Reversible hydrogen storage in the lithium borohydride-calcium hydride coupled system

Article abstract of DOI:10.1016/j.jallcom.2007.09.125

We report large reversible hydrogen storage in a new coupled system, LiBH₄/CaH₂, via the reaction 6LiBH₄ + CaH₂ ? 6LiH + CaB₆ + 10H₂ having a theoretical hydrogen capacity of 11.7 wt% and an estimated reaction enthalpy of ΔH = 59 kJ/mol H₂. Samples that include 0.25 mol (18.2 wt%) TiCl₃ reproducibly store 9.1 wt% hydrogen, corresponding to 95% of the available hydrogen. H₂ is the only evolved gas detected by mass spectrometry. X-ray diffraction confirms that the sample cycles between LiBH₄ and CaH₂ in the hydrogenated state and LiH and CaB₆ in the dehydrogenated state.

Full text of DOI:10.1016/j.jallcom.2007.09.125

Journal of Alloys and Compounds 464 (2008) L1–L4
Letter
Reversible hydrogen storage in the lithium borohydride—calcium
hydride coupled system
∗
F.E. Pinkerton , M.S. Meyer
Materials and Processes Laboratory, General Motors Research and Development Center,
MC 480-106-224, 30500 Mound Road, Warren, MI 48090-9055, United States
Received 1 August 2007; received in revised form 23 September 2007; accepted 29 September 2007
Available online 5 October 2007
Abstract
We report large reversible hydrogen storage in a new coupled system, LiBH
4
/CaH
2
, via the reaction 6LiBH
4
+ CaH
2
↔ 6LiH + CaB
6 2
+ 10H
having a theoretical hydrogen capacity of 11.7 wt% and an estimated reaction enthalpy of ꢀH = 59 kJ/mol H
18.2 wt%) TiCl reproducibly store 9.1 wt% hydrogen, corresponding to 95% of the available hydrogen. H is the only evolved gas detected by
in the hydrogenated state and LiH and CaB in the
2
. Samples that include 0.25 mol
(
3
2
mass spectrometry. X-ray diffraction confirms that the sample cycles between LiBH
dehydrogenated state.
4
and CaH
2
6
©
2007 Elsevier B.V. All rights reserved.
Keywords: Energy storage materials; Hydrogen absorbing materials; Gas–solid reactions; X-ray diffraction
Development of practical on-board hydrogen storage systems
based on solid hydride materials remains a significant technical
and scientific challenge for fuel cell vehicles [1]. In the last few
years, much of the focus in hydride research has been on light
metal alanates [2–6] and borohydrides [7–10] because they offer
relatively large gravimetric hydrogen densities. Built on earlier
work of Reilly and Wiswall [11], Vajo and Olson [12] recently
adopted a promising strategy for generating new hydride storage
systems by coupling existing binary and complex hydrides that
form mixed compounds or alloys upon dehydrogenation. Vajo et
al. reported reversible hydrogen storage exceeding 9 wt% in the
LiBH4/MgH2 system when catalyzed with TiCl3 [13] according
to the reaction:
Here we report synthesis and hydrogen cycling of the new,
coupled complex hydride system LiBH4/CaH2, described by
6
LiBH4 + CaH2 ↔ 6LiH + CaB + 10H2,
(2)
6
with 11.7 wt% theoretical hydrogen capacity. The theoretical
volumetric hydrogen density is also relatively large at 0.09 kg
H /l (materials basis). Reversible hydrogen storage of 9.1 wt%
2
has been achieved in samples that also incorporate 0.042 mol
of TiCl per LiBH ; including the additional TiCl weight, this
3
4
3
represents 95% of the theoretically available hydrogen in the
system. Our results are consistent with the enthalpy of reaction
ꢀH ∼ 60 kJ/mol H recently predicted by Alapati et al. [15] for
2
this system, and with thermodynamic estimates.
Samples of 6LiBH4 + CaH2 + 0.25TiCl3 were prepared by
ball milling. TiCl3 was included in the sample as a catalyst
to increase the kinetics of hydrogen release and reabsorption,
however it is not known at this time whether it actually had
any effect on hydrogen cycling. All sample handling was done
in an Ar inert gas glove box to prevent atmospheric exposure.
Powders of LiBH4 (Aldrich, 95% purity), CaH2 (Aldrich, 95%
purity), and TiCl3 (Aldrich, 99.999% purity) were mixed in a
2
LiBH4 + MgH2 ↔ 2LiH + MgB2 + 4H2
(1)
The formation of MgB2 as a reaction product of LiBH4 and
MgH2 significantly changes the thermodynamic pathway and,
most importantly, greatly improves rehydrogenation to form
LiBH4. Similar effects were observed in the MgH2/Si system
given by 2MgH2 + Si → Mg2Si + 2H2, although rehydrogena-
tion was proven much more difficult [14].
6:1:0.25 molar ratio having a total weight of 1 g, loaded into
a hardened steel ball mill jar along with one large and two
small steel balls, and ball-milled for 1 h in a SPEX 8000 Mixer/
Mill.
∗
Corresponding author. Tel.: +1 586 986 0661; fax: +1 586 986 3091.
E-mail address: frederick.e.pinkerton@gm.com (F.E. Pinkerton).
0
925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.09.125

