International Journal of Food Science and Technology 1999, 34, 287–290
Short Communication
Water sorption isotherms of roasted coffee and coffee
roasted with sugar
Emilio Cepeda*, Raul Ortiz de Latierro, María J. San José & Martín Olazar
Dpt. de Ingeniería Química, Facultad de Farmacia, Universidad del País Vasco, Aptdo 450, 01006 Vitoria, Spain
Keywords BET equation, GAB equation, monolayer moisture content, torrefacto,water activity.
Introduction
Coffee roasted with sugar (in Spanish ‘torrefacto’
coffee) is used in Spain and Portugal by mixing it
with conventionally roasted coffee to improve the
taste and the colour of the solutions with extract-
ed hot water. ‘Torrefacto’ coffee is prepared from
ground coffee and sugar. In the process, the gran-
ulated crystalline sugar is poured over the hot cof-
fee beans in a rotary oven at 140–150 °C when the
coffee is almost roasted. The sugar is partially
dehydrated by the heat and sticks to the surface
of the roasted coffee. The surface of the beans
roasted with sugar coffee is black and glossy and
the flavour of its extracted solutions is very bitter.
In Spain, it is used in mixtures (20–50%) with con-
ventionally roasted coffee, in the preparation of
coffee drinks. For coffee beans roasted with sugar
the maximum content of permitted sugar is 12%
(percentage of dry matter) and burnt grains of
5.5% (Código Alimentario Español, 1997).
Knowledge of coffee’s water sorption behaviour
is needed for predicting its hygroscopic properties
and for designing and optimizing operations of
units for preservation, storage, packaging and
mixing. Values of sorption isotherms for roasted
coffee were reported by Ayerst (1965), Quast &
Teixeira Neto (1976), Hayakawa et al. (1978), and
Weisser (1986), but there are no similar data for
coffee roasted with sugar. The purpose of the pre-
sent study was to obtain water sorption isotherms
for roasted and coffee roasted with sugar in the
*Correspondent: Fax: +34 945 130756;
e-mail: iqpcelee@vf.ehu.es
© 1999 Blackwell Science Ltd
287
temperature range of 25–40 °C; to correlate exper-
imental data with the Guggenheim-Anderson-de
Boer (GAB) model; and to calculate the mono-
layer moisture from the BET equation.
Materials and methods
Conventionally roasted and coffee roasted with
sugar was prepared by a local manufacturer
(Cafés la Brasileña, Vitoria). The coffee roasted
with sugar was obtained by the addition of 15%
(weight) of crystalline sugar to the ground coffee
at an oven temperature of 145 °C. Later the cold
beans were milled and sieved. The powder
obtained had the following particle size distribu-
tion: 200 mm, 21%; 200–400 mm, 35%; 400–800
mm, 42%; .800 mm, 2%. During this study, the
coffee was stored in sealed flasks at 224 °C.
The moisture content of the samples was mea-
sured by means of a vacuum oven method (70 °C,
16 h, 130 mbar; DIN 10 764, p.2). The sorption
isotherms were measured using the standardized
COST 90 sorbostat (Wolf et al., 1984), consisting
of preserve jars, and a support for the weighing
bottles (25 3 25 mm, DIN 12605) in which the
sample material (0.3–0.5 g) is exposed to atmos-
pheres of various humidities. The jars were main-
tained at the required temperature in a
thermostat. Saturated salt solutions were used as
the humidity source. The water activities of the
solutions covered the activity range of 0.1–0.9.
These salt solutions were prepared with reagent
grade salts and distilled water. The relative humid-
ity data of the salt solutions were obtained from
Greenspan (1977).
288 Water sorption isotherms of roasted coffees E. Cepeda et al.
Samples of coffee taken for adsorption mea-
surements were pre-dried for 7 days in a vacuum
desiccator containing anhydrous P2O5. For de-
sorption measurements the samples were condi-
tioned for one month to a relative humidity of
0.9026 (25 °C), by placing the weighing bottles in
desiccators with aqueous saturated BaCl2 solu-
tions. Samples were weighed at 24 h intervals until
no further change in weight occurred. This took
about three weeks. No microbial spoilage was
observed.
Equilibrium sorbed water was determined by
sample weighing using an analytical balance with
a precision of 0.1 mg. The percentage difference in
equilibrium moisture content between duplicate
samples was generally , 1%, and the average of
the two values was taken.
Results and discussion
The adsorption and desorption isotherms for
roasted and coffee roasted with sugar are present-
ed in Figs 1 and 2. The isotherms for roasted cof-
fee gave the characteristic S-shaped curve of water
adsorption isotherm, of type II according to the
B.E.T. classification, and the system showed a sig-
nificant hysteresis loop in the aw range 0.2–0.6.
The water activity data for roasted coffee deter-
mined in this work are slightly different to the
Figure 1 Adsorption and
desorption isotherms of roasted
coffee at 25 °C (d), 30 °C (.)
and 40 °C (j). Filled symbols for
adsorption; hollow symbols for
desorption.
International Journal of Food Science and Technology 1999, 34, 287–290
data of Weisser (1986), and this can be attributed
to a difference in the coffee variety or to the
preparation method.
The coffee roasted with sugar gave type II
isotherms up to an aw of 0.75, but at higher aw type
III isotherms were obtained due to the presence of
sugar. The pure sugar presents an adsorption
isotherm of type III, with a constant moisture
content of 0.05 % (dry basis) up to an a w of 0.75
(Leiras & Iglesias, 1991), and at higher aw the
sugar passes the dissolution point and adsorbs
more water. Consequently when the relative
humidity is high the coffee roasted with sugar
absorbs more water that the roasted coffee.
However, for low water activities both types of
coffee have similar behaviour in respect to water
adsorption.
The experimental data have been correlated to
the three parameter G.A.B. equation (van den
Berg & Bruin, 1981):
W 5 XmCKaw
W5 (1)
(1 2 Kaw) (1 2 Kaw 1 CKaw)
where aw is water activity; W is water content on
a dry basis at equilibrium; Xm is the water content
when each sorption site contains one water mole-
cule (monolayer); C and K are constants which
are related to the energies of interaction between
the first and further sorbed molecules at the indi-
© 1999 Blackwell Science Ltd
 Water sorption isotherms of roasted coffees E. Cepeda et al. 289

