Effect of Ti-doping on the dehydrogenation kinetic parameters of lithium aluminum hydride

Article abstract of DOI:10.1016/j.jallcom.2005.09.067

The effect of Ti-doping on the dehydrogenation kinetics of lithium aluminum hydride has been investigated in this paper. For the decomposition of lithium aluminum hydride into trilithium hexahydridoaluminate apparent activation energies of 81 and 89 kJ/mol are found for the un-doped and Ti-doped sample, respectively. For the decomposition of trilithium hexahydridoaluminate into lithium hydride apparent activation energies of 108 and 103 kJ/mol are found. The differences in apparent activation energies between the un-doped and the Ti-doped sample are within experimental uncertainty, suggesting that the effect of Ti-doping on the kinetic parameters of dehydrogenation of lithium aluminum hydride, is mainly a prefactor effect.

Full text of DOI:10.1016/j.jallcom.2005.09.067

Journal of Alloys and Compounds 419 (2006) 40–44
Effect of Ti-doping on the dehydrogenation kinetic
parameters of lithium aluminum hydride
Anders Andreasena,b∗
a
Materials Research Department, Risø National Laboratory, DK-4000 Roskilde, Denmark
Interdisciplinary Research Center for Catalysis, Department of Chemical Engineering,
b
Technical University of Denmark, DK-2800 Lyngby, Denmark
Received 6 September 2005; received in revised form 27 September 2005; accepted 29 September 2005
Available online 16 November 2005
Abstract
The effect of Ti-doping on the dehydrogenation kinetics of lithium aluminum hydride has been investigated in this paper. For the decom-
position of lithium aluminum hydride into trilithium hexahydridoaluminate apparent activation energies of 81 and 89 kJ/mol are found for
the un-doped and Ti-doped sample, respectively. For the decomposition of trilithium hexahydridoaluminate into lithium hydride apparent
activation energies of 108 and 103 kJ/mol are found. The differences in apparent activation energies between the un-doped and the Ti-doped
sample are within experimental uncertainty, suggesting that the effect of Ti-doping on the kinetic parameters of dehydrogenation of lithium
aluminum hydride, is mainly a prefactor effect.
© 2005 Elsevier B.V. All right reserved
Keywords: Hydrogen storage materials; High energy ball milling; X-ray diffraction; Thermal analysis
1
1. Introduction
LiH + Al → LiAl + H2
(R3)
2
Since the discovery of reversible solid state hydrogen stor-
age in NaAlH4 [1], the alanates and borohydrides, in general,
have been subject to an intense research effort. Until present
date, main focus has been on NaAlH4 although it does not
fully meet the DOE targets for gravimetric hydrogen storage
density and thermodynamics. LiAlH4, on the other hand, has
a higher gravimetric hydrogen density and thermodynamics
allow hydrogen desorption at 1 bar even below 100 ◦C, but it
has not been investigated as thoroughly as NaAlH4.
Hydrogen desorbs from LiAlH4 in a three-step decompo-
sition, (R1) –(R3) [2–4]:
(R1) –(R3) proceeds with a theoretical hydrogen release
of 5.3, 2.6, and 2.6 wt.%, respectively.
In thermal analysis, (R1) occurs around 112–220 ◦C [5–
7] and is usually initiated by the melting of LiAlH4 (en-
dothermic), unless very low heating rates are applied, which
is subsequently transformed into solid Li3AlH6 and Al dur-
ing release of hydrogen gas (exothermic). (R2) takes place
around 127–260 ◦C (endothermic) [3,4]. (R3) (endothermic)
proceeds at elevated temperatures around 350–400 ◦C [4].
Both ball milling and doping with VCl3, TiCl3, or
TiCl3·1/3AlCl3 results in a lowering of the decomposition
temperature of both (R1) and (R2) [4–6,8]. Pure LiAlH4 does
not decompose during ball milling at ambient temperatures
[6], whereas LiAlH4 doped with catalytic additives rapidly
decomposes during milling [4,6,9,10]. Slow decomposition
after doping has even been monitored during storage under
ambient conditions [11].
