must be in
good agreement.
Category 2
For data with the number of numerical observations > 3 and < 7
An assigned value is based on the mean when 70% of values have a z-score of |Z| < 3
and a minimum of 4 observations have |Z| < 2. Otherwise the value is indicative. i.e. for
small datasets, n > 3 and n < 7, there need to be very good agreement and a
maximum of one extreme value before an assigned value can be given.
Category 3
For data with the number of numerical observations < 4
No assigned value is given. Normally the median value is given as an indicative value.
Category 4
For data with the high Total Error% >100% in combination with bad performance
No assigned value is given.
z - score =
The values for the PE and CE are set by the QUASIMEME Scientific Assessment Group
and are monitored annually. The values are based on the following criteria:
- Consistency of the required standard of performance to enable participating
laboratories to monitor their assessment over time.
14
- Achievable targets in relation to the current state of the art and the level of
performance needed for national and international monitoring programmes.
The assessment is based on the standard ISO 17043 and z-scores. The QUASIMEME
model is designed to provide a consistent interpretation over the whole range of
concentration of analytes provided, including an assessment where LCVs are reported.
The proportional error is set at 12.5% for all matrices. This applies to all determinands.
The constant error (CE) has been set for each determinand or determinand group. This
value was initially set to reflect the limit of determination, but is at present more
closely related to the overall laboratory performance. The magnitude of the CE is set to
provide a constant assessment in terms of z-score regardless of concentration. The CE
is set at 0.025. Therefore at low concentrations the level of accuracy required to obtain
a satisfactory z-score is less stringent than at high concentrations.
Following usual practices e.g. ISO 17043, the z-scores can be interpreted as follows for
laboratories which take part in QUASIMEME to assure the quality of their data for use in
international marine monitoring programmes:
|Z| < 2 Satisfactory performance
2 <|Z| < 3 Questionable performance
|Z| > 3 Unsatisfactory performance
|Z| > 6 frequently points to gross errors (mistakes with units during reporting,
calculation or dilution errors, etc.).
The following schematic presentation illustrates the interpretation of the z-scores:
Questionable
performance
Unsatisfactory
performance
Satisfactory
performance
AV-3*TE
Assigned
Value (AV)
AV-2*TE
Unsatisfactory
performance
AV+3*TE
AV+2*TE
TE : total error
It is not possible to calculate a z-score for LCVs as LCVs represent a cut-off value
rather than continuous data. However, Quasimeme provides a simple quality criterion:
LCV/2 < (concentration corresponding to |z|=3): LCV consistent with assigned value
15
LCV/2 > (concentration corresponding to |z|=3): LCV inconsistent with assigned value,
i.e. LCV reported by laboratory much higher than numerical values reported by other
laboratories.
Z-score key:
S Satisfactory
Q Questionable
U Unsatisfactory
LCV key:
C Consistent
I Inconsistent
No data:
B - Blank
17
Results
The submitted results have been evaluated statistically and whenever the data met the
requirements (as described in chapter 2), an assigned value was established. Z-scores
were calculated based on the assigned value. Summary statistics are presented in
Table 3. A summary of the assigned values and the percentage of satisfactory to
unsatisfactory z-scores are presented in Table 4. Whenever less than values (LCV) were
submitted, the percentage of consistent and inconsistent LCVs with the assigned value
is given. The submitted data is presented in Appendix 2. Tables with individual zscores are presented in Appendix 3, consistencies of the individual results are
presented in Appendix 4 and z-score plots in Appendix 5.
Table 3
Determinand
Max**
Model
Between
Lab CV%
Percentage
n>LOQ
in PMF1
1,1,1,3,9,10-HCD
129
138
NA
138
148
51
810
22
57
1,1,1,3,11,12-HCDD
132
177
NA
177
197
122
2800
46
76
1,1,1,3,11,13,13,13OCTD
96
111
NA
111
120
90
1500
28
78
12000
NA
18500
56
76
14
Total CP
Table 4
Determinand
Total
Assigned Indicative
Error
value
Value
%
% data
received
% |Z|<2
(Satisfactory)
% 3>|Z|>2 % 6>|Z|>3
% |Z|>6
(Questi(Unsatis(Extreme)
onable)
factory)
1,1,1,3,9,10-HCD
138
NA
12.9
50
57
14
29
1,1,1,3,11,12-HCDD
177
NA
12.8
50
57
14
14
14
1,1,1,3,11,13,13,13OCTD
111
NA
13.0
50
71
14
14
Total CP
NA
18500
100
19
Discussion
In total 15 laboratories from all over the world participated in the present study. Of
those laboratories, 11 submitted data, of which three laboratories handed in two data
sets obtained with either two different analyses methods(see chapter 2.3), or with
different calibration solutions (see chapter 2.3.1). Seven participants were already
working on CP analysis for over 3 years, while two laboratories were analysing CPs
between 1 and 3 years. Only one laboratory had an experience in CP analysis of
shorter than one year. The experience of one participant was unclear. No significant
difference is observed between the reported concentrations of the less experienced
participants and the more experienced participants (Figure 5).
