ChapterPDF Available

Abstract and Figures

Coffee is a relatively young caffeinated beverage, known in Western Europe since the 17th century. Initially consumed in Europe by the aristocracy, coffee has developed into one of the world's most popular beverages. The unique aroma is without doubt one of the key drivers for its rise in consumption. Over the past two decades, research on coffee aroma has mostly focused on two areas. (1) Identification and quantification of aroma relevant volatile compounds. This led to the publication of lists of sensory relevant (impact/key) compounds, together with their chemical and sensory properties. (2) Elucidation of the formation and degradation mechanisms of coffee aroma compound. After a period of consolidation of all these insights, the focus is shifting to new areas: (i) time-resolved studies of coffee aroma formation, release and degradation; (ii) study of different types of interactions and the development of an increasingly holistic approach to aroma perception; (iii) prediction of sensory profiles from instrumental data.
Content may be subject to copyright.
Chapter 33
Coffee Volatile and Aroma
Compounds – From the Green
Bean to the Cup
Chahan Yeretzian*, Sebastian Opitz, Samo Smrke and
Marco Wellinger
Zurich University of Applied Sciences, Institute of Chemistry and
biotechnology, 8820 Wädenswil, Switzerland
*e-mail: yere@zhaw.ch
33.1
Introduction
Many hands in many countries on many continents from people who have
never met are needed to make a cup of coffee people who pour their
hearts into creating the cup that we enjoy. From the hands of the farmer
and farmworkers that plant, cultivate and harvest the cherries to the ones
that dry, sort and process them in the origin countries. From the hands that
pack and ship the beans to the hands that roast them, arriving in the hands
that grind and extract the brew, and finally into the hands of those who con-
sume it. From the seed to the cup, coffee is a product that goes through the
hands of many, seemingly unrelated, people. Together, these hands com-
pose this complex and fascinating value chain that crafts one of the world's
most consumed and popular beverages. According to the International Cof-
fee Organisation (ICO), in 2011 the coffee sector employed approximately
Coffee: Production, Quality and Chemistry
Edited by Adriana Farah
© The Royal Society of Chemistry 2019
Published by the Royal Society of Chemistry, www.rsc.org
726
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
727
26 million farmers and coffee workers,1 and it is estimated that around
125 million people worldwide depend on coffee for their livelihoods.2 If we
include coffee-related jobs such as the truck driver or the port worker who
transport the coffee, as well as those serving coffee in cafes and restaurants
or working at factories such as paper cup manufacturers, this figure would
increase further. At the consumer end, over 2.25 billion cups of coffee are
consumed in the world every day.3 These numbers indicate that coffee is a
crop of major economic importance to producing as well as to consuming
countries. Coffee is today the most widely commercialized tropical product
on the international market.1
Coffee is a lifestyle product that has penetrated the daily routine of many
people and that is continuing to expand into new and emerging markets.
While there are different factors that can be put forward to explain the suc-
cess story of coffee, such as the mildly stimulating effect of caffeine, the
social facilitation over a cup of coffee or the emerging positive health impli-
cations, the unique and complex flavour profile is probably the main reason
why so many enjoy a cup of coffee. This chapter will specifically address the
aroma of coffee.
33.2
Coffee Aroma From Seed to Cup
The aroma of a cup of coffee is the perceptible result of a series of transfor-
mations from the seed to the cup.4 Figure 33.1 outlines various factors that
impact the aroma of coffee. genetic predispositions, environmental and cli-
matic factors, harvest and post-harvest practices, sorting, grading, storage
and transport, roasting, grinding and extraction and finally consumption
practices all impact to varying degrees on the final result. This journey can
schematically be condensed into three stages: origin transformation
release. All three play a specific role in the final aroma.
Origin the predispositions: the green coffee variety (genetics) with its
specific set of chemical precursors in the green bean forms the basis for
the aroma compounds that will later develop and are released during
consumption to the sensory receptors, which triggers a sensation and
leads to a perception of the aroma. The agronomy, climate, harvest
practices and post-harvest treatment already at this stage start to mod-
ulate the genetic predisposition of the coffee bean and will affect the
composition of the green bean as it goes into roasting.
Transformation linking the seed to the cup: even the best green coffee
will be ruined if not properly processed. The craft and science of creat-
ing good coffee is to master every step when transforming the coffee.
however, the smell and taste of a green bean provide no clue as to what
it might turn into once roasted. The first and most significant step in
the transformation process is roasting. This unlocks the potential of a
green bean and creates the coffee aroma. however, grinding, degassing
and storage are also very important.
728
Chapter 33
Figure 33.1 every step matters in obtaining a good cup of coffee: the aroma of cof-
fee is the result of a long chain of transformations, which link the seed
to the cup. It is influenced by genetic, agricultural, chemical and tech-
nical factors. Most importantly, the many people and the care taken
by all those involved in these transformations are the most important
factors in shaping the final quality in the cup.
release connecting to the consumer: at the end of the long journey of
the coffee bean, both in distance and in time, the coffee finally comes
into contact with the consumer. The way we extract, prepare and con-
sume our coffee greatly affects the flavour we perceive.
33.3
The Sensory Experience of Coffee
people's motives for drinking coffee vary widely. Some just want the caf-
feine kick to keep them going. More and more people, though, are currently
drinking coffee not for the energy rush, but for the flavour a trend that
was strongly promoted by the advent of the specialty coffee movement in the
late 1960s. More recently, reported health benefits from moderate coffee con-
sumption are becoming an additional reason for people to choose to drink
coffee a trend fuelled by insights gained from research in the polyphenols
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
729
Figure 33.2 Our senses represent our windows to the world and to our cup of cof-
fee. While all are relevant and contribute to the holistic sensory expe-
rience while drinking coffee, some are considered more relevant than
others. In this chapter, we specifically focus on volatile organic com-
pounds that contribute towards the sense of olfaction.
and antioxidants of coffee.5 Finally, and something which has always been
the case, coffee has been a fabulous facilitator of social interactions a trend
that was propelled by the global rise of coffee shops.
In this chapter, we focus on the aroma of coffee, which is an important
facet of the multidimensional sensory experience, as shown in Figure 33.2.
Indeed, drinking coffee elicits a multi-sensory experience which involves all
our senses, including olfaction, taste, touch, trigeminal sensation, vision
and possibly hearing. Aroma is an odour; it is often referred to as smell and
is sensed by receptors in the nose. An important distinction is made between
the orthonasal and the retronasal aroma. When we have coffee in front of us,
we can smell it. We call this the orthonasal, or above-the-food, aroma. Vola-
tile organic compounds (VOCs) released from coffee enter our nasal cavity
during inhalations and reach the olfactory epithelium. The olfactory centre
resides high along the roof of the nasal cavity, just below and between the
eyes. It covers an area of 9 cm2 (3 × 3 cm) and contains some 1020 mil-
lion receptor cells.6–8 In order for these centres to be stimulated, the odour
molecules have to be inhaled into the nose and carried up to the roof of the
nose. The second type of aroma of significance to consumers is the aroma
released in the mouth when drinking coffee. While coffee is in the mouth
and also after swallowing, volatile compounds are released from the drink
and transported by various airflows (mouth movement, respiration) from the
oral cavity to the pharynx, passing the soft palate (velum palatinum). When
exhaling, volatiles are swept by the airflow coming from the oral cavity and
the lungs and are released through the nose. during their transport from
the oral cavity through the pharynx to the nasal cavity, VOCs pass along the
730
Chapter 33
olfactory epithelium and may trigger an olfactory perception. This aroma is
termed the retronasal, or in-mouth, aroma, and is part of the flavour of coffee
along with the taste aspect, as perceived during eating or drinking.9,10
Taste, or gustation, is the sensation of saltiness, sweetness, sourness,
bitterness and umami (savouriness). Flavour is the combination of taste
and smell. Another important additional sensation to describe the fla-
vour of coffee is the body. It can be light, like a dry light wine, or it can
be heavy, like a red wine. The perception of irritants is mediated not by
taste and smell receptors, but by other chemosensitive receptors. The
perceptual characteristics of chemical irritation, also called chemesthe-
sis, are mediated by non-specific, multimodal somatosensory fibres and
are a property of the skin.11 These chemical senses are complemented by
the physical senses of sight or vision, hearing and touch (somatosensa-
tion). Figure 33.2 summarizes the multidimensional space of the sensory
experience.
besides the various modalities of a sensory experience, as outlined in Fig-
ure 33.2, a complete sensory experience can be decomposed into four hierar-
chical levels, as outlined in Figure 33.3.
Figure 33.3 The sensory experience of coffee is a truly holistic one where objective
and measurable aspects of the product interact with processes occur-
ring at the individual/personal level in the mouth. These are comple-
mented by subjective processes in the brain and are finally modulated
by psychological and sociological elements such as mood, memories
and social context. All four levels collectively contribute to the overall
sensory experience elicited when consuming coffee.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
731
Starting at the level of the product itself, its chemical composition and
physical properties, the product can be integrated into an increasingly larger
context, including the specific consumer, her/his physiological, neurological
and psychological characteristics.
1.
Chemical and physical properties: this first and best understood fac-
tor of the sensory experience is exclusively related to the product itself,
without reference to the specific consumer and mode of consumption.
These are properties that can be precisely measured by instrumental/
analytical techniques and focus on the concentration of aroma and
taste compounds in the product, on compounds that elicit cooling and
pungent impressions, on temperature and texture (e.g. viscosity, % of
total dissolved solids, viscosity) and on colour.
2.
The consumption process: a specific cup of coffee might have a pre-
cisely measurable composition of aroma compounds, yet these com-
pounds may be released in-mouth and transported to the sensory
receptors differently from person to person, leading to inter-individual
differences in perception. Individual characteristics in consumption
patterns (volume of sip, breathing rhythm, swallowing pattern, move-
ment of the coffee in the mouth, etc.) and in-mouth physiognomy and
physiology (shape of nasal cavity, amount and composition of saliva)
will modulate the sensory experience and lead to inter-individual dif-
ferences in the sensory experience, even if exactly the same coffee is
consumed.12–20 Indeed, a given coffee will not “taste” the same for dif-
ferent people. This second factor, therefore, concerns the process of
drinking and all aspects related to the physiological and physical envi-
ronment in the mouth that leads to the liberation of the aroma and its
transport from the oral to the nasal cavity, where the olfactory recep-
tors are located. by recognizing and exploring such factors involved in
a person's perception of coffee aroma, it was possible to develop novel
analytical approaches that pave the way for the science of individual-
ized aroma.9,10,21
gierczynski et al. studied the effect of textural modifications of solid
milk gels on in vivo aroma release. The aroma perception was inves-
tigated with a panel of 14 subjects. great inter-individual differences
were observed on aroma-release data, and the consequences of these
differences on aroma perception were studied.22 While this may be
related to difference in consumption behaviours and in in-mouth phys-
iognomy and physiology, the fact cannot be excluded that neurological
differences may be at play as well.
