Content uploaded by Gianmichele Arrighetti
Author content
All content in this area was uploaded by Gianmichele Arrighetti on Feb 06, 2019
Content may be subject to copyright.
Research Paper
Insights into the physicochemical properties of coffee oil
Sonia Calligaris1, Marina Munari1, Gianmichele Arrighetti2and Luisa Barba2
1Dipartimento di Scienze degli Alimenti, University of Udine, Udine, Italy
2Institute of Crystallography, National Council of Research, Trieste, Italy
The lipid fraction of roasted coffee is an interesting ingredient that could be used in a large number of food
formulations. Coffee oil has peculiar flavouring as well as nutraceutical characteristics. The feasibility of
the use of coffee oil as ingredient greatly depends not only on its chemical characteristics but also on its
physical properties. The crystallisation and melting properties of the coffee oil extracted from Arabica
roasted coffee powder were determined by using synchrotron X-ray diffraction coupled with differential
scanning calorimetry. The fatty acid composition and the flavour profile were also assessed by using GC
and GC-MS analyses, respectively. The main fatty acids found in coffee oil are linoleic and palmitic acid.
Significant amounts of stearic and oleic acid are also present. These chemical characteristics are linked to
the phase transition behaviour. The crystallisation of coffee oil occurs at 6.5 60.3 7C, independently of
the cooling rate applied (from 0.5 to 10 7C/min). A unique crystalline structure was identified: a double
chain length (2L) b’ structure (55.29 Å). The sole formation of the b’ form indicates that this metastable
crystal is the only one that one should expect in foods containing coffee oil stored below 7 7C.
Keywords: DSC/XRD / Oil / Physicochemical properties / Roasted coffee
Received: February 24, 2009; accepted: June 24, 2009
DOI 10.1002/ejlt.200900042
1270 Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277
1 Introduction
Lipids in green coffee beans are mainly located in the endo-
sperm while only a small amount is found in the outer layer.
The lipid content ranges from 10 to 14%, depending on the
coffee origin: in green Arabica coffee it averages some 15% on
a dry basis, whilst in Robusta it is about 10% [1, 2]. Lipids
extracted from coffee beans contain about 75% of triacyl-
glycerols (TG) with a high percentage of unsaponificables,
including about 19% of total free and esterified diterpene
alcohols, about 5% of total free and esterified sterols, and
very low quantities of other substances such as tocopherols.
Among the different diterpenes, cafestol and kahweol have
been widely studied due to their potential anticarcinogenic
effects [3–5]. It is well known that the chemical and physico-
chemical characteristics of green coffee beans greatly change
during the roasting process. The main changes are associated
with the development of the Maillard reaction, which allows
beverages with particular characteristics of flavour, colour
and texture to be produced. In fact, the heat treatment at
very high temperatures (around 200 7C) induces the forma-
tion of a number of volatile compounds with a wide range of
functional groups [6, 7]. The majority of the components of
the coffee aroma is liposoluble and can be extracted along
with the lipids from the roasted coffee beans. Different
extraction methods have been proposed: solvent extraction,
supercritical carbon dioxide extraction, and mechanical
extraction under pressure [8, 9]. As reported by Sarrazin et
al. [8], aroma recovery greatly depends on the extraction
methodology applied.
Even if the roasting process of green coffee induces dra-
matic changes in the coffee beans, the literature data evidence
that the lipid fraction of coffee is stable just after processing
has been completed and during storage [10, 11]. In particular,
Anese et al. [10] evidenced that the roasting does not affect the
oxidation level of the coffee lipid fraction. The stability of
coffee oil was attributed to the presence of lipid-soluble dark-
coloured Maillard reaction products. The latter are indicated
as having strong antioxidant properties through different
mechanisms, e.g. chain breaking, oxygen scavenging or metal
chelating [12–16].
