Effect of Nb2O5 content on hydrogen reaction kinetics of Mg

Article abstract of DOI:10.1016/S0925-8388(03)00530-9

The effect of Nb₂O₅ concentration on the kinetics of the magnesium hydrogen sorption reaction at 300°C and 250°C is investigated. Absorption and desorption kinetics of nanocrystalline magnesium catalyzed with 0.05, 0.1, 0.2, 0.5, and 1 mole% Nb₂O₅ are determined at 300 and 250°C. Fastest kinetics are obtained using 0.5 mole% Nb₂O₅. At 300°C, absorption and desorption of 7 wt.% of hydrogen are facilitated in 60 and 90 s, respectively. At 250°C, more than 6 wt.% are absorbed in 60 s and desorbed again in 500 s. Respective reaction rates and activation enthalpies are calculated. The activation energy for desorption varies exponentially with Nb₂O₅ concentration and reaches a limit of 61 kJ/mole at 1 mole%. The rate-limiting step for a particular parameter set is deduced by fitting the experimentally obtained transformed phase fraction and time data with analytical kinetics rate expressions. The results show that there is a change in the rate-limiting step with catalyst content.

Full text of DOI:10.1016/S0925-8388(03)00530-9

Journal of Alloys and Compounds 364 (2004) 242–246
Effect of Nb2O5 content on hydrogen reaction kinetics of Mg
Gagik Barkhordarian∗, Thomas Klassen, Rüdiger Bormann
Institute for Materials Research, GKSS Research Centre, Geesthacht GmbH, Max-Planck-Strasse, D-21502 Geesthacht, Germany
Received 18 April 2003; accepted 29 April 2003
Abstract
The effect of Nb2O5 concentration on the kinetics of the magnesium hydrogen sorption reaction at 300 ◦C and 250 ◦C is investigated.
Absorption and desorption kinetics of nanocrystalline magnesium catalyzed with 0.05, 0.1, 0.2, 0.5, and 1 mole% Nb2O5 are determined
at 300 and 250 ◦C. Fastest kinetics are obtained using 0.5 mole% Nb2O5. At 300 ◦C, absorption and desorption of 7 wt.% of hydrogen are
facilitated in 60 and 90 s, respectively. At 250 ◦C, more than 6 wt.% are absorbed in 60 s and desorbed again in 500 s. Respective reaction
rates and activation enthalpies are calculated. The activation energy for desorption varies exponentially with Nb2O5 concentration and reaches
a limit of 61 kJ/mole at 1 mole%. The rate-limiting step for a particular parameter set is deduced by fitting the experimentally obtained
transformed phase fraction and time data with analytical kinetics rate expressions. The results show that there is a change in the rate-limiting
step with catalyst content.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Hydrogen absorbing materials; Metal hydrides; Gas–solid reactions; Kinetics
1. Introduction
have a substantial effect, thus capacity is only slightly re-
duced.
Hydrogen is the ideal means of storage, transport and con-
version of energy for a comprehensive clean-energy concept.
Regarding the use of hydrogen as fuel for the zero-emission
vehicle, the main problem is the storage of hydrogen. Metal
hydrides offer a safe alternative to storage in compressed
or liquid form. In addition, metal hydrides have the high-
est storage capacity by volume. Mg hydride has also a high
storage capacity by weight and is therefore favored for mo-
bile applications. However, light metal hydrides have not
been considered competitive because of their rather sluggish
sorption kinetics. Filling a tank could take several hours.
In recent years, significant progress has been made using
nanocrystalline Mg produced by high energy milling [1–8]
and adding suitable catalysts [9,10]. For technical applica-
tion, sufficiently fast kinetics have been achieved at 300 ◦C
[11]. Among the catalysts used in different studies, transi-
tion metal oxides show a high potential [9–11]. In addition,
oxides are cheap, and amounts as low as 0.2 mole% already
In a previous paper [11], the superior catalytic effect of
Nb2O5 was demonstrated and compared to other metal ox-
ide catalysts. In the present study, the influence of Nb2O5
content on the hydrogen sorption kinetics of magnesium is
investigated, at reaction temperatures of 250 and 300 ◦C.
The final goal is to identify the optimum amount of catalyst,
which is required to guarantee superior kinetics, while giv-
ing maximum reversible storage capacity. Different amounts
ranging from 1 down to 0.05 mole% are investigated. The
rate-limiting step for transformation kinetics is determined
by fitting various analytical kinetic rate expressions to the
obtained experimental data.
2. Experimental
The initial MgH2 powder was purchased from Gold-
schmidt AG, with 95% purity, the rest being magnesium.
Milling was carried out in a Fritsch P5 planetary ball mill
and a ball to powder weight ratio of 400 g/40 g. In the
first phase, this powder was milled for 20 h, yielding a
homogeneous brownish colored magnesium hydride pow-
der. Afterwards, 0.05, 0.1, 0.2, 0.5, and 1 mole% Nb2O5
∗
Corresponding author. Tel.: +49-415-287-2541;
fax: +49-415-287-2636.
E-mail address: gagik.barkhordarian@gkss.de (G. Barkhordarian).
0925-8388/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0925-8388(03)00530-9

