The decomposition pathways for LiBD4-MgD2 multicomponent systems investigated by in situ neutron diffraction

Article abstract of DOI:10.1016/j.jallcom.2008.05.030

Complex hydride-binary hydride multicomponent hydrogen storage systems offer great potential for practical hydrogen stores because their dehydrogenation thermodynamics can be tailored through a destabilisation mechanism whilst maintaining a high hydrogen

Full text of DOI:10.1016/j.jallcom.2008.05.030

Journal of Alloys and Compounds 472 (2009) 559–564
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
The decomposition pathways for LiBD4–MgD2 multicomponent systems
investigated by in situ neutron diffraction
T.E.C. Pricea, D.M. Granta, I. Telepenia, X.B. Yub, G.S. Walkera,∗
a School of Mechanical, Materials and Manufacturing Engineering, University of Nottingham, University Park,
Nottingham, Notts NG7 2RD, UK
b Shanghai Institute of Microsystem and Information Technology, Shanghai 200050, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Complex hydride–binary hydride multicomponent hydrogen storage systems offer great potential for
practical hydrogen stores because their dehydrogenation thermodynamics can be tailored through a
destabilisation mechanism whilst maintaining a high hydrogen storage capacity. One promising system
is based on lithium tetrahydridoborate and magnesium hydride which in a 0.3:1 molar ratio has a storage
capacity of 9.8 wt.%. In situ neutron diffraction of the LiBD4–MgD2 in a 0.3:1 molar ratio was performed on
samples under both a dynamic vacuum and under a self-generated deuterium atmosphere to elucidate
the reaction pathway under each condition. Under vacuum the decomposition occurred in three steps:
firstly the MgD2 started to decompose forming Mg at 300 ◦C, secondly the Mg catalysed decomposition
of LiBD4 starting at the relatively low temperature of 350 ◦C, forming LiD which subsequently reacted
with the Mg to form the alloys Mg0.70Li0.30 and Mg0.816Li0.184 in step 3. Under a self-generated deuterium
atmosphere the MgD2 decomposed in a similar manner, evolving a pressure in excess of 30 bar, the rate
of decomposition was limited by the equilibrium pressure of the system, which itself was dependent on
the increasing temperature; hence MgD2 decomposition occurred over a larger temperature range. The
evolved deuterium pressure was greater than the equilibrium pressure of LiBD4 at experimental temper-
atures (ca. 730 ◦C) thereby preventing decomposition without destabilisation. Only on heating past the
melting point of magnesium (650 ◦C) and subsequent cooling and re-crystallisation was any LiD observed.
This suggests that the destabilisation decomposition mechanism became the predominant reaction in the
molten phase. We postulate that under controllable pressure regimes below the plateau pressure the MgD2
would destabilise the LiBD4 and offer a reversible high capacity storage material with two different cycling
routes depending on the partial pressure of deuterium.
Received 10 March 2008
Received in revised form 29 April 2008
Accepted 11 May 2008
Available online 26 June 2008
Keywords:
Fuel cells
Metal hydrides
Phase transitions
Neutron diffraction
Thermal analysis
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
Recent studies have focused on low-Z complex hydrides such as
alanates, amides, and borohydrides which have high gravimetric
The success of hydrogen fuel cells in automotive applications is
dependent on a means of compact on-board storage of hydrogen.
Safe, mobile hydrogen storage is a major challenge for this to be
achieved. Currently the specification laid out by the U.S. Depart-
ment of Energy (DOE) is for a mass storage capacity of 9 wt.% and
volumetric capacity of 81 kg H2 m−3 by 2015 [1]. Storage solutions
such as pressurized gas and liquid hydrogen storage have been
explored but the safety and capacity of these systems is limited
and will not meet the 2015 targets. Hydrogen storage in metal
hydrides has proved a popular field due to the high volumetric and
gravimetric storage of these materials.
and volumetric storage capacities [2–6]. By analyzing the prop-
erties of these systems and the stages of dehydrogenation with
specific knowledge of the chemical changes occurring, dehydro-
genation steps can be identified, leading to new reversible storage
materials being developed.
