Content uploaded by María Molina Calle
Author content
All content in this area was uploaded by María Molina Calle on Apr 20, 2016
Content may be subject to copyright.
RESEARCH PAPER
HS–GC/MS volatile profile of different varieties of garlic
and their behavior under heating
María Molina-Calle
1,2,3
&Feliciano Priego-Capote
1,2,3
&María D. Luque de Castro
1,2,3
Received: 18 December 2015 /Revised: 1 March 2016 /Accepted: 8 March 2016
#Springer-Verlag Berlin Heidelberg 2016
Abstract Garlic is one of the most used seasonings in the
world whose beneficial health effects, mainly ascribed to
organosulfur compounds, are shared with the rest of the
Allium family. The fact that many of these compounds are
volatile makes the evaluation of the volatile profile of garlic
interesting. For this purpose, three garlic varieties—White,
Purple,andChinese—cultivated in the South of Spain were
analyzed by a method based on a headspace (HS) device
coupled to a gas chromatograph and mass detector (HS–GC/
MS). The main temperatures in the HS were optimized to
achieve the highest concentration of volatiles. A total number
of 45 volatiles were tentatively identified (among them 17
were identified for the first time in garlic); then, all were clas-
sified, also for the first time, and their relative concentration in
three garlic varieties was used to evaluate differences among
them and to study their profiles according to the heating time.
Chinese garlic was found to be the richest variety in sulfur
volatiles, while the three varieties presented a similar trend
under preset heating times allowing differentiation between
varieties and heating time using principal component analysis.
Keywords Garlic varieties .Headspace .GC/MS .Sulfur
volatiles .Flavor .Heating times
Abbreviations
GC Gas chromatography
HS Headspace
MF Molecular feature
MS Mass spectrometry
PCA Principal component analysis
SPME Solid phase microextraction
Introduction
Garlic (Allium sativum) is widely used as a seasoning in cui-
sines worldwide, especially in Asia, Africa, and Europe.
Historically, garlic was used by the Egyptians in several ther-
apeutic formulas [1]. Also, Greeks and Romans used garlic as
a healing agent, which led to the expansion of its use to the
whole Mediterranean region [2]. Spain assumed this historical
heritage and became one of the major producers of garlic in
the European Union, increasing its production to reach 173
million tons of garlic in 2013 [3]. Research to explain gar-
lic’s properties, known since antiquity, started in the second
half of the nineteenth century when Louis Pasteur assessed its
antibacterial properties [4]. More recent studies have reported
that garlic extracts act as antioxidant and antimicrobial agents
and also produce beneficial effects against cardiovascular dis-
eases, as reviewed by Corzo-Matínez et al. [5]. Nowadays,
attention has focused on the cancer preventive properties of
Electronic supplementary material The online version of this article
(doi:10.1007/s00216-016-9477-0) contains supplementary material,
which is available to authorized users.
*Feliciano Priego-Capote
q72prcaf@uco.es
*María D. Luque de Castro
qa1lucam@uco.es
1
Department of Analytical Chemistry, University of Córdoba, Annex
Marie Curie Building, Campus of Rabanales, Carretera Nacional IV
Km. 396, 14014 Córdoba, Spain
2
Maimónides Institute of Biomedical Research (IMIBIC), Reina Sofía
University Hospital, University of Córdoba, 14014 Córdoba, Spain
3
University of Córdoba, ceiA3 Agroalimentary Excellence Campus,
Campus of Rabanales, 14071 Córdoba, Spain
Anal Bioanal Chem
DOI 10.1007/s00216-016-9477-0
garlic [6–10]. For example, Jin et al. showed the protective
association between garlic intake and lung cancer [11], while
Huang et al. demonstrated the antiproliferative effect on colon
cancer cells of diallyl disulfide, one of the most characteristic
compounds in garlic [12].
The beneficial properties of garlic have been related to the
presence of organosulfur compounds, mainly thiosulfinates
and sulfur volatiles, which are shared with the rest of the
Allium family. Thiosulfinate derivatives of S-alk(en)yl-L-cys-
teine are unstable compounds and undergo chain reactions
that transform them into volatiles [13], which are responsible
for the characteristic garlic aroma and its potential anticancer
activity, as demonstrated by Yang et al. and Seki et al. for
diallyl disulfide and diallyl trisulfide [14,15].
Methods for identification of volatile compounds in garlic
have relied on sample preparation procedures based on solid
phase microextraction (SPME) or headspace separation (HS)
prior to gas chromatography–mass spectrometry (GC–MS)
[16–21]. Sample preparation procedures proposed by
Warren et al. for extraction of thiol volatiles were based on
either in-needle or in-fiber derivatization with N-
phenylmaleimide, which endowed the products with high se-
lectivity [16]. Lee et al. compared different procedures for
extraction of volatiles (viz., steam distillation, simultaneous
distillation and solvent extraction, solid-phase trapping sol-
vent extraction, and HS–SPME), resulting HS–SPME the
most suitable because of the absence of solvent and the rela-
tively low temperature that hindered degradation of thermola-
bile volatiles [20]. Kim et al. evaluated the volatile com-
pounds in garlic subjected to different processing conditions:
autoclaving, high temperature aging (black garlic), crushing,
and roasting at different temperatures. The resulting garlic
products were analyzed by HS–SPME–GC/MS and 26 vola-
tiles were identified and their relative quantifies in the samples
were determined [17]. Finally, in a recent work by Radulović
et al. 78 organosulfur compounds were detected in the essen-
tial oil of wild garlic using GC−FID and GC−MS [22].
