Aluminium hydride: A reversible material for hydrogen storage

Article abstract of DOI:10.1039/b901878f

Aluminium hydride has been synthesized electrochemically, providing a synthetic route which closes a reversible cycle for regeneration of the material and bypasses expensive thermodynamic costs which have precluded AlH₃ from being considered as

Full text of DOI:10.1039/b901878f

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Aluminium hydride: a reversible material for hydrogen storagew
Ragaiy Zidan,* Brenda L. Garcia-Diaz, Christopher S. Fewox, Ashley C. Stowe,
Joshua R. Gray and Andrew G. Harter
Received (in Berkeley, CA, USA) 29th January 2009, Accepted 10th April 2009
First published as an Advance Article on the web 29th May 2009
DOI: 10.1039/b901878f
Aluminium hydride has been synthesized electrochemically,
providing a synthetic route which closes a reversible cycle for
regeneration of the material and bypasses expensive thermo-
dynamic costs which have precluded AlH3 from being considered
as a H2 storage material.
heat of formation and heat of fusion of LiCl.8 In addition,
ꢀ117 kJ molꢀ1 is required to regenerate LiAlH4. The large
amount of energy required to regenerate AlH3 from spent
aluminium and the alkali halide makes this chemical synthesis
route economically impractical for a reversible AlH3 storage
system.
For these reasons, our research has pursued a more
economically and thermodynamically cost effective synthetic
and regeneration route to produce alane. We believe our
findings will not only impact the hydrogen storage community
but will also have broad influence on the synthetic endeavors
of others, especially those looking to bypass unfeasible
thermodynamics. We present a cycle that uses electrolysis
and catalytic hydrogenation of spent Al(s) to avoid both the
high hydriding hydrogen pressure of aluminium and the
formation of stable by-products such as LiCl. The cycle,
shown in Fig. 1, utilizes electrochemical potential to drive
the formation of alane and alkali hydride (e.g. LiH, NaH and
KH) from an ionic alanate salt. The starting alanate is
regenerated by direct hydrogenation of spent aluminium with
the byproduct alkali hydride (e.g. NaH) in the presence
of a titanium catalyst under moderate hydrogen pressure
(B100 bars), a well-studied reaction.9–13 In comparison to
chemical methods, using the electrochemical cycle should, in
principle, cost the heat of formation of NaAlH4, which
Discovering efficient and economic methods for storing
hydrogen is critical to realizing the hydrogen economy. The
US Department of Energy (DOE) is supporting research to
demonstrate viable materials for on-board hydrogen storage.
The DOE goals are focused on achieving a storage system of
6 mass% H2 and 45 kg H2 mꢀ3 by 2010 and developing a
system reaching 9 mass% H2 and 81 kg H2 mꢀ3 by 2015.1
Researchers worldwide have identified a large number of
compounds with high hydrogen capacity that can fulfill these
gravimetric and volumetric requirements. Unfortunately, the
majority of these compounds fail to meet the thermodynamic
and kinetic requirements for on-board storage systems.
Alane has the gravimetric (10.1 mass% H2) and volumetric
(149 kg H2
m
ꢀ3) density needed to meet the 2010 DOE
goals.2,3 In addition, rapid hydrogen release from alane can
be achieved using only the waste heat from a fuel cell or a
hydrogen internal combustion engine.4 The main drawback to
using alane in hydrogen storage applications is unfavorable
hydriding thermodynamics. The direct hydrogenation of
aluminium to alane requires over 105 bars of hydrogen
pressure at room temperature as shown in eqn (1). The
impracticality of using high hydriding pressure has precluded
alane from being considered as a reversible hydrogen storage
material.
is ꢀ115 kJ molꢀ1
.
The electrolysis reaction is carried out in an electro-
chemically stable, aprotic, polar solvent such as THF or ether.
NaAlH4 is dissolved in this solvent, forming the ionic solution
(Na+–AlH4ꢀ–THF) which is used as an electrolyte. Though
not directed at the regeneration of alane, elaborate research
105 bar 25 ꢁ
Al þ H2 ꢀ! AlH3
C
3
2
ð1Þ
The typical formation route of alane is through the chemical
reaction of lithium alanate with aluminium chloride in
diethyl ether. This reaction yields dissolved alane etherate,
AlH3ꢂEt2O, and precipitates lithium chloride. Alane can then
be separated from the ether by heating in vacuo.3,5–7 The
synthesis of AlH3 by this method results in the formation of
alkali halide salts such as LiCl. The formation of these salts
becomes a thermodynamic sink because of their stability.
For a cyclic process, lithium metal must be recovered from
lithium chloride by electrolysis of a LiCl–KCl melt at 600 1C
and costing at least ꢀ429 kJ molꢀ1 of energy equivalent to the
Energy Security Directorate, Savannah River National Lab,
P. O. Box A, Aiken, SC 29808, United States of America.
E-mail: Ragaiy.Zidan@srnl.doe.gov
w Electronic supplementary information (ESI) available: Material
preparation and characterization. See DOI: 10.1039/b901878f
Fig. 1 Proposed reversible fuel cycle for alane. All components of the
electrochemical process can be recycled to continually afford a viable
solid state storage material for the hydrogen economy.
ꢃc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 3717–3719 | 3717

