Carbon 96 (2016) 753e758

Contents lists available at ScienceDirect

Carbon
journal homepage: www.elsevier.com/locate/carbon

Towards a better understanding of water adsorption hysteresis in

activated carbons by scanning isotherms

my Guillet-Nicolas b, Ga

bor Dobos c, Matthias Thommes b,
Leticia F. Velasco a, *, Re
Peter Lodewyckx a
a
b
c

Department of Chemistry, Royal Military Academy, Renaissancelaan 30, 1000 Brussels, Belgium
Quantachrome Instruments, 1900 Corporate Drive, Boynton Beach, FL 33426, USA
Department of Atomic Physics, Budapest University of Technology and Economics, H-1521 Budapest, Hungary

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 16 July 2015
Received in revised form
6 October 2015
Accepted 7 October 2015
Available online xxx

The occurrence of water sorption hysteresis associated with the lling of micro and narrow mesopores
(particularly for pores of widths smaller than ca. 3 nm, where nitrogen and argon isotherms at their
boiling temperatures, i.e. 77 K and 87 K, respectively, are known to be reversible) provides additional
opportunities for textural characterization. In this work systematic water scanning desorption isotherms
within the hysteresis loop were carried out on well-characterized activated carbons with varied textural
features and surface chemistry. Accurate micro-mesopore analysis was obtained by means of nitrogen,
argon and carbon dioxide adsorption experiments coupled with advanced density functional theory
methods (i.e., NLDFT, QSDFT). The obtained results indicated that water adsorption/desorption phenomena for pores of different sizes take place independently from each other. This investigation constitutes a starting point for the interpretation of water adsorption hysteresis by means of scanning
desorption measurements.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Although the potential use of water as a probe for surface
chemistry and pore structure characterization of porous carbon
materials has led to a lot of interest in the last years [1e9], much
remains to be done to improve our understanding of the involved
mechanisms. In fact, and even though several theories have been
proposed to describe the adsorption isotherm of water on activated
carbons, still none can be used satisfactorily for all types of activated carbons and in the whole range of relative pressure [2].
The use of water adsorption is attractive because: i) it can be
performed at room temperature, ii) water has a very small kinetic
diameter (0.28 nm) [6] which allows it to enter pores even smaller
than the ones accessible to carbon dioxide [10] and iii) it is sensitive
to surface chemistry. Currently, the most accepted mechanism that
can better describe the adsorption of water on activated carbons
claims that water molecules are rst adsorbed on primary active
sites on the surface of the carbon. Then the adsorbed water

* Corresponding author.
E-mail address: leticia.fernandez@mil.be (L.F. Velasco).
http://dx.doi.org/10.1016/j.carbon.2015.10.017
0008-6223/ 2015 Elsevier Ltd. All rights reserved.

molecules act as secondary sites for further adsorption of water to

form a cluster and as the relative pressure increases these clusters
grow and nally merge to ll the micropores [11]. So although at
low relative pressures the controlling factor in water adsorption is
the surface chemistry of the carbon, at higher relative pressures
both surface chemistry and pore structure play an important role
for the water adsorption/desorption mechanism [2,12]. Bearing this
in mind, it is clear that adsorption of water on activated carbons is
quite different from that of non-associating simple uids, such as
nitrogen, carbon dioxide, or hydrocarbons [13]. The main underlying differences arise from the strong waterewater interactions,
the weak waterecarbon interactions, and the formation of
hydrogen bonds with oxygen and nitrogen groups on the surface of
the material. Hence, the models developed to describe the behaviour of simple uids in conned spaces cannot be applied to water
in porous carbons [1].
Moreover, water adsorption in activated carbons exhibits a
hysteresis loop that cannot be explained by capillary condensation
theory. The hysteresis indicates that during the adsorption and/or
desorption process the water molecules form different molecular
structures [14]. The properties of the carbon material, such as pore
shape, pore connectivity and surface chemistry, but also

754

L.F. Velasco et al. / Carbon 96 (2016) 753e758

Fig. 1. Comparison between the water (273 K) and the nitrogen (77 K) isotherms
obtained for the same activated carbon (BPL).