On the basis of these observations, roasted coffee oil could
be considered an interesting matrix for a wide number of food
formulations, in which it could be used as flavouring ingre-
Correspondence: Sonia Calligaris, Dipartimento di Scienze degli Ali-
menti, University of Udine, Via Sondrio 2/a, 33100 Udine, Italy.
E-mail: sonia.calligaris@uniud.it
Fax: 139 043 2558130
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277 Coffee oil physicochemical properties 1271
dient (i.e. ice creams, ready-to-drink beverages, instant cof-
fee) or as nutraceutical able to improve the health-protecting
capacity of food products. The possibility of an efficientuse of
coffee oil as ingredient greatly depends not only on its chemi-
cal characteristics but also on its physicochemical properties.
Knowledge of the phase transition behaviour of coffee oil is
crucial to evaluate the feasibility of its use as ingredient in a
complex food, which has to be stored under defined condi-
tions. At the moment, very little literature data are available on
the chemical characteristics of coffee oil, whereas no study
provides information on its thermal and structural properties.
The crystallisation behaviour of complex lipids can be
studied by combining thermo-analytical techniques, among
which differential scanning calorimetry (DSC) is the most
widely used, and diffraction techniques such as X-ray dif-
fraction (XRD) [17]. The latter is the most direct technique to
study polymorphism arising from the different lateral packing
of fatty acid chains and of longitudinal stacking of molecules
in lamellae [18]. The two levels of organisation are easily
identifiable from the short and long spacing observed by XRD
at wide (WAXD) and small angle (SAXD), respectively. In
particular, the use of synchrotron radiation, which provides an
X-ray flux 103–106times more intense than that generated by
usual X-ray sources, allows step-by-step recordings as a
function of temperature [17, 19].
The aim of this paper was to study the crystallisation and
melting properties of the coffee oil extracted from Arabica
roasted coffee powder. In particular, the phase transition be-
haviour of TG in coffee oil was evaluated by synchrotron
XRD coupled with DSC. The fatty acid composition and the
flavour profile of coffee oil were also assessed.
2 Materials and methods
2.1 Coffee oil preparation
The lipid fraction of commercial roasted coffee powder
(100% Coffea Arabica) was obtained by solid-liquid extraction
using chloroform/methanol (Carlo Erba, Milan, Italy) mix-
tures (2 : 1 wt/wt) by stirring at room temperature for 3 h.
The ratio between the coffee and solvent mixture was 1 : 6 on
weight basis. After filtration through filter paper (Whatman
No. 1), the oil was separated from the solvent by evaporation
with a Rotavapor (mod. 4001; Heidolph Instruments, Milan,
Italy) at 40 7C.
2.2 Analytical determinations
2.2.1 Fatty acid content
Analysis of the fatty acid composition of the coffee oil was
carried out according to the European Official Methods of
Analysis [20].
2.2.2 Headspace solid-phase micro-extraction
sampling
The manual holder and the solid-phase micro-extraction
(SPME) fibre Sableflex 2 cm-50/30 mm DVB/CAR/PDMS
film were purchased from Supelco (Bellefonte, PA, USA).
Before sampling, the fibre was reconditioned for 30 min in the
GC injection port at 240 7C. Aliquots of 3 g of coffee oil were
inserted in 10-mL capacity vials, immediately sealed with
butyl septa and metallic caps. Vials were equilibrated at 60 7C
in a thermostatic bath for 30 min. An optimisation of the
experimental conditions had previously been realised. The
SPME fibre was exposed to the coffee oil headspace for 5 min.
