Thermodynamic destabilisation of MgH2 and NaMgH3 using Group IV elements Si, Ge or Sn

Article abstract of DOI:10.1016/j.jallcom.2014.10.086

The addition of Group IV elements of Si, Ge or Sn to Mg-based hydrides has led to the successful destabilisation of MgH₂ or NaMgH₃, resulting in hydrogen release at lower temperatures. This is the first time a direct comparison has b

Full text of DOI:10.1016/j.jallcom.2014.10.086

Journal of Alloys and Compounds 623 (2015) 109–116
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jalcom
Thermodynamic destabilisation of MgH and NaMgH using Group IV
2
3
elements Si, Ge or Sn
Anna-Lisa Chaudhary a,b, , Mark Paskevicius , Drew A. Sheppard , Craig E. Buckley
⇑
a
a
a
a
Department of Imaging and Applied Physics, Fuel and Energy Technology Institute, Curtin University, Australia
Institute of Materials Research, Materials Technology, Helmholtz-Zentrum Geesthacht, Max-Planck-Strasse 1, D-21502 Geesthacht, Schleswig-Holstein, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 August 2014
Received in revised form 7 October 2014
Accepted 15 October 2014
Available online 23 October 2014
The addition of Group IV elements of Si, Ge or Sn to Mg-based hydrides has led to the successful desta-
bilisation of MgH
direct comparison has been made with all the samples prepared and characterised using identical condi-
tions. Pure MgH desorbs hydrogen at a pressure of 1 bar at 282 °C, a temperature too high for typical
mobile applications. The addition of Group IV metals to MgH causes the formation of intermetallic com-
pounds (Mg Si, Mg Ge and Mg Sn) upon hydrogen release. Theoretical calculations show promising ther-
2 3
or NaMgH , resulting in hydrogen release at lower temperatures. This is the first time a
2
2
2
2
2
Keywords:
modynamic equilibrium conditions for these systems. Experimentally, these conditions were difficult to
achieve, however, hydrogen desorption results show that Ge has the most significant effect in allowing
low temperature hydrogen release, followed by Sn, then Si. It was found that Si also has a beneficial effect
Magnesium hydride
Sodium magnesium hydride
Desorption
Destabilisation
3
on NaMgH , reducing the desorption temperature.
Reaction kinetics
Thermodynamics
Ó 2014 Elsevier B.V. All rights reserved.
1
. Introduction
Hydrogen based energy systems have the capacity to store
namic stability of MgH
2
is more difficult to overcome, where the
thermodynamics are known to be:
S = 133.4 J/mol H /K [16]. At thermodynamic equilibrium these
thermodynamic properties equate to a 1 bar H desorption tem-
perature of 282 °C, too high for typical mobile applications.
2
DH = 74 kJ/mol H and
D
2
clean, sustainable energy now and for future generations. As global
energy demands rise, so do concerns over climate change and
depleting fossil fuel resources [1], the shift towards energy produc-
tion from renewable resources has gained considerable momen-
tum [2,3], Hydrogen storage systems can be combined with fuel
cells for on demand electricity or for use in mobile transport to
address these environmental energy issues [4,5]. Finding an effi-
cient and safe way to store hydrogen is one of the challenges to
hydrogen fuel usage [6,7].
The most promising option being explored is the solid-state
storage of hydrogen in metal hydrides [8–10]. Magnesium is an
attractive choice as it is inexpensive, abundant and has a high
hydrogen storage capacity of 7.7 wt.% hydrogen in the form of
2
MgH2 ! Mg þ H ðgÞ
2
ð1Þ
NaMgH is another viable hydrogen storage compound with a stor-
3
age capacity of 6 wt.% H . NaMgH undergoes a two-step desorption
2
3
process that releases ca. 4 wt.% and 2 wt.% of hydrogen gas in each
step [17,18]. The first step (Reaction (2)) is the most relevant to
practical applications because of its lower operating temperature
and the benefit of restricting molten Na metal formation. Similar
to MgH , NaMgH also contains strongly bound hydrogen [19]
2 3
resulting in high thermodynamic stability. A recent study measured
the thermodynamic properties of the first decomposition step of
2
MgH (Reaction (1)) [11–13]. The drawbacks, however, include
NaMgH
3
as
D
H = 86.6 kJ/mol H
2
and
D
S = 132.2 J/mol H
2
/K, giving
slow reaction kinetics and the strong binding energy between
magnesium and hydrogen [14]. The kinetic issues in the Mg–H sys-
tem have largely been overcome with a range of additives intro-
duced via ball milling, that provide particle size refinement and
enable rapid reaction kinetics [15]. However, the high thermody-
2
an operational temperature of 382 °C at 1 bar H [20].
NaMgH3 ! Mg þ NaH þ H ðgÞ
2
ð2Þ
2 3
Thermodynamic destabilisation of both MgH and NaMgH can be
achieved by introducing another reactive element to allow for the
formation of a more energetically favourable intermetallic upon
hydrogen desorption. Si is a well-known additive that has been used
⇑
Corresponding author at: Institute of Materials Research, Materials Technology,
Helmholz Zentrum Geesthacht, Germany. Tel.: +49 (0) 4152872647.
to destabilize MgH
2
[21–24] (Reaction (3)), significantly reducing
ꢀ1
E-mail address: anna-lisa.chaudhary@hzg.de (A.-L. Chaudhary).
the enthalpy of reaction to 38.9 kJ mol H
2
[24]. This reduction in
http://dx.doi.org/10.1016/j.jallcom.2014.10.086
0
925-8388/Ó 2014 Elsevier B.V. All rights reserved.

