Journal of Chromatography A, 1253 (2012) 144153

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A
j our nal home page: www. el sevi er . com/ l ocat e/ chr oma
Analytical method for the identication and assay of 12 phthalates in cosmetic
products: Application of the ISO 12787 international standard
CosmeticsAnalytical methodsValidation criteria for analytical results using
chromatographic techniques
Pascal Gimeno

, Annie-Franc oise Maggio, Claudine Bousquet, Audrey Quoirez, Corinne Civade,

Pierre-Antoine Bonnet
Agence nationale de scurit du mdicament et des produits de sant (ANSM), Direction des Contrles (DC), Unit Physico Chimie, 635 rue de la Garenne, 34740 Vendargues, France
a r t i c l e i n f o
Article history:
Received 26 March 2012
Received in revised form26 June 2012
Accepted 28 June 2012
Available online 4 July 2012
Keywords:
Phthalates
Cosmetics
Extraction
ISO 12787
Validation criteria
Gas chromatographymass spectrometry
a b s t r a c t
Esters of phthalic acid, more commonly named phthalates, may be present in cosmetic products as
ingredients or contaminants. Their presence as contaminant can be due to the manufacturing process,
to raw materials used or to the migration of phthalates from packaging when plastic (polyvinyl chlo-
ride PVC) is used. 8 phthalates (DBP, DEHP, BBP, DMEP, DnPP, DiPP, DPP, and DiBP), classied H360
or H361, are forbidden in cosmetics according to the European regulation on cosmetics 1223/2009. A
GC/MS method was developed for the assay of 12 phthalates in cosmetics, including the 8 phthalates
regulated. Analyses are carried out on a GC/MS system with electron impact ionization mode (EI). The
separation of phthalates is obtained on a cross-linked 5%-phenyl/95%-dimethylpolysiloxane capillary
column 30m 0.25 mm (i.d.) 0.25 mm lm thickness using a temperature gradient. Phthalate quan-
tication is performed by external calibration using an internal standard. Validation elements obtained
on standard solutions, highlight a satisfactory system conformity (resolution > 1.5), a common quanti-
cation limit at 0.25 ng injected, an acceptable linearity between 0.5 g mL
1
and 5.0g mL
1
as well
as a precision and an accuracy in agreement with in-house specications. Cosmetic samples ready for
analytical injection are analyzed after a dilution in ethanol whereas more complex cosmetic matrices,
like milks and creams, are assayed after a liquid/liquid extraction using ter-butyl methyl ether (TBME).
Depending on the type of cosmetics analyzed, the common limits of quantication for the 12 phtha-
lates were set at 0.5 or 2.5 g g
1
. All samples were assayed using the analytical approach described in
the ISO 12787 international standard CosmeticsAnalytical methodsValidation criteria for analytical
results using chromatographic techniques. This analytical protocol is particularly adapted when it is not
possible to make reconstituted sample matrices.
2012 Elsevier B.V. All rights reserved.
1. Introduction
Phthalates are esters of phthalic acid (Fig. 1). These compounds
are often present in cosmetic products such as perfumes and
toiletries. Their presencecanresult of their useas ingredient, of con-
taminants in raw materials or of the migration of phthalates from
packaging when plastic is used. 8 phthalates, classied H360 May
damage fertility or the unborn child or H361 Suspected of damaging
fertility or the unborn child, are regulated by the Europeancosmetic
regulation 1223/2009 (Annex II) [1ac] (dibutyl phthalate (DBP);
diethylhexyl phthalate (DEHP); butylbenzyl phthalate (BBP);

Corresponding author. Tel.: +33 467913951; fax: +33 467913983.

