Effect of moisture and water activity on textural properties of raw and roasted coffee beans

Abstract

The effect of moisture and water activity (Aw) on the textural properties of raw and roasted (light, medium and dark) coffee beans was investigated. Water–coffee interactions were described by the sorption isotherm and the mechanical properties of the samples equilibrated in the 0–0.95 Aw range were studied by uniaxial compression were studied. Roasted beans presented a lower strength and toughness than raw ones and, independently from roasting, resulting textural properties were highly dependent on Aw. At increasing hydration degree, an increase in strength and toughness was observed until a critical Aw value, different for raw (0.75) and roasted (ca. 0.86) beans related to a water, was reached. This was plasticization effect. Above these critical Aw values, a progressive softening of the bean matrix occurs. In roasted coffee beans, this plasticization determined the loss of the characteristic brittleness. The critical Aw value, above which plasticization occurred, was determined by modeling the compressive modulus using the Fermi's distribution function.
Coffee beans are a highly hygroscopic matrix and could readily take up moisture when exposed to the environment during storage. The characteristic brittleness and fragility induced by roasting is the primary attribute of roasted coffee beans. It is of great interest for process purposes to limit the textural changes of beans during their shelf-life, mainly in relation to the grinding step carried out on roasted coffee beans before the extraction of coffee brew. The investigation of the effect of water content and state on textural properties of both raw and roasted coffee beans could have implications regarding the work to be applied during the grinding process.

Full text

EFFECT OF MOISTURE AND WATER ACTIVITY ON TEXTURAL
PROPERTIES OF RAW AND ROASTED COFFEE BEANS
PAOLA PITTIA1,3, MARIA CRISTINA NICOLI2and GIAMPIERO SACCHETTI1
1Dipartimento di Scienze degli Alimenti
University of Teramo, Via Carlo R. Lerici
64023 Mosciano S. Angelo (TE), Italy
2Dipartimento di Scienze degli Alimenti
University of Udine, Via Marangoni 97
33100 Udine, Italy
Received for Publication December 9, 2005
Accepted for Publication August 16, 2006
ABSTRACT
The effect of moisture and water activity (Aw) on the textural properties
of raw and roasted (light, medium and dark) coffee beans was investigated.
Water–coffee interactions were described by the sorption isotherm and the
mechanical properties of the samples equilibrated in the 0–0.95 Awrange were
studied by uniaxial compression were studied. Roasted beans presented a
lower strength and toughness than raw ones and, independently from roasting,
resulting textural properties were highly dependent on Aw. At increasing
hydration degree, an increase in strength and toughness was observed until a
critical Awvalue, different for raw (0.75) and roasted (ca. 0.86) beans related
to a water, was reached. This was plasticization effect. Above these critical Aw
values, a progressive softening of the bean matrix occurs. In roasted coffee
beans, this plasticization determined the loss of the characteristic brittleness.
The critical Awvalue, above which plasticization occurred, was determined by
modeling the compressive modulus using the Fermi’s distribution function.
PRACTICAL APPLICATIONS
Coffee beans are a highly hygroscopic matrix and could readily take up
moisture when exposed to the environment during storage. The characteristic
brittleness and fragility induced by roasting is the primary attribute of roasted
coffee beans. It is of great interest for process purposes to limit the textural
3Corresponding author. TEL: +39-0861-266895; FAX: +39-085-8071509; EMAIL: ppittia@unite.it
Journal of Texture Studies 38 (2007) 116–134. All Rights Reserved.
© 2007, The Author(s)
Journal compilation © 2007, Blackwell Publishing
116

changes of beans during their shelf-life, mainly in relation to the grinding step
carried out on roasted coffee beans before the extraction of coffee brew. The
investigation of the effect of water content and state on textural properties of
both raw and roasted coffee beans could have implications regarding the work
to be applied during the grinding process.
KEYWORDS
Aw, coffee beans, texture, water antiplasticization, water plasticization
INTRODUCTION
Roasting is an important step in coffee processing during which green
beans are subjected to heat treatments at temperatures up to 200–250C
depending on the desired degree of roasting. The heat-induced reactions and
phenomena determine marked changes in the chemical, chemico-physical,
physical and structural properties of the raw bean. These changes determine
the peculiar sensorial and textural characteristics of the roasted coffee (Lerici
and Nicoli 1990; Illy and Viani 1995; Clarke and Vitzum 2001). As regards
texture, it is known that during roasting, coffee beans lose their strength and
toughness and become brittle and fragile. Uniaxial compression carried out on
single coffee beans at different roasting degree showed changes in the shape of
the force–deformation curve upon roasting. These changes could be described
by a progressive decrease in force, energy and strain at the fracture point upon
roasting (Pittia et al. 2001).
The textural characteristics of roasted coffee could be related to the
effects of some chemical and physical changes induced on the raw bean
components by the severe thermal process. In particular, both the modification
of the coffee composition, caused by nonenzymatic browning and pyrolysis,
and the decrease of density could be implied in the determination of the
peculiar brittleness of the roasted product.
The reaching of a certain degree of brittleness is very important in the
grinding process, which is carried out on roasted coffee beans before
the extraction of coffee brew.
As coffee beans have a highly hygroscopic matrix, they could readily take
up moisture when exposed to the atmospheric environment during storage.The
investigation of water content and state on the textural properties of coffee
beans could be of great interest in relation to the energy that has to be applied
during grinding. Moreover, roasted coffee beans could be included in food
products like sweets and chocolate where brittleness is considered a primary
attribute in these products (Szczesniak 1990).
117TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