1
10
A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
enthalpy results in a theoretical desorption temperature of 25 °C at
bar with a H capacity of 5 wt.% [24]. Attempts to achieve thermo-
Table 2
2
Ge addition to MgH , crystallite size from Rietveld refinement.
1
2
dynamic equilibrium of the Mg–Si–H system experimentally have
been limited by reaction kinetics [22,25]. The destabilisation of
Sample
Phase
Structure
wt.% Cryst.
size (nm)
NaMgH
tem (Reaction (4)), with the desorption temperature of pure NaM-
gH being reduced from 350 to 250 °C [19]. The theoretical
hydrogen capacity of NaMgH is also reduced from 4 wt.% to
.13 wt.% once Si is added.
After Si, the next Group IV elements are Ge and Sn, which also
form intermetallics with Mg [26–28]. Ge has similar properties to
Si and has also been studied as a destabilising additive to MgH
3
has also been achieved with the addition of Si to the sys-
Ge
Ge
Cubic, 227, Fd-3m
Trigonal, 152, P3121
97.4 227 ± 4
2.6 58 ± 8
R
2
R
Mg
R
wp = 8.7% (as supplied) GeO
2
MgH
wp = 9% (cryomilled)
Ge
wp = 6.7% (desorbed)
2
+ Ge
2
b-MgH Tetragonal, 136, P42/mnm 38.1 5.6 ± 0.2
Ge
Mg
Ge
3
Cubic, 227, Fd-3m
Ge Cubic, 225, Fm-3m
Cubic, 227, Fd-3m
61.9 35 ± 1
95.9 73 ± 1
4.1 78 ± 3
3
2
2
3
2
Table 3
Sn addition to MgH
[
29] (Reaction (5)). Adding Ge to the system results in a reduced
2
, crystallite size from Rietveld refinement.
2
theoretical storage capacity of 3.22 wt.% H . Walker et al. [29]
found that the enthalpy of desorption with the addition of Ge
Sample
Phase Structure
wt.% Cryst.
size (nm)
ꢀ1
ꢀ1
was reduced by 60 kJ mol
H
2
to a value of
D
H = 14 kJ mol
2
H .
The Mg–Ge–H system is investigated further herein by directly
comparing changes in dehydrogenation properties with its sister
Si system.
Sn Rwp = 8.1% (as supplied) Sn
Tetragonal, 141, I41/amd 100 547 ± 16
Tetragonal, 136, P42/ 26.4 7.0 ± 0.6
MgH2 mnm
2MgH
2
+ Sn
b-
R
Mg
wp = 6.2% (cryomilled)
Sn
Rwp = 7.8% (desorbed)
Sn
Mg
Sn
Tetragonal, 141, I41/amd 73.6 85 ± 3
Sn can also be added to MgH
molecular mass with the stoichiometric addition of Sn reduces the
hydrogen storage capacity to 2.36 wt.% H . Experimentally, the
addition of Sn to MgH has already proven to be an effective desta-
bilising element for MgH [30–32] however, there has been no
2
(Reaction (6)) and the increase in
2
2
Sn Cubic, 225, Fm-3m
Tetragonal, 141, I41/amd
93.6 147 ± 5
6.4 94 ± 7
2
2
2
The hydrides were mixed in stoichiometric ratios from Reactions (3)–(6), with
Si, Ge or Sn in a SpexSamplePrep 6850 Freezer Mill (USA) at liquid nitrogen temper-
atures (77 K). The powders and stainless steel rod impactor were placed into a
direct comparison made between Si, Ge and Sn with consistent
preparation and characterisation techniques. Therefore, this study
1
4.3 cm3 milling vial constructed from stainless steel and sealed with stainless steel
aims to destabilise hydrogen desorption from MgH
2
using these
end caps. The mill was programmed for a total grinding time of 30 min with a 2 min
cooling interval for each minute of grinding.
Group IV elements with the same mixing process and desorption
experimentation. Destabilisation of NaMgH
gated by the same processes by adding Si.
3
will also be investi-
X-ray diffraction (XRD) measurements were conducted using a D8 Advance
(Bruker, Germany) X-ray diffractometer with a copper anode tube (k = 1.5418 Å)
and LynxEye detector. Scans were taken at a 2h range of 10–100° with a 0.02° step
size and 0.7 s exposure times per step. While in the glovebox, the sample was sealed
within an airtight XRD holder made from a poly(methyl methacrylate), or PMMA,
dome to prevent exposure to air and moisture during the measurements. Bruker
Diffracplus EVA version 16 and Diffracplus TOPAS version 4.2 were used to identify
crystalline compounds present and for Rietveld refinement respectively. An instru-
mental parameter file was used to eliminate instrumental line broadening for all
analyses. Crystallite size values were taken from the LVol-IB (volume weighted
mean column height) that incorporates Lorentzian and Gaussian convolutions vary-
2
2
2
2
MgH þ Si $ Mg Si þ 2H
2
ðgÞ
ð3Þ
ð4Þ
ð5Þ
ð6Þ
2
2
NaMgH þ Si ! Mg Si þ 2NaH þ 2H
2
ðgÞ
3
2
MgH þ Ge $ Mg Ge þ 2H
2
ðgÞ
ðgÞ
2
2
ꢀ
1
ing in 2h as a function of cos(h) and tan(h) respectively. This method provides a
volume weighted average crystallite size. Uncertainties were reported from TOPAS
MgH þ Sn $ Mg Sn þ 2H
2
2
2
(
bootstrap method of error determination). It should be noted that the grey plots at
the bottom of each XRD figure (Figs. 2–5) are an indication of the difference
between the raw data collected on the XRD equipment and Rietveld refinement.
