The effect of H2 partial pressure on the reaction progression and reversibility of lithium-containing multicomponent destabilized hydrogen storage systems

Article abstract of DOI:10.1021/ja204381n

It is known that the reaction path for the decomposition of LiBH ₄:MgH₂ systems is dependent on whether decomposition is performed under vacuum or under a hydrogen pressure (typically 1-5 bar). However, the sensitivity of this multicomponent hydride system to partial pressures of H₂ has not been investigated previously. A combination of in situ powder neutron and X-ray diffraction (deuterides were used for the neutron experiments) have shed light on the effect of low partial pressures of hydrogen on the decomposition of these materials. Different partial pressures have been achieved through the use of different vacuum systems. It was found that all the samples decomposed to form Li-Mg alloys regardless of the vacuum system used or sample stoichiometry of the multicomponent system. However, upon cooling the reaction products, the alloys showed phase instability in all but the highest efficiency pumps (i.e., lowest base pressures), with the alloys reacting to form LiH and Mg. This work has significant impact on the investigation of Li-containing multicomponent systems and the reproducibility of results if different dynamic vacuum conditions are used, as this affects the apparent amount of hydrogen evolved (as determined by ex situ experiments). These results have also helped to explain differences in the reported reversibility of the systems, with Li-rich samples forming a passivating hydride layer, hindering further hydrogenation.

Full text of DOI:10.1021/ja204381n

ARTICLE
pubs.acs.org/JACS
The Effect of H2 Partial Pressure on the Reaction Progression and
Reversibility of Lithium-Containing Multicomponent Destabilized
Hydrogen Storage Systems
Tobias E. C. Price,† David M. Grant,‡ David Weston,† Thomas Hansen,§ Lene M. Arnbjerg,^
Dorthe B. Ravnsbæk,^ Torben R. Jensen,^ and Gavin S. Walker*,†
†Energy and Sustainability Research Division, and ‡Materials, Mechanics and Structures Division, Engineering Faculty,
The University of Nottingham, University Park, Nottingham, NG7 2RD, U.K.
§Institut Laue Langevin, 6 rue Jules Horowitz, BP 156, 38042 Grenoble cedex 9, France
^Center for Materials Crystallography, iNANO and Department of Chemistry, Aarhus University, Langelandsgade 140,
DK-8000 Aarhus C, Denmark
ABSTRACT: It is known that the reaction path for the decomposition of LiBH4:
MgH2 systems is dependent on whether decomposition is performed under
vacuum or under a hydrogen pressure (typically 1À5 bar). However, the
sensitivity of this multicomponent hydride system to partial pressures of H2 has
not been investigated previously. A combination of in situ powder neutron and
X-ray diffraction (deuterides were used for the neutron experiments) have shed
light on the effect of low partial pressures of hydrogen on the decomposition of
these materials. Different partial pressures have been achieved through the use of different vacuum systems. It was found that all the
samples decomposed to form LiÀMg alloys regardless of the vacuum system used or sample stoichiometry of the multicomponent
system. However, upon cooling the reaction products, the alloys showed phase instability in all but the highest efficiency pumps (i.e.,
lowest base pressures), with the alloys reacting to form LiH and Mg. This work has significant impact on the investigation of Li-
containing multicomponent systems and the reproducibility of results if different dynamic vacuum conditions are used, as this affects
the apparent amount of hydrogen evolved (as determined by ex situ experiments). These results have also helped to explain differences
in the reported reversibility of the systems, with Li-rich samples forming a passivating hydride layer, hindering further hydrogenation.
