Potential of gas chromatography-orthogonal acceleration time-of-flight mass spectrometry (GC-oaTOFMS) in flavor research

Article abstract of DOI:10.1021/jf0261203

Gas chromatography-orthogonal acceleration time-of-flight mass spectrometry (GC-oaTOFMS) is an emerging technique offering a straightforward access to a resolving power up to 7000. This paper deals with the use of GC-oaTOFMS to identify the flavor components of a complex seafood flavor extract and to quantify furanones formed in model Maillard reactions. A seafood extract was selected as a representative example for complex food flavors and was previously analyzed using GC-quadrupole MS, leaving several molecules unidentified. GC-oaTOFMS analysis was focused on these unknowns to evaluate its potential in flavor research, particularly for determining exact masses, N-Methyldithiodimethylamine, 6-methyl-5-hepten-2-one, and tetrahydro-2,4-dimethyl-4H-pyrrolo- [2,1-d]-1,3,5-dithiazine were successfully identified on the basis of the precise mass determination of their molecular ions and their major fragments. A second set of experiments was performed to test the capabilities of the GC-oaTOFMS for quantification. Calibration curves were found to be linear over a dynamic range of 103 for the quantification of furanones. The quantitative data obtained using GC-oaTOFMS confirmed earlier results that the formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone was favored in the xylose/glycine model reaction and 2(or 5) -ethyl-4-hydroxy-5(or2)-methyl-3(2H)-furanone in the xylose/alanine model reaction. It was concluded that GC-oaTOFMS may become a powerful analytical tool for the flavor chemist for both identification and quantification purposes, the latter in particular when combined with stable isotope dilution assay.

Full text of DOI:10.1021/jf0261203

2708 J. Agric. Food Chem. 2003, 51, 2708−2713
Potential of Gas Chromatography−Orthogonal Acceleration
Time-of-Flight Mass Spectrometry (GC-oaTOFMS) in Flavor
Research
LAURENT B. FAY,*,† ANTHONY NEWTON,§ HERVEÄ SIMIAN,† FABIEN ROBERT,†
†
DAVID DOUCE,§ PETER HANCOCK,§ MARTIN GREEN,§ AND IMRE BLANK
Nestle´ Research Centre, Nestec Ltd., Vers-chez-les-Blanc, P.O. Box 44,
1000 Lausanne 26, Switzerland, and Micromass U.K. Ltd., Floats Road,
Wythensawe, Manchester M23 9LZ, United Kingdom
Gas chromatography-orthogonal acceleration time-of-flight mass spectrometry (GC-oaTOFMS) is
an emerging technique offering a straightforward access to a resolving power up to 7000. This paper
deals with the use of GC-oaTOFMS to identify the flavor components of a complex seafood flavor
extract and to quantify furanones formed in model Maillard reactions. A seafood extract was selected
as a representative example for complex food flavors and was previously analyzed using GC-
quadrupole MS, leaving several molecules unidentified. GC-oaTOFMS analysis was focused on these
unknowns to evaluate its potential in flavor research, particularly for determining exact masses.
N-Methyldithiodimethylamine, 6-methyl-5-hepten-2-one, and tetrahydro-2,4-dimethyl-4H-pyrrolo-
[2,1-d]-1,3,5-dithiazine were successfully identified on the basis of the precise mass determination
of their molecular ions and their major fragments. A second set of experiments was performed to
test the capabilities of the GC-oaTOFMS for quantification. Calibration curves were found to be linear
over a dynamic range of 103 for the quantification of furanones. The quantitative data obtained using
GC-oaTOFMS confirmed earlier results that the formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone
was favored in the xylose/glycine model reaction and 2(or 5)-ethyl-4-hydroxy-5(or 2)-methyl-3(2H)-
furanone in the xylose/alanine model reaction. It was concluded that GC-oaTOFMS may become a
powerful analytical tool for the flavor chemist for both identification and quantification purposes, the
latter in particular when combined with stable isotope dilution assay.
KEYWORDS: GC-oaTOFMS; flavor; furanones
INTRODUCTION
spectral libraries containing about 400000 mass spectra, iden-
tification of unknown flavor molecules remains a challenge.
Researchers have tried to overcome the complexity of flavor
mixtures by increasing the resolving power of their analytical
techniques using, for example, high-resolution capillary gas
chromatography (2) or GC-tandem mass spectrometry (3, 4).
