Synergistic effects of co-dopants on the dehydrogenation kinetics of sodium aluminum hydride

Article abstract of DOI:10.1016/j.jallcom.2004.08.069

A systematic analysis of the effect of co-dopants on the dehydrogenation kinetics of freshly doped and ball milled NaAlH₄ samples was carried out with chlorides of Ti, Zr and Fe as the catalysts. Numerous samples of NaAlH₄ when co-doped with binary and ternary combinations of Ti, Zr and Fe at 4 mol% total catalyst content exhibited synergistic behavior, with respect to improving the dehydrogenation kinetics of the first decomposition reaction (i.e., NaAlH₄ → Na₃AlH₆) over that of a sample of NaAlH₄ doped with 4 mol% Ti or Zr as single catalysts. In general, the dehydrogenation kinetics improved with the amount of Ti present in a co-doped sample, whether it was a binary or ternary system, with the top five performers all having at least 2 mol% Ti as one of the co-dopants. The binary combination of 3 mol% Ti-1 mol% Fe exhibited the best synergistic performance, with dehydrogenation rates 3.7, 2.0 and 1.5 times that of 4 mol% Ti alone at 90, 110 and 130 °C, respectively. The binary co-doped Zr-Fe system exhibited more pronounced synergistic effects than did the binary co-doped Ti-Fe system; however, their performance was always worse because Ti is a better single catalyst than Zr. The least synergism was exhibited by the binary co-doped Ti-Zr system, where it was surmised that the superior electron sharing ability of electron rich Fe was responsible for it being a better promoter of Ti and Zr than Zr was of Ti. This supposition was further supported by the systematic trends observed with the ternary co-doped systems, with their synergistic effects seemingly limited to binary combinations of the Ti-Fe and Zr-Fe systems. The effects of Ti, Zr and Fe as co-dopants on the second decomposition reaction (i.e., Na₃AlH₆ → NaH) were not as pronounced as their effects on the first reaction; but synergisms were still observed, especially with all three binary Zr-Fe co-doped systems and to a lesser extent only with the 3 mol% Ti-1 mol% Fe system. A future study will consider the effects of these co-dopants on the dehydrogenation/rehydrogenation kinetics after cycling.

Full text of DOI:10.1016/j.jallcom.2004.08.069

Journal of Alloys and Compounds 391 (2005) 245–255
Synergistic effects of co-dopants on the dehydrogenation kinetics
of sodium aluminum hydride
J. Wanga, A.D. Ebnera, R. Zidanb, J.A. Rittera,∗
a
Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208, USA
b
Savannah River National Laboratory, Aiken, SC 29804, USA
Received 22 August 2004; accepted 24 August 2004
Available online 27 October 2004
Abstract
A systematic analysis of the effect of co-dopants on the dehydrogenation kinetics of freshly doped and ball milled NaAlH4 samples was
carried out with chlorides of Ti, Zr and Fe as the catalysts. Numerous samples of NaAlH4 when co-doped with binary and ternary combinations
of Ti, Zr and Fe at 4 mol% total catalyst content exhibited synergistic behavior, with respect to improving the dehydrogenation kinetics of
the first decomposition reaction (i.e., NaAlH4 → Na3AlH6) over that of a sample of NaAlH4 doped with 4 mol% Ti or Zr as single catalysts.
In general, the dehydrogenation kinetics improved with the amount of Ti present in a co-doped sample, whether it was a binary or ternary
system, with the top five performers all having at least 2 mol% Ti as one of the co-dopants. The binary combination of 3 mol% Ti–1 mol% Fe
exhibited the best synergistic performance, with dehydrogenation rates 3.7, 2.0 and 1.5 times that of 4 mol% Ti alone at 90, 110 and 130 ◦C,
respectively. The binary co-doped Zr–Fe system exhibited more pronounced synergistic effects than did the binary co-doped Ti–Fe system;
however, their performance was always worse because Ti is a better single catalyst than Zr. The least synergism was exhibited by the binary
co-doped Ti–Zr system, where it was surmised that the superior electron sharing ability of electron rich Fe was responsible for it being a
better promoter of Ti and Zr than Zr was of Ti. This supposition was further supported by the systematic trends observed with the ternary
co-doped systems, with their synergistic effects seemingly limited to binary combinations of the Ti–Fe and Zr–Fe systems. The effects of Ti,
Zr and Fe as co-dopants on the second decomposition reaction (i.e., Na3AlH6 → NaH) were not as pronounced as their effects on the first
reaction; but synergisms were still observed, especially with all three binary Zr–Fe co-doped systems and to a lesser extent only with the
3 mol% Ti–1 mol% Fe system. A future study will consider the effects of these co-dopants on the dehydrogenation/rehydrogenation kinetics
after cycling.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Hydrogen storage materials; Nanostructures; Chemical synthesis; High-energy ball milling; Thermal analysis
1. Introduction
doping NaAlH4 with a transition metal, such as Ti, signif-
icantly lowers the temperature at which reasonable rates of
dehydrogenation occur. More importantly, in the same work
[2], they were the first to show that Ti as a dopant causes
reasonable rates of rehydrogenation to occur at temperatures
and pressures suitable for commercial applications. This ex-
citing, pioneering work with NaAlH4 sparked a great deal
of interest worldwide [3–39], because, for the first time, it
revealed that a complex hydride has the potential to be used
as a reversible hydrogen storage material in a fashion similar
to metal hydrides.
Hydrogen storage in a solid-matrix has become the focus
of intense research, because it is considered by many people
to be the only viable option that has any possibility of meet-
ing the recently established performance targets set for auto-
motive applications [1]. One of the more promising classes
of hydrogen storage materials being studied is the complex
hydride, which includes the very popular NaAlH4 system.
Bogdanovic and Schwickardi [2] were the first to show that
At standard conditions, the dehydrogenation of NaAlH4
is thermodynamically favorable, but it is kinetically slow and
∗
Corresponding author. Tel.: +1 803 777 3590; fax: +1 803 777 8265.
E-mail address: ritter@engr.sc.edu (J.A. Ritter).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.08.069

246
J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
takes place at temperatures well above 200 ◦C in a two-step
process involving the following reactions [2]:
TiCl3. It should be noted, however, that prior to this work,
Bogdanovic et al. [20] also reported on the marked effect of
using Ti nanoparticles as a dopant for NaAlH4.