A widely investigated system is that of LiBH4 combined with
MgH2; however, there have been inconsistencies in the reported
reactions and the reversibility throughout the literature, depend-
ing on the decomposition environment, in particular when the
decomposition was performed under vacuum.5,6,8À11
In situ structural characterization by synchrotron radiation
powder X-ray diffraction (SR-PXD) or neutron diffraction
(PND) has been key in probing the interplay between the
components, varying the stoichiometry and reaction environ-
ment within the LiBH4:MgH2 system.4,7,12À16 For decomposi-
tions performed under a hydrogen partial pressure at or above 1
bar, the system forms MgB2 as shown in Scheme 1, as confirmed
by in situ PND for both the 2:1 ratio and a magnesium-rich
system with a 0.3:1 ratio4 and in situ SR-PXD of the 2:1 ratio by
Bosenberg et al.12
Under a dynamic vacuum environment, both LiBH4:MgH2
ratios were confirmed to decompose through a reaction origin-
ally proposed by Yu et al., Scheme 2,3 forming the R- and β-alloys
of MgxLi1-x (where x = 0.816 and 0.70, respectively).4,7
Although there is consistency for the results reported by
different laboratories for the reaction under p(H2) g 1
1. INTRODUCTION
A concerted investigation into alternatives to fossil fuels for
mobile power generation has been undertaken worldwide in
order to reduce our anthropogenic effects on the environment.
The potential of hydrogen as an energy vector for mobile and
small scale energy generation has led to challenging H2 capacity
targets. For automotive applications, a system target of 5.5 wt.%
H2 capacity by 2015 has been set by the US Department of
Energy.1 Light metal complex hydrides have been the source of
many studies due to their high volumetric and gravimetric
capacity. Complex hydrides based on [AlH4]À and [BH4]À
anions ionically bonded to a light metal cation such as Li+,
Mg2+, Na+ give some of the highest achievable hydrogen storage
capacities (viz., 18.5 wt.% for LiBH4).2 However, the high
stability of many complex hydrides causes decomposition tem-
peratures to be significantly higher than that required for
automotive applications, e.g., LiBH4 does not fully decompose
until >600 ꢀC.3 Development of systems based on these complex
hydrides for automotive applications requires the operating
temperatures to be significantly reduced which may be achieved
by thermodynamic destabilization of the system.3À7 Addition of
a destabilization reagent which leads to the formation of more
stable end products will reduce the enthalpy for dehydrogena-
tion, leading to a reduction in the decomposition temperature.
Received: May 12, 2011
Published: July 14, 2011
r
2011 American Chemical Society
13534
dx.doi.org/10.1021/ja204381n J. Am. Chem. Soc. 2011, 133, 13534–13538
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Reaction Pathway when the Decomposition is
under a Pressure
Scheme 2. Reaction Pathway when the Decomposition Is
under a Dynamic Vacuum
bar,5,6,8À12,16 this is not the case for groups investigating the
decomposition reaction under a dynamic vacuum. Some groups
have not found evidence for the alloys, reporting instead forma-
tion of Mg metal and LiH, and for the 2:1 ratio, finding no
reversibility for these end products.5,6,8 In contrast, we have
shown that hydrogenation of the alloy-containing end products
have excellent reversibility and very fast kinetics.4 Investigations
in the literature have used varying pumping systems, of differing
H2 pumping efficiencies, which will affect the ultimate base
pressure and hence the partial pressure of hydrogen into which
the decompositions were performed. The results presented
below represent a collation of data from our investigations into
0.3:1, 0.23:1, 0.44:1, and 2:1 ratios of MgH2:LiBH4, through
PND and PXD in situ investigations. This has allowed us to
investigate the effect of different pumping systems and the
quality of the vacuum they produce upon the reaction pathways
of these multicomponent systems.
Figure 1. PXD and PND results for products of LiBH4:MgH2 samples
decomposed under a dynamic vacuum followed by subsequent cooling
to 260 ꢀC. The sample stoichiometry and type of vacuum system used
(VSA, VSB, or VSC) is indicated on the plot.
10À1 and 10À2 mbar measured next to the pump separated from the
manifold by a low conductance pipe (<10À4 lsÀ1).
The PND experiments were performed using the neutron diffract-
ometer D2020 at the Institute Laue-Langevin (ILL) in Grenoble, France
(λ = 2.42 Å, flux =4.2 Â 107 n sÀ1 cmÀ2). Data were analyzed using the
Large Array Manipulation Program (LAMP) version 6. Samples of 1 g
mass were packed into 316 L stainless steel vessels of 7 mm diameter i.d.,
with a length of 500 mm and sealed under argon gas. Deuteriding was
performed via a gas manifold system connected to the sample, allowing
application of >50 bar D2 pressure at temperatures between 350 ꢀC
(0.23:1 and 0.44:1) and 400 ꢀC (0.3:1). The PND manifold system
is denoted vacuum system B (VSB) for experiments on 0.44:1 and
0.23:1 samples. VSB was pumped by a multistage roots pump (Adixen
ACP-28) which had an ultimate pressure measured at the pump head of
>3 Â 10À2 mbar and connected to the manifold via piping of ca. 0.5 lsÀ1
conductance. A different vacuum system (VSC) was used on the 0.3:1
samples. In this case a scroll pump (Varian SH-110) of ultimate pressure
of 6 Â 10À2 mbar was connected directly to the manifold by a short
2. EXPERIMENTAL METHODS
Three ratios of the LiBH4:MgH2 system were prepared for this study.