Although this study claims that the sulfur-volatile content in
garlic was evaluated, a liquid extract was obtained and subse-
quently analyzed, instead of using a sampling device to obtain
thevolatilefraction.
The present research aimed to address the absence of in-
depth studies on the volatile non-sulfurous compounds profile
of garlic and on the behavior of volatile compounds in garlic
when subjected to different heating times. Our specific objec-
tives were (i) to optimize a cheap and easy analytical method
for garlic volatiles based on HS–GC/MS without the need for
an SPME step; (ii) to identify the compounds that make up the
volatile profile of garlic, especiallythose which have so far not
received enough attention; (iii) to differentiate the volatile
profiles of three garlic varieties cultivated in the South of
Spain; and (iv) to study differences in the influence of preset
heating times on the volatile profile of each variety.
Materials and methods
Samples
Fresh garlic from three varieties, Purple (var. Rocambole),
White (var. Porcelain), and Chinese (var. Turban), were pro-
vided by La Abuela Carmen (Montalbán, Spain). Four bulbs
of each variety were stored at −20 °C and, prior to analysis, the
cloves were separated and chopped, located into the HS vial,
and encapsulated. The analysis was carried out immediately
after pretreatment to avoid enzymatic degradation of the sam-
ple. Each sample was analyzed in duplicate.
Apparatus
Ten-milliliter HS vials sealed with 20-mm aluminum vial caps
(Análisis Vínicos, Tomelloso, Spain) and 20-mm silicone/
PTFE septa (Análisis Vínicos) placed in the rack of a 7694E
headspace autosampler from Agilent (Palo Alto, CA, USA)
were used. An Agilent 7890B gas chromatograph coupled to
an Agilent 5977A mass spectrometer was used for appropriate
separation and detection of volatile compounds. A Factor VF-
5 ms fused silica capillary column (30 m × 0.25 mm I.D.,
0.25 μm film thickness, Varian) completed the experimental
setup.
HS–GC/MS analysis
Vial, loop, and transfer line temperatures were set at 103, 113,
and 123 °C, respectively. The vial was heated without shaking
for 10 min and pressurized for 12 s at 30 psi; the loop filling
and equilibration times were set at 12 and 3 s, respectively;
and the injection time was set at 45 s. The sample was
chopped and 2 g of it was placed in the HS vial. Three milli-
liters of the headspace content was injected into the chromato-
graph. The injector temperature was set at 180 °C and injec-
tion was performed in a 1/5 split mode. The gasflow was set at
1 mL/min. The oven temperature was programmed as follows:
initial temperature 40 °C (held for 5 min) and increased at
10 °C/min to 250 °C (held for 5 min). The total analysis time
took 31 min, and 3 min extra time was necessary for re-
establishing and equilibrating the initial conditions. The single
quadrupole mass spectrometer was operated in the scan mode,
for which the instrumental parameters were set as follows:
transfer line, source, and quad temperatures were kept at
250, 230, and 180 °C, respectively; electron energy was set
at 70 eV; data acquisition was set between m/z 30 and 500; and
the solvent delay was programmed for 2 min.
Experimental design
Optimization of the principal parameters for HS was support-
ed on a Box–Behnkendesignbyapplicationofresponse
M. Molina-Calle et al.
surface methodology. This design evaluated three levels for
each studied factor: vial temperature (X
1
), vial equilibration
time (X
2
), and injection time (X
3
). The resulting experimental
design included 15 experiments by defining the minimum,
maximum, and average values of each parameter: 60, 90,
and 120 °C for vial temperature; 5, 7.5, and 10 min for vial
equilibration time; and 12, 3, and 45 s for injection time.
Seven of the most intense peaks in different sections of the
chromatogram were selected as response variables, Y
k
(x). For
the first peak, the minimum response was considered as opti-
mal because this peak is generated by compounds of low
molecular mass and high volatility, suspected of being gener-
ated by degradation. The experimental data were fitted to a
second-order polynomial model by the following equation:
Yi¼β0þβmXmþβnXnþβmmX2
mþβmnXmXnþβnn X2
n
A desirability function approach was used to select the
condition for each factor that resulted in the best concentration
of volatiles in the HS according to the following equation:
D¼d1Y1
ðÞ
d2Y2
ðÞ
…dnYn
ðÞðÞ
1=n
where d
i
(Y
i
) is a desirability function for each response Y
i
(x).
When the d
i
(Y
i
) function is equal to 1, this represents a
completely desirable or, in this case, the ideal response value,
obtained by the surface response model. The results were
processed using the software Statgraphics Centurion XVI
(StatPoint Technologies 2011, USA).
Data processing and statistical analysis
Qualitative Analysis software (version 7.0, Agilent
Technologies, Santa Clara, CA, USA) was used to process
all data obtained by GC/MS. Treatment of the raw data files
started by deconvolution of potential molecular features
(MFs) by the algorithm included in the software, which con-
sidered all ions exceeding 2000 counts for the absolute area
parameter. The NIST Mass Spectral Search Program v. 11.0
(NIST, Washington, DC, USA) was used for spectral search
(Mainlib and Replib libraries). Tentative identification was
supported on correlation between experimental and database
spectra above 650 in normal search mode.
Peaks corresponding to identified entities were integrated
in all samples and the resulting areas were compiled in a
matrix using a .csv file. The matrix was exported to the
Mass Profiler Professional (MPP) software package (Version
12.0, Agilent Technologies, Santa Clara, CA, USA), where
the data set was treated by principal component analysis
(PCA) to evaluate the behavior of the three garlic varieties
under preset heating times.
