3438 J. Agric. Food Chem.

2010, 58, 3438–3445

DOI:10.1021/jf9039364

Adsorption Equilibria of Water Vapor on Cork

SONIA LEQUIN,†,‡,§ DAVID CHASSAGNE,†,‡ THOMAS KARBOWIAK,‡ RÉGIS GOUGEON,†,‡
LAURENT BRACHAIS,‡ AND JEAN-PIERRE BELLAT*,§
†
Institut Universitaire de la Vigne et du Vin, Institut Jules Guyot, 1 Rue Claude Ladrey, Université de
Bourgogne, F-21078 Dijon, France, ‡EA 581 EMMA, Dijon Agrosup, 1 Esplanade Erasme, Université de
Bourgogne, F-21078 Dijon, France, and §Laboratoire Interdisciplinaire Carnot de Bourgogne, ICB UMR
5209 CNRS, 9 Avenue Alain Savary, Université de Bourgogne, F-21078 Dijon, France

We report here for the first time a complete thermodynamic study of water vapor adsorption on
Downloaded via UNIV INDUSTRIAL DE SANTANDER on September 30, 2019 at 23:31:20 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

crude cork powder and plate. Adsorption-desorption isotherms were accurately measured by
thermogravimetry at 283, 298, and 313 K in a large range of relative pressure. Adsorption enthalpies
were determined by calorimetry as a function of loading. Adsorption-desorption isotherms exhibit a
hysteresis due to the swelling of the material. The influence of the presence of lenticels on the
adsorption properties of cork is found to be negligible. A detailed analysis and interpretation of
adsorption data allow proposal of an adsorption mechanism in two steps. (i) First, water adsorbs on
hydrophilic sites constituted by hydroxyl and methoxyl groups. (ii) Then water adsorption continues
by clusters formation around the hydrophilic sites.

KEYWORDS: Cork; water; adsorption; thermogravimetry; calorimetry; solid state NMR

INTRODUCTION bottle industry. But cork will have to wait an additional century to
Cork, constituting the outer bark of oak tree Querc-us suber L., give rise to a relevant interest in wine scientific research because of
is a natural and renewable material with peculiar mechanical, cork taint (9). More recently, different potential applications were
physical, and chemical properties. It reveals an alveolar and discovered (therapeutic shoes (10), antiaging skin, medicine
anisotropic structure without intercellular spaces (1-3) but with effects like tumor inhibition (11-13), ...). Another possible
lenticular channels named lenticels. These lenticels constitute the interesting use of cork is the biosorption of pollutants. Its good
main porosity of cork (4). The structure of cork cell wall is made performances and low cost make it a good candidate for biosorp-
up of three layers. The thin middle lamella or internal primary tion of heavy metals such as Cu(II), Zn(II), Cr(VI), and Ni(II) (14,
wall is composed of lignin. Lignin represents about 25% of the 15), pesticides (16), etc.
total material weight and offers the mechanical support and Cork and wood are very similar materials. The topochemistry
rigidity of cell walls. Without lignin cork cells would totally of the cell wall is quite the same except for the secondary wall of
collapse. The thick secondary cell wall is made up of alternating wood which is constituted of cellulose. As for wood, physical
suberin and wax lamella. Suberin, the main component of cork properties of cork are strongly dependent on its water content.
(about 50%), is a polyester whose structure is not yet well-known. Surprisingly, if the interaction of water with wood has been
It is composed of long chains of fatty acids, hydroxyl fatty acids, largely studied (17-24), data relative to adsorption of water by
and phenolic acids linked by ester groups (5-7). Finally, a thin cork are very scarce (25-27). To our knowledge, the only data
layer of celluloses and hemicelluloses with a proportion of about reported in the literature are those of Abdulla et al. (23), Gonzales
10 wt % constitutes the tertiary wall. Adrados et al. (24), and Gil et al. (25) These authors measured the
Cork is known to be light, compressible, and impermeable to adsorption isotherms of water vapor on cork at several tempera-
liquids and gases, allowing its use in a lot of application fields. tures, modeled the adsorption isotherms, and estimated the
Some of them have been known for a long time, such as thermal isosteric adsorption heat. However, they did not study desorption
and vibration insulation, wall and floor covering, or cork object and no accurate measurements are given for adsorption enthal-
making (cigarette box, desk pads, mouse pad, mats, shoe soles, pies and molar adsorbate entropies as a function of the adsorbed
memo board, shuttlecocks, ...) (8). The main sector of cork amount. Therefore, a more detailed study was necessary in order
industry remains the production of natural cork stoppers. Cork to acquire a better knowledge concerning the interactions of
has been used since antiquity for sealing alcoholic beverages, water with cork. The objective of this study was thus to perform a
particularly wines. However, production of cork stoppers really complete thermodynamic study of the adsorption of water vapor
started during the 19th century with the development of the glass on cork. Adsorption and desorption were accurately studied,
and a few explanations are put forward concerning the hysteresis
*To whom correspondence should be addressed. Telephone: þ33 phenomenon observed in desorption. Special attention was
(0) 3 80 39 59 29. Fax: þ33 (0) 3 80 39 61 32. E-mail: jean-pierre. paid to the role of lenticels in water adsorption. Original solid
[email protected]. state NMR experiments were also conducted to attempt the