2.2.3 GC-MS
The SPME coating containing the headspace volatile com-
pounds was immediately inserted into the GC injection port,
pushed out of its housing, and thermally desorbed for 5 min at
250 7C. A HGRC Mega 2 Series gas chromatograph (Fisons
Instruments, Milan, Italy) and a thermal conductivity detector
(Fisons HWD Control; Fisons Instruments) were used. The
separation was done by using a capillary column (CP Wax
52 CB, 50 m60.32 mm60.40 mm film thickness; Chrom-
pack, Middelburg, The Netherlands). The injector tempera-
ture was set at 200 7C and helium (1.7 mL/min linear speed)
was the carrier gas. The oven temperature was maintained at
60 7C for 6 min and then raised at 5 7C/min up to 200 7C. The
chromatograms were integrated using Chromcard (Ver. 1.18,
1996; CE Instrument, Milan, Italy) chromatography data
system software.
The MS analysis was performed using a Varian Saturn
mass spectrometer (ion trap detector) (Varian, Palo Alto, CA,
USA) operated in the electron impact ionisation mode
(70 eV). The ion source temperature was set at 250 7C. Each
sample was analysed in triplicate.
2.2.4 Identification of volatile compounds
The identification of volatile compounds was carried out by
comparison of their mass spectra with those of pure reference
compounds and the Wiley library, and also by comparing their
retention times with those of standard compounds and data
from the literature.
2.2.5 DSC
Calorimetric analyses were made using a TA4000 differential
scanning calorimeter (Mettler-Toledo, Greifensee, Switzer-
land) connected to GraphWare software TAT72.2/5 (Mettler-
Toledo). Heat flow calibration was achieved using indium
(heat of fusion 28.45 J/g). Temperature calibration was car-
ried out using hexane (m.p. –93.5 7C), water (m.p. 0.0 7C)
and indium (m.p. 156.6 7C). Samples were prepared by care-
fully weighing 10–15 mg of the coffee oil in 160-mL alumi-
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
1272 S. Calligaris et al. Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277
nium DSC pans, which were closed without hermetic sealing.
An empty pan was used as reference. Samples were heated
under nitrogen flow (0.5 mL/min) at 40 7C for 10 min to
destroy the crystallisation memory, cooled to –30 7C and then
heated from –30 to 40 7C. The scanning rate was 0.5, 2, 5 and
10 7C/min. The start and the end of the melting transition
were taken as onset (Ton) and offset (Toff) points of transition,
which are the points at which the extrapolated baseline inter-
sects the extrapolated tangent of the calorimetric peak in the
transition state. Results were normalised to account for the
weight variation of the samples. Total peak enthalpy was
obtained by integration. The programme STAR ever. 8.10
(Mettler-Toledo) was used to plot and analyse the thermal
data.
2.2.6 XRD analysis
XRD patterns were recorded at the XRD beam-line at the
Elettra storage ring in Trieste. The X-ray beam emitted by the
wiggler source on the Elettra 2 GeV electron storage r ing was
monochromatised by an Si(111) double crystal mono-
chromator, focused on the sample and collimated by a double
set of slits, giving a spot size of 0.260.2 mm. The sample
consisted of a drop of oil kept in the photon flux by means of a
nylon loop of 0.7 mm. The temperature of the sample was
varied by means of a 700 series cryocooler (Oxford Cryosys-
tems, Oxford, UK) with an accuracy of ,17C. The temper-
ature profile was the same as that of the DSC experiments
(heating at 40 7C for 10 min, cooling to –30 7C and then
heating to 20 7C at a scanning rate of 2 7C/min). In order to
collect data under the best analytical conditions at both the
wide- and small-angle signal, experiments were carried out at
two different photon energies and sample distances from the
detector (1.41 Å, 93.3 mm and 0.85 Å, 300 mm). A MarRe-
search 165 mm CCD detector assembly was used. As the
intensity of the synchrotron radiation beam decreases in time,
each diffraction pattern was collected rotating the sample for
0.017at variable rotation speed, under the condition that the
dose of photons absorbed by the sample was the same for
every step. Several hundreds of bi-dimensional patterns col-
lected with the CCD were calibrated and integrated using the
software FIT2D [21], resulting in two series of powder-like
Table 1. Fatty acid composition of coffee oil.