3NaAlH4 → Na3AlH6 + 2Al + 3H2
2Na3AlH6 → 6NaH + 2Al + 3H2
(1)
(2)
Surprisingly, in the midst of this substantial research ef-
fort with the NaAlH4 system, with recent efforts focusing
also on elucidating the mechanism behind Ti promoted re-
versibility of the NaAlH4 system [16], a systematic study
on the use of co-dopants to improve the performance of this
complex hydride has not been carried out. Although the two
aforementioned works revealed interesting and even syner-
gistic effects between Ti–Zr [4] and Ti–Fe [6], they were both
quite limited in their scope of discovery. Therefore, the ob-
jective of this work is to report on the first systematic study
of the doping of NaAlH4 with co-dopants.
The effects of three different metal dopants, namely TiCl3,
FeCl3 and ZrCl4 on the dehydrogenation kinetics of NaAlH4
were studied individually and in binary and ternary combina-
tions of each up to 4 mol% total metal dopant. Temperature
programmed desorption (TPD) was carried out on all samples
to determine the initial effect of the dopant or co-dopants just
after doping and ball milling. Constant temperature desorp-
tion (CTD) was also carried out on selected (most) samples
to determine initial dehydrogenation rates and corresponding
activation energies. In all cases, dehydrogenation kinetics are
reported from freshly doped and balled milled samples. A
future study will report on the effects of co-dopants on the
dehydrogenation kinetics after carrying out several rehydro-
genation/dehydrogenation cycles.
The first reaction releases 3.7 wt% hydrogen, while the sec-
ond reaction releases 1.8 wt%. Note that the theoretical hy-
drogen capacity of NaAlH4 is 7.5 wt%, whereas its useful
capacity is only 5.6 wt%. This difference is due to the for-
mation of NaH in Eq. (2), the decomposition temperature of
which is too high to be useful for hydrogen generation. Fur-
thermore, it is now generally accepted that Ti-doped NaAlH4
provides, atbest, around3 wt%reversiblehydrogenstorageat
temperatures of around 100 ◦C, that this hydrogen essentially
comes from the first reaction depicted in Eq. (1), and that the
effect of Ti-doping on the second reaction depicted in Eq. (2)
is not sufficient to warrant any useful hydrogen at temper-
atures below about 130 ◦C [12,13]. As a result, researchers
have been focusing mainly on improving the performance of
the Ti-doped NaAlH4 system by finding ways to improve the
kinetics of the first reaction, with the second reaction being
only an ancillary consideration. A brief history of some of
the more important findings is summarized below.
Soon after the initial work by Bogdanovic and
Schwickardi [2], where it was shown that solution chemistry
techniquescanbeusedtoeffectivelydopeNaAlH4 withTiCl3
or Ti(BuO)4, several other important advancements occurred
with respect to improving the performance of this model
complex hydride system. For example, Jensen and coworkers
[3,4,9,11,13,15] showed that the decomposition temperature
of Ti-doped NaAlH4 can be lowered even further by using
various grinding and high-energy ball milling processes to
prepare the doped sample. Zidan et al. [4] later revealed a
unique behavior between Ti and Zr when co-doping NaAlH4
with both metals, wherein Ti lowered the decomposition tem-
perature of the first reaction as usual, but now Zr lowered
the decomposition temperature of the second reaction even
lower than when using Ti. A short time later, Bogdanovic et
al. [6] reported a synergistic effect when co-doping NaAlH4
with Ti and Fe, wherein the decomposition temperature of
the first reaction was lowered even further compared to using
either Ti or Fe alone as the dopant. Then, in a more recent
study, Bogdanovic et al. [8] showed that the capacity losses
realized upon cycling can be alleviated by the addition of
a small amount of Al powder to the Ti-doped sample. Di-
rect synthesis of Ti-doped NaAlH4 from NaH, Al and TiCl3
[18], and dry doping techniques [12,13], both utilizing high-
energy ball milling, have also been developed. TiCl3 was also
shown to be the best dopant among a host of metal chlorides
covering numerous transition and other metals and their cor-
responding oxidation states [19]. Probably the most exciting
result of late has been reported by Fictchner et al. [17,39].
They showed that Ti13 clusters as a catalyst precursor pro-
vide the best dehydrogenation kinetics to date when NaAlH4
is doped with similar amounts of these clusters compared to
2. Experimental
TiCl3 (Aldrich, 99.99%, anhydrous), FeCl3 (Aldrich,
99.99%, anhydrous) and ZrCl4 (Aldrich, 99.9%), the cata-
lyst precursors, were used as received. Crystalline NaAlH4
(Fluka) was purified from a THF (Aldrich, 99.9%, an-
hydrous) solution and vacuum-dried while being filtered
through 0.7 ␮m filter paper to remove any residual Al pow-
der and other solids from the solution. The purified NaAlH4
was mixed with the catalyst precursor(s) in THF (a wet dop-
ing procedure) to produce a doped sample containing up
to 4 mol% metal dopant (relative to either Na or Al). The
THF was evaporated while the NaAlH4 and the catalyst were
mixed manually for about 30 min using a mortar and pestle,
or until the sample appeared to be completely dry. A sample
was then ball milled for 2 h using a SPEX 8000 high-energy
mill. In each case, 0.5–1.0 g of sample was subjected to high-
energy ball milling in a 65 cm3 SS vial using a single SS ball
(8.2 g) 1.3 cm in diameter. All procedures were carried out in
a nitrogen glove box.
A thermogravimetric analyzer (TGA) also located in a ni-
trogen glove box was used to determine the dehydrogenation
kineticsintermsofthewt%hydrogenreleasedatatmospheric
pressure using TPD and CTD modes. For TPD runs, the sam-
pleswereheatedto250 ◦Catarampingrateof5 ◦C/minunder
1 atm of He flowing at approximately 60 cm3/min, using an

J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
247
initial 1 min delay to ensure an environment of pure He. For
CTD runs, a similar procedure was followed except that the
samples were heated rapidly to the desired temperature and
then held at that temperature for the desired time. Approxi-
mately 10 mg of sample were used in each TPD or CTD run.