1
2
LiAlH4 → Li3AlH6 + Al + H2
(R1)
(R2)
3
3
1
3
1
1
Li3AlH6 → LiH + Al + H2
3
2
∗
Present address: MAN B&W A/S, Basic Research and Emission, Tegl-
holmsgade 41, DK-2450 Copenhangen SV, Denmark. Tel.: +45 3385 2902;
fax: +45 3385 1041.
The kinetic parameters of (R1) and (R2) in terms of appar-
ent activation energies and prefactors, including the influence
E-mail address: andr1976@gmail.com.
0925-8388/$ – see front matter © 2005 Elsevier B.V. All right reserved
doi:10.1016/j.jallcom.2005.09.067

A. Andreasen / Journal of Alloys and Compounds 419 (2006) 40–44
41
of catalytic doping, have only been evaluated to a limited ex-
tent [5,12,13]. Further, no solid conclusions about the influ-
ence of doping on the kinetic parameters can be drawn due to
contradicting results [5,13]. This serves as a main motivation
for studying the influence of Ti-doping on the dehydrogena-
tion kinetics of both (R1) and (R2).
2. Experimental
Lithium aluminum hydride, LiAlH4 (purity 95% minimum, typ-
ically 97%), was obtained in powder form. Doping of LiAlH4 was
carried out by ball milling with 2 mol% TiCl3·1/3AlCl3 (purity 76–
78% TiCl3). All chemicals were purchased from Alfa Aesar (John-
son Matthey). All materials handling was carried out in the presence
of a drying agent within an argon-filled glove box.
Ball milling was utilized with a Retsch PM 100 planetary ball
mill using a wolfram carbide (WC) vial with a WC-coated aluminum
lid sealed with a rubber O-ring. Three WC balls with a diameter of
approx. 20 mm and a weight of approx. 60 g each were used for all
ball milling experiments. Typically 4–5 g of sample was ball milled
giving a ball-to-sample mass ratio of 36–45:1.
Fig. 1. XRPD pattern of as-received LiAlH4.
3.2. Dehydrogenation kinetics of as-received LiAlH4
The as-received LiAlH4 powder has been investigated by
DSC applying different heating rates cf. Fig. 2. DSC traces
shows the following features: an exothermic peak at 150 ◦C,
an endothermic peak around 170 ◦C, an exothermic peak at
178–198 ◦C, and an endothermic peak around 227–247 ◦C.
The first exotherm has previously been assigned to the
interaction of LiAlH4 with surface hydroxyl impurities [3],
and the first endotherm corresponds to the melting of LiAlH4
[2,3]:
X-ray powder diffraction (XRPD) was performed with a Bragg-
Brantano STOE diffractometer (50 kV, 300 mA, Cu K␣12 radiation
¯
˚
with λ = 1.5418 A). Powdered samples were pressed into 13 mm
diameterpelletswithahightofafewmmusingapressingtoolplaced
in the glove box. While still in the glove box, a pellet was placed
in a specially designed air tight sample holder with an aluminum
foil X-ray window. The holder was assembled and transported to
the diffractometer for subsequent XRPD analysis.
LiAlH4(s) → LiAlH4(l)
(1)
The position of the melting point of LiAlH4 is rather insen-
sitive to the choice of heating rate. The melting of LiAlH4 is
immediately followed by the second exotherm corresponding
to the decomposition of liquid LiAlH4 [2,3]:
A Seiko DSC 120 U instrument was used for differential scan-
ning calorimetry (DSC) experiments. Typically 3–7 mg of sample
was loaded in an alumina crucible in the glove box. The crucible
was placed in a sealed plastic bottle in order to protect it against
air exposure during transportation from the glove box to the DSC
instrument. However, a few seconds of air exposure could not be
avoided when the sample was transferred from the plastic bottle
to the DSC apparatus. An empty crucible was used as a reference.
