Hydrogen production and crystal structure of ball-milled MgH2-Ca and MgH2-CaH2 mixtures

Article abstract of DOI:10.1016/j.jallcom.2003.12.013

The structure and hydrolysis reaction of ball-milled MgH₂-Ca and MgH₂-CaH₂ mixtures was investigated in function of milling time and component proportion. The nanocomposites formed by milling show faster hydrolysis reactio

Full text of DOI:10.1016/j.jallcom.2003.12.013

Journal of Alloys and Compounds 376 (2004) 180–185
Hydrogen production and crystal structure of ball-milled
MgH2–Ca and MgH2–CaH2 mixtures
J.-P. Tessiera, P. Palaua,1, J. Huotb,2, R. Schulzb, Daniel Guaya,∗
a
Lab. de Recherche Mat. Avances, INRS Énergie, Matériaux et Télécommunications, Université de Québec,
1650 Boul. Lionel-Boulet, C.P. 1020, Varennes, Que., Canada J3X 1S2
Expertise Chimie et Matériaux, Institut de Recherche d’Hydro-Québec, 1800 Boul. Lionel-Boulet, Varennes, Que., Canada J3X 1S1
b
Received 15 November 2003; received in revised form 8 December 2003; accepted 8 December 2003
Abstract
The structure and hydrolysis reaction of ball-milled MgH2–Ca and MgH2–CaH2 mixtures was investigated in function of milling time and
component proportion. The nanocomposites formed by milling show faster hydrolysis reaction as well as higher hydrogen yield compared
to conventional polycrystalline materials. This improvement is due to the formation of calcium hydride upon milling and to the fine mixture
that results from extensive milling. When the milling is performed on MgH2–20.3 mol% CaH2 mixtures, the reaction yield reaches 80% after
30 min of hydrolysis.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Chemical hydrides; Nanocrystalline materials; Hydrolysis; Irreversible hydrides
1. Introduction
production by steam hydrolysis of NaBH4 was also investi-
gated and found to meet the purity requirement for fuel cell
applications [7]. However, the price and availability of these
types of compounds may make their widespread use diffi-
cult. Therefore, there is a need for the development of an in-
expensive chemical hydride which has high hydrogen yield,
and fast but controllable hydrogen release. Furthermore, the
residual products should be environmentally friendly.
In a previous paper it was shown that, compared to con-
ventional materials, MgH2–Ca nanocomposites have hydrol-
ysis kinetics greatly enhanced as well as much higher yield
of reaction [8]. In this paper, we present a systematic in-
vestigation of the structure and hydrolysis characteristics of
ball-milled MgH2–Ca and MgH2–CaH2 mixtures. Depend-
ing on the phases present in the synthesized nanocompos-
ites, the chemical reactions of interest are:
For widespread utilization of hand-portable proton ex-
change membrane fuel cell (PEMFC), a compact, safe and
inexpensive source of hydrogen must be designed. Hydroly-
sis reactions could potentially serve this purpose. However,
for most compounds, the reaction is either too violent (for ex-
ample: LiAlH4) or too slow to be convenient. Also, because
of the practical non-reversibility of this means of generating
hydrogen, the chosen materials should be of low cost.
Chemical hydrides such as sodium borohydride (NaBH4)
or lithium aluminum hydride (LiAlH4) can be used as hy-
drogen generators [1]. However, one important drawback of
this method is the need of using a catalyst to activate the re-
action [2–4]. Recently, Matthews and coworkers [5,6] have
investigated novel complex hydrides stabilized with organic
ligands as a source of hydrogen for PEMFCs. Hydrogen
Mg + 2H2O → Mg(OH)2 + H2
MgH2 + 2H2O → Mg(OH)2 + 2H2
Ca + 2H2O → Ca(OH)2 + H2
CaH2 + 2H2O → Ca(OH)2 + 2H2
(1)
(2)
(3)
(4)
∗
Corresponding author. Tel.: +1-450-929-8141;
fax: +1-450-929-8102.
E-mail addresses: huot.jacques@ireq.ca (J. Huot),
guay@inrs-emt.uquebec.ca (D. Guay).