3. Results and discussion
The effects of three different catalyst species (Ti, Zr and
Fe) were studied individually and with all possible combi-
nations of their binary and ternary mixtures. NaAlH4 doped
with 4 mol% metal was chosen as the base condition, because
this amount of dopant generally has been shown to provide a
significant effect on the dehydrogenation kinetics of NaAlH4,
at least with Ti as the dopant [12]. Hence, 15 different cata-
lyst combinations doped at 4 mol% total metal were studied:
3 systems with single catalysts of Ti, Zr and Fe; 9 systems
with binary combinations of Ti, Zr and Fe; and 3 systems with
ternary combinations of Ti, Zr and Fe. For comparison, an un-
doped sample of NaAlH4 was also prepared and tested in the
same manner as the doped samples, along with three samples
of NaAlH4 doped with 1, 2 and 3 mol% Ti. Table 1 displays
the different combinations of the single and co-dopant sys-
tems investigated. TPD runs were carried out on all samples
listed in Table 1; CTD runs were carried out on most, but
not all, samples as indicated by the dashes (–) and explained
later. Representative systems that revealed important effects,
trends and findings are displayed in Figs. 1–8.
Fig. 1. Temperature programmed desorption (TPD) curves at 5 ◦C min−1
for NaAlH4 samples just doped and ball milled with 1, 2, 3 and 4 mol%
catalyst consisting of Ti.
At the outset, it was important to study the effect of the
catalyst concentration on the dehydrogenation kinetics, to
verify that in all cases the kinetics increase with increasing
catalyst concentration. In this way, a synergistic effect like
that reported in the literature could easily be observed from
a TPD or CTD curve exhibiting a better performance with
co-dopants than with single dopants for the same total metal
content, as explained later. Ti was chosen to establish this sin-
gle dopant concentration effect, wherein it was assumed that
similar trends would be exhibited by the Zr and Fe systems.
The TPD results for NaAlH4 catalyzed with 1 to 4 mol%
Ti are shown in Fig. 1. The 4 mol% Ti sample exhibits the
best behavior with respect to the dehydrogenation kinetics,
while the 1 mol% Ti sample has the highest hydrogen
capacity. These trends generally agree with that reported
Table 1
Single doped, and binary and ternary co-doped, NaAlH4 systems studied
during dehydrogenation just after wet doping and ball milling; and activation
energies and pre-exponential factors obtained from fitting initial intrinsic
constant temperature desorption (CTD) rate data to the Arrhenius equation
(Eq. (4)) at 90, 110 and 130 ◦C
NaAlH4 sample
A (wt% h−1
)
EA (kJ mol−1
)
No dopant
1 mol% Ti
2 mol% Ti
3 mol% Ti
4 mol% Ti
4 mol% Zr
4 mol% Fe
–
–
2.08 × 1012
4.85 × 1011
3.26 × 1011
8.95 × 1010
3.26 × 1014
–
89.4
82.7
80.4
75.6
105.6
–
1 mol% Ti, 3 mol% Zr
2 mol% Ti, 2 mol% Zr
3 mol% Ti, 1 mol% Zr
1 mol% Ti, 3 mol% Fe
2 mol% Ti, 2 mol% Fe
3 mol% Ti, 1 mol% Fe
1 mol% Zr, 3 mol% Fe
2 mol% Zr, 2 mol% Fe
3 mol% Zr, 1 mol% Fe
1 mol% Ti, 2 mol% Zr, 1 mol% Fe
1 mol% Ti, 1 mol% Zr, 2 mol% Fe
2 mol% Ti, 1 mol% Zr, 1 mol% Fe
7.01 × 1011
1.76 × 1011
1.72 × 1011
7.15 × 1011
2.29 × 1011
7.05 × 1010
–
1.30 × 1012
2.30 × 1012
3.78 × 1012
8.99 × 1011
1.06 × 1011
83.9
78.1
77.9
84.5
79.1
74.5
–
87.6
88.9
80.7
85.2
76.0
Fig. 2. Temperature programmed desorption (TPD) curves at 5 ◦C min−1
for NaAlH4 samples just doped and ball milled individually with 4 mol%
catalyst consisting of Ti, Zr or Fe.
Note that (–) indicates the dehydrogenation kinetics were considered to be
too slow for these single or co-dopant systems at the temperatures of 90, 110
and 130 ◦C to obtain reasonable activation energies.

248
J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
Fig. 3. Temperature programmed desorption (TPD) curves at 5 ◦C min−1 for
NaAlH4 samples just doped and ball milled with 4 mol% catalyst consisting
of single and binary combinations of (a) Ti–Fe and (b) Zr–Fe.
Fig. 5. Temperature programmed desorption (TPD) curves at 5 ◦C min−1 for
NaAlH4 samples just doped and ball milled with 4 mol% catalyst consisting
of binary and ternary combinations of Ti, Zr and Fe with (a) Ti, (b) Zr and
(c) Fe fixed at 1 mol%.
elsewhere [12]. The conflicting trend of kinetics increasing
but capacity decreasing with increasing catalyst concentra-
tion can be explained based on the study by Sandrock et al.
[12]. They found that the TiCl3 was completely reduced by
Na in the NaAlH4 to form NaCl and most likely zero-valent
Ti. This solid-state reaction can be written as [12]:
Fig. 4. Temperature programmed desorption (TPD) curves at 5 ◦C min−1 for
NaAlH4 samples just doped and ball milled with 4 mol% catalyst consisting
of single and binary combinations of Ti–Zr.
(1 − x)NaAlH4 +xTiCl3 → (1 − 4x)NaAlH4 + 3xNaCl
+ xTi + 3xAl + 6xH2
(3)

J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
249
Fig. 6. Temperature programmed desorption (TPD) curves at 5 ◦C min−1
for the six best samples of NaAlH4 just doped and ball milled with 4 mol%
catalyst consisting of binary and ternary combinations of Ti, Zr and Fe.
where x is the mole fraction of TiCl3 in the NaAlH4 (similar
reactions can be written for FeCl3 and ZrCl4). Eq. (3) shows
that the hydrogen capacity depends on the amount of TiCl3
in the sample. Theoretically, after doping with 4 mol%
Ti, the hydrogen capacity should decrease to 4.6 wt%; the
experimental value obtained here is very close to this value
at 4.5 wt%. The fact that the presence of chlorine in the
catalyst produces a side reaction that diminishes the capacity
is the main driving force behind finding catalysts, like Ti
nanoclusters [17,39], that do not contain species that react
with sodium.
Fig. 2 shows the TPD results for NaAlH4 catalyzed
individually with 4 mol% Ti, 4 mol% Zr and 4 mol% Fe,
and a TPD curve for an un-doped but ball milled sample of
NaAlH4. It is very clear that Ti is the best catalyst, followed
by Zr and then Fe; these trends generally agree with that
reported elsewhere [4,6]; and as mentioned above, it is
expected that the concentration dependence exhibited by
Ti would qualitatively be the same for Zr and Fe (Fig. 1).