Fatty acid composition [wt-%]
16:0 34.3 60.4
18:0 6.5 60.1
18:1 8.5 60.2
18:2 46.1 60.3
18:3 1.2 60.2
20:0 2.1 60.3
20:1 0.2 60.1
22:0 0.3 60.1
patterns. The high-brilliance source allowed to record weak
structures not otherwise detectable, which helped in the pro-
cess of indexing the patterns.
2.3 Data analysis
Each coffee oil sample was analysed in triplicate. All results are
shown as mean and standard deviation. The indexing of the
XRD patterns obtained by the two crystalline phases was
performed using the programmes Winplotr [22] and Check-
cell [23].
3 Results and discussion
Coffee oil extracted from coffee powder is a brown viscous
liquid. The colour of the product is mainly due to the presence
of liposoluble Maillard reaction products separated during the
oil extraction. Table 1 shows the fatty acid composition of the
coffee oil. The main fatty acid is linoleic acid (L), followed by
palmitic acid (P). Significant amounts of stearic (S) and oleic
acid (O) are also present, whereas the percentages of linolenic
(Li) and arachidonic (A) acid are about 1–2%. Finally, gado-
leic (G) and behenic (B) acid are found only in traces. It
should be remembered that the roasting process is indicated to
cause only slight changes in the fatty acid composition [11].
The results are in agreement with previous literature data on
crude green coffee [2]. As reported by Folstar [24], the fatty
acids of the coffee oil are organised predominantly in two TG:
PLP (about 28.1%) and PLL (about 27.5%). Significant
quantities of SLP (about 8.6%), LLL (about 6.7%), POP
(5.9%) and SLL (4.2%) are also present.
In accordance with literature data, the oil extracted from
the roasted coffee powder is rich in aroma compounds [8].
Table 2 shows the volatile compounds identified in the head-
space of the coffee oil, including aldeydes, ketones, furans,
pyrroles, pyrazines, pyridine and phenolic compounds. All
these compounds are typical of roasted coffee flavour [25, 26].
The possibility to use coffee oil as flavouring ingredient in
foods greatly depends on its physicochemical properties. The
latter have been assessed by applying DSC and synchrotron
XRD analysis. Figures 1 and 2 show the crystallisation and
melting curves of coffee oil obtained during cooling and sub-
sequent heating at 0.5, 2, 5 and 10 7C/min from 20 to –40 7C
and vice versa. It can be noted that a single exothermic event
was recorded during cooling, and a single endothermic one
was observed during heating. Tables 3 and 4 show the onset
temperature (Ton), the offset temperature (Toff) and the en-
thalpy (DH) associated with crystallisation and melting of
coffee oil. Ton of the crystallisation and melting curves is not
affected by the scanning rate. On the contrary, slight differ-
ences in Toff of crystallisation and melting are observed as a
function of the scanning rate. These changes take place con-
comitantly with a decrease in the transition enthalpy as the
scanning rate increases. This result is attributable to the fact
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277 Coffee oil physicochemical properties 1273
Table 2. Main volatile compounds identified in the headspace of
coffee oil.
Peak Compound
1 Acetaldehyde
2 2-Methylfuran
3 2-Methylbutanal
4 2,5-Dimethylfuran
5 2,3-Pentanedione
6 2,3-Hexanedione
7 1-Methylpyrrole
8 Pyridine
9 Pyrazine
10 2-Methoxymethylfuran
11 2-Methylpyrazine
12 2,5-Dimethylpyrazine
13 2,6-Dimethylpyrazine
14 2-Ethylpyrazine
15 2,3-Dimethylpyrazine
16 2-Ethyl-6-methylpyrazine
17 2-Ethyl-5-methylpyrazine
18 2-Ethyl-3-methylpyrazine
19 2,6-Diethylpyrazine
29 2-Methyl-3,5-diethylpyrazine
21 Furfural
22 2-Acetylfuran
23 2-Furanmethanol acetate
24 5-Methylfurfural
25 2-Methyldihydrofuranone
26 2-Furanmethanol
27 2-Methyl-1-pyrrole
28 2-Methoxyphenol
29 2-Acetylpyrrole
30 Difurfuryl ether
31 1-Pyrrole-2-carboxaldehyde
32 4-Ethyl-2-methoxyphenol
33 4-Methylphenol
34 4-Ethylphenol
35 2-Methoxy-4-vinylphenol
36 1-Furanyl-2-formylpyrrole
that a lower portion of oil crystallised at the higher cooling
rates, probably because the TG under such conditions do not
have enough time to organise [27].