The materials used were LiBH4 (Acros Organics, 95%), 7Li11BD4
(Katchem, 98%), and MgH2 (Alfar Aesar, 98%), the main impurities
being limited to oxide and/or unhydrogenated products as characterized
by PXD.14 The latter was ball milled and subsequently cycled in
deuterium (99.8%, D2, BOC) at least two times to form MgD2. All
handling procedures were conducted under an inert atmosphere (<0.3
ppm H2O/O2). Mixtures (1.5 g) of LiBD4:MgD2 (0.3:1, 0.23:1, and
0.44:1 molar ratios) or LiBH4:MgH2 (0.3:1) were mechanically milled
for 1 h under Ar gas at 300 rpm using a Fritsch Rotary P5 ball mill.
In situ experiments performed using PXD and PND are described
below, the gas manifold systems are described in detail, the vacuum
pumps attached to the manifold systems were commercial pumps which,
like most commercial pumps, have little information on the pumping
characteristics of hydrogen or deuterium. However, the base pressures
achievable, directly above the pump, in hydrogen and deuterium are
likely to be up to 2 orders of magnitude higher than that achievable in N2
or Ar.17 The pressure above the sample will also be affected by the
conductance of the manifold system. This is discussed in relation to each
vacuum system as detailed below.
In situ SR-PXD data were collected at the synchrotron MAX II, Lund,
Sweden, in the research laboratories MAX-Lab at beamline I711
equipped with a MAR165 CCD detector system, using a selected
wavelength of λ = 1.09719 Å.18 The sample cell was developed
specifically for studies of gas/solid reactions at high pressures and
temperatures. The samples were mounted in a sapphire single crystal
tube (1.09 mm o.d., 0.79 mm i.d., Al2O3) in an argon-filled glovebox
(<0.1 ppm O2, H2O).19 The gas manifold allowed high pressure
hydrogen to be applied and evacuation of the sample via a rotary pump
during X-ray data acquisition and is denoted as vacuum system A (VSA).
For the VSA system, the ultimate pressure was measured to be between
flexible pipe of conductance ca. 2 lsÀ1
.
Diffraction peak areas were measured in order to follow the relative
phase composition during the in situ experiments. The data presented
has been collected using hydrogenated and deuterided samples. In order
to simplify the text in both Results and Discussion, the word hydrogen
and hydride will be used for both hydrogenated and deuterided samples,
while the symbol H will denote both H and D. Figures will reflect the
true isotope of the samples.
3. RESULTS
Mg-rich LiBH4:MgH2 samples with a molar ratio of 0.3:1,
0.23:1, and 0.44:1 were heated to above 500 ꢀC until their fully
dehydrogenated state was reached, i.e., until all the LiH had been
decomposed to form LiÀMg alloys. The decomposition pro-
ducts were subsequently cooled, and the stability of the end
products was assessed. Figure 1 shows the diffraction results for
samples after decomposition at elevated temperatures and after
subsequent cooling under vacuum to 260 ꢀC. For the 0.3:1
sample, when decomposed under a vacuum generated by VSC, the
end products at 560 ꢀC contained R-alloy and β-alloy, and these
phases were retained upon cooling to 260 ꢀC; the diffraction lines
had all shifted to smaller d-spacing as expected from the thermal
expansion coefficient for these phases. However, for a 0.3:1
sample decomposed under vacuum generated by VSA, even
though the decomposition products at 515 ꢀC had the same two
phases, R-alloy and β-alloy, after cooling to 260 ꢀC the peak
13535
dx.doi.org/10.1021/ja204381n |J. Am. Chem. Soc. 2011, 133, 13534–13538

Journal of the American Chemical Society
ARTICLE
Scheme 3. Reaction Pathway during Partial Hydrogenation
of the Mg Alloys
upon decomposition of the 0.3:1 system, because the ratio of Li
to Mg is in the middle of a two-phase region in the alloy phase
diagram. For the 0.23:1 sample, only an R-alloy phase was
formed because the Li content was lower and remained within
the single phase R region of the phase diagram. The more Li-rich
0.44:1 sample had sufficiently high Li content to take the alloy
composition across into the single-phase β region of the phase
diagram. These reactions were confirmed by the decomposition
experiments regardless of the vacuum system used, so long as the
sample was kept at an elevated temperature, > 500 ꢀC, as shown
in Figure 1.
