Synthesis of NaAlH4-based hydrogen storage material using milling under low pressure hydrogen atmosphere

Article abstract of DOI:10.1016/j.jallcom.2006.05.025

The present study highlights the advantages of milling NaH/Al under moderate hydrogen pressure as a favourable production step for NaAlH₄-based hydrogen storage materials. Firstly, it is demonstrated that NaAlH₄ can be obtained by applying a moderate hydrogen pressure (6-12 bar) during milling of NaH and Al with and without the presence of an inexpensive catalyst (TiCl₄). The yield of NaAlH₄ depends critically on process parameters, such as hydrogen pressure and milling time. A fully converted product is capable of reversible hydrogen storage without any activation procedure. Under optimized conditions, a capacity of 4.2 wt.% was achieved and kinetics in the first desorption are comparable to NaAlH₄ doped with TiCl₃. Secondly, the synthesis has been optimized towards shorter milling times. By applying a few absorption/desorption cycles to material that was partially converted during milling, almost full reversible storage capacity can be reached. In addition, kinetics is extremely enhanced. For example, such material exhibits an optimum capacity already after two sorption cycles at 100 bar and 125 °C and allows to absorb 80% of the reversible hydrogen content within a few minutes.

Full text of DOI:10.1016/j.jallcom.2006.05.025

Journal of Alloys and Compounds 430 (2007) 350–355
Synthesis of NaAlH -based hydrogen storage material using milling
4
under low pressure hydrogen atmosphere
∗
Nico Eigen , Mirko Kunowsky, Thomas Klassen, R u¨ diger Bormann
GKSS Research Center GmbH, Department of Nanotechnology, Max-Planck-Str. 1, D-21502 Geesthacht, Germany
Received 14 March 2006; received in revised form 3 May 2006; accepted 5 May 2006
Available online 17 July 2006
Abstract
The present study highlights the advantages of milling NaH/Al under moderate hydrogen pressure as a favourable production step for NaAlH
4
-
based hydrogen storage materials. Firstly, it is demonstrated that NaAlH can be obtained by applying a moderate hydrogen pressure (6–12 bar)
during milling of NaH and Al with and without the presence of an inexpensive catalyst (TiCl ). The yield of NaAlH depends critically on process
parameters, such as hydrogen pressure and milling time. A fully converted product is capable of reversible hydrogen storage without any activation
procedure. Under optimized conditions, a capacity of 4.2 wt.% was achieved and kinetics in the first desorption are comparable to NaAlH doped
with TiCl . Secondly, the synthesis has been optimized towards shorter milling times. By applying a few absorption/desorption cycles to material
4
4
4
4
3
that was partially converted during milling, almost full reversible storage capacity can be reached. In addition, kinetics is extremely enhanced.
◦
For example, such material exhibits an optimum capacity already after two sorption cycles at 100 bar and 125 C and allows to absorb 80% of the
reversible hydrogen content within a few minutes.
©
2006 Elsevier B.V. All rights reserved.
Keywords: Alanate; Hydrogen absorbing materials; Mechanochemical processing; X-ray diffraction; Kinetics
1
. Introduction
ditions, 4.5 wt.% out of the theoretical capacity of 5.6 wt.% are
reached [3]. These materials, however, are intricately prepared
[4,5]. Na or NaH and Al are diluted in THF and a hydrogen pres-
Hydrogen is the ideal means of storage, transport and con-
◦
version of energy for a comprehensive clean-energy mobility
concept. Regarding the use of hydrogen as fuel for the zero-
emission vehicle, one of the main problems is the storage of
hydrogen. Metal hydrides offer a safe alternative to storage in
compressed or liquid form. In addition, metal hydrides have the
highest storage capacity by volume [1]. Magnesium hydride also
has a high storage capacity by weight, however the temperature
sure of 140 bar at 150 C has to be applied. Subsequently, the
NaAlH4 present in solution is filtered, precipitated and dried.
Since the solubility of NaAlH4 is very low, a high amount of
solvent is necessary.