The sample was heated from room temperature to 300 ◦C under
an argon atmosphere with a flow of approx. 850 ml/min applying
different heating rates.
1
2
LiAlH4(l) → Li3AlH6(s) + Al(s) + H2(g)
(2)
3
3
The second endotherm corresponds to the decomposition of
Li3AlH6 as described by (R2) [2,3,6,17].
By performing a Kissinger analysis [18], i.e. an analysis
of the sensitivity of the peak positions, in terms of Tm the
temperature of the peak maximum, to the applied heating
rate, β, the apparent activation energy, EA, can be obtained
3. Results and discussion
3.1. Initial materials characterization
Fig. 1 shows the XRPD pattern of the as-received LiAlH4.
All observed peaks except one (marked with an asterisk)
can by indexed as belonging to a monoclinic unit cell
˚
˚
(P21/c) with unit cell parameters a = 4.83 A, b = 7.83 A,
◦
˚
c = 7.92 A, and β = 112.3 in agreement with previous ob-
servations of the crystal structure of LiAlH4/LiAlD4 [14].
The unexplained peak around 2θ = 35 ◦corresponds well
with the strongest reflection of LiCl, suggesting that LiCl
is present as an impurity as also suggested by Hauback et al.
[4,14]. LiCl might originate from the preparation procedure
[15]. No other impurities, e.g. hydroxides [4] or oxides [16]
are found, suggesting that these are either absent or in an
X-ray amorphous state.
Fig. 2. Dehydrogenation of as-received LiAlH4 investigated by DSC.

42
A. Andreasen / Journal of Alloys and Compounds 419 (2006) 40–44
Fig. 4. XRPD of LiAlH4 doped with 2 mol% TiCl3·1/3AlCl3 ball milled for
1 min at 100 rpm and 5 min at 400 rpm, respectively.
However, this work is not conclusive about the chemical state
of Ti in LiAlH4. The sample ball milled for 5 min at 400 rpm
displayed a grey color indicating a significant decomposition
of LiAlH4.
Fig. 3. Kissinger plot for dehydrogenation of as-received LiAlH4.
from the following equation:
Fig. 4 shows XRPD patterns of 2 mol% TiCl3·1/3AlCl3-
doped samples ball milled for 1 min at 100 rpm and 5 min at
400 rpm. Thesampleballmilledfortheshortesttimeandlow-
est speed shows no indication of any decomposition during
the doping procedure. The diffraction pattern closely resem-
bles that of the pure LiAlH4 shown in Fig. 1 with no signs
of metallic Al/LiH in the diffraction pattern. Increasing the
milling intensity leads to an almost complete decompostion
of LiAlH4 into Li3AlH6 cf. Fig. 4. This clearly demonstrates
the effectiveness of TiCl3·1/3AlCl3 as a doping material and
the fact that LiAlH4 is thermodynamically unstable at am-
bient conditions (although kinetically stable in the absence
of a catalyst). This is in qualitative agreement with previous
observations [3,4,6].
ꢀ
ꢁ
β
d ln
Tm2
EA
R
ꢀ
ꢁ
= −
(3)
1
T
d
m
ꢀ
ꢁ
β
Tm2
Thus, EA can be obtained as the slope in a plot of ln
1
T
versus . A Kissinger analysis is applied to the exotherm
m
(R1) and the endotherm (R2) cf. Fig. 3. From the Kissinger
analysis the apparent activation energy for decomposition of
LiAlH4 into Li3AlH6, Al and H2 (R1) is found to be 81
4 kJ/mol. The apparent activation energy for decomposition
of Li3AlH6 into LiH, Al and H2 (R2) is found to be 108
8 kJ/mol.
