CO2 Activation of Olive Bagasse for Hydrogen
Storage
Najoua Bader and Ouederni Abdelmottaleb
Research Laboratory: Process Engineering and Industrials Systems National School of Engineers of Gabes, University of Gabes,
St. Omar Ibn Khattab, Gabes 6029, Tunisia; najoua.bader@gmail.com (for correspondence)
Published online 31 October 2016 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/ep.12514
Hydrogen is considered as the most promising future fuel,
as its combustion generates only water vapor besides energy.
However, finding efficient, safe and low-cost storage methods
is the basic impediment for the adoption of hydrogen as fuel.
For this purpose, an ultramicroporous activated carbon was
successfully synthesized from low-cost biomass residues; olive
bagasse. The carbon material was prepared throw physical
activation, using carbon dioxide as activating agent, under
controlled thermal conditions. Then, it was tested as hydro-
gen storage material at cryogenic and room temperature
conditions. The textural characterizations showed a highly
tailored porous texture adequate for gas adsorption. The
totality of the created microporosity is of narrow range
(d < 0.7 nm). Interestingly, the measured median pore width
was about 0.65 nm, likely the optimum pore size for hydro-
gen adsorption. This feature accorded the AC with exception-
al H2 storage capacity, achieving 3.34 wt % at liquid
nitrogen temperature. It has outperformed Zeolites and
MOFs. V C 2016 American Institute of Chemical Engineers Environ
Prog, 36: 315–324, 2017
Keywords: olive bagasse, CO2 activation, microporous
materials, narrow microporosity, hydrogen storage
INTRODUCTION
Hydrogen is the cleanest, sustainable and renewable ener-
gy carrier, and a hydrogen energy system is expected to pro-
gressively replace the existing fossil fuels in the future, the
latter are being depleted very fast and cause severe environ-
mental problems. In particular, one potential use of hydro-
gen lies in powering zero-emission vehicles via a proton
exchange membrane fuel cell to reduce atmosphere pollu-
tion. Hydrogen can be produced not only from the fossil
fuels, such as coal and natural gas, but also from wind, solar,
thermal, hydroelectric, biomass or municipal solid wastes
with no consumption of non-renewable resources and no
pollution of any kinds.
Moreover, hydrogen exhibits the highest heating value
per mass of all chemical fuels. However, one of the major
drawback of hydrogen as energy carrier is its low energy
density on a volumetric basis. For example, one liter of gaso-
line (31.7 MJ/L, 8.8 kW h/L) contains approximately six times
as much energy as a liter of H2 compressed to 70 MPa (4.7
MJ/L, 1.3 kW h/L) [1]. Hence, hydrogen storage is the bottle-
neck that holds up the implementation of “hydrogen econo-
my”. The classical H2 storage technologies show some
C 2016 American Institute of Chemical Engineers
V impact of the chemicals, which need to be eliminated after
Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
important drawbacks such as low density and high pressures
operations for compression [2], boil-off excessive energy con-
sumption for liquefaction [3]. To overcome these issues,
researchers have focused their attention on exploring
material-based hydrogen storage, due to their ability to
increase the density of H2 under safe operation conditions.
These materials are typically categorized according to hydro-
gen storage mechanism: (i) physisorption of molecular H2 on
microporous surfaces; (ii) chemisorption of atomic H2 on the
lattice framework of metal hydrides. Chemisorption is less
attractive due to relatively high temperatures required for
hydrogenation and release of H2. Therefore, a lot of attention
has been focused during last decades on studies of H2 physi-
sorption on different adsorbents including metal organic
framework (MOF) [4,5], Zeolites [6], and carbon-based mate-
rials [7,8]. Compared to crystalline materials (MOFs and Zeo-
lites) carbon-based adsorbents are more advantageous. This
is related to their low cost, good chemical, mechanical and
thermal stability, easier regeneration, low densities, and wide
diversity of bulk and pore structures [9]. Among them, we
find activated carbons (ACs) which are the result of chemical
or physical activations of any carbon-rich material. They
have been intensively studied as H2 storage materials, espe-
cially after the non reproducible spectacular results of Baker
and Rodriguez [10] and Dillon et al. [11]. ACs are easy to
obtain and their porous structure is more developed than
carbon nanotubes and carbon nanofibers. Indeed, recent
studies showed that the cryogenic hydrogen storage capaci-
ties of ACs are quite promising. Yang et al. reported a stor-
age capacity of 6.9 wt % at 77 K and 20 bar for a zeolite-like
carbon material [12]. Wang et al. prepared a kind of activated
carbons which exhibits high specific surface area and
achieved a large hydrogen uptake of 7.08 wt % at 77 K and
20 bar [13]. However, Fierro et al. described a hydrogen stor-
age density of 6 wt % at 77 K and 40 bar for an activated car-
bon derived from a high-rank mineral coal [14]. The porous
texture is the main feature determining the hydrogen adsorp-
tion capacity of ACs. Many authors have stated that ACs
should have a highly developed internal surface and a large
volume of micropores [15] with diameters in the range of
0.6–0.8 nm [16–18] for an efficient performance as a hydro-
gen storage material at both atmospheric and high pressures.
Hence tailoring pore structures of carbon’s surface is crucial.
Recently, KOH activation of carbonaceous precursor has
attracted a great attention as it has often formed porous car-
bons with high surface areas and pore volume [19,21]. How-
ever, its main drawbacks are the price and the environmental
January 2017 315
the treatment by thorough washing of the resulting carbon.
In contrast, the physical activation considered to be more
environmentally friendly, and is commonly carried out in a
two step process. The carbonaceous raw material is first
pyrolyzed in inert atmosphere, generally below 8008C, and
the resulting char is then partly gasified between 700 and
10008C with an activating agent that usually consists in steam
and CO2 [22].
As Mediterranean country, olive cultivation is particularly
widespread throughout Tunisia. The mean annual olive oil
production for the decade 2004–2014 is 180 thousand tonnes
