Page 1
This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and
education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.
Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies areencouraged to visit:
http://www.elsevier.com/copyright
Page 2
Author's personal copy
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ijhydene
Adsorption of hydrogen in nickel and rhodiumexchanged zeolite X
K.P. Prasantha, Renjith S. Pillaia, H.C. Bajaja, R.V. Jasraa,�, H.D. Chungb,T.H. Kimb, S.D. Songc
aDiscipline of Inorganic Materials and Catalysis, Central Salt and Marine Chemicals Research Institute (CSMCRI),
G.B. Marg, Bhavnagar 364 002, IndiabHydrogen System Center, KIER, Daejon, Republic of KoreacHanseo University, Chungnam, Republic of Korea
a r t i c l e i n f o
Article history:
Received 12 July 2007
Received in revised form
26 October 2007
Accepted 30 October 2007
Available online 3 December 2007
Keywords:
NiNaX
RhNaX
NaX
Hydrogen adsorption
Grand Canonical Monte Carlo
simulation
a b s t r a c t
Adsorption of hydrogen in zeolite NaX and its nickel and rhodium exchanged forms were
investigated at 77.4 K using a static volumetric adsorption system up to 1 bar, and at 303
and 333 K in a gravimetric adsorption system up to 5 bar. Hydrogen adsorption at 77.4 K for
NaX and the nickel and rhodium exchanged zeolite X was found to be reversible with
pressure. However, chemisorption of hydrogen was observed at 303 and 333 K. The highest
adsorption capacities for hydrogen in NaX, NiNaX and RhNaX were observed to be 24.28,
34.51 and 51.85 molecule/u.c at 333 K and 5 bar. Hydrogen uptake capacities of the nickel
and rhodium exchanged zeolite X at 303 and 333 K were found increasing initially up to 70%
of nickel and 55% of rhodium exchange, beyond which it showed a decreasing trend. The
observed decrease in adsorption capacity at higher Ni and Rh cation exchange is explained
in terms of partial degradation of the zeolite structure during cation exchange and the high
temperature vacuum dehydration process. Grand Canonical Monte Carlo simulations were
also performed to study the adsorption of hydrogen in NaX as well as nickel and rhodium
exchanged zeolites X at 77.4, 303 and 333 K and the simulated data were compared with
experimentally measured values.
& 2007 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights
reserved.
1. Introduction
The decreasing oil reserves and growing environmental
pollution awareness have necessitated search for alternate
fuels and zero emission vehicles. Fuel cell driven car with the
proton exchange membrane (PEM) fuel cell and hydrogen as a
fuel seems to be a promising alternate [1]. Though hydrogen
is a clean fuel giving water as the only by-product on its
combustion in air, on-board storage of hydrogen is still a
challenge. Various methods for storing hydrogen including
gaseous, liquid and solid state storage have been suggested
[2,3]. The important methods for storing hydrogen in
solid materials include: (i) storage in metals and alloys;
(ii) storage in complex hydrides (alanates, borides);
(iii) storage by trapping (e.g., clatharates); and (iv) storage in
microporous materials (carbons, metal-organic frameworks,
polymers, zeolitic materials, etc.). Among the best known
microporous solids, the zeolites have the advantage of high
stability, low cost and a variety of porous structures. Zeolites
are highly crystalline alumino-silicate materials with crystal-
lographically well-defined channels and cavities [4,5].
A general chemical formula of zeolites is represented as
ARTICLE IN PRESS
0360-3199/$ - see front matter & 2007 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.ijhydene.2007.10.047
�Corresponding author. Tel.: +91 278 2471793; fax: +91 278 2567562, +91 278 2566970.E-mail address: [email protected] (R.V. Jasra).
I N T E R N AT I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5
Page 3
Author's personal copy
Mx=n½ðAlO2ÞxðSiO2Þy� �wH2O, where M is an exchangeable
cation with valency n, w is the number of water molecules
and ratio y=x is having a value X1. Mainly alkali and alkaline
earth cations are present in the zeolite in the form of
exchangeable extra framework cations.
Recent studies have shown that zeolites can be potential
materials for reversible hydrogen storage [6–15]. Du and Wu
reported hydrogen adsorption capacity of 2.55 wt% for NaX at
77 K and 70 bar pressure [7]. Langmi et al. reported the
hydrogen storage capacities of 1.54, 1.79, 1.81 and 2.19 wt%
obtained at 77 K and 15 bar for zeolites NaA, NaX, NaY and
CaX respectively [8]. Zecchina et al. obtained a hydrogen
capacity of 1.28 wt% on H-SSZ-13 zeolite at 77 K and 0.92 bar
[9]. Vitillo et al. predicted a theoretical maximal hydrogen
storage capacity of 2.86 wt% for FAU and RHO zeotypes [10].
However, experimental data on hydrogen adsorption in
zeolites at room temperature are sparse. Weitkamp et al.
reported hydrogen uptake capacity of 0.023 and 0.030 wt% in
NaX, NaY and NaA zeolites, respectively, at 573 K and 100 bar
[11]. Kayiran et al. reported hydrogen uptake capacity of 0.12
and 0.10 wt% for KA and NaA zeolites at 293 K and 10 bar
pressure [12]. It has been reported that the amount of
hydrogen adsorbed on zeolites depends on its framework
structure, composition and acidity [13,14]. The cations in the
framework are considered as the adsorption sites for hydro-
gen. Hydrogen adsorption studies in zeolites are largely
confined to alkali and alkaline earth metal cations as
exchangeable extra framework cations with few studies
reported in transition metal exchanged zeolites [15,16]. In
order to examine the role of framework structure and
exchangeable cations in hydrogen adsorption, we report
hydrogen adsorption in zeolite X exchanged with Ni2+ and
Rh3+ ions in this paper. Nickel and rhodium are known to
have affinity for hydrogen and form hydrides on reaction with
hydrogen and are used as catalysts for various reactions such
as hydrogenation and hydroformylation [17,18].