Another feature that is noteworthy concerns the very weak
effect of Fe, which is only marginally better than the
un-doped sample of NaAlH4, with 2 wt% hydrogen not
being realized until about 215 ◦C compared to 170 and
140 ◦C for Zr and Ti, respectively. Finally, it is important
to mention that the pronounced effect reported for Zr on the
second reaction is not observed in Fig. 2 [4].
Fig. 7. Constant temperature desorption (CTD) curves at (a) 90 ◦C, (b)
110 ◦C and (c) 130 ◦C for the six best samples of NaAlH4 just doped and ball
milled with 4 mol% catalyst consisting of binary and ternary combinations
of Ti, Zr and Fe.
It is also interesting to point out that all but two of the TPD
curves in Figs. 1 and 2 exhibit two different dehydrogenation
regions that are characteristic of the dehydrogenation reac-
tions depicted in Eqs. (1) and (2) occurring in the lower and
higher temperature regions, respectively. The first reaction
dominates in the lower temperature region and essentially
goes to completion (providing at most but usually less than
3.7 wt% hydrogen) before the second reaction begins to take
off in the higher temperature region, which also goes to com-
pletion (providing at most but usually less than 1.8 wt% hy-
drogen). These characteristic regions are not observed with
the Fe-doped and un-doped samples, which suggests that the
better catalysts, such as Ti and Zr, have profound effects on
the dehydrogenation kinetics of both the first and second re-
actions, with the effect being more pronounced for the first
reaction. These results also suggest that the weaker dopant,
Fe, behaves similarly to the un-doped case, with both reac-
tions proceeding at more comparable but far slower rates and
at the higher temperatures compared to the Ti and Zr doped
samples.

250
J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
With the metal dopant concentration fixed in all cases at
this metal is co-doped with either Ti or Zr. For the Ti–Fe sys-
tem (Fig. 3a), synergistic effects are observed immediately
by all three co-doped samples in the temperature range of in-
terest, i.e., below about 120 ◦C, and the effect improves with
decreasing concentrations of Fe; this is considered to be an
extraordinary result. Moreover, it is quite remarkable that the
binary sample with 3 mol% Ti–1 mol% Fe exhibits improved
performance over that of the sample doped with 4 mol% Ti
for almost the entire temperature range up to the plateau in
the high temperature region, indicating a minimal but still
synergistic effect of the 3 mol% Ti–1 mol% Fe system even
on the second reaction (Eq. (2)).
Even more distinct synergistic effects are observed with
the samples containing binary mixtures of Zr and Fe as shown
in Fig. 3b; however, these results are generally not as good
as those obtained with the samples containing Ti and Fe,
simply because Ti is a better single metal catalyst than Zr
(Fig. 2). For this Zr–Fe system, the two binary samples con-
taining 2 mol% Zr–2 mol% Fe, and 3 mol% Zr–1 mol% Fe
show better kinetic behavior than the 4 mol% Zr sample at all
temperatures between 70 and 200 ◦C. In contrast, the sample
containing 1 mol% Zr–3 mol% Fe does not begin to exhibit
4 mol%, a synergistic effect (as defined here) is realized any-
time the TPD curve obtained from a co-doped sample lies
above or to the left of the TPD curve obtained from the sam-
ple doped with only 4 mol% of the more active single metal
catalyst. This definition is more restrictive than the usual one
that would declare a synergy when a co-doped system outper-
forms a linear combination (i.e., a simple physical mixture)
of the performances of the single metal catalysts involved
in the binary or ternary system. Examples of synergistic ef-
fects associated with binary co-doped samples of NaAlH4 are
shown in Fig. 3. Fig. 3a shows the TPD results for NaAlH4
catalyzed with 4 mol% total metal ranging from 4 mol% Ti to
4 mol% Fe and all binary combinations of Ti–Fe. Similarly,
Fig. 3b shows the TDP results for the Zr–Fe system. The in-
serts in these figures reveal the synergistic effects in the very
low temperature region between 50 and 100 ◦C, where the
first dehydrogenation reaction begins and the initial intrinsic
rates are realized.
A comparison of the results in Figs. 2 and 3 shows that
the weak influence exhibited by Fe alone on the dehydro-
genation kinetics of NaAlH4 is profoundly different when
Fig. 8. Arrhenius plots showing the long-time linear dehydrogenation rates of the first decomposition reaction (Eq. (1)) as a function of temperature for samples
of NaAlH4 just doped and ball milled with 4 mol% catalyst consisting of combinations of Ti, Zr and Fe containing (a) 3 mol%, (b) 2 mol%, (c) 1 mol% and (d)
0 mol% Ti. For comparison, samples of NaAlH4 just doped and ball milled with 4 mol% Ti from this work and taken from Sandrock et al. [12] are included in
each plot.

J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
251
any synergistic behavior until around 160 ◦C. Also, notice
that the synergistic effect from all three co-doped samples
is much more pronounced for the first reaction, having only
a marginal synergistic effect on the second reaction. Nev-
ertheless, this synergism associated with the second reac-
tion is particularly marked in the sample containing 1 mol%
Fe–3 mol% Zr, where the TPD run is more comparable to the
samples containing Ti and Fe (Fig. 3a).
fests as improved dehydrogenation kinetics. In other words,
it is hypothesized that the observed synergistic behavior with
the Ti–Fe system results from a highly collaborative elec-
tronic interaction between these two metal species through
the intrinsic ability of Ti (or even Zr) to share electrons (by
virtue of its low electronic density and also its low ionization
energies) with the electron rich Fe species (or even Cu or
others). The result is enhanced dehydrogenation kinetics, as
shown in Fig. 3 with the Ti–Fe and Zr–Fe co-doped binary
systems.
As mentioned above, synergistic effects of the co-dopants,
Ti–Fe, have already been reported for the metal catalyzed
NaAlH4 system [6]. However, in that study an alkoxide of
Fe was used as the dopant precursor, whereas chlorides of
Zr and Fe were used here. This difference in the precursors
indicates that the synergism is in no way associated with the
anionic species (i.e., the chloride or alkoxide) of the cata-
lyst precursor. On the contrary, this result strongly suggests
that it is in some way due to an interaction between the two
metals and the alanate structure. Bogdanovic et al. [6] at-
tempted to explain this behavior through the formation of a
TiFe alloy. However, these authors could only find evidence
suggesting that Fe and Ti each reacted separately with the
Al species, i.e., no alloying between Ti and Fe occurred dur-
ing the doping process and first decomposition reaction. A
plausible explanation as to why a synergism between Ti–Fe
should be expected with the NaAlH4 system is offered below.