Therefore, alternative routes for the production of NaAlH4
were investigated. Dymova et al. milled AlH3 with NaH in 1:1
ratiotoachieveNaAlH4 [6]. Indeed, millinginargonatmosphere
with and without toluene as milling agent lead to NaAlH4 in
relatively short times [6]. However, AlH3 itself is produced from
alkali alanate, such that this synthesis cannot serve as a feasible
production route for NaAlH4. In addition, Dymova et al. showed
that synthesis is possible from Na, Al and H2 at high temperature
◦
for desorption at 1 bar has to be at least 300 C.
Therefore, research has recently focused on alkali alanates
as potential storage materials. Bogdanovi c´ and Schwickardi
showed in 1996 that hydrogen can be reversibly stored and
released in sodium aluminum hydride, when it is ball milled
with a suitable catalyst [2]. Sodium aluminum hydrides with Ti
colloides can absorb hydrogen in a few minutes at 100 bar and
◦
(280 C) and pressure (175 bar), where Na is present in liquid
state [7]. However, due to the high temperature and the high
pressure this process is technically not favourable.
◦ ◦
00 C and also release hydrogen at 100 C [3]. Under such con-
1
Fortunately, it could be demonstrated that hydrogenating a
composite of NaH and Al is also possible in solid state when
adding a titanium-based catalyst [8–10]. Firstly, these compo-
nents were mixed, leading to a composite capable of reversible
∗
Corresponding author. Tel.: +49 4152 872647; fax: +49 4152 872625.
E-mail address: nico.eigen@gkss.de (N. Eigen).
0
925-8388/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2006.05.025

N. Eigen et al. / Journal of Alloys and Compounds 430 (2007) 350–355
351
absorption and desorption. The mixingis possiblein a glass reac-
tor or in a high energy ball mill under inert atmosphere. However,
full hydrogen capacity is reached only after numerous activation
cycles [8]. The highest initial capacity of about 3.5 wt.% was
reached, if the mixing is carried out in pentane. However, still
a number of cycles are necessary to reach a maximum capac-
ity [9,10]. The best results showed a maximum capacity of only
4
.0 wt.%afterabout20cycles [10]. Itwasalsoshownthattheini-
tial capacity can be increased by milling doped NaH/Al under
about 1 bar hydrogen atmosphere instead of inert atmosphere
and by using Ti powder instead of Ti-based compounds as cat-
alyst [11]. However, sorption kinetics of the material produced
in this way is rather slow [11].
Recently, it has been shown, that NaAlH4 can be formed by
milling NaH/Al with TiCl3 under hydrogen atmosphere with an
initial pressure of 83 bar [12]. The material exhibits an initial
capacity of about 4 wt.% and very fast sorption kinetics without
activation. However, the pressure used during milling is too high
to be applied in industrial milling equipment. Recent attempts to
form NaAlH4 by milling under hydrogen pressure of less than
Fig. 1. Temperature in the milling vials without powder as a function of milling
time at a rotational speed of 350 rpm in the fan-cooled planetary ball mill.
was cooled by fans. The temperature inside the vials was estimated by milling
without powder for different milling times and subsequent measurement of the
temperature by a thermocouple. The values were confirmed by measuring the
pressure before and after different intervals of milling without powder.
All handling including milling was carried out in a glove box with purified
argon atmosphere. Milling was performed with or without the addition of 2
or 5 mol.% TiCl4 (Fluka, Buchs, Switzerland) as catalyst. Since TiCl4 is very
volatile in dry argon atmosphere, it was added with a pipette and the milling
vessel was immediately closed. Subsequently, hydrogen was introduced by a
device, that allows charging the milling vessel through a non-return valve with
a defined initial hydrogen pressure and, additionally, allows measuring the pres-
sure after milling. Thus, it was confirmed that the pressure in the vial never
decreased to less than 80% of the initial value.
8
.5 bar failed and led only to partial formation of Na3AlH [13].
6
The reason for only partial formation of sodium alanate using
moderate pressures remains unclear.