The effect of both water content and water activity (Aw) on the textural
properties of raw and roasted coffee beans was investigated. The interactions
between water and coffee beans matrix were studied by means of sorption
isotherms. The mechanical properties of raw and differently roasted coffee
samples equilibrated at various Awvalues (between 0 and 0.95) were
determined.
MATERIALS AND METHODS
Materials
Raw Coffea arabica var. Santos beans and a blend of roasted beans
containing 50% C. arabica var. Santos, 30% C. arabica (Colombia) and 20%
C. canephora var. Robusta (Ivory Coast) were used.
The beans that were roasted at three different degrees of browning were
kindly supplied by the Nestlè Research Centre (Lausanne, Switzerland).
Weight losses upon roasting of these samples were 14.5, 16.2 and 18.9%.
Methods
Color. A tristimulus colorimeter (Chromameter-2 Reflectance, Minolta,
Osaka, Japan) equipped with a CR-200 measuring head was used. Standard
Commission Internationale de l’Eclairage (CIE) conditions (CIE 1986) with
illuminant “C” (6774K) and 2° standard observer were used. The instrument
was calibrated on a white tile (L*=98.82; a*=-0.18; b*=-0.31) before each
measurement. Color, expressed in L*, a* and b* parameters, was measured on
samples of coffee beans placed on suitable cells.
Density. The density was determined by a picnometer, according to
Lerici et al. (1980).
Moisture. The moisture was determined by gravimetrical analysis after
drying in a vacuum oven at 75C for 12 h.
Water Sorption Isotherm. Green and roasted coffee beans were pre-
liminarily dehydrated by vacuum drying and by exposure to P2O5. The dried
samples were put in desiccators and hydrated at different relative humidity
(RH%) conditions over P2O5and saturated salt solution at room temperature
(22 ⫾1C). The RH% was adjusted at 0, 33.0, 44.4, 53.8, 76.0, 85.8 and 94.0%
using P2O5and saturated solution of MgCl2,K
2CO3, Mg(NO3)2, NaCl, KCl and
K2NO3(Labuza et al. 1985). The water content was determined from weight
gain and the sorption isotherm was constructed at 20C using steady state water
118 P. PITTIA, M.C. NICOLI and G. SACCHETTI

contents and Awvalues determined by anAqualab (Decagon Devices, Pullman,
WA) hygrometer.
Textural Properties. Determined by an Instron Universal Testing
Machine, model 4301 (UTM, Instron International Limited, High Wycombe,
U.K.) equipped with a 1,000 N load cell. Uniaxial compression was carried out
at a rate of 0.83 cm/s until failure occurred; working temperature was
20 ⫾1C.
Measurements were performed at least on 25 beans randomly taken from
each coffee type, which were positioned individually between two parallel
plates of the dynamometer on their longest side and with the flat side up.
Mechanical properties of coffee beans at different moisture and Awwere
characterized with the following empirical parameters obtained from the
force–displacement curve (Borges and Peleg 1997):
(1) fracture force (N), corresponding to the force at the major failure event.
It was considered an empirical measure of strength;
(2) fracture energy (J), corresponding to the area under the force–
displacement curve until the first breaking event occurred. It was consid-
ered as an empirical index of toughness;
(3) strain at fracture (%), corresponding to the deformation at the first break-
ing point and used as index of the deformability. It was expressed as
percentage to keep into account the actual dimensions of the different
coffee samples.
Compressive modulus (N/m2) was determined according to Chang et al.
(2000) and computed using the following equation [dN/dY]·(h/A), where
dN/dY corresponds to the initial slope of the force–deformation curve,
h=height of the specimen (m), A =sample area exposed to stress.
The mean value of bean height and area exposed to stress was evaluated
on 25 specimens of the raw and roasted coffee samples at the different levels
of moisture content and used to this purpose.
Data Modelling. The adsoption isotherms were described by the GAB
equation according to Van den Berg and Bruin (1981).
WXA
AA A
m
=⋅⋅⋅
−⋅
()
−⋅ + ⋅⋅
()
Cg K
KKCgK
w
ww w
11 (1)
where Wis water content on dry basis, Xm is the water content of the mono-
layer, Awis water activity and Cg and Kare constants that are related to the
119TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