E-mail addresses: pascal.gimeno@ansm.sante.fr, pascal.gimeno@afssaps.sante.fr
(P. Gimeno).
di(2-methoxyethyl) phthalate (DMEP); di-n-pentyl phthalate
(DnPP); diisopentyl phthalate (DiPP), n-pentyl isopentyl phthalate
(DPP) and diisobutyl phthalate (DiBP)).
Some phthalates, particularly those with a low molecular
weight, can be introduced into cosmetics and perfume as ingre-
dients, for example DEP and DMP are used as solvents or perfume
xatives [24], DEP can also be used as alcohol denaturing [3,5].
Phthalates can also come from plastic containers. Indeed, phtha-
lates are commonly used as plasticizers of polyvinyl chloride (PVC)
and are not chemically bound to it. As a consequence, these com-
pounds can be extracted fromcontainers by cosmetic matrix. This
contamination can either come from the cosmetic packaging or
fromrawmaterials containers.
In 2012, no ofcial or standard method exists for the deter-
mination of phthalates in cosmetic products and only few
analytical methods are proposed in the literature [2,514].
0021-9673/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.chroma.2012.06.090
P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153 145
O
O
R
1
R
2
O
O
R1, R2 : Alkyl group
Fig. 1. Esters of phthalic acid.
HPLC/UV is the most commonly described method for the
separation, identication and quantication of phthalates in
cosmetic products [2,68,1012]. All of these methods use
C18 columns, either 250mm4.6mm 5m [2,6,8,11,12] or
150mm4.6mm3.5m[7,10]. Phthalate separationis obtained
usingeither a gradient elutionmode [2,6,7,10] or anisocratic condi-
tion [8,11,12]. The most commonly used elution solvents are EtOH
[10], MeOH [7,11,12] or ACN [8] with an aqueous buffer adjusted
to pH 2.8 [8] or water [7,1012]. Two papers use a more complex
elution gradient with 4 solvents (H
2
O/CH
3
CN/2-propanol/MeOH)
[2,6]. If most of these papers focus on the identication and quan-
tication of the phthalates the most commonly found in cosmetics
(DEP, DBP, DEHP and BBP), some papers also discuss the assay of
less usedphthalates likedi-n-heptyl phthalate(DnHP) [8] or di-allyl
phthalate (DAP) [12]. Attention has to be payed to the few num-
ber of phthalates simultaneously assayed by these HPLC methods.
Indeed, most papers propose analytical methods for the separa-
tion of 4 or 5 phthalates among which only 2 or 3 are regulated
(BBP, DEHP and DBP). Only one HPLC/UV method [10] proposes the
separation of up to 7 phthalates (DMP, DEP, DPP, DiBP, BBP, DBP
and DEHP) among which 5 are regulated by the European cosmetic
regulation.
Although less described for cosmetics, gas chromatography
(GC) is also mentioned for the analysis of phthalates in cosmetic
products [7,9]. A 5% dimethylpolysiloxane column is used for the
separation of up to 7 phthalates (di-n-octyl phthalate (DnOP), DPP,
DCHP, DEP, BBP, DBP, DEHP) among which only 4 are regulated. If
one paper [9] describes a gas chromatography (GC)/ame ioniza-
tion detection (FID) for the quantication of phthalates and GC/MS
for the identication, another paper [7] compares results obtained
using GC/MS andHPLC/UVfor the assay of 7phthalates. Whenmass
spectrometry is employed for the analysis of phthalates, a capillary
column of 30m0.25mm0.25m is used, the phthalate ion-
ization is performed by electronic impact (EI) and the detection
obtained using SIMmode (m/z =149). The sample is injected either
in splitless mode [7] or in one fth split mode [9] using He as carrier
gas. Regardless analytical method used (HPLC or GC), no reported
study deals with the simultaneous quantication of the 8 regulated
phthalates in cosmetic products.
Among analytical methods foundinthe literature for the extrac-
tion of phthalates from cosmetic products [2,512], two papers
[11,12] propose the use of innovative techniques, SMMEs (polymer
monolith microextraction) using a specic extraction tool [11] and
a new liquid/liquid extraction (LLE) approach with USAEME SFO
(ultrasound-assisted emulsication microextraction with droplet
solidication of oating organic) [12]. Later extraction protocol is
based on an optimized version of the conventional liquid/liquid
extraction process (LLE). A more conventional extraction process
based on the extraction of several cosmetic samples with hexane
on a celite column [2,6] is mentioned. Papers discussing the pos-
sibility of dispersing different cosmetic samples in methanol and
extract phthalates by ultrasounds [79] are also available. A paper
describes the evaporation of the solvent under reduced pressure
after the extraction step, the purication of the residue is per-
formed using C18 SPE columns [7]. Optimization of the extraction
solvent, of the extraction conditions and of the purication step are
also discussed [7]. It is noteworthy that this method requires large
amounts of solvent during the purication step. Extraction with
ethanol/water is also reported [10] but only for the assay of signif-
icant phthalates levels in the sample (phthalate amount >0.1%) so
this method does not appears suitable for trace quantication.
As for cosmetics the sample treatment is often reduced to a
suspension of the sample in an organic solvent followed by sonica-
tion/centrifugation/ltration prior to analysis, but as liquid/liquid
(LLE) extractionprocess is also mentioned[12], it was decidedprior
to the selection of the sample treatment to extend the literature
study to the assay of phthalates on other complexes matrices like
food and environment. A paper presents an extraction process for
the determination of phthalates from milk products [15]. Phtha-
lates are extracted by a mixture of ter-buthyl methyl ether (TBME)
and hexane from liquid samples, before a LC/ESI/MSMS analysis.
5 phthalates, including 3 regulated by the European cosmetic reg-
ulation 1223/2009, are assayed. LLE extraction techniques are also
given for the extraction of phthalates fromdifferent environmental
matrices using single dropmicro-extraction(SDME)/dispersive liq-
uid liquid micro-extraction (DLLME) [16] or water-soluble organic
solvents added with inorganic salts [17]. Another paper discusses
the analysis of phthalates in landll leaching by GC/MS [18]. This
document, published in 2002, uses a conventional LLE extraction
process and compares different organic solvents for the extraction
of phthalates from water. Ethyl acetate appears to be the most
interesting solvent (recoveries >80% regardless of the phthalate
considered). Those papers [12,1518] highlight the possibility of
extracting phthalates fromcosmetics after suspension/dissolution
in water by an immiscible organic solvent.
After a presentation of the analytical method used for the assay
of 12 phthalates, including the 8 phthalates regulated, this paper
proposes to apply the approach for result validation described
in the ISO 12787 international standard CosmeticsAnalytical
methodsValidation criteria for analytical results using chromato-
graphic techniques [19] to the analysis of different cosmetics.
Cosmetic products are various and complex, their diversity and
complexityareduetothelargeextent of their formulationmatrices.
Analytical methods developed to assess the quality of cosmetics
are given to be pertinent for their intended use, widely usable,
comprehensible and transferable. Nevertheless, their application
to cosmetic matrices requires the use of specic validation criteria
in order to assess the results obtained. The ISO12787 international
standard denes validation criteria to which analytical results
obtained fromthe analysis of cosmetic products should comply in
order to give condence in performance, reliability and quality of
the nal result. It proposes an analytical approach for chromato-
graphic analyses in order to obtain an assay result and different
validation elements which can be determined for each sample (or
sample groups) submitted to the analysis. If the same matrix or
very similar ones are used, validation results obtained on each cri-
teria can be extended to all other analyzed samples. Once those
validation elements for analytical results are determined, other
samples assays are carried out using an external calibration curve
for quantication taking into account validation criteria previously
obtained. This analytical protocol is particularly adapted when it
is not possible to make reconstituted sample matrices (i.e. sold or
in-house prepared blank sample matrix).
2. Experimental
2.1. Materials and methods
2.1.1. Gas chromatography and mass spectrometry
All analyses are carried out on a Varian 3800 gas chromato-
graph interfaced with a Varian 1200 quadripole mass spectrometer
146 P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153
Table 1
Selected ions for SIMmode acquisition.
SIM Time from Ions (m/z) Dwell time
1st group 2.5min 105, 135, 149, 163, 177, 194 25ms
2nd group 5.0min 59, 104, 149, 205, 223 30ms
3rd group 8.2min 104, 149, 219, 237, 310, 312 25ms
4th group 10.5min 57, 71, 91, 97, 149, 206 25ms
5th group 16.5min 149, 167, 225, 226, 249, 279; 25ms
6th group 20.0min 149, 167, 279 50ms
(Palo Alto, USA). The column is an Agilent HP-5MS capillary
column (crosslinked poly 5% diphenyl/95% dimethylsiloxane);
30m0.25mm (i.d.) 0.25mm lm thickness. The oven tem-
perature is programmed as follows: 100

C ramped to 200

C at
30

Cmin
1
, then to 260

C at 3

Cmin
1
, and then to 320

C at
30

Cmin
1
heldfor 5min. One microliter of eachextract is injected
in split mode one twentieth using a Varian 8400 autosampler.
The injection port and transfer line temperatures are both xed at
300

C. The ion source temperature is set at 230

C, and the helium

carrier gas owrate xed at 1mL min
1
. The mass spectrometer is
tuned on electron impact ionization (EI) at 70eV. The solvent delay
is xed at 2.5min and the run time at 30min. The acquisition is
performed on full-scan (m/z =40350) and on SIMmode according
to Table 1. On SIMmode three different ions are monitored for each
phthalate according to literature (NIST library) and after a previous
injection using the full-scan mode. Retention times of each group
are determined after the injection of a standard solution.
2.1.2. Reference standards
12 phthalates are analyzed by the method and are listed in
Table 2. Phthalates marked in bold are the 8 phthalates reg-
ulated by the European cosmetic regulation 1223/2009. DBP
(99.0%), DEHP (99.0%), BBP (98.0%), DMEP (99.4%), DCHP (99.7%),
DEP (99.5%), DMP (99.0%), DiBP (99.0%), DnOP (98.0%) and 4,4