operational variables like temperature or equilibration time, inuence the position, the extension and width of the water hysteresis loop [12,15e17].
Nowadays, there is a lack of methods to perform a complete and
integrated characterization of the textural and surface properties of
activated carbons in the micropore region (pore size < 2 nm). In fact
nitrogen and argon adsorption at their boiling temperatures (77
and 87 K, respectively) are reversible up to pore diameters of ca.
3 nm [18], whereas water is one of the few adsorbents to exhibit
hysteresis even in micropores. Hence, it is worthwhile to investigate the potential of water hysteresis scanning isotherms for
obtaining more information about the textural and surface features
of these materials (this can be visualized in Fig. 1 and will be further
discussed in the results section). Hysteresis scanning isotherms are
measured by reversing the direction of the gas pressure variation in
the adsorption or desorption process only in the relative pressure
range where hysteresis occurs, and once the initial hysteretic
adsorption/desorption curve (the so-called boundary adsorption/
desorption isotherm) has been obtained.
Scanning hysteresis behaviour within the context of typical
capillary condensation into mesopores was already studied for the
rst time by Van Bemmelen in 1897 [19], and a few decades later by
Everett and others (ref [20] and references therein), who pointed
out that the description of the scanning curves cannot only be
achieved based on the models of adsorption in individual pores and
that it is necessary to consider a cooperative nature of capillary
condensation/evaporation, including pore network effects.
Although for some years this methodology seems to have been set
aside, very recently it has gained the attention of the adsorption
scientic community because of the availability of ordered mesoporous molecular sieves which serve as model substances. Specifically, this methodology was applied to N2, Ar and Xe isotherms in
order to obtain new useful information about the pore network

geometry, including its connectivity and pore size distribution, of

porous materials with different structures, which cannot be
revealed from the main adsorption and desorption isotherms [21].
In the case of water adsorption, the study of the scanning curves
within the hysteresis loop has been limited to theoretical studies
[22] or to isolated experiments on very specic samples [6,23].
Therefore, it seemed necessary to carry out a systematic study
by water scanning desorption isotherms of well-characterized
activated carbons and thus investigate the potential of this technique to get new useful information regarding the textural and
surface characterization of this kind of materials.
To attain this goal, and in order to have a complete picture of the
process, three commercial activated carbons with different textural
properties and hydrophilicity were fully characterized and then
submitted to water desorption scanning measurements, as
described in the experimental section. The varied features of these
materials, especially their different surface oxygen content, made it
possible to distinguish between the inuence of the pore structure
and the surface chemistry on the observed water scanning
desorption performance.
Summing up, this work constitutes a rst systematic approach
to explore the usefulness of water desorption scanning isotherms
for (i) investigating the underlying mechanisms that govern the
water sorption processes in carbon micropores, (ii) the potential of
water desorption hysteresis scanning as a complimentary tool for
the pore structure and surface chemistry characterization of activated carbons.
2. Experimental
The selected activated carbons were manufactured by Blcher
GmbH (Blcher), Cabot Norit (110362) and Calgon Carbon (BPL) and
a comparison between their textural properties and oxygen content
is presented in Table 1. The choice of these three activated carbons
allowed us to perform the following sequential study: i) rst of all, a
highly hydrophobic activated carbon with a highly developed
porosity (Blcher) was studied in order to isolate the effect of the
textural properties of the material on its water scanning behaviour,
and ii) the contribution of the surface chemistry was evaluated by
the comparison of the results obtained with two activated carbons
(Norit and BPL) with similar textural properties and different (high)
oxygen contents.
Complete and scanning water desorption isotherms were carried out at 293 K using a gravimetric water sorption analyser
(Aquadyne DVS, Quantachrome Instruments). Scanning isotherms
are normally performed by repeatedly loading and unloading the
adsorbent in the relative pressure range where hysteresis occurs.
However, in the case of water sorption on surfaces with polar
surface functionality, irreversible changes to the surface chemistry
cannot be ruled out and, therefore, we used a fresh sample from the
same homogeneous batch to carry out each water sorption measurement. Thus, after the initial adsorption/desorption isotherm
was obtained, another fresh sample was taken and the adsorption
branch was measured only up to a relative pressure which is
smaller than the saturation pressure. Consequently only a portion
of the pore system was lled with the probe molecule before the