3.3. Catalytic doping and stability of LiAlH4
3.4. Dehydrogenation kinetics of Ti-doped samples
Before ball milling the red/purple TiCl3·1/3AlCl3 was
gently mixed with the white/grayish LiAlH4 in the milling
vial resulting in a reddish powder. The color remained after
ball milling for 1 min at 100 rpm, however, the next day the
reddish color had disappeared and the powder appeared very
much like the un-doped starting LiAlH4 indicating that a re-
duction of Ti3+ in TiCl3 had taken place. A possible reduction
mechanism is outlined below [4]:
Fig. 5 shows DSC experiments performed with the sam-
ples ball milled for 1 min at 100 rpm with different applied
heating rates. The DSC traces mainly consists of two en-
49LiAlH4 + TiCl3 · 1/3AlCl3 → 46LiAlH4 + Ti + 3Al
+ 3LiCl + 1/3AlCl3 + 6H2
(4)
The reduced Ti species may possibly be distributed as metal-
lic Ti across the surface of the powdered particles [16] or
present as a TixAly alloy as proposed by Balema et al. for
TiCl4-doped LiAlH4 [10] and as also proposed for NaAlH4
[19–21]. The fact that the color of the sample ball milled for 1
min at 100 rpm did not change i.e. no darkening due to finely
dispersed Ti and Al particles, may indicate that Ti is incor-
porated into the LiAlH4 as suggested by Fichtner et al. [22].
Fig. 5. DSC for LiAlH4 doped with 2 mol% TiCl3·1/3AlCl3 ball milled for
1 min at 100 rpm at different apllied heating rates.

A. Andreasen / Journal of Alloys and Compounds 419 (2006) 40–44
43
higher activation energy for Na3AlH6 decomposition [28]).
This is similar to the reported values for (R1) and (R2) of the
un-doped samples of LiAlH4 in this study. The apparent ac-
tivation energies for dehydrogenation of un-doped NaAlH4
and Na3AlH6 is approx. 120 kJ/mol for both reactions [26].
Thus, apparently a lowering of the apparent activation en-
ergies of (R1) and (R2) in NaAlH4 is a consequence of Ti-
doping.
Early studies on the kinetics of the isothermal decomposi-
tion of un-doped LiAlH4 by McCarthy et al. [12] revealed an
apparent activation energy of (R1) to be approx. 100 kJ/mol,
although it is unclear to which degree the fitted data had
been influenced by (R2). Recent studies by Blanchard et al.
[13] shows a value of 102 kJ/mol for (R1). McCarthy et al.
[12] also reported an apparent activation energy of (R2) from
constant heating rate experiments of approx. 100 kJ/mol. The
apparent activation energies found in this work seems to be
in good agreement with both the studies of McCarthy et al.
[12] and Blanchard et al. [13].
Fig. 6. Kisinger analysis for LiAlH4 doped with 2 mol% TiCl3·1/3AlCl3
ball milled for 1 min at 100 rpm.
In addition, Blanchard et al. [13] also studied the in-
fluence of doping with catalytic additives on the apparent
activation energy of (R1). They found a reduction of ap-
prox. 5–10 kJ/mol upon doping with either TiCl3·1/3AlCl3
or VCl3. Chen et al. [5] have investigated the dehydro-
genation kinetics of TiCl3·1/3AlCl3 doped LiAlH4 and they
found apparent activation energies of 43 and 55 kJ/mol for
(R1) and (R2), respectively. The low values of apparent ac-
tivation energies reported by Chen et al. are in contradic-
tion to the ones reported in this study for Ti-doped LiAlH4
and also in contradiction to the results of Blanchard et al.
[13].
To summarize, in this study no significant effect on the ap-
parent activation energies are observed as a consequence of
Ti-doping, although a minor change cannot be ruled out due
to an uncertainty of 5–10 kJ/mol in the present experiments.