Due to its low thermodynamical stability, higher pressures
might be required to form NaAlH4. For example, thermodynam-
ical studies at low temperature indicate that at 1 bar hydrogen
◦
pressure, NaAlH4 is thermodynamically stable only up to 33 C
[
14]. However, depending on process parameters, much higher
The product was characterized by X-ray diffraction (XRD) using a diffrac-
tometer (Siemens D5000) with Cu K␣ radiation and a secondary monochro-
mator. To prevent the reaction with air, the specimen was encapsulated with a
polyamide film (capton foil) during XRD. The data were collected in the range
temperatures can occur during milling, eliminating the driv-
ing force for its formation. This drawback may not be crucial
as it could be shown that sodium alanate, only partially con-
verted during milling under hydrogen atmosphere, can be fur-
ther hydrogenated without activation by applying a subsequent
hydrogenation/dehydrogenation cycle reaching about 3.5 wt.%
of hydrogen storage capacity [13].
◦
◦
◦
between 28 and 50 in steps of 0.02 .
The sorption kinetics were characterized by a volumetric method in a Siev-
ert’s type apparatus (HERA, Que., Canada). The powder samples of about
125 mg were filled in a sample holder under purified argon.
The aim of this work is to develop a simple process route
for hydrogen storage materials with optimum storage capacity
and kinetics and to obtain a material that does not require a
high number of activation cycles. In a first step, NaAlH4 will
be produced by milling of NaH and Al under hydrogen atmo-
sphere at moderate pressures. Then, it will be evaluated, which
suitable precursor is to be formed during milling to obtain opti-
mum reversible storage capacity of the material and fast sorption
kinetics. In view of a low cost production route, instead of TiCl3
used in most studies, the less expensive TiCl4 is used as catalyst
in the present study.
3. Results and discussion
Fig. 1 shows the temperature of the vessels without powder
as a function of the milling time. The temperature increases in
about 1.5 h from room temperature to about 45 C and remains
◦
◦
constant at this value. At 45 C the equilibrium pressure for
NaAlH4 is about 2 bar [14]. The little higher pressure of 6 bar
hydrogen pressure was chosen for the first experiment to guar-
antee a sufficient driving force for the reaction.
Fig. 2 shows the X-ray diffraction patterns of the 1:1 stoichio-
metric mixture of NaH and Al. The unmilled mixture shows the
characteristic peaks of NaH and Al. After 5 h of milling, NaH
and Al are still the dominating phases. The significant broaden-
ing of the peaks of both the NaH and the Al phase indicates that
the crystallite sizes of both phases have been strongly reduced.
However, no phase reaction has occurred yet. After 20 h milling,
2
. Experimental
Milling experiments were conducted using commercial NaH (95%),
Sigma–Aldrich, Steinheim, Germany, and aluminum (99.5%), Johnson Matthey,
Karlsruhe, Germany, as initial materials. The materials were mixed with a spat-
ula and milled in a planetary ball mill, type Pulverisette 5, Fritsch, Germany,
in stainless steel vials using 10 mm stainless steel balls. Milling was performed
at a rotational speed of 350 rpm and a ball-to-powder-ratio (BPR) of 50:1 and
additionally, at a rotational speed of 230 rpm and a BPR of 10:1.
first traces of Na3AlH are observed and the phase fraction
6
increases continuously up to the maximum investigated milling
time of 100 h. After 60 h of milling, NaAlH4 is detected addi-
tionally. This demonstrates that NaAlH4 can be formed even
without a catalyst by milling under hydrogen atmosphere of
6 bar. The slow conversion, however, indicates that a high yield
The heat produced during milling leads to elevated temperatures in the inte-
rior of the milling vial. To keep the temperature during milling low, the mill

3
52
N. Eigen et al. / Journal of Alloys and Compounds 430 (2007) 350–355
Fig. 2. X-ray diffraction patterns of NaH–Al samples after different times of
milling at 6 bar hydrogen pressure.
Fig. 3. X-ray diffraction patterns of NaH–Al + 5 mol.% TiCl4 after different
times of milling at 6 bar hydrogen pressure.
der might temporally exceed the temperature range, in which
NaAlH4 is stable at the surrounding hydrogen pressure, leading
only to a partial conversion. To prove the assumption that local
temperatures impede a complete conversion, an experiment with
increased pressure (12 bar) was conducted. The phase evolution
is shown in Fig. 4. The conversion is faster and the amount of
NaAlH4 formed is significantly higher, if 12 bar are used instead
of NaAlH4 can probably not be reached in reasonable milling
times.