Systems based on lithium tetrahydridoborate are of interest
because of its large hydrogen capacity of 18.5 wt.%. However, the
system requires a large energy input to release the stored hydrogen
(ꢀH = 66.9 kJ mol−1 H2 [7]). Work into the catalysis of the system
by Au and Jurgensen using oxide and chloride catalysts has yielded
capacities of 9 wt.% [8] with decomposition beginning at 200 ◦C.
Zu¨ ttel et al. have looked at the effect of large quantities of SiO2
(75%) yielding 3.5 mass% hydrogen release for the whole system
starting at 200 ◦C [7]. LiBH4 is a highly reactive compound [9] and
this reactivity offers the potential for destabilisation of LiBH4.
∗
Corresponding author. Tel.: +44 115 9513752; fax: +44 115 9513800.
E-mail address: gavin.walker@nottingham.ac.uk (G.S. Walker).
0925-8388/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2008.05.030

560
T.E.C. Price et al. / Journal of Alloys and Compounds 472 (2009) 559–564
Work by Vajo et al. into lithium tetrahydridoborate destabilised
with magnesium hydride in a 2:1 ratio has a theoretical capacity
of 11.4 wt.% according to Eq. (1) with an experimental capacity in
the range of 8–10 wt.% being achieved [10]. By utilising a hydride
as a destabilising agent, the system maintains a high overall hydro-
gen storage capacity. The advantage of magnesium hydride is that
it has a reasonable theoretical capacity of 7.6 wt.% and we have
shown for other compositions of LiBH4–MgH2 that not only can the
Mg destabilise the LiBH4 but also the LiBH4 can reduce the MgH2
dehydrogenation temperature [11].
Bosenberg et al. have shown that during decomposition under
a hydrogen pressure, the MgH2 decomposes with the magnesium
component subsequently forming MgB2 with the boron from the
LiBH4 [12]. Vajo et al. found that dehydrogenation under vac-
uum did not result in any MgB2 formation, only formation of Mg
metal [10]. Because they were unable to rehydrogenate these prod-
ucts they suggested it was the presence of MgB2 which enabled a
reversible reaction shown by the following equation:
Fig. 1. DTA–TG data for the as prepared magnesium deuteride sample, run under
an Ar carrier gas with a heating rate of 10 ◦C min−1
.
2LiBH4 + MgH2 → 2LiH + MgB2 + 4H2
(1)
For neutron diffraction, ND, studies the samples were analysed at the ILL
facility in Grenoble, France using the D20 neutron diffractometer (ꢁ 2.4 A, flux
˚
Our work has sought to further understand LiBH4–MgH2 and
the specific chemical changes during each dehydrogenation stage,
specifically looking at the role of magnesium hydride during the
destabilisation reaction. Increasing the MgH2 content will reduce
the dehydrogenation temperatures for both starting phases [11] but
this of course leads to a decrease in the overall hydrogen capacity
of the system.
108 ns−1 cm−2). Data were analysed using Large Array Manipulation Program (LAMP)
version 6. Samples of 1 g mass were loaded into stainless steel 316 L vessels under
an inert gas (Ar). An internal temperature calibration of a blank sample vessel was
used to calibrate any variance from the in situ sample temperatures measured on the
outside of the vessel during the neutron diffraction experiments. Two samples were
analysed for each pressure condition, dynamic vacuum and in a sealed evacuated
vessel.
A 0.3:1 molar ratio is a suitable compromise yielding a mul-
ticomponent system with a hydrogen capacity of 9.8 wt.%. Our
initial ex situ X-ray diffraction (XRD) experiments showed alter-
native steps during dehydriding to that reported by Vajo et al.,
it was shown that for dehydrogenation under vacuum the MgH2
first decomposes to Mg and H2, and subsequently lithium from the
decomposition of LiBH4 diffuses into the magnesium forming firstly
a HCP ␣-phase alloy, Mg0.814Li0.186, then as the lithium content
increases the BCC ␤-phase alloy, Mg0.70Li0.30, is formed, creating
a two phase (␣ + ␤) composition [11]. This reaction is reversible,
and most significantly reversible through alternative end products
not relying on the presence of MgB2. However, there is still uncer-
tainty whether LiH is formed during the dehydrogenation and the
exact order of events during decomposition.