These results allow coming to the hypothesis that one sin-
gle crystal form develops in the coffee oil during crystal-
lisation, independently of the cooling rate applied. This indi-
cates that the main TG molecules, which have long fatty acid
chains from 16 to 18 carbons, crystallise together in a unique
crystal structure during a unique thermal event.
To study the polymorphic structure of the coffee oil crys-
tals, synchrotron XRD analysis was performed at both small
and wide angles. Figures 3 and 4 show the patterns recorded
at small and wide angles, respectively, during the diffraction
experiment performed at a wavelength of 0.85 Å. The results
are reported as a function of temperature during cooling and
heating at 2 7C/min. In order to evidence the correspondence
between DSC and XRD events, white lines at temperatures
corresponding to the DSC thermal events (Ton,Toff and tem-
perature corresponding to the peak) are also reported in
Figs 3 and 4.
From 60 to 6.5 7C, two bumps at 4.69 and 23.61 Å are
observed. These bumps can be associated with the short-
range organisation of the TG molecules in the liquid phase, as
previously reported by other authors [27, 28].
At about 6.5 7C, in agreement with the DSC data, the
crystallisation of coffee oil is put in evidence by the appear-
ance of a number of diffraction peaks. In particular, wide-
angle diffraction peaks emerge at 4.17, 3.73 and 2.51 Å con-
comitantly with small-angle peaks at 54.80, 27.60, 18.41,
13.81 and 9.22 Å. The intensity of these peaks increases pro-
gressively. Once the temperature reaches –40 7C, after a 10-
min pause, the samples was heated at 2 7C/min. At –15 7C, in
correspondence with the endothermic DSC peak, the inten-
sity of all the XRD peaks decreases, indicating that the oil
starts to melt. The peaks disappear at about 6.5 7C(Toff of the
DSC exothermic peak). It is evident that no polymorphic
transformation is observed during the heating of the samples.
The interplanar distances at 4.17 and 3.73 Å are typical of
the organisation of the acylglycerol chain in an orthorhombic
perpendicular b’ subcell [15]. The small-angle diffraction
peaks (54.80, 27.60, 18.41, 13.81 and 9.22 Å), along with the
wide-angle peak, 2.51 Å, correspond to a double chain length
organisation, named 2L, with a parameter cof 55.29 Å.
The refining of the reticular parameter cagainst the five
peak positions at low resolution, performed by using the soft-
ware Checkcell, converged to a value that agreed very well (Dy
= 0.00097) with the position of the high-resolution peak, pur-
posely excluded from the refinement in order to validate the
value chosen as cparameter.
Since the occurrence of the b’ polymorph can be expected
when one of the three fatty acid chains of a TG is somehow
different from the other two [18], the formation of the b’
crystal in coffee oil could be related to the predominance of
this type of TG (i.e. PLP, PLL, SLL, POP, SLL). In addition,
considering the fatty acid composition of coffee oil, 18 can be
considered as the mean number of carbon atoms in the fatty
acid chains. Since 1.52 Å is reported as an average carbon-
carbon distance in the zigzag plane of the acylglycerol chain
[29], the mean length of one fatty acid chain is 1.52618 =
27.36 Å. Thus, the length corresponding to the two fatty acid
chains is 54.72 Å. This value is an excellent confirmation of
the experimental data, indicating that the b’ L2 structure is
mainly constituted by palmitic (16:0), stearic (18:0), oleic
(18:1) and linoleic acid (18:2).