Molecular simulation of adsorption phenomena in zeolite is
emerging as a rapid, cost-effective method for the evaluation
of potential adsorbents. The grand canonical Monte Carlo
simulation technique is particularly adapted to calculate the
equilibrium adsorption isotherm measurements. For exam-
ple, Razmus and Hall [19] used Monte Carlo simulation to
reproduce the experimental single component equilibrium
adsorption isotherms of nitrogen, oxygen and argon in zeolite
5A. Watanabe et al. [20] used grand canonical Monte Carlo
simulation method to study air separation properties of
zeolite A, X and Y. Song et al. [21] used grand canonical
Monte Carlo procedure for the study of the effect of cation
type, available volume, surface area, temperature and pres-
sure on the hydrogen uptake on model zeolites. Berg et al. [22]
studied the hydrogen uptake isotherms in the microporous
sodalite-type structures by means of grand canonical Monte
Carlo calculations. Grand canonical Monte Carlo simulations
are suitable for establishing a correlation between the
microscopic behavior of the zeolite and adsorbate system
with the macroscopic properties such as adsorption iso-
therms which are measured experimentally [23]. In the
present paper, grand canonical Monte Carlo simulations were
performed to study the adsorption of hydrogen in NaX as well
as nickel and rhodium exchanged zeolites X at 77.4, 303 and
333 K and simulated data were compared with experimentally
measured values.
2. Experimental
2.1. Materials
Zeolite X powder procured from Zeolites and Allied Products,
Bombay, India, has a chemical composition Na88Al88
Si104O384 � 220H2O. Nickel chloride and rhodium chloride
purchased from E Merck India Ltd., Bombay, India, were used
as the starting materials for the adsorbent preparation.
Hydrogen with 99.9995% purity was procured from Hydrogas
India Pvt. Ltd., Bombay, India. Hydrogen was further purified
by passing through a molecular sieve trap prior to its use for
the adsorption isotherm measurements.
2.2. Preparation of cation exchanged zeolite X
Nickel and rhodium cations were introduced into sodium
form of zeolite X by the conventional cation exchange
method from aqueous solution. Typically, the zeolite NaX
was treated with 0.05 M aqueous solution of nickel chloride or
rhodium chloride with solid/liquid ratio of 1:80 at 353 K for
4 h. The residue was filtered, washed with hot distilled water,
until the washings were free from chloride ions as tested by
silver nitrate solution and dried in air at room temperature.
Zeolite samples having different amounts of nickel and
rhodium were prepared by subjecting repeated ion exchange
into the zeolites. The extent of nickel and rhodium ions
exchanged on different zeolite X samples were determined by
atomic emission spectroscopy (AES) and inductively coupled
plasma (ICP) analysis using Perkin-Elmer optical emission
spectrometer (Optima 2000 DV). The following terminology is
used to describe the ion exchanged samples: the first letters
show the exchanged cation and the number at the end shows
the percentage of sodium cation present in NaX exchanged
with this cation. For example, NiNaX-49 means 49% of Na+
ions present in NaX are replaced with Ni2+ ions and RhNaX-29
means 29% of Na+ ions present in NaX are replaced with Rh3+
ions.
3. Characterization
3.1. X-ray powder diffraction
The X-ray powder diffraction measurements at ambient
temperature were carried out using a PHILIPS X’pert MPD
system in the 2y range of 5–651 at a scan speed of 0.11 s�1
using CuKa1 radiation ðl ¼ 1:54056 AÞ. The diffraction patterns
of the starting material shows it is highly crystalline showing
reflections in the 2y range 5–351 typical of zeolite X.
Percentage crystallinity of the transition metal ion exchanged
zeolites were determined from the X-ray diffraction pattern
by summation of the intensities of 10 major peaks at 2yvalues 6.1, 10.0, 15.5, 20.1, 23.4, 26.7, 29.3, 30.5, 31.0 and 32.1.
The sodium form of the zeolite X was considered as an
arbitrary standard for the calculations.
ARTICLE IN PRESS
I N T E R N A T I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5736
Page 4
Author's personal copy
3.2. Diffuse reflectance infrared Fourier transform (DRIFT)spectroscopy
DRIFT spectroscopic studies of adsorbed H2 at 303 K in nickel
and rhodium exchanged zeolite X samples were carried out in
a Perkin-Elmer Spectrum GX FT-IR instrument equipped with
the selector DRIFT accessory (Graseby Specac, P/N 199900
series) incorporating an environmental chamber (EC) assem-
bly (Graseby Specac, P/N 19930 series). The zeolite samples
were activated in situ at 673 K under a vacuum of 5�
10�3 mmHg at a heating rate of 10 K min�1 using an automatic
temperature controller (Graseby Specac, P/N 19930 series)
connected with the EC. The sample was kept at 673 K for
60 min and then cooled to 303 K, and the IR spectra were
collected under vacuum. H2 was purges on the sample at
1 atm pressure, kept for 60 min, and the IR spectra were
collected. Typically 60 scans were co-added at a resolution of
4 cm�1.
3.3. Adsorption isotherm measurements
Hydrogen adsorption measurements at 77.4 K were done by
static volumetric system (Micromeritics Instrument Corpora-
tion, USA, model ASAP 2010) up to 1 bar pressure. The
samples were activated by heating at a rate of 1 K min�1 to
673 K under vacuum (5� 10�3 mmHg) and the temperature
and vacuum were maintained for about 8 h before the
sorption measurements. The amount of activated sample
was determined from the weight of the samples before as
well as after activation and prior to start of adsorption
measurements.
