CPDW project
Assessment of migration of non-suspected
compounds from products in contact with
drinking water by GC-MS
H. James1, M. Bondant2, E.J. Hoekstra3, S. Langer4, T. van Leerdam5, Th. Noij5,
E. Stottmeister6, E. Veschetti7
Sample ID: Test mix
Acquired on 26-Aug-2002 at 22: 46: 34
MD4878
Scan EI+
T IC
8.92e7
RT
Scan
44.503
2190
100
28.735
1244
%
17.968
598
8.034
2
9.317
79
33.569
1534
14.501
390
0
rt
10.000
2003
15.000
20.000
25.000
30.000
35.000
40.000
45.000
50.000
1
WRc, United Kingdom
2
IRH Environnement, France
3
European Commission, DG JRC
4
Sveriges Provnings- och Forskningsinstitut, Sweden
5
Kiwa N.V., The Netherlands
6
Umweltbundesamt, Germany
7
Istituto Superiore di Sanità, Italy
EUR 20833 EN
2
July 2003
Assessment of migration of nonsuspected
compounds
from
products in contact with drinking
water by GC-MS
Development of a harmonised test to be used in
the European Acceptance Scheme concerning
CPDW
European Commission
Contract number
EVK1-CT2000-00052
Authors
H. James (WRc), M. Bondant (LHRSP), E.J. Hoekstra (EC
DG JRC), S. Langer (SP), T. van Leerdam (Kiwa), Th. Noij
(Kiwa), E. Stottmeister (UBA), E. Veschetti (ISS)
3
4
Summary
A harmonised test method is needed for the determination of compounds amenable to
GCMS in migration waters of materials that come into contact with drinking water. Only
France and UK apply such test in their current national acceptance scheme. Since the
methods did not differ significantly and the UK method was well documented, BS 6920
Part 4 was chosen as starting point.
The participants of the project initially tested the practicality of BS 6920 Part 4 and made
several recommendations for the later stages of the project.
In the next stage the French and the UK method were compared in order to ascertain
whether there were significant differences in the results obtained for equivalent samples.
The main difference is the pH at which the migration water is extracted. Spiked blank
water samples were extracted with dichloromethane at pH 2, pH 7, and at pH 10 + 2
(initially an extraction at pH 10 followed by pH adjustment and extraction at pH 2, and
combination of both extracts. The results show that the French two-stage extraction at pH
10 and pH 2 is marginally better, particularly with respect to the recovery of basic
compounds.
For various reasons, e.g. ease of automation, reduction in solvent usage, solid phase
extraction (SPE) is now preferred to solvent extraction for the analysis of many organic
compounds in water. It was considered appropriate to establish whether this offered a
viable alternative for this particular application. Two of the more promising types of SPE
material were compared to ascertain whether either were suitable. Several compounds,
which are recovered with good efficiently using solvent extraction, were poorly recovered
(if at all) when using SPE. Solvent extraction was therefore selected as the most
appropriate extraction technique and a comparison of the preferred solvent extraction with
the preferred SPE procedure was not carried out. Further R&D work would be necessary
to produce a SPE procedure, which provides comparable recoveries to solvent extraction.
In the final stage of the project all participating laboratories analysed extracts from
migration waters from epoxy resin, polyester resin, cement with organic additive and
EPDM rubber. For each material one laboratory prepared the migration water using the
two-stage solvent extraction at pH 10 and pH 2 and using chlorinated and non-chlorinated
water. The laboratory prepared the extracts and distributed them to the other laboratories
for GCMS analysis. The results suggest that provided the level of interest is set at about 5
µg/l, different laboratories (once they had gained some experience of this type of work)
would produce comparable data with a inter-laboratory reproducibility in the range of 2050%, calculated as the coefficient of variation of the mean.
Arising from this project the following conclsions can be drawn:
• inter-laboratory exercises are required to allow laboratories to gain experience with the
GCMS method. A proficiency testing scheme is required to ensure that approved
laboratories are, and continue to be, competent.
• Care must be taken to ensure that high purity solvent is used for the solvent extraction,
and that blanks and/or migration waters are not contaminated, either during the
preparation of the migration waters or when the solvent extraction is carried out.
Specifications need to be set for acceptable blanks.
• The greatest discrepancies in the data were noted for very volatile compounds, which
are more volatile than the internal standard d6-benzene. If such compounds are of
interest than an alternative technique such as purge and trap GCMS is more
appropriate for their determination.
5
• To facilitate the use of the GCMS assessment procedure developed, a mass spectral
library of all of the compounds on the Drinking Water Positive List (DWPL) which are
amenable to GCMS analysis using the agreed protocol needs to be produced to ensure
consistency of data between laboratories. Infospec, a software package developed by
Kiwa, allows both chromatographic and mass spectral data to be included for
compounds of interest, and all of the features necessary for the purposes of the EAS
are already present. This software package could be the basis for the EAS mass
spectral library.
• As the GCMS method provides estimates of the concentrations of compounds
detected, dedicated analytical methods may still be required for some compounds
where the concentrations found give rise to concern and accurate quantification is
necessary.
• To facilitate the operation of the EAS it is recommended that the appropriate regulatory
body sets pass/fail criteria for the GCMS assessment. These criteria should include a
concentration below which any compound detected is considered insignificant.
The overall performance of the method is such that it is recommended that GCMS
analysis for unsuspected compounds should be incorporated into the overall EAS.
6
Contents
SUMMARY
5
CONTENTS
7
INTRODUCTION
9
STAGE 1 – HARMONISATION OF THE BASIC GCMS PROCEDURE
11
STAGE 2 – COMPARISON OF UK AND FRENCH SOLVENT EXTRACTION
PROCEDURES
17
STAGE 3 – SELECTION OF POSSIBLE SOLID PHASE EXTRACTION (SPE)
PROCEDURE
23
STAGE 4 – COMPARISON OF PREFERRED SOLVENT EXTRACTION PROCEDURE
AND PREFERRED SPE PROCEDURE
29
STAGE 5 – INVESTIGATION OF OVERALL PREFERRED GCMS ASSESSMENT
PROCEDURE
31
STAGE 6 – ESTABLISHMENT OF A GCMS DATABASE
55
CONCLUSIONS
57
RECOMMENDATIONS
59
APPENDIX 1 – BS6920: PART 4
61
APPENDIX 2 – PROPOSED MODIFIED VERSION OF BS6920: PART 4
83
APPENDIX 3 – DATA FROM STAGE 5
107
7
8
Introduction
Objectives
GCMS analysis is undertaken to determine whether organic chemicals not specified in the
formulation of CPDW materials migrate into drinking waters. The harmonisation of the
GCMS assessment needs to be studied to decide on the comparability of data produced
by different laboratories using different equipment, when all laboratories use the same
protocol. Initially an assessment is required of the rigidity of the protocol in order to ensure
good comparability of data, and the agreed (and if necessary, modified) protocol needs to
be applied to real migration waters from CPDW materials to allow its effectiveness to be
assessed.
Initial work programme
Only France and the UK routinely apply GCMS to detect unsuspected organic chemicals
(unspecified compounds) in migration waters from CPDW materials in their national
approval schemes. Although the two procedures are similar in many ways, there are some
important differences that need to be harmonised. One should bear in mind that the
current GCMS assessment procedures (France and UK) are a necessary compromise
between generating data on products and excessive costs. Additional advanced
techniques could be applied to address some of the limitations of the GCMS approach,
but the harmonised methodology must be practical and not limited to just a few research
laboratories.
The work programme initially proposed was as follows:
Stage 1 – Harmonisation of the basic GCMS procedure
In this stage, all participating laboratories assessed the practicality of the GCMS protocol
described in the UK BS6920: Part 4. A range of compounds was to be chosen for this
purpose.
Stage 2 – Selection of the solvent extraction procedure
As the procedures routinely used by France and the UK were slightly different, the two
procedures were compared to ascertain which was the better of the two.
Stage 3 – Selection of possible SPE procedure
As solid phase extraction (SPE) is increasingly used for the analysis of many organic
compounds in water, and for various reasons (e.g. ease of automation, reduction in
solvent usage) is now preferred to solvent extraction, it was considered appropriate to
establish whether this offered a viable alternative for this particular application. Two
different types of SPE material were compared to ascertain whether either were suitable.
Stage 4 – Comparison of preferred solvent extraction and SPE procedures
Assuming that one of the SPE procedures studied in Stage 3 was comparable to the
preferred solvent extraction procedure (from Stage 2), then it would be necessary to
directly compare the different extraction techniques to assess which was the most
appropriate for the analysis of migration waters from CPDW materials. If neither of the
SPE procedures tested in Stage 3 gave results that were comparable to the preferred
solvent extraction procedure, then this stage would not be undertaken.
Stage 5 – Investigation of the overall preferred GCMS assessment procedure
In this stage all participating laboratories were to analyse extracts from migration waters
from epoxy resin, GRP, cement with organic additive and rubber. Four participants were
to prepare the migration waters, using chlorinated and unchlorinated water, and carry out
the extractions to prepare extracts. These extracts were to be circulated to all of the
participants for GCMS analysis.
9
Stage 6 – Establishment of a GCMS database and preparation of final report
To facilitate the use of the GCMS assessment procedure developed, the most useful
design and operation of a GCMS database relating to materials in contact with drinking
water will be discussed and recommended. Its value compared to existing databases
would be assessed.
Scientific/Technical progress
BS6920: Part 4 (see Appendix 1) was taken as a starting point, as it was the only
available well-documented procedure. This standard has to be complied with by test
laboratories submitting data to the body which provides expert advice to Government
Authorities in England and Wales on approval issues (The Committee on Products and
Processes for use in Public Water Supply (CPP)). This standard deals only with the
analytical procedures, as the preparation of migration waters from CPDW materials is
considered to be a separate issue.
BS6920: Part 4 has been written so as to be non-specific, to allow freedom to laboratories
approved to undertake this type of work to choose the type of equipment used from a
supplier of their choice. For example any type of mass spectrometer may be used,
provided it meets or exceeds specific performance requirements relating to mass range,
scan speed and mass spectrometric resolution. As the main purpose of the GCMS
analysis is the identification of unsuspected compounds that leach from materials being
tested, some minimum performance standards are set for the performance of the GC
system in order to provide sufficient GC resolution to allow good quality mass spectra to
be produced. Isotopically-labelled internal standards are used to provide reference
compounds to allow the quantities of leached compounds to be estimated. The provision
of accurate quantitative data should be a separate exercise, if the general survey GCMS
shows that unsuspected compounds of particular toxicological interest are leached.
The internal standards, consisting of a mixture of isotopically-labelled compounds (so that
they can only be present in the migration water due to deliberate addition), are used to
check the initial GC column performance and to ensure continued good performance
during use. They were chosen to represent a range of volatilies (c. 80°C to 400°C) and a
range of compound types. They are spiked into the migration waters at specified
concentrations between 1 µg/l and 8 µg/l.
The sensitivity of the GCMS system is set to ensure a detectable response for the lowest
concentration internal standards, but not so high that the internal standards present at the
highest concentration overload the system (i.e. saturated mass spectra are produced).
The GC conditions are set to ensure that the most volatile internal standard (d6-benzene)
is separated from the extraction solvent (dichloromethane), and the retention time for d62squalane must be between 35 and 45 minutes. These requirements are imposed in an
attempt to ensure that inappropriate conditions cannot be used. The peak shapes
obtained on the total ion current (TIC) chromatogram for d5-phenol and d8-naphthalene
are checked to ensure that the asymmetry factor for these compounds is between 0.67
and 2.00 to provide a quality assurance check on the GC column during use.
10
Stage 1 – Harmonisation of the basic GCMS
procedure
The main purpose of Stage 1 was to provide all of the participating laboratories in WP3
with an opportunity to familiarise themselves with the suggested GCMS conditions in
BS6920: Part 4, using a range of compounds relevant to extracts from materials migration
waters.
The composition of the reference mixtures used for this stage was based on a list of some
typical compounds detected by WRc-NSF during the use of BS6920: Part 4 for the
analysis of migration waters from materials submitted to the UK regulatory body for
approval. The agreed list was as follows (compound classes are given in brackets):
Iso-propylbenzene (aromatic hydrocarbon)
Fluoranthene (PAH)
Dimethyl phthalate (phthalate)
2,6-Di-t-butyl-p-benzoquinone (quinone)
Benzoic acid (aromatic acid)
Benzyl alcohol (aromatic alcohol)
n-Butanol (aliphatic alcohol)
Acetophenone (aromatic ketone)
Methyl methacrylate (aliphatic ester)
bis-Phenol A diglycidyl ether (aromatic glycidyl ether)
bis-Phenol A (phenol)
N-methylpyrrolidinone (aromatic nitrogen-containing)
Tri-n-butylphosphate (phosphate)
Di-n-butylamine (aliphatic amine)
Hexadecane (aliphatic hydrocarbon)
Some of these compounds are fairly volatile i.e. they will elute at the beginning of the
GCMS run while others are less volatile, so the range of interestin terms of volatility is
adequately covered. However, it was recognised that some of these compounds due to
their polar nature might prove to be difficult to elute from some GC columns, particularly if
they were present at low concentrations in the reference mixtures.
From the UK experience of distributing mixtures of compounds for analysis by GCMS, it
was considered preferable to circulate two separate standard solutions – one containing
the acidic and neutral compounds, the other containing the basic compounds, to avoid
any possibility of reactions between the different compound types. These were to be
referred to as REFmix 1A and REFmix 1B. Additionally, both of these mixtures would
contain all of the isotopically-labelled internal standards.
It was agreed that the mixture of isotopically labelled standards used in BS 6920: Part 4
should be used for this stage of the work package. This mixture consists of the following
compounds:
d6-Benzene
d5-Chlorobenzene
d10-p-Xylene
d6-Phenol
d8-Naphthalene
d21-2,6-Di-t-butyl-4-methylphenol (BHT)
d34-Hexadecane
d10-Phenanthrene
d62-Squalane
11
It should be noted that for both d6-phenol and d21-BHT, the deuterium-label on the
phenolic group is immediately replaced by hydrogen in the presence of water, so these
two compounds are detected as d5-phenol and d20-BHT. This conversion takes place even
in standards prepared in high purity dichloromethane.
Most of these standards are available at relatively low-cost from various suppliers (e.g.
Sigma-Aldrich). d21-BHT and d62-squalane are less widely available, but can be obtained
from C/D/N Isotopes (www.cdniso.com) or other sources. The use of these standards
does not contribute significantly to the cost of the overall analysis.
The internal standards were chosen to cover a range of volatilities and compound types
and essentially provide internal analytical quality control. For example, loss or (resulting
in non-detection) of the more volatile standards indicates that problems were encountered
during the concentration of the extract from the migration water. Other compounds in the
extract with similar volatility would therefore have also been lost, and consequently the
validity of the analysis should be questioned.
The range of concentrations at which the internal standards are used (equivalent to 1–8
µg/l) was considered adequate for the range of interest for materials testing. Compounds
which are detected at concentrations far in excess of this range in all probability would be
required by the appropriate regulatory body to be quantified using specific methods
provided that they are of concern for toxicological reasons.
Regarding equipment specifications, which have deliberately been non-prescriptive in
BS6920: Part 4, it was agreed that it was not necessary to specify the GCMS equipment
to be used, although the experience of the participating laboratories in carrying out the
work for this stage might result in some suggestions in this respect. All modern mass
spectrometers meet the performance requirements in terms of the mass range to be
covered, the scan speed and the mass spectrometric resolution. The GC system could be
fitted with any type of injector, provided it allowed all of the internal standards to be
reliably detected during the GCMS run.
Other suggestions which were put forward as part of the harmonisation work were as
follows:
• As BS6920: Part 4 does not specify a recovery for the most volatile internal standard
(d6-benzene) the harmonised procedure should specify a minimum recovery of 50%.
This would ensure that the concentration of the solvent extract was carried out
carefully.
• The GC column specification should recommend a capillary column with a film
thickness of 0.25 µm.
• The frequency of the inclusion of the GCMS system check-standard, which is specified
in BS6920: Part 4 as being run after every six extracts (migration waters or blanks)
should only be specified as being run regularly, with the recommendation that it should
be after every six extracts.
• The way in which the retention times of the compounds detected in extracts is
recorded in the data (results) tables could be improved, so rather than using absolute
retention times a system for normalising retention times could be used. One of the
participating laboratories (Kiwa) already use such a system.
The incorporation of these suggestions into the harmonised procedure would be
considered at a later stage of the work package, after participants had gained more
experience of using BS6920: Part 4.
12
The reference mixtures circulated (solutions in dichloromethane) were as follows
Compound
REFmix 1A
ng/µl
2.00
2.00
2.00
2.00
2.00
2.01
1.87
2.08
1.97
2.20
1.96
2.00
1.91
bis-phenol A
benzoic acid
fluoranthene
2,6-di-t-butyl-1,4-benzoquinone
bis-phenol A diglycidyl ether
acetophenone
methyl methacrylate
benzyl alcohol
tri-n-butyl phosphate
diethyl phthalate
iso-propyl benzene (Cumene)
1-butanol
hexadecane
N-methylpyrrolidinone
di-n-butylamine
REFmix 1B
ng/µl
2.05
1.86
Internal standards
d62-squalane
d5-chlorobenzene
d10-phenanthrene
d5-phenol
d6-benzene
d21-2,6-di-t-butyl-4-methylphenol
d34-hexadecane
d8-naphthalene
d10-p-xylene
15.72
4.35
4.00
16.00
3.52
17.52
1.63
2.00
1.70
15.72
4.35
4.00
16.00
3.52
17.52
1.63
2.00
1.70
The results can be summarised as follows:
• Generally all of the isotopically-labelled internal standards were detected, although
some laboratories had initial difficulties with d6-benzene when using split/splitless
injection. At first sight this is surprising as this could be achieved using equipment
available over twenty years ago, but it may reflect the fact that many laboratories do
not regularly attempt to detect volatile compounds when analysing solvent extracts
using GCMS. If volatile compounds are of interest then they tend to be analysed using
static headspace or purge and trap (dynamic headspace) techniques, so GCMS
operators may need to be prepared to optimise their normal instrumental operating
parameters in order to meet the requirement to detect d6-benzene.
• Some participants commented on problems with the length of time required to cool
their GC ovens to the required start temperature (30°C). Given that all GCMS systems
are usually fitted with auto-injectors and can be run on a 24 hour per day basis, this is
not seen as a serious problem, but if necessary (e.g. in the southern countries of
Europe) it could be addressed by suggesting oven cooling options (e.g. liquid CO2 or
liquid N2; both of these options are relatively low cost) when purchasing GC
equipment, or by using the GCMS system in an air-conditioned laboratory.
• The only compound in the reference mixtures that caused serious difficulties, i.e. it
was not detected by the majority of participants was di-n-butylamine. This was not
unexpected, as special GC columns are generally required to analyse aliphatic amines
successfully, and it is not a compound that has been detected previously in CPDW
migration waters.
• Several other compounds (bis-phenol A, BADGE, N-methylpyrrolidinone, benzoic
acid) were only generally detected after concentrating the reference mixtures.
• Some participants had encountered difficulties with the specification for the mass
range to be scanned (at least 20-650 Daltons) and would prefer a mass range of 35-
13
650 Daltons (which means that the residual air (mainly nitrogen and oxygen) which is
normally present in all mass spectrometers, is not detected and does not appear in the
mass spectra recorded).
The following conclusions could be drawn at this stage of the overall work package:
• The mixture of isotopically-labelled compounds used both as a GC column
performance check and as internal standards for the analysis of migration waters is
satisfactory.
• It does not appear to be necessary to specify a particular GC column or the GC
injection volume, although some optimisation of the GC operating conditions may be
necessary to ensure that d6-benzene is satisfactorily resolved from the solvent peak.
• The restrictions regarding GC column length (minimum 50 metres) and GC
temperature programming rate (not greater than 12°c/min) do not require amendment.
• The concerns regarding the mass range to be scanned could be addressed by
specifying a minimum mass range of 35-650 Daltons, which would allow laboratories
to use a wider mass range (e.g. 20-700 Daltons) if thought appropriate.
• Some compounds (e.g. benzoic acid, BADGE) which are known to leach from
materials were not detected by all participants. This suggests that if regulators suspect
that polar compounds ( acids or bases) are likely to leach from some materials, and
would be concerned by this, specific analyses to provide good quantitative information
on these types of compounds would need to be requested.
• Given that the concentrations of the various compounds in the reference mixtures
were at the lower end of the range of interest, the overall results were encouraging.
(Most of the compounds were present in the reference mixtures at concentrations of
about 2 ng/l – this would correspond to a compound present in a migration water at a
concentration of 2 µg/l, assuming an extraction efficiency of 50% and a final extract
volume of 500 µl).
A typical total ion current (TIC) chromatogram for the reference mixture Ref 1A is shown
in Figure 1.1 below.
14
Figure 1.1 Typical TIC chromatogram of reference mixture Ref 1A
15
16
Stage 2 – Comparison of UK and French solvent
extraction procedures
The UK method uses dichloromethane (DCM) extraction after adjusting the pH of the
migration water to pH2. Following concentration the extract is run on GCMS. The French
method also uses DCM extraction, but includes an additional extraction at pH10, with the
extract from this additional extraction being combined with the extract from the pH2
extraction before concentration, so that only one GCMS run is required.
The two laboratories taking part in Stage 2 were WRc and LHRSP.
Reference mixtures (REFmix 2A and REFmix 2B; prepared and circulated by JRC)
containing the same compounds as those in the mixtures used for Stage 1 were used to
test the performance of the two methods. The concentrations of the compounds in these
two mixtures were as follows:
Compound
REFmix 2A
ng/µl
REFmix 2B
ng/µl
10.00
10.00
10.00
10.00
10.00
10.07
9.36
10.42
9.83
11.01
8.52
9.61
9.91
bis-phenol A
benzoic acid
fluoranthene
2,6-di-t-butyl-1,4-benzoquinone
bis-phenol A diglycidyl ether
acetophenone
methyl methacrylate
benzyl alcohol
tri-n-butyl phosphate
diethyl phthalate
iso-propyl benzene (Cumene)
1-butanol
hexadecane
N-methylpyrrolidinone
di-n-butylamine
10.24
9.69
WRc analysed six extracts, as follows:
•
blank water spiked with both Ref2A and Ref2B mixtures (150 µl of each) and the
internal standards mixture, extracted with DCM at pH2;
•
blank water spiked with internal standards only, extracted with DCM at pH2;
•
blank water spiked with both Ref2A and Ref2B mixtures (150 µl of each) and internal
standards mixture, extracted with DCM at pH7;
•
blank water spiked with internal standards only, extracted with DCM at pH7;
•
blank water spiked with both Ref2A and Ref2B mixtures (150 µl of each) and internal
standards mixture, extracted with DCM firstly at pH10 and then at pH2; both extracts
were combined before drying, concentration and GCMS analysis;
•
blank water spiked with internal standards only, extracted with DCM firstly at pH10
and then at pH2; both extracts were combined before drying, concentration and
GCMS analysis.
17
The blank water used was groundwater, known to contain minimal concentrations of
organic compounds detectable by GCMS, obtained from a borehole within the laboratory.
The results obtained are summarised in Table WP3.1 below.
For several compounds (isopropylbenzene, benzyl alcohol , acetophenone, 2,6-di-t-butylp-benzoquinone, diethyl phthalate, hexadecane, tri-n-butyl phosphate, fluoranthene,
bisphenol A), the recoveries were similar (within +/- 10%) at all of the pH values.
Generally the recovery efficiencies are considered acceptable, although the recovery of
fluoranthene (32%-35%) was surprisingly low. The recovery of BADGE (4%-13%) was
generally poor, N-methylpyrrolidone was only detected in the combined pH (2+10)
extraction, and was poorly recovered, and benzoic acid was not detected. Not surprisingly
(as it could not be detected in the GCMS runs on the Ref1A GCMS standard solution), din-butylamine was not detected.
From experience, it is not surprising that the more polar compounds were not detected
when they are present at relatively low concentrations (1-2 µg/l). Comparison of data from
the general survey GCMS method with that from specific methods for these types of
compounds suggests that they may not be detected by the former method until their
concentrations are in the range 5-10 µg/l.
BADGE was detected in the GCMS runs on the extracts, although it was not detected in
the GCMS runs of the standard solutions analysed in Stage 1. This seems to be due to
the fact that a J&W DB5-ms column was used when the standard solutions were run, but
a J&W DB-1 column was used when the extracts were run.
