Interaction of hydrogen with an Mg-Al alloy

Article abstract of DOI:10.1016/j.jallcom.2005.01.119

The interaction of hydrogen with an Mg-Al alloy pre-exposed to air have been studied with in situ time resolved X-ray powder diffraction. Phase fractions as a function of time are derived from series of consecutive diffraction patterns allowing kinetic analysis. The apparent activation energy for dehydrogenation of the Mg-Al alloy is found to be 160 kJ/mol. This is not significantly higher than for pure and fully activated Mg. It is suggested that the addition of Al improves the resistance towards oxygen contamination.

Full text of DOI:10.1016/j.jallcom.2005.01.119

Journal of Alloys and Compounds 404–406 (2005) 323–326
Interaction of hydrogen with an Mg–Al alloy
A. Andreasena,b,∗, M.B. Sørensenc, R. Burkarlc, B. Møllerc, A.M. Molenbroekd,
A.S. Pedersena, J.W. Andreasena,e, M.M. Nielsene, T.R. Jensenc,∗∗
a
Materials Research Department, Risø National Laboratory, DK-4000 Roskilde, Denmark
b
Interdisciplinary Research Center for Catalysis, Department of Chemical Engineering, Technical University of Denmark,
DK-2800 Lyngby, Denmark
iNano, Department of Chemistry, Aarhus University, DK-8000 Aarhus C, Denmark
Haldor Topsøe A/S, DK-2800 Kgs. Lyngby, Denmark
Danish Polymer Centre, Risø National Laboratory, DK-4000 Roskilde, Denmark
c
d
e
Received 31 May 2004; received in revised form 25 January 2005; accepted 31 January 2005
Available online 11 July 2005
Abstract
The interaction of hydrogen with an Mg–Al alloy pre-exposed to air have been studied with in situ time resolved X-ray powder diffraction.
Phase fractions as a function of time are derived from series of consecutive diffraction patterns allowing kinetic analysis. The apparent
activation energy for dehydrogenation of the Mg–Al alloy is found to be 160 kJ/mol. This is not significantly higher than for pure and fully
activated Mg. It is suggested that the addition of Al improves the resistance towards oxygen contamination.
© 2005 Elsevier B.V. All rights reserved.
PACS: 61.10.Nz; 81.05.Bx; 82.20.Pm
Keywords: Intermetallics; Hydrogen storage materials; Gas–solid reactions; X-ray diffraction
1. Introduction
Desorption of hydrogen leads to complete reaction forming
an Mg–Al compound, suggesting reversibility upon hydro-
genation/dehydrogenation. These findings are in agreement
with those of Bouaricha et al. [3]. Moreover, the addition of
aluminum may add improved heat transfer properties to the
hydride bed. This has prompted us to study the interaction of
hydrogen with an Mg–Al alloy.
Magnesium hydride has a high theoretical gravimetric
hydrogen density (7.6 wt.%) but, suffers from several draw-
backs e.g.: (i) thermodynamics dictate heating to above 280–
300 ◦C for desorption of hydrogen from MgH2, making it
unsuitable for low-temperature applications, (ii) kinetics of
hydrogenation/dehydrogenation may be regarded as slow and
(iii)magnesiumisverysensitivetogaseousimpuritiessuchas
oxygen, creating an oxide shell retarding the kinetics. Ther-
modynamics and kinetics may be improved to some degree
by alloying. The price being a reduced hydrogen capacity.
Alloying with Al have been reported to improve both
thermodynamics [1,2] and kinetics [3]. From X-ray powder
diffraction studies of the hydrogenated Mg–Al compound
disproportionation into MgH2 and Al could be concluded [1].
2. Experimental
The Mg–Al alloy subject to our investigations was pre-
pared by arc melting approximately 5 g of a mixture of
magnesium and aluminum according to the stoichiometry
Mg17Al12 (␥-phase [4]) in an Edmund Buhler Arc Melting
system. The magnesium (7.9 mm rod from Goodfellow) and
aluminum (5–15 mm ingots from Sigma–Aldrich) used were
99.9% purity. The Mg–Al sample was melted repeatedly in an
Argon atmosphere until the sample appeared homogeneous.
Subsequently, the Mg–Al sample was ball milled for 10 min
using WC balls.
∗
∗∗Corresponding author. Tel.: +45 4677 5776; fax: +45 4677 5758.
Co-corresponding author. Tel.: +45 8942 3894; fax: +45 8619 6199.
E-mail addresses: anders.andreasen@risoe.dk (A. Andreasen);
trj@chem.au.dk (T.R. Jensen).
0925-8388/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2005.01.119