Sorption kinetics of sodium alanate can be improved by a
suitable catalyst, which should be homogeneously and finely
dispersed in the material. Therefore, it is reasonable to add the
catalyst before milling. Fig. 3 shows X-ray diffraction patterns of
the initial mixture and milled material both containing 5 mol.%
TiCl4. As expected, the initial mixture shows NaH and Al and
small peaks of NaCl, which indicates that TiCl4 reacted with a
small amount of NaH. Already after 1 h of milling, a first small
of 6 bar. However, after 60 h, Al and minor amounts of Na3AlH6
are still present and no further conversion is achieved. This may
originate from some temperature peaks that still impede a full
conversion. In addition, a stoichiometric mixture of NaH and
Al was used as initial material. As part of the NaH reacts with
TiCl4 to NaCl, a surplus of Al remains in the final product, thus
decreasing the capacity of the material.
peak of Na3AlH and one of NaAlH4 is detectable. After 5 h of
6
milling, the peaks of Al have significantly broadened, as it was
also observed during milling without catalyst. However, due to
the presence of a catalyst significant amounts of Na3AlH have
6
Therefore, to optimize conversion of the storage material, a
further experiment was performed with significantly less milling
intensity to prevent high temperatures. The milling intensity was
decreased by using a lower ball to powder ratio (10:1 instead of
50:1) and less rotational speed (230 rpm instead of 350 rpm).
Additionally, the amount of catalyst was reduced to 2 mol.%
TiCl4 to decrease the amount of NaCl in the product. Moreover,
to eliminate the excessive aluminium, the ratio of NaH to Al was
adjusted to 1.08:1 according to the reaction,
already formed, proving that the catalyst leads to a much faster
formation of Na3AlH . The formation of NaAlH4, however, is
6
still very slow and after 20 h no significant further conversion is
observed.
◦
Although a maximum temperature of 45 C was measured
at the bottom of the milling vials, locally higher temperatures
may occur. Particularly, in powder particles which are trapped
between the grinding balls and the wall of the vial during a col-
lision, the temperature can increase significantly due to the heat
released by deformation of powder particles and the friction
of particles against each other. Additionally, heat is produced
by the exothermic formation of alanate phase. Thus, the pow-
1.08NaH + Al + 0.02TiCl4 + 1.46H2
→ NaAlH4 + 0.08NaCl + 0.02Ti,

N. Eigen et al. / Journal of Alloys and Compounds 430 (2007) 350–355
353
Fig. 4. X-ray diffraction patterns of NaH–Al + 5 mol.% TiCl4 after different
times of milling at 12 bar hydrogen pressure.
Fig. 5. X-ray diffraction patterns of NaH–Al + 2 mol.% TiCl4 after different
times of milling with low milling intensity at 12 bar hydrogen pressure.
assuming, in a first approach, that Ti is present as pure metal
after milling.
when compared to the kinetics observed on TiCl3 doped NaAlH4
as measured by Sandrock et al. [15]. The first decomposition
reaction is completed in about 1 h.
In Fig. 5, the X-ray diffraction patterns of this mixture are
shown after different milling times. Due to the lower content
of catalyst and the lower milling intensity, the conversion rate
is quite slow. However, after 240 h of milling the conversion is
almost complete.
In order to clarify, to what extent capacity and kinetics change
upon further absorption–desorption cycles, a cycling test was
◦
carriedoutat125 C, usingasamplewith2 mol.%TiCl4 content,
prepared under 12 bar hydrogen atmosphere with low milling
intensity. It can be seen in Fig. 7 that the material desorbs about
In Fig. 6, the first desorption of the samples produced by
milling with low milling intensity is plotted. Desorption was
◦
performed at 125 C into vacuum (fixed volume). The change
in desorption rate, occurring after 1 h, can be attributed to the
transition from the first decomposition reaction,
3
NaAlH4 → Na3AlH + 2Al + 3H2,
6
to the second decomposition reaction,
Na3AlH → 3NaH + Al + 1.5H2.