L2
F.E. Pinkerton, M.S. Meyer / Journal of Alloys and Compounds 464 (2008) L1–L4
X-ray diffraction (XRD) was performed using a Bruker AXS
General Area Detector Diffractometer System (GADDS) using
Cu K␣ radiation. The XRD samples were loaded under Ar into
1
.0 mm diameter quartz capillary tubes and sealed with clay for
transfer to the diffractometer.
Hydrogen cycling was carried out using a Cahn Model 2151
high pressure thermogravimetric analyzer (TGA). The as-milled
powder (217 mg) was loaded into an open stainless steel sam-
ple bucket in the Ar glove box and protected during transfer to
the TGA by covering the sample with anhydrous pentane. The
TGA was purged with He gas while the pentane evaporated. The
sample was cycled three times starting from the hydrogenated
state (left side of reaction (2)). Samples were dehydrogenated
◦
in 0.13 MPa flowing He gas by heating to 450 C at either 5
◦
◦
or 2 C/min, followed by soaking at 450 C until weight loss
stopped. Rehydrogenation was accomplished in 8.3 MPa flow-
ing H2 gas by heating to 400 C at 5 C/min (first hydrogenation)
◦
◦
◦
◦
or 25 C/min (second dehydrogenation) and soaking at 400 C.
For the third hydrogenation, the temperature was increased at
◦
2
2
C/min through a series of constant temperature steps at 250,
◦
75, 300, 350, and finally 400 C. After the first desorption, the
startingweightpercentofeachabsorptionordesorptionscanwas
taken to be the ending weight percent from the previous scan.
The sample was temporarily removed from the TGA under He
gas and a small portion of the sample was extracted for XRD
analysis after the first dehydrogenation, after the subsequent first
rehydrogenation, and at the end of the third cycle. We monitored
the composition of the TGA exhaust gas by residual gas analyzer
(
RGA) mass spectrometry to detect H2 release during dehydro-
genation, and to test for the presence of H2O, CH4, NH3, B2H6,
N2/CO, O2, and CO2.
4 2 3
Fig. 1. X-ray diffraction patterns from 6LiBH + CaH + 0.25TiCl : (a) as-
milled, (b) in the desorbed state, and (c) in the rehydrogenated state after three
cycles.
The XRD pattern of the as-milled powder, shown in Fig. 1(a),
contains the expected LiBH4 and CaH2 phases, indicating that
no reaction between the major phases occurred during milling.
The TiCl3, however, clearly reacted with a small portion of the
LiBH4 to form LiCl. No diffraction peaks corresponding to Ti-
containing species were observed. A small quantity of oxide
impurity was also present.
to be asymptotically increasing toward its starting value. The
absorption was terminated after 550 min with the weight at 99%.
XRD obtained at this stage (not shown) showed only peaks that
could be indexed as LiBH4 and CaH2, along with CaO and a
small quantity of LiCl. The sample was replaced in the TGA
and cycled again, as shown by the dash–dot lines in Fig. 2(a)
and (b). The reabsorption kinetics were somewhat slower in the
second cycle, and the experiment was terminated after 735 min.
A third desorption–reabsorption cycle is shown in Fig. 3. For
this reabsorption experiment, a number of temperatures were
Dehydrogenation and rehydrogenation are shown in
Figs. 2 and 3 for three hydrogen cycles of 6LiBH4 + CaH2 +
0
.25TiCl3. Starting from the as-milled state in Fig. 2(a), the
sample began to lose significant weight in He at temperatures
◦
as low as 150 C. Most of the weight loss, however, occurred
◦
◦
after the sample reached 350 C, and was complete after soak-
tried between 250 and 400 C. Some absorption did occur at
◦
ing at 450 C for ∼30 min. It should be noted that this is above
the lower temperatures, but with poor kinetics. The sample was
◦
◦
the melting temperature of pure LiBH4 (288 C). Only H2 was
held at 400 C until the total heating time reached 2700 min; its
observed in the RGA of the evolved gas to within our detec-
tion limits. The total weight loss was 9.1 wt%. After correcting
for the additional TiCl3 content, the weight loss relative to the
LiBH4/CaH2 content was 11.1 wt%, compared to the theoretical
hydrogen capacity of 11.7 wt%, or 95% of the hydrogen pre-
dicted by reaction (2). After dehydrogenation, XRD confirms
that the original LiBH4 and CaH2 phases have disappeared and
final weight was very close to the original value prior to cycling.
A final XRD pattern obtained after this cycle, Fig. 1(c), clearly
demonstrates that the sample returned to the original mixture of
LiBH4 and CaH2. The CaO lines are somewhat stronger, most
likely due to incidental exposure during sample extraction and
reinsertion in the TGA and impurities in the flowing gas stream
over the long time scale of the experiment. Although the sample
weight gain due to the increased oxide is not known quantita-
tively, and probably accounts for the observation that the third
rehydrogenation shown in Fig. 3(b) appears to be approach-
ing a maximum value slightly greater than 100%, it is clear
in their place are diffraction peaks belonging to CaB and LiH,
6
as shown in Fig. 1(b) [16].
First cycle rehydrogenation is shown as the solid curve in
Fig. 2(b). After several hundred minutes, the weight appears