-
dibromobiphenyle (ISTD) were purchased fromSigmaAldrich (St.
Quentin Fallavier, France), DPP (99.0% mixed of isomers), DiPP
(99.5%) and DnPP (99.0%) were obtained fromCIL Cluzeau Info Labo
(Sainte-Foy-La-Grande, France). Sodiumsulphate anhydrous came
from Panreac Quimica Sau (Barcelona, Spain). All organic solvents
(more than 99%) were obtained from different industrial sources
(VWR, Fluka, ChromaNorm).
Table 2
Analyzed phthalates.
Phthalates CAS number Legislation
DBP (dibutyl phthalate) 84-74-2 [A1,2,
3
] [B]
DEHP (diethylhexyl phthalate) 117-81-7 [A1,2,
3
] [B]
BBP (butylbenzyl phthalate) 85-68-7 [A1,2,
3
] [B]
DMEP (di(2-methoxyethyl) phthalate) 117-82-8 [A1,2,
3
]
DnPP (di-n-pentyl phthalate) 131-18-0 [A1,2,
3
]
DiPP (diisopentyl phthalate) 605-50-5 [A1,2,
3
]
DPP (n-pentyl isopentyl phthalate) 84777-06-0 [A1,2,
3
]
DiBP
a
(diisobutyl phthalate) 84-69-5 [A1,2,
3
] [B]
DCHP (dicyclohexyl phthalate) 84-61-7 [B]
DEP (diethyl phthalate) 84-66-2 [B]
DMP (dimethyl phthalate) 131-11-3 [B]
DnOP (di-n-octyl phthalate) 117-84-0 [B]
[A
1
] European regulation on cosmetics 1223/2009 30/11/2009 Annex-II (entrance
675, 677, 678, 1151 and 1152) and article 15.
[A
2
] European regulation 1272/2008 Annex VI.
[A
3
] European regulation 790/2009 Annexes II and V.
[B] Opinion on phthalates in cosmetic products. Scientic Committee on Consumer
Products (SCCP) 21/03/2007.
a
DiBP is regulated by the European cosmetic regulation 1223/2009 since the 1st
December 2010.
2.1.3. Standard solutions
The preparation and storage of standard solutions are per-
formed only with glass material to prevent the extraction of
phthalates fromplastics. The glassware is appropriately decontam-
inated before use [2,2022] by rinsing glassware several times with
an organic solvent (ethanol). During assay, ethyl acetate [23] is reg-
ularly injected to check and prevent carry over. Standard solutions
are prepared in ethanol by weighing 20.0mg of each phthalates
in a 20.0mL volumetric ask. These stock solutions can be used
for at least 7 days if they are stored in a refrigerator (see Section
3.2). A calibration curve ranging from 0.5gmL
1
to 5.0gmL
1
is prepared by diluting a 1000gmL
1
stock standard solution in
ethanol. 4,4

-Dibromobiphenyl (retention time around 9.0min) is

used as internal standard (ISTD) at a concentration of 10gmL
1
.
4,4

-Dibromobiphenyl is already used as internal standard in our

laboratoryandinliterature[24] for theassayof allergens incosmet-
ics. Inorder tohaveacommonISTDfor theassayof bothcompounds
(allergens and phthalates), 4,4

-Dibromobiphenyl was also used as

internal standard for the assay of phthalates (phthalate may be
extracted fromcosmetics in a similar way as allergens see Section
3.1).
2.1.4. Cosmetic samples
10 cosmetics samples, 6 perfumes and 4 complex cosmetics
matrices (shampoo, cream, body milk and shower gel), all taken
fromthe market, were used to test the analytical method proposed
for a quantitative analysis of phthalates. Samples identied Perf-1
to 6 were selected after a previous screening because of the pres-
ence of one of the 8 regulated phthalates at a concentration higher
than 5gg
1
(limit of report xed for phthalates in this type of
matrix after a previous dilution1/10 inethanol). Samples identied
Cos-1 to 4, initially blank of phthalates (blk), have been supple-
mented with one or several phthalates (between 1 and 10gg
1
)
to verify the accuracy and precision of the method. Supplementa-
tion was performed randomly using the 8 regulated phthalates and
the 4 different cosmetic matrices. The purpose was to obtain sam-
ples with 13 phthalates at concentration ranging from 1gg
1
,
limit of report xed for phthalates in these complexes matrices to
10gg
1
corresponding to the highest calibration level in order to
avoid extra dilutions.
2.1.5. Sample preparation
2.1.5.1. Standard protocol used for cosmetic samples ready for ana-
lytical injection (samples Perf-1 to 6). This preparation is used for
the quanticationof phthalates incosmetic samples readyfor ana-
lytical injection such as perfumes or lotions. 1mL of the sample is
diluted in a 10mL volumetric ask with ethanol after addition of
4,4

-dibromobiphenyl as internal standard (ISTD) at 10gmL

1
. In
case of excessive concentration of phthalates, an appropriate extra
dilution of the sample in ethanol is performed.
2.1.5.2. Standard protocol used for complex cosmetic matrixes (sam-
ples Cos-1 to 4). This extraction process is used for complex cos-
metic matrices like creams, milks, oils. 1g of sample is suspended
P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153 147
in 10mL of water and extracted with 10mL of TBME. After extrac-
tion, a centrifugation and a drying step with sodium sulphate is
performed. 5mL of the TBME extraction phase are evaporated to
dryness under a gentle nitrogen ow. The residue after evapora-
tion is dissolved in 1mL of ethanol (containing the ISTD). In case
of excessive concentration of phthalates, further dilutions with
ethanol are performed. As for ready for analytical injection sam-
ples 4,4

-dibromobiphenyl is used as internal standard (ISTD) at a

xed nal concentration of 10gmL
1
.
2.2. ISO 12787 standard protocol: analyses on standard solutions
and cosmetic samples
2.2.1. General
All assays were performed using the analytical approach
described in the ISO 12787 standard. This standard is divided
into three different parts. The rst part consists of determining,
using standard solutions, the main characteristics of the analytical
method used before performing tests on samples. The valida-
tion criteria on standard solutions checked during this rst step
are: the analyte limit of detection (LoD) and/or limit of quan-
tication (LoQ); the conformity of the chromatographic analysis
(Rs: resolution, As: assimetry); the linear range of the analyte
signal and the standard accuracy. This rst step is carried out
once at the beginning of the analytical programme and should be
performed again or at least checked if any analytical parameter
of the method is changed (calibration solvent, injection volume,
chromatographic column type, separation conditions, calibration
range, etc.). The second part, called screening, consists of eval-
uating the quantity of analyte in the sample, if present, in order
to adapt the sample assay protocol. According to the sample
screening result, two different protocols for the assay of the
analyte in the sample are proposed in the 3rd part of the ISO
standard.
If the analyte is not detected or detected with a S/N ratio less
than 10, assays are performed using spiked recoveries (see Sec-
tions 2.2.4.1 and 2.2.4.2 for spiked preparations). The aimof these
assays is to check that the non-detection, or the detection of the
analyte with a S/N ratio less than 10 in the analyzed sample, is not
due to an analytical problem(suppression of the signal due to the
presence of an interference compound in the cosmetic matrix, bad
extraction of the analyte from the cosmetic) but to the absence of
this analyte, at a known concentration limit, in the analyzed sam-
ple. If the analyte is detected with a S/N ratio higher than 10 (see
Sections 2.2.4.1 and 2.2.4.2 for spiked preparations), assays are per-
formed using a specic protocol based on the preparation of spiked
and unspiked samples. The aim of these trials is to determine the
analyte concentration in the sample as well as several validation
parameters, like the matrix effect, the extraction yield, the accu-
racy, andthe condence interval. These parameters are determined
by performing statistical analyses on six preparations of the same
sample: 3unspikedpreparations, 2samples spikedwiththeanalyte
of interest at the beginning of the analytical procedure (PrEMS) and
1 sample taken through the entire extraction procedure and spiked
withthe analyte of interest at the endof the extractionimmediately
before, or very close to, detection (PoEMS). Recoveries obtained
from PoEMS and/or PrEMS lead to the determination of different
validation criteria on the analytical results. The PoEMS recovery
relative to calibration standards shows whether or not there is a
matrix effect. The difference between the PoEMS and the PrEMS
recoveries, relative to calibration standards, gives the extraction
yield of the analytical process. The recovery obtained for the PrEMS
relative to the PoEMS gives the accuracy of the analytical result. A
RSD or a condence interval on the assay result can be obtained by
a statistical analysis of the replicates.
2.2.2. Validation on standard solutions for the 12 phthalates
analyzed (1st part)
2.2.2.1. Detection and quantication limits in solvent. As 8 of the
12 analyzed phthalates are forbidden in cosmetics products and
as according to the European regulation on cosmetics 1223/2009
(Paragraph 7, Article 17 and Annex I) forbidden substance in cos-
metic may be present at trace levels, phthalates must be quantied
at the lower possible amount (trace elements determination). In
order to estimate this quantication amount, an evaluation of the
limits of detection (LoD) and quantication (LoQ) in standard solu-
tion was performed. For the LoQ evaluation, in agreement with the
ISO12787, small quantities of phthalates are injected in the GC/MS
systemranging from0.025ng to 0.5ng (0.025, 0.050, 0.10, 0.25 and
0.50ng). The limit of detection (LoD) is given by the y intercept
value obtained during the linearity assay (Section2.2.2.3). The limit
of quantication (LoQ) is evaluated as 3LoD and checked inject-
ing a standard solution at the estimated LoQ concentration value.
The S/N peak to peak at the LoQ must be >10.
2.2.2.2. Analytical conformity/specicity. The analytical conformity
of the methodis checkedby injecting a solvent calibrationstandard
at the higher expected level of the calibration curve (5.0gmL
1
).
In order to ensure the absence of interference peak in the solvent,
the dilution solvent (ethanol) is also injected. The specicity of the
detection is checked using the MS spectra (m/z ions and ratio) and
the phthalates retention times.
2.2.2.3. Calibration linearity standard accuracy. Three indepen-
dent calibration curves ranging from 0.5gmL
1
to 5.0gmL
1
(0.5, 1.0, 1.5, 2.0, 3.0, 4.0 and 5.0gmL
1
) are prepared by diluting
3 different standard stock solutions in ethanol and are injected. The
different calibration levels are uniformly distributed along the cali-
brationrange. For the determinationof lowconcentrationanalytes,
the rst calibration level corresponds to the quantication limit in
solvent. The upper end is usually signied by a change in instru-
ment response. The method accuracy on standard solutions could
be estimated, for each calibration level, as the mean concentration
bias obtained on this level (three values for each calibration level).
The calibration linearity was checked on standard solutions using
statistic software(AVAAide la ValidationAnalytiqueversion3.1.
distributed by Qualilab (45160 Olivet France) or Statgraphic Centu-
rionversion16.1.05 (32 bits) distributedby StatPoint Technologies,
Inc. (USA)) 4,4

