Quantification of 3-monochloropropane-1,2-diol esters in edible oils by large-volume injection coupled to comprehensive gas chromatography-time-of- flight mass spectrometry

Article abstract of DOI:10.1002/cplu.201300351

3-Monochloropropane-1,2-diol (3-MCPD) esters are a group of process contaminants formed during the refining of edible oils and fats. A method for the determination of 3-MCPD esters in such oils and fats based on large-volume injection-comprehensive gas ch

Full text of DOI:10.1002/cplu.201300351

CHEMPLUSCHEM
COMMUNICATIONS
DOI: 10.1002/cplu.201300351
Quantification of 3-Monochloropropane-1,2-diol Esters in
Edible Oils by Large-Volume Injection Coupled to
Comprehensive Gas Chromatography–Time-of-Flight Mass
Spectrometry
Sjaak de Koning*[a]
3-Monochloropropane-1,2-diol (3-MCPD) esters are a group of
process contaminants formed during the refining of edible oils
and fats. A method for the determination of 3-MCPD esters in
such oils and fats based on large-volume injection–compre-
hensive gas chromatography–time-of-flight mass spectrometry
(LVI–GCꢀGC–ToF MS) is presented. The simplified method for
sample preparation consists of alkaline hydrolysis followed by
extraction of 3-MCPD from the lipid matrix and derivatization
using phenylboronic acid. The limit of detection was
0.8 mgkgꢀ1, reported as free 3-MCPD. The repeatability was
better than 2.7% relative standard deviation for levels around
0.5 mggꢀ1 3-MCPD. The GCꢀGC–ToF MS system was shown to
be stable over 100 analyses.
the analysis of 3-MCPD esters are the coelution of compounds
of interest with large amounts of matrix constituents, the sen-
sitivity, and system stability.
The current methods for 3-MCPD ester analysis in edible oils
and fats actually measure the total 3-MCPD content of the oil
or fat after hydrolysis. The procedures consist of a number of
subsequent steps starting with hydrolysis, the removal of the
fatty acids (as their methyl esters), extraction of the free
3-MCPD with salting out, derivatization with phenylboronic
acid, preconcentration by solvent evaporation, and finally GC–
MS analysis.[5] Deuterium-labeled [D5]3-MCPD or esters thereof
are used as internal standards. Potential problems in the pro-
cedure are 1) degradation of the 3-MCPDs during (alkaline) hy-
drolysis resulting in higher detection limits, 2) formation of ad-
ditional 3-MCPDs is possible if chloride salts are used in the
salting out extraction steps, and 3) the stability of the mass
spectrometer owing to strong source contamination. Limits of
detection (LODs) are in the range of 0.5 ppm 3-MCPD ester.
Several studies have been published in which large-volume
injection (LVI) methods were used for the GC determination of
trace pollutants.[6] The LVI technique enables significant im-
provement of the sensitivity of the analytical methods. Rather
than using splitless injections of 1–2 mL, with LVI it is possible
to inject sample volumes of over 100 mL. Another reason to
use LVI can be to simplify sample preparation, for example, by
taking out concentration steps such as solvent evaporation or
salting out.
3-Monochloropropane-1,2-diol (3-MCPD) is carcinogenic, highly
suspected to be genotoxic in humans, has male antifertility ef-
fects, and is a chemical byproduct that may be formed in
foods. It is primarily created in foods by protein hydrolysis
through adding hydrochloric acid to speed up the reaction of
(soy) protein with lipids at high temperatures. In another
method, 3-MCPD can also occur in foods that have been in
contact with materials containing epichlorohydrin-based wet-
strength resins—used in the production of some tea bags and
sausage casings. It has been found in some East Asian and
Southeast Asian sauces, such as oyster, hoisin, and soy sauces.
The use of hydrochloric acid rather than traditional slow fer-
mentation is a far cheaper and faster method but unavoidably
creates carcinogens.
In 2006 Zelinkovꢁ et al.[1] reported the detection of 3-MCPD
fatty acid esters (3-MCPD esters) in edible oils. In native or un-
refined fats and oils, no or only traces of 3-MCPD esters were
detectable,[2,3] but in nearly all refined fats and oils, concentra-
tions of 3-MCPD esters in the range of 0.2–20 mgkgꢀ1 are pres-
ent. There are several methods available for the determination
of 3-MCPD esters, from which gas chromatography coupled to
mass spectrometry (GC–MS) is the most common technique.[4]
Key challenges when using chromatographic separation for
About a decade ago, a new chromatographic technique for
the characterization of complex samples became commercially
available: comprehensive two-dimensional gas chromatogra-
phy (GCꢀGC), first reported by Phillips et al.[7] GCꢀGC has
a much increased peak capacity and offers significantly im-
proved detection limits through chromatographic optimiza-
tion.[8,9] Owing to the high peak capacity and the numerous
compounds that are resolved in a GCꢀGC separation, the use
of a mass spectrometer is highly desirable for identification
and confirmation purposes. Dallꢂge et al.[10] reported that only
MS instruments that can acquire a minimum of 50 full spectra
per second allow reliable identification, and subsequent quan-
tification, of the classical narrow peaks in the two-dimensional
chromatogram. At present, time-of-flight mass spectrometry
(ToF MS) is the method of choice because it provides full mass
range spectra at high data acquisition rates.
[a] Dr. S. de Koning
Da Vinci Laboratory Solutions
Cairostraat 10, 3047 BC Rotterdam (The Netherlands)
E-mail: sjaak.de.koning@davinci-ls.com
Part of a Cluster Issue on “Two-Dimensional Gas Chromatography“.
To view the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cplu.v79.6/issuetoc.
Herein, a feasibility study is presented that focuses on the
use of LVI coupled to GCꢀGC–ToF MS for efficient, more relia-
ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 776 – 780 776

