RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 2004; 18: 19891994

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rcm.1582

Determination of free fatty acids in chocolate by liquid

chromatography with tandem mass spectrometry
Daniela Perret*, Alessandra Gentili, Stefano Marchese, Manuel Sergi and Lidia Caporossi
Laboratorio Chimico per la Sicurezza, Dipartimento di Chimica, Universita La Sapienza, Piazzale Aldo Moro 5, P.O. Box 34, Posta 62,
00185 Roma, Italy
Received 15 April 2004; Revised 8 July 2004; Accepted 8 July 2004

This paper describes a rapid extraction method, based on a matrix solid-phase dispersion technique
using diatomaceous earth as solid support and 50:50 (v/v) chloroform/methanol as extracting
solvent, that can determine 11 free fatty acids in chocolate. The extraction procedure is followed
by reversed-phase liquid chromatography/tandem mass spectrometry (LC/MS/MS) using a nor-
mal-bore (4.6 mm i.d.) C-18 column and an electrospray interface operating in the negative ion
mode. The tandem mass spectra of selected compounds show that charge-remote fragmentation
(CRF) mechanisms are occurring; the intensities of the CRF reactions increase with the carbon num-
ber and degree of unsaturation of the fatty acids. Average recoveries, evaluated by the standard
addition method, vary between 79103%, and the estimated quantification limits are less than
153 ng/g. The proposed method has been used to analyse nine chocolate samples from various price
ranges, bought from supermarkets. Copyright # 2004 John Wiley & Sons, Ltd.

Chocolate is made from different components of cocoa beans
after processing cocoa pods from the cocoa tree (Teobroma
cacao) that have been harvested and the beans removed and
fermented. Chocolate liquor is prepared by finely grinding
the cocoa beans and is the basis for all chocolate products.
Cocoa powder is made by removing part of the cocoa butter
from the liquor; cocoa butter (CB) is a yellowish fat solid at
room temperature and is obtained by hydraulic pressing of
the treated beans. The pressed butter has a distinctive flavour
and is used directly for making chocolate, together with sugar,
other flavours (like vanilla), and often milk (in milk chocolate).
A directive of the European Community (EU Directive
2000/36/CE) allows the use of other vegetable fats, the
cocoa butter equivalents (CBEs), in addition to cocoa butter
for the production of chocolate. The CBEs are individual
mixtures of vegetable fats of tropical trees (palm, shea, illipe,
sal, kokum, mango) not containing lauric acid, with chemical
and physical properties similar to those of CB.1 4 CBEs may
be used for technological and economic reasons but they
cannot exceed 5% of the final product.
The quantification of various fats in chocolate is of great
interest to research and development laboratories involved in
process control and quality control during manufacture.
The main components of fat in chocolate are triglycerides
and fatty acids, followed by minor components such as
tochopherols, trienols and sterenes.3 Most of the fatty acids
are bound as esters, with only small amounts occurring as
*Correspondence to: D. Perret, Laboratorio Chimico per la
Sicurezza, Dipartimento di Chimica, Universita La Sapienza,
Piazzale Aldo Moro 5, P.O. Box 34, Posta 62, 00185 Roma, Italy.
E-mail: daniela.perret@uniroma1.it
non-esterified or free fatty acids (FFAs). The fatty acid
composition in chocolate includes saturated and unsaturated
compounds; the major fatty acids, including palmitic (16:0,
2530%), stearic (C18:0, 3137%), oleic (C18:1, 3138%) and
linoleic (C18:2, 25%), are accompanied by small amounts of
other saturated and polyunsaturated C14 C22 acids.
Several methods, including high-performance liquid
chromatography (HPLC) with detection by UV,5 and gas
chromatography (GC) combined with flame ionisation
detection (FID),48 have been proposed for detecting fatty
acids in chocolate, in CB and in CBEs. The majority of fatty
acid profiles are determined by GC of the fatty acid methyl
esters (FAMEs) employing capillary columns; the most
common derivatisation methods involve base-catalysed
transesterification of lipid-bound fatty acids with sodium
methoxide and/or esterification of the FFAs to FAMEs, after
preliminary fractionation into lipid classes. In contrast with
GC, use of HPLC allows the fatty acids to be analysed as
underivatised compounds9 or converted to a large number
of different derivatives;6 of the multitude of derivatives
reported, phenacyl esters and their substituted analogs are
used most frequently.
The recommended methods for the quantitative extraction
of fat from chocolate are labour-intensive and require large
amounts of solvent and time. For example, the method
recommended by AOAC10 utilises an acid digestion step
followed by Soxhlet extraction with petroleum ether. The
replacement of Soxhlet for the extraction of fat has been
largely applied in the field of environmental contaminants
and, in more recent years, in the field of food analysis.11
In particular, more rapid alternative methods, such as
microwave extraction and supercritical fluid extraction