The sample behavior upon cooling was as expected when VSC
was used to provide the dynamic vacuum conditions, i.e., there
was no change to the R- and β-alloy end products. This suggests
that VSC provided a sufficiently low partial pressure of H2,
preventing hydrogenation of the Li upon cooling. In contrast, the
sample cooled using VSA led to a significant change in the phases,
with a loss of the β-alloy and a concomitant increase in the
amount of R-alloy. This implies that the Li content available for
the alloy system had decreased, which can only be explained by Li
reacting with H2 to form LiH. However, no LiH was observed in
this PXD experiment, which is due to the fact that both Li and H
have small X-ray scattering cross-sections and that the amount of
LiH formed was beyond the detection limit of the experiment. In
contrast, PND is inherently more sensitive to both elements,22
allowing detection of a LiH phase for the 0.44:1 and 0.23:1
samples when cooled using VSB. Reaction upon cooling of the
β-alloy with a low pressure of H2 resulted in the formation of
R-alloy and LiH, Scheme 3. The reaction of the R-alloy during
cooling under a low pressure of H2 caused formation of Mg and
LiH, Scheme 3. It is possible that had the 0.44:1 sample been
given enough time, all the Li within the remaining R-alloy would
have reacted with the residual H2 (so long as there was enough
residual H2 within the manifold).
Figure 2. Sample composition as characterized by diffraction peak area
analysis during the cooling of LiBH4:MgH2 samples while under vacuum
generated by either a VSA, VSB, or VSC vacuum system.
intensity had decreased for the β-alloy with an increase in
intensity for the R-alloy pattern. For the more Li-rich 0.44:1
sample decomposed under a vacuum generated by VSB, only the
β-alloy was identified at 530 ꢀC, but after cooling to 260 ꢀC the
β-alloy was no longer evident and new phases of LiH and R-alloy
had appeared in the NPD patterns. A more Li-deficient sample,
0.23:1 ratio, was also investigated under a vacuum generated by
VSB, which formed the R-alloy evident at 530 ꢀC, but once again
after cooling to 260 ꢀC the sample had formed LiH; however, this
time the remaining metal was Mg and not an alloy.
Changes in the relative amounts of compounds was investi-
gated using integrated diffracted intensities measured during
cooling of the samples from >520 to 260 ꢀC; see Figure 2. No
change in phase was detected upon cooling the decomposed
0.3:1 sample when VSC was used to generate the vacuum.
However, using VSA led to the β-alloy pattern reducing in
intensity at ca. 430 ꢀC with a concomitant rise in intensity of
an R-alloy pattern. No XRD pattern was observed for a LiH
phase during the cooling phase (nor in fact during the preceding
decomposition). The decomposition products of 0.23:1 and
0.44:1 samples cooled under a VSC generated vacuum showed
that in the more Li-rich 0.44:1 sample, the β-alloy pattern
decreased in intensity, along with the appearance of an R-alloy
phase at 500 ꢀC and a LiH phase at 430 ꢀC. By 260 ꢀC, no β-alloy
was detected, leaving only the R-alloy and LiH diffraction
patterns for this sample. Upon cooling the more Mg-rich
0.23:1 sample, the initial R-alloy pattern increased in intensity
at ca. 490 ꢀC, along with the formation of a LiH phase, both
patterns increasing in intensity as the sample was cooled to
260 ꢀC.