First and foremost, alloying between Ti and Fe cannot be
ruled out here, because the samples were prepared using a
high-energy ball mill, compared to the far milder solution
chemistry technique employed at that time by Bogdanovic
et al. [6]. Although it is unclear whether alloying actually
occurs, it is reasonable to assume that the two metals must
come into close proximity to each other to have enough in-
teraction with the alanate to foster a synergistic effect. For-
mation of a TiFe alloy presents one possibility; on the other
hand, other scenarios should be considered. A plethora of
information in the literature supports the ability of Fe to en-
hance the performance of Ti, whether through the use of Fe
to enhance the hydriding ability of Ti [40–44], or even to en-
hance the performance of Ti as a catalyst in the hydrogena-
tion/dehydrogenation of organic compounds [45,46], possi-
bly through the formation of redox pairs. It is probably not
a coincidence that Ti, Zr and V (in this order) show the best
activities with complex hydrides [19], which exactly coin-
cides with the ability of these metals to form hydrides [47].
It is also probably not a coincidence that Anton [19] cate-
gorizes Fe and other related metals such as Cu and Co as
only moderately active catalysts for the dehydrogenation of
NaAlH4, which is consistent with their diminished ability to
form hydrides [47]. However, it is well known that alloying
some of these metals together produces stable and cyclable
metal hydrides, like TiFe [40–44].
Based on thisplausible explanationof the Ti–Fe andZr–Fe
synergisms, it is not too surprising that the Ti–Zr co-doped
system exhibits far smaller or essentially no synergistic be-
havior, as shown by the TPD curves in Fig. 4. Similar to the
Ti–Fe and Zr–Fe systems, the concentration of Zr was var-
ied from 0 to 4 mol% in steps of 1 mol% and balanced with
Ti to maintain a total catalyst concentration of 4 mol%. The
insert shows that the initial rates of dehydrogenation in the
50–100 ◦C range. In this low temperature region all the initial
dehydrogenation rates are very similar. In general, however,
as the temperature increases beyond 100 ◦C the performance
improves with increasing concentrations of Ti. Clearly, the
synergistic effect on the first reaction is quite limited and es-
sentially non-existent on the second reaction for all three co-
doped samples. It is important to point out that these results
are in stark contrast to that reported in the literature about the
significant and positive effect of Zr improving the kinetics of
the second reaction, whether alone or co-doped with Ti [4].
This difference could be due to Zidan et al. [4] using Zr(OPr)4
as the catalyst precursor compared to this study using ZrCl3.
However, based on the fact that the use of an alkoxide versus
a chloride catalyst precursor has no effect of the synergistic
effect associated with the Ti–Fe system (as explained above),
it is more likely that this difference is due to the improvement
reported by Zidan et al. [4] being observed only after several
dehydrogenation/rehydrogenation cycles were carried out. In
contrast, all the TPD results in Fig. 4 correspond to the dehy-
drogenation of freshly doped and ball milled samples, i.e., un-
cycled samples that have undergone only a dehydrogenation
step. A future study will consider the effects of co-dopants on
the dehydrogenation/rehydrogenation kinetics after cycling,
in part, to substantiate these earlier findings [4], which are
considered to be extremely important because they remain to
be the only results that show a significant effect of a metal
catalyst on the second reaction, at least beyond that observed
with Ti alone and as reported here for the co-doped Zr–Fe
and Ti–Fe systems.
The dehydrogenation behavior of NaAlH4 when co-doped
with three different metal catalysts is shown in Fig. 5.
Fig. 5a shows the TPD curves with each system having
1 mol% Ti present and the Zr–Fe contents varying from
3 mol% Zr–0 mol% Fe to 2 mol% Zr–1 mol% Fe to 1 mol%
Zr–2 mol% Fe to 0 mol% Zr–3 mol% Fe. Fig. 5b and c dis-
play similar results for Zr and Fe with each system in Fig. 3b
having 1 mol% Zr present and the Ti–Fe contents varying sys-
tematically, andwitheachsystemin Fig. 3chaving1 mol%Fe
With these thoughts in mind, the synergistic effect could
simplyandmoreprobabilisticallybeduetoaweakinteraction
between these active metal species that further destabilizes
the electronic structure of the alanate, thereby facilitating en-
hanced electron or atomic hydrogen transport, which mani-

252
J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
present and the Ti–Zr contents varying systematically. The in-
serts show the initial dehydrogenation rates in the 50–100 ◦C
range. In general, for all three systems, binary co-doped sys-
tems perform better than ternary co-doped systems. The sole
exception is for the ternary system with 1 mol% Zr–2 mol%
Ti–1 mol% Fe (Fig. 5b), which outperforms the binary sys-
tem with 1 mol% Zr–3 mol% Ti in the 50–125 ◦C tempera-
ture range, probably because of the presence of Fe. Also, in
general, the performance of a system improves with its Ti
content, whether it is present in a binary or ternary system. It
is interesting to observe the following additional trends with
these binary and ternary co-doped systems.
For the system with Ti present in 1 mol% (Fig. 5a), the
performance improves only marginally when Zr replaces Fe,
suggesting that Fe and Zr promote the catalytic effect of low
levelsofTiinasimilarmanner. ForthesystemwithFepresent
in 1 mol% (Fig. 5c), the improvement in the performance is
slightly better when Ti replaces Zr, suggesting that Ti is the
better catalyst of the two metals, which is consistent with
the results shown in Fig. 2. The most improvement in the
performance is observed for the system with Zr present in
1 mol% (Fig. 5b). In this case, when Ti replaces Fe the per-
formance improves substantially, again indicating the supe-
rior ability of Ti to catalyze the first dehydrogenation reac-
tion compared to Fe. These systematic trends between the
binary and ternary co-doped systems may be indicating that
the synergistic effects are limited to the interaction between
binary combinations of these metals. From a physical point
of view, this makes sense based on the simple argument that
the probability of a binary system coming together to form
a stable catalytically active complex that promotes the first
dehydrogenation reaction of NaAlH4 is much greater than
that associated with a ternary system.