View Online
and extensive studies on the electrochemical properties of this
type of electrolyte have been reported.14,15 Although attempts
in the past were made to synthesize alane electrochemically,16–18
none have shown isolated material or a characterized alane
product.
Thermodynamic properties of the above electrolyte, along
with cyclic voltammetry were the basis for conducting
this electrochemical process. Thermodynamic calculations
were made to determine the reduction potentials for possible
electrochemical reactions of NaAlH4 in an aprotic solution
(THF) with an aluminium electrode. From the half reaction
potentials, the cell voltage for alane formation was calculated
and a theoretical cyclic voltammogram was constructed
(see Fig. 2).
Fig. 3 Aluminium electrode before (left) and after reaction (right) in
NaAlH4–THF. The electrode is consumed when the voltage is held
at +1.5 V versus SHE.
evolution of hydrogen is suppressed and the reaction is
expected to consume the Al electrode as in eqn (3):
The onset potential of the NaAlH4 decomposition reaction
to Al is shifted to approximately ꢀ2.05 V vs. SHE indicating
greater than 250 mV of overpotential. A limiting current of
9.7 mA was reached near ꢀ1.5 V. At least two separate
reaction mechanisms can produce alane at the aluminium
electrode. One possible mechanism is the oxidation of the
alanate ion to produce alane, an electron, and hydrogen as
shown in eqn (2):
ꢀ
3AlH4 + Al ꢀ! 4AlH3ꢂnTHF + 3eꢀ
(3)
The different conditions of reactions (2) and (3) depend
on their Gibbs free energy (DG). DG for reaction (2) is
ꢀ167.1 kJ molꢀ1 AlH3 and for reaction (3) is ꢀ113.7 kJ molꢀ1
AlH3. Since Na+ is the counter-ion of AlH4ꢀ, the respective
equilibrium potentials are ꢀ1.73 for reaction (2) and ꢀ1.57 V
for reaction (3) vs. SHE. Experimental observations confirm
that under the conditions of reaction (3), the anode is
consumed as shown in Fig. 3.
ꢀ
ꢀ
1
2
AlH4 ꢀ! AlH3ꢂnTHF + H2m + e
(2)
Another possible mechanism is the reaction of AlH4ꢀ with the
aluminium anode to form alane. In this reaction route, the
Once the correct operating voltage for the formation of
alane was calculated, constant voltage experiments were
performed. During these experiments, the current was steady
and increased slightly with time. The electrochemical production
of alane is found not to be slowed by the formation of
AlH3. The surface area of the electrodes and the cell current
are observed to be the rate limiting factors. In contrast to
previous reports, no visible signs of alane formation are
observed and the alane produced by our method is completely
dissolved in solution as a THF adduct.18 During electrolysis,
dendritic material was deposited on the platinum counter
electrode. This material was collected and determined to be
Na3AlH6 from XRD data.
Experiments were conducted to determine the feasibility of
plating sodium at the platinum cathode to complete the cycle.
The platinum cathode and aluminium anode potentials
were ꢀ2.89 V and ꢀ1.31 V, respectively. Plating of Na metal
was observed at the cathode while alane was produced at the
aluminium anode. In this case, no dendrites were observed at
the platinum cathode as sodium was reduced. Furthermore,
bubbling hydrogen at the platinum cathode can form NaH
and increase the efficiency by eliminating dendrite formation
and the need to reduce Na+ to sodium metal.
With an over potential of B0.3 V the alane process is 25%
of the H2 value gained from alane. This value is less than that
required in compression and liquefaction of H2 gas (30–35%).
Once our understanding of the electrochemical processes took
shape, our attention turned to harvesting AlH3 from the
solution. Alane recovered from the electrochemical cells
was characterized by powder X-ray diffraction, Raman
spectroscopy, and TGA. The methods of chemical separation
and results of characterization are discussed herein.
Fig. 2 Experimental and hypothetical cyclic voltammograms for
the electrochemical formation of alane. (a) A hypothetical cyclic
voltammogram was formulated from the equilibrium potential data
for possible reactions and the anticipated state of each species generated.
(b) Bulk electrolysis experiment at an aluminium wire electrode for a cell
containing a 1.0 M solution of NaAlH4 in THF at 25 1C.
ꢃc
This journal is The Royal Society of Chemistry 2009
3718 | Chem. Commun., 2009, 3717–3719

