Potent odorants of raw Arabica coffee. Their changes during roasting

Article abstract of DOI:10.1021/jf990609n

Aroma extract dilution analysis of raw Arabica coffee revealed 3- isobutyl-2-methoxypyrazine (I), 2-methoxy-3,5-dimethylpyrazine (II), ethyl 2- methylbutyrate (III), ethyl 3-methylbutyrate (IV), and 3-isopropyl-2- methoxypyrazine (V) as potent odorants. T

Full text of DOI:10.1021/jf990609n

868
J. Agric. Food Chem. 2000, 48, 868−872
P oten t Od or a n ts of Ra w Ar a bica Coffee. Th eir Ch a n ges d u r in g
Roa stin g
Michael Czerny and Werner Grosch*
Deutsche Forschungsanstalt fu¨r Lebensmittelchemie,Lichtenbergstrasse 4, D-85748 Garching, Germany
Aroma extract dilution analysis of raw Arabica coffee revealed 3-isobutyl-2-methoxypyrazine (I),
2-methoxy-3,5-dimethylpyrazine (II), ethyl 2-methylbutyrate (III), ethyl 3-methylbutyrate (IV), and
3-isopropyl-2-methoxypyrazine (V) as potent odorants. The highest odor activity value was found
for I followed by II, IV, and V. It was concluded that I was responsible for the characteristic, peasy
odor note of raw coffee. Twelve odorants occurring in raw coffee and (E)-â-damascenone were also
quantified after roasting. The concentration of I did not change, whereas methional, 3-hydroxy-
4,5-dimethyl-2(5H)-furanone, vanillin, (E)-â-damascenone, and 4-vinyl- and 4-ethylguaiacol increased
strongly during the roasting process.
Keyw or d s: Raw coffee; key odorants; effect of roasting; aroma extract dilution analysis; quantitative
analysis
INTRODUCTION
roasting. Therefore, the aim of the present investigation
was to close this gap. To this end, the potent odorants
were evaluated by aroma extract dilution analysis
(AEDA), and then their concentrations were compared
before and after roasting.
During roasting, the peasy, green smell of raw coffee
changes into the pleasant aroma characteristic of the
Arabica coffee to which 2-furfurylthiol, 4-vinylguaiacol,
several alkylpyrazines, furanones, acetaldehyde, pro-
panal, methylpropanal, and 2- and 3-methylbutanal
contribute as key odorants (Czerny et al., 1999).
Vitzthum et al. (1976) were the first to analyze the
volatiles of raw coffee by combining instrumental with
sensory methods. They identified four 3-alkyl-2-meth-
oxypyrazines and concluded, on the basis of the results
of high-resolution gas chromatography-olfactometry
(HRGCO), that only 3-isopropyl-2-methoxypyrazine and
the corresponding isobutyl derivative contributed to the
characteristic green coffee smell. In some charges of
roasted East African coffee a peasy off-flavor, which is
also known as potato taste (Bouyjou et al., 1999), was
caused by a 5-fold increase of 3-isopropyl-2-methoxy-
pyrazine (Becker et al., 1988).
Guyot et al. (1983) analyzed the odorants causing a
fruity by-note in raw coffee. In a fraction containing six
esters ethyl 2-methylbutyrate and isoamyl acetate were
identified as strong odorants that contributed to the
fruity note. Recently, Holscher and Steinhart (1995)
have reviewed comprehensively the volatiles identified
in raw coffee. Furthermore, they analyzed the volatiles
by HRGCO. Eleven odorants were intensively perceived
by using this procedure. Identification experiments
indicated that the methoxypyrazines mentioned above
also belonged to this fraction. In addition, the authors
concluded that some oxidation products of unsaturated
fatty acids such as (E)- and (Z)-2-nonenal, (E,Z)-2,6-
nonadienal, and (E,E)-2,4-decadienal due to their high
odor intensity during HRGCO might play a role in the
overall aroma.
MATERIALS AND METHODS
Coffee. The raw material of Coffea arabica var. tipica was
from Colombia. A portion of the beans was roasted with a
Neotec RFBS fluidized bed roaster. The roast degree of the
beans was characterized by the color value of 12.0 (Color Tester
LK 100, Dr. Lange, Berlin, Germany). The raw and roasted
materials were packed in 300 g portions and stored at -35
°C. For extraction, the beans were frozen in liquid nitrogen,
ground, and sieved (diameter of the pores ) 2 mm) with an
ultracentrifugation mill (type ZM1; Retsch, Haan, Germany).