-dibromobiphenyl is usedas internal standard(ISTD)

at a xed concentration of 10gmL
1
.
2.2.3. Sample screening (2nd part)
Before assays, a screeningof samples is performed, inagreement
with the 2nd step of the ISO 12787 standard, in order to estimate
the amount of phthalates in analyzed samples if present. A cal-
ibration curve, in agreement with the linearity range previously
determined (ranging from0.5gmL
1
to 5.0gmL
1
), is prepared
and injected. A control standard is performed at 2.0gmL
1
inde-
pendently fromthe standard solutions used for the preparation of
the calibration curve. This control standard is used to check the cal-
ibration curve preparation. Each sample (Perf-1 to 6 and Cos-1 to
4) is prepared according to the sample preparation protocol given
in Section 2.1.5 depending on the cosmetic matrix and injected.
2.2.4. Assay of phthalates (3rd part)
For each extraction batch, a whole reagent blank is carried out
to check that the materials used for extraction and analysis had not
contaminated the samples. For each sample, spiked preparations
are performed using the minimum possible amount of a solvent
compatible with the sample preparation to introduce the analyte
in the cosmetic. Depending on the sample type, spiked samples
are prepared either mixing the analyte solution with the sample
148 P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153
Fig. 2. Example of chromatograms obtained for a blank reagent, the sample Cos-2
(cream) and a 0.5gL
1
standard solution.
(Cos-1 to 4) or allowing dispersion into liquid samples (Perf-1 to
6). SIMchromatograms obtained for a blank reagent and for Cos-2
(cream) are given as example in Fig. 2.
2.2.4.1. Assayfor perfume samples (Perf-1to6). For eachassay, three
unspiked preparations and two spiked preparations of the sample
at the estimated concentration value (evaluated during the screen-
ing step) are made according to Section 2.1.5.1. In order to check
the absence of interference peak with the ISTD used, an additional
unspiked preparation without internal standard is performed.
Phthalate quantications are performed by external calibration
using an internal standard and using a multi-point/single levels
standard addition method (using the 3 unspiked and the 2 spiked
preparations) [25,26]. An additional independent assay is per-
formed, using a multi-point/multi-levels standardadditionmethod
(using 2 unspiked preparations and 2 spiked preparations at 4 con-
centration levels) [27]. Results were compared to those obtained
using the single-point standard addition approach, as proposed in
the ISO 12787 standard. For information, a prediction interval (PI)
using the Fieller theorem (allows the calculation of a condence
interval for the ratio of two means) [28,29] based on data from
the multi-point/multi-levels standard addition method is deter-
mined. For this determination statistic software, like Statgraphic
was used. Phthalates recoveries are calculatedby external standard
quantication from spiked preparations. In order to check recov-
eries obtained for phthalates not detected, an additional spiked
preparation at 5gg
1
, corresponding after sample preparation
(Section 2.1.5.1) to the limit of report xed for ready for analytical
injection samples (Section 3.1.1), with all phthalates analyzed is
systematically performed for all samples studied.
2.2.4.2. Assay for cosmetic samples other than perfumes (Cos-1 to 4).
For eachassay, three unspiked preparations of the overloaded sam-
ple and two spiked preparation with all phthalates analyzed at
1gg
1
of the overloaded samples before extraction are prepared
according to Section 2.1.5.2. An additional spiked preparation at
1gg
1
of the overload samples after extraction is performed
according to Section 2.1.5.2. In order to check the absence of inter-
ference peak with the ISTD used, a preparation without internal
standards is also carried one. 1gg
1
corresponds, after sam-
ple preparation (Section 2.1.5.2), to the limit of report xed for
cosmetic samples (Section3.1.1). Phthalate quantications are per-
formedby external calibrationusing aninternal standardandusing
a multi-point/single levels standard addition method (using the
3 unspiked and the 2 spiked preparations). Phthalates recover-
ies are calculated by external standard quantication fromspiked
preparations (before or after extraction) depending on validation
elements determined (matrix effect, extraction yield, etc.). The
accuracy and reliability of the assay are determined from the
amount of phthalate initially introduced in the blank matrix.
3. Results and discussion
3.1. Method development
As no paper deals with the simultaneous quantication of the
8 regulated phthalates in cosmetic products, and as most of the-
ses compounds can be considered as volatiles analytes, a GC/MS
method was developed for the assay of at least the 8 phthalates
regulated. GC/MS is a commonly used analytical technique for the
determination of phthalates in the environment. The development
of the chromatographic method proposed was rst focused on the
separation and identication of the 8 phthalates regulated by the
European cosmetic regulation and then extended, when possible,
to the assay of phthalates discussed by the Scientic Committee
on Consumer Products (SCCP) in document Opinion on phthalates
in cosmetic products [13]. The chromatographic conditions devel-
oped allow the determination of 12 phthalates (DBP, DEHP, BBP,
DMEP, DnPP, DiPP, DPP, DiBP, DCHP, DEP, DMP, DnOP), including
the 8 phthalates regulated, but are not suitable for the quan-
tication of di-isononylphthalate (DiNP) and di-isodecylphthalate
(DiDP). According to SCCP [13], the possible presence of DiNP and
DiDP in cosmetics does not seemto be a problemfor human health.
AGC/MS method using positive chemical ionization with ammonia
as collision gas is proposed in literature for the determinations of
both phthalates in cosmetics products [30]. In order, as far as possi-
ble, to develop for cosmetic standardization an easy and currently
applicable analytical method for control laboratories, this ioniza-
tion mode, which requires a specic equipment, was not preferably
considered.
In agreement with literature and in order to eliminate when
possible, interfering compound and limit the contamination of
the apparatus, an extraction protocol based on the dissolu-
tion/suspension of the cosmetic in water followed by a LLE using
TBME was tested. This protocol, already used in our laboratory for
the extraction of volatile fragrance allergens in cosmetics, is pro-
posed in literature [31] for the extraction of allergens in foods.
Such extraction mode could be relevant for the assay of phtha-
lates because, phthalates, which are esters of phthalic acid, may
be extracted fromcosmetics in a similar way as allergens contain-
ing similar organic function like esters of benzoic acid. In order
to have a common extraction process for the assay of allergens
and phthalates, TBME, not studied for the extraction of phthalates
in landll leaching [18] but used with hexane for the determina-
tion of phthalates frommilk products [15], was compared to ethyl
acetate (EA), supposed to be the most interesting solvent for the
extraction of phthalates from water [18]. In this aim, three inde-
pendent extractions of the matrix Cos-1 (shampoo), supplemented
at 1gg
1
(report limit) with phthalates classied H360, were
P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153 149
Table 3
S/N ratio obtained for each phthalate on the quantier ion.
S/N ratio
Phthalates Quantier ion (m/z) Solvent 0.025gmL
1
0.050gmL
1
0.10gmL
1
0.25gmL
1
0.50gmL
1
DMP 163 0 <10 17 33 85 158
DEP 149 0 <10 18 41 102 192
DiBP 149 0 <10 16 46 87 204
DBP 149 0 <10 16 34 82 166
DMEP 59 0 <10 <10 <10 10 36
DiPP 149 0 <10 <10 16 31 74
DPP 149 0 <10 <10 <10 28 49
DnPP 149 0 <10 16 23 74 124
BBP 149 0 <10 <10 <10 12 34
DCHP 149 0 <10 <10 <10 15 42
DEHP 149 0 <10 <10 <10 17 36
DnOP 149 0 <10 <10 <10 20 45
S/N ratio are obtained using the instrument software on peak to peak ratio mode.
performed using either EA or TBME as extraction solvent according
to the sample preparation given in Section 2.1.5.2. Results obtained
point out similar recoveries using TBME or EA, average recoveries
obtained are between 70 and 130% regardless of the phthalate and
the extraction method used. These recoveries were calculated rel-
ative to the amount supplemented. The main advantage of TBME,
comparedtoEA, is its lower boilingpoint (55.2

