Documente Academic
Documente Profesional
Documente Cultură
Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma
a r t i c l e
i n f o
Article history:
Received 15 June 2011
Received in revised form 1 August 2011
Accepted 4 August 2011
Available online 10 August 2011
Keywords:
Pressurized liquid extraction
Simultaneous extraction and derivatization,
GCMS
Bakery foods
Chloropropanols
BSTFA
a b s t r a c t
3-Chloropropane-1,2-diol (3-MCPD) and 1,3-dichloro-2-propanol (1,3-DCP) were determined for the rst
time in bakery foods using pressurized liquid extraction (PLE) combined with in situ derivatization and
GCMS analysis. This one-step protocol uses N,O-bis(trimethylsilyl)triuoroacetamide (BSTFA) as silylation reagent. Initially, screening experimental design was applied to evaluate the effects of the variables
potentially affecting the extraction process, namely extraction time (min) and temperature ( C), number
of cycles, dispersant reagent (diatomaceous earth in powder form and as particulate matter with high
pore volume Extrelut NT) and percent of ush ethyl acetate volume (%). To reduce the time of analysis
and improve the sensitivity, derivatization of the compounds was performed in the cell extraction. Conditions, such as the volume of BSTFA, temperature and time for the in situ derivatization of analytes using
PLE, were optimized by a screening design followed to a Doehlert response surface design. The effect of
the in-cell dispersants/adsorbents with diatomaceous earth, Florisil and sodium sulfate anhydrous was
investigated using a BoxBehnken design. Using the nal best conditions, 1 g of sample dispersed with
0.1 g of sodium sulfate anhydrous and 2.5 g diatomaceous earth was extracted with ethyl acetate. 1 g
of Florisil, as clean-up adsorbent, and 70 L of BSTFA were used for 3 min at 70 C. Under the optimum
conditions, the calibration curves showed good linearity (R2 > 0.9994) and precision (relative standard
deviation, RSD 2.4%) within the tested ranges. The limits of quantication for 1,3-DCP and 3-MCDP,
1.6 and 1.7 g kg1 , respectively, are far below the established limits in the European and American legislations. The accuracy, precision, linearity, and limits of quantication provided make this analytical
method suitable for routine control. The method was applied to the analysis of several toasted bread,
snacks, cookies and cereal samples, none of which contained chloropropanols at concentrations above
the legislation levels.
2011 Elsevier B.V. All rights reserved.
1. Introduction
3-Chloropropane-1,2-diol (3-MCPD) and 1,3 dichloro-2propanol (1,3-DCP), the best-known chloropropanols, are
contaminants detected in acid-hydrolyzed vegetable proteins
(acid-HVP) as well as in foods such as bakery products with
components that are not subjected to acid hydrolysis during their
manufacture [16]. 3-MCPD esters are now found widespread
in thermally processed foodstuffs [7]. Both 3-MCPD and 1,3-DCP
were recently set by IARC into group 2B as probably carcinogenic
to humans [810]. The Food Chemical Codex (FCC) limited the
presence of 3-MCPD and 1,3-DCP in acid-HVP to not more than 1
Corresponding authors. Tel.: +34 981 563 100; fax: +34 981 547 141.
E-mail addresses: rosaantonia.lorenzo@usc.es (R.A. Lorenzo), tuchi.carro@usc.es
(A.M. Carro).
1
These authors have contributed equally to this work.
0021-9673/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2011.08.004
6879
injected in the splitless mode, and then the splitter was opened
after 2 min. The GC oven temperature were programmed as follows:
initial temperature 50 C for 2 min; then increased at 3 C min1 to
100 C and nally from 100 to 280 C at a rate of 40 C min1 (kept
for 10 min). The injector temperature was hept constant at 280 C.
The total run time of analysis was 33 min.
The mass spectrometer was operated in electron impact (EI),
70 eV of ion energy, with a 10 min solvent delay. The temperatures of the interface, EI ionization source and the quadruple mass
analyzer were set at 280 C, 230 C and 150 C, respectively.