Ch em ica ls. Compounds 1-7, 10-12, 14, 15, 18, and 20
(Table 1), cellulose, chlorine, 3-chloro-2,5-dimethylpyrazine,
[2H3]methanol (CD3OH), sodium, and sodium methoxide in
methanol (∼30%) were obtained from Aldrich (Steinheim,
Germany). Compounds 8 and 19 were purchased from Lan-
caster (Mu¨hlheim/Main, Germany), and (E)-â-damascenone
(22) was a gift of Haarmann & Reimer (Holzminden, Ger-
many). Reference substances 13 and ethyl (S)-2-methylbu-
tyrate were synthesized according to the methods of Ullrich
and Grosch (1988) and Fuhrmann (1998), respectively.
Syn th eses. 2-Methoxy-3,5-dimethylpyrazine (9). A refluxed
solution of 2,6-dimethylpyrazine (1.08 g, 10 mmol) in carbon
tetrachloride was treated with chlorine for 1 h. After the
reaction mixture had cooled to 20 °C, methylene chloride (50
mL) and, subsequently, an aqueous solution of sodium carbon-
ate (0.5 mol/L, 20 mL) were added. The organic phase was
separated, dried over anhydrous Na2SO4, and evaporated to
dryness. The residue was taken up in methanol (2 mL) and,
after addition of sodium methoxide in methanol (2 mL, ∼10
mmol), the mixture was refluxed for 2 h. Distilled water (20
mL) was added to the mixture, and the organic compounds
were extracted with diethyl ether (5 × 20 mL). The extract
was dried over anhydrous Na2SO4 and concentrated to 10 mL
by distilling off the solvent on a Vigreux column (40 × 1 cm).
Compound 9 was purified by thin-layer chromatography as
reported by Mayer et al. (1999) for [2H3]-2-ethenyl-3,5-di-
methylpyrazine. The zone appearing in the Rf range 0.60-
0.70 was extracted from the plate and then further purified
However, the studies undertaken until now have not
clearly identified the potent odorants occurring in raw
coffee or those that change their concentrations during
* Corresponding
Lrz.tu-muenchen.de).
author
(e-mail
werner.grosch@
10.1021/jf990609n CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/18/2000

Raw Coffee Aroma
J. Agric. Food Chem., Vol. 48, No. 3, 2000 869
Ta ble 1. P oten t Od or a n ts (F D F a ctor g 16) in Ra w Ar a bica Coffee
RI on
previously
no.
compound
fractiona
DB-5
FFAP
odor description
green
sweaty
sweaty
fruity
fruity
cooked potato-like
sweaty
mushroom-like
earthy
FD factor
identifiedb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
hexanalc
NBF
AF
AF
Id
800
821
844
848
851
909
911
979
1055
1097
1097
1109
1149
1161
1184
1248
1259
1285
1324
1412
1076
1616
1657
1044
1060
1440
1725
1296
1417
1419
1543
2185
1496
1525
1516
16
16
32
256
256
64
16
16
512
128
16
1-3
4
4, 5
1
1
5
4
5
butyric acidc
2-/3-methylbutyric acidc
ethyl 2-methylbutyratec
ethyl 3-methylbutyratec
methionalc
Id
IId
AF
NBF
pentanoic acidc
1-octen-3-onee
2-methoxy-3,5-dimethylpyrazinec,f
3-isopropyl-2-methoxypyrazinec,f
linaloolc
peasy
sweet
spicy
5-7
1-3
I
AF
3-hydroxy-4,5-dimethyl-2(5H)furanonec (sotolon)
(Z)-2-nonenalc,f
64
64
fatty, cardboard-like
fatty, cardboard-like
peasy
sweet, minty
sweet, minty
sweet
clove-like
vanilla-like
sweet
5
5
(E)-2-nonenalc
Id
128
4096
256
256
64
64
128
64
3-isobutyl-2-methoxypyrazinec
unknown
NBFd
1-3, 5-7
unknown
4-ethylguaiacolc
IVd
IVd
AFd
2015
2198
2559
2068
4-vinylguaiacolc
3, 6
3
vanillinc
unknown
a
Fraction in which the compound was detected: NBF, neutral-basic fraction; AF, acidic fraction; I, II, and IV, fractions obtained by
b
column chromatography. References: 1, Guyot et al. (1983); 2, Spadone and Liardon (1988); 3, Spadone et al. (1990); 4, Wo¨hrmann et
al. (1997); 5, Holscher and Steinhart (1995); 6, Vitzthum et al. (1976); 7, Becker et al. (1988). c The compound was identified by comparing
it with the reference substance on the basis of the following criteria: retention indices (RI) on the capillaries detailed in the table, MS-EI,
and odor quality perceived at the sniffing port. The fraction was analyzed by MD-HRGCMS. e 8 was identified only by comparing its RI
d
values and odor properties with the corresponding properties of an authentic sample. f The compound was identified in an extract obtained
by procedure B using MD-HRGCMS.