Total CP
50000
45000
40000
35000
30000
(ng/mL)
25000
20000
Indicative value
15000
Target value
10000
5000
0
experience
< 1 year
experience
1-3 years
experience
> 3 years
experience
unknown
Participating laboratories
4.1
Laboratory performance
As shown in Table 3, the model between lab CVs varied between the compounds from
22-56%. Satisfactory z-scores (i.e. Z|2|) were obtained from the majority (57-71%)
(depending on the determinand) of the participants that submitted results (see Table
4).
Comparing the assigned values (or indicative value in case of total CP) with the target
values (see Table 3 and Figure 6) shows that all assigned values (and indicative values)
are 1.11.5 fold higher than the target values, which might be due to coeluting CPs,
since the deviation is not observed by Participant CPP3 who used high resolution (HR)
MS.
Due to the deviation between assigned value and target value some participants, who
reported concentrations which agreed well with the target concentrations, received
higher, though still satisfactory, z-scores. For example participants CPP3, CPP10 and
CPP15m2 received z-scores of -1.9, -2.0 and -1.8 respectively for 1,1,1,3,11,12-HCDD,
Discussion
while the reported concentrations were 135, 131, and 136 ng/mL and the target value
was 132 ng/mL.
300
1,1,1,3,9,10-HCD (n=21)
250
200
(ng/mL)
150
Assigned value
Target value
100
50
0
Participating laboratories
300
1,1,1,3,11,12-HCDD (n=21)
250
200
Assigned value
150
(ng/mL)
20
Target value
100
50
Participating laboratories
200
21
1,1,1,3,11,13,13,13-OCTD (n=21)
180
160
140
120
Assigned value
(ng/mL)
100
Target value
80
60
40
20
0
Participating laboratories
50000
Total CP (n=42)
45000
40000
35000
30000
(ng/mL)
25000
20000
Indicative value
15000
Target value
10000
5000
0
Participating laboratories
Figure 6 Comparison of assigned values (or indicative value) with the target values.
In Figure 7 a comparison is given between the different ionization techniques used and
the reported concentration of total CP. Although only limited data is available per
method and no hard conclusions can be made, it looks like EI with positive polarity
gives results, which are closer to the target value. Only one participant reported results
on the concentrations of total CPs obtained with ECD, which resulted in a deviation of
only 4.7% of the target value. Two of the participants, marked with a asterisk in figure
7 used HR MS.
Discussion
Total CP
50000
45000
40000
35000
HR MS
30000
25000
(ng/mL)
22
20000
15000
Indicative value
HR MS
Target value
10000
5000
CI negative
quantified with
in-house std.
CI negative
quantified with
ampoule B
ECNI
EI positive
ECD
Participating laboratories
4.2
23
the quantification (see chapter 2.3.1), which resulted in a better performance of those
participants and a lower model between lab CV (see Figure 8).
Total CP
50000
CPP4 M1
45000
40000
35000
30000
(ng/mL)
25000
20000
CPP4 M2
CPP8 M1
Indicative value
15000
Target value
10000
CPP8 M2
5000
0
Quantiification with
in-house mixture
Participating laboratories
25
Conclusions
For individual CP analysis the between lab CVs varied between 22 and 46%. Since no
other intercomparison studies were known for the analysis of individual CPs, no
comparison could be made with previous intercomparison studies. For total CP a
between lab CV of 56% was found.
A larger number of laboratories participated in this international interlaboratory study
than in previous intercomparison studies on total SCCP, showing that an increasing
interest in quantitative analysis of CPs, as well as the intention to evaluate laboratory
performance so as to improve the data produced in the field. This study showed that
even though a variety of quantification techniques were used the difference between
the median value and the target value has decreased compared to a study of Tomy et
al. (1999), in which all participants used the same quantification technique.
