Journal of Materials Processing Technology 162–163 (2005) 169–177

An overview of advanced materials for hydrogen storage

Elena David
National Institute of Cryogenics and Isotopic Technologies, Rm. Valcea, O.P.4, P.O. Box 10, Code 240050, Romania

Abstract

In a future sustainable energy system based on renewable energy, environmentally harmless energy carriers like hydrogen will be of crucial
importance. One of the major impediments for the transition to a hydrogen-based energy system is the lack of satisfactory hydrogen storage
alternatives. In the last years, the possible to store hydrogen in various materials was extensively studied.
This paper is a preliminary study with the focus on advanced nanostructured materials such as solids of large surface area based on carbon
structures, metals and different types of metal alloys, other intermetallic compounds, etc. as possibilities for hydrogen storage. The newest
materials used for hydrogen storage are light metal alloys. We have so far focused in this review almost exclusively on experimental studies.
Also there are presented the most important characteristics of these materials such as mechanical strength, porosity and affinity to hydrogen,
and also the recent developments in the search for innovative materials with high hydrogen-storage capacity and our contribution in this field.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Carbon structures; Metal alloys; Light metal hydrides; Hydrogen storage

1. Introduction
The current interest in hydrogen is primarily due to envi-
ronmental concerns of the harmful emissions from the fossil
fuels used presently. Also, a demand for more efficient power
sources has increased the interest in different kinds of new
technologies, such as fuel cells using hydrogen or hydrocar-
bons as fuel [1,2]. It is widely believed that hydrogen will
within a few years become the fuel that powers most vehicles
and portable devices, i.e. hydrogen will become the means of
storing and transporting energy. The reason is the depletion of
oil and the relatively facile production of hydrogen from the
various renewable sources of energy – hydroelectric, wind,
solar, geothermal – with water being the only raw material
needed. To release the energy, hydrogen can be burned in an
efficient and clean way in a fuel cell to form water again, or
made to drive an electrochemical cell as in the commonly
used nickel hydride battery.
While hydrogen has many obvious advantages, there re-
mains a problem with storage and transportation. Pressurised
hydrogen gas takes a great deal of volume compared with,
for example, gasoline with equal energy content, about 30
E-mail address: david@icsi.ro.
times bigger volume at 100 bar gas pressure. Condensed hy-
drogen is about 10 times denser, but is much too expensive
to produce and maintain [3]. There are also obvious safety
concerns with the use of pressurised or liquefied hydrogen in
vehicles. The results obtained and presented by many stud-
ies show that three kinds of materials are competitive for to
be used in hydrogen storage processes. These are materials
based on carbon structures, metals and metal alloys. We will
refer especially at these kinds of materials.
2. Materials based on carbon structures
The best performance in hydrogen storage was achieved
with materials based on carbon structures of highest effective
porosity. The two forms of carbon that is the most known to
us are diamond and graphite. In diamond each atom is fully
coordinated symmetrically in space in all three dimensions.
Graphite, on the other hand, is build up of a two-dimensional
hexagonal sheet of carbon atoms, with long distance be-
tween each sheet. However, there are also other forms of car-
bon structures such as fullerenes and nanotubes, that are the
newest advanced carbon structures, with special properties.
Fig. 1 shows the structure of graphite, diamond, fullerenes
and of single-wall carbon nanotube.