by preparative gas chromatography according to the method
of Guth and Grosch (1991). Compound 9, which eluted on the
stainless steel column (3 m × 2 mm, packed with SE-30 on
Chromosorb WHP, 80-100 mesh) in the range of 12.6-14.0
min, was trapped: MS-EI 138 (100%, M+), 42 (70%), 54 (67%),
109 (62%), 82 (43%), 120 (40%), 39 (39%), 107 (33%), 68 (31%),
52 (23%), 95 (22%), 40 (12%), 55 (12%), 66 (12%), 123 (10%);
MS-CI 139 (100%, M+ + 1, 138 (5%); 1H NMR (CDCl3, 360
MHz) δ 2.44 (3H), 2.47 (3H), 3.95 (3H), 7.78 (1H). The position
of the methoxy group at the pyrazine ring was confirmed by
comparing the 1H NMR data with those of 3-methoxy-2,5-
dimethylpyrazine.
µm film thickness, J &W Scientific, Folsom, CA). The samples
were applied on the column at 40 °C. After 2 min, the
temperature was raised at a rate of 6 °C/min to 230 °C and
held for 5 min. The flow rate of the carrier gas helium was 2
mL/min. The effluents eluting at the end of the capillary were
split (1:1 by volume) into an FID and a sniffing port (Blank et
al., 1989).
High -Resolu tion Ga s Ch r om a togr a p h y-Ma ss Sp ec-
tr om etr y (HRGCMS). MS analyses were performed with a
MAT-95 S spectrometer (Finnigan, Bremen, Germany) in
tandem with the capillaries described above. Mass spectra in
the electron impact mode (MS-EI) were generated at 70 eV.
3-Methoxy-2,5-dimethylpyrazine. 3-Chloro-2,5-dimethylpyra-
zine (1.42 g, 10 mmol), sodium methoxide in methanol (2 mL,
∼10 mmol), and methanol (2 mL) were refluxed for 2 h, and
the obtained reaction product was purified as reported for
compound 9. MS-EI 138 (100%, M+), 54 (66%), 109 (50%), 42
(47%), 39 (32%), 107 (31%), 120 (31%), 123 (27%), 95 (24%),
Mu ltid im en sion a l High -Resolu tion Ga s Ch r om a tog-
r a p h y-Ma ss Sp ectr om etr y (MD-HRGCMS). MD-HRGC-
MS was performed with the moving capillary stream switching
(MCSS) system (Fisons Instruments, Mainz-Kastell, Germany)
as described by Reiners and Grosch (1998) using the following
modification: The effluents of the MCSS outlet were split into
an FID and a sniffing port as described above. The effluent
ranges (Table 2), at which the mixture of the analyte and the
labeled internal standard appeared on the precolumn (DB-
FFAP), were estimated by nasal appraisal. These cuts of the
effluent were transferred onto the main column. Retention
indices (RIs) of the analytes on the main column (DB-5) were
determined on the basis of the retention times of the n-alkanes
C6-C16 after preseparation on the precolumn. Mass chromato-
grams were registrated in the electron impact mode at 70 eV
and in the chemical ionization mode at 115 eV using methanol
as reagent gas. Abundances were monitored for the ions given
in Table 2. The calibration factors reported in Table 2 were
determined by analyzing mixtures containing defined amounts
of the analyte and the internal standard (Sen et al., 1991).
1
68 (20%), 55 (12%); MS-CI 139 (100%, M+ + 1), 138 (5%); H
NMR (CDCl3, 360 MHz) δ 2.41 (3H), 2.43 (3H), 3.96 (3H), 7.81
(1H).
The internal standards used for stable isotope dilution
assays (IDAs) were labeled either with deuterium (d) or with
carbon-13 (c).
[2H3]-2-Methoxy-3,5-dimethylpyrazine (d -9) was synthesized
as unlabeled 9 but using sodium [2H3]methoxide, which was
prepared by the reaction of sodium (0.20 g, 8.7 mmol) and
[2H3]methanol (CD3OH, 1.60 g, 44 mmol) at room temperature.
