Comparative study of dehydrogenation of sodium aluminum hydride wet-doped with ScCl3, TiCl3, VCl3, and MnCl2

Article abstract of DOI:10.1016/j.jallcom.2008.03.093

A comparative study of the dehydrogenation of pure, ScCl₃-, TiCl₃-, VCl₃-, and MnCl₂-doped sodium alanate is reported. The samples wet-doped with transition metal halides exhibit significant lowering of both fir

Full text of DOI:10.1016/j.jallcom.2008.03.093

Journal of Alloys and Compounds 471 (2009) L16–L22
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
Letter
Comparative study of dehydrogenation of sodium aluminum
hydride wet-doped with ScCl3, TiCl3, VCl3, and MnCl2
Mehraj-ud-din Naika, Sami-ullah Rathera, Renju Zachariaa, Chang Su Sob,
Sang Woon Hwanga, Ae Rhan Kimb, Kee Suk Nahma,b,∗
a Nanomaterial Research Center and School of Chemical Engineering and Technology, Chonbuk National University, Chonju 561-756, Republic of Korea
b Department of Hydrogen and Fuel Cells Engineering, Chonbuk National University, Chonju 561-756, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
A comparative study of the dehydrogenation of pure, ScCl3-, TiCl3-, VCl3-, and MnCl2-doped sodium
alanate is reported. The samples wet-doped with transition metal halides exhibit significant lowering
of both first and second dehydrogenation temperatures when compared with the pure sodium alanate.
The reduction of the dehydrogenation temperature, ꢀTdec, is maximum for the VCl3-doped alanates: 37
and 55 ◦C, respectively, for the decomposition steps yielding Na3AlH6 and NaH. Among the transition
metal halides, the order in which the decomposition temperature is lowered is Mn < Sc < Ti < V. The reduc-
tion in the dehydrogenation temperature of alanates and the observed order are explained by considering
the Pauling’s electronegativity of Sc, Ti, V, Mn, and Na. Transition metals inductively diminishes the bond
strength of the adjacent Al–H bond and renders its easy cleavage to form molecular hydrogen. The tran-
sition metal-induced instability shows a quadratic dependence on the electronegativity, in accordance
with the Pauling’s electronegativity equation.
Received 31 October 2007
Received in revised form 24 March 2008
Accepted 25 March 2008
Available online 7 May 2008
Keywords:
Phase transitions
Metal hydride
Precipitation
X-ray diffraction
Oxidation
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
xNaAlH4 + yTiCl3 → ( x − 3y)NaAlH4
+ yTi + 3yNaCl + 3yAl + 6yH2
(1)
The development of safe solid-state hydrogen storage materi-
als that exhibit sufficiently fast hydrogenation–dehydrogenation
kinetics at ambient temperatures is perceived as the most vex-
ing challenge for the wider utilization of hydrogen fuel. Complex
chemical hydrides, such as NaAlH4, LiBH4, LiNH2, etc., possess
high hydrogen content, owing to which can exceed the long-term
DOE hydrogen storage target [1]. Nevertheless, complex chemi-
cal hydrides are well known for their thermodynamic stability,
extremely limited, and irreversible kinetics that render them unac-
ceptable for automobile applications [2]. In 1997, Bogdanovic´ and
Schwickardi achieved a significant breakthrough by improving the
kinetics of alanate dehydrogenation by doping the prototypical
alanate, NaAlH4 with titanium chloride (TiCl3) and titanium ortho-
butoxide (Ti-o-Bun) [3].
According to this reaction, it is clear that in a pre-mixed compo-
sition, 3 mol of sodium alanate is rendered unusable for each mole
of TiCl3 used. Additionally, the above scheme indicates separation
of Ti and Al phases, while it has been argued that Ti-doped NaAlH4
produces Al3Ti or other Ti–Al alloy phases [5]. This is because
the formation of Al3Ti from a mixture of Al and Ti is thermody-
namically favorable with the free energy change of 136 kJ mol−1
[6]. Direct doping of NaAlH4 with Al3Ti, indeed, improved its de-
/hydriding kinetics suggesting the likelihood of catalysis by Al3Ti
[7]. Likewise, the X-ray diffraction (XRD) analysis of ball-milled
TiCl3-doped NaAlH4 shows evidences of in situ Ti–Al alloy for-
mation [8]. However, the existence of pure Al in zero-valent state
has been shown using the photoelectron spectra, which indicates
the absence of Al3Ti alloy phase [9]. Thus, presently the origin
and mechanism of transition metal-mediated alanate dehydro-
genation are intensely debated. The application of the transition
metal halides other than TiCl3 and Ti-based compounds is not
investigated extensively and is restricted to few studies [10–12].