Hydrogen adsorption measurements at 303 and 333 K were
performed in a gravimetric adsorption analysis system
(Gravimetric Sorption Analyzer, VTI Co., USA) up to 5 bar
pressure. The samples were activated in situ by controlled
heating up to 673 K (heating rate 1 K min�1) under high
vacuum (1� 10�3 mmHg) for 8 h before the sorption measure-
ments. After activation, the samples were allowed to cool
down to the desired temperature and the temperature was
maintained during the analysis using an external water
circulator (PolyScience, USA). Hydrogen with a definite
pressure steps was introduced into the sample chamber.
The increase in the weight of the sample due to hydrogen
adsorption was accurately measured using a microbalance
(Cahn D-200, USA) connected to the sample holder. Deso-
rption of hydrogen on heating the zeolite samples was
performed in gravimetric sorption analyzer. The samples
were heated at a rate of 1 K min�1 up to 423 K after the
adsorption analysis, without removing the samples from
sample holder.
3.4. Surface area measurements
The Langmuir surface area of the cation exchanged zeolites
was calculated by fitting the hydrogen adsorption isotherm
data at 77.4 K in Langmuir isotherm model. The cross-
sectional area of one hydrogen molecule is taken as
0:142 nm2 [24] for the calculation of Langmuir surface area
based on dry weight of adsorbent samples.
4. Computational methodology
NaX, NiX and RhX were modeled using the recently refined
structure reported by Zhu and Seff [25], Bae and Seff [26] and
Busch et al. [27], respectively. Four aluminum atoms of the
unit cell were converted randomly to silicon atoms in order to
get the desired number of 88 aluminum atoms per unit cell
close to Si/Al ratio of 1.2 of the sample used for experimental
measurements. The zeolite framework was built in accor-
dance with Lowenstein’s Al–O–Al avoidance rule. For NaX, the
distribution of the extra framework cations was taken from
Zhu et al., with 28 Na+ ions in SI0 sites, 28 Na+ ions in SII sites,
28 in SIII0 sites, 6-ring and 12-ring, respectively. The distribu-
tion of extra framework cations for NiX was taken as follows:
Ni2+ ions are present at two different SIII0 sites, 20 six
coordinate octahedral Ni2+ ions are found in first SIII0 sites
and 10 Ni2+ ions which are three coordinate occupy second
SIII0 sites, 8 Ni2+ ions in SI0 sites, 4 Ni2+ ions in SII sites and 2
Ni2+ ions in SI sites. In case of RhX the extra framework
cations are located in SIII sites. All structures were energy
minimized for starting simulation with low energy structures
by Cerius2 minimizer using the force field by Ramsahye et al.
[28] with the constraint that the unit cell remains cubic. Only
extra framework cations were considered to be movable
during energy minimization.
The force field used for the gas adsorption simulations was
a modified version of the Cerius2 Watanabe–Austin potential
energy model [20,29]. The total energy of the zeolite frame-
work and adsorbed molecules (U) is expressed as the sum of
the interactions energy between the adsorbate and zeolite
(UAZÞ and that between the adsorbates ðUAAÞ molecules.
U ¼ UAZ þ UAA. (1)
Both UAZ and UAA are written as the sums of pair-wise
additive potentials, uij, in the form
uij ¼ 4�ijsij
rij
!12
�sij
rij
!624
35þ qiqj
rij
!, (2)
where the first term is the repulsion–dispersion LJ potential,
with �ij and sij corresponding to the parameter sets for each
interacting pair that is obtained from �i and si of each species
by using the Lorentz Berthelot mixing rule (i.e., a geometric
combining rule for the energy and an arithmetic one for the
atomic size: �ij ¼ ð�i�jÞ1=2 and sij ¼ ðsi þ sjÞ=2). The second term
is the coulombic contribution between point charges qi and qj
separated by distance rij.
Electric multipoles of hydrogen molecules are described by
a linear electric quadrupole corresponding to the distribution
of three effective charges. For hydrogen, we used a three-
point charge model [20,29,30], where the two outer sites that
are separated by a distance of l ¼ 0:7414 A have a charge q ¼
�0:21e [31] and the third midpoint has a point charge of �2q.
The hydrogen molecule is also considered as the LJ interac-
tion site and the LJ potential parameters are taken from the
literature as given in Table 1 [30,32]. The zeolite framework is
assumed to be semi-ionic, with the partial charges for silicon
and oxygen fixed as þ2:4e and �1:2e, respectively, as usually
considered values [33,34]. The charges on sodium were taken
as þ0:85e to reproduce the heat of adsorption reported [7] at
ARTICLE IN PRESS
I N T E R N AT I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5 737
Page 5
Author's personal copy
lower pressure range of the adsorption at 77.4 K. We chose
þ1:60e and þ2:35e charges for Ni2+ and Rh3+, respectively [33].
The values of partial charges on aluminum in sodium form of
zeolite X and the cation exchanged zeolites were taken as
þ1:55e and þ1:60e, respectively, in order to take into account
the partial charge transfer from the framework, with the
constraint of a global charge of zero for the faujasite system.
The force field LJ parameters [20,33,35] used for the simula-
tions for zeolite framework atoms are given in Table 1.
Absolute adsorption isotherms were then computed using a
grand canonical Monte Carlo calculation algorithm, as
implemented in the sorption module of Cerius2 software
suite [36]. The evolution of the total energy over the Monte
Carlo steps was plotted in order to monitor the equilibration
conditions. The zeolite structure was assumed to be rigid
during the sorption process and the extra framework cations
were maintained fixed in their initial optimized positions.
