JFS C: Food Chemistry and Toxicology

Analysis of the Headspace Volatiles

of Freshly Brewed Arabica Coffee
Using Solid-Phase Microextraction
M. AKIYAMA, K. MURAKAMI, M. IKEDA, K. IWATSUKI, A. WADA, K. TOKUNO, M. ONISHI, AND H. IWABUCHI
C: Food Chemistry & Toxicology

ABSTRACT: Headspace volatiles of freshly brewed drip coffee were investigated by gas chromatography/mass spec-
trometry (GC/MS) and gas chromatography/olfactometry (GC/O, CharmAnalysisTM ) analyses. For this purpose, a
solid-phase microextraction (SPME) sampling method for the headspace volatiles of freshly brewed drip coffee
was developed. SPME fiber coated with divinylbenzene (DVB)/carboxen/polydimethylsiloxane (PDMS) was selected
from 6 types, and sampling time was determined at 2 min. The headspace coffee volatiles stayed constant in propor-
tion for the first 2 min to keep the freshness of the brewed coffee aroma. Using this sampling method, the headspace
volatiles of freshly brewed drip coffee (Ethiopian arabica coffee, roast degree: L value; 23) were examined by GC/MS


and GC/O analyses. From the GC/O results, 1-(3,4-dihydro-2H-pyrrol-2-yl)-ethanone (nutty-roast odor) and 4-(4 -
hydroxyphenyl)-2-butanone (raspberry ketone, sweet-fruity odor) were newly detected as components in the aroma
of coffee.
Keywords: freshly brewed drip coffee, gas chromatography/mass spectrometry (GC/MS), gas chromatogra-
phy/olfactometry (GC/O), headspace volatiles, solid-phase microextraction (SPME)

Introduction a short period of time. For collecting aromatic volatiles, methods

B y using mainly solvent extraction and headspace sampling

methods, more than 800 odorous constituents have so far
been identified in the numerous previous analytical studies on
brewed and ground coffee (Semmelroch and Grosch 1995, 1996; Cz-
erny and others 1996, 1999; Nijssen and others 1996; Grosch 1998;
Mayer and others 2000; Sanz and others 2002). Also, studies to un-
derstand the changes of aroma freshness occurring during stor-
age of ground coffee powder and the aroma staling occurring while
keeping the coffee brew warm have been performed (Holscher and
Steinhart 1992; Hofmann and Schieberle 2002). In spite of intensive
studies, the key to the delightful aromatic character of coffee is not
fully understood. In the coffee making process, grinding of roasted
coffee beans and brewing are the important steps where very at-
tractive aroma is released. The aromas released during the grinding
of roasted coffee beans and from freshly brewed coffee are partic-
ularly attractive and pleasant. The perception of freshness and the
desirable rich aromatic odor are indispensable factors for the deli-
ciousness of coffee. Therefore, the study of the fresh and pleasant
overall aromas themselves could help food chemists and flavorists
to create new and more desirable coffee flavors for processed foods
and beverages.
The analysis of the freshly brewed coffee volatiles that linger in
the air and reach the human nose could be a direct way to under-
stand the yet unrecognized contributing factors that attract peo-
ple to pleasant coffee aroma. Because this delightful aroma does
not linger for long, volatiles to be analyzed must be collected in
MS 20070121 Submitted 2/16/2007, Accepted 5/23/2007. Au-
thors Akiyama, Murakami, Ikeda, and Iwatsuki are with Mori-
naga Milk Industry Co., Ltd., 5-1-83, Higashihara, Zama,
Kanagawa, 228-8583, Japan. Authors Wada, Tokuno, Onishi, and
Iwabuchi are with San-Ei Gen F.F.I., Inc., 1-1-11, Sanwa-cho, Toyon-
aka, Osaka, 561-8588, Japan. Direct inquiries to author Akiyama (E-mail:
m akiyam@morinagamilk.co.jp).
such as purge-and-trap (Miszczak and others 1995), solid-phase
microextraction (Arthur and Pawliszyn 1990), direct sampling by
gastight syringe (Holscher and Steinhart 1992) are known. Among
these, the headspace SPME method has been widely used in the
study of aroma as a rapid, simple, inexpensive, and reproducible
method (Yang and Peppard 1994; Bicchi and others 1997; Steenson
and others 2002; Jung and Ebeler 2003; Charles-Bernard and others
2005; López-Galilea and others 2006). This SPME method enables
rapid sampling of coffee aroma, but sampling conditions need to
be optimized to ensure reproducibility. In the course of our studies
on the pleasant coffee volatiles, findings on the coffee volatiles re-
leased during grinding roasted coffee beans have previously been
reported (Akiyama and others 2003a, 2003b, 2005). In this article,
another attractive fresh aroma released from freshly brewed drip
coffee was investigated using SPME.
The purpose of this study was to optimize the conditions for the
short-time SPME sampling, and to evaluate the pleasant and desir-
able aroma released from freshly brewed drip coffee.
Materials and Methods
Coffee sample
Arabica (Coffea arabica) coffee beans from Ethiopia (commod-
ity name Sidamo grade 2) were supplied by Unicafe Inc. (Tokyo,
Japan). These were roasted to medium roast (L value, 23) using
a Probat (Emmerich, Germany) G-12 roaster, packed in 1-kg por-
tions, and stored at −20 ◦ C until use. The degree of roasting is rep-
resented as an L value, which was determined by measuring ground
roasted coffee (particle size: < 500 µm) using a color meter ZE-2000
(Nippon Denshoku Industries Co., Ltd., Tokyo, Japan). All roasted
coffee beans were held at room temperature for 2 h before grinding
to allow their temperatures to equalize.