This suggests that (a) the effect of Ti-doping is different in
LiAlH4 compared to NaAlH4, and (b) since the dehydrogena-
tion temperatures shift downwards in the DSC experiments
upon doping this must be accounted for mainly in terms of
an increase in the prefactor, since the activation energies are
unaffected by the doping.
dotherms, one with maximum around 150 ◦C and the other
with maximum at approx. 200 ◦C. The first endotherm can
be assigned to the dehydrogenation of solid LiAlH4 into
Li3AlH6 according to (R1). Clearly the catalytic doping
procedure has lowered the decomposition temperature of
LiAlH4 below its melting point in agreement with the find-
ings reported in Ref. [4,5]. In this context, it should be
mentioned that the melting of LiAlH4 prior to decompo-
sition can also be avoided without doping by applying ei-
ther isothermal conditions [23] or very low heating rates [7].
However, this is of course only due to a change in reaction
conditions.
The small feature between the two dominating endother-
mic peaks, consisting of an endothermic and an exothermic
peak, can be attributed to the melting of unreacted LiAlH4
and subsequent decomposition into Li3AlH6. The second of
the two dominating endothermic peaks corresponds to the
decomposition of Li3AlH6 into LiH according to (R2). The
decomposition temperature of Li3AlH6 has also been low-
ered significantly compared to the pure LiAlH4. To further
investigate the effect of Ti-doping on the dehydrogenation
kinetics of LiAlH4 a Kissinger analysis, as outlined previ-
ously for the un-doped sample, is performed on the data
shown in Fig. 5. Fig. 6 shows the resulting Kissinger anal-
ysis and the analysis reveals apparent activation energies of
89 9 kJ/mol and 103 1 kJ/mol for (R1) and (R2), respec-
tively. The found apparent activation energies are very close
to those for the un-doped sample (81 4 kJ/mol and 108
8 kJ/mol).
In order to get a rough estimate of the increase in prefactor,
ν, upon doping the Redhead equation is used (assuming first
order desorption)
ꢂ ꢃ
ꢂ
ꢃ
EA
ν
EA
=
exp
−
(5)
RTm
β
RTm
This equation can be used to solve for ν for a given set of
parameters of β, EA and Tm. According to Figs. 2 and 5
both (R1) and (R2) is lowered by approx. 40 ◦C upon dop-
ing. Plugging in the numbers of β, EA and Tm in the above
equation, one finds the prefactor to be increased approx. by
one order of magnitude for both (R1) and (R2) after doping
with Ti.
3.5. Effect of Ti-doping on the kinetic parameters
Reported values of apparent activation energies for dehy-
drogenation of Ti-doped NaAlH4 and Na3AlH6 are found to
be in the range of 73–100 and 97–100 kJ/mol, respectively
[24–27], (with only a recent study indicating a significant

44
A. Andreasen / Journal of Alloys and Compounds 419 (2006) 40–44
[6] V.P. Balema, V.K. Pecharsky, K.W. Dennis, J. Alloys Compd. 313
4. Conclusion
(2000) 69–74.
[7] H.W. Brinks, B.C. Hauback, P. Norby, H. Fjellva˚g, J. Alloys Compd.
351 (2003) 222–227.
[8] M.D. Resan, M. Hampton, J.K. Lomness, D.K. Slattery, Int. J. Hydro-
gen Energy 30 (2005) 1417–1421
[9] V.P. Balema, K.W. Dennis, V.K. Pecharsky, Chem. Commun. (17)
(2000) 1665–1666.
[10] V.P. Balema, J.W. Wiench, K.W. Dennis, M. Pruski, V.K. Pecharsky, J.
Alloys Compd. 329 (2001) 108–114.
[11] H.W. Brinks, A. Fossdal, J.E. Fonneløp, B.C. Hauback, J. Alloys
Compd. 397 (2005) 291–295.
[12] M. McCarthy Jr., J.N. Maycook, V.R.P. Verneker, J. Phys. Chem. 72
(1968) 4009–4014.
[13] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, J. Muller, J. Alloys
Compd. 404–406 (2005) 743–747.
[14] B.C. Hauback, H. Brinks, H.W. Fjellva˚g, J. Alloys. Compd. 346 (2002)
184–189.