324
A. Andreasen et al. / Journal of Alloys and Compounds 404–406 (2005) 323–326
Fig. 2. XRPD of the as-prepared ball milled sample.
␥-phase (Mg17Al12). However, content of the ⑀ (Mg42Al58
[4]) and ␤-phase (Mg2Al3 [4]) can not be ruled out. Using
the Scherrer equation the crystallite size is approximated to
be 10 nm.
The initial hydrogenation of the ball milled Mg–Al al-
loy showed almost complete hydrogenation within approx-
imately 17 h. The total hydrogen uptake was approximately
3.1 wt.% corresponding to an approximate stoichiometry of
Mg0.41Al0.59 assuming hydrogenation of Mg only. This sto-
ichiometry is close to both that of the ␤ and the ⑀-phase. The
hydrogen uptake is somewhat lower than expected from the
initial stoichiometry. The Mg–Al alloy is transformed com-
pletely into MgH2 and metallic Al cf. Fig. 3 at t = 0. The
peak width is reduced probably due to sintering at the ele-
vated temperature during hydrogenation and corresponds to
a crystallite size of approximately 70 nm.
The decomposition of MgH2 + Al studied with in situ time
resolved XRPD is illustrated by 72 consecutive XRPD pat-
terns in Fig. 3 showing the disappearance of the MgH2 (1 1 0),
(1 0 1), (2 2 0) and the Al (1 1 1) and (2 0 0) reflections and the
corresponding appearance of several reflections in the range
2θ = 35–43 ◦C of an Mg–Al alloy. During the heating of the
sample the MgH2 and Al reflections shift towards lower 2θ
values due to the thermal expansion of the crystal lattices.
The phase fraction of MgH2 have been calculated from
the in situ XRPD data by numerically integrating the (1 1 0)
reflection of MgH2. Dehydrogenation curves i.e. phase frac-
tion of MgH2 versus time have been constructed for all ex-
periments using normalized integrated intensities. The results
are shown in Fig. 4. All dehydrogenation curves show a sig-
moidal shape and faster dehydrogenation with higher tem-
peratures, although, the shape of the dehydrogenation curve
at 390 ◦C seems to deviate slightly from the others.
Fig. 1. Schematic drawing illustrating the experimental setup.
The ball milled sample of Mg–Al alloy was initially hy-
drogenated in a Sartorius high pressure balancing unit de-
scribed in detail elsewhere [5]. The sample was hydrogenated
by applying a hydrogen pressure (99.9997% purity from Air
Liquide) of approximately 30 bar and a temperature of ap-
proximately 350 ◦C.
The X-ray instrument used for time resolved in situ X-
ray powder diffraction (XRPD) was built around a Rigaku
¯
˚
rotating anode (Cu K␣12 radiation, λ = 1.5418 A, 50 kV and
300 mA). The intensity of the diffracted beam was recorded
with a curved position sensitive detector, INEL CPS 120,
covering 120◦ in 2θ with a resolution of ca. 0.03◦. The ac-
quisition time per powder pattern was chosen to be 150 s.
The in situ reactor cell is described in detail elsewhere [6].
The powder sample was loaded between plugs of quartz glass
wool in a quartz capillary tube (0.7 mm o.d.) in order to fix
the bed and allow a gas flow (Ar, 10 mL/min) through the
sample during data acquisition as shown in Fig. 1. The sam-
ples were heated by a stream of hot nitrogen gas (20 ◦C/min)
to a constant temperature in the range 350–400 ◦C and the
dehydrogenation was followed under isothermal conditions.
A Johnson–Mehl–Avrami (JMA) type nucleation and
growth rate equation has been fitted to the dehydrogenation
curves in Fig. 4
α(t) = exp (−(kt)n)
(1)
3. Results and discussion
where α(t) is the time-dependent phase fraction. Assuming
an Arrhenius expression for the rate constant k cf. Eq. (2)
The XRPD of the as prepared ball milled sample is shown
in Fig. 2. No metallic magnesium or aluminum was observed.
The powder pattern shows relatively wide reflections and the
peak positions correspond well with those observed for the
ꢀ
ꢁ
EA
RT
k = A exp
−
(2)