Cvs 77.1

C) leading
to an easier and faster evaporation of the extraction solvent during
the concentration step.
3.2. Preliminary validation criteria using the ISO 12787 standard
In agreement with the analytical approach described in the ISO
12787 standard, preliminary assay using standard solutions were
performed in order to check some general analytical elements like:
phthalates stability in solution, apparatus conformity (injection
repeatability, detector calibration, etc.). For phthalates, the detec-
tor calibration has been achieved via the performance of a mass
spectrometer auto-tune. The apparatus conformity was checked
by injecting 6 times a standard solution containing a mix of the 12
phthalates at 5.0gmL
1
. The RSDobtainedfor these solutions was
<3% for each phthalate considered in agreement with in-house
validationcriteria. In-house criteria were established fromdiffer-
ent documents in the eld of cosmetic and agro chemistry [3239].
An evaluation of phthalates stability in EtOH, by comparing ini-
tial solutions to fresh prepared solutions, was performed, results
pointed out an acceptable conservation of phthalates in stock solu-
tions for at least 7 days if stored in a refrigerator (absence of light
at approximately 4

C). Result obtained during this study point

out relative bias in concentration <10% for concentrations ranging
from 1.0 to 5.0gmL
1
regardless phthalate considered (<20% if
the stock solution is stored 15 days). Nevertheless, for the deter-
mination of validation criteria, standards and samples were daily
prepared. Results obtained for the assay of the 12 phthalates in
the 10 samples considered, using the standard analytical method
described in Section 2.2.1 are discussed following the ISO 12787
standard protocol.
3.3. Validation data obtained on standard solutions (ISO 12787
1st part)
3.3.1. Detection and quantication limits on the quantier ion
For the LoD, an estimation is achieved using the y intercept
value of each calibration curve obtained during the linearity assay
[4042]. The LoD is estimated from the y intercept (a) and the
standard deviation of the y-intercepts (SDa) according to the
formula: LoD=a +3SDa. According to this mathematical approx-
imation, the detection limit of each phthalate on standard solution
was determined between 0.01gmL
1
and 0.1gmL
1
(0.1ng
injected). The common LoDon the quantier ion for the 12 selected
phthalates was estimated at 0.1gmL
1
for 1L injection in split
one twentieth. This limit of detection (LoD) evaluated on stan-
dard solutions corresponds, according to sample preparation, to a
quantication limit roughly of 1.0gmL
1
(i.e. 0.9gg
1
consid-
ering a mean density for perfumes around d=0.85) for ready for
analytical injection samples and of 0.2gg
1
for cosmetic sam-
ples. S/N ratios obtained for rst calibration levels are presented in
Table 3. The LoQ is evaluated using standard solutions. According
to assay performed, the quantication limit of each phthalate was
determinedbetween0.05gmL
1
and0.25gmL
1
(0.050.25ng
injected) corresponding to an S/Nratio10. A common LoQon the
quantier ionfor the 12selectedphthalates was set at 0.25gmL
1
on standard solutions using the same analytical conditions. This
limit of quantication (LoQ) evaluated on standard solutions cor-
responds, according to sample preparation, to a quanticationlimit
of 2.5gmL
1
(i.e. 2.1gg
1
considering a mean density for per-
fumes around d=0.85) for ready for analytical injection samples
and of 0.5gg
1
for cosmetic samples. The report limits were xed
at 5.0gg
1
for ready for analytical injection samples and at
1.0gg
1
for cosmetic samples.
3.3.2. Analytical conformity/specicity
The specicity of the detection is ensured by monitoring three
different ions for each phthalate using a mass spectrometer in
SIM mode (Table 4). Results obtained do not point out any inter-
ference of the dilution solvent with neither phthalates nor the
ISTD used. Nevertheless, a weak specicity due to the low quali-
er ions intensity (low ion ratio) is observed as general issue for
benzene derivatives such as phthalates. The maximum permitted
tolerances for relative ion intensities should be in agreement with
the 2002/657 Commission Decision [43]. The system compliance
is checked by injecting a standard solution at 5.0gmL
1
(corre-
sponding to the higher calibration level). To ensure the absence of
interfering peaks or of a carryover, the dilution solvent (ethanol)
is also injected. Resolution factors between the nearest phthalates
(DCHP and DEHP), and DMEP and ISTD must be Rs >1.5. For infor-
mation, a chromatographic prole obtained in SIM mode for a
0.5gmL standard solution is given in Fig. 2.
3.3.3. Calibration linearity standard accuracy
Calibration linearity is checked on standard solutions using
statistical software (Table 5). Results obtained point out a not sig-
nicant (NS) variance homogeneity for all phthalates tested. This
variance homogeneity allows a valid statistical process for lin-
earity criteria. Determination coefcients (R
2
) obtained for each
phthalates (>0.9900), Fishers tests, performed to verify the good
150 P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153
Table 4
Specicity data.
Phthalates RT (min) Ion 1 (m/z) Ion 2 (m/z) Ion 3 (m/z) Ion ratio 2/1% Ion ratio 3/1%
DMP 3.8 163 194 135 6 4
DEP 4.6 149 177 105 21 9
DiBP 6.8 149 104 223 7 5
DBP 7.9 149 205 223 3 3
DMEP 8.4 59 149 104 13 10
DiPP 9.4 149 237 104 7 5
DPP 10.1 149 237 219 5 2
DnPP 10.7 149 237 104 2 3
BBP 14.4 149 91 206 73 19
DCHP 17.9 149 167 249 31 3
DEHP 18.5 149 167 279 31 2
DnOP 22.5 149 279 167 2 1
Table 5
Calibration linearity.
Validation paramaters DBP DEHP BBP DMEP DnPP DiPP
Determination coefcient R
2
0.9909 0.9915 0.9931 0.9914 0.9915 0.9905
Slope 0.2103 0.0898 0.0657 0.0949 0.2583 0.1344
y-Intercept 0.0670 0.0422 0.0171 0.0337 0.0931 0.0408
Variance homogeneity (Cochran) NS NS NS NS NS NS
Aberrant values (Dixon) NO NO NO NO NO NO
y-Intercept comparison with 0 (Student) S S S S S S
