Int. Agrophys.

, 2020, 34, 161-171

doi: 10.31545/intagr/117532

Comparison of sorption properties of black pepper of different fineness levels

using selected models
Aneta Ocieczek1* , Tomasz Pukszta1 , and Vitorina Chilumbo2
1
Department of Commodity Science and Quality Management, Gdynia Maritime University, Morska 83, 81-225 Gdynia, Poland
2
Academy of Fishery and Marine Sciences, Moçâmedes, Namibe, Angola

Received September 11, 2019; accepted January 27, 2020

A b s t r a c t. The study assessed pepper’s tendency to absorb world’s production) and Brazil (16% of the world’s produc-
water vapour from the environment. This property depends on its tion), which indicates that the global popularity of the spice
fineness level, and should be regarded as the basic factor deter- is associated with the need for long-distance transport.
mining its storage stability. The scope of the study included the
Therefore, the quality of this commodity is significantly
determination of the water content and activity of the analysed
material; the determination of adsorption isotherms at 20°C affected by maritime transport conditions. According to the
as well as a mathematical description of sorption isotherms Observatory of Economic Complexity (OEC), the import
using the Brunauer, Emmett, Teller, Guggenheim, Anderson, of black pepper to Poland in 2017 amounted to 2.9% of
De Boer, and Peleg models. Another aim of the study was to the world’s production, which accounted for 70% of the
assess these models for describing surface phenomena and to tonnage of Poland’s import of spices and 80% of its value.
evaluate selected parameters of the pepper particle surface micro- The quality of spices is undoubtedly determined by the
structure as well as the parameters characterizing the energy
cultivation and harvesting methods as well as by the pack-
phenomena accompanying adsorption as factors used to predict
stability during storage. An analysis of the results demonstrated aging used and the storage methods and conditions applied,
that commercial samples of black pepper of different fineness of which temperature, humidity and light exposure are of
levels were characterized by a different water content and activity. particular importance. An excessively high storage tempe-
Pepper with a low fineness level exhibited a higher monomolecu- rature can cause the rapid oxidation of essential oils, while
lar layer volume and a greater sorption-specific surface area. The excessively high humidity can promote the development
analysed pepper samples did not differ significantly in terms of of moulds (in particular) but also of bacterial flora. On the
porosity or capillary capacity. The GAB and Peleg models empiri-
cally described the determined sorption isotherms of ground black
other hand, light, mainly in the UV range, can cause chang-
pepper in the full range of water activity very well. es to the chemical components of the spice. As a result of
K e y w o r d s: sorption models, storage stability, fineness, sur- these changes, the quality of the pepper may deteriorate
face phenomena, black pepper significantly.
Dried spices are currently widely available, have a high
INTRODUCTION degree of popularity and are among the spices, which are
Black pepper (Piper nigrum L.) is one of the best known most frequently chosen by consumers. Their end user value
and most frequently used spices worldwide. Major world is determined by both the time of harvesting and the me-
exporters of whole grain black pepper include Vietnam thod and conditions of drying, which are intended to reduce
(39% of the world’s production), Indonesia (17% of the the water content to the minimum level set by the relevant

© 2020 Institute of Agrophysics, Polish Academy of Sciences

*Corresponding author e-mail: a.ocieczek@wpit.umg.edu.pl
162 A. OCIECZEK et al.