and reached 299 thousand tonnes in 2015 [23]. Therefore,
olive pomace with high olive stones content is a very abun-
dant agricultural by-product in Tunisia, and many results
obtained made clear that this lignocellulosic precursor is a
very adequate raw material to obtain efficient activated car-
bons [24,25].
The main goal of the present study is to prepare an AC
from olive stones by physical activation with CO2 for hydro-
gen storage purpose. The morphology and textural proper-
ties of the produced carbon material were investigated, using
N2 and CO2 adsorption, X-ray diffraction, and scanning elec-
tron microscopy. Finally, its hydrogen sorption property was
performed in three different international laboratories; using
verified volumetric automated and laboratory-made devices
which ensure high reproducibility.
MATERIALS AND METHODS
Activated Carbons Preparation
Carbon materials were prepared from olive bagasse. First,
the raw precursor was washed abundantly with hot distilled
water to obtain grains of olive stones sized to about 1–3 mm.
Then, dried olive stones were activated through physical
process. First, the carbonization of the above solid was car-
ried out in a quartz tubular reactor with a length of 115 cm
using nitrogen as the sweep gas. The reactor was heated
externally by a horizontal electrical furnace, whose tempera-
ture is controlled using a thermocouple feedback installed
inside the reactor. The temperature was maintained at 6008C
for 2 h, while the heating rate was 108C min21 under nitro-
gen flow (100 cm3.min21). Activation of the resultant char
was, then, conducted in the same tubular furnace, under a
CO2 flow of 100 cm3 min21. The temperature of activation
was set at 9008C for 8 h, under heating rate of 108C min21.
The granular carbon material was denoted as OB-CO2. One
other activated carbon was prepared for comparison pur-
pose. The carbonization step is the same as described above.
However, the activating gas was water steam under a flow of
100 cm3 min21, and was left for 2 h at 7008C. The resultant
material was also in granular form and labeled as OB-H2O.
Material Characterization
The textural properties of activated carbon were deter-
mined by N2 and CO2 sorption at 77K and 273K, respective-
ly, using an automated manometric system ASAP2020
(Micromeritics Instrument Corporation). Before each mea-
surement, about 100 mg of the sample were degassed for
24 h at 2508C and 10 26Torr. From N2 adsorption isotherm,
the specific surface area was estimated by applying the Bru-
nauer, Emmett and Teller (BET) within the range of relative
pressure (P/P8) from 0.01 to 0.10. The total pore volume was
determined from the amount of nitrogen adsorbed at P/P85
0.99. While, t-plot method was applied to measure the total
micropore area (Sinternal) and total micropore volume (pore
size smaller than 2 nm). Moreover, the total micropore vol-
ume was determined by the application of Dubinin Radush-
kivish (DR) equation, and the pore size distributions (PSD)
were calculated by density functional theory (DFT)
316 January 2017 Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
considering slit-shaped pores and by using Horvath-Kawazoe
(HK) method. Adsorption of CO2 at 273 K was used to assess
the narrowest micropores (pore size smaller than 0.7 nm)
[26]. Elemental analysis of olive stones, its charred and activat-
ed samples were carried out using ELTRA elemental analyzers;
CS-800 and ONH-2000. Scanning electron micrographs of the
charred and the corresponding activated carbon were
recorded by Dual Beam TM – SEM/FIB (Nova 600 NanoLab). It
combines ultra-high resolution field emission scanning elec-
tron microscopy (SEM) and precise focused ion beam (FIB).
X-ray diffraction (XRD) experiment for OB-CO2 sample was
performed using a Philips X’ Pert diffractometer (Co_Ka radia-
tion, k 5 1. 7903 A8) equipped with position sensitive detector
and monochromator in the diffracted beam. XRD pattern was
recorded for 2u in the range [58;1358], under operating condi-
tions of 40 kV and 40 mA, at a step of 0.058 and 10 s per step.
Measurement of Hydrogen Storage Capacities
H2 storage was measured in three independent laborato-
ries in volumetric devices. In all experiments, high-purity H2
( 99.9999%) was used for the uptake measurements. H2
uptakes were calculated in gravimetric basis (wt %).
At 77K and Up to 1 Bar
Determination of H2 storage at liquid nitrogen tempera-
ture and up to 1 bar was carried out at Alicante University
(Laboratory of Advanced Materials, Spain). The automated
apparatus was the one used for N2 and CO2 adsorption
measurements. In each adsorption test 100 mg of the pre-
pared AC sample were degassed at 2508C under vacuum pri-
or to adsorption measurements for at least 6 h.
At Temperature Ranged from 77 K to Room Temperature and up to
25 Bar
At these conditions, the measurements were performed at
Max Planck Institute for Intelligent Systems in Stuttgart
(Hydrogen Storage Group, Germany). During the adsorption
tests a commercial Sieverts type apparatus (PCT Pro 2000
with microdoser, HyEnergy) was used. This apparatus mea-
sures automatically adsorption and desorption isotherms in a
sample volume of 1.3 mL. Before each measurement, the
sample (about 100 mg) was heated under vacuum at 2008C
for 24 h. the relative excess uptake in wt % is calculated by:
nexc 3M
exc:uptake51003 (1)
nexc 3M1m
M denotes the molar mass of hydrogen (M 5 2.01588 g.
mol21); m denotes sample mass.
The absolute hydrogen adsorption is calculated by the
multiplication of a correction factor:
!
P3M
nabsolute 5nexc 3 11 (2)
Z3R3qliq 3T
qliq : hydrogen liquid density(assumed to 0.07 g mL21); R:
ideal gas constant(8.314472 J K21 mol21); Z: correction factor
for the non ideal gas given by Ref. 27; P in bar and T in K.
A 77 K and Room Temperature and up to 200 Bar
These measurements were carried out with a laboratory-
built volumetric device at Institut de Recherche sur l’Hydro-
gène (Universite de Quebec a Trois Rivières, Canada). Before
measurements, the carbon sample was outgassed for 12 h at
2508C. The apparatus consists of calibrated reference cell
which was attached to sample cell via Swagelok quick con-
nect. The volume of the reference cell measured at ambient
temperature was found to be 6.60 mL. Once the sample was
Table 1. Elemental analysis of the stones, the charred and
the resultant activated carbon sample.