G. Barkhordarian et al. / Journal of Alloys and Compounds 364 (2004) 242–246
243
were added to the milled magnesium powder and the
mixture was milled for additional 100 h. Kinetics were
characterized at 250 or 300 ◦C under 8.4 bar hydrogen
or vacuum during absorption and desorption, respectively,
using a volumetric Sieverts apparatus designed by Hy-
dro Quebec/HERA Hydrogen Storage Systems [12]. All
milling experiments, preparation and transfer for character-
ization has been done under argon atmosphere to exclude
any contact of the powder with ambient air. The reaction
rates were calculated using the time and hydrogen content
values between 20 and 80% of the respective maximum
capacity.
3. Results
Fig. 2. Effect of Nb2O5 concentration on desorption kinetics of MgH2 at
a reaction temperature of 300 ◦C and under vacuum conditions.
Fig. 1 shows the first 60 s of absorption of magnesium with
different Nb2O5 contents at 300 ◦C and 8.4 bar hydrogen
pressure. Although differences are rather small, absorption
kinetics is slightly better using higher Nb2O5 content. In
general, about 7 wt.% of hydrogen is absorbed within 1 min.
Differences in the desorption kinetics are more pro-
nounced (Fig. 2). In general, an increase of Nb2O5 content
again has a positive effect on desorption kinetics. How-
ever, a saturation limit is reached for 0.5 mole% of Nb2O5,
and kinetics does not accelerate further, if more catalyst is
present. With 0.5 mole% Nb2O5, magnesium hydride des-
orbs the full 7 wt.% of hydrogen within 90 s. To our knowl-
edge, this is the fastest desorption kinetics for magnesium
hydride reported so far.
It is worth noting that low catalyst contents also have
a significant effect on kinetics: e.g. with only 0.1 mole%
Nb2O5, full desorption takes place in 240 s. Even using only
0.05 mole% of Nb2O5, full desorption is possible within
500 s.
Similar results are obtained at 250 ◦C (Figs. 3 and 4).
Absorption kinetics is almost independent of the amount of
Fig. 3. Effect of Nb2O5 concentration on absorption kinetics of magnesium
at a reaction temperature of 250 ◦C and a hydrogen pressure of 8.4 bar.
Fig. 1. Effect of Nb2O5 concentration on absorption kinetics of magnesium
at a reaction temperature of 300 ◦C and a hydrogen pressure of 8.4 bar.
Fig. 4. Effect of Nb2O5 concentration on desorption kinetics of MgH2 at
a reaction temperature of 250 ◦C and under vacuum conditions.