standards (7-14%). Food dehydration is among the most a glassy state should be protected from water adsorp-
common processes applied to improve food stability as tion, e.g. by choosing the appropriate packaging material
it significantly reduces microbial activity and minimizes (Lewicki, 2006).
enzymatic, chemical and physical changes during storage The concept of water content equilibrium is an impor-
(Babetto et al, 2011; Triwahyudi et al., 2015). The dehy- tant factor in the process of drying agricultural products.
dration of food prolongs its shelf-life while preserving A knowledge of water content equilibrium may be used to
its sensory characteristics, thus making it similar to fresh determine the duration of product drying, the parameters
products (Sacilik and Unai, 2005; Triwahyudi et al., 2015). of the drying equipment, and the costs associated with the
Drying is probably one of the oldest methods of food drying process (Akanbi et al., 2006; Chowdhury et al.,
preservation. It involves the removal of water until its con- 2006; Triwahyudi et al., 2015). Research into the shelf-
tent reaches a final level, which provides a product with life of food includes the application of static methods for
microbiological stability and guarantees its expected stor- determining the water content equilibrium, which involves
age life. In view of the phenomena, which occur after the leaving a sample to equilibrate in humid air, the degree of
completion of the drying process, it should be noted that saturation is usually determined by an appropriate saturated
a dry product is not in a thermodynamic equilibrium state chemical solution used to maintain a stable relative humid-
(Newman and Zografi, 2019). Water activity and mobil- ity environment.
ity are determined by water content and the temperature Besides research into the equilibrium water content
of the material. At higher water content values or higher of food, research into the water activity of these products
temperatures, both the activity and mobility of molecules has been carried out. It has been known for more than half
which are in a liquid state are high, and the material is sus- a century that the relative pressure of vapour above a prod-
ceptible to physical, chemical and biological degradation. uct, related to this product’s water activity, can play a role
For this reason, products should be dried to a final water much more significant to the quality and shelf-life of the
content level that will guarantee their expected shelf-life
food than the total amount of water contained in it. The
(Lewicki, 2006). The provision of adequate barrier packag-
preference for water activity (aw) over water content as
ing materials for the storage of dried materials is of equal
a reference parameter for most practical applications relat-
significance.
ed to food processing and storage is a result of the fact that
The quality of dried plant spices characterized by rich
water activity is a determinant of microbial growth which
aroma and flavour is determined by their complex and var-
affects most chemical, enzymatic and physical degradation
ied chemical composition. These properties are constantly
reactions. It also has an effect on water migration in food;
changing, primarily due to the surface phenomena occur-
along with the equilibrium water content (water vapour
ring during their storage.
sorption isotherms), it allows for the estimation of the mon-
In the retail trade in Poland, two basic forms of black
pepper are found: whole grained and ground. The process olayer used to determine the optimal water content of food
of pepper grinding expands its surface area, which may products preserved by drying and is relatively straightfor-
result in an increase in the dynamics of surface phenomena ward to measure using modern techniques, its measurement
related to the effects of water vapour. is also non-invasive to the analysed product.
The most stable form of natural substances (including The aim of the study was to compare the sorption pro-
food) is the crystalline state, for example, the typical native perties of black pepper of different fineness levels, which
form of starch granules. Slade and Levine (1991) found are a significant factor in the determination of hygrosco-
that solids in a glassy state also exhibit a favourable degree picity. The results obtained were used to compare the
of stability. At least some of the dried products that are in analysed samples in terms of storage stability. Another aim
a glassy state should remain in this state during storage. The was to assess the usefulness of selected mathematical sorp-
change from a glassy state to a rubbery state at a constant tion models in order to describe the properties of pepper.
water content occurs within a certain temperature range Moreover, selected parameters of pepper particle surface
specific to each product, and is characterized by a mean microstructure as well as the parameters characterizing the
temperature referred to as the glass-transition temperature energy phenomena accompanying adsorption are factors
(Tg). At a temperature lower than Tg, the product is in which may be used to predict stability during storage.
a glassy state, whereas when the temperature is exceeded, The achievement of the aim mentioned above was
the product transitions to a rubbery state, its structure dis- based on the verification of the research hypothesis, which
integrates and the taste and aroma compounds are released assumed that the pepper fineness level has a significant
(Roos, 1995). The Tg temperature level is closely corre- effect on its sorptivity in terms of the interaction with water
lated with water content, since water acts as a plasticizer, vapour present in the environment, which in turn has an
and an increase in its content contributes significantly to influence on its hygroscopicity which determines stability
a reduction in Tg. This means that a product remaining in during storage.
 COMPARISON OF SORPTION PROPERTIES OF BLACK PEPPER 163

MATERIAL AND METHODS The GAB equation was in the following form:
The study material comprised two commercial black
pepper samples differing in fineness level, both were pur-
chased at a store in the city of Gdańsk: “Prymat” brand (2)
black pepper of a high fineness level (D) and “Prymat”
brand black pepper of a low fineness level (G). where: CG – Guggenheim energy constant; K – a con-
The water content was determined by thermal drying stant, which corrects the properties of molecules forming
to a constant weight at a temperature of 373-378 K (100- the multilayer compared to the liquid phase (Figura and
105°C) under normal pressure. Teixeira, 2007).
Water activity (aw) at a temperature of 293.15 K (20˚C) Peleg’s equation was in the form:
was determined with an accuracy of ±0.003 using an
(3)
AquaLab apparatus (version AS4 2,14.0 2017, series
4TE and 4TEV, manufactured by Decagon Devices, Inc., where: A, B, C, D – constants (Figura a nd Teixeira, 2007).
Pullman, WA, USA). The parameters of selected equations were determined
Sorption isotherms were analysed by the static desic- using non-linear regression and a Monte Carlo algorithm,
cator method using saturated solutions of the following which prevented the estimation process from being halted
substances: NaOH H2O (0.0698); LiCl H2O (0.1114); by the local minimum. The objective function was the min-
KC2H3O7 1.5 H2O (0.231); MgCl2 6 H2O (0.3303); K2CO3 imization of the sum of squares deviation (SSD) (Ocieczek
2 H2O (0.440); Na2Cr2O7 2 H2O (0.548); KJ (0.6986); NaCl and Kostek, 2009). The calculations were performed using
(0.7542); KCl (0.8513); KNO3 (0.932); K2Cr2O7 (0.9793). an Excel 2010 spreadsheet. Standard error values for the
Testing was conducted within the environmental water determined parameters of particular equations were esti-
activity range of 0.07 to 0.98 and at 293.15 K (20°C). The mated using the SolverAid macro command based on
duration of the system equilibrium determination was 90 the Hessian matrix. The usefulness of the model for the
days from the time at which the samples were placed in description of experimental data were assessed based on
desiccators. The samples incubated in desiccators with
the root mean square (RMS) expressed in % (Pałacha and
a water activity of more than 0.7 were protected from micro-
Sas, 2016):
floral development. In the first phase of the experiment,
pepper samples intended for the determination of sorption ve − vo 2
isotherms were placed in the amount of approximately ∑( ve
)
(4)
1 g ± 0.1 mg in weighing bottles with a diameter of approx. RMS = ⋅100% ,
N
35 mm in a desiccator containing P2O5 as a drying agent at
room temperature for three weeks in order to minimize the where: N – number of data, ve – experimental equilibrium
moisture content (~2%) of the samples. The weights of the water content (g H2O 100 g-1 d.m.); vo – forecasted equilib-
samples were then determined and before they were placed rium water content (g H2O 100 g-1 d.m.).
in desiccators with saturated solutions of appropriate sub- The specific surface area of adsorbent was calculated
stances. Based on the initial product weight (determined using the following equation:
after 3 weeks of incubation in a desiccator with P2O5) and
changes to the water content, the equilibrium water con- (5)
tents were calculated and adsorption isotherms were plotted
where: asp – specific surface area of sorption (m2 g-1); N
using EXCEL 2010 software. The measurement of water
– Avogadro’s number (6.023 1023 molecules mol-1); M –
activity in the samples was performed 90 days after placing
molecular mass of water (18 g mol-1); ω – cross-sectional
them in desiccators using an AquaLab apparatus.
area of water (1.05 10-19 m2 molecule-1) (Pérez-Alonso et
The diversity of the course of sorption isotherms within
al., 2006).
the entire aw range was statistically analysed using a t-Stu-
The size and volume of the capillaries of the material
dent test for the differences between the means for paired
studied was determined for the condensation area using
samples, with the differences at a significance level not
the Kelvin equation which assumes that their shape is
exceeding a P value of 0.05, which is regarded as statisti-
cylindrical:
cally significant.
The BET equation was in the following form:
(6)
(1)
where: aw – water activity (–), v – equilibrium water content where: σ – surface tension of liquid at temperature T (N m-1),
(g H2O 100 g-1 d.m.); vm – water content of the monolay- rk – capillary radius (nm); R – universal gas constant
er (g H2O 100 g-1 d.m.); C – energy constant (Figura and ((kJ(mol K)-1); T – process temperature (K), v – molar volu-
Teixeira, 2007). me of the adsorbate (m3 mol-1) (Figura and Teixeira, 2007).
164 A. OCIECZEK et al.