View Online
ꢀ
absences, the space group was assigned as R3c which is
consistent with a-alane.19
In conclusion, we have demonstrated the feasibility of a
recyclable and reversible hydrogen storage material. We have
used the nature of the alane molecule and its tendency to form
complexes to our advantage, helping in the isolation of a pure,
highly crystalline compound. This generation cycle of alane
provides a clean, facile route to a high capacity H2 storage
material while avoiding unrecoverable thermodynamic costs.
We anticipate this work will impact other fields including
those of thin films, adduct based syntheses, and the recycling
and regeneration of other materials.
Work presented here is supported by the DOE office of
EERE (Energy Efficiency and Renewable Energy) Contract
Number EB4202000. Our gratitude is extended to Dr Robert
Lascola for his expertise in characterizing materials using
Raman spectroscopy. Special thanks are given to Joseph G.
Wheeler for whose assistance and guidance we are extremely
grateful.
Notes and references
1 Hydrogen, Fuel Cells & Infrastructure Technologies Program.
Multi-Year Research, Development and Demonstration Plan:
Planned Program Activities for 2005–2015. http://www1.eere.
energy.gov/hydrogenandfuelcells/mypp/ (2007).
2 S. Beattie, T. Humphries, L. Weaver and S. McGrady, in 2008
APS March Meeting, American Physical Society, New Orleans,
Louisiana, 2008.
Fig. 4 XRD patterns for products recovered from an electrochemical
cell. (a) Alane separated from reaction mixture as the THF adduct.
When heated under vacuum to remove THF, the solid partially
decomposes, losing hydrogen and affording aluminium. (b) Alane is
recovered using triethylamine.
3 J. Graetz and J. J. Reilly, J. Phys. Chem. B, 2005, 109,
22181–22185.
4 G. Sandrock, J. Reilly, J. Graetz, W. M. Zhou, J. Johnson and
J. Wegrzyn, J. Alloys Compd., 2006, 421, 185–189.
5 A. E. Finholt, A. C. Bond and H. I. Schlesinger, J. Am. Chem.
Soc., 1947, 69, 1199–1203.
6 F. M. Brower, N. E. Matzek, P. F. Reigler, H. W. Rinn,
C. B. Roberts, D. L. Schmidt, J. A. Snover and K. Terada,
J. Am. Chem. Soc., 1976, 98, 2450–2453.
7 T. Kato, Y. Nakamori, S. Orimo, C. Brown and C. M. Jensen,
J. Alloys Compd., 2007, 446, 276–279.
8 U. Wietelmann and R. J. Bauer, in Ullmann’s Encyclopedia of
Industrial Chemistry 2005, Wiley-VCH, Weinheim, 2005.
9 B. Bogdanovic and M. Schwickardi, J. Alloys Compd., 1997, 253,
1–9.
The separation of alane from the etherate adduct is well
established and affords the pure AlH3 compound. Separation
of the AlH3ꢂTHF adduct is however not as straight forward
as that of the etherate and has proven more sensitive to
decomposition if isolated in the same manner as the etherate.
For this reason we have explored the used of adducts in
purifying the product obtained from the cell. The procedure
involves precipitation of the unconsumed sodium alanate
using a non-polar hydrocarbon followed by filtration and
introduction of triethylamine (TEA). The adduct free alane
is then recovered by heating the neat liquid AlH3ꢂTEA
in vacuo.
10 R. A. Zidan, S. Takara, A. G. Hee and C. M. Jensen, J. Alloys
Compd., 1999, 285, 119–122.
11 C. M. Jensen, R. Zidan, N. Mariels, A. Hee and C. Hagen, Int. J.
Hydrogen Energy, 1999, 24, 461–465.
12 B. Bogdanovic, R. A. Brand, A. Marjanovic, M. Schwickardi and
J. Tolle, J. Alloys Compd., 2000, 302, 36–58.
13 G. Sandrock, K. Gross and G. Thomas, J. Alloys Compd., 2002,
339, 299–308.
14 H. Senoh, T. Kiyobayashi, N. Kuriyama, K. Tatsumi and
K. Yasuda, J. Power Sources, 2007, 164, 94–99.
15 H. Senoh, T. Kiyobayashi and N. Kuriyama, Int. J. Hydrogen
Energy, 2008, 33, 3178–3181.
16 S. Adhikari, J. J. Lee and K. R. Hebert, J. Electrochem. Soc., 2008,
155, C16–C21.
Powder X-ray diffraction pattern data for two different
separation methods are shown in Fig. 4. When alane is
separated by simply heating the AlH3ꢂTHF product after
removing left over starting materials, the diffraction pattern
shows the presence of aluminium metal as well as a-alane. The
TEA adduct is more stable and less susceptible to thermal
degradation. Separation using the TEA method affords only
the pure a-alane phase as seen in the diffraction pattern; no
aluminium is present. Indexing of this pattern was performed
and the unrefined unit cell parameters were found to be
a = 4.446 A and c = 11.809 A. Based on the systematic
17 N. M. Alpatova, T. N. Dymova, Y. M. Kessler and O. R. Osipov,
Russ. Chem. Rev., 1968, 37, 99–114.
18 H. Clasen, Ger. Pat., 1141 623, 1962.
19 J. W. Turley and H. W. Rinn, Inorg. Chem., 1969, 8, 18–22.
ꢃc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 3717–3719 | 3719