Resan et al. compared the catalysis by FeCl3 with that of TiCl3
and TiCl4 and found it inferior to the latter [10]. Blanchard et
al. found that VCl3 reduces the LiAlH4 decomposition tempera-
ture by 60 ◦C, slightly better than TiCl3·(1/3)(AlCl3), which reduces
A general scheme of reaction between sodium alanate and TiCl3,
as shown below involves the liberation of hydrogen [4]:
∗
Corresponding author at: Nanomaterial Research Center and School of Chem-
ical Engineering and Technology, Chonbuk National University, Chonju 561-756,
Republic of Korea. Tel.: +82 63 270 2311; fax: +82 63 270 3909.
E-mail address: nahmks@chonbuk.ac.kr (K.S. Nahm).
0925-8388/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2008.03.093

M.-u.-d. Naik et al. / Journal of Alloys and Compounds 471 (2009) L16–L22
L17
the temperature by 50 ◦C [11]. Furthermore, they found that VCl3
induces less decomposition during ball milling when compared
with the Ti-based additive. An extensive study of hydrogen desorp-
tion kinetics of various transition metal halides ball-milled with
NaAlH4 is performed by Anton [12]. Moreover, the ball milling of
pure NaAlH4 activates it so as to allow desorption of hydrogen at
a significant rate. Using the Vegard’s law, it has been observed
that cations having a radius in the range of 0.73–0.80 A should
hold the highest potential for enhanced hydrogen desorption rate
[12].
begin our discussion on the dehydrogenation of doped alanates by
referring to the thermogravimetric profiles of alanates presented
in Fig. 2, where (a), (b), (c), (d), and (e) represent the TGA data of
pure, ScCl3, TiCl3, VCl3, and MnCl2-doped sodium alanate, respec-
tively. In the inset, the differential of the corresponding profiles are
provided. For pure sodium alanate, the onset of the first decompo-
sition occurs at 169.4 ◦C, much less compared to 187 ◦C, previously
reported by Zhang et al. [13]. The decomposition rate of the first
step yielding Na3AlH6 (Eq. (2)), increases dramatically at 196.82 ◦C
and remains constant until 203 ◦C.
˚
In this letter, we compare the kinetic enhancement and degree of
dehydrogenation of TiCl3-, VCl3-, ScCl3-, and MnCl2-doped sodium
aluminum hydride with that of un-doped pure sodium alanate. The
alanates are wet-doped using condensed-phase reaction involving
predetermined quantities of alanate and transition metal halides
in tetrahydrofuran (THF) medium, under moisture-free Schlenk
tube technique. The dehydrogenation of alanates and evolution of
alanate phases are investigated using thermogravimetric and XRD
analysis of the samples. The decomposition temperature of transi-
tion metal-doped alanates is correlated to the alanate–hydrogen
bond strength and shows a quadratic dependence on the elec-
tronegativity of the transition metals, in agreement with the
Pauling’s electronegativity relation.