It should be noted that the same crystalline structure was
the only one identified even after a flash freezing of the coffee
oil (data not shown). As is well known, flash freezing of the
lipid matrix is generally applied to induce the formation of the
less thermostable crystal form (aform) [15]. Under our
experimental conditions, the sole presence of the b’ form
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
1274 S. Calligaris et al. Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277
Figure 1. DSC crystallisation curves of coffee oil at
scanning rates of –0.5, –2, –5 and –10 7C/min.
Figure 2. DSC melting curves of coffee oil at scan-
ning rates of 0.5, 2, 5 and 10 7C/min.
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277 Coffee oil physicochemical properties 1275
Figure 3. WAXD patterns as a function
of temperature recorded during cooling
and subsequent heating of coffee oil at
a scanning rate of 2 7C/min. Each dif-
fraction pattern is represented as
intensity (counts) vs. interplanar dis-
tance d(Å) as a function of tempera-
ture. As temperature decreases, the
patterns present a darker colour. The
white lines represent the patterns
recorded in correspondence with Ton,
Toff and Tat peaks of DSC thermal
events.
Figure 4. SAXD patterns as a function
of temperature recorded during cooling
and subsequent heating of coffee oil at
a scanning rate of 2 7C/min. Each dif-
fraction pattern is represented as
intensity (counts) vs. interplanar dis-
tance d(Å) as a function of tempera-
ture. As temperature decreases, the
patterns present a darker colour. The
white lines represent the patterns
recorded in correspondence with Ton,
Toff and Tat peaks of DSC thermal
events.
indicates that this metastable crystal is the only one that
should be expected in foods containing coffee oil stored
below 7 7C.
In conclusion, the use of DSC analysis in combination
with a high-flux X-ray source for the XRD technique allows
the identification and the characterisation of the crystalline
structures formed during the crystallisation of coffee oil. In
particular, the TG organise in a double chain length structure
with an orthorhombic perpendicular subcell: b’ 2L (55.29 Å).
From a technological point of view, coffee oil crystals
could be found in products stored at the temperatures nor-
mally applied for chilled (4 7C) and frozen (–18 7C) storage.
Under such conditions, coffee oil is not completely crystal-
lised and part of it is in the amorphous state, as shown by the
presence of the amorphous signal during the whole XRD
experiment. In addition, the crystallised fraction is expected
not to undergo polymorphic transformation during storage
below the phase transition temperature. It is interesting to note
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
1276 S. Calligaris et al. Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277
Table 3. Onset temperature (Ton), offset temperature (Toff), and en-
thalpy (DH) associated with the crystallisation of coffee oil as a
function of the scanning rate.
Scanning rate
[7C/min]
Ton [7C]
crystallisation
Toff [7C]
crystallisation
DH[J/g]
0.5 6.8 60.6a–14.6 60.9a49.4 62.1a
2 6.6 60.4a–15.4 60.5a49.1 61.5a
5 6.5 60.3a–16.6 60.4b45.0 61.2b
10 6.5 60.4a–21.6 61.2c41.0 61.9c
a,bç Data with different letters in the same column are significantly
different (p.0.05).
Table 4. Onset temperature (Ton), offset temperature (Toff), and en-
thalpy (DH) associated with the melting of coffee oil as a function of
the scanning rate.
Scanning rate
[7C/min]
Ton [7C]
melting
Toff [7C]
melting
DH[J/g]
0.5 –15.1 60.2a5.71 60.7a49.1 61.1a
2 –15.6 60.6a7.6 60.5a51.1 61.3a
5 –14.8 60.4a9.36 60.9b45.8 61.2b
10 –15.0 60.3a10.7 61.0b44.0 61.6b
a,b Data with different letters in the same column are significantly dif-
ferent (p.0.05).
that the partial crystallisation of oil may induce changes in the
overall perceived aroma of the product. In fact, an increase in
the headspace concentration of the aroma compounds, due to
the increase in their concentration in the liquid phase sur-
rounding fat crystals, could be expected [30]. This informa-
tion could be useful in the research and development process
of new food formulations containing coffee oil as ingredient.