C388 JOURNAL OF FOOD SCIENCE—Vol. 72, Nr. 7, 2007 

C 2007 Institute of Food Technologists
doi: 10.1111/j.1750-3841.2007.00447.x
Further reproduction without permission is prohibited
Volatiles of freshly brewed coffee . . .
Apparatus used for brewing drip coffee
The electronic grinder (Model CG-4B; Melitta Japan, Tokyo) and
the electronic coffee maker (Model HCD-6GJ; Toshiba Corporation,
Tokyo, Japan) were purchased from a local store and used for grind-
ing roasted coffee and for brewing by the paper drip mode. The
brewing by coffee maker was chosen to standardize brewing con-
ditions. The coffee server was reconstructed for sampling by SPME
fiber as shown in Figure 1. The roasted coffee beans were ground by
Melitta CG-4B grinder suitable for the so-called paper drip method.
The ground coffee (particle sizes: 400 to 800 µm) was immediately
placed in a Toshiba HCD-6GJ coffee maker and then extracted by
paper drip mode using purified water. The glass pot containing
brewed coffee was open to the atmosphere and the temperature
of brewed coffee was kept at about 80 ◦ C that kept the headspace
temperature at about 70 ◦ C.
SPME device
SPME devices were purchased from Sigma-Aldrich Co. (St. Louis,
Mo., U.S.A.). The following 6 types of SPME fibers were used for
a comparison of SPME fiber types: (1) PDMS with 100-µm thick-
ness, (2) PDMS/DVB with 65-µm thickness, (3) carboxen/PDMS
with 75-µm thickness, (4) polyacrylate with 85-µm thickness,
(5) DVB/carboxen/PDMS with 50/30-µm thickness, and (6) car-
bowax/DVB with 65-µm thickness. Unless otherwise mentioned,
the type of DVB/carboxen/PDMS fiber was used for sampling
headspace volatile compounds.
GC/MS parameters
GC/MS analyses were performed in triplicate on a Hewlett-
Packard (HP) 5973 (Palo Alto, Calif., U.S.A.) with a fused silica cap-
illary column DB-WAX (60 m × 0.25 mm, 0.25-µm film thickness,
J&W Scientific, Folsom, Calif., U.S.A.). The flow of the helium carrier
gas was 1.6 mL/min. The temperature program was set at an initial
50 ◦ C for 2 min, followed by an increase of 3 ◦ C/min to 220 ◦ C, and
held at 220 ◦ C for 20 min. The injection port, equipped with a 0.75-
mm i.d. liner (Sigma-Aldrich Co.) was maintained at 250 ◦ C. The
inlet was operated in the splitless mode, and the injection purge on
the GC was off for an initial 1 min for the manual SPME injection.
GC/MS analyses were conducted under the same conditions where
quantitative comparisons of data were necessary.
GC/O parameters
GC/O was conducted in triplicate using CharmAnalysisTM on an Before headspace sampling, the SPME fiber was recondi-
HP6890 GC modified by DATU Inc. (Geneva, N.Y., U.S.A.) (Acree tioned in the GC injection port according to the supplier’s
Figure 1 --- Sampling apparatus
used to trap headspace volatile
compounds of freshly brewed
coffee
C: Food Chemistry & Toxicology
and others 1984). A fused silica capillary column DB-WAX (15 m ×
0.32 mm, 0.25-µm film thickness; J&W Scientific) was used. The
flow of the helium carrier gas was 3.2 mL/min. The temperature
program was set at an initial 40 ◦ C, followed by an increase at
6 ◦ C/min to 230 ◦ C and held at 230 ◦ C for 20 min. The injection and
detector ports were maintained at 250 ◦ C and the injection purge on
the GC was off for the initial 1 min. The retention time of each com-
pound was converted to Kovats indices using C 6 to C 28 n-alkanes
(Tokyo Kasei Kogyo Co., Ltd., Tokyo, Japan).
Identification of the volatile compounds
2-Methylbutanal, 3-methylbutanal, 2,3-pentanedione, pyridine,
2-methylpyrazine, 2,5-dimethylpyrazine, 2-furanecarboxaldehyde,
2-acetylfuran, 3-methylbutyric acid, phenol, 4-ethyl-2-
methoxyphenol, and 1H-indole were purchased commercially
from Tokyo Kasei Kogyo Co., Ltd.; 2-ethyl-6-methylpyrazine
and E-linalool oxide (furanoid) were from CTC Organ-
ics (Atlanta, Ga., U.S.A.); 1-(2-furanylmethyl)-1H-pyrrole,
linalool, and geraniol were obtained from Sigma-Aldrich
Co.; and 2-ethyl-3,5-dimethylpyrazine, furfuryl acetate, 5-
methyl-2-furancarboxaldehyde, 2-furanmethanol, 4-ethenyl-2-
methoxyphenol, 2-((methylthio)methyl)furan, and 4-ethylphenol
were obtained from Oxford Chemicals Ltd. (Hartlepool, U.K.). We
synthesized 2,3-dihydro-6-methylthieno[2,3c]furan and 1-(1H-
pyrrol-2-yl)-ethanone by the procedure described in the literature,
respectively (Buchi and others 1971; Kimpe and others 1993).
Volatile compounds were identified by comparing their mass
spectra and Kovats indices using C 6 to C 28 n-alkanes to those of
standard compounds and to those from the literature (Baltes and
Bochmann 1987). Also, tentative identifications of some potent
odorants found only by the GC/O analyses were made by compar-
ing their Kovats indices and aroma properties to those of standard
compounds and to those from the literature (Holscher and others
1990). When compounds were newly detected only by GC/O
analyses, GC/O analyses were conducted using 2 kinds of columns
with different polarities (DB-WAX and HP-5 (15 m × 0.32 mm,
0.25-µm film thickness; Agilent, Santa Clara, Calif., U.S.A.)), and
their Kovats indices and odor characters were compared with
those of standard compounds to verify that they were exactly
identified.
General SPME procedure
Vol. 72, Nr. 7, 2007—JOURNAL OF FOOD SCIENCE C389
 Volatiles of freshly brewed coffee . . .

recommendations (data sheet T7941231; Sigma-Aldrich Co.). After Headspace SPME sampling
sampling, the fiber was placed into the injection port of the GC/MS from freshly brewed coffee
or GC/O and thermally desorbed for 10 min at 250 ◦ C. Each SPME Coffee extract was brewed with an electric coffee maker. Freshly
sampling was conducted in triplicate. ground bean powder (24 g) was placed in filter paper (UCC
Ueshima Coffee Co., Ltd., Kobe, Japan) inside a funnel over a glass
Headspace SPME sampling using pot. From 420 mL of ion-exchange water in a tank, about 360 mL of
6 types of SPME fibers coffee was extracted in about 7 min. When coffee extraction was
The headspace volatile compounds were collected using the completed, sampling was started by exposing an SPME fiber to
following different 6 fibers, PDMS, PDMS/DVB, carboxen/PDMS, headspace in each glass pot; the collected volatiles were then an-
polyacrylate, DVB/carboxen/PDMS, and carbowax/DVB, under the alyzed by GC/MS and GC/O.
same static condition for the choice of fiber suitable for sampling
coffee volatiles. Brewed coffee (10 mL) was transferred into a Pyrex Liquid–liquid extraction method
glass bottle (volume about 21 mL; Sogo Laboratory Glass Works Co., Coffee brew was made in the same procedure as above from
C: Food Chemistry & Toxicology

Ltd., Kyoto, Japan), sealed with a Teflon-coated silicone cap (Sigma-
Aldrich Co.), and during equilibration at 50 ◦ C a fiber was exposed
to the headspace of coffee brew for 30 min by using the same sam-
pling conditions as described by Bicchi and others (2002). After
SPME sampling, the fiber was placed into the injection port of the
GC/MS, and thermally desorbed for 10 min at 250 ◦ C.
freshly ground coffee (24 g) and ion-exchange water (420 mL).
When coffee extraction was completed, the glass pot was quickly
placed in an ice–water bath and cooled to room temperature (un-
der 25 ◦ C). Cooled coffee (300 mL) was moved to a 1-L flask, 3-
heptanol (50 µg/mL in ethanol, 1 mL) was added as an internal
standard, and the solvent (diethyl ether, 150 mL) was added slowly.