energies of interaction between the first and further adsorbed molecules at the
individual sorption sites.
Compressive modulus was modeled as a function of Awusing the general
form of the Fermi’s distribution that, according to Wollny and Peleg (1994), is
described by the following equation:
YY
e
AAA
w
b
wwc
()
−
=
+
0
1
(2)
where Y(Aw) is the compressive modulus, Y0is the compressive modulus in the
dry state (Aw=0.1), Awc is the Awvalue where a drastic textural change takes
place, eis exponential and bis a dimensionless constant.
Data Processing and Statistical Analysis. Mean values were compared
by one-way analysis of variance, followed by Duncan’s multiple range test at
a significance level of p<0.05. Nonlinear regression analysis was performed
using the algorithm “Quasi-Newton” supplied by the STATISTICA (StatSoft,
Inc., Tulsa, OK) package. The standard significance associated with the error
in the estimation of the parameters was determined by a t-test.
RESULTS AND DISCUSSION
The values of moisture, Aw, density and colorimetric parameters (L*, a*
and b*) of the coffee samples under study are shown in Table 1. Significant
differences in the chemical, physical and physico-chemical properties between
the green and the differently roasted coffee samples are caused by the heat-
induced reactions during processing (Massini et al. 1991; Severini et al. 1992;
Clarke and Vitzum 2001). With increasing degree of roasting and weight loss,
the moisture and density decreased and the brown color increased. Browning
is, in turn, described by a decrease of L* as well as of a* and b* parameters.
Samples were classified as light, medium and dark roasted on the basis of the
weight loss, density and color parameters (Da Porto et al. 1991; Nicoli et al.
1997; Anese et al. 2000; Romani et al. 2003).
In the Awrange considered in this study, the isotherms show an S-shaped
curve, typical of porous products, which correspond to that of a type II
isotherm according to the Brunauer, Deming, Deming and Teller (BDDT)
classification (Brunauer et al. 1940).
Isotherms were fitted by the GAB equation and the results are reported in
Table 2. The monolayer value of green coffee was calculated to be 0.0434 g/g
of solids and those of roasted coffee were about 0.035 g/g of solids in
120 P. PITTIA, M.C. NICOLI and G. SACCHETTI

TABLE 1.
MOISTURE, COLOR AND DENSITY OF RAW AND DIFFERENTLY ROASTED COFFEE SAMPLES
Coffee sample Moisture (% w/w) AwDensity (g/mm3) Color
L*a*b*
Raw 7.95 ⫾0.21a0.523 ⫾0.01a1.154 ⫾0.002a50.3 ⫾2.71a2.7 ⫾0.34d11.6 ⫾1.05a
Roasted light 1.81 ⫾0.05b0.253 ⫾0.01b0.597 ⫾0.001b31.1 ⫾0.79b10.8 ⫾0.39a10.4 ⫾0.99a
Roasted medium 1.60 ⫾0.06b0.214 ⫾0.08c0.560 ⫾0.002c26.0 ⫾0.71c8.8 ⫾0.18b5.8 ⫾0.40b
Roasted dark 1.29 ⫾0.05c0.207 ⫾0.09c0.503 ⫾0.001d24.3 ⫾0.42c6.5 ⫾0.11c1.1 ⫾0.20c
Values with the same letter in a column are not significantly different at a P>0.05 level.
121TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

accordance with other authors (Ortolà et al. 1993; Gutiérrez et al. 1997;
Cepeda et al. 1999). Even if the monolayer values are consistent with those
found in literature, it is noticeable that the Cg values of raw and light roasted
coffee are lower than those of medium and dark roasted ones and this could
lead to a much bigger error in estimating the monolayer capacity (Lewicki
1997).
The upward concavity of the S-shaped curve at Aw>0.825 occurred in the
range of 0.20 and 0.17 g water/g of solids, respectively, for green and roasted
coffee (Fig. 1). Above 0.825, at increasing Awvalues, roasted coffee was able
to retain more water than the raw coffee. This part of the isotherm corresponds
to bulk phase water that is physically entrapped in the cellular structure
(Fennema 1996).
TABLE 2.
REGRESSION PARAMETERS OF GAB EQUATION APPLIED
TO SORPTION ISOTHERMS
Coffee sample Xm
% dry basis
CgKR
2
Raw 4.34a** 4.29b** 0.960d** 0.999
Light roasted 3.58b** 3.05b* 0.991c** 0.999
Medium roasted 3.56b** 6.00ab* 0.998b** 0.999
Dark roasted 3.54b** 6.76a* 0.999a** 0.999
* Parameter significant at P<0.05 level; ** parameter significant
at P<0.01 level.
Values with the same letter in a column are not significantly different
at a P>0.05 level.
FIG. 1. ADSORPTION ISOTHERMS OF RAW AND ROASTED COFFEE BEANS
122 P. PITTIA, M.C. NICOLI and G. SACCHETTI