A. Andreasen et al. / Journal of Alloys and Compounds 404–406 (2005) 323–326
325
Fig. 3. Time resolved in situ XRPD of the dehydrogenation of MgH2 + Al. The graphic consists of 72 consecutive diffraction patterns stacked chronologically
from left to right. Bright areas correspond to a high detector count rate (reflections), whereas dark areas correspond to low detector count rates (background).
The isothermal reaction temperature is T = 400 ◦C and acquisition time is t = 150 s.
where A is a pre-exponential factor, EA the apparent activa-
tion energy and R is the universal gas constant, we can extract
the apparent activation energy by plotting ln k versus 1/T cf.
Fig. 5 and find EA/R as the slope. As shown in the figure the
data points fit the Arrhenius expression Eq. (2) fairly well.
The apparent activation energy is found to be 160 kJ/mol
H2. This value is very close to the activation barrier of 160–
166 kJ/mol H2 for the dehydrogenation of pure magnesium
hydride recently found by Fernandez and Sanchez [7,8]. The
sample in this investigation differs from the one investigated
by Fernandez and Sanchez not only in composition but
also in pretreatment. Special precautions were taken not to
oxidize the sample in Ref. [7,8] prior to kinetic measure-
ments and the sample was also fully activated by several
adsorption/desorption cycles before measurements. In this
investigation the sample has been exposed to air both before
the initial hydrogenation and before the time resolved XRPD
study. Recently, we investigated the dehydrogenation of pure
magnesium [9] with the same setup used here and with the
samepretreatmentprocedure. Wefoundasubstantialincrease
in the apparent activation enthalpy of dehydrogenation to
approximately 250–300 kJ/mol (probably due to oxygen con-
tamination/magnesium oxide formation). This comparison
suggests that alloying with Al creates a compound less sensi-
tive towardsoxygencontamination which requireslittle ifany
pretreatment in order to activate the sample. No crystalline
magnesium oxide is observed neither in the as-prepared
ball milled sample nor in the hydrogenated/dehydrogenated
sample. However, oxygen was detected using Energy
Dispersive X-ray Spectroscopy suggesting the presence
of an X-ray amorphous oxide. This is consistent with the
observations of Scotto-Sheriff et al. [10]. Compared to a
crystalline oxide layer, the presence of an amorphous oxide
layer is likely to offer improved hydrogen diffusion [11,12].
There may be other kinetic improvements associated
with alloying Mg with Al. The grain boundaries between
MgH2/Al/Mg–Al may provide favorable diffusion paths as
proposed for Mg/Mg2Cu [13].
Fig. 4. Experimental dehydrogenation curves for pure MgH2 in MgH2 + Al
determined by integrated intensity of the MgH2 (1 1 0) reflection from time
resolved in situ XRPD data.

326
A. Andreasen et al. / Journal of Alloys and Compounds 404–406 (2005) 323–326
cellence Towards a hydrogen based society. TRJ thanks The
Danish National Research Council for a Steno stipend. R.
Christensen (Haldor Topsøe A/S) is greatly acknowledged
for assisting with in situ XRPD experiments.
References
[1] A. Zaluska, L. Zaluski, J. Stro¨m-Olsen, Appl. Phys. A 72 (2001) 157–
165.
[2] M. Bououdina, Z.X. Guo, J. Alloys Compd. 336 (2002) 222–
231.
[3] S. Bouaricha, J.P. Dodelet, D. Guay, J. Huot, S. Boily, R. Schulz, J.
Alloys Compd. 297 (2000) 282–293.
[4] H.L. Su, M. Harmelin, P. Donnadieu, C. Baetzner, H.J. Seifert, H.L.
Lukas, G. Effenberg, F. Aldinger, J. Alloys Compd. 247 (1997) 57–
65.
[5] A. Pedersen, J. Kjøller, B. Larsen, B. Vigeholm, Int. J. Hydrogen En-
ergy 8 (3) (1983) 205–211.
[6] B. Clausen, G. Steffensen, B. Fabius, J. Villadsen, R. Feidenhans’l, H.
Topsøe, J. Catal. 132 (1991) 524–535.
Fig. 5. Arrhenius plot of the logarithmic rate constant (derived from fitting
the dehydrogenation data from time resolved in situ XRPD) vs. reciprocal
temperature for the dehydrogenation of MgH2 + Al.
[7] J.F. Fernandez, C.R. Sanchez, J. Alloys Compd. 340 (2002) 189–
198.
[8] J.F. Fernandez, C.R. Sanchez, J. Alloys Compd. 356–357 (2003) 348–
352.
4. Conclusion
[9] T.R. Jensen, A. Andreasen, J.W. Andreasen, T. Vegge, K. Sta˚hl, F.
Besenbacher, A. Molenbroek, M.M. Nielsen, A.S. Pedersen, Int. J.
Hydrogen Energy, submitted for publication.
[10] S. Scotto-Sheriff, E. Darque-Ceretti, G. Plassart, M. Aucouturier, J.
Mater. Sci. 34 (1999) 5081–5088.
[11] L. Schlapbach (Ed.), Hydrogen in intermetallic compounds II. Surface
and dynamic properties, applications, Vol. 67 of Topics in Applied
Physics, Springer-Verlag, 1992.
The interaction of hydrogen with an Mg–Al alloy has
been studied by in situ time resolved X-ray powder diffrac-
tion. The activation energy of dehydrogenation is found to be
160 kJ/mol. Further, we suggest that Mg–Al is less sensitive
than Mg to oxygen contamination.
[12] A.B. Belonoshko, A. Rosengren, Q. Dong, G. Hultquist, C. Leygraf,
Phys. Rev. B. 69 (2004) 024302.
Acknowledgments
[13] M. Au, J. Wu, Q. Wang, Int. J. Hydrogen Energy 20 (1995) 141–
150.
This work has received financial support from The Dan-
ish Technical Research Council through the Center of Ex-