1
Present address: Pechiney Centre de Recherches de Voreppe, 725
rue A. Berges, BP 27, 38341 Voreppe Cedex, France.
2
Co-corresponding author.
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2003.12.013

J.-P. Tessier et al. / Journal of Alloys and Compounds 376 (2004) 180–185
181
2. Experimental details
The nanocomposites were synthesized by first mixing the
raw materials inside an argon filled glove box. The mix-
tures were then milled in a Spex 8000 model shaker mill
using a vial and balls made of stainless steel, with a ball to
powder weight ratio of 8:1. Magnesium hydride from Th.
Goldschmidt (95 wt.% MgH2, 5 wt.% Mg), calcium (Alfa,
98.8%), and CaH2 powder (Aldrich, 95%) were used. In an
earlier work on ball-milled magnesium hydride, it was found
that the milled powder contains a significant (8 wt.%) frac-
tion of magnesium oxide [9]. Because this oxide does not
participate in the hydrolysis reaction, it was subtracted in all
calculations of the theoretical yields of the reaction.
The hydrolysis reactions were carried out in a 1 l glass re-
actor with four openings used for hydrogen exhaust, water
addition, thermometer, and purge gas. Water addition was
made using a 250 ml ampoule provided with a bypass duct
to prevent gas flow during water addition. The exhaust gas
was flown through a condenser followed by drierite in or-
der to remove all water vapors. The dry hydrogen gas was
then directed to a flow meter (ADM 2000, Agilent Tech-
nologies) where the flow was measured. In each experiment,
a mass of about 250 mg of nanocomposite was loaded into
the flask and allowed to react with 100 ml of deionized wa-
ter. The water addition was performed at t = 1 min in order
to establish a background reading of the flow meter. This
background (≈0.29 ml min−1) was subsequently subtracted
from the raw data. For all measurements, a magnetic stirrer
was used to continuously stir the mixture. In order to eas-
ily compare the different measurements, hydrogen releases
is given as reacted fraction, F, defined as the ratio of the
volume of released hydrogen over the theoretical volume of
hydrogen that should be released assuming that all material
is hydrolyzed. For each experiment, the mass of reactant is
250 5 mg and water volume is 100 ml.
The pH was measured with a pH meter from VWR Scien-
tific model 2000, using an Accumet 13-620-108 electrode.
The specific surface area was measured with an Autosorb 1
gas sorption system from Corporation Quantachrome using
the BET method with nitrogen gas.
X-ray powder diffraction was carried out on a Bruker AXS
D8 diffractometer using Cu K␣ radiation. Crystallite size of
magnesium hydride was evaluated using TOPAZ 2 software
to fit the (1 1 0) peak of magnesium hydride.
Fig. 1. Hydrogen released during hydrolysis of (A) polycrystalline MgH2,
(B) MgH2 milled 10 h, MgH2–X mol% Ca nanocomposite milled 10 h
with (C) X = 4.7 mol%, (D) X = 8.9 mol%, (E) X = 20.3 mol%, and
(F) pre-milled MgH2–20.3 mol% Ca milled 10 h.
this first step is given by examination of the pure MgH2
curves. In this case, the reaction stops after about 1 min.
Independent measurements of the solution pH show that it
reaches 10.5 in the first few minutes of hydrolysis. From
the potential–pH equilibrium diagram, the concentration
of the soluble species (Mg2+) of reaction (1) is only 10−4
molar at pH 10.5 [10]. Therefore, after only 1 min of hy-
drolysis, the pH of the solution is already high enough
to reduce the solubility of Mg2+ to negligible value. This
may explain the sudden stop of the initial burst of the re-
action, as the formation of insoluble magnesium hydroxide
at the surface of the grains will prevent the solution from
reaching un-reacted MgH2. There is almost twice as much
hydrogen liberated from MgH2 milled during 10 h com-
pared to polycrystalline MgH2. This most probably arises
as a consequence of the former material having a higher
effective surface area than the later one (9.7 and 1.2 m2 g−1
respectively).