Data acquisitions were carried out over the mass range m/z
60400, in SIM mode (Single Ion Monitoring) to maximize sensitivity. To identify the compounds 445 L of 1 g mL1 standard
solution of each individual analyte was introduced into an amber
vial, 55 L of derivatization reagent was added, the mixture was
shaken vigorously in a vortex and heated to 60 C for 70 min for the
derivatization reaction. Then 1 L was injected in the CG/MS to nd
the higher m/z and characteristic ions for each compound. Specic
conditions for each analyte are listed in Table 1. Quantication was
accomplished by the relative areas vs. internal standards, which
were added to samples at the spiked step. For analyzing chloropropanols by gas chromatography if is necessary to carry out this
kind of reaction in order to modify their low volatility and high
polarity and improve the sensibility [8,31]. Trimethylsilyl (TMS)
derivatives were prepared according to a previous work dealing
with the BSTFA derivatization of chloropropanols [32]. The reaction
was carried out with standard solutions in ethyl acetate, demonstrating the suitability of this solvent to accomplish derivatization.
The TMS derivatives were stable for more than one month.
2.4. Extraction and derivatization procedure
Extractions were accomplished using a pressurized liquid
extractor ASE 200 from Dionex (Sunnyvale, CA, USA), equipped
with a 24-sample carousel, 11 mL stainless steel cells, and 40 mL
collection vials. Two cellulose lters (Dionex) were placed at the
bottom of the PLE cell And aA layer of the adsorbent clean-up (0.6 g
of Florisil) was placed on the lters. 1 g sample was mixed with
1.5 g of anhydrous sodium sulfate (drying agent) and 2 g of DE (dispersant agent), using a mortar and a pestle, and the mixture was
introduced into the cell lling the remaining volume with clean
sand (5070 mesh particle size, SigmaAldrich). Then, another cellulose lter was placed on the top. In all experiments, isotopically
labelled analogues (30 ng mL1 ) were added as surrogate standards
to each sample before extraction. The extraction pressure was set
to 1500 psi and ethyl acetate was employed as extraction solvent.
The cell was tightly closed and placed into the carousel of the ASE
system. Extractions were performed by preheating the cell before
lling with the solvent (preheat method). Analytes were recovered in one extraction cycle of 5 min, at 130 C. The ush volume
of ethyl acetate was 80% and the purge time was set to 90 s. The
PLE extract (ca. 20 mL), was collected in vials sealed with rubber
septa, evaporated using a gentle stream of nitrogen in a Turbo Vap
II concentrator (Zymark, Hopkinton, MA, USA) and adjusted to a
nal volume of 1 mL. 445 L of extract was derivatized with BSTFA
as described in Section 2.1 and 1 L was analyzed by GC/MS.
In the simultaneous extractionderivatization experiments,
70 L of BSTFA was added to the sample before the addition of
the drying agent (0.1 g of anhydrous sodium sulfate) to avoid the
susceptibility towards hydrolysis of silylation agents [29,33] and
the dispersion sorbent (2.5 g of DE). 1 g of Florisil, was added as
clean-up adsorbent. Thereafter, the PLE procedure was carried out.
Analytes were recovered in one extraction cycle of 3 min, at 70 C.
The ush volume and the purge time were as previously described.
The PLE extracts were collected and evaporated in a Turbo Vap II
concentrator and adjusted to a nal volume of 1 mL. The extract
6880
Table 1
Retention times (tR ), quantier ion and qualier ions for the chloropropanols by GCMS.
d5 -1,3-DCP-TMS
1,3-DCP-TMS
d5 -3-MCPD-TMS
3-MCPD-TMS
154
93.185
11.75
151
93.190
12.00
116
239
17.45
119
244
17.85
6881
Fig. 1. Graphics of total effects obtained using a screening design 21 32 //9, for 3-MCPD (A) and 1,3-DCP (B). Global desirability response surface plot using a Doehlert design
(C) and a BoxBenhken design (D).
Concentration of the IS along the calibration curve was maintained constant at 30 ng mL1 for d5 -1,3-DCP and for d5 -3-MCPD.
The response functions were found to be linear with determination coefcient (R2 ) higher than 0.999. Limits of detection (LODs) and quantication (LOQs) of the overall method
(extractionderivatization by PLE following GCMS) were calculated as the concentrations giving a signal-to-noise ratio of 3 or
10 (S/N = 3 or 10), respectively according to the European norm EN
14573 [19] and the Eurachems recommendations [38]. LODs were
0.5 g kg1 for both chloropropanols while LOQs were 1.6 g kg1
for 1,3-DPC and 1.7 g kg1 for 3-MCPD. These values are lower
than the quantication limit of 3 g kg1 for both chloropropanols
in soy sauce and other food products reported by Abu-El-Haj et al.