The TPD curves from the five binary and ternary co-doped
NaAlH4 samples that exhibit the best kinetic behavior are
shown in Fig. 6, along with the sample doped with 4 mol%
Ti. It is noteworthy that at least 2 mol% Ti is present in each of
the co-doped samples. A distinct difference is also observed
between the samples doped with Fe compared to those with
no Fe present. At temperatures below 125 ◦C, the TPD curves
for the three samples containing Fe clearly lie above those of
the sample doped with 4 mol% Ti. In contrast, the two sam-
ples that do not contain any Fe essentially overlap the TPD
curve of the 4 mol% Ti doped sample. The marked syner-
gistic effect of Fe is quite apparent here, especially in the
100 ◦C range, as shown in the insert; nevertheless, the syner-
gistic effect of Zr is still noticeable by the overlapping of the
TPD curves of the co-doped Ti–Zr samples with that of the
Ti sample below 125 ◦C. It is worth pointing out that at tem-
peratures higher than about 140 ◦C, the obvious synergistic
effects become diminished except for the 3 mol% Ti–1 mol%
Fe sample, the TPD curve of which essentially overlaps
that of the 4 mol% sample. These results further substanti-
ate that the effects of co-dopants on the second reaction are
much less pronounced, except for this 3 mol% Ti–1 mol% Fe
sample.
The CTD runs obtained at 90, 110 and 130 ◦C for the
same six samples discussed above (that exhibit the best ki-
netics) are given in Fig. 7a, b and c, respectively. Similarly
to the TPD curves, any co-doped CTD curve that lies above
the CTD curve for the sample doped with only 4 mol% Ti
(in this case) indicates a synergistic effect, as defined here.
These independently obtained results corroborate very well
the synergistic effects observed in the co-doped TPD curves
shown in Fig. 6. For example, all the samples containing Fe
in combination with the other two metals (i.e., Ti and/or Zr)
all perform better than those without Fe present, with the co-
doped 3 mol% Ti–1 mol% Fe system again exhibiting the best
dehydrogenation kinetics at all three temperatures. The de-
hydrogenation rates for this Ti–Fe co-doped sample are 3.7,
2.0 and 1.5 times that of the sample doped only with 4 mol%
Ti at 90, 110 and 130 ◦C, respectively. This result is quite
remarkable when considering the fact that Fe is almost in-
active as a dehydrogenation catalyst for NaAlH4, even when
NaAlH4 is doped with 4 mol% Fe, as shown in Fig. 2. An
equally intriguing result is the fact that all the co-doped sam-
ples exhibit faster dehydrogenation rates than the Ti-doped
sample within the time frame of the predominantly linear
regions, again for all three temperatures. The two worst sam-
ples, exhibiting the slowest rates at all three temperatures,
bothhaveZrpresent, whichasexplainedabove, furthershows
that Zr is not as good of a promoter for Ti as Fe is for both Ti
and Zr.
To determine the activation energy associated with the first
reaction (Eq. (1)), the initial intrinsic decomposition rate (k)
obtained from a CTD run at each temperature (T) of interest
(90, 110 and 130 ◦C) was evaluated by fitting most of the
CTD curve to a sixth order polynomial and then determining
the derivative of this time-dependent function (i.e., slope) at
the time set equal to zero. The three slopes and corresponding
temperatures were then correlated with the Arrhenius equa-
tion to determine A, the pre-exponential factor and Ea, the
activation energy, according to:
ꢀ
ꢁ
−Ea
k = A exp
(4)
RT
where R is the universal gas constant. A plot ln k versus 1/T
for each system produced a very linear relationship from
which the activation energies and pre-exponential factors
were obtained for all systems, except for the three samples
that exhibited dehydrogenation kinetics that were considered
to be too slow at the temperatures of interest. The results
are given in Table 1. With the exception of the sample con-
taining 4 mol% Zr, the activation energies ranged between
75 and 90 kJ mol−1 for all of the samples, which are very
close to the values obtained by Sandrock et al. [12], which
ranged between 72 and 80 kJ mol−1 for NaAlH4 doped with
different concentrations of Ti. The activation energy for the
sample doped with 4 mol% Zr is also similar to that reported
by them for un-doped NaAlH4 [12]. However, the interesting
and possibly intuitive trend of decreasing activation energy

J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
253
with increasing catalyst concentration, as seen here, was not
observed by Sandrock et al. [12].
In this kind of plot, a synergistic effect is realized by
a hydrogen desorption rate from a co-doped sample be-
ing higher than that corresponding to a sample doped only
with 4 mol% Ti or Zr at a given temperature. It is clear
that any co-doped sample containing at least 3 mol% Ti
(Fig. 8a) or at least 2 mol% Ti (Fig. 8b) exhibits synergis-
tic effects. Again, the best co-doped system, by far, is the
3 mol% Ti–1 mol% Fe binary system. In stark contrast, only
the binary system with 1 mol% Ti–3 mol% Zr and the ternary
system with 1 mol% Ti–2 mol% Zr–1 mol% Fe, which are
both co-doped with 1 mol% Ti (Fig. 8c) exhibit any syn-
ergy. The synergistic effects exhibited by the Zr–Fe sys-
tem are quite pronounced with respect to 4 mol% Zr alone
(Fig. 8d); but only the binary system with 3 mol% Zr–1 mol%
Fe, which has no Ti present, exhibits a better performance
with respect to 4 mol% Ti alone. The fact that the Ti–Fe
and Zr–Fe systems exhibit the most pronounced synergis-
tic and kinetic behaviors, again suggests very strongly that
the presence of Fe, an element with a propensity to share
its electrons, promotes the catalytic activity of Ti and Zr
more than the presence of Zr does in most of the systems
containing Ti.
Overall, since the differences in the activation energies
observed between these relatively active catalyst systems are
minor and most probably due to using only three experi-
mental temperatures, two important findings can be gleaned
from these results. First, despite the different concentrations
and nature of the catalysts, when the catalysts are suffi-
ciently active, the first dehydrogenation reaction apparently
goes through the same transition state with a very similar
rate determining step; the kinetic differences observed be-
tween the different catalysts are just due to a difference in
their entropic states, as generally indicated by differences
in their pre-exponential factors. In other words, a good cat-
alytic system does not require as much organization (i.e.,
as low of an entropic state) of the reactants for the reac-
tion to occur, while they all seem to go through the same
limiting path, possibly through a metal (Ti or Zr) or metal
alloy (e.g., Ti–Al) structure. As explained above, the pres-
ence of Fe promotes this active species by destabilizing its
electronic structure even further, thereby facilitating electron
or atomic hydrogen transport. Second, the higher activation
energies observed for the samples with poor catalytic activ-
ity (e.g., the sample with 4 mol% Zr) or no catalytic activity
(e.g., the un-doped NaAlH4 sample [12]) go through a dif-
ferent reaction pathway (i.e., a different limiting step). Pos-
sible evidence for this supposition has been reported unwit-
tingly by Gross et al. [9], when they found two unidentifiable
structures in an XRD pattern that were only observed in un-
doped samples of NaAlH4 just after the first decomposition
reaction.