18
TableWP3.1 Results from comparison of solvent extraction procedures (WRc)
Compound name
A
Methyl methacrylate
iso-Propylbenzene
N-Methylpyrrolidinone
Benzyl alcohol
Acetophenone
2,6-Di-t-butyl-p-benzoquinone
Diethyl phthalate
Hexadecane
Tri-n-butyl phosphate
Fluoranthene
bis-Phenol A
BADGE
Di-n-butylamine
n-Butanol
Benzoic acid
1.30
1.47
0.00
0.65
1.23
0.77
1.41
1.73
1.37
1.39
0.40
0.06
nd*
nq**
nd*
Calculated concentration (micrograms/litre)
Actual
(uncorrected)
concentration
(µg/l)
pH2
pH7
pH2+10
B
A
B
A
B
Mean
Mean
Mean
%Recovery
(corrected)
pH2
pH7 pH2+10
1.29
1.51
0.00
0.63
1.23
0.78
1.44
1.79
1.51
1.43
0.54
0.07
nd*
nq**
nd*
72
78
0
41
63
65
108
77
98
33
31
4
nd*
nq**
nd*
nd* - not detected
nq** - detected, but not quantified
due to co-eluitng interference
19
1.30
1.49
0.00
0.64
1.23
0.77
1.43
1.76
1.44
1.41
0.47
0.06
nd*
nq**
nd*
1.85
1.44
0.00
0.68
1.20
0.70
1.33
1.62
1.22
1.37
0.47
0.22
nd*
nq**
nd*
1.39
1.44
0.00
0.71
1.27
0.73
1.30
1.69
1.30
1.39
0.48
0.16
nd*
nq**
nd*
1.62
1.44
0.00
0.69
1.23
0.72
1.27
1.65
1.26
1.38
0.47
0.19
nd*
nq**
nd*
2.06
1.47
0.16
0.92
1.21
0.73
1.34
1.71
1.39
1.47
0.51
0.12
nd*
nq**
nd*
1.38
1.55
0.03
0.71
1.12
0.75
1.42
1.84
1.51
1.54
0.49
0.17
nd*
nq**
nd*
1.72
1.51
0.09
0.82
1.16
0.74
1.38
1.77
1.45
1.51
0.50
0.14
nd*
nq**
nd*
1.40
1,28
1.54
1.56
1.51
1.50
1.65
1.49
1.47
1.50
1.50
1.50
1.45
1.44
1.50
90
76
0
44
64
60
96
72
86
32
32
13
nd*
nq**
nd*
95
79
6
52
60
61
105
77
99
35
33
10
nd*
nq**
nd*
The results were from LHRSP were generally comparable with those obtained by WRc.
Methyl methacrylate, iso-propylbenzene, benzyl alcohol, acetophenone, 2,6-di-t-butyl-pbenzoquinone, diethyl phthalate, hexadecane, tri-n-butyl phosphate and fluoranthene
were all recovered at all of the pH values used, but the extractions carried out at pH 2+10
generally resulted in better recoveries. Di-n-butylamine, N-methylpyrrolidinone, BADGE,
benzoic acid and n-butanol were not detected.
These data are summarised in Table WP3.2 and WP3.3 below.
Table WP3.2 Results (raw data) from comparison of solvent extraction procedures
(LHRSP)
Compound
Area pH2
Area pH7
Area pH(2+10)
pH2 A pH2 B Mean pH7 A pH7 B Mean pH
pH
Mean
2+10A 2+10B
d6-Benzene
Methyl methacrylate
6710
964
3059
2224
4885
1594
5017
1705
6676
2527
5847
2116
5794
1886
6181
2040
5988
1963
d5-Chlorobenzene
d10-p-Xylene
iso-Propylbenzene
di-n-butylamine
610
305
522
nd
246
123
200
nd
428
214
361
590
291
511
nd
778
443
721
nd
684
367
616
580
315
502
nd
695
346
538
nd
638
331
520
d5-Phenol
N-Methylpyrrolidinone
Benzyl alcohol
Acetophenone
Benzoic acid
1377
nd
423
571
nd
845
nd
358
297
nd
1111
526
nd
388
405
nd
618
nd
580
520
nd
572
611
nd
634
501
nd
595
nd
674
558
nd
603
d8-Naphthalene
2,6-Di-t-butyl-p-benzoquinone
948
477
450
239
699
358
1978
402
1456
596
1717
499
912
498
1026
414
969
456
d21-BHT
d34-Hexadecane
Diethyl phthalate
Hexadecane
Tri-n-butyl phosphate
7106
1025
1703
1532
1178
3920
540
1610
794
834
5513
783
1657
1163
1006
6713
717
1268
993
761
7972
1049
1091
858
1062
7343
883
1180
926
912
5741
809
1523
1189
710
6252
890
1496
1275
823
5997
850
1510
1232
767
d10-Phenanthrene
Fluoranthene
bis-Phenol A
999
982
1366
602
515
1113
801
768
960
864
696
787
749
659
954
807
580
627
1240
256
487
372
279
355
26239 16780 21510 21778 28930 25354 20604 21626
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
742
604
317
21115
d62-Squalane
BADGE
n-butanol
391
434
20
484
463
654
530
Table WP3.3 Results from comparison of solvent extraction procedures (LHRSP)
Compound name
Methyl methacrylate
Iso-Propylbenzene
N-Methylpyrrolidinone
Benzyl alcohol
Acetophenone
2,6-Di-t-butyl-pbenzoquinone
Diethyl phthalate
Hexadecane
Tri-n-butyl phosphate
Fluoranthene
Bis-Phenol A
BADGE
Di-n-butylamine
n-Butanol
Benzoic acid
Calculated concentration (micrograms/litre)
A
PH2
B
Mean
A
B
Mean
A
1.58
1.71
0.00
0.45
0.60
0.47
9.04
1.63
0.00
0.80
0.66
0.44
5.31
1.67
0.00
0.62
0.63
0.45
2.89
1.76
0.00
0.20
0.20
0.56
3.25
1.63
0.00
0.40
0.36
0.57
3.07
1.69
0.00
0.30
0.28
0.56
3.25
1.59
0.16
0.70
0.55
0.62
2.94
1.55
0.03
0.66
0.54
0.47
3.09
1.57
0.10
0.68
0.55
0.54
1.66
1.49
1.15
1.97
0.40
nd*
nd*
nd*
nd*
2.98
1.47
1.54
1.71
0.54
nd*
nd*
nd*
nd*
2.32
1.48
1.35
1.84
0.47
nd*
nd*
nd*
nd*
1.77
1.38
1.06
1.72
0.47
nd*
nd*
nd*
nd*
1.04
0.82
1.01
1.99
0.48
nd*
nd*
nd*
nd*
1.40
1.10
1.04
1.85
0.47
nd*
nd*
nd*
nd*
1.88
1.47
0.88
1.67
0.51
nd*
nd*
nd*
nd*
1.68
1.43
0.92
1.59
0.49
nd*
nd*
nd*
nd*
1.78
1.45
0.90
1.63
0.50
nd*
nd*
nd*
nd*
nd* - not detected
21
pH7
Actual
concentration
(µg/l)
pH2+10
B
Mean
%Recovery
pH2
pH7
pH2+10
1.40
1.28
1.54
1.56
1.51
1.50
379
130
0
40
42
30
219
132
0
19
19
38
221
123
6
43
36
36
1.65
1.49
1.47
1.50
1.50
1.50
1.45
1.44
1.50
141
99
92
123
31
nd*
nd*
nd*
nd*
85
74
71
123
32
nd*
nd*
nd*
nd*
108
97
61
109
33
nd*
nd*
nd*
nd*
Stage 3 – Selection of possible solid phase
extraction (SPE) procedure
Stage 3 involved a comparison of different SPE procedures.
SPE involves passing a water sample (generally 10 – 1000 ml) through a column or
cartridge containing a small amount (0.1 – 5 g) of a solid sorbent, onto which compounds
of interest are adsorbed. They are then eluted with a small volume of a suitable solvent.
This extract may need concentration prior to analysis, but usually the concentration factor
is less than 10 ×.
Since SPE was generally accepted in the mid 1980’s, it has become the preferred method
of extraction for many analyses. The advantages over, for example, liquid-liquid extraction
(LLE) include the following:
•
SPE is less labour-intensive and can be automated.
•
SPE can be used on-line in an automated analysis (particularly when the
detection/quantification is by LCMS; it is not as widely used when
detection/quantification is by GCMS). In this case sample volumes can be reduced to
as little as 10 ml.
•
Less organic solvent is required to elute compounds from SPE cartridges than to carry
out LLE; in some cases no concentration of extracts is required prior to analysis.
However SPE is used mainly for target compound analysis and, with the exception of
certain styrene-divinylbenzene resins (e.g. Amberlite XAD-2), little information is available
on the utility of the vast majority of the available SPE adsorbents for general survey
analyses.
As there is no generally agreed “best” SPE cartridge for the type of analysis required, two
laboratories each investigated a different type of SPE cartridge. It was agreed that Kiwa
would investigate a styrene-divinylbenzene type cartridge, and that SP would check the
performance of a C18 cartridge. In both cases, many alternative cartridges are available
from various manufacturers, and the decision as to which should be investigated was left
to the laboratory carrying out the work.
Kiwa used LiChrolut EN (an ethylvinylbenzene-divinylbenzene copolymer) SPE cartridges
(200 mg). They were conditioned with DCM (5 ml), methanol (5 ml) and MilliQ water (5 ml;
pH adjusted to pH2 or pH10 depending on the pH of the sample to be extracted –
separate extractions were carried out at each of these pH values). After appropriate pH
adjustment, addition of the Ref 2A and Ref 2B mixtures and the internal standards,
samples (500 ml) was passed through the cartridges at ∼10 ml/minute. The spiking level
for the Ref2A and Ref2B mixtures was 1 - 2 µg/l. It was noted that the colour of the
LiChrolut EN changed to a dark brown after passage of the samples which had been pH
adjusted to pH2. After the sample had passed through the cartridge, the cartridge was
dried for 2 hours using a moderate flow of nitrogen, and then eluted with DCM (5 x 1 ml).
The DCM eluates were combined and concentrated using nitrogen blow down to 500 µl.
The concentrated extracts were then run on GCMS.The results obtained by Kiwa are
shown in Table WP3.4 below.
The basic compounds (N-methylpyrrolidone and di-n-butylamine) were not detected in any
of the concentrated extracts. In the case of di-n-butylamine this was not surprising as it
could not be detected in standard solutions. Butanol, methyl methacrylate and the internal
standard d6-benzene were not detected because they eluted too close to the solvent peak
in the GCMS run, although some optimisation of the operating parameters of the GC
system could have resolved this problem. The recoveries for the remaining compounds
23
were generally in the range 60-110%, although several for several of the hydrocarbons
(hexadecane, d34-hexadecane, d10-phenanthrene and fluoranthene) recoveries were lower
(6-62%). This is thought to be due to poor recovery during the elution step of the SPE
procedure.
Given that the BS6920: Part 4 protocol was the basis for the harmonisation exercise, the
fact that three internal standards (d6-benzene, d34—hexadecane and d62-squalane) could
not be detected in the extract from the pH 10 extraction indicates that these extraction
conditions are not satisfactory. The results from the extractions at pH2 were more
promising, and for some of the compounds in the reference mixtures the recoveries were
comparable to those obtained when using solvent extraction.
24
TableWP3.4 Results from SPE extraction using styrene-divinylbenzene cartridges (Kiwa)
ret. time compound name
Recoveries
(min:sec)
6.68
7.06
8.5
10.14
12.05
13.1
16.79
24.61
25.24
27.17
27.82
28.06
29.15
33.14
39.37
41.2
47.54
1-butanol
methyl methacrylate
Chlorobenzene-d5
para-xylene-d10
iso-propyl benzene
Phenol-d5
benzyl alcohol
acetophenone
naphthalene-d8
2,6-di-t-butyl-p-benzoquinone
2,6-Di-(tert-butyl)-4-methylphenol-d21
Hexadecane-d34
hexadecane
diethyl phthalate
tr-n-butyl phosphate
Phenantrene-d10
fluorantheen
bis-phenol A
2,6,10,15,19,23-Hexamethyltetracosane-d62
25
blank pH2
pH 2 + add avg
n=2
blank pH10
pH 10 + ADD 1
n=1
73%
77%
61%
64%
61%
95%
67%
88%
74%
90%
42%
52%
29%
93%
75%
30%
6%
114%
114%
57%
107%
59%
61%
60%
86%
87%
84%
74%
84%
37%
111%
86%
41%
34%
31%
84%
94%
67%
33%
27%
18%
85%
59%
22%
7%
SP initially used IST Isolute C18 cartridges with a sample volume of 500 ml. The cartridges
were conditioned with methanol (6 ml) and water prior to use. Following passage of the
sample, the cartridges were extracted with methanol. Unfortunately, nothing was detected
when the extracts were analysed by GCMS, probably because the extracts were not
concentrated sufficiently prior to analysis and due to the fact that methanol is not a
particularly good solvent for gas chromatography with capillary columns. Following
discussions, SP agreed to repeat their SPE work and concentrate the extract before the
GCMS analysis.
The solid phase extraction was then performed using 3M Empore High Perfomance
Extraction Disks, C18 (Octadecyl), diameter 47 mm. Test water samples were prepared by
adding 200µl of the standard solutions Ref 2A and 2B and the deuterated internal
standards (100 µl) to 1 liter water (MilliQ). The pH of the samples was adjusted prior to
addition of the organic compounds to pH 2 and 10, respectively. The resulting
concentrations of the test compounds were between 1 - 2 µg/liter and the concentrations
of the internal standards were between 1 - 8 µg/liter.
The extraction was performed following the procedure described in the manufacturer’s
instructions. The disks were activated by washing with 10 ml of DCM followed by 10 ml of
methanol. The samples were pumped through the discs at 100 ml per minute and then
dried by applying vacuum (about 15 minutes). The organic compounds were eluted from
the extraction disks using 3 portions of 5 ml of DCM, which were than combined, and the
extract was concentrated to 1 ml using a gentle stream of air. The extracts were then
analysed by GC/MS.
As can be seen from Table WP3.5 below not all of the internal standards were detected in
the GCMS chromatogram; d10-phenanthrene, d5-phenol and d6-benzene were not detected
at all. Benzoic acid, bis-phenol A, diglycidyl ether (BADGE) , methyl methacrylate, benzyl
alcohol, 1-butanol, N-methylpyrrolidinone and di-n-butyl amine were not detected in the
extracts. The other compounds were recovered with yields between 13% and 150%.
The results from the SPE investigations were not entirely unexpected as one disadvantage
of SPE with respect to GCMS is that the solvent used to elute the adsorbed compounds
from the cartridge needs to be sufficiently polar to do so efficiently e.g. ethyl acetate,
methanol or acetonitrile. These solvents have boiling points of at least 65°C. The best
solvents for GC purposes are generally less polar solvents such as pentane,
dichloromethane or diethyl ether, which also have lower boiling points, so that it is possible
to concentrate these solvents without loosing volatile compounds extracted from the
sample.
26
Table WP3.5 Results from SPE (C18 discs) extraction (SP)
corrected for recovery
of the internal standards
MS area
Concentration
detected
ret. time Compound
minutes
µg/Liter
41,43 bis-phenol A
0,38
benzoic acid
0
40,84 fluoranthene
1,56
29,81 2,6-di-t-butyl-1,4-benzoquinone
0,49
bis-phenol A diglycidyl ether
0
21,72 acetophenone
1,52
11,05 methyl methacrylate
1,24
20,79 benzyl alcohol
1,47
32,29 tri-n-butyl phosphate
0,94
31,91 diethyl phthalate
1,37
17,78 isopropyl benzene
1,83
9,49 1-butanol
1,34
31,47 hexadecane H34
1,57
20,89 N-methylpyrrolidinone
1,31
18,39 di-n-butyl amine
1,06
27
pH = 2
1853018
8345626
1348259
2533800
pH = 10
pH = 2
0,56
1273522
2,35
0,5
1784585
1,4
2041125
5459949
3951742
4007119
µg/Liter
pH = 10
recoveries
recoveries
%
pH = 2
147
%
pH = 10
0,35
150
101
71
0,73
92
48
5919342
0,56
0,8
2,19
2,43
60
58
120
133
966560
1,49
0,21
95
13
Stage 4 – Comparison of preferred solvent
extraction procedure and preferred
SPE procedure
The original intention of Stage 4 was to decide whether solid phase extraction or solvent
extraction should be used in the harmonised procedure, and if necessary check the
performance of SPE in a direct comparison with solvent extraction when analysing
extracts from real migration waters.
On the basis of the two sets of data produced by the laboratories who undertook the SPE
work, it appeared that the Lichrolut EN cartridges provide the best option for SPE for
typical compounds found in materials migration waters, although the recoveries of
hydrocarbons were poor and N-methylpyrrolidinone and BADGE were not detected in the
extracts at either of the pH values used. Recoveries of many of the compounds in the
reference mixtures were higher when the pH of the sample was pH 2 than when the pH
was pH 10, but the general concensus was that SPE did not provide an improvement over
solvent extraction. A comparison of the data obtained from Stages 2 and 3 is shown in
Table WP3.6.
It was therefore decided that there was no point in carrying out a comparison of SPE with
solvent extraction on real migration waters.
As noted earlier, almost all of the information available on SPE relates to its use for target
compound analysis, where conditions are optimised for one or a few compounds with
similar chemical properties. For this type of analysis rinses, it can be incorporated into the
overall procedure to remove interferences prior to elution of the compounds of interest
and the elution conditions can be tailored to remove the compounds of interest while
leaving other interferences on the cartridge. For the type of general survey analysis
required to detect unsuspected compounds in CPDW migration waters, none of the
compounds present can be considered to be interferences, as all are of potential interest.
Therefore all GCMS-amenable compounds should be retained efficiently in the adsorption
stage by the SPE cartridge used, and all should be effectively removed in the elution step.
At present it is clear that considerable work is required to decide which of the many SPE
cartridges currently available from a variety of manufacturers would best meet these
requirements.
It was considered impractical to attempt to undertake this within the scope of the current
project.
However it was considered that it should be recommended that if further work was to be
carried out at a later date a more thorough investigation of SPE should be undertaken as
the advantages of this technique are significant in terms of time and effort, and therefore
cost.
29
Table WP3.6 Comparison of SPE and solvent extraction data
Compound name
d6-Benzene
n-Butanol
Methyl methacrylate
d5-Chlorobenzene
d10-p-Xylene
Isopropylbenzene
Di-n-butylamine
d5-Phenol
N-methylpyrrolidinone
Acetophenone
Benzyl alcohol
d8-Naphthalene
Benzoic acid
2,6-Di-t-butyl-p-benzoquinone
d21-BHT
Diethyl phthalate
d34-Hexadecane
Hexadecane
Tri-n-butylphosphate
d10-Phenanthrene
Fluoranthene
bis-Phenol A
d62-Squalane
BADGE
KIWA SPE (Lichrolut EN)
data
pH2
pH10
nd
nd
nd
61%
64%
61%
nd
95%
det.
88%
67%
74%
det.
90%
42%
93%
52%
29%
75%
30%
6%
114%
114%
nd
nd
nd
nd
59%
61%
60%
nd
86%
det.
84%
87%
74%
det.
84%
37%
85%
nd
18%
59%
22%
7%
nd
nd
nd
nd = not detected
det = detected but no recovery determined
30
SP SPE (Empore C18 discs)
data
pH2
pH10
nd
nd
nd
nd
36%
120%
nd
nd
nd
92%
nd
42%
nd
101%
43%
58%
35%
95%
60%
51%
150%
147%
33%
nd
Nd
Nd
Nd
Nd
Nd
133%
Nd
Nd
Nd
48%
Nd
31%
Nd
71%
53%
Nd
15%
13%
Nd
58%
Nd
Nd
57%
Nd
LHRSP Solvent Ext. data
WRc Solvent Ext. data
pH2
pH7
pH(2+10)
pH2
pH7
pH(2+10)
det
nd
379%
det
det
130%
nd
det
nd
42%
40%
det
nd
30%
det
141%
det
99%
92%
det
123%
31%
det
nd
det
nd
219%
det
det
132%
nd
det
nd
19%
19%
det
nd
38%
det
85%
det
74%
71%
det
123%
32%
det
nd
det
nd
221%
det
det
123%
nd
det
6%
36%
43%
det
nd
36%
det
108%
det
97%
61%
det
109%
33%
det
nd
20%
det
72%
54%
59%
78%
nd
33%
nd
63%
41%
84%
nd
65%
52%
108%
80%
77%
98%
65%
33%
31%
75%
4%
34%
det
90%
93%
83%
76%
nd
15%
nd
64%
44%
117%
nd
60%
63%
96%
82%
72%
86%
72%
32%
32%
68%
13%
16%
det
95%
59%
63%
79%
nd
30%
6%
60%
52%
97%
nd
61%
58%
105%
75%
77%
99%
70%
35%
33%
72%
10%
Stage 5 – Investigation of overall preferred
GCMS assessment procedure
The purpose of this stage was to assess the comparability of data produced by different
laboratories using the same protocol to analyse identical extracts produced from real
migration waters.
Four laboratories prepared migration waters from different types of materials for
distribution to all participants. It was considered essential to carry out this stage of the
work in this way to ensure that all participants were analysing the same extracts. The
migration waters (chlorinated and unchlorinated) were prepared from:
Factory applied epoxy resin coated onto ductile iron pipe (ID 25 cm) (WRc)
Cement blocks prepared with an organic additive (Kiwa)
EPDM rubber hoses (ID 1.25 cm) (UBA)
Polyester resin (as used in GRP pipes) (LHRSP)
Extracts were to be prepared from the migration waters (chlorinated and unchlorinated)
and from the test water (chlorinated and unchlorinated) used for the leaching tests. It was
proposed that the procedure described in prEN 12873-1 should be followed, with the third
migration water (72 hour stagnation period, following two earlier 72 hour stagnation
periods) being used for the GCMS analysis.
Originally it was intended to add the compounds present in the reference mixtures Ref 2A
and Ref 2B to all of the migration waters prior to extraction, but following discussions
within WP3 it was decided that these would only be added to the migration waters from
the cement with organic additive and the epoxy resin materials as it was considered that
would already be sufficient compounds to detect in the migration waters from the rubber
and polyester resin materials.
It was necessary to prepare sufficient migration water (at least 7 litres) to produce extracts
for all seven participants, and it was suggested that those laboratories preparing the
migration waters should extract the migration waters and blanks as 1 litre aliquots using
the agreed solvent extraction procedure (i.e. following pH adjustment, extractions using
DCM at pH2 and pH10, with the extracts combined prior to concentration). During the
concentration of the separate solvent extracts from each of the leaching experiments, the
extracts were combined (at an appropriate point) and concentrated to the correct volume
(e.g. if 7 litres of a migration water had been extracted the final volume of the combined
concentrated extract would have been 3.5 ml; if 10 litres of migration water had been
extracted the final volume would be 5.0 ml). Aliquots (500 µl) of the concentrated extracts
were then places in suitable containers (CERTAN capillary bottles) for distribution. Each
participating laboratory received 14 extracts to analyse, rather than the 16 originally
agreed, as insufficient extracts were prepared from the test water and chlorinated test
water used to produce the migration waters for the cement with organic additive although
data was provided to all participants on the compounds detected in the test water and
chlorinated test water from analyses carried out by Kiwa.
This stage of the work generated considerable data (which is given in Appendix 3), as with
the exception of the migration waters from the factory applied epoxy resin material, large
numbers of compounds were detected in all of the extracts by all of the participants.
Correlating this data has proved to be an onerous task, particularly as a high proportion of
the compounds detected could not be identified, or were only tentatively identified by a
few (in some cases, one) of the participants.
31
Table WP3.7. Relative retention times (normalised to d8-naphthalene) for compounds detected by all participants in extract from unchlorinated
migration water from epoxy resin material
CAS No.
Compound
Relative retention times
1076-43-3
80-62-6
3114-55-4
41051-88-1
98-82-8
4165-62-2
98-86-2
1146-65-2
719-22-2
N/A
d6 Benzene
Methyl methacrylate
d5 Chlorobenzene
d10 p-Xylene
Isopropyl benzene
d5 Phenol
Acetophenone
d8 Naphthalene
2,6-Di-t-butyl-p-benzoquinone
d20 2,6-Di-t-butyl-4-methyl phenol
WRc
0.21
0.28
0.50
0.54
0.64
0.71
0.83
1.00
1.34
1.37
SP
0.30
0.37
0.58
0.61
0.69
0.74
0.86
1.00
1.23
1.24
Kiwa
0.16
0.28?