,
With calcium addition, the fast initial step is followed by
a constant hydrogen flow regime and the reacted fraction
increases with time. In this regime, the calcium addition al-
lows a sustained reaction of magnesium hydride, thus giv-
ing a faster kinetics and higher yield. From Fig. 1, we could
see that the second slope is proportional to the calcium con-
tent, with higher calcium proportion producing higher flow
rates and higher yield. However, at the highest calcium con-
tents (8.9 and 20.3 wt.%), a third regime appears where the
flow rate seems to be inversely proportional to the calcium
content and the reacted fraction seems to level off. More-
over, the inflexion point, where the hydrogen release curve
switches from the second to the third regime, is appearing
sooner for higher calcium proportion. For calcium content
of 8.9 and 20.3 mol%, the inflexion points are estimated to
be 50 and 35 min, respectively. The curve with 4.7 mol% Ca
does not show the third regime.
3. Results and discussion
3.1. Effect of Ca proportion
The effect of Ca proportion in 10 h ball-milled MgH2–Ca
nanocomposites is shown in Fig. 1. For comparison,
ball-milled MgH2 without Ca addition and polycrystalline
MgH2 are also shown. All curves showed a very fast ini-
tial reaction which lasts from 30 s to 1 min. The nature of

182
J.-P. Tessier et al. / Journal of Alloys and Compounds 376 (2004) 180–185
Table 1
ball-milling [11]. The diffraction peaks are broad, pointing
to the fact that these compounds are in a nanocrystalline
form. More details about this material are given elsewhere
[9]. For small proportion of calcium (4.7 mol%), the diffrac-
tion pattern shows a reduction of the ␥-MgH2 metastable
phase and the appearance of a shoulder near 37 degrees.
This shoulder corresponds to the most intense peak of the
magnesium pattern. For higher calcium content (8.9 mol%),
the ␥-MgH2 phase is almost totally extinct and the magne-
sium peaks are more apparent. At the highest calcium con-
tent prepared in this investigation (20.3 mol%), the ␥-MgH2
phase is absent, the magnesium peaks are better defined and
peaks corresponding to CaH2 are clearly present, with the
main peaks located close to 30.5◦. Therefore, upon milling
of a MgH2–Ca mixture, hydrogen atoms are transferred from
MgH2 to Ca, leading to the formation of CaH2 and Mg. The
nanocomposite thus formed should be written MgH2–CaH2
with inclusion of Mg phase. This reaction was expected as
the enthalpy of formation of CaH2 is −188.7 kJ mol−1 [12]
compared to −71.2 kJ mol−1 for MgH2 [13]. The fact that
␥-MgH2 is absent from the nanocomposite with the higher
Ca content suggests that this phase reacts preferentially with
Ca to form CaH2 and Mg. This is consistent with the fact
that ␥-MgH2 is less stable than ␤-MgH2 [14].
Specific surface area of MgH2–Ca mixtures ball-milled 10 h
Composition
Specific surface area (m2 g−1
)
Un-milled MgH2
MgH2
MgH2–4.7 mol% Ca
MgH2–8.9 mol% Ca
MgH2–20.3 mol% Ca
1.2
9.7
7.1
5.8
3.4
One consequence of ball-milling is the variation of spe-
cific surface area of the milled material. In Table 1, the spe-
cific surface area of ball-milled MgH2–Ca composites with
various calcium contents is reported. Magnesium hydride
being a brittle material, milling without additive produced
an almost 10-fold increase of the specific surface area. When
calcium is added, it acts as a binder thus reducing the spe-
cific surface area in direct proportion of the calcium content.
From Table 1, it is clear that the specific surface area is not
responsible for the enhanced reaction kinetic and reaction
yield of MgH2–Ca mixtures compared to ball-milled MgH2.
However, the leveling off of the reacted fraction at high Ca
content may be related to the reduction of the surface area.
3.2. Phase composition
To support this assertion, an experiment was performed
where MgH2 was first ball-milled for 5 h in order to syn-
thesize a large fraction of the ␥-MgH2 metastable phase.