NaAlH4
→
1 Na3AlH6
+
2 Al + H2
(2)
3
3
At 226 ◦C, the TG profile indicates a shift in the rate of decom-
position implying the onset of the second decomposition, yielding
NaH:
Na3AlH6 → 3NaH + Al + 3 H2
(3)
2
The amount of hydrogen released in the first step is 1.15 wt.%,
which is equivalent to 31% of the theoretical maximum. The sec-
ond stretch of decomposition that starts at 226 ◦C continues until
285.6 ◦C. The derivative of the TG profile suggests four distinct peaks
in this temperature range. The peak at 251 ◦C can readily be ascribed
to the phase transition of ␣-Na3AlH6 to ␤-Na3AlH6, while the 265 ◦C
peak corresponds to the decomposition of Na3AlH6 yielding NaH
[3]. In order to confirm the temperature evolution of Na3AlH6 and
NaH phases, the XRD spectra of sodium alanate heated to five dif-
ferent temperatures, i.e., RT, 140, 180, 220, and 260 ◦C are compared
in Fig. 3. Initially at room temperature, the XRD spectrum exhibits
the characteristic peaks of NaAlH4. After subjected to 140 ◦C, its
partial decomposition leads to the formation of Na3AlH6 and Al,
which is evident from the XRD spectrum. However, the decrease
in intensity of NaAlH4 peaks indicates the decomposition of par-
tial amount of NaAlH4 at this stage. At 180 ◦C the peak at 2ꢃ = 37◦
shows a broad feature which matches with the XRD peak positions
of both NaAlH4 and Na3AlH6 (JCPDS file no. 20-1072). Thus, even
at 180 ◦C, the first decomposition of the sodium alanate is appar-
ently incomplete. However, with the further rising of temperature,
a slight shift of this peak to lower 2ꢃ and typical sharpening indi-
cate the completion of the first step of reaction. More importantly,
as the temperature increases to 220 ◦C, the main peaks of NaAlH4
at 2ꢃ = 31.5◦ (1 0 3) and 35.5◦ (2 0 0) disappeared implying the com-
pletion of the first decomposition step. Also, very high intensity of
the aluminum peak (1 1 1) at 220 ◦C, here, implies that the formed
Na3AlH6 partially decomposes yielding Al and NaH. This is not sur-
prising, as in TGA the onset of decomposition is close to 220 ◦C.
The small peak recorded at 2ꢃ = 35.26◦ (1 0 4) matches well with
the main peak of aluminum oxide (JCPDS file no. 46-1212), which
presumably formed by the oxidation of aluminum when negligible
amount of O2 is present in the reaction medium. The formation
of Al2O3 has been previously observed by Herberg et al. using
the Al27 nuclear magnetic resonance spectroscopy [14]. The XRD
spectra of transition metal-doped alanates, however, do not show
any peak of transition metal halides; which is not unexpected as
the metal halides are present only in very small quantities. Nev-
ertheless, the presence of TiCl3 (JCPDS file no. 18-1396) is clearly
evident with a higher dopant concentration of 15 wt.%, shown in
Fig. 4.
2. Experimental
2.1. Purification and doping of sodium alanate
TiCl3, VCl3, MnCl2, THF, and n-hexane were sure-sealed reagents with purity
of 99.95% and were purchased from Aldrich. Sodium alanate of purity 97% was
obtained from Aldrich and was purified by re-crystallizing from THF solvent using
n-hexane. For doping of alanates with transition metals, 20, 25, 10, and 5 mmol of
sodium alanate were reacted with 0.5 mmol of TiCl3, VCl3, 0.25 mmol of MnCl2, and
0.1 mmol of ScCl3, respectively. The mole ratios corresponding to the used reagents
are 2.4 mol% (TiCl3) and 2.0 mol% (VCl3, MnCl2, and ScCl3). The reaction was per-
formed at 35 ◦C for nearly 20 h under inert atmosphere using standard Schlenk
technique, so as to avoid the contamination of air- and moisture-sensitive chem-
icals. After the reaction, solution was filtered and traces of solvent in the precipitate
were removed by heating the samples at 80 ◦C under vacuum.
2.2. Characterization of pure and doped alanates
The thermogravimetric analysis of pure and doped alanates was recorded
using thermogravimetric analyzer TQ50 under N2 atmosphere. The TGA was col-
lected using a ramp rate of 5 ◦C/min. The powder XRD spectra of alanates were
recorded using Rigaku D/MAX 2500 X-ray diffractometer equipped with Cu K␣
radiation source (ꢁ = 0.15418 nm). The morphology of as-purchased and doped
alanates was compared using
a field emission scanning electron microscope,
FESEM (Hitachi S-4700). Fourier transform infrared (FTIR) spectra of the samples
were recorded using a Varian FTS 1000 spectrometer, using the dry KBr pellet
method.