The work presented here uses in situ neutron diffraction (ND) to
elucidate the reaction pathway for the decomposition under vac-
uum of the LiBD4–MgD2 (0.3:1 ratio) system. This is also the first
report of the decomposition of 0.3:1 system under a self-generated
atmosphere, with the intention to investigate whether reaction
(1) proceeds for these non-stoichiometric compositions, forming
MgB2.
3. Results
In order to calculate the thermodynamic properties for
the destabilised system the enthalpy changes for the LiBH4
orthorhombic–hexagonal phase transition and fusion of the
hexagonal phase were measured by DSC and found to be
4.38 0.04 kJ mol−1 and 8.03 0.07 kJ mol−1
.
The DTA–TG of the as prepared MgD2 in Fig. 1 shows the weight
loss and enthalpy data associated with the decomposition of the
MgD2. The MS analysis of the evolved gases only detected deu-
terium (m/z 4) with the intensity of this signal corresponding with
the weight loss recorded by TG analysis. The weight loss associated
with MgD2 decomposition was 10.8% compared to 14.2% maximum
theoretical capacity. XRD of the sample shown in Fig. 2 shows a
sharp diffraction pattern which corresponded to a ␤-MgD2 phase
with only very weak patterns for MgO and Mg. Reduced capacity
2. Experimental
Materials used were LiBH4 (Acros Organics, 95%), LiBD4 (Katchem, 98%) and
MgH2 (Alfar Aesar, 98%) which was ball milled and subsequently deuterided (99.8%,
D2, BOC) to form MgD2 through multiple deuteriding cycles. All handling procedures
were conducted under an inert atmosphere. 1.5 g mixtures of LiBD4–MgD2, with a Li
to Mg molar ratio of 0.3:1, were mechanically milled for 1 h under Ar gas at 300 rpm
using a Fritsch Rotary P5 ball mill.
Samples were characterised by differential thermal analysis, thermogravime-
try coupled with mass spectrometry (DTA–TG–MS, TA instruments STD 600 with
a Hiden Analytical HPR-20 mass spectrometer) using a heating rate of 10 ◦C min−1
under1 atm argonand apurgerateof100 cm3 min−1. Differentialscanning calorime-
try (DSC) was performed using a TA Q10 DSC with a 5 ◦C min−1 ramp rate under an
argon atmosphere; the samples were in hermitically sealed aluminium pans. Pow-
der XRD was performed on a Bruker D8 Advance, step size was 0.02 and dwell time
˚
0.5 s, a copper source with wavelength 1.542 A was used. Powders were analysed on
a Si single crystal wafer, covered by an amorphous polymer tape to avoid oxidation
during the measurement.
Fig. 2. XRD plot of as prepared MgD2.

T.E.C. Price et al. / Journal of Alloys and Compounds 472 (2009) 559–564
561
Fig. 3. DTA–TG data for the as prepared LiBD4–MgD2 sample (0.3:1 ratio), ball milled
for 1 h. Run under an Ar carrier gas with a heating rate of 10 ◦C min−1
.
is principally due to incomplete deuteriding, i.e. MgD2 + Mg. This
partial deuteriding was taken into account in calculating the mass
of each phase for the multicomponent samples.
The DTA–TG data collected for the ball-milled LiBD4–MgD2
(0.3:1 ratio) sample is shown in Fig. 3. Endotherms are visible at
125 ◦C and 279 ◦C before three weight loss stages. The first weight
loss of 2.1% occurred with an endotherm at 335 ◦C, the second
weight loss of 5.1% occurred with an endotherm at 359 ◦C, the last
weight loss, 3.2%, occurred with an endothermic peak at 428 ◦C. A
steady increase in sample weight occurred from 450 ◦C to 600 ◦C.
A final endotherm is visible at 587 ◦C. The associated MS data col-
lected during the DTA–TG experiments showed no m/z 2 evolved
from the sample. All weight loss was therefore attributed to deu-
terium (m/z 4).
Neutron diffraction data is shown in Fig. 4. The sample under
vacuum conditions shows diffractions patterns of the starting
phases (the low temperature orthorhombic LiBD4 phase and MgD2
phases) and from the sample vessel (316 L stainless steel).