pubs.acs.org/JAFC Published on Web 03/01/2010 © 2010 American Chemical Society

Article J. Agric. Food Chem., Vol. 58, No. 6, 2010 3439
Table 1. Temperature of Adsorption, Water Vapor Pressure, and Relative cork powder at different temperatures. Mathematical equations of these
Pressure Ranges Investigated by Thermogravimetry for Cork Plate and Cork models are briefly recalled below.
Powder BET Model. The Brunauer, Emmet, and Teller (BET) model is well
cork sample T, K p range, hPa ps, hPa p/ps range adapted to describe a mono- and multilayer gas adsorption in a relative
pressure range between 0.05 and 0.4 (31). This model is defined by the
powder 313 0.4-20 73.6 0.005-0.27 following relation:
plate and powder 298 0.5-30 31.5 0.0015-0.95
powder 283 0.2-10 12.1 0.017-0.90 ma Cðp=ps Þ
¼ ð1Þ
mm ð1 - p=ps Þ½1 þ ðC -1Þðp=ps Þ
identification of water adsorption sites. Results obtained in this
work will be very useful for the development of different applica- where ma is the full amount of adsorbed water (mmol 3 g-1), mm is the
tions mentioned above, notably for those concerning wine which amount of water adsorbed on the monolayer (mmol 3 g-1), p/ps is the
contains about 90% of water. relative pressure, and C is a constant related to adsorption energies of the
first and subsequent adsorbed layers. The adsorption energy of subsequent
EXPERIMENTAL PROCEDURES layers is assumed to be equal to the pure adsorptive liquefaction energy.
C is related to the temperature by the relation
Material. Raw cork planks, from Quercus suber L. oak trees in the
Mora (Portugal) production area, were supplied by Bouchons Trescases
CBET ¼ CoBET e -ðΔHBET -ΔHliq Þ=ðRT Þ ð2Þ
S.A. (Boulou, France). Planks were neither washed nor surface treated
(with paraffin or silicone) prior to use. Uniform cork pieces in plate where ΔHBET is the adsorption enthalpy of the monolayer and ΔHliq is the
geometry, 35 mm long, 10 mm wide, and 1 mm thick for gravimetric study liquefaction enthalpy.
of water vapor adsorption, were manually cut from the planks. The matter GAB Model. The Guggenheim-Anderson-de Boer (GAB) model is a
present in the lenticels (lenticular cells) was extracted from the planks by refined extension of the BET theory (32). It is expressed as the following
scraping with a cutter blade. Cork powder, with and without lenticels, was equation:
also used. Powder is made by grating cork stopper or cork plate with a
rasp. After sieving, the mean particle size was lower than 500 μm. ma CKðp=ps Þ
Thermogravimetry. Adsorption of water vapor on natural cork was ¼ ð3Þ
mm ½1 - K ðp=ps Þ½1 -K ðp=ps Þ þ CK ðp=ps Þ
investigated by thermogravimetry with a homemade McBain balance,
under controlled temperature and pressure. The adsorption-desorption In this model, the properties of the subsequent layers are discriminated
isotherm was measured step by step using a static method. Once a plateau from those of the pure liquid bulk. The C constant is related to adsorption
of mass was recorded, the next equilibrium was reached by changing the energies of the first and second layers, whereas the K constant is related
water vapor pressure. Two samples with different geometries were studied: to the adsorption energies of the second and subsequent layers which
cork plate and cork powder. The mass of the cork sample was around lie somewhere between the monolayer adsorption energy and the pure
40 mg. Prior to each experiment, the cork sample was outgassed in situ adsorptive liquefaction energy. These constants are defined by the
under vacuum (10-5 hPa) at 298 K until a plateau of mass is reached. This relations (33)
needed 12 h for the powder and 24 h for the plate. After this treatment the
water content did not exceed 1 wt %. The adsorption temperatures and CGAB ¼ CoGAB e -ðΔHGAB1 - ΔHGABn Þ=ðRT Þ ð4Þ
the water vapor pressure ranges investigated for cork plate and cork
powder are reported in Table 1. The corresponding ranges of relative
pressure p/ps, ps being the saturation water vapor pressure at the adsorp- KGAB ¼ CoGAB e -ðΔHGABn - ΔHliq Þ=ðRT Þ ð5Þ
tion temperature, are also indicated. The experimental accuracy is
(0.01 mmol 3 g-1 for the adsorbed amount, (0.5 K for the temperature, where ΔHGAB1 is the adsorption enthalpy of the first layer and ΔHGABn is
and (0.01 hPa for the pressure. that of the subsequent layers.
Differential Calorimetry Coupled with Manometry. Adsorption The GAB model was recently reconsidered by Pradas et al. (34) from
enthalpy of water vapor on cork was measured with a differential thermodynamic and statistical mechanic statements. They showed that the
calorimeter (Thian-Calvet Setaram C80) coupled with manometry. This GAB equation takes into account the possibility of an incomplete
device has been described in detail in previous studies (28-30). It allows occupation of the first adsorption layer even at saturation. Thus, the
measuring the molar adsorption enthalpy ΔHa(ma) of water, so-called the GAB equation appears more appropriate to describe an adsorption
adsorption heat, as a function of the adsorbed amount ma. Adsorption was process by clusters formation around localized adsorption sites rather
realized at 298 K on about 640 mg of cork powder, previously outgassed in than by formation of a homogeneous monolayer followed by multilayer
situ under vacuum at 10-5 hPa for 72 h. The water vapor pressure ranged condensation on the surface as considered in BET theory.
from 0.1 to 25 hPa.
High Resolution Solid State NMR. NMR experiments were carried RESULTS AND DISCUSSION
out at 293 K on cork samples either outgassed under vacuum and kept
Adsorption Isotherms. The adsorption-desorption isotherms
under argon for analysis or previously equilibrated at 293 K under water
relative pressures of 0.11 and 0.98 using lithium chloride and potassium of water vapor on plate and powder cork at 298 K are shown in
dichromate saturated salt solution, respectively. These NMR experiments Figure 1. Both curves display a type II shape of the IUPAC
were run on a Bruker DSX 300 spectrometer operating at frequencies of classification (35), typical of the adsorption on nonporous or
300.1 and 75.5 MHz for 1H and 13C, respectively. All spectra were acquired macroporous solids.
with a Bruker double-channel 4 mm MAS probe at 10 kHz spinning rate. The slope of adsorption isotherms, when the relative pressure
The 1H-13C spectra were obtained by cross-polarization (CP-MAS), with tends to zero, is rather low. This indicates a weak adsorption
spinal-64 proton dipolar decoupling. The experimental conditions were as affinity of water for cork. Indeed, the Henry constant given by
follows: a 1H 90° pulse of 3.6 μs, a contact time of 1 ms, and a recycle delay this slope is about 5 and 10 mmol 3 g-1 for plate and powder,
of 4 s between consecutive pulses. Hartmann-Hahn matching for the respectively. These values are slightly higher than those reported
1
H-13C CP-MAS experiments was set on adamantane for 1H and 13C
for activated carbons (1-4 mmol 3 g-1), which are hydrophobic
radiofrequency fields of ∼60 kHz. Chemical shifts for 1H and 13C spectra
were referenced to the signal of water (4.87 ppm) and to the methylene materials (36, 37). However they are 10 times lower than those
signal of adamantane (29.47 ppm), respectively. observed on silica gel (30-40 mmol 3 g-1), which is hydrophilic
Modeling of Adsorption Isotherms. The modeling of adsorption and often used as desiccant (38, 39). Thus, cork can be considered
isotherms was performed by the well-known BET and GAB equations. as a rather hydrophobic material. This is in agreement with
Modeling was only performed on adsorption isotherms determined on wettability measurements of water on crude cork, which showed
3440 J. Agric. Food Chem., Vol. 58, No. 6, 2010 Lequin et al.
Table 2. BET and GAB Parameters and Adsorption Enthalpies for Adsorption
of Water Vapor on Cork Powder at Different Temperatures
p/ps range mm, ΔHBET, ΔHGAB1, ΔHGABn,
of validity T, K mmol 3 g-1 C K kJ 3 mol-1 kJ 3 mol-1 kJ 3 mol-1