The Rwp values (Tables 1–4), or weighted profile R-factor value, also gives an indi-
cation of the accuracy of the simulated model. This discrepancy index uses an algo-
rithm to optimize the model function so that a minimum of the weighted sum of
squares differences between the experimental and computed intensities is calcu-
lated [34]. As a general rule, Rwp values 5% or less indicates an acceptable goodness
of fit [34], however, this value largely depends on an over estimation of uncertain-
ties and should only be used as a guide [34].
Scanning electron microscopy (SEM) of the materials was performed using a
Zeiss Neon 40EsB (Zeiss, Germany). Specimens were prepared by distributing a
small amount of powder onto carbon tape then coating with a 2–4 nm layer of high
atomic elements, either gold or platinum, to produce a conductive layer and reduce
charging of the sample during its interaction with the electrons in the SEM. All sam-
ples were exposed to air for a short period of time when transferred from the coat-
ing instrument as well as loading into the SEM chamber.
2
. Experimental
All material handling was undertaken in an argon atmosphere glovebox (Uni-
labGlovebox, mBraun, Germany). An automatic gas purifier unit controlled the oxy-
gen and moisture levels to limit any risk of contamination (O < 1 ppm,
O < 1 ppm). Magnesium hydride (H storage grade, 95%), silicon powder
ꢀ325 mesh, 99%), germanium (>99.999%) and tin (>99%) were supplied by
Sigma–Aldrich. NaMgH synthesis is described in a previous publication [20].
For direct comparison with MgH , theoretical equilibrium pressures of MgH
mixed with Si, Ge and Sn were calculated using thermodynamic data for MgH from
2
H
(
2
2
3
2
2
2
Bogdanovi c´ et al. [33]and thermodynamic data for the other compounds from the
software program HSC (HSC Chemistry 6.12 software, Outotech Research). This soft-
ware allows the user to enter in known values (from literature) for enthalpy and
entropy and calculate theoretical thermodynamic properties over a specified tem-
perature range and pressure.
Table 1
Pure MgH
Sample
MgH
2
and Si addition to MgH
2
crystallite size from Rietveld refinement.
Phase
Structure
wt.%
Cryst. size (nm)
2
b-MgH
Mg
2
Tetragonal, 136, P42/mnm
Hexagonal, 194, P63/mmc
Cubic, 227, Fd-3m
Tetragonal, 136, P42/mnm
Orthorhombic, 60, Pbcn
Hexagonal, 194, P63/mmc
Cubic, 227, Fd-3m
Cubic, 225, Fm-3m
Cubic, 227, Fd-3m
Cubic, 225, Fm-3m
96.3
3.7
100
45.2
21.6
1.0
32.2
97.1
2.4
86 ± 1
93 ± 12
211 ± 3
7.6 ± 0.1
2.4 ± 0.2
58 ± 13
67 ± 1
62 ± 1
100 ± 13
2.3 ± 0.1
Rwp = 8.1% (as supplied)
Si Rwp = 7% (as supplied)
Si
2
MgH
2
+ Si
b-MgH
2
Rwp = 4.8% (cryomilled)
c-MgH
2
Mg
Si
Mg
2
Si
Mg
Si
2
Si
Rwp = 7.9% (desorbed)
MgO
0.5

A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
111
Table 4
3
Si addition to NaMgH , crystallite size from Rietveld Refinement.
3
3
00x10
(
(
A)
B)
Sample
Phase
Structure
wt.% Crystallite
size (nm)
NaMgH
3
NaMgH
NaH
3
Orthorhombic, 62, 76.3 31 ± 1
Pnma
Cubic, 225, Fm-3m 2.2 7.7 ± 1.4
Rwp = 9.3% (synthesised from
Mg Si
2
MgH + NaH)
2
MgH
2
Tetragonal, 136,
3.7 2.5 ± 0.7
P42/mnm
Cubic, 225, Fm-3m 17.8 1.3 ± 0.1
200
100
0
- MgH2
MgH2
MgO
-
2
NaMgH
3
+ Si
NaMgH
3
Orthorhombic, 62, 60.9 10.7 ± 0.1
Pnma
Cubic, 227, Fd-3m 21.4 81 ± 1
Cubic, 225, Fm-3m 17.7 1.4 ± 0.1
Si
Mg
Rwp = 4.5% (cryomilled)
Si
MgO
Mg
2
Si, NaH
Mg
2
Si
Cubic, 225, Fm-3m 47.7 44 ± 1
Cubic, 227, Fd-3m 27.4 40 ± 1
Cubic, 227, Fd-3m
Rwp = 6.2% (desorbed)
NaH
Si
MgO
7.7 76 ± 3
Cubic, 225, Fm-3m 17.2 2.2 ± 0.1
Hydrogen desorption properties were analysed using a manometric Sieverts
apparatus where the sample cell was placed in a furnace and isothermal measure-
ments taken over time (see [16] for more details). Each sample was held for 24 h at
(
C)
5
0 °C and then at 50 °C increments through to a maximum temperature of 350 °C. If
(A)
B)
there was no hydrogen desorption detected after 2 h at 350 °C, the run was stopped.