The Ewald summation method [30] was used for calculating
the electrostatic interactions and the short range interac-
tions, with a cutoff distance of 12 A. The Lennard-Jones
potential for the adsorbate–zeolite interactions and both the
Lennard-Jones and coulombic terms of the adsorbate–adsor-
bate interactions were calculated using the minimum image
convention [29] with a real space potential cut-off distance of
12.0 A. The simulations were performed at a temperature of
77.4, 303 and 333 K using one unit cell of each zeolite for 3–5
million Monte Carlo steps.
5. Results and discussion
Powder X-ray diffraction patterns of nickel and rhodium
exchanged zeolite X samples are shown in Fig. 1. The XRD
patterns show the retention of the zeolite structure after ion
exchange. But loss of crystallinity was observed during the
ion exchange and the dehydration process; the loss of
crystallinity increases with increase in ion exchange levels.
The percentage crystallinity of the ion exchanged zeolite
samples was determined from the X-ray powder patterns and
the values are given in Table 2.
It was reported that the stability of the nickel exchanged
zeolite decreases with increase in nickel ion exchange [25].
The hydrated Ni2+ and Rh3+ ions hydrolyze within the zeolite
cavity, with de-alumination of the zeolite framework [37]. The
crystal structure of the fully dehydrated nickel exchanged
zeolite X shows hydronium ions in the secondary coordina-
tion spheres of hydrolyzed Ni2+ ions. This hydronium ion
acidity is likely to be responsible for the loss of crystallinity up
on further dehydration of both nickel and rhodium ex-
changed zeolite X.
The Langmuir surface area values, crystallinity and chemi-
cal composition of the ion exchanged zeolite X samples are
shown in Table 2. The surface areas of the nickel and rhodium
exchanged zeolite X decreases continuously as the percen-
tage of cation exchange level increases. The decrease in
surface areas may be attributed to the loss of crystallinity of
the zeolite samples as the cation exchange level increases.
Fig. 2 shows the relation between the percentage of cation
exchange, Langmuir surface area and hydrogen adsorption
capacity at 77.4 K for both nickel and rhodium exchanged
zeolite X samples.
Zeolite X is a synthetic aluminum rich analogue of the
naturally occurring mineral faujasite. The 14-hedron with 24
vertices known as the sodalite cavity or b-cage may be viewed
as its principal building block. These b-cages are connected
ARTICLE IN PRESS
10 20 30 40 50 60
NiNaX-75
NaX
Inte
nsity (
a.u
)
2θ
NiNaX-70
NiNaX-65
NiNaX-73
NiNaX-49
10 20 30 40 50 60
RhNaX-73
2θ
RhNaX-55
RhNaX-69
RhNaX-44
Inte
nsity (
a.u
)
RhNaX-29
NaX
Fig. 1 – X-ray powder diffraction pattern of (a) NaX and NiNaX and (b) NaX and RhNaX zeolites.
Table 1 – Lennard-Jones parameters used for adsorba-te–adsorbate and adsorbate–zeolite interactions
Atom type s (A) � ðkcal mol�1Þ References
H 2.81 0.0171 [30]
Si 0.076 0.0370 [20]
Al 1.140 0.0384 [20]
O 3.040 0.3342 [20]
Na 1.746 0.1008 [33]
Ni 2.833 0.015 [35]
Rh 2.929 0.053 [35]
I N T E R N A T I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5738
Page 6
Author's personal copy
tetrahedrally at six-rings by bridging oxygen to give double
six-rings (D6Rs, hexagonal prisms) and concomitantly, an
interconnected set of even larger cavities (super cage)
accessible in three dimensions through 12-ring (24-mem-
bered) windows. The Si and Al atoms occupy the vertices of
these polyhedra. The oxygen atoms lie approximately mid-
way between each pair of Si and Al atoms but are displaced
from those points to give near-tetrahedral angles about Si and
Al. Single six rings (S6Rs) are shared by sodalite and super
cage and may be viewed as the entrances to the sodalite units.
Each unit cell has eight sodalite units, eight super cages, 16
D6Rs, 16 12-rings and 32 S6Rs.
Exchangeable cations that balance the negative charge of
the aluminosilicate framework are found within the zeolite
cavities. These are usually found at the following sites shown
in Fig. 3, site I at the center of the D6R, I0 in the sodalite cavity
on the opposite side of one of the D6Rs six-rings from site I, II0
inside the sodalite cavity near an S6R, II at the center of the
S6R or displaced from this point into a super cage, III in the
super cage on a twofold axis opposite to a four-ring between
two 12-rings and III0 somewhat or substantially off III (off the
twofold axis) on the inner surface of the super cage. During
the cation exchange process, two sodium ions are replaced by
ARTICLE IN PRESS
0
100
200
300
400
500
600
700
0 49 65 70 73 75
% of ion exchange
Langm
uir s
urf
ace a
rea (
m2/g
)
0
20
40
60
80
100
120
140
160
180
200H
2 A
dsorp
tion (
mole
cule
/u.c
)Langmuir Surface area
H2 Adsorption at 1bar,77K
0
100
200
300
400
500
600
700
0 29 44 55 69 73
% of ion exchange
Langm
uir s
urf
ace a
rea (
m2/g
)
-10
10
30
50
70
90
110
130
150
H2 a
dso
rptio
n (
mo
lecu
le/u
.c)
Langmuir Surface area
H2 Adsorption at 1bar, 77K
Fig. 2 – Correlation between percentage of ion exchange, Langmuir surface area and hydrogen adsorption capacity of
(a) nickel exchanged zeolite X and (b) rhodium exchanged zeolite X.