Results and discussion
Optimization of HS parameters
The optimization study focused on the three main HS param-
eters that afforded the highest concentration of volatile com-
pounds in the headspace of the vial. As described in
BMaterials and methods^, a response surface methodology
was used for this purpose. Independent studies were carried
out for each selected peak of the chromatogram (see
Electronic Supplementary Material (ESM) Fig. S1), and the
results were combined in a desirability response surface with
capability to define the optimum conditions to obtain the max-
imum signal for peaks 2 to 7 (identified as diallyl sulfide, 1,3-
dithiane, diallyl disulfide, allyl 1-propenyl disulfide, 3-vinyl-
1,2-dithiacyclohex-5-ene, and diallyl trisulfide, respectively)
and the minimum signal for peak 1 (the sum of acetone and 2-
propen-1-ol). This criterion was applied to obtain the maxi-
mum signal from compounds of interest in garlic with the
greatest variety of them, and also with the minimum signal
from less important compounds produced by degradation of
higher compounds. The resulting desirable conditions are
listed in Table 1. Figure 1shows the response surface reported
by this study. As can be seen in the figure, the desirable tem-
perature in the vial was 103 °C, slightly above the boiling
point of water, which suggested that volatile compounds of
garlic are stored in the aqueous intracellular medium. The
injection time adopted a high value (0.75 min), without
reaching the maximum studied value. Finally, the desirable
value of vial equilibration time was 10 min, the maximum
value studied. Therefore, the heating kinetics was studied to
understand the behavior of the target compounds at vial equil-
ibration times longer than 10 min. Figure 2shows the evolu-
tion of the intensity of the target compounds to reach the
equilibration time, where an overall decreased intensity of
the target peaks after 60 min of equilibration time can be
observed. Nevertheless, at 30 min of heating time compounds
such as 3-dithiane, diallyl sulfide, 3-vinyl-1,2-dithiacyclohex-
5-ene, and diallyl trisulfide experienced a slight increase of
intensity; however, the increase for the peak corresponding
to the sum of acetone and 2-propen-1-ol was more noticeable.
These two last compounds are characterized by low molecular
mass and high volatility, and they can be considered as final
Table 1 Optimization of the main HS parameters by a desirability
study to maximize the concentration of volatilecompounds in the HS vial
Factor Low High Optimum
Vial temperature (°C) 60.0 120.0 103
Vial equilibration time (min) 5.0 10.0 10.0
Injection time (s) 12 48 45
HS–GC/MS volatile profile of different varieties of garlic
degradation products. Thus, a heating time of 10 min was
considered the best to obtain the volatile profile of garlic at
the working temperature.
Tentative identification of volatile compounds
The headspace of the vial containing the chopped garlic was
generated under the desirability conditions and analyzed by
GC/MS as described in BMaterials and methods^. Tentative
identification of the peaks was carried out by comparison of
the resulting spectra with the theoretical spectra stored in the
National Institute of Standards and Technology (NIST) data-
base, establishing a cut on the match factor of 650 (65 %
correlation between the experimental and theoretical spectra).
Also, a standard alkanes mixture was analyzed to calculate the
experimental RT index to support the identification by an al-
kanes calibration model. Probably as a result of the different
matrixes, i.e., gas in the case of the HS of the samples and
liquid in the case of the standard alkanes mixture, a mismatch
between the experimental and the theoretical RT index (ob-
tained from the NIST) was produced. To overcome it, a cor-
relation plot was generated to obtain a corrected RT index by
extrapolation of the experimental value in the plot (see ESM
Fig. S2). Table 2lists the tentatively identified compounds, the
main parameters that support identification, and also high-
lights those identified for the first time. The tentatively iden-
tified compounds were classified into derivatives of S-
alk(en)yl-L-cysteine, flavor compounds, and others. Flavor
compounds were classified as a function of their flavor, as
listed in http://www.thegoodscentscompany.com/. Twenty-
one volatile organosulfur compounds were tentatively identi-
fied in the content of the headspace. These compounds are
originated from S-alk(en)yl-L-cysteine by a series of chain
reactions. This derivative of L-cysteine is oxidized, generating
S-alk(en)yl-L-cysteine-S-oxides among which alliin is the
most abundant in garlic. When garlic is chopped, the enzyme
alliinase is activated and transforms alliin and the rest of
alk(en)yl-L-cysteine-S-oxides into sulfenic acid intermediates.
In the aqueous intracellular medium, these intermediates are
rapidly condensed into thiosulfinates, the major thiosulfinate
present in garlic being allicin, formed from alliin. These com-
pounds are very unstable and their subsequent reactions lead
to a wide variety of sulfur derivatives, most of them volatile
[13]. Within this group of volatiles, diallyl mono-, di-, and
trisulfide and 1,3-dithiane are four of the most characteristic
sulfur volatiles present in garlic. The sulfur volatiles presented
a similar fragmentation pattern, as can be seen in Table 2,
where the most intense fragments of each compound are
listed. Nevertheless, isomers such as diallyl and di(1-
propenyl) presented a distinguishable fragmentation pattern,
as shown Fig. 3; the di(1-propenyl) isomer presented a num-
ber of fragments at higher m/z than the diallyl isomer, while
the former provided higher intensity of the major fragments.
