Catalytic effect of Al3Ti on the reversible dehydrogenation of NaAlH4

Article abstract of DOI:10.1016/j.jallcom.2006.01.006

Al₃Ti was directly utilized as catalyst precursor to examine its catalytic activity in the reversible dehydrogenation of NaAlH₄. It was observed that Al₃Ti possessed considerable catalytic effect on the de-/hydriding react

Full text of DOI:10.1016/j.jallcom.2006.01.006

Journal of Alloys and Compounds 424 (2006) 365–369
Catalytic effect of Al Ti on the reversible dehydrogenation of NaAlH4
3
a
a,∗
b
c
c
a
X.D. Kang , P. Wang , X.P. Song , X.D. Yao , G.Q. Lu , H.M. Cheng
a
Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, PR China
b
Analysis and Testing Center, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, PR China
c
ARC Centre for Functional Nanomaterials, School of Engineering, The University of Queensland, Brisbane, QLD 4072, Australia
Received 13 November 2005; received in revised form 29 December 2005; accepted 1 January 2006
Available online 7 February 2006
Abstract
Al
Ti possessed considerable catalytic effect on the de-/hydriding reactions of NaAlH
increased with increasing ball-milling time. Moreover, it was noted that dehydriding kinetics and cycling performance of the Al
were highly dependent on the starting material, 1:1 NaH/Al mixture or NaAlH . The combined property and phase examinations suggest that Al
may act as active species to catalyze the reversible de-/hydrogenation of NaAlH
2006 Published by Elsevier B.V.
3
Ti was directly utilized as catalyst precursor to examine its catalytic activity in the reversible dehydrogenation of NaAlH
. The catalytic enhancement by Al Ti doping significantly
Ti-doped NaAlH
Ti
4
. It was observed that
Al
3
4
3
3
4
4
3
4
.
©
Keywords: Hydrogen storage materials; NaAlH4; Catalysis; Kinetics; X-ray diffraction
1
. Introduction
gested to act as active catalysts. Among them, Al3Ti is the
◦
most thermodynamically stable phase (ꢀG = −136 kJ/mol),
f
◦
◦
f
Upon doping with a Ti-based catalyst, NaAlH4 may undergo
followed by TiH2 (ꢀG = −86 kJ/mol) and AlTi alloy (ꢀG =
f
reversible dehydrogenation following Eq. (1) under moderate
temperatures.
−72 kJ/mol). The speculation that Al3Ti acts as active species
was also supported by theoretical calculation, X-ray absorption
and electron-microscopy studies [19,21]. Actually, the forma-
tion of Al3Ti has already been experimentally confirmed in the
mechanically milled 4:1 LiAlH4/TiCl4 and 3:1 NaAlH4/TiCl3
mixtures. And the pronounced catalytic enhancement achieved
in TiCl4 or TiCl3-doped hydrides has been partially attributed to
the in situ formed nano/micro-crystalline Al3Ti during milling
NaAlH4 ꢀ 1/3Na3AlH + 2/3Al + H2
6
ꢀ
NaH + Al + 3/2H2
(1)
The two-step reversible reactions of NaAlH4 yield a theoretical
hydrogen capacity of 5.6 wt.%. The relatively high capacity, in
combination with the moderate operation temperature and large
reversibility, makes catalytically enhanced Ti–NaAlH4 rather
attractive as a potential hydrogen storage medium for onboard
application [1–11].
While the catalytic enhancement arising upon doping
NaAlH4 with a Ti catalyst has been well established, the
mechanism by which active Ti-species catalyze the dehydrid-
ing/rehydriding reaction of NaAlH4 is still unclear. Even the
nature of active Ti–species is still a subject of speculation and
controversy. Several Ti-containing species, including zerova-
lent Ti [1,12–15], Al–Ti Alloy [9,16–23], Ti hydride(s) [23,24]
and Ti cation with variable valances [25,26], have been sug-
[9,16]. At a typical doping level of <5 mol%, however, direct
phase identification has provided no convincing evidence about
the existence of crystalline Al3Ti. Therefore, it was gener-
ally speculated that Al3Ti presents in the hydride matrix in
an X-ray amorphous state [1,12,17–19,23]. On the other hand,
the hypothesis about Al3Ti acting as active species was chal-
lenged by the property measurements. It was reported that direct
doping the hydrides with Al3Ti did not result in a consider-
able catalytic enhancement [16,29]. Therefore, whether Al3Ti
can explain the nature of active Ti-species is still an open
question.