0924-0136/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmatprotec.2005.02.027
170 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177
Fig. 1. Structure of (a) graphite, (b) diamond, (c) fullerenes and (d) single-
wall carbon nanotube.
Fullerenes are single or multiple layers of graphite
wrapped together into very stable ball (Fig. 1c) or tube mole-
cules – single-wall carbon nanotube (SWCT), Fig. 1d, re-
spectively.
2.1. Fullerenes
Fullerenes are a new class of carbon aromatic compounds
with unusual structural, chemical and physical properties
which, in turn, will lead to novel and unexpected applica-
tions. Fullerenes are synthesised carbon molecules usually
shaped like a football, such as C60 and C70 and they are able
to hydrogenate through the reaction [4]:
C60 +xH2 O + xe− ↔ C60 Hx + xOH− (1)
According to theoretical calculations, the most stable of these
are C60 H24 , C60 H36 and C60 H48 , latter of which is equal to
6.3 wt.% of hydrogen adsorbed [4,5]. An experimental study
made by Chen et al. shows that more than 6 wt.% of hydrogen
can be adsorbed on fullerenes at 180 ◦ C and at about 25 bar
[6]. Usually the bonds between C and H atoms are so strong
that temperatures over 400 ◦ C are needed to desorb the hydro-
gen [5], but Chen et al. were able to do this at a temperature
below 225 ◦ C [6]. Despite the quite high hydrogen-storing
ability, the cyclic tests of fullerenes have shown poor prop-
erties of storing hydrogen [7].
In our laboratory, it was obtained a mixture of C60 and C70
fullerenes by electric arc discharge method in which carbon
is vaporised between two graphite electrodes. In Fig. 2 is
presented the image of the C70 fullerene surface obtained by
transmission electron microscopy (TEM) and SEM image of
single-wall carbon nanotubes (SWCTs).
Our experimental study concerning hydrogen storage
showed that more than 0.6 wt.% of hydrogen can be adsorbed
on fullerenes at 200 ◦ C and at about 12 bar.
2.2. Carbon nanotubes
Nanotubes are the strongest carbon fibres that are currently
known, single-wall nanotube (SWNT; see Fig. 1c) and multi-
Fig. 2. (a)TEM image of C70 fullerene. Each white feature corresponds to
a column of C70 molecules lying parallel to the electron beam. The arrow
indicates that defects occur in clusters separating large perfect crystallites.
These defects are the hosts for hydrogen storage. (b) SEM image of single-
wall carbon nanotubes (SWCTs).
wall nanotubes (MWNT). A single-wall nanotube can be up
to 100 times stronger than that of steel with the same weight.
The Young’s modulus of SWNT is up to 1 TPa, which is 5
times greater than steel (230 GPa) while the density is only
1.3 g/cm3 [8]. That means that materials made of nanotubes
are lighter and more durable. Another property of carbon
nanotubes is their ability to quickly adsorb high densities of
hydrogen at room temperature and atmospheric pressure [9].
An SEM image of SWCTs is shown in Fig. 2b.
Nanotubes are manufacturing from pure carbon. Pure car-
bons only have two covalent bonds sp2 and sp3 , the former
constitutes graphite and the latter constitutes diamond. sp2
is a strong bond within a plane but weak between planes.
sp2 is composed of one s-orbital and two p-orbitals. When
more sp2 bonds together, they form six-fold structures, like
honey comb pattern, which is a plane structure, the same
structure as graphite. Graphite is stacked layer by layer so
it is only stable for one single sheet. Viewing these layers
perpendicularly shows the honey comb patterns of graphite.
Wrapping these patterns back on top of themselves, joining
the edges and close one end while leave one end open, it
forms a tube of graphite, which is known as nanotube [10].
In Fig. 3 it is presented schematically the nanotube form-
ing. Carbon nanotubes were discovered in 1991 accidentally
when synthesising fullerenes [11].
Hydrogen can be stored into the nanotubes by chemisorp-
tion or physisorption. The methods of trapping hydrogen are
not known very accurately [11,12] but density functional cal-
culations have shown some insights into the mechanisms
Fig. 3. The forming mechanism of nanotube.
 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177 171

The density functional calculations have shown that theo-

retically in proper conditions a single-wall nanotube can ad-
sorb over 14 wt.% [11,12] and a multi-wall nanotube about
7.7 wt.% of hydrogen [12]. Dillon and co-workers [14] re-
ported the first experimental result of high hydrogen uptake
by a nanotube. They estimated that hydrogen could achieve
a density of 5–10 wt.%. Chen et al. [15] reported that alkali-
doped nanotubes are able to store even 20 wt.% under ambi-
ent pressure, but are unstable or require elevated temperatures
[15,16]. The result has shown to be in a great disagreement
with other results and has been thought to be incorrect [17].
Recent results on hydrogen uptake of single-wall nanotubes
are promising. At 0.67 bar and 327 ◦ C, about 7 wt.% of hy-
drogen have been adsorbed and desorbed with a good cycling
stability [18].
Another result at ambient temperature and pressure shows
that 3.3 wt.% hydrogen can be adsorbed and desorped repro-
ducibly and 4.2 wt.% hydrogen with a slight heating [19]. The
hydrogen amounts storaged depend of the number of layers
Nl and of tube diameter.
The effect of nanotube type (SWNT or MWNT) on the
storage capacity for hydrogen has been investigated theoreti-
cally [20]. The calculated storage hydrogen mass as a function
of number of layers and tube diameter is presented in Fig. 5.
Hydrogen condenses as a monolayer at the surface of the
nanotube or condenses in the cavity of the tubes. The con-
densed hydrogen has the same density as liquid hydrogen at
−253 ◦ C.
The arc discharge method produces a good-quality multi-
wall and single-wall carbon nanotubes.
The studies performed by us within a project the system
used for to obtain carbon structures of fullerenes was used
for to produce carbon nanotubes. For this it was utilised two
graphite electrodes to generate an arc by a high d.c. cur-
rent. After arc discharging for a period of time, a carbon rod
builds up at the cathode. Carbon nanotube bundles and amor-

Fig. 4. Hydrogen adsorption in a nanotube: (a) exterior adsorption with H/C

coverage 1.0, (b) interior adsorption with coverage 1.0, (c) interior adsorption
with coverage 1.2 and (d) interior adsorption with coverage 2.4 [12].