Figure 2 Desorption and

adsorption isotherms of coffee
roasted with sugar at 25 °C (d),
30 °C (.) and 40 °C (j). Filled
symbols for adsorption; hollow
symbols for desorption.

vidual sorption sites. In Table 1 the values of the ed data and N is the number of data. It is gener-
constants calculated by the Marquard method of ally considered that a good fit is obtained when
non-linear optimization are presented. To evaluate E , 10% (Lomauro et al., 1985; Aguerre et al.,
the goodness of fit of each equation, the mean rel- 1989). In all the cases E , 5% (Table 1), and the
ative percentage deviation modulus E was used correlation according to this criterion is very
(Aguerre et al., 1989). This modulus is defined by satisfactory.
the following equation: Monolayer moisture content is important for
N
commercial purposes (Iglesias & Chirife, 1976) as
2 mcal
E% 5
100
N Sm
i51
exp

mexp
(2) it is related to several physical and chemical dete-
riorations of dehydrated foods, such as lipid oxi-
dations, enzyme activity, non enzymatic browning
Where mexp and mcal are experimental and calculat- reactions, aroma retention and textural character-

Table 1 G.A.B. constants and mean relative deviation modulus (eqn 2) for each temperature

C K Xm E C K Xm E
g100 g21 (db) % g100 g21 (db) %

Roasted coffee
Adsorption Desorption
25 °C 04.186 0.941 4.203 4.93 18.56 0.862 4.522 1.42
30 °C 03.520 0.940 3.923 4.77 14.41 0.860 4.189 2.71
40 °C 02.531 0.937 3.441 2.23 08.901 0.856 3.591 1.91

Coffee roasted with sugar

Adsorption Desorption
25 °C 11.70 0.994 3.445 1.61 19.41 0.919 4.151 3.75
30 °C 09.088 0.990 3.288 0.74 13.83 0.909 3.914 3.40
40 °C 05.610 0.982 3.008 2.26 07.256 0.891 3.500 0.93

db 5 dry basis.