The first change occurred during the initial heating ramp to
150 ◦C as the orthorhombic LiBD4 phase disappeared and the higher
temperature hexagonal phase appeared. This pattern then disap-
peared at approximately 260 ◦C, as the temperature passed through
the melting point of LiBD4. The magnesium deuteride diffraction
line can be clearly charted and its disappearance at 300 ◦C coincides
with the appearance of a magnesium pattern (2ꢂ = 50.9◦, 54.6◦,
58.4◦). Following the appearance of a magnesium phase, a lithium
deuteride diffraction line (2ꢂ = 60.6◦) appeared at approximately
360 ◦C, and subsequently disappeared by 450 ◦C coinciding with
a shift in the magnesium lines 2ꢂ = 54.6◦, 2ꢂ = 58.4◦ to higher 2ꢂ
values of 55.3◦ and 58.6◦. This is more clearly seen in Fig. 5.
The shifted diffraction line positions matched those of the ␣-
alloy, Mg0.816Li0.184 found by Yu et al. via ex situ XRD [11]. As the
temperature rose, a diffraction line at 2ꢂ = 56.7◦ appeared, this
matched that of the ␤-alloy, Mg0.70Li0.30, also found by Yu et al.
On heating up to 730 ◦C all the diffraction patterns due to the sam-
ple were lost as the temperature passed through the melting point
of the alloys at 588 ◦C. On cooling, diffraction lines for both of the
Mg–Li alloys re-appeared.
Fig. 4. In situ ND results for LiBD4–MgD2 sample (0.3:1 ratio) ball milled for 1 h and
heated in a sealed vessel, i.e. under a self-generated atmosphere, (top) and under a
dynamic vacuum (bottom).
Mg and no LiD diffraction pattern appeared. On heating to 730 ◦C
the Mg diffraction patterns disappeared as the temperature passed
through the melting point of Mg at 650 ◦C. On cooling to 330 ◦C
a lithium deuteride phase was identified (2ꢂ = 59.9◦) along with
magnesium and magnesium deuteride.
The sample was then heated up to 430 ◦C where upon the mag-
nesium deuteride phase was lost but those of magnesium and
lithium deuteride were retained. On cooling to 330 ◦C the magne-
sium deuteride diffraction pattern re-appeared.
The diffraction data from the sample within a closed vessel is
shown in Fig. 4. The initial phases and phase changes are similar
to that of the sample under vacuum, the orthorhombic–hexagonal
phase transition followed by the melting of LiBD4 and the loss of
MgD2 and formation of Mg (ca. 340 ◦C). However, unlike the sample
under vacuum, no changes were observed in the line positions for
Fig. 5. In situ ND results for LiBD4–MgD2 sample (0.3:1 ratio) ball milled for 1 h and
heated under a dynamic vacuum.

562
T.E.C. Price et al. / Journal of Alloys and Compounds 472 (2009) 559–564
4. Discussion
The intensity of the ␤-alloy line compared with that for the ␣-
alloy was larger than expected from the phase diagram, the 0.3:1
ratio should yield equal proportions of both phases. The difference
in intensity indicated a greater amount of the lithium rich ␤-alloy
than expected. This could be due to magnesium from the alloys
reacting with boron from the decomposed LiBD4, forming MgB2,
reducing the quantity of available magnesium. Thus leading to a
larger amount of the lithium rich ␤-alloy being formed. This agrees
with the loss of the ␣-alloy pattern by 545 ◦C which corresponds
with the high temperatures required for the formation of MgB2
found by Yu et al. [11]. However, no MgB2 phase was identified
which could be due to the fact that in this experiment the boron was
not isotopically pure leading to higher background levels and lower
intensities for neutron diffraction from boron containing phases.
The reaction path for the decomposition occurred over 4 steps
as shown in Eqs. (2)–(5):
The first feature at 125 ◦C in the DTA–TG of the as prepared
LiBD4–MgD2 shown in Fig. 3 was a small endotherm associated
with the phase change of low temperature orthorhombic LiBD4 to
high temperature hexagonal LiBD4, this temperature agrees with
the literature: 108–112 ◦C [7]. The second feature was an endotherm
at 279 ◦C associated with the melting of LiBD4, agreeing with the
literature value of 268–286 ◦C [7].