3.
The neurological make-up of individuals: besides individual differ-
ences in consumption patterns and in-mouth characteristics, the
sensory experience is also modulated by individual differences in the
initial sensation at the level of individual receptors, as well as in final
and conscious perception. Sensation refers to the process of sensing
732
Chapter 33
our environment through touch, taste, sight, sound and smell.11,23,24
This information is sent to our brains in raw form where perception
comes into play. perception is the way we interpret these sensations
and, therefore, make sense of everything around us. It is affected by
a complex transduction process from activation of the multiple sen-
sory receptors to the final response pattern at the level of the cen-
tral nervous system, where interactions from other senses may be
integrated.25–27 Sensation is hence the process that allows our brains
to take in information via our five senses, which can then be experi-
enced and interpreted (i.e. turned into perception) by the brain. The
individual sensory experience is, therefore, affected not only by the
composition of sensory active compounds in the coffee and modu-
lated by the individual's in-mouth characteristics and consumption
habits. Individuals also differ in their neurological and physiological
make-up. In the same way as the in-mouth environment, this make-up
will lead to an individually modulated sensory experience. hence, the
same coffee may elicit distinctively different sensory experiences for
different people.28–31 A well-documented example is the perception
of bitterness (sensitivity to 6-n-propylthiouracyl (prOp) or phenylth-
iocarbamide (pTC)). It was established that individuals may vary in
the extent to which they perceive bitter tasting compounds32 and that
this inheritable trait is related to the density of fungiform papillae on
the tongue.33,34 There are reports that prOp sensitivity affects mac-
ronutrient selection.35 referring to smell, individual differences in
sensitivity to the odour of 4,16-androstadien-3-one (a pheromone)
were observed, and a bimodal sensitivity distribution was reported. A
smaller group of individuals showed a high sensitivity to androstadi-
enone (supersmellers).36 Whether this specific sensitivity to androsta-
dienone is due to exposure-induced changes at the receptor level or to
genetic factors37 remains to be determined. Furthermore, the inabil-
ity of some people to perceive specific odours, termed anosmia, is
also well established.38,39
4.
psychology and cognition: finally, individual history, past experiences,
expectations, product familiarity,40,41 psycho-social and cognitive fac-
tors42 such as culture, mood, conditioning and social context can all
affect the way a person experiences a cup of coffee (40,43–50). he or she
catches an aroma of a dish and is suddenly immersed in a flurry of vivid
memories, often emanating from childhood. What is it about smells
that can trigger memories so strong and real it feels like the person has
been transported back in time? This is known as “odour-evoked auto-
biographical memory” or the proust phenomenon, after the French
writer Marcel proust. In his famous novel In Search of Lost Time, the
narrator dips a madeleine cookie into a cup of tea and is transported
back into time as long-forgotten memories of his childhood come
flooding back. Indeed, research shows that odours are especially effec-
tive as reminders of past experience, much more so than cues from
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
733
other senses, such as sights or sounds.51,52 This might have to do with
the way the brain processes odours and memories. Smells are routed
through the olfactory bulb, which is the smell-analysing region in the
brain. This is closely connected to the amygdala and hippocampus,
brain regions that handle memory and emotion. The close connection
may explain why a scent might be tied to vivid memories in the brain,
and then come flooding back when one is exposed to that particular
odour trigger.
In the following sections, we look in detail at the subject of aroma anal-
ysis of coffee and explore the aroma of coffee from the green bean to the
cup. In Figure 33.4, pTr-ToF-MS (proton-Transfer-reaction Time-of-Flight
Mass-Spectrometry) profiles of green coffee, roasted whole bean coffee, roast
and
ground
(
r
&
g
)
coffee
and
fi
nally
of a coffee bre
w
are
sho
w
n,
all plotted on
an identical intensity scale. pTr-ToF-MS is an emerging analytical technique
that has first been applied by Yeretzian and co-workers (since 1997) to coffee
aroma analysis53–55 and is today an established technique in the field.56 While
green coffee shows a distinctively different profile of much lower intensities
than roasted coffee, grinding the roasted beans leads to a strong increase of
the volatile intensities.
In Section 33.4, we will first address the VOC profile of green/raw cof-
fee. While the actual aroma of the coffee beverage is formed during roast-
ing, the VOC profile of green coffee might reveal beans that will negatively
impact the quality of roasted coffee.57 From Section 33.5 on, we will move
into discussing the aroma of roasted coffee and “what makes coffee smell so
good”.58,59
33.4
Dynamic Headspace Analysis of Green Bean
Volatile Compounds
Volatile fingerprinting of green coffee beans has been used to identify poten-
tial markers for defective and low-quality beans.60,61 differences in green
bean volatile profiles were observed between coffees of different varieties
that were harvested from the same farm at different degrees of ripeness.62
Climate conditions, as well as the terroir, are also reflected in the volatile
profiles of the green beans.63 hence, the complex composition of green bean
VOCs seems to indirectly relate to coffee properties that are important for the
aroma in roasted coffee.
We studied the green coffee volatile composition by means of dynamic
headspace pTr-ToF-MS. Compared to traditional gC-MS, pTr-ToF-MS can
achieve better sensitivity and reproducibility of the analysis. A highly repro-
ducible way of measuring volatiles from coffee (green or roasted) is to perform
a dynamic measurement of the headspace composition. In this experiment,
the headspace over a sample is constantly flushed by nitrogen gas and the
volatiles, dynamically released from the sample, are continuously measured
Figure 33.4 pTr-ToF-MS mass spectra profiles of coffee as green whole beans, roasted whole beans, roast and ground beans and coffee
brew. The volatiles of ground coffee clearly show the strongest overall intensities.
734
Chapter 33
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
735
by a direct pTr-ToF-MS analysis. In the case of green coffee samples, this
improves the reliability of the measurements, since volatiles formed by sur-
face oxidation are rapidly depleted and only those from the inside of the bean
are measured. We found that measuring volatile composition over ground
green coffee leads to irreproducible results, most likely from uncontrollable
oxidation on the surface of the particles during sample preparation, as was
also observed when analysing ground green coffee with gC-MS.62
In contrast, VOC profiles of whole green beans resulted in highly repro-
ducible profiles when using dynamic headspace pTr-ToF-MS. Volatile pro-
files of green whole beans of
C. arabica
l.
(12 samples) and
C. canephora
pierre, cv. robusta (10 samples) were measured. A very large diversity of
volatile profiles was found, ranging from 87 to 170 mass peaks detected
in a green coffee sample with intensity over a threshold of 1 ion count
per second. The number of molecular mass peaks that could be reliably
assigned in all green coffee samples was relatively small only 17 masses
were present in the profiles of every sample. The instrumental analysis was
performed in a highly automatized way, and the data were analysed using
a non-targeted approach. Successful assignment of peaks in all samples is
needed if the quantitative data are statistically analysed,
e.g.
with principal
component analysis or building a model to relate the instrumental data to
other coffee properties.
The composition of green coffee volatiles is very different from that
of roasted coffee. One study has reported 12 alcohols, 6 acids, 6 esters, 5
ketones, 4 pyrazines, 3 furans and 3 aldehydes detected in coffee beans from
3 farms in a single region.60 Another study reported 68 gC peaks, detected
in green coffee samples from a single farm (2 different varieties), with 48 of
those identified by comparing to a library of gC-MS spectra.62 differences
are seen when comparing these compounds to roasted coffee aroma, since
green coffee does not contain any of the important coffee aroma compounds,
although similar groups of compounds can be observed. Indeed, most of the
important coffee aroma compounds are formed during thermal treatment
(roasting) of coffee and their content depends critically on the time tempera-
ture roast profile and roast degree. The groups that are not present in green
coffee, but which are important for roasted coffee aroma, are furanones, phe-
nols and sulfur-containing compounds.
The mass peaks and corresponding tentatively identified compounds in
22 green coffee samples of our study are presented in Table 33.1. The results
show that small unsaturated hydrocarbons, alcohols and aldehydes were
common to all samples analysed by dynamic headspace pTr-ToF-MS. The
detected intensities of dynamic headspace signals were highly reproducible
for repetitive analysis of a green coffee sample, but were strongly variable
between samples. Using principle component analysis, we were able to dif-
ferentiate all of the samples, in spite of the fact that the number of mass
peaks included in the analysis was low.22
The key result of the pTr-ToF-MS profiling of green coffee volatiles shows
that there is a very strong diversity and differences in volatile intensities
3 8
2 5
2 7
4 7
4 9
5 9
736
Chapter 33
Table 33.1 list of mass peaks detected in all 22 evaluated coffee samples (12 ara-
bica and 10 robusta) applying dynamic headspace of green whole beans
with pTr-ToF-MS for volatile analysis. The mass accuracy in ppm is
presented as standard deviation of a total of 100 measurements. Com-
pounds were tentatively identified based on molecular formula and
information obtained by gC-MS.60,62
exact
molecular mass
(protonated)/
m
/
z
Mass accuracy/
ppm
Molecular
formula
Compound name (tentative
identification)
31.018 3.4 Ch3O+ Formaldehyde
33.034 2.6 Ch
5
O
+
Methanol
39.032 6.2
Not identified Not identified
41.039 1.0 C h
+
Cyclopropane or propadiene
3 6
43.054 12.1 C h
+
propene or cyclopropane
45.034 0.8 C h O
+
Acetaldehyde
47.050 3.4 C h O
+
ethanol
51.044 1.6 C h O
+
ethylene glycol
2 6 2
55.054 21.8 C h
+
Unsaturated hydrocarbon
57.035 7.0 C h O
+
Acrolein or cyclopropanone
3 5
57.071 1.2 C h
+
Unsaturated hydrocarbon
59.050 1.2 C h O
+
Acetone, propanal
3 7
69.071 3.8 C h
+
Unsaturated hydrocarbon
71.086 13.2 C h
+
Unsaturated hydrocarbon
5 11
83.086 9.3 C h
+
Unsaturated hydrocarbon
87.044 17.8
6 11
+ butyrolactone
C
4
h
7
O
2
101.059 19.5 C h O
+
Not identified
5 9 2
when comparing coffees that do not come from the same farm or region.
While the overall intensity of the signals from the dynamic volatile profiles
is low for green coffee (compared to roasted coffee), the spectrum itself is
highly complex. In Figure 33.5, a mass section of the spectrum from two
different arabica coffees is presented, where large differences in the signal
intensities are shown for peaks having relatively strong signal intensities.