Calcium was then added in a proportion of 20.3 mol%
and the mixture was further milled for 10 h. The X-ray
powder diffraction pattern of that compound is shown in
Fig. 2. There is a higher proportion of CaH2 and Mg in
The phase composition of ball-milled MgH2–Ca mixtures
is revealed by the X-ray powder diffraction patterns shown
in Fig. 2. The sample without calcium shows the character-
istic diffraction peaks of ␤-MgH2 as well as those indicat-
ing the presence of metastable ␥-MgH2 formed by energetic
Fig. 2. X-ray powder diffraction patterns of MgH2–X mol% Ca nanocomposite milled 10 h with (A) X = 0.0 mol%, (B) X = 4.7 mol%, (C) X = 8.9 mol%,
(D) X = 20.3 mol%, and (E) pre-milled MgH2–20.3 mol% Ca milled 10 h.

J.-P. Tessier et al. / Journal of Alloys and Compounds 376 (2004) 180–185
183
Fig. 3. Hydrogen released during hydrolysis of MgH2–8.9 mol% Ca nanocomposite ball-milled for various durations.
the pre-milled sample, as indicated by the higher intensity
of the diffraction peaks located at ∼30.5◦ and ∼37.0◦, re-
spectively. The hydrolysis curve of this material is shown
in Fig. 1. It can be seen that pre-milled nanocomposite
has a much bigger initial burst of hydrogen but afterward,
for hydrolysis time between 1 and 40 min, the reaction
is slower than in MgH2–20.3 mol% Ca milled 10 h. For
longer hydrolysis times, the reaction is faster for pre-milled
materials. The higher initial burst could be explained by
the higher proportion of CaH2 in the pre-milled sample as
evidenced in the X-ray powder diffraction pattern.
The evolution of the solution pH during the hydrolysis
reaction of MgH2–8.9 mol% Ca milled 450 min is shown
on Table 2. The pH increases very quickly and stabilizes to
values superior to 11 after a few min of hydrolysis. In the
same conditions, the end values of pH for hydrolysis of pure
MgH2 and pure Ca are, respectively, 10.5 and 12.3.
The X-ray powder diffraction patterns of ball-milled
MgH2–8.9 mol% Ca are shown in Fig. 4. After 60 min of
milling, the structure is nanocrystalline as indicated by
the broad peaks in the diffraction pattern. Peaks could be
indexed to ␤-MgH2 and ␥-MgH2 magnesium hydrides.
Magnesium and calcium hydride peaks are also relatively
strong, indicating that the decomposition of magnesium
hydride occurs at the beginning of the milling process. The
relative intensities of the diffraction peaks do not vary with
milling time, suggesting that the sample’s composition does
not evolve with time after the initial 60 min of milling.
On these premises, a variation of composition with milling
time could hardly explain the difference of reaction yield
and reaction kinetics observed between samples milled for
different periods of time (see Fig. 3).
3.3. Effect of milling time
A systematic study of hydrogen release as a function
of milling time was performed for the ball-milled mixture
MgH2–8.9 mol% Ca. The results are shown in Fig. 3. As
revealed by Fig. 3A, for milling time shorter than 120 min,
the reaction stops after the initial burst. The milling time
has some effect on the magnitude of this initial burst, the
reacted fraction increases with milling time until a maxi-
mum is reached for 30 min of milling. Further milling up to
120 min only slightly reduced the initial hydrogen release.
Longer milling times are shown in Fig. 3B. For 300 min
milling, the initial burst is followed by a second regime of
constant flow rate. When milling is performed for 450 min
or more, this second regime gives way to a third regime
of smaller flow rate. The transition from the second to the
third regime is around 50 min for the 450 min milled sample
and about 40 min for both 600 and 900 min milled samples.
Therefore, the transition point seems to be function of cal-
cium content (see Fig. 1) and milling time (see Fig. 3), with
an optimum value for each of them.