The spectrum of the diallyl isomer shows a major fragment at
m/z 41.1, generated by rupture of the bond between the sulfur
atom and the alpha carbon. This fragment was also present in
the spectrum corresponding to di(1-propenyl) isomer but its
Fig. 1 Response surface associated with the desirability conditions for
the main HS parameters. Optimization was carried out by monitoring
seven representative compounds by HS–GC/MS
Fig. 2 Heating kinetics of
monitored compounds, expressed
as peak area versus heating time
M. Molina-Calle et al.
Tab le 2 Volatiles identified in the headspace generated from garlic
Compound CAS RT (min) RT index Theoretical RT index Match R. Match Mass (Da) Formula Principal fragments
Derivatives from S-alk(en)yl-L-cysteine
Dimethyl sulfide 75-18-3 1.67 553 455 813 815 58.0 C
4
H
6
O 32.0, 41.0, 43.0, 56.0
Allyl mercaptan 870-23-5 2.07 574 562 817 831 74.0 C
3
H
6
S 39.1, 41.1, 72.1, 74.0
Allyl methyl sulfide 10152-76-8 3.12 629 660 813 820 88.0 C
4
H
8
S 39.0, 45.0, 73.0, 88.0
Dimethyl disulfide 624-92-0 4.07 679 722 892 910 94.0 C
2
H
6
S
2
45.0, 55.0, 84.0, 93.9
Diallyl sulfide 592-88-1 7.42 854 849 820 881 114.1 C
6
H
10
S 39.0, 45.0, 73.0, 114.0
Allyl n-propyl sulfide 27817-67-0 7.81 875 859 828 867 116.1 C
6
H
12
S 32.0, 43.0, 44.0, 55.0
Allyl 1-propenyl sulfide 122156-02-9 8.18 894 114.1 C
6
H
10
S 32.0, 45.0, 99.0, 114.0
Di(1-propenyl) sulfide 65819-74-1 8.27 899 884 750 800 114.1 C
6
H
10
S 32.0, 41.0, 45.0, 99.0
1,3-Dithiane 505-23-7 8.85 912 1002 768 772 120.2 C
4
H
8
S
2
39.0, 41.0, 45.0, 119.9
Allyl methyl disulfide 2179-58-0 9.08 941 911 654 698 120.0 C
4
H
8
S
2
32.0, 41.0, 72.0, 115.9
Methyl 1-propenyl disulfide 10152-77-9 9.43 960 928 750 795 120.0 C
4
H
8
S
2
45.0, 72.0, 75.0, 119.9
Dimethyl trisulfide 3658-80-8 10.09 995 972 902 927 126.0 C
2
H
6
S
3
45.0, 63.9, 78.9, 125.9
Diallyl disulfide 2179-57-9 12.25 1108 1099 865 872 146.0 C
6
H
10
S
2
39.0, 41.1, 81.0, 146.0
Allyl 1-propenyl disulfide 122156-03-0 12.52 1122 146.0 C
6
H
10
S
2
41.1, 45.0, 73.0, 146.0
Di(1-propenyl) disulfide 53925-82-9 12.63 1127 1103 816 819 146.0 C
6
H
10
S
2
41.1, 45.0, 73.0, 146.0
Allyl methyl trisulfide 34135-85-8 13.28 1162 1161 656 715 152.0 C
4
H
8
S
3
41.0, 45.0, 73.0, 87.0
3-Vinyl-1,2-dithiacyclohex-4-ene 62488-52-2 14.10 1205 1134 778 926 144.0 C
6
H
8
O
2
79.0, 97.0, 111.0, 144.0
3-Vinyl-1,2-dithiacyclohex-5-ene 62488-53-3 14.48 1224 1155 900 956 144.0 C
6
H
8
O
2
71.0, 72.0, 111.0, 144.0
Benzothiazole* 95-16-9 14.66 1234 1208 863 925 135.0 C
7
H
5
NS 32.0, 44.0, 108.0, 134.9
Diallyl trisulfide 2050-87-5 15.72 1292 1350 854 875 178.0 C
6
H
10
S
3
39.0, 41.1, 73.0, 113.0
Di(1-propenyl) trisulfide 115321-81-8 16.15 1312 178.0 C
6
H
10
S
3
41.0, 45.0, 73.0, 114.0
Diallyl tetrasulfide 2444-49-7 18.95 1459 1510 812 826 210.0 C
6
H
10
S
4
39.0, 41.0, 73.0, 146.0
Flavor compounds
Green/floral flavor
2-Butenal 4170-30-3 2.53 598 615 943 946 70.0 C
4
H
6
O 39.0, 41.0, 69.0, 70.0
3-Penten-2-one* 625-33-2 3.98 674 714 799 927 84.1 C
5
H
9
O 41.0, 43.0, 69.0, 84.0
2-Methyl-4-pentenal 108-11-2 4.19 685 732 796 928 98.1 C
6
H
10
O 41.0, 43.0, 45.0, 55.0
2-Hexanol* 626-93-7 4.38 695 780 720 779 102.1 C
6
H
14
O 39.0, 55.0, 83.0, 84.0
5-Hexen-2-one* 109-49-9 5.01 728 744 908 909 98.1 C
6
H
10
O 39.0, 43.0, 55.0, 83.0
5-Hexenal* 764-59-0 5.35 746 796 875 918 98.1 C
6
H
10
O 39.0, 41.0, 54.0, 80.0
Hexanal* 66-25-1 5.74 767 803 890 892 100.1 C
6
H
12
O 41.0, 44.0, 56.1, 72.0
4-Heptenal 6728-31-0 7.09 837 870 840 842 112.1 C
7
H
12
O 41.0, 53.0, 67.0, 68.0
2,4-Hexadienal 142-83-6 8.77 925 877 833 902 96.1 C
6
H
8
O 53.0, 67.0, 96.0, 81.0
Benzaldehyde* 100-52-7 9.91 985 920 909 918 106.0 C
7
H
6
O 51.0, 77.0, 92.0, 106.0
3-Methyl-benzaldehyde* 620-23-5 12.05 1097 1060 884 897 120.1 C
8
H
8
O 65.0, 91.0, 119.0, 120.0
Roasted flavor
3-Methyl butanal* 590-86-3 2.59 601 643 930 930 86.0 C
5
H
10
O 39.0, 41.0, 69.0, 70.0
2-Methyl butanal* 96-17-3 2.70 607 643 912 912 86.0 C
5
H
10
O 41.0, 43.0, 57.1, 58.0
2-Pentenal 1576-87-0 2.96 621 697 885 890 84.1 C
5
H
8
O 39.0, 41.0, 55.0, 83.0
Furfural* 98-01-1 6.72 818 831 960 961 96.0 C
5
H
4
O
2
39.0, 67.0, 95.0, 96.0