For the 0.23:1 sample, the loss of Li from the R-alloy was
inferred from the lack of change in the d-spacing for this phase
upon cooling. One would have expected a significant decrease in
the d-spacing due to thermal lattice contraction, but the expected
lattice contraction from cooling was negated by the loss of Li
from the lattice, which under isothermal conditions would result
in a lattice expansion. Other supporting evidence is provided by
the increase in peak area for the R-alloy pattern upon cooling,
which is due to loss of Li from the alloy, which reduces the
attenuation effect of Li presence due to its lower scattering cross
section than that for Mg.23
The partial pressure over the sample for each of the vacuum
systems employed was estimated to be of the order of 6 Â 10À2
mbar up to 3 mbar for all three vacuum systems based on the
poorer ultimate pressures expected for hydrogen.17 However,
from the results, VSC was clearly the only vacuum system with a
base pressure lower than the equilibrium pressure for the
hydrogenation of the LiÀMg alloys at 260 ꢀC (i.e., preventing
LiH formation upon cooling). Although all three vacuum sys-
tems had similar projected ultimate pressures for hydrogen, the
scroll pump in VSC was directly attached to the manifold,
4. DISCUSSION
The decomposition of the LiBH4:MgH2 samples under a
dynamic vacuum forms LiÀMg alloys. As reported previously,3,4
the LiH is destabilized by the presence of Mg, leading to the
dehydrogenation of LiH at temperatures significantly below
600 ꢀC. The resulting MgxLi1-x alloy(s) formed, as dictated by
the binary MgÀLi phase diagram,21 is illustrated in Scheme 2.
From the results presented in Figure 1, the R-alloy (viz.,
Mg0.816Li0.184) and β-alloy (viz., Mg0.70Li0.30) were both formed
13536
dx.doi.org/10.1021/ja204381n |J. Am. Chem. Soc. 2011, 133, 13534–13538

Journal of the American Chemical Society
ARTICLE
reducing outgassing from excess piping which may explain its
lower base pressure. All three vacuum systems could deliver a
base pressure higher than the hydrogenation equilibrium pres-
sure at 500 ꢀC, explaining why the alloys were formed above
500 ꢀC regardless of the vacuum system employed. As the
samples cooled, the plateau pressure for the alloys will eventually
fall below the H2 partial pressure that a given vacuum system can
attain. No LiH was detected when VSC was used even upon
cooling the sample to room temperature. This indicates that the
temperature at which the plateau pressure is at a value below the
back pressure from VSC must be lower than 260 ꢀC and at a
temperature where the kinetics are too slow for any significant
hydrogenation to occur. Both VSA and VSB caused a higher H2
partial pressure, enabling hydrogenation of the Li to occur upon
cooling. The temperature at which hydrogenation starts for a
given H2 partial pressure gives an indication of the scale of
thermodynamic destabilization for LiH by Mg.
the reactions that occur during an experiment. It also shows a
surprising effect related merely to the pumping efficiency of a
system and the care needed in interpreting reaction paths only
based on ex situ characterization techniques. The hydrogenation
effect observed due to the partial pressure of H2 while cooling a
sample can be expected for other Li-containing systems and
those containing similarly reactive metals. We have not discussed
static vacuum conditions, because such experiments will not be
under vacuum once the sample starts to dehydrogenate and the
partial pressure normally quickly rises to several bar depending
on the size of the sample and the manifold; thus, virtually all the
dehydrogenation occurs under a significant hydrogen pressure.
5. CONCLUSIONS
The results presented here demonstrate the sensitivity of
LiBH4:MgH2 systems to subtle changes in hydrogen partial
pressure under dynamic vacuum conditions, influencing whether
the decomposition products are either retained upon cooling or
form LiH. The work also shows that LiH was thermodynamically
destabilized by Mg, forming the LiÀMg alloys, reducing
ΔH by 29 kJ molÀ1 H2.
These results offer an explanation for the apparent disparity
between reported reaction pathways in the literature for samples
decomposed under inert conditions. Clearly, inert conditions for
H2-sensitive materials require either decomposition under an
inert flowing gas or a dynamic vacuum system achieving a
sufficiently low H2 partial pressure. These findings highlight
the important role of in situ structural characterization when
following phase progressions, particularly when evolved phases
may be dependent on partial pressures of H2 and changing
temperature. The in situ experiments show how the end pro-
ducts, formed at temperature, can differ from those after cooling;
hence, ex situ experiments can in some instance unintentionally
lead to the wrong reaction pathway being proposed. These
results also show the importance of continuity between reaction
environments when investigating samples through a range of
techniques, not just diffraction experiments, e.g. using a flowing
inert carrier gas such as DSC and TGA.