6
As expected, the amount of hydrogen released in the first step of
desorption continuously increases with the milling time. Since
the second decomposition reaction is very slow, desorption was
◦
accelerated by increasing the temperature to 150 C for the 240 h
milled sample. Thus, a total capacity of 4.2 wt.% could be mea-
sured. Considering the incomplete conversion into NaAlH4, this
value is very favourable when compared to the highest capac-
ity of 4.5 wt.%, which could so far be achieved on TiCl3 doped
NaAlH4 [15]. Also, desorption kinetics seem to be satisfying
◦
Fig. 6. First desorption at 125 C into vacuum of NaH–Al + 2 mol.% TiCl4 after
different times of milling at 12 bar hydrogen pressure in comparison with second
desorption of NaAlH4 with 2 mol.% TiCl3 [15].

3
54
N. Eigen et al. / Journal of Alloys and Compounds 430 (2007) 350–355
◦
Fig. 8. First hydrogenation/dehydrogenation cycles at 125 C of
NaH–Al + 5 mol.% TiCl4 after 10 h of milling in 6 bar hydrogen atmo-
sphere: (a) absorption at 100 bar and (b) desorption in vacuum. The experiment
was started with absorption.
◦
Fig. 7. Cycling experiment at 125 C of almost fully converted NaAlH4, pro-
duced by 180 h reactive milling of NaH–Al + 2 mol.% TiCl4 in 12 bar hydrogen
atmosphere: (a) desorption in vacuum and (b) absorption at 80 bar. The experi-
ment was started with desorption.
more than 1 wt.% of hydrogen is absorbed additionally to the
amount of hydrogen, that was already introduced during reac-
tive milling. The lack of any abrupt change in the absorption
◦
3
.8 wt.% of hydrogen at 125 C. Nearly all hydrogen released
in the first desorption, can be re-absorbed at 80 bar of hydrogen
pressure. Also, in the following cycles, no significant change in
the reversible storage capacity is found. However, kinetics slow
downslightlyfromthefirsttothesecondcycleinbothdesorption
and absorption, as can be deduced from the sorption slopes in
Fig. 7. After the second cycle, kinetics remains unchanged until
theendoftheexperiment. Thecurvesfromsecondtofourthcycle
are identical within the accuracy of the experimental method.
This shows that the NaAlH4-based storage material synthesized
by reactive milling does not have to be activated by a cycling
procedure to achieve a high reversible capacity and optimum
kinetics.
rate in this first absorption confirms that Na3AlH has already
6
been formed during milling. In the second and third absorp-
tion, a reversible capacity of about 3.3 and 3.5 wt.% can be
detected. This amount compares well with the reversible capac-
ity observed by Sandrock et al., who found 3.0 wt.% H2 for a
sample doped with 6 mol.% TiCl3 and 3.9 wt.% H2 for a sample
doped with 4 mol.% TiCl3. In addition, our results demonstrate
that maximum possible capacity is already obtained in the third
absorption. From a technical point of view, the activation can
be carried out in the tank shell, because the first two activation
◦
cycles can be carried out at 100 bar and 125 C only.
In view of an industrial production of hydrogen storage mate-
rial, the milling times to achieve a full conversion into NaAlH4
during milling are by far too high. Therefore, in a second
approach, sorption properties of material are investigated, that
is only partially converted into NaAlH4. Fig. 8 shows the first
cycles of hydrogenation/dehydrogenation of a sample milled for
Fig. 8 also shows that after the first absorption, desorption
and absorption both exhibit extraordinarily fast kinetics: 80% of
the maximum reversible hydrogen content is absorbed in about
3 min. Desorption is so fast that during establishing the vacuum,
the sample desorbs already a high amount of hydrogen, leading
to incomplete desorption curves. By considering the time of
about 2 min for establishing the vacuum, it can be estimated
that the first step of desorption is completed in about 5 min.
This shows that for material, produced by partial hydrogenation
10 h in 6 bar of hydrogen atmosphere. A relatively high catalyst
content of 5 mol.% TiCl4 was used to improve reaction kinet-
ics. Although the first hydrogenation of the sample is very slow,

N. Eigen et al. / Journal of Alloys and Compounds 430 (2007) 350–355
355
during reactive milling and subsequent hydrogenating, kinetics
is more than an order of magnitude faster than the material that is
fully converted into NaAlH4 during reactive milling. The faster
kinetics can be in part explained by the higher catalyst content
and the slightly higher pressure. However, it was reported, that
increasing the catalyst content from 2 to 6 mol.% TiCl3, leads
only to about three times faster kinetics [15]. Therefore, further
studies have to evaluate the microscopic reasons for the fast
kinetics.