The conflict of interest statement
The authors have declared no conflict of interest.
References
[1] A. J. Wilson, M. Petracco, E. Illy: Some preliminary investiga-
tions of oil biosynthesis in the coffee fruit and its subsequent re-
distribution within green and roasted beans. In: Proceedings of
the 17th ASIC Colloquium, ASIC, Paris (France) 1997, pp. 92–
99.
[2] K. Speer, I. Kolling-Speer: Lipids. In: Coffee: Recent Develop-
ments. Eds. R. J. Clarke, O. G. Vitzthum, Blackwell Science,
Oxford (UK) 2001, pp. 33–49.
[3] C. Cavin, D. Holzhaeuser, G. Scharf, A. Constable, W. W.
Hubber: Cafestol and kahweol, two coffee specific diterpenes
anticarcinogenic activity. Food Chem Toxicol. 2002, 40, 1155–
1158.
[4] B. Ross, M. J. Caslake, A. F. H. Stalenhoef: The coffee diter-
pene cafestol increases plasma triacylglycerol by increasing the
production rate of large VLDL apolipoprotein B in the heal-
thy normolipidemic subjects. Am J Clin Nutr. 2001, 73, 45–
48.
[5] W. W. Huber, W. Rossmanith, M. Grusch, E. Haslinger, S.
Prustomersky, B. Peter-Vorosmarty, W. Parzefall, G. Scharf,
R. Schulte-Hermann: Effects of coffee and its chemopreven-
tive components kahweol and cafestol on cytochrome P450
and sulfotransferase in rat liver. Food Chem Toxicol. 2008, 46,
1230–1238.
[6] M. Czerny, F. Mayer, W. Grosch: Sensory study on the char-
acter impact odorants of roasted Arabica coffee. J Agric Food
Chem. 1999, 47, 695–699.
[7] M. Czerny, W. Grosch: Potent odorants of raw Arabica coffee.
Their changes during roasting. J Agric Food Chem. 2000, 48,
868–872.
[8] C. Sarrazin, J. L. Le Quesre, C. Gretsh, R. Liardon: Repre-
sentatives of coffee aroma extracts: A comparison between
different methods. Food Chem. 2000, 70, 99–102.
[9] J. M. A. Araujo, D. Sandi: Extraction of coffee diterpenes and
coffee oil using supercritical carbon dioxide. Food Chem. 2006,
101, 1087–1094.
[10] M. Anese, T. De Pilli, R. Massini, C. R. Lerici: Oxidative sta-
bility of the lipid fraction in roasted coffee. Ital J Food Sci.
2000, 12, 457–462.
[11] M. A. Vila, S. Andeuza, P. De Pena, C. Cid: Fatty acid evolu-
tion during the storage of ground, roasted coffees. J Am Oil
Chem Soc. 2005, 82, 639–646.
[12] M. C. Nicoli, M. Anese, L. Manzocco, C. R. Lerici: Anti-
oxidant properties of coffee brews in relation to the roasting
degree. Lebensm Wiss Technol. 1997, 30, 292–297.
[13] M. Anese, M. C. Nicoli: Antioxidant properties of ready-to-
drink coffee brews. J Agric Food Chem. 2003, 51, 942–946.
[14] B. Caemmerer, L. W. Kroh: Antioxidant activity of coffee
brews. Eur Food Res Technol. 2006, 223, 469–474.
[15] P. Parras, M. Martinez-Tomè, A. M. Jimenez, M. A. Murcia:
Antioxidant capacity of coffees of several origins brewed fol-
lowing three different procedures. Food Chem. 2007, 102,
582–592.