The relative simplicity of measuring volatiles in green beans (when com-
paring to the analysis of non-volatile components) opens a wide range of pos-
sible applications for rapid analysis of volatile composition of green beans,
such as controlling consistency of quality, origin of green beans, monitor-
ing deterioration and changes during storage or during shipment of the
green beans. Most importantly, green coffee VOC profiles may reveal specific
defects and quality issues that will affect the cup profile.
33.5
Roasted Coffee Aroma Compounds
Among the various sensory modalities, aroma (smell) is of paramount
importance to the quality of coffee. In order for coffee to be sensed by our
nose, aromatic VOCs are released from e.g. the brew and reach the olfactory
epithelium, a region in the upper part of the nasal cavity which contains the
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
737
Fig
ur
e
33.5
d
y
namic
p
T
r
-
T
oF-MS
mass spectra
in
the
range
of
101108
m
/
z
f
or
two arabica coffees. For three of the mass peaks shown in the mass
window, only the molecular formula could be assigned and, for one,
it was possible to tentatively identify the compound as dimethylpyri-
dine, based on identification with gC-MS.
nerve endings that allow us to smell. but what are these volatile aromatic
compounds that compose the coffee aroma?
Scientific efforts to elucidate the origin of the rich and distinctive aroma
of coffee, and ultimately to understand “What Makes that Coffee Smell so
good”,58,59 can be traced back to 1880 when bernheimer identified a few vol-
atile compounds in coffee.64 but the first significant level of progress can
most likely be attributed to reichstein and Staudinger, who, in 1926, identi-
fied and patented several important aroma active compounds in coffee.65,66
Mainly fuelled by progress in analytical techniques, in particular gas chroma-
tography (gC), the number of publications on coffee aroma and the number
of identified coffee VOCs has rapidly increased since then. Today, around
1000 VOCs have been reported in coffee, which includes compounds from
both green and roasted coffee.67
For many years, scientists have concentrated on identifying the VOCs in
coffee. however, in the 1970s it had already become clear that only a small
fraction of these volatiles perhaps 5% are odoriferous and hence relevant
to the aroma. As a result, the focus shifted towards these few sensory relevant
aroma-active compounds in the headspace of coffee (hS, the air-space above
the coffee).
Several instrumental methods have been developed to identify and quan-
tify the odour-relevant volatiles, assess their odour impact and character-
istics, and recombine coffee aromas from the identified and quantified
main aroma compounds. given the extensive differences in coffee genet-
ics, geographical origins, cultivation practices and processing techniques,
it is not surprising that publications on coffee aroma composition differ
738
Chapter 33
with respect to the relative importance given to the different aroma com-
pounds. Consequently, different studies often report slightly different lists
of VOCs responsible for the aroma of the particular coffee being studied.
The different strategies and analytical techniques used for measuring VOC
compositions are an additional source of variability in the ranking of the
main aroma compounds. hence, in order to ‘eliminate’ variability due to
differences between, for example, species, in Table 33.2 we focus on just the
Coffea arabica.
Studies by grosch and his co-workers in the mid-1990s concluded that
fewer than 30 VOCs are important for the aroma of roasted coffee.
70,72,74,9699
In follow-up omission experiments, they suggested that the actual number
of indispensable coffee aroma compounds could be as small as nine.
75,77,78,100
based on this detailed work, a condensed list of coffee aroma VOCs has been
compiled in Table 33.3. Instead of grouping compounds into chemical fam-
ilies, as in Table 33.2, compounds are grouped in Table 33.3 into sensory
families. besides concentrations in roast and ground coffee, concentrations
in the liquid coffee extract and the extraction yields have also been included.
The last column in Table 33.3 marks the 12 compounds that were considered
by grosch and his co-workers to be particularly important with their corre-
sponding structural formulas shown in Figure 33.6.
75,77,100
Omitting these
compounds from a coffee aroma model (individually or as groups) leads to a
significant difference in the coffee aroma profile.
33.6
Analytical Techniques for Coffee Aroma
Analysis
33.6.1
Gas Chromatography
The “work-horse” for coffee aroma analysis has been gas chromatography
(gC). Over the past 30 years it has led to the elucidation, characterization and
quantification of compounds that are relevant to coffee aroma. One potential
weakness of such an approach is that it is slow, making it unsuitable for mon-
itoring fast processes in real time. Yet, in spite of these and other drawbacks,
gC analysis is still considered an excellent approach for understanding and
reconstituting the flavour of coffee. The ultimate proof that we “understand”
the aroma of coffee is when we are able to reconstitute it.77,100
33.6.2
Olfactometry When the Human Nose Becomes a
Detector
Many aroma active compounds in coffee only appear in the hS at very low
concentrations, some of which can hardly be detected by instrumental
detectors. The only detector capable of sensing these highly potent coffee
aroma compounds is our nose. It is of course also the most relevant detec-
tor, when it comes to aroma compounds. hence, aroma chemists use their
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
739
nose (or a panel of “sniffers”) to detect compounds that are eluted at the
end of the gC column. Trained sniffers also provide a sensory description
of the compound and rate the perceived intensity. In combination with the
chromatographic retention time and mass spectral profile, the aroma note
often allows the compound to be unambiguously identified. Techniques
which combine separation by gas chromatography and the human nose as
a detector are called gC-Olfactometry gC-O103 and are explained in the
following section.
Two major gC-O screening techniques have been developed: one by grosch
and his co-workers (called the Aroma extract dilution Analysis)
78,104
and the
other
b
y
A
cree
and
his
co
-w
ork
ers (called
C
h
A
r
M
analysis).
105–113 These
techniques are shown schematically in Figure 33.7. both evaluate a dilution
series of an original aroma extract using gC-O. The occurrence of an aroma in
each dilution is noted. As the dilution increases, the compounds with lower
odour potency successively cease to be sensed, and only the most potent are
detected at higher dilutions. The number of times each odorant was sensed
across the dilutions is then summed up. The greater the number of dilutions
in which an odorant is sensed, the higher is the odour potency in both AedA
and
C
h
A
r
M
analysis.
This leads to plot
s
of
dil
ution
or
C
h
A
r
M
val
ues.
b
oth
A
ed
A
and
C
h
A
r
M
analysis
origi
nally
proposed
that
the
l
arger
the
dil
ution
or
C
h
A
r
M
val
ue,
the
more
i
mpor
t
ant
the
contribution of
the
respectiv
e aroma
compound to the overall aroma. While this interpretation has evolved, both
techniques are still widely used to estimate the relative contribution of var-
ious VOCs to the aroma of coffee.
4,106,108,109,111, 114
An alternative technique,
called gC-SnIF, was introduced by Chaintreau and his co-workers.
115
In this
method, the intensities of the aromagram peaks were based on the detection
frequencies of the odorants perceived by a panel of assessors at the sniffing
port of a gC-O. This approach allowed standard deviations to be calculated
and hence led to more quantitative data-analysis than previous gC-O meth-
ods. For a review of gC-O, with a specific discussion of coffee, please refer to
ref. 116.
An interesting application of gC-O was the identification of the chem-
icals responsible for the mouldy/earthy off-notes found in lots of green
Mexican coffees.4,114 gC-profiles obtained from a reference sample (a
sample free of off-notes, assessed by expert tasters) and a defective
mouldy sample, showed minor differences and no indication was found
as to what the compounds responsible for the off-note could be. The same
samples were then subjected to a gC-O sniffing analysis. This led to the
recognition of several differences between the extracts, which could be
related to the off-flavour and resulted in the identification of a series of
compounds that caused the off-flavour (Figure 33.8). In particular, peaks
that were related to earthy, green, chemical and mouldy attributes could
be identified as 2-methyl isoborneol, 2,4,6-trichloroanisole, geosmin and
various pyrazines. The chemical identification allowed the development
of a hypothesis concerning the source of the off-flavour and, subsequently,
propose mitigation strategies.
Table 33.2 Flavour active VOCs in arabica coffee (adapted from ref. 68). The first column lists 72 VOCs, classified into 15 different groups
of chemical compounds. data compiled from ref. 27, 40 and 6995.