BET Model

0-0.4 283 1.4 15.6 -60

0-0.4 298 1.3 15.6 -60
0-0.4 313 1.3 9.6 -60

GAB Model

0-0.99 283 1.5 18.0 0.8 -75 -50

0-0.95 298 1.7 11.8 0.7 -75 -50
0-0.27 313 1.2 9.7 1.1 -75 -50

Figure 1. Adsorption-desorption isotherms of water vapor on dry cork

plate (circles) and dry cork powder (squares) at 298 K: open symbols,
adsorption; closed symbols, desorption.

that water is a nonwetting liquid for the cork surface (40). This
hydrophobicity of the cork surface is attributed to the presence of
suberin (41).
When the relative pressure is approaching 1, the adsorbed
amount is around 6 mmol 3 g-1. This corresponds to 11 wt %. This
value is in good agreement with those found by other authors on
cork granules of 0.5-1 mm (25) and on little pieces of cork (26).
This adsorption capacity is rather low compared to microporous
activated carbons or silica gels that can adsorb up to 40-50 wt %.
It is nevertheless higher than that observed for very hydrophobic
materials such as pure siliceous materials (2.2 mmol 3 g-1) (42).
Adsorption isotherms on powder and on plate exhibit a Figure 2. Modeling of adsorption isotherm of water vapor on dry cork at
hysteresis loop. The adsorption-desorption process is not rever- 283 K: (O) experiment; (- - -) BET model; (-) GAB model.
sible, especially on powder, for which the amplitude of the
hysteresis is more noticeable. However, in both cases, desorption more restricted swelling in block cork. Grating destroys the cell
is total; no water remains adsorbed after pumping under vacuum. structure of wood by breaking bonds between microfibers. The
This means that no water molecule is chemically bonded to the material then becomes more elastic and consequently more sensi-
surface. Water is only physisorbed. So the hysteresis pheno- tive to swelling during water adsorption. Since the amplitude of the
menon cannot be due to chemisorption. This kind of hysteresis hysteresis loop is higher in cork powder than in cork plates, a
has already been observed for water sorption on U.S. hard- similar phenomenon in cork could be postulated. However, this
woods, eucalyptus, pine, beech, spruce, and mahogany heart- increase in of the adsorption capacity of cork after grating might
wood (21, 43-45). Studies of water adsorption on lignin and also be due to the creation of new hydrophilic sites by breaking
cellulose, some of the major chemical components of wood, also chemical bonds.
reported the existence of a hysteresis loop in the adsorp- Modeling of Adsorption Isotherms. Values of BET and GAB
tion-desorption isotherms (46, 47). This hysteresis phenomenon parameters (mm, C, and K) are collected in Table 2. The notion of
is attributed to swelling effects that take place during hydration. monolayer is replaced below by that of “equivalent monolayer”
In these lignocellulosic materials the adsorption mechanism is because water molecules can form clusters on cork surface. The
described in two steps (19). (i) First, water adsorbs on the surface, modeling of adsorption isotherms is shown in Figure 2. The best
where it forms clusters around hydrophilic sites rather than a fit is obtained with the GAB model which gives a more accurate
uniform monolayer. (ii) Then water diffuses into the cell wall. description of the adsorption process throughout the range of
During this diffusion step, mechanical strengths occur between pressures explored.
hydrated and dry layers, leading to a swelling of the material. This The equivalent monolayer lies between 1.2 and 1.7 mmol 3 g-1
swelling depends on the water vapor pressure, the cell wall according to the temperature and the model used. This corre-
thickness, and the geometry of the sample (block or powder). sponds to 2.16 - 3 wt %. These values are in good agreement with
In the case of cork, which is also a lignocellulosic material, the those found in the literature on cork granules (25) (2.4 wt %) and
hysteresis phenomenon has certainly the same origin. Cork cell cellulose (48) (3.3 wt %) at 298 K.
walls are probably hydrated with swelling effects as for wood. As expected, the value of the C constant decreases when the
It is noteworthy that adsorbed amounts are higher on cork temperature increases. It is not the case for the K constant, the
powder than on cork plate, whatever the relative pressure. This value of which is higher at 313 K than at lower temperatures. The
difference could be attributed to the increase of the external surface value of K at this temperature is probably wrong, the range of
area after grating. However, the difference in adsorbed amounts is relative pressure investigated being too much restricted to accu-
not so important compared to the difference in the external surface rately determine this value.
between these two geometries (about 0.04 m2 3 g-1 for the plate and Contribution of Matter Contained in Lenticels. Lenticels of cork
about 163 m2 3 g-1 for the powder). In another study, Nakano (20) are described as quasi-cylindrical channels composed of died
also observed a slight difference between powder and block wood cell walls and intercellular spaces. They cross cork along the
from Todomatsu (A. sachalinensis Fr. Schm.). He explained it by a radial direction without structural arrangement (Figure 3). Their
Article J. Agric. Food Chem., Vol. 58, No. 6, 2010 3441