Table 1 --- Comparison of the adsorption capabilities of typical volatile compounds of brewed coffee sampled by
different types of SPME fibers under the static condition
(A)
Peak area (× 103 )
PDMS Polyacrylate Carbowax/DVB Carboxen/PDMS PDMS/DVB DVB/Carboxen/PDMS
Total ion Alcohols 5112 9627 10047 13702 39699 36708
Aldehydes 3984 3728 5237 540494 24046 17305
Furans 30199 122579 124348 1187801 410400 653510
Ketones 902 2185 2366 282720 8598 11918
Phenols 7848 27914 33509 3933 45146 26727
Pyrazines 9601 5671 6448 187959 53780 73009
Pyridines 2953 1389 1291 169880 13690 13688
Pyrroles 8963 22552 28198 15073 69455 59971
Specific ion (m/z ) Furans 842 1625 2800 4999 8974 7910
Phenols 263 2294 2138 2463 4447 3527
Pyrroles 624 2472 3416 9007 8992 8529
(B)
Total ion
Alcohols Linalool Ketones 2,3-Butanedione
Geraniol 2,3-Pentanedione
Aldehydes 2- and 3-Methylbutanals Phenols 4-Ethenyl-2-methoxyphenol
Furans 2-Furanmethanol Pyrazines 2-Ethyl-3,5-dimethylpyrazine
E -Linalool oxide (furanoid) 2-Methylpyrazine
5-Methyl-2-furancarboxaldehyde 2,5-Dimethylpyrazine
2-Furancarboxaldehyde 2-Ethyl-6-methylpyrazine
Furfuryl acetate Pyridines Pyridine
2-Acetylfuran Pyrroles 1-(2-Furanylmethyl)-1H -pyrrole
1-(2-Furanylmethyl)-1H -pyrrole-2-carboxaldehyde

(C)
Specific ion
Furans 2-((Methylthio)methyl)furan (m/z 128)
Furfuryl formate (m/z 126)
2,3-Dihydro-6-methylthieno[2,3c ]furan (m/z 140)
Phenols 4-Ethenylphenol (m/z 120)
2-Methoxyphenol (m/z 109)
Phenol (m/z 94)
4-Ethyl-2-methoxyphenol (m/z 152)
Pyrroles 1H -Indole (m/z 117)
1-Methyl-1H -pyrrole-2-carboxaldehyde (m/z 109)
1-(1H -Pyrrol-2-yl)-ethanone (m/z 109)
1H -Pyrrole-2-carboxaldehyde (m/z 94)
Volatiles were collected by exposing a SPME fiber to the headspace of coffee brew for 30 min as described in the text. Sampled coffee volatiles were divided into 8
kinds of compound groups (B), and the total peak areas of those groups are shown in (A). Furthermore, the compounds, which were detected by these specific
ions, were divided into 3 kinds of compound groups (C), and the total peak areas of those groups are shown in (A).

C390 JOURNAL OF FOOD SCIENCE—Vol. 72, Nr. 7, 2007

Volatiles of freshly brewed coffee . . .
The mixture was stirred gently with a magnetic stirrer for 30 min
at room temperature (under 25 ◦ C), then separated in separatory
funnel, and dried over anhydrous sodium sulfate. The extract was
carefully concentrated in a rotary evaporator to a final volume of
about 0.5 mL.
Reduced-pressure steam distillation
extraction of brewed coffee
Freshly ground coffee (100 g) was placed in filter paper (UCC
Ueshima Coffee Co., Ltd.) on a plastic cone (coffee dripper 104D;
Kalita Co., Ltd., Tokyo, Japan) positioned on a 2-L glass flask. Boil-
ing water (about 1250 mL) was poured onto the coffee grounds and
the yielded coffee brew (about 1 L) was quickly cooled to room tem-
perature. One liter of the brew was poured into a 2-L 2-necked flask
in a water bath that was connected with a 500-mL receiver flask
through a water-jacketed condenser tube. A receiver flask was con-
nected through 2 sequential cold traps with a pump system. The
pressure was reduced to about 670 Pa and the temperature of the
water bath was maintained at 40 ◦ C to collect about 150 mL of dis-
tillate. The distillate was extracted with 200 mL ether, and 30 mL out
of the ether extract were concentrated to about 1 mL to be analyzed
by GC/MS.
GC/O evaluation of samples
During CharmAnalysisTM , a sniffer sniffs the air coming from the
olfactometer and holds down the mouse button while he or she de-
tects some odor in the air, and the computer records the percep-
tion and the time and duration of the sensation. Odor intensities
Figure 2 --- Comparison of total ion chromatograms of volatile compounds of brewed Ethiopian coffee obtained
using reduced-pressure steam distillation method (A) and the SPME sampling method (B, C). (B) SPME fiber:
DVB/carboxen/PDMS, sampling time: 2 min (C) SPME fiber: PDMS/DVB, sampling time: 2 min.
of volatile components obtained by GC/O dilution analyses were
recorded as charm values (Acree and others 1984), and the rela-
tive intensities of component odorants were represented in terms
of the odor spectrum values (OSVs) (Acree 1997). Each charm value
was rounded off to 2 significant figures in order to reflect the ac-
tual resolution of the dilution analysis. Acidic, buttery-oily, green-
blackcurrant, green-earthy, nutty-roast, phenolic, smoke-roast, soy
sauce, sweet-caramel, and sweet-fruity were the odor descriptions
used in all GC/O experiments to describe potent coffee odorants.
These descriptions were chosen from the results of a single prelim-
inary free choice GC/O analysis using a lexicon of words commonly
used for coffee evaluation (Akiyama and others 2003a).
C: Food Chemistry & Toxicology
Results and Discussion
Selection of SPME fiber types
The SPME technique is well known as a simple, rapid, sensitive,
and high reproducible sampling method. However, in order to ob-
tain significant data using this technique, it is important to set up
sampling conditions that are the most suitable for an experiment
system. Therefore, types of SPME fibers and exposure time of the
fiber to headspace were examined in the course of the development
of sampling method.
Some types of SPME fibers are in the market at present. The
headspace SPME concentration capability is significantly influ-
enced by the fiber coating composition, as illustrated by Bicchi and
others (2000) and Roberts and others (2000). Also, comparison of
3 fibers (PDMS, PDMS/DVB, and carboxen/PDMS) for ground
Vol. 72, Nr. 7, 2007—JOURNAL OF FOOD SCIENCE C391
 C: Food Chemistry & Toxicology