Materials C H O S N
Stones 47.5 4.36 35.3 0.83 2.02
char 89.6 2.38 6.48 0.00 ca 0.33
OB-CO2 90.2 0.48 6.25 0.00 ca 0.48

Figure 1. N2 adsorption-desorption isotherms at 77K of CO2

and steam activated carbons.
loaded on the system, it was typically left for 24 h outgassing
at room temperature using the turbomolecular pump. For
the measurements at 77 K, the sample cell was immersed
into liquid nitrogen dewar flask to a predetermined height.
In order to maintain isothermal conditions, entire Sievert’s
system is placed inside a thermostated laboratory whose
temperature is maintained constant within 60.58C. The rela-
tive excess uptake in wt % was calculated by:
madsorbate
Hydrogen uptake5 3100 (3)
madsorbent
Isosteric Heat of Adsorption Measurements
In cryogenic hydrogen storage applications, the heat of
adsorption is a very important characteristic and has to be
identified because hydrogen uptake at the low pressure
range is strongly governed by this value. The apparent isos-
teric enthalpy of adsorption of hydrogen into the carbon
pores, DHads, can be estimated for different surface coverage
using Vant’ Hoff’s equation:
! vation by the creation of mesoporosity (two broad peaks at
oðlnPÞ DHads
  5 (4)
o T1 R
u tered in narrow range as demonstrated by Figure 2b. The
where P: pressure in bar; T in K; R: molar gas constant and h
is the surface coverage. For calculations we used adsorption
isotherms at 77 K (liquid nitrogen) and 87 K (liquid argon)
and up to 25 bar.
RESULTS AND DISCUSSION
Surface Chemistry and Porosity Development During
Activations
As can be seen from the chemical analyses presented in
Table 1, the raw biomass shows high carbon and low sulfur
Figure 2. PSD using DFT method of CO2 and steam activat-
ed samples (a); H-K pore size distribution of OB-CO2 sample
(b).
contents. The carbonization of the stones followed by CO2
activation increase the carbon content up to 90% (wt/wt),
decrease the hydrogen and oxygen contents and eliminate
sulphur from the carbon matrix. The nitrogen content
remains almost unaltered during the activation process which
can be indicative of the inclusion of nitrogen into the aro-
matic structure and not in functional groups on the layer
edges.
The N2 adsorption desorption isotherms of OB-CO2 and
OB-H2O are shown in Figure 1. One can conclude that both
the activating agents generate microporous activated carbons,
because of the Type I characteristic of the isotherms [28].
However, the main difference lays in the N2 adsorption
capacity revealing the difference on the created surface area.
On the other hand, by observing the PSD (Figure 2a) of the
two samples one can note that CO2 activation is more effec-
tive on microporosity creation than water steam. Moreover,
the figure shows a heterogeneous PSD after water steam acti-
d > 2 nm). In contrast, CO2 activation leads to highly micro-
porous activated carbons with well controlled porosity cen-
plot is perfectly mono-modal with one maximum centered
on 0.65 nm. Recently, Nabais et al. [29] also stated that CO2
activation of lignocellulosic precursor (coffee endocarp)
leads to samples with higher BET surface areas and pore vol-
umes when compared with samples produced by steam acti-
vation and with similar burn-off values. The different textural
parameters deduced from N2 and CO2 adsorption at 77K and
273K, respectively are gathered in Table 2. For the charred
sample the nitrogen sorption measurement was also per-
formed but no data were collected, indicative of diffusion
restrictions on the adsorption of N2 at 77 K. However, at a
much temperature carbon dioxide can penetrate due to an

Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep January 2017 317
Table 2. Textural parameters of char, CO2 and steam-activated olive stones deduced from N2 and CO2 adsorption isotherms.

SBET Sinternal VT pore VDR-N2 Vmeso VDR-CO2 DHK

Materials (m2/g) (m2/g) (cm3/g) (cm3/g) (cm3/g) % micro (cm3/g) (nm)
Char 296* – – 0.03 – – 0.17 –
OP-CO2 1185 1078 0.52 0.43 0.09 83 0.47 0.65
OP-H2O 351 – 0.18 0.15 0.03 83 0.15 –

*Measured with CO2 adsorption isotherm at 273 K; DHK: median pore width calculated by HK method; % micro 5 VDR-N2/VT pore.

Table 3. Comparison of our results with previous works on CO2 activation of different raw materials.

N2 pyrolysis CO2 activation SBET VT pore Vmicro VDR-CO2

Precursors conditions conditions (m2/g) (cm3/g) (cm3/g) (cm3/g) References
Eucalyptus T 5 8008C T 5 8008C 1190 0.52 0.48 0.35 Tancredi et al.
wood t52 h Flow rate 5 200 mL/min 1996 [30]
Flow rate 5 100 mL/min T rate 5 10 K/min
T rate 5 10 K/min
Furnace 5 horizantal
Bituminous T 5 7508C T 5 9008C 754 – 0.31 – Ahmadpour
coal t52 h Flow rate 5 150 mL/min et al. 1995 [32]
Flow rate 5 100 mL/min T rate 5 158C/min
T rate 5 38C/min Furnace 5 vertical
Furnace 5 vertical
Olive-mill T 5 3008C; 8008C T 5 8408C 793 – – 0.26 Moreno Castilla
waste t 5 3 h; t 5 2h Flow rate 5 300 mL/min et al. 2001[33]
Flow rate 5 300 mL/min
T rate 5 108C/min
Quercus T 5 4508C T 5 8408C 1201 0.67 0.51 0.011 Sanchez et al.
agrifolia t52 h t52 h 2001 [31]
wood T rate 5 208C/min Flow rate 5 500 mL/min
Furnace 5 horizantal
Olive stones T 5 6008C T 5 8508C 778 1.4 0.4 – Roman et al.
t51 h t52 h 2008 [34]
Flow rate 5 150 mL/min Flow rate 5 40 mL/min
Coffee T 5 6008C T 5 7008C 1287 0.64 0.53 – Nabais et al.
endocarp t51 h Flow rate 5 85 mL/min 2008 [29]
Flow rate 5 85 mL/min
T rate 5 108C/min
Olive stones T 5 6008C T 58008C 909 0.44 – 0.29 Plaza et al.
2009 [35]
Flow rate 5 10 mL/min
oak Intermediate pyrolysis T 5 9008C 724 0.32 0.27 – Jung and Kim
(pyrolysis plant) t 51h 2014 [36]
Flow rate 5 10 NL/min flow rate 510 NL/min
T rate 5 108C/min
724
Olive stones T 5 6008C T 5 9008C 1185 0.52 0.42 0.47 This study
t52 h t58 h
Flow rate 5 100 mL/min Flow rate 5 100 mL/min
T rate 5 108C/min T rate 5 108C/min
Furnace 5 horizantal Furnace 5 horizantal