Figure 3. Scanning electronic microscopy pictures of (a) lenticels in axial (A), radial (R), and tangential (T) sections of cork, (b) the inside of lenticels, (c) cork
cell in axial section, and (d) cork cell in radial section.

reported for comparison. Adsorption-desorption isotherms of

water vapor on lenticular cells have the same shape as those for
cork. It similarly exhibits a hysteresis loop that is, however, more
pronounced. The adsorbed amounts are higher than those
measured for cork. At low relative pressures, the Henry constant
(28 mmol 3 g-1) and the CGAB constant (18.6) are higher than
those for cork. This indicates that the interactions with water are
stronger in the case of lenticular cells. The value of the equivalent
monolayer determined from GAB model (3.2 mmol 3 g-1) is twice
higher than for cork (1.7 mmol 3 g-1). Likewise, the adsorption
capacity close to saturation reaches 8 mmol 3 g-1 versus 6 mmol 3 g-1
for cork. These results show that lenticular cells are more hydro-
philic than cork. Despite this, no noticeable difference in water
vapor adsorption capacities is observed between powder with and
without lenticels (Figure 4). This can be explained by the low
Figure 4. Adsorption-desorption isotherms of water vapor on lenticels lenticels content in this superior quality cork.
(circles), dry cork powder without lenticels (squares), and dry cork powder Adsorption Enthalpies and Entropies. Calorimetric adsorption
with lenticels (triangles) at 298 K: open symbols, adsorption; closed enthalpies of water vapor on cork powder containing lenticels are
symbols, desorption. given in absolute value versus loading in Figure 5. The values of
adsorption enthalpies of the first and subsequent layers were also
diameter is a few millimeters, and their length can reach several calculated from eqs 4 and 5 by plotting ln(constant) versus 1/T.
centimeters, depending on the cork thickness. They allow gas They are reported in Table 2.
exchange between tree core and outside environment (49). The When loading tends to zero, the adsorption enthalpy tends to
chemical composition of lenticular cells is close to that of wood 65 kJ 3 mol-1. This value is of the same order of magnitude as
cell wall. Lenticular cells are mainly constituted of lignin (about those previously given by the BET and GAB theories (Table 2).
47 wt %), polysaccharides (about 21 wt %), and extractives, The net adsorption enthalpy at zero loading, which is the
mainly polar extractives (about 32 wt %) (50). Lenticels are difference between adsorption and liquefaction enthalpies, is
overgrown by microorganisms (Figure 3, lower left-hand SEM equal to 21 kJ 3 mol-1. It is similar to the value given by Abdulla
picture). Lenticels create the macroporosity of cork. The ratio of et al.(25), using the isosteric method. This indicates that cork-
lenticels in cork is used as the main indicator of cork quality in water interactions are rather strong compared to water-water
industrial applications (3). Therefore, it is of relevant interest to interactions occurring in the liquid bulk. However, this interac-
know if the lenticular cells play an important role in the adsorp- tion energy is close to those usually involved in physisorption. As
tion process of water on cork. the adsorbed amount increases, the adsorption enthalpy con-
Adsorption-desorption isotherms of water vapor on matter tinuously decreases and tends to the liquefaction enthalpy of
present in lenticels at 298 K are displayed in Figure 4. Those water for loading which lies between 1.5 and 2 mmol 3 g-1. This
acquired for cork powder with and without lenticels are also loading is very close to the “equivalent monolayer” determined by
3442 J. Agric. Food Chem., Vol. 58, No. 6, 2010 Lequin et al.