3-Ethyl pyridine* 536-78-7 9.86 982 937 879 914 107.1 C
7
H
9
N 51.0, 77.0, 105.0, 106.0
Sweet flavor
1-Hydroxy-2-propanone* 116-09-6 3.72 660 698 832 873 74.0 C
3
H
6
O
2
31.0, 43.0, 45.0, 74.0
2-Acetylfuran* 1192-62-7 8.80 927 878 943 946 110.0 C
6
H
6
O
2
39.0, 42.0, 81.0, 108.0
Various flavor
2-Ethyl-2-butenal (pungent)* 19780-25-7 6.30 796 791 935 938 98.1 C
6
H
10
O 41.0, 55.0, 69.0, 98.0
5-Ethyl-2-methyl-pyridine (nutty)* 104-90-5 11.21 1053 1000 845 882 121.1 C
8
H
11
N 45.0, 77.0, 106.0, 121.0
Other compounds
Acetone* 67-64-1 1.67 569 455 813 816 58.0 C
3
H
6
O 40.0, 41.0, 43.0, 56.0
2-Propen-1-ol 107-18-6 1.86 579 552 945 947 58.0 C
3
H
6
O 31.1, 39.1, 57.0, 58.0
2-Methylene-4-pentenal 108-11-2 4.26 689 763 829 834 96.1 C
6
H
8
O 39.0, 41.0, 67.0, 95.0
*Compounds tentatively identified for the first time in garlic
HS–GC/MS volatile profile of different varieties of garlic
relative intensity was lower than that of the diallyl isomer. The
di(1-propenyl) isomer provided one other intense fragment at
m/z 73.0 generated by rupture of the sulfur–sulfur bond; thus,
the presence of a double bond between the alpha and beta
carbons in the di(1-propenyl) isomer stabilizes the alpha car-
bon–sulfur bond. The spectra of both compounds were also
characterized by the occurrence of characteristic fragments
generated by cyclization. Two of these fragments, m/z 39.0
and 45.0, are produced by intramolecular fragmentation of
both rings. Additionally, the two rings can be opened to gen-
erate the m/z 81.0 and 104.9 fragments from the diallyl isomer
and the di(1-propenyl) isomer, respectively. Table 2does not
list the theoretical RT index, match, and R of the allyl 1-
propenyl isomer. Match is not listed because these compounds
are not present in the NIST database; however, their spectra
are characterized by the same fragments as diallyl and di(1-
propenyl) isomers. Also, Radulovićet al. identified these
compounds and obtained the same elution order for the three
isomers as in this study. The same applies to the identification
of di(1-propenyl) trisulfide.
The structural stability caused by the presence of one dou-
ble bond between the alpha and beta carbons can also be
observed in the fragmentation of the 3-vinyl-dithiacyclohex-
4-ene and 3-vinyl-dithiacyclohex-5-ene isomers (see ESM
Fig. S3). The first isomer, originated from the cyclization of
the diallyl disulfide, is fragmented by rupture of the alpha
carbon–sulfur and sulfur–sulfur bonds, generating the frag-
ments m/z 79.0, 97.0, and 111.0. On the other hand, 3-vinyl-
dithiacyclohex-5-ene results from cyclization of allyl 1-
propenyl disulfide, yielding as major fragment the m/z 72.0
generated by rupture of the heterocycle through the sulfur–
sulfur bond. This fragmentation leads to two fragments at
m/z 72.0, which explain the high intensity of the peak in the
spectra of 3-vinyl-dithiacyclohex-5-ene.
Tentatively identified compounds that provide flavor to
garlic constitute the second group of relevant volatile compo-
nents. Eleven compounds with green or floral flavors were
identified in the headspace from garlic samples, which also
included five hexane derivatives. Five compounds with
roasted flavor were also identified, among which furfural
(which presents a roasted almond flavor) is the most remark-
able. This compound, characteristic of the Maillard reaction,
results from the degradation of a pentose sugar; as does a
derivative from furfural, 2-acetylfuran, also identified in the
headspace, which provides a characteristic sweet flavor, quite
similar to 1-hydroxy-2-propanone. Finally, two compounds
with different flavor were tentatively identified in the head-
space: 2-ethyl-2-butenal, with a pungent flavor, and 5-ethyl-2-
methyl-pyridine, characterized by a nutty flavor.
The last group encompasses three compounds without fla-
vor characteristics (acetone, 2-methylene-4-pentenal, and 2-
propen-1-ol), the first of these being tentatively identified for
the first time in the headspace from garlic, all of them with a
high volatility as their peaks appeared in the chromatogram
within the first 4 min.