Copyright # 2004 John Wiley & Sons, Ltd.

1990 D. Perret et al.
(SFE), have been applied for fat extraction from foodstuffs
and other biological products.11
The aim of the present work was to develop an analytical
method, based on matrix solid-phase dispersion (MSPD) and
determination by liquid chromatography with tandem mass
spectrometry (LC/MS/MS), for the identification of the
major free fatty acids (saturated and polyunsaturated) in
samples of dark and milk chocolate. MSPD is a very fast and
simple technique for food extraction; many authors have
demonstrated the applicability of this procedure to a large
number of fruits, vegetables and animal tissues, using
different solid supports.1217 A fine dispersion of the sample
matrix onto a solid support such as silica, alumina,
diatomaceous earth, C-18-bonded silica, or other sorbents,
is obtained by blending the sample and the sorbent with
mortar and pestle. After blending, this material is packed into
a minicolumn and the analytes are eluted by a suitable
extractant. Compared with the classical sample treatment
procedures, MSPD offers several advantages including
simplification of the analytical protocol and the reduction
of solvent consumption.
Therefore, the objective of the current study was to
investigate the possibility of application of the MSPD
procedure, using diatomaceous earth as solid support, to
propose an alternative fat extraction method that is both
economical and time-saving.
EXPERIMENTAL
Chemicals and reagents
Tetradecanoic acid (myristic acid, C14:0), hexadecanoic
acid (palmitic acid, C16:0), octadecanoic acid (stearic acid,
C18:0), eicosanoic acid (arachidic acid, C20:0), docosanoic
acid (behenic acid, C22:0), cis-9-hexadecenoic acid (palmito-
leic acid, C16:1), cis-9-octadecenoic acid (oleic acid, C18:1),
cis,cis-9, 12-octadecadienoic acid (linoleic acid, C18:2), cis,
cis,cis-9, 12, 15-octadecatrienoic acid (linolenic acid,
C18:3), cis-11-eicosenoic acid (arachidenic acid C20:1) and cis,-
cis,cis, cis-5,8,11,14-eicosatetraenoic acid (arachidonic acid,
C20:4) were purchased from Labservice Analytica (Bologna,
Italy). Stock solutions of the individual standards were pre-
pared by dissolving each analyte in chloroform at 1 mg/mL;
composite working standard solutions were prepared by
mixing appropriate amounts of each standard solution and
diluting in methanol. All standard solutions were stored
at
188C before use.
Chloroform, dichloromethane, acetone, formic acid, and
ammonia were of analytical grade and were purchased from
Carlo Erba (Milan, Italy). For use in LC, distilled water was
further purified by passage through a Milli-Q Plus apparatus
(Millipore, Bedford, MA, USA). Methanol-plus for HPLC
was obtained from Carlo Erba. The diatomaceous earth
Spe-ed Matrix was purchased from LabService (Bologna,
Italy). Six dark chocolates and three milk chocolates from
various price ranges, obtained from local markets, were used
as test samples.
Sample preparation and extraction
The chocolate bars were grated using a kitchen grater
and 10 mg of the grated sample were added to 100 mL of
Copyright # 2004 John Wiley & Sons, Ltd.
chloroform and the mixture was mixed with 1 g of diatomac-
eous earth. Before use, the diatomaceous earth was
washed sequentially with acetone and 50:50 (v/v) dichloro-
methane/methanol to minimise interferences in the subse-