However, capillary gas chromatography coupled with high-
resolution mass spectrometry (HR-MS) has rarely been used.
Capillary GC-HR-MS has been performed up to now either with
double-focusing magnetic sector instruments (5) or with FTICR
machines (6), which are very expensive and need well-trained
operators. Recently, gas chromatography-orthogonal accelera-
tion time-of-flight mass spectrometry (GC-oaTOFMS) has been
shown as an emerging technique offering a straightforward
access to a resolving power up to 7000 (7). For GC-MS analysis
orthogonal acceleration time-of-flight provides many advantages
over conventional scanning mass spectrometers. In oaTOFMS
ions formed in a continuous ion source are accelerated focused
into a parallel ion beam. As the ions traverse an orthogonal
sampling region a sudden voltage pulse is applied, ejecting a
In 1955, Gohlke and McLafferty (cited in ref 1) realized the
first coupling between a gas chromatograph and a mass
spectrometer, offering to the analytical community one of their
most powerful spectroscopic techniques. Nowadays, gas chro-
matography-mass spectrometry (GC-MS) is the method of
choice for the analysis of complex mixtures. In flavor research,
GC-MS is the workhorse of any flavor chemist, and each year
several hundred publications discuss results obtained by GC-
MS. However, as flavor research deals with key odorants that
usually occur in trace amounts, often embedded in extracts
containing volatile compounds at much higher concentrations,
GC-MS suffers sometimes from a lack of selectivity and/or
sensitivity, making identification of unknown compounds dif-
ficult in such complex mixtures. Despite the availability of mass
* Author to whom correspondence should be addressed (telephone
+41-21-785-86-09; fax +41-21-785-85-44; e-mail laurent-bernard.fay@
rdls.nestle.com).
† Nestec Ltd.
§ Micromass U.K. Ltd.
10.1021/jf0261203 CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/22/2003

GC-oaTOFMS in Flavor Research
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2709
The tube was sealed with a screw cap and heated at 90 °C for 1 h in
an oil bath under stirring with a magnetic stirrer. The reaction mixture
was rapidly cooled with tap water. Then, water (100 mL) and the labeled
internal standards (73.0 µg of 13C2-HDMF and 18.9 µg of 2H3-EHMF)
were added to the dark brown reaction mixture, which was then
saturated with NaCl (35 g). The pH was adjusted to 4 (aqueous HCl,
2 mol/L), and the neutral compounds were continuously extracted with
diethyl ether (50 mL) overnight using a rotation perforator (Normag,
Weinheim, Germany). The organic phase was separated, dried over
sodium sulfate at 4 °C, and concentrated to 1 mL using a Vigreux
column (50 cm × 1 cm) and a microdistillation device. All experiments
were performed in duplicate.
Low-Resolution GC-MS. Conventional GC-MS analysis was
performed using a gas chromatograph (HP-5890, Agilent, Geneva,
Switzerland) equipped with a splitless injector heated at 260 °C and
coupled with a quadrupolar mass spectrometer (HP-5970, Agilent)
operated in electron impact ionization mode at 70 eV. Acquisitions
were carried out over a mass range of 10-300 Da. Separation was
performed on a 100% dimethyl polysiloxane stationary phase (Agilent
Ultra-1 PONA, 50 m × 0.20 mm i.d., 0.5 µm film thickness). Helium
was used as the carrier gas with a constant flow rate of 0.6 mL/min.
The oven was programmed as follows: 20 °C (0.5 min), 70 °C/min to
60 °C, 4 °C/min to 240 °C. The temperature of the transfer line was
held at 280 °C during the chromatographic run. The same chromato-
graphic equipment and conditions were used to detect odorous
compounds by GC-olfactometry (14). Flame ionization detection was
performed in parallel to GC-O.
High-Resolution GC-oaTOFMS. All experiments were performed
using a Micromass GCT mass spectrometer (Manchester, U.K.) operated
in electron impact ionization mode at 70 eV. The GCT is a benchtop,
orthogonal acceleration, reflectron, time-of-flight mass spectrometer
capable of elevated resolution (7000 full-width at half-maximum height
definition). Acquisitions were carried out over a mass range of 50-
450 Da with an acquisition rate of one spectrum per second at a
resolution of 7000 (fwhm). The source temperature was held at 180
°C. Exact masses were determined using a lock mass at m/z 201.9609
obtained after continuous infusion of chloropentafluorobenzene during
the GC program.