Table 2
Solution pH as a function of hydrolysis time of MgH2–8.9 mol% Ca
milled 450 min
Hydrolysis time (min)
pH
0
1
8
19
30
45
78
6.8
10.8
11.2
11.5
11.3
11.4
11.1

184
J.-P. Tessier et al. / Journal of Alloys and Compounds 376 (2004) 180–185
Fig. 5. Hydrogen released during hydrolysis of MgH2–20.3 mol% CaH2
nanocomposite ball-milled for 5 min and 10 h.
Fig. 4. X-ray powder diffraction patterns of MgH2–8.9 mol% Ca
ball-milled for different milling times.
reaction that stops after the initial reaction burst. A reaction
yield of ∼0.3 is consistent with CaH2 being fully hydrolyzed
(F = 0.2), while the F value of polycrystalline MgH2 is
close to 0.1 (see Fig. 1). On the other hand, the hydrolysis
reaction of a composite with the same initial composition
but milled for a longer period of time proceeds at a much
faster rate after the initial burst. In that later case, the reac-
tion yield reaches ∼0.8 after 30 min of reaction. Since this
compound contains only 20.3 mol% of CaH2, such a high
reaction yield cannot be explained by assuming that only
CaH2 is hydrolyzed (in that case, the reaction yield would
have been 0.2). Indeed, close to 75% of MgH2 present in
the MgH2 + 20.3 mol% CaH2 nanocomposite milled 10 h
must have reacted to get a reaction yield close to 0.8. This
is quite surprising considering that the solution pH reaches
values larger than 11.0 after only one or 2 min of reaction.
In conventional materials, reduced solubility of Mg2+ and
formation of a passivation layer effectively stop the hydrol-
ysis reaction after a few minutes of reaction. However, in
nanocrystalline materials with the same composition, the re-
action proceeds almost to completion even at this low solu-
The evolution of crystallite size of the ␤-MgH2 phase
as a function of milling time is shown in Table 3. Starting
from the polycrystalline material, an important reduction of
crystallite size is achieved after the first hour of milling.
The final crystallite size is reached after 7.5 h of milling.
It is interesting to note that higher reaction yield and faster
reaction kinetics are reached as crystallite sizes are reduced
during extensive milling.
The idea that compounds with similar composition but
different structure (nanostructure) can exhibit largely vary-
ing reaction yield and reaction kinetics is best illustrated
by the hydrolysis of less complex systems. A mixture of
MgH2 + 20.3 mol% CaH2 was milled for 5 min and 10 h.
Since Ca is present in the initial mixture in the form of a
hydride, no reaction with MgH2 is expected. This is con-
firmed by the X-ray powder diffraction pattern of the later
compound (not shown) that displays only the characteris-
tic peaks of MgH2 and CaH2. Milling for duration as short
as 5 min does not change the structure of the material and
serves only to mix the two raw components, while milling
for 10 h reduces the crystallite size of both hydrides and fa-
vors the development of an intimate mixture down to the
nanometer level.
bility level of Mg2+
.
It seems that there are two requirements for the hydroly-
sis reaction of MgH2-based compounds to proceed through
near completion. Firstly, MgH2 should be mixed with a Ca
compound, either in the form of a metal or a hydride. In the
first case, prolonged milling of MgH2 with Ca leads to the
in situ formation of CaH2. It is believed that the synergetic
effect occurring in MgH2 +CaH2 milled for 10 h is the same
as that responsible for the higher reaction yield and faster
reaction kinetics observed in MgH2 + Ca nanocomposite.
Secondly, as evidenced by the data of Fig. 5, the mixing
between MgH2 and CaH2 should be thorough and should
occur at the nanometer level.