Temperature programmed desorption (TPD) was also undertaken using a Stan-
ford Research Systems Residual Gas Analyzer (RGA) 300 under vacuum. This
(
ꢀ
5
method uses a turbo vacuum pump to achieve a low pressure atmosphere (10
-
(
C)
mbar) and identifies gases released from the system according to the atomic mass
unit (amu) of the gas by a mass spectrometer. 20–30 mg of each sample was loaded
into a sample cell and the system was outgassed at room temperature until all
traces of argon were undetectable. An N-type thermocouple was attached to the
outside the cell within proximity of the sample being analysed. The samples were
2
0
30
40
50
60
70
80
ꢀ
1
heated from room temperature up to 600 °C at a rate of 2 °C min
.
Fig. 2. XRD patterns of Si based experiments (A) MgH
Mg Si, 24 h in 50–350 °C increments (B) Cryomilled with MgH
2
and Si desorbed to form
for 30 min (C) Si as
2
2
supplied from Sigma–Aldrich.
3
. Results and discussion
Theoretical comparisons between all compounds discussed in
this article can be found in Fig. 1. Since NaMgH
3
enthalpy data
Fig. 3. XRD patterns of Ge based experiments (A) MgH
2
and Ge desorbed to form
Fig. 1. Calculated equilibrium pressures for MgH
added Group IV elements.
2
and NaMgH
3
with and without
Mg
as supplied from Sigma–Aldrich.
2
Ge, 24 h in 50–350 °C increments (B) Cryomilled with MgH for 30 min (C) Ge
2

1
12
A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
was not available, the formation enthalpy and entropy were deter-
mined at 419 °C based on the values for NaH, Mg and H at that
temperature combined with enthalpy of hydrogen desorption,
6.6 kJ/mol H (from NaMgH via the reaction of NaMgH
M NaH + Mg + H [20]). The heat capacity of NaMgH is not exper-
imentally known but was approximated as the sum of the heat
capacities of its individual components: NaH and MgH . The verac-
ity of this approached was confirmed with the isostructural phase,
NaMgF . There is a difference of less than 3.5% between the heat
capacity of NaMgF [35] and the sum of the heat capacities for
NaF and MgF over the temperature range of 40–400 °C. The result-
ing standard enthalpy, , and entropy, , of formation at 25 °C
are ꢀ143.0 kJ/mol and ꢀ208.0 J/mol K, respectively. The value
is in good agreement with the value calculated by Bouhadda [36],
2
8
2
3
3
-
2
3
2
3
3
2
D
H
f
DS
f
DH
f
ꢀ
151.8 kJ/mol, using Density Functional Theory.
The results show that when mixed with either Si or Sn at tem-
peratures less than 50 °C, hydrogen equilibrium pressure of less
than 10 bar are achievable. However, at room temperature, MgH
2
combined with Ge, results in a hydrogen equilibrium pressure
above 500 bar. These predictions indicate an improvement in the
thermodynamic behaviour of hydrogen release when MgH
2
is com-
bined with Si, Ge or Sn.
3 2
The comparison between NaMgH /Si with MgH /Si shows that
the equilibrium pressure for the former is much lower than the lat-
ter at higher temperatures (Fig. 1). This indicates that reaction
kinetics could be improved with an increase in temperature with-
out thermodynamic restrictions (high pressures) coming into
effect. This change in the reaction pathway would also lead to bet-
ter hydrogen reabsorption into Mg
the addition of Group IV elements to MgH
2
Si. These results indicate that
or NaMgH can lead
2
3
Fig. 5. XRD patterns of NaMgH
form Mg Si and NaH, 24 h in 50–350 °C increments (B) Cryomilled with Si for
0 min (C) NaMgH synthesised from MgH and NaH.
3 3
based experiments (A) NaMgH and Si desorbed to
2
3
3
2
to thermodynamically favourable desorption temperatures, thus
creating a more practical hydrogen storage system.
Prior to all desorption experiments, the materials as supplied by
the manufacturer were characterised in terms of phases present
and crystallite size (Table 1). XRD and Rietveld refinement results
show that the MgH contained approximately 3.7 wt.% Mg which
2
would result in a lower than expected hydrogen capacity. The
raw Si had no detectable contaminants. Ge had a small percentage
of contamination, GeO
evidence of this phase in subsequent analysis as Mg/MgH
reduce GeO to form Ge and MgO. Sn, like Si, had no detectable
2
, about 2.6 wt.%. However, there was little
2
should
2
contamination and had a different crystallite structure to both Si
and Ge with a much larger crystallite size of 547 ± 16 nm. The syn-
3
thesized NaMgH had a similar level of minor oxidation as
Sheppard et al. [20] with some MgO forming during the synthesis
process.
After cryomilling each of the reactants in the correct stoichiom-
etric ratios (Reactions (2)–(6), XRD was used to calculate the
resulting crystallite sizes of each alloy present. From an initial crys-
tallite size of 85 ± 1 nm, the size of b-MgH
after cryomilling with Si, Ge and Sn producing sizes of
.6 ± 0.1 nm, 5.6 ± 0.2 nm and 7.0 ± 0.6 nm respectively and pure
b-MgH cryomilled under the same conditions results in a crystal-
lite size of 6.3 ± 0.1 nm. Each of the Group IV elements also reduced
in crystallite size, but not to the same extent as MgH . It is common
for a portion of pure MgH to form metastable orthorhombic
-MgH during milling.[37] It was interesting to observe that
cryomilling MgH with Si (6.5 MohsScale) led to the formation of
-MgH , however, this polymorph was not evident when milling
2
was greatly reduced
7
2
2
2
c
2
2
c
2
Fig. 4. XRD patterns of Sn based experiments (A) MgH
Mg Sn, 24 h in 50–350 °C increments (B) Cryomilled with MgH
supplied from Sigma–Aldrich.