GC analyses were performed using a gas chromatograph (HP-6890,
Agilent) equipped with a splitless injector heated at 280 °C and a DB-
5MS (J&W Scientific, Folsom, CA) capillary column (5% phenyl, 95%
dimethyl polysiloxane stationary phase, 20 m × 0.18 mm i.d., 0.18
µm film thickness). Helium was used as the carrier gas with a constant
flow rate of 1.0 mL/min. The oven program was 40 °C (2 min) and
then 4 °C/min to 250 °C. The temperature of the transfer line was held
at 250 °C during the chromatographic run.
portion of the beam orthogonally. This packet of ions is then
accelerated into the time-of-flight drift region. Ions of different
m/z values have different velocities and hence arrive at a detector
at different times relative to the orthogonal acceleration pulse.
By precisely recording these arrival times a time-of-flight mass
spectrum is produced. The axial ion beam is typically sampled
at between 10000 and 100000 times per second; individual time-
of-flight spectra are generally summed before storage to disk.
As the initial energy spread of the ions is generally very low in
the orthogonal direction compared with the axial direction, the
spread of ion arrivals for a particular m/z value is minimized
and high mass resolution may be obtained. The addition of a
reflectron device, to increase the flight time of the ions, may
further improve the mass resolution of the oaTOF analyzer. This
elevated resolution reduces mass interferences, increasing the
selectivity of the instrument. The oaTOF analyzer simulta-
neously samples ions of all m/z values, unlike scanning
instruments where ions are detected or ejected sequentially. This
provides a particular advantage for GC-MS analysis, in which
analyte concentration and composition change rapidly as
components elute from the GC column. Spectra produced are
representative of the composition of the analyte regardless of
how quickly the concentration of the analyte changes. Spectral
“skew”, which may be apparent with rapidly eluting analytes
using scanning instruments, is avoided using oaTOFMS. The
high duty cycle of the oa-TOFMS (typically 30%) results in
significantly improved sensitivity for full mass range data,
compared to conventional scanning instruments. In addition, the
precise and stable relationship between an ion’s arrival time
and the square root of its mass allows good mass accuracy
with only a single internal reference mass. Finally, benchtop
GC-oaTOFMS systems like the one used in this study are easy
to operate and do not need highly experienced users.
This paper will present identification and quantification
experiments that have been performed with complex seafood
flavor extracts and model Maillard reactions.
MATERIALS AND METHODS
Chemicals. D-Xylose, glycine, L-alanine, methylmercaptan, dimeth-
ylamine, phenyl isocyanate, and thionyl chloride of highest purity
(>99%) were obtained from Fluka (Buchs, Switzerland). 4-Hydroxy-
2,5-dimethyl-3(2H)-furanone (HDMF) and 2(or 5)-ethyl-4-hydroxy-
5(or 2)-methyl-3(2H)-furanone (EHMF) were from Fluka and Givaudan
(Du¨bendorf, Switzerland), respectively. 4-Hydroxy-2(or 5)-[13C]methyl-
5(or 2)-methyl-3(2H)-[2(or 5)-13C]furanone (13C2-HDMF) and 2(or 5)-
([2,2,2-2H3]eth-1-yl)-4-hydroxy-5(or 2)-methyl-3(2H)-furanone (2H3-
EHMF) were synthesized as recently described (8). The isotopic content
of the labeled compounds was 99%. Anhydrous sodium sulfate,
dipotassium hydrogen phosphate dihydrate, and diethyl ether were from
Merck (Darmstadt, Germany). The organic solvents were purified by
slow distillation on a Vigreux column (1 m × 1 cm).
Quantification Experiments. The calibration curve for 13C2-HDMF
was established with standard mixtures containing defined amounts of
unlabeled and labeled compounds in different ratios following the
procedure described by Guth and Grosch (15). Samples for establishing
the calibration curve were injected four times and for quantifying
HDMF in the Maillard model reactions, twice.