C392
Table 2 --- Changes of peak area of typical coffee volatile compounds by exposure time of SPME fiber (DVB/carboxen/PDMS) and of linear relationship between
peak area of typical coffee volatile compounds and exposure time of the SPME fiber
Total ion peak area
0.5 min 1 min 2 min 4 min 8 min 16 min
Retention Peak areaa Peak areaa Peak areaa Peak areaa Peak areaa Peak areaa
time Compound (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b
1 5.628 2- and 3-Methylbutanals 1083 --- 643 1 474 0.82 392 0.66 266 0.62 184 0.56
2 6.760 2,3-Butanedione 841 --- 677 1 559 0.92 470 0.83 322 0.85 255 0.75
3 8.908 2,3-Pentanedione 2452 --- 2403 1 2097 0.96 1785 0.98 1192 0.99 795 0.92
4 13.790 Pyridine 7569 --- 9979 1 11444 0.89 12485 0.79 6994 0.07 3813 0.55
5 17.234 2-Methylpyrazine 6724 --- 9790 1 13198 0.97 17144 0.95 14371 0.45 12364 0.10
6 19.781 2,5-Dimethylpyrazine 3931 --- 5754 1 7905 0.98 11906 0.99 12176 0.78 10491 0.39
7 22.617 2-Ethyl-6-methylpyrazine 3028 --- 4289 1 5875 0.98 8275 0.98 8212 0.73 6675 0.27
8 25.154 E -Linalool oxide (furanoid) 1039 --- 1541 1 2159 0.98 3008 0.97 2746 0.62 1881 0.06
9 25.899 2-Furancarboxaldehyde 35203 --- 51346 1 72451 0.99 101015 0.97 108042 0.80 102232 0.52
10 26.061 2-Ethyl-3,5-dimethylpyrazine 440 --- 687 1 998 0.98 1731 1.00 1897 0.84 2113 0.72
Volatiles of freshly brewed coffee . . .

11 27.767 2-Acetylfuran 5224 --- 7677 1 10768 0.98 15529 0.98 16298 0.79 13903 0.39
12 29.073 Furfuryl acetate 10834 --- 14017 1 19153 1.00 26425 0.99 23730 0.60 16104 0.03
13 29.595 Linalool 813 --- 1058 1 1577 1.00 2418 1.00 2617 0.83 2539 0.58

JOURNAL OF FOOD SCIENCE—Vol. 72, Nr. 7, 2007

14 30.724 5-Methyl-2-furancarboxaldehyde 24183 --- 36454 1 52498 0.99 82988 0.99 104399 0.93 120594 0.83
15 34.374 2-Furanmethanol 32001 --- 46721 1 60338 0.96 79151 0.95 79909 0.72 84754 0.60
16 41.016 1-(2-Furanylmethyl)-1H -pyrrole 1960 --- 2999 1 4730 1.00 8099 1.00 10054 0.92 9088 0.60
17 41.753 Geraniol 500 --- 965 1 1609 0.99 3144 1.00 4096 0.93 4798 0.82
18 53.950 4-Ethenyl-2-methoxyphenol 3072 --- 6040 1 11372 1.00 25821 1.00 46585 1.00 75570 0.98
19 55.828 1-(2-Furanylmethyl)-1H -pyrrole-2- 497 --- 1015 1 1853 1.00 4092 1.00 7130 0.99 11508 0.98
carboxaldehydec
Specific ion (m/z) peak area

0.5 min 1 min 2 min 4 min 8 min 16 min

Retention Compound Peak areaa Peak areaa Peak areaa Peak areaa Peak areaa Peak areaa
time (selected ion, m/z ) (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b (× 103 ) R 2b
20 26.971 2-((Methylthio)methyl)furan (128) 43 --- 49 1 61 1.00 89 1.00 67 0.35 44 0.01
21 27.281 Furfuryl formate (126) 62 --- 65 1 75 0.98 84 0.96 31 0.38 11 0.74
22 34.987 3-Methylbutyric acid (60) 2265 --- 3413 1 4007 0.87 4174 0.66 3177 0.04 2738 0.06
23 37.519 2,3-Dihydro-6- 163 --- 227 1 344 1.00 576 1.00 501 0.63 404 0.21
methylthieno[2,3c ]furan
(140)
24 42.239 2-Methoxyphenol (109) 226 --- 403 1 598 0.97 977 0.99 1200 0.90 1200 0.68
25 46.389 1-(1H -Pyrrol-2-yl)-ethanone (109) 139 --- 243 1 415 1.00 715 1.00 931 0.93 1289 0.91
26 47.489 Phenol (94) 183 --- 304 1 429 0.97 675 0.99 848 0.92 842 0.69
27 48.296 1H -Pyrrole-2-carboxaldehydec 151 --- 270 1 419 0.99 734 1.00 1060 0.97 1474 0.94
(94)
28 48.398 4-Ethyl-2-methoxyphenol (152) 33 --- 68 1 101 0.96 214 0.99 301 0.96 407 0.92
29 60.057 4-Ethenylphenol (120) 30 --- 55 1 107 1.00 216 1.00 392 1.00 678 0.99
30 62.088 1H -Indole (117) 47 --- 87 1 160 1.00 324 1.00 585 1.00 850 0.96
a
Peak areas are an average value of 3 measurements. Fractions have been regarded as within the range of error in the integration process and rounded off.
b
Coefficient of determination.
c
Tentative identifications were achieved by comparing mass spectra and retention indices with literature data (Baltes and Bochmann 1987).
Volatiles of freshly brewed coffee . . .

coffee headspace volatiles and for brewed coffee headspace
volatiles was reported (Roberts and others 2000; Akiyama and
others 2003a).
Selection of the most appropriate SPME fiber is very impor-
tant for sampling volatiles in the same proportion as the targeted
headspace volatiles. In this study, 6 types of fibers were tested:
PDMS, carbowax/DVB, PDMS/DVB, polyacrylate, carboxen/PDMS,
and DVB/carboxen/PDMS. A comparison of the adsorption capa-
bilities of these 6 fibers is displayed in Table 1, which clearly shows
the different adsorption affinities of volatile compounds. Sampled
coffee volatiles were divided into 8 kinds of compound groups
(Table 1[B]), and the total peak areas of those groups are shown in
Table 1(A). Furthermore, the compounds which were detected by
most components and their ability of collecting aldehydes was
particularly low. Carboxen/PDMS was very sensitive for most
components, but not sensitive to phenols that are important coffee
volatiles. In total, PDMS/DVB, which was adapted for sampling
coffee volatiles released during grinding (Akiyama and others
2003a, 2003b, 2005), and DVB/carboxen/PDMS showed good
sensitivity for coffee volatiles. To choose the most appropriate
fiber type, GC/MS chromatograms sampled by PDMS/DVB and
DVB/carboxen/PDMS fibers were compared with a chromatogram
obtained by conventional reduced-pressure steam distillation,
whose profile was thought to be close to “real headspace.”
Two important key compounds—2,3-pentanedione and 2-
furancarboxaldehyde—were less potently acquired by PDMS/DVB

C: Food Chemistry & Toxicology

these specific ions were divided into 3 kinds of compound groups
(Table 1[C]), and the total peak areas of those groups are shown in
Table 1(A).
PDMS, which is a traditional fiber with good stability, poly-
acrylate, and carbowax/DVB did not show high sensitivity for
compared with the total ion chromatogram of the extract of
steam distillation (Figure 2). The chromatogram obtained by
DVB/carboxen/PDMS showed high similarity with the extract of
steam distillation, and DVB/carboxen/PDMS was finally chosen
for the analysis of coffee volatiles in this study.

