Temporal and spatial imaging of hydrogen storage materials: Watching solvent and hydrogen desorption from aluminium hydride by transmission electron microscopy

Article abstract of DOI:10.1039/b807091a

An in situ thermal desorption study of solvated aluminum hydride (alane) by transmission electron microscopy and selected area diffraction has permitted characterisation of the structural and morphological changes during desorption of solvent and hydrogen

Full text of DOI:10.1039/b807091a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Temporal and spatial imaging of hydrogen storage materials: watching
solvent and hydrogen desorption from aluminium hydride by transmission
electron microscopyw
a
a
b
a
Shane D. Beattie,* Terry Humphries, Louise Weaver and G. Sean McGrady*
Received (in Cambridge, UK) 29th April 2008, Accepted 11th June 2008
First published as an Advance Article on the web 28th July 2008
DOI: 10.1039/b807091a
An in situ thermal desorption study of solvated aluminum
hydride (alane) by transmission electron microscopy and selected
area diffraction has permitted characterisation of the structural
and morphological changes during desorption of solvent and
hydrogen in real-time; this powerful technique for studying
hydrogen storage materials complements several others already
employed.
by multiple techniques. The chemical and structural changes
that occur in metal hydride systems during hydrogen desorption
have been studied thoroughly using a range of analytical
4
techniques including isothermal desorption, differential scann-
6
,7
8
ing calorimetry (DSC), thermal gravimetric analysis (TGA),
9,10
powder X-ray diffraction (XRD),
11
neutron diffraction, and
12,13
solid-state NMR spectroscopy (SSNMR).
All of these ana-
lytical techniques report information about the bulk properties
of the sample, either at the macroscopic level, or as a statistical
average at the atomic level. They are insensitive to material
changes at the micron to nanometer scale. Scanning electron
There are extensive efforts underway worldwide to develop
inexpensive and practical hydrogen storage materials for the
implementation of a hydrogen-based energy economy. Per-
haps the most promising candidates are light metal hydrides
microscopy (SEM) has been used by several groups to study the
14–16
1
,2
and their complexes. Amongst these, the binary hydride of
aluminium, alane (AlH , is increasingly recognised as a
leading candidate, satisfying most of the requirements laid
size and morphology of metal hydride particles,
but this
3
)
x
technique has never been used to characterise the hydrogen
uptake or release characteristics of these materials in situ.
Transmission electron microscopy (TEM) is uniquely suited
to the study of localised changes in a material at the nano-
meter scale. Furthermore, it is possible to study phase transi-
tions, by in situ heating and imaging. As hydrogen gas is
evolved, and the material undergoes a transition, structural
and morphological changes can be followed as a function of
temperature and time. Although metal hydrides have been
3
out by the US Department of Energy. Alane demonstrates
impressive gravimetric and volumetric hydrogen content
ꢀ
3
around 10 wt% H and 150 kg H m ), acceptable H2
(
release kinetics, and a low dehydrogenation temperature
4
(
3
60–150 1C). Conventionally, AlH is prepared by the method
5
of Brower et al., involving formation of an ethereal inter-
mediate (eqn (1)) that is converted to the polymeric hydride on
annealing (eqn (2)). Alane is a kinetically stable hydride, with
1
6,17
previously studied using TEM,
study to image–in real time—the local changes occurring
during heating and desorption of H
Powdered AlH O was prepared using the method
ꢁnEt
this appears to be the first
5
an equilibrium dissociation pressure of B10 bar at ambient
temperature, and the removal of the ether solvent (eqn (2))
.
2
converts the material from a gelatinous paste into a fine, dry
3
2
5
1,2
5
powder. In contrast to many metal hydride systems, the
dehydrogenation of alane is a chemically simple process,
yielding aluminium metal and hydrogen gas in a single step
developed by Brower et al. (refer to ESIw). The powder was
sprinkled onto a copper/carbon grid and transferred to a TEM
heating stage. This transfer process involved brief exposure to
the atmosphere. An image of the as-prepared sample is
displayed in Fig. 1. The particles are coated with a thin
amorphous layer of (presumably) aluminum oxide, resulting
from exposure to the atmosphere during the transfer process.
Residual diethyl ether is present in the sample. A selected area
diffraction pattern (SADP) of the as-prepared material is
shown as the inset of Fig. 1; the SADP indicates amorphous
(
eqn (3)).
Et2O
3
LiAlH4 þ AlCl3 ! 4AlH3 ꢁ nEt2O þ 3LiCl
AlH O - 4AlH + nEt
ꢁnEt
AlH ) - xAl + 1.5xH
ð1Þ
(2)
(3)
4
3
2
3
2
O
(
3
x
2
However, from a material standpoint, the dehydrogenation
process is much more complicated, and requires characterisation
AlH
ether-solvated alane by Graetz and Reilly.
3
ꢁnEt
2
O, in agreement with the recent XRD study of
4
The sample was heated in vacuo to 80 1C inside the TEM
chamber. After 90 min, small flecks were evident in some of
the particles. A dark-field image (Fig. 2) revealed small crystal-
lites emerging, appearing as bright spots embedded in the
amorphous matrix. A SADP was obtained close to one of
these bright-spot crystallites, and is shown as the inset to
Fig. 2. The SADP reveals that the material is still largely
a
Department of Chemistry, University of New Brunswick, 30 Dineen
Drive, Fredericton, NB, Canada E3B 6E2. E-mail: sbeattie@unb.ca;
smcgrady@unb.ca
Microscopy and Microanalysis Facility, University of New
b
Brunswick, 10 Bailey Drive, Fredericton, NB, Canada E3B 5A3
w Electronic supplementary information (ESI) available: Synthesis of
AlH
3
ꢁnEt
2
O and a-AlH
3
; SADP ring measurements, powder XRD
pattern of a-AlH
3
. See DOI: 10.1039/b807091a
4
448 | Chem. Commun., 2008, 4448–4450
This journal is ꢂc The Royal Society of Chemistry 2008