The decomposition associated with the MgD2 resulted in two
large endotherms with weight losses of 2.1% at 335 ◦C and 5.1%
at 359 ◦C. The fact that a single endotherm was detected for the
decomposition of the as prepared MgD2 by DTA–TG (Fig. 1) indi-
cates that this feature was caused during milling with LiBD4 rather
than being due to a property of the as prepared MgD2. The next
endotherm occurred at 428 ◦C and was accompanied by a weight
loss of 3.2%, this weight loss was due to the first decomposition of
LiBD4 as shown in Eq. (3). The weight increases as the temperature
rose was unfortunately due to a small air contamination resulting
in some oxidation of the sample. Finally at 587 ◦C there was an
endotherm as the Li–Mg alloy melts (cf. 588 ◦C [13]).
The neutron diffraction data in Fig. 4 can be analysed in con-
junction with the enthalpy and weight loss data shown in the
DTA–TG results. The starting materials of LiBD4 and MgD2 were
clearly identified at the start of the experiment. The first phase
change was the LiBD4 orthorhombic to hexagonal phase transition
which can be seen during the initial temperature ramp to 150 ◦C
and correlates with the endotherm at 125 ◦C on the DTA data. The
next feature was the disappearance of the LiBD4 pattern at approxi-
mately 260 ◦C due to the melting of the material, which agrees with
the DTA–TG temperature of 279 ◦C. At 300 ◦C the magnesium deu-
teride decomposed as identified by the loss of the MgD2 diffraction
pattern and the appearance of a Mg diffraction pattern, agreeing
with the endotherms seen at 335 ◦C and 359 ◦C on the DTA–TG
results.
Step 1
◦
MgD2∼−30→0 C Mg + D2
(2)
(3)
Step 2
◦
∼400 C,Mg
0.3LIBD4 −→
0.3LiD + 0.3B + 0.45D2
Step 3
◦
Mg + 0.3LiD −→
∼400 C 0.78Mg0.816Li0.184 + 0.52Mg0.70Li0.30 + 0.15D2
(4)
Step 4
0.78Mg0.816Li0.184 + 0.52Mg0.70Li0.30 + 0.30B
◦
>−45→0 C 0.39Mg0.816Li0.184 + 0.76Mg0.70Li0.30 + 0.15MgB2
(5)
The appearance of a LiD phase at 360 ◦C (diffraction line at a
2ꢂ = 60.6◦) obviously indicates the decomposition of LiBD4. This
binary deuteride was not retained, the intensity for this reflection
reached a maximum and then subsequently diminished until it was
lost altogether by 425 ◦C. The formation of LiD agrees with DTA–TG
data which shows a weight loss and endotherm starting at 360 ◦C.
The loss of the LiD phase was accompanied by a shift to higher 2ꢂ
values for the magnesium pattern, the most notable shift was for the
(0 0 2) reflection at a 2ꢂ value of ca. 55◦. This decrease in d-spacing
was due to the lithium from the decomposition of LiD diffusing into
the magnesium forming a Mg–Li ␣-alloy, as more lithium diffused
into the magnesium the composition reached the solubility limit
for the ␣-phase of Mg0.816Li0.184. As more lithium diffused into the
magnesium the system moved into the two-phase region of the
equilibrium phase diagram forming the ␤-alloy with a composi-
tion of Mg0.70Li0.30 alongside the ␣-alloy. As the intensity for the
␤-alloy pattern increased, that of the ␣-alloy reduced and seemed
to be lost by 545 ◦C. The loss of all neutron diffraction lines at ca.
600 ◦C was due to the Mg alloy melting, which correlates with the
endotherm at 587 ◦C shown in the DTA–TG data. On cooling below
600 ◦C, both ␣- and ␤-alloys solidified, althoughonlyaweakdiffrac-
tion was identified for the ␣-alloy with only one line identifiable. As
the material cooled, the diffraction lines shifted◦to their room tem-
with the reaction for step 3 occurring before the completion of step
2. Because steps 2 and 3 do not appear to be concerted reactions but
are sequential, this means that the low decomposition temperature
for the LiBD4 is due to a catalytic effect from the Mg, rather than
the Mg acting as a destabilising agent for LiBD4. The Mg is however
acting as a destabilising agent for LiD, forming Mg–Li alloys. The
overall reaction by 550 ◦C was
MgD2 + 0.30LiBD4 → 0.39Mg0.7Li0.3 + 0.76Mg0.816Li0.184
+ 0.15MgB2 + 1.60D2.