F.E. Pinkerton, M.S. Meyer / Journal of Alloys and Compounds 464 (2008) L1–L4
L3
Fig. 2. (a) H2 desorption profiles and (b) subsequent absorption profiles for the
first two cycles of 6LiBH4 + CaH2 + 0.25TiCl3. The sample started from the as-
milled state at 100% weight in the upper panel. First cycle data is represented
by solid lines, and the second cycle is shown as dash–dotted lines. The dashed
curves show the corresponding temperature histories; the dash–double dot line
in panel (b) is the temperature history of the second absorption.
Fig. 3. (a) Third desorption and (b) reabsorption cycle for
LiBH4 + CaH2 + 0.25TiCl3.
6
time. Another possibility is the presence of TiCl3 in our sample,
although we stress that the effect of Ti, if any, has not yet been
assessed in the present work.
The thermodynamics of the LiBH4/CaH2 system were esti-
mated from the enthalpy and entropy values for LiBH4,
CaH2, LiH, and CaB6 in the Outokumpu HSC chemistry
thermodynamics database [18]. For reaction (2), HSC chem-
istry gives ꢀH = 66.2 kJ/mol H2 and ꢀS = 100 J/K mol H2 at
room temperature. These values, however, do not include the
orthorhombic-to-hexagonal structural transformation of LiBH4
that the third H2 absorption is very close in magnitude to the
original H2 desorption of 9.1 wt%. Curiously, the LiCl lines
are no longer observed in the hydrogenated state. A similar
phenomenon was observed by Vajo et al. in the LiBH4/MgH2
system catalyzed with TiCl3 [13]. We also note that Barkhordar-
ian et al. [17] recently attempted to hydrogenate a ball-milled
mixture of 6LiH + CaB (starting on the dehydrogenated side
6
◦
of reaction (2)) at 400 C, but were unsuccessful in forming
◦
◦
a hydride even by exposure to 35 MPa H2 gas for 1440 min. In
contrast, we have successfully cycled reaction (2) multiple times
using 8.3 MPa H2 pressure for hydrogenation. It is unclear why
the Barkhordarian et al. sample did not hydrogenate, given that
their experimental conditions appear similar to ours. One differ-
ence is that we began on the hydrogenated side of reaction (2)
at 113 C [19] or melting of the LiBH4 component at 288 C.
Recently, Z u¨ ttel et al. [20] have reported the enthalpies of
the LiBH4 structural phase transition (ꢀHs = 4.18 kJ/mol) and
of melting (ꢀHm = 7.56 kJ/mol); when these values are com-
bined with HSC chemistry the estimated enthalpy of reaction
◦
(2) at 400 C is ꢀH = 59.2 kJ/mol H2. Similar calculation
(
6LiBH4 + CaH2) whereas Barkhordarian et al. started from the
of ꢀS using entropies from HSC chemistry and including
the entropy changes ꢀSs = ꢀHs/Ts and ꢀSm = ꢀHm/Tm for
the structural transformation and melting, respectively, gives
dehydrogenated state (6LiH + CaB ). We also note that Barkhor-
darian et al. milled for 24 h, much longer than our 1 h milling
6