key odorant CAS Id#
Aldehyde
Concentration
a
(ppb) Aroma descriptor
Sensory
thresholdb (ppb)
2-Methylbutanal 96-17-3 20700 rancid, almond-like, toasty 1.3
2-
Methylpropanal 78-84-2 Toasty, caprylic, cheesy, dark chocolate,
ethereal, fruity, malty, pungent
3-
Methylbutanal 590-86-3 18600 Fruity, almond-like, toasty, ethereal,
chocolaty, peachy, fatty
0.35
(E)-2-nonenal 18829-
56-6
19 Fatty, green, cucumber, citrus 0.08
Acetaldehyde 75-07-0 139000 pungent, ethereal, fresh, lifting, penetrating, fruity,
musty
4-
Methoxybenzaldehyde 123-11-5 Sweet, powdery, vanilla, anise, woody,
coumarin, creamy
0.7
27
Ester
ethyl-2-methylbutyrate 7452-79-1 3.9 Fruity, berry 0.5
ethyl-3-methylbutyrate 117442-
70-3
14 Fruity 0.6
Furan
Furfural 98-01-1 588019370 Sweet, brown, woody, bready, caramellic 280
2-((Methylthio)methyl)
furan/2-furfuryl methyl sulfide
2-Furanmethanol acetate/fur-
fulyl acetate
1438-91-1 Smoke, roast, onion, garlic, sulfuraceous, pungent,
vegetable, horseradish
623-17-6 2452040040 Onion, garlic, sulfury, pungent, vegetable,
horseradish
2-Methyl furan 534-22-5 burnt, ethereal (mild), gasoline, acetone, chocolate
740
Chapter 33
phenylacetaldehyde
122-78-1
Sweet, fruity, honey, floral, fermented
propanal
123-38-6
17400
ethereal, pungent, earthy, alcoholic
10
Acid
2-Methylbutyric acid
116-53-0
25000
Acidic, fruity, dirty, cheese
10
3-Methylbutyric acid
503-74-2
1806032180
Cheesy, dairy, acidic, sour, pungent, fruity, stinky
700
5-Methyl-2-furancarboxyalde-
hyde/2-methyl furfural
620-02-0 Sweet, caramellic, bready, brown,
coffee-like
6000
Furfuryl formate 13493-
97-5
Furfuryl methyl ether 13679-
46-4
ethereal
roasted coffee
Furfuryl disulfide 4437-20-1 Sulfury, coffee, roasted chicken, meaty, onion,
cabbage
Sulfur-containing compounds
75-2
Methional 3268-49-3 213240 boiled potato-like, musty, tomato, earthy,
vegetable, creamy
0.2
Thiols
3-Mercapto-3-methyl butyl
formate
50746-
10-6
130 green blackcurrant, herbal, fruity, roasted, sweaty 0.0035
Thiophene
3-Methylthiophene 616-44-4 Fatty, winey
Thiazole
2,4-dimethyl-5-ethylthiazole 38205-
61-7
nutty, roasty, meaty, earthy
Furanone
dihydro-2-methyl-3(2H)-furanone 3188-00-9 758030000 Sweet, bread, buttery, nutty 0.005
(
continued
)
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
741
dimethyl trisulfide
3658-80-8
28
Sulfurous, cooked onion, savory, meaty, cabbage-like
0.001
bis(2-methyl-3-furyl)disulfide
28588-
Meaty, roasted scallion, onion, sulfury
0.00076
2-Furfurylthiol
98-02-2
10805080
r
oast
y
(coffee-lik
e), su
lfurous
0.01
2-Methyl-3-furanthiol
28588-
6068
Sulfury, meaty, fishy, metallic, boiled
0.007
74-1
3-Mercapto-3-methylbutylacetate
50746-
7.5
r
oast
y
,
fr
uit
y
,
sulfurous, s
w
eet
09-3
3-Methyl-2-butene-1-thiol
5287-45-6
13
Sulfury, smoky, leeky, onion
0.0003
Methanethiol
74-93-1
4550
rotten eggs, meat or fish, cabbage, garlic, cheesy
0.02
Table 33.2 (continued)
key odorant CAS Id#
Concentration
a
(ppb) Aroma descriptor
Sensory
thresholdb (ppb)
2-ethyl-4-hydroxy-5-meth-
yl-3(2H)-furanone
(homofuraneol)
27538-
10-9
16800 Sweet, caramel, candy 20
3-hydroxy-4,5-dimethyl-2(5H)-fu- 28664- 1.111470 extremely sweet, strong caramel, maple, burnt sugar, 20
ranone (sotolone) 35-9 coffee
4-
hydroxy-2,5-dimethyl-3(2H)-fu- 3658-77-3 10930112000 Sweet, candy, caramel, strawberry, sugar 10
ranone (furaneol)
5-
ethyl-3-hydroxy-4-methyl-
2(5H)-furanone (abhexon)
5-ethyl-4-hydroxy-2-methyl-
3(2H)-furanone
Ketone
144831-
60-7
27538-
09-6
104160 Seasoning-like, caramel-like 7.5
17300 Sweet, caramel, bready, maple, brown sugar, burnt 1.15
1-
Octen-3-one 4312-99-6 herbal, mushroom, earthy, musty, dirty 0.0036
2,3-hexadione 3848-24-6 burnt, buttery, caramel, chocolate cream, creamy,
fruity, oily, pear, sweet
2,3-butanedione 431-03-8 4840050800 buttery, creamy, fatty, oily, sweet, vanilla 0.3
2,3-pentanedione 600-14-6 354039600 buttery, caramel, creamy, penetrating, sweet 20
4-(4-hydroxyphenyl)-2-butanone 5471-51-2 1 Sweet, fruity, berry, jam, raspberry, ripe, floral (rasp-
berry ketone)
1–10
1-(2-Furanyl)-2-butanone 4208-63-3 rummy
Norisoprenoid
(E)-β-damascenone 23726-
93-4
195255 honey-like, fruity, apple, rose, honey, tobacco, sweet 0.00075
Phenolic compounds
742
Chapter 33
guaiacol
90-05-1
200017970
phenolic, burnt, smoke, spice, vanilla, woody
2.5
4-ethyl guaiacol
2785-89-9
80024800
Spicy, smoky, bacon, phenolic, clove
25
4-Vinyl guaiacol
7786-61-0
800064800
Spicy, dry woody, fresh amber, cedar, roasted peanut
0.75
Vanillin
121-33-5
22904800
Sweet, vanilla, creamy
25
Pyrazine
2,3-diethyl-5-methylpyrazine
(hazulnut pyrazine)
18138-
04-0
medical
7395 nutty-roast, musty, meaty, vegetable, roasted
hazelnut
0.09
2-
ethyl-3,5-dimethylpyrazine 27043-
05-6
55330 nutty-roast 0.04
2-ethyl-3,6-dimethyl-pyrazine
(3,6-cocoa pyrazine)
2-Methoxy-3,5-dimethylpyrazine
(3,5-cocoa pyrazine)
2-Methoxy-3,2-methylpropylpyr-
azine
13360-
65-1
13925-
07-0
24683-
00-9
2570–5980 potato, cocoa, roasty, nutty 8.6
1.1 earthy, burnt, almonds, roasted nuts, coffee 0.006
green, pea green, bell pepper
2-
Methoxy-3-isopropylpyrazine
25773-
40-4
2.4 earthy, pea, beany 0.002
3-
ethenyl-2-ethyl-5-methylpyr-
azine
6,7-dihydro-5-methyl-5
H
-cyclo-
pentapyrazine
181589-
32-2
23747-
48-0
earthy
nutty-roast, earthy, baked potato, peanut, roasted
ethylpyrazine 13925-
00-3
peanut butter, musty, nutty, woody, roasted cocoa 4000
Pyridine
pyridine 110-86-1 2128065520 Fishy 77
pyrrole 109-97-7 Sweet, nutty, ethereal
1-Methyl pyrrole 96-54-8 Smoky, woody, herbal, negative notesdefective
beans
Terpene
aThe second column shows the range of reported concentrations in ppb (µg g1) in roasted arabica coffee. For some VOCs no information on concentration
was reported.
bIn cases where different sensory detection thresholds were published, the one included in the table corresponds to the lowest one reported.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
743
2,3-dimethylpyrazine
5910-89-4
25806100
nutty, coffee, peanut butter, walnut, caramel, leather
800
2,5-dimethylpyrazine
123-32-0
455011730
Cocoa, roasted nuts, roast beef, woody, grass,
80
linalool
78-70-6
Flowery, citrus, orange, terpy, waxy, rose
0.17
limonene
138-86-3
Citrus, herbal, terpene, camphor
4
geraniol
106-24-1
Sweet, floral, fruity, rose, waxy, citrus
1.1
Table 33.3 Quantitative composition of roasted coffee, ground coffee and coffee extract/beverage, for the 28 most potent coffee aroma
compounds, determined for a medium-roasted columbian arabica coffee. (a) Mass-fraction of the aroma compound in roast
and
ground
coffee
po
w
der
.
101,102
(b)
C
oncentration of aroma compounds
in
the
extract/be
v
erage
(54
g
of
r
&
g
extracted
with
1
l
of hot water).74,101 (c) Mass-fraction of aroma compounds in the extract/beverage (calculated from ref. 74 and 101; brew ratio
=
18.5
:
1
l
water
to
54
g
r
&
g
).
(d)
e
xtraction
yie
ld
.
101
(e)
C
ompounds that caused a perce
iv
ed
(significant)
c
hange of aroma,
when omitted from the aroma model.75 The numbers in brackets behind the compound names correspond to the numbers
of the structures shown in Figure 33.6.
powder/µg
Concentration in
Yield per mass of
coffee used/(µg
extraction-
none
744
Chapter 33
Compound name
kg1 (a) bre
w/(µg
l
1
)
(b)
kg1) (c) yield/% (d) Aroma model (e)
Sweet, caramel notes
2-Methylpropanal (1)
24000
760
14000
59
x
2-Methylbutanal (2)
26000
870
16000
62
x
3-Methylbutanal (3)
17000
570
11000
62
x
2,3-butandione
49000
2100
39000
79
2,3-pentandione
35000
1600
30000
85
4-hydroxy-2,5-dimethyl-3(2H)-furanone
140000
7200
130000
95
2(5)-ethyl-4-hydroxy-5(2)-methyl-3(2H)-fura-
16000
800
15000
93
Vanillin
4100
210
3900
95
Earthy notes
2-ethyl-3,5-dimethylpyrazine (4)
400
17
310
79
x
2-ethenyl-3,5-dimethylpyrazine (5)
53
1
19
35
x
2,3-diethyl-5-methylpyrazine (6)
100
3.6
67
67
x
2-ethenyl-3-ethyl-5-methylpyrazine
15
0.2
4
25
3-Isobutyl-2-methoxypyrazine (7)
120
1.5
28
23
x
Roasty, sulfury notes
sotolone
5-ethyl-3-hydroxy-4-methyl-2(5
H
)-furanone/
furaneol (12)
104
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
745
2-Furfurylthiol (8)
1700
17
320
19
x
2-Methyl-3-furanthiol
60
1.1
20
34
Methional
250
10
185
74
3-Mercapto-3-methylbutyl format
130
5.7
105
81
3-Methyl-2-buten-1-thiol
13
0.6
11
85
Methanthiol
4400
170
3100
72
dimethyl trisulfide
28
Phenolic
guaiacol
2400
120
2200
73
x
4-ethylguaiacol
1800
48
900
49
4-Vinylguaiacol (9)
45000
740
14000
30
Fruity, flowery
Acetaldehyde (10)
120000
4700
87000
73
x
propanal (11)
17400
x
(E)-β-damascenone
260
1.6
30
11
Sharp
3-hydroxy-4,5-dimethyl-2(5H)-furanone/
1900
80
1500
78
x
746
Chapter 33
Figure 33.6 Twelve compounds that are considered of particular importance to
the aroma of coffee. The numbers in brackets correspond to the num-
bers behind the names in Table 33.2/first column. Omitting these
compounds, individually or as part of compound groups, from a cof-
fee aroma model, leads to a significant difference in the aroma profile.
They are therefore considered to be of particular importance to the
aroma of coffee.75
Figure 33.7 Schematic description of gas Chromatography-Olfactometry, gC-O.
Aroma
extract
dil
ution
analysis
(A
ed
A)
and
C
h
A
r
M
analysis
are the
two most common realizations of gC-O.
The example in Figure 33.8 demonstrates some of the strengths of gC-O:
the identification of VOCs responsible for olfactory off-notes in a defective
sample. It compensates the lack of sensitivity to low concentration flavour
active compounds of other detection systems. In this study, it was clear that
instrumental detection failed to recognize the defects documented in the
sensory profile.4,114
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
747
Figure 33.8 Comparison of a gC chromatogram on the top with the gC-O profile
analysed by one sniffer on the bottom for a sample with a mouldy/
earthy defect. All sniffed signals without sensory descriptors represent
signals that are typical to coffee and are not related to the off-note.
33.7
Trends and New Developments in Coffee
Aroma Analysis
Over the past 30 years, our understanding of the aroma of coffee has grown
steadily. Today, we believe that the list of aroma active compounds is essen-
tially complete. nevertheless, research into coffee aroma is still in its infancy.
Major efforts are currently being made to develop a range of novel analytical
technologies and strategies. We would like to highlight three major trends
that we believe will shape the future of coffee aroma research. These are (i)
development of time-resolved analytical technologies (Section 33.7.1), (ii)
progress in individualized aroma science with the development of analytical
techniques that allow the capturing of such differences and, consequently,
the developing of a better understanding of inter-individual difference in
aroma release, sensation and perception (Section 33.7.2) and (iii) mathemat-
ical/statistical models to predict sensory profiles from instrumental mea-
surements (Section 33.7.3).