The present study aims at clarification of the catalytic
effect of Al3Ti on the reversible dehydrogenation reactions of
∗
4 3
NaAlH . For this purpose, Al Ti was prepared and directly
Corresponding author. Tel.: +86 24 2397 1622; fax: +86 24 2389 1320.
E-mail address: pingwang@imr.ac.cn (P. Wang).
utilized as a dopant precursor to prepare doped NaAlH4, and
0
925-8388/$ – see front matter © 2006 Published by Elsevier B.V.
doi:10.1016/j.jallcom.2006.01.006

3
66
X.D. Kang et al. / Journal of Alloys and Compounds 424 (2006) 365–369
the property and structure of the Al3Ti-doped NaAlH4 were
investigated.
2
. Experimental section
The starting materials NaH (95%, 200 mesh, <74 ␮m), Al powder (99.95+%,
2
00 mesh, <74 ␮m) were all purchased from Sigma–Aldrich Corp. and used
as received. Raw NaAlH4 was purchased from Albemarle Corp. and purified
following a procedure as described in [27] before usage. Al3Ti was prepared by
arc-melting stoichiometric mixtures of pure Al (99.9%) and Ti (99.8%) metals
under Ar atmosphere in a water-cooled copper hearth. To ensure composition
homogeneity, the alloy ingot was turned over and re-melted for three times
during arc-melting. The as-prepared Al3Ti(D022) ingot was smashed and then
mechanically milled for 10 h under Ar atmosphere by using a high-energy SPEX
8
000 mill to prepare a metastable Al3Ti(L12) dopant precursor.
The powder mixtures of NaH + Al + 4 mol% Al3Ti and NaAlH4 + 4 mol%
Al3Ti were mechanically milled under Ar atmosphere by using a Fritsch 7 Plan-
etary mill at 400 rpm in a vial together with eight balls (10 mm in diameter)
made of stainless steel. The ball-to-powder weight ratio was around 40:1. All
sample operations were performed in an Ar-filled glove box equipped with a
recirculation system to keep the H2O and O2 levels < 1 ppm.
Fig. 1. XRD patterns of the conventionally prepared Al3Ti (D022) (a) and the
metastable Al3Ti (L12) formed after ball-milling for 10 h (b).
Hydriding/dehydriding behaviors of the samples were examined by using a
carefullycalibratedSievert’stypeapparatus. Atypicalcyclicexperimententailed
◦
◦
absorption at 120 C and desorption at 150 C with an initial pressure condition
about 11 MPa and <100 Pa, respectively. To minimize H2O/O2 contamination,
the high-purity hydrogen gas (with 99.999%) was further purified using a hydro-
gen storage alloy system. Hydrogen capacity was determined with respect to the
total sample weight, including the catalyst.
10 h, respectively, followed by hydrogenation with an initial H-
pressure of about 11 MPa for over 10 h. For the 1 h sample,
the long period of hydrogen exposure only resulted in a partial
restoration of Na3AlH . While for the sample milled for 10 h, a
6
largely complete restoration of NaAlH4 from the NaH/Al mix-
ture has been achieved in the hydriding process.
Theincreasedhydrogenationlevelwithincreasingthemilling
time was further quantified by dehydriding measurements. As
The samples were characterized by powder X-ray diffraction (XRD, Rigaku
D/MAX-2500, Cu K␣ radiation) and scanning electron microscope (SEM,
LEO Supra 35) equipped with an energy dispersive X-ray (EDX) analysis unit
(
Oxford). All the sample preparation was operated in the Ar-filled glove-box.
To minimize the H2O/O2 contamination during the XRD measurement, small
amount of grease was used to cover the surface of the samples. The SEM sam-
ples were prepared by spreading the dry powder on conducting tape supported
on a copper pole. The subsequent sample transferring into SEM equipment was
performed in a specially designed Ar-filled device.
seen in Fig. 3, the sample milled for 1 h only released about
◦
0
.6 wt.% hydrogen after being kept at 150 C for over 10 h.
When the milling time was increased to 10 h, the sample was
observed to release 4.0 wt.% hydrogen within about 10 h. This
value corresponds to about 80% of the theoretical hydrogen
capacity of the doped sample containing 4 mol% Al3Ti dopant.