0.53
0.56
0.66
0.73
0.85
1.00
1.30?
1.32
UBA
0.30
0.35
0.54
0.57
0.65
0.72
0.85
1.00
1.29
1.32
JRC
0.22
0.29
0.51
0.55
0.64
0.71
0.83
1.00
1.29
1.32
ISS
0.29
0.40
0.63
0.65
0.72
0.75
0.90
1.00
1.20
1.19
LHRSP
0.17
0.23
0.43
0.46
0.56
0.66
0.80
1.00
1.38
1.42
Mean
0.227
0.319
0.515
0.548
0.641
0.712
0.846
1.000
1.308
1.329
Stdev
0.0595
0.0649
0.0509
0.0491
0.0440
0.0274
0.0301
0.0000
0.0578
0.0591
CV(%)
26
20
10
9
7
4
4
0
4
4
84-66-2
15716-08-2
Diethyl phthalate
d34 Hexadecane
1.44
1.46
1.32
1.28
1.41
1.38
1.41
1.39
1.38
1.36
1.28
1.22
1.53
1.51
1.416
1.396
0.0707
0.0809
5
6
544-76-3
126-73-8
1517-22-2
Hexadecane
Tri-n-butyl phosphate
d10 Phenanthrene
1.48
1.50
1.63
1.30
1.34
1.52
1.40
1.45
1.59
1.41
1.45
1.59
1.40
1.42
1.59
1.23
1.29
1.62
1.54
1.59
1.79
1.423
1.434
1.618
0.0821
0.0997
0.0894
6
7
6
2096-44-0
N/A
Fluoranthene
d62 Squalane
1.87
2.26
1.73
2.04
1.81
2.15
1.82
2.12
1.82
2.21
1.62
1.74
2.08
2.49
1.821
2.212
0.1401
0.1554
8
7
32
The most significant problem which has become apparent is that the use of different GC
columns and different GC programming rates may make the comparison of data from
different laboratories difficult, unless the compounds detected have been positively
identified (i.e. there is no doubt regarding their identity). It was suggested at a WP3
meeting that the use of either relative retention times (corrected for the GCMS system
“dead volume”) normalised to one of the internal standards (d8-naphthalene was chosen
as its retention time was approximately midway between the most and least volatile
internal standards) or elution temperatures would make this task easier. However this did
not prove to be the case.
Table WP3.7 shows the results obtained for compounds which all participants detected in
the extract from the non-chlorinated migration water from the epoxy resin material. It
should be noted that the mixture of reference compounds was added to this migration
water to ensure that there were compounds present which would be detected. The relative
retention times observed by the participating labpratories are given, and the means,
standard deviations and coefficients of variation (CVs: relative standard deviations
expressed as percentages) are given. As can be seen the CVs for the more volatile
compounds detected are relatively high (26% for d6-benzene), and generally they are
between 4 – 8%. While this may not initially appear to be a significant variation, if the data
are examined more closely it can be seen that, for example, the relative retention time
found by ISS for tri-n-butyl phosphate (1.29) is the same as that found by UBA for 2,6-di-tbutyl-p-benzoquinone. This is not a problem in this instance as the mass spectrum of
each of these compounds is quite distinctive, and there is therefore no doubt as to the
identity of the compounds concerned. However if they had not been identified it would be
difficult to decide whether these were two compounds which happened to have similar
mass spectra or whether they were the same compound, which due to the different GC
conditions, eluted at different relative retention times on the CGMS systems in the
different laboratories.
A similar problem was encountered when elution temperature was investigated as a
potentially useful way of identifying which of the unidentified compounds detected by
different laboratories were the same. The results of this exercise are shown in Table
WP3.8
Table WP3.8 Elution temperatures reported for some compounds detected
in extract from non-chlorinated migration water from epoxy resin
Compound
d6 Benzene
Methyl methacrylate
n-Heptane
Toluene
Diacetone alcohol
d5 Chlorobenzene
d10 p-Xylene
Isopropyl benzene
d5 Phenol
2-Ethylhexanol
d8 Naphthalene
WRc
45.47
54.67
55.33
66.67
80.13
83.07
88.93
101.73
111.20
122.13
148.40
Elution temperatures
SP
Kiwa
JRC
75.40
91.04
88.40
105.00
61.70
78.70
125.60
129.40
142.60
152.20
199.10
ISS
52.00
114.60
141.41
134.96
140.63
154.80
164.29
111.00
115.00
208.50
181.00
134.00
The GC columns and operating conditions used by the laboratories whose data has been
included in Table WP3.8 above were as follows:
33
•
•
•
•
•
WRc – 60 m DB-1, 0.32 mm ID, phase thickness 0.25µm; initial temperature 30°C
held for 4 min, then linearly programmed at 8°C/min to 300°C and held for 20 min; on
column injection at 30°C; injection volume 1 µl.
Kiwa – 25 m Rtx-5 Amine, 0.32 mm ID, phase thickness 1.00 µm; initial temperature
32°C held for 5 min, then linearly programmed at 7°C/min to 290°C held for 7 min;
PTV injector with on-column interface liner; initial temperature 35°C ramped to 300°C
at 8°C/sec, splitless time 2 min 30 sec; injection volume 1 µl.
SP – 50 m BPX-5, 0.32 mm ID, phase thickness 1.00 µm; initial temperature 50°C
held for 4 min then linearly programmed at 8°C/min to 300°C.
JRC – 60 m CP-Sil 5 CB Low Bleed/MS, 0.25 mm ID, phase thickness 0.25 µm; initial
temperature 45°C held for 4 min, then linearly programmed at 8.5°C/min to 300°C held
for 20 min; PTV injector with on-column interface liner; initial temperature 45°C (for
0.02 min) ramped at 12°C/sec to 300°C (for 4 min), splitless time 1 min
ISS – 50m SGE HT8, 0.22 mm ID, phase thickness 0.25 µm; 35°C for 10 min, then
linearly programmed at 8°C/min to 300°C; split/splitless injector at 300°C, splitless
time 0.50 min.
As can be seen from the data in Table WP3.8, the retention temperature for d6-benzene
varies between 45.47°C and 91.04°C, while for d8-naphthalene the range is 141.40 –
208.50°C, so again this parameter is not useful in deciding which unidentified compounds
detected by different laboratories are in fact the same compound.
Thus although there was a conclusion following the completion of Stage 1 of the work viz.
there is no need to specify the type of GC column that should be used, there may be a
need to reconsider this in the light of experience with real migration waters.
Obviously when GCMS is being used to analyse extracts, the main parameter which is
taken into account deciding whether two unidentified compounds detected when using two
different GCMS systems are one and the same is the mass spectrum obtained on each of
the systems. It is therefore essential that full mass spectra of all of the compounds
leaching from materials being assessed are provided in the report describing the results of
the GCMS assessment.
Because of the sheer volume of data produced during this comparative exercise, it can
only be summarised within the main body of this report. All of the data provided by each
laboratory is given in Appendix 3.
In order to illustrate the approach used in assessing the data, one of the simpler data sets
(the results from the extract from the epoxy resin migration water using chlorinated test
water) has been used. The reference mixtures of compounds used in Stages 1, 2 and 3
(in this case referred to as Ref 3A and Ref 3B) were added to the migration water prior to
the extraction at concentrations of about 7-8 µg l-1. Table WP3.9 shows the data submitted
by each laboratory in terms of the names of identified compounds and MS data for
unknown compounds detected, together with the quantities of each compound detected
when calculated using the appropriate internal standards. The rows containing the entries
compounds utilised as internal standards are highlighted in turquoise, while those
containing entries for compounds deliberately added to the migration waters are
highlighted in gray, except where these compounds were not detected, when a yellow
highlight is used.
The total numbers of compounds detected (after blank subtraction) varied between 20
(ISS) and 31 (UBA). Of the 42 compounds reported in total, 13 (31%) were only reported
by a single laboratory while16 (38%), which included the nine isotopically-labeled internal
standards, were reported by all laboratories.
34
Table WP3.9. Compounds detected by each participating laboratory in extract from epoxy
resin migration water using non-chlorinated test water (blank subtracted).
Epoxy resin - non-chlorinated migration water
Compound
Hexane
d6 Benzene
n-Butanol
Unknown 43,49, ,85 ?
Methyl methacrylate
Toluene
Diacetone alcohol
4,4-Dimethyl-2-pentanone
d5 Chlorobenzene
d10 p-Xylene
Isopropyl benzene
d5 Phenol
Unknown 73,193,281,??
1-Octanol
N-methyl pyrrolidinone
Benzyl alcohol
2-Ethylhexanol
Acetophenone
Benzoic acid
Octanoic acid
d8 Naphthalene
Nonanoic acid
Unknown (170,220,238,188)
Unknown 174,192,66,46
2,6-Di-t-butyl-p-benzoquinone
d20 2,6-Di-t-butyl-4-methyl phenol
Dodecanoic acid
Diethyl phthalate
d34 Hexadecane
Hexadecane
Tri-n-butyl phosphate
Unknown 41,207,57,39
n-Butyl benzene sulphonamide
d10 Phenanthrene
Di-isobutyl phthalate
n-Hexadecanoic acid
Di-n-butyl phthalate
Fluoranthene
Bisphenol A
Di-(2-ethylhexyl) phthalate
d62 Squalane
Di(HCl) derivative of BADGE
WRc
NR
2.0
1.8
NR
7.0
NR
1.6
NR
2.0
1.0
8.5
8.0
NR
NR
1.1
6.1
5.9
9.4
0.0
NR
1.0
1.2
NR
NR
9.7
8.0
NR
8.4
1.0
15.3
6.7
NR
NR
2.0
NR
NR
NR
7.8
1.7
1.0
8.0
1.2
SP
NR
2.0
1.9
NR
6.1
NR
NR
1.7
2.0
1.0
6.2
8.0
NR
1.2
2.5
5.2
NR
7.8
0.0
NR
1.0
NR
1.1
NR
8.3
8.0
NR
8.5
1.0
9.5
9.1
NR
NR
2.0
NR
NR
NR
7.4
2.0
NR
8.0
NR
Concentrations found (µg/l)
Kiwa
UBA
JRC
ISS
NR
NR
10.7
NR
2.0
2.0
2.0
2.0
0.0
0.5
1.0
0.0
1.2
NR
NR
NR
6.2
5.1
9.7
8.2
4.1
NR
NR
4.9
1.1
1.2
1.1
NR
NR
NR
NR
NR
2.0
2.0
2.0
2.0
1.0
1.0
1.0
1.0
8.6
10.7
11.0
7.4
8.0
8.0
8.0
8.0
NR
NR
4.3
NR
NR
NR
NR
NR
0.0
0.0
1.4
1.1
5.8
6.2
12.8
0.0
4.9
NR
3.0
NR
7.9
10.7
17.1
5.4
0.0
2.8
0.0
0.0
NR
2.4
NR
1.0
1.0
1.0
1.0
1.0
NR
2.0
1.0
NR
NR
2.1
NR
NR
NR
1.8
NR
NR
10.2
15.0
15.4
14.3
8.0
8.0
8.0
8.0
NR
1.2
NR
NR
9.0
15.0
17.5
8.3
1.0
1.0
NR
1.0
12.2
1.3
18.2
12.1
8.2
5.4
14.8
7.5
NR
NR
1.5
NR
NR
1.4
NR
NR
2.0
2.0
2.0
2.0
NR
1.4
NR
NR
NR
1.4
NR
NR
NR
1.9
NR
NR
9.9
7.6
12.4
8.5
5.0
4.0
11.1
0.0
NR
NR
NR
NR
8.0
8.0
8.0
8.0
NR
NR
NR
NR
LHRSP
NR
2.0
0.4
NR
7.4
NR
1.0
NR
2.0
1.0
7.1
8.0
NR
NR
0.0
5.4
NR
7.0
0.5
NR
1.0
NR
NR
NR
11.4
8.0
NR
8.0
1.0
13.2
8.7
NR
NR
2.0
NR
NR
NR
6.8
1.3
NR
8.0
NR
Several of the compounds in the reference mixtures Ref 3A and Ref 3B were not detected
by all participants. As difficulties were encountered when some of these compounds were
run on GCMS as standard solutions in Stage 1, this was not surprising. Generally these
35
were the more polar compounds (e.g. benzoic acid was only detected by two laboratories;
N-methylpyrrolidinone was detected by four laboratories).
Of the compounds (13) which were only detected by a single laboratory, most (11) were
detected at apparent concentrations of less than 2.0 µg/l. It is difficult to determine why so
many compounds were detected by only a single laboratory when each laboratory was
analysing the same extract. The only stage at which some discrepancies could have
occurred would have been in the transfer of the extract from the CERTAN bottles in which
they were distributed to the vials used on the autosamplers on the various GCMS systems
in the participating laboratories. Obviously the autosampler vials used would be one
potential source of contamination, but generally these are extremely clean when supplied
and do not cause problems in this type of analysis. Also assuming that similar vials were
used for the extracts from the blank (chlorinated test water), blank subtraction should have
(in effect) removed this type of contamination problem, unless of course it is intermittent.
When those compounds which were only detected at apparent concentrations greater
than 5 µg/l (as reported by at least one laboratory) are considered (see Table WP3.10).
then a simpler picture emerges and only a few compounds have not been detected by all
participants.
In terms of the apparent quantities of the compounds detected (i.e. the quantities present
have been calculated relative to the appropriate isotopically-labeled internal standard), in
those cases where a compound has been detected by all participants the coefficients of
variation (CVs; relative standard deviation expressed as a % of the mean) were in the
range 21-46%. Considering that this was an inter-laboratory comparison, this is an
exceptionally good result. For those compounds which were not detected by all
participants, the data are less satisfactory (CVs in the range 63-108%; non-detections
have been included as zeros when calculating means and standard deviations), but are
probably acceptable.
Similar data are obtained from the epoxy resin migration water produced using chlorinated
test water (Table WP3.11). When compounds are detected by all seven participants, the
CVs for the apparent quantities detected are in the range 14-48% - again, quite
acceptable. However, as the number of detections falls, the CVs become larger – e.g. for
benzyl alcohol and bis-phenol A, detected by six participants the CVs are 85% and 89%
respectively, while for benzoic acid, detected by only two participants, the CV is 201%.
On the basis of these data it appears that provided a compound can be detected by all
laboratories, the quantities reported will be similar, which from a harmonisation point of
view, seems to be acceptable.
However the most important factor is that all laboratories should detect the same
compounds, and as this is much more dependent on the GC injector and GC column than
on the MS system used, as noted previously, it may be necessary to be more
presecriptive regarding these aspects of BS 6920: Part 4 if it is to be used as the basis of
a European standard.
36
Table WP3.10. Compounds detected at concentrations > 5 micrograms/litre in extract from epoxy resin migration water using nonchlorinated test water
Compound
Hexane
d6 Benzene
Methyl methacrylate
d5 Chlorobenzene
d10 p-Xylene
Isopropyl benzene
d5 Phenol
N-methyl pyrrolidinone
Benzyl alcohol
2-Ethylhexanol
Acetophenone
d8 Naphthalene
2,6-Di-t-butyl-p-benzoquinone
d20 2,6-Di-t-butyl-4-methyl phenol
Diethyl phthalate
d34 Hexadecane
Hexadecane
Tri-n-butyl phosphate
d10 Phenanthrene
Fluoranthene
Bisphenol A
d62 Squalane
Concentrations found (micrograms/litre)(blank subtracted)
WRc
NR
2.0
7.0
2.0
1.0
8.5
8.0
1.1
6.1
5.9
9.4
1.0
9.7
8.0
8.4
1.0
15.3
6.7
2.0
7.8
1.7
8.0
SP
Kiwa
NR
NR
2.0
2.0
6.1
6.2
2.0
2.0
1.0
1.0
6.2
8.6
8.0
8.0
2.5
0.0
5.2
5.8
NR
4.9
7.8
7.9
1.0
1.0
8.3
10.2
8.0
8.0
8.5
9.0
1.0
1.0
9.5
12.2
9.1
8.2
2.0
2.0
7.4
9.9
2.0
5.0
8.0
8.0
Internal standard
UBA
NR
2.0
5.1
2.0
1.0
10.7
8.0
0.0
6.2
NR
10.7
1.0
15.0
8.0
15.0
1.0
1.3
5.4
2.0
7.6
4.0
8.0
JRC
10.7
2.0
9.7
2.0
1.0
11.0
8.0
1.4
12.8
3.0
17.1
1.0
15.4
8.0
17.5
1.0
18.2
14.8
2.0
12.4
11.1
8.0
Number of
detections
ISS
LHRSP
NR
NR
1
2.0
2.0
7
8.2
7.4
7
2.0
2.0
7
1.0
1.0
7
7.4
7.1
7
8.0
8.0
7
1.1
0.0
4
0.0
5.4
6
NR
NR
3
5.4
7.0
7
1.0
1.0
7
14.3
11.4
7
8.0
8.0
7
8.3
8.0
7
1.0
1.0
7
12.1
13.2
7
7.5
8.7
7
2.0
2.0
7
8.5
6.8
7
0.0
1.3
6
8.0
8.0
7
Compound added to migration
water
Mean
Stdev
CV%
7.1
1.5
21
8.5
1.8
21
0.9
5.9
0.9
3.7
108
63
9.3
3.8
41
12.0
2.8
24
10.7
3.9
36
11.7
8.6
5.3
3.0
46
35
8.6
3.6
1.9
3.7
22
104
Not detected
Table WP3.11 Compounds detected at concentrations > 5 micrograms/litre in extract from epoxy resin migration water using chlorinated
test water
37
Compound
d6 Benzene
n-Butanol
Methyl methacrylate
d5 Chlorobenzene
d10 p-Xylene
Isopropyl benzene
d5 Phenol
N-methyl pyrrolidinone
Benzyl alcohol
Acetophenone
Benzoic acid
d8 Naphthalene
2,6-Di-t-butyl-p-benzoquinone
d20 2,6-Di-t-butyl-4-methyl phenol
Diethyl phthalate
d34 Hexadecane
Hexadecane
Tri-n-butyl phosphate
d10 Phenanthrene
Di-isobutyl phthalate
Di-n-butyl phthalate
Fluoranthene
Bisphenol A
d62 Squalane
Concentrations found (micrograms/litre)(blank subtracted)
WRc
2.0
1.8
7.1
2.0
1.0
8.1
8.0
0.8
5.1
8.1
0.0
1.0
10.4
8.0
8.3
1.0
14.7
3.8
2.0
NR
NR
7.1
1.2
8.0
SP
Kiwa
2.0
2.0
2.1
0.0
8.2
5.3
2.0
2.0
1.0
1.0
7.6
8.0
8.0
8.0
2.9
0.0
5.9
4.0
8.7
7.2
0.0
0.0
1.0
1.0
7.5
12.1
8.0
8.0
8.6
8.6
1.0
1.0
9.4
10.9
8.6
7.2
2.0
2.0
1.5
6.4
1.0
NR
6.4
7.6
1.4
4.2
8.0
8.0
Internal standard
UBA
2.0
0.7
6.3
2.0
1.0
11.0
8.0
0.0
4.9
2.1
3.2
1.0
9.9
8.0
9.4
1.0
7.7
3.1
2.0
2.0
8.2
7.2
3.4
8.0
JRC
ISS
LHRSP
2.0
2.0
2.0
1.1
0.0
0.0
12.9
7.2
12.7
2.0
2.0
2.0
1.0
1.0
1.0
10.4
6.5
9.7
8.0
8.0
8.0
1.6
1.7
0.0
17.3
0.0
6.4
15.4
6.9
11.4
0.0
1.0
0.0
1.0
1.0
1.0
6.6
11.2
10.6
8.0
8.0
8.0
7.4
7.3
6.1
1.0
1.0
1.0
10.7
10.1
10.6
8.3
6.5
7.1
2.0
2.0
2.0
1.1
1.2
NR
NR
1.2
NR
11.3
7.1
7.5
7.3
0.0
1.7
8.0
8.0
8.0
Compound added to migration
water
Number of Mean
detections
7
4
7
7
7
7
7
4
6
7
2
7
7
7
7
7
7
7
7
5
3
7
6
7
Table WP3.12. Compounds detected in extract from rubber migration water using non-chlorinated test water
38
Stdev
CV%
0.81
8.52
0.88
3.04
109
36
8.75
1.63
19
1.00
6.23
8.53
0.61
1.12
5.31
4.11
1.22
112
85
48
201
9.75
1.99
20
7.97
1.11
14
10.59
6.38
2.12
2.12
20
33
7.75
2.74
1.61
2.45
21
89
Not detected
Compound
Concentration (micrograms/litre; blank subtracted)
WRc
SP
Kiwa
UBA
JRC
1.1
NR
NR
NR
NR
5.8
3.1
NR
2.2
NR
3.3
1.4
NR
NR
NR
4.8
2.1
NR
NR
1.4
2.3
1.2
NR
1.6
1.1
2.0
2.0
2.0
2.0
2.0
1.4
NR
NR
NR
NR
NR
9.2
NR
NR
NR
1.1
NR
1.0
NR
NR
NR
2.3
NR
NR
NR
NR
NR
NR
1.0
NR
1.6
1.2
NR
NR
1.2
NR
2.7
NR
3.2
1.2
2.6
27.1
3.1
1.5
4.7
NR
NR
NR
NR
1.1
NR
NR
NR
NR
1.8
2.0
2.0
2.0
2.0
2.0
NR
NR
NR
1.2
NR
1.0
1.0
1.0
1.0
1.0
1.3
1.2
NR
1.1
2.3
8.0
8.0
8.0
8.0
8.0
6.7
10.2
9.3
3.8
5.6
6.7
NR
1.0
NR
5.6
2.9
3.0
2.4
NR
4.2
7.1
6.6
4.6
NR
8.1
NR
NR
NR
4.0
NR
Rubber, non-chlorinated migration water (continued)
WRc
SP
Kiwa
UBA
JRC
2-Hydroxy-2-phenylpropyl-4-methylbenzoic acid
NR
NR
NR
6.7
NR
Methyl ethyl ketone
Chloroform & Ethyl acetate
Trimethyloxirane
1,2-Dichloroethane
1,1,1-Trichloroethane
d6 Benzene
Carbon tetrachloride
Cyclohexane
n-Heptane
2,2,4,4-Tetramethylpentane
2.4-Pentadienenitrile
Pyridine
1-Methylpiperidine
Toluene
Tetrachloroethylene
Butenoic acid
d5 Chlorobenzene
3-Methylpyridine
d10-p-Xylene
Tetramethylurea
d5 Phenol
2-Ethylhexanol
Benzyl alcohol
Acetophenone
Alpha, alpha-dimethylbenzenemethanol
Unknown 105,77,51,50
39
No. of detections
ISS
NR
NR
NR
NR
NR
2.0
NR
NR
NR
NR
NR
NR
NR
2.2
NR
NR
2.0
NR
1.0
1.1
8.0
NR
NR
3.2
4.6
NR
LHRSP
NR
NR
NR
NR
NR
2.0
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
2.0
NR
1.0
NR
8.0
10.1
NR
3.0
7.9
NR
1
3
2
3
4
7
1
1
2
1
1
3
3
6
1
1
7
1
7
5
7
6
3
6
6
1
ISS
NR
LHRSP
NR
No. of detections
1
ester
3,5,5-Trimethyl-2-cyclohexenone + Unknown
NR
NR
NR
NR
1.1
42,39,53,55
Unknown 42,41,55,27
1.4
NR
NR
NR
NR
1-Piperidinecarboxaldehyde
7.4
10.1
NR
5.0
2.1
d8 Naphthalene
1.0
1.0
1.0
1.0
1.0
Tetramethylthiourea + ethylmethylpyrazine isomer
1.6
1.1
NR
NR
NR
Caprolactam
48.6
25.3
29.0
NR
23.2
Unknown 88,135,42,45
NR
NR
NR
NR
NR
4,7-Methano-1H-indene, 3a,4,7,7a-tetramethyl
NR
5.1
NR
NR
NR
Unknown 72,84,69,156
NR
2.8
NR
NR
3.1
Unknown 91,117,83,150
NR
NR
8.5
NR
3.5
Unknown 114,56,55,85
NR
NR
NR
8.8
NR
Dicyclopentadiene alcohol
6.6
NR
NR
NR
NR
Unknown 72,69,156,156?
NR
NR
5.0
NR
NR
5-Amino-3-oxo-4-hexanoic acid, methyl ester
NR
NR
NR
1.2
NR
Unknown 72,84,69,41 (M+ 156?)