Table 3 --- Peak area against sampling timing for typical coffee volatile compounds used for optimization of SPME
conditions
Total ion peak area
Peak areaa (× 103 )
Retention 0 to 2 2 to 4 4 to 6 6 to 8 8 to 10 10 to 12 12 to 14 14 to 16
time Compound min min min min min min min min
1 5.628 2- and 3-Methylbutanals 464 480 472 409 326 342 320 325
2 6.760 2,3-Butanedione 514 493 473 472 417 388 391 401
3 8.908 2,3-Pentanedione 1954 1905 1831 1642 1655 1686 1657 1631
4 13.790 Pyridine 11244 9758 9072 8612 8635 8519 8351 8279
5 17.234 2-Methylpyrazine 12616 12411 11534 11185 10757 10606 10085 10489
6 19.781 2,5-Dimethylpyrazine 7878 7881 7331 7021 6672 6542 6567 6395
7 22.617 2-Ethyl-6-methylpyrazine 5795 5486 5176 4807 4586 4615 4475 4389
8 25.154 E -Linalool oxide (furanoid) 2027 1902 1792 1709 1614 1602 1533 1475
9 25.899 2-Furancarboxaldehyde 69391 68561 65620 64914 64949 63979 62419 60720
10 26.061 2-Ethyl-3,5-dimethylpyrazine 1023 1007 868 839 859 808 827 819
11 27.767 2-Acetylfuran 9769 9622 9632 9513 9335 8826 8295 8330
12 29.073 Furfuryl acetate 18965 17121 15492 13583 12861 11729 10676 10317
13 29.595 Linalool 1547 1323 1230 1088 1015 1047 932 868
14 30.724 5-Methyl-2-furancarboxaldehyde 50922 50309 47958 47925 47573 46890 45770 44634
15 34.374 2-Furanmethanol 59724 59249 56974 56931 57613 57011 57519 56477
16 41.016 1-(2-Furanylmethyl)-1H -pyrrole 4716 4312 3881 3800 3337 3446 3311 3098
17 41.753 Geraniol 1580 1401 1214 1101 1045 1022 976 944
18 53.950 4-Ethenyl-2-methoxyphenol 11249 10487 10093 10073 10054 10068 9985 9953
19 55.828 1-(2-Furanylmethyl)-1H -pyrrole-2- 1882 1856 1732 1639 1618 1617 1562 1478
carboxaldehydeb
Total 273261 265563 252375 247264 244923 240743 235652 231024
Specific ion (m/z) peak area
Peak areaa (× 103 )
Retention Compound 0 to 2 2 to 4 4 to 6 6 to 8 8 to 10 10 to 12 12 to 14 14 to 16
time (selected ion, m/z ) min min min min min min min min
20 26.971 2-((Methylthio)methyl)furan (128) 59 56 52 46 41 33 31 30
21 27.281 Furfuryl formate (126) 56 19 11 5 0 0 0 0
22 34.987 3-Methylbutyric acid (60) 4491 3757 3432 3283 3149 3130 3088 2992
23 37.519 2,3-Dihydro-6- 347 296 260 239 210 176 167 156
methylthieno[2,3c ]furan
(140)
24 42.239 2-Methoxyphenol (109) 593 568 531 526 516 491 478 452
25 46.389 1-(1H -Pyrrol-2-yl)-ethanone (109) 425 402 372 358 360 354 347 323
26 47.489 Phenol (94) 427 408 383 369 363 354 346 328
27 48.296 1H -Pyrrole-2-carboxaldehydeb 416 400 382 377 347 338 341 330
(94)
28 48.398 4-Ethyl-2-methoxyphenol (152) 109 100 93 91 86 88 84 81
29 60.057 4-Ethenylphenol (120) 107 102 95 96 95 96 96 92
30 62.088 1H -Indole (117) 169 161 146 135 133 133 134 130
Total 7200 6270 5756 5525 5299 5193 5111 4914
a
Peakareas are an average value of 3 measurements. Fractions have been regarded as within the range of error in the integration process and rounded off.
b
Tentative identifications were achieved by comparing by comparing mass spectra and retention indices with literature data (Baltes and others 1987).

Vol. 72, Nr. 7, 2007—JOURNAL OF FOOD SCIENCE C393

 Volatiles of freshly brewed coffee . . .

SPME exposure time
As the purpose of this study was to capture headspace volatiles of
just-brewed coffee, sampling by SPME fiber was started when cof-
fee extraction by the drip method was completed. To make analysis
more effective, it is favorable to collect volatiles in larger quantity,
but the longer sampling time could lead to the change of fresh cof-
fee aroma, or the constitutional difference from “real headspace”
caused by competition phenomena on the SPME fiber. To deter-
mine appropriate exposure time, the SPME fiber was exposed to
headspace at 0.5, 1, 2, 4, 8, and 16 min, respectively. The sampled
amounts of 30 important coffee volatiles are listed in Table 2. In
the case of 4-min sampling, 2- and 3-methylbutanals, pyridine, and
C: Food Chemistry & Toxicology
3-methylbutyric acid showed smaller increases in peak abundance
and lost linearity (coefficient of determination, R2 < 0.80), which
led to the change of component proportion among 30 odorants.
From these results, 2 min was chosen for sampling time.
Timing for sampling fresh coffee aroma
As shown in Table 2, the R2 values of some volatiles such as
pyrazines (2-methylpyrazine, 2,5-dimethylpyrazine, and 2-ethyl-
6-methylpyrazine), trans-linalool oxide (furanoid), 2-acetylfurane,
furfuryl acetate, 2-((methylthio)methyl)furan, 3-methylbutyric
acid, and 2,3-dihydro-6-methylthieno[2,3c]furan decreased in the
case when a fiber was exposed for 16 min. These decreases of the