View Article Online
Fig. 3 TEM image and SADP (inset) of a fully decomposed sample
of AlH
3
ꢁnEt
2
O.
Fig. 1 TEM image and associated SADP (inset) of a freshly prepared
sample of AlH O.
ꢁnEt
3
2
1
8
to Al metal (refer to ESIw for ring measurements), confirming
the decomposition reaction (eqn (3)) to be complete. The more
diffuse features towards the centre of this SADP indicate the
amorphous, but nascent diffuse rings indicate a degree of
crystallinity starting to evolve. The spacing of these rings
corresponds to Al metal (refer to ESIw for ring measure-
ments). The number and widespread distribution of the
Al crystallites indicates that the dehydrogenation of alane
presence of a residual amorphous phase; as-yet unidentified.
4,6
1
8
Previous studies by Graetz and Reilly
indicate that
AlH ꢁnEt O loses ether to produce a-AlH before decompos-
3
2
3
(
eqn (3)) has occurred simultaneously in localized regions
ing to Al. We were unable to identify definitively this solvent-
free a-AlH3 phase during the experiments described above.
However, a faint SADP was obtained, with features that may
throughout the bulk material. These observations agree with
the conclusions drawn by Graetz and Reilly on the basis of a
4
DSC study; namely, that the decomposition of alane is
correspond to the a-AlH (Fig. S1 in ESIw). In order to clarify
3
controlled by nucleation and growth of Al particles in two
and three dimensions, and that the kinetics are not dependent
on catalysis by surface oxide nor limited by hydrogen diffusion
through an oxide barrier.
the origin of these features, a follow-up series of TEM experi-
ments was conducted, this time using ether-free a-AlH (refer
3
to ESIw for details), whose identity was confirmed ex situ by
1
9
powder XRD (Fig. S2 in ESIw). Although the a-AlH sample
3
After prolonged heating for 8 h at 80 1C, the image shown in
Fig. 3 was obtained. The image clearly reveals an agglomeration
of Al crystallites (average size B50 nm) in an amorphous
matrix. The SADP displayed as an inset to Fig. 3 indicates a
high degree of crystallinity, with the ring spacing corresponding
showed well defined peaks in the ex situ powdered XRD
pattern, it was difficult to find a SADP corresponding to a-
AlH with the TEM. After several inconclusive attempts, it
3
was realised that a-AlH is sensitive to the high energy TEM
3
electron beam. This is perhaps unsurprising, as previous
Fig. 2 Dark-field image and associated SADP (inset) after heating
3
Fig. 4 TEM image and SADP (inset) of a-AlH after short exposure
AlH
3
ꢁnEt
2
O for 90 min at 80 1C.
to the TEM beam.
This journal is ꢂc The Royal Society of Chemistry 2008
Chem. Commun., 2008, 4448–4450 | 4449

