Odorants of Raw Coffee Beans
J. Agric. Food Chem., Vol. 55, No. 14, 2007 5769
ester caused a fruity, silage-like off-flavor when added to a
coffee beverage. Lopez et al. (7) found that properly fermented
Arabica coffee from Colombia with a water content of 10-
vinylguaiacol) (Lancaster, M u¨ hlheim/Main, Germany). (E)-â-Dama-
scenone was a gift from Symrise (Holzminden, Germany). 2-Methyl-
3
-(methyldithio)furan was synthesized according to the method given
in ref 13.
1
2%, which was stored for 9 months at 20 °C, kept a good
Syntheses. 2-Methoxy-5-Vinylphenol. A suspension of an instant-
ylide mixture of methyltriphenylphosphonium bromide with sodium
beverage quality. However, in samples with a higher humidity
molds proliferated in the last months of storage, leading to a
strong “mold-like” taste in the beverage (7).
Illy and Viany (8) detected an off-flavor reminiscent of rotten
fish in immature green beans, and Full et al. (9) later suggested
amide (1.5 g ) 3.6 mmol of methyl triphenylphosphoniumbromide)
(14) and dry diethyl ether (10 mL) was vigorously stirred for 15 min
in a round-bottom flask covered with a piece of cotton wool.
4-Methoxy-3-hydroxybenzaldehyde (0.2 g, 1.3 mmol) dissolved in dry
diethyl ether (2 mL) was added, and the mixture was stirred for 45
min at room temperature. After the addition of a saturated aqueous
4
-heptenal as a key odorant for this aroma defect. On the other
hand, Becker et al. (10) had reported that a pea-like off-flavor
in roasted East African coffee was caused by unusual high
concentrations of 3-isopropyl-2-methoxypyrazine. This off-
flavor, which is also called “potato taste”, is obviously caused
by insects (especially by the variegated coffee bug), because
methoxypyrazine-producing bacteria can penetrate into hurt
coffee cherries (11). Methoxypyrazines did not decrease during
coffee roasting (4), so it seems probable that in higher
concentrations these might affect the aroma of the coffee
beverage. In addition, also the oxidation of coffee lipids and
the generation of (E)-2-nonenal has been proposed as a source
of off-odors in coffee beverages (12), and it was shown that at
a level of 8 µg/L of this compound caused a woody off-note.
Because the literature survey indicates that the storage
conditions of raw beans seemed to be an important parameter
in off-odor development in coffee beverages, the present study
was aimed at (i) characterizing the key odorants in authentic
fresh raw coffee beans by application of the odor activity value
concept and (ii) monitoring changes in the concentrations of
selected odorants occurring during storage of the same batch
of coffee beans under defined conditions.
NH
aqueous phase was extracted with diethyl ether (2 × 20 mL). The
organic layers were combined, dried over anhydrous Na SO , and
4
Cl solution (50 mL), the organic phase was separated and the
2
4
evaporated to dryness at 40 °C. The crude material was taken up in
dichloromethane (1 mL) and purified by flash chromatography (glass
column, 40 cm × 2.5 cm) on silica gel 60 (particle size ) 0.035-
0
.070 mm) by stepwise elution (100 mL each) with solvent mixtures
(v/v by volume) of increasing polarity: A, pentane; B, pentane/
dichloromethane, 95:5; C, pentane/dichloromethane, 90:10; D, pentane/
dichloromethane, 85:15; E, pentane/dichloromethane, 80:20; F, pentane/
dichloromethane, 75:25; G, pentane/dichloromethane, 70:30; H, pentane/
dichloromethane, 60:40; and I, pentane/dichloromethane, 50:50. The
target compound was eluted in fractions G and H. After removal of
the solvent at 40 °C, 155 mg (yield ) 78%) of a white powder was
obtained (15): MS-CI, m/z (%) 151 (100), 137 (18), 152 (9); MS-EI,
m/z (%) 135 (100), 150 (92), 107 (25), 77 (18), 136 (7), 151 (7), 79
1
(
6), 51 (5), 78 (5); H NMR (360 MHz, d
4
-methanol, TMS), δ 3.84 (s,
3
H), 5.05 (dd, J ) 1.2 and 10.9 Hz, 1Hâ), 5.56 (dd, J ) 1.2 and 17.7
Hz, 1HR), 6.58 (dd, J ) 10.9 Hz and J ) 17.7, 1H), 6.83-6.84 (m,
2H), 6.91-6.92 (m, 1H).
2
Isotopically Labeled Compounds: [ H
3
]-2-Methoxy-4-Vinylphenol.
2
Starting from [ H
prepared as reported in ref 16, [ H
synthesized following the procedure described above for the unlabeled
3
]-4-hydroxy-3-methoxybenzaldehyde (100 mg),
2
3
]-2-methoxy-4-vinylphenol was
MATERIALS AND METHODS
2
-methoxy-5-vinylphenol. Yield ) 0.1 g (100%). The MS-EI and MS-
Coffee Samples. Raw Arabica coffee beans from Colombia were
provided by a German coffee trading company. The time between raw
coffee production and the storage experiments was about 3 weaks. To
adjust defined water contents, aliquots of the raw beans with a water
content of 11.75% (30 kg) were carefully dried in hot air to 6.2%. To
reach a water content of 13.5%, water (10 L) was added to another
aliquot (30 kg) of the raw coffee and the suspension was stirred at
CI data were in agreement with data reported in ref 16.