Table 2 – Crystallinity, Langmuir surface area and chemical composition of nickel and rhodium exchanged zeolite X
Zeolite sample % Crystallinity Langmuir surface area ðm2 g�1Þ Unit cell formula (on dry basis)
NaX 100 665 Na88Al88Si104O384
NiNaX-49 74 515 Ni21.7Na44.5Al88Si104O384
NiNaX-65 60 398 Ni28.7Na30.6Al88Si104O384
NiNaX-70 45 322 Ni30.8Na26.2Al88Si104O384
NiNaX-73 12 268 Ni32.5Na22.9Al88Si104O384
NiNaX-75 10 256 Ni33.1Na21.7Al88Si104O384
RhNaX-29 54 465 Rh8.5Na62.5Al88Si104O384
RhNaX-44 34 251 Rh12.9Na49.2Al88Si104O384
RhNaX-55 25 116 Rh16.1Na39.5Al88Si104O384
RhNaX-69 18 84 Rh20.2Na27.3Al88Si104O384
RhNaX-73 14 51 Rh21.4Na23.6Al88Si104O384
Fig. 3 – The framework structure of zeolite X, near the center
of the each line segment is an oxygen atom. The numbers
1–4 indicate the different oxygen atoms. Silicon and
aluminum atoms alternate at the tetrahedral intersections,
except that Si substitutes for Al at about 4% of the Al
positions. Extra framework cation positions are labeled
with roman numerals.
I N T E R N AT I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5 739
Page 7
Author's personal copy
one nickel ion and three sodium ions are replaced by one
rhodium ion.
Hydrogen adsorption isotherms on nickel and rhodium
exchanged zeolite X were measured at 77.4 K and are shown
in Fig. 4. All the hydrogen adsorption isotherms obtained at
77.4 K are Type I in nature, showing adsorption of hydrogen
inside microporous zeolite cavities. The hydrogen adsorption
capacity was found decreasing as the ion exchange levels of
nickel and rhodium increase. At 77.4 K NaX shows a
maximum hydrogen adsorption capacity of 114.7 molecule/
u.c (1.7 wt%) which is comparable to the previous literature
data. For example Kazansky et al. [14] have reported 1.22 wt%
of hydrogen adsorbed at 77.4 K and 0.61 bar. Langmi et al. [8]
have reported hydrogen adsorption capacity of 1.79 wt% and
Du and Wu [7] have reported 2.28 wt% for NaX at 77.4 K and
15 bar. The adsorption of hydrogen at 77.4 K in all the nickel
and rhodium exchanged NaX was highly reversible with
decrease in pressure. Fig. 5 shows the simulated adsorption
isotherms of hydrogen at 77.4 K up to 1 bar in fully nickel and
rhodium exchanged zeolite NaX and also the experimental
ARTICLE IN PRESS
0.0 0.2 0.4 0.6 0.8 1.0 1.20
20
40
60
80
100
120
NaX NiNaX-49 NiNaX-65
NiNaX-70 NiNaX-73 NiNaX-75
mole
cule
/u.c
Pressure (bar)
0.0 0.2 0.4 0.6 0.8 1.0 1.20
20
40
60
80
100
120
mole
cule
/u.c
Pressure (bar)
NaX
RhNaX-69 RhNaX-73RhNaX-55
RhNaX-44RhNaX-29
Fig. 4 – Experimental hydrogen adsorption isotherms at 77.4 K up to 1 bar for (a) nickel exchanged zeolite X and (b) rhodium
exchanged zeolite X.
0.0 0.2 0.4 0.6 0.8 1.0 1.20
20
40
60
80
100
120
140
160
180
200
220
240
Closed symbol - Adsorption
Open symbol - Desorption
RhX simNiX simNaX sim
RhNaX-44 expNiNaX-70 expNaX exp
mole
cule
/u.c
Pressure (bar)
Fig. 5 – Simulated and experimental hydrogen adsorption isotherms of NaX, NiNaX and RhNaX at 77.4 K up to1 bar.
I N T E R N A T I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5740
Page 8
Author's personal copy
adsorption–desorption isotherms in nickel and rhodium
exchanged NaX with higher cation exchange levels. We used
force field based simulation for the adsorption isotherm
generation in which nonbonded interactions like van der
Waals and coulombic interactions were considered.
The simulated hydrogen adsorption capacities at 77.4 K up
to 1 bar are higher compared to that of the experimentally
measured values in all the three zeolite samples. The unit cell
structures used for simulation studies are perfectly crystal-
line. However, the experimentally used Ni and Rh exchanged
zeolite samples are not always perfectly crystalline. This
could be the main reason for observance of higher hydrogen
adsorption capacity from simulation calculations compared
to experimental results. Probably a completely ion exchanged
nickel or rhodium zeolite X without any loss of crystallinity
could have shown hydrogen adsorption capacities compar-
able to that of simulated data. However, preparation of such a
sample is not practically feasible as zeolite samples lose
crystallinity on cation exchange. Another possible reason for
higher simulated capacities for hydrogen may be the poor
estimation of LJ parameters used for the simulation of
hydrogen adsorption in the Ni and Rh exchanged zeolites
ARTICLE IN PRESS
0 2 50
1
2
3
4
5
6
7
8
9
10
11
12
NaX
NiNaX-49
NiNaX-65
NiNaX-70
NiNaX-73
NiNaX-75
mole
cule
/u.c
Pressure (bar)
0
5
10
15
20
25
30
35
NaX
NiNaX-49
NiNaX-65
NiNaX-70
NiNaX-73
NiNaX-75
mole
cule
/u.c
Pressure (bar)
1 3 4 60 2 51 3 4 6
Fig. 6 – Experimental hydrogen adsorption isotherms of nickel exchanged zeolite X (a) at 303 K up to 5 bar and (b) 333 K up to
5 bar.