Table 4 --- Reproducibility (3 times) of peak area by each length of SPME fiber and linearity of peak area with different SPME exposure
length
Total ion peak area
2.5 mm 5 mm 10 mm 20 mm
Retention Peak areaa Peak areaa Peak areaa Peak areaa
time Compound (× 103 ) RSDb % (× 103 ) RSDb % (× 103 ) RSDb % (× 103 ) RSDb % R 2c
1 5.628 2- and 3-Methylbutanals 401 3.10 854 5.65 1753 4.18 3666 4.14 0.998
2 6.760 2,3-Butanedione 308 4.54 775 1.47 1540 3.55 2927 4.76 0.998
3 8.908 2,3-Pentanedione 1124 5.95 2578 3.97 5472 2.44 12054 3.00 0.993
4 13.790 Pyridine 10057 2.73 19673 1.69 37548 4.19 70328 4.31 0.996
5 17.234 2-Methylpyrazine 7890 3.68 16179 3.75 31843 5.09 58985 2.45 0.996
6 19.781 2,5-Dimethylpyrazine 5262 3.40 9576 3.56 19341 3.59 35373 2.66 0.994
7 22.617 2-Ethyl-6-methylpyrazine 4883 3.27 8397 4.44 17108 5.88 33366 4.60 0.999
8 25.154 E -Linalool oxide (furanoid) 1832 4.74 3523 2.63 7139 4.66 13130 3.23 0.996
9 25.899 2-Furancarboxaldehyde 66214 2.15 121662 3.45 220645 2.74 353842 4.31 0.944
10 26.061 2-Ethyl-3,5- 854 2.49 1665 4.82 3320 4.71 6269 5.32 0.998
dimethylpyrazine
11 27.767 2-Acetylfuran 9138 5.13 16448 0.73 31185 4.52 58659 2.86 0.995
12 29.073 Furfuryl acetate 18830 4.13 35522 0.66 69959 4.88 108953 1.52 0.947
13 29.595 Linalool 1258 3.37 2449 1.30 4812 3.79 9014 4.54 0.997
14 30.724 5-Methyl-2- 43407 2.79 77543 2.52 142022 2.75 230325 4.57 0.950
furancarboxaldehyde
15 34.374 2-Furanmethanol 47127 5.45 86974 2.22 155346 5.08 249039 2.96 0.940
16 41.016 1-(2-Furanylmethyl)-1H - 4016 5.65 7945 1.76 15402 4.80 26365 2.21 0.981
pyrrole
17 41.753 Geraniol 1073 4.02 2027 0.71 4527 3.37 8268 3.83 0.996
18 53.950 4-Ethenyl-2- 12078 2.46 23205 2.62 44662 3.01 67678 4.66 0.929
methoxyphenol
19 55.828 1-(2-Furanylmethyl)-1H - 1975 4.30 3541 0.95 8246 1.68 13689 4.17 0.981
pyrrole-2-
carboxaldehyded
Specific ion (m/z) peak area
2.5 mm 5 mm 10 mm 20 mm
Retention Compound Peak areaa Peak areaa Peak areaa Peak areaa
time (selected ion, m/z ) (× 103 ) RSDb % (× 103 ) RSDb % (× 103 ) RSDb % (× 103 ) RSDb % R 2c
20 26.971 2-((Methylthio)methyl)furan 55 4.79 115 2.45 233 3.43 440 3.64 0.998
(128)
21 27.281 Furfuryl formate (126) 52 4.20 94 3.49 161 4.33 277 0.70 0.962
22 34.987 3-Methylbutyric acid (60) 3143 1.39 5064 1.99 11922 4.98 28274 3.28 0.982
23 37.519 2,3-Dihydro-6- 271 2.73 522 2.03 1094 4.67 1962 4.63 0.994
methylthieno[2,3c ]furan
(140)
24 42.239 2-Methoxyphenol (109) 515 2.98 791 4.50 1819 3.91 3592 4.50 0.997
25 46.389 1-(1H -Pyrrol-2-yl)-ethanone 310 0.88 465 4.71 1125 4.43 2095 3.37 0.994
(109)
26 47.489 Phenol (94) 401 1.87 606 3.53 1322 5.68 2716 3.94 0.997
27 48.296 1H -Pyrrole-2-carboxaldehyded 311 3.40 519 4.75 1179 4.25 2237 4.82 0.997
(94)
28 48.398 4-Ethyl-2-methoxyphenol 93 1.36 164 4.67 377 4.76 768 3.80 0.997
(152)
29 60.057 4-Ethenylphenol (120) 94 4.31 136 3.80 296 3.50 507 0.35 0.978
30 62.088 1H -Indole (117) 134 3.80 220 1.90 440 2.70 719 2.00 0.962
a
Peak areas are an average value of 3 measurements. Fractions have been regarded as within the range of error in the integration process and rounded off.
b
Relative standard deviation
c
Coefficient of determination
d
Tentative identifications were achieved by comparing mass spectra and retention indices with literature data (Baltes and Bochmann 1987).

C394 JOURNAL OF FOOD SCIENCE—Vol. 72, Nr. 7, 2007

Volatiles of freshly brewed coffee . . .

R2 values indicate the changes of increase rate of volatiles. This was
thought to be caused by the loss of fresh character in coffee aroma.
To verify the influence of aroma change, sampling time was fixed
as 2 min and sampling was started at different timings (0 to 2, 2 to
4, 4 to 6, 6 to 8, 8 to 10, 12 to 14, and 14 to 16 min) after finishing
brewing as shown in Table 3. Total abundance of 30 odorants be-
came smaller as sampling timing was started later. In addition, the
aroma profile was found to change in the proportion of odorants as
time passed from just finishing the extraction. Because the purpose
of our study was to evaluate fresh coffee aroma, 0 to 2 min after
finishing brewing was adopted as the optimum sampling timing.
Reproducibility of headspace SPME sampling
GC/O (CharmAnalysisTM ) (Deibler and others 1999). We had already
adopted this method for GC/O analysis and verified the reliance in
a previous study (Akiyama and others 2003a).
By cutting 20-mm SPME fiber, 2.5-mm, 5-mm, and 10-mm fibers
were prepared, respectively, and GC/O dilution analysis was con-
ducted with these 3 kinds of fibers together with the original length.
Before applying these fibers to GC/O analysis, coffee volatiles were
sampled 3 times for each fiber to be analyzed by GC/MS (Table 4).
GC peak areas were reproducible at less than 6% relative standard
deviation (RSD) for each fiber length, and main coffee odorants
produced linear plots of GC peak area compared with exposed fiber
length (R2 > 0.90). This novel SPME method was proved to be rapid,
simple, reproducible, and useful in the GC/MS and GC/O analyses

C: Food Chemistry & Toxicology

and linearity with different SPME length of fresh coffee headspace volatiles.
The GC/O dilution analysis by varying fiber length was per-
formed in this study. The dilution analysis using various fiber GC/O analysis
lengths was developed by Deibler and others (1999) and success- To investigate potent odor compounds, intensities, and pro-
fully applied to the preparation of the samples of brewed coffee for files of freshly brewed drip coffee, the GC/O dilution analysis of