quent analysis. A 6 mL glass cartridge was filled with the
mixture, with two polyethylene frits to keep the packing in
place. Elution was performed with 20 mL of 50:50 (v/v)
chloroform/methanol under moderate vacuum from a water
pump and 50 mL of the solution were injected into the
LC/MS/MS system.
The analytes were quantified by the standard addition
method; extracts containing high levels of fatty acids were
diluted to bring the concentrations within the calibrated range.
For recovery studies, 10 mg of chocolate were placed in a
porcelain mortar and spiked with appropriate volumes of the
working standard solution, taking care to uniformly spread it
on the sample. The samples were kept for 1 h in the dark at
room temperature before the MSPD procedure.
LC/MS/MS analysis
The LC apparatus was a series 200 binary pump (Perkin-
Elmer, Norwalk, CT, USA) equipped with a Rheodyne 7125
injector with a 50 mL loop and a series 200 vacuum degasser
(Perkin-Elmer). Analytes were chromatographed on an
Alltima 250  4.6 mm column filled with 5-mm C-18
reversed-phase packing (Alltech, Deerfield, IL, USA). The
flow rate of the mobile phase was 1 mL/min; methanol was
selected as phase A and water as phase B, and both solvents
contained 0.25 mmol/L HCOOH. Gradient elution was per-
formed by linearly increasing the percentage of organic
modifier from 95 to 100% in 25 min. A methanolic solution
of ammonia (40 mmol/L, freshly prepared each day) was
added post-column to the LC column effluent, at a flow rate
of 0.11 mL/min, by an isocratic LC pump model 2510 (Varian,
Walnut Creek, CA, USA). A total of 0.2 mL/min of the LC
column effluent was diverted to the electrospray ionisation
(ESI) source.
A PE Sciex API 2000 tandem triple-quadrupole mass
spectrometer (Perkin Elmer), equipped with a TurboIon-
Spray source operated in the NI (negative ionisation) mode,
was used for this work. The ionspray voltage was
5500 V.
Nitrogen gas from a liquid nitrogen dewar at 7 bar was used
as nebuliser, drying, curtain and collision gases. The settings
for the nebuliser, drying and curtain gases were 30, 60 and 30
(instrument units). The TurboIonSpray probe temperature
was maintained at 3508C and the gas pressure in the collision
cell was set at 4 mTorr. For each analyte, selected reaction
monitoring (SRM) transitions were chosen for the quantita-
tion after observing the collision-induced dissociation (CID)
spectra obtained by full-scan product ion experiments.
Mass axis calibration of each mass-resolving quadrupole
(Q1 and Q3) was performed by infusion of a poly(propylene
glycol) solution at 10 mL/min. Unit mass resolution was
established and maintained in each mass-resolving quadru-
pole by keeping a full width at half maximum of approxi-
mately 0.7 Th. All the source and instrument parameters for
monitoring analytes were optimised by standard solutions of
100 pg/mL (containing 4 mmol/L of ammonia) infused at
10 mL/min by a syringe pump. Data acquisition was divided
into four periods (Table 1), and in each period individual ion
Rapid Commun. Mass Spectrom. 2004; 18: 19891994
 Determination of free fatty acids in chocolate by LC/MS/MS 1991

Table 1. Experimental conditions for the SRM LC/MS/MS determination of selected analytes
Declustering Collision Retention
Compound Period (min) SRM transition (m/z) potentiala (V) potentialb (V) time (min) Dwell time (ms)