33.7.1
Time-resolved Analytical Techniques
A major new development in analytical technologies represents the time-re-
solved methods based on direct injection mass spectrometry using optical
and laser ionization
117123
and chemical ionization.
10,56,124138
here, we will
discuss two applications of time-resolved approaches, one on coffee roasting
748
Chapter 33
and one on extraction. both are dynamic processes requiring a time resolu-
tion of approximately one second.
33.7.2
Analysis of Aroma Formation During Roasting
The crucial step in creating the coffee aroma is roasting. One possibility
for studying the formation of VOCs during roasting is to take samples at
different times during the roasting process and analyse these off-line by gas
chromatography.101,139–143 however, these techniques are not only time-con-
suming, but often require complex sample preparation processes before
analysis, with the additional risk that irregularities in these processes might
affect the outcome of the analysis. In contrast, on-line measurements of
the roaster off-gas provide a direct insight into the dynamics of VOC forma-
tion in real time, are very sensitive and do not require sample preparation
(hence avoiding potential distortion of information). An already well-proven
technology for on-line analysis of coffee roasting is proton-transfer-reac-
tion mass spectrometry (pTr-MS).53–55,125,144 Furthermore, alternative tech-
niques based on resonant laser ionization coupled to time-of-flight mass
analysis117,121–123 or ion trap mass spectrometry145 have also been applied to
explore the coffee roasting process. More recently, with single-photon ion-
ization time-of-flight mass spectrometry (SpI-ToF-MS), single bean roasting
has been performed on arabica as well as on robusta beans.123 Combining
soft ionization via protontransfer reaction with high mass resolution of
a time-of-flight instrument (pTr-ToF-MS)146 provides the advantage of a
fast analytical technique to record information about VOCs formed during
roasting in just one single mass spectrum. On-line analysis of coffee roast-
ing with pTr-ToF-MS allows the monitoring of the formation of VOCs in
real-time.136,138,147,148
Two examples for on-line analysis by pTr-MS of roaster off-gas during cof-
fee roasting are provided below. The first example represents a Colombian
arabica coffee which was roasted using different roaster gas temperatures in
order to achieve the same dark roast degree (CTn 67 on the neuhaus scale).
The temperature of the roaster gas was set to 228 °C, 238 °C, 248 °C and 258
°C (isothermal roasting) for four roasting trials and more than 50 VOCs were
measured simultaneously during each trial. Figure 33.9 shows the experi-
mental setup, while Figure 33.10 shows the timeintensity profiles for two
selected compoundsfurfural and 5-methylfurfuralduring the four differ-
ent isothermal roasting trials. While the coffee reaches the same dark roast
degree of CTn 67 in all four trails, the roasting times differed. At 228 °C (Fig-
ure 33.10, green line), the roasting time corresponded to 25 minutes (1500
seconds). In contrast, roasting the coffee at 238 °C, 248 °C and 258 °C (in the
black, pink and blue traces) meant that the time taken to reach CTn 67 was
reduced to 14 min, 10 min and 7.5 min. respectively. besides the obvious
reduction in roasting time with increased temperature, the timeintensity
profiles for the two selected compounds differed in two major aspects: (i)
a strong increase in the intensity of the VOCs in the off-gas with increasing
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
749
Figure 33.9 experimental setup for on-line measurements of the roaster off-gas
during roasting on a 200 g per batch fluidized bed sample roaster from
neuhaus neotec. The gas from the roaster off-gas is sampled through
a dust filter and diluted with nitrogen gas to reduce the temperature
and humidity of the gas. All sampling lines are heated to 50 °C to pre-
vent condensation. While most of the sampled gas is pumped through
the Mass Flow Controller (MFC) to the exhaust of the pump, a small
fraction is sampled via the air inlet into the drift tube of the pTr-MS.
In the drift-tube, VOCs are ionized by proton transfer from h3O+ and
mass analysed by a quadruple mass filter.
Figure 33.10 The pTr-MS ion traces that are shown were monitored on-line and in
real-time in the off-gas of a fluidized bed roaster (200 g batch size). Two
compounds are shown (tentative assignment) during four different
roasting trials for different roaster gas temperatures; (a) shows the trace
for the compound furfural, (b) shows the compound 5-methyl-furfural.
750
Chapter 33
Figure 33.11 Three different arabica coffees were roasted to medium roast degree
(103 pt on the probat scale). Coffees were roasted either by applying
a lower roasting temperature (lower burner intensity), or by apply-
ing a higher roasting temperature (medium burner intensity). The
roasting process can be visualized by following the formation of the
two volatiles, which peak at different time points. On the one hand,
there is a difference in roasting time for the three coffees, where cof-
fee from guatemala is roasted the fastest and followed by the coffee
from Colombia and Yirga Cheffe. On the other hand, lower burner
capacity prolongs the roasting time for all coffees and the difference
in roast time is no longer visible between the three coffees. (a) shows
the trace for a compound with mass formula [Ch3O2]+, which is ten-
tatively assigned to formic acid. (b) on the right shows the compound
with mass formula [Ch5O]+, which is tentatively assigned to methanol.
temperature, which corresponds to a concomitant increase in the com-
pound's concentration in the roasted coffee.138 Clearly, short-time high-tem-
perature (SThT) roasting generates coffee that exhibits significantly higher
aroma
intensit
y
,
in
comparison to
long-ti
me lo
w-temperat
ure
(
l
T
l
T)
roasti
ng
to the same roast degree. The impact of the timetemperature profiles (for
identical roast degree) has been confirmed in gas chromatographic138 and
sensory (unpublished) studies. (ii) Modifying the timetemperature roasting
profile may alter the dynamics of the compound formation and hence modu-
late the formation of intermediates. by exploring the formation of a range of
phenolic compounds on-line, it was possible to demonstrate the sequential
formation of compounds, from precursors via various intermediate stages to
the final VOC.123
The second example examines the roasting of three distinct single-ori-
gin arabica coffees. Figure 33.11 shows the timetemperature profiles of
two exemplary VOCs: formic acid as an example of a volatile organic acid
formed during roasting, and methanol as an example of a VOC with alco-
hol functionality. All coffees were roasted to the same medium roast degree
(103 pt Colorette 3b, probat), using either a medium or a low burner inten-
sity/temperature. The left frame shows the pTr-ToF-MS ion traces of formic
acid. roasting at medium burner intensity, the first to reach the target roast
degree was the Central American coffee from guatemala, where the peak in
formic acid intensity already happens after 10 minutes. The second fastest
coffee was the one from South America (Colombia), while the Yirga Cheffe
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
751
took the longest time with formic acid peaking only after 12 minutes. In a
subsequent set of experiments, the same coffees were roasted at low burner
intensity. besides extending the roasting times in general, the first to reach
the target roast degree of 103 pt was now the Yirga Cheffe. These examples
demonstrate, for instance, that the aroma development strongly depends on
the timetemperature roasting and differs among arabica and robusta cof-
fee.147 In a recent publication, the implication of such observations on split
versus mixed roasting was discussed.138
33.7.3
Extraction Kinetics of Coffee Aroma Compounds
Using proton-transfer-reaction time-of-flight mass-spectrometry (pTr- ToF-
MS), we investigated the extraction dynamic of 95 ion traces in real time
(time resolution: 1 second) during espresso coffee preparation.149 Fifty-two of
these ions were tentatively identified. This was achieved by on-line sampling
of the VOCs close to the coffee extract flow, at the exit of the extraction hose of
the espresso machine (single serve capsules). The results show considerable
differences in the extraction kinetics for different compounds, which led to
an evolution of the volatile profiles in the extract flow and consequently to an
evolution of the aroma profile in the cup.
The timeintensity profiles in Figure 33.12 show different extraction
dynamics for the analyzed VOCs (left frame). Four example compounds
are shown: methylpropanal, pyridine, methyl furan and guaiacol. The time
to reach the maximum intensity ranged from 2 to 20 seconds. Once the
maximum had been reached, the intensity fell at different rates, depend-
ing on the compound. This decrease in intensity provides information on
Figure 33.12
Time intensity profiles for one specific capsule (lungo, 42 second
extraction time, 110
ml
extraction volume) showing differences in
extraction kinetics. left frame: data normalized to the maximum
intensity of each of the four VOCs. right frame: integration of the
area under the curve at each time point as a percentage of the total
area at the end of the extraction. The shaded ribbons show the 95%
confidence interval.
752
Chapter 33
the extraction kinetics of the compounds. A fast decrease implies that the
compound is already extracted over a relatively short time period, whereas
a slow decrease implies that the compound is extracted over a longer
period. The implication of such observations is that the aromatic com-
position varies as the extraction progresses, with some compounds being
fully extracted within the first few seconds of the extraction while other
compounds extract much slower and are starting to dominate the extract
at prolonged extractions. Using this analytical approach, we are currently
exploring the impact of extraction pressure, the temperature and the min-
eral content of the extraction water.
This technique was also applied to study the time-resolved extraction
dynamics with a professional machine (with porta filter) and is currently
being applied to a fully automatic machine extraction. Such studies allow
both a novel and detailed insight to be obtained into the extraction process
and an understanding of how the various instrumental (e.g. temperature and
mineral composition of the extraction water; pressure of extraction; shape
of porta filter, pre-infusion, …) and coffee (variety, roast degree, grind, …)
parameters and the brew ratio affect the extraction process.
33.7.4
Moving Towards an Individualized Aroma Science – In-
mouth Coffee Aroma
by looking at the actual flavour experience in more detail that occurs whilst
coffee is being drunk, it becomes clear that it is a highly dynamic experi-
ence, quickly changing and evolving in the mouth. In order to develop a bet-
ter understanding of the aroma perceived by a consumer, it is important to
develop techniques that capture the temporal evolution of the aroma during
the actual process of consumption.
10,124,150–154
Coffee aroma evolves in the mouth during drinking and leaves a typical
after-odour, or finish, in the mouth for several minutes after swallowingwe
have termed this dynamic evolution of the in-mouth aroma “the melody of
coffee”.10,21 The nose-space technique allows these dynamic processes to be
visualized, and provides a vivid insight into aroma release and its temporal
evolution in the mouth.
Figure 33.13 shows the nose-space profiles from four different assessors
(four male assessors aged 3343 and labelled A to d) when drinking the same
coffee.10 The nose-space profile is the time-resolved representation of the
intensity of volatile aroma compounds in the air exhaled through the nose.