(6)
The neutron diffraction data for the sample heated in a closed
vessel showed the same initial phase progressions as the sample
under vacuum, with the LiBD4 phase transformation and sub-
sequent melting. The MgD2 decomposition started at a higher
temperature (ca. 340 ◦C) accompanied by the formation of Mg,
but this reaction was limited by the increasing pressure. Complete
decomposition of the MgD2 would result in a pressure in excess of
30 bar, which would require a temperature in excess of 400 ◦C to
complete the reaction, this agrees with the ND data which shows
MgD2 did not completely decompose until 425 ◦C, a much higher
temperature than needed for the sample under vacuum.
Pressure was also the explanation for the lack of LiBD4 decom-
position as no evidence of LiD was detected whilst heating up. To
understand this better one needs to consider the thermodynamics
of the LiBH4–MgH2 system. In calculating the changes in enthalpy
and entropy for LiBH4 dehydrogenations, many authors have used
the thermodynamic data for the low temperature orthorhom-
bic form of LiBH4. However, given the temperature for a 1 bar
◦
˚
˚
perature values (˛: 59.9 , d = 2.4336 A; ˇ: 57.7 , d = 2.4869 A [14].
The reappearance of the ␣-alloy pattern is most likely explained
by the re-crystallisation after melting of the alloys which led to a
more well-defined crystalline ␣-alloy, i.e. it was believed that the
␣-alloy was present before melting, but it was either too dispersed
and/or had insufficient long range order to give any identifiable ND
pattern.

T.E.C. Price et al. / Journal of Alloys and Compounds 472 (2009) 559–564
563
Fig. 7. A van’t Hoff plot showing the expected equilibrium pressures for the dehydro-
genation of LiBH4 in both its orthorhombic and molten states (-o and -l suffix used to
identify the respective state), MgH2 and the destabilised system LiBH4–MgH2 with
a 2:1 molar ratio. The experimental data reported by Vajo et al. [18] are also plotted
with a line of best fit for comparison.
whilst heating up with MgD2 in a sealed vessel. After the MgD2
decomposed the deuterium pressure would be >30 bar and it can be
seen from Fig. 7 that the LiBD4 equilibrium pressure is much lower
than 30 bar for the temperatures investigated, at the melting point
of Mg (650 ◦C) it would have been 7.4 bar. Through this tempera-
ture programmed experiment, the decomposition must be under
kinetic control, for if it were in equilibrium the thermodynamics
for the destabilised reaction as illustrated in Fig. 7 shows that the
equilibrium pressure for the destabilised reaction is in excess of
100 bar and one would expect decomposition of the LiBD4. Even
extrapolating the experimental data from Vajo et al. [18] indicates
that the destabilised reaction has an equilibrium pressure in excess
of 30 bar at temperatures above 450 ◦C.
On cooling (having melted the Mg phase), a LiD phase formed
immediately from the melt along with Mg metal. The sample had
been taken to 730 ◦C, but this is still not high enough to decompose
the LiBD4(l), which has an equilibrium pressure just above 13.6 bar
at this temperature. As has already been stated, the decomposition
through the destabilised reaction is possible. Hence it would sug-
gest that when the Mg was in the molten state, the reaction was no
longer kinetically limited, enabling the decomposition of the LiBD4
through the destabilised reaction shown in Eq. (1). This is a reason-
able hypothesis as mass transport effects will be greatly diminished
when both reactants are in a mobile liquid state.
Fig. 6. Change in enthalpy for the LiBH4–MgH2 system (not to scale).
plateau pressure, T (1 bar), for the LiBH4–MgH2 destabilised sys-
tem is above the orthorhombic to hexagonal phase transformation,
this enthalpy change needs to be taken into account. However,
the reactions are often being investigated at temperatures above
the melting point of LiBH4 and hence this thermodynamic event
should also be taken into account. In order to calculate thermody-
namic data for the LiBH4–MgH2 destabilised reaction, the enthalpy
for the LiBH4 phase transformation and fusion were measured.