3. Results and discussion
The morphology of the alanates is displayed in Fig. 1(A), where
the panels (a), (b), (c), (d), and (e) represent the FESEM images
and EDS of pure, ScCl3-, TiCl3-, VCl3-, and MnCl2-doped sodium
alanates, respectively. One can see that particle size of alanates
becomes finer after the doping. The histograms showing the par-
ticle size distributions (PSDs) of pure, ScCl3-, TiCl3-, VCl3-, and
MnCl2-doped sodium alanates are constructed by sampling sev-
eral FESEM images and are provided in Fig. 1B(a)–(e). The pure,
ScCl3-, TiCl3-, VCl3-, and MnCl2-doped sodium alanate particles
formed possess broad diameter distribution that peaked at 13.51,
5.01, 8.72, 4.50, and 5.66 ␮m, with a standard deviation, ꢂ = 5.82,
3.39, 4.89, 2.26, and 3.60 ␮m. EDS of each sample shows the pres-
ence of sodium, aluminum, and transition metal dopants such as
Sc, Ti, V, and Mn as shown in Fig. 1(a)–(e), respectively. In addi-
tion, the major elemental composition present in the EDS profiles
were Na and Al in the ratio of approximately 2:3, respectively. We
On adding TiCl3, the first decomposition starts prematurely at
139 ◦C (Fig. 2, trace c). The first decomposition step proceeds until
175.3 ◦C. The weight loss recorded on completion of the first decom-
position is 2.97 wt.%. Note that there is a loss of hydrogen in doped
sample according to alanate decomposition Eq. (1), which essen-
tially implies that only (x − 3y) moles of usable alanate per y moles
of the dopant. Therefore, in the determination of the hydrogen con-
tent of doped alanates, we need to account for the weight of the

L18
M.-u.-d. Naik et al. / Journal of Alloys and Compounds 471 (2009) L16–L22
dopant as well as the partial loss of hydrogen. This is done by deter-
mining the weight of the sodium alanate available in the doped
sample that remains after the doping reaction, and by replacing Ti
by V, Sc, and Mn appropriately in Eq. (1).
and
HS2
(1/2)(x − 3y)H2
=
× 100,
(5)
(x − 3y)NaAlH4 + y Ti + 3y NaCl + 3y Al
The theoretical maximum hydrogen contents that can be
released in the first and second decomposition steps of TiCl3-doped
alanate are given by
respectively. Here x and y are the number of moles of sodium
alanate and TiCl3 used. The maximum hydrogen content that is
obtained by the initial decomposition of 20 mmol sodium alanate
mixed with 0.5 mmol of TiCl3, i.e., HS1 is 3.24 wt.%. From the ther-
mal decomposition profile of TiCl3-doped alanate, the completion
of the first decomposition yielded 2.97 wt.%; implying 90% of the
(x − 3y)H2
HS1
=
× 100
(4)
(x − 3y)NaAlH4 + y Ti + 3y NaCl + 3y Al
Fig. 1. (A) FESEM and EDS images of pure and doped sodium alanates detailing the morphology, and elemental analysis: (a) pure sodium alanate, (b) ScCl3-doped sodium
alanate, (c) TiCl3-doped sodium alanate, (d) VCl3-doped sodium alanate, and (e) MnCl2-doped sodium alanate. (B) Provide the histograms of particle size distribution
corresponding to the samples of pure sodium alanate (a), ScCl3-doped sodium alanate (b), TiCl3-doped sodium alanate (c), VCl3-doped sodium alanate (d), and MnCl2-doped
sodium alanate (e). The average diameter of these particles is found to be 13.51 5.0, 5.01 2.2, 8.72 3.5, 4.50 2.0, and 5.66 2.5 ␮m.

M.-u.-d. Naik et al. / Journal of Alloys and Compounds 471 (2009) L16–L22
L19
Fig. 1. (Continued) .
Fig. 2. Thermogravimetric (TG) profiles of pure and doped sodium alanate performed under N2 atmosphere. The TG data of (a) pure sodium alanate, (b) ScCl3-doped sodium
alanate, (c) TiCl3-doped sodium alanate, (d) VCl3-doped sodium alanate, and (e) MnCl2-doped sodium alanate are represented. The data are collected using a temperature
ramp of 5 ◦C/min. The inset displays the differential thermal analysis (DTA) of the samples, where profiles (a), (b), (c), (d), and (e) correspond to the DTA plots of pure and
scandium, titanium, vanadium, and manganese chloride-doped sodium alanate, respectively.
theoretical maximum hydrogen content. Comparing this with the
hydrogen released from pure alanate suggests an improvement by
a factor of three. In the second decomposition step, the maximum
content estimated using Eq. (3) is 1.7 wt.%, while experimentally
we obtained 1.19 wt.%, indicating the release of 73% of the the-
oretical content. The total amount of hydrogen released in both
steps is equivalent to 4.16 wt.%, which is slightly lower than the
amount of hydrogen released by the pure alanate, i.e., 4.43 wt.%.