Maxima represent exhalation while minima are periods of inhalation. The
time axis (in seconds) runs through all four experiments without interrup-
tion and shows the relative time of analysis for the four assessors. Five dif-
ferent mass signals are shown, differentiated by the colour. All experiments
were conducted following a strict protocol. The assessor introduced the two
ends of the nose-piece into his two nostrils. Initially, and during three com-
plete breathing cycles, the assessor had no coffee in his mouth; the pTr-MS
signals during this phase represent VOC that naturally occur in the assessor's
Figure 33.13 Top frame: schematic of the nosepiece (exhaled nose air sampler) for
sampling breath-by-breath the air exhaled through the nose during
drinking coffee. The nosepiece is connected to a pTr-MS, for on-line
monitoring of the concentration of volatile organic compounds in
the exhaled breath air. bottom four frames: nose-space profile of
four different assessors (A, b, C, d) while drinking an espresso cof-
fee. Five different pTr-MS ion-mass trace intensity in the exhaled air
are shown. All four assessors drank the same coffee following the
identical breathing and drinking protocol, while their breath-air
exhaled through the nostrils was measured on-line by pTr-MS. Four
characteristic phases of the consumption process are outlined. (1)
The three first exhalations were done with the nosepiece connected
to the two nostrils while the assessors had no coffee in their mouths.
Very low signal intensities are observed in the exhaled air. before the
fourth exhalation, the assessors take a sip of coffee. (2) “exhalation
just after first sip” corresponds to the nose-space profiles of the exha-
lation just after taking coffee into the mouth. during the subsequent
two exhalations, the assessors keep the coffee in their mouths. (3)
“Swallow-breath” corresponds to the nose-space exhalation follow-
ing just after swallowing of the coffee. “Finish” corresponds to the
sequence of nose-space profiles after swallowing when the assessors
no longer
hav
e
any
coffee left
in
the
i
r
mouths.
Fiv
e
m
/
z
ion traces are
sho
w
n
and
q
ualified
b
y
the
i
r
m
/
z
val
ues. The
y
hav
e tent
ativ
e
ly
been
assigned as:
m
/
z
73:
2-butanone;
m
/
z
75:
methyl
acetate;
m
/
z
81:
p
y
ra
-
zi
ne
or/and
a fragment of
furur
yl
alcohol
;
m
/
z
83:
2-methylfuran;
m
/
z
87: 2-,3-methylbutanal (57%); diacetyl (43%).10 The time resolution
used for recording the nose-space spectra was 0.5 s.
754
Chapter 33
breathed air (background). After the third exhalation, assessors take a sip
of approximately 10
ml
coffee. The first exhalation just after the first sip
showed high intensities for assessors A and b, while assessor C had very low
intensities. This indicated that significant inter-individual differences occur
in the transfer of VOC from the oral cavity to the nose space (nasopharynx).
possible explanations for such variations are difference (i) in the release of
VOC in the mouth due to differences in mouth flora or temperature, (ii) in
the opening of the velum (uvula) which connects the mouth cavity to the
nasopharynx or (iii) in air flows and breathing patterns. After three breath-
ing cycles with coffee in the mouth, assessors swallow the coffee. The first
exhalation just following the swallowing is termed the swallow breath and in
general shows relatively high intensities. This is explained by the wetting of
the trachea with coffee which releases VOC into the airflow of the first exhala-
tion. For some assessors (e.g. C and d), the swallow breath shows the highest
intensities in VOC of all exhalations. After the swallow breath, assessors have
no coffee in their mouths but still show significant intensities in the exhaled
air that slowly decreases in intensity over time. This is often called the after-
odour or finish.
Assessor A drank the coffee first, from seconds 770 to 970; assessor b
drank the coffee between seconds 2430 to 2630, and so forth. It is obvious
that the four nose-space profiles are different. The main differences pertain
to: (i) the absolute intensities of the three characteristic moments during the
in-mouth consumption experiencethe first sip, the swallow-breath and the
after-odour or finish and (ii) the relative intensities of the individual VOCs
(the balance or VOC profile) for each of these characteristic moments. For
assessor C, the coffee aroma is essentially composed of the swallow-breath
aroma and the persisting after-odour. Comparing assessors A and C with b
and d, we notice that their first sip and swallow-breath aromas are different
in terms of aroma balance.
The nose-space experiment demonstrates that the coffee aroma that
reaches the olfactory receptors can vary over a large range between asses-
sors, even if exactly the same coffee is drunk. The fraction and composition
of VOCs that are actually released from a substance in the mouth and trans-
ported to the olfactory receptors depends not just on the composition of
the substance, but is also strongly modulated by anatomic and physiologic
characteristics of the person, and may be further modulated by the person's
consumption and breathing habits. This is a demonstration of what we may
all already know a given coffee or given food does not taste the same to
everybody.
Analyses of mouth- and nose-spaces have been used abundantly in research
over the past few decades for medical purposes as well as for sensory analy-
sis. All of these sampling methods have been using the analysis of the respi-
ratory flow to determine the content of volatiles present in the lungs, the
mouth or the nasal cavity. With regard to sensory analysis, this means that
the compounds of interest that are accumulating in the mouth and nasal
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
755
Figure 33.14 Sampling setup with direct mouth-space sampling setup.
cavity become diluted by the respiratory flow, leading to low signal inten-
sities and that fact that only a limited number of flavour active compounds
can be monitored. hence, one aspect of the research in our laboratory is to
increase the signal intensity from VOCs present in the mouth cavity.
To this end we have tested, compared and optimized three different sam-
pling methods: (i) nose-space sampling of the exhalation stream (with nose-
piece). (ii) Mouth-space sampling of the exhalation stream, using a buffered
end-Tidal breath sampling inlet (beT) by Ionicon. (iii) direct mouth-space:
drawing sample gas directly from the mouth cavity while breathing through
the nose mouthpiece from beT (non-re-breathing mouthpieces) coupled
to a sampling and dilution lance155 a schematic of the setup is shown in
Figure 33.14.
The measurements of seven different coffees have revealed the same order
of magnitude in differences between the three sampling methods. Figure
33.15 shows two series of graphs. The upper series depicts the intensities
f
or
m
/
z
153.055
and
m
/
z
153.091,
tent
ativ
e
ly
identified as
vanilli
n
and
ethyl
-
guaiacol respectively, which represent signals with the lowest intensity that
were still distinguishable from the background. The lower series shows the
intensities
of
m
/
z
73.065,
m
/
z
87.080
and
m
/
z
101.060,
tent
ativ
e
ly
identified
as 2-methylpropanal, 2-/3-methylbutanal and 2,3-pentadione, respectively,
which represent signals with high intensities.
In summary, the signal area of our new direct mouth-space sampling
method was at least a factor of five higher in intensity, compared to the indi-
rect mouth-space sampling and a factor of at least 20 higher compared to
nose-space sampling.
756
Chapter 33
Figure 33.15 Comparison of signal intensities of low intensity compounds (to the
left) and high intensity compounds (to the right) for three different
sampling methods: upmost panels: nose-space; middle panels: beT
mouth-space; lowest panels: direct mouth-space.
The results from the comparison of the three different sampling meth-
ods have shown a massive improvement in signal intensity of the newly
developed direct mouth-space sampling compared to the other two (con-
ventional) sampling methods. The improvement in signal intensity can
be directly linked to the fact that this approach uses a direct sampling of
the volatiles present in the mouth-cavity instead of indirectly sampling
of these in the exhalation stream. It therefore enables the sensitivity of
the measurement to be increased under otherwise equal conditions such
as time resolution and a given measurement setup. Thereby, the direct
mouth-space method enables a more sensitive dynamic measurement of
volatiles during the aftertaste or lingering phase following the ingestion
of food or drinks.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
757
Figure 33.16 predicting sensory profiles from instrumental measurements: a
three-step approach.
33.7.5
Predicting Sensory Profile From Instrumental
Measurements
Improved strategies and methods for the correlation of sensory and instru-
mental analysis are being developed with the ultimate objective of pre-
dicting the sensory profile from instrumental measurements. While this
represents a truly challenging endeavour, it is often considered the holy
grail of flavour science.132 –134
The strategy presented in ref. 132 can be schematically described as a three-
step process, as outlined in Figure 33.16. In the first step, a range of coffees
are analysed by instrumental techniques (e.g. pTr-MS and/or gC-MS) and are
assessed by a trained sensory panel. The second step concerns the develop-
ment of a mathematical/statistical model that predicts the sensory profiles
based on measured instrumental data. Thirdly, in order to validate the pre-
dictive model, a series of coffees were measured by instrumental methods,
and the sensory profiles predicted based on the model were developed in
step 2. Subsequently, the same coffees were profiled by the sensory panel and
the profiles compared to the predicted ones. If the match was considered sat-
isfactory, the model was successfully validated and can be applied to predict
the sensory profiles of coffee, based on measured instrumental data.
In Figure 33.17, the application of a model was applied to two capsules that
was developed specifically for single-serve coffees.132,133 predicted sensory
profiles were generated using a formerly established predictive model based
on pTr-ToF-MS measurements (coloured area) and then superimposed on
758
Chapter 33
Figure 33.17 Comparison of predicted sensory profiles, based on instrumental
measurements (pTr-MS) and sensory profiles for two selected cof-
fees. The coffees analysed here were two single-serve coffee capsules.
the sensory profiles of the same capsules created by a sensory panel (red
line). Clearly, a very good match was achieved.
33.8
What Next?
In the past, research in coffee aroma tended to focus on the identification,
quantification and qualification of main coffee aroma compounds, and it is
believed that essentially all relevant compounds have been identified. Con-
sequently, the focus is shifting towards new fields. We see three major trends
(among others) that we believe will dominate research in coffee aroma in the
years to come.
Technological and analytical progress in instrumentation, and on-line
techniques with high time resolution and very high sensitivity will certainly
be one of the most prominent and relevant instrumental developments.
Understanding individual coffee flavour perception and preferences is a
second major field of research that will attract significant attention. novel
tools and strategies will be developed to measure the volatile aroma com-
pounds delivered breath-by-breath to the nose at an individual level. Under-
standing the basis of the differences in aroma delivery during coffee drinking
and sensation/perception will contribute to the development of individual-
ized aroma science.
predicting the sensory profile of coffee from instrumental measurements
is possibly the most significant challenge in flavour science and will certainly
attract major attention and much effort for many more years to come.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
759
Today, flavour science is moving into a discipline which is truly multidis-
ciplinary and which requires a new breed of scientists.21 What was once the
playground of food and flavour scientists and analytical chemists is today a
complex scientific platform where experts from biology, psychophysics, psy-
chology, organic chemistry, analytics, material sciences, physics, mathemat-
ics and health meet collaboratively with food and flavour scientists.
Acknowledgements
This work would not have been possible without the help and support of
many colleagues and friends with whom we have been able to share our
fascination for our work on the science of coffee aroma. They are: Werner
lindinger, Alfons Jordan, Martin graus, Imre blank, philippe pollien, Santo
Ali, Christian lindinger, ralf zimmermann, ralph dorfner, Alexia n. glöss,
barbara Schönbächler, Flurin Wieland and many more.