RESULTS AND DISCUSSION the same time, it should be emphasized that the identified
Water contained in food is present in various fractions, water activity level prevented the product from maintaining
as determined by the resultant force of interactions occur- microbiological stability (and thus storage stability as well)
ring between the components of a solid and water. The type during storage.
and rate of the processes occurring in the product (and thus The sorption properties of dried spices, i.e. products
its shelf-life) are determined by the water fraction dominant with hygroscopic properties, clearly determine their stor-
in the product. Most methods for preserving food, includ- age stability. Sorption isotherms are an effective tool for
ing dried spices, involve a reduction in the amount of water investigating these properties. A mathematical description
contained in a product in order to decrease water activity. of these curves using theoretical or empirical models in
Hygroscopicity is a characteristic of products with connection with a physical interpretation of the significance
a high degree of dehydration which, while moving towards of parameters of some of these models is an appropriate
a dynamic equilibrium with the commonly more humid approach to research into the inherent properties of all types
environment, absorb water. The hygroscopicity of each of hygroscopic powders. Knowledge of the isotherm shape
substance is primarily determined by the state of water indicates the mechanism which determines the way water is
present in this substance. The preliminary assessment of bound in a product, the product sensitivity to water vapour
the hygroscopicity of pepper included the determination of in the atmosphere, and it also helps to predict changes dur-
water content and activity (Table 1). ing the storage of a particular product.
Pepper with a low fineness level (G) was characterized Adsorption isotherms of products D and G correspond-
by a statistically significantly higher water content on a dry- ed to type II in the classification proposed by Brunauer
matter basis compared to pepper with a high fineness level (Fig. 1), which is an unequivocal confirmation that the
(D) (F-Test = 0.9371; p value = 0.000). It may be assumed ground pepper particles are porous bodies.
that the reduction in water content was due to the grind- In isotherms of this type, three areas expressing three
ing process accompanied by an increase in temperature and mechanisms influencing the course of adsorption phenom-
thus an increase in water vapour pressure that promotes the enon can be distinguished. Isotherms of both analysed
phenomenon of its desorption. Pepper of a low fineness pepper samples (D) and (G) were constant within the entire
level (G) was also characterized by a higher water activity, water activity range, which indicates that no crystallization
which was determined by its higher content (Table 1). At of any components occurred during the surface absorption
of water vapour. In accordance with Gondek and Lewicki
Ta b l e 1. Water content and activity in the analysed black pepper (2005), it was assumed that the isotherms of ground pepper
samples would also not exhibit the hysteresis phenomenon either,
Water
whose presence is associated with at least the partial sus-
Water content ceptibility of adsorbent particles to dissolution and the
SD activity SD
(g H2O 100 g-1 d.m.) crystallization of sugars contained in it. Roman et al. (1982)
(-)
Product G and Tsami et al. (1990) are of the opinion that the water
activity value at which hysteresis is observed is inversely
13.62 0.33 0.680 0.003
proportional to the sugar content. Ground black pepper is
Product D
primarily characterized by the presence of crystalline starch
10.62 0.35 0.445 0.003 with a negligible number of amorphous areas determined
by the peripheral damage to starch granules during grain
25
G D
Water content (g H2O 100 g‐1 d.m.)