Table 1
Textural parameters, calculated from the N2 adsorption isotherms at 77 K, and oxygen content, measured by XPS, of the selected activated carbons.
Activated carbon

SBET (m2/g)

Vmeso (cm3/g)

Vmicro (cm3/g)
NLDFT

Oxygen content (wt. %)

Blcher
110362
BPL

2317
1076
1002

1.61
0.12
0.15

0.47
0.39
0.37

1.9
16.9
13.4

L.F. Velasco et al. / Carbon 96 (2016) 753e758

755

starting of the desorption isotherm. This procedure, called

desorption scanning, was repeated for several cycles with each
material.
The pore structure of the samples was also extensively characterized by means of high-resolution N2 (77.4 K) and CO2 (273.15 K)
adsorption experiments using an Autosorb-1 sorption instrument
(Quantachrome Instruments) coupled with advanced data reduction methods based on density functional theory (i.e. NLDFT and
QSDFT). Before each sorption measurement the samples were
outgassed at 423 K for 12 h under vacuum (10
5e10
6 Torr). Also,
the oxygen content of the samples was measured by X-ray photoelectron spectroscopy (XPS) using an XR3E2 twin anode X-ray
source (300 W, VG Microtech) and a Clam2 hemispherical electron
energy analyser. The base pressure of the analysis chamber was
about 5  10
9 mbar. Samples were analysed using an Mg Ka
(1253.6 eV) anode, without monochromation.
3. Results and discussion
3.1. Inuence of textural properties on the scanning desorption
behaviour
First of all, and in order to eliminate the inuence of specic
interactions between water and the oxygen-containing groups of
the material on the isotherm, an activated carbon (Blcher) with a
very low oxygen content (1.9 wt. %) and a highly developed porosity
(SBET 2317 m2/g) was studied. As mentioned before, this will
allow to isolate the effect of the textural properties of the activated
carbon on its water scanning behaviour and to simplify the interpretation of the scanning desorption isotherms.
Regarding the textural characterization, a type I/IV [24] nitrogen
(77 K) isotherm was obtained for Blcher carbon (Fig. 2a). The
corresponding pore size distribution was calculated by assuming
slit/cylindrical pore geometry and using the QSDFT method. This
kernel takes into account the heterogeneity and surface roughness
of the carbon and allows one to obtain a more reliable pore size
distribution [25,26]. CO2 adsorption at 273 K was also measured
(Fig. 2b), thus overcoming the problems of restricted diffusion of
nitrogen and argon at cryogenic temperatures, allowing this
adsorbate to enter the narrow microporosity [27]. The coupled N2
and CO2 pore size calculations (Fig. 2c) indicate that this material
presents both micropores (with a maximum at 1 nm) and narrow
mesopores (in the range 3e5 nm).
The water adsorption isotherm (Fig. 3) reveals three separate
parts: i) a region up to a relative pressure of 0.6, where almost no
adsorption takes place, due to the highly hydrophobic character of
this sample, ii) then a steep rise of the water adsorption is observed,

Fig. 3. Water scanning desorption isotherms for Blcher activated carbon.

which is associated with the lling of the micropores and narrow

mesopores, and iii) a second adsorption step at relative pressures
>0.9, associated with the lling of the some wider mesopores. Both
micropore and mesopore lling are accompanied by hysteresis
which might be caused by different mechanisms of adsorption/
pore lling and desorption [16]. Based on these results, both nitrogen and water isotherms lead to the same description of the kind
of porosity of this activated carbon. However, and as a consequence
of the relatively weak waterecarbon interactions, the micro and
mesopore lling during water adsorption is shifted signicantly to
higher relative pressures as compared to nitrogen adsorption at
77 K, which corresponds to a complete wetting situation. Actually,
and as it can be inferred from the Blcher water isotherm, this
adsorbate can only ll a part of the narrow mesoporosity of this
activated carbon. Specically, the steep rise of the water uptake at
the nal part of the isotherm (p/p0 > 0.9) indicates that the mesopore system is not being completely lled by water.
Regarding the water scanning desorption isotherms (Fig. 3), it is
clearly observed that in all the cases the rst point of the desorption
scanning isotherm perfectly matches with the envelope curve, thus
pointing out the homogeneity of the sample and the repeatability
of the experiments. Also the shape of the hysteresis scanning curves
resembles the shape of the boundary curve. Desorption scanning
curves starting at relative pressures larger than 0.9 (e.g. cycles 0.91
and 0.96) follow essentially exactly the boundary curve whereas