4.1
NR
NR
NR
NR
Unknown 97,41,43,39
1.2
NR
NR
NR
NR
Dimethyl aniline
NR
NR
1.5
NR
NR
3-t-Butyl-4-hydroxyanisole
NR
3.0
NR
NR
NR
Unkown 161,203,218,175
NR
12.7
NR
NR
NR
Iso-propylphenyl ketone
1.6
NR
NR
NR
NR
Unknown 165,180,137,179
NR
NR
NR
2.7
2.6
2,6-Di-t-butyl-4-hydroxy-4-methyl-2,52.7
NR
NR
NR
NR
cyclohexadiene-1-one
2,6-Di-t-butyl-4-bromomethylphenol
NR
NR
NR
7.1
NR
N,N-Dimethylbenzamide
NR
NR
NR
NR
NR
Bicyclo[3.1.0]hexane-2-one, 1,5-di-t-butyl-3,3NR
NR
3.3
NR
NR
dimethyl
Rubber, non-chlorinated migration water (continued)
WRc
SP
Kiwa
UBA
JRC
2,6-Di-t-butyl-4-methylene-2,5-cyclohexadiene-12.7
NR
7.9
NR
16.6
40
NR
NR
1
NR
1.1
1.0
1.6
56.0
1.3
NR
4.4
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
5.3
1.0
1.3
20.1
NR
NR
NR
4.2
NR
NR
2.6
NR
NR
NR
NR
NR
NR
NR
NR
NR
1
6
7
4
6
1
1
3
3
1
1
2
1
1
1
1
1
1
1
2
1
NR
1.5
NR
NR
1.5
NR
1
2
1
ISS
2.3
LHRSP
NR
No. of detections
4
one
d20 2,6-Di-t-butyl-4-methyl phenol
Unknown 161,41,203,91
BHT
Unknown 84,172,88,41
Unknown 72,44,41,56
3-(3,4-methylenedioxyphenyl)ethynyl cyanide
d34 Hexadecane
Unknown 91,66,149,149?
Diethyl phthalate
Unknown 72,44,184,41
Unknown 41,175,128,69
Unknown 84,41,42,69
Unknown 219,220,263,262
N',N'-dimethyl-N-cycloheptylurea?
Unknown 175,69,41,175??
Piperidine-1,1'-carboxybis
Piperidine carboxylic acid, ethyl ester
bis-(Pentamethyleneurea)
2,6-Di-t-butyl-4-ethylphenol
2,6-Di-t-butyl-4-(dimethylaminomethyl)phenol
2,6-di-t-butyl-4-hydroxybenzaldehyde
Unknown 235,220,236,291
d10 Phenanthrene
N-Benzoylcyclohexylamine
Unknown 220,298,302,288
Unknown 57,41,219,287
3,5-Di-t-butyl-4-hydroxybenzaldehyde + alcohol
Unknown 212,213,84,128
Unknown 191,57,234,234??
41
8.0
8.0
8.0
8.0
8.0
NR
NR
NR
NR
NR
NR
1.5
1.8
1.5
1.0
1.8
NR
1.9
NR
NR
2.0
NR
NR
NR
NR
NR
NR
NR
1.1
NR
1.0
1.0
1.0
1.0
1.0
NR
NR
1.4
NR
NR
NR
1.4
1.4
NR
NR
NR
1.4
NR
NR
1.5
1.5
NR
NR
NR
NR
NR
NR
NR
NR
1.2
11.1
17.2
NR
NR
28.3
NR
NR
1.5
NR
NR
NR
NR
1.5
NR
NR
NR
1.9
NR
1.0
NR
NR
1.1
NR
NR
NR
3.4
NR
2.5
NR
NR
4.6
5.8
NR
NR
NR
NR
NR
15.6
29.8
NR
NR
NR
NR
NR
4.5
NR
NR
NR
10.0
NR
2.0
2.0
2.0
2.0
2.0
NR
NR
NR
1.0
NR
NR
NR
NR
2.3
NR
NR
NR
NR
NR
NR
NR
NR
1.6
NR
NR
NR
NR
NR
3.5
NR
Rubber, non-chlorinated migration water (continued)
WRc
SP
Kiwa
UBA
JRC
NR
NR
3.1
NR
NR
8.0
9.7
2.0
NR
NR
NR
1.0
NR
NR
NR
NR
NR
NR
NR
NR
1.9
NR
NR
NR
4.5
NR
NR
2.0
NR
NR
1.0
NR
NR
8.0
NR
3.3
NR
NR
NR
1.0
NR
NR
1.5
1.1
NR
24.1
NR
NR
1.4
NR
NR
6.2
NR
NR
NR
2.0
NR
NR
NR
NR
3.0
7
1
6
2
1
1
7
1
2
3
2
1
4
1
1
4
1
2
3
3
1
1
7
1
1
1
1
2
ISS
NR
LHRSP
NR
No. of detections
1
Unknown 84,41,212,42
Unknown 84,163,219,301
Cyclohexylamine, N-ethyl
Unknown 57,103,45,41
Unknown 112,153,225
Di-n-butyl phthalate
Unknown 84,41,56,69
Unknown 84,41,69,56
Unknown 41,69,56,112
Unknown 66,131,39,98 (M+ 228?)
Unknown 84,219,163,41
2-Aminodiphenylsulphone
Unknown 219,86,304,220
Unknown 57,45,29,202
Di-(2-ethylhexyl) phthalate
d62 Squalane
Squalene
Unknown 436,57,438
2.4
NR
NR
NR
NR
NR
NR
8.7
NR
NR
12.2
NR
NR
1.4
40.0
8.0
1.2
NR
1.9
21.8
1.0
1.3
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
34.7
8.0
NR
NR
1.8
NR
NR
NR
NR
NR
NR
NR
4.7
1.0
19.0
NR
1.81?
NR
6.0
8.0
NR
NR
Internal standard
42
NR
NR
NR
NR
NR
3.2
NR
NR
10.7
NR
32.4
2.1
1.7
NR
16.0
8.0
2.3
5.6
NR
NR
NR
NR
NR
2.4
NR
6.8
NR
NR
66.2
NR
NR
NR
26.4
8.0
1.1
NR
NR
10.9
NR
NR
NR
NR
2.0
3.5
NR
NR
NR
NR
NR
NR
29.8
8.0
2.0
NR
NR
NR
NR
NR
9.4
NR
NR
NR
2.1
NR
40.2
NR
NR
NR
NR
8.0
NR
NR
3
2
1
1
1
2
1
3
3
1
5
1
2
1
6
7
4
1
The data from the extract from the rubber migration water (non-chlorinated) (see Table
WP3.12) was considerably more complex than that from the epoxy resin migration waters.
In total 94 compounds were detected at concentrations > 1 µg/l, of which 49 (52%) were
only reported by one laboratory. Only the internal standards were detected by all
participants, and only 15 compounds were reported by at least four laboratories (i.e. more
than half of those participating).
Thirty two compounds were detected by at least one laboratory at an apparent
concentration > 5 µg/l, and twelve compounds were similarly detected at > 10 µg/l. Of
these latter twelve compounds two were only detected by a single laboratory. Table
WP3.13 shows the compounds detected at apparent concentrations > 10 µg/l. Although
this table is considerably simpler than the original data in Table WP3.12, there are still
some discrepancies, as follows:
• Toluene was detected by six laboratories, but five laboratories reported concentrations
in the range 1.5 - 4.7 µg/l while the sixth reported a concentration of 27.1 µg/l (i.e. an
order of magnitude greater than the mean of the other five laboratories). One
laboratory did not detect this compound. Toluene is occasionally detected as a
contaminant in blanks, and it should be noted that in the migration water obtained
when using chlorinated test water no laboratory reported this compound (after blank
subtraction) at a concentration > 1 µg/l. It therefore appears that the single high result
could be due to contamination, rather than the presence of toluene in the migration
water.
• Although 2-ethylhexanol was detected by six laboratories (mean 7.6 µg/l), it was its
presence was not reported by the remaining laboratory. This could be due to a
chromatographic problem.
• Caprolactam was detected by six laboratories (mean 33.7 µg/l), but not reported by
the remaining laboratory. Again this may be due to a chromatographic problem
• It appears likely that the unidentified compound (Unknown 219,220,263,262) may be
the
same
as
the
tentatively
identified
compound
2,6-di-t-butyl-4(dimethylaminomethyl)phenol, as four laboratories reported the former compound and
the remaining three laboratories reported the latter compound. The relative retention
times were not available, but since the compounds are tabulated in increasing order of
retention time, these two compounds do have similar retention times. This illustrates
the need to provide a mass spectrum for each compound believed to leach from the
material being tested.
• Two unidentified compounds (Unknown 84,163, 219, 301 and Unknown 84,219, 163,
41) may also be identical as no single laboratory reported both, but all laboratories
reported one or the other. Again copies of the mass spectra are required to resolve
this issue
Generally, the variation in the concentrations of the compounds reported by each
laboratory were similar to those found for the epoxy resin migration waters i.e. in the
range 35 – 70%. However it should be noted that the only compounds which were
reported by all laboratories (with the possible exception of compounds discussed above
as discrepancies) were the internal standards.
A comparison of the data from the extract from the rubber migration water (chlorinated)
shows that again a high percentage (62%) of the total number of compounds reported
(92) were reported by only a single laboratory. The corresponding numbers for the nonchlorinated migration water were 52% and 94. Sixteen compounds were reported by at
least four laboratories, compared to fifteen in the non-chlorinated migration water.
When only those compounds apparently present at a concentration > 10µg/l are
considered (see Table WP3.14), fewer compounds were reported than for the nonchlorinated migration water (see Table WP3.13), but there are similar discrepancies to
those mentioned above e.g. two of the compounds reported were apparently only
detected by single laboratories.
43
Table WP3.13 Compounds detected at concentrations > 10 µg/l in extract of migration water from rubber using non-chlorinated test water
Compound
Concentration (micrograms/litre; blank subtracted)
d6 Benzene
Toluene
d5 Chlorobenzene
d10-p-Xylene
d5 Phenol
2-Ethylhexanol
1-Piperidinecarboxaldehyde
d8 Naphthalene
Caprolactam
Unknown 161,203,218,175
d20 2,6-Di-t-butyl-4-methyl phenol
d34 Hexadecane
Unknown 219,220,263,262
2,6-Di-t-butyl-4-(dimethylaminomethyl)phenol
Unknown 235,220,236,291
d10 Phenanthrene
Unknown 84,163,219,301
Unknown 41,69,56,112
Unknown 84,219,163,41
Di-(2-ethylhexyl) phthalate
d62 Squalane
WRc
2.0
2.6
2.0
1.0
8.0
6.7
7.4
1.0
48.6
NR
8.0
1.0
11.1
NR
NR
2.0
NR
NR
12.2
40.0
8.0
Number of compounds detected at > 10 ug/l
4
44
SP
2.0
27.1
2.0
1.0
8.0
10.2
10.1
1.0
25.3
12.7
8.0
1.0
17.2
NR
NR
2.0
21.8
NR
NR
34.7
8.0
Kiwa
2.0
3.1
2.0
1.0
8.0
9.3
NR
1.0
29.0
NR
8.0
1.0
NR
15.6
NR
2.0
NR
4.7
19.0
6.0
8.0
UBA
2.0
1.5
2.0
1.0
8.0
3.8
5.0
1.0
NR
NR
8.0
1.0
NR
29.8
10.0
2.0
NR
10.7
32.4
16.0
8.0
JRC
2.0
4.7
2.0
1.0
8.0
5.6
2.1
1.0
23.2
NR
8.0
1.0
28.3
NR
NR
2.0
NR
NR
66.2
26.4
8.0
ISS
2.0
2.2
2.0
1.0
8.0
NR
1.1
1.0
56.0
NR
8.0
1.0
NR
4.5
NR
2.0
10.9
NR
NR
29.8
8.0
7
3
4
4
3
Internal standard
Detected by fewer than 50% of participants
LHRSP
2.0
NR
2.0
1.0
8.0
10.1
5.3
1.0
20.1
NR
8.0
1.0
24.1
NR
NR
2.0
NR
2.1
40.2
NR
8.0
4
No. of
detections
7
6
7
7
7
6
6
7
6
1
7
7
4
3
1
7
2
3
5
6
7
Mean
RSD
CV%
2.8
1.20
42.7*
7.6
4.7
2.65
3.33
34.8
70.5
33.7
14.88
44.2
20.2
7.59
37.6
34.0
25.5
21.08
12.53
62.0
49.2
* - outlier ignored
Table WP3.14 Compounds detected at concentrations > 10 µg/l in extract of migration water from rubber using chlorinated test water
Compound
Concentration >10 (micrograms/litre; blank subtracted)
No. of
detections
Mean
RSD
CV%
d6 Benzene
d5 Chlorobenzene
d10-p-Xylene
d5 Phenol
WRc
2.0
2.0
1.0
8.0
SP
2.0
2.0
1.0
8.0
Kiwa
2.0
2.0
1.0
8.0
UBA
2.0
2.0
1.0
8.0
JRC
2.0
2.0
1.0
8.0
ISS
2.0
2.0
1.0
8.0
LHRSP
2.0
2.0
1.0
8.0
7
7
7
7
1-Piperidinecarboxaldehyde
12.7
14.6
10.2
10.0
17.6
2.7
11.8
7
11.4
4.65
40.9
d8 Naphthalene
Caprolactam
2,6-Di-t-butyl-p-benzoquinone
2,6-Di-t-butyl-4-hydroxy-4-methyl-2,5cyclohexadiene-1-one
d20 2,6-Di-t-butyl-4-methyl phenol
Unknown 161,43,203,91
d34 Hexadecane
Unknown
d10 Phenanthrene
Di-(2-ethylhexyl) phthalate
d62 Squalane
1.0
84.9
1.1
1.0
76.4
NR
1.0
56.0
NR
1.0
34.3
3.4
1.0
144.2
21.0
1.0
241.5
NR
1.0
84.0
NR
7
7
3
103.0
69.77
67.7
2.4
NR
NR
NR
10.2
NR
NR
2
8.0
NR
1.0
NR
2.0
56.0
8.0
8.0
NR
1.0
NR
2.0
29.0
8.0
8.0
NR
1.0
NR
2.0
24.3
8.0
8.0
NR
1.0
NR
2.0
16.4
8.0
8.0
NR
1.0
14.0
2.0
26.3
8.0
8.0
26.2
1.0
NR
2.0
21.2
8.0
8.0
NR
1.0
NR
2.0
NR
8.0
7
1
7
1
7
6
7
28.9
13.98
48.4
3
3
3
3
3
3
2
Number of compounds detected at >10 ug/l
Internal standard
Detected by fewer than 50% of participants
45
The data obtained from the analyses of the polyester resin was again extremely complex.
For the epoxy resin and the rubber migration waters the extracts from the blanks and
chlorinated blanks (i.e. the test water and chlorinated test water) were relatively
uncontaminated. The results for the epoxy resin blank are shown in Table WP3.15, from
which it can be seen that (apart from the internal standards) only nine compounds were
detected at concentrations > 1 µg/l. The four compounds detected which are more volatile
than benzene (i.e. the first four compounds in the Table) are regularly detected as
contaminants in the dichloromethane solvent used for the extractions, and di-(2ethylhexyl)phthalate is a ubiquitous contaminant.
However the extract produced from the test water used for the leaching of the polyester
resin contained many contaminants. The results obtained for this extract are shown in
Table WP3.16, from which it can be seen that 72 compounds were detected at
concentrations > 1 µg/l. Although some of these may be contaminants in the
dichloromethane solvent, many are unlikely to be from this source, and it is known that
contamination problems were being encountered by the laboratory who carried out the
leaching tests and extractions when the migration waters and extracts were prepared.
Although this may be an extreme case caused by unusual circumstances, it does
demonstrate that on some occasions test waters or extracts can become unacceptably
contaminated. It is known that the UK regulatory body would not be prepared to accept
leaching data produced at a time when blank contamination of this degree was occurring,
and it may be necessary to set some minimum standard for acceptable blanks in a
harmonised procedure.
A further potential problem arising from heavily contaminated blanks is that it becomes
difficult to ascertain whether compounds detected in the blanks are also present in the
migration waters. This was noted earlier, and a procedure is suggested in BS6920: Part 4
for dealing with quantification when the concentrations of contaminants in blanks are low,
but this was not designed for heavily contaminated blanks. The obvious solution is (as
noted above) to set a minimum standard for blanks in terms of contaminant levels.
BS6920: Part 4 was originally designed to provide good estimates of compounds detected
in migration waters at concentrations in the range 1-10 µg/l. Data already presented
shows that for this range, the variation in the quantitative estimates produced by different
laboratories is relatively low (20-60%). The data in Table WP3.16 also illustrate another
factor which is that even when compounds are apparently present at very much higher
concentrations (> 100 µg/l) there is not very much more inter-laboratory variation. For
example for toluene, the mean concentration reported was 328 µg/l and the CV was 57%;
for styrene the comparable figures are 145 µg/l and 60%. As noted previously, regulatory
bodies may well demand specific analyses for compounds detected at high concentrations
so that more accurate data are obtained.
46
Table WP3.15 Compounds detected (concentrations > 1 µg/l) in extract from test water (blank) for epoxy resin
Epoxy resin - non-chlorinated blank
Compound
1,2-Dichloroethylene
Methylethyl ketone
Chloroform & Ethyl acetate
Tetrahydrofuran
d6 Benzene
Acetone
Methyl-t-butyl ketone
Toluene
d5 Chlorobenzene
d10 p-Xylene
d5 Phenol
2-Ethyl hexanol
d8 Naphthalene
d20 2,6-Di-t-butyl-4-methyl phenol
d34 Hexadecane
d10 Phenanthrene
Di-(2-ethylhexyl) phthalate
d62 Squalane
Concentrations found (>1 micrograms/litre)
WRc
NR
NR
16.5
28.4
2.0
NR
2.3
24.7
2.0
1.0
8.0
1.9
1.0
8.0
1.0
2.0
1.1
8.0
SP
3.1
1.2
10.7
25.6
2.0
NR
1.6
22.0
2.0
1.0
8.0
5.0
1.0
8.0
1.0
2.0
1.3
8.0
Kiwa
NR
NR
15.5
17.0
2.0
1.4
NR
23.7
2.0
1.0
8.0
4.2
1.0
8.0
1.0
2.0
1.2
8.0
Internal standard
47
UBA
NR
NR
25.3
NR
2.0
NR
NR
23.1
2.0
1.0
8.0
2.9
1.0
8.0
1.0
2.0
NR
8.0
JRC
NR
NR
20.4
34.3
2.0
NR
2.1
33.8
2.0
1.0
8.0
3.5
1.0
8.0
1.0
2.0
NR
8.0
ISS
NR
NR
NR
NR
2.0
NR
2.3
NR
2.0
1.0
8.0
NR
1.0
8.0
1.0
2.0
1.4
8.0
LHRSP
NR
NR
6.1
18.4
NR
NR
NR
26.8
2.0
1.0
8.0
15.0
1.0
8.0
1.0
2.0
NR
8.0
Table WP3.16 Compounds detected (concentrations > 1 µg/l) in extract from test water (blank) for polyester resin
Compound
WRc
NR
NR
NR
NR
NR
181.3
NR
NR
2.0
44.5
58.8
NR
NR
8.4
10.0
58.5
15.3
1.1
125.3
2.5
16.3
1.0
6.8
45.8
8.0
NR
2.0
5.7
2-Methoxy-2-methylpropane
3-Methylpentane
Hexane isomer
Di-isopropyl ether
Ethyl acetate
C6H12 isomer
Chloroform
Unknown 83,56,69
d6-benzene
Carbon tetrachloride + C6H12 isomer
C6H12 isomer
2,2-Dimethylhexane
2,2-Dimethylheptane
Ethyl acrylate + Bromodichloromethane
3-methyl-1H-pyrazole
5-Methylisoxazole + C8H18 isomer
Methyl methacrylate
Dimethyl disulphide
Toluene
1,3-Dichloropropane
Dibromochloromethane
1,2-Dibromoethane
4-Methyl-3-penten-2-one
Tetrachloroethylene
Diacetone alcohol
4,4-Dimethyl-2-pentanone
d5-chlorobenzene
Xylene isomer
48
Concentrations detected (micrograms/litre)
SP
UBA
JRC
ISS
1.3
NR
NR
NR
11.9
NR
NR
NR
140.8
NR
NR
NR
106.5
NR
NR
NR
4.1
NR
NR
NR
188.1
NR
410.9
NR
95.8
91.4
435.3
NR
NR
87.1
NR
NR
2.0
2.0
2.0
2.0
41.4
53.6
167.6
NR
NR
NR
NR
NR
15.9
NR
NR
NR
NR
12.6
17.2
NR
5.3
NR
7.0
NR
11.9
14.4
14.7
14.6
47.4
49.4
104.9
55.5
10.5
10.7
NR
15.1
NR
NR
NR
NR
235.7
365.6
630.6
186.7
NR
NR
NR
1.5
11.6
13.1
32.2
23.5
NR
NR
NR
1.3
5.6
7.1
14.6
NR
32.6
61.8
141.6
65.7
NR
8.6
2.0
NR
8.8
NR
NR
11.0
2.0
2.0
2.0
2.0
4.1
5.5
4.5
7.9
LHRSP
NR
NR
NR
NR
NR
15.2
11.3
NR
2.0
149.4
NR
NR
NR
NR
NR
NR
NR
NR
434.4
NR
15.8
NR
NR
61.4
NR
NR
2.0
NR
Polyester resin - non-chlorinated blank (continued)
WRc
SP
UBA
d10-p-Xylene
1.0
1.0
1.0
Bromoform
2.8
1.9
NR
Butylpropenoate
NR
4.3
NR
Xylene isomer
3.4
4.4
2.6
Xylene isomer
3.4
NR
4.3
Unknown 55,56,73,27
NR
NR
NR
Styrene
77.2
101.3
161.8
Xylene isomer
8.2
NR
NR
1,1,2,2-Tetrachloroethane
8.2
12.9
7.1
C3-Alkylbenzene isomer
1.1
1.0
1.3
Benzaldehyde
2.6
2.8
NR
Chlorotoluene isomer
1.9
2.1
2.1
Benzoyl bromide
NR
NR
4.9
C3-Alkylbenzene isomer
1.0
NR
NR
C3-Alkylbenzene isomer
1.0
1.1
NR
d5-phenol
8.0
8.0
8.0
Unknown
NR
NR
3.7
Butyl methacrylate + C3-alkylbenzene isomer
21.1
3.5
NR
C3-Alkylbenzene isomer
27.2
21.1
43.7
Dichlorobenzene isomer (m-?)
13.0
NR
18.0
Dichlorobenzene isomer (p-?)
27.2
8.7
43.7
C12H26 isomer
2.0
NR
NR
C3-Alkylbenzene isomer
1.0
NR
NR
C12 H24 isomer
1.0
NR
NR
Dichlorobenzene isomer (o-?)
3.2
2.1
6.2
2-Ethylhexanol
9.6
9.1
NR
C4-Alkylbenzene isomer
2.0
1.4
3.2
C4-Alkylbenzene isomer
3.4
2.2
6.2
C4-Alkylbenzene isomer
3.9
2.1
4.1
2-Methyl-1-propenyl benzene
NR
NR
NR
49
JRC
1.0
2.6
2.8
4.3
NR
NR
312.6
NR
14.2
NR
3.9
3.3
NR
NR
NR
8.0
NR
NR
45.4
20.4
57.5
NR
1.3
NR
5.6
11.5
2.0
4.9
2.8
1.7
ISS
1.0
NR
NR
13.9
NR
12.2
80.2
NR
13.6
NR
4.5
3.3
NR
NR
3.2
8.0
NR
0.8
44.4
22.7
43.6
NR
NR
NR
5.7
NR
5.8
6.4
NR
NR
LHRSP
1.0
NR
NR
NR
NR
NR
134.9
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
32.8
NR
11.6
NR
NR
NR
NR
NR
NR
NR
NR
NR
Polyester resin - non-chlorinated blank (continued)
WRc
SP
UBA
JRC
Unknown 41,39,43,57
NR
NR
NR
1.3
Trichlorobenzene isomer
2.9
1.8
3.8
NR
1-Dodecanol
NR
1.3
NR
NR
Ethyl benzoate
NR
2.5
NR
NR
Benzoic acid
NR
NR
NR
NR
d8-naphthalene
1.0
1.0
1.0
1.0
2-(2-hydroxypropoxy)propanol
NR
1.0
NR
NR
2-Methyl-3-hydroxy-2,4,4-trimethylpentyl
NR
1.1
NR
NR
propanoate
2,6-Di-t-butyl-p-benzoquinone
NR
1.3
1.6
NR
Unknown 41,39,67,91
NR
NR
NR
1.4
d20-BHT
8.0
8.0
8.0
8.0
d34-hexadecane
1.0
1.0
1.0
1.0
Diethyl phthalate
NR
NR
1.2
NR
Propanoic acid, 2-methyl-1-(1,1-dimethylethyl)NR
1.8
NR
NR
2-methyl-1,3-propandiyl ester
Unknown 43,71,41,55
NR
NR
NR
1.7
d10-Phenanthrene
2.0
2.0
2.0
2.0
Di-isobutyl phthalate
1.1
3.0
5.2
2.3
Di-n-butyl phthalate
NR
1.3
2.5
NR
d62-Squalane
8.0
8.0
8.0
8.0
Internal standard
50
ISS
NR
3.2
3.2
NR
3.8
1.0
NR
NR
LHRSP
NR
NR
NR
NR
NR
1.0
NR
NR
NR
NR
8.0
1.0
NR
NR
NR
1.7
8.0
1.0
NR
NR
NR
2.0
2.3
NR
8.0
NR
2.0
7.3
NR
8.0
Table WP3.17 Compounds detected by WRc-NSF in extract from non-chlorinated migration water from cement with organic additive
Contact Name:
Client:
Client Reference:
WRc-NSF Reference:
WRc-NSF Contract No:
Ret.