Here, it should be noted that the increasingly favorable de-
3
. Results and discussion
/hydriding performance arising upon increasing the milling time
Al3Ti has two types of crystal structure. Upon mechanical
milling, Al3Ti may undergo polymorphic modification from
tetragonal D022- to metastable cubic L12-structure [9]. In view
of that mechanical milling has been generally adopted as a
predominant technique in preparation of doped hydrides, we
selected cubic L12-type Al3Ti as dopant to prepare doped
NaAlH4. Metastable cubic L12-Al3Ti was prepared by mechan-
ically milling the arc-melted Al3Ti under Ar atmosphere by
using a high-energy SPEX 8000 mill. As shown in Fig. 1, a
complete transformation from tetragonal D022- to cubic L12-
phase had been achieved after 10 h milling. The XRD pattern
of this metastable phase is characterized by low-intensity and
wide diffraction peaks, indicating the grain pulverization and
the introduction of micro-strain during milling process. Accord-
ing to Scherrer equation, the grain size of L12-Al3Ti after 10 h
milling was estimated to be about 5 nm.
NanocrystallineAl3Tiwasfoundtopossesscatalyticfunction
on the de-/hydriding reactions of NaAlH4. Moreover, the cat-
alytic enhancement by Al3Ti doping becomes more pronounced
with increasing milling time. Fig. 2 gives the XRD patterns of
the samples prepared by mechanical milling 1:1 NaH/Al mix-
ture with 4 mol% Al3Ti under Ar atmosphere for 1, 5, and
Fig. 2. The XRD patterns of the samples prepared by mechanically milling
NaH + Al + 4 mol% Al3Ti under Ar atmosphere for (a) 1 h; (b) 5 h; (c) 10 h.
The inset gives the magnified patterns of the marked region in Fig. 2 and the
corresponding multi-peaks lines fitted by Lorentz function.

X.D. Kang et al. / Journal of Alloys and Compounds 424 (2006) 365–369
367
Even in this case, the diffraction signal of Al3Ti only appears
as a weak and broad peak at the higher angle side of Al peaks
◦
(
at 2θ = 39.3 ). In addition, no reaction between NaH/Al mix-
ture (or sodium aluminum hydride) and nanocrystalline Al3Ti
is expected to occur during the milling process. Therefore, the
improved catalytic enhancement arising upon increasing milling
time should be correlated to the increasingly favorable disper-
sion of Al3Ti in the hydride matrix. This was evidenced by the
SEM morphology observation combined with EDX analyses.
As seen in Fig. 4, a more homogeneous distribution of the Ti-
containing particles (bright ones) in the NaH/Al matrix has been
achieved after increasing the milling time from 1 to 10 h. EDX
analyses confirm that the Al/Ti atomic ratio in the bright particles
is close to 3:1. Recently, there have been several experimental
efforts to clarify the catalytic activity of Al3Ti in the reversible
dehydrogenation of NaAlH4, but with no positive results and
conclusions being reported [16,29]. This is in great contrast to
our findings. One possible reason is that the applied milling time
in the literatures, typically around 1 h, was insufficient to achieve
◦
Fig. 3. The typically dehydriding profiles at 150 C of NaH + Al + 4 mol% Al3Ti
mixtures milled under Ar atmosphere for 1, 5 and 10 h, respectively.
were also observed in our studies on the metallic Ti-doped
NaAlH4 [27,28].
a favorable distribution of Al Ti phase in the hydride matrix.
3
We further examined the sample prepared by milling 1:1
Al3Ti is highly stable according to thermodynamic consid-
eration. It was therefore highly expected that Al3Ti retained its
phase stability during milling process. However, definite iden-
tification of Al3Ti in doped samples by XRD is difficult due to
the combined effects of co-existence of large amount of Al, the
close peak position of Al3Ti to that of Al and the nanostructure
of Al3Ti. To minimize the interference of Al phase, we exam-
ined the samples in the hydrogenated state, as given in Fig. 2.
NaH/Al with Al Ti under Ar atmosphere for 10 h, and found
that its cycling performance was highly stable. As shown in
3
Fig. 5, other than a slightly decreased H-capacity from the
NaAlH /Na AlH + Al decomposition step in the first cycle, the
4
3
6
dehydriding profiles of the following cycles are almost identical.
The samples after the first and 5 cycles, both in the hydro-
genated state, were examined by XRD. It was observed from
Fig. 6 that the phase stability of Al Ti was well maintained in
3
Fig. 4. Back scattering electron (BSE) images of NaH + Al + 4 mol% Al3Ti mechanically milled under Ar atmosphere for 1 h (top, left), 10 h (top, right), and the
representative EDX results. It should be noted that, despite the special caution taken, a significant amount of oxygen contamination has been introduced into the
samples during sample transfer.

3
68
X.D. Kang et al. / Journal of Alloys and Compounds 424 (2006) 365–369
Fig. 5. The cyclic stability of the NaH + Al + 4 mol% Al3Ti mixture milled under
Ar atmosphere for 10 h.