20

15

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water activity (‐)
Fig. 1. Water vapour adsorption isotherms determined at 20°C on both black pepper samples (G and D).
 COMPARISON OF SORPTION PROPERTIES OF BLACK PEPPER 165

grinding. Nevertheless, it should be taken into account that selection of the most suitable sorption model include the
grinding leading to damage in “crystalline” materials to degree of fitting to experimental data and the simplicity of
which natural starches belong can result in an increase in the model.
hygroscopicity, and thus contribute to a reduction in physico- Knowledge of the shape and location of water adsorp-
chemical stability. This, in turn, can potentially decrease tion isotherms in mathematical terms allows for, in the case
storage shelf-life (Newman and Zografi, 2019). of theoretical models, the estimation of selected parameters
Ahlneck and Zografi (1990) demonstrated that heavily of the material surface microstructure and energy phe-
dehydrated materials characterized by the presence of at nomena accompanying the sorption process. To this end,
least a partially amorphous and, thus, physically differen- BET and GAB models whose parameters have a specific
tiated structure of the matrix, are prone to intense water physical meaning are applied. The determination of the
absorption particularly by these especially amorphous monolayer volume, which is a parameter of both the BET
and GAB model, enables the estimation of the specific sorp-
areas. In such a situation, even a seemingly insignificant
tion capacity. On the other hand, the determination of the
increase in water content will actually form a water pool of
energy constant serves to identify the difference between
importance to product stability, located in a small area of
the chemical potential of the particles of adsorbate in the
the sample, which is referred to as the “amplification” of
pure state and in the first adsorption layer (BET) or the dif-
the water effect. Even a statistically insignificant increase
ference between the chemical potential of the particles of
in the water amount can then change the molecular mobi- adsorbate in the first adsorption layer and in the higher lay-
lity in these unstructured regions and initiate changes that ers (GAB) (Timmermann et al., 2001). These parameters
result in both physical and chemical instability (Buckton, are regarded as useful for predicting the optimal storage
1997) and, consequently, reduce the shelf-life or determine conditions and the storage stability of (particularly dehy-
the differentiation of technological characteristics. drated) food.
The course of the isotherms of both analysed pepper The BET and GAB equations are an extension of the
samples was particularly similar at the sites correspond- multimolecular concept proposed by Langmuir and they
ing to monomolecular adsorption (tcalc.= 1.803; tcrit.= 2.228) allow for a description of sigmoidal isotherms commonly
and multilayer adsorption (tcalc.= 2.153; tcrit.= 2.228), while observed in the context of food and other material of bio-
differing, though to a statistically insignificant extent, logical origin.
(tcalc.= 1.674; tcrit.= 2.228) within the capillary condensa- The BET equation is the most frequently applied
tion area. It may be assumed that greater differences in the model used to investigate surface phenomena occurring
course of both sorption isotherms in the capillary condensa- in food, and is a major breakthrough in the interpretation
tion area were due to the expansion of the pepper surface of multilayer sorption isotherms, particularly type II and
resulting from its different fineness levels and not due to III. Moreover, it is an effective method for estimating the
the differences in the chemical composition of the powders. amount of water bound to the polar sites of dehydrated
While aiming at expressing the relationship between food. This model assumes that the sigmoidal shape of the
water activity and its content in food in mathematical isotherm is due to multilayer adsorption. According to this
terms, various regression models have been developed. assumption, each adsorbed molecule becomes an adsorp-
The parameters of these models are partial regression coef- tion centre for the next adsorbate molecule. The forces
ficients, which describe each of the three sorption isotherm occurring during the formation of subsequent layers are
analogous to those, which cause vapour-to-liquid conden-
zones. In many cases, known models that exhibit a use-
sation. The model can be used to calculate the content of
fulness for describing the sorption properties of only one
water bound in the monomolecular layer on the assump-
particular food product are ineffective for describing oth-
tion that the adsorption surface is homogeneous and that
ers. Some of the models also exhibit an adequate predictive
no interactions take place between the adsorbed water mol-
power only within certain water activity ranges (Andrade ecules. At the same time, it should be stressed that the BET
et al., 2011). equation, regardless of whether it has been determined by
Some of the mathematical sorption isotherm models the analytical or numerical method, accurately describes
were derived from the theoretical foundations of the the course of the sorption phenomenon within a limited
adsorption mechanism description, while others are either water activity range (0.05-0.5) (Rizvi, 1995).
empirical in nature or simplified versions of more complex Figures 2a and b show the course of the sorption iso-
models. The following are most commonly used to describe therms for both tested black pepper samples, empirically
sorption phenomenon in food products: the Langmuir equa- determined and theoretically calculated based on the BET
tion, the BET equation, the GAB model, the Oswin model, equation. A graphic interpretation was applied to the entire
the Smith model, the Halsey model, the Henderson model, water activity range to reflect the limitations of the BET
the Iglesias-Chirife equation and the Peleg model (Sahin equation in the description of surface phenomena to a nar-
and Gülüm, 2006). The criteria taken into account in the rower range.
166 A. OCIECZEK et al.

a 70
G emp.

Water content (g H2O  100 g‐1 d.m.)
60 G BET
50

40

30

20

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water activity (‐)

b 60
D emp.
Water content (g H2O  100 g‐1 d.m.)

50 D BET

40

30

20

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water activity (‐)

Fig. 2. Empirical and theoretical (BET) isotherm of water vapour adsorption by: a – pepper G and b – pepper D particles.