[11–13]. Calculations indicate that hydrogen can be adsorbed

at the exterior of the tube wall by H C bonds with a H/C cov-
erage 1.0 or inside the tube by H H bonds with a coverage
up to 2.4 [11,12] as shown in Fig. 4 modified from [12]. The
adsorption into the interior wall of the tube is also possible
but not stable. The hydrogen relaxes inside the tube form-
ing H H bonds [11]. The numbers in the figure tell the bond
lengths in 10−10 m. Multi-wall nanotubes, in which two or
more single tubes are rounded up each other with van der
Waal’s attraction, can adsorb hydrogen between the single-
wall nanotubes. The hydrogen causes the radius of the tubes
to increase and thus makes a multi-wall nanotube less stable
[12]. In nanotube bundles hydrogen can also be adsorbed in Fig. 5. Calculated storage hydrogen mass% as a function of the number of
the middle of different tubes. layers Nl and as a function of the tube diameter.
172 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177
Fig. 6. Pore structure of activated carbon – granular form.
phous carbon both form at cross-section of the rod. Helium
gas was present to increase the speed of carbon deposition.
This method does not produce clean results due to the ex-
istence of these amorphous carbons. Some parameters that
are critical in this process are: the pressure of the helium, the
temperature and the d.c. current. Efficient cooling is neces-
sary to form homogenous deposition of carbon nanotubes.
Our studies will try to perfect the arc discharge method by
using graphite electrodes that contain metal catalysts.
Even though the price of the nanotubes is still high, they
have a good potential in hydrogen storing and will remain
as materials that are competitive for to be used in hydrogen-
storage processes.
2.3. Activated carbons
Bulky carbon with high surface area, the so-called acti-
vated carbon, is other carbon structure able to adsorb hy-
drogen in its microscopic pores. The main problems are that
only some of the pores are small enough to catch the hydro-
gen atom and that high pressures must be applied in order
to get the hydrogen into the pore. The pores in activated car-
bons are scattered over a wide range of size and shape. The
pores are classified by their sizes usually into three groups:
(i) macropores having average diameter more than 50 nm,
(ii) mesopores with diameter 2–50 nm and (iii) micropores
having average diameter less than 2 nm. These are further
divided into supermicropores (0.7–2.0 nm) and ultramicrop-
ores of diameter less than 0.7 nm. The pore structure in wood-
derived carbons is a memory of the texture of the precursor
wood and hence differ based on the type of wood/plant (see
Figs. 6 and 7). Basically this consists of slit-like voids and
macropores. In some wood carbon, these are of similar shape
and size and are equally distributed, whereas in coal they
are of very different shapes and sizes and are organised in
circular fashion. These macropores are in turn connected to
Table 1
Texture analysis and H2 storage capacities at (−196 ◦ C) and 12 bar
Nr. crt. Material Pore volume (cm3 /g) SBET (m2 /g)
1 Graphite 0.04 7 2.26
2 ACSPh 0.85 1620
3 ACWPh 1.49 1850
Fig. 7. Pore structure of activated carbon – fibrous form.
mesopores and micropores. This is shown in Fig. 6. Fig. 7 in-
cludes the structure of activated carbon fibres which, in con-
trast to wood-based activated carbon, predominantly consists
of micropores directly exposed to the surface of the fibres.
About 5.2 wt.% of hydrogen adsorbed into the activated
carbon has been achieved at cryogenic temperatures and in
pressures of about 45–60 bar [20]. In ambient temperature
and pressure of 60 bar, the storage capacity has been only
approximately 0.5 wt.% [20,25].
In our recent studies, we developed chemically activated
carbons (ACWPh, ACSPh) [23]. These were produced by
mixing an chemical agent (phosphoric acid) with a young
carbonaceous material obtaining from walnut shells (CW)
and sawdust (CS) and carbonising the resultant mixture at
550 ◦ C under inert atmosphere.
The resulting very porous carbon structure is filled with
activation agent, which removed from carbon by washing.
The characteristics of these materials are presented in Table 1.
Such as can see from Table 1, activated carbon (ACWPh) has
a pore volume of 1.49 cm3 /g, which is 37 times greater than
graphite.
About 2.2–2.8 wt.% H2 of hydrogen adsorbed into these
activated carbons has been achieved at cryogenic tempera-
tures and in pressures of about 12–15 bar. For microporous
carbons the details of the pores (size and shape) apparently
affect the specific H2 uptake to a large extent.
In addition to general-purpose active carbons, advanced
active carbons with specific control on pore structure have
been developed over the past few decades for specific ap-
plications. Research and development efforts are continuing
for more and more efficient applications of these materials.
Carbon molecular sieves (CMS) are a special class of ac-
tive carbons having small pore sizes with a sharp distribution
in a range of micropores, as compared with other activated
carbons.
The pore size in CMSs is comparable to the size of adsor-
bate molecules, such as nitrogen and hydrogen [24].