Signicant slope existence (Fisher) HS HS HS HS HS HS
Validity of calibration curve (Fisher) NS NS NS NS NS NS
Validation paramaters DPP DCHP DEP DMP DiBP DnOP
Determination coefcient R
2
0.9924 0.9909 0.9936 0.9933 0.9935 0.9930
Slope 0.0737 0.0905 0.1564 0.1622 0.1945 0.1476
y-Intercept 0.0220 0.0155 0.0419 0.0394 0.0481 0.0427
Variance homogeneity (Cochran) NS NS NS NS NS NS
Aberrant values (Dixon) NO NO NO NO NO NO
y-Intercept comparison with 0 (Student) S S S S S S
Signicant slope existence (Fisher) HS HS HS HS HS HS
Validity of calibration curve (Fisher) NS NS NS NS NS NS
NS: non signicant; S: signicant; HS: highly signicant.
correlation between experimental values and the regression model
used (NS), as well as mean relative bias in agreement with in-
house specications (see Table 6) point out an acceptable linear
correlation for all phthalates on the calibration range considered
(0.55.0gmL
1
). For this study, the 0 value has not been forced
through the origin and the comparison of y-intercept with zero
(Students test) is signicant (S). The linearity may probably not be
acceptable on a wider concentration range. Residual plots show
a general symmetric repartition of bias obtained for the differ-
ent calibration levels using the linear regression model regardless
phthalate considered. Mean relative bias (absolute bias/calibration
level 100) are summarized in Table 6.
3.4. Sample screening (ISO 12787 2nd part)
The aim of this step is to evaluate the quantity of phtha-
lates in each analyzed sample. This step is a requirement before
samples assays according to the analytical approach described
in the ISO 12787 standard (step 2). After having checked
the determination factors (R
2
>0.9900) and accepted the bias
obtained for the check standard using the standard calibration
curve (<20%), the phthalates amount in each sample was esti-
mated using a standard calibration curve. Screening results are
given in Table 7. For samples Cos-1 to 4, no phthalates were
detected.
Table 6
Mean relative bias (%).
Calibration level DBP DEHP BBP DMEP DnPP DiPP DPP DCHP DEP DMP DiBP DnOP
0.5gmL
1
26.2 21.8 19.1 35.2
b
18.7 12.6 13.2 11.4 20.2 19.1 21.7 21.4
1.0 gmL
1
3.1 1.3 0.0 4.8 5.3 4.2 3.7 6.3 0.6 0.6 1.9 1.4
1.5 gmL
1
4.5 1.9 3.6 8.0 2.2 2.3 1.0 1.2 0.9 1.5 2.5 4.7
2.0 gmL
1
8.0 8.2 4.1 Nr
a
8.3 4.7 5.1 0.2 4.7 4.2 4.7 5.5
3.0 gmL
1
0.0 0.9 1.1 1.9 0.8 0.6 3.6 0.9 1.7 2.9 0.3 2.1
4.0 gmL
1
0.0 0.3 1.9 1.7 0.8 0.2 2.7 2.4 0.1 1.4 0.4 0.6
5.0 gmL
1
1.3 1.1 Nr
a
2.3 0.9 1.0 1.2 1.0 1.2 0.7 1.0 Nr
a
In-house validation criteria: rst calibration level (close to LoQ) <30%/other calibration levels <10%.
a
Non retained calibration level (signicant variance homogeneity).
b
Calibration level near the DMEP LoQ (0.25gmL
1
).
P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153 151
Table 7
Assay of phthalates in samples Perf-1 to 06.
Sample Phthalate Screening Assay Additional
independent assay
Screening amount
b
External standard
quantication
Multi-point/single level
standard addition method
(3 unspiked+2 spiked
samples)
Multi-point/multi-
levels standard
addition method (5
levels in duplicate)
Perf-1 BBP 100gg
1
126gg
1
(RSD=5.6%. n=3)
Mean recovery:
100%
(90%. 109%)
112gg
1
90gg
1
PI: [59125]
a
Perf-2 BBP 300gg
1
312gg
1
(RSD=3.9%. n=3)
Mean recovery:
91%
(85%. 96%)
323gg
1
350gg
1
PI: [293419]
a
Perf-3 DEHP 100gg
1
103gg
1
(RSD=4.7%. n=3)
Mean recovery:
113%
(109%. 118%)
118gg
1
143gg
1
PI: [111179]
a
Perf-4 DEHP 300gg
1
301gg
1
(RSD=5.1%. n=3)
Mean recovery:
96%
(95%. 97%)
273gg
1
340gg
1
PI: [220474]
a
Perf-5 DEHP 40gg
1
39gg
1
(RSD=3.2%. n=3)
Mean recovery:
96%
(89%. 104%)
38gg
1
42gg
1
PI: [2169]
a
Perf-6 DEHP 10gg
1
12gg
1
(RSD=9.7%. n=3)
Mean recovery:
86%
(88%. 83%)
13gg
1
18gg
1
PI: [1223]
a
a
Predictive interval (PI) calculated using the Fieller theorem approach [30,33].
b
Each other phthalates <5gg
1
.
3.5. Assay of phthalates in samples (ISO 12787 3rd part)
3.5.1. Results obtained for the assays of perfumes
In agreement with screening results, spiked preparations were
realizedfor samples Perf-1to4at 100gg
1
andfor samples Perf-5
to 6 at 10gg
1
with the phthalate detected from the screen-
ing step (BBP or DEHP depending on the sample considered). In
order to check recoveries obtained for phthalates not detected,
additional spiked preparations at 5gg
1
with all phthalates ana-
lyzed were systematically performed for all samples analyzed.
Results obtained for the assay of phthalates detected in sam-
ples Perf-1 to 6 are given in Table 7. This table compares for
each sample: screening results, quantication results obtained by
external calibration or using a multi-point/single level standard
addition method and an independent assay performed using a
multi-point/multi-levels standard addition method. A prediction
interval (PI) obtained from the multi-point/multi-levels standard
addition method is also given. Phthalates recoveries are calcu-
latedby external standardquanticationfromspikedpreparations.
Validation elements obtained for perfumes samples (Perf-1 to 6)
show analyte amounts determined by external calibration close
to those obtained by both standard addition methods with a RSD
for n=3 less or around 5% excepted for sample Perf-6 with a RSD
around10%for a DEHPamount of 10gg
1
. Recoveries obtainedon
spiked preparations, using external standard quantication, range
from80%to 120%. Validationcriteria obtained (RSDand recoveries)
are in agreement with in-house specications. These results point
out a satisfactory correlation between external and standard addi-
tion quantications showing the absence of a matrix effect. The
analyte amount obtained using multi-point/single level standard
addition method is included in the prediction interval obtained
with a second independent assay using multi-point/multi-levels
standard addition method. These results may allow the use of
multi-point/single level standard addition method for the quanti-