References
1.
O. Monteiro da Silva and C. A. Moreira leite, Income Content of the World
Coffee Exports, Ad-minister, 2013, pp. 6982.
2.
b
.
l
e
wi
n,
d
.
g
io
vannucci
and
p
.
V
arangis, Coffee Mar
k
et
s:
New
Par
a
-
digms in Global Supply and Coffee Markets Supply and Demand
, 2004.
3.
S
.
p
onte,
The
`
l
atte
r
e
v
olution'?
r
egu
lation,
Mark
et
s
and
C
onsump
-
tion in the global Coffee Chain,
World Development
, 2002,
30
, 10991122.
4.
e. Cantergiani, h. brevard, Y. krebs, A. Feria-Morales, r. Amado and C.
Yeretzian, Characterisation of the aroma of green Mexican coffee and
identification of mouldy/earthy defect,
Eur. Food Res. Technol.
, 2001,
212
, 648657.
5.
b. A. goodman,
Coffee Consumption and Health
, nova Science publish-
ers, Inc., hauppauge, new York, 2012.
6.
S. Firestein, A code in the nose, Sci. STKE, 2004, 2004, pe15.
7.
S. Firestein, how the olfactory system makes sense of scents,
Nature
,
2001,
413
, 211–218.
8.
r.
C. Araneda, A. d. kini and S. Firestein, The molecular receptive range
of an odorant receptor,
Nat. Neurosci.
, 2000,
3
, 12481255.
9.
J. b. german, C. Yeretzian and V. b. Tolstoguzov, Olfaction, where nutri-
tion, Memory and Immunity Intersect, in
Flavours: Chemistry, Technol-
ogy and Resources
, ed. r. g. berger, Springer Verlag, berlin, germany,
2007, pp. 2541.
10.
C. Yeretzian, p. pollien, C. lindinger and S. Ali, Individualization of Fla-
vor preferences: Toward a Consumer-centric and Individualized Aroma
Science, Compr. Rev. Food Sci. Food Saf., 2004, 3, 152159.
11.
J. delwiche, The impact of perceptual interactions on perceived flavor,
Food Quality and Preference, 2004, 15, 137146.
760
Chapter 33
12.
A. buettner, Influence of human salivary enzymes on odorant concen-
tration changes occurring in vivo. 1. esters and thiols, J. Agric. Food
Chem., 2002, 50, 32833289.
13.
A. buettner and J. beauchamp, Chemical input - Sensory output: diverse
modes of physiology-flavour interaction, Food Quality and Preference,
2010, 21, 915924.
14.
A. buettner, A. beer, C. hannig and M. Settles, Observation of the
swallowing process by application of videofluoroscopy and real-time
magnetic resonance imaging-consequences for retronasal aroma stim-
ulation, Chem. Senses, 2001, 26, 12111219.
15.
A. buettner, A. beer, C. hannig, M. Settles and p. Schieberle, physiolog-
ical and analytical studies on flavor perception dynamics as induced by
the eating and swallowing process, Food Quality and Preference, 2002,
13, 497504.
16.
A. buettner, S. Otto, A. beer, M. Mestres, p. Schieberle and T. hum-
mel, dynamics of retronasal aroma perception during consumption:
Cross-linking on-line breath analysis with medico-analytical tools to
elucidate a complex process, Food Chem., 2008, 108, 12341246.
17.
A. buettner and p. Schieberle, in
Changes in the Concentrations of Key
Fruit Odorants Induced by Mastication
, ed. d. d. roberts and A. J. Taylor,
ACS, Washington, dC, 2000, pp. 8798.
18.
A. buettner and p. Schieberle, exhaled Odorant Measurement
(eXOM) A new Approach to Quantify the degree of In-Mouth
release of Food Aroma Compounds, Lebensm.-Wiss. Technol., 2000,
33, 553–559.
19.
A. buettner and p. Schieberle, Influence of mastication on the concen-
trations of aroma volatilessome aspects of flavour release and flavour
perception, Food Chem., 2000, 71, 347354.
20.
M. Mestres, n. Moran, A. Jordan and A. buettner, Aroma release and
retronasal perception during and after Consumption of Flavored Whey
protein gels with different Textures. 1. in Vivo release Analysis, J. Agric.
Food Chem., 2004, 53, 403409.
21.
I. blank, M. Wust and C. Yeretzian, expression of multidisciplinary fla-
vor science: research highlights from the 12th Weurman Symposium, J.
Agric. Food Chem., 2009, 57, 98579859.
22.
I. gierczynski, h. laboure and e. guichard, In Vivo Aroma release of
Milk gels of different hardnesses: Inter-individual differences and
Their Consequences on Aroma perception,
J. Agric. Food Chem.
, 2008,
56
, 16971703.
23.
p. Mombaerts, F. Wang, C. dulac, S. k. Chao, A. nemes, M. Mendelsohn,
J. edmondson and
r.
Axel, Visualizing an olfactory sensory map,
Cell
,
1996,
87
, 675686.
24.
p. Mombaerts, Molecular biology of odorant receptors in vertebrates,
Annu. Rev. Neurosci.
, 1999,
22
, 487–509.
25.
r.
J. Stevenson, J. prescott and
r.
A. boakes, Confusing tastes and
smells: how odours can influence the perception of sweet and sour
tastes,
Chem. Senses
, 1999,
24
, 627635.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
761
26.
J. prescott, V. Johnstone and J. Francis, Odour-taste interactions: effects
of attentional strategies during exposure,
Chem. Senses
, 2004,
29
,
331340.
27.
J. S. ribeiro, r. F. Teófilo, F. Augusto and M. M. C. Ferreira, Simulta-
neous optimization of the microextraction of coffee volatiles using
response surface methodology and principal component analysis, Che-
mom. Intell. Lab. Syst.
, 2010,
102
, 45–52.
28.
M. damm, J. Vent, M. Schmidt, p. Theissen, h. e. eckel, J. lotsch and
T. hummel, Intranasal volume and olfactory function, Chem. Senses,
2002, 27, 831839.
29.
k. zhao, p. W. Scherer, S. A. hajiloo and p. dalton, effect of Anatomy on
human nasal Air Flow and Odorant Transport patterns: Implications
for Olfaction, Chem. Senses, 2003, 29, 365379.
30.
J. T. hornung, d. J. Smith, d. b. kurtz, T. White and d. A. leopold, effect
of nasal dilaors on nasal structures, sniffing strategies, and olfactory
ability, Rhinology, 2001, 39, 8487.
31.
d. A. leopold, The relationship between nasal Anatomy and human
Olfaction, Laryngoscope, 1988, 98, 12321238.
32.
l.
M. bartoshuk, V. b. duffy and I. J. Miller, pTC/prOp tasting:
anatomy, pschophysics, and sex effects,
Physiol. Behav.
, 1992,
56
,
11651171.
33.
l.
h. Snyder, Inherited Taste deficiency,
Science
, 1931,
74
, 151152.
34.
U. k. kim, e. Jorgenson, h. Coon, M. leppert, n. risch and d.
drayna, positional cloning of the human quantitative trait locus
underlying taste sensitivity to phenylthiocarbamide,
Science
, 2003,
299
, 1221–1225.
35.
M. M. kamphuis and M. S. Westerterp-plantenga, prOp sensitivity
affects macronutrient selection, Physiol. Behav., 2003, 79, 167172.
36.
J. n. lundström, T. hummel and M. J. Olsson, Individual differences in
Sensitivity to the Odor of 4,16-Androstadien-3-one, Chem. Senses, 2003,
28, 643650.
37.
k. k. Yee and C. J. Wysocki, Odorant exposure increases olfactory sen-
sitivity: olfactory epithelium is implicated, Physiol. Behav., 2001, 72,
705711.
38.
S. Assouline, M. I. Shevell, r. J. zatorre, M. Jones-gotman, M. d. Schloss
and k. Oudjhane, Children Who Can't Smell the Coffee: Isolated Con-
genital Anosmia,
J. Child Neurol.
, 1998,
13
, 168172.
39.
n. d. Abolmaali, V. hietschold, T. J. Vogl, k.-b. hüttenbrink and T. hum-
mel, Mr evaluation in patients with Isolated Anosmia Since birth or
early Childhood,
Am. J. Neuroradiol.
, 2002,
23
, 157164.
40.
W. grosch, Specificity of the human nose in perceiving food odorants,
in
Weurman Flavour Research Symposium, Freising
, ed. p. Schieberle and
k.-h. engel, deutsche Forschungsanstalt für lebensmittelchemie, gar-
ching, Freising, 2000, pp. 213219.
41.
C.
r
e
v
erdy
,
p
.
Sc
hlich,
e
.
p
.
k
öster
,
e
.
g
inon
and
C.
l
ange,
e
ff
ect
of
sen
-
sory education on food preferences in children, Food Quality and Prefer-
ence, 2010, 21, 794804.
762
Chapter 33
42.
F. grabenhorst, e. T. rolls and A. bilderbeck, how Cognition Modulates
Affective responses to Taste and Flavor: Top-down Influences on the
Orbitofrontal and pregenual Cingulate Cortices, Cereb. Cortex, 2008,
18, 15491559.
43.
e. d. Capaldi, Conditioned food preferences, in Why We Eat what We
Eat: The Psychology of Eating, ed. e. d. Capaldi, American psychological
Association, Washington, dC, 1996.
44.
p. rozin, Towards a Psychology of Food Choice, Institut danone, brux-
elles, 1998.
45.
F. lucas and A. Sclafani, Flavor preferences conditioned by high-fat
versus high-carbohydrate diets vary as a function of session length,
Physiol. Behav., 1999, 66, 389395.
46.
C. perez, F. lucas and A. Sclafani, Increased flavor acceptance and pref-
erence conditioned by the postingestive actions of glucose, Physiol.
Behav., 1998, 64, 483492.
47.
J. A. Mennella and g. k. beauchamp, Flavor experiences during formula
feeding are related to preferences during childhood, Early Hum. Dev.,
2002, 68, 7182.
48.
A. M. d. hirsch, S. g. harris, J. Fawcett, and A. hirsch, What Flavor Is
Your Personality? Discover Who You Are by Looking at what You Eat, Source-
books Trade naperville, Illinois, 2001.
49.
J. Mojet and e. p. koster, Texture and flavour memory in foods: an inci-
dental learning experiment, Appetite, 2002, 38, 110117.
50.
C. Sulmont, S. Issanchou and e. p. koster, Selection of odorants for
memory tests on the basis of familiarity, perceived complexity, pleas-
antness, similarity and identification, Chem. Senses, 2002, 27, 307317.
51.
S. Chu and J. downes, proust nose best: odors are better cues of autobi-
ographical memory, Memory and Cognition, 2002, 30, 511518.