Fig. 2. a) N2 (77 K) and b) CO2 (273 K) isotherms for Blcher; c) pore size distribution obtained by the slit/cylindrical N2-QSDFT and the slit CO2-NLDFT kernels.

756

L.F. Velasco et al. / Carbon 96 (2016) 753e758

scanning curves starting at lower relative pressures deviate from

the boundary desorption curve and are shifted to slightly higher
relative pressures.
Based on these results, the shape of the hysteresis loop associated with the pore lling of micropores seems to depend on the
degree of micropore lling. On the other hand, as described, this is
not observed when desorption scanning is measured starting at a
relative pressure where essentially all the micro- and narrow
mesopores have been lled with water, which is in agreement with
previous observations [6].
3.2. Inuence of surface chemistry on the scanning desorption
behaviour
After studying the effect of the textural properties of the activated carbon on the scanning desorption behaviour, we have
selected another activated carbon (sample 110362) with a high
oxygen content (16.9 wt. %) in order to evaluate the contribution of
the surface chemistry.
The N2 adsorption isotherm at 77 K (Fig. 4a) of this activated
carbon belongs to type I in the IUPAC classication [24] with an
appreciable contribution of type IV. It presents a steep raise at
low relative pressures, followed by an open knee, which is characteristic of a microporous material with the presence of narrow
and wide micropores. Also, the narrow hysteresis loop observed
in the nitrogen adsorption isotherm at relative pressures above
0.5 indicates a non-negligible development of the mesoporosity.
Additionally, the presence of ultramicropores (pore widths 
0.7 nm according to IUPAC) was conrmed by the information
obtained from high resolution CO2 (273 K) sorption experiments
(Fig. 4b). When coupling the DFT pore size distributions obtained
with N2 and CO2, it is possible to obtain a precise pore size/volume
distribution of the sample (Fig. 4c) over the complete range of
micro- (including the narrowest ultramicropores) and mesopores.
Regarding the water adsorption isotherm (Fig. 5), in this case,
and as a consequence of its high surface oxygen content, a rise in
the water uptake is observed even at low relative pressures (p/
p0 < 0.5). In this region the non-specic waterecarbon interactions
are less important than the specic interactions between water and
the oxygen containing groups and the amount of water adsorbed is
proportional to the number of oxygen groups present on the surface of the activated carbon. At higher relative pressures, where
waterewater interactions become predominant, there is a sharp
rise of the curve, related to the pore lling of the micropore system.
The third region of the water isotherm (p/p0 > 0.7) is less steep than
the second and it accounts for the lling of the wide micropores
and the narrow mesopores present in this activated carbon.

Fig. 5. Water scanning desorption isotherms for Norit 110362 activated carbon.

The scanning desorption isotherms (Fig. 5) show a very similar

prole than the one obtained for the Blcher carbon, with a shift
in the shape of the hysteresis loop as the micropore system is
getting lled. This tendency is even more noticeable in this sample, when comparing the rst (p/p0 0.55) and the last (p/
p0 0.7) scanning desorption isotherms. The similarity between
the observed scanning hysteresis behaviour obtained at medium
and high relative pressures with two activated carbons of
completely different surface chemistry suggests that the shape of
water sorption hysteresis loops observed in carbon micropores is
dominated by the pore structure, and not affected by differences
in surface chemistry. However it should be stated that the width
of hysteresis loop for the Blcher carbon is signicantly wider
than the hysteresis loop observed for Norit and shifted to much
higher relative pressures, which mainly reects the major differences in surface chemistry, although differences in pore size distributions can contribute here as well (the effect of surface
chemistry on the position and width of the hysteresis loop has
been recently demonstrated by using two periodic mesoporous
organosilica model materials of identical pore size/geometry but
dened differences in surface chemistry [12]).
However, and in order to corroborate this result, another activated carbon (BPL) with intermediate oxygen content and similar
textural properties to the more hydrophilic carbon (110362), was
investigated.