L4
F.E. Pinkerton, M.S. Meyer / Journal of Alloys and Compounds 464 (2008) L1–L4
ꢀ
S = 85.6 J/(K mol H2). The thermodynamic equilibrium tem-
References
◦
perature in 1 bar of H2 gas is predicted to be T(1 bar) = 418 C.
Alapati et al. [15] obtained a similar estimate for the reaction
enthalpy, ꢀH ∼ 60 kJ/mol H2, using ab initio density functional
theory (DFT) calculations (neglecting the enthalpy of melt-
ing). The high temperatures required for dehydrogenation in
the absence of H2 gas, and rehydrogenation in an 8.3 MPa H2
overpressure, are clear indications that the limited kinetics at
lower temperatures dominates hydrogen cycling in these sam-
ples. Experimental determination of the reaction enthalpy and
entropy would clearly be valuable, and such a study is planned
for the near future.
[
1] US Department of Energy—Energy Efficiency and Renewable Energy:
Hydrogen, Fuel Cells and Infrastructure Technologies Program, Hydrogen
storage system requirements (2006) available at: http://www.eere.energy.
gov/hydrogenandfuelcells/storage/current technology.html.
[
2] B. Bogdanovi c´ , M. Schwickardi, J. Alloys Compd. 253/254 (1997) 1.
[3] B. Bogdanovi c´ , R.A. Brand, A. Marjanovic, M. Schwickardi, J. T o¨ lle, J.
Alloys Compd. 302 (2000) 36.
[
[
[
[
4] C.M. Jensen, K. Gross, J. Appl. Phys. A 72 (2001) 213.
5] G. Sandrock, K. Gross, G. Thomas, J. Alloys Compd. 339 (2002) 299.
6] F. Sch u¨ th, B. Bogdanovi c´ , M. Felderhoff, Chem. Commun. (2004) 2249.
7] A. Z u¨ ttel, S. Rentsch, P. Fischer, P. Wenger, P. Sudan, Ph. Mauron, Ch.
Emmenegger, J. Alloys Compd. 356/357 (2003) 515.
[
[
8] Y. Nakamori, S. Orimo, J. Alloys Compd. 370 (2004) 271.
9] S. Orimo, Y. Nakamori, A. Z u¨ ttel, Mater. Sci. Eng. B 108 (2004) 51.
Our results demonstrate that LiBH4/CaH2 forms a coupled
complex hydride system according to reaction (2). By itself,
CaH2 is stable up to 600 C. LiBH4 in isolation can be
[
[
10] F.E. Pinkerton, G.P. Meisner, M.S. Meyer, M.P. Balogh, M.D. Kundrat, J.
Phys. Chem. B 109 (2005) 6.
11] J.J. Reilly, R.H. Wiswall, Inorg. Chem. 6 (1967) 2220.
◦
decomposed to LiH + B + 3/2H2 in the temperature regimes
used here, but reversibility of this reaction has proven extremely
challenging [7]. By coupling the two, we have succeeded in both
destabilizing CaH2 and promoting full reversibility of LiBH4.
The predicted T(1 bar) for this coupled reaction is too high for
its use as an on-board hydrogen storage medium; this system
nevertheless provides another excellent example of the power of
coupled reactions in designing practical hydride-based storage
systems.
[12] J.J. Vajo, G.L. Olson, Scripta Mater. 56 (2007) 829.
[
[
13] J.J. Vajo, S.L. Skeith, F. Mertens, J. Phys. Chem. B 109 (2005) 3719.
14] J.J. Vajo, F. Mertens, C.C. Ahn, R.C. Bowman Jr., B. Fultz, J. Phys. Chem.
B 108 (2004) 13977.
[
[
15] S.V. Alapati, J.K. Johnson, D.S. Sholl, J. Phys. Chem. B 110 (2006) 8769.
16] The GADDS diffraction pattern for this sample consisted of a mixture
of diffraction rings and individual diffraction spots, so the intensities of
the various diffraction peaks do not follow those predicted for a random
powder diffraction pattern. Although no direct morphological examination
was made, the presence of individual diffraction spots suggests that some
material was present in the form of large crystallites. This did not impact
reversibility, but could be a factor in the slow kinetics of this system.
17] G. Barkhordarian, T. Klassen, M. Dornheim, R. Bormann, J. AlloysCompd.
Acknowledgements
[
4
40 (2007) L18.
The authors wish to thank Michael Balogh and Richard Speer
for providing the X-ray diffraction data, John Vajo and Jan
Herbst for valuable discussions, and James Spearot and Mark
Verbrugge for their encouragement and support.
[
[
18] Outokumpu HSC Chemistry, version 5.1 (ChemSW Inc., 2002).
19] J-Ph. Souli e´ , G. Renaudin, R. Cern y´ , K. Yvon, J. Alloys Compd. 200 (2002)
ˇ
346.
[20] A. Z u¨ ttel, A. Borgschulte, S. Orimo, Scripta Mater. 56 (2007) 823.