C18:3 011 277.3/127.0
65 30 6.5 250

C14:0 227.3/93.0
55 30 7.0 250
C16:1 253.3/253.3
65 13 7.2 250
C20:4 303.3/59.0
80 35 7.3 250
C18:2 279.3/97.0
80 35 7.8 250
C16:0 255.3/237.0
60 25 9.6 250
C18:1 281.3/99.0
80 35 9.8 250
C18:0 1117 283.3/265.0
80 30 13.2 750
C20:1 309.3/97.0
75 35 13.4 750
C20:0 1721 311.3/183.0
100 42 18.0 1500
C22:0 2130 339.3/183.0
100 50 23.8 1500
a
The declustering potential is the difference between the orifice and skimmer voltages (the skimmer is grounded and not user-controlled).
b
The collision potential is the potential difference between the collision cell quadrupole and high-pressure entrance quadrupole.

optics and MS/MS tuning parameters were optimised for

each SRM transition in order to enhance the sensitivity. The
mass spectrometry data handling system used was the PE
Sciex package Multiview 1.4.

RESULTS AND DISCUSSION

Fragmentation study and MS/MS optimisation
Preliminary studies of fragmentation were performed to find
the best instrumental conditions affording the unequivocal
identification of the analytes in real samples. In product-ion
scan mode, maximum transmission of the precursor ion for
each analyte was achieved by use of a low declustering
energy to minimise up-front CID. In the NI mode, the ESI
spectra of the analytes all contained the [MH]
anion as
base peak, and these were therefore selected as precursor
ions in the MS/MS experiments.
For myristic (C14:0), palmitic (C16:0) and stearic acids (C18:0),
the CID MS/MS spectra were dominated by the precursor
ions and the fragment ions [M18]
produced by loss of H2O
from the carboxyl group (Fig. 1). For arachidic (C20:0) and
behenic (C22:0) acids, fragment ions at m/z 183 were observed
(Fig. 2). As reported in the literature,18 22 the formation of
these ions can be rationalised by charge-remote fragmenta-
tions (CRF), a class of gas-phase decompositions that occur
physically remote from the charge site.23 These reactions are
analytically useful because they allow determination of the
positions of double bonds and branching in the aliphatic Figure 1. Negative ion ESI-MS/MS spectrum of [MH]
ion
chain.24 of stearic acid (C18:0).
The CID spectra of [MH]
ions of unsaturated com-
pounds contained an ion series with an inter-peak spacing of general in the spectra of the monounsaturated fatty acids,
14 Th, representing CRF cleavages of consecutive CC single occurred at m/z 97; this ion could be generated by the CRF
bonds in the fatty acid chain24 (Fig. 3). These ion series were process (cleavage of C5 C6) followed by further loss of H2
interrupted by gaps at the locations of the double bonds. For from the charge-carrying fragment.
example, the MS/MS spectrum of arachidenic acid (Fig. 4)
includes an ion series at m/z 71, 85, 99, 113, 127, 141 and 155; Recovery studies
this series terminated with the member representing The first step in the development of the MSPD method was
cleavage of the allylic CC single bond, on the side of the the evaluation of a suitable matrix/diatomaceous earth ratio
carboxylic group. The double-bond position was determined for sample extraction. A ratio of 1:100 was shown to be satis-
from the gap between m/z 155 and 209; the fragment ion at factory, but the sample had to be previously dissolved in
m/z 209 was produced by allylic cleavage of the single bond chloroform to facilitate the dispersion in the solid phase.
on the side of the alkyl chain. However, the most abundant Several solvents (methanol, acetonitrile, dichloromehane,
ion in the MS/MS spectra of the arachidenic acid, and in chloroform) and their mixtures were tested in recovery