Figure 5. Adsorption enthalpy (absolute value) of water vapor on cork Figure 6. Molar entropy of water adsorbed on cork powder at 298 K as a
powder measured by calorimetry at 298 K versus loading. ΔHliq is the function of loading: (a) standard molar entropy of gas Sg°m = 188.84
liquefaction enthalpy of water. J 3 mol-1 3 K-1; (b) standard molar entropy of liquid Sl°m = 69.95
BET and GAB models (Table 2). This first part of the calorimetric J 3 mol-1 3 K-1; (c) standard molar entropy of solid Ss° m = 41.50

curve corresponds to the adsorption of water molecules on J 3 mol-1 3 K-1.

hydrophilic sites. The first molecules are adsorbed on the more Molecular Investigation of Water Adsorption. In order to better
energetic sites and the following ones on sites of decreasing characterize the possible adsorption sites and the interactions that
energy. The fact that the adsorption enthalpy decreases in this can occur between the cork surface and water vapor, high
region suggests that the adsorbate-adsorbent interactions de- resolution solid state 13C NMR as a selective technique for a
crease and that the adsorbate-adsorbate interactions are negli- molecular scale investigation was used. Figure 7 shows the
gible. In other words, at low loading, the adsorbed molecules are 1
H-13C (CP-MAS) NMR spectra of partially hydrated cork
too far from each other to allow molecular interactions. (pre-equilibrated under the relative pressure of 0.11) and com-
At loading higher than 2 mmol 3 g-1, the adsorption enthalpy pletely hydrated cork (pre-equilibrated under the relative pressure
sharply increases up to about 70 kJ 3 mol-1 and then tends back to of 0.98). From the literature on NMR of cork (15, 52-55), most
the liquefaction enthalpy of liquid bulk as expected when of the distinct peaks of cork can be identified. In particular, the
approaching saturation. This exothermic effect shows the emer- low frequency peak at 30.1 and 32.9 ppm can be attributed to the
gence of strong molecular interactions in the adsorbed phase. carbon of the aliphatic methylene region of suberin (15, 52-54).
This phenomenon could be caused by the aggregation of water As no water molecule is assumed to bind to this chemical
molecules around the hydrophilic sites. The presence of this function, these two most intense and well-defined peaks unam-
exothermic peak at high loading reinforces the hypothesis that biguously attributed to -(CH2)n- of suberin were arbitrarily
the adsorption process occurs by formation of water clusters as chosen to normalize the relative intensity of both spectra. The
advanced by Hartley et al.(19) for wood. We think that it cannot peak at 21.3 ppm can be assigned to the acetate function of
take place as a monolayer and multilayer condensation mecha- hemicellulose (53, 54). The peak at 56.2 ppm corresponds to the
nism because in this case the adsorption enthalpy should remain methoxy group from both lignin and suberin and also a small
close to the liquefaction enthalpy of water after completion of the contribution from hemicellulose (15, 52-54). The peak at
monolayer, i.e., for loading above 2 mmol 3 g-1. 64.4 ppm can be tentatively attributed to the carbon number 6
The molar entropy of the adsorbed phase is calculated from of the carbohydrate unit from cellulose (53). The two peaks at
adsorption isotherm and calorimetric data by using the following 72.3 and 74.6 ppm are attributed to the carbons 2, 3, and 5 of the
equation derived from the Clapeyron relation (51): carbohydrate structure from cellulose, hemicellulose, and partly
Sm
a
ðma Þ ¼ ΔH a ðma Þ=T -R lnðp=po Þma þ Sm
g°
ð6Þ to side chain C-OH groups of lignin (53, 54). The peaks at 82.4,
105.1, 130.1, and 173 ppm correspond respectively to the carbon 4
where Sg°m is the standard molar entropy of water vapor at 298 K of the carbohydrate that could be attached to suberin, the
and p° the standard pressure (105 Pa). anomeric carbon number of the carbohydrate that could also
The evolution of the molar entropy of water adsorbed on cork be attached to suberin, the aromatic carbons of suberin and
powder as a function of loading at 298 K is shown in Figure 6. lignin, and the carbonyl groups from suberin, hemicellulose, and
Standard molar entropy values for gas, liquid, and solid water at lignin (ester groups and uronic acids) (53, 54). The region of the
298 K are also reported. spectrum between 62 and 105 ppm therefore contains overlapping
At low loading, the molar entropy of the adsorbate is close to signals from the three major constituents of cork.
that of the solid. This means that the first water molecules Although not all the peaks can be unambiguously attributed, it
adsorbed on the strongest hydrophilic sites lose degrees of free- is possible to draw information on the interactions between water
dom. They are “frozen” on their adsorption sites and are in a and cork by comparing the relative intensity of a given 13C peak
solid-like physical state. As loading increases, the molar entropy of cork at low or high water loading (Figure 7). From adsorption
increases to the molar entropy of liquid bulk. Water molecules are data given in Figure 1, the 0.11 relative water pressure almost
progressively adsorbed on less energetic sites and are more corresponds to the equivalent monolayer, whereas 0.98 lies near
mobile. Then above 2 mmol 3 g-1, a decrease in the molar entropy the saturation. The formation of clusters therefore occurs bet-
which gets closer to the molar entropy of the solid is observed. ween these two limits from aggregation of water molecules
This sudden decrease of the molar entropy at high loading is around the hydrophilic sites already occupied.
attributed to the formation of clusters around the hydrophilic If we compare the spectrum of partially hydrated cork (in
sites which are more structured than the liquid bulk. orange) and that of completely hydrated cork (in blue), we can see
Article J. Agric. Food Chem., Vol. 58, No. 6, 2010 3443

Figure 7. 1H-13C CPMAS spectrum of partially hydrated cork in orange (pre-equilibrated at water relative pressure of 0.11) and of completely hydrated cork
(pre-equilibrated at water relative pressure 0.98) in blue. Spectra were obtained with a contact time of 1 ms and 1200 scans (p/ps = 0.98) and 3700 scans
(p/ps = 0.11).