Hydrolysis curves of MgH2 +20.3 mol% CaH2 are shown
in Fig. 5. The sample milled for 5 min shows a hydrolysis
Table 3
Crystallite size of magnesium hydride as a function of milling time of
MgH2–8.9 mol% Ca mixture
Milling time (h)
Crystallite size (nm)
1
2
5
7.5
15
42
35
24
17
16
The exact nature of the synergistic effect between MgH2
and CaH2 is not clear. Under our experimental conditions,
calcium hydride is more readily hydrolyzed than MgH2 and
this reaction may be beneficial to open up the structure of

J.-P. Tessier et al. / Journal of Alloys and Compounds 376 (2004) 180–185
185
MgH2 and favors the interaction between freshly exposed
MgH2 surface and the electrolyte. Also, the fact that the
initial burst of hydrogen in the most active compound is
followed by a period of time where the hydrogen flow is
significant points to the fact that the passivation layer that is
formed at the surface of MgH2 grains must be different from
that formed on pure MgH2. The presence of CaH2 and Ca2+
ions in the vicinity of MgH2 must influence the composi-
tion, the structure and the permeability of the passivation
layer, thereby allowing the hydrolysis reaction on MgH2
to proceed to a greater extent than observed in the case of
pure MgH2. The nanostructure that develops in the com-
posite upon extensive milling is essential in obtaining high
reaction yields and fast kinetics of the hydrolysis reaction.
MgH2 + 20.3 mol% CaH2 milled during 10 h, the reaction
yield is close to 0.8 after 30 min of hydrolysis. In that later
case, 75% of the MgH2 initially present in the composite is
decomposed after 30 min of hydrolysis.
References
[1] H.I. Schlesinger, H.C. Brown, A.E. Finholt, J.C. Gilbreath, H.R.
Hoekstra, E.K. Hyde, J. Am. Chem. Soc. 75 (1953) 215–219.
[2] S.C. Amendola, M. Binder, S.L. Sharp-Goldman, M.T. Kelly, US
Patent 6534033 B1, 2003.
[3] S.C. Amendola, S.L. Sharp-Goldman, M.S. Janjua, N.C. Spencer,
M.T. Kelly, P.J. Petillo, M. Binder, Int. J. Hydrogen Energy 25 (2000)
969–975.
[4] S. Suda, Y.-M. Sun, B.-H. Liu, Y. Zhou, S. Morimitsu, K. Arai,
N. Tsukamoto, M. Uchida, Y. Candra, Z.-P. Li, Appl. Phys. A 72
(2001) 209.
[5] J.E. Stearns, M.A. Matthews, D.L. Reger, J.E. Collins, Int. J. Hy-
drogen Energy 23 (1998) 469–474.
4. Conclusion
[6] R. Aiello, M.A. Matthews, D.L. Reger, J.E. Collins, Int. J. Hydrogen
Energy 23 (1998) 1103–1108.
[7] R. Aiello, J.H. Sharp, M.A. Matthews, Int. J. Hydrogen Energy 24
(1999) 1123–1130.
[8] J. Huot, G. Liang, R. Schulz, J. Alloys Compd. 353 (2003) L12–L15.
[9] J. Huot, G. Liang, S. Boily, A.V. Neste, R. Schulz, J. Alloys Compd.
293–295 (1999) 495–500.
[10] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solu-
tion, NACE International, Cebelcor, 1974.
[11] J. Huot, G. Liang, R. Schulz, Appl. Phys. A 72 (2001) 187–195.
[12] D.R. Lide (Editor-in-Chief), Handbook of Chemistry and Physics,
77th ed., 1996–1997.
[13] J.J. Reilly, R.H. Wiswall, K.C. Hoffman, C.H. Waide, BNL 18867
(1974).
A systematic investigation of the structure and hydrolysis
characteristics of ball-milled MgH2–Ca and MgH2–CaH2
mixtures was carried out. It was found that when a mixture
of magnesium hydride and calcium is milled, calcium hy-
dride is formed by the transfer of hydrogen from magnesium
hydride. The resulting nanocomposite Mg–MgH2–CaH2
has faster hydrolysis kinetics and higher yield compared
to ball-milled pure magnesium hydride. Thus, calcium
addition allows a faster and more complete hydrolysis re-
action of magnesium hydride. Even if calcium hydride is
formed during milling, faster hydrolysis reaction and higher
yield are achieved when calcium hydride is used directly
as starting component instead of calcium. In the case of
[14] A.R. Yavari, J.F.R. de Castro, G. Vaughan, G. Heunen, J. Alloys
Compd. 353 (2003) 246–251.