2
and Sn desorbed to form
with either Ge or Sn. One explanation is perhaps the presence of
Ge (6 Mohs Scale) and Sn (4 Mohs) relieve the impact pressure
2
2
for 30 min (C) Sn as

A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
113
on MgH
NaMgH
2
thus preventing the formation of
with Si mixed the reagents and reduced the crystallite
c
-MgH
2
. Cryomilling
spread in particle size distribution ranging from 100 nm through to
a few microns.
3
size as expected without effecting the structure of either NaMgH
or Si.
In order to gauge homogeneity of mixing and morphological
information of the cryomilled powders, each milled sample con-
3
This result is similar to previous research into ball milled MgH
and Ge, where the authors found a large particle size distribution
ranging from 10 m particles to the sub-micron range [29].
The hydrogen desorption behaviour of as-milled samples was
observed using a manometric Sieverts apparatus (Fig. 7) and tem-
perature programmed desorption (TPD) instrument (Fig. 8). All of
the mixtures showed little or no hydrogen desorption at the lower
temperatures of 50 °C and 100 °C.
2
l
2
taining MgH and Group IV elements was analysed with SEM.
Backscattered electron (BSE) images of these samples are shown
in Fig. 6a, c and e. The BSE images highlight elements of heavier
atomic mass as brighter regions. All samples were viewed under
similar conditions (magnification, spot size, electron voltage and
working distance) to easily compare differences or similarities.
2
As indicated in Fig. 7a desorption of Si/MgH began at 250 °C
and, as expected, an increase in temperature lead to an increase
in reaction kinetics although the sample was not fully desorbed
until 350 °C. These results are in agreement with Paskevicius
The milled MgH
well mixed. Analysis is made difficult because there is little differ-
ence in contrast MgH and Si as they have similar atomic mass.
However, the slightly brighter regions can be attributed to Si
particles. Si particles appear to be slightly larger in size to MgH
2
/Si sample (Fig. 6a) shows that the reagents are
2
et al. [22] where MgH
the range of 250 °C and 350 °C. XRD post reaction resulted in the
presence of Mg Si with near complete conversion, 97 wt.%
(Fig. 2a). Again the theoretical hydrogen content of 5 wt.% was
not reached due to the impurity of Mg detected in the MgH
Sieverts apparatus desorption obtained a final H release of
2
ball milled with Si for 24 h desorbed in
2
,
2
which is expected due to the hardness of Si and consequently the
difficulty in particle size refinement during milling. With the
higher atomic masses of Ge and Sn, the contrast between MgH
2
.
2
2
and Ge (Fig. 6c) and Sn (Fig. 6e) is more apparent. The micrographs
show that each of the reactants is well dispersed after mixing with
the cryomill. Similar to Si, Ge and Sn particle sizes are larger than
4.81 wt.%, 96% of the theoretical value. BSE SEM on the sample
after Sieverts apparatus desorption (Fig. 6b) correlates with near
2
full conversion to Mg Si, showing a homogenous mixture in terms
MgH
2
. All the micrographs of the cryomilled materials show a large
of atomic weight (no contrast variation) as well as particle size
2 2 2 2 2 2
Fig. 6. SEM BSE images (accelerating voltage 20 kV) of (A) 2MgH + Si (B) Mg Si (C) 2MgH + Ge (D) Mg Ge (E) 2MgH + Sn (F) Mg Sn.

1
14
A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
(
average size less than 1
l
m). The desorption temperature during
from the Sieverts desorption was 2.91 wt.% H
the theoretical value of 3.22 wt.%. The marginally lower experi-
mental value is attributed to both Ge and MgH containing small
amounts of Mg and GeO , as indicated by the XRD prior to desorp-
tion (Figs. 2b and 3b). MgH
has a purity of ꢁ95% therefore all
2
, slightly lower than
TPD measurements was also higher for the Si mixture than either
Ge or Sn, where desorption occurred as a single event at a temper-
2
ature of 350 °C. This value was marginally lower than MgH
milled under the same conditions (desorption temperature of
70 °C), again restating the fact that this reaction is kinetically lim-
2
cryo-
2
2
3
desorption reactions would not reach theoretical desorption values
due to the presence of impurities. The BSE image from SEM
(Fig. 6d) after desorption also shows a more homogeneous mor-
phology of smaller particles when compared to the pre-desorption
image. Also, there is no significant brightness contrast indicating
that an almost full conversion to a single phase has taken place
as differences in the atomic mass are not observed.
ited despite the thermodynamic destabilisation effect of Si addi-
tion. The single decomposition peak indicates that Mg diffusion
into the Si matrix occurs in one kinetically limited step, implying
that the particle size distribution was narrower. Another study that
involved TPD also gave single peak desorption, however at a lower
temperature (290 °C), as a different heating rate was used and the
sample was under a helium atmosphere, not under vacuum [23].
An important point to note for the Group IV element samples
2
Temperature programmed desorption from the MgH and Ge
mixture also began at a lower temperature (260 °C) compared to
all other materials (Fig. 8c). It was unexpected that this TPD would
contain two decomposition events. This indicates that the decom-
position is two-step, either due to an intermediate decomposition
product, or a two-stage kinetic process. To understand this phe-
nomenon, a second TPD was performed and halted at 300 °C (after
the first peak was fully resolved) and the sample cell quenched in
liquid nitrogen to prevent any further reaction. The resultant XRD
is shown in Fig. 9a. There are no unknown structures other than
added to MgH
than pure MgH
MgH
2
was that all desorption temperatures were lower
2
(365 °C).