View Article Online
particles. Furthermore, the rate of decomposition is not con-
trolled by diffusion of H
2
to the surface. Amorphous
AlH ꢁnEt O loses the solvent to produce a-AlH , which then
3
2
3
proceeds rapidly to lose H , leaving behind crystalline Al with
2
a particle size of ca. 50 nm. In spite of the sensitivity of the
material to the electron beam valuable information has been
obtained about each of the decomposition steps involved in
the dehydrogenation and the structural evolution of a-AlH
at the nanoscopic level – in real-time. Although we have been
successful in studying AlH by this technique, the sensitivity of
3
–
3
the sample to the high energy TEM electron beam may prove
to be a more serious problem with other metal hydrides.
We thank NSERC, CFI, NBIF and HSM Systems for
financial support of this research.
Notes and references
Fig. 5 TEM image and SADP (inset) of the same area as Fig. 4 after
z A time-resolved movie of the decomposition of a-AlH
3
to Al is
prolonged exposure to the TEM beam.
available on the McGrady group website: http://www.unb.ca/freder-
icton/science/chem/smcgrady/news/index.html
studies have shown the decomposition of a-AlH to Al to be
3
1
4
60
accelerated by exposure to UV light, and to 1 MeV Co
1
W. Grochala and P. P. Edwards, Chem. Rev., 2004, 104, 1283.
2 S. Orimo, Y. Nakamori, J. R. Eliseo, A. Zuttel and C. M. Jensen,
Chem. Rev., 2007, 107, 4111.
S. Satyapal, J. Petrovic, C. Read, G. Thomas and G. Ordaz, Catal.
Today, 2007, 120, 246.
J. Graetz and J. J. Reilly, J. Phys. Chem. B, 2005, 109, 22181.
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.
2
0
¨
g-irradiation. Although the sample decomposes within sec-
onds on exposure to the electron beam, ultimately, a SADP
3
that clearly corresponds to a-AlH
The SADP in Fig. 4 contains contributions from both a-
AlH and metallic Al (refer to ESIw for ring measurements). It
3
was obtained (Fig. 4).
4
5
3
is difficult to identify unambiguously certain features, as both
materials have similar d-spacings. However, a-AlH has a
6 J. Graetz and J. J. Reilly, J. Alloys Compd., 2006, 424, 262.
3
7
8
9
M. Resana, M. D. Hamptona, J. K. Lomnessa and D. K. Slattery,
Int. J. Hydrogen Energy, 2005, 30, 1417.
J. Wang, A. D. Ebner, T. Prozorov, R. Zidan and J. A. Ritter,
J. Alloys Compd., 2005, 395, 252.
B. Bogdanovic, R. A. Brand, A. Marjanovic, M. Schwickardi and
˚
unique reflection at 3.28 A, which is clearly observed in the
˚
SADP of Fig. 4. The 3.28 A reflection was observed previously
(
refer to ESI,w Fig. S1) during the initial stages of AlH ꢁnEt O
3
2
decomposition, however, the SADP was diffuse and it was
difficult to definitively identify the a-AlH phase.
¨
J. Tolle, J. Alloys Compd., 2003, 350, 246–255.
0 E. H. Majzoub and K. J. Gross, J. Alloys Compd., 2003, 356–357,
3
1
Fig. 5 shows the same area as Fig. 4 and corresponding
SADP following further exposure to the electron beam. Of
particular note in Fig. 5 are the larger grains, the absence of
363.
11 S. Singh, S. W. H. Eijt, J. Huot, W. A. Kockelmann, M. Wage-
maker and F. M. Mulder, Acta Mater., 2007, 55, 5549.
2 V. P. Tarasov, Yu. B. Muravlev, S. I. Bakum and A. V. Novikov,
1
˚
the earlier spots at 3.28 A, and the strong Al doublet rings at
Dokl. Phys. Chem., 2003, 393, 353.
13 O. J. Zogait, M. Punkkinen, E. E Ylinent and B. Stalinski,
˚
2
These features unambiguously point to the disappearance of
.338 and 2.024 A (refer to ESIw for ring measurements).
J. Phys.: Condens. Matter, 1990, 2, 1941.
4 P. J. Hereley, S. O. Christofferson and J. A. Todd, J. Solid State
Chem., 1980, 35, 391.
5 B. Bogdanovic, R. A. Brand, A. Marjanovic, M. Schwickardi and
J. Tolle, J. Alloys Compd., 2000, 302, 36–58.
¨
6 A. Leon, O. Kircher, H. Rosner, B. Decamps, E. Leroy, M.
Fichtner and A. P. Guegan, J. Alloys Compd., 2006, 414, 190.
7 M. Felderhoff, K. Klementiev, W. Grunert, B. Spliethoff, B.
1
1
1
1
a-AlH and the emergence of metallic Al in its place.
3
To the best of our knowledge, this is the first real-time, in
situ study by TEM of hydrogen desorption from a metal
hydride.z The high resolution of the technique, in combination
with the ability to obtain localised diffraction patterns, has
allowed an in situ temperature and time resolved study of the
structural and morphological changes that occur during des-
Tesche, J. M. Bellosta von Colbe, B. Bogdanovic
Pommerin, F. Schuth and C. Weidenthalera, Phys. Chem. Chem.
Phys., 2004, 6, 4369.
´
, M. Hartel, A.
¨
olvation and hydrogen desorption. Decomposition of a-AlH
occurs spontaneously at localized sites throughout the bulk
3
1
1
8 JCPDS-ICCD PDF-2 Database #4-787, 1996.
9 JCPDS-ICCD PDF-2 Database #14-1436, 1996.
material, and is not initiated by nucleation at the surface of the
20 P. J. Herley and R. H. Irwin, J. Phys. Chem. Solids, 1978, 39, 1013.
4
450 | Chem. Commun., 2008, 4448–4450
This journal is ꢂc The Royal Society of Chemistry 2008