Density (cm3 /g) Hardness (GN/m2 ) Median pore H2 storage
diameter (Å) capacity (wt.%)
30.15 – 0
0.47 16.06 3–300 2.2
0.43 14.78 3–200 2.8
 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177
In addition, the temperature of adsorption governs the rate
of adsorption of a gas because of the activated diffusion of
adsorption molecules in micropores.
3. Metals and alloys for hydrogen storage
Hydrogen is a highly reactive element and is known to
form hydrides and solid solutions with thousands of metals
and alloys. Most of the natural elements adsorb hydrogen un-
der proper conditions. Metal hydrides are composed of metal
atoms that constitute of a host lattice and hydrogen atoms
that are trapped in interstitial sites, such as lattice defects.
The trap site can be a vacancy or a line defect. In the case of a
line defect, a string of hydrogen atoms may accumulate along
the defect. Such a string increases the lattice stress, especially
if two adjacent atoms recombine to form molecular hydro-
gen [25]. Since adsorption of hydrogen increases the size of
lattices [25,26], the metal is usually ground to a powder in or-
der to prevent the decrepitation of metal particles. There are
two possible ways of hydriding a metal, direct dissociative
chemisorption and electrochemical splitting of water. These
reactions are:
M + X/2 · H2 ↔ MHX
M + X/2 · H2 O + X/2 · e− ↔ MHx + X/2·OH−
where M represents the metal. In electrochemical splitting
there has to be a catalyst, such as palladium, to break down
the water.
The necessary condition for hydrogen storage is that the
thermodynamic and kinetic conditions are fulfilled. In that
case, a metal exposed to hydrogen gas absorbs hydrogen
until equilibrium is established. There are a number of re-
action steps that kinetically may hinder a hydrogen-storing
system to reach its thermodynamical equilibrium of hydro-
gen storage within a reasonable time. The reaction rate of a
metal–hydrogen system is therefore a function of pressure
and temperature.
Metal and hydrogen usually form two different kinds of
hydrides, ␣-phase and ␤-phase hydride. In ␣-phase there is
only some hydrogen adsorbed and in ␤-phase the hydride is
fully formed. A schematic of phase transition is presented
in Fig. 8. When charging, hydrogen diffuses from the sur-
face of the particle trough the ␤-phase to the phase-transition
interface and forms additional ␤-phase hydride. When dis-
charging, hydrogen from the phase-transition interface dif-
fuses through the ␣-phase to the surface of the particle where
it is recombined into the form of molecular hydrogen [27].
Hydrogen storage in metal hydrides is a complex process
consisting of a lot of mechanistic steps and depends on impor-
tant parameters. A metal surface has to be able to dissociate
the hydrogen molecule and to allow hydrogen atoms to move
easily in order to be able to store hydrogen. Metals differ in
the ability to dissociate hydrogen, this ability being depen-
dent on surface structure, morphology and purity. Normally
173
Fig. 8. Schematic of phase transition in metal hydrides.
the hydrides are divided into high-temperature hydrides and
low-temperature hydrides depending on the temperature of
absorption/desorption process. In the low-temperature hy-
drides, the hydrogen normally is bound through covalent
bonding and the metal hydride consists of high-molecular-
weight material. In the high-temperature hydrides, the hydro-
(2) gen is normally ionic bound, and the metal hydride consists
of low-molecular-weight material. The hydrogen-storage ca-
(3)
pacities are higher for the high-temperature hydrides. The
most common examples of hydrogen-storing alloys are Fe–Ti
hydrides, La–Ni hydrides, Ti–Zr–V series of hydrides, etc.
Metals can absorb hydrogen in atomic form and thereby
act as hydrogen “sponges” [28]. Around 50 metallic elements
of the periodic table can absorb hydrogen in great quantity
and the possible choices of hydrogen storage materials are,
therefore, enormous.
Many scientific and engineering studies have been carried
out of the absorption/desorption of hydrogen in metals and
development of such storage devices. Daimler-Benz, for ex-
ample, produced in the early 1980s a car fuelled by hydrogen
where the storage tank was a chunk of Fe–Ti metal alloy [29].
The volume of this storage device is less than a factor of 2
greater than the equivalent gasoline tank, but the problem is
that the hydride is 20 times heavier [30].
The only successful commercial large-scale application of
metal hydrides as hydrogen storage so far is the metal hydride
battery, which has supplied battery power to many small elec-
trical appliances such as mobile phones and portable comput-
ers. Hydride systems based on existing metal hydrides cannot
store large amounts of hydrogen and the development of new
kinds of hydride materials is required.
3.1. Light metal alloys for hydrogen storage
If metal hydrides are to become important energy carri-
ers in mobile vehicles, the mass of the system needs to be
reduced from today’s devices, such as FeTi. This puts strong
constraints on the chemical elements which can be used. A
174 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177