T: temperature, t: time, T rate: temperature rate: furnace: conception of the used furnace.

activated adsorption process and BET surface area and the
micropore volume were measured. Consequently, the
carbonized material must be essentially microporous, with
surface area of about 300 m2/g basically in microporosity
only slightly larger than the carbon dioxide diameter.
Actually, carbonization of lignocellulosic materials implies a
set of reactions such as depolymerization, cracking, and
dehydration of lignin and cellulose, thus leading to a char
with a rudimentary pore structure, and other products such
as tar and volatile liquids and gas. After steam and CO2 gasi-
fication this porosity will be more accessible to N2 and
almost all the textural parameters are measured. These two
agents extract carbon atom from the structure of the porous
carbon according to the following stoichiometric equations

318 January 2017 Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
Figure 3. SEM micrographs of the outer surface (a, b) and transverse sections (c, d) of the activated carbon OB-CO2.
[22], thus the porosity has been widened and became acces-
sible to the nitrogen:
C1CO2 52CO DH5 1159 kJ mol21
C1H2 O5CO1H2 DH5 1117 kJ mol21
The two reactions are endothermic which facilitates the
accurate control of experimental conditions in the furnace.
Let’s start with CO2-activated sample. According to Table 2, it
can be seen that the prepared carbon exhibits a relatively
large surface area, and the porosity is almost totally devel-
oped internally (Sinternal 5 91% SBET). As expected, the carbon
is highly microporous, and the microporosity represents 83%
of the total developed porosity.
The most surprising is the almost equality registered
between VDR-N2 and VDR-CO2 with high values (0.47cm3/g)
which reveals a uniform PSD localized in a narrow range.
This means that all the operation conditions are optimized
(gasification temperature, time, burn off degree, etc . . ..)
even though with prolonged activation time (8 h). To show
the exceptional carbon characteristics a comparison of our
results with others is presented in Table 3. The overview of
the table indicates that our result is among the best achieved
in terms of carbon’s porosity. However a direct comparison
Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
of the volume of narrow micropores (VDR-CO2) shows that
OB-CO2 sample has the highest measured value. On the oth-
er hand the difference between VDR-CO2 and the micropore
volume measured with N2 adsorption (indicative of porosity
homogeneity) is the lowest in this study. All these observa-
tions evidence that this prepared sample has a tailored and
well ordered porosity. Therefore it is noteworthy to stat that
carbon’s porosity could be tuned via simple two-step CO2
activation process. Actually, this well-ordered pore system of
the OB-CO2 is the result of contribution of many factors.
First, it is the nature of the raw material, because Tancredi
et al. [30] and Sanchez et al. [31]who activated materials
based on wood, found very low VDR-CO2 values even though
the overall created specific surface and pore volume are sim-
ilar to our result. This means that the composition of the raw
material is crucial factor which affect the porosity develop-
ment. Whereas, the results are worst with gasification of coal
[32].
The second reason is the use of horizontal furnace. Actu-
ally, the horizontal conception limits the external mass-
transport, and consequently, will favor the internal burning
of the char particle and the creation of micropores rather
than their widening. Also, we shouldn’t ignore the effect of
low heating rate (108C min21) on the development of a uni-
form porosity [37]. Finally, one other important reason is the
January 2017 319
Figure 4. XRD pattern of OB-CO2 sample.
activating agent itself: carbon dioxide. This is clearly elabo-
rated after the comparison with the steam activated sample
(see Table 2). However, OB-H2O sample is very similar to
the char, which reveals the incomplete steam gasification
process. Actually, this is very surprising because even though
water steam is less effective as activating agent than carbon
dioxide, we didn’t expect such poorly developed porosity.
When the activating gas comes in contact with the char, it
will react with both the exterior and the interior of the parti-
cle, the latter requiring previous diffusion of the activating
gas. At the activation temperature used in the preparation of
OB-H2O, 7008C, the diffusion rate of water molecules to the
interior of the particle is sufficiently high to expect a devel-
opment of the whole range of porosity, at least in the early
stages of activation. However, the progress of the reaction
brings about the production of CO and H2, which are inhibi-
tors of the reaction (especially H2). The increase in the con-
centration of inhibitors in the interior of the particle will
favor the reaction of steam at the exterior, with the subse-
quent destruction of porosity [38]. Hence there is a widening
of the microporosity, since VDR-N2 increases and VDR-CO2
remains almost constant. To prevent this inhibition the tem-
perature must be further enhanced, in order to minimize the
concentration of the inhibitors. Consequently, the creation of
microporosity is maximized in the OB-CO2 sample (activa-
tion temperature 5 9008C).
Morphology Observations and Structural Analyses: SEM
and XRD
Micrographs of the outer surface and transverse sections
of OB-CO2 sample are shown in Figure 3. As can be seen,
the micrographs of sections of the outer surface (Figures 3a
and 3b) are significantly different from their corresponding
transverse sections (Figures 3c and 3d). On the former two
images, except for a few cracks (Figure 3b), the surface is
plain and does not manifest any pores at this level of magni-
fication and resolution of the SEM. Therefore a transverse
section is recommended in order to see the form of the
microporosity developed during CO2 activation. On the latter
two images (Figures 3c and 3d) the pores are clearly seen as
flat and irregular holes with diameter  5mm at the level of
magnification and resolution of the SEM used in this work.
The XRD pattern of the same sample is plotted in Figure 4. It
indicates a typical amorphous carbon and shows abroad
asymmetric peaks at corresponding to 2u 258 and 2u 508.
However, there are some sharp peaks at 2u 5 30, 35 and 458,
which proves a little high degree of graphitization. This is
due to the ordered distribution of crystallites during nitrogen
carbonization step [39]. Subsequently, the high activation
320 January 2017 Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
Figure 5. Low-pressure excess adsorption-desorption iso-
therms of H2 on OB-CO2 and OB-H2O samples at 21968C.
[Color figure can be viewed at wileyonlinelibrary.com]
temperature (9008C) leads to some edge orientation and
reduce the concentration of nonparallel single layers.
Hydrogen Storage Capacity
Cryogenic Hydrogen Storage
Figure 5 shows the hydrogen adsorption-desorption iso-
therms of CO2 and steam activated carbons at 77K and up to
1 bar. The two investigated carbon materials show a type I
adsorption isotherm with an initial steep increase of the
hydrogen uptake at low pressures and an almost horizontal
plateau at higher pressures. This type of isotherms is charac-
teristic for monolayer adsorption on microporous solids [40].
No multilayers of hydrogen are formed at 77K, because the
interaction strength between single layers is too weak at tem-
perature higher than the critical temperature of H2 [41].
Moreover, the sorption was completely reversible, indicating
physisorption of hydrogen on pore walls of the carbons. As
expected, OB-CO2 adsorbs more than OB-H2O sample,
thanks to its highly developed porosity. Thus, there are more
active sites for H2 adsorption. At 1 bar, the adsorption capac-
ity of OB-CO2 sample reaches 1.69 wt %, which is a consid-
erable amount especially when compared to values reached
by non-carbonaceous porous materials, as presented in Table
4. The overview of the Table 4 reveals that at the adopted
operation conditions OB-CO2 has outperformed a large
number of nanoporous materials such as Zeolites, polymers
and MOFs
In addition, these materials are high cost, non sustainable
and their production processes are not reproducible. Surpris-
ingly, OB-CO2 sample has outperformed nanoporous materi-
als with higher specific surface area [46,48,49], which reveals
that surface area is not the only parameter that needs to be
regarded. Even though, many investigators have shown that,
at 77K and subatmospheic pressures, the hydrogen uptake
was governed by the microporosity and the specific surface
area of the tested adsorbents [13,52,53]. Actually, the only
exceptional feature of OB-CO2 sample is its ultramicroporous
character, with a well ordered pore system centered at
0.65 nm. This observation clearly points out that the smallest
pores are much more efficient in H2 sorption due a stronger
interaction with H2 molecules. Therefore, the pore size is
also a key factor that governs cryogenic H2 adsorption at cer-
tain pressure conditions. On the other hand, this result justi-
fies previous works which stated the existence of an
optimum pore size for H2 adsorption [16,54–56]. These
Table 4. Comparison of physical properties and hydrogen uptake capacities of our CO2 activated carbon with non carbona-
ceous and porous materials.