We subsequently estimated the hydrogen contents in ScCl3-, VCl3-
, and MnCl2-doped NaAlH4 by using the actual amount of these
halides and replacing Ti in Eqs. (2) and (3) by Sc, V, and Mn, respec-
tively. The results of these estimations are compared in Table 1.
The numbers in the parentheses indicate the lowered decomposi-
tion temperature with respect to the pure alanate. The onset of the
decomposition temperatures are obtained from the tangents to the
TGA profiles.
Comparing the decomposition temperature as well as the actual
amount of hydrogen released of pure alanate, it is evident that
doping of all four metal halides had catalyzed the alanate decom-
position. The decomposition temperature is lowered by up to 51 ◦C
when NaAlH4 was wet-doped with TiCl3, which closely agrees with
the previously reported Tdec of Ti-o-Bun-doped sodium alanate [13].
Among the four transition metal halides, we found VCl3 is best
suited in lowering the initial decomposition temperature; which

L20
M.-u.-d. Naik et al. / Journal of Alloys and Compounds 471 (2009) L16–L22
Table 1
Comparison of the hydrogen content of pure and doped sodium alanate
Maximum capacity
% of H2 released compared to maximum capacity
Eq. (2)
Eq. (3)
First decomposition
First and second decomposition
Pure NaAlH4
ScCl3–NaAlH4
TiCl3–NaAlH4
VCl3–NaAlH4
MnCl2–NaAlH4
3.7
3.3
3.3
3.3
3.3
1.8
1.7
1.7
1.7
1.7
31.1%
80.3%
89% (26 4 ◦C)
90% (32 4 ◦C)
92% (37 4 ◦C)
84% (35 4 ◦C)
77% (31 5 ◦C)
83% (51 4 ◦C)
78% (55 4 ◦C)
80% (35 5 ◦C)
The driving force for this reaction is the precipitation of highly
ionic NaCl in the THF medium. Furthermore, a qualitative picture
of this reaction can be easily developed by utilizing the Pearson’s
hard soft acid–base (HSAB) concept [15]. The HSAB principle out-
lines that hard acids react faster with hard bases and form stronger
bonds while soft acids react faster with soft bases to form stronger
bonds [15]. In the Pearson’s scale, hardness of hypervalent tran-
sition metal ions, such as Sc3+ is higher than that of monovalent
Na+ ion (ꢄ = 24.6 eV vs. 21.1 eV) [15]. Thus, intrinsically hard com-
−
plex anion, AlH4 prefers to bond with the transition metal ion
and forms transition metal alanate, M(AlH4)3. Once, the transition
metal alanate is formed, its thermal decomposition can be readily
understood in terms of the dissociation of Al–H bonds, the pairing
of H atoms to form molecular hydrogen and the electronic con-
tribution of transition metals. In Ref. [16], Grochala and Edwards
outlined a general dehydrogenation mechanism of alanates. This
mechanism involves the dissociation of Al–H bonds and the pair-
ing of four ligated H atoms to form two hydrogen molecules [16].
A simple way to accommodate the electronic contribution of the
transition metal counter ions in the cleavage of Al–H bonds is to
consider the transition metal electronegativity. The Pauling’s elec-
tronegativity ꢅP of transition metals increases from 1.36 to 1.63
as we move across the periodic table from Sc to V (see Table 2).
The rising electronegativity of transition metals inductively dimin-
ishes the electron density between Al and H, which renders an easy
cleavage and forms molecular hydrogen.
Fig. 3. Thermal evolution of XRD patterns of pure sodium alanate collected at RT,
140, 180, 220, and 260 ◦C.
lowers the Tdec of first and second decomposition by up to 37 and
55 ◦C, respectively. The order in which the initial decomposition
temperature is lowered is Mn < Sc < Ti < V.