very promising approach is to use magnesium. The hydride,

MgH2 , can store up to 7.6 wt.% of hydrogen [28,29].
The automobile industry has set 5–5.6 wt.% as a target for
efficient hydrogen storage. But there are three problems with
pure Mg: (i) The rate at which hydrogen absorbs and desorbs
is too low because diffusion of hydrogen atoms through the
hydride is slow. Problem (ii) is that the hydrogen molecules
do not readily dissociate at the surface of Mg to generate the
hydrogen atoms that diffuse into the metal. Transition metals
can catalyse this bond-breaking/formation even at the sur-
face, but not main group elements [30]. Problem (iii) is that
hydrogen atoms bind too strongly with the Mg atoms, i.e. the
enthalpy of formation of the hydride is too large, so that the
hydride needs to be heated to very high temperature, around
350 ◦ C, in order to release hydrogen gas at high enough pres-
sure (over 1 atm) [28].
A useful metal hydride should release the hydrogen in the
temperature range between 50 and 100 ◦ C. It is important to
distinguish between problems (i) and (ii) although they both
have to do with the kinetics of absorption and desorption
processes. These problems could be solved efficiently by dif-
ferent means, but problem (iii) remains unsolved. Our studies
performed till now were focused on these three problems.
The first problem could be reduced by forming a compos-
ite of small Mg crystals agglomerated together as opposed
Fig. 9. Transmission electron microscopy (TEM) micrograph of the powder
to a large chunk of Mg crystal. The clusters formed by ball showing the nanocrystalline structure of the particles (a), and a scanning
milling are micrometer sized such as seen in the scanning electron microscopy (SEM) micrograph of a powder of Mg particles which
electron microscopy (SEM) and the transmission electron mi- have small Ni clusters attached on the surface in order to catalyse the break
crograph (TEM) images (Fig. 9). The size of the magnesium up of hydrogen molecules during adsorption (b).
grains in this composite is 40–200 nm. Hydrogen absorption
and desorption of this material was found to be reasonably
fast. ysisorption). If the kinetic energy is large enough for the H H
The dark regions represent crystalline domains, while the bond to stretch slightly (one activation barrier), its molecular
lighter regions are less dense and likely amorphous. The scan- bond can shift partly towards the surface and become more
ning electron microscopy (SEM), shown in Fig. 9b, reveals tightly bound to the surface (chemisorption). If the kinetic
the crystalline structure of the particles. Another way to im- energy is so high that the H H bond can be split (another acti-
proving the kinetics is to mix in another chemical component, vation barrier) and turned completely towards the surface, the
for example, by mixing in Ni. Channels with lower binding hydrogen atoms can be dissociated and atomic chemisorbed
energy and smaller migration barriers for the hydrogen are to the surface, but only if there are two free sites at the surface
apparently formed and maintained through out the hydride, such is shown in Fig. 10. Using Mg as example, the activation
thus enabling fast loading and unloading. This may occur barrier for dissociation is approximately 0.5 eV [31].
because holes where the hydrogen atom sites are now sur- Furthermore, the transfer of the dissociated hydrogen atom
rounded by two types of atoms forming bonds of different into the interior of the solid structure includes another barrier
strength with the hydrogen atom. The alloy hydride can also of approximately 0.4 eV and inside the metal the hydrogen
have different diffusivity because the alloy has a different lat- atom can move by diffusion at yet another activation bar-
tice structure than pure magnesium. The problem is that this rier of approximately 0.1 eV. The kinetic energy depends on
metal, Ni, is heavy and greatly reduces the weight percent of temperature and the activation barriers therefore set a lower