The following isotopically labeled compounds were prepared as
2
2
described previously: [ H
3
]-methyl 2-methylbutanoate (17), [ H
2
]-(E,E)-
]-(E)-2-
2
2
2,6-nonadienal (16), [ H4-6]-(E)-â-damascenone (18), [ H
]-2-methoxy-3-isopropylpyrazine (20), [ H ]-ethyl
3 3
2-methylbutanoate (21), [ H
2
2
2
nonenal (19), [ H
2
3
]-ethyl 3-methylbutanoate (21).
Dichloromethane, n-pentane, ethanol (Lichrosolv), dry diethyl ether,
cellulose, anhydrous sodium sulfate, sodium hydrogen carbonate,
sodium chloride p.A., hydrochloric acid p.A. (32%), and ammonium
chloride p.A. were purchased from Merck. Dichloromethane and
pentane were freshly distilled before use. Silica gel 60 (particle size )
0.035-0.070 mm) and the instant-ylide (methyltriphenylphosphonium
bromide with sodium amide) were from Fluka (Sigma-Aldrich Chemie).
Isolation and Fractionation of Volatiles. Raw coffee beans (30 g)
were frozen with liquid nitrogen and powdered in a ultracentrifugal
mill (type ZM1; Retsch, Haan, Germany; diameter of the pores )
2 mm). The powder was extracted by stirring with dichloromethane
(200 mL) at ambient temperature for 1 h. After filtration, the residue
was extracted twice with dichloromethane (100 mL each) for another
90 min. The extracts were combined and concentrated to 100 mL using
a Vigreux column (50 cm × 1 cm). After concentration to 100 mL,
the volatiles were isolated by SAFE distillation (22).
The distillate was separated into the acidic and the neutral-basic
volatiles as reported previously (23). The fractions containing either
the neutral and basic volatiles (NBF) or the acidic volatiles (AF) were
dried over anhydrous sodium sulfate and concentrated to 0.3 mL using
a Vigreux column (50 cm × 1 cm) followed by microdistillation (24).
For compound identification, the NBF fraction was further separated
by column chromatography on silica gel (25).
45 °C for 20 min. No measurable extraction of green coffee constituents
occurred during this short period.
To maintain a defined oxygen atmosphere, aliquots (2 kg) of the
three coffee samples were kept in airtight alumina bags at 2 or 20%
oxygen atmosphere, respectively. The total headspace volume was 200
mL. The sealed bags were stored at either 12 or 40 °C, respectively.
Thus, the following green coffee samples varying in one parameter
were obtained: series A with 13.5% water content included samples
stored at 20% oxygen and 12 °C, 2% oxygen and 40 °C, and 20%
oxygen and 40 °C. The respective B series with 11.75% water content
and the C series with 6.2% water content were stored under the same
conditions. Aliquots of the fresh raw coffee beans were stored at
-
30 °C as the reference. Storage conditions were chosen according to
advice by the coffee trade.
Chemicals: Reference Aroma Compounds. The following com-
pounds were obtained from commercial sources given in parentheses:
(R/S)-methyl 2-methylbutanoate, (R/S)-ethyl 2-methylbutanoate, ethyl
3-methylbutanoate, hexanal, 2-ethyl-3,5-dimethylpyrazine, 2-methoxy-
3-isopropylpyrazine, 3-(methylthio)propanal (methional), (E)-2-nonenal,
2-methoxy-3-isobutylpyrazine, (R/S)-2-methylbutanoic acid, 3-meth-
ylbutanoic acid, (E,E)-2,4-nonadienal, 4-hydroxy-2,5-dimethyl-3(2H)-
furanone (4-HDF), 3-hydroxy-4,5-dimethyl-2(5H)-furanone (3-HDF),
4
3
-methoxy-3-hydroxybenzaldehyde, 3-methoxyphenol, and 4-hydroxy-
-methoxybenzaldehyde (vanillin) (Aldrich, Sigma-Adrich Chemie,
High-Resolution Gas Chromatography-Olfactometry (HRGC-
O); Mass Spectrometry. For HRGC-O a gas chromatograph type Trace
GC 2000 series (Thermo Finnigan, Egelsbach, Germany) was used.
Helium at a pressure of 110 kPa served as the carrier gas. Samples
were applied by cold-on-column injection onto capillaries DB-FFAP
Taufkirchen, Germany); acetic acid (Merck, Darmstadt, Germany);
-phenylethanol (Fluka, Sigma-Aldrich Chemie, Taufkirchen, Ger-
many); methyl 3-methylbutanoate and 2-methoxy-4-vinylphenol (4-
2