To understand the transition metal halide-mediated dehydro-
genation of sodium aluminum hydride, the mechanism of the
alanate decomposition is concisely considered in the following. Ini-
tially, the reaction between the NaAlH4 and metal halide, which
leads to the formation of the transition metal alanate, M(AlH4)3
(M = Sc, Ti, V) is envisaged:
This is, in principle, in accordance with the kinetic stabilization
of hydrides when hydrogen atoms are substituted with electron
donating groups, as observed by Brynda et al. [17]. A correlation
between the electronegativity and the bond strength is given by
the Pauling’s definition of electronegativity [18]:
ꢀ
Edalanate–H − [EdAl–Al + EdH
]
ꢅalanate − ꢅH = eV0.5
(7)
2
3NaAlH4 + MCl3 → M(AlH4)3 + 3NaCl ↓
(6)
2
Here ꢅalanate and ꢅH represent the Pauling electronegativity of
–MAlH3 species and H, respectively, while, EdAl–H, EdAl–Al and EdH
are the bond dissociation energies of Al–H, Al–Al, and H–H, respec2-
tively. The electronegativity of transition metal alanates –MAlH3,
ꢅalanate are determined using the super-atom approximation and
ꢁ
employing, ꢅalanate = [Vc × ꢅc +
Niꢅi]/N, where Vc, and ꢅc are
i
the valence and electronegativity of the central atom in MAlH3
species. Ni and ꢅi are the number and electronegativities of ith
species attached to the Al, and N is the sum of the valence of the
Table 2
Comparison of the Pauling’s electronegativities of ions
Ion
ꢅP
ꢅalanate
Na+
AlH4
Sc3+
Ti3+
V3+
0.93
1.95
1.36
1.54
1.63
1.55
1.77
−
1.83
1.85
1.87
Fig. 4. XRD pattern of sodium alanate treated with 15 mol% of TiCl3. The reaction is
performed for 2 h at RT.
Mn2+

M.-u.-d. Naik et al. / Journal of Alloys and Compounds 471 (2009) L16–L22
L21
Fig. 6. The FTIR spectra of pure and TiCl3-doped sodium alanate in the region of
Al–H stretching vibrations.
[21,22]. In TiCl3-doped sodium alanate we observe red-shift of
this band by ∼20–1635 cm−1. This lowering of ꢆAl–H confirms the
decreased covalent character of Al–H bond when the samples are
doped with TiCl3 [22].
4. Conclusions
To summarize, we compared the dehydrogenation of ScCl3,
TiCl3, VCl3, and MnCl2-doped sodium alanate with un-doped
pure sodium alanate. The doped samples made using wet-doping
method in THF indicate the reduction in the initial dehydrogenation
temperature by as much as 37 ◦C and the second dehydrogenation
by as much as 55 ◦C. Among the transition metal halides, the order
in which the initial decomposition temperature is lowered is found
to be Mn < Sc < Ti < V. The significant decrease in the initial dehy-
drogenation temperature when V is used, is attributed to the high
electronegativity of the metal, which inductively diminishes cova-
lent bond strength of Al–H bond and renders an easy cleavage to
form molecular hydrogen. The dopant-mediated decrease in bond
strength is, indeed, manifested by a red-shift of Al–H vibrational
frequency in FTIR spectra. The instability induced by the transition
metal doping shows a quadratic dependence on electronegativity,
in accordance with the Pauling’s electronegativity equation.
Fig. 5. (a) Monotonic dependence of first decomposition temperature on the
alanate–hydrogen bond strength for pure and transition metal-doped sodium
alanate. (b) The quadratic dependence of the first decomposition temperature on
the electronegativity difference of MAlH3 and H, i.e., ꢅalanate − ꢅH is provided.
central atom and the number of substituents connected to the cen-
tral atom. In Table 2, ꢅalanate estimated using Vc = 3, ꢅc = 1.61, NH = 3,
ꢅH = 2.20, NM = 1 and the individual electronegativities of Sc, Ti, V,
and Mn are presented. Utilizing the group electronegativity val-
ues provided in Table 2, we estimate the bond strength of Al–H
in transition metal alanates. In Fig. 5(a), the observed monotonic
dependence of the Edalanate–H on the decomposition temperature
is depicted. This monotonic variation agrees well with Nakamori et
al.’s density functional calculations on metal borohydrides [19]. The
stability of alanates induced by the difference in the electronega-
tivity ꢅalanate − ꢅH takes a quadratic dependence according to Eq.
(7). Note that here we neglect the influence of transition metals on
the Al–Al dissociation energy and use the bond dissociation energy
of 1.93 eV [20]. We do not know the magnitude of uncertainty aris-
ing from this assumption; however, we presume that it is small
due to typically large inter-atomic separation between Al atoms in
transition metal alanates. In Fig. 5(b), the quadratic dependence of
first decomposition temperature on ꢅalanate − ꢅH is presented, as
expected from Eq. (7).
Acknowledgments
This work was supported by the Korea Research Foundation
Grant funded by the Korean Government (MOEHRD) KRF-2005-
210-D00037. Authors gratefully acknowledge Mr. Bae Tae Sung,
Korea Basic Science Institute (KBSI), Jeonju for the assistance pro-
vided for recording the field emission scanning electron microscopy
(FESEM) data analysis.