Fig. 7. The cyclic stability of NaAlH4 + 4 mol% Al3Ti mechanically milled
under Ar atmosphere for 10 h.
the de-/hydriding cycles. This is consistent with the recent stud-
ies using X-ray absorption spectroscopy [14,19,21].
TiF /TiCl -doped samples, the typical dehydriding periods
for fulfilling the two steps are within 10 min and 2–3 h,
respectively.
3
3
These results suggest that Al3Ti may act as active species and
catalyze the reversible de-/hydriding reactions of NaAlH4, thus
providing evidence to support the previous speculation about the
catalytic function of Al3Ti. However, before this finding could
be generalized to account for the nature of active Ti-species in
the Ti–NaAlH4 system, further investigations are still required to
achieve a comprehensive understanding of the following exper-
imental findings:
(2) The variation of starting materials for Al Ti doping leads
3
to substantially different hydrogen storage performance. In
great contrast to the highly stable cycling performance of the
1:1 NaH/Al mixture doped with Al Ti (as seen in Fig. 5),
3
the sample prepared by milling NaAlH with Al Ti was
4
3
observed to undergo serious capacity degradation during
the first several cycles. As shown in Fig. 7, the H-capacity
decreased from the initial 3.2 to about 2.0 wt.% after four
cycles, especially with a degradation of ∼0.8 wt.% in the
first two cycles. In addition, the dehydriding kinetics of the
sample doped from NaAlH4 was found to be significantly
lower than that of the sample doped from the 1:1 NaH/Al
mixture.
(
1) Catalytic effect of Al3Ti was significantly weaker than that
of Ti(III) or Ti(IV) dopants on the sorption properties of
NaAlH4. As seen in Fig. 5, the Al3Ti-doped hydrides take
about 1 h to fulfill the first dehydriding step, and another
◦
0 h to complete the second step at 150 C. While for the
1
(
3) Recently, we have found that mechanical milling the 1:1
NaH/Al mixture (or NaAlH4) together with metallic Ti pow-
der resulted in the in situ formation of nanocrystalline Ti
hydride, which might act as active species to enhance the
reversible dehydrogenation of NaAlH4 [30]. This finding
further complicated the situation of the Ti–NaAlH4 system.
In view of thermodynamic consideration, both Al3Ti and
TiH2 are more stable than zerovalent Ti. Experimentally,
the phase stability of both phases in the de-/hydriding cycles
was evidenced by the XRD examinations [30]. Therefore,
it is hard to generalize a common active Ti-species origi-
nated from different Ti-sources. Rather these findings sug-
gest a possibility that there exist different active Ti-species
that can catalyze the reversible de-/hydriding reactions of
NaAlH4.
Further research efforts to answer these “puzzling” questions
may provide valuable hints for understanding the key aspects
of the catalytically enhanced Ti–NaAlH4 system, including the
effect of distribution of catalyst particles, the nucleation/growth
processes of the parent hydride with the presence of catalyst
particles, and the catalysis mechanism.
Fig. 6. The XRD patterns of the NaH + Al + 4 mol% Al3Ti mixture milled under
Ar atmosphere for 10 h. (a) in the hydrogenated state in the first cycle; (b) in the
hydrogenated state in the sixth cycle. The inset gives the magnified patterns of
the marked region in Fig. 6 and the corresponding multi-peaks lines fitted by
Lorentz function.

X.D. Kang et al. / Journal of Alloys and Compounds 424 (2006) 365–369
369
4
. Conclusions
[9] V.P. Balema, J.W. Wiench, K.W. Dennis, M. Pruski, V.K. Pecharsky, J.
Alloys Compd. 329 (2001) 108.
[
[
10] P. Chen, Z. Xiong, J. Luo, J. Lin, K.L. Tan, Nature 420 (2002)
DirectlydopingthehydridewithAl3Tithroughlongperiodof
302.
mechanical milling results in a considerable catalytic enhance-
ment on the reversible dehydrogenation of NaAlH4. This finding
supports the previous hypothesis about the catalytic function of
Al3Ti. However, the catalytic effect from Al3Ti doping was sub-
stantially weaker than from Ti-halide doping. In addition, dehy-
driding kinetics and cycling performance of the Al3Ti-doped
hydrides was found to be highly dependent on the starting mate-
rial, 1:1 NaH/Al mixture or NaAlH4. Therefore, further investi-
gations on distribution of the catalyst particles and its influence
on the H-storage performance of the material, and the nucle-
ation/growth processes involved in de-/hydriding reactions of
NaAlH4 are still required to firmly establish the catalytic mech-
anism of Al3Ti in the reversible dehydrogenation of NaAlH4.