that the relative intensities of the peaks at 56.2, 64.4, 72.3, and can easily involve hydrogen bonds with water molecules. Then the
74.6 ppm are increased when water loading increases. The only adsorption continues by formation of water clusters around the
factor that could induce such an increase is a better efficiency of hydrophilic sites.
the cross-polarization from nearby protons. Such a better effi-
ciency can result from a reduced mobility of these nearby protons ABBREVIATIONS USED
and/or from an increased density of nearby protons. In the CBET, constant in BET equation; CGAB, constant in GAB
present situation, additional protons would necessarily come equation; CoBET, Arrhenius type constant to express tempera-
from additional bound water molecules in the vicinity of these ture dependence of CBET; CoGAB, Arrhenius type constant to
specific carbon sites. The water proportion introduced in the express temperature dependence of CGAB; K, constant in GAB
system increases from 0.55 to 5.3 mmol 3 g-1 for relative pressures equation; m, amount adsorbed (mmol 3 g-1); mm, amount of
from 0.11 to 0.98, respectively, as displayed by the adsorption water adsorbed in the monolayer (mmol 3 g-1); p, pressure
isotherm in Figure 1. This suggests that water interacts specifically (hPa); ps, saturation water vapor pressure (hPa); p°, standard
with some of the carbon sites offered by the cork structure. The pressure (Pa); Sam(ma), molar entropy of the adsorbed phase;
four identified carbon sites that could be implied in the sorption Sg°
m , standard molar entropy of water vapor at 298 K
process for water display attached chemical functions such as (J 3 mol-1 3 K-1); Sl°m, standard molar entropy of liquid water
hydroxyl or methoxyl that can easily involve hydrogen bonds at 298 K (J 3 mol-1 3 K-1); Ss°
m, standard molar entropy of solid
with water molecules. As the chemical structure and spatial water at 298 K (J 3 mol-1 3 K-1); T, temperature (K); amount
distribution of cork constituents remain unknown, the location adsorbed wt (%); ΔHa(ma), molar adsorption enthalpy at the
of the particular adsorption sites cannot be unambiguously loading ma (kJ 3 mol-1); ΔHBET, adsorption enthalpy
identified. However, these results support the hypothesis of a of the monolayer (kJ 3 mol-1); ΔHGAB1, adsorption enthalpy
specific physicochemical interaction in which the cellulosic frac- of the first layer (kJ 3 mol-1); ΔHGABn, adsorption enthalpy
tion seems to play a significant contribution in the water sorption of the subsequent layers (kJ 3 mol-1); ΔHliq, liquefaction
process. enthalpy (kJ 3 mol-1).
Adsorption Mechanism of Water on Cork. This complete
thermodynamics study and the detailed characterization of
adsorbed water by NMR show that the adsorption of water on ACKNOWLEDGMENT
cork does not follow a monolayer and multilayer condensation We thank Christian Paulin, from Laboratoire Interdiscipli-
process, as it could be expected considering the type II adsorption naire Carnot de Bourgogne (Dijon), for assistance regarding
isotherm. It rather occurs according to a localized adsorption experimental techniques. We also thank Jésus Raya and Jér^ome
mechanism with aggregation of water molecules, as already Hirschinger from the Laboratoire de Résonance Magnétique et
emphasized for wood. First, water molecules are adsorbed on de Biophysique des Membranes at University of Strasbourg for
hydrophilic sites identified as hydroxyl or methoxyl groups that accessing the NMR spectrometer.
3444 J. Agric. Food Chem., Vol. 58, No. 6, 2010 Lequin et al.
LITERATURE CITED (23) Rawat, S. P. S.; Khalid, D. P. Clustering of water molecules during
(1) Lequin, S.; Karbowiak, T.; Brachais, L.; Chassagne, D.; Bellat, J. P. adsorption of water in wood. J. Polym. Sci., Polym. Phys. 1998, 36
Adsorption equilibria of sulphur dioxide on cork. Am. J. Enol. (4), 665–671.
Viticult. 2009, 60 (2), 138–144. (24) Zhang, Y.; Zhang, S. Y.; Chui, Y. H. Water vapor adsorption and
(2) Gibson, L. J.; Easterling, K. E.; Ashby, M. F. The structure and volumetric swelling of melt-impregnated wood-polymer composites.
mechanics of cork. Proc. R. Soc. London, Ser. A 1981, 377, 99–117. J. Appl. Polym. Sci. 2006, 102 (3), 2668–2676.
(3) Silva, S. P.; Sabino, M. A.; Fernandes, E. M.; Correlo, V. M.; Boesel, (25) Abdulla, G.; Belghit, A.; Allaf, K. Impact of instant controlled
L. F.; Reis, R. L. Cork: properties, capabilities and applications. Int. pressure drop treatment on moisture adsorption isotherm of cork
Mater. Rev. 2005, 50, 345–365. granules. Drying Technol. 2009, 27 (2), 237–247.
(4) Pereira, H. The Formation and Growth of Cork. In Cork; Elsevier (26) Adrados, J. R. G.; Haro, R. M. C. Variacion de la humedad de