RESULTS AND DISCUSSION
Identification of Unknown Flavor Molecules. Seafood
flavors are known to be very complex, composed of many
volatile compounds that occur in a wide concentration range
(16-18). In general, a number of volatile compounds can be
identified by conventional GC-quadrupole MS on the basis of
retention index and fragmentation pattern. However, in the case
of unknowns or compounds for which no reference mass
spectrum is available, positive identification remains a challenge
to the flavor chemist. Therefore, a seafood sample was selected
as a representative example for complex food flavors, which
had previously been analyzed by conventional GC-quadrupole
MS. Using this technique, several molecules remained unknown.
GC-oaTOFMS analysis was focused on these unknowns to
evaluate its potential in flavor research, particularly for deter-
mining exact masses.
Seafood Flavor Extract. This was prepared from a commercially
available liquid seafood sample using the SAFE technique (9), which
allows the careful isolation of volatiles from complex mixtures. The
aqueous seafood flavor concentrate (50 mL) was adjusted to pH 9 with
NaOH (20 mL, 2 N). Methylene chloride was added (50 mL), and the
nonvolatile compounds were separated from this mixture by evaporation
of the volatiles, water, and organic solvent under high vacuum. The
organic phase was separated from the aqueous layer, dried with Na2-
SO4, and concentrated to 0.6 mL by microdistillation (10) to obtain
the SAFE extract. Chemical synthesis of N-methyldithiodimethylamine
was performed by adapting procedures described by Mukaiyama et al.
(11) and Kulikovskaya et al. (12).
Maillard Reaction Samples. Sample preparation was performed
as recently described (13) with some modifications. In a 15 mL Pyrex
tube, xylose (750 mg) and glycine (375 mg) or alanine (446 mg) were
dissolved in a phosphate buffer (5 mL, 0.2 mol/L K2HPO4, pH 6.0).

2710 J. Agric. Food Chem., Vol. 51, No. 9, 2003
Fay et al.
Figure 1. Total ion chromatogram obtained after GC-oaTOFMS analysis
of a seafood extract. Numbers correspond to compounds identified by
exact mass measurement.
Figure 3. Mass spectra of compound 4 identified as N-dimethyldithiodi-
methylamine (A), compound 7 identified as 6-methyl-5-hepten-2-one (B),
and compound 13 identified as tetrahydro-2,4-dimethyl-4H-pyrrolo[2,1-d]-
1,3,5-dithiazine (C).
The deviations observed between the measured mass and the
calculated mass of the proposed structure are on average below
1 mDa. These results are in agreement with the data recently
published with hybrid instruments (quadrupole-oaTOFMS)
working with an electrospray source and showing that good mass
accuracies were obtained in MS and MS/MS operation of the
TOF instrument (19, 20). The mass accuracy obtained in our
experiments is sufficient to obtain the “true” elemental composi-
tion identified in the first hit in >90% of the cases. This result
is valid because low molecular weight molecules were inves-
tigated and, moreover, because the identity of these molecules
can be confirmed by other techniques such as GC-O and GC
coupled with specific detectors indicating the presence of
specific elements. However, this shows that the GC-oaTOFMS
technology could be a valuable tool for flavor identification as
the compounds of interest have molecular weights around 200
Da with fragment ions in the low mass range.
The mass spectrum of compound 4 (Figure 3A) is a typical
example of the information that can be obtained by GC-
oaTOFMS. The mass spectrum displayed an intense molecular
ion measured at m/z 123.0176 and small fragment ions at m/z
107.9945 and 76.0228. The molecular ion measured at 123.0176
Da is in agreement with the theoretical formula C3H9NS2. The
two fragment ions measured at m/z 107.9945 and 76.0228 allow
the theoretical formulas C2H6NS2 (theoretical mass 107.9942,
difference of 0.3 mDa) and C2H6NS (theoretical mass 76.0221,
difference of 0.7 mDa), respectively, to be proposed. From these
data, the new structure N-methyldithiodimethylamine was
proposed.
Figure 2. Chemical structures of volatile compounds identified in the
seafood aroma extract.
Figure 1 shows the GC-oaTOFMS chromatogram obtained.
Some unknown compounds of interest were examined either
after selection by GC-O on the basis of their sensory charac-
teristics or because of their weak abundance to test the GC-
oaTOFMS sensitivity. Their exact mass spectra and the elemen-
tal composition of the major mass ions were determined. For
each measured mass ion, a theoretical composition was calcu-
lated within a window of (2 mDa. The results are summarized
in Table 1, and the structures of all compounds studied are
presented in Figure 2.