Considering the limitations of the BET model, its The water content corresponding to the monomolecular
parameters were estimated within the range indicated as layer is the minimum and, at the same time, the optimal
those corresponding to a good fit to the empirical data. water content due to the durability of the dried product.
The values of the deviation square sums and of the errors Water in an amount equivalent to the monomolecular coat-
with which they were determined indicate that the BET ing is characterized by low activity, does not freeze at
model accurately describes the process of water vapour -40°C, is not available as a solvent, has no softening prop-
adsorption on the ground black pepper surfaces and with erties and does not contribute to hydrolytic degradation but
a similar accuracy, irrespective of the sample fineness level rather protects against oxidative changes both in the aque-
(Table 2). This finding is also confirmed by the determined ous and lipid phases. A knowledge of this parameter allows
RMS values, which indicate that the BET model is cha- for the determination of the final drying point and thus, the
racterized by a very good fit to empirical data (RMS values optimization of the costs of technological processes and it
lower than 10%). is also of significance in food storage and transport (Gal,
1983). According to Karel (1975), most food products are
Ta b l e 2. BET equation parameters of the analysed black pepper characterized by a monomolecular layer ranging from 4
samples to 11 kg per 100 kg of dry matter. The monolayer (vm) is
regarded as an indicator of the availability of polar sites for
Parameters Product G Product D water vapour irrespective of which component is a source
of hydrophilic groups. The vm is determined by the number
SSD (-) 1.2425±0.6436 0.7849±0.5115
of particular components, which are abundant in polar sites
RMS (%) 5.0439 4.1476 and by their physical state, by determining the amount of
-1
νm (g H2O 100 g d.m.) 3.154±1.221 2.901±1.297 water, which is firmly bound to the food matrix (Ocieczek,
C (-) 1.431±0.375 1.478±0.442 2012). Excess water, as compared to the monomolecular
layer, results in a critical water content which, if exceeded,
 COMPARISON OF SORPTION PROPERTIES OF BLACK PEPPER 167

causes various undesirable changes in the product that products is perceived to be a continuous transition from
impair its quality and enable microbial development. The bound water to free water. In order to provide a description
monolayer values determined are based on a BET transfor- of the phenomenon of sorption on a solid surface that is
mation ranging from 2.90 to 3.15 g H2O per 100 g of dry more precise than the BET model, the GAB model contains
matter of the product, and were lower than those determined a third constant (K) which specifies the chemical potential
by Karel (1975). The differences in monolayer values for difference between the molecules forming the multilayer
the analysed pepper samples may indicate the differentia- (described as the 2nd area of the sorption isotherm) and the
tion of the dry matrix state, which is determined by the chemically pure water molecules.
fineness level. The structure of starch granules, which are The advantage of the GAB model is that a physical
a dominant component of black pepper grains, may under- interpretation related to the course of the sorption phe-
go changes due to their disaggregation during grinding. nomenon was assigned to its parameters. It also enabled
The results obtained indicate that the differences in mon- an analysis of the effects of temperature on isotherms
olayer extension should not only be related to the pepper using Arrhenius-type equations (Al-Muhtaseb et al., 2002).
grain fineness level. It appears that the state of the particles A disadvantage of the GAB model is that it understates the
formed in this way (which are primarily comprised of starch predicted equilibrium water contents corresponding to high
granules of different levels of damage to the semi-crystal- levels (aw > 0.93) of water activity (Fig. 3a and b). Above
line structure due to the applied grinding method) is also of these values, the GAB model loses its great predictive
significance. The presence of amorphous areas formed in power. It should therefore be stressed that the method using
this way can intensify the phenomenon of surface absorp- saturated substance solutions in the determination of sorp-
tion, which results in a local increase in hygroscopicity and tion isotherms should not be applied within the full water
the “amplification” of the water effect. This phenomenon activity range if the data are to be subject to a mathemati-
was concisely described by Buckton (1997). cal description of the sorption curve using the GAB model.
The estimated values of the monolayer vm provided
The GAB equation parameters identified while taking into
a basis for the calculation of the specific sorption surface,
account the above-described limitations along with the
which ranged from 101.9 (D) to 110.8 m2 g-1 (G). On the
error values are provided in Table 3.
other hand, the BET model was not applied to calculate
A comparison of the SSD and RMS values allows
the general capillary capacity because the model exhibits
for the conclusion to be reached that the GAB model, by
a constant tendency towards a significant over-estimation
describing the course of the phenomenon of sorption on the
of the equilibrium water content after exceeding the point
analysed pepper sample surfaces within a wider range aw
corresponding to the beginning of the capillary condensa-
compared to the BET model, also allowed for the empirical
tion phenomenon (Fig. 2a and b). This area provides a basis
data to be reflected better than by using the BET model.
for this estimation.
This is because it is assumed that the root mean square
The energy constant C reflects the difference between
the enthalpy of desorption from the monolayer and the (RMS) value at a level lower than 10% indicates a good
enthalpy of fluid adsorbent evaporation. Low constant C fit of the model to sorption data within the entire analysed
values indicate that the process observed on the surface of range aw (Pałacha and Sas, 2016).
analysed pepper samples was by nature a physical adsorp- The vm values determined based on the GAB model
tion process. were higher than those for the BET model, and fell within
Another theoretical model used to describe both the the range of 6.41 to 8.03 g H2O per 100 g of dry matter of
sorption properties of black pepper and selected surface the product, which corresponded to the range provided by
microstructure parameters was the GAB model, which has Karel (1975).
many advantages over other models. One of the key proper- The Guggenheim energy constant CG ranged from
ties of the GAB model is that it has a realistic theoretical 16.59 for pepper with large particles (G) to 26.17 for pep-
background, because it is an improved version of the physi- per with small particles (D). These values indicate that the
cal adsorption theory by Langmuir and includes aspects GAB model describes the results of the study very well as
of the BET model. The GAB model is an extension of the the CG parameter took on values greater than 5.67, which,
multilayer adsorption theory and a modified version of the according to Lewicki (1997), indicates that the model and
BET model. It postulates that the state of sorbate particles its parameters were properly selected for the description
in the second layer is identical to that in higher layers but of the empirical data. In turn, according to Diosady et al.
differs from the liquid state. The water molecules, which (1996), strong exothermic interactions between the adsor-
form the second and subsequent layers, usually fill small bent and the adsorbate decrease the process temperature
capillaries and are more loosely bound to the solid matrix and increase the CG value. In the context of these records, it
than the monomolecular water layer. Thus, they contribute may be concluded that the process of surface water absorp-
to the progressive loosening of the structure and expose tion for both analysed black pepper samples was physical
additional polar sites. This fraction of water found in food in nature.
168 A. OCIECZEK et al.

a 18
G emp.