CHEMPLUSCHEM
COMMUNICATIONS
www.chempluschem.org
Figure 1. TIC chromatogram of 25 mL palm-oil extract by LVI–GC–ToF MS. The 3-MCPD derivative shows strong coelution with matrix compounds and is not
visible in the chromatogram.
ble, and more sensitive 3-MCPD ester analysis in edible oils
and fats.
MS analysis of a palm-oil extract. The arrow in Figure 1 points
to the retention time at which the 3-MCPD derivative elutes.
However, it immediately becomes clear that the 3-MCPD deriv-
ative coelutes with a lot of matrix. As expected the standard
3-MCPD derivative and the labeled [D5]3-MCPD derivative coe-
lute. In the next set of experiments the analysis was repeated
using LVI–GCꢀGC–ToF MS. Figure 2 shows the TIC contour
The analytical method as described in the literature[5] is
a single-dimension GC procedure. To be able to compare this
procedure with that of the comprehensive GC approach, first
a run was performed using GC–ToF MS. Figure 1 shows the
total-ion-current (TIC) chromatogram of a 25-mL LVI–GC–ToF
Figure 2. TIC contour plot of 25 mL palm-oil extract by GCꢀGC–ToF MS. The 3-MCPD derivative (inside red circle) is very well separated from the matrix com-
pounds.
ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 776 – 780 777

CHEMPLUSCHEM
COMMUNICATIONS
www.chempluschem.org
plot of the LVI–GCꢀGC–ToF MS analysis and the peak high-
lighted in the red circle is that of the 3-MCPD derivative.
Here it must be mentioned that despite the chirality of
3-MCPD, only one peak is seen. The selected columns for both
the first- and second-dimension separation do not provide
chiral separation. The peak for 3-MCPD ester contributes the
total amount of all stereoisomers of 3-MCPD. By comparing
the GC–ToF MS and GCꢀGC–ToF MS chromatograms, one can
immediately see that the high baseline at which the 3-MCPD
derivative elutes originates from tailing of
Table 1. Quantitative analysis of 3-MCPD.
Sample
Weight [mg]
Area
Area
3-MCPD
Amount
[a]
[D5]3-MCPD
3-MCPD [mggꢀ1
]
A
B
C
B[b]
100.2
102.4
101.3
104.4
13121818
18810221
16923181
103775824
1979261
9301420
5706487
90231127
0.15
0.48
0.33
0.83
[a] Reported as free 3-MCPD. [b] NaCl used in the sample preparation.
a major, unspecified matrix compound. The
peakarea3 ꢀ MCPD
amount½D5ꢁ3 ꢀ MCPD ðmgÞ
sampleweight ðgÞ
peak at about 1000 seconds that is strongly
ð1Þ
Amount3 ꢀ MCPD ¼
ꢂ
peakarea½D5ꢁ3 ꢀ MPCD
tailing is from the derivatization agent,
which was used in excess. The signal at
Additionally, one of the samples, the sample coded B, was
also prepared using NaCl during the sample preparation. On
comparing the data from the two techniques it could be con-
cluded, based on the peak area of [D5]3-MCPD, that the extrac-
tion efficiency is about five times better when using NaCl.
However, the reported amount of 3-MCPD is approximately
double, most likely because of formation of 3-MCPD esters
during sample preparation through the influence of chloride
ions. Based on these results, it can be concluded that it is
better not to use NaCl during sample preparation from the
perspective of method precision, even though this results in
a loss of method sensitivity. Fortunately, this loss in sensitivity
can be easily overcome if using LVI. In this method, 25 mL of
extract is injected compared to 1 mL in the standard method.
To get an impression of the LOD and the limit of quantifica-
tion (LOQ), the extract of palm oil sample 123 was injected ten
times. The analytical data are shown in Table 2. This resulted in
an average reported amount of 0.47 mggꢀ1 of 3-MCPD with an
average signal-to-noise ratio (S/N) of 1767. Although it is not
about 2000 seconds could not be identified owing to the satu-
ration of the mass spectrum.
For quantification the extract is spiked with a known
amount of [D5]3-MCPD, thus allowing quantification by stable
isotope dilution. However, a drawback of this type of quantifi-
cation is that the 3-MCPD and [D5]3-MCPD derivatives coelute
in the first and even almost in the second dimension of the
GCꢀGC separation (see Figure 3; for an explanation of the
ˇ
slightly shorter retention time of the [D5] derivative, see ZꢁÅek
et al.[11]). The resulting mixed spectrum makes proper identifi-
cation, at least of the 3-MCPD derivative, impossible. However,
a powerful advantage of using ToF MS is the ability to perform
deconvolution of the mass spectra. Despite the (almost) coelu-
tion of the 3-MCPD and the [D5]MCPD derivatives, pure spectra
of the analytes can be reconstructed (see Figure 4), now
making clear identification of both compounds possible.
Using this strategy, three different palm oil samples were an-
alyzed (for results see Table 1). The amounts of 3-MCPD are cal-
culated using Equation (1):
Figure 3. Zoomed extracted ion current of base slice of the 3-MCPD derivative (m/z 147, orange), coeluting with the [D5]3-MCPD derivative (m/z 150, green).
ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 776 – 780 778