Because of its sensitivity, GC-oaTOFMS allows exact mass
measurements of low intensity ions, for example, ions at m/z
61, 83, and 67 for compounds 4, 9, and 10, respectively. From
our past experience, this has never been achieved so easily using
a double-focusing sector field instrument. The latter, however,
works at a higher resolution power (from 10000 to 35000) and
is more accurate as compared to the TOF instrument.
Straighforward measurement of the exact mass of low-
intensity ions enables the analyst to propose an elemental
composition of each ion obtained in the mass spectrum and,
therefore, to pursue structure identification on the basis of not
only the molecular ion but also all fragments. This is a valuable
piece of information as fragmentation is directly linked to the
chemical structure, thus allowing a higher degree of confidence
in the identification of unknowns.
To assess this proposal, we synthesized the N-methyldithiodi-
methylamine following the route depicted in Figure 4. Bis-
(dimethylamine) sulfoxide I was obtained by reacting dimethyl-
amine with thionyl chloride in chloroform. In the presence of
phenyl isocyanate, I resulted in the betaine II, which reacted

GC-oaTOFMS in Flavor Research
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2711
Table 1. Investigated Peaks Detected in the Seafood Flavor Extract and Measured at a Resolution of 7000 Using the GC-oaTOFMSa
measured
mass
relative ion
intensity (%)
proposed
composition
theoretical
mass
deviation
(mDa)
peak
identification
odor descriptor
4
123.0176
107.9945
76.0228
61.0002
100.0
18.4
31.0
1.5
C H NS
2
123.0176
107.9942
76.0221
60.9986
0.0
0.3
0.7
1.6
methyldithiodimethylamine
sulfury and alliaceous
3
9
C H NS
2
6
2
C H NS
2
6
CH NS
3
6
7
116.0943
72.0446
39.7
100.0
C H N O
C H NO
3 6
116.0950
72.0449
−0.7
−0.3
tetramethylurea
odorless
5
12 2
126.1045
111.0816
108.0931
84.0602
69.0350
12.7
40.2
19.4
37.3
100.0
C H O
126.1045
111.0810
108.0939
84.0575
69.0340
0.0
0.6
−0.8
2.6
6-methyl-5-hepten-2-one
fruity and blue cheese
8
14
C H O
7
11
C H
8
12
C H O
5
8
C H O
1.0
4
5
9
113.0834
84.0572
83.0506
67.0552
56.0261
55.0187
100.0
41.9
5.9
7.4
15.9
35.7
C H NO
113.0841
84.0575
83.0497
67.0548
56.0262
55.0184
−0.7
−0.3
0.9
0.4
−0.1
0.3
ꢀ-caprolactam
odorless
6
11
C H O
5
8
C H O
5
7
C H
5
7
C H O
3
4
C H O
3
3
10
12
135.0138
108.0037
83.0506
67.0552
100.0
21.7
5.9
C H NS
135.0143
108.0034
83.0497
67.0548
−0.5
0.3
0.9
benzothiazole
slightly rubbery
floral and fruity
7
5
C H S
6
4
C H O
5
7
7.4
C H
0.4
5
7
194.1683
176.1576
161.1331
151.1122
136.1254
133.1017
126.1036
125.0969
121.1022
108.0922
107.0864
93.0706
12.2
10.7
10.4
91.5
10.6
8.3
C H O
194.1671
176.1565
161.1330
151.1123
136.1252
133.1017
126.1045
125.0966
121.1017
108.0939
107.0861
93.0704
1.2
1.1
0.1
−0.1
0.2
0.0
−0.9
0.3
0.5
−1.7
0.3
0.2
geranyl acetone
13 22
C H
13 20
C H
12 17
C H O
10 15
C H
10 16
C H
10 13
6.2
C H O
8
14
48.4
17.2
11.6
52.7
23.1
100.0
C H O
8 13
C H
9
13
C H
8
12
C H
8
11
C H
7
9
69.0706
C H
69.0704
0.2
5
9
13
189.0653
40.0
C H NS
2
189.0646
0.7
tetrahydro-2,4-dimethyl-4H-
pyrrolo[2,1-d]-1,3,5-dithiazine
musty and roasty, crustacean-like
8
15
129.0616
128.0541
97.0906
96.0822
95.0737
94.0662
70.0667
69.0591
68.0510
60.0037
58.9960
57.9884
56.9808
54.0358
28.1
10.0
60.5
26.2
29.8
34.7
41.6
33.7
31.3
100.0
93.7
34.5
19.5
11.44
C H N O
9 2 2
189.0664
129.0612
128.0534
97.0891
96.0813
95.0735
94.0657
70.0657
69.0578
68.0500
60.0034
58.9955
57.9877
56.9799
54.0344
−1.1
0.4
0.7
1.5
0.9
0.2
0.5
1.0
1.3
1.0
0.3
0.5
0.7
0.9
1.4
10
C H NS
6
11
C H NS
6
10
C H N
6
11
C H N
6
10
C H N
6
9
C H N
6
8
C H N
4
8
C H N
4
7
C H N
4
6
C H S
2
4
C H S
2
3
C H S
2
2
C HS
2
C H N
3
4
a For each peak, a tentative identification was proposed considering all ions measured.