244
G. Barkhordarian et al. / Journal of Alloys and Compounds 364 (2004) 242–246
Table 1
Reaction rates calculated between 20 and 80% of maximum capacity
Nb2O5 mole%
(mole%)
Absorption rate at 250 ◦C
(wt.%/s)
Absorption rate at 300 ◦C
(wt.%/s)
Desorption rate at 250 ◦C
(wt.%/s)
Desorption rate at 300 ◦C
(wt.%/s)
0.05
0.1
0.2
0.5
0.23
0.3
0.3146
0.38
0.146
0.152
0.185
0.205
0.00269
0.00841
0.01154
0.0137
0.0157
0.03132
0.05
0.102
catalyst used, and even 0.05 mole% leads to fast absorption
of about 6 wt.% of hydrogen within 1 min.
expose the specimen to ambient air at any time, otherwise
the formation of additional oxides can not be excluded. In
this regard, it is also worth noting that not only oxides but
also nitrides show a good catalytic effect [10], which may
add to the observed catalytic effect, if commercially pure
metals are used.
In one of our previous studies [10], we have already shown
that V2O5 has a catalytic effect, while ultrapure V, processed
in parallel under the same conditions under purified argon at-
mosphere, only shows a rather small influence. What is even
more striking: if the MgH2 milled with ultrapure metallic V
is exposed to limited amounts of air, kinetics significantly
improve [10]. Obviously, oxygen (or nitrogen) does indeed
play a role in the catalysis of the sorption reaction.
In order to elucidate the mechanism of catalysis of tran-
sition metal oxides, tentative activation energies for the
desorption reaction of magnesium catalyzed with different
Nb2O5 contents are calculated using the kinetic rate con-
stants at the two different temperatures (Fig. 5). Obviously,
the activation energy varies exponentially with the Nb2O5
content and reaches a limit of 62 kJ/mole at 0.2 mole%
Nb2O5. As will be discussed in the next paragraph, this can
be explained by a change in the rate-limiting step of the
desorption kinetics. The calculated value of 62 kJ/mole is in
good agreement with the results for pure metallic Nb [13],
where an activation energy of 64 kJ/mole was determined.
This may be taken as a further indication that identical
phases are present and identical processes take place.
Desorption kinetics are again more critical and the opti-
mum kinetics are achieved with catalyst contents of 0.5 or
1 mole%. Analogous to the results at 300 ◦C, the desorption
curve for 0.2 mole% Nb2O5 is only slightly slower by about
20%. All reaction rates are listed in Table 1 for quantita-
tive evaluation. Compared to other catalysts [13], Nb2O5 is
superior also at 250 ◦C for both absorption and desorption.
4. Discussion
The results demonstrate that even small amounts of ox-
ide catalysts result in a large improvement of the reaction
kinetics.
The calculated kinetic rates in Table 1 show that the ab-
sorption rate at 250 ◦C is nearly two times faster than at
300 ◦C, independent of the catalyst content. This result can
simply be explained by the higher thermodynamic driving
force for the absorption reaction at lower temperatures at
identical hydrogen pressures, according to the Van ‘t Hoff
equation. This is a further proof that the kinetic barriers are
efficiently reduced by Nb2O5 additions. Thus, the activation
energy is lower and the barriers can be overcome also at
lower temperatures. Consequently, the influence of the ther-
modynamic driving force increases and the reaction rate at
250 ◦C is faster than at 300 ◦C.
Compared to studies using pure metal catalysts [13], strik-
ing differences are observed with respect to the amount of
catalyst required for equally fast kinetics. For full desorption
within 250 s at 300 ◦C, 5 mole% of pure Nb is necessary,
whereas only 0.1 mole% of Nb2O5 is sufficient to desorb
even more hydrogen in the same time. At first glance, this
already proves that the catalytic effect of the metals in a pure
state is much weaker than in form of the respective oxide.
However, the question may be posed, whether pure metallic
Nb is catalytically active at all. Considering the high affin-
ity of niobium to oxygen, it may be reasonable to conclude
that at least a large part—if not all—of the reported cat-
alytic effect of pure metals is due the presence of oxides
in the samples. If one assumes, that the Nb used in other
studies has technical purity, i.e. the material contains about
1 wt.% oxygen, an amount of 5 mole% Nb would include
0.0581 mole% of Nb2O5. In addition, if the catalytic effect
of pure metals is studied, special care has to be taken not to
Fig. 5. Variation of the activation energy for the desorption reaction with
Nb2O5 content.