Materials SBET (m2.g21) T (K) P (bar) H2 uptakes (wt %) References

Zeolites (NaA) 725 77 15 1.54 Langmi H. et al., 2003 [42]
298 0.28
Zeolites (CaX) 669 77 15 2.19 Langmi H. et al., 2005 [43]
Prussian Blue 870 77 1 1.4 Kaye et Long, 2005 [44]
Polymers of intrinsic microporosity (PIM) 830 77 1 1.43 Neil et al., 2006 [45]
Hypercrosslinked polystyrene gel 1930 77 1.2 1.55 Germain et al., 2006 [46]
Nanoporous polymers 1043 77 60 3.7 Yuan et al., 2009 [47]
298 70 0.43
Zr-MOF 1367 77 1 1.5 Ren et al., 2014 [48]
MOF (Mil-101) 1715 77 1 1.65 Ren et al., 2014 [49]
Zr-fumarate MOF 849 77 1 1.38 Ren et al., 2015 [50]
Fly ash derived Zeolites X 404 77 1 0.7 Musoyka et al., 2015 [51]
CO2 activated olive stones (OB-CO2) 1185 77 1 1.69 This study
77 25 2.59
77 200 3.34
298 200 0.63

Figure 6. Excess hydrogen adsorption-desorption isotherms

of OB-CO2 samples at 77K and up to 25 bar.
optimum values can hold two layers of adsorbed hydrogen
into slit pores.
Figure 6 plots the excess H2 adsorption-desorption iso-
therms of OB-CO2 sample at higher pressure (up to 25 bar).
The type I isotherm is more pronounced than in Figure 4, as
at higher pressure the adsorption sites are more fully saturat-
ed. At 25 bar the H2 adsorption capacity of the carbon
reaches 2.59 wt %. The enhancement on H2 adsorption after
the increase of pressures from 1 to 25 bar is about 53%
which is a moderate improvement. This is probably related
to the very narrow microporosity of this sample, which satu-
rate at very low pressure. This is the case of carbide derived
carbons which have shown an exceptional H2 adsorption
capacity but they have been saturated at only 1 bar [57].
The high-pressure excess and absolute adsorption iso-
therms at 77K for OB-CO2 sample are illustrated by Figure 7.
With the increase of testing pressure, excess adsorption
reached the maximum value (2.84 wt %) at approximately 30
bar and then decreased step by step, which is totally differ-
ent from absolute hydrogen storage. Theoretically, when the
gas densities in the pore and the bulk gas are increasing at
the same rate with respect to pressure, the excess adsorption
will achieve a maximum value. Consequently, continuous
increase in pressure has no positive effect upon the amount
adsorbed. At higher pressure over the critical point, the gas
density in the sample pores remains unchanged while the
bulk gas density keeps increasing, resulting in a negative
gain in the excess adsorption [58]. Absolute adsorption was
Figure 7. Excess and absolute H2 adsorption isotherms of
OB-CO2 at 77K and high pressures. [Color figure can be
viewed at wileyonlinelibrary.com]
also calculated using Eq. (2), because it quantifies the
amount of hydrogen present in the adsorbed layer and in
the adsorbent pores. It is crucial measurement for the design
of storage system tanks. As shown in Figure 7 the absolute
adsorption isotherm exhibits Type I or Langmuirian shape.
The difference between excess and absolute adsorption
become more significant at high pressure (from 60 bar),
because there is compression contribution and the benefit of
adsorption is lost [59]. At 200 bar, the absolute adsorption
amount of our lab-made carbon reaches 3.34 wt %, that is a
considerable amount. These data highlight that hydrogen
storage on our biomass-based carbon can be an interesting
option, especially in wind and solar hybrid systems, wherein
weight and volume are not limiting factors. Noting that at
200 bar we still working on safe conditions and the H2 stor-
age system is controllable.
Room Temperature Hydrogen Storage
The Figure 8 illustrates the H2 adsorption-desorption iso-
therms of OB-CO2 at room temperature (298K). In contrast
to low temperature measurements, at room temperature the
adsorption isotherm is far from saturation and is described
with a linear equation, according to the Henry-type behavior.
Only fractional coverage takes place and the adsorbed mole-
cules behave like two-dimensional ideal gas. The overall

Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep January 2017 321

Figure 8. High pressure adsorption-desorption isotherms of
H2 on OB-CO2 at room temperature (298K).
Figure 9. Isosteric heat of adsorption for OB-CO2 versus
surface coverage.
adsorption capacity is very low and doesn’t exceed 1 wt %.
Also at 200 bar and room temperature there is no compres-
sion contribution and H2 is stored by simply physisorption
process. According to Fierro et al. [59], at room temperature
there is compression contribution only when the applied
pressures are above 300 bar.
Isosteric Heat of Adsorption
In Figure 9 the isosteric enthalpy of adsorption is shown
in dependence of the surface coverage, which is the hydro-
gen uptake normalized to the uptake at 25 bar and 77K. The
isosteric enthalpy of adsorption starts at 4 kJ.mol21 and it
decreases slowly up to 0.6 surface coverage with an average
of 3.54 kJ.mol21. Then it decreases strongly to achieve 0.06
kJ.mol21 when the carbon surface is totally covered with H2
molecules. This reveals the effect of pore size on the carbon-
H2 binding energy. Only the narrow micropores accord the
carbon with strong affinity to H2 molecules. After their filling
the binding energy between carbon surface and H2
decreases dramatically.
CONCLUSIONS
Olive bagasse whose disposal represents a serious prob-
lem in the oil industry, is valorized as a hydrogen sorbent
material. Carbon dioxide activation, at 9008C for 8 hours, of
the charred olive stones donates narrow microporosity to the
framework of the prepared activated carbon. Furthermore, a
homogeneous PSD is observed which is rarely seen in the
disordered activated carbon. This porous feature allows our
biosorbent to store at 77 K up to 3.34 under safe operational
322 January 2017 Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
conditions. Moreover, it was clearly shown that under 1 bar
only the narrow micropores contribute to H2 adsorption.
ACKNOWLEDGMENT
The authors gratefully acknowledge Dr. F. Reinoso-
Rodriguez (University of Alicante, Spain), Dr. R. Chahine
(IRH, Universite du Quebec a Trois Rivière, Canada) and Dr.
M. Hirscher (Max Planck Inst. for Intelligent Systems, Stutt-
gart, Germany) for accepting the research internship of Miss
N. Bader and University of Gabes for funding.
ABBREVIATIONS
BET Brunauer, Emmett and Teller
DFT Density functional theory
DR Dubinin Radushkivish
FIB Focused ion beam
HK Horvath-Kawazoe
MOF Metal organic framework
PSD Pore size distributions
SEM Scanning electron microscopy
XRD X-ray diffraction
LITERATURE CITED
1. Dresselhaus, M.S., & Thomas, I.L. (2001). Alternative
energy technologies, Nature, 414, 332–337.
2. Zhou, L. (2005). Progress and problems in hydrogen stor-
age methods, Renew Sustain Energy Reviews, 9, 395–408.
3. Z€uttel, A. (2003). Materials for hydrogen storage, Materi-
als Today, 6, 24–33.
4. Panella, B., Hirscher, M., P€ utter, H., & M€
uller, U. (2006).
Hydrogen adsorption in metal–organic frameworks: Cu-
MOFs and Zn-MOFs compared, Advances in Functional
Materials, 16, 520–524.
5. Rowsell, J.L.C., Millward, A.R., Park, K.S., & Yaghi, O.M.
(2004). Hydrogen Sorption in Functionalized Metal-
Organic Frameworks, Journal of the American Chemical
Society, 126, 5666–5667.
6. Li, Y., & Yang, R.T. (2006). Hydrogen storage in low silica
type X Zeolites, Journal of Physical Chemistry B, 110,
17175–17181.
7. Y€ur€
um, Y., Taralp, A., & Veziroglu, T.N. (2009). Storage
of hydrogen in nanostructured carbon materials, Interna-
tional Journal of Hydrogen Energy, 34, 3784–3798.
8. Yang, S.J., Jung, H., Kim, T., & Park, C.R. (2012). Recent
advances in hydrogen storage technologies based on
nanoporous carbon materials, Progress in Natural Sci-
ence: Materials International, 22, 631–638.
9. Schlapbach, L., & Z€ uttel, A. (2001). Hydrogen-storage
materials for mobile applications, Nature, 414, 353–358.
10. Chambers, A., Park, C., Baker, R.T.K., & Rodriguez, N.M.
(1998). Hydrogen Storage in Graphite Nanofibers, Journal
of Physical Chemistry B, 102, 4253–4256.
11. Dillon, A.C., Jones, K.M., Bekkedahl, T.A., Kiang, C.H.,
Bethune, D.S., & Heben, M.J. (1997). Storage of hydrogen
in single-walled carbon nanotubes, Nature, 386, 377–379.
12. Yang, Z., Xia, Y., & Mokaya, R. (2007). Enhanced hydro-
gen storage capacity of high surface area zeolite-like car-
bon materials, Journal of the American Chemical Society,
129, 1673–1679.
13. Wang, H., Gao, Q., & Hu, J. (2009). High hydrogen stor-
age capacity of porous carbons prepared by using acti-