with methylmercaptan to form N,N-dimethyl N-phenyl urea III,
N-methylthiosulfoxy dimethylamine IV, and N-methyldithiodi-
methylamine V. Although the target compound V was not the
main product obtained, its spectrum was identical to that of
compound 4 (data not shown).
mDa). The exact mass measurement indicated that this odd-
electron fragment is correlated with a formula CH3NS (theoreti-
cal mass of 60.9986, difference of 1.6 mDa) and could be
formed by a methyl group rearrangement leading to the odd-
electron ion [CH3sNdS]•+.
These results confirmed the chemical structure of N-meth-
yldithiodimethylamine 4 as proposed after interpretation of the
fragmentation pattern and considering the elemental composition
obtained by GC-oaTOFMS analysis. Moreover, and because of
its sensitivity, GC-oaTOFMS allowed the detection of a weak
fragment at m/z 61.0002. Sulfur-containing compounds often
exhibit such a typical mass peak corresponding to the fragment
C2H5S (theoretical mass of 61.011196, difference of 10.99
Compound 7 was described as fruity by the GC-O analysis,
and its mass spectrum is presented in Figure 3B. The measure-
ments of the ions at m/z 126.1045, 111.0816, 108.0931, and
84.0602 were the basis for suggesting the structures C8H14O,
C7H11O, C8H12, and C5H8O, respectively. They support the
identification of compound 7 as 6-methyl-5-hepten-2-one. For
the ion measured at m/z 69.0350, the closest theoretical formula
is C4H5O (69.0340; mass difference of 1.0 mDa), whereas the

2712 J. Agric. Food Chem., Vol. 51, No. 9, 2003
Fay et al.
Table 2. Quantification of HDMF and HEMF in Model Maillard
Reactionsa
Maillard system
HDMF
HEMF
xylose/glycine
xylose/L-alanine
16.6 ± 4.0
4.7 ± 0.5
0.3 ± 0.1
8.5 ± 0.6
a Results are expressed in µg/mmol sugar and are the mean of six
measurements.
mainly due to the limited repeatability of Maillard reaction
samples. A particular advantage of monitoring exact masses for
quantification purposes is the possibility to discriminate inter-
fering compounds having the same nominal masses. Further-
more, not only can the molecular mass be used but also smaller
fragments having a characteristic elemental composition for the
given compound.
Figure 4. Synthesis pathway of N-methyldithiodimethylamine.
In conclusion, GC-oaTOFMS may become a powerful
analytical tool for the flavor chemist for both identification and
quantification purposes, in particular when combined with the
IDA method.
alternative structure C5H9 can be ruled out (69.0704; mass
difference of 35 mDa). The presence of oxygen in this fragment
suggests the migration of the double bond to the conjugated
position followed by allylic cleavage. Compound 7 is probably
isomerized prior to the formation of the even-electron ion at
m/z 69 (M-57).
ACKNOWLEDGMENT
We thank Cynthia Parker for linguistic proofreading and
Ste´phanie Devaud for expert technical assistance.