0
5
10
15
20
25
30
35
40
45
50
55
mole
cule
/u.c
Pressure (bar)
00
5
10
15
20
25
30
35
40
45
50
55
RhNaX-73RhNaX-69RhNaX-55
RhNaX-44RhNaX-29NaX
RhNaX-73RhNaX-69RhNaX-55
RhNaX-44RhNaX-29NaX
mole
cule
/u.c
Pressure (bar)
1 2 3 4 5 60 1 2 3 4 5 6
Fig. 7 – Experimental hydrogen adsorption isotherms of rhodium exchanged zeolite X (a) at 303 K up to 5 bar and (b) 333 K up
to 5 bar.
I N T E R N AT I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5 741
Page 9
Author's personal copy
that were also taken from literature in order to match the
heat of adsorption of hydrogen in zeolite samples
ð�10 KJ mol�1Þ with that of literature data.
The experimental hydrogen uptake capacities at 303 and
333 K up to 5 bar pressure for various amounts of nickel and
rhodium exchanged zeolite X samples are shown in Figs. 6
and 7 , respectively. The figures show that NaX is having
hydrogen adsorption capacities of 7.5 and 25.3 molecule/u.c at
303 and 333 K, respectively, at 5 bar. The hydrogen adsorption
capacity increases with increase in cation exchange levels up
to 70% for nickel and 55% for rhodium loading inside zeolite
X, and then shows a decreasing tendency. This may be due to
the structural breakage of the zeolite occurring at higher
cation loading levels as evidenced from XRD studies. As the
temperature increases from 303 to 333 K hydrogen adsorption
capacity in NaX increases. In case of nickel exchanged zeolite
X the hydrogen adsorption capacity increases from 303 to
333 K but rhodium exchanged zeolite X shows not much
variation in hydrogen uptake capacity at 303 and 333 K. The
experimental hydrogen uptake values in nickel and rhodium
exchanged zeolite X at 77.4 K and 1 bar and also at 303 and
333 K temperatures and at 5 bar are given in Table 3. The
hydrogen adsorption studies in nickel and rhodium ex-
changed zeolite X have been done by repeated measurements
of the hydrogen isotherms at 303 K. The isotherms are not
completely reversible showing the chemisorption of hydrogen
ARTICLE IN PRESS
Table 3 – Experimental hydrogen uptake values of nickeland rhodium exchanged zeolite X
Zeolite sample Maximum hydrogen uptake values(molecule/u.c)
77 K, 1 bar 303 K, 5 bar 333 K, 5 bar
NaX 114.65 7.52 24.28
NiNaX-49 88.80 8.42 28.72
NiNaX-65 72.07 9.12 30.97
NiNaX-70 59.82 11.31 34.51
NiNaX-73 49.31 9.21 26.96
NiNaX-75 50.68 8.82 25.36
RhNaX-29 79.11 27.23 34.85
RhNaX-44 43.63 42.91 38.66
RhNaX-55 22.19 49.29 51.85
RhNaX-69 16.81 35.61 30.07
RhNaX-73 10.40 29.40 27.07
0 40
5
10
15
20
25
30
35
40
45
50
55
RhNaX-55(I)
RhNaX-55(II)
RhNaX-55(III)
NiNaX-70(I)
NiNaX-70(II)
NiNaX-70(III)
RhNaX-55 desorption
NiNaX-70 desorption
mole
cule
/u.c
Pressure (bar)
1 2 3 5 6
Fig. 8 – Repeated hydrogen adsorption isotherms of NiNaX-70 and RhNaX-55 at 303 K up to 5 bar pressure. Desorption
isotherm of hydrogen during the first run is also shown. The roman numerals in bracket denote the number of runs.
0 2 4 50
10
20
30
40
50
mole
cule
/u.c
Pressure (bar)
RhNaX expt 303K
RhNaX expt 333K
NiNaX expt 303K
NiNaX expt 333K
NiX sim 303K
NiX sim 333K
RhX sim 303K
RhX sim 333K
1 3
Fig. 9 – Simulated and experimental hydrogen adsorption
isotherms of nickel exchanged zeolite X and rhodium
exchanged zeolite X at 303 and 333 K up to 5 bar.
I N T E R N A T I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5742
Page 10
Author's personal copy
in these samples (Fig. 8). Furthermore, a decrease in the
hydrogen uptake was observed in second hydrogen adsorp-
tion run. This decrease could be due to the hydrogen
consumed for initial reduction of nickel/rhodium ion to
metal. The first hydrogen adsorption run was followed by
the activation of the sample under vacuum at 673 K for 6 h.
Thus activated sample was brought to 303 K under vacuum
and the second hydrogen adsorption run was carried out.
Following this, the third adsorption run was carried out after
activating the sample as done in the second run. Hydrogen
adsorption data obtained in the second and third adsorption
runs show comparable values for both nickel and rhodium
exchanged zeolites (Fig. 8). The data in Fig. 8 show that the
hydrogen consumed for reduction of metal ions is limited and
small compared to chemisorbed hydrogen.