MS-EI 141 (100%, M+), 122 (28%), 139 (27%), 82 (27%), 54
(25%), 111 (24%), 42 (23%), 109 (14%), 108 (13%), 112 (13%),
123 (12%), 96 (9%), 140 (9%); MS-CI 142 (100%, M+ + 1), 143
(6%), 141 (4%).
The following internal standards were synthesized according
to the literature cited: d -4, d -5 (Guth and Grosch, 1993), d -6
(Sen and Grosch, 1991), d -10 (Semmelroch, 1995), c-12 (Blank
et al., 1993), d -13, d -14 (Guth and Grosch, 1990), d -15
(Semmelroch and Grosch, 1996), d -18-20 (Semmelroch et al.,
1995), and d -22 (Sen et al., 1991).
High -Resolu tion Ga s Ch r om a togr a p h y-Olfa ctom etr y
(HRGCO). HRGC was performed with a type 5160 gas
chromatograph (Carlo Erba, Hofheim, Germany) using fused
silica capillaries DB-5 and DB-FFAP (30 m × 0.32 mm, 0.25
En a n tioselective An a lysis of Eth yl 2-Meth ylbu tyr a te.
Ethyl (R)- and (S)-2-methylbutyrate were separated by MD-
HRGCMS (effluent range ) 1030-1050) using a BGB-176
capillary column (30 m × 0.32 mm, 0.25 µm film thickness;
BGB Analytik, Rothenfluh, Switzerland) instead of the DB-5
main column. Mass chromatograms were recorded in the
chemical ionization mode as reported above. The distribution
of the enantiomers was calculated on the basis of the relative
abundances found for the ion m/z 131.

870 J. Agric. Food Chem., Vol. 48, No. 3, 2000
Czerny and Grosch
Ta ble 2. Cu ts, Selected Ion s, a n d Ca libr a tion F a ctor s for
MD-HRGCMS
were isolated according to procedure B and then analyzed by
MD-HRGCMS using the conditions shown in Table 2.
Compounds 18-20. Ground raw (25 g) and roasted coffees
(0.5 g) were extracted by isolation method C after addition of
known amounts of the internal standards d -18, d -19, and d -20.
The samples were analyzed by MD-HRGCMS (Table 2).
Od or Th r esh old Va lu es. The values were determined in
water (Semmelroch et al., 1995) and in air by HRGCO on the
capillary OV-1701 (J &W Scientific) according to the method
of Ullrich and Grosch (1988). Odor thresholds in cellulose were
determined by adding an ethereal solution of the odorant to
cellulose, which was purified according to the method of Czerny
et al. (1999). The mixture was shaken for 15 min, and aliquots
were diluted with increasing amounts of purified cellulose.
After shaking (15 min), each aliquot (1 g) was filled into a glass
beaker (diameter ) 40 mm, capacity ) 45 mL), and its odor
was compared in a triangle test with two samples of odorless
cellulose. The samples were presented in order of decreasing
concentration. Threshold values were calculated according to
a German Health Organization (1993) method.
RI range selected ion
selected ion calibra-
of cut on
of analyte
(m/z)
of int std
(m/z)
tion
analytea precolumnb
int stdc
factord
4
5
6
1030-1050
1050-1070
1430-1450
1400-1430
1400-1430
2170-2220
1510-1530
1480-1505
1510-1530
1990-2030
2160-2210
2510-2550
1820-1850
131
131
105
139
153
129
141
123
167
153
151
153
191
d -4
d -5
d -6
d -9
d -10
c-12
d -13
d -14
d -15
d -18
d -19
d -20
d -22
134
134
108
142
156
131
143
125
0.87
0.85
1.05
1.00
1.06
1.00
1.06
0.98
0.95
0.72
1.12
1.01
0.75
9
10
12
13
14
15
18
19
20
22
170
156-158e
154
156
195-197e
a
b
Numbering refers to Table 1. Retention index (RI) range of
the effluent from the precolumn which was transferred onto the
d
main column. c Abbreviations: c, carbon-13; d, deuterium. The
calibration factor refers to 1:1 (by weight) mixture for the labeled
and unlabeled compounds (Guth and Grosch, 1990; Semmelroch
et al., 1995). e The sum of the relative abundances of the ions was
calculated.