vated carbon, Journal of the American Chemical Society,
131, 7016–7022.
14. Fierro, V., Szczurek, A., Zlotea, C., Mar^eche, J.F.,
Izquierdo, M.T., Albiniak, A., Latroche, M., Furdin, G., &
Celzard, A. (2010a). Experimental evidence of an upper
limit for hydrogen storage at 77 K on activated carbons,
Carbon, 48, 1902–1911.
15. Chahine, R., & Bose, T.K. (1994). Low-pressure adsorp-
tion storage of hydrogen, International Journal of Hydro-
gen Energy, 19, 161–164. 35.
16. Georgiev, P.A., Ross, D.K., Albers, P., & Ramirez-Cuesta,
A.J. (2006a). The rotational and translational dynamics of
molecular hydrogen physisorbed in activated carbon: A
direct probe of microporosity and hydrogen storage per- 36.
formance, Carbon, 44, 2724–2738.
17. Cabria, I., Lopez, M.J., & Alonso, J.A. (2007). The opti-
mum average nanopore size for hydrogen storage in car-
bon nanoporous materials, Carbon, 45, 2649–2658. 37.
18. Gogotsi, Y., Portet, C., Osswald, S., Simmons, J.M.,
Yildirim, T., Laudisio, G., & Fischer, J.E. (2009). Impor-
tance of pore size in high-pressure hydrogen storage by
porous carbons, International Journal of Hydrogen Ener- 38.
gy, 34, 6314–6319.
19. Sevilla, M., Mokaya, R., & Fuertes, A.B. (2011). Ultrahigh
surface area polypyrrole-based carbons with superior
performance for hydrogen storage, Energy & Environ- 39.
mental Science, 4, 2930–2936.
20. Sevilla, M., Alam, N., & Mokaya, R. (2010a). Enhancement
of hydrogen storage capacity of zeolite-templated car-
bons by chemical activation, Journal of Physical Chemis- 40.
try C, 114, 11314–11319.
21. Sevilla, M., Foulston, R., & Mokaya, R. (2010b). Superacti- 41.
vated carbide-derived carbons with high hydrogen stor-
age capacity, Energy & Environental Science, 3, 223–227.
22. Marsh, H., and Reinoso, F.R. (2006). Activated carbon, 42.
UK: Elsevier.
23. Office National de l’Huile (ONH), Repub. of Tunisia,
Available at: http://www.onh.com.tn/index.php/en/
la-production. (Accessed 20.10.2015).
24. Martınez, M.L., Torres, M.M., Guzman, C.A., & Maestri, 43.
D.M. (2006). Preparation and characteristics of activated
carbon from olive stones and walnut shells, Industrial
Crops and Products, 23, 23–28.
25. Rios, R.V.R.A., Martınez-Escandell, M., Molina-Sabio, M.,
& Rodrıguez-Reinoso, F. (2006). Carbon foam prepared 44.
by pyrolysis of olive stones under steam, Carbon, 44,
1448–1454.
26. Cazorla-Amoros, D., Alcaniz-Monge, J., de la Casa-Lillo,
M.A., & Linares-Solano, A. (1998). CO2 as an adsorptive 45.
to characterize carbon molecular sieves and activated car-
bon, Langmuir, 14, 4589–4569.
27. Ye, Y. (2001). Interaction of hydrogen with novel carbon
materials. phd. California Institute of Technology.
28. Sing, K.S.W. (1985). Reporting physisorption data for gas/
solid systems with special reference to the determination
of surface area and porosity (Recommendations 1984), 46.
Pure & Applied Chemistry, 57, 603–619.
29. Nabais, J.M.V., Nunes, P., Carrott, P.J.M., Ribeiro Carrott,
M.M.L., Garcıa, A.M., & Dıaz-Dıez, M.A. (2008). Production 47.
of activated carbons from coffee endocarp by CO2 and
steam activation, Fuel Process Technology, 89, 262–268.
30. Tancredi, N., Cordero, T., Rodrıguez-Mirasol, J., &
Rodrıguez, J.J. (1996). Activated carbons from Uruguayan 48.
eucalyptus wood, Fuel, 75, 1701–1706.
31. Sanchez, A.R., Elguezabal, A.A., & de La Torre Saenz, L.
(2001). CO2 activation of char from Quercus agrifolia
wood waste, Carbon, 39, 1367–1377.
32. Ahmadpour, A., & Do, D.D. (1996). The preparation of 49.
active carbons from coal by chemical and physical activa-
tion, Carbon, 34, 471–479.
33. Moreno-Castilla, C., Carrasco-Marın, F., L opez-Ram on,
M.V., and., & Alvarez-Merino, M.A. (2001). Chemical and
physical activation of olive-mill waste water to produce 50.
activated carbons, Carbon, 39, 1415–1420.
34. Roman, S., Gonzalez, J.F., Gonzalez-Garcıa, C.M., &
Zamora, F. (2008). Control of pore development during
Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
CO2 and steam activation of olive stones, Fuel Process
Technology, 89, 715–720.
Plaza, M.G., Pevida, C., Arias, B., Fermoso, J., Casal,
M.D., Martın, C.F., Rubiera, F., & Pis, J.J. (2009). Develop-
ment of low-cost biomass-based adsorbents for postcom-
bustion CO2 capture, Fuel, 88, 2442–2447.
Jung, S.H., & Kim, J.S. (2014). Production of biochars by
intermediate pyrolysis and activated carbons from oak by
three activation methods using CO2, Journal of Analytical
and Applied Pyrolysis, 107, 116–122.
Gonzalez, M.T., Rodrıguez-Reinoso, F., Garcıa, A.N., &
Marcilla, A. (1997). CO2 activation of olive stones carbon-