For LiH, a temperature of 500 ꢀC yields a plateau pressure of
3.5 Â 10À2 mbar H2, calculated from thermodynamic data in the
literature;24 therefore, the pressure attained by all the vacuum
systems would need to be below this plateau pressure in order for
LiH to decompose by 500 ꢀC. None of the pumping systems
used are able to achieve such a low pressure of hydrogen and are
several orders of magnitude greater; thus, the LiH has been
destabilized by Mg, leading to a lower ΔH of dehydrogenation. A
rough calculation of the enthalpy of dehydrogenation can be
made from the equation ΔH = ΔST À RT ln p; assuming the LiH
entropy is the same for the destabilized system and that the
minimum H2 partial pressure for the vacuum systems is esti-
mated as 3 mbar H2, then this would equate to a reduction in ΔH
from 181 kJ molÀ1 H2 to e152 kJ molÀ1 H2 to enable
dehydrogenation above 500 ꢀC. This equates to a heat of
formation of 2.6 kJ molÀ1 for Mg0.816Li0.184 which agrees well
with the values reported in the literature 3.3 kJ molÀ1 25
.
These results help us to understand the apparent disparity
within the literature for reported decomposition products under
vacuum and inert gas conditions for the LiBH4:MgH2 system. In
addition to experiments run under dynamic vacuum where the
H2 partial pressure will be dependent on the pump and manifold,
many results are reported for products formed after temperature-
programmed experiments which use a flowing inert gas (e.g.,
differential scanning calorimetry and thermal gravimetric ana-
lysis). Experiments using a flowing inert gas continually remove
any hydrogen produced during the experiment; thus, there will
be no H2 partial pressure and hence would be analogous to an
ultrahigh vacuum system. This helps to explain the difference in
the results for Yu et al.3 who identified the MgxLi1-x alloys in the
products from DSC run under an Ar carrier gas compared to Vajo
et al.5 who found LiH and Mg products after decomposing under
a vacuum. The results shown above would suggest that Vajo et al.
also formed the alloy products at 450 ꢀC, but these hydrided
upon cooling due to the residual H2 partial pressure.
’ AUTHOR INFORMATION
Corresponding Author
Gavin.walker@nottingham.ac.uk
’ ACKNOWLEDGMENT
This work was supported by the Energy Programme of the
Research Councils UK (RCUK) and through a Carbon Vision
Leadership award to GSW by the Carbon Trust and RCUK.
’ REFERENCES
(1) Dilich, S. Annual Merit Review and Peer Evaluation Meeting, US
Department of Energy, Arlington, VA, May, 2009.
(2) Schlesinger, H. I.; Brown, H. C. J. Am. Chem. Soc. 1940, 62, 3429.
(3) Yu, X. B.; Grant, D. M.; Walker, G. S. Chem. Commun.
(Cambridge, U. K.) 2006, 3906.
(4) Price, T. E. C.; Grant, D. M.; Legrand, V.; Walker, G. S. Int. J.
Hydrogen Energy 2010, 35, 4154.
(5) Vajo, J. J.; Skeith, S. L.; Mertens, F. J. Phys. Chem. B 2005,
109, 3719.
Another discrepancy between the two results is that when the
end products for a 2:1 sample were LiH and Mg, these could not
be hydrogenated to reform the LiBH4,5 but our earlier work has
clearly shown that LiBH4 can be reformed from the end products
containing Mg + LiH or MgxLi1-x alloys.4 However, for the high
Li-content 2:1 system, formation of a LiH passivating layer at the
surface of Mg during hydrogenation caused a significant decrease
in the reaction kinetics, yielding minimal MgH2 formation
under 90 bar H2 at 400 ꢀC.4 These results indicate the impor-
tance of in situ characterization techniques in fully understanding
(6) Pinkerton, F. E.; Meyer, M. S.; Meisner, G. P.; Balogh, M. P.;
Vajo, J. J. J. Phys. Chem. C 2007, 111, 12881.