Fig. 4. a) N2 (77 K) and b) CO2 (273 K) isotherms for sample 110362; c) pore size distribution obtained by the slit/cylindrical N2-QSDFT and the slit CO2-NLDFT kernels.

L.F. Velasco et al. / Carbon 96 (2016) 753e758

757

Fig. 6. a) N2 (77 K) and b) CO2 (273 K) isotherms for sample BPL; c) pore size distribution obtained by the slit N2-QSDFT and the slit CO2-NLDFT kernels.

The N2 (77 K) adsorption isotherm of sample BPL is shown in

Fig. 6a. The shape of the isotherm, in accordance with previous
works [28], indicates that BPL is a predominantly microporous
material, with narrow and wide micropores (the latter pointed out
by the opening of the knee of the adsorption isotherm). Moreover,
high resolution N2 (77 K) and CO2 (273 K) sorption experiments
(Fig. 6a and b) have allowed us to obtain an accurate pore size
distribution of this activated carbon (Fig. 6c) by applying the slit N2QSDFT and the slit CO2-NLDFT kernels. This pore size distribution is
similar to the one obtained for sample 110362, but with a slight
difference: the absence of a maximum at 2e3 nm.
The prole of the BPL water adsorption isotherm (Fig. 7) is also
very similar to the one obtained for sample 110362. However, and
as it was expected, the water uptake at the rst stages of the
sorption measurement is larger for the most hydrophilic activated
carbon (110362). On the other hand, at medium and high relative
pressures both adsorption isotherms converge, and only at p/
p0 > 0.85, there is a more pronounced rise in the water uptake for
sample 110362, in accordance with its higher mesopore volume.
Also, and in consonance with previous works [12], it is observed
that the hysteresis loop for Norit (110362) is narrower than for BPL,
which might be caused by small differences in pore size distributions, and of course the signicant differences in surface chemistry
(BPL being more hydrophobic), which might affect the clustering
which is assumed to be the mechanism for lling the carbon micro-

Fig. 7. Water scanning desorption isotherms for BPL activated carbon.

and narrow mesopores [12,14].

Regarding the scanning desorption isotherms (Fig. 7), the same
trend in the evolution of the shape of the curves as the micropore
system is getting lled is observed. However, it is important to
remark that in this case, and opposite to what it was observed for
the other two materials, the scanning desorption measurement
carried out at p/p0 0.8, where the wide micropores/narrow
mesopores are getting lled, does not perfectly overlap with the
initial desorption isotherm. This is probably due to the more
complex porous structure of sample BPL, since pore network effects
may affect the pore evaporation/desorption of water from narrow
mesopores [21].
It is also noteworthy to mention that for the two carbons with
the higher oxygen content, all the scanning desorption curves
converge at the same relative pressure (0.4). This would indicate
indeed that micropores ll and empty successively with size, i.e.
independent from each other.
Also, water sorption scanning isotherms at low relative pressures (Fig. 8) were performed trying to reveal more information
regarding the rst steps of the adsorption process. It is shown that,
for this specic activated carbon (BPL), the scanning water
desorption curves overlap with the adsorption isotherm up to a
relative pressure of 0.25. Taking into account that there is a change

Fig. 8. Water scanning desorption isotherms at low relative pressures obtained for BPL
activated carbon.