The present study also suggests that the overall performance of participants in the
analysis of total CPs has improved compared to the interlaboratory study of Pellizzato
et al. (2009), although the present study only dealt with a solution of CPs, while the
study of Pellizzato et al. (2009) dealt with an oil extract.
Obviously, further developmental work in laboratories and more interlaboratory
comparison exercises are needed to improve the analysis of CPs. Analytical standards
of CP congeners that occur in environmental samples are badly needed.
A second step of this exercise including analysis of clean extracts will be organised
shortly.
27
Acknowledgements
The Chlorinated Paraffins Sector Group of Eurochlor is gratefully acknowledged for
their financial support of this interlaboratory study, and their technical input.
29
References
Cofino, W.P., Stokkum, I.H.M. van, Steenwijk, J. Van & Wells, D.E. (2005) A new model for
the inference of population characteristics from experimental data using uncertainties.
Part II. Application to censored datasets. Anal. Chim. Acta, 533, 31-39.
Cofino, W.P., Wells, D.E., Arise, F., Stokkum, I. van , Wegener, J.W. & Peerboom, R. (2000). A
new model for the inference of population characteristics from experimental data
using uncertainties. Application to interlaboratory studies. Chemometrics. Intell. Lab.
Syst. 53, 37-55.
Pellizzato, F., Ricci, M., Held , A., Emons, H., Bhmer, W., Geiss, S., Iozza, S., Mais, S.,
Petersen, M. & Lepom, P. (2009). Laboratory intercomparison study on the analysis of
short-chain chlorinated paraffins in an extract of industrial soil. Trends Anal. Chem.
28, 1029-1035.
Thompson, M., Ellison, S.L.R. & Wood, R. (2006). The international harmonised protocol for
proficiency testing of analytical chemistry laboratories. IUPAC Technical report.
Interdivisional working party for harmonization of quality assurance schemes. Pure
Appl. Chem. 78 , 145-196.
Thompson, M.& Wood, R. (1993). International harmonised protocol for proficiency testing
of (chemical) analytical laboratories. J. AOAC Int, 76, 926-940.
Tomy, G.T., Westmore, J.B., Stern, G.A., Muir, D.C.G. & Fisk, A.T. (1999). Interlaboratory
study on quantitative methods of analysis of C10-C13 polychloro-n-alkanes. Anal.
Chem., 71, 446-451.
Wells, D.E., Cofino, W.P. & Scurfield, J.A. (2004). The application of the Cofino model to
evaluate laboratory performance study data using bandwidth estimator. UK. Report
nr. 04/04. FRS Marine Laboratory, 375 Victoria Road, Aberdeen 388 AB11 9DB.
Wells, D.E. & Scurfield, J.A. (2004). Assessment rules for the evaluation of the QUASIMEME
laboratory performance studies data. UK, version 2. FRS Marine Laboratory, 375
Victoria Road, Aberdeen AB11 9DB, pp.390 -391.
31
Appendices
1.
2.
3.
4.
5.
6.
List of participants
Results and graphical representation
Numerical z-score values per matrix
Consistency of data
Graphical output of the Cofino Model statistics
Additional method information
32
Appendix 1
List of Participants
Appendix 2
33
Participant code:
Date Received:
Date Analyzed:
Weight Received:
Weight Analyzed:
Concentration in ng/mL
1,1,1,3,9,10-HCD
1,1,1,3,11,12-HCDD
1,1,1,3,11,13,13,13-OCTD