13537
dx.doi.org/10.1021/ja204381n |J. Am. Chem. Soc. 2011, 133, 13534–13538

Journal of the American Chemical Society
ARTICLE
(7) Walker, G. S.; Grant, D. M.; Price, T. E. C.; Legrand, V.; Yu, X.
J. Power Sources 2009, 189, 902.
(8) Nakagawa, T.; Ichikawa, T.; Hanada, N.; Kojima, Y.; Fujii, H.
J. Alloys Compd. 2007, 446À447, 306.
(9) Fan, M.-Q.; Sun, L.-X.; Zhang, Y.; Xu, F.; Zhang, J.; Chu, H.-l.
Int. J. Hydrogen Energy 2008, 33, 74.
(10) Mao, J. F.; Wu, Z.; Chen, T. J.; Weng, B. C.; Xu, N. X.; Huang,
T. S.; Guo, Z. P.; Liu, H. K.; Grant, D. M.; Walker, G. S.; Yu, X. B. J. Phys.
Chem. C 2007, 111, 12495.
(11) Wang, P.-J.; Fang, Z.-Z.; Ma, L.-P.; Kang, X.-D.; Wang, P. Int. J.
Hydrogen Energy 2008, 33, 5611.
(12) Bosenberg, U.; Doppiu, S.; Mosegaard, L.; Barkhordarian, G.;
Eigen, N.; Borgschulte, A.; Jensen, T. R.; Cerenius, Y.; Gutfleisch, O.;
Klassen, T.; Dornheim, M.; Bormann, R. Acta Mater. 2007, 55, 3951.
(13) Mosegaard, L.; Moller, B.; Jorgensen, J. E.; Filinchuk, Y.;
Cerenius, Y.; Hanson, J. C.; Dimasi, E.; Besenbacher, F.; Jensen, T. R.
J. Phys. Chem. C 2008, 112, 1299.
(14) Price, T. E. C.; Grant, D. M.; Telepeni, I.; Yu, X. B.; Walker,
G. S. J. Alloys Compd. 2009, 472, 559.
(15) Price, T. E. C.; Grant, D. M.; Walker, G. S. In Materials
Innovations in an Emerging Hydrogen Economy; George, G.; Wicks,
J. S., Eds.; John Wiley & Sons Ltd,: Hoboken, New Jersey, USA,
2009; p 97.
(16) B€osenberg, U.; Kim, J. W.; Gosslar, D.; Eigen, N.; Jensen, T. R.;
von Colbe, J. M. B.; Zhou, Y.; Dahms, M.; Kim, D. H.; G€unther, R.; Cho,
Y. W.; Oh, K. H.; Klassen, T.; Bormann, R.; Dornheim, M. Acta Mater.
2010, 58, 3381.
(17) In, S. R. Vacuum 2009, 84, 689.
(18) Cerenius, Y.; Stahl, K.; Svensson, L. A.; Ursby, T.; Oskarsson,
A.; Albertsson, J.; Liljas, A. J. Synchrotron Radiat. 2000, 7, 203.
(19) Jensen, T. R.; Nielsen, T. K.; Filinchuk, Y.; Jørgensen, J. E.;
Cerenius, Y.; Gray, E.; Webb, C. J. Appl. Crystallogr. 2010, 43, 1456.
(20) Hansen, T. C.; Henry, P. F.; Fischer, H. E.; Torregrossa, J.;
Convert, P. Meas. Sci. Technol. 2008, 034001.
(21) Butts, D. A.; Gale, W. F.; Totemeier, T. C. In Smithells Metals
Reference Book, 8th ed.; Butterworth-Heinemann: Oxford, 2004; p 1.
(22) Munter, A. Neutron News 1992, 3, 29.
(23) Sears, V. Neutron News 1992, 3, 26.
(24) Lide, D. CRC Handbook of Chemistry and Physics, 88th ed.;
CRC: Boca Raton, FL, 2007.
(25) Liang, C. P.; Gong, H. R. J. Alloys Compd. 2009, 489, 130.
13538
dx.doi.org/10.1021/ja204381n |J. Am. Chem. Soc. 2011, 133, 13534–13538