CHEMPLUSCHEM
COMMUNICATIONS
www.chempluschem.org
Figure 4. Deconvoluted peak true spectra of the 3-MCPD derivative (top) and [D5]3-MCPD derivative (bottom).
an exact calculation, it was estimated that the S/N behaves lin-
early with the amount of 3-MCPD. Based on this estimation
the LOD=(3/1767)ꢀ0.47=0.00080 mggꢀ1 and the LOQ=(10/
1767)ꢀ0.47=0.00267 mggꢀ1. The % relative standard deviation
(RSD) for the calculated amount of 3-MCPD was found to be
2.7%.
tion, a total of 100 analyses were performed. After 100 runs
the following data were obtained:
*
Weight (g):
0.1024
17579599
8551150
0.48
*
*
*
*
Area [D5]3-MCPD:
Area 3-MCPD:
Amount 3-MCPD (mggꢀ1):
S/N 3-MCPD:
System stability is another topic of interest when using the
modified method. It is known from the field that this is often
a problem and that the ion source of, for example, quadrupole
MS instruments requires frequent maintenance and cleaning,
usually after only some ten analyses. To test the Pegasus
system stability using the demonstrated method/instrumenta-
1798
From these results it can be concluded that after 100 analy-
ses, the system was still performing like in the first run.
ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 776 – 780 779