hydrogen in the hydride. The second problem, the dissocia- temperature limit below which reactions are too slow to be
tion of hydrogen molecules on the surface, could be solved useful. The depth of the potential well determines the stabil-
by adding a small amount of a catalytically active metal on ity of the storage, i.e. the temperature range of operation and
the surface of the Mg clusters, as shown in Fig. 9b. the possibility to maintain storage at lower pressure.
When hydrogen is bonded chemically in a solid structure, The formation of a hydride can be illustrated by a p–c–T
such as a metal hydride, it has to overcome certain activation diagram, where the hydrogen pressure (p) is plotted as a func-
barriers during the process. Near the surface it can adsorb tion of atomic ratio of hydrogen (c) at a given temperature (T),
in a weak van de Waal’s interaction with the surface elec- such as presented in Fig. 11. The p–c–T diagram illustrates
trons through a small polarisation of its electronic charge (ph- the formation of solid solution (␣-phase) at low H-pressures
 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177
Fig. 10. Molecular processes involved when storing hydrogen.
and atomic ratio [32], a pressure plateau where there is a
continuing transformation from the solid solution ␣-phase to
the metal hydride ␤-phase at constant pressure [33] and the
growth of the ␤-phase with increasing hydrogen pressures
and atomic ratio. The p–c–T diagram may show hysteresis
when coming from high to low hydrogen content (decom-
position) as compared to low to high hydrogen content (for-
mation). This occurs if, for example, strain that has been
accumulated in the sample is relaxed by plastic deformation.
In Fig. 11 the first plateau corresponds to the absorption of
hydrogen into pure magnesium and the second plateau de-
notes the absorption of hydrogen into the Mg2 Ni alloy. The
absorption/desorption on hydrogen into this alloy is very fast
and it is clear that microstructure plays an important role in
the performance of the material. To ensure that large volumes
of hydrogen can be stored it is essential to use small granules
of the base material to make a large surface area available.
When the hydride is formed in alloy, it has been seen that
a surface passivation layer may be formed, consisting of a
Fig. 11. Characteristics of p–c–T diagram for hydrogen absorp-
tion/desorption processes on MgNi alloy.
175
combination of the metal oxides, hydroxides, carbon–oxygen
compounds and water. This surface passivation layer acts as
a diffusion barrier and decreases the numbers of dissociation
sites for H2 , and thus reduce the hydrogen uptake rate. Poi-
soning effects can also occur, where gaseous impurities in the
reacting atmosphere are inhibiting the hydrogen dissociation.
Very strong (SO2 , H2 S and CH3 SH) and strong gaseous poi-
sons (O2 , CO2 and CO) has to be avoided, in contrast to CH4 ,
C2 H4 and N2 that have no measurable effect on the hydrogen
absorption kinetics.
The third problem with pure Mg as a hydrogen-storage
device is the most serious one, thanks to the high enthalpy of
formation of the hydride. Some improvement has been shown
to occur by adding Ni to Mg. The desorption temperature of
the Mg2 Ni alloy is about 50–60 ◦ C lower than for pure Mg,
but this is still far from low enough. The addition of the heavy
transition metal greatly reduces the storage capacity, down to
3.8% from 7.6% in pure Mg. A better solution would be to add
a light metal that has low affinity to hydrogen, for example,
aluminium or another light main group element that competes
with hydrogen for the valence electrons of Mg. This strategy
has apparently not been tried and will be pursued in a new
proposed project.
3.2. Alanates and other hydrides
Some of the lightest elements in the periodic table, for
example, lithium, boron, sodium and aluminium, form sta-
ble and ionic compounds with hydrogen [34]. The hydrogen
content reaches values of up to 18 mass% for LiBH4 .
However, such compounds desorb the hydrogen only at
temperatures from 80 to 600 ◦ C, and the reversibility of the