ized under different experimental conditions, Carbon, 35,
159–162.
Rodrıguez-Reinoso, F., Molina-Sabio, M., & Gonzalez,
M.T. (1995). The use of steam and CO2 as activating
agents in the preparation of activated carbons, Carbon,
33, 15–23.
Rodrıguez-Reinoso, F., Martın-Martınez, J.M., Molina-
Sabio, M., Perez-Lledo, I., & Prado-Burguete, C. (1985). A
comparison of the porous texture of two CO2 activated
botanic materials, Carbon, 23, 19–24.
Lowell, S., and Shields, J.E. (1991). Powder surface area
and porosity, Dordrecht: Springer Netherlands.
Hirscher, M. (2010). Handbook of hydrogen storage: New
materials for future energy storage, Weinheim: Wiley-
VCH.
Langmi, H.W., Walton, A., Al-Mamouri, M.M., Johnson,
S.R., Book, D., Speight, J.D., Edwards, P.P., Gameson, I.,
Anderson, P.A., & Harris, I.R. (2003). Hydrogen adsorp-
tion in zeolites A, X, Y and RHO, Journal of Alloys and
Compounds, 356–357, 710–715.
Langmi, H.W., Book, D., Walton, A., Johnson, S.R., Al-
Mamouri, M.M., Speight, J.D., Edwards, P.P., Harris, I.R.,
& Anderson, P.A. (2005). Hydrogen storage in ion-
exchanged zeolites, Journal of Alloys and Compounds,
404–406, 637–642.
Kaye, S.S., & Long, J.R. (2005). Hydrogen storage in the
dehydrated prussian blue analogues M3[Co(CN)6]2
(M 5 Mn, Fe, Co, Ni, Cu, Zn), Journal of the American
Chemical Society, 127, 6506–6507.
McKeown, N.B., Gahnem, B., Msayib, K.J., Budd, P.M.,
Tattershall, C.E., Mahmood, K., Tan, S., Book, D.,
Langmi, H.W., & Walton, A. (2006). Towards polymer-
based hydrogen storage materials: engineering ultrami-
croporous cavities within polymers of intrinsic micropo-
rosity, Angewandte Chemie International Edition, 45,
1804–1807.
Germain, J., Hradil, J., Frechet, J.M.J., & Svec, F. (2006).
High surface area nanoporous polymers for reversible
hydrogen storage, Chemical Materials, 18, 4430–4435.
Yuan, S., Kirklin, S., Dorney, B., Liu, D.J., & Yu, L.
(2009). Nanoporous polymers containing stereocontorted
cores for hydrogen storage, Macromolecules, 42, 1554–
1559.
Ren, J., Langmi, H.W., North, B.C., Mathe, M., &
Bessarabov, D. (2014a). Modulated synthesis of
zirconium-metal organic framework (Zr-MOF) for hydro-
gen storage applications, International Journal of Hydro-
gen Energy, 39, 890–895.
Ren, J., Musyoka, N.M., Langmi, H.W., Segakweng, T.,
North, B.C., Mathe, M., & Kang, X. (2014b). Modulated
synthesis of chromium-based metal-organic framework
(MIL-101) with enhanced hydrogen uptake, International
Journal of Hydrogen Energy, 39, 12018–12023.
Ren, J., Musyoka, N.M., Langmi, H.W., North, B.C.,
Mathe, M., Kang, X., & Liao, S. (2015). Hydrogen storage
in Zr-fumarate MOF, International Journal of Hydrogen
Energy, 40, 10542–10546.
January 2017 323
51. Musyoka, N.M., Ren, J., Langmi, H.W., North, B.C., &
Mathe, M. (2015). A comparison of hydrogen storage
capacity of commercial and fly ash-derived zeolite X
together with their respective templated carbon deriva-
tives, International Journal of Hydrogen Energy, 40,
12705–12712.
52. Kunowsky, M., Marco-Lozar, J.P., Oya, A., & Linares-
Solano, A. (2012). Hydrogen storage in CO2-activated
amorphous nanofibers and their monoliths, Carbon, 50,
1407–1416.
53. Texier-Mandoki, N., Dentzer, J., Piquero, T., Saadallah, S.,
David, P., & Vix-Guterl, C. (2004). Hydrogen storage in
activated carbon materials: Role of the nanoporous tex-
ture, Carbon, 42, 2744–2747.
54. Celzard, A., Fierro, V., Mareche, J.F., & Furdin, G., (2007).
Advanced preparative strategies for activated carbons
designed for the adsorptive storage of hydrogen, Adsorp-
tion Science & Technology, 25, 129–142.
324 January 2017 Environmental Progress & Sustainable Energy (Vol.36, No.1) DOI 10.1002/ep
55. Patchkovskii, S., Tse, J.S., Yurchenko, S.N., Zhechkov, L.,
Heine, T., & Seifert, G. (2005). Graphene nanostructures
as tunable storage media for molecular hydrogen, Pro-
ceedings of National Academy of Sciences U S A, 102,
10439–10444.
56. Rzepka, M., Lamp, P., & de la Casa-Lillo, M.A. (1998).
Physisorption of hydrogen on microporous carbon and
carbon nanotubes, Journal of Physical Chemistry B, 102,
10894–10898.
57. Dash, Ranjan Kumar Nanoporous Carbons Derived from
Binary Carbides and their Optimization for Hydrogen
Storage. Drexel University.
58. Menon, P.G. (1968). Adsorption at high pressures, Chemi-
cal Reviews, 68, 277–294.
59. Fierro, V., Zhao, W., Izquierdo, M.T., Aylon, E., &
Celzard, A. (2010b). Adsorption and compression contri-
butions to hydrogen storage in activated anthracites,
International Journal of Hydrogen Energy, 35, 9038–9045.