cation of phthalate in perfumes. The advantage of the single level
standard addition method (using the 3 unspiked and the 2 spiked
preparations) compared to multi levels standard addition method
is to reduce the number of sample preparation without removing
condence in performance, reliability and quality of the nal result
thanks to the evaluation of validation criteria. For phthalates not
detected in perfumes, mean recoveries, using 5gg
1
spiked sam-
ples, range from70% to 130%. These recoveries allowto check that
the absence of signal in the non-spiked sample is not due to an ana-
lytical problemduringassayor toasuppressionof theanalytesignal
in the cosmetic matrix because of the presence of an interference
compound.
3.5.2. Results obtained for the assay of Cos-1 to 4
Samples Cos-1 to 4, initially blank of phthalates, have been sup-
plementedbefore the assay randomly withone or several regulated
phthalates (between 1 and 10gg
1
) in order to verify the accu-
racyandprecisionof the method(see Table 8). These supplemented
samples were then assayed using the ISO 12787 standard proto-
col. According to Table 8, expected phthalates amounts in samples
Cos-1 to 4 range between 1 and 10gg
1
for supplemented phtha-
lates or are <1gg
1
for not added phthalates. Consequently, as
1gg
1
could be used as a common spiked level for the 12 stud-
ied phthalates, spiked preparations at 1gg
1
with all phthalates
analyzed were performed to check recoveries for all samples ana-
lyzed. Results obtained for the assay of phthalates in samples Cos-1
152 P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153
Table 8
Assay of phthalates in samples Cos-1 to 4.
Sample Added phthalate Added amount
a
Assay (external
quantication)
Multi-point/single level
standard addition method
Matrix effect (ME)
Extraction yield (EY)
Mean recovery (MR)
Cos-1
(shampoo)
DBP 4.4gg
1
4.7gg
1
(RSD=1.5%. n=3)
Accuracy: 106%
3.9gg
1
ME: 103%
EY: 126%
MR: 130%
DiBP 2.2gg
1
2.8gg
1
(RSD=0.7%. n=3)
Accuracy: 128%
2.1gg
1
ME: 104%
EY: 122%
MR: 127%
DEHP 5.3gg
1
5.6gg
1
(RSD=4.7%. n=3)
Accuracy: 106%
5.9gg
1
ME: 130%
EY: 96%
MR: 125%
Cos-2
(cream)
DMEP 5.9gg
1
6.3gg
1
(RSD=3.9%. n=3)
Accuracy: 107%
6.5gg
1
ME: 78%
EY: 122%
MR: 96%
DiPP 2.0gg
1
2.1gg
1
(RSD=3.4%. n=3)
Accuracy: 106%
2.6gg
1
ME: 89%
EY: 101%
MR: 90%
Cos-3
(body milk)
BBP 7.9gg
1
8.7gg
1
(RSD=3.6%. n=3)
Accuracy: 110%
8.5gg
1
ME: 92%
EY: 112%
MR: 104%
DnPP 5.2gg
1
4.2gg
1
(RSD=16.5%. n=3)
Accuracy: 80%
3.6gg
1
ME: 96%
EY: 122%
MR: 118%
Cos-4 (shower gel) DPP 1.0gg
1
1.0gg
1
(RSD=11.3%. n=3)
Accuracy: 101%
0.8gg
1
ME: 107%
EY: 116%
MR: 124%
a
Initially blank samples supplemented with one or several phthalates.
to 4 are given in Table 8. This table compares for each sample, the
knownamount of phthalate initially introducedinthe blank matrix
and quantication results obtained by external calibration or using
a multi-point/single level standardadditionmethod. Validationcri-
teria obtained for each sample (matrix effect, extraction yield and
mean recovery) are calculated according to the ISO 12787 stan-
dard by external standard quantication fromspiked preparations
(previous or after extractiondependingonvalidationcriteria deter-
mined). The accuracy and reliability of the assay are checked from
the determination of the amount of phthalate initially introduced
in the blank matrix. Whatever phthalate or sample considered,
results obtained highlight an accuracy and a relative standard devi-
ation (RSD) of the assay, determined fromthe amount of phthalate
initially introduced in the blank matrix, between 80% and 130%
with a RSD for n=3 usually <5%, except for DnPP in Cos-3 and
DPP in Cos-4 respectively at RSD=17% and RSD=11%. Such relative
high RSD results might be link to slight but acceptable interaction
with other matrix components in such body milk and shower gel.
In addition to above validation criteria, extraction yield, matrix
effect and mean recovery could be evaluated thanks to the ISO
12787 standard approach. Regardless phthalate or cosmetic sam-
ple considered, extraction yields are between 70 and 130% which
could be considered as acceptable regarding sample preparation
(liquid/liquid extraction) and trace level analysis. Furthermore, no
signicant matrix effect is pointed out when comparing sample
preparation spiked after (overload) and before (spiked) extraction.
This last criteria are in agreement with the satisfactory correlation
obtained between phthalate amount using external and standard
addition quantications. Average recoveries obtained fromspiked
preparations, before extraction, are between 70 and 130% and
are directly correlated with extraction yield and matrix effect
values.
For phthalates not detected in these cosmetic products, mean
recoveries range from70% to 130%. These recoveries allowto check
the performance of phthalate extractions fromthe cosmetic matrix
using TBME as extraction solvent and showthat the use of the sam-
ple preparation proposed for the assay of these cosmetic samples
is adequate for it intended use.
4. Conclusion
For safety reason, the ANSM laboratories (National Drug and
Health Products Safety Agency) are particularly involved in the
development of new standards at the European (CEN) or Inter-
national (ISO) levels for the assay of cosmetic products. ANSM
participated in the elaboration of the ISO 12787 international
standard CosmeticsAnalytical methodsValidation criteria for
analytical results using chromatographic techniques.
The analytical method proposed in this paper allows the sep-
aration and quantication of 12 phthalates by GC/MS including
the 8 phthalates regulated by the European regulation on cosmet-
ics 1223/2009, 30/11/2009 (DBP, DEHP, BBP, DMEP, DnPP, DiPP,
DPP and DiBP). Validation criteria obtained using the analytical
approach described in the ISO 12787 standard on standard solu-
tions highlight a satisfactory conservation of stock solutions for at
least 7 days if they are stored in a refrigerator (absence of light at
approximately 4