It is interesting to note, that the obtained sorption kinetics
under our optimized conditions are almost comparable to the
best material based on NaAlH4 that was prepared in a compli-
cated procedure by synthesizing NaAlH4 at high pressure in sol-
vent, purifying and subsequently doping with Ti-clusters [3,16].
Therefore, the applied method of reactive milling under mod-
erate hydrogen pressure and subsequent hydrogenation, which
does not necessitate the use of any solvent and utilizes TiCl4 as
a cost-effective catalyst, is very promising for an industrial pro-
duction route for hydrogen storage material based on NaAlH4
and other related materials.
mum capacity after only two cycles of sorption. Moreover, such
material exhibits extraordinarily fast kinetics. Consequently, this
production method has a high potential for a low cost indus-
trial production process for hydrogen storage materials based
on alanates or other related materials.
Acknowledgement
The authors appreciate the financial support of the Euro-
pean Community within the framework of the integrated project
“StorHy—Hydrogen Storage Systems for Automotive Applica-
tion”, project no. 502667.
References
[1] L. Schlapbach, A. Z u¨ ttel, Nature 414 (2001) 23.
[
[
2] B. Bogdanovi c´ , M. Schwickardi, J. Alloys Compd. 253–254 (1997) 1.
3] M. Fichtner, J. Engel, O. Fuhr, O. Kircher, O. Rubner, Mater. Sci. Eng. B
108 (2004) 42.
[
[
4] E.C. Ashby, G.J. Brendel, H.E. Redman, Inorg. Chem. 3 (1963) 499.
5] B. Bogdanovi c´ , R.A. Brand, A. Marjanovi c´ , M. Schwickardi, J. T o¨ lle, J.
Alloys Compd. 302 (2000) 36.
[
[
[
[
6] T.N. Dymova, D.P. Aleksandrov, V.N. Konoplev, T.A. Silina, N.T. Kuznet-
zov, Russ. J. Coord. Chem. 19 (1993) 607.
7] T.N. Dymova, N.G. Eliseeva, S.I. Bakum, Y.M. Dergachev, Dokl. Akad.
Nauk. SSSR 215 (1974) 1369.
8] K.J. Gross, E.H. Majzoub, S.W. Spangler, J. Alloys Compd. 356–357
(2003) 423.
9] B. Bogdanovi c´ , M. Schwickardi, Appl. Phys. A 72 (2001) 221.
4
. Conclusions
By reactive milling of NaH and Al under moderate hydrogen
pressure, NaAlH4 can be formed with and without the addi-
tion of a TiCl4 as catalyst. The addition of catalyst speeds up
the conversion and leads to a hydrogen storage material with
sorption kinetics and a hydrogen storage capacity comparable
to conventionally prepared material. Favourably, no activation
process is required to obtain good capacity and sorption kinet-
ics. However, complete conversion of NaH and Al into NaAlH4
during reactive milling is very slow at moderate pressures and
consequently requires very long milling times. Therefore, in a
subsequent approach, it was demonstrated that a partial con-
version into NaAlH4 is sufficient to obtain a storage material
that is capable of reversibly storing hydrogen with nearly opti-
[
10] S.S. Srinivasan, H.W. Brinks, B.C. Hauback, D. Sun, C.M. Jensen, J. Alloys
Compd. 377 (2004) 283.
11] P. Wang, C.M. Jensen, J. Phys. Chem. B 108 (2004) 15827.
12] J.M. Bellosta von Colbe, M. Felderhoff, B. Bogdanovi c´ , F. Sch u¨ th, C.
Weidenthaler, Chem. Commun. 37 (2005) 4732.
[
[
[
[
13] P. Wang, X.D. Kang, H.M. Cheng, Appl. Phys. Lett. 87 (2005) 071911.
14] K.J. Gross, G.J. Thomas, C.M. Jensen, J. Alloys Compd. 330–332 (2002)
683.
[
[
15] G. Sandrock, K. Gross, G. Thomas, J. Alloys Compd. 339 (2002) 299.
16] M. Fichtner, O. Fuhr, O. Kircher, J. Rothe, Nanotechnology 14 (2003)
778.