11] A. Z u¨ ttel, S. Rentsch, P. Fischer, P. Wenger, P. Sudan, P. Muron, C.
Emmenegger, J. Alloys Compd. 356–357 (2003) 515.
[12] B. Bogdanovi c´ , R.A. Brand, A. Marjanovic, M. Schwickardi, J. T o¨ lle, J.
Alloys Compd. 302 (2000) 36.
[
13] J.M. Bellosta, V. Colbe, B. Bogdanovi c´ , M. Felderhoff, A. Pommerin, F.
Sch u¨ th, J. Alloys Compd. 370 (2003) 104.
[14] A. Leon, O. Kircher, J. Rothe, M. Fichtner, J. Phys. Chem. B 108 (2004)
1
6372.
[15] F. Sch u¨ th, B. Bogdanovi c´ , M. Felderhoff, Chem. Commun. (2004)
249.
16] E.H. Majzoub, K.J. Gross, J. Alloys Compd. 356–357 (2003)
63.
2
[
[
3
17] B. Bogdanovi c´ , M. Felderhoff, M. Germann, M. H a¨ rtel, A. Pommerin,
F. Sch u¨ th, C. Weidenthaler, B. Zibrowius, J. Alloys Compd. 350 (2003)
246.
[
18] C. Weidenthaler, A. Pommerin, M. Felderhoff, B. Bogdanovi c´ , F. Sch u¨ th,
Phys. Chem. Chem. Phys. 5 (2003) 5149.
Acknowledgements
[
19] M. Felderhoff, K. Klementiev, W. Gr u¨ nert, B. Spliethoff, B. Tesche, J.M.
Bellosta, von. Colbe, B. Bogdanovi c´ , M. H a¨ rtel, A. Pommerin, A. Sch u¨ th,
C. Weidenthaler, Phys. Chem. Chem. Phys. 6 (2003) 4369.
20] H.W. Brinks, C.M. Jensen, S.S. Srinivasan, B.C. HauBack, D. Blanchard,
K. Murphy, J. Alloys Compd. 376 (2004) 215.
The financial supports for this research from the Hundred
Talents Project of Chinese Academy of Sciences, the National
Natural Science Foundation of China (Project 50571099), the
Australian Research Council through ARC LP0562609 and
the ARC Centre for Functional Nanomaterials are gratefully
acknowledged.
[
[
[
21] J. Graetz, J.J. Reilly, J. Johnson, AYu. Ignatov, T.A. Tyson, Appl. Phys.
Lett. 85 (2004) 500.
22] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, Mater. Sci. Eng. B
108 (2004) 54.
[
[
23] V.P. Balema, L. Balema, Phys. Chem. Chem. Phys. 7 (2005) 1310.
24] K.J. Gross, E.H. Majzoub, S.W. Spangler, J. Alloys Compd. 356–357
References
(
2003) 423.
[
[
[
[
1] B. Bogdanovi c´ , M. Schwickardi, J. Alloys Compd. 253–254 (1997) 19.
2] L. Schlapbach, A. Z u¨ ttel, Nature 414 (2001) 353.
3] C.M. Jensen, K.J. Gross, Appl. Phys. A 72 (2001) 213.
4] K.J. Gross, G.J. Thomas, C.M. Jensen, J. Alloys Compd. 330–332 (2002)
[25] D. Sun, T. Kiyobayashi, H.T. Takeshita, N. Kuriyama, C.M. Jensen, J.
Alloys Compd. 337 (2002) L8.
[26] J. Iniguez, T. Yildirim, T.J. Udovic, E.H. Majzoub, M. Sulic, C.M. Jensen,
Phys. Rev. B 70 (2004) 060101.
6
83.
[27] P. Wang, C.M. Jensen, J. Alloys Compd. 379 (2004) 99.
[28] P. Wang, C.M. Jensen, J. Phys. Chem. B 108 (2004) 15827–15829.
[29] M. Resan, M.D. Hampton, J.K. Lomness, D.K. Slattery, Int. J. Hydrogen
Energy 30 (2005) 1417.
[30] P. Wang, X.D. Kang, H.M. Cheng, J. Phys. Chem. B 109 (2005)
20131–20136.
[
[
[
[
5] A.M. Seayad, D.M. Antonelli, Adv. Mater. 16 (9–10) (2004) 765.
6] W. Grochala, P. Edwards, Chem. Rev. 104 (2004) 1283.
7] E. David, J. Mater. Process Technol. 162–163 (2005) 169.
8] A. Zaluska, L. Zaluski, J.O. Str o¨ m-Olsen, J. Alloys Compd. 307 (2000)
1
57.