C), a common quantication limit (LoQ) for all

phthalates considered at 0.25gmL
1
which complies with most
assay purposes and an acceptable linearity from 0.5gmL
1
to
5.0gmL
1
. This method was applied to cosmetic products ready
for analytical injection (Perf-1 to 6), and more complex cosmetic
samples (Cos-1 to 4). For perfume samples (Perf-1 to 6), recoveries
obtained on spiked preparations, using external standard quanti-
cation, range from80%to 120%witha RSDonassay generally less or
about 5%. For more complex cosmetic samples (Cos-1 to 4), results
obtained highlight, regardless phthalate or sample considered, an
accuracyanda relative standarddeviation(RSD) of the assay, deter-
mined from the amount of phthalate initially introduced in the
blank matrix, between 80% and 130% with a RSD usually <5%. For
phthalates not detected in these cosmetic products, mean recover-
ies range from70% to 130%.
The chromatographic method proposed in this paper was dis-
cussed on cosmetic standardization meetings and proposed as a
working document for the possible elaboration of a new standard
for the assay of phthalates in cosmetic samples ready for analyti-
cal injection at the CEN(EuropeanCommittee for Standardisation)
level [44]. For cosmetic samples ready for analytical injection
P. Gimeno et al. / J. Chromatogr. A 1253 (2012) 144153 153
(perfumes, lotions), the sample treatment was limited for phtha-
lates quantication to a dilution in ethanol and for phthalates
identication to a direct injection of the cosmetic samples in the
GC/MS system.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.chroma.2012.06.090.
References
[1] (a) European regulation on cosmetics 1223/2009 30/11/2009 Annex-II
(entrances 675, 677, 678, 1151 and 1152) and article 15.;
(b) European regulation 1272/2008 Annex VI.;
(c) European regulation 790/2009 Annexes II and V.
[2] J.C. Hubinger, D.C. Havery, J. Cosmet. Sci. 57 (2006) 127.
[3] A. Gmez-Hens, M.P. Aguilar-Caballos, Trends Anal. Chem. 22 (2003) 847.
[4] FDA, Phthalates Cosmetic Products, April 19, 2001, Updated March 31, 2005
and February 7, 2008.
[5] E.W. Godly, A.E. Mortlock, Analyst 98 (1973) 493.
[6] J.C. Hubinger, J. Cosmet. Sci. 61 (2010) 457.
[7] H.Y. Shen, H.L. Jiang, H.L. Mao, et al., J. Sep. Sci. 30 (2007) 48.
[8] H.J. Koo, B.M. Lee, J. Toxicol. Environ. Health 67 (2004) 1901.
[9] H. Chen, C. Wang, X. Wang, et al., Int. J. Cosmet. Sci. 27 (2005) 205.
[10] D. De Orsi, L. Gagliardi, R. Porr, et al., Anal. Chim. Acta 555 (2006) 238.
[11] R. Su, X. Zhao, Z. Li, et al., Anal. Chim. Acta 676 (2010) 103.
[12] F. Kamarei, H. Ebrahimzadeh, Y. Yamini, Microchem. J. 99 (2011) 26.
[13] SCCP/1016/06, Opinion on Phthalates in Cosmetic Products, Scientic
Committee on Consumer Products, 2007, http://ec.europa.eu/health/
ph risk/committees/04 sccp/docs/sccp o 106.pdf.
[14] L. Sanchez-Prado, M. Llompart, J. Pablo Lamas, et al., Talanta 85 (2011) 370.
[15] L.K. Sorensen, Rapid Commun. Mass Spectrom. 20 (2006) 1135.
[16] C. Mahugo-Santana, Z. Sosa-Ferrera, M.E. Torres-Padron, et al., Trends Anal.
Chem. 30 (2011) 731.
[17] Y. Cai, Y. Cai, Y. Shi, et al., Microchim. Acta 157 (2007) 73.
[18] S. Jonsson, H. Boren, J. Chromatogr. A 963 (2002) 393.
[19] ISO 12787:2011, CosmeticsAnalytical methodsValidation criteria for ana-
lytical results using chromatographic techniques.
[20] A. Fankhauser-Noti, K. Grob, Anal. Chim. Acta 582 (2007) 353.
[21] F. David, P. Sandra, B. Tienpont, et al., The Handbook of Environment Chemistry.
Analytical Methods Review, vol. 3Q, 2003, p. 9 (Chapter 2).
[22] B. Tienpont, F. David, E. Dewulf, et al., Chromatographia 61 (2005) 365.
[23] ISO 18856:2004, Water quality determination of selected phthalates using
gas chromatography/mass spectrometry.
[24] A. Chaintreau, E. Cicchetti, N. David, et al., J. Chromatogr. A 1218 (2011) 7869.
[25] A. Garrido Frenich, J.L. Martnez Vidal, J.L. Fernndez Moreno, et al., J. Chro-
matogr. A 1216 (2009) 4798.
[26] M. Antler, E.J. Maxwell, D.A. Duford, et al., Anal. Chem. 79 (2007) 688.
[27] J.C. Miller, J.N. Miller, Statistics for Analytical Chemistry, 2nd ed., Halsted Press,
NewYork, 1992, pp. 117120.
[28] European Pharmacopoeia 7.0, 5.3 Statistical analysis of results of biological
assays and tests.
[29] J.P. Buonaccorsi, Encycolopedia of Environmetrics, 2001.
[30] C. George, H. Prest, LCGC N. Am. 20 (2) (2002) 143.
[31] D. Smith, K. Lynam, GC/MS Identication of Flavour and Fragrance Allergens
in Some Common Snack Foods using an Agilent J&W DB-17ms Capillary GC
Column, Application Note Agilent Technologies, 2009.
[32] European Commission: Directorate General Health and Consumer Protection
SANCO/825/00 rev. 8.1 16/11/2010 Guidance Document on Pesticide Residue
Analytical Methods.
[33] Technical Guideline CORESTA Guide N

5, February 2008, Technical Guideline

for Pesticide Residues Analysis on Tobacco and Tobacco Products Sub-Group
on Pesticides Residues.
[34] M. Krigbaum, Evaluation of Automated Solid Phase Extraction of Agro-
chemical and Industrial Organic Compounds from Drinking Water using
USEPA Method 525.2, Weber Consulting (http://www.weber.hu/PDFs/
SPE/EPA Method525 2AutomatedSPEAgrochemicals.pdf).
[35] P. Van Toan1, Z. Sebesvari1, I. Rosendahl, et al., Residues of Recently Used
Pesticides in Water of the Mekong Delta, Vietnam, United Nations University,
Institute for Environment and Human Security, Bonn, Germany, 2009.
[36] EPA Method 525.2, U.S. Environmental Protection Agency, OHIO 45268.
[37] 2002/657/EC: Commission Decision of 12 August 2002 implementing Council
Directive 96/23/EC concerning the Performance of Analytical Methods and the
Interpretation of results (notied under document number C (2002) 3044).
[38] Document No. SANCO/10684/2009: Method Validation and Quality Control
Procedures for Pesticides Residues Analysis in Food and Feed, Supersedes Doc-
ument No. SANCO/3131/2007 Implemented by 01/01/2010.
[39] J. Mark Green, Anal. Chem. 68 (May) (1996) 305A.
[40] G. Boatto, M. Nieddu, G. Dessi, et al., J. Chromatogr. A 1159 (2007) 198.
[41] M. Nieddu, G. Boatto, G. Dessi, J. Chromatogr. B 1159 (2007) 578.
[42] Compendium of International Methods of Analysis OIV Estimation of the
detection and quantication limits of a method of analysis (Resolution Oeno
7/2000).
[43] 2002/657/EC: Commission Decision of 12 August 2002 implementing Council
Directive 96/23/EC concerning the performance of analytical methods and the
interpretation of results (notied under document number C (2002) 3044).
[44] CEN/TC 392 N73 NWI GC/MS method for the identication and assay of 12
phthalates in ready to inject cosmetic samples.