52.
J. Willander and M. larsson, Smell your way back to childhood: autobi-
ographical odor memory, Bull. Psychon. Soc., 2006, 13, 240244.
53.
C. Yeretzian, A. Jordan, h. brevard, and W. lindinger, in Time-Resolved
Headspace Analysis by Proton-Transfer-Reaction Mass-Spectrometry Fla-
vour Release, Washington, DC, 2000, ed. d. d. roberts and A. J. Taylor,
ACS, Washington, dC, 2000, pp. 5872.
54.
C. Yeretzian, A. Jordan, h. brevard, and W. lindinger, in On-line Moni-
toring of Coffee Roasting by Proton-Transfer-Reaction Mass-Spectrometry,
Flavour Release, Washington, DC, 2000, ed. d. d. roberts and A. J. Taylor,
ACS, Washington, dC, 2000, pp. 112123.
55.
C. Yeretzian, h. brevard, A. Jordan, A. hansel, and W. lindinger, On-line
analysis of coffee roasting by proton-transfer-reaction/mass-spectrom-
etry (pTr/MS), ed. n. M. ACS, in 218th ACS National Meeting, new Orle-
ans, 1999.
56.
F. biasioli, C. Yeretzian, F. gasperi and T. d. Märk, pTr-MS monitoring
of VOCs and bVOCs in food science and technology, Trends Anal. Chem.,
2011, 30, 968977.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
763
57.
C. r. eloy dias and C. Yeretzian, Investigating Coffee Samples by raman
Spectroscopy for Quality Control - preliminary Study, Int. J. Exp. Spectro-
scopic Tech.
, 2016,
1
, 1–5.
58.
W. grosch, Warum riecht kaffee so gut?,
Chemie in unserer Zeit
, 1996,
30, 126133.
59.
T. h. parliment and h. d. Stahl, What Makes That Coffee Smell So good,
CHEMTECH, 1995, 25, 3847.
60.
A. T. Toci and A. Farah, Volatile fingerprint of brazilian defective
coffee seeds: corroboration of potential marker compounds and
identification of new low quality indicators, Food Chem., 2014, 153,
298314.
61.
A. T. Toci and A. Farah, Volatile compounds as potential defective coffee
beans' markers, Food Chem., 2008, 108, 11331141.
62.
S. Smrke, I. kroslakova, A. n. gloess and C. Yeretzian, differentia-
tion of degrees of ripeness of Catuai and Tipica green coffee by chro-
matographical and statistical techniques, Food Chem., 2015, 174,
637–642.
63.
b
.
b
ertrand
,
r
.
b
ou
l
anger
,
S
.
d
usser
t
,
F
.
r
ibe
y
re,
l
.
b
er
thiot
,
F
.
d
escroix
and T. Joet, Climatic factors directly impact the volatile organic com-
pound fingerprint in green Arabica coffee bean as well as coffee bever-
age quality,
Food Chem.
, 2012,
135
, 25752583.
64.
O. bernheimer, zur kenntniss der röstproducte des Caffees,
Monat-
shefte für Chemie
, 1880,
1
, 456467.
65.
T. reichstein and h. Staudinger, Method for isolating the aromatic prin-
ciple contained in roasted coffee, Br. Pat., 246454, 1926.
66.
h. Staudinger and T. reichstein, producing artificial coffee oil,
Br. Pat.
260960, 1928.
67.
l.
M. nijssen, C. A. Visscher, h. Maarse,
l.
C. Willemsense, and M. h.
boelens,
Volatile Compounds in Food. TNO Nutrition and Food Research
Institute
, zeist, The netherlands, 1996, vol. 7th.
68.
W. b. Sunarharum, d. J. Williams and h. e. Smyth, Complexity of coffee
flavor: A compositional and sensory perspective, Food Res. Int., 2014,
62, 315325.
69.
I. blank and W. grosch, evaluation of potent odorants in dill seed
&
dill
herb (
Anethum graveolens
l.) by aroma extract dilution analysis,
J. Food
Sci.
, 1991,
56
, 63–67.
70.
I. blank, A. Sen and W. grosch, potent odorants of the roasted pow-
der and brew of arabica coffee, Z. Lebensm.-Unters. Forsch., 1992, 195,
239245.
71.
C. Cerny and W. grosch, Quantification of character-impact odour
compounds of roasted beef, Zeitschrift fuer Lebensmittel-Untersuchung
und-Forschung, 1993, 196, 417422.
72.
p. Semmelroch and W. grosch, Analysis of roasted Coffee powders and
brews by gas Chromatography-Olfactometry of headspace Samples,
Lebensm.-Wiss. Technol., 1995, 28, 310313.
764
Chapter 33
73.
M. Czerny,
r.
Wagner and W. grosch, detection of odor-active ethe-
nylalkylpyrazines in roasted coffee, J. Agric. Food Chem., 1996, 44,
3268–3272.
74.
p. Semmelroch and W. grosch, Studies on Character Impact Odorants
of Coffee brews, J. Agric. Food Chem., 1996, 44, 537543.
75.
M. Czerny, F. Mayer and W. grosch, Sensory study on the character
impact odorants of roasted arabica coffee, J. Agric. Food Chem., 1999,
47, 695699.
76.
M. Czerny and W. grosch, potent odorants of raw Arabica cof-
fee. Their changes during roasting, J. Agric. Food Chem., 2000, 48,
868–872.
77.
F. Mayer, M. Czerny and W. grosch, Sensory study of the character
impact aroma compounds of a coffee beverage, Eur. Food Res. Technol.,
2000, 211, 272276.
78.
W. grosch., Chemistry III: Volatile compounds, in
Coffee: Recent Devel-
opments
, ed.
r.
J. Clarke and O. g. Vitzthum, blackwell Science, london,
2001, pp 6889.
79.
h. d. belitz, W. grosch, and p. Schieberle,
Food Chemistry
, Springer, ber-
lin heidelberg, 4th edn, 2009.
80.
M. Czerny, M. Christlbauer, M. Christlbauer, A. Fischer, M. granvogl,
M. hammer, C. hartl, n. hernandez and p. Schieberle, re-investiga-
tion on odour thresholds of key food aroma compounds and devel-
opment of an aroma language based on odour qualities of defined
aqueous odorant solutions, Eur. Food Res. Technol., 2008, 228, 265273.
81.
M. Akiyama, k. Murakami, n. Ohtani, k. Iwatsuki, k. Sotoyama, A.
Wada, k. Tokuno, h. Iwabuchi and 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.
82.
M. Akiyama, k. Murakami, M. Ikeda, k. Iwatsuki, A. Wada, k. Tokuno,
M. Onishi and h. Iwabuchi, Analysis of the headspace Volatiles of
Freshly brewed Arabica Coffee Using Solid-phase Microextraction, J.
Food Sci., 2007, 72, C388C396.
83.
M. Akiyama, k. Murakami, Y. hirano, M. Ikeda, k. Iwatsuki, A. Wada,
k. Tokuno, M. Onishi and h. Iwabuchi, Characterization of headspace
Aroma Compounds of Freshly brewed Arabica Coffees and Studies on a
Characteristic Aroma Compound of ethiopian Coffee, J. Food Sci., 2008,
73, C335C346.
84.
T. Michishita, M. Akiyama, Y. hirano, M. Ikeda, Y. Sagara and T.
Araki, gas Chromatography/Olfactometry and electronic nose Anal-
yses of retronasal Aroma of espresso and Correlation with Sensory
evaluation by an Artificial neural network, J. Food Sci., 2010, 75,
S477–S489.
85.
W. holscher and h. Steinhart, Investigation of roasted coffee freshness
with an improved headspace technique, Zeitschrift fuer Lebensmittel-Un-
tersuchung und-Forschung, 1992, 195, 3338.
Coffee Volatile and Aroma Compounds From the Green Bean to the Cup
765
86.
J. S. ribeiro, F. Augusto, T. J. Salva and M. M. Ferreira, prediction models
for Arabica coffee beverage quality based on aroma analyses and che-
mometrics, Talanta, 2012, 101, 253260.
87.
k. kumazawa and h. Masuda, Identification of Odor-Active 3-Mercap-
to-3-methylbutyl Acetate in Volatile Fraction of roasted Coffee brew
Isolated by Steam distillation under reduced pressure, J. Agric. Food
Chem., 2003, 51, 30793082.
88.
r. J. Clarke and O. g. Vitzthum,
Coffee: Recent Developments
, blackwell
Sciences ltd., 2001.
89.
M. A. gianturco, A. S. giammarino and p. Friedel, The volatiles con-
stituents of coffee III The structure of two heterocyclic compounds
and the synthesis of tetrahydrofuranones, Tetrahedron, 1964, 20, 1763
1772.
90.
r. Silwar, h. kamperschroer and r. Tressl, gaschromatographic-mass
spectrometric investigation of coffee aromaquantitative determina-
tion of steam-volatile aroma constituents,
Chem., Mikrobiol., Technol.
Lebensm.
, 1987,
10
, 176187.
91.
g. A. burdock., Fenaroli's Handbook of Flavor Ingredients, CrC press/Tay-
lor & Francis, boca raton, 6th edn, 2010.
92.
p. Salo, determining the Odor Thresholds for Some Compounds in
Alcoholic beverages, J. Food Sci., 1970, 35, 9599.
93.
J. barrett, S. A. halsey and T.
l.
peppard, Flavour Stability Of Ales brewed
Using hydrolysed Maize SyrupSensory And Analytical Aspects, J I
Brewing, 1983, 89, 356360.
94.
M. larsen and
l.
poll, Odour thresholds of some important aroma
compounds in strawberries,
Zeitschrift fuer Lebensmittel-Untersuchung
und-Forschung
, 1992,
195
, 120123.
95.
r.
g. buttery and
l.
C. ling, 2-ethyl-3,5-dimethylpyrazine and 2-eth-
yl-3,6-dimethylpyrazine: Odor thresholds in water solution,
Food Sci-
ence and Technology (LEB)
, 1997,
30
, 109–110.
96.
I. blank, A. Sen and W. grosch, in
Aroma impact compounds of arabica and
robusta coffees. Qualitative and quantitative investigations,
ASIC-14eme
Colloque Scientifique International sur le Café, paris, 1992, ASIC, paris,
1991, pp. 117129.
97.
A
.
Sen,
g
.
l
askaw
y
,
p
.
Sc
hie
berle
and
W
.
g
rosc
h,
Quantit
ativ
e
d
eter
-
mination of á-damascenone in Foods Using a Stable Isotope dilution
Assay, J. Agric. Food Chem., 1991, 39, 757759.
98.
W. grosch, determination of potent Odourants in Food by Aroma
extract dilution Analysis (AedA) and Calculation of Odour Activity Val-