reaction is not yet clear for all systems. Bogdanovic and
Schwickardi [35] showed in 1996 that the decomposition
temperature of NaAlH4 can be lowered by doping the hy-
dride with TiO2 .
The same group showed the reversibility of the reaction for
several desorption/absorption cycles. This is a good example
of the potential of such hydrides, which were discovered more
than 50 years ago.
Several points have to be clarified. First is the high des-
orption temperature due to the poor kinetics of the system
or due to the thermodynamic stability of the compound. The
kinetics can be improved by applying an appropriate cata-
lyst to the system and apparently also by ball milling and the
introduction of defects.
Second, the conditions for a reversible reaction, for exam-
ple, formation and stabilisation of clusters of an intermetallic
compound of the remaining metals on desorption, and third
is the desorption reaction and the intermediate reaction prod-
ucts (decomposition in several steps).
These questions can be answered by the role of light hy-
drides in an on-board fuel cell. The exhaust gas of a fuel cell
is water vapour, which could be collected and reused for on
board hydrogen production. The common experiment [34] –
shown in many chemistry classes – where a small piece of
176 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177
sodium floating on water produces hydrogen, demonstrates
such a process. Sodium is transformed to sodium hydroxide
in this reaction. The reaction is not reversible, but sodium hy-
droxide could later be removed and reduced in a solar furnace
back to sodium. Each sodium atom produces one hydrogen
atom, so the corresponding gravimetric hydrogen density of
the sodium reaction is slightly more than 4 mass%.
Lithium used in the same way would deliver up to
14 mass% of hydrogen. The alkali metals as a hydrogen
source are easy to handle, and a car could be refilled within
a few minutes. To deliver the necessary 4 kg of hydrogen
using the water produced in the fuel cell would take 28 kg
of lithium. After using up all the hydrogen the tank would
contain 99 kg of lithium hydroxide, ready to be recycled
[35].
Besides the alloys described above, there are several other
families of intermetallics having a capability of hydrogen
absorption, none of which has attained commercial interest.
These include, for example, A2 B, AB3 , A2 B7 , A3 B, etc. The
A elements are usually Ti, Zr, Hf, Th or a lanthanide (atomic
number 57–71). The B elements can be a variety of transition
and non-transition metals such as V, Cr, Mn and Fe. Some of
these have good hydrogen-storage capacities but do not have
favorable absorption/desorption characteristics.
One promising, but quite expensive group is alloys based
on vanadium. Especially, a combination of V–Ti–Fe is an
attractive alternative. For example, (V0.9 Ti0.1 )0.95 Fe0.05 has
a maximum hydrogen-storage capacity of 3.7 wt.% and a
reversible hydrogen-storage capacity of 1.8 wt.% and still
with good absorption properties [36], but the system is too
heavy.
4. Conclusions
Research is focused on solid-state hydrogen storage using
gas on solid adsorption in materials such as high surface area
carbons, or absorption in the interstices of a metal. In solid-
state storage, gas on solids and metal hydrides are the options
which are safer technologies and they provide high storage
capacity than physical storage systems. There is research go-
ing on determining the hydrogen adsorption/desorption prop-
erties of available materials based on carbon structures and
absorption/desorption on metals and metal alloys. Basic con-
cept on mechanisms of solid-state storage is essential in or-
der to get more knowledge about high-performance storage
materials such as storage capacities, rates of charge and dis-
charge, thermal and mechanical effects of available materi-
als, increasing capability of manufacturing of new storage
materials, like metal alloys and high surface area activated
carbons. Materials systems, such as fullerenes, carbon nan-
otubes, light metal alloys, are very promising new materials