Science B.V.: Amsterdam, 2007; pp 7-31. equilibrio del corcho en plancha con la humedad relativa. Modelos
(5) Gandini, A.; Pascoal Neto, C.; Silvestre, A. J. D. Suberin: a de regression no lineal para las isotermas de adsorcion. Invest.
promising renewable resource for novel macromolecular materials. Agrar., Sist. Recur. For. 1994, 3 (2), 199–209.
Prog. Polym. Sci. 2006, 31 (10), 878–892. (27) Gil, L.; Cortic-o, P. Cork hygroscopic equilibrium moisture content.
(6) Graca, J.; Santos, S. Linear aliphatic dimeric esters from cork Holz Roh.- Werkst. 1998, 56, 355–358.
suberin. Biomacromolecules 2006, 7 (6), 2003–2010. (28) Moise, J.-C.; Bellat, J.-P. Effect of preadsorbed water on the
(7) Cordeiro, N.; Belgacem, M. N.; Silvestre, A. J. D.; Pascoal Neto, C.; adsorption of p-xylene and m-xylene mixtures on BaX and BaY
Gandini, A. Cork suberin as a new source of chemicals: 1. Isolation zeolites. J. Phys. Chem. 2005, 109 (36), 17239–17244.
and chemical characterization of its composition. Int. J. Biol. (29) Weber, G.; Benoit, F.; Bellat, J.-P.; Paulin, C.; Mougin, P.; Thomas,
Macromol. 1998, 22 (2), 71–80. M. Selective adsorption of ethyl mercaptan on NaX zeolite. Micro-
(8) Pereira, H. Cork Products and Uses. In Cork; Elsevier Science B.V.: porous Mesoporous Mater. 2008, 109 (1-3), 184–192.
Amsterdam, 2007; pp 243-261. (30) Simonot-Grange, M. H.; Bertrand, O.; Pilverdier, E.; Bellat, J.-P.;
(9) Buser, H. R.; Zanier, C.; Tanner, H. Identification of 2,4,6-trichloro- Paulin, C. Differential calorimetric enthalpies of adsorption of
anisole as a potent compound causing cork taint in wine. J. Agric. p-xylene and m-xylene on y faujasites at 25°C. J. Therm. Anal.
Food. Chem. 1982, 30 (2), 359–362. Calorim. 1997, 48 (4), 741–754.
(10) Reiber, G. E.; Smith, D. G.; Wallace, C.; Sullivan, K.; Hayes, S.; (31) Brunauer, S.; Emmett, P. H.; Teller, E. Adsorption of gases in
Vath, C.; Maciejewski, M. L.; Yu, O.; Heagerty, P. J.; LeMaster, J. multimolecular layers. J. Am. Chem. Soc. 1938, 60 (2), 309–319.
Effect of therapeutic footwear on foot reulceration in patients with (32) Anderson, R. B. Modifications of the Brunauer, Emmett and Teller
diabetes: a randomized controlled trial. JAMA, J. Am. Med. Assoc. equation. J. Am. Chem. Soc. 1946, 68 (4), 686–691.
2002, 287 (19), 2552–2558. (33) Quirijns, E. J.; van Boxtel, A. J. B.; van Loon, W. K. P.; van Straten,
(11) Moiteiro, C.; Marcelo Curto, M. J.; Mohamed, N.; Bailen, M.; G. Sorption isotherms, GAB parameters and isosteric heat of
Martinez-Diaz, R.; Gonzalez-Coloma, A. Biovalorization of friede- sorption. J. Sci. Food. Agric. 2005, 85 (11), 1805–1814.
lane triterpenes derived from cork processing industry byproducts. (34) Pradas, M. M.; Sanchez, M. S.; Ferrer, G. G.; Ribelles, J. L. G.
J. Agric. Food. Chem. 2006, 54 (10), 3566–3571. Thermodynamics and statistical mechanics of multilayer adsorption.
(12) Moiteiro, C.; Manta, C.; Justino, F.; Tavares, R.; Curto, M. J. M.; J. Chem. Phys. 2004, 121 (17), 8524–8531.
Pedro, M.; Nascimento, M. S. J.; Pinto, M. Hemisynthetic secofrie- (35) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti,
delane triterpenes with inhibitory activity against the growth of R. A.; Rouquérol, J.; Siemieniewska, T. Reporting physisorption
human tumor cell lines in vitro. J. Nat. Prod. 2004, 67 (7), 1193–1196. data for gas/solid systems. Pure Appl. Chem. 1985, 57 (4), 603–
(13) Moiteiro, C.; Justino, F.; Tavares, R.; Marcelo-Curto, M. J.; 619.
Florencio, M. H.; Nascimento, M. S. J.; Pedro, M.; Cerqueira, F.; (36) Rutherford, S. W.; Coons, J. E. Equilibrium and kinetics of water
Pinto, M. M. M. Synthetic secofriedelane and friedelane derivatives adsorption in carbon molecular sieve: theory and experiment.
as inhibitors of human lymphocyte proliferation and growth of Langmuir 2004, 20, 8681–8687.
human cancer cell lines in vitro. J. Nat. Prod. 2001, 64 (10), 1273– (37) Qi, N.; LeVan, M. D. Adsorption equilibrium modeling for water on
1277. activated carbons. Carbon 2005, 43 (11), 2258–2263.
(14) Chubar, N.; Carvalho, J. R.; Correia, M. J. N. Cork biomass as (38) Ng, K. C.; Chua, H. T.; Chung, C. Y.; Loke, C. H.; Kashiwagi, T.;
biosorbent for Cu(II), Zn(II) and Ni(II). Colloids Surf., A 2003, 230 Akisawa, A.; Saha, B. B. Experimental investigation of the silica
(1-3), 57–65. gel-water adsorption isotherm characteristics. Appl. Therm. Eng.
(15) Villaescusa, I.; Fiol, N.; Cristiani, F.; Floris, C.; Lai, S.; Nurchi, 2001, 21, 1631–1642.
V. M. Copper(II) and nickel(II) uptake from aqueous solutions by (39) Wang, X.; Zimmermann, W.; Ng, K.; Chakraboty, A.; Keller, J.