Simulation studies of the adsorption of hydrogen in zeolite
NaX, NiX and RhX were also carried out at 303 and 333 K up to
5 bar pressure and the adsorption isotherms are shown in
Fig. 9. The simulation studies at 303 and 333 K for NaX, NiX
and RhX showed negligible hydrogen adsorption capacities as
compared to experimental data which means our potential
energy functions based on nonbonded van der Waals and
coulombic forces fail to explain adsorption behavior in NaX,
nickel exchanged and rhodium exchanged zeolite X at 303
and 333 K. For simulation studies, we calculated the hydrogen
adsorption isotherms of NaX, NiX and RhX at 303 and 333 K
using the same potential functions as used at 77.4 K. The
negligible hydrogen adsorption capacities at 303 and 333 K
predicted by simulation studies compared to the experimen-
tal results show that interactions other than nonbonded
ARTICLE IN PRESS
2200 2000 1800 1600 1400 1200
1690 cm-1 B
A
Tra
nsm
itta
nce (
a.u
.)
wavenumber (cm-1)
2200 2000 1800 1600 1400 1200
Tra
nsm
itta
nce (
a.u
.)
wavenumber (cm-1)
B
A1933 cm-1
Fig. 10 – (a) DRIFT spectra of nickel exchanged zeolite X, A: activated at 673 K and cooled to 303 K under vacuum, B: H2
adsorbed in A at 1 bar and 303 K. (b) DRIFT spectra of rhodium exchanged zeolite X, A: activated at 673 K and cooled to 303 K
under vacuum, B: H2 adsorbed in A at 1 bar and 303 K.
20 40 60 80 100 12035
30
25
20
15
10
5
0
H2 D
eso
rptio
n (
mo
lecu
le/u
.c)
Temperature °C40 60 80 100 120
50
40
30
20
10
0
H2 D
esorp
tion (
mole
cule
/u.c
)
Temperature °C
Fig. 11 – Hydrogen desorption from (a) NiNaX-70 hydrogen absorbed at 333 K and 5 bar and (b) RhNaX-55 hydrogen absorbed
at 333 K and 5 bar in heating up to 393 K.
I N T E R N AT I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5 743
Page 11
Author's personal copy
interactions between hydrogen molecule and zeolite struc-
ture are predominant. Hydrogen molecules are known to
have chemical interactions with nickel and rhodium and the
observed hydrogen sorption at 303 and 333 K is due to the
chemisorption of hydrogen on nickel and rhodium metal
which are generated by the reduction of nickel and rhodium
cations inside the zeolite framework on hydrogen adsorption.
The DRIFT spectra of the hydrogen adsorbed on NiNaX-70
and RhNaX-55 at 303 K are shown in Fig. 10. The M–H
stretching frequencies at 1690 and 1933 cm�1 confirm the
formation of hydride in nickel exchanged zeolite X and
rhodium exchanged zeolite X, respectively. The mechanism
of hydride formation inside the zeolite framework could be as
follows:
Mnþþ nH2$M2Hn þ nHþ. (3)
In accordance with this scheme, the dissociation of hydrogen
results in the formation of H+ ions and metal hydride species.
Indeed, the vibration frequencies of the metal–hydrogen bond
observed in IR and Raman are usually located in the
frequency region 160022300 cm�1. A similar type of mechan-
ism was proposed by Baba et al. [38] for the dissociative
adsorption of hydrogen on silver exchanged zeolites.
The H+ formed during hydride formation would compensate
the total framework charge of the zeolite structure. The
hydrogen adsorbed as hydrides inside the zeolites can be
desorbed fully by heating the zeolites up to 393 K. Fig. 11
shows the desorption of hydrogen from NiNaX-70 and
RhNaX-55.
6. Conclusions
Hydrogen adsorption measurements were carried out at
77.4 K up to 1 bar and 303 and 333 K up to 5 bar in NaX and
nickel and rhodium exchanged zeolite X using volumetric and
gravimetric equipments, respectively. The highest hydrogen
adsorption was obtained for NaX than for nickel and rhodium
exchanged zeolite X at 77.4 K up to 1 bar. Physisorption of
hydrogen was obtained at 77.4 K but chemisorption of
hydrogen takes place at 303 and 333 K forming hydrides with
nickel and rhodium. The chemisorbed hydrogen as hydrides
can be desorbed by heating the zeolites up to 393 K. The grand
canonical Monte Carlo simulation studies also support the
physisorption of hydrogen at 77.4 K and chemisorption of
hydrogen at 303 and 333 K, respectively.
Acknowledgments
This work was supported by the Korea Research Foundation
Grant funded by the Korean Government (MOEHRD, Basic
Research Promotion Fund) (KRF-2006-D00003). The authors
express their thanks to Korea Research Foundation, South
Korea, for financial support and Director’s KIER and CSMCRI
for encouragement.
R E F E R E N C E S
[1] Schlapbach L. Hydrogen as a fuel and its storage for mobilityand transport. MRS Bull 2002;27:675–9.
[2] Schlapbach L, Zuttel A. Hydrogen storage materials formobile applications. Nature 2001;414:353–8.
[3] Zuttel A. Materials for hydrogen storage. Mater Today2003;6:24–33.
[4] Breck DW. Zeolite molecular sieves. New York: Wiley-Inter-science; 1978.
[5] Barrer RM. Zeolites and clay minerals as sorbents andmolecular sieves. London: Academic Press; 1978.
[6] Palomino GT, Carayol MR, Arean CO. Hydrogen adsorption onmagnesium-exchanged zeolites. J Mater Chem2006;16:2884–5.
[7] Du X, Wu E. Physisorption of hydrogen in A, X and ZSM-5types of zeolites at moderately high pressures. Chin J ChemPhys 2006;19:457–62.
[8] Langmi HW, Book D, Walton A, Johnson SR, Al-Mamouri MM,Speight JD, et al. Hydrogen storage in ion exchanged zeolites.J Alloys Compd 2005;404–406:637–42.
[9] Zecchina A, Bordiga S, Vitillo JG, Ricchiardi G, Lamberti C,Spoto G, et al. Liquid hydrogen in protonic chabazite. J AmChem Soc 2005;127:6361–6.
[10] Vitillo JG, Ricchiardi G, Spoto G, Zechina A. Theoreticalmaximal storage of hydrogen in zeolitic frameworks. PhysChem Chem Phys 2005;7:3948–54.