References
[1] S. Satyapal, J. Petrovic, C. Read, G. Thomas, G. Ordaz, Catal. Today 120 (2007)
246–256.
[2] J.J. Vajo, G.L. Olson, Scripta Mater. 56 (2007) 829–834.
[3] B. Bogdanovic´, M. Schwikardi, J. Alloys Compd. 253-254 (1997) 1–9.
[4] A.G. Haiduc, H.A. Stil, M.A. Schwarz, P. Paulus, J.J.C. Geerlings, J. Alloys Compd.
393 (2005) 252–263.
[5] M. Felderhoff, K. Klementiev, W. Gru¨ nert, B. Spliethoff, B. Tesche, J.M.B. von
Colbe, B. Bogdanovic´, M. Ha¨rtel, A. Pommerin, F. Schu¨ th, C. Weidenthaler, Phys.
Chem. Chem. Phys. 6 (2004) 4369–4374.
[6] E.H. Majzoub, J.L. Herberg, R. Stumpf, S. Spangler, R.S. Maxwell, J. Alloys Compd.
394 (2005) 265–270.
Additionally, the evidence for the weakening of Al–H bond
in doped alanates stems from the red-shift of the Al–H vibra-
tional frequency in FTIR spectra. In Fig. 6, we reproduce the FTIR
spectra of pure and Ti-doped sodium alanate in the frequency
region 1200–2000 cm−1. The broad band which peaks around
∼1655 cm−1, is the characteristic of Al–H deformation vibrations
[7] X.D. Kang, P. Wang, X.P. Song, X.D. Yao, G.Q. Lu, H.M. Cheng, J. Alloys Compd.
424 (2006) 365–369.

L22
M.-u.-d. Naik et al. / Journal of Alloys and Compounds 471 (2009) L16–L22
[8] C. Weidenthaler, A. Pommerin, M. Felderhoff, B. Bogdanovic´, F. Schu¨ th, Phys.
Chem. Chem. Phys. 5 (2003) 5149–5153.
[17] M. Brynda, G. Bernardinelli, C. Dutan, M. Geoffroy, Inorg. Chem. 42 (2003)
6586–6588.
[9] A. Le´on, D. Schild, M. Fichtner, J. Alloys Compd. 404–406 (2005) 766–770.
[10] M. Resan, M.D. Hampton, J.K. Lomness, D.K. Slattery, Int. J. Hydrogen Energy 30
(2005) 1413–1416.
[11] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, Mater. Sci. Eng. B 108 (2004)
54–59.
[18] IUPAC Compendium of Chemical Technology, 2nd ed., Royal Society of Chem-
istry, Cambridge, UK, 1997, the online version of the IUPAC Compendium of
Chemical Technology mostly corresponds to the second edition (1997), com-
piled by A.D. MacNaught, A. Wilkinson.
[19] Y. Nakamori, K. Miwa, A. Ninomiya, H. Li, N. Ohba, S. Towata, A. Zu¨ ttel, S. Orimo,
[12] D.L. Anton, J. Alloys Compd. 356/357 (2003) 400–404.
[13] S. Zhang, A. Taniguchi, Q. Xu, N. Takeichi, H.T. Takeshita, N. Kuriyama, T. Kiy-
obayashi, J. Alloys Compd. 413 (2006) 218–221.
Phys. Rev. B 74 (2006) 045126.
[20] D.R. Lide (Ed.), Handbook of Physics and Chemistry, 73rd ed., CRC Press, Cleave-
land, OH, 1992.
[14] J.L. Herberg, R.S. Maxwell, E.H. Majzoub, J. Alloys Compd. 417 (2006)
39–44.
[21] M. Fichtner, J. Engel, O. Fuhr, A. Glolss, O. Rubner, R. Ahlrichs, Inorg. Chem. 42
(2003) 7060–7066.
[15] R.G. Parr, R.G. Pearson, J. Am. Chem. Soc. 105 (1983) 7512–7516.
[16] W. Grochala, P.P. Edwards, Chem. Rev. 104 (2004) 1283–1316.
[22] S. Gomes, G. Renaudin, H. Hagemann, K. Yvon, M.P. Sulic, C.M. Jensen, J. Alloys
Compd. 390 (2005) 305–313.