Copyright # 2004 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2004; 18: 19891994
1992 D. Perret et al.
Figure 2. Negative ion ESI-MS/MS spectra of [MH]
ions
of arachidic (C20:0) and behenic acids (C22:0).
studies. Tests performed with pure solvents were unsatisfac-
tory; recoveries with acetonitrile and dichloromethane were
low for all compounds, methanol was ineffective for ex-
traction of stearic, arachidic and behenic acids, while chloro-
form gave low recoveries (60%) for palmitoleic and oleic acids
(data not shown). Evaluation of several methanol/chloro-
form mixtures (20:80, 50:50, and 80:20, all v/v), showed that
methanol/chloroform 50:50 (v/v) was the most efficient
extracting solution and provided the cleanest extracts.
Extraction recoveries were evaluated by the standard
addition method. Known amounts of each analyte were
added to the chocolate samples; for this purpose four samples
(two dark chocolates and two milk chocolates) were screened
Figure 3. Cleavage of consecutive CC single bonds in the chains of representative unsaturated
fatty acids.
Copyright # 2004 John Wiley & Sons, Ltd.
Figure 4. Negative ion ESI-MS/MS spectrum of the [MH]

ion of arachidenic acid (C20:1).
for their endogenous levels of the selected free fatty acids. The
amounts of analytes added to the samples for recovery
studies corresponded to about 50150% of the estimated
endogenous levels of fatty acids. The results are summarised
in Table 2; different spike levels were used based on the
estimated content of the fatty acids in the screened samples.
The recoveries exceeded 80% in most cases, with relative
standard deviations (RSDs) ranging between 5 and 11%.
Figures 5 and 6 show typical LC-SRM chromatograms
obtained by analysing the targeted free fatty acids in a
working standard solution and in a dark chocolate sample.
Limits of detection
Limits of detection (LODs) and limits of quantification
(LOQs) of the method were calculated for the analytes by
Rapid Commun. Mass Spectrom. 2004; 18: 19891994
 Determination of free fatty acids in chocolate by LC/MS/MS 1993
a b
Table 2. Recovery (and RSD ) of selected analytes at
different spiked levels
Analyte 80 ng/mg 100 ng/mg 120 ng/mg

Linolenic acid 96 (6) 95 (5) 96 (6)

Arachidonic acid 89 (8) 91 (7) 86 (6)
Arachidenic acid 79 (9) 82 (8) 86 (10)
Arachidic acid 97 (6) 98 (6) 97 (7)
Behenic acid 85 (8) 84 (9) 86 (8)
Analyte 2 mg/mg 4 mg/mg 6 mg/mg
Myristic acid 99 (6) 101 (8) 98 (10)
Palmitic acid 98 (7) 96 (6) 97 (7)
Palmitoleic acid 81 (8) 79 (8) 84 (7)
Stearic acid 91 (6) 93 (5) 88 (7)
Oleic acid 93 (8) 92 (7) 95 (8)
Linoleic acid 103 (9) 98 (11) 100 (9)
a
Mean values from five determinations.
b
Relative standard deviation.

Figure 6. Composite LC-SRM chromatogram resulting from

the analysis of a dark chocolate sample. 5: C18:2; 6: C16:0; 7:
C18:1; 8: C18:0.

Table 3. Limits of detection (LODs) and quantification

(LOQs) of the method for selected analytes
Analyte LODs (ng/g) LOQs (ng/g)

Linolenic acid 41 123

Arachidonic acid 40 120
Myristic acid 42 126
Palmitoleic acid 33 99
Linoleic acid 51 153
Palmitic acid 20 60
Oleic acid 20 60
Arachidenic acid 8 24
Stearic acid 7 21
Arachidic acid 10 30
Behenic acid 11 33