Compound 13 (Figure 3C) is a very potent and well-known
compound often present in seafood. It was easily identified as
tetrahydro-2,4-dimethyl-4H-pyrrolo[2,1-d]-1,3,5-dithiazine be-
cause of the typical fragmentation of dithiazines. The various
mass peaks generated by the fragmentation of these byciclic
dithiazines have been described by Werkhoff (21). The ions at
m/z 129.0616, 97.0906, and 70.0667 correspond to the formulas
C6H11NS, C6H11N, and C4H8N. Unlike the spectrum obtained
from quadrupole mass spectrometers (from Werkhoff’s article
or from our own acquisition), where no clear mass peak indicates
the alkyl group in 2-position, a specific peak (m/z 60.0037) of
the spectrum from the GC-TOF shows the side chain associated
with the fragment CH-S of the ring. Surprisingly, this ion is
even the base peak of the mass spectrum obtained with the GC-
TOFMS and confirms the presence of a methyl group in the
2-position.
LITERATURE CITED
(1) Gohlke, R. S.; McLafferty, F. W. Early gas chromatography/
mass spectrometry. J. Am. Soc. Mass Spectrom. 1993, 4, 367-
371.
(2) Cartoni, G. P.; Goretti, G.; Neri, B.; Russo, M. V. Evaluation
of small diameter capillary columns for GC. J. Chromatogr.
1989, 475, 145-151.
(3) Fay, L. B.; Blank, I.; Cerny, C. In FlaVor Science: Recent
DeVelopments; Taylor, A. J., Mottram, D. S. Eds.; Royal Society
of Chemistry: Cambridge, U.K., 1996; pp 271-276.
(4) Sellier, N.; Cazaussus, A.; Budzinski, H.; Lebon, M. Structure
determination of sesquiterpenes in Chinese vetiver oil by gas
chromatography-tandem mass spectrometry. J. Chromatogr.
1991, 557, 451-458.
Quantitative Experiments. The potential of the orthogonal
acceleration TOFMS technology for quantitative measurements
has recently been investigated using small molecules separated
by reversed phase liquid chromatography and ionized into an
atmospheric pressure ionization interface (22). With this in-
strumental setup, it has been shown that excellent sensitivity
and a linear dynamic range of nearly four decades were obtained.
As capillary GC requires faster acquisition rates than HPLC
(because of their respective peak widths), we evaluated our GC-
oaTOFMS for quantitation. This work has been carried out with
furanones because of their low sensory threshold and therefore
their important contribution to many flavors, even if present at
low concentrations. We applied the isotope dilution assay (IDA)
as quantitation method (8, 15, 23), which is based on the use
of labeled internal standards to minimize losses of labile aroma
compounds, such as furanones, during sample preparation.
Calibration curves were found to be linear over a dynamic range
of 103 with correlation coefficients >0.9997 and all values
comparable to those determined with quadrupole instruments
(8). The method was applied to Maillard model reaction
mixtures. The obtained data confirmed earlier results (8); that
is, higher amounts of HDMF were found in the system xylose/
glycine, and the formation of HEMF was favored in the xylose/
alanine model reaction (Table 2). The variation coefficients are
from about 5 to 30%, depending on the concentration, which is
(5) Perkins, G.; Pullen, F.; Thompson, C. Automated high-resolution
mass spectrometry for the synthetic chemist. J. Am. Soc. Mass
Spectrom. 1999, 10, 546-551.
(6) Woodward, W. S. FT-ICR: a new weapon in the arsenal.
Spectrosc. Eur. 1998, 10, 14-18.
(7) Newton, A.; Hancock, P.; Green, M.; Bateman, R. H.; White,
S. Proceedings of the 48th ASMS Conference on Mass Spec-
trometry and Allied Topics, Long Beach, CA; ASMS: 2000.
(8) Blank, I.; Fay, L. B.; Lakner, F. J.; Schlosser, M. Determination
of 4-hydroxy-2,5-dimethyl-3(2H)-furanone and 2(or 5)-ethyl-4-
hydroxy-5(or 2)-methyl-3(2H)-furanone in pentose sugar-based
Maillard model systems by isotope dilution assays. J. Agric. Food
Chem. 1997, 45, 2642-2648.
(9) Engel, W.; Bahr, W.; Schieberle, P. Solvent assisted flavor
evaporation: a new and versatile technique for the careful and
direct isolation of aroma compound from complex food matrices.
Eur. Food Res. Technol. 1999, 209, 234-241.