CHEMPLUSCHEM
COMMUNICATIONS
www.chempluschem.org
the extract by solvent evaporation (not needed in the case of LVI
analysis), the n-hexane layer was separated and injected into the
GCꢀGC–ToF MS system.
Table 2. Analytical data for 3-MCPD analysis.
Sample[a]
Area
Area
Amount
3-MCPD [mggꢀ1
S/N
3-MCPD
[b]
[D5]3-MCPD
3-MCPD
]
B:1
B:2
B:3
B:4
B:5
B:6
B:7
B:8
B:9
B:10
18810221
20186242
20314729
19742597
19607596
19160417
18324274
18015697
21695114
19660950
9301420
9268336
9837198
9622891
9287701
9154108
8998403
8450171
10796004
9816324
0.48
0.45
0.47
0.48
0.46
0.47
0.48
0.46
0.49
0.49
1548
1634
1661
1940
1866
1589
1613
1615
2077
2125
System parameters
The system used for sample analysis was a Pegasus 4D GCꢀGC–
ToF MS instrument (LECO, St. Joseph, MI, USA), equipped with
a quad-jet thermal modulator (LECO), a second-dimension oven
(LECO), and a CIS4 programmed-temperature vaporizing injector
(Gerstel, Mꢂlheim a/d Ruhr, Germany) containing a baffled glass
liner. The Pegasus 4D instrument was controlled by ChromaTOF
(LECO) data acquisition and processing software. The CIS4 injector
was controlled by Maestro (Gerstel) software. The first-dimension
column was an Rxi-1 SIL ms column of size 30 mꢀ0.25 mm with
a film thickness of 0.25 mm (Restek, Bellefonte, PA, USA) and the
second-dimension column was an Rxi-17 SIL ms 1 mꢀ0.18 mm
column with a film thickness of 0.18 mm (Restek). Sample volumes
of 25 mL were injected at a temperature of 408C. After solvent
evaporation, the injector was switched to splitless and heated to
3008C at 58Csꢀ1. The first-dimension GC oven started at a tempera-
ture of 408C, at which it was held for 1 min. Following this, it was
heated to 1908C at 68Cminꢀ1 and then to 2808C (30 min hold) at
208Cminꢀ1. The second-dimension oven was programmed follow-
ing the first-dimension oven with an offset of +58C. The modula-
tion time was set to 4 s. Data were acquired in the range of 50–
500 m/z using an acquisition rate of 200 spectrasꢀ1. Helium was
Average
SD
%RSD
0.47
0.01
2.74
1767
LOD
LOQ
0.00080
0.00267
[a] Weight was 102.4 mg for each sample. [b] Reported as free 3-MCPD.
Experimental Section
Sample preparation
The homogenized sample (ca. 100 mg) was weighed into a screw-
capped tube, and tert-butyl methyl ether/ethyl acetate (8:2 v/v,
500 mL), internal standard solution (250 mL, free [D5]3-MCPD,
20 mgmLꢀ1 in tert-butyl methyl ether), and 0.5m NaOCH3 solution
(1 mL) were added. After 10 min of gentle shaking, isohexane
(3 mL), glacial acetic acid (100 mL), and NaCl solution (3 mL,
200 gLꢀ1) were added (the addition of NaCl solution was omitted
for the LVI analysis). After 1 min of agitation, the upper layer was
removed with a pipette and discarded. The aqueous layer was ex-
tracted with a new portion of isohexane; the upper layer was
again removed and discarded.
used as the carrier gas at a flow rate of 1 mLminꢀ1
.
Keywords: esters · fats and oils · gas chromatography · mass
spectrometry · sample preparation
ˇ
ˇ
[1] Z. Zelinkovꢁ, B. Svejkovskꢁ, J. Velꢄsek, M. Dolezal, Food Addit. Contam.
2006, 23, 1290.
[2] R. Weißhaar, Eur. J. Lipid Sci. Technol. 2008, 110, 183.
[3] W. Seefelder, N. Varga, A. Struder, G. Williamson, F. P. Scanlan, R. H. Sta-
dler, Food Addit. Contam. 2008, 25, 391.
[4] S. Abu-El-Haj, M. J. Bogusz, Z. Hassan, M. Al Tufail, Food Control 2007,
18, 81.
[5] Ester-bond 3-chloropropane-1,2-diol (3-MCPD esters) and 3-MCPD form-
ing substances, DFG Standard Method Section C-Fats, C-III 18(09).
[6] S. de Koning, M. van Lieshout, H.-G. Janssen, U. A. Th. Brinkman, J. Mi-
crocolumn Sep. 2000, 12, 153.
For derivatisation (see Scheme 1), derivatisation reagent (250 mL;
5 g phenylboronic acid dissolved in 19 mL acetone and 1 mL
water) was added to the aqueous phase.[2] Next the tube was
[7] J. B. Phillips, Z. Liu, J. Chromatogr. Sci. 1991, 29, 227.
[8] Ph. Marriott, R. Shellie, J. Ferguson, R. Ong, P. Morrison, J. Flavour Fragr.
2000, 15, 225.
[9] C. Meinert, U. J. Meierhenriech, Angew. Chem. 2012, 124, 10610; Angew.
Chem. Int. Ed. 2012, 51, 10460.
Scheme 1. Derivatization reaction of 3-MCPD.
[10] J. Dallꢂge, J. Beens, U. A.Th. Brinkman, J. Chromatogr. A 2003, 1000, 69.
ˇ
[11] P. ZꢁÅek, J. Kindl, O. Hovora, I. Valterovꢁ, unpublished results.
heated to 808C for 20 min under gentle shaking. After cooling to
room temperature, the phenylboronate derivative of 3-MCPD was
extracted by shaking with n-hexane (3 mL). After concentration of
Received: October 14, 2013
Published online on January 29, 2014
ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2014, 79, 776 – 780 780