Table 5 --- Potent odorants found in the headspace volatiles of freshly brewed coffee using the SPME sampling method
Description Component Retention indices Charm valuea (OSVb )
Acidic Acetic acid 1431 18 (22)
3-Methylbutyric acid 1677 380 (100)
Buttery-oily 2- and 3-Methylbutanals 916 100 (51)
2,3-Butanedione 976 220 (76)
2,3-Pentanedione 1055 40 (32)
(Z )-2-Nonenalc 1491 62 (40)
(E )-2-Nonenal 1517 150 (63)
Unknown 1530 66 (42)
2,6-(E ,Z )-Nonadienal 1568 8 (15)
Unknown 1787 13 (18)
Unknown 1934 89 (48)
Unknown 1982 21 (24)
Unknown 1987 130 (58)
Green-blackcurrant 3-Mercapto-3-methylbutyl formate 1501 130 (58)
Green-earthy 2-Methoxy-3-(1-methylethyl)pyrazine 1422 48 (36)
2-Methoxy-3-(2-methylpropyl)pyrazine 1513 97 (50)
Nutty-roast 1-(3,4-Dihydro-2H -pyrrol-2-yl)-ethanone 1321 20 (23)
Unknown 1407 4 (10)
2-Ethyl-3,5-dimethylpyrazine 1449 200 (73)
2,3-Diethyl-5-methylpyrazine 1482 180 (69)
Unknown 1574 110 (54)
6,7-Dihydro-5-methyl-5H -cyclopentapyrazine 1599 63 (41)
6,7-Dihydro-5H -cyclopentapyrazinec 1647 20 (23)
5,6,7,8-Tetrahydroquinoxaline 1728 55 (38)
Phenolic 2-Methoxyphenol 1840 200 (73)
4-Ethyl-2-methoxyphenol 2013 140 (61)
4-Ethenyl-2-methoxyphenol 2173 52 (37)
3-Methyl-1H -indole 2468 140 (61)
Smoke-roast 3-Methyl-2-butene-1-thiol 1100 200 (73)
2-Furanemethanethiol 1417 180 (69)
2-((Methylthio)methyl)furan 1471 140 (61)
Unknown 1539 120 (56)
Unknown 1718 85 (47)
Soy sauce 3-(Methylthio)propanal 1436 180 (69)
Sweet-caramel 2-Hydroxy-3-methyl-2-cyclopenten-1-one 1820 20 (23)
4-Hydroxy-2,5-dimethyl-3(2H )-furanone 2026 250 (81)
2-Ethyl-4-hydroxy-5-methyl-3(2H )-furanone 2056 30 (28)
3-Hydroxy-4,5-dimethyl-2(5H )-furanone 2183 270 (84)
Sweet-fruity Linalool 1545 140 (61)
Benzeneacetaldehyde 1616 100 (51)
(E )-Beta-damascenone 1796 200 (73)
Vanillin 2545 33 (30)

4-(4 -Hydroxyphenyl)-2-butanone 2977 110 (54)
Other 1-Octen-3-one 1289 8 (15)
Dimethyl trisulfide 1354 16 (21)
Total charm value 4839
a
Each charm value is rounded off to 2 significant figures to reflect the actual resolution of the dilution analysis, and represented as average value of 3
measurements.
b
Odor spectrum value (OSV) is the normalized charm value modified with an approximate Stevens’ law exponent (n = 0.5).
c
Tentative identifications were achieved by comparing mass spetra and retention indices with literature data (Holscher and others 1990).

Vol. 72, Nr. 7, 2007—JOURNAL OF FOOD SCIENCE C395

 Volatiles of freshly brewed coffee . . .
headspace volatiles from Ethiopian arabica drip coffee (L value, 23)
was conducted by using SPME fibers with 4 different lengths. Dilu-
tion of injected amount was controlled by fiber length, which lim-
ited the dilution value into 4 steps (20 , 21 , 22 , and 23 ); however, this
procedure was extremely useful to be able to evaluate the relative
intensity of each component by GC/O analysis.
The charm values (Acree and others 1984) and OSVs (Acree 1997)
of 45 odorants detected by GC/O analysis are listed in Table 5.
An OSV is the normalized charm value modified with an approxi-
mate Stevens’ law exponent (n = 0.5) to the most potent odorant
detected. OSVs are calculated by the formula of OSV = ((charm
value)/(charm value) max )1/2 × 100, and approximate the relative
importance of odorants by accounting for the exponential nature
C: Food Chemistry & Toxicology
of olfactory psychophysics; charm values indicate the true odor ac-
tivity measurement and are a linear function of concentration.
3-Methylbutyric acid (acidic odor), 2,3-butanedione (buttery-
oily odor), 2-ethyl-3,5-dimethylpyrazine (nutty-roast odor),
2-methoxyphenol (phenolic odor), 3-methyl-2-butene-1-thiol
(smoke-roast odor), 4-hydroxy-2,5-dimethyl-3(2H)-furanone
(sweet-caramel odor), 3-hydroxy-4,5-dimethyl-2(5H)-furanone
(sweet-caramel odor), and (E)-beta-damascenone (sweet-fruity
odor) showed large charm values and these odorants were found to
make a major contribution to the fresh aroma of Ethiopian arabica
drip coffee. Among the odorants identified in this experiment,
1-(3,4-dihydro-2H-pyrrol-2-yl)-ethanone (nutty-roast odor) and
 Bicchi C, Drigo S, Rubiolo P. 2000. The influence of fibre coating in headspace-solid
4-(4 -hydroxyphenyl)-2-butanone (raspberry ketone, sweet-fruity
odor) have not been reported in published literature to the best
of our knowledge. 1-(3,4-Dihydro-2H-pyrrol-2-yl)-ethanone and

4-(4 -hydroxyphenyl)-2-butanone are therefore newly detected
coffee aroma components in this GC/O study of fresh drip cof-
fee aroma. These 2 compounds were also detected by GC/O
analyses using a nonpolar column (HP-5) in the same Kovats
indices and aroma properties with those of standard compounds.
Their indices in HP-5 were 922 for 1-(3,4-dihydro-2H-pyrrol-
 in roasted coffee. J Agric Food Chem 44:3268–72.
2-yl)-ethanone and 1554 for 4-(4 -hydroxyphenyl)-2-butanone.

4-(4 -Hydroxyphenyl)-2-butanone was detected as a small peak in
the GC/MS chromatogram of the ether extract of freshly brewed
coffee in the same retention index as an authentic sample. In
 Hofmann T, Schieberle P. 2002. Chemical interaction between odor-active thiols and
particular, 4-(4 -hydroxyphenyl)-2-butanone had a comparatively
high OSV (54), which approximated the relative importance
 Holscher W, Steinhart H. 1992. Investigation of roasted coffee freshness with an im-
of odorants, in GC/O analysis. This result suggests that 4-(4 -
hydroxyphenyl)-2-butanone makes a considerable contribution to
the characteristic aroma of freshly drip-brewed Ethiopian arabica
coffee.
Conclusions
T o investigate the fresh and pleasant aroma of freshly brewed
drip coffee, a short-time (2-min) sampling method to capture
the headspace volatiles just after brewing coffee using a SPME fiber
(fiber type; DVB/carboxen/PDMS) was developed in this study. Us-
ing this SPME sampling method, the headspace volatiles of freshly
brewed drip coffee (Ethiopian arabica coffee, roast degree: L value,
23) was examined by GC/MS and GC/O (CharmAnalysisTM ) anal-
yses. The GC/MS results proved that this sampling method was
reproducible and that the composition of the headspace volatiles
of freshly brewed coffee changed shortly after brewing. Also, from
the GC/O results, 1-(3,4-dihydro-2H-pyrrol-2-yl)-ethanone (nutty-
 Semmelroch P, Grosch W. 1996. Studies on character impact odorants of coffee brews.
roast odor) and 4-(4 -hydroxyphenyl)-2-butanone (raspberry ke-
tone, sweet-fruity odor) are reported for the 1st time as components
 oil and corn oil compounds. J Food Sci 67:71–6.
in coffee aroma. In particular, 4-(4 -hydroxyphenyl)-2-butanone
had a comparatively high OSV (54), which approximated the rel-
ative importance of odorants, in GC/O analysis. This result sug-
C396 JOURNAL OF FOOD SCIENCE—Vol. 72, Nr. 7, 2007