During roasting, it is known that both volume increase of the bean and
development of pores or structure alteration occur (Massini et al. 1991; Ortolà
et al. 1998; Schenker et al. 2000). It could be suggested that the higher
porosity and exposed surface to the surrounding atmosphere of the roasted
bean could be responsible for the higher quantity of physically segregated
water above 0.825 Awvalue.
The degree of roasting did not show to affect water–matrix interactions in
the monolayer hydration shell because the adsorption isotherms of the light,
medium and dark roasted coffee beans were almost similar in the Awrange
below 0.825.
Textural properties of raw and roasted coffee samples equilibrated at
different Awvalues, and thus having different moisture content, were studied
by uniaxial compression according to Pittia et al. (2001). In Figs. 2 and 3,
examples of the force–deformation curves of green and roasted coffee beans at
three different degrees of hydration are shown.
In all cases, at low moisture (0 <Aw<0.44), the force–deformation rela-
tionship was irregular and jagged (Figs. 2 and 3). Furthermore, a peak stress
was seen at the first main fracture, which was followed by a series of fractures
at lower stresses until compression was completed. The degree of jaggedness
of this compressive force–deformation relationship could be a manifestation of
the failure events that produce the sensation of crispness in brittle products
(Vickers 1988; Suwonsichon and Peleg 1998; Vincent 1998).
At medium and high moisture content (Aw>0.52), this characteristic
behavior showed a progressive change, suggesting an evolution of the fracture
mechanism toward plastic failure, usually associated with polymeric materials
in the leathery or rubbery state (Figs. 2 and 3). After a bioyield point, stress
0
20
40
60
80
100
120
140
160
180
0 0,5 1 1,5 2 2,5
deforma tion (mm)
Force (N)
FIG. 2. FORCE–DEFORMATION CURVES OF RAW COFFEE BEANS AT DIFFERENT WATER
ACTIVITY (Aw) VALUES
solid line: Aw=0.11; dotted line: Aw=0.52; dashed line: Aw=0.91.
123TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

continued to increase to a maximum with increasing deformation because of
compressing or densification of the material.
The change in the trend of the force–deformation graph occurred both in
the raw and roasted coffee beans, although it happened at a different degree of
hydration according to the coffee sample and in particular at Aw>0.44 and
>0.52 for raw and roasted coffee, respectively (Figs. 2 and 3).
In Fig. 4, the fracture force (a), energy (b) and strain (c) are compared
between raw and dark roasted coffee beans as a function of Awvalue.
Roasted coffee beans, at Awvalues between 0 and 0.25, showed a lower
fracture force and energy (i.e., strength and toughness, respectively [see
Fig. 4a,b]) than the raw ones. This level of hydration corresponds to what is
generally found in roasted coffee (see also Table 1). Textural characteris-
tics shown by coffee samples in this Awrange refer to a material that is
essentially nondeformable, that breaks easily and that abruptly releases
energy (Szczesniak 1990). This highly brittle and fragile texture of the
roasted beans is confirmed also by their relatively low fracture strain
(Fig. 4c).
The raw bean showed much higher fracture force values than dark roasted
coffee which could be explained by the natural presence of a certain amount of
some structural polysaccharides, especially those derived from the polymer-
ization of mannan (Trugo 1985). The roasting process causes a dramatic
decrease of hardness and a significant decrease of fracture energy and strain,
which could be explained by porosity and structure alteration caused by
chemical and physical changes of the cell wall compounds (Massini et al.
1991; Ortolà et al. 1998; Schenker et al. 2000).
Upon moisture sorption, fracture force and energy of both raw and
roasted coffee beans increased progressively to a maximum value, after which
0
5
10
15
20
25
30
35
40
0 0,2 0,4 0,6 0,8 1 1,2 1,4
deformation (mm )
Force (N)
FIG. 3. FORCE–DEFORMATION CURVES OF DARK ROASTED COFFEE BEANS
AT DIFFERENT WATER ACTIVITY (Aw) VALUES
solid line: Aw=0.11; dotted line: Aw=0.52; dashed line: Aw=0.91.
124 P. PITTIA, M.C. NICOLI and G. SACCHETTI

a decrease of these textural parameters was observed. The hydration degree at
which the increase of strength and toughness started as well as the reaching of
the maximum was different for raw and roasted coffee. In particular, the
toughening upon moisture adsorption of raw coffee beans started at Awⱖ0.33
and the maximum was evidenced at 0.75, while in dark roasted coffee beans,
a
b
c
0
50
100
150
200
250
00,20,40,60,8 1
a
w
Fracture force (N)
0
0,03
0,06
0,09
0,12
0,15
00,20,40,60,81
a
w
Fracture energy (J)
0
10
20
30
40
50
0 0,2 0,4 0,6 0,8 1
aw
Fracture strain (%)
FIG. 4. FRACTURE FORCE (a), ENERGY (b) AND STRAIN (c) OF RAW AND ROASTED
COFFEE BEANS AS A FUNCTION OF THE WATER ACTIVITY VALUE
䉫:raw;䊏: dark roasted.
125TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