Comparison of garlic varieties
Purple,White, and Chinese garlic varieties were analyzed by
the HS–GC/MS method described in BMaterials and
methods^. Table 3shows the percentage referred to as the total
area of the tentatively identified compounds in the three vari-
eties. Within the organosulfur compounds, it is worthy
distinguishing that diallyl disulfide and diallyl trisulfide are
major garlic volatiles which represented a relative concentra-
tion of 51.72 % for Chinese garlic, 43.37 % for White garlic,
and 42.32 % for Purple garlic. Also, Chinese garlic was the
variety with the highest number of sulfur volatiles, allyl
Fig. 3 MS spectra and
fragmentation patterns of diallyl
disulfide (a) and di(1-propenyl)
disulfide (b)
M. Molina-Calle et al.
methyl disulfide being the only sulfur volatile not detected in
this sample; allyl methyl disulfide was only detected in White
garlic. On the other hand, benzothiazole and di(1-propenyl)
trisulfide were only detected in Chinese garlic.
Concerning green or floral flavor compounds, White garlic
must be highlighted because it provided the major variety of
these volatiles; however, Purple garlic presented the highest
concentration of 2-butenal, the major compound of this group
in all the samples. Comparing the total concentration of these
volatiles (11.87 % for Chinese garlic, 14.38 % for White gar-
lic, and 15.01 % for Purple garlic) it can be said that the
Purple variety provided a major green and floral flavor.
Roasted and sweet flavor compounds were not detected or
detected at a concentration below 1 %. Only 3-methyl butanal
was detected in the three varieties, while 2-methyl butanal was
only detected in White garlic, and 2-pentenal was detected in
Chinese and White garlic, but at a concentration below 0.1 %.
Finally, 2-ethyl-2-butenal, with a pungent flavor, was detected
in the three varieties, but at a high concentration in White and
Purple garlics.
Acetone, 2-propen-1-ol, and 2-methylene-4-pentenal were
also identified in the varieties under study, with a different
relative concentration depending on the variety. Acetone and
2-porpen-1-ol are clearly the major compounds within this
group, the sum of both representing more than 5 % in all cases.
It should be mentioned that the concentration of acetone was
higher in White garlic, while 2-propen-1-ol was more concen-
trated in Purple garlic.
Heating kinetics of the three garlic varieties
The pool of each sample was divided into eight 2-g
portions, each of them placed in a headspace vial and
subjected to 103 °C for preset times (10, 20, 30, 40,
50, 60, 90, or 120 min), then analyzed by HS–GC/MS
as described in BMaterials and methods^. Peaks corre-
sponding to tentatively identified compounds were
Table 3 Relative concentration (percentage of the total area) of the
compounds identified in the three garlic varieties: Purple,Chinese,and
White
Compound Garlic varieties
Chinese White Purple
Derivatives from S-alk(en)yl-L-cysteine
Dimethyl sulfide 0.20 0.35 0.06
Allyl mercaptan 0.14 –
a
1.4
Allyl methyl sulfide 1.1 2.3 3.6
Dimethyl disulfide 1.7 3.1 1.2
Diallyl sulfide 3.5 6.7 4.8
Allyl n-propyl sulfide 0.05 0.06 –
Allyl 1-propenyl sulfide 0.08 0.11 0.06
Di(1-propenyl) sulfide 0.12 0.29 0.16
1,3-Dithiane 3.0 2.9 5.4
Allylmethyldisulfide –0.04 –
Methyl 1-propenyl disulfide 0.33 0.33 0.33
Dimethyl trisulfide 0.13 0.18 0.74
Diallyl disulfide 26.4 24.8 21.0
Allyl 1-propenyl disulfide 2.8 1.7 0.90
Di(1-propenyl) disulfide 7.3 5.0 2.6
Allylmethyltrisulfide 5.0 4.6 8.8
3-Vinyl-1,2-dithiacyclohex-4-ene 0.38 0.19 0.26
3-Vinyl-1,2-dithiacyclohex-5-ene 1.1 0.55 0.92
Benzothiazole 0.05 ––
Diallyl trisulfide 25.3 18.5 21.4
Di(1-propenyl) trisulfide 0.38 ––
Diallyl tetrasulfide 0.64 0.32 0.21
Flavor compounds
Green/Floral flavor
2-Butenal 7.8 11.0 12.2
3-Penten-2-one 0.05 0.08 0.04
2-Methyl-4-pentenal 0.24 0.40 0.25
2-Hexanol –0.13 –
5-Hexen-2-one 0.32 0.24 0.18
5-Hexenal 0.48 0.55 0.29
Hexanal 0.47 0.25 0.28
4-Heptenal 1.8 1.5 1.0
2,4-Hexadienal 0.64 0.23 0.74
Benzaldehyde 0.03 0.04 –
3-Methyl-benzaldehyde –––
Roasted flavor
3-Methyl butanal 0.29 0.53 0.48
2-Methyl butanal –0.16 –
2-Pentenal 0.06 0.08 –
Furfural –––
3-Ethyl pyridine –––
Sweet flavor
1-Hydroxy-2-propanone –––
2-Acetylfuran –––
Table 3 (continued)
Compound Garlic varieties
Chinese White Purple
Other flavors
2-Ethyl-2-butenal (pungent) 0.81 1.4 1.9
5-Ethyl-2-methyl-pyridine (nutty) –––
Other compounds
Acetone 1.4 4.1 1.9
2-Propen-1-ol 4.5 4.9 5.3
2-Methylene-4-pentenal 0.38 0.33 0.29
a
Compounds that were not detected in the sample
HS–GC/MS volatile profile of different varieties of garlic
integrated and the resulting area was used to quantify
the compounds as a relative percentage of the total
sum of the peak areas. The results from the heating
kinetics for each type of sample are shown in ESM
Tab le S1. As can be seen, the three garlic varieties
are rich in sulfur volatiles, the sum of which is more
than50%ofthetotalpeakareaforallthestudied
heating times. It should be emphasized that diallyl di-
sulfide and diallyl trisulfide are major compounds in
the samples, reaching percentages above 18 % in all
instances. The literature on garlic shows that diallyl
mono-, di, tri-, and tetrasulfide are four of the most
common volatiles generated from S-alk(en)yl-L-cyste-
ine; therefore, their evolution with the heating time
and temperature is key to the presence and concentra-
tion of sulfur volatiles [13]. Figure 4shows a plot of
percentage of total area versus heating time for each of
the allyl sulfides discussed above. The three garlic va-
rieties showed similar behavior: while diallyl sulfide
and diallyl tetrasulfide increased their relative concen-
tration in the headspace with the heating time, diallyl
disulfide and diallyl trisulfide significantly decreased
during the same interval, even to one half of their
initial concentration in some cases.