Water content (g H2O  100 g‐1 d.m.)
16
G GAB
14

12

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water activity (‐)

b 25
D emp.
Water content (g H2O  100 g‐1 d.m.)

D GAB
20

15

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water  activity (‐)

Fig. 3. Empirical and theoretical (GAB) water vapour adsorption isotherm of samples: a – G and b – D.

Ta b l e 3. The GAB equation parameters of black pepper samples cal value of the constant CG greater than 5.5. Moreover,
and aw corresponding to νm according to Caurie (2006), the K value enables the dif-
ferentiation of monomolecular (K ≤ 0.5) and multilayer
Parameters Product G Product D
(K > 0.5) adsorption. On the other hand, Chirife and Iglesias
SSD (-) 0.208 ± 0.161 0.338 ± 0.220 (1992) concluded that the K value is determined by the type
RMS (%) 1.4775 3.3212 of the product being described. Where absorption occurs on
-1
νm (g H2O 100 g d.m.) 8.028 ± 0.270 6.410 ± 0.367 the surface of protein structures, the K parameter values fall
K (-) 0.612 ± 0.016 0.772 ± 0.025 within the range of 0.82 to 0.86, while for starch structures
CG (-) 16.60 ± 2.01 26.17 ± 12.72 they ranged from 0.7 to 0.77. On the other hand, Pérez-
Alonso et al. (2006) indicated that for products in which the
The value of K, a constant, which corrects the properties sorption properties are largely determined by proteins, the
of molecules forming the multilayer as compared to the liq- K constant values fall within a wider range of 0.84 to 1.00,
uid phase, took on values ranging from 0.6125±0.0157 (G) while for the dominant role of fibre, the K constant takes
to 0.7718±0.0248 (D) (Table 3). The obtained results indi-
on values ranging from 0.717 to 0.893. Taking into account
cate that this equation was appropriately used to describe
the specificity of the analysed material, the obtained K
the experimental data, since Lewicki (1997) demonstrated
that one of the conditions for the correct application of the values should be regarded as the result of interactions
GAB equation to describe the experiment is the K value occurring during the absorption of water molecules by the
falling within the range of 0.24 to 1. Where the numeri- heterogenous structure of ground pepper grains rather than
cal value of the constant K falls within the indicated range, a simple result of the interactions occurring between pro-
maintaining the error of calculation of the water content teins and starch. Therefore, the sorption properties of the
of the monolayer at a level of ±15.5% requires a numeri- analysed black pepper samples of different fineness levels
 COMPARISON OF SORPTION PROPERTIES OF BLACK PEPPER 169

were determined by the significant percentage of the starch expectations. This is because the grinding process results in
fraction as well as by their physical diversity related to the a significant expansion of the specific surface area, which
fineness level. usually contributes to an increase in the dynamics of surface
Based on the previously calculated GAB equation phenomena related to the effects of water vapour, which is
parameters or, more precisely, the extension of the mon- referred to as hygroscopicity. As a result of grinding, pep-
olayer (vm) and the range of points describing the adsorption per should become a highly hygroscopic product, due to
isotherms in the area of capillary condensation, selected the increase in the area of contact with the environment and
parameters of the surface microstructure of the black pep- a significant reduction in water content during the grinding
per samples of different fineness levels were evaluated process, which is accompanied by an increase in tempera-
(Table 4). ture that promotes drying. At the same time, it should be
Large-particle pepper (G) was characterized by a great- stressed that in the process of pepper grinding, the starch
er sorption specific surface area than very fine-particle matrix (characterized by an ordered, crystalline structure)
pepper (D), which may intuitively appear to be contrary to is primarily expanded. This, in turn, does not necessarily
result in a significant increase in sorptivity. It should also
Ta b l e 4. Structural characteristics of the tested products be noted that large-particle pepper (G) was characterized
Specific surface General capacity of by a higher content of water which, by interacting with the
Capillary radius protein fraction, may bring about structural changes in this
of sorption asp capillaries
at aw = 0.65 (nm) fraction that lead to the exposure of new, previously hidden
(m2 g-1 d.m.) (mm3 100 g-1 d.m.)
Product G hydrophilic groups and thus contribute to an increase in the
specific sorption surface. The presence of water in a product
282 65.48 1.48
measurably contributes to shaping its quality (the product’s
Product D
characteristics) which may be perceived in the context of
225 67.28 1.50 inter alia hygroscopicity or stability during storage.

a 18
G emp.
Water content (g H2O 100 g‐1 d.m.)

16
G Peleg
14

12

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water activity (‐)

b 25
D emp.
Water content (g H2O 100 g‐1 d.m.)

D Peleg
20

15

10

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Water activity (‐)

Fig. 4. Empirical and theoretical (Peleg) water vapour adsorption isotherm of samples: a – G and b – D.
170 A. OCIECZEK et al.