gests that 4-(4 -hydroxyphenyl)-2-butanone makes a considerable
contribution to the characteristic aroma of freshly drip-brewed
Ethiopian arabica coffee. This short-time sampling method will al-
low us to study the fresh aroma characteristics of various types of
freshly brewed drip coffees, which would be of future use in coffee
flavor creation.
References
Acree TE. 1997. GC/olfactometry. Anal Chem News Features 69:170A–5A.
Acree TE, Barnard J, Cunningham DG. 1984. A procedure for the sensory analysis of
gas chromatographic effluents. Food Chem 14:273–86.
Akiyama M, Murakami K, Ohtani N, Iwatsuki K, Sotoyama K, Wada A, Tokuno K,
Iwabuchi H, Tanaka K. 2003a. Analysis of volatile compounds released during the
grinding of roasted coffee beans using solid-phase microextraction. J Agric Food
Chem 51:1961–9.
Akiyama M, Murakami K, Ohtani N, Iwatsuki K, Sotoyama K, Wada A, Tokuno K,
Iwabuchi H, Tanaka K. 2003b. Characterization of flavor compounds during grind-
ing of roasted coffee beans. In: Deibler KD, Delwiche J, editors. Handbook of fla-
vor characterization: sensory analysis, chemistry, and physiology. 1st ed. New York:
Marcel Dekker Inc. p 231–58.
Akiyama M, Murakami K, Ikeda M, Iwatsuki K, Kokubo S, Wada A, Tokuno K, Onishi M,
Iwabuchi H, Tanaka K. 2005. Characterization of flavor compounds released during
grinding of roasted robusta coffee beans. Food Sci Technol Res 11:298–307.
Arthur CL, Pawliszyn J. 1990. Solid phase microextraction with thermal desorption
using fused silica optical fibers. Anal Chem 62:2145–8.
Baltes W, Bochmann G. 1987. Model reaction on roast aroma formation. III. Mass
spectrometric identification of pyrroles from the reaction of serine and threonine
with sucrose under the condition of coffee roasting. Z Lebensm Unters Forsch
184:478–84.
Bicchi CP, Panero OM, Pellegrino GM, Vanni AC. 1997. Characterization of roasted
coffee and coffee beverages by solid-phase microextraction-gas chromatography
and principal component analysis. J Agric Food Chem 45:4680–6.
phase microextraction-gas chromatography (HS-SPME-GC) analysis of aromatic
and medicinal plants. J Chromatogr A 892:469–85.
Bicchi C, Iori C, Rubiolo P, Sandra P. 2002. Headspace sorptive extraction (HSSE), stir
bar sorptive extraction (SBSE), and solid phase microextraction (SPME) applied to
the analysis of roasted arabica coffee and coffee brew. J Agric Food Chem 50:449–
59.
Buchi G, Degen P, Gautschi F, Willhalm B. 1971. Structure and synthesis of Kahweo-
furan, a constituent of coffee aroma. J Org Chem 36:199–200.
Charles-Bernard M, Roberts DD, Kraehenbuehl K. 2005. Interactions between volatile
and nonvolatile coffee components. 2. Mechanistic study focused on volatile thiols.
J Agric Food Chem 53:4426–33.
Czerny M, Wagner R, Grosch W. 1996. Detection of odor-active ethenylalkylpyrazines
Czerny M, Mayer F, Grosch W. 1999. Sensory study on the character impact odorants
of roasted Arabica coffee. J Agric Food Chem 47:695–9.
Deibler KD, Acree TE, Lavin EH. 1999. Solid phase microextraction application in gas
chromatography/olfactometry dilution analysis. J Agric Food Chem 47:1616–8.
Grosch W. 1998. Flavour of coffee. A review. Nahrung 42:344–50.
melanoidins involved in the aroma staling of coffee beverages. J Agric Food Chem
50:319–26.
proved headspace technique. Z Lebensm Unters Forsch 195:33–8.
Holscher W, Vitzthum OG, Steinhart H. 1990. Identification and sensorial evaluation
of aroma-impact-compounds in roasted Colombian coffee. Café Cacao 34:205–12.
Jung D, Ebeler SE. 2003. Headspace solid-phase microextraction method for the study
of the volatility of selected flavor compounds. J Agric Food Chem 51:200–5.
Kimpe NG, Stevens CV, Keppens MA. 1993. Synthesis of 2-acetyl-1-pyrroline, the prin-
cipal rice flavor component. J Agric Food Chem 41:1458–61.
López-Galilea I, Fournier N, Cid C, Guichard E. 2006. Changes in headspace volatile
concentrations of coffee brews caused by the roasting process and the brewing pro-
cedure. J Agric Food Chem 54:8560–6.
Mayer F, Czerny M, Grosch W. 2000. Sensory study of the character impact aroma
compounds of a coffee beverage. Euro Food Res Technol 211:272–6.
Miszczak A, Forney CF, Prange RK. 1995. Development of aroma volatiles and color
during postharvest ripening of Kent strawberries. J Am Soc Hortic Sci 120:650–5.
Nijssen LM, Visscher CA, Maarse H, Willemsens LC, Boelens MH. 1996. Volatile com-
pounds in food. In Qualitative and quantitative data, 7th ed. Zeist, The Netherlands:
TNO Nutrition and Food Research Inst. 72:1–72. 23.
Roberts DD, Pollin P, Milo C. 2000. Solid phase microextraction method development
for headspace analysis of volatile flavor compounds. J Agric Food Chem 48:2430–7.
Sanz C, Czerny M, Cid C, Schieberle P. 2002. Comparison of potent odorants in a fil-
tered coffee brew and in an instant coffee beverage by aroma extract dilution anal-
ysis (AEDA). Euro Food Res Technol 214:299–302.
Semmelroch P, Grosch W. 1995. Analysis of roasted coffee powders and brews by gas
chromatography-olfactometry of headspace samples. Food Sci Tech 28:310–3.
J Agric Food Chem 44:537–43.
Steenson DF, Lee JH, Min DB. 2002. Solid phase microextraction of volatile soybean
Yang X, Peppard T. 1994. Solid-phase microextraction for flavor analysis. J Agric Food
Chem 42:1925–30.