Figure 5. Composite LC-SRM chromatogram of a standard

mixture of free fatty acids (100 ng of each injected). 1: C18:3; Table 4. Analytical precision (RSDa) of the method
2: C14:0; 3: C16:1; 4: C20:4; 5: C18:2; 6: C16:0; 7: C18:1; 8: C18:0;
Analyte Intra-day precision Inter-day precision
9: C20:1; 10: C20:0; 11: C22:0.
Linolenic acid 5 10
use of the criterion reported by Parker;25 the LOD was set at Arachidonic acid 6 11
three times the noise level of the baseline in the chromato- Myristic acid 6 12
gram (i.e., a definition based on S/N 3) and the LOQ was Palmitoleic acid 5 9
Linoleic acid 7 10
set at three times the LOD. LODs were estimated from the
Palmitic acid 6 12
LC-SRM chromatograms resulting from analyses of a dark Oleic acid 5 9
chocolate sample in which the amount of each analyte was Arachidenic acid 5 10
evaluated by the standard addition method. The resulting Stearic acid 6 12
data reported in Table 3 indicate that the method is adequate Arachidic acid 7 11
Behenic acid 5 9
for analysing selected free fatty acids in chocolate samples at
a
ng/g levels. Relative standard deviation.

Copyright # 2004 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2004; 18: 19891994
1994 D. Perret et al.

Table 5. Free fatty acid (FFA) composition (% of total fraction) of dark and milk chocolate samples
Dark chocolate samples* Milk chocolate samples*

Sample n8 1 2 3 4 5 6 7 8 9

Extracted FFAs (wt%) 0.3 0.4 0.6 0.7 0.8 1.3 1.3 1.7 2.3
C14:0 7.4 16.6 3.9 5.8 25.9 3.9 21.6 31.0
C16:0 33.7 24.1 29.7 32.3 14.7 29.5 23.2 21.3 14.6
C16:1 3.4 2.3 0.9 2.0 1.4 1.7 1.6 2.5
C18:0 23.6 16.3 39.7 32.9 23.2 26.8 58.0 38.5 34.9
C18:1 20.2 29.0 17.9 27.4 43.0 11.8 8.5 10.2 10.4
C18:2 7.8 9.1 6.8 7.3 8.6 3.7 3.1 4.7 6.0
C18:3 1.0 1.1 0.5 0.3 0.4 0.7 0.2 0.5
C20:0 1.7 1.2 0.6 2.5 0.4 1.8
C20:1 0.2 0.1
C20:4
C22:0 1.0 0.3 0.2 0.8

* Mean of two determinations.