RESULTS AND DISCUSSION
AEDA revealed 21 potent odorants in raw coffee
showing FD factors of 16 and higher (Table 1). Hexanal
(1), butyric acid (2), the mixture of 2- and 3-methylbu-
tyric acid (3), pentanoic acid (7), 1-octen-3-one (8),
sotolon (12), 3-isobutyl-2-methoxypyrazine (15), and
vanillin (20) were identified in the neutral-basic and
acid fractions without further enrichment (Table 1). In
the cases of the methoxypyrazines 9 and 10 and in that
of (Z)-2-nonenal (13), the identification was successful
when an extract containing the volatile compounds was
consecutively separated on two capillary columns using
MD-HRGCMS. This technique provided also clear MS-
EI signals for the esters 4 and 5, methional (6), and
nonenal 14 as well as for the phenols 18 and 19 when
these odorants were at first enriched by column chro-
matography in the fractions listed in Table 1. For
identification of linalool (11), a purification by column
chromatography was sufficient. Only the sweet, minty
odorants 16 and 17 and the sweet-smelling odorant 21
were not identified. Enantioselective analysis of ethyl
2-methylbutyrate (4) confirmed the results of Wo¨hr-
mann et al. (1997) that 95% of 4 was the (S)-enantiomer.
AEDA indicated for the peasy smelling 3-isobutyl-2-
methoxypyrazine (15) the highest FD factor. With an
8-fold lower FD factor, 2-methoxy-3,5-dimethylpyrazine
(9) was identified as another very odor-active compound.
A comparison with the literature (references in Table
1) indicates that the earthy-smelling pyrazine 9 as well
as sotolon (12) and 4-ethylguaiacol (19) was identified
for the first time in raw coffee. AEDA of coffee powder
(data not shown) revealed compound 9 as an aroma
compound, but its FD factor was low in comparison to
those of other odorants of coffee powder.
2-Methoxy-3,5-dimethylpyrazine (9) has been detected
as a metabolite of aerobic Gram-negative bacteria
(Mottram et al., 1984), which were isolated from a
machine cutting-fluid emulsion. There are indications
that the methoxypyrazines occurring in raw coffee are
also produced by bacteria which enter the fruits through
holes caused by insects such as the variegated coffee
bug (Anestiopsis orbitalis) (Bouyjou et al., 1999).
The odor threshold of 3 µg/L in water published by
Calabretta (1973) for pyrazine 9 is much too high. We
found threshold values of 0.0004 µg/L (water), 0.006 µg/
kg (cellulose), and 0.000001 µg/L (air). However, the
latter very low odor threshold increases drastically to
56 ng/L (air) when the methoxy group at position 2 of
1H NMR Sp ectr oscop y. 1H NMR spectra were recorded
with an AM 360 MHz spectrometer (Bruker, Karlsruhe,
Germany). The substances were dissolved in CDCl3 containing
tetramethylsilane (TMS) as internal standard.
Isola tion of Vola tiles. Procedure A. The ground sample
(25 g) was suspended into CH2Cl2 (4 × 100 mL) and then
stirred for a total period of 4 h. The extract was concentrated
to 100 mL by distilling off the solvent on a Vigreux column
(40 × 1 cm) and distilled in vacuo (5 mPa, 50 °C) using the
apparatus described by Sen et al. (1991) and J ung et al. (1992).
The condensate was separated in the neutral-basic and the
acidic volatiles (Rychlik and Grosch, 1996). Each fraction was
concentrated to 0.1 mL by distilling off the solvent on a Vigreux
column (40 × 1 cm) and by microdistillation (Bemelmans,
1979). The neutral-basic volatiles were fractionated at 12 °C
on a water-cooled column (30 × 1 cm) packed with a slurry of
silica gel 60 in pentane. After application of the sample,
stepwise elution was performed with the following pentane/
diethyl ether mixtures (100 mL each): 9:1 (v/v, fraction I), 8:2
(v/v, fraction II), 7:3 (v/v, fraction III), and 1:1 (v/v, fraction
IV). Finally, the column was eluted with diethyl ether (100
mL, fraction V). After concentration to ∼100 µL by distilling
off the solvents on a Vigreux column (40 × 1 cm) and by
microdistillation (Bemelmans, 1979), each fraction was ana-
lyzed by HRGCMS.
Procedure B. The condensate obtained by distillation in
vacuo according to procedure A was concentrated to 0.1 mL
as reported above and then analyzed by MD-HRGCMS.