the same events took place at higher Awvalues (0.44 and 0.84, respectively). At
high Awvalues, however, a dramatic decrease in both strength and toughness
occurred in both raw and roasted coffee beans.
Also, the deformation at the fracture point showed a progressive increase
with increasing water uptake. The maximum relative deformation at fracture
was reached at the same Awvalue of the maximum fracture force and energy,
although, after this point, no clear decreases in this textural parameter was
observed (Fig. 4c).
Roasting degree showed a limited influence on the textural parameters
under study (Fig. 5). The trends of fracture force, energy and strain as a
function of Aware similar and no significant differences were observed
between the mean data of the differently roasted coffee beans. This result
could be caused by data dispersion, as indicated by the error bars of the
textural parameters, and is related to the nature of the samples under study. In
fact, the use of a blend of different species (Arabica and Robusta) and varieties
(Arabica and Arabica cv. Santos) produces a heterogeneous coffee material
where each coffee type could exert a significant different mechanical behavior
even if it underwent a similar roasting process. This is a consequence of the
specific chemical composition and related effects of the heat treatment (Pittia
et al. 2001). However, a progressive increase of all these textural indices was
evidenced with decreasing the degree of roasting. This is likely related to the
higher moisture and density and/or to a lower thermal degradation of the
biopolymers of the less roasted coffee beans.
The changes of the mechanical parameters as a consequence of the
increased moisture uptake and, in particular, the increase of fracture deform-
ability, clearly indicate a change in the textural properties of the coffee beans.
At high water activity values, dried coffee beans that are more (raw) or less
(roasted) hard and brittle become soft and ductile upon moisture adsorption.
For roasted coffee, it means that the brittle and fragile structure of the bean
gradually loses its characteristic crunchiness and crispiness to become, at high
Awvalues, plastic and viscous.
The food polymeric science approach attributed the decrease in hardness
upon moisture adsorption to a glass transition of the material, which is trig-
gered by lowering the glass transition temperature (Tg) to below the ambient
temperature (Slade and Levine 1993; Peleg 1994). Support for this mechanism
is related to the known plasticizing effect of water on various bio and synthetic
polymers which is indeed manifested by lowering Tg(Harris and Peleg 1996).
Because of the water effect on the physical properties and state of a polymeric
matrix, at increasing moisture content a decrease in toughness is generally
observed (Levine and Slade 1988; Slade and Levine 1991).
However, in all coffee samples, at low and intermediate Awvalues, the
increase in moisture induces a progressive higher stiffness and toughness and
126 P. PITTIA, M.C. NICOLI and G. SACCHETTI

a)
c)
b)
0
10
20
30
40
50
60
70
80
90
00,20,40,60,81
a
w
Fracture force (N)
0
5
10
15
20
25
30
35
40
0 0,2 0,4 0,6 0,8 1
aw
Fracture strain (%)
0
0,01
0,02
0,03
0,04
0,05
0,06
00,20,40,60,81
a
w
Fracture energy (J)
FIG. 5. FRACTURE FORCE (a), ENERGY (b) AND STRAIN (c) OF LIGHT, MEDIUM AND
DARK ROASTED COFFEE BEANS AT DIFFERENT HYDRATION DEGREE
䊉: light; 䉱: medium; 䊏: dark roasted.
127TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

this effect seems to be opposite to those observed as a consequence of the
plasticization effect of water. In coffee beans, water showed to act as plasti-
cizer only at a high hydration degree and above specific Awvalues, which, in
turn, were different for the raw and the differently roasted samples.
This contrasting behavior has also been observed in other food matrices
and was attributed to an “antiplasticizer” effect of water (Seow et al. 1999).
It has been suggested that this phenomenon takes place in systems at
low-medium Awvalues provided that they are in an amorphous state. Dif-
ferent are the hypothetical causes: reduction of the free volume of the
polymeric system; polymer–diluent interactions, which create steric hin-
drance and decrease segmental mobility; and stiffening action because of
the presence of rigid plasticizer molecules adjacent to polar groups of the
polymer (Seow et al. 1999). In this case, with increasing hydration, the
material becomes harder and less flexible even though a decrease of its Tgis
evident. The different coffee matrix–water relationship observed in the sorp-
tion isotherm seems to affect the occurrence of the antiplasticizing effect on
the raw and roasted coffee. In particular, from Awvalues >0.25, raw coffee
was able to adsorb more water than roasted ones and this could explain the
marked increase of toughness and strength even at this quite low moisture
level.
The present work did not aim to define the conditions in which the
transition phase of coffee occurs. However, Geiger et al. (2002) developed a
state diagram of Colombian coffee beans by means of thermo-mechanical
analysis. On the basis of their data, a phase transition from amorphous to a
rubbery state takes place at ambient temperature (20–25C) when beans had a
moisture content of about 18–20%. Adsorption isotherms of our samples
(Fig. 1) evidence that this condition corresponds to coffee having an Awvalue
of about 0.8; this suggests an amorphous state of our coffees in the Awrange
between 0 and 0.75–0.80.
In coffee beans, water exhibits both an antiplasticizer and a plasticizer
effect depending on the hydration degree. As long as, upon moisture adsorp-
tion, coffee beans do not reach a water content and a correspondent Awvalue
at which a phase transition is induced, an antiplasticizer effect is evident.
Above this condition, water present in the coffee matrix acts as plasticizer
and could decrease toughness, crispness and hardness, increase flowability
and softness and favor viscous behavior (Nelson and Labuza 1993; Roos
1995).
For raw coffee beans, our results indicate that the critical Awvalue is
between 0.538 and 0.760. In the roasted coffee, this is slightly higher (between
0.760 and 0.940).
In order to highlight the influence of Awon textural properties of raw and
roasted coffee beans, compressive moduli were determined from the force–
128 P. PITTIA, M.C. NICOLI and G. SACCHETTI