G. Barkhordarian et al. / Journal of Alloys and Compounds 364 (2004) 242–246
245
Table 2
Rate limiting step in desorption reactions with various Nb2O5 concentrations
Used kinetic equation
Best fit found for reaction
Rate limiting step
Interface controlled
1 − (1 − α)1/2 = kt
At 250 ◦C and 0.2 mole%
At 300 ◦C and 0.2 mole%
At 250 ◦C and 0.5 mole%
At 300 ◦C and 0.5 mole%
At 250 ◦C and 1 mole%
At 250 ◦C and 1 mole%
[−ln(1 − α)]1/2 = kt
1 − (2α/3) − (1 − α)2/3 = kt
α = kt
–
–
–
–
At 300 ◦C and 0.05 mole%
At 300 ◦C and 0.1 mole%
Surface controlled
[−ln(1 − α)]1/3 = kt
At 250 ◦C and 0.05 mole%
At 250 ◦C and 0.1 mole%
Three dimensional growth controlled
The reaction of metallic elements with hydrogen consists
of five distinct steps, physisorption, chemisorption (recom-
bination), surface penetration, diffusion, hydride formation
(decomposition). The slowest step determines the overall ki-
netic rate, and the experimentally measured kinetics repre-
sents the kinetics of the slowest part in each phase of the
reaction. The kinetic curve at each step has a characteris-
tic form, which can be formulated as equations relating the
transformed phase fraction to time. Thus, it is possible to
deduce the rate-limiting step of the kinetics, if a good fit
of experimental data with a specific kinetic equation can be
obtained. Using the above mentioned method, the rate lim-
iting step of the reactions for different Nb2O5 contents are
identified and the results are listed in Table 2. In detail, the
results can be interpreted as follows:
rate-determining anymore and the reaction becomes in-
terface controlled again.
At both temperatures investigated, i.e. at 250 and at
300 ◦C, the rate-limiting step changes for catalyst additions
of around 0.2 mole%. However, at 250 ◦C, the rate-limiting
process changes from ‘nucleation and growth’ to ‘interface-
controlled’, while at 300 ◦C, rate-limiting process changes
from ‘surface-contolled’ to ‘interface-controlled’. The
change in the rate-limiting process at 0.2 mole% catalyst is
in good agreement with the saturation in activation enthalpy
(Fig. 5).
5. Conclusion
The effect of Nb2O5 catalyst concentration on the kinet-
ics of hydrogen sorption reaction of nanocrystalline Mg is
investigated at 250 and 300 ◦C. Fastest kinetics are obtained
using 0.5 mole% Nb2O5. At 300 ◦C, absorption and desorp-
tion of 7 wt.% of hydrogen are facilitated in 60 and 90 s, re-
spectively. At 250 ◦C, more than 6 wt.% is absorbed in 60 s
and desorbed again in 500 s. Absorption kinetics are nearly
independent of the catalyst content, even 0.05 mole% Nb2O5
yield the full hydrogen absorption of 7 wt.% within 60 s.
Tentative activation energies for different catalyst contents
were calculated, and it was found that the activation energy
decreases exponentially with catalyst additions, reaching the
lower saturation limit of 61 kJ/mole at 0.2 mole%. For this
catalyst content, a change in the rate-limiting process was
detected by fitting the experimentally transformed fraction
and time data using analytical rate expressions.
1. At 250 ◦C and with catalyst contents lower than 0.2
mole%, the reaction is three dimensional growth con-
trolled. This can be explained by the slower diffusion
at this low temperature, as well as the slow hydrogen
draining due to a lack of catalyst and longer diffusion
paths.
2. At 250 ◦C and with catalyst contents above 0.2 mole%,
the recombination rate is high enough and the kinetic
rate-limiting step is changed to interface-controlled. As
the Nb2O5 content increases, the catalytic effect reaches
a limit at 1 mole%, see Figs. 3 and 5, because the recom-
bination rate of hydrogen atoms is not the rate determin-
ing step anymore.
3. For a reaction temperature of 300 ◦C, with a catalyst con-
tent lower than 0.2 mole%, the kinetic is clearly surface
controlled. Due to easier diffusion of hydrogen atoms at
300 ◦C, transport of hydrogen does not play an important
role anymore. Now, the reaction rate is only determined
by the slow gas–solid reaction due to the low catalyst
content.
References
[1] R. Schulz, S. Boily, L. Zaluski, A. Zaluska, P. Tessier, J.O.
Ström-Olsen, in: Proceedings of the International Conference on
Composite Materials and Energy—ENERCOMP 95 (8–10 May
1995)—Innovations in Metallic Materials, Technomic Publishing,
Lancaster, PA, 1995, p. 529.
4. Again at 300 ◦C, if the catalyst content reaches values
of 0.2 mole% or more, the recombination rate of hy-
drogen atoms increases up to a limit, where it is not

246
G. Barkhordarian et al. / Journal of Alloys and Compounds 364 (2004) 242–246
[7] G. Liang, S. Boily, J. Huot, A. Van Neste, R. Schulz, J. Alloys
[2] P. Tessier, L. Zaluski, Z.-H. Yan, M.L. Trudeau, R. Bormann, R.
Schulz, J.O. Ström-Olsen, Mater. Res. Soc. Symp. Proc. 286 (1993)
209.
Comp. 268 (1998) 302.
[8] K.J. Gross, P. Spatz, A. Züttel, L. Schlapbach, J. Alloys Comp. 240
[3] M.L. Trudeau, R. Schulz, L. Zaluski, S. Hosatte, D.H. Ryan, C.B.
Doner, P. Tessier, J.O. Ström-Olsen, A. Van Neste, in: P.H. Shingu
(Ed.), Proceedings of an International Symposium on Mechanical
Alloying, Trans Tech Publications, Kyoto, 1991, p. 537.
[4] L. Zaluska, S. Hosatte, P. Tessier, D.H. Ryan, J.O. Ström-Olsen,
M.L. Trudeau, R. Schulz, Z. Phys. Chem. 183 (1994) 45.
[5] L. Zaluski, A. Zaluska, P. Tessier, J.O. Ström-Olsen, R. Schulz, J.
Alloys Comp. 227 (1995) 53.
(1996) 206.
[9] W. Oelerich, T. Klassen, R. Bormann, J. Alloys Comp. 315 (2001)
237.
[10] W. Oelerich, T. Klassen, R. Bormann, J. Alloys Comp. 322 (2001)
L5–L9.
[11] G. Barkhordarian, T. Klassen, R. Bormann, Scripta Mater. 49 (2003)
213.
[12] R. Schulz, S. Boily, J. Huot, Can. Pat. 2207149, 1999.
[13] J. Huot, J.F. Pelletier, L.B. Lurio, M. Sutton, R. Schulz, J. Alloys
Comp. 348 (2003) 319.
[6] G. Liang, S. Boily, J. Huot, A. Van Neste, R. Schulz, J. Alloys
Comp. 267 (1998) 302.