2
with Ge was the first mixture to begin hydrogen release
at 150 °C; although this step was kinetically slow, as indicated by
the shallow incline at 150 °C in Fig. 7b. The majority of hydrogen
desorbed from this mixture at higher temperatures (200–250 °C)
with faster reaction rates (steeper inclines Fig. 7b) until it com-
pletely desorbed at 300 °C. XRD from the decomposed sample is
given in Fig. 3a and reveals almost complete conversion to Mg
No evidence of MgH peaks could be detected using XRD, however,
traces of Ge still remained. The total quantity of hydrogen released
2
Ge.
2 2
the expected MgH , Ge and Mg Ge. That is, there are no unex-
pected intermetallics present, such as MgGe. The second desorp-
tion peak can be explained by the wide range in particle size that
2
2 2 2 2 3
Fig. 7. Time and temperature relationship with wt.% of H desorbed from (A) 2MgH + Si (B) 2MgH + Ge (C) 2MgH + Sn (D) 2NaMgH + Si.

A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
115
the sample was fully desorbed. Fig. 4a shows the resultant XRD
after desorption with an almost full conversion to 94 wt.% Mg Sn.
Again the Sieverts measurements gave a slightly lower hydrogen
release of 2.15 wt.% H when compared to the theoretical storage
capacity of 2.36 wt.%. SEM results give an interesting change in
morphology of the reacted Mg Sn that was not seen in any other
sample (Fig. 6f). Although largely homogenous in atomic mass
no contrast differences), well defined obelisk shaped particles of
various sizes were detected, typically quite large, ꢁ5 m. This is
possibly due to the fact that the melting point of Sn is only about
32 °C. The XRD after halting the TPD at 300 °C (Fig. 9b) showed
that some Sn remained so this could have melted and agglomer-
ated. This is in contrast to the MgH /Sn morphology in the as-syn-
thesised ball milled sample, which is more spherical in structure
38]. Similar to the reaction containing Ge, Sn with MgH also
2
-
6
2
4
00x10
2
(
3
2
1
00
00
00
0
l
(
(
A)
B)
2
2
[
2
resulted in two-step decomposition during the TPD experiment
likely due to the wide particle size distribution. The initial hydro-
gen release peak occurred at 270 °C with the second at 360 °C as
shown in Fig. 8d. A double desorption peak was also displayed dur-
(
C)
(
(
D)
E)
2
ing a TPD experiment [31] with balled milled MgH and Sn with
the addition of cyclohexane. This different preparation technique
of ball milling with cyclohexane reduced the temperature of the
peaks to 217 °C and 257 °C [31]. The authors of this study [31] con-
cluded that the formation of the two desorption events was due to
1
00
200
300
400
500
600
the existence of two types of hydrogen species in the Sn/MgH
2
Temperature (ºC)
composite, however, our XRD results contradict this statement.
Similar to the Ge system, the TPD was repeated and stopped at
300 °C and the XRD (Fig. 9b) showed no unexpected intermediate
phases. Therefore, the two-step decomposition can be most likely
attributed to the variation in particle size, resulting in different dif-
Fig. 8. Mass spectra of desorbed gases from: (A) Pure MgH
MgH + G0065 (D) 2MgH + Sn (E) 2NaMgH + Si.
2
2
(B) 2MgH + Si (C)
2
2
2
3
2
fusion rates for Mg and Sn to form Mg Sn. Both of the Ge and Sn
samples mixed with MgH2 appeared to have less kinetic limita-
tions than Si with MgH2.
causes two-step kinetic behaviour. The particle size distributions
must be significantly different enough such that diffusion of Mg
and Ge occurs in two stages, one temperature (260 °C) for the
smaller particle range, approximately 100 nm or below, and a
higher temperature of 350 °C for the larger (micron sized) parti-
cles. This is in contradiction to Walker et al. [29] who attributed
a single peak in DSC measurements to be the thermodynamic
Sieverts desorption results for the NaMgH /Si system are shown
3
in Fig. 7d. Initially, desorption occurred at a consistently slow rate
until 250 °C, where the rate increased. The majority of desorption
however, occurred at 300 °C, 80 °C lower than pure NaMgH3 at
1 bar of pressure_ENREF_3. Similar, to the previous mixtures, a
lower experimental value was reached for the total quantity of
hydrogen desorbed, 2.34 wt.%, compared to the theoretical value
of 3.13 wt.%. TPD results (Fig. 8e) show overlapping peaks occur-
ring at maximum temperatures of 320 °C and 340 °C. As with Ge
2
event of MgH dehydrogenation despite a large variation in particle
size.
2
Sn with MgH began desorption at 200 °C as detected using the
Sieverts apparatus. However, full desorption did not occur after
2
2
4 h at this temperature. Once the temperature increased to
50 °C and then 300 °C, the rate of reaction slowed and by 350 °C
and Sn addition to MgH , this could be the result of differing par-
2
ticle sizes.
2 2
Fig. 9. XRD taken after heating to before (25 °C), during (300 °C) and after (600 °C) RGA analysis (A) 2MgH + Ge (B) 2MgH + Sn.

1
16
A.-L. Chaudhary et al. / Journal of Alloys and Compounds 623 (2015) 109–116
13] R. Schulz, J. Huot, G. Liang, S. Boily, G. Lalande, M.C. Denis, J.P. Dodelet, Recent
[
4
. Conclusions
developments in the applications of nanocrystalline materials to hydrogen
technologies, Mater. Sci. Eng. A – Struct. 267 (1999) 240–245.
[14] J. Huot, G. Liang, R. Schulz, Mechanically alloyed metal hydride systems, Appl.