The change in enthalpy measured for the phase transformation
(4.38 kJ mol−1) and the enthalpy of fusion (8.03 kJ mol−1) are in
good agreement with values given in a review by Zu¨ ttel et al.
[15] (4.18 kJ mol−1 and 7.56 kJ mol−1, respectively). The change
in entropy was calculated using the temperatures for these two
events (110 ◦C and 279 ◦C, respectively) and were found to be
11.4 J K−1 mol−1 and 14.5 J K−1 mol−1, respectively. Using standard
thermodynamic database information [16,17] the enthalpy changes
for the destabilised reaction were calculated and are given in Fig. 6.
A van’t Hoff plot can be constructed using this data to show the
effect the solid and liquid state of the LiBH4 has on the equilibrium
pressure for the dehydrogenation. Data is presented in Fig. 7 for the
dehydrogenation of hexagonal-LiBH4, molten-LiBH4 and the desta-
bilised LiBH4–MgH2 system for both solid and liquid LiBH4 states.
Also plotted for comparison is the data from Vajo et al. [10] and for
MgH2. The enthalpy and T (1 bar) data are given in Table 1.
The intention was to investigate redeuteriding of the material at
lower temperatures, using the self-generated deuterium pressure.
However, on cooling to 330 ◦C only a small amount of MgD2 was
formed even though thermodynamically one would expect most
of the Mg to react. This is likely to be because after melting, the
sample will no longer be a fine powder but have a very low surface
area and a large grain size limiting the rate of deuteriding. There
was no noticeable difference in the ND intensity for LiD during this
low temperature cycling, presumably because this phase was now
The van’t Hoff plot clearly shows that the equilibrium pressure
is lower for LiBH4 in the molten state compared to the solid state.
This data can be used to explain why the LiBD4 did not decompose
Table 1
Calculated enthalpy and entropy data plus the corresponding T (1 bar) values for the dehydrogenation of LiBH4, MgH2 and mixtures of both hydrides, forming LiH and
accompanying products
Dehydrogenation
ꢀH298 (kJ mol−1(H2))
ꢀS298 (J K−1 mol−1(H2))
T (1 bar) (◦C)
2LiBH4(ortho) → 2LiH + 2B + 3H2
2LiBH4(l) → 2LiH + 2B + 3H2
66.9
58.6
75.3
46.0
43.8
39.8
97.3
80.1
132.3
104.0
98.3
414
459
296
169
173
164
MgH2 → Mg + H2
2LiBH4(ortho) + MgH2 → 2LiH + MgB2 + 4H2
2LiBH4(hex) + MgH2 → 2LiH + MgB2 + 4H2
2LiBH4(l) + MgH2 → 2LiH + MgB2 + 4H2
91.0

564
T.E.C. Price et al. / Journal of Alloys and Compounds 472 (2009) 559–564
precipitates within a Mg matrix and therefore physically separated
from the boron, preventing any reaction to form LiBD4.
Hence, for the sample under a self-generated atmosphere, the
reactions that occurred were
decomposition could continue through the destabilisation mecha-
nism limited until then by kinetics. No alloy formation was detected
because the deuterium pressure stabilised the LiD phase. Both sets
of conditions (under vacuum and under deuterium pressure) show
potential for cycling. By controlling the partial pressure both routes
may allow reversibility which will be reported at a later date.
Step 1
◦
MgD2>−30→0 CMg + D2
(7)
(8)
Acknowledgements
Step 2
Funding and support from EPSRC Supergen UKSHEC. We would
also like to thank Dr. Bachir Ouladdiaf and Dr. Vincent Legrand for
their expertise and for time on the D20 neutron diffractometer at
Institut Laue-Langevin., ILL expt number 5-25-129.
◦
0.3LiBD4 + 0.15Mg>−65→0 C0.3LiD + 0.15MgB2 + 0.45D2
These in situ experiments have helped identify the reaction path
for the decomposition under vacuum or an inert carrier gas, proving
the dual role of the Mg as a catalyst for the decomposition of LiBD4
and as a destabilising agent for LiD. Under these conditions all the
deuterium can be liberated, utilising the full capacity of the system
and our earlier work showed this to be a reversible system [11].