© 1999 Blackwell Science Ltd International Journal of Food Science and Technology 1999, 34, 287–290
290 Water sorption isotherms of roasted coffees E. Cepeda et al.
Table 2 Monolayer moisture
Temperature (°C)
contents, Xm (g 100 g21 dry basis),
calculated from the B.E.T.
Roasted coffee
equation
Roasted with sugar coffee
SD 5 Standard Deviation.
istics. The estimated values from the G.A.B equa-
tion are approximate because the parameters Xm,
C and K are highly correlated (Maroulis et al.,
1988). In this case, the B.E.T. equation is widely
used to determine the monolayer moisture con-
tents (Van den Berg & Bruin, 1981) The B.E.T.
equation correspond to the G.A.B. equation with
K 5 1. The values shown in Table 2 were calcu-
lated from the B.E.T. equation using data for low
water activity (0.1–0.5). There was only slight dif-
ferences in the monolayer moisture contents
between the roasted and coffee roasted with sugar,
because internally the two types of coffee are iden-
tical and the sugar adsorbs little water at low
humidities. The monolayer values decreased
slightly as temperature increased. This suggests
that shelf-life will not decrease to a great degree as
temperature increases, because the amount of
water that the product can tolerate is similar.
However, storage at low temperature is recom-
mended as this assists aroma preservation.
References
Aguerre, R. J., Suarez, C. & Viollaz, P. E. (1989). New
B.E.T. type multilayer sorption isotherms. Part II.
Modelling water sorption in foods. Lebensmittel-
Wissenschaft und Technologie, 22, 192–195.
Ayerst, G. (1965). Determination of water activity of
some hygroscopic food materials by a dew-point
method. Journal of the Science of Food and Agriculture,
16, 71–78.
van den Berg, C. & Bruin, S. (1981). Water activity and
its estimation in food systems: theoretical aspects. In:
International Journal of Food Science and Technology 1999, 34, 287–290
25 30 40 SD
3.49 3.22 3.12 0.002
3.36 3.22 3.07 0.002
Water Activity: Influence on Food Quality, (edited by
L. B. Rockland & G. F. Stewart). Pp. 1–61. New
York: Academic Press.
Código Alimentario Español y Disposiciones
Complementarias (1997). 3.25.05, Madrid: Ed. Tecnos.
Greenspan, L. (1977). Humidity fixed points of binary
saturated aqueous solutions, Journal of Research of the
National Bureau of Standards,-a- Physics and
Chemistry, 81A, 89–96.
Hayakawa, K. I., Matas, J. & Hwang, M. P. (1978).
Moisture sorption isotherms of coffee products. Journal
of Food Science, 43, 1026–1027.
Iglesias, H. A. & Chirife, J. (1976). B.E.T. monolayer
values in dehydrated foods and food components.
Lebensmittel-Wissenschaft und Technologie, 9, 123–127.
Leiras, M. C. & Iglesias, H. A. (1991). Water vapour
sorption isotherms of two cake mixes and their
components. International Journal of Food Science and
Technology, 26, 91–97.
Lomauro, C. J., Bakshi, A. S. & Labuza, T. P. (1985).
Evaluation of food moisture sorption isotherms
equations. Part I. Fruit, vegetable and meat products,
Lebensmittel-Wissenschaft und Technologie, 18, 111–117.
Maroulis, Z. B., Tsami, E., Marinos-Kouris, D. &
Saravacos, G. D. (1988). Application of the G.A.B.
model to the moisture sorption isotherms for dried
fruits. Journal of Food Engineering, 7, 63–78.
Quast, D. G. & Teixeira Neto, R. O. (1976). Moisture
problems of foods in tropical climates. Food
Technology, 30, 98, 102, 104–98, 102, 105.
Weisser, H. (1986). Influence of temperature on sorption
isotherms. In: Food Engineering and Process
Applications (eds. M. LeMaguer, & P. Jelen). Pp.
186–200. London: Elsevier App, Sci.
Wolf, W., Spiess, W. E. L., Jung, G., Weisser, H., Bizot,
H. & Duckworth, R. B. (1984). The water-vapour
sorption isotherms of microcrystalline cellulose (MCC)
and of purified potato starch. Results of a collaborative
Study. Journal of Food Engineering, 3, 51.
© 1999 Blackwell Science Ltd