The data generated based on the GAB model were also The RMS values of the sorption isotherm estimation
used to calculate the general capillary capacity of the test- using the Peleg model ranged from 1.53 to 3.52%, which
ed material as the total volume of water adsorbed by the proves that the model described the experimental data very
material within the water activity range of 0.65 to 0.93 was well, irrespective of the differences in the course of the sur-
due to the model’s usefulness in describing sorption curves face absorption phenomenon, which are determined by the
within a range wider than that of the BET model (Figura physical diversity of the analysed pepper samples. A com-
and Teixeira, 2007; Ocieczek, 2012). It was demonstrated parison with Peleg’s equation parameters indicates that the
that irrespective of the pepper fineness level, the surface analysed pepper samples differed significantly from each
structure did not differ significantly in terms of porosity, other in terms of their sorption properties.
which was reflected in the comparable values of the gen- A comparison of the statistics used to assess the predic-
eral capillary capacity (Table 4). At the same time, under tive power of particular models allowed for an unequivocal
conditions initiating capillary condensation, capillaries of demonstration that both the GAB and Peleg models met the
a similar radius were filled on the surface of the pepper par- criterion of a very good model fit to the experimental data.
ticles irrespective of the fineness level. Moreover, taking into account the theoretical nature of the
In summary, it may be assumed that the size of pores as GAB model, one may indicate its greater usefulness by
reflected by their radius was the factor determining the gen- conducting research into the surface phenomena occurring
eral capillary capacity of the analysed products. The same in dehydrated food, which determine its storage stability.
conclusions were reached by Ocieczek and Ruszkowska
(2018) in a study concerning the sorptivity of quinoa CONCLUSIONS
seeds. The simultaneously obtained results indicate that 1. Isotherms of water vapour sorption on the surface
irrespective of the pepper fineness level, its particle sur- of black pepper particles of different fineness levels were
face structure is similar. This, in turn, may be interpreted
characterized by a course typical of the type II isotherm
as evidence of the dominant role of the starch fraction with
according to the Brunauer classification.
an ordered structure in shaping the sorption properties of
2. Commercial samples of black pepper of different
ground pepper and the significant role of water in the modi-
fineness levels (G and D) were characterized by water con-
fication of the surface structure through its interaction with
tents, which varied to a statistically significant extent, the
the hydrophilic protein fraction.
content of water was determined by its activity and was
In order to describe the sorption isotherms of pepper
probably the main reason for the differentiation of the sorp-
of different fineness levels, the four-parameter Peleg model
tion properties of the samples. Pepper with a low fineness
was also applied. Figures 4a and b show the course of the
level exhibited a higher monomolecular layer volume and,
sorption isotherms for both tested black pepper samples,
consequently, a greater sorption specific surface area. The
empirically determined and theoretically calculated based
analysed pepper samples did not differ significantly in
on the Peleg equation. This model is described as an empir-
terms of porosity or capillary capacity.
ical equation devoid of a theoretical background (Andrade
3. The GAB and Peleg models described the empirical-
et al., 2011). The results of numerous studies, including this
paper, indicate that the Peleg model exhibits a similar use- ly determined sorption isotherms of ground black pepper
fulness for describing water vapour sorption as that of the very well within the full range of water activity. The cal-
GAB model (Table 5). culated root mean square for these models oscillated at
a level, which was undoubtedly lower than 10%.

Conflict of interest: The Authors do not declare con-

Ta b l e 5. Peleg’s equation parameters flict of interest.

Parameters Product G Product D REFERENCES

SSD (-) 2.4032 ± 0.6329 0.2161 ± 0.1898 Ahlneck C. and Zografi G., 1990. The molecular basis of mois-
ture effects on the physical and chemical stability of drugs
in the solid state. Int. J. Pharm., 62, 87-95.
RMS (%) 3.5222 1.5306
https://doi.org/10.1016/0378-5173(90)90221-o
Akanbi C.T., Adeyemi R.S., and Ojo A., 2006. Drying charac-
A 8.39 ± 1.92 15.22 ± 0.55
teristics and sorption isotherm of tomato slices. J. Food
Eng., 73(2), 157-163.
B 0.40 ± 5.80 0.61 ± 0.04
https://doi.org/10.1016/j.jfoodeng.2005.01.015
Al-Muhtaseb A.H., McMinn W.A.M., and Magee T.R.A., 2002.
D 8.71 ± 1.92 23.66 ± 2.22
Moisture sorption isotherm characteristics of food pro-
ducts: A review. Food Bioprod. Process., 80(2), 118-128.
E 1.03 ± 9.13 8.76 ± 0.97
https://doi.org/10.1205/09603080252938753
 COMPARISON OF SORPTION PROPERTIES OF BLACK PEPPER 171