Concerning tentatively identified compounds contrib-
uting to flavor, the headspace from the three varieties
was rich in compounds with green or floral flavor,
while sweet flavor compounds were never detected in
Chinese garlic and not detected during the first 20 min
of heating in Purple and White garlic. 2-Butenal was
the major green flavor compound detected in all garlic
samples; however, its concentration drastically decreased
within 10–20 min heating, probably because of degra-
dation. Within the group of roasted flavor compounds,
furfural was the most significant compound, detected in
the three varieties of garlic samples after 20 min of
heating, which reveals activation of Maillard reactions
in garlic when heated for more than 10 min. 2-
Acetylfuran and 5-methyl furfural, compounds with a
sweet flavor, are also products of the Maillard reaction,
as derivatives from furfural. However, they were only
detected within the 50–60minofheatinginWhite gar-
lic. Concerning the rest of compounds with various
flavors, only two compounds were relatively quantified:
(a) 2-ethyl-2-butenal was detected in all three varieties
and its concentration gradually decreased during heating
in all instances (thus, it can be assumed that the pun-
gent flavor provided by this compound is characteristic
of fresh garlic); (b) 5-ethyl-2-methyl-pyridine was only
detected in Purple garlic after 20 min heating and its
concentration progressively decreased with increasing
heating time.
Finally, the remaining compounds that do not provide fla-
vor were relatively quantified in all varieties; the behavior of
acetone was remarkable and its concentration increased
strongly after 10–20 min heating (more than 10 % of the total
peak area).
A matrix was created from the obtained data as de-
scribedintheBMaterial and methods^section and treat-
ed by PCA to evaluate differences among samples and
the contribution of each compound. Figure 5shows the
PCA plot of heating kinetics of the three garlic varie-
ties: Purple (P), Chinese (C), and White (W) garlic.
Component 3, plotted on the z-axis, clearly differentiates
Fig. 4 Evolution of the relative
concentration (percentage of the
total peak area) of diallyl sulfide,
diallyl disulfide, diallyl trisulfide,
and diallyl tetrasulfide in the three
varieties during the heating time
M. Molina-Calle et al.
between garlic varieties, while components 1 and 2,
plotted on the x- and y-axis, respectively, mark the evo-
lution of the samples according to the heating time. The
compounds with a major contribution to component 3
were dimethyl sulfide, acetone, and 3-penten-2-one;
thus, the concentration of these compounds in the HS
allows the differentiation of White garlic from the other
varieties. In contrast, allyl methyl sulfide, dimethyl tri-
sulfide, and methyl 2-propenyl trisulfide presented the
largest negative contribution to component 3, being re-
sponsible for the separation of Purple garlic samples in
the PCA. Concerning the separation as a function of the
heating time, compounds with high values on the x-axis
and low values on the y-axis are responsible for differ-
entiation of garlic samples heated for 10 min. 2-Butenal,
allyl mercaptan, and benzothiazole had higher values of
component 1, while hexanal, 3-vinyl-1,2-dithiacyclohex-
4-ene, and 2-butenal were the compounds with the low-
est values of component 2. The presence of 2-butenal
among the most significant compounds is consistent in
this case with its behavior discussed above; thus, the
concentration of this compound in the HS, indicative
of garlic freshness, corresponds to low degradation at
short heating times. On the other hand, compounds with
a low value of component 1 and a high value of com-
ponent 2 allow differentiation of samples with longer
heating times. Diallyl tetrasulfide, 1-propenyl allyl sul-
fide, and 2-methyl-4-pentenal presented the lowest value
of component 1, while 2-propen-1-ol, dimethyl trisul-
fide, and 2-pentenal provided the highest value of
component 2. Although the identified volatile com-
pounds classified as others are not characteristic of gar-
lic and do not provide any flavor, they seem to be
important for differentiation of both garlic varieties
and fresh–heated garlic, since most of them are respon-
sibleforthevariancealongthethreecomponents.
Conclusions
An HS–GC/MS method is proposed for the identification of
the volatile profile of garlic. The proposed method was opti-
mized to obtain the highest number of volatile compounds.
The optimized method was applied to garlic samples, which
led to tentative identification of 45 volatile compounds (17 of
which were identified for the first time in garlic) by study of
the fragmentation pattern of the most important sulfur com-
pounds. Comparison of the volatile profile of three garlic va-
rieties—Chinese,White, and Purple—showed that Chinese
garlic is the richest in sulfur volatiles, and Purple garlic has
the highest concentration of flavor volatiles. The proposed
method was also used to study the behavior of the three garlic
varieties under different heating times. The three varieties
showed similar behavior, flavor compounds being the most
illustrative of the evolution of garlic with the heating time. The
concentration of green and floral volatiles, freshness indica-
tors in garlic, drastically decreases with the heating time, while
roasted and sweet flavor compounds appear or increase their
concentration with increasing heating time. On the other hand,
sulfur volatiles presented both trends depending on the
Fig. 5 PCA of Purple (P),
Chinese (C), and White (W) garlic
samples heated for different times
(numbers express minutes of
heating)
HS–GC/MS volatile profile of different varieties of garlic
compound. Finally, these results allowed PCA differentiation
among both the varieties and the heating times; it could be
seen that all varieties had a similar trend when subjected to
preset heating times, although they can be clearly distin-
guished among themselves.