758

L.F. Velasco et al. / Carbon 96 (2016) 753e758

in the slope of the water adsorption isotherm beyond this relative

pressure, this result indicates that the sorption mechanism governed by the specic interactions between water and the surface
oxygen containing groups and the subsequent cluster formation is
reversible. On the other hand, at higher relative pressures, it is
observed again that the water desorption scanning curves follow a
different trend from the boundary isotherm, thus indicating that
pore structure starts affecting the water adsorption mechanism.
4. Conclusions
In this work, we present for the rst time a systematic study by
water scanning desorption isotherms of three well-characterized
activated carbons with varied textural properties and surface
chemistry.
The gathered results indicate that in these activated carbons
water adsorptionedesorption phenomena for pores of different
sizes occur independently from each other (i.e. the wider the pore
the higher the relative pressure of pore lling) and also demonstrate that the shape of the hysteresis loops (both boundary curves
and scanning curves) is signicantly affected by the pore structure.
Furthermore, measurements carried out on activated carbons
with different hydrophilic/hydrophobic character have enabled us
to discriminate between the effect of the porous structure and the
surface chemistry of the samples on their scanning desorption
performance. The shape of the hysteresis curve and scanning hysteresis loops seems not to depend on the surface chemistry (but
rather on pore structure as mentioned). However, the relative
pressure where pore lling/desorption occurs as well as the width
of the boundary and scanning hysteresis loops are affected by
surface chemistry, i.e. reecting the effect of surface chemistry on
the water adsorption mechanism. In this respect, denitely more
experimental and theoretical/molecular simulation work is needed
here in order to arrive at more detailed understanding of the
interplay between surface chemistry and pore structure effects on
water adsorption in nanoporous carbons.
Summarizing, hysteresis scanning within the water sorption
hysteresis loop can be used as a potential technique to get further
insights into the water adsorption mechanisms. Nevertheless,
further experimental work and theoretical modelling are still
needed in order to better understand the underlying mechanisms
and to fully utilize the scanning of water hysteresis loops as a
complimentary structural and surface characterization tool for
nanoporous carbons.
Acknowledgements
The authors wish to thank Norit, Blcher GmbH and Calgon
Carbon, who kindly provided them with the activated carbon
samples.
References
[1] J.K. Brennan, T.J. Bandosz, K.T. Thomson, K.E. Gubbins, Water in porous carbons, Rev. Colloids Surf. A Phys. Eng. Asp. 187e188 (2001) 539e568.
[2] D. Mowla, D.D. Do, K. Kaneko, Adsorption of water vapor on activated carbon:
a brief overview, Chem. Phys. Carbon 28 (2003) 229e262.

[3] T. Horikawa, T. Sekida, J. Hayashi, M. Katoh, D.D. Do, A new adsorptiondesorption model for water adsorption in porous carbons, Carbon 49 (2011)
416e424.
[4] T. Kimura, H. Kanoh, T. Kanda, T. Ohkubo, Y. Hattori, Y. Higaonna, R. Denoyel,
K. Kaneko, Cluster-associated lling of water in hydrophobic carbon micropores, J. Phys. Chem. B 108 (2004) 14043e14048.
[5] A.M. Slasli, M. Jorge, F. Stoeckli, N.A. Seaton, Modelling of water adsorption by
activated carbons: effects of microporous structure and oxygen content,
Carbon 42 (10) (2004) 1947e1952.
[6] M. Thommes, C. Morlay, R. Ahmad, J.P. Joly, Assessing surface chemistry and
pore structure of active carbons by a combination of physisorption (H2O, Ar,
N2, CO2), XPS and TPD-MS, Adsorption 17 (2011) 653e661.
[7] T.X. Nguyen, S.K. Bhatia, How water adsorbs in hydrophobic nanospaces,
J. Phys. Chem. C 115 (2011) 16606e16612.

th, K. La

szlo, Water adsorption by carbons. Hydrophobicity and hydro[8] A. To

n (Ed.), Novel Carbon Adsorbents, Elsevier, Amsterphilicity, in: J.M.D. Tasco
dam, 2012, pp. 147e171.
~ ero, M. Vaclavikova, I. Berezovska,
[9] P. Lodewyckx, E. Raymundo-Pin