Andrade R.D., Lemus R.M., and Pérez C.C., 2011. Models of Ocieczek A., 2012. Hydration properties of machine wheat flours
sorption isotherms for food: uses and limitations. Vitae, as a discriminant of the usable quality (in Polish). Scientific
J. Faculty of Pharmaceutical Chemistry, 18(3), 325-334. Works of the Gdynia Maritime University, Gdynia, Poland.
Babetto A.C., Freire F.B., Barrozo M.A.S., and Freire J.T., Ocieczek A. and Kostek R., 2009. Sorptive properties of type 2000
2011. Drying of garlic slices: Kinetics and nonlinearity wheat and rye flours. Acta Agrophysica, 14(2), 393-402.
measures for selecting the best equilibrium moisture con- Ocieczek A. and Ruszkowska M., 2018. Comparing sorption
tent equation. J. Food Eng., 107, 347-352. properties of grains of selected Quinoa Varieties (Chenopo-
https://doi.org/10.1016/j.jfoodeng.2011.07.007 dium quinoa Willd.) (in Polish). Żywność, Nauka,
Buckton G., 1997. Characterisation of small changes in the phy-
Technologia, Jakość, 25, 3(116), 71-88.
sical properties of powders of significance for dry powder
Pałacha Z. and Sas A., 2016. Sorption properties of selected spe-
inhaler formulations. Adv. Drug Deliv. Rev., 26, 17-27.
https://doi.org/10.1016/s0169-409x(97)00507-3 cies of rice (in Polish). Acta Agroph., 23 (4), 681-694.
Caurie M., 2006. The derivation of the GAB adsorption equation Pérez-Alonso C., Berstain C.I., Lobato-Calleros C., Rodriguez-
from the BDDT adsorption theory. Int. J. Food Sci. Technol., Huezo M.E., and Vernon-Carter E.J., 2006. Thermo-
41(2), 173-179. dynamic analysis on the sorption isotherms of pure and
https://doi.org/10.1111/j.1365-2621.2005.01044.x blended carbohydrate polymers. J. Food Eng., 77(4), 753-
Chirife J. and Iglesias H.A., 1992. Estimation of precision of 760. https://doi.org/10.1016/j.jfoodeng.2005.08.002
isosteric heat of sorption determined from the temperature Rizvi S.S.H., 1995. Thermodynamic properties of food in dehy-
dependence of food isotherms. LWT, 25(1), 83-84. dration. In: Engineering Properties of Foods (Eds M.A.
Chowdhury M.M.I., Huda M.D., Hossain M.A., and Hassan Rao, S.S.H. Rizvi). Marcel Dekker Inc., New York-Basel-
M.S., 2006. Moisture sorption isotherms for mungbean Hong Kong.
(Vigna radiata L). J. Food Eng., 74, 462-467. Roman G.N., Urbicain M.J., and Rotstein E., 1982. Moisture
https://doi.org/10.1016/j.jfoodeng.2005.03.036 equilibrium in apples at several temperatures. Experimental
Diosady L.L., Rizvi S.S.H., Cai W., and Jagdeo D.J., 1996. data and theoretical considerations. J. Food Sci., 47,
Moisture sorption isotherms of canola meals, and applica- 1484-1492.
tions to packing. J. Food Sci., 61, 204-208.
https://doi.org/10.1111/j.1365-2621.1982.tb04966.x
https://doi.org/10.1111/j.1365-2621.1996.tb14760.x
Roos Y.H., 1995. Phase transitions in foods. CA: Academic Press
Figura L.O. and Teixeira A.A., 2007. Food Physics. Physical
Properties, Measurement and Applications. Springer- San Diego.
Verlag, Berlin Heidelberg. Sacilik K. and Unai G., 2005. Dehydration characteristics of
Gal S., 1983. The need for, and practical applications of sorp- Kastamonu garlic slices. Biosys. Eng., 92(2), 201-215.
tion data. In: Physical Properties of Foods (Eds R. Jowit, https://doi.org/10.1016/j.biosystemseng.2005.06.006
F. Escher, B. Hallstrom, H.F.T. Meffert, W.E.L. Spiess, Sahin S. and Gülüm S., 2006. Physical properties of foods. New
G. Vos). Applied Science Published, New York. York, United States: Springer.
Gondek E. and Lewicki P.P., 2005. Moisture sorption isotherms Slade L. and Levine H., 1991. Beyond water activity: Recent
of dried and candided fruits (in Polish). Acta Sci. Pol., advanced based on alternative approaches to the assessment
Technol. Aliment., 4(1), 63-71. of food quality and safety. Critical Reviews in Food Science
Karel M., 1975. Water activity and food preservation. In: Physical and Nutrition, 30, 115-360.
Principles of Food Preservation. Principles of Food Science. https://doi.org/10.1080/10408399109527543
Part 2 (Eds M. Karel, O.R. Fennema. D.B. Lund). Marcel Timmermann E.O., Chirife J., and Iglesias H.A., 2001. Water
Dekker, New York. sorption isotherms of food and foodstuffs: BET Or GAB
https://doi.org/10.1201/9780203911792 parameters? J. Food Eng., 48, 19-31.
Lewicki P.P., 1997. The applicability of the GAB model to food
https://doi.org/10.1016/s0260-8774(00)00139-4
water sorption isotherms. Int. J. Food Sci. Technol., 32(6),
Triwahyudi S., Rahardjo B., Nelwan L.O., and Wulandani D.,
553-557. https://doi.org/10.1111/j.1365-2621.1997.tb02131.x
Lewicki P.P., 2006. Design of hot air drying for better foods. 2015. Mathematical modeling of equilibrium moisture con-
Food Sci. Technol., 17, 153-163. tent of local cardamon (Amomum cardamomum Wild). Int.
https://doi.org/10.1016/j.tifs.2005.10.012 J. Scientific Eng. Technol., 4, 2, 40-44.
Newman A. and Zografi G., 2019. An examination of water https://doi.org/10.17950/ijset/v4s2/203
vapor sorption by multicomponent crystalline and amor- Tsami E., Marinos-Kouris D., and Maroulis Z.B., 1990. Water
phous solids and its effects on their solid-state properties. sorption isotherms of raisins, currants, figs, prunes and apri-
J. Pharmaceutical Sci., 108, 1061-1080. cots. J. Food Sci., 55, 1594-1597.
https://doi.org/10.1016/j.xphs.2018.10.038 https://doi.org/10.1111/j.1365-2621.1990.tb03578.x