they have a good potential in hydrogen storage. Although the
lightweight Mg-class hydrides show high hydrogen-storage
capacities even after extensive cycling they cannot for the
present be regarded as a solution for vehicular application
mainly because of the slow kinetics in the lower temperature
range.
There is reason for hope that one day much better
hydrogen-storage materials will be discovered and devel-
oped, rather as we have seen a revolution in high-temperature
superconducting materials or hard permanent magnets. We
must bear in mind that to develop a sustainable future en-
ergy policy requires us to focus not only on the scien-
tific and technical challenge, but also on vital adaptations
by the socioeconomic system and a change in attitudes to
energy.
References
[1] G. Sandrock, K.J. Gross, G. Thomas, J. Alloys Compd. 339 (2002)
299.
[2] B. Bogdanovic, R.A. Brand, A. Marjanovic, M. Schwickardi, J.
Tolle, J. Alloys Compd. 302 (2000) 36.
[3] C.M. Jensen, R. Zidan, N. Mariels, A. Hee, C. Hagen, Int. J. Hy-
drogen Energy 24 (1999) 461.
[4] M. Beltowska-Brzezinska, A. Czerwinski, J. Kleperis, M. Kopczyk,
J. Skowronski, G. Wójcik, J. Solid State Electrochem. 5 (2001)
229–249.
[5] G. Sandrock, A. Panoramic, J. Alloys Compd. 293–295 (1999)
877–888.
[6] F. Chen, R. Li, W. Loutfy, R. Murphy, J. Wang, Proceedings of the
1998 U.S. DOE Hydrogen Program Review, USA, 1998.
[7] C. Ahn, B. Fultz, J. Vajo, Y. Ye, J. Zinck, Appl. Phys. Lett. 77 (14)
(2000) 2171–2173.
[8] Saito, et al., Appl. Phys. Lett. 60 (18) (1992) 4.
[9] M.S. Fuhrer, IEEE Trans. Nanotechnol. 1 (2002) 78.
[10] J. Kong, et al., Science 287 (2000) 622.
[11] S. Lee, Y. Lee, Appl. Phys. Lett. 76 (20) (2000) 2877–2899.
[12] D. Bae, J. Bok, Y. Choi, Y. Choi, T. Frauenheim, N. Kim, S. Lee,
Y. Lee, K. Nahm, K. Park, Y. Park, S. Yu, Syn. Met. 113 (2000)
209–216.
[13] S. An, T. Frauenheim, W. Kim, S. Lee, Y. Lee, Y. Park, G. Seifert,
Syn. Met. 121 (2001) 1189–1190.
[14] T. Bekkedahl, D. Bethune, A. Dillon, M. Heben, K. Jones, C. Kiang,
Nature 386 (1997) 377–379.
[15] P. Chen, J. Lin, K. Tan, X. Wu, Science 285 (1999).
[16] H. Cheng, C. Liu, Q. Yang, Carbon 39 (2001) 1447–1454.
[17] R. Yang, Carbon 38 (2000) 623–626.
[18] J. Alleman, A. Dillon, T. Gennett, K. Jones, P. Parilla, Proceedings
of the 2000 DOE/NREL Hydrogen Program Review, USA, 2000.
[19] H. Cheng, H. Cong, Y. Fan, C. Liu, M. Liu, Science 286 (1999).
[20] G. Stan, W.M. Cole, J. Low Temp. Phys. 110 (1999) 539–544.
[23] E. David, V. Stanciu, A. Armeanu, A. Lazar, C. Sandru, Acta Tech-
nica Napocensis 45 (2002) 197–202.
[24] E. David, A. Talaie, V. Stanciu, A.C. Nicolae, J. Mater. Process.
Technol., in press.
[25] T. Carter, L. Cornish, Eng. Failure Anal. 8 (2001) 113–121.
[26] P. Fischer, K. Yvon, in: Schlapbach (Ed.), Hydrogen in Intermetallic
Compounds I, Springer-Verlag, Germany, 1988, p. 75.
[27] K. Chuang, N. Cui, J. Luo, J. Electroanal. Chem. 503 (2001) 92–98.
[28] R. Griessen, T. Riesterer, Hydrogen in intermetallic compounds I, in:
L. Schlapbach (Ed.), Topics in Applied Physics, vol. 63, Springer,
Berlin, 1988, p. 219.
[29] H. Buchener, R. Povel, Int. J. Hydrogen Energy 7 (1982) 259.
[30] G. Liang, J. Huot, S. Boily, A. Van Neste, R. Schulz, J. Alloys
Compd. 292 (1999) 247–252.
[31] K.J. Gross, J. Alloys Compd. 261 (1997) 276–280.
[32] H. Jensen, Z. Takara, J. Alloys Compd. 285 (1999) 119–122.
 E. David / Journal of Materials Processing Technology 162–163 (2005) 169–177 177

[33] C.E. Thomas, I.F. Kuhn, B.D. James, F.D. Lomax, G.N. Baum, Int.
J. Hydrogen Energy 23 (6) (1998) 507–516.
[34] L. Schlapbach, A. Zuttel, Nature 414/415 (2001) 353–358.
[35] B. Bogdanovic, M. Schwickardi, J. Alloys Compd. 253 (1997) 1–9.
[36] T. Raissi, A. Banerjee, K. Sheinkopf, Proceedings of 31st Interso-
ciety Energy Conversion Engineering Conference, vol. 4 (9), 1996,
pp. 2280–2285.
Further reading
[21] K. Bohmhammel, J. Garche, L. Jörissen, T. Schliermann, W. Schütz,
R. Ströbel, V. Trapp, G. Wolf, J. Power Sources 84 (1999) 221–
224.
[22] T. Bose, R. Chanine, E. Poirier, Int. J. Hydrogen Energy 26 (2001)
831–835.