Time
4.80
5.13
5.33
5.93
6.30
6.92
8.18
8.60
9.32
9.97
10.65
11.38
11.43
11.50
11.62
12.07
12.20
12.40
12.68
12.97
Ret.
Corr. RT
Temp.
36.40
2.30
39.07
2.63
40.67
2.83
45.47
3.43
48.40
3.80
53.33
4.42
63.47
5.68
66.80
6.10
72.53
6.82
77.73
7.47
83.20
8.15
89.07
8.88
89.47
8.93
90.00
9.00
90.93
9.12
94.53
9.57
95.60
9.70
97.20
9.90
99.47
10.18
101.73
10.47
H A James
WRc-NSF Ltd
WP3-Kiwa
N/A
12702-0
Sample Code:
Sample Type:
Data System Code:
Associated Blank:
Sample Volume:
RRT
Compound
0.14
0.16
0.17
0.21
0.23
0.27
0.35
0.37
0.42
0.46
0.50
0.54
0.55
0.55
0.56
0.59
0.59
0.61
0.62
0.64
Methyl ethyl ketone
Ethyl acetate
Tetrahydrofuran
d6-Benzene
n-Butanol
1,4-Dioxane
2-Butanol
Toluene
4-Methyl-3-penten-2-one (Mesityl oxide)
Furural
d5-chlorobenzene
d10-p-Xylene
Bromoform
Xylene isomer
Cyclohexanone
Di-n-butyl ether
1,3,5-Trioxane
2-Butoxyethanol
N,N-Diethylformamide
3-Methyl-4-heptanone
51
Chlorinated migration water
Migration water
T20260.19
Kiwa blank
1 litre
Con.L** Peak
Area
P
P
P
P
P
P
P
P
P
P
P
P
P
T
P
P
T
T
T
T
7.61
5.83
6.79
0.67
910.00
0.43
12.14
1.89
0.58
0.57
1.71
0.62
0.40
0.43
0.27
9.01
0.83
17.81
2.46
0.43
Number of samples:
Date Received
Date Analysed:
Page :
Method Ref:
Conc. Internal
(ug/l) Standard
8.9
6.8
7.9
2.0
1064.3
0.5
14.2
2.2
0.7
0.7
2.0
1.0
0.5
0.5
0.3
10.5
1.0
20.8
2.9
0.5
Cl
Cl
Cl
I.S.
Cl
Cl
Cl
Cl
Cl
Cl
I.S.
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
1
N/A
1 of 4
Org 042 $
Origin of Peak
Test material
Test material
Test material
Internal Standard
Test material
Test material
Test material
Test water
Test material
Test material
Internal Standard
Internal Standard
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Cement with organic additive, non-chlorinated migration water (continued)
RRT Compound
Con.L** Peak
Conc. Internal
Ret.
Time
13.42
14.65
14.70
14.78
14.93
15.03
15.42
15.57
16.02
16.85
17.87
18.83
18.92
20.78
22.40
23.22
Ret.
Corr. RT
Temp.
105.33
10.92
115.20
12.15
115.60
12.20
116.27
12.28
117.47
12.43
118.27
12.53
121.33
12.92
122.53
13.07
126.13
13.52
132.80
14.35
140.93
15.37
148.67
16.33
149.33
16.42
164.27
18.28
177.20
19.90
183.73
20.72
24.38
24.82
25.93
26.30
27.68
193.07
196.53
205.47
208.40
219.47
21.88
22.32
23.43
23.80
25.18
1.34
1.37
1.43
1.46
1.54
28.23
223.87
25.73
1.58
28.80
29.15
228.40
231.20
26.30
26.65
1.61
1.63
Area
0.67
0.74
0.75
0.75
0.76
0.77
0.79
0.80
0.83
0.88
0.94
1.00
1.01
1.12
1.22
1.27
Benzaldehyde
d5 Phenol
Dipropylene glycol methyl ether isomer
Dipropylene glycol methyl ether isomer
Hexanoic acid
1-(2-methoxypropoxy)-2-propanol
Benzyl alcohol
2-Ethylhexanol
Acetophenone
Heptanoic acid
2-Ethylhexanoic acid
d8-Naphthalene
Octanoic acid
Nonanoic acid
Decanoic acid
C2- (dimethyl or ethyl) phenol ethoxylate
isomer
Dodecanol
d20-BHT
Dodecanoic acid
Methyl phenol diethoxylate
C2- (dimethyl or ethyl) phenol
diethoxylate isomer
C2- (dimethyl or ethyl) phenol
diethoxylate isomer
C3 - phenol diethoxylate isomer
d10-Phenanthrene
52
P
P
T
T
P
T
P
P
P
P
P
P
P
P
T
T
P
P
P
T
T
T
T
P
(ug/l)
0.3
0.4
41.70
48.8
0.8
0.9
6.60
7.7
0.91
5.76
12.67
1.1
6.7
14.9
24.62
29.0
23.33
2.77
1.82
27.4
3.3
2.1
7.37
6.80
52.51
17.32
8.7
8.0
61.8
20.4
28.52
25.0
2.66
2.28
2.3
2.0
Origin of Peak
Standard
Cl
I.S.
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
BHT
I.S.
BHT
BHT
BHT
BHT
Test material
Internal Standard
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Internal Standard
Test material
Test material
Test material
Test material
BHT
I.S.
BHT
BHT
Ph
Test material
Internal Standard
Test material
Test material
Test material
Ph
Test material
Ph
I.S.
Test material
Internal Standard
Ret.
Time
30.55
31.37
31.58
31.80
32.17
32.58
32.83
33.95
34.07
34.95
35.18
35.48
35.83
37.08
37.77
37.95
38.17
38.47
38.77
39.35
40.05
40.17
40.98
41.28
41.62
41.95
Ret.
Corr. RT
Temp.
242.40
28.05
248.93
28.87
250.67
29.08
252.40
29.30
255.33
29.67
258.67
30.08
260.67
30.33
269.60
31.45
270.53
31.57
277.60
32.45
279.47
32.68
281.87
32.98
284.67
33.33
294.67
34.58
>300
35.27
>300
35.45
>300
35.67
>300
35.97
>300
36.27
>300
36.85
>300
37.55
>300
37.67
>300
38.48
>300
38.78
>300
39.12
>300
39.45
Cement with organic additive, non-chlorinated migration water (continued)
RRT Compound
Con.L** Peak
Conc. Internal
Area
1.72
1.77
1.78
1.79
1.82
1.84
1.86
1.93
1.93
1.99
2.00
2.02
2.04
2.12
2.16
2.17
2.18
2.20
2.22
2.26
2.30
2.31
2.36
2.37
2.39
2.42
Methyl phenol triethoxylate isomer
Di-isobutyl phthalate
C2 - phenol triethoxylate isomer
C2 - phenol triethoxylate isomer
C2 - phenol triethoxylate isomer
C3 - phenol triethoxylate isomer
C3 - phenol triethoxylate isomer
Methyl phenol tetraethoxylate isomer
Methyl phenol tetraethoxylate isomer
C2 - phenol tetraethoxylate isomer
C2 - phenol tetraethoxylate isomer
C2 - phenol tetraethoxylate isomer
C3 - phenol tetraethoxylate isomer
Methyl phenol pentaethoxylate isomer
C2 - phenol pentaethoxylate isomer
C2 - phenol pentaethoxylate isomer
C2 - phenol pentaethoxylate isomer
C2 - phenol pentaethoxylate isomer
C3 - phenol pentaethoxylate isomer
d62-Squalane
Methyl phenol hexaethoxylate isomer
Methyl phenol hexaethoxylate isomer
C2 - phenol hexaethoxylate isomer
C2 - phenol hexaethoxylate isomer
C2 - phenol hexaethoxylate isomer
C3 - phenol hexaethoxylate isomer
53
T
P
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
P
T
T
T
T
T
T
(ug/l)
41.27
1.37
36.2
1.2
66.25
58.1
2.16
1.00
1.9
0.9
57.80
50.7
146.34
64.8
4.10
57.53
1.8
25.5
91.20
40.4
2.80
18.06
1.2
8.0
32.25
14.3
69.70
30.9
1.46
0.6
Origin of Peak
Standard
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Sq
Sq
Sq
Sq
Sq
Sq
Sq
Sq
Sq
Sq
I.S.
Sq
Sq
Sq
Sq
Sq
Sq
Test material
Test water
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Test material
Internal Standard
Test material
Test material
Test material
Test material
Test material
Test material
Ret.
Time
43.57
44.75
45.20
45.72
48.75
50.28
Ret.
Temp.
>300
>300
>300
>300
>300
>300
Corr. RT
Cement with organic additive, non-chlorinated migration water (continued)
RRT Compound
Con.L** Peak
Conc. Internal
Area
41.07
42.25
42.70
43.22
46.25
47.78
2.51
2.59
2.61
2.65
2.83
2.93
Methyl phenol heptaethoxylate isomer
C2 - phenol heptaethoxylate isomer
C2 - phenol heptaethoxylate isomer
C2 - phenol heptaethoxylate isomer
Methyl phenol octaethoxylate isomer
C2 - phenol octaethoxylate isomer
T
T
T
T
T
T
(ug/l)
18.15
8.0
33.29
14.7
8.57
11.55
3.8
5.1
Origin of Peak
Standard
Sq
Sq
Sq
Sq
Sq
Sq
Test material
Test material
Test material
Test material
Test material
Test material
Internal standards used: Bz=d6-Benzene, Cl=d5-Chlorobenzene, Xy=d10-p-Xylene, Po=d5-Phenol, Na=d8-Naphthelene,BHT=d20 2,6-Di-t-butyl-4-methyl
phenol, Hx=d34-Hexadecane, Ph=d10-Phenanthrene and Sq=d62 Squalane
**Con.L = Confidence level of identification: P=Positive, T=Tentative and U=Unknown
54
It was considered almost impossible to correlate the data produced by all of the
laboratories for the extracts from the migration waters from the cement with organic
additive. There were two reasons for this – firstly, large numbers of unknown compounds
were detected, secondly, there were many co-eluting compounds present in the extracts
rendered the quantification of individual compounds difficult. Therefore, only the data
produced by one laboratory has been included here (although all of the data tables are
given in the Appendix 3). It was noted, after the extract had been circulated to participants
that the migration water which had been extracts was the first 72 hour migration water,
rather than the third 72 hour migration water, as intended. It is likely that this first migration
water contained very much higher concentrations of compounds leached from this
material than would have been the case had the third migration water been used.
As can be seen from Table WP3.17, about 70 compounds were detected, many of which
were tentatively identified as alkylphenol ethoxylates (from dimethyl (or ethyl)phenol
monoethoxylate to dimethyl (or ethyl)phenol octaethoxylate). It is assumed that these
compounds are used in the formulation for their surfactant properties.
Stage 6 – Establishment of a GCMS database
Kiwa introduced the Infospec software package to participants. It allows both
chromatographic and mass spectral data to be included for compounds of interest, and all
of the features necessary for the purposes of the EAS are already present. As currently
configured, users can add information and a central organisation can assess whether
such additions are useful or appropriate. The utility of this latter feature for the purposes of
the EAS needs to be considered, as it may be better to make it a read-only database with
only a central organisation allowed to introduce additional data.
55
56
Conclusions
The conclusions that can be drawn from the work undertaken are as follows:
Stage 1 – Harmonisation of the basic GCMS procedure
The aim of this stage was to allow the participants to become accustomed to the GCMS
operating conditions as laid out in BS6920: Part4, using a range of compounds previously
detected in extracts from materials migration waters.
Generally this Stage progressed satisfactorily. The majority if the compounds present in
the reference mixtures were detected.
Some initial difficulties in satisfactorily resolving the most volatile internal standard (d6benzene) from the solvent used (dichloromethane) were reported by some participants.
This may be because nowadays analyses for volatile compounds are carried out using
headspace techniques coupled to GCMS.The relatively low initial GC temperature
required may cause some problems when ambient laboratory temperatures approach or
exceed 30°C, so if rooms in which GCMS equipment is operated cannot be temperature
controlled (e.g. at 25°C or below), it may be necessary to use GCs fitted with cooling
options (e.g. utilising liquid CO2 or liquid N2) which allow sub-ambient operation, in order
to minimise the time required to cool the GC oven temperature from the final to the initial
temperature.
Although they were selected for their suitability to analyse a wide range of organic
compounds, none of the GC columns used by participants were ideal as no participants
detected all of the compounds present in the reference mixtures. However this was not
surprising as some of the compounds are quite polar and would normally either be
derivatised prior to analysis involving GC or analysed using other techniques such as
liquid chromatography (LC). None of the GC columns used was significantly better than
any of the others used, so at the end of this Stage it did not appear to be necessary to
specify a particular column in a harmonised procedure.
Stage 2 – Selection of the solvent extraction procedure
The work in this Stage was intended to provide data to allow a decision to be taken on
which of the two available alternative solvent extraction procedures (UK and France) was
the best. The major difference was that the UK procedure involves extraction at a single
pH (pH2) whereas in the French procedure samples are extracted at two different pH
values (pH 2 and pH 10) and the extracts combined prior to concentration and analysis.
The results of this exercise showed that the French procedure was marginally better than
the UK procedure, particularly with respect to the recovery of basic compounds.
Stage 3 – Selection of possible SPE procedure
For various reasons (e.g. it is easily automated, less solvent is used), for specific (targetcompound) analysis SPE is nowadays the method of choice for the GCMS analysis of
compounds in drinking waters. Although SPE has not been widely used for the type of
investigative analysis needed for materials testing, it was considered worthwhile doing
some work to establish whether there was a possibility that it could be used in preference
to solvent extraction, particularly as the use of chlorinated solvents is being discouraged.
The outcome of this Stage was not sufficiently encouraging to allow a suggestion that
SPE should replace solvent extraction, as several compounds which are recovered with
good efficiently using solvent extraction were poorly recovered (if at all) when using SPE.
However it may be worth investigating other adsorbents used for SPE as a future
57
research project, as there are now numerous types available from various manufacturers,
and it was impossible to investigate the utility of all of these within the timescale and
budget of this project.
Stage 4 – Comparison of preferred solvent extraction procedure with preferred SPE
procedure
As neither of the SPE procedures tested was comparable to the preferred solvent
extraction procedure, it was decided that it would not be worthwhile carrying out the
proposed work for this Stage.
Stage 5 – Investigation of the overall preferred GCMS assessment procedure
Each of the seven participating laboratories analysed 14 solvent extracts prepared from
migration waters and blanks as follows:
•
•
•
•
epoxy resin - migration water, chlorinated migration water, test water blank,
chlorinated test water blank;
polyester resin - migration water, chlorinated migration water, test water blank,
chlorinated test water blank;
rubber - migration water, chlorinated migration water, test water blank, chlorinated test
water blank;
cement with organic additive – migration water, chlorinated migration water (extracts
from the test water blank and chlorinate test water blank were not circulated).
The compounds detected in the extracts were tabulated and quantified using the
procedure set out in BS6920: Part 4. The data tables produced by each laboratory are
given in Appendices 3A – 3G.
The collation of the data proved to be extremely time-consuming for the following reasons:
• with the exception of the test water blanks and chlorinated test water blanks for the
epoxy resin and rubber materials, all of the solvent extracts contained many
compounds (up to 100) detectable by GCMS;
• because all participants used different GC columns, it was very difficult to ascertain
which compounds which had not been identified by one or more laboratories were
identical;
Based on experience gained of analysing migration waters from materials submitted to the
UK regulatory body, the extracts from the rubber and cement with organic additive
products were extremely complex, and in retrospect may not have been the most
appropriate extracts to circulate to laboratories who previously had little or no experience
of this type of work.
However, the results suggest that provided the level of interest is set at about 5 µg/l,
different laboratories (once they had gained some experience of this type of work) would
produce comparable data.
Care must be taken to ensure that solvent of high purity is used for the solvent extraction,
and that blanks and/or migration waters are not contaminated, either during the
preparation of the migration waters or when the solvent extraction is carried out.
The greatest discrepancies in the data was noted for very volatile compounds (those
which are more volatile than the internal standard d6-benzene). If such compounds are of
interest than an alternative technique such as purge and trap GCMS is more appropriate
for their determination.
58
Recommendations
As follows:
• More inter-laboratory exercises are required to allow laboratories to gain experience.
• A proficiency testing scheme is required to ensure that approved laboratories are (and
continue to be) competent.
• Specifications need to be set for acceptable blanks (both extraction solvent blanks and
migration test water blanks).
• A mass spectral library of all of the compounds on the Drinking Water Positive List
(DWPL) which are amenable to GCMS analysis using the agreed protocol needs to be
produced to ensure consistency of data between laboratories; this library will require
regular updating as new compounds are added to the DWPL.
• As the GCMS method provides estimates of the concentrations of compounds
detected, dedicated analytical methods may still be required for some compounds
where the concentrations found give rise to concern and accurate quantification is
necessary.
• Additional R&D is required to establish an effective alternative SPE extraction
procedure.
• Consideration should be given to implementing a separate test method for very volatile
compounds using purge and trap GCMS analysis.
• To facilitate the operation of the EAS it is recommended that the appropriate regulatory
body sets pass/fail criteria for the GCMS assessment. These criteria should include a
concentration below which any compound detected is considered insignificant.
• The overall performance of the method is such that it is recommended that GCMS
analysis for unsuspected compounds should be incorporated into the overall EAS.
59
60
Appendix 1 – BS6920: Part 4
(Annexes relating to leachate preparation omitted)
61
Introduction
Chemicals that leach from materials used in contact with public water supplies can cause
health concerns for consumers. Potential health effects of these chemicals are assessed in
three stages:
1) preparation of leachates by exposing a portion of the material to water under controlled
conditions;
2) analysis of the leachates;
3) toxicological evaluation of the concentrations of substances identified.
Analysis of organic substances present in the leachates usually involves two types of
analytical methods:
a screening method which enables a qualitative and semi-quantitative assessment to be
made of unspecified organic compounds;
accurate quantitative methods for the determination of specific target compounds known to
be present in the chemical formulations of the materials.
This part of BS 6920 describes the analytical procedures based upon gas chromatography
and mass spectrometry (GCMS) used to screen leachates for unspecified organic
compounds derived from finished products such as pipes, protective coatings, membranes
etc.
This method is suitable for leachates from all non-metallic materials that may be used in
contact with water for human consumption, and which may be the subject of an application
for approval by the national regulator. In addition, annex A gives general guidance on the
preparation of test samples and their leachates.
NOTE 1 At the time of publication, products for use in contact with public water
supply must be approved by the Secretary of State for the Environment under the
provisions of regulation 25(1)(a) of the Water Supply (Water Quality) Regulations
1989 [1], unless any of the subsections 25(1)(b), (c) or (d) applies. The methods
described in this section of BS 6920 may form the basis of leaching test
specifications issued by the committee which advises the Secretary of State on the
approval of substances and products for use in contact with public water supplies.
Equivalent provisions apply in Scotland and Northern Ireland.
NOTE 2 The results of these tests are assessed by reference to existing
toxicological data concerning the chemicals identified. Where little or no
information exists, it may be necessary for the substance itself or the material
leachate to be submitted for toxicity testing.
1 Scope
This part of BS 6920 describes a method for identifying organic chemicals which are
amenable to GCMS analysis using the techniques described and which may leach from a
product into water intended for human consumption. A method of calculating the
concentrations of the organic substances identified is also provided.
62
This part of this British Standard is not applicable to the toxicological evaluation of
chemicals.
NOTE The method to be used for the preparation of leachates will be specified by
the National Regulator. Annex A provides a suitable method for preparing these
leachates.
Terms and definitions
For the purposes of this British Standard, the following terms and definitions apply.
2.1 amu
atomic mass unit; defined as 1/12 of the mass of a single atom of carbon-12 in the gas phase
(i.e. unbound), at rest and in its ground state
2.3 asymmetry factor
(As), measure of the adsorption of a compound during gas chromatographic analysis. It
may be derived from the equation
As = (a+b)/2b
where:
is the distance from the leading edge of the peak at the point on the baseline at
which a perpendicular dropped from the peak maximum crosses it;
is the corresponding distance from the trailing edge of the peak.
a
b
Provide details of the method of calculation of peak asymmetries in the final test report.
NOTE Some manufacturer’s GCMS software packages allow the calculation of
peak asymmetries to be produced automatically.
2.3 electron impact ionization
ionization by a beam of electrons
2.4 GCMS
analytical instrument comprising a gas chromatograph (GC) linked to a mass spectrometer
(MS)
2.5 GCMS general survey analysis
acquisition of a series of mass spectra (up to several thousand) during the course of a gas
chromatographic run, by operating the mass spectrometer in a continuous cyclic scanning
mode over a wide m/z range
2.6 internal standards
organic compounds added to the leachate at a known concentration prior to the
commencement of the analysis
63
NOTE Internal standards are added for the following reasons (a) to demonstrate
that the analysis has been undertaken successfully, and (b) to provide a reference to
allow other substances detected to be quantified. Ideally, the internal standards
should not be present in the leachate; for this reason, isotopically labelled standards
are preferred.
2.7 laboratory blank
water sample known to contain negligible levels of contamination, to which internal
standards have been added and which is then analysed in the same way as the leachate
NOTE Laboratory blanks are used to check for potential contamination of either leachates
or solvent extracts which may occur within the laboratory during analysis.
2.8 leachate
aqueous solution that results from leaving test water in contact with the test material under
the specified test conditions - see annex A
2.9 mass spectrometric resolution
measure of the capability of the mass spectrometer to correctly distinguish two mass
spectral peaks, having similar mass to charge m/z values, as separate peaks; when z =1, this
is generally denoted by m2/(m2 − m1) where m2 has the higher m/z value and m1 has the
lower m/z value
NOTE A mass spectrometer set up so that the resolution is 650 will satisfactorily
resolve and assign the correct masses to mass spectral peaks at m/z 649 and
m/z 650.
2.10 m/z
mass-to-charge ratio of an ion
NOTE As most ions produced by electron impact ionisation are singly charged,
this ratio usually corresponds to the mass of an ion. However, exceptionally, ions
may possess multiple charges.
2.11 solvent extract
solution containing compounds partitioned from the leachate into the extraction solvent (in
this case dichloromethane)
2.12 total ion current
(TIC), sum of all the separate ion currents carried by the individual ions contributing to a
mass spectrum
2.13 TIC chromatogram
graphical representation of the TIC versus time
64
3 Principle
A mixture of internal standards is added to each of the test leachates prior to solvent
extraction with dichloromethane. The solvent extract is concentrated and analysed by
GCMS to determine the identity and approximate concentrations of organic chemicals that
may be present.
The mass spectrometer is used in a repetitive full-scan mode and the mass spectra
produced are recorded by, and stored on, the GCMS data system.
Wherever possible, each compound detected is identified and may be quantified by
reference to the responses obtained for the internal standards.
NOTE 1 The number and duration of leaching (migration) periods, the nature of
the test samples(s) and the test surface area to volume ratio are specified by the
national regulator based upon the nature of the product and its proposed use.
Annex A provides a typical ‘model’ for the preparation of leachates based upon the
current requirements of the national regulator.
NOTE 2 The methods used to identify organic compounds from their mass spectra
do not form part of this method, but further information on this subject is given in
annex D.