In this paper the effect on Ti-doping on the dehydro-
genation kinetics of lithium aluminum hydride has been
investigated. The decomposition of lithium aluminum
hydride into trilithium hexahydridoaluminate (R1) has an
apparent activation energy of approx. 85 kJ/mol and is
insensitive to Ti-doping. The decomposition of trilithium
hexahydridoaluminate into lithium hydride (R2) has an
apparent activation energy of 105 kJ/mol and is also insen-
sitive to the doping. These results suggest that the effect of
Ti-doping on the kinetic parameters of dehydrogenation of
lithium aluminum hydride is mainly a prefactor effect. A
possible explanation to this phenomenon could be that the
addition of TiCl3 increases the number of nucleation sites
by introducing defects or vacancies, e.g. by withdrawal of Li
atoms from the LiAlH4 lattice during the reduction of TiCl3
to either metallic Ti on the surface or Ti incorporated into the
lattice.
[15] E. Wiberg, E. Amberger, Hydrides of the Elements of Main Groups
I–IV, Elsevier publishing company, Amsterdam, 1971.
[16] C.M. Andrei, J.C. Walmsley, H.W. Brinks, R. Holmestad, B.C. Blan-
chard, D. Hauback, G.A. Botton, J. Phys. Chem. B 109 19 (2005)
4350–4356.
[17] L. Zaluski, A. Zaluska, J.O. Stro¨m-Olsen, J. Alloys Compd. 290 (1999)
71–78.
[18] H. Kissinger, Anal. Chem. 29 (11) (1957) 1702–1706.
[19] H.W. Brinks, C.M. Jensen, S.S. Srinivasan, B.C. Hauback, D. Blan-
chard, K. Murphy, J. Alloys Compd. 376 (2004) 215–221.
[20] B. Bogdanovic´, M. Felderhoff, M. Germann, M. Ha¨rtel, A. Pommerin,
F. Schu¨th, C. Weidenthaler, B. Zibrowius, J. Alloys Compd. 350 (2003)
246–255.
Acknowledgment
This work has received financial support from the The
Danish Technical Research Council through the Center of
Excellence Towards a hydrogen based society.
[21] E.H. Majzoub, K.J. Gross, J. Alloys Compd. 356/357 (2003) 363–367.
[22] M. Fichtner, P. Canton, O. Kircher, A.Le´on, J. Alloys Comps. 404–406
(2005) 732–737.
[23] A. Andreasen, T. Vegge, A.S. Pedersen, J. Solid State Chem. 178 (2005)
3672–3678.
References
[1] B. Bogdanovic´, M. Schwickardi, J. Alloys Compd. 253/254 (1997)
1–9.
[24] G. Sandrock, K. Gross, G. Thomas, C. Jensen, D. Meeker, S. Takara,
J. Alloys Compd. 330–332 (2002) 686–701.
[2] J.A. Dilts, E.C. Ashby, Inorg. Chem. 11 (6) (1972) 1230–1236.
[3] T.N. Dymova, D.P. Aleksandrov, V.N. Konoplev, T.A. Silina, A.S.
Sizareva, Russ. J. Coord. Chem. 20 (4) (1994) 279–285.
[4] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, Mater. Sci. Eng.
B 108 (2004) 54–59.
[25] W. Luo, K.J. Gross, J. Alloys Compd. 385 (2004) 224–231.
[26] G. Sandrock, K.J. Gross, G. Thomas, J. Alloy Compd. 339 (2002)
299–308.
[27] T. Kiyobayashi, S.S. Srinivasan, D. Sun, C.M. Jensen, J. Phys. Chem.
A 107 (39) (2003) 7671–7674.
[5] J. Chen, N. Kuriyama, Q. Xu, H.T. Takeshita, T. Sakai, J. Phys. Chem.
B 105 (2001) 11214–11220.
[28] O. Kircher, M. Fichtner, J. Alloys Compd. 404–406 (2005) 339–
342.