cork wastes: a NMR and potentiometric study. Polyhedron 2002, 21 Investigation on the isotherm of silica gelþwater systems. J. Therm.
(14-15), 1363–1367. Anal. Calorim. 2004, 76 (2), 659–669.
(16) Domingues, V.; Alves, A.; Cabral, M.; Delerue-Matos, C. Sorption (40) Gomes, C. M. C. P. S.; Fernandes, A. C.; de Almeida, B. d. J. V. S.
behaviour of bifenthrin on cork. J. Chromatogr., A 2005, 1069 (1), The Surface tension of cork from contact angle measurements.
127–132. J. Colloid Interface Sci. 1993, 156 (1), 195–201.
(17) Barkas, W. W. Wood-water relationships, part III. Molecular (41) Cordeiro, N.; Pascoal Neto, C.; Gandini, A.; Belgacem, M. N.
sorption of water by sitka spruce wood. Proc. Phys. Soc. 1937, 49 Characterization of the cork surface by inverse gas chromatography.
(3), 237–242. J. Colloid Interface Sci. 1995, 174 (1), 246–249.
(18) Spalt, H. A. Water-Vapor Sorption by Woods of High Extractive (42) Flanigen, E. M.; Bennett, J. M.; Grose, R. W.; Cohen, J. P.; Patton,
Content. Proceedings: Wood Moisture Content, Temperature and R. L.; Kirchner, R. M.; Smith, J. V. Silicalite, a new hydrophobic
Humidity Relationships, Blacksburg, VA, October 29, 1979; Department crystalline silica molecular sieve. Nature 1978, 271, 512–516.
of Agriculture, Forest Service: Washington, D.C., 1979; pp 55-61. (43) Christensen, G. N. Sorption and swelling within wood cell walls.
(19) Hartley, I. D.; Kamke, F. A.; Peemoeller, H. Cluster theory for water Nature 1967, 213 (5078), 782–784.
sorption in wood. Wood Sci. Technol. 1992, 26 (2), 83–99. (44) Malmquist, L.; S€oderstr€om, O. Sorption equilibrium in relation to
(20) Nakano, T. Effects of cell structure on water sorption for wood. the spatial distribution of molecules. Application to sortpion of
Holzforschung 2003, 57 (2), 213–218. water by wood. Holzforschung 1996, 50 (5), 437–448.
(21) Okoh, K.; Skaar, C. Moisture sorption isotherms of the wood and (45) Arévalo, R.; Hernandez, R. E. Influence of moisture sorption
inner bark of ten southern U.S. hardwoods. Wood Fiber Sci. 1980, 12 on swelling of mahogany (Swietenia macrophylla King) wood.
(2), 98–111. Holzforschung 2001, 55 (6), 590–594.
(22) Mandla, A. T.; Agnes, R. D.; Williams, R. S. Correlation of water (46) Christensen, G. N.; Kelsey, K. E. The sorption of water vapour by
vapor adsorption behavior of wood with surface thermodynamic the constituents of wood III. The swelling of lignin. Aust. J. Appl. Sci.
properties. J. Appl. Polym. Sci. 1999, 73 (3), 399–407. 1959, 30.
Article J. Agric. Food Chem., Vol. 58, No. 6, 2010 3445
(47) Repellin, V.; Guyonnet, R. Evaluation of heat-treated wood swelling (53) Pascoal Neto, C.; Rocha, J.; Gil, A.; Cordeiro, N.; Esculcas, A. P.;
by differential scanning calorimetry in relation to chemical composi- Rocha, S.; Delgadillo, I.; De Jesus, J. D. P.; Correia, A. J. F. 13C
tion. Holzforschung 2005, 59 (1), 28–34. solid-state nuclear magnetic resonance and Fourier transform infra-
(48) Hollenbeck, G. R.; Peck, G. E.; Kildsig, D. O. Application of red studies of the thermal decomposition of cork. Solid State Nucl.
immersional calorimetry to investigation of solid-liquid interac- Magn. Reson. 1995, 4 (3), 143–151.
tions: Mmicrocrystalline cellulose-water system. J. Pharm. Sci. (54) Gil, A. M.; Lopes, M.; Rocha, J.; Pascoal Neto, C. A 13C solid state
1978, 67 (11), 1599–1606. nuclear magnetic resonance spectroscopic study of cork cell wall
(49) Groh, B.; H€ubner, C.; Lendzian, K. Water and oxygen permeance of structure: the effect of suberin removal. Int. J. Biol. Macromol. 1997,
phellems isolated from trees: the role of waxes and lenticels. Planta 20 (4), 293–305.
2002, 215 (5), 794–801. (55) Perra, B.; Haluk, J.-P.; Metche, M. IR, 1H and 13C NMR spectro-
(50) Pereira, H. Cork: Biology, Production and Uses, 1st ed.; Elsevier: scopic studies of suberin from beech bark (Fagus sylvatica L.).
Amsterdam, 2007; p 336. Holzforschung 1995, 49, 99–103.
(51) Bellat, J.-P.; Simonot-Grange, M. H. Adsorption of gaseous
p-xylene and m-xylene on NaY, KY and BaY zeolites. Part 2: Received for review November 12, 2009. Revised manuscript received
Modeling. Enthalpies and entropies of adsorption. Zeolites 1995,
February 10, 2010. Accepted February 15, 2010. We gratefully
15, 219–227.
acknowledge the BIVB (Bureau Interprofessionnel des Vins de
(52) Lopes, M. H.; Sarychev, A.; Pascoal Neto, C.; Gil, A. M. Spectral
editing of 13C CP/MAS NMR spectra of complex systems: applica- Bourgogne) and the Regional Council of Burgundy for financial
tion to the structural characterisation of cork cell walls. Solid State support of this work (Project 2008-9201AAO024S00120), and the
Nucl. Magn. Reson. 2000, 16 (3), 109–121. Trescases company for providing cork stoppers.