Procedure C. The sample of ground raw coffee (25 g) or
ground roasted coffee (0.5 g) was suspended in a mixture of
water, CH2Cl2, and methanol (4:5:10, v/v/v, 100 mL), and then
the suspension was stirred for 3 h. After filtration, the residue
was extracted again with the solvent mixture (100 mL) and
CH2Cl2 (100 mL) for 1 h each. The extracts were combined,
and the organic phase was separated, treated with water (100
mL), and dried over anhydrous Na2SO4. The extract was
concentrated to 100 mL by distilling off the solvent over a
Vigreux column. Isolation and concentration of the volatiles
were performed as reported for procedure A.
Ar om a Extr a ct Dilu tion An a lysis (AEDA). The fractions
containing the neutral-basic and the acidic volatiles (procedure
A) of raw coffee were stepwise diluted with CH2Cl2 (1:2, v/v).
AEDA was performed by HRGCO using the capillaries DB-5
und DB-FFAP (Ullrich and Grosch, 1987).
Qu a n tifica tion of Od or a n ts. Compounds 4-6, 9, 10, 12-
15, and 22. After addition of known amounts of the internal
standards d -4, d -5, d -6, d -9, d -10, c-12, d -13, d -14, d -15, and
d -22 to the extraction solvent, the analytes and their standards

Raw Coffee Aroma
J. Agric. Food Chem., Vol. 48, No. 3, 2000 871
Ta ble 3. Con cen tr a tion s, Od or Th r esh old s, a n d Od or Activity Va lu es of P oten t Od or a n ts in Ra w a n d Med iu m -Roa sted
Ar a bica Coffee
concnb
roasted coffee
3.9
14
213
1.1
2.4
1870
OAVd
no.a
odorant
raw coffee
odor thresholdc
raw coffee
4
5
6
ethyl 2-methylbutyrate
ethyl 3-methylbutyrate
methional
2-methoxy-3,5-dimethylpyrazine
3-isopropyl-2-methoxypyrazine
3-hydroxy-4,5-dimethyl-2(5H)furanone
(Z)-2-nonenal
2.4
22
22
0.5
2.3
0.7
<0.3
12
97
21
117
82
<0.3
0.5
0.6
9
0.006
0.1
2.1
nae
15
0.2
35
80
100
0.15
4.8
37
2.4
83
23
<1
9
10
12
13
14
15
18
19
20
22
<0.3
19
97
4060
39000
3290
255
(E)-2-nonenal
<1
490
<1
1.5
<1
<2
3-isobutyl-2-methoxypyrazine
4-ethylguaiacol
4-vinylguaiacol
vanillin
(E)-damascenone
a
b
Numbering refers to Table 1. Values in micrograms per kilogram raw and medium roasted Coffea arabica var. tipica. The data are
means of at least two assays. c Nasal odor threshold in µg/kg cellulose. The odor acitivity values (OAV) were calculated by dividing the
concentration by the odor threshold values in cellulose, which were obtained for 4, 5, 9, 10, 12, 14, and 15 in this study and for 6, 18-20,
and 22 from Mayer (unpublished). e na, not analyzed.
d
the pyrazine ring system and the methyl group at
position 3 exchange places, leading to 3-methoxy-2,5-
dimethylpyrazine.
is characteristic of coffee but, in addition, the odorants
which mask the peasy odor note caused by pyrazine 15.
Only when the concentrations of the methoxypyrazines
are too high in the raw beans might a peasy off-flavor
break through even after roasting as observed by Becker
et al. (1988).
The identified odorants showing FD factors of 64 and
higher as well as (E)-â-damascenone (22) were quanti-
fied before and after roasting of coffee. To eliminate the
simplifications associated with AEDA (Grosch, 1994),
odor activity values (OAVs; ratio of concentration to odor
threshold) were calculated for the odorants of raw coffee
using odor threshold values that were determined with
cellulose as base. The results in Table 3 reveal 3-isobu-
tyl-2-methoxypyrazine (15) showing an OAV of 490 as
the predominant odorant of raw coffee. The OAVs of the
next following odorants, 2-methoxy-3,5-dimethylpyra-
zine (9) and ethyl 3-methylbutyrate (5), were 83 and
37, respectively, much lower than the OAV value of
pyrazine 15. Roasting did not change the concentrations
of the pyrazines 10 and 15 (Table 3). This finding
confirmed the assumption of Vitzthum et al. (1976) that
methoxypyrazines are stable during roasting. Other
odorants such as methional (6), furanone 12, and the
phenols 18-20, which are potent odorants of roasted
coffee (Blank et al., 1992), increased strongly (Table 3).
The concentration of 3-isobutyl-2-methoxypyrazine
(15) in raw coffee agreed with the value found by
Holscher and Steinhart (1995) and was near the range
of 50-70 µg/kg reported by Spadone and Liardon (1988).