4 Reagents
4.1 General, only reagents of analytical grade shall be used, except where specified
otherwise. All reagents shall be of sufficient purity to ensure that they do not give rise to
interferences during the GCMS analysis.
NOTE Contamination can arise from various sources e.g. plastics or rubber
materials. The use of procedural blanks and laboratory blanks assists in detecting
and identifying the source of any contamination.
4.2 Reagent water having a conductivity of < 2 mS/m, a total organic carbon content of
< 0.2 mg/l carbon, and free from organic contaminants which may interfere with the
GCMS analysis of the extracts.
NOTE Suitable water may be prepared by reverse osmosis, deionization or
distillation
4.3 Hydrochloric acid, concentrated (30 % m/V).
4.4 Hydrochloric acid solution, prepared as follows.
Slowly add (0.5 ± 0.01) l of concentrated hydrochloric acid (see 4.3) to (0.5 ± 0.01) l of
reagent water (see 4.2).
NOTE 1 Care is needed in preparing this solution which can generate heat.
NOTE 2 This solution should be replaced on a monthly basis.
4.5 Nitric acid, concentrated (65 % m/V).
65
4.6 Nitric acid solution, prepared as follows.
Slowly add (0.5 ± 0.01) l of concentrated nitric acid (see 4.5) to (0.5 ± 0.01) l of reagent
water (see 4.2).
NOTE Care is needed in preparing this solution which may generate heat.
4.7 Sulfuric acid, concentrated (98 % m/V).
4.8 Sulfuric acid solution (0.5 mol/l), Slowly add (14.0 ± 0.5) ml of sulfuric acid (see 4.7)
to (300 ± 5) ml of reagent water (see 4.2) and make up to (500 ± 5) ml with reagent water.
NOTE Care is needed in preparing this solution which may generate heat.
4.9 Chromic acid (5% m/V), prepared as follows.
Dissolve (50 ± 1) g of chromium (VI) oxide in (1 ± 0.02) l of sulfuric acid (see 4.7).
NOTE Chromic acid is a storage hazard; it may burst a sealed container due to
carbon dioxide release. It is a powerful oxidant and may give potentially explosive
reactions with oxidizable materials. It may ignite on contact with acetone or
alcohols. When heated to decomposition, it emits acrid smoke and irritating fumes.
4.10 Sodium hydroxide solution (0.5 mol/l), prepared as follows.
Dissolve (20.0 ± 0.1) g of sodium hydroxide pellets in test water and make up to 1 l.
NOTE Replace this solution on a monthly basis.
4.11 Dichloromethane, glass distilled grade.
NOTE Other grades may be suitable but it is necessary to demonstrate that any
impurities present do not interfere with the detection of compounds of interest or
the internal standards, or introduce unacceptable contamination.
4.12 Acetone, glass distilled grade.
NOTE Other grades may be suitable but it is necessary to demonstrate that any
impurities present do not interfere with the detection of compounds of interest or
the internal standards, or introduce unacceptable contamination.
4.13 Ascorbic acid solution, prepared as follows.
Dissolve (4.0 ± 0.1) g of ascorbic acid in (1.0 ± 0.01) l test water. Prior to use, extract this
solution with dichloromethane (2 × 200 ml).
NOTE 1 This solution should be replaced on a monthly basis.
4.14 Internal standards, use the following isotopically-labelled compounds:
d6-benzene;
d21-2,6-di-t-butyl-4-methylphenol (d21-BHT);
66
d5-chlorobenzene;
d34-hexadecane;
d8-naphthalene;
d10-phenanthrene;
d5-phenol;
d62-squalane;
d10-p-xylene.
4.15 Internal standards stock solutions, prepared as follows.
Make up in acetone (see 4.12) as follows:
d6-benzene
d21-BHT
d5-chlorobenzene
d34-hexadecane
d8-naphthalene
d10-phenanthrene
d5-phenol
d10-p-xylene
(2.0 ± 0.05) mg/ml;
(8.0 ± 0.20) mg/ml;
(2.0 ± 0.05) mg/ml;
(1.0 ± 0.02) mg/ml;
(1.0 ± 0.02) mg/ml;
(2.0 ± 0.05) mg/ml;
(8.0 ± 0.20) mg/ml;
(1.0 ± 0.02) mg/ml.
NOTE 1 Due to its volatility, it is difficult to make standard solutions of d6benzene by weighing; it is recommended that suitable volumes (measured using
micro-syringes) of d6-benzene, based on its density (0.950), are used.
Make up the following stock solution in dichloromethane (see 4.11):
d62-squalane
(8.0 ± 0.20) mg/ml.
NOTE 2 The stock solutions are stable for at least six months, provided they are
stored in the dark at (−18 ± 5) °C.
4.16 Internal standards intermediate solution, prepared as follows.
Place (2.5 ± 0.025) ml of the d62-squalane stock solution in a 25 ml volumetric flask.
Gently evaporate the dichloromethane until it is completely removed using nitrogen blow
down. Verify the complete removal of the dichloromethane to incipient dryness.
NOTE 1 One way of verifying this is to evaporate to constant mass.
Then place (2.5 ± 0.025) ml of each of the remaining individual standard stock solutions
(see 4.15) into the volumetric flask, ensuring that the d6-benzene stock solution is added
last, and make up to the graduation mark with acetone (see 4.12).
NOTE 2 The d6-benzene stock solution is added last in order to minimize any
potential evaporative losses of this standard.
NOTE 3 This solution is stable for at least six months provided it is stored in the
dark at (−18 ± 5) °C.
67
4.17 Internal standards GC column test solution, prepared as follows.
Add (200 ± 5) µl of the internal standards intermediate solution (see 4.16) to (8 ± 0.5) ml
of dichloromethane (see 4.14) in a 10 ml volumetric flask then make up to the graduation
mark with dichloromethane (see 4.11).
NOTE 1 To avoid potential losses of the most volatile internal standard (d6benzene) it is recommended that a syringe fitted with a needle of sufficient length
to allow the tip to be introduced below the meniscus of the dichloromethane prior
to expelling the syringe contents is used.
NOTE 2 This solution should be renewed every three months or sooner, if during
its use, an indication is obtained that the concentrations of any of the internal
standards have changed.
4.18 Internal standards spiking solution, prepared as follows.
Add (1.00 ± 0.01) ml of the internal standards intermediate solution (see 4.16) to (8 ±
0.5) ml acetone (see 4.15) in a 10 ml volumetric flask, then make up to the graduation
mark with acetone (see 4.12).
NOTE This solution should be renewed every three months, or sooner if during its
use any indication is obtained that the concentrations of any of the internal
standards have changed.
4.19 Sodium sulfate (anhydrous), prepared as follows.
Remove any organic contaminants by heating at (500 ± 50) °C for ≥ 4 h, and store so that
rehydration is minimized and re-contamination cannot occur.
5 Apparatus
5.1 Vessels, containers etc., constructed of a material, such as glass or
polytetrafluoroethylene (PTFE) that is inert under the specified test conditions.
5.2 Cleaning, in accordance with either the following procedure or with an alternative
procedure demonstrated to be equally effective in removing all detectable traces of organic
compounds from the apparatus.
Laboratory glassware shall be cleaned by washing with a biodegradable laboratory
detergent specially designed for the removal of organic materials, followed by rinsing with
hydrochloric acid solution (see 4.6) or chromic acid (see 4.9) and finally by thoroughly
rinsing with reagent water (see 4.2).
NOTE Hydrochloric acid (see 4.4) should not be used to clean any stainless steel
apparatus and equipment.
5.3 1 litre borosilicate glass bottles, of nominal capacity 1 l, fitted with either groundglass stoppers or screw-top caps with PTFE or PTFE-faced liners.
68
5.4 Concentration apparatus, which is specifically designed to allow the solvent extract
(see 3.11) to be reduced in volume from 200 ml to 50-500 µl.
NOTE Various commercial equipment (such as Kuderna-Danish or Turbovap
apparatus) may be suitable, but during this operation (which may proceed in several
steps) it is essential that losses of volatile compounds are minimized. The response
of the most volatile internal standard (d6-benzene) should be checked to ensure that
losses of this compound in the concentration step do not exceed 50 %. One method
of checking for losses is given in D.2.
5.5 Instrumental
5.5.1 Capillary gas chromatograph, with temperature gradient facility interfaced to a
mass spectrometer (see 5.5.4).
5.5.2 GC capillary column, fused silica having a length of at least 50 m, with an internal
diameter of 0.25 mm or 0.32 mm, coated with a non-polar bonded phase which performs
equivalently to dimethylsilicone gum (e.g. DB-1 or BP-1).
NOTE Other column types capable of giving equivalent or better performance
characteristics in this test may be used. Suitable validation data should be produced
to justify the substitution.
5.5.3 Carrier gas (for GCMS system), helium (99.999 % purity, or better).
5.5.4 Mass spectrometer, which is capable of operating in the electron impact ionization
mode covering the m/z range 20-650 provided that the mass spectrometric resolution (see
2.9)
is sufficient to allow unit mass resolution at the highest mass recorded with a scan
cycle of ≤ 1 s.
5.5.5 Mass spectrometry data system, which is capable of acquiring data from the mass
spectrometer under the conditions in 5.5.4, and which can produce total ion current (TIC)
chromatograms, background-subtracted averaged mass spectra and TIC peak areas. It shall
also be capable of producing hard-copy outputs of TIC chromatograms and mass spectra.
5.5.6 Mass spectral library - if a mass spectral library is not available on the mass
spectrometry data system (see 5.5.5) alternative hard-copy documents shall be available.
NOTE Examples of alternative hard-copy documents are the ‘Eight Peak Index of
Mass Spectra’[2] or the ‘Registry of Mass Spectral Data’ [3].
6 Leachate
NOTE The leachates for analysis may be prepared in accordance Annex A or the
specific requirements of the National Regulator.
6.1 Storage of test sample leachates
Start the analysis of leachate samples as soon as possible following collection. If necessary
store the leachate samples in the absence of light at (4 ± 2) °C for a maximum of 48 h
before solvent extraction is undertaken.
69
6.2 Procedural blank tests
Carry out a blank test procedure with each batch of test samples using the same test
conditions (test water, test temperature, leaching periods, type of test vessel, stoppers etc.)
as specified for the test samples (annex A) but omitting the test sample(s).
When only glass stoppers or stainless steel plates are used to seal pipe test pieces, a glass
container is suitable for the procedural blank; when other articles, such as stoppers,
connectors or sealants are used (e.g. PTFE, silicone sealant etc.) include these in the
procedural blank at the same test surface area to volume (S/V) ratio as the exposure in the
test containers.
7 Method of analysis
7.1 Extraction procedure
SAFETY NOTE During the solvent extraction step, excess pressure (which can build up
in the stoppered separating funnel) should be dissipated via the separating funnel tap when
the separating funnel is inverted. As dichloromethane vapour is hazardous, dissipation of
pressure should be undertaken with the separating funnel in a fume hood.
NOTE If any leachates have been prepared using chlorinated water the residual
chlorine should be neutralized by the addition of ascorbic acid solution (see 4.13).
Transfer (1 ±0.01) l of the leachate sample (see 7.4) to a 2 l separating funnel.
Add (100 ± 2) µl of the internal standards spiking solution (see 4.18) into the sample using
a syringe, ensuring that the tip of the syringe needle is below the surface of the sample.
Insert the stopper and swirl the contents of the separating funnel to mix.
Check the pH of the sample and adjust to (2 ± 0.2), if necessary, by dropwise addition of
either sulfuric acid solution (see 4.8) or sodium hydroxide solution (see 4.10) as
appropriate.
Add dichloromethane (100 ± 5) (see 4.11) ml to the spiked pH adjusted sample in the
separating funnel and insert the stopper. Shake the separating funnel for a total of 3 min ±
20 s. Remove the dichloromethane (lower layer) into a flask (capacity at least 250 ml).
Add a further (100 ± 5) ml of dichloromethane to the separating funnel and repeat the
extraction step. Add the second aliquot of dichloromethane to the flask in which the initial
solvent extract is stored, so that the two extracts are combined.
Dry this combined solvent extract, then transfer it to the apparatus to be used for
concentration of the extract and reduce it to a small volume (100-500 µl). Store the
concentrated extract in a freezer at (−18 ± 5) °C or below until the GCMS analysis is
carried out.
NOTE 1 Do not shake the contents of the separating funnel vigorously following
addition of the internal standards spiking solution; doing so (in the absence of the
extracting solvent, which is added later) will result in losses of the more volatile
internal standards.
70
NOTE 2 Various methods may be suitable for drying solvent extracts e.g. freezing
or addition of small amounts of sodium sulfate (see 4.19). Any of these may be
used provided that they do not affect adversely the performance of this method.
7.2 GCMS analysis
SAFETY NOTE GCMS systems typically operate from a nominal mains voltage (220240 V AC; exceptionally, some operate from a ‘3-phase’ 415 V AC supply). However,
certain parts or components of some mass spectrometers (which utilize a magnetic field for
mass resolution) may be at a very high electrical potential (up to 8 kV) relative to earth;
other mass spectrometers utilize radio-frequency radiation and DC voltages for mass
separation. Due care is necessary in the operation of GCMS systems.
7.2.1 Mass spectrometer operating parameters
Follow the manufacturer’s instructions for the mass spectrometer to set the following
parameters:
Ionization mode:
Electron energy:
Mass range:
Scan speed:
Scan mode:
electron impact (EI);
70 eV;
to include 20-650 amu;
≥ 1 scan per second;
repetitive.
7.2.2 Setting up the mass spectrometry data system
Follow the manufacturer’s instructions relating to optimizing the performance and
sensitivity of the mass spectrometer, mass calibration and data acquisition and processing.
7.2.3 Initial tuning and mass calibration of the mass spectrometer.
Follow the manufacturer’s instructions. All of the major reference peaks in the mass range
covered in the calibration table held on the MS data system shall be found in the scan(s)
used for calibration purposes.
NOTE Major reference peaks are those having an intensity > 5 % of that of the
base peak (which by convention is assigned an intensity of 100 %) of the calibrant
used.
7.2.4 Setting up the GCMS system
Install the GC column according to the manufacturers’ instructions and verify its
performance (e.g. in terms of separation number and adsorption) against the column
performance data supplied by the manufacturer.
NOTE Proprietary standard solutions are available for this purpose (see D.4).
Provided the general performance of the column is satisfactory, use the internal standards
GC column test solution (see 4.17) to establish the initial performance of the column for
the internal standards. Use the same GC temperature program for this purpose as that used
for the GCMS analysis of the concentrated solvent extracts (see 8.1).
71
Ensure the temperature programming rate does not exceed 12 °C/min at any time.
Ensure that all of the internal standards are detected on the TIC chromatogram.
Ensure the d6-benzene is separated from the solvent peak and ensure the retention time of
d62-squalane is between 35 min and 45 min.
Ensure that the asymmetry factors, As, (see 2.2) for the peaks obtained for d5-phenol and
d8-naphthalene are within the range 0.67–2.0. If this requirement is not met, investigate the
cause and correct before continuing with the analysis. If necessary, install a new GC
column.
Adjust the sensitivity of the mass spectrometer so that the mass spectra obtained for the
internal standards present at the highest level (16 ng/µl) in the internal standards GC
column test solution are not saturated.
Inspect the mass spectra obtained from the GCMS system performance test to ensure that
they correspond closely to mass spectra previously acquired for these internal standards on
the same GCMS system under identical operating conditions. Ensure that the m/z value of
the base peak is consistent, and that the intensities of all peaks having an intensity > 10 %
of the base peak do not vary by more than 30 % of their intensity when compared to
previously acquired spectra.
If the internal standards GC column test solution has not previously been analysed, analyse
it once a day on the GCMS system on five separate days to obtain typical spectra of the
internal standards.
Check the mass spectrometer mass calibration by inspecting the high mass ions
(> m/z 300) in the mass spectra for d62-squalane to ensure that they are correctly mass
measured. If this is not the case, recalibrate the mass spectrometer before continuing with
the analysis.
NOTE The requirements for the GCMS GC run-programme can be complied with
by using a GC column of length 50-60 m with an internal diameter of 0.32 mm,
coated with a bonded phase equivalent to OV-1, an initial temperature of 30 °C for
4 min, linearly programmed at 8 °C/min to a final temperature of 300 °C and
maintaining this for 20 min. Other conditions may also be suitable.
7.2.5 GCMS operating conditions for analysis of solvent extracts
Analyse concentrated solvent extracts using identical conditions to those used for checking
the performance of the GCMS system using the internal standards GC column test
solution.
Check the performance of the GCMS system at the end of every batch of concentrated
solvent extracts run, or after every sixth concentrated solvent extract if batch sizes are
greater than six.
Check the criteria in 7.2.4 to ensure that the performance of the system has not
deteriorated.
72
If the system is not in accordance with any of 7.2.4, stop the analysis, investigate and
correct the cause of the failure before continuing with the analysis.
7.2.6 Production of required outputs from the GCMS data system
Ensure the following outputs are obtained for each of the GCMS runs carried out on
concentrated solvent extracts.
–
a hard copy of the TIC trace (covering the mass range scanned);
NOTE If a solvent delay is included as part of the data acquisition, the TIC trace
will not include a peak for the solvent - this is acceptable;
–
the retention times correct to ± 1 s, or the GCMS data system scan number, of the peak
maximum of every peak detected on the TIC chromatogram, including the internal
standards;
–
the peak areas of every detected peak, including the internal standards;
–
hard copies of a mass spectrum obtained for each of the compounds detected which are
considered to originate from the sample; this shall be the best spectrum obtainable,
normally obtained by background subtraction and averaging of several mass spectra.
NOTE Compounds detected which are not considered to arise from the sample or
which are not internal standards, are included in the above requirements. However,
an indication should be given as to which of the compounds detected fall into this
category, along with their probable origin e.g. contaminants in the solvent used for
the solvent extraction, or compounds present in the test water.
8 Expression of results
8.1 Reporting of results
Tabulate the results from the GCMS analysis and include, for every peak detected on the
TIC chromatogram, provided its peak area exceeds 50 % of that of d8-naphthalene.
a) the retention time or scan number;
b) its identity;
c) its estimated concentration (expressed in µg/l);
d) the internal standard used for quantification;
e) its probable origin.
NOTE If the peak for d8-naphthalene is obscured the peak for d34-hexadecane may
be used as an alternative.
8.2 Identification of compounds detected
Use three categories to define the confidence level associated with the identities of the
detected compounds, as follows.
a) A positive identification (P) indicates that the mass spectrum and GC retention
time are the same as those obtained from a pure standard of the compound run
73
under the identical GCMS conditions on the GCMS system used to analyse the
concentrated solvent extract.
b) A tentative identification (T) indicates that a possible identity has been obtained
either from computerized library searching of a mass spectral data base, or from
manual searching of a printed mass spectral data base, or by interpretation from
first principles by a mass spectroscopist. However, a pure standard has either not
been run under identical GCMS conditions on the GCMS system used to analyse
the concentrated solvent extract or is not available.
c) An unknown (U) is any compound not covered by either of the above categories.
The four most intense peaks in the mass spectrum, in decreasing order of
intensity, with the base (100%) peak being emphasized by underlining (e.g. 147,
43, 71, 91), should be noted in the tabulation of results together with the
retention time or scan number.
NOTE Further information on identification can be found in C.1.
8.3 Quantification of compounds detected
Quantify each detected compound by comparing its response to the nearest internal
standard (in terms of GC retention time) present at 2 µg/l or 8 µg/l, with the exception of
d5-phenol, which is not used for quantification.
Use the following equation to provide the concentration of a compound D of each
compound detected in a leachate sample:
[ D] =
PD × I
PS
where:
[D]
PD
PS
I
is the concentration of a compound D (in µg/l);
is the peak area of a compound D;
is the peak area of the internal standard;
is the internal standard concentration (in µg/l).
Make no attempt to adjust [D] for extraction efficiency.
8.4 Quality assurance (QA) and quality control (QC) procedures
8.4.1 The mass calibration of the mass spectrometer
Verify on each occasion that a batch of concentrated solvent extracts is analysed. Use the
calibrant normally used for mass calibration for this purpose. Recalibrate the mass
spectrometer if any of the calibrant masses are incorrectly assigned.
8.4.2 The performance of the GCMS system
Check the performance of the GCMS system on each occasion that a batch of concentrated
solvent extracts is to be run by analysing the internal standards GC column test solution
(see 5.17). Compare the response (peak area) obtained for each internal standard to that
obtained when setting up the GCMS system (8.2.4). Provided that the peak areas are within
74
30 %, and the asymmetry factors are in accordance with 8.2.4, consider the performance
acceptable.
8.4.3 The performance of the method
Consider the performance of the method acceptable provided the following criteria are
satisfied.
a) All of the internal standards are detected in the GCMS TIC chromatogram;
b) The recoveries of the internal standards d8-naphthalene, d10-phenanthrene and
d62-squalane are > 50 %.
NOTE 1 The absence of any of the internal standards in the GCMS TIC
chromatogram indicates that either the extraction step has not been carried out
correctly, or the concentration of the solvent extract has not be carried out correctly,
or the GCMS system is not functioning correctly.
NOTE 2 A procedure for calculating the recoveries of internal standards is given
in D.3.
9 Test report
9.1 General
The test report shall include the following particulars:
— a title (e.g. “Test Report”) and the date of issue of the report;
— a reference to this British Standard, i.e. BS 6920-4;
— name and address of laboratory, and location where the tests were carried out if
different from the address of the laboratory;
— unique identification of the test report (such as serial number), and on each page an
identification in order to ensure that the page is recognized as a part of the test report,
and a clear identification of the end of the test report;
— name and address of the client placing the order;
— the name(s), function(s) and signature(s) or equivalent identification of person(s)
authorizing the test report;
9.2 Test results
Report of the GCMS analysis undertaken on the leachates, including the following:
— a copy of the TIC chromatograms (2.13) for the internal standards GC column test
solution obtained on each analytical occasion;
— a data table listing the following for each sample and GC column test mix:
the peak asymmetry values for d5-phenol, d8-nathalene, and the percentage
recoveries for d8-naphthalene, d10-phenanthrene and d62-squalane;
75
— limits of detection for the deuterated internal standards and a description of the
procedures used to obtain them;
— description and results of the method validation (method performance) for the GCMS
method;
— results from the GCMS examination of each solvent extract reported in a tabular
format, together with a copy of the TIC chromatogram for each solvent extract;
— data tables listing the following:
1) all peaks detected, including internal standards which were “spiked” at
concentrations equivalent to 1 µg/l or greater in the leachates;
2) those peaks considered not to originate from the product being tested
with an indication of their possible origins;
3) the scan number or retention time of each peak listed and the identity of
the compound:
4) the estimated concentration of each peak considered to originate from the
test material in µg/l, together with the internal standard used to derive this
estimate;
5) hose peaks which cannot be identified reported as “unknowns” with their
four major ions (in decreasing order of intensity);
6) a print-out (or copy) of the mass spectrum for each compound detected
which is considered to originate from the product being tested;
7) a description of the basis on which peaks are identified (see 8.2).
NOTE In cases where compounds detected in procedural or laboratory blanks are
also detected in solvent extracts from the leachates, if the apparent concentrations
in the blanks and extracts are low (< 2 µg/l) and do not differ by more that 25 % (of
the highest concentration), the concentrations and differences are not considered
significant and no concentrations should be indicated in the data tables. In cases
where the apparent concentrations of such compounds are lower in a solvent extract
than in the procedural or laboratory blank, no concentration should be indicated in
the data tables. When the apparent concentrations of such compounds are higher in
the solvent extracts than in the procedural blanks, and > 2 µg/l, a ‘blank subtracted’
concentration should be reported, i.e. the apparent concentration in the procedural
or laboratory blank is subtracted from the apparent concentration in the solvent
extract.
76
Annex A (informative)
Additional
procedural
details
A.1 Outline of general approach for identification of compounds detected
The data acquired during the GCMS run for each solvent extract is normally stored on the
mass spectrometry data system as a discrete data file which may be inspected either while
the run is proceeding, or after the run has been completed.
The data is usually initially displayed on a data system visual display unit (VDU) as a total
ion current (TIC) chromatogram or reconstructed ion chromatogram (RIC). Each
compound detected should appear as a peak on the TIC or RIC trace, and the mass spectra
produced by each compound can be displayed on the VDU using the appropriate
commands.
Normally, the mass spectrum initially chosen for display will be that produced when the
concentration of the compound of interest is at its maximum (i.e. at the top of the peak)
However, if it is suspected that the eluting peak is a mixture (i.e. two or more compounds
are not satisfactorily separated by the GC column), or if the mass spectrum is saturated
(due to the dynamic range of the mass spectrometer being exceeded), other spectra may be
chosen for display.