guin, G. Dobos, Suggested improvements in the parameters
M. Thommes, F. Be
used for describing the low relative pressure region of the water vapour
isotherms of activated carbons, Carbon 60 (2013) 556e558.
[10] P. Lodewyckx, The effect of water uptake in ultramicropores on the adsorption
of water vapour in activated carbon, Carbon 48 (2010) 2549e2553.
[11] M.M. Dubinin, V.V. Serpinsky, Isotherm equation for water-vapor adsorption
by microporous carbonaceous adsorbents, Carbon 5 (1981) 402e403.
ba, Combining nitrogen, argon and
[12] M. Thommes, J. Morell, K.A. Cychosz, M. Fro
water adsorption for advanced characterization of ordered mesoporous carbons (CMKs) and periodic mesoporous organosilicas (PMOs), Langmuir 29
(2013) 14893e14902.
[13] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity, second ed.,
Academic Press Inc, London, 1982.
[14] T. Ohba, K. Kaneko, Cluster-associated lling of water molecules in slit-shaped
graphitic nanopores, Mol. Phys. 105 (2007) 139e145.
[15] M. Nakamura, T. Ohba, P. Branton, H. Kanoh, K. Kaneko, Equilibrium-time and
pore width dependent hysteresis of water adsorption isotherm on hydrophobic microporous carbons, Carbon 48 (2010) 2549e2553.
[16] T. Ohba, K. Kaneko, Kinetically forbidden transformations of water molecular
assemblies in hydrophobic micropores, Langmuir 27 (2011) 7609e7613.
[17] T. Horikawa, N. Sakao, D.D. Do, Effects of temperature on water adsorption on
controlled microporous and mesoporous carbonaceous solids, Carbon 56
(2013) 183e192.
[18] M. Thommes, Physical adsorption characterization of ordered and amorphous
mesoporous materials, in: G.Q. Lu, X.S. Zhao (Eds.), Nanoporous Materials:
Science and Engineering, Imperial College Press, Oxford, 2004, pp. 317e364.
[19] J.M. Van Bemmelen, Die absorption das Wasser in den Kolloiden, besonders in
ure, Z Anorg. Allg. Chem. 13 (1897) 233e356.
dem Gel der Kieselsa
[20] D.H. Everett, Adsorption hysteresis, in: E.A. Flood (Ed.), The Solid-Gas Interface, Dekker, New York, 1967, pp. 1055e1113.
[21] R. Cimino, K.A. Cychosz, M. Thommes, A.V. Neimark, Experimental and
theoretical studies of scanning adsorption-desorption isotherms, Colloids
Surf. A Physicochem Eng. Asp. 437 (2013) 76e89.
[22] M. Hefti, M. Mazzotti, Modeling water vapor adsorption/desorption cycles,
Adsorption 20 (2014) 359e371.
[23] J. Miyawaki, K. Kaneko, Adsorption properties of porous carbons: inuence of
preadsorbed water on gas adsorption behaviour, in: Carbon Materials for
Today and Future, Turkish-japanese Joint Carbon Symposium, Istanbul,
Turkey, 2010.
[24] M. Thommes, K. Kaneko, A.V. Neimark, J.P. Olivier, F. Rodrguez-Reinoso,
J. Rouquerol, K.S.W. Sing, Physisorption of gases, with special reference to the
evaluation of surface area and pore size distribution (IUPAC Technical Report),
Pure Appl. Chem. 87 (9/10) (2015) 1051e1069.
[25] M. Thommes, K.A. Cychosz, A.V. Neimark, Novel carbon adsorbents, in:

n (Ed.), Advanced Physical Adsorption Characterization of NanoJ.M.D. Tasco
porous Carbons, Elsevier, Amsterdam, 2012, pp. 107e145.
[26] A.V. Neimark, Y. Lin, P.I. Ravikovitch, M. Thommes, Quenched solid density
functional theory and pore size analysis of micro-mesoporous carbons, Carbon 47 (2009) 1617e1628.
[27] J. Garrido, A. Linares-Solano, J.M. Martn Martnez, M. Molina-Sabio,
F. Rodrguez-Reinoso, R. Torregrosa, Use of nitrogen vs. carbon dioxide in the
characterization of activated carbons, Langmuir 3 (1987) 76e81.
[28] B.P. Russell, M.D. Levan, Pore size distribution of BPL activated carbon
determined by different methods, Carbon 32 (5) (1994) 845e855.