(10) Bemelmans, J. M. H. Review of isolation and concentration
techniques. In Progress in FlaVor Research; Land, D. G.,
Nursten, H. E., Eds.; Applied Sciences: London, U.K., 1979;
pp 79-88.
(11) Mukaiyama, T.; Takei, H.; Shimizu, H. The reaction of N,N′-
sulfinylbis(dialkylamines). Bull. Chem. Soc. Jpn. 1967, 40, 939-
941.
(12) Kulikovskaya, E. A.; Kuznetsova, T. G.; Gritsaev, E. I.; Slizhov,
Yu. E.; Dozmorov, S. V. Synthesis of some aliphatic amino
sulfur compounds. Tr. Tomsk. Gos. UniV. 1973, 249, 31-33.

GC-oaTOFMS in Flavor Research
J. Agric. Food Chem., Vol. 51, No. 9, 2003 2713
(13) Blank, I.; Fay, L. B. Formation of 4-hydroxy-2,5-dimethyl-3(2H)-
furanone and 4-hydroxy-2(or 5)-ethyl-5(or 2)-methyl-3(2H)-
furanone through Maillard reaction based on pentose sugars. J.
Agric. Food Chem. 1996, 44, 531-536.
(14) Blank, I. Gas chromatography-olfactometry in food aroma
analysis. In Techniques for Analyzing Food Aroma; Marsili, R.,
Ed.; Dekker: New York, 1997; pp 293-329.
(15) Guth, H.; Grosch, W. Deterioration of soya-bean oil: Quantifica-
tion of primary flavor compounds using a stable isotope dilution
assay. Lebensm. Wiss. Technol. 1990, 23, 513-522.
(16) Kubota, K.; Watanabe, K.; Kobayashi, A. Novel dithiazine
compounds in volatile components from cooked Sakuraebi
(Sergia lucens Hansen). Agric. Biol. Chem. 1988, 52, 1537-
1540.
(17) Kubota, K.; Nakamoto, A.; Moriguchi, M.; Kobayashi, A.; Ishii,
H. Formation of pyrrolidino[1,2-e]-4H-2,4-dimethyl-1,3,5-dithi-
azine in the volatiles of boiled short-necked clam, clam, and
corbicula. J. Agric. Food Chem. 1991, 39, 1127-1130.
(18) Kawai, T.; Ishida, Y. Comparison of volatile components of dried
squid to reaction products formed from the mixture of hydrogen
sulfide, ammonia, and aldehydes. J. Agric. Food Chem. 1989,
37, 1026-1031.
2. Applications using the dual-electrospray ion source. Anal.
Chem. 2001, 73, 2605-2612.
(20) Hau, J.; Stadler, R.; Jenny, T. A.; Fay, L. B. Tandem mass
spectrometric accurate mass performance of time-of-flight and
Fourier transform ion cyclotron resonance mass spectrometry:
a case study with pyridine derivatives. Rapid Commun. Mass
Spectrom. 2001, 15, 1840-1848.
(21) Werkhoff, P.; Gu¨nthert, M.; Hopp, R. Dihydro-1,3,5-dithiaz-
ines: unusual flavor compounds with remarkable organoleptic
properties. Food ReV. Int. 1992, 8 (3), 391-442.
(22) Clauwaert, K. M.; Van Bocxlaer, J. F.; Major, H. J.; Claereboudt,
J. A.; Lambert, W. E.; Van den Eeckhout, E. M.; Van Peteghem,
C. H.; De Leenheer, A. P. Investigation of the quantitative
properties of the quadrupole orthogonal acceleration time-of-
flight mass spectrometer with electrospray ionisation using 3,4-
methylenedioxymethamphetamine. Rapid Commun. Mass Spec-
trom. 1999, 13, 1540-1545.
(23) Schieberle, P.; Grosch, W. Quantitative analysis of aroma
compounds in wheat and rye bread crusts using stable isotope
dilution assay. J. Agric. Food Chem. 1987, 35, 252-257.
(19) Wolff, J. C.; Eckers, C.; Sage, A. B.; Giles, K.; Bateman, R.
Accurate mass liquid chromatography/mass spectrometry on
quadrupole orthogonal acceleration time-of-flight mass analyzers
using switching between separate sample and reference sprays.
Received for review November 12, 2002. Revised manuscript received
January 31, 2003. Accepted February 16, 2003.
JF0261203