The suggestion that (E)-â-damascenone (22) belongs to
the intense odorants of raw coffee (Holscher and Stein-
hart, 1995) was not confirmed, as its concentration,
when present in raw coffee, was below the detection
limit of 0.3 µg/kg. However, after roasting, 255 µg/kg of
22 was found. Most likely, 22 was produced by a heat-
induced degradation of carotenoids (Holscher and Stein-
hart, 1995). Also, the high values reported for (E)-2-
nonenal (14) (280 µg/kg; Holscher and Steinhart, 1995)
and 4-vinylguaiacol (19) (2.3-7.5 mg/kg; Spadone and
Liardon, 1988) in raw coffee were not confirmed. Most
likely, the conventional quantitative methods used in
these studies were not suitable for an accurate deter-
mination of 14, 19, and 22 in raw coffee.
CONCLUSION
AEDA, quantification of potent odorants, and calcula-
tion of their OAVs demonstrated that the aroma of raw
coffee is primarily caused by 3-isobutyl-2-methoxypyra-
zine. 2-Methoxy-3,5-dimethylpyrazine was identified for
the first time as an additional aroma-active compound
of raw as well as roasted coffee. The high odor activity
of this compound was demonstrated by determination
of its odor threshold. Some potent odorants of roasted
coffee, for example, methional and phenols, are still
present in the raw beans. During roasting, the concen-
trations of these aroma compounds increase drastically
and mask, together with other odorants that are exclu-
sively formed by the roasting process [e.g., (E)-â-
damascenone, 2-furfurylthiol], the peasy odor of 3-isobu-
tyl-2-methoxypyrazine.
LITERATURE CITED
Becker, R.; Do¨hla, B.; Nitz, S.; Vitzthum, O. G. Identification
of the peasy off-flavour note on central african coffees. In
12th International Conference on Coffee, Montreux, 1987;
ASIC: Paris, France, 1988; pp 203-215.
Bemelmans, J . M. H. In Progress in Flavour Research; Land,
D. G., Nursten, H. E., Eds.; Applied Science: London, U.K.,
1979; pp 79-88.
Blank, I.; Fischer, K.-H.; Grosch, W. Intensive neutral odou-
rants of linden honey. Z. Lebensm. Unters. Forsch. 1989,
189, 426-433.
Blank, I.; Sen, A.; Grosch, W. Potent odorants of the roasted
powder and brew of Arabica coffee. Z. Lebensm. Unters.
Forsch. 1992, 43, 337-343.
Blank, I.; Schieberle, P.; Grosch, W. In Progress in Flavour
Precursor Studies; Schreier, P., Winterhalter, P., Eds.;
Allured Science: Carol Stream, IL, 1993; pp 103-109.
Bouyjou, B.; Decazy, B.; Fourny, G. Removing the “potato
taste” from Burundian Arabica. Plantations, recherche,
de´veloppement 1999, 6, 113-116.
Recently, a sensory study indicated that the major
methoxypyrazine of raw coffee, 3-isobutyl-2-methoxy-
pyrazine (15), was not perceived in a model mixture
which imitated the aroma of roasted coffee (Czerny et
al., 1999). Therefore, we suggest that roasting of the raw
material generates not only the pleasant aroma which
Calabretta, P. J . Pyrazinesspotent flavour and aroma agents.
Food Manuf. Ingred. Surv. 1973, March, 7-10.

872 J. Agric. Food Chem., Vol. 48, No. 3, 2000
Czerny and Grosch
Czerny, M.; Mayer, F.; Grosch, W. Sensory study on the
character impact odorants of roasted Arabica coffee. J . Agric.
Food Chem. 1999, 47, 695-699.
Fuhrmann, E. Influence of the cultivar on the aroma of apples.
Relationship between chemical structure and odor activity
of esters contributing to the aroma. Thesis, Technical
University of Munich, 1998 (in German).
German Health Organisation (Bundesgesundheitsamt). Of-
ficial Collection of Analytical Methods according to §35
LMBG; Beuth Verlag: Berlin, Germany, 1993; Vol. I/1,
Methods 00.90-7 and 00.90-9 (in German).
Grosch, W. Review: Determination of potent odorants in foods
by aroma extract dilution analysis (AEDA) and calculation
of odour activity values (OAVs). Flavour Fragrance J . 1994,
9, 147-158.
Guth, H.; Grosch, W. Deterioration of soya-bean oil: quanti-
fication of primary flavour compounds using stable isotope
dilution assay. Lebensm.-Wiss. -Technol. 1990, 23, 513-522.