Linear dynamic range CONCLUSIONS

The instrumental dynamic range was established by measur-
ing peak areas for the fatty acids over a 5-fold range of concen-
tration, using composite working standard solutions. The
slope, intercept and correlation coefficient were calculated
by linear regression analysis. Measurements were obtained
in triplicate for each amount injected. The instrumental
response of the ESI-MS/MS system was linearly dependent
on the amounts of the analytes injected up to 200 ng; the
regression coefficients for the calibration curves were not
less than 0.997.
Analytical precision
The precision of the method was evaluated using a set of sam-
ples of identical matrices, all using the described extraction
procedure. The intra-day precision was measured by repeat-
ing three independent tests and measuring peak areas of the
fatty acids within the same day, while the inter-day precision
was determined on three different days. Table 4 shows the
results obtained: in general the analytical precisions were
<7% (intra-day) and <12% (inter-day).
Application of the developed method
The proposed method was applied to the determination of
free fatty acids in some dark and milk chocolates obtained
from local supermarkets. The results are shown in Table 5.
The content of the different FFAs varied between 0.3 and
2.3% (wt%) and were greater in milk chocolate samples
than in dark; this difference, in particular with regard to
the content of stearic and myristic acids, could reflect the
presence of the milk fats. In general the FFA composition in
the analysed samples was dominated by stearic (C18:0, 16.3
58.0%), oleic (C18:1, 8.543.0%), and palmitic (C16:0, 14.6
33.7%) acids, followed by myristic (C14:0, 3.931.0%) and lino-
leic (C18:2, 3.19.1%) acids. Linolenic (C18:3), arachidic (C20:0),
arachidenic (C20:1), and behenic (C20:0) acids, present as minor
components (below 3%), were detected in some samples,
while the arachidonic acid (C20:4) content was always below
the LOD. The FFA composition observed in commercial
samples resembled that typical of CB and CBEs;5 the vari-
ability of the FFA profiles for the chocolates probably reflects
the variation in CB and CBEs and the different mixing ratios
used for their preparation.4
A simple and rapid method for the determination of FFAs
in chocolate by LC/ESI-MS/MS was developed. The MSPD
procedure with diatomaceous earth as solid support offers
the advantages of speed and of reduction of solvent con-
sumption; the time necessary for the extraction is about
20 min. The extracts appear to be sufficiently clean and do
not require manipulation or cleanup before the final analysis
by LC/MS/MS. The resulting simplification of the analytical
protocol for sample preparation is nevertheless possible only
because of the high specificity and sensitivity of the LC/MS/
MS system used subsequently in the analysis.
REFERENCES
1. Padley FB, Timms RE. JAOCS 1980; 9: 286.
2. Simoneau C, Hannaert P, Anklam E. Food Chem. 1999; 65:
111.
3. Barcarolo R, Anklam E. J. AOAC Int. 2001; 84: 1485.
4. Spangerberg JE, Dionisi F. J. Agric. Food Chem. 2001; 49: 4271.
5. Lipp M, Anklam E. Food Chem. 1998; 62: 99.
6. Gutnikov G. J. Chromatogr. B 1995; 671: 71.
7. Lipp M, Simoneau C, Ulberth F, Anklam E, Crews C,
Brereton P, de Greyt W, Schwack W, Wiedmaier C. J. Food
Composition Anal. 2001; 14: 399.
8. Nesaretnam K, Mohd Ali AR. J. Sci. Food Agric. 1992; 60: 15.
9. Batta AK, Dayal V, Colman RW, Sinha AK, Shefer S, Salen
G. J. Chromatogr. 1984; 284: 257.
10. AOAC International (1996) Official method 963.5. Fat in cacao
products, Soxhlet extraction method, final action 1973. In
Official Methods of AOAC International, (16th edn) vol. II.
11. Simoneau C, Naudin C, Hannaert P, Anklam E. Food Res.
Int. 2000; 33: 733.
12. Barker SA. J. Chromatogr. A 2000; 885: 115.
13. Barker SA. J. Chromatogr. A 2000; 880: 63.
14. Torres CM, Pico Y, Manes J. Chromatographia 1995; 41: 685.
15. Torres CM, Pico Y, Manes J. J. Chromatogr. A 1996; 754: 301.
16. Tekel J, Hatrik S. J. Chromatogr. A 1996; 754: 397.
17. Perret D, Gentili A, Marchese S, Sergi M, DAscenzo G.
J. AOAC Int. 2002; 85: 724.
18. Claeys M, Nizigiyimana L, Van den Heuvel H, Derrick PJ.
Rapid Commun. Mass Spectrom. 1996; 10: 770.
19. Kerwin JK, Wiens AM, Ericsson LH. J. Mass Spectrom. 1996;
31: 184.
20. Gross ML. Int. J. Mass Spectrom. 2000; 200: 611.
21. Fong-Fu H, Turk J. J. Am. Soc. Mass Spectrom. 1999; 10: 600.
22. Seto C, Grosset JS, Waddel DS, Curtis JM, Boyd RK. J. Am.
Soc. Mass Spectrom. 2001; 12: 571.
23. Cheng C, Gross ML. Mass Spectrom. Rev. 2000; 19: 398.
24. Voinov VG, Claeys M. Int. J. Mass Spectrom. 2001; 205: 57.
25. Parker GA. J. Assoc. Off. Anal. Chem. 1991; 74: 868.

Copyright # 2004 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2004; 18: 19891994