Phys. A – Mater. 72 (2001) 187–195.
A direct comparison of the addition of Si, Ge and Sn showed
interesting results when each mixture was prepared under identi-
cal conditions. Each of the Group IV elements added to MgH and Si
to NaMgH , successfully lowered the decomposition temperature,
[
15] M.P. Pitt, M. Paskevicius, C.J. Webb, D.A. Sheppard, C.E. Buckley, E.M. Gray, The
synthesis of nanoscopic Ti based alloys and their effects on the MgH system
benchmark, Int. J. Hydrogen Energy 37
2
2
3
compared with the MgH
2 2 5
+ 0.01Nb O
however, it appears reaction kinetics play a larger role than ther-
modynamics in the decomposition reaction. It appears that the sys-
tem that contained Ge and MgH had the fastest reaction kinetics
2
(2012) 4227–4237.
[
16] M. Paskevicius, D.A. Sheppard, C.E. Buckley, Thermodynamic changes in
mechanochemically synthesized magnesium hydride nanoparticles, J. Am.
Chem. Soc. 132 (2010) 5077–5083.
since it started desorption at the lowest temperature of 150 °C,
although it undertook two-step decomposition. Sn also had a sig-
[17] K. Ikeda, S. Kato, Y. Shinzato, N. Okuda, Y. Nakamori, A. Kitano, H. Yukawa, M.
Morinaga, S. Orimo, Thermodynamical stability and electronic structure of a
3
perovskite-type hydride, NaMgH , J. Alloys Comp. 446–447 (2007) 162–165.
nificant effect on MgH
with comparison to the other Group IV elements, had a lesser effect
on MgH with desorption initiated at 250 °C. This study also pro-
vided information on the influence of Si on the NaMgH compound.
2
with desorption occurring at 200 °C. Si,
[
18] D. Pottmaier, E.R. Pinatel, J.G. Vitillo, S. Garroni, M. Orlova, M.D. Baró, G.B.M.
Vaughan, M. Fichtner, W. Lohstroh, M. Baricco, Structure and thermodynamic
properties of the NaMgH
2011) 2317–2326.
3
perovskite: a comprehensive study, Chem. Mater. 23
2
(
3
[
[
19] H. Wu, W. Zhou, T.J. Udovic, J.J. Rush, T. Yildirim, Crystal chemistry of
Reduction in desorption temperatures for all materials was
observed, however kinetic limitations ensured that thermody-
namic equilibrium conditions were not achieved. This indicates
that kinetics have a significant role in each of the systems studied.
Since differences in particle size were considered significant, the
differences in kinetics are most likely due to different orders of
reaction for nucleation and possibly growth due to particle size.
perovskite-type hydride NaMgH : implications for hydrogen storage, Chem.
3
Mater. 20 (2008) 2335–2342.
20] D. Sheppard, M. Paskevicius, C. Buckley, Thermodynamics of hydrogen
desorption from NaMgH
3
and its application as a solar heat storage medium,
Chem. Mater. 23 (2011) 4298–4300.
[21] A.-L. Chaudhary, D.A. Sheppard, M. Paskevicius, M. Saunders, C. Buckley,
Mechanochemical synthesis of amorphous silicon nanoparticles, RSC Adv. 42
(
2014) 21979–21983.
[
[
[
22] M. Paskevicius, D.A. Sheppard, A.L. Chaudhary, C.J. Webb, E.M.A. Gray, H.Y.
Tian, V.K. Peterson, C.E. Buckley, Kinetic limitations in the Mg–Si–H system,
Int. J. Hydrogen Energy 36 (2011) 10779–10786.
Acknowledgements
23] M. Polanski, J. Bystrzycki, The influence of different additives on the solid-state
2
reaction of magnesium hydride (MgH ) with Si, Int. J. Hydrogen Energy 34
AC acknowledges Curtin University for granting the Postgradu-
ate Scholarship and Research Scholarship (CUPS and CURS) as well
as the Australian Commonwealth Scientific and Research Organisa-
tion (CSIRO) for providing funding for the project. AC also acknowl-
edges Elaine Miller for her assistance with the SEM at Curtin
University. CEB, MP and DAS acknowledge the financial support
of the Australian Research Council for ARC Linkage Grant
LP120100435, and CEB acknowledges the ARC for ARC LIEF Grants
LE0775551 and LE0989180.
(2009) 7692–7699.
24] J.J. Vajo, F. Mertens, C.C. Ahn, R.C. Bowman, B. Fultz, Altering hydrogen storage
properties by hydride destabilization through alloy formation: LiH and MgH
destabilized with Si, J. Phys. Chem. B 108 (2004) 13977–13983.
2
[25] A.-L. Chaudhary, D.A. Sheppard, M. Paskevicius, C.J. Webb, E.M. Gray, C.E.
Buckley, Mg Si nanoparticle synthesis for high pressure hydrogenation, J. Phys.
2
Chem. C 118 (2014) 1240–1247.
[
[
[
26] N.O. Folland, F. Bassani, Selection rules and Mg
Chem. Solids 29 (1968) 281–290.
2
Si, Mg
2
Ge and Mg
2
Sn, J. Phys.
27] J.J. Martin, Thermal conductivity of Mg
Solids 33 (1972) 1139–1148.
2
Si, Mg
2
Ge and Mg
2
Sn, J. Phys. Chem.
28] D. Zhou, J. Liu, S. Xu, P. Peng, Thermal stability and elastic properties of Mg
X = Si, Ge, Sn, Pb) phases from first-principle calculations, Compd. Mater. Sci.