deformation graphs, and the curves obtained by plotting the compressive
modulus as a function of Aware reported in Fig. 6.
The compressive modulus that could be considered an index of the
stiffness or consistency has been previously used by other authors to evaluate
mechanical properties’ variation of food matrices caused by moisture adsorp-
tion (Borges and Peleg 1997; Chang et al. 2000). This compressive modulus
was chosen in order to take into consideration also the changes in the dimen-
sions of the beans caused by water adsorption (Table 3). With increasing
hydration, an increase of the dimensions of the beans was seen up to a certain
moisture content and Aw, which was different for raw and roasted coffee
samples, above which a decrease was observed.
The relationship between compressive modulus and Awin all coffee
samples (raw and differently roasted) showed a sigmoid shape that could be
0
1
2
3
4
5
6
7
0 0,2 0,4 0,6 0,8 1
aw
compr. modulus (N/m
-2
)*10
-6
0
0,5
1
1,5
2
2,5
3
0 0,2 0,4 0,6 0,8 1
aw
compr. modulus (N/m
-2
)*10
-6
a)
b)
FIG. 6. COMPARISON OF COMPRESSIVE MODULUS OF RAW AND DARK ROASTED (a)
AND DIFFERENTLY ROASTED (b) COFFEE BEANS
䉫:raw;䊉: light; 䉱: medium; 䊏: dark roasted.
129TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

described by the Fermi’s distribution function as previously reported for
certain cereal products (Wollny and Peleg 1994; Harris and Peleg 1996;
Borges and Peleg 1997). The compressive modulus as a function of Awdoes
not evidence an antiplasticization effect of water, while quite clear is the
hydration degree at which the water plasticization effect occurs (Fig. 6).
In Table 4, the estimated values of the Fermi’s model constants are
reported for each coffee sample under study. The Awc values obtained by the
model suggest that a 50% decrease in stiffness occurred at 0.77 and 0.82–0.84
for raw and differently roasted coffee, respectively. These estimated values are
in the Awrange at which fracture force and energy started a steep decrease for
both raw and roasted coffee (Fig. 4). Furthermore, the Awrange over which
stiffness was lost is, in general, relatively narrow as indicated by the low b
TABLE 3.
AREA OF THE PLANE SURFACE AND HEIGHT OF THE RAW AND DARK ROASTED
COFFEE BEANS
AwRaw Dark roasted
Area (mm2) Height (mm) Area (mm2) Height (mm)
0.05 49.7 ⫾4.7 3.6 ⫾0.3 73.8 ⫾6.4 4.9 ⫾0.3
0.33 53.6 ⫾3.1 3.6 ⫾0.3 77.1 ⫾7.5 4.8 ⫾0.5
0.44 54.0 ⫾3.9 3.8 ⫾0.3 78.1 ⫾6.9 4.9 ⫾0.3
0.52 53.1 ⫾3.7 3.9 ⫾0.3 76.6 ⫾4.5 4.9 ⫾0.3
0.76 59.3 ⫾4.8 3.9 ⫾0.3 72.6 ⫾5.1 4.9 ⫾0.2
0.86 68.6 ⫾4.3 4.3 ⫾0.4 72.2 ⫾8.1 4.8 ⫾0.4
0.91 67.9 ⫾5.2 4.3 ⫾0.3 72.0 ⫾3.8 4.7 ⫾0.2
TABLE 4.
REGRESSION PARAMETERS OF FERMI’S EQUATION
APPLIED TO COMPRESSIVE MODULUS CHANGES
CAUSED BY HYDRATION
Coffee sample Y0(N/m2)Awc bR
2
Raw 4.6a** 0.77b** 0.053a* 0.967
Light roasted 2.1b** 0.83a** 0.051a* 0.925
Medium roasted 1.9b** 0.82a** 0.048a* 0.935
Dark roasted 1.6c** 0.84a** 0.036a* 0.970
* Parameter significant at P<0.05 level; ** significant at P<0.01
level.
Values with the same letter in a column are not significantly different
at a P>0.05 level.
130 P. PITTIA, M.C. NICOLI and G. SACCHETTI