All of the CTD runs were analyzed further to obtain the
more useful hydrogen discharge rates of the first reaction as
a function of temperature and catalyst system. These rates
correspond to the slope of the long-time linear region of each
CTD curve, beyond that associated with the initial intrinsic
kinetic region used to obtain the activation energies. It should
not be a surprise that the long-time linear and initial intrin-
sic kinetic regions may differ due to diffusion or transport
limitations arising in the system once the reaction progresses
past the surface and into the interior of the NaAlH4 parti-
cles. The results of this analysis are displayed in Fig. 8, in
terms of the hydrogen desorption rate plotted as a function of
both the inverse temperature and the temperature. Fig. 8a, b,
c and d correspond to samples of NaAlH4 co-doped with 3,
2, 1 and 0 mol% Ti, respectively. For comparison, each figure
also presents the results for samples of NaAlH4 doped indi-
vidually with 4 mol% Ti from this study and from Sandrock
et al. [12]. Notice that in all cases a justifiably linear rela-
tionship results, suggesting that even the long-time apparent
kinetics are fundamental in nature and possibly related to an
activated diffusion process. Observe also that the results from
Sandrock et al. [12] agree very well with those obtained in
this study, with some minor discrepancy appearing only in the
lower temperature region where the measurement becomes
more problematic.
4. Conclusions
The first systematic analysis of the effect of co-dopants
on the dehydrogenation kinetics of NaAlH4 has been car-
ried out with chlorides of Ti, Zr and Fe as the catalysts.
Fifteen different catalyst combinations doped at 4 mol% to-
tal metal were studied: three systems with single catalysts
of Ti, Zr and Fe; nine systems with binary combinations
of Ti, Zr and Fe; and three systems with ternary combina-
tions of Ti, Zr and Fe. For comparison, an un-doped sam-
ple of NaAlH4 was also prepared and tested in the same
manner as the doped samples, along with three samples of
NaAlH4 doped with 1, 2 and 3 mol% Ti. In this way, a syn-
ergistic effect was defined and identified by the performance
of a co-doped system being better than that obtained from
NaAlH4 doped with 4 mol% Ti or Zr alone, i.e., the best sin-
gle metal catalyst of the two or three catalysts in the binary
or ternary system. Temperature programmed desorption and
constant temperature desorption runs were carried out gravi-
metrically on each sample, just after being prepared by a
wet doping procedure followed by high-energy ball milling.
Important findings about the possible improvement of the
dehydrogenation kinetics of co-doped NaAlH4 include the
following.
Numerous samples of NaAlH4 when co-doped with vari-
ous combinations of Ti, Zr and Fe exhibited synergistic be-
havior, with respect to improving the dehydrogenation kinet-
ics of the first decomposition reaction (i.e., NaAlH4 decom-
posing to form Na3AlH6). These are very exciting results
when considering the fact that as a dopant for NaAlH4, Ti
was found to be a much better catalyst than Zr, which was
found to be a much better catalyst than Fe, which was found

254
J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
to be almost inactive. In general, the dehydrogenation kinet-
ics improved with the amount of Ti present in a co-doped
sample, whether it was a binary or ternary system. Of the
12 different co-doped samples, the top 5 performers all had
at least 2 mol% Ti as one of the co-dopants. The best syn-
ergistic performance was exhibited by the sample co-doped
with a binary combination of 3 mol% Ti–1 mol% Fe, with
dehydrogenation rates 3.7, 2.0 and 1.5 times that of 4 mol%
Ti alone at 90, 110 and 130 ◦C, respectively. Although the
binary co-doped Zr–Fe systems exhibited more pronounced
synergistic effects (with respect to 4 mol% Zr alone) than
did the binary co-doped Ti–Fe systems effects (with respect
to 4 mol% Ti alone), their overall performance was always
worse simply because Ti is a better single metal catalyst than
Zr. The least synergistic effects were exhibited by the binary
co-doped Ti–Zr systems, where based on evidence from the
literature, it was surmised that the superior electron sharing
ability of electron rich Fe was responsible for it being a bet-
ter promoter of Ti and Zr than Zr was of Ti. This supposition
was further supported by the fact that the synergistic effects
exhibited by the ternary co-doped systems behaved more like
simple physical mixtures of binary combinations of the Ti–Fe
and Zr–Fe systems.
References
[1] J.A. Ritter, A.D. Ebner, J. Wang, R. Zidan, Mater. Today 9 (2003)
18.
[2] B. Bogdanovic, M. Schwickardi, J. Alloys Compd. 253 (1997) 1.
[3] C.M. Jensen, R. Zidan, N. Mariels, A. Hee, C. Hagen, Int. J. Hy-
drogen Energy 24 (1999) 461.
[4] R.A. Zidan, S. Takara, A.G. Hee, C.M. Jensen, J. Alloys Compd.
285 (1999) 119.
[5] K.J. Gross, S. Guthrie, S. Takara, G. Thomas, J. Alloys Compd. 297
(2000) 270.
[6] B. Bogdanovic, R.A. Brand, A. Marjanovic, M. Schwickardi, J.
Tolle, J. Alloys Compd. 302 (2000) 36.
[7] C.M. Jensen, K.J. Gross, Appl. Phys. A: Mater. Sci. Process. 72
(2001) 213.
[8] B. Bogdanovic, M. Schwickardi, Appl. Phys. A: Mater. Sci. Process.
72 (2001) 221.
[9] K.J. Gross, G.J. Thomas, C.M. Jensen, J. Alloys Compd. (2002)
683.
[10] B. Bogdanovic, G. Sandrock, MRS Bull. 27 (2002) 712.
[11] K.J. Gross, G. Sandrock, G.J. Thomas, J. Alloys Compd. (2002)
691.
[12] G. Sandrock, K. Gross, G.J. Thomas, J. Alloys Compd. 339 (2002)
299.
[13] G. Sandrock, K. Gross, G. Thomas, C. Jensen, D. Meeker, S. Takara,
J. Alloys Compd. (2002) 696.
[14] G.P. Meisner, G.G. Tibbetts, F.E. Pinkerton, C.H. Olk, M.P. Balogh,
J. Alloys Compd. 337 (2002) 254.
The effect of these binary and ternary co-dopants of Ti, Zr
and Fe on the second decomposition reaction (i.e., Na3AlH6
decomposing to form NaH) was not as pronounced as their ef-
fect on the first reaction. But synergisms were still observed,
especially with all three binary Zr–Fe co-doped systems and
to a lesser extent with only the 3 mol% Ti–1 mol% Fe system.