More work needs to be done for the decomposition of this system
under deuterium pressures, but it has been shown that the presence
of deuterium alters the reaction pathway, halting the decomposi-
tion at the formation of LiD, progressing via a similar reaction to
that first reported by Vajo et al. [10]. The reversibility still needs to
be investigated, but one would expect this to be comparable to that
found for the 2:1 molar ratio [10,12].
References
[1] DOE Hydrogen Storage Targets, 2007, Acc 05-02-2008. http://www.eere.
energy.gov/hydrogenandfuelcells/mypp/pdfs/storage.pdf.
[2] M. Aoki, T. Noritake, G. Kitahara, Y. Nakamori, S. Towata, S. Orimo, J. Alloys
Compd. 428 (2006) 307–311.
[3] P.A. Chater, W.I.F. David, S.R. Johnson, P.P. Edwards, P.A. Anderson, Chem. Com-
mun. (2006) 2439–2441.
[4] Z.P. Li, B.H. Liu, K. Arai, S. Suda, J. Electrochem. Soc. 150 (2003) A868–A872.
[5] S. Orimo, Y. Nakamori, A. Zuttel, Mater. Sci. Eng. B 108 (2004) 51–53.
[6] S.-I. Orimo, Y. Nakamori, N. Ohba, K. Miwa, M. Aoki, S.-I. Towata, A. Zuttel, Appl.
Phys. Lett. 89 (2006) 021920–21923.
[7] A. Zuttel, S. Rentsch, P. Fischer, P. Wenger, P. Sudan, P. Mauron, C. Emmenegger,
J. Alloys Compd. 356–357 (2003) 515–520.
[8] M. Au, A. Jurgensen, J. Phys. Chem. B 110 (2006) 7062–7067.
[9] L. Mosegaard, B. Moller, J.E. Jorgensen, Y. Filinchuk, Y. Cerenius, J.C. Hanson, E.
Dimasi, F. Besenbacher, T.R. Jensen, J. Phys. Chem. C 112 (2008) 1299–1303.
[10] J.J. Vajo, S.L. Skeith, F. Mertens, J. Phys. Chem. B 109 (2005) 3719–3722.
[11] X.B. Yu, D.M. Grant, G.S. Walker, Chem. Commun. (2006) 3906–3908.
[12] U. Bosenberg, S. Doppiu, L. Mosegaard, G. Barkhordarian, N. Eigen, A.
Borgschulte, T.R. Jensen, Y. Cerenius, O. Gutfleisch, T. Klassen, M. Dornheim,
R. Bormann, Acta Mater. 55 (2007) 3951–3958.
[13] E.A. Brandes, G.B. Brook (Eds.), Smithells Metals Reference Book, 7th ed., But-
terworth Heinemann, London, 1992, pp. 344.
[14] F.H. Herbstein, B.L. Averbach, Acta Metall. 4 (1956) 407–413.
[15] A. Zuttel, P. Wenger, S. Rentsch, P. Sudan, P. Mauron, C. Emmenegger, J. Power
Sources 118 (2003) 1–7.
5. Conclusions
In situ neutron diffraction analysis of LiBD4–MgD2 has shown
that under vacuum the presence of Mg, formed from the decom-
position of MgD2, catalysed the decomposition of LiBD4 forming
LiD, B and D2. The presence of Mg subsequently destabilised the
LiD through a concerted reaction which resulted in the formation
of two alloys Mg0.816Li0.184 and Mg0.70Li0.30. The formation of these
two alloys rather than the MgB2 phase suggests a different progres-
sion of destabilisation under vacuum conditions than that proposed
by Vajo et al. For the experiment undertaken in a sealed vessel, the
deuterium pressure generated was too high (>30 bar) to allow LiBD4
decomposition until the system was in the molten phase and the
[16] M.W. Chase, NIST-JANAF Thermochemical Tables, 4th ed., American Institute of
Physics, New York, 1998.
[17] D.R. Lide, CRC Handbook of Chemistry and Physics, CRC press, London, 2002.
[18] J.J. Vajo, T.T. Salguero, A.F. Gross, S.L. Skeith, G.L. Olson, J. Alloys Compd. 446–447
(2007) 409–414.