values. It is interesting to observe that the latter parameter becomes progres-
sively lower starting from raw coffee to the dark roasted product.
Harris and Peleg (1996) observed that for cellular cereal products, the
mechanical behavior seems to be determined by phenomena at molecular and
structural level and by their interactions. The difference in the Fermi’s model
constants of differently roasted coffee beans could thus be because of the
structural and compositional modifications determined by roasting (Clarke and
Vitzum 2001). Roasted coffee beans, as a consequence of the severe heat
treatment, present a significantly different composition and structure with
respect to raw coffee. Changes in the concentration of hydrophilic macromol-
ecules (low and high molecular saccharides, proteins), formation of Maillard
reaction and pyrolysis products, as well as the modification of the cellular
characteristics of the raw coffee matrix (porosity, cell size distribution, cell
wall thickness) are all factors that could be implied in the different textural
behavior of the raw and roasted beans before and after moisture adsorption and
determine a markedly different critical Awvalues.
CONCLUSIONS
Textural properties of coffee beans depend strictly on the hydration
degree as well as on roasting.
With increasing hydration, water acts as antiplasticizer until a critical Aw
value, which is different for raw and roasted beans, is reached. Above this
value, water acts as plasticizer and fracture force, energy and strain decrease
with increasing water content in the bean. This effect is evident both for raw
and roasted coffee beans, even if the former has a higher strength and tough-
ness than the latter because of chemical and physical changes in the beans
during roasting. The existence of a critical Awvalue above which a dramatic
change in texture take place was also studied by fitting the compressive
modulus as a function of Awby the Fermi’s equation. The moisture level at
which the dramatic change in texture occurs was similar to that at which a
phase transition occurred and was different for raw and differently roasted
coffee beans. In roasted coffee beans, this critical Awis very close to that where
the upward concavity of the water sorption starts while this correspondence
was not observed in the raw ones. The different mechanical behavior of raw
and roasted coffee beans as a function of moisture and Awcould only be partly
explained by the water–matrix interactions evidenced in the correspondent
sorption isotherm as it depends also on the specific chemical, structural and
cellular (porosity, open or closed cell, cell wall thickness) properties of the
beans, either raw or roasted.
131TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

REFERENCES
ANESE, M., DE PILLI, T., MASSINI, R. and LERICI, C.R. 2000. Oxidative
stability of the lipid fraction in roasted coffee. Ital. J. Food Sci. 4(12),
457–462.
BORGES, A. and PELEG, M. 1997. Effect of water activity on the mechanical
properties of selected legumes and nuts. J. Sci. Food Agric. 75, 463–
471.
BRUNAUER, S., DEMING, L.S., DEMING, W.E. and TELLER, E. 1940. On
a theory of the van der Waals adsorption of gases. J. Am. Chem. Soc. 62,
1723–1732.
CIE. 1986. Colorimetry, 2nd Ed., Publ. 15, Suppl. 2. Central Bureaux of the
Commission Internationale de l’Eclairage, Wien, Austria.
CEPEDA, E., ORTIZ DE LATIERRO, R., SAN JOSÈ, M. and OLAZAR, M.
1999. Water sorption isotherms of roasted coffee and coffee roasted with
sugar. Int. J. Food Sci. Tech. 34, 287–290.
CHANG, Y.P., CHEAH, P.B. and SEOW, C.C. 2000. Variations in flexural and
compressive fracture behaviour of a brittle cellular food (dried bread) in
response to moisture sorption. J. Texture Studies 31, 525–540.
CLARKE, R.J. and VITZUM, O.G. (eds.) 2001. Coffee. Recent Develop-
ments. Blackwell Science, Oxford, U.K.
DA PORTO, C., NICOLI, M.C., SEVERINI, C., SENSIDONI, A. and
LERICI, C.R. 1991. Study on physical and physico-chemical changes in
coffee beans during roasting. Note 2. Ital. J. Food Sci. 3, 197–207.
FENNEMA, O. (ed.) 1996. Food Chemistry. Marcel Dekker, Inc., New York,
NY.
GEIGER R., PERREN R., SCHENKER S. and ESCHER F. 2002. Mechanism
of volume expansion in coffee beans during roasting. In Proceedings 19th
Int. Conference on Coffee Sci. ASIC, May 14–19, 2001, ASIC, Paris,
France.
GUTIÉRREZ, C., ORTOLÀ, M.D. and CHIRALT, A. 1997. Interacciones
agua-café. Alim. Equip. Technol. 16(1), 121–128.
HARRIS, M. and PELEG, M. 1996. Patterns of textural changes in brittle
cellular cereal foods caused by moisture sorption. Cereal Chem. 73,
225–231.
ILLY, A. and VIANI, R. (eds.) 1995. Espresso Coffee, 1st Ed., Academic
Press, London, U.K.
LABUZA, T.P., KAANANE, A. and CHEN, J.Y. 1985. Effect of temperature
on the moisture sorption isotherms and water activity shift of two dehy-
drated foods. J. Food Sci. 50, 385–391.
LERICI, C.R. and NICOLI, M.C. 1990. Il caffè: Aspetti chimici, tecnologici e
qualità della bevanda. Riv. Soc. It. Sci. Alim. 19, 1–16.
132 P. PITTIA, M.C. NICOLI and G. SACCHETTI