[11] Weitkamp J, Fritz M, Ernst S. Zeolites as media for hydrogenstorage. Int J Hydrogen Energy 1995;20:967–70.
[12] Kayiran SB, Darkim FL. Synthesis and ionic exchanges ofzeolites for gas adsorption. Surf Interface Anal 2003;34:100–4.
[13] Regli L, Zechina A, Vitillo JG, Cocina D, Spoto G, Lamberti C,et al. Hydrogen storage in chabazite zeolite frameworks. PhysChem Chem Phys 2005;7:3197–203.
[14] Kazansky VB, Borovkov VY, Serich A, Karge HG. Lowtemperature hydrogen adsorption on sodium forms offaujasites: barometric measurements and drift spectra.Micropor Mesopor Mater 1998;22:251–9.
[15] Langmi HW, Walton A, Al-Mamouri MM, Johnson SR, Book D,Speight JD, et al. Hydrogen adsorption in zeolites A, X, Y andRHO. J Alloys Compd 2003;356:710–5.
[16] Nishimiya N, Kishi T, Mizushima T, Matsumoto A, TsutsumiK. Hyperstoichiometric hydrogen occlusion by palladiumnanoparticles included in NaY zeolite. J Alloys Compd2001;319:312–21.
[17] Bauer HJ, Wagner FE. Hydride formation, magnetic andtransport properties of nickel and nickel based alloys. Pol JChem 2004;78:463–514.
[18] Wang X, Andrews L. Infrared spectra of rhodium hydrides insolid argon, neon, and deuterium with supporting densityfunctional calculations. J Phys Chem A 2002;106:3706–13.
[19] Razmus DM, Hall CK. Prediction of gas adsorption in 5Azeolites using Monte Carlo simulation. AIChE J1991;37:769–79.
[20] Watanabe K, Austin N, Stapleton MR. Investigation of the airseparation properties of zeolite types A, X and Y by MonteCarlo simulations. Mol Simul 1995;15:197–221.
[21] Song MK, No KT. Molecular simulation of hydrogen adsorp-tion in organic zeolite. Catal Today 2007;120:374–82.
[22] Berg AWC, Bromley ST, Wojdel JC, Jancen JC. Adsorptionisotherms of H2 in microporous materials with the SODstructure: a grand canonical Monte Carlo study. MicroporMesopor Mater 2006;87:235–42.
[23] Fuchs AH, Cheetham AK. Adsorption of guest molecules inzeolitic materials: computational aspects. J Phys Chem B2001;105:7375–82.
ARTICLE IN PRESS
I N T E R N A T I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5744
Page 12
Author's personal copy
[24] Zuttel A, Sudan P, Mauron P, Wenger P. Model for hydrogenadsorption on carbon nanostructures. Appl Phys A2004;78:941–6.
[25] Zhu L, Seff K. Reinvestigation of the crystal structure ofdehydrated sodium zeolite X. J Phys Chem B 1999;103:9512–8.
[26] Bae D, Seff K. Crystal structure of zeolite X nickel (II)exchanged at pH 4.3 and partially dehydrated, Ni2(NiOH)35
(Ni4AlO4)2(H3O)46Si101Al91O384. Micropor Mesopor Mater2004;40:219–32.
[27] Busch F, Jaeger N, Ekloff GS. Growth of rhodium dispersionsin zeolite NaX: identification of a precursor state by Rietveldrefinement. Zeolites 1996;17:244–9.
[28] Ramsahye NA, Bell RG. Cation mobility and the sorption ofchloroform in zeolite NaY: molecular dynamics study. J PhysChem B 2005;109:4738–47.
[29] Cerius2 User Guide: Forcefield-based simulations. MolecularSimulations Inc, San Diego, April 1997.
[30] Allen MP, Tildesley DJ. Computer simulation of liquids.Oxford, UK: Clarendon; 1987.
[31] Weinberger B, Darkrim FL, Kariyan SB, Gicquel A, Levesque D.Molecular modeling of H2 purification on Na–LSX zeolite andexperimental validation. AIChE J 2005;51:142–8.
[32] Murad S, Gubbins KE. Molecular dynamics simulation ofmethane using a singularity free algorithm. In: Lykos P,editor. Computer modeling of matter ACS symposium series.Washington: American Chemical Society; 1978. p. 62–71.
[33] Maurin G, Llewellyn P, Poyet T, Kuchta B. Influence ofextraframe work cations on the adsorption of X-faujasitesystems: micro calorimetry and molecular simulations.J Phys Chem B 2005;109:125–9.
[34] Kramer GJ, Farragher NP, van Beeset BWH, van Santen RA.Interatomic forcefields for silicas, aluminophosphates, andzeolites derivation based on ab initio calculations. Phys Rev B1991;43:5068–80.
[35] Rappe AK, Casewit CJ, Colwell KS, Goddard III WA, Skiff WM.UFF, a full periodic table forcefield for molecular mechanicsand molecular dynamics simulations. J Am Chem Soc1992;114:10024–35.
[36] Cerius2, v.4.2. Accelrys, Inc., San Diego, CA; 1999.[37] Olson DH. Crystal structure of the zeolite nickel faujasite.
J Phys Chem 1968;72:4366–73.[38] Baba T, Komatsu N, Sawada H, Yamaguchi Y, Takahashi T,
Sugisawa H, et al. 1H magic angle spinning NMR evidence fordissociative adsorption of hydrogen on Ag+-exchanged A andY- zeolites. Langmuir 1999;15:7894–6.
ARTICLE IN PRESS
I N T E R N AT I O N A L J O U R N A L O F H Y D R O G E N E N E R G Y 3 3 ( 2 0 0 8 ) 7 3 5 – 7 4 5 745