51 (2012) 409–414.
2
X
(
References
[
[
29] G. Walker, M. Abbas, D. Grant, C. Udeh, Destabilisation of magnesium hydride
by germanium as a new potential multicomponent hydrogen storage system,
Chem. Commun. 47 (2011) 8001–8003.
[
[
[
1] D.M. Abrams, R.J. Wiener, A model of peak production in oil fields, Am. J. Phys.
8 (2010) 24–27.
7
30] H. Imamura, K. Tanaka, I. Kitazawa, T. Sumi, Y. Sakata, N. Nakayama, S.
2] D.J. Arent, A. Wise, R. Gelman, The status and prospects of renewable energy
for combating global warming, Energy Econ. 33 (2011) 584–593.
3] N.A. Kelly, T.L. Gibson, M. Cai, J.A. Spearot, D.B. Ouwerkerk, Development of a
renewable hydrogen economy: optimization of existing technologies, Int. J.
Hydrogen Energy 35 (2010) 892–899.
Ooshima, Hydrogen storage properties of nanocrystalline MgH
nanocomposite synthesized by ball milling, J. Alloys Comp. 484 (2009) 939–
42.
2 2
and MgH /Sn
9
[
[
[
31] H. Imamura, K. Yoshihara, M. Yoo, I. Kitazawa, Y. Sakata, S. Ooshima,
Dehydriding of nanocomposite formed by ball milling of with Sn, Int. J.
Hydrogen Energy 32 (2007) 4191–4194.
32] H.C. Zhong, H. Wang, L.Z. Ouyang, M. Zhu, Microstructure and hydrogen
storage properties of Mg–Sn nanocomposite by mechanical milling, J. Alloys
Comp. 509 (2011) 4268–4272.
33] B. Bogdanovi c´ , K. Bohmhammel, B. Christ, A. Reiser, K. Schlichte, R. Vehlen, U.
Wolf, Thermodynamic investigation of the magnesium–hydrogen system, J.
Alloys Comp. 282 (1999) 84–92.
[
[
[
[
4] J. Brouwer, On the role of fuel cells and hydrogen in a more sustainable and
renewable energy future, Curr. Appl. Phys. 10 (2010) S9–S17.
5] P.P. Edwards, V.L. Kuznetsov, W.I.F. David, N.P. Brandon, Hydrogen and fuel
cells: towards a sustainable energy future, Energy Policy 36 (2008) 4356–4362.
6] M. Hirscher, Handbook of Hydrogen Storage: New Materials for Future Energy
Storage, 2009.
7] U. Department of Energy, Targets for OnBoard Hydrogen Storage Systems for
Light-Duty Vehicles, Office of Energy Efficiency and Renewable Energy and The
FreedomCAR and Fuel Partnership, 2009.
[
[
34] B.H. Toby, R factors in Rietveld analysis: how good is good enough?, Powder
Diffr 21 (2006) 67–70.
35] L. Topor, A. Navrotsky, Y. Zhao, D.J. Weidner, Thermochemistry of fluoride
perovskites: heat capacity, enthalpy of formation, and phase transition of
[
[
8] U. Bösenberg, C. Pistidda, M. Tolkiehn, N. Busch, I. Saldan, K. Suarez-Alcantara,
A. Arendarska, T. Klassen, M. Dornheim, Characterization of metal hydrides by
in-situ XRD, Int. J. Hydrogen Energy 39 (2014) 9899–9903.
9] L.H. Jepsen, M.B. Ley, Y.-S. Lee, Y.W. Cho, M. Dornheim, J.O. Jensen, Y. Filinchuk,
J.E. Jørgensen, F. Besenbacher, T.R. Jensen, Boron–nitrogen based hydrides and
reactive composites for hydrogen storage, Mater. Today 17 (2014) 129–135.
3
NaMgF , J. Solid State Chem. 132 (1997) 131–138.
[
[
[
36] Y. Bouhadda, Y. Boudouma, N.-E. Fennineche, A. Bentabet, Ab initio
calculations study of the electronic, optical and thermodynamic properties
of NaMgH , for hydrogen storage, J. Phys. Chem. Solids 71 (2010) 1264–1268.
3
37] J. Huot, G. Liang, S. Boily, A. Van Neste, R. Schulz, Structural study and
hydrogen sorption kinetics of ball-milled magnesium hydride, J. Alloys Comp.
[
10] M.B. Ley, L.H. Jepsen, Y.-S. Lee, Y.W. Cho, J.M. Bellosta von Colbe, M. Dornheim,
M. Rokni, J.O. Jensen, M. Sloth, Y. Filinchuk, J.E. Jørgensen, F. Besenbacher, T.R.
Jensen, Complex hydrides for hydrogen storage – new perspectives, Mater.
Today 17 (2014) 122–128.
2
93–295 (1999) 495–500.
38] T. Aizawa, R. Song, Mechanically induced reaction for solid-state synthesis of
Mg Si and Mg Sn, Intermetallics 14 (2006) 382–391.
[
[
11] D.A. Sheppard, M. Paskevicius, C.E. Buckley, The mechanochemical synthesis of
magnesium hydride nanoparticles, J. Alloys Comp. 492 (2010) L72–L74.
12] A.Y. Yermakov, N.V. Mushnikov, M.A. Uimin, V.S. Gaviko, A.P. Tankeev, A.V.
Skripov, A.V. Soloninin, A.L. Buzlukov, Hydrogen reaction kinetics of Mg-based
alloys synthesized by mechanical milling, J. Alloys Comp. 425 (2006) 367–372.
2
2