However, the striking effect of Zr on the second reaction that
was reported in the literature for samples of NaAlH4 doped
with Zr and Zr–Ti and cycled was not observed here with
similarly doped but un-cycled samples. This important dif-
ference suggests that cycling may be the key to activating the
co-dopants, like Ti–Fe and especially Zr–Fe, and enhancing
the dehydrogenation kinetics of the second reaction. In fact,
it is considered to be extremely important to further substan-
tiate these findings through cycling, because Zr remains to
be the only catalyst to reportedly have any significant effect
on the second reaction, at least beyond that observed with Ti
alone and as reported here for the Zr–Fe and Ti–Fe systems.
Therefore, a future study will consider the effects of these co-
dopants on the dehydrogenation/rehydrogenation kinetics af-
ter cycling, to verify whether the synergistic effects observed
here, especially with the 3 mol%Ti–1 mol% Fe system, sur-
vive the effects of cycling, and to substantiate the report in
the literature on the positive influence of Zr on improving the
dehydrogenation kinetics of the second reaction.
[15] G.J. Thomas, K.J. Gross, N.Y.C. Yang, C.M. Jensen, J. Alloys
Compd. 702 (2002) 330–332.
[16] D. Sun, T. Kiyobayashi, H.T. Takeshita, N. Kuriyama, C.M. Jensen,
J. Alloys Compd. 337 (2002) 8.
[17] M. Fichtner, O. Fuhr, O. Kircher, J. Rothe, Nanotechnology 14
(2003) 778.
[18] K.J. Gross, E.H. Majzoub, S.W. Spangler, J. Alloys Compd. (2003)
356–357.
[19] D.L. Anton, J. Alloys Compd. 400 (2003) 356–357.
[20] B. Bogdanovic, M. Felderhoff, S. Kaskel, A. Pommerin, K.
Schlichte, F. Schueth, Adv. Mater. (Weinheim, Germany) 15 (2003)
1012.
[21] B. Bogdanovic, M. Felderhoff, M. Germann, M. Hartel, A. Pom-
merin, F. Schuth, C. Weidenthaler, B. Zibrowius, J. Alloys Compd.
350 (2003) 246.
[22] T. Kiyobayashi, S.S. Srinivasan, D. Sun, C.M. Jensen, J. Phys. Chem.
A 107 (2003) 7671.
[23] C. Weidenthaler, A. Pommerin, M. Felderhoff, B. Bogdanovic, F.
Schueth, PCCP 5 (2003) 5149.
[24] M.P. Balogh, G.G. Tibbetts, F.E. Pinkerton, G.P. Meisner, C.H. Olk,
J. Alloys Compd. 350 (2003) 136.
[25] D. Sun, S.S. Srinivasan, T. Kiyobayashi, N. Kuriyama, C.M. Jensen,
J. Phys. Chem. B 107 (2003) 10176.
[26] E.H. Majzoub, K.J. Gross, J. Alloys Compd. 363 (2003) 356–
357.
[27] A. Zuttel, P. Wenger, P. Sudan, P. Mauron, S.-I. Orimo, Mater. Sci.
Eng. B: Solid-State Mater. Adv. Technol. B108 (2004) 9.
[28] A. Zuettel, Naturwissenschaften 91 (2004) 157.
[29] J.M. Bellosta von Colbe, B. Bogdanovic, M. Felderhoff, A. Pom-
merin, F. Schuth, J. Alloys Compd. 370 (2004) 104.
[30] H.W. Brinks, C.M. Jensen, S.S. Srinivasan, B.C. Hauback, D. Blan-
chard, K. Murphy, J. Alloys Compd. 376 (2004) 215.
[31] D. Sun, S.S. Srinivasan, G. Chen, C.M. Jensen, J. Alloys Compd.
373 (2004) 265.
Acknowledgements
[32] J. Graetz, J.J. Reilly, J. Johnson, A.Y. Ignatov, T.A. Tyson, Appl.
Phys. Lett. 85 (2004) 500.
[33] S.-C. Chung, H. Morioka, J. Alloys Compd. 372 (2004) 92.
[34] M.E. Arroyo y de Dompablo, G. Ceder, J. Alloys Compd. 364 (2004)
6.
Financial support provided by SCUREF/WSRC/DOE un-
der contract WEST052, KG09725-O, and the NRO under
contract NRO-00-C-0134 is greatly appreciated.

J. Wang et al. / Journal of Alloys and Compounds 391 (2005) 245–255
255
[35] A. Aguayo, D.J. Singh, Phys. Rev. B: Condens. Matter Mater. Phys.
69 (2004), 155103/1.
[41] G. Busch, L. Schlapbach, F. Stucki, P. Fischer, A.F. Andresen, Int.
J. Hydrogen Energy 4 (1979) 29.
[36] V. Ozolins, E.H. Majzoub, T.J. Udovic, J. Alloys Compd. 375 (2004)
1.
[37] P. Vajeeston, P. Ravindran, R. Vidya, H. Fjellvaag, A. Kjekshus,
Cryst. Growth Des. 4 (2004) 471.
[38] P. Ravindran, P. Vajeeston, H. Fjellvag, A. Kjekshus, Comput. Mater.
Sci. 30 (2004) 349.
[39] M. Fichtner, J. Engel, O. Fuhr, O. Kircher, O. Rubner, Mater. Sci.
Eng. B108 (2004) 42.
[42] G. Sicking, B. Jungblut, Surf. Sci. 127 (1983) 255.
[43] C.-H. Chiang, Z.H. Chin, T.-P. Perng, J. Alloys Compd. 307 (2000)
259.
[44] T. Liu, H. Shao, X. Li, Intermetallics 12 (2004) 97.
[45] M.W.Davis, US Patent 4,560,816 (1985).
[46] N. Nishimiya, T. Wada, A. Matsumoto, K. Tsutsumi, J. Alloys
Compd. 313 (2000) 53.
[47] R. Speiser, in: W.M. Mueller, J.P. Blackledge, G.G. Libowitz (Eds.),
Metal Hydrides, Academic Press, New York, 1968, p. 94.
[40] L. Schlapbach, A. Seiler, F. Stucki, Mater. Res. Bull. 13 (1978) 1031.