An obvious indication that a mass spectrum is saturated, or overloaded, is provided by the
presence of more than one peak in a mass spectrum at an intensity of 100 %. Mass spectra
from scans obtained before or after the intensity maximizes should be inspected to obtain a
representative mass spectrum for the compound of interest, although if a single spectrum is
chosen it should be ascertained that it is not distorted (‘skewed’).
Mass spectra may be averaged across a peak (provided it is considered that the peak is due
to a single compound) to minimize any distortion of the spectra which can occur if the
concentration of a compound entering the mass spectrometer changes significantly during
the course of a single mass spectrometer scan. This can occur when a GC peak is very
sharp e.g. only 2 s – 3 s wide. However, before averaging several spectra through a peak,
each spectrum should be checked to ascertain whether any are saturated; if any are, due
allowance should be made when assessing the resulting averaged spectrum.
A background subtraction should also be performed, either on a mass spectrum from a
single scan or on an averaged spectrum, in order to remove spurious peaks such as those
produced by residual air in the mass spectrometer, or from GC column bleed.
The mass spectrum obtained for each peak detected is generally initially inspected visually.
Depending on the experience of the mass spectroscopist, it may be possible to identify the
compound giving rise to the spectrum without recourse to reference mass spectra held in
libraries (on the data system, or in reference books).
77
If the mass spectrum is not visually recognized, a library search is usually carried out on
the data system. It is recommended that a reverse searching procedure should be used. The
closeness of the match between the unknown and the chosen library spectra is usually
expressed in terms of three parameters - fit, purity and reverse fit. However, the best match
chosen by the data system does not necessarily lead to the identification of the unknown,
and the mass spectroscopist has to apply his/her judgement, taking into account such
factors as the GC retention time, in order to decide whether the identification suggested by
the computerized library search is accepted.
If there is any doubt concerning such an identification, it should be noted as a tentative
identification and, if it is necessary to confirm the identification, a pure standard of the
compound in question should be obtained and run on the GCMS system in order to check
the mass spectrum obtained and the GC retention time. The same principles apply to
potential identifications resulting from manual inspection of mass spectral reference
collections in books such as ‘The Eight Peak Index of Mass Spectra’ [2].
If it is suspected that a TIC peak is a mixture of two or more compounds, mass
chromatography may be of use in deciding whether this is the case, and by careful choice
of mass spectra it may be possible to produce spectra corresponding to each co-eluting
component. However, where two compounds have identical retention times this may not be
possible, and further progress is dependent on the experience of the mass spectroscopist.
It is inevitable that a significant proportion of the compounds detected in many general
survey GCMS runs will only be tentatively identified, and that some will be unidentified,
as the reference collections of mass spectra currently available represent a very small
proportion (< 10 %) of the known organic compounds that are amenable to GCMS
analysis.
A.2 Checking suitability of apparatus used for concentrating solvent extracts
It is necessary to be able to reduce the volume of the dichloromethane solvent extracts
from about 200 ml to 50 µl – 500 µl without significant losses of volatile components
which may have been present in the leachate sample.
To verify that this can be satisfactorily achieved, it is recommended that a 500 µl portion
of the internal standards GC column test solution (see 4.17) is diluted to 200 ml with
dichloromethane, and the resulting solution concentrated to 500 µl, using appropriate
apparatus or equipment. This concentrate should be run on the GCMS system under
exactly the same conditions as used when using the GC column test standard solution for
checking for satisfactory GC performance, and the TIC or RIC trace compared to a TIC or
RIC trace obtained when the GC column test standard is run. Provided the loss of the most
volatile internal standard, d6-benzene, is not more than 50 % the technique used for the
concentration of the solvent extracts is considered satisfactory.
78
A.3 Procedure for calculation of recoveries of internal standards
The concentrations of the various internal standard solutions, the volume of the leachate
analysed, and the volumes injected onto the GCMS system when the final volume of the
concentrated solvent extract is 500 µl are such that the TIC chromatograms generated for
the internal standards GC column test solution (see 4.17) and the concentrated extract (see
7.1) are directly comparable, so that the following equation can be used to calculate %
recoveries:
R=
Pe × 100
Ps
where:
R
Pe
Ps
is the recovery of internal standard (in %);
is the peak area of the internal standard chosen for the comparison, in
extract;
is the peak area of the internal standard chosen for the comparison, in
standard.
A.4 Standard solutions for checking GC column performance
Several chromatography supply companies produce mixtures specifically designed to
evaluate the performance of GC columns, in terms of parameters such as column efficiency
and adsorptive or ‘active’ sites. These are sometimes referred to as ‘grob mixtures’. If the
GC column used is from a manufacturer who does not provide a suitable test
chromatogram, the column should be evaluated before use with solvent extracts of
leachates, using this type of test mixture.
A.5 Performance testing data for this protocol
A tabulated summary of the data obtained during within-laboratory and inter-laboratory
performance testing of the analytical procedures described in this part of BS 6920 is given
in Tables C.1 and C.2. Competent laboratories intending to use these procedures should be
able to produce comparable data.
79
Table D.1 — Within-laboratory performance data for internal standards
Sample1)
Data
system
code
0030S021
0030S031
0030S041
0030S051
0030S061
0030S071
0030S091
0030S101
0030S121
0030S131
0030S141
0030S161
0030S171
0030S181
0030S191
0030S201
d6-benzene
d5-chlorobenzene
d10-p-xylene
d5-phenol
Peak area
d8-naphthalene
d21-BHT2)
d34-hexadecane
d10-phenanthrene
d62-squalane
PE-TW (1)
7211
8522
2981
9844
2739
56153
2768
14583
90672
PE-TW-Cl (1)
5588
10633
1815
10635
3700
56423
4668
19328
100599
GRP-TW (1)
5754
9923
3685
13097
2753
61581
3160
16134
99354
GRP-TW-CL (1)
8994
10401
3820
14973
3369
57335
4567
21121
101328
BIT-TW (1)
5603
6327
1318
7156
2302
48171
3638
16048
85730
BIT-TW-CL (1)
6916
11074
2969
9627
4402
57669
4469
16901
54348
PB-TW (1)
12190
13005
2887
10916
4029
65931
4728
15546
102366
PB-TW-CL (1)
9215
10259
2671
8351
3744
48374
4523
14240
78070
PE-TW (2)
6498
9329
3311
7335
3868
53707
3531
14276
85766
PE-TW-CL (2)
4665
9183
3188
6197
2929
62658
4435
17728
113155
GRP-TW (2)
5766
7986
2087
12357
2531
50362
5212
16965
87439
GRP-TW-CL (2)
5410
11762
3473
13804
2375
61799
6283
20480
100379
BIT-TW (2)
4714
6008
4194
9691
2534
39768
2037
13363
36799
BIT-TW-CL (2)
5702
9184
3345
14796
2609
47853
3479
15293
54239
PB-TW (2)
7207
11073
3163
6846
2765
49139
4875
14024
88095
PB-TW-CL (2)
6444
7862
2702
7097
2175
54553
6013
15490
85893
Mean
6742
9533
3038
10170
3052
54467
4274
16345
85265
SD
1955
1890
681
2934
695
6882
1125
2311
20579
% RSD
29 %
20 %
22 %
29 %
23 %
13 %
26 %
14 %
24 %
1)
Samples referred to as follows: PE = polyethylene; GRP = glass-reinforced polyester; BIT = bitumen lined ductile iron; PB = procedural blank; TW = test water; TW-CL = chlorinted test water; (1) = batch 1; (2) =
batch 2
2)
BHT = 2,6-di-t-butyl-4-methylphenol
80
Table D.2 — Summary of variation (% RSD) for internal standards in interlaboratory performance testing1)
% RSD for peak areas of internal standards
d5-chlorobenzene
d10-p-xylene
d5-phenol
d8-naphthalene
d21-BHT
1
1
15
37
24
18
23
2
19
27
51
42
16
2
1
19
26
24
21
11
2
20
19
35
19
14
32)
1
15
68
32
27
36
2
35
27
63
44
53
4
½
10
8
12
12
14
1)
Each laboratory analysed leachates from three samples in duplicate, together with procedural blanks. Each batch consisted of eight analyses.
2)
Laboratory 3 did not follow instructions regarding conditions for GCMS analysis
Laboratory
Batch number
d6-benzene
16
25
30
15
81
d34-hexadecane
52
21
19
32
24
84
15
d10-phenanthrene
15
19
14
15
10
36
9
d62-squalane
26
21
19
30
33
96
47
Bibliography
BS 6068-2.26:1986, Water quality — Part 2: Physical, chemical and biochemical methods
— Section 2.26: Method for determination of free chlorine and total chlorine: colorimetric
method using N,N-diethyl-1,4-phenylenediamine, for routine control purposes.
BS 6920-2.1:1996, Suitability of non-metallic products for use in contact with water
intended for human consumption with regard to their effect on the quality of the water —
Part 2: Methods of test — Section 2.1: Samples for testing.
BS EN ISO 3696:1995, Water for analytical laboratory use — Specification and test
methods.
Other publications
[1] The Water Supply (Water Quality) Regulations 1989. Statutory Instrument 1989 No.
1147 (and amendments SI 1989 No. 1384, SI 1991 No. 1837). London: HMSO
[2] The Eight Peak Index of Mass Spectra, 4th Edition, 1991. ISBN 0 85 186417 1,
Nottingham: The Royal Society of Chemistry.
[3] Registry of Mass Spectral Data, 3rd Edition, 1989. Eds. F.W. McLafferty and
D.B. Stauffer. ISBN 0 471 62886 7. New York: John Wiley & Sons.
82
Appendix 2 – Proposed
BS6920: Part 4
83
modified
version
of
The GC-MS identification of water leachable organic
substances from materials in contact with water intended
for human consumption
84
Foreword
This CEN Standard has been prepared by CEN TC 164 WG3 … based on the co-normative
research project EVK1-CT 2000-00052.
This Standard describes methods of identification only, and should not be used or quoted
as a specification. References to this standard should indicate that the methods of
identification used are in accordance with EN ….
Contents
1 Introduction
Chemicals that leach from materials used in contact with public water supplies can cause
health concerns for consumers. Potential health effects of these chemicals are assessed in
three stages:
1. preparation of leachates by exposing a portion of the material to water under controlled
conditions
2. analysis of the leachates
3. toxicological evaluation of the concentrations of substances identified
Analysis of organic substances present in the leachates usually involves two types of
analytical methods:
•
a screening method which enables a qualitative and semi-quantitative assessment to be
made of unspecified organic compounds;
• accurate quantitative methods for the determination of specific target compounds
known to be present in the chemical formulations of the materials.
This standard describes the analytical procedures based upon gas chromatography and
mass spectrometry (GCMS) used to screen leachates for unspecified organic compounds
derived from finished products such as pipes, protective coatings, membranes etc.
This method is suitable for leachates from all materials that may be used in contact with
water for human consumption, and which may be the subject of an application for approval
by the national regulator.
NOTE 1 The results of these tests are assessed by reference to existing
toxicological data concerning the chemicals identified. Where little or no
information exists, it may be necessary for the substance itself or the material
leachate to be submitted for toxicity testing.
2 Scope
This standard describes a method for identifying organic chemicals which are amenable to
GCMS analysis using the techniques described and which may leach from a product into
water intended for human consumption. A method of calculating the concentrations of the
organic substances identified is also provided.
85
This standard is not applicable to the toxicological evaluation of chemicals.
NOTE The method to be used for the preparation of leachates is specified by
separate CEN standards.
3 Normative references
EN ISO 3696:1995, Water for analytical laboratory use — Specification and test methods.
PrEN 12873 all parts
4 Terms and definitions
For the purposes of this standard, the following terms and definitions apply.
4.1 amu
atomic mass unit; defined as 1/12 of the mass of a single atom of carbon-12 in the gas phase
(i.e. unbound), at rest and in its ground state
4.2 asymmetry factor
(As), measure of the adsorption of a compound during gas chromatographic analysis. It
may be derived from the equation
As = (a+b)/2b
where:
is the distance from the leading edge of the peak at the point on the baseline at
which a perpendicular dropped from the peak maximum crosses it;
is the corresponding distance from the trailing edge of the peak.
a
b
NOTE Some manufacturer’s GCMS software packages allow the calculation of
peak asymmetries to be produced automatically.
4.3 electron impact ionization
ionization by a beam of electrons
4.4 GCMS
analytical instrument comprising a gas chromatograph (GC) linked to a mass spectrometer
(MS)
4.5 GCMS general survey analysis
acquisition of a series of mass spectra (up to several thousand) during the course of a gas
chromatographic run, by operating the mass spectrometer in a continuous cyclic scanning
mode over a wide m/z range
86
4.6 injection standard
organic compound added to the final solvent extract prior to analysis
NOTE An injection standard is added to calculate proper recoveries of the internal
standards
4.7 internal standards
organic compounds added to the leachate at a known concentration prior to the
commencement of the analysis
NOTE Internal standards are added for the following reasons (a) to demonstrate
that the analysis has been undertaken successfully, and (b) to provide a reference to
allow other substances detected to be quantified. Ideally, the internal standards
should not be present in the leachate; for this reason, isotopically labelled standards
are preferred.
4.8 laboratory blank
water sample known to contain negligible levels of contamination, to which internal
standards have been added and which is then analysed in the same way as the leachate
NOTE Laboratory blanks are used to check for potential contamination of either
leachates or solvent extracts which may occur within the laboratory during analysis.
4.9 leachate
aqueous solution that results from leaving test water in contact with the test material under
the specified test conditions
4.10 procedural blank
aqueous solution that results from leaving test water in contact with tanks or containers
identical to those used to prepare the leachate in the absence of the test material.
NOTE Procedural blanks are used to check for potential contamination of leachates
which may arise during the leaching procedure. For example the tanks or containers
used may themselves leach compounds into the test water, or aerial contamination
may occurr if volatile compounds are present in the laboratory atmosphere. Further
details are provided in the CEN standard regrading the preparation of leachates.
4.11 mass spectrometric resolution
measure of the capability of the mass spectrometer to correctly distinguish two mass
spectral peaks, having similar mass to charge m/z values, as separate peaks; when z =1, this
is generally denoted by m2/(m2 − m1) where m2 has the higher m/z value and m1 has the
lower m/z value
87
NOTE A mass spectrometer set up so that the resolution is 650 will satisfactorily
resolve and assign the correct masses to mass spectral peaks at m/z 649 and
m/z 650.
4.12 m/z
mass-to-charge ratio of an ion
NOTE As most ions produced by electron impact ionisation are singly charged,
this ratio usually corresponds to the mass of an ion. However, exceptionally, ions
may possess multiple charges.
4.13 solvent extract
solution containing compounds partitioned from the leachate into the extraction solvent (in
this case dichloromethane)
4.14 total ion current
(TIC), sum of all the separate ion currents carried by the individual ions contributing to a
mass spectrum
4.15 TIC chromatogram
graphical representation of the TIC versus time
5 Principle
A mixture of internal standards is added to each of the test leachates prior to solvent
extraction with dichloromethane. The solvent extract is concentrated and analysed by
GCMS to determine the identity and approximate concentrations of organic chemicals that
may be present.
The mass spectrometer is used in a repetitive full-scan mode and the mass spectra
produced are recorded by, and stored on, the GCMS data system.
Wherever possible, each compound detected is identified and may be quantified by
reference to the responses obtained for the internal standards.
NOTE 1 The methods used to identify organic compounds from their mass spectra
do not form part of this method, but further information on this subject is given in
annex D.
6 Reagents
6.1 General, only reagents of analytical grade shall be used, except where specified
otherwise. All reagents shall be of sufficient purity to ensure that they do not give rise to
interferences during the GCMS analysis.
88
NOTE Contamination can arise from various sources e.g. plastics or rubber
materials. The use of procedural blanks and laboratory blanks assists in
detecting and identifying the source of any contamination.
6.2 Reagent water, having a conductivity of < 2 mS/m, a total organic carbon content of
< 0.2 mg/l carbon, and free from organic contaminants which may interfere with the
GCMS analysis of the extracts.
NOTE Suitable water may be prepared by reverse osmosis, deionization or
distillation
6.3 Hydrochloric acid, concentrated (30 % m/V).
6.4 Hydrochloric acid solution prepared as follows.
Slowly add (0.5 ± 0.01) l of concentrated hydrochloric acid (see 6.3) to (0.5 ± 0.01) l of
reagent water (see 6.2).
NOTE 1 Care is needed in preparing this solution which can generate heat.
NOTE 2 This solution should be replaced on a monthly basis.
6.5 Sulfuric acid concentrated (98 % m/V).
6.6 Sulfuric acid solution (0.5 mol/l) Slowly add (14.0 ± 0.5) ml of sulfuric acid (see 5.5)
to (300 ± 5) ml of reagent water (see 5.2) and make up to (500 ± 5) ml with reagent water.
NOTE Care is needed in preparing this solution which may generate heat.
6.7 Sodium hydroxide solution (0.5 mol/l) prepared as follows.
Dissolve (20.0 ± 0.1) g of sodium hydroxide pellets in reagent water (5.2) and make up to
1 l.
NOTE Replace this solution on a monthly basis.
6.8 Dichloromethane glass distilled grade.
NOTE Other grades may be suitable but it is necessary to demonstrate that any
impurities present do not interfere with the detection of compounds of interest or
the internal standards, or introduce unacceptable contamination.
6.9 Acetone glass distilled grade.
NOTE Other grades may be suitable but it is necessary to demonstrate that any
impurities present do not interfere with the detection of compounds of interest or
the internal standards, or introduce unacceptable contamination.
6.10 Ascorbic acid solution prepared as follows.
Dissolve (4.0 ± 0.1) g of ascorbic acid in (1.0 ± 0.01) l test water. Prior to use, extract this
solution with dichloromethane (2 × 200 ml).
89
NOTE 1 This solution should be replaced on a monthly basis.
6.11 Internal standards use the following isotopically-labelled compounds:
d6-benzene;
d21-2,6-di-t-butyl-4-methylphenol (d21-BHT);
d5-chlorobenzene;
d34-hexadecane;
d8-naphthalene;
d10-phenanthrene;
d5-phenol;
d62-squalane;
d10-p-xylene.
6.12 Internal standards stock solutions prepared as follows.
Make up in acetone (see 6.9) as follows:
d6-benzene
d21-BHT
d5-chlorobenzene
d34-hexadecane
d8-naphthalene
d10-phenanthrene
d5-phenol
d10-p-xylene
(2.0 ± 0.05) mg/ml;
(8.0 ± 0.20) mg/ml;
(2.0 ± 0.05) mg/ml;
(1.0 ± 0.02) mg/ml;
(1.0 ± 0.02) mg/ml;
(2.0 ± 0.05) mg/ml;
(8.0 ± 0.20) mg/ml;
(1.0 ± 0.02) mg/ml.
NOTE 1 Due to its volatility, it is difficult to make standard solutions of d6benzene by weighing; it is recommended that suitable volumes (measured using
micro-syringes) of d6-benzene, based on its density (0.950), are used.
Make up the following stock solution in dichloromethane (see 6.8):
d62-squalane
(8.0 ± 0.20) mg/ml.
NOTE 2 The stock solutions are stable for at least six months, provided they are
stored in the dark at (−18 ± 5) °C.
6.13 Internal standards intermediate solution prepared as follows.
Place (2.5 ± 0.025) ml of the d62-squalane stock solution in a 25 ml volumetric flask.
Gently evaporate the dichloromethane until it is completely removed using nitrogen blow
down. Verify the complete removal of the dichloromethane to incipient dryness.
NOTE 1 One way of verifying this is to evaporate to constant mass.
Then place (2.5 ± 0.025) ml of each of the remaining individual standard stock solutions
(see 6.12) into the volumetric flask, ensuring that the d6-benzene stock solution is added
last, and make up to the graduation mark with acetone (see 6.9).
90
NOTE 2 The d6-benzene stock solution is added last in order to minimize any
potential evaporative losses of this standard.
NOTE 3 This solution is stable for at least six months provided it is stored in the
dark at (−18 ± 5) °C.
6.14 Internal standards GC column test solution prepared as follows.
Add (200 ± 5) µl of the internal standards intermediate solution (see 6.13) to (8 ± 0.5) ml
of dichloromethane in a 10 ml volumetric flask then make up to the graduation mark with
dichloromethane .
NOTE 1 To avoid potential losses of the most volatile internal standard (d6benzene) it is recommended that a syringe fitted with a needle of sufficient length
to allow the tip to be introduced below the meniscus of the dichloromethane prior
to expelling the syringe contents is used.
NOTE 2 This solution should be renewed every three months or sooner, if during
its use, an indication is obtained that the concentrations of any of the internal
standards have changed.
6.15 Internal standards spiking solution prepared as follows.
Add (1.00 ± 0.01) ml of the internal standards intermediate solution (see 6.13) to (8 ±
0.5) ml acetone in a 10 ml volumetric flask, then make up to the graduation mark with
acetone .
NOTE This solution should be renewed every three months, or sooner if during its
use any indication is obtained that the concentrations of any of the internal
standards have changed.
6.16 Sodium sulfate (anhydrous) prepared as follows.
Remove any organic contaminants by heating at (500 ± 50) °C for ≥ 4 h, and store so that
rehydration is minimized and re-contamination cannot occur.
6.17 Injection standard, use 1,2,3,4-tetrachloronaphthalene.
6.18 Injection standard stock solution prepared as follows.
Weigh 25.00 ±0.02 mg 1,2,3,4 tetrachloronaphthalene and add to approximately 20 ml
dichloromethane contained in a 25 ml volumetric flask. Shake to dissolve and then make
up to the graduation mark with dichloromethane. This solution contains 1.00 mg/ml and is
stable for 3 months provided it is stored in a freezer at -18°±4°C.
6.19 Injection standard spiking solution prepared as follows.
Take 1.00 ml of the injection standard stock solution and add to approximately 5 ml of
dichloromethane contained in a 10 ml volumetric flask. Make up to the graduation mark
with additional dichloromethane. This solution contains 100 ng/µl and is stable for at least
3 months provided it is stored in a freezer at -18°±4°C.
91
7 Apparatus
7.1 Vessels, containers etc. constructed of a material, such as glass or
polytetrafluoroethylene (PTFE) that is inert under the specified test conditions.
7.2 Cleaning in accordance with either the following procedure or with an alternative
procedure demonstrated to be equally effective in removing all detectable traces of organic
compounds from the apparatus.
Laboratory glassware shall be cleaned by washing with a biodegradable laboratory
detergent specially designed for the removal of organic materials, followed by rinsing with
hydrochloric acid solution (see 6.4) and finally by thoroughly rinsing with reagent water
(see 6.2).
NOTE Hydrochloric acid (see 6.4) should not be used to clean any stainless steel
apparatus and equipment.
7.3 [check should be in 12873]Concentration apparatus which is specifically designed to
allow the solvent extract (see 4.13) to be reduced in volume from 200 ml to 50-500 µl.
NOTE Various commercial equipment (such as Kuderna-Danish or Turbovap
apparatus) may be suitable, but during this operation (which may proceed in several
steps) it is essential that losses of volatile compounds are minimized. The response
of the most volatile internal standard (d6-benzene) should be checked to ensure that
losses of this compound in the concentration step do not exceed 50 %. One method
of checking for losses is given in A.2.
7.4 Instrumental
7.4.1 Capillary gas chromatograph, with temperature gradient facility interfaced to a
mass spectrometer (see 7.4.4).
7.4.2 GC capillary column, fused silica coated with a non-polar bonded phase which
performs equivalently to dimethylsilicone gum (e.g. DB-1 or BP-1).
NOTE Other column types capable of giving equivalent or better performance
characteristics in this test may be used. Suitable validation data should be produced
to justify the substitution.
7.4.3 Carrier gas (for GCMS system), helium (99.999 % purity, or better).
7.4.4 Mass spectrometer, which is capable of operating in the electron impact ionization
mode covering the m/z range 20-650 provided that the mass spectrometric resolution (see
4.11) is sufficient to allow unit mass resolution at the highest mass recorded with a scan
cycle of ≤ 1 s.
7.4.5 Mass spectrometry data system, which is capable of acquiring data from the mass
spectrometer under the conditions in 7.4.4, and which can produce total ion current (TIC)
chromatograms, background-subtracted averaged mass spectra and TIC peak areas. It shall
also be capable of producing hard-copy outputs of TIC chromatograms and mass spectra.
92
7.4.6 Mass spectral library - if a mass spectral library is not available on the mass
spectrometry data system (see 7.4.5) alternative hard-copy documents shall be available.