Guth, H.; Grosch, W. Detection of furanoid fatty acids in
soyabean oilscause for the light-induced off-flavour. Fat.
Sci. Technol. 1991, 93, 249-255.
Guth, H.; Grosch, W. Quantitation of potent odorants of virgin
olive oil by stable-isotope dilution assays. J . Am. Oil Chem.
Soc. 1993, 70, 513-518.
Guyot, B.; Cros, E.; Vincent, J .-C. Characterization and
identification of the compounds of the volatile fraction of a
sound green Arabica coffee and a malodorous Arabica coffee.
In 10th International Conference on Coffee, Salvador, 1982;
ASIC: Paris, France, 1983; pp 253-269 (in French).
Holscher, W.; Steinhart, H. Aroma compounds in green coffee.
In Food Flavors: Generation, Analysis and Process Influ-
ence; Charalambous, G., Ed.; Elsevier Science: London,
U.K., 1995; pp 785-803.
J ung, H. P.; Sen, A.; Grosch, W. Evaluation of potent odorants
in parsley leaves by aroma extract dilution analysis. Leb-
ensm.-Wiss. -Technol. 1992, 25, 55-60.
Mayer, F.; Czerny, M.; Grosch, W. Influence of provenance and
roast degree on the composition of potent odorants in roasted
Arabica coffees. Eur. Food Res. Technol. 1999, 209, 242-
250.
Mottram, D. S.; Patterson, R. L. S.; Warrilow, E. 2,6-Dimethyl-
3-methoxypyrazine: a microbiologically-produced compound
with an obnoxious musty odour. Chem. Ind. 1994, pp 448-
449.
Reiners, J .; Grosch, W. Odorants of virgin olive oil with
different flavor profiles. J . Agric. Food Chem. 1998, 46,
2754-2763.
Rychlik, M.; Grosch, W. Identification and quantification of
potent odorants formed by toasting of wheat bread. Leb-
ensm.-Wiss. -Technol. 1996, 29, 515-525.
Semmelroch, P. Studies to objectify the aroma difference
between Arabica and Robusta coffee. Thesis, Technical
University Munich, 1995 (in German).
Semmelroch, P.; Grosch, W. Studies on character impact
compounds of coffee brews. J . Agric. Food Chem. 1996, 44,
537-543.
Semmelroch, P.; Laskawy, G.; Blank, I.; Grosch, W. Determi-
nation of potent odorants in roasted coffee by stable isotope
dilution assay. Flavour Fragrance J . 1995, 10, 1-7.
Sen, A.; Grosch, W. Synthesis of six deuterated sulfur contain-
ing odorants to be used as internal standards in quantifica-
tion assays. Z. Lebensm. Unters. Forsch. 1991, 192, 541-
547.
Sen, A.; Laskawy, G.; Schieberle, P.; Grosch, W. Quantitative
determination of -damascenone in foods using stable isotope
dilution assay. J . Agric. Food Chem. 1991, 39, 757-759.
Spadone, J . C.; Liardon, R. Identifcation of specific volatile
components in Rio coffee beans. In 12th International
Conference on Coffee, Montreaux, 1987; ASIC: Paris, France,
1988; pp 194-202.
Spadone, J . C.; Takeoka, G.; Liardon, R. Analytical investiga-
tion of Rio off-flavor in green coffee. J . Agric. Food Chem.
1990, 38, 226-233.
Ullrich, F.; Grosch, W. Identification of the most intense
volatile flavour compounds formed during autoxidation of
linoleic acid. Z. Lebensm. Unters. Forsch. 1987, 184, 277-
282.
Ullrich, F.; Grosch, W. Flavour deterioration of soja-bean oil:
identification of intense odour compounds formed during
flavour reversion. Fat. Sci. Technol. 1988, 90, 332-336.
Vitzthum, O. G.; Werkhoff, P.; Ablanque, E. Flu¨chtige In-
haltsstoffe des Rohkaffees. In 7th International Conference
on Coffee, Hamburg, 1975; ASIC: Paris, France, 1976; pp
115-123.
Wo¨hrmann, R.; Hojabr-Kalali, B.; Maier, H. G. Volatile minor
acids in coffee. I. Contents of green and roasted coffee. Dtsch.
Lebensm.-Rundsch. 1997, 93, 191-194.
Received for review J une 7, 1999. Revised manuscript received
November 30, 1999. Accepted December 15, 1999.
J F990609N