Total CP
Target
value
Assigned
value
129
132
96
12000
138
177
111
NA
Indicative Model
value
Mean Median
NA
NA
NA
18500
138
177
111
18180
148
197
120
18500
MIN
MAX
51
122
90
8100
810
2800
1500
49300
Model
Model
Between percentage
Lab CV%
in PMF1
n>LOQ
22
46
28
56
57
76
78
76
7
7
7
14
34
Participant code:
CPP1
1
Date Received:
Date Analyzed:
Weight Received:
Weight Analyzed:
08/08/2011
10-14/10/2011
7.0712g 7.3768g 7.1024g
7.0716g 6.8108g 7.1030g
Concentration in ng/mL
1,1,1,3,9,10-HCD
1,1,1,3,11,12-HCDD
1,1,1,3,11,13,13,13-OCTD
Total CP
NA
NA
NA
CPP1
2
NA
NA
NA
CPP1
3
NA
NA
NA
CPP2
1
CPP2
2
CPP2
3
CPP3
1
CPP3
2
CPP3
3
02/08/2011
12/08/2011
7.2569g 7.2209g 7.3395g
7.2566g 7.2214g 7.3400g
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
CPP6
1
CPP6
2
CPP6
3
CPP4
m1
1
CPP4
m1
2
CPP4
m1
3
04/08/2011
11/09/2011
7.2241g 7.2393g 7.0773g
200 l
200 l
CPP4
m2
1
CPP4
m2
2
CPP4
m2
3
04/08/2011
11/09/2011
7.2241g 7.2393g 7.0773g
200 l
200 l
138
132
139
NA
NA
NA
NA
NA
NA
135
130
139
NA
NA
NA
NA
NA
NA
91
96
94
NA
NA
NA
NA
NA
NA
11400 12200 12050 49300 48500 40300 21600 20300 17100
Participant code:
CPP5
1
CPP5
2
CPP5
3
Date Received:
Date Analyzed:
Weight Received:
Weight Analyzed:
Concentration in ng/mL
1,1,1,3,9,10-HCD
1,1,1,3,11,12-HCDD
1,1,1,3,11,13,13,13-OCTD
Total CP
ND: not detected
NA: not analyzed
CPP7
1
01/11
1.2267g
0.3597g
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
CPP7
2
CPP7
3
04/08/2011
28/10/2
27/10
011
1.4115g 1.2328g
0.5237g 0.3420g
CPP8
m1
1
CPP8
m1
2
CPP8
m1
3
05/08/2011
28/10/2011
7.2778g 7.3531g 7.1863g
2 l
2 l
2 l
51
51
60
640
810
700
220
210
210
2600 2700 2800
110
130
120
1400 1500 1300
12000 12000 12000 18000 17000 17000
CPP8
m2
1
CPP8
m2
2
CPP8
m2
3
05/08/2011
12/09/2011
7.2778g 7.3531g 7.1863g
2 l
2 l
2 l
NA
NA
NA
8700
NA
NA
NA
8100
NA
NA
NA
8200
Participant code:
CPP9
1
CPP9
2
CPP9
3
Date Received:
Date Analyzed:
Weight Received:
Weight Analyzed:
20/08/2011
20/10/2011
7.2042g 7.3329g 7.2475g
7.2042g 7.3329g 7.2475g
Concentration in ng/mL
1,1,1,3,9,10-HCD
1,1,1,3,11,12-HCDD
1,1,1,3,11,13,13,13-OCTD
Total CP
NA
NA
NA
NA
NA
NA
NA
NA
NA
31900 33500 33080
CPP10
1
CPP10
2
CPP10
3
CPP11
1
CPP11
2
CPP11
3
CPP12
1
01/08/2011
22/09/2011
1
1 ampoule
1 ampoule
1
1 ampoule ampoule
1 ampoule
ampoule
126.6
129.2
92.20
22600
123.5
131
97.02
23130
126.6
133.3
97.9
23030
NA
NA
NA
10300
NA
NA
NA
11000
35
CPP12 CPP12
3
2
14/11/2011
NA
NA
NA
11100
NA
NA
NA
18000
NA
NA
NA
26000
NA
NA
NA
28000
CPP15
m1
3
CPP15
m2
1
CPP15 CPP15
m2
m2
2
3
Participant code:
CPP13 CPP13
1
2
Date Received:
Date Analyzed:
Weight Received:
Weight Analyzed:
15/11
7.2471 g
1.3518 g
Concentration in ng/mL
1,1,1,3,9,10-HCD
1,1,1,3,11,12-HCDD
1,1,1,3,11,13,13,13-OCTD
Total CP
145.2
292.4
153.0
11942
CPP13
3
CPP14
1
CPP14
2
CPP14
3
CPP15
m1
1
09/11/2011
11/11
12/11
7.3842 g
7.1574 g
1.4382 g
1.3758 g
147.0
270.8
133.8
12399
152.95
254.2
148.3
11504
CPP15
m1
2
01/08/2011
13/10/2011
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
308
196
118
11200
308
200
124
11200
01/08/2011
13/10/2011
301
195
123
11900
153
147
101
31200
143
122
90
30200
159
139
101
31600
36
37
38
Appendix 3
Appendix 4
39
Consistency of data
40
Appendix 5
41
42
43
44
Appendix 6
45
46