Acknowledgments La Abuela Carmen (Montalbán, Spain) is thanked
for the sample supply. The Spanish Ministerio of Economy and
Competitiveness (MINECO) is thanked for financial support through
project CTQ2012-37428. F.P.C. is also grateful to the Ministerio de
Ciencia e Innovación (MICINN) for a Ramón y Cajal contract (RYC-
2009-03921).
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
References
1. Block E. The chemistry of garlic and onions. Sci Am. 1985;252(3):
114–9.
2. Griffiths G, Trueman L, Crowther T, Thomas B, Smith B. Onions-
A global benefit to health. Phytoher Res. 2002;16(7):603–15.
3. FAOSTAT. Food and Agriculture Organization of the United
Nations Statistics Division. http://faostat3.fao.org/browse/Q/QC/
E. Accessed 25 Nov 2015.
4. Pasteur L. Mémoire surla fermentation appelée lactique. Ann Chim
Phys. 1858;52:404–18.
5. Corzo-Martínez M, Corzo N, Villamiel M. Biological properties of
onions and garlic. Trends Food Sci Technol. 2007;18(12):609–25.
6. Schäfer G, Kaschula CH. The immunomodulation and anti-
inflammatory effects of garlic organosulfur compounds in cancer
chemoprevention. Anticancer Agents Med Chem. 2014;14(2):233–
40.
7. Seki T, Hosono T, Suda S, Kimura K, Ariga T. Anticancer property
of allyl sulfides derived from garlic (Allium sativum L.). J Food
Drug Anal. 2012;20:309–12.
8. Tsubura A, Lai Y-C, Kuwata M, Uehara N, Yoshizawa K.
Anticancer effects of garlic and garlic-derived compounds for
breast cancer control. Anticancer Agents Med Chem. 2015;11(3):
249–53.
9. Miroddi M, Calapai F, Calapai G. Potential beneficial effects of
garlic in oncohematology. Mini-Rev Med Chem. 2015;11(6):461–
72.
10. Raghu R, Lu K-H, Sheen L-Y. Recent research progress on garlic
(dà suàn) as a potential anticarcinogenic agent against major diges-
tive cancers. J Tradit Complement Med. 2012;2(3):192–201.
11. Jin Z-Y, Wu M, Han R-Q, Zhang X-F, Wang X-S, Liu A-M, et al.
Raw garlic consumption as a protective factor for lung cancer, a
population-based case–control study in a Chinese population.
Cancer Prev Res. 2013;6(7):711–8.
12. Huang Y-S, Xie N, Su Q, SuJ, Huang C, Liao Q-J. Diallyl disulfide
inhibits the proliferation of HT-29 human colon cancer cells by
inducing differentially expressed genes. Mol Med Rep. 2011;4(3):
553–9.
13. Lanzotti V. The analysis of onion and garlic. J Chromatogr A.
2006;1112(1–2):3–22.
14. Yang J-S, Kok L-F, Lin Y-H, Kuo T-C, Yang J-L, Lin C-C, et al.
Diallyl disulfide inhibits WEHI-3 leukemia cells in vivo.
Anticancer Res. 2006;26(1A):219–25.
15. Seki T, Hosono T, Hosono-Fukao T, Inada K, Tanaka R, Ogihara J,
et al. Anticancer effects of diallyl trisulfide derived from garlic. Asia
Pac J Clin Nutr. 2008;17:249–52.
16. Warren JM, Parkinson DR, Pawliszyn J. Assessment of thiol com-
pounds from garlic by automated headspace derivatized in-needle-
NTD-GC-MS and derivatized in-fiber-SPME-GC-MS. J Agric
Food Chem. 2013;61:492–500.
17. Kim NY, Park MH, Jang EY, Lee J. Volatile distribution in garlic
(Allium sativum L.) by solid phase microextraction (SPME) with
different processing conditions. Food Sci Biotechnol. 2011;20(3):
775–82.
18. Keles D, Taskin H, Baktemur G, Kafkas E, Büyükalaca S.
Comparative study on volatile aroma compounds of two different
garlic types (Kastamonu and Chinese) using gas chromatography
mass spectrometry (HS-GC/MS) technique. Afr J Tradit
Complement Altern Med. 2014;11(3):217–20.
19. Clemente JG, Williams JD, Cross M, Chambers CC. Analysis of
garlic cultivars using head space solid phase microextraction/gas
chromatography/mass spectroscopy. Open Food Sci J. 2011;6:1–4.
20. Lee S-N, Kim N-S, Lee D-S. Comparative study of extraction tech-
niques for determination of garlic flavor components by gas chro-
matography–mass spectrometry. Anal Bioanal Chem. 2003;377(4):
749–56.
21. Calvo-Gómez O, Morales-López J, López MG. Solid-phase
microextraction −gas chromatog raphic −mass spectrometric anal-
ysis of garlic oil obtained by hydrodistillation. J Chromatogr A.
2004;1036:91.
22. RadulovićNS, MiltojevićAB, StojkovićMB, BlagojevićPD. New
volatile sulfur-containning compounds from wild garlic (Allium
ursinum L., Liliceae). Food Res Int. 2015;78:1–10.
M. Molina-Calle et al.