[16] R. Amarowicz: Antioxidant activity of Maillard reaction
products. Eur J Lipid Sci Technol. 2009, 111, 109–111.
[17] G. Keller, F. Lavigne, L. Forte, K. Andrieux, M. Dahim, C.
Loisel, M. Ollivon, C. Bourgaux, P. Lesieur: DSC and X-ray
diffraction coupling. Specifications and application. J Ther-
mal Anal. 1998, 51, 783–791.
[18] K. Sato: Solidification and phase transformation behaviour of
fats – a review. Fett/Lipid 1999, 12, 467–474.
[19] S. Calligaris, G. Arrighetti, L. Barba, M. C. Nicoli: Phase
transition of sunflower oil as affected by the oxidation level.
J Am Oil Chem Soc. 2008, 85, 591–598.
[20] European Community: Regulation 2568/91. Off J Eur Com-
munity 1991, L 248, 28–29.
[21] A. P. Hammersley, S. O. Svensson, M. Hanfland, A. N. Fitch,
D. Hausermannm: Two-dimensional detector software: From
real detector to idealised image or two-theta scan. High Press
Res. 1996, 14, 235–248.
[22] T. Roisnel, J. Rodriguez-Carvajal: WinPLOTR: A Windows
tool for powder diffraction patterns analysis. In: Materials
Science Forum, Proceedings of the Seventh European Powder
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com
Eur. J. Lipid Sci. Technol. 2009, 111, 1270–1277 Coffee oil physicochemical properties 1277
Diffraction Conference (EPDIC 7), Eds. R. Delhez, E. J. Mit-
tenmeijer, Barcelona (Spain) 2000, pp. 118–123.
[23] J. Laugier, B. Bochu: CHECKCELL: A software performing
automatic cell/space group determination. Collaborative
Computational Project Number 14 (CCP14). Laboratoire
des Materiaux et du Génie Physique de l’Ecole Supérieure de
Physique de Grenoble (France) 2000.
[24] P. Folstar: Lipids. In: Coffee.Vol. 1. Eds. R. J. Clarke, R
Macrae, Elsevier Applied Science Publishers, New York, NY
(USA) 1985, pp. 203–222.
[25] M. Akiyama, K. Murakami, N. Ohtani, K. Iwatsuki, K.
Sotoyama, A. Wada, K. Tokuno, H. Iwabuchi, K. Tanaka:
Analysis of volatile compounds released during the grinding
of roasted coffee beans using solid-phase microextraction.
J Agric Food Chem. 2003, 51, 1961–1969.
[26] I. Lopez-Galilea, N. Fournier, C. Cid, E. Guichard: Changes
in headspace volatile concentration of coffee brews caused by
roasting process and the brewing procedure. J Agric Food
Chem. 2006, 54, 8560–8566.
[27] C. Lopez, F. Lavigne, P. Lesieur, G. Keller, M. Ollivon:
Thermal and structural behaviour of anhydrous milk fat.
2. Crystalline forms obtained by slow cooling. J Dairy Sci.
2001, 84, 2402–2421.
[28] K. Larsson: Molecular organization in lipids. In: Food Emul-
sions. Eds. S. E. Friberg, K. Larsson, Marcel Dekker, New
York, NY (USA) 1997, pp 111–140.
[29] A. van Langevelde, K. van Malssen, E. Sonneveld, R. Peschar,
H. Schenk: Crystal packing of a homologous series of
b’-stable triacylglycerols. J Am Oil Chem Soc. 1999, 76, 603–
609.
[30] D. D. Roberts, P. Pollien, B. Watzke: Experimental and mod-
elling studies showing the effect of lipid type and level on fla-
vour release from milk-based lipid emulsions. J Agric Food
Chem. 2003, 51, 198–195.
©2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.ejlst.com