LERICI C.R., DALLA ROSA M., MAGNANINI E. and FINI P. 1980. Pro-
cessi di trasformazione del caffè: Aspetti chimici, fisici e tecnologici.
Nota 4: Evoluzione di alcuni caratteri fisici del caffè durante la torrefazi-
one. Ind. Bevande 9, 375–381.
LEVINE, H. and SLADE, L. 1988. Water as plasticizer: Physico-chemical
aspects of low-moisture polymeric systems. In Water Science Reviews 3,
(E. Franks, ed.) pp. 79–185, Cambridge University Press, Cambridge,
U.K.
LEWICKI, P.P. 1997. The applicability of the GAB model to food water
sorption isotherms. Int. J. Food Sci. Tech. 32, 553–557.
MASSINI, R., NICOLI, M.C., CASSARÀ, A. and LERICI, C.R. 1991. Study
on physical and physico-chemical changes of coffee beans during roast-
ing. Note 1. Ital. J. Food Sci. 2, 123–134.
NELSON, K.A. and LABUZA, T.P. 1993. Glass transition theory and the
texture of cereal foods. In The Glassy State of Foods, (J.M.V. Blanshard
and P.J. Lillford, eds.) pp. 513–518, Nottingham University Press, Lough-
borough, U.K.
NICOLI, M.C., ANESE, M., MANZOCCO, L. and LERICI, C.R. 1997.
Antioxidant properties of coffee brews in relation to the roasting degree.
Lebensm. –Wiss. Technol. 30, 292–297.
ORTOLÀ, M.D., GUTIERREZ, C., CHIRALT, A. and FITO, P. 1993. Control
cinetico y termodinamico de las interactions agua-cafè tostato. In Pro-
ceedings 15th Int. Colloquium on Coffee ASIC, (ed.) pp. 672–681, ASIC,
Paris, France.
ORTOLÀ, M.D., LONDONO, L., GUTIERREZ, C.L. and CHIRALT, A.
1998. Influence of roasting temperature on physicochemical properties of
different coffees. Food Sci. Technol. Int. 4, 59–66.
PELEG, M. 1994. A model of mechanical changes in biomaterials at and
around their glass transition. Biotechnol. Prog. 10, 385–388.
PITTIA, P., DALLA ROSA, M. and LERICI, C.R. 2001. Textural changes of
coffee beans as affected by roasting conditions. Lebensm. – Wiss.
Technol. 33, 1–8.
ROMANI, S., PINNAVAIA, G.G. and DALLA ROSA, M. 2003. Influence of
roasting levels on ochratoxin a content in coffee. J. Agric. Food Chem.
51, 5168–5171.
ROOS, Y.H. 1995. Mechanical properties. In Phase Transitions in Foods, (ed.)
pp. 247–270, Academic Press, San Diego, CA.
SCHENKER, S., HANDSCHIN, S., FREY, B., PERREN, R. and ESCHER, F.
2000. Pore structure of coffee beans affected by roasting conditions. J.
Food Sci. 65, 452–457.
SEOW, C.C., CHEAH, P.B. and CHANG, Y.P. 1999. Antiplasticisation by
water in reduced-moisture food systems. J. Food Sci. 64, 576–581.
133TEXTURE OF COFFEE BEANS AND WATER ACTIVITY

SEVERINI C., NICOLI M.C., MASTROCOLA, D. and LERICI, C.R. 1992.
Influence of heating rate on some physical physico-chemical properties of
coffee beans. In Proceedings of 14th Int. Sci. Colloquium on Coffee, (ed.)
pp.641–648, ASIC, San Francisco, CA.
SLADE, L. and LEVINE, H. 1991. Beyond water activity: Recent advances
based on alternative approach to the assessment of food quality and
safety. Crit. Rev. Food Sci. Nutr. 30, 115–360.
SLADE, L. and LEVINE, H. 1993. The glassy state phenomenon in
food molecules. In Glassy State in Foods, (J.M.V. Blanshard and P.J.
Lillford, eds.) pp. 35–101, Nottingham University Press, Loughborough,
U.K.
SUWONSICHON, T. and PELEG, M. 1998. Instrumental and sensory detec-
tion of simultaneous brittleness loss and moisture toughening in three
puffed cereals. J. Texture Studies 29, 255–274.
SZCZESNIAK, A.S. 1990. Texture: Is still overlooked food attribute. Food
Technol. 44(9), 86–95.
TRUGO, L.C. 1985. Carbohydrates. In Coffee. Vol. 1: Chemistry (R.J. Clarke
and R. Macrae, eds.) pp. 83–114, Elsevier Applied Science Publisher,
London, U.K.
VAN DEN BERG, C. and BRUIN, S. 1981. Water activity and its estimation
in food systems: Theoretical aspects. In Water Activity: Influence on Food
Quality, (L.B. Rockland and G.F. Stewart, eds.) pp. 1–61, Academic
Press, New York, NY.
VICKERS, Z.M. 1988. Instrumental measurement of crispness and their cor-
relation with sensory assessment. J. Texture Studies 19, 1–14.
VINCENT, J.F.V. 1998. The quantification of crispness. J. Sci. Food Agric. 78,
162–168.
WOLLNY, M. and PELEG, M. 1994. A model of moisture-induced plastici-
zation of crunchy snacks on Fermi’s distribution function. J. Sci. Food
Agric. 64, 467–473.
134 P. PITTIA, M.C. NICOLI and G. SACCHETTI