NOTE Examples of alternative hard-copy documents are the ‘Eight Peak Index of
Mass Spectra’[2] or the ‘Registry of Mass Spectral Data’ [3].
8 Leachate
NOTE The leachates for analysis should be prepared in accordance with
appropriate CEN standards.
8.1 Storage of leachates and procedural blanks [check terminology with 12873] and
laboratory blanks
Start the analysis of leachate samples as soon as possible following collection. If necessary
store the leachate samples procedural blanks and laboratory blanks under the same
conditions, i.e. in the absence of light at (4 ± 2) °C for a maximum of 7 days [there is a ISO
standard for storing samples, check] before solvent extraction is undertaken.
9 Method of analysis
9.1 Extraction procedure
SAFETY NOTE During the solvent extraction step, excess pressure (which can build up
in the stoppered separating funnel) should be dissipated via the separating funnel tap when
the separating funnel is inverted. As dichloromethane vapour is hazardous, dissipation of
pressure should be undertaken with the separating funnel in a fume hood.
NOTE If any leachates have been prepared using chlorinated water the residual
chlorine should be neutralized by the addition of ascorbic acid solution (see 5.10).
Transfer (1 ±0.01) l of the leachate sample to a 2 l separating funnel.
Add (100 ± 2) µl of the internal standards spiking solution (see 6.15) into the sample using
a syringe, ensuring that the tip of the syringe needle is below the surface of the sample.
Insert the stopper and swirl the contents of the separating funnel to mix.
Check the pH of the sample and adjust to (2 ± 0.2), if necessary, by dropwise addition of
either sulfuric acid solution (see 6.6) or sodium hydroxide solution (see 6.7) as appropriate.
Add dichloromethane (100 ± 5) (see 6.8) ml to the spiked pH adjusted sample in the
separating funnel and insert the stopper. Shake the separating funnel for a total of 3 min ±
20 s. Remove the dichloromethane (lower layer) into a flask (capacity at least 250 ml).
Add a further (100 ± 5) ml of dichloromethane to the separating funnel and repeat the
extraction step. Add the second aliquot of dichloromethane to the flask in which the initial
solvent extract is stored, so that the two extracts are combined.
Change the pH of the sample to (10 ± 0.2) by addition of sodium hydroxide solution (see
6.7).
93
Add a further (100 ± 5) ml of dichloromethane (see 6.8) to the separating funnel and
repeat the extraction step. Add the aliquot of dichloromethane to the flask in which the
initial two solvent extracts at pH of 2 is stored, so that the extracts are combined. Add a
last (100 ± 5) ml of dichloromethane to the separating funnel, repeat the extraction step
and add the aliquot of dichloromethane to the other three.
Dry this combined solvent extract, then transfer it to the apparatus to be used for
concentration of the extract and reduce it to a small volume (100-500 µl). Store the
concentrated extract in a freezer at (−18 ± 5) °C or below until the GCMS analysis is
carried out.
NOTE 1 Do not shake the contents of the separating funnel vigorously following
addition of the internal standards spiking solution; doing so (in the absence of the
extracting solvent, which is added later) will result in losses of the more volatile
internal standards.
NOTE 2 Various methods may be suitable for drying solvent extracts e.g. freezing
or addition of small amounts of sodium sulfate (see 6.16). Any of these may be
used provided that they do not affect adversely the performance of this method.
9.2 GCMS analysis
SAFETY NOTE GCMS systems typically operate from a nominal mains voltage (220240 V AC; exceptionally, some operate from a ‘3-phase’ 415 V AC supply). However,
certain parts or components of some mass spectrometers (which utilize a magnetic field for
mass resolution) may be at a very high electrical potential (up to 8 kV) relative to earth;
other mass spectrometers utilize radio-frequency radiation and DC voltages for mass
separation. Due care is necessary in the operation of GCMS systems.
9.2.1 Mass spectrometer operating parameters
Follow the manufacturer’s instructions for the mass spectrometer to set the following
parameters:
Ionization mode:
Electron energy:
Mass range:
Scan speed:
Scan mode:
electron impact (EI);
70 eV;
minimum 35-650 amu;
≥ 1 scan per second;
repetitive.
9.2.2 Setting up the mass spectrometer and data system
Follow the manufacturer’s instructions relating to optimizing the performance and
sensitivity of the mass spectrometer, mass calibration and data acquisition and processing.
9.2.3 Initial tuning and mass calibration of the mass spectrometer
Follow the manufacturer’s instructions. All of the major reference peaks in the mass range
covered in the calibration table held on the MS data system shall be found in the scan(s)
used for calibration purposes.
94
NOTE Major reference peaks are those having an intensity > 5 % of that of the
base peak (which by convention is assigned an intensity of 100 %) of the calibrant
used.
9.3 Setting up the GCMS system
Install the GC column according to the manufacturers’ instructions and verify its
performance (e.g. in terms of separation number and adsorption) against the column
performance data supplied by the manufacturer.
NOTE Proprietary standard solutions are available for this purpose (see A.4).
Provided the general performance of the column is satisfactory, use the internal standards
GC column test solution (see 6.15) to establish the initial performance of the column for
the internal standards. Use the same GC temperature program for this purpose as that used
for the GCMS analysis of the concentrated solvent extracts (see 9.1).
Ensure that:
1. the temperature programming rate does not exceed 12 °C/min at any time.
2. all of the internal standards are detected on the TIC chromatogram.
3. d6-benzene is separated from the solvent peak and that the retention time of
d62-squalane is between 35 min and 45 min.
4. the asymmetry factors, As, (see 4.2) for the peaks obtained for d5-phenol and
d8-naphthalene are within the range 0.67–2.0 [broaden slightly to 0.5-3?]. If this
requirement is not met, investigate the cause and correct before continuing with the
analysis. If necessary, install a new GC column.
5. the mass spectra obtained for the internal standards present at the highest level (16 ng/µ
l) in the internal standards GC column test solution are not saturated by adjusting the
sensitivity of the mass spectrometer.
6. the mass spectra obtained from the GCMS system performance test correspond closely
to mass spectra previously acquired for these internal standards on the same GCMS
system under identical operating conditions.
7. the m/z value of the base peak is consistent, and that the intensities of all peaks having
an intensity > 10 % of the base peak do not vary by more than 30 % of their intensity
when compared to previously acquired spectra.
8. the high mass ions (> m/z 300) in the mass spectra for d62-squalane are correctly mass
measured after the mass spectrometer mass calibration. If this is not the case, recalibrate
the mass spectrometer before continuing with the analysis.
If the internal standards GC column test solution has not previously been analysed, analyse
it once a day on the GCMS system on five separate days to obtain typical spectra of the
internal standards.
NOTE The requirements for the GCMS GC run-programme can be complied with
by using a GC column of length 50-60 m with an internal diameter of 0.32 mm,
coated with a bonded phase equivalent to OV-1, an initial temperature of 30 °C for
4 min, linearly programmed at 8 °C/min to a final temperature of 300 °C and
maintaining this for 20 min. Other conditions may also be suitable.
95
9.4 GCMS operating conditions for analysis of solvent extracts
Analyse concentrated solvent extracts using identical conditions to those used for checking
the performance of the GCMS system using the internal standards GC column test
solution.
Check the performance of the GCMS system at the end of every batch of concentrated
solvent extracts run, or regularly during the batch, e.g. after every sixth concentrated
solvent extract if batch sizes are greater than six.
Check the criteria in 9.2.4 to ensure that the performance of the system has not
deteriorated.
If the system is not in accordance with any of 9.2.4, stop the analysis, investigate and
correct the cause of the failure before continuing with the analysis.
9.5 Production of required outputs from the GCMS data system
Ensure the following outputs are obtained for each of the GCMS runs carried out on
concentrated solvent extracts.
•
a hard copy of the TIC trace (covering the mass range scanned);
NOTE If a solvent delay is included as part of the data acquisition, the TIC trace
will not include a peak for the solvent - this is acceptable;
•
the retention times correct to ± 1 s of the peak maximum of every peak detected on the
TIC chromatogram, including the internal standards;
• the peak areas of every detected peak, including the internal standards;
• hard copies of a mass spectrum obtained for each of the compounds detected which are
considered to originate from the sample; this shall be the best spectrum obtainable,
normally obtained by background subtraction and averaging of several mass spectra.
NOTE Compounds detected which are not considered to arise from the sample or
which are not internal standards, are included in the above requirements. However,
an indication should be given as to which of the compounds detected fall into this
category, along with their probable origin e.g. contaminants in the solvent used for
the solvent extraction, or compounds present in the test water.
10 Expression of results
10.1 Reporting of results
Tabulate the results from the GCMS analysis and include the following information, for
every peak detected on the TIC chromatograms obtained for extracts of the leachates,
procedural blanks and laboratory blank which are present at ≥ 1.0 µg/l:, .
1. retention time
2. relative retention time with respect to d8-naphthalene
3. compound name for tentative and positive identifications; in the case of unknown
compounds the m/z values of the 4 most intense masses in decreasing order of intensity
shall be listed
96
4. confirmation level, using the following abbreviations – P(=positive), T(=tentative) or
U(=unknown) (see 10.2)
5. peak area
6. CAS number for positive identifications
7. estimated concentration (expressed in µg/l to one place of decimals) for the leachates
and blanks; the concentrations reported for the leachates should be corrected (i.e. blank
subtracted) with respect to the concentrations detected in the procedural blanks
8. for those compounds detected in both the leachates and the procedural blanks, calculate
and report the ratio of non-corrected concentrations in leachates and procedural blanks
for the leachates results
9. internal standard used for quantification
10. probable origin
NOTE If the ratio of the concentration of a compound detected in both a leachate
and the corresponding procedural blank is <1, this indicates that the compound was
detected at a higher concentration in the procedural blank than in the leachate; in
this case the blank-subtracted concentration will be a negative number and should
be reported as 0.0
9.2 Identification of compounds detected
Use three categories to define the confidence level associated with the identities of the
detected compounds, as follows.
1. A positive identification (P) indicates that the mass spectrum and GC retention time are
the same as those obtained from a pure standard of the compound run under the
identical GCMS conditions on the GCMS system used to analyse the concentrated
solvent extract.
2. A tentative identification (T) indicates that a possible identity has been obtained either
from computerized library searching of a mass spectral data base, or from manual
searching of a printed mass spectral data base, or by interpretation from first principles
by a mass spectroscopist [reference to CEN/ISO document Theo circulated at the start
of the project; Theo will check reference or send electronic file]. However, a pure
standard has either not been run under identical GCMS conditions on the GCMS system
used to analyse the concentrated solvent extract or is not available.
3. An unknown (U) is any compound not covered by either of the above categories. The
four most intense peaks in the mass spectrum, in decreasing order of intensity (e.g. 147,
43, 71, 91), should be noted in the tabulation of results.
NOTE Further information on identification can be found in A.1
9.3 Quantification of compounds detected
Quantify each detected compound by comparing its response to the nearest, in terms of GC
retention time, internal standard. Only those internal standards present at concentrations ≥2
µg/l (with the exception of d5-phenol and d6-benzene; i.e. d5-chlorobenzene, d21-BHT,
d10-phenanthrene or d62-squalane. are used for this purpose.
Use the following equation to provide the concentration of a compound D of each
compound detected in a leachate sample:
97
[ D] =
PD × I
PS
where:
[D]
PD
PS
I
is the concentration of a compound D (in µg/l);
is the peak area of a compound D;
is the peak area of the internal standard;
is the internal standard concentration (in µg/l).
NOTE 1 In cases where an internal standard used for quantification co-elutes with
another compound, the next nearest internal standard has to be used.
NOTE 2 The concentration calculated on the basis of an internal standard assumes
that both the compound and the internal standard have equal losses during leachate
extraction and extract preparation and have an identical response on the GCMS TIC
trace. Adjustment of concentration [D] for any of these factors is not necessary.
10 Quality assurance (QA) and quality control (QC) procedures
10.1 The mass calibration of the mass spectrometer
Verify on each occasion that a batch of concentrated solvent extracts is analysed. Use the
calibrant normally used for mass calibration for this purpose. Recalibrate the mass
spectrometer if any of the calibrant masses are incorrectly assigned.
10.2 The performance of the GCMS system
Check the performance of the GCMS system on each occasion that a batch of concentrated
solvent extracts is to be run by analysing the internal standards GC column test solution
(see 6.14). Compare the response (peak area) obtained for each internal standard to that
obtained when setting up the GCMS system (9.2.4). Provided that the peak areas are within
30 %, and the asymmetry factors are in accordance with 9.2.4, consider the performance
acceptable.
10.3 The performance of the method
Consider the performance of the method acceptable provided the following criteria are
satisfied.
•
All of the internal standards are detected in the GCMS TIC chromatogram;
The recoveries of the internal standards d8-naphthalene, d10-phenanthrene and
d62-squalane are > 50 %.
NOTE 1 The absence of any of the internal standards in the GCMS TIC
chromatogram indicates that either the extraction step has not been carried out
correctly, or the concentration of the solvent extract has not be carried out correctly,
or the GCMS system is not functioning correctly.
98
NOTE 2 A procedure for calculating the recoveries of internal standards is given
in A.3.
11 Test report
11.1 General
The test report shall include the following particulars:
•
•
•
a title (e.g. “Test Report”) and the date of issue of the report
a reference to this standard
name and address of laboratory, and location where the tests were carried out if
different from the address of the laboratory
• unique identification of the test report (such as serial number), and on each page
an identification in order to ensure that the page is recognized as a part of the
test report, and a clear identification of the end of the test report;
• name and address of the client placing the order;
11.2 Test results
Report of the GCMS analysis undertaken on the leachates, including the following:
•
•
•
•
•
•
a copy of the TIC chromatograms (3.15) for the internal standards GC column test
solution obtained on each analytical occasion;
a data table listing the following for each sample and GC column test mix:
! the peak asymmetry values for d5-phenol, d8-nathalene, and the
percentage recoveries for d8-naphthalene, d10-phenanthrene and d62squalane;
limits of detection for the deuterated internal standards and a description of the
procedures used to obtain them;
description and results of the method validation (method performance) for the GCMS
method;
results from the GCMS examination of each solvent extract reported in a tabular
format, together with a copy of the TIC chromatogram for each solvent extract;
data tables listing the following:
! all peaks detected, including internal standards, which were “spiked”
or calculated to be present at concentrations equivalent to 1 µg/l or
greater in the leachates
! those peaks considered not to originate from the product being tested
with an indication of their possible origins
! retention time of each peak listed and the identity of the compound
! the calculated concentration of each peak in µg/l, together with the
internal standard used to derive this estimate and an indication of the
origin of the compound
! those peaks which cannot be identified reported as “unknowns” with
their four major ions (in decreasing order of intensity)
! a print-out (or copy) of the mass spectrum for each compound
detected which is considered to originate from the product being
tested
! a description of the basis on which peaks are identified (see 10.2)
99
Annex B (informative)
Additional
procedural
details
B.1 Outline of general approach for identification of compounds detected
The data acquired during the GCMS run for each solvent extract is normally stored on the
mass spectrometry data system as a discrete data file which may be inspected either while
the run is proceeding, or after the run has been completed.
The data is usually initially displayed on a data system visual display unit (VDU) as a total
ion current (TIC) chromatogram or reconstructed ion chromatogram (RIC). Each
compound detected should appear as a peak on the TIC or RIC trace, and the mass spectra
produced by each compound can be displayed on the VDU using the appropriate
commands.
Normally, the mass spectrum initially chosen for display will be that produced when the
concentration of the compound of interest is at its maximum (i.e. at the top of the peak)
However, if it is suspected that the eluting peak is a mixture (i.e. two or more compounds
are not satisfactorily separated by the GC column), or if the mass spectrum is saturated
(due to the dynamic range of the mass spectrometer being exceeded), other spectra may be
chosen for display.
An obvious indication that a mass spectrum is saturated, or overloaded, is provided by the
presence of more than one peak in a mass spectrum at an intensity of 100 %. Mass spectra
from scans obtained before or after the intensity maximizes should be inspected to obtain a
representative mass spectrum for the compound of interest, although if a single spectrum is
chosen it should be ascertained that it is not distorted (‘skewed’).
Mass spectra may be averaged across a peak (provided it is considered that the peak is due
to a single compound) to minimize any distortion of the spectra which can occur if the
concentration of a compound entering the mass spectrometer changes significantly during
the course of a single mass spectrometer scan. This can occur when a GC peak is very
sharp e.g. only 2 s – 3 s wide. However, before averaging several spectra through a peak,
each spectrum should be checked to ascertain whether any are saturated; if any are, due
allowance should be made when assessing the resulting averaged spectrum.
A background subtraction should also be performed, either on a mass spectrum from a
single scan or on an averaged spectrum, in order to remove spurious peaks such as those
produced by residual air in the mass spectrometer, or from GC column bleed.
The mass spectrum obtained for each peak detected is generally initially inspected visually.
Depending on the experience of the mass spectroscopist, it may be possible to identify the
compound giving rise to the spectrum without recourse to reference mass spectra held in
libraries (on the data system, or in reference books).
100
If the mass spectrum is not visually recognized, a library search is usually carried out on
the data system. It is recommended that a reverse searching procedure should be used. The
closeness of the match between the unknown and the chosen library spectra is usually
expressed in terms of three parameters - fit, purity and reverse fit. However, the best match
chosen by the data system does not necessarily lead to the identification of the unknown,
and the mass spectroscopist has to apply his/her judgement, taking into account such
factors as the GC retention time, in order to decide whether the identification suggested by
the computerized library search is accepted.
If there is any doubt concerning such an identification, it should be noted as a tentative
identification and, if it is necessary to confirm the identification, a pure standard of the
compound in question should be obtained and run on the GCMS system in order to check
the mass spectrum obtained and the GC retention time. The same principles apply to
potential identifications resulting from manual inspection of mass spectral reference
collections in books such as ‘The Eight Peak Index of Mass Spectra’ [1].
If it is suspected that a TIC peak is a mixture of two or more compounds, mass
chromatography may be of use in deciding whether this is the case, and by careful choice
of mass spectra it may be possible to produce spectra corresponding to each co-eluting
component. However, where two compounds have identical retention times this may not be
possible, and further progress is dependent on the experience of the mass spectroscopist.
It is inevitable that a significant proportion of the compounds detected in many general
survey GCMS runs will only be tentatively identified, and that some will be unidentified,
as the reference collections of mass spectra currently available represent a very small
proportion (< 10 %) of the known organic compounds that are amenable to GCMS
analysis.
B.2 Checking suitability of apparatus used for concentrating solvent extracts
It is necessary to be able to reduce the volume of the dichloromethane solvent extracts
from about 200 ml to 50 µl – 500 µl without significant losses of volatile components
which may have been present in the leachate sample.
To verify that this can be satisfactorily achieved, it is recommended that a 500 µl portion
of the internal standards GC column test solution (see 5.17) is diluted to 200 ml with
dichloromethane, and the resulting solution concentrated to 500 µl, using appropriate
apparatus or equipment. This concentrate should be run on the GCMS system under
exactly the same conditions as used when using the GC column test standard solution for
checking for satisfactory GC performance, and the TIC or RIC trace compared to a TIC or
RIC trace obtained when the GC column test standard is run. Provided the loss of the most
volatile internal standard, d6-benzene, is not more than 50 % the technique used for the
concentration of the solvent extracts is considered satisfactory.
B.3 Procedure for calculation of recoveries of internal standards
The concentrations of the various internal standard solutions, the volume of the leachate
analysed, and the volumes injected onto the GCMS system when the final volume of the
concentrated solvent extract is 500 µl are such that the TIC chromatograms generated for
the internal standards GC column test solution (see 4.17) and the concentrated extract (see
101
7.1) are directly comparable, so that the following equation can be used to calculate %
recoveries:
R=
Pe × 100
Ps
where:
R
Pe
Ps
is the recovery of internal standard (in %);
is the peak area of the internal standard chosen for the comparison, in
extract;
is the peak area of the internal standard chosen for the comparison, in
standard.
B.4 Standard solutions for checking GC column performance
Several chromatography supply companies produce mixtures specifically designed to
evaluate the performance of GC columns, in terms of parameters such as column efficiency
and adsorptive or ‘active’ sites. These are sometimes referred to as ‘grob mixtures’. If the
GC column used is from a manufacturer who does not provide a suitable test
chromatogram, the column should be evaluated before use with solvent extracts of
leachates, using this type of test mixture.
B.5 Performance testing data for this protocol
A tabulated summary of the data obtained during within-laboratory and inter-laboratory
performance testing of the analytical procedures described in this part of BS 6920 is given
in Tables A.1 and A.2. Competent laboratories intending to use these procedures should be
able to produce comparable data.
102
Table A.1 — Within-laboratory performance data for internal standards
Sample1)
Data
system
code
0030S021
0030S031
0030S041
0030S051
0030S061
0030S071
0030S091
0030S101
0030S121
0030S131
0030S141
0030S161
0030S171
0030S181
0030S191
0030S201
D6-benzene
d5-chlorobenzene
d10-p-xylene
d5-phenol
Peak area
d8-naphthalene
d21-BHT2)
d34-hexadecane
d10-phenanthrene
d62-squalane
PE-TW (1)
7211
8522
2981
9844
2739
56153
2768
14583
90672
PE-TW-Cl (1)
5588
10633
1815
10635
3700
56423
4668
19328
100599
GRP-TW (1)
5754
9923
3685
13097
2753
61581
3160
16134
99354
GRP-TW-CL (1)
8994
10401
3820
14973
3369
57335
4567
21121
101328
BIT-TW (1)
5603
6327
1318
7156
2302
48171
3638
16048
85730
BIT-TW-CL (1)
6916
11074
2969
9627
4402
57669
4469
16901
54348
PB-TW (1)
12190
13005
2887
10916
4029
65931
4728
15546
102366
PB-TW-CL (1)
9215
10259
2671
8351
3744
48374
4523
14240
78070
PE-TW (2)
6498
9329
3311
7335
3868
53707
3531
14276
85766
PE-TW-CL (2)
4665
9183
3188
6197
2929
62658
4435
17728
113155
GRP-TW (2)
5766
7986
2087
12357
2531
50362
5212
16965
87439
GRP-TW-CL (2)
5410
11762
3473
13804
2375
61799
6283
20480
100379
BIT-TW (2)
4714
6008
4194
9691
2534
39768
2037
13363
36799
BIT-TW-CL (2)
5702
9184
3345
14796
2609
47853
3479
15293
54239
PB-TW (2)
7207
11073
3163
6846
2765
49139
4875
14024
88095
PB-TW-CL (2)
6444
7862
2702
7097
2175
54553
6013
15490
85893
Mean
6742
9533
3038
10170
3052
54467
4274
16345
85265
SD
1955
1890
681
2934
695
6882
1125
2311
20579
% RSD
29 %
20 %
22 %
29 %
23 %
13 %
26 %
14 %
24 %
1)
Samples referred to as follows: PE = polyethylene; GRP = glass-reinforced polyester; BIT = bitumen lined ductile iron; PB = procedural blank; TW = test water; TW-CL = chlorinted test water; (1) = batch 1; (2) =
batch 2
2)
BHT = 2,6-di-t-butyl-4-methylphenol
103
Table A.2 — Summary of variation (% RSD) for internal standards in interlaboratory performance testing1)
% RSD for peak areas of internal standards
d5-chlorobenzene
d10-p-xylene
d5-phenol
d8-naphthalene
d21-BHT
1
1
15
37
24
18
23
2
19
27
51
42
16
2
1
19
26
24
21
11
2
20
19
35
19
14
2)
3
1
15
68
32
27
36
2
35
27
63
44
53
4
½
10
8
12
12
14
1)
Each laboratory analysed leachates from three samples in duplicate, together with procedural blanks. Each batch consisted of eight analyses.
2)
Laboratory 3 did not follow instructions regarding conditions for GCMS analysis
Laboratory
Batch number
d6-benzene
16
25
30
15
104
d34-hexadecane
52
21
19
32
24
84
15
d10-phenanthrene
15
19
14
15
10
36
9
d62-squalane
26
21
19
30
33
96
47
Bibliography
[1] The Eight Peak Index of Mass Spectra, 4th Edition, 1991. ISBN 0 85 186417 1,
Nottingham: The Royal Society of Chemistry.
[2] Registry of Mass Spectral Data, 3rd Edition, 1989. Eds. F.W. McLafferty and
D.B. Stauffer. ISBN 0 471 62886 7. New York: John Wiley & Sons.
105
106
Appendix 3 – Data from Stage 5
see separate report
107