[39] using a clean-up alumina column, dichloromethane extraction,
heptauorobutyryl anhydride derivatization, and isotope dilution
GCMS. Moreover, the obtained limits are much lower than those
established in the legislation [11,12] and than the limits proposed
by the European Commission for the ofcial control of the levels of
3-MCPD in food samples (LOD < 5 g kg1 and LOQ <10 g kg1 )
[40]. On the other hand, in order to evaluate the precision of
the proposed method, six independent analyses were performed
using spiked toasted bread samples at two concentration levels (20 ng g1 and 2 g g1 for both chloropropanols). The results
obtained showed relative standard deviation (RSD%) of 2.4 and
0.9% for 20 ng g1 and 2 g kg1 , respectively for 1,3-DCP and 1.4
and 1.03% for 20 g kg1 and 2 mg kg1 ,respectively, for 3-MCPD.
These values are fully compatible with the requirements established by the European Commission for 3-MCPD [40]. Accuracy
studies, expressed as % of recovery, were carried out by applying
the optimized extractionderivatization by PLEGCMS method to
the extraction of a real bakery samples. Due to the possible matrix
6882
Table 2
Extraction recoveries (n = 5 replicates) in fortied (20 ng g1 ) bakery samples by extractionderivatization by PLEGCMS. Results of the determination of 1,3-DCP-TMS and
3-MCPD-TMS in bakery samples (n = 4 replicates) are calculated for dry weight.
Sample*
1,3-DCP-TMS
3-MCPD-TMS
Concentrations of 1,3-DCP
nonspiked bakery samples
(concentration, ng g1 SD)
Concentrations of 3-MCPD
nonspiked bakery samples
(concentration, ng g1 SD)
% Recovery
RSD%
% Recovery
RSD%
91.8
97.8
94.1
5.2
1.2
1.5
93.7
109
86.2
1.6
3.2
1.6
37.0 2.9
43.9 1.8
23.4 1.5
3.6 0.2
11.8 1.3
<LOQ
94.1
6.3
91.7
2.0
12.8 0.4
n.d.
13.4 0.1
99.28 2.9
2.8 0.1
6.8 0.7
n.d.
4.5 0.1
4. Conclusions
A
counts
100%
260
93.0
75%
220
50%
180
1,3-D C P -T MS
151.0
25%
140
185.0
0%
90
100
110
130
150
170
190 m/z
60
20
9.50
75%
counts
160
140
120
100
80
60
40
cereals, toasted bread, snacks and biscuits, were analyzed in quadruplicate and quantication was accomplished using the standard
addition method (Table 2). The presence of chloropropanols was
evidenced in the samples analyzed, but the levels found did not
exceed the limit regulated for soy sauces and related products
[11,12]. These results are in agreement with the fact that the moisture content of bakery products inuences the chloropropanol
formation process, since glycerol appears to be a precursor in foods
with a low water content (<15%) [41]. Values shown in Table 2,
are in the same order of magnitude as those obtained for bakery
products reported by Hamlet [8] and Leon et al. [20]. Although
the information on 1,3-DCP in retail foods is scarce at present,
data from different reports suggest that 1,3-DCP may be present
in retail ready-to-eat processed foods including bakery products
without 3-MCPD and that can be related with the presence of
dimethylamineepichlorohydrin copolymer used in sugar rening
[42] at levels substantially higher than those of 3-MCPD [8]. Fig. 2
shows the SIM GCMS chromatogram obtained for a snack sample.
The spectrum obtained by MS conrmed the identity of the analytes
detected in the sample (Fig. 2).
50%
3-MC P D -T MS
25%
239.0
0%
The suitability of the PLE technique for the extraction of chloropropanols from bakery foods has been demonstrated for rst time.
The use of DE as sample dispersant and Florisil as clean-up sorbent, combined with simultaneous in-cell derivatization, allowed
to obtain a ready to inject GC extract providing recoveries, better
than the derivatization out of the cell in a longer step. The main
advantages of the proposed methodology are the rapid and automated sample pre-treatment and the reduction of the sample size
and handling. The sensitivity of the method is suitable for the legislation requirements. The results obtained will be useful for future
studies aimed to determine the presence of chloropropanols in
foods, in the total dietary intake, and their persistence and toxicity
taking into account that it is unknown whether they are transformed into the human body and degraded or subjected to other
processes. Application of the method to other matrices, such as
biological ones, would also provide an interesting insight into the
uptake and bioaccumulation of chloropropanols in tissues or body
uids.
Acknowledgments
6883