Reversible hydrogen decomposition of KAlH4

Article abstract of DOI:10.1016/S0925-8388(02)01307-5

We report that potassium aluminum tetrahydride, KAlH₄, exhibits a rapid reversible hydrogen decomposition reaction without any aid of external catalysts. This is significantly different from the reaction mechanism of NaAlH₄ and LiAlH₄, in which the homogeneous doping of the transition metal catalyst is essential for the reversibility and fast kinetics. The reversible dehydrogenation and hydrogenation were clearly observed without a catalyst over a temperature range 250-340°C under <10 bar of hydrogen. Hydrogen storage ability was above 3.5 wt.% (81.4% of the theoretical value of 4.2 wt.%). KAlH₄ may be a competitive option with a lightweight but high-temperature metal hydride system such as Mg₂NiH₄. These findings will also provide renewed guidelines for understanding the reaction mechanism of the catalyzed alanates as reversible hydrogen storage materials.

Full text of DOI:10.1016/S0925-8388(02)01307-5

Journal of Alloys and Compounds 353 (2003) 310–314
L
www.elsevier.com/locate/jallcom
Reversible hydrogen decomposition of KAlH4
1
2
*
Hiroyuki Morioka , Kenichi Kakizaki , Sai-Cheong Chung, Atsuo Yamada
Materials Laboratories, Sony Corporation, 2-1-1 Shinsakuragaoka, Hodogaya, Yokohama 240-0036, Japan
Received 4 November 2002; accepted 11 November 2002
Abstract
We report that potassium aluminum tetrahydride, KAlH4, exhibits a rapid reversible hydrogen decomposition reaction without any aid
of external catalysts. This is significantly different from the reaction mechanism of NaAlH4 and LiAlH4, in which the homogeneous
doping of the transition metal catalyst is essential for the reversibility and fast kinetics. The reversible dehydrogenation and hydrogenation
were clearly observed without a catalyst over a temperature range 250–340 8C under ,10 bar of hydrogen. Hydrogen storage ability was
above 3.5 wt.% (81.4% of the theoretical value of 4.2 wt.%). KAlH4 may be a competitive option with a lightweight but high-temperature
metal hydride system such as Mg2NiH4. These findings will also provide renewed guidelines for understanding the reaction mechanism of
the catalyzed alanates as reversible hydrogen storage materials.
2002 Elsevier Science B.V. All rights reserved.
Keywords: Metal hydrides; Energy storage materials; Gas–solid reactions; X-ray diffraction
1. Introduction
Schwickardi [1]. The reaction mechanism is based on
reversible decomposition, in which all of the elements
migrate during the reaction. Jensen and Gross [2] have
proposed that neutral cluster of AlH3, leaving the (NaH)
neutral clusters, are the main mobile species. This is
significantly different from the reaction mechanisms of
conventional metal hydrides using interstitial crystallo-
graphic sites for hydrogen atoms, hydrogen being the only
mobile species. Another feature of interest in Bogdanovic
and Schwickardi’s pioneering work is that they added a
catalyst to NaAlH4 to enhance the kinetics and reversibili-
ty of the reaction. The major role of the catalyst is believed
to be to cut Al31–H2 ionic bonds to form an Al-rich
surface and H2. These findings reported in 1997 spurred
the interest of the scientific community and a number of
groups have started to work on this material [1–21].
Thermolysis of NaAlH4 releases hydrogen and it con-
sists of three steps of dehydrogenation. The first two steps
are used for the reversible hydrogen releasing reaction.
Intermetallic compounds, such as in BCC, Laves, and
AB5 phases, are well known as hydrogen storage materi-
als. However, for practical use under ambient temperature
and pressure, their low energy density per unit weight is
recognized as a fatal disadvantage. For instance, the
hydrogen capacity of the most popular LaNi5-based alloys
operating at moderate temperature does not exceed 1.4
wt.%. Whereas an Mg-based alloy, for example Mg2Ni,
shows a hydrogen capacity as high as 3.6 wt.% but its
operating temperature is too high (ca. 330 8C). In conven-
tional hydrogen storage metal alloys, the reversible hydro-
genation and dehydrogenation is based on the topotactic
reaction, in which the structure of the host metal frame-
work is allowed to change only by atomic displacement,
with no diffusive rearrangement. The host must be stable
enough to tolerate repeated hydrogen insertion and ex-
traction over a finite range of solid solution.
In 1997, NaAlH4 was introduced as a novel lightweight
reversible hydrogen storage material by Bogdanovic and
3NaAlH4 → Na3AlH6 1 2Al 1 3H2 (H2: 3.7 wt.%)
Na3AlH6 → 3NaH 1 Al 1 3/2H2 (H2: 1.9 wt.%)
3NaH → 2Na 1 3/2H2 (H2: 1.9 wt.%)
(1)
(2)
(3)
*
Corresponding author.
E-mail address: hiroyuki.morioka@jp.sony.com (H. Morioka).
1Co-corresponding author.
2Present address: Department of Electronic Chemistry, Tokyo Institute
of Technology, 4259 Nagatsuta, Midori, Yokohama, 226-8502 Japan.
In step 1, NaAlH4 decomposes to Na3AlH6, followed by
0925-8388/02/$ – see front matter
doi:10.1016/S0925-8388(02)01307-5
2002 Elsevier Science B.V. All rights reserved.

H. Morioka et al. / Journal of Alloys and Compounds 353 (2003) 310–314
311
decomposition to NaH in step 2. Due to its high binding
energy, NaH is not applicable for a reversible hydrogen
source. Before the decomposition of NaH, theoretically 5.6
wt.% of hydrogen is reversibly discharged from NaAlH4.
Other alanates, such as Na2LiAlH6, LiAlH4, and Li3AlH6,
were also reported to exhibit reversible hydrogenation/
dehydrogenation upon doping with a titanium catalyst for
kinetic enhancement. However, potassium aluminum tetra-
hydride, KAlH4, which is also in the family of
A1(AlH4)2-type ionic complex hydrides, has not yet been
explored with an eye to hydrogen storage applications.
Although KAlH4 shows an apparent disadvantage in terms
of its gravimetric hydrogen capacity as compared to
NaAlH4 and LiAlH4, the thermodynamic and kinetic
parameters for KAlH4 will shed light on the reaction
mechanism of the catalyzed alanates.
the phases. Because of the hygroscopic nature of KAlH4,
the sample was set in an argon environment and was
placed in a small sealed container with clear plastic
windows that allowed X-rays to irradiate the sample.
2.3. Gas identification
Prior to the following experiments, a thermal pro-
grammed desorption (TPD) measurement was performed
in a high vacuum to confirm that desorbed gas is hydrogen.
The gas was identified using a quadrupole mass spectrome-
ter equipped in an RGA detector (SRS RGA100). The
temperature of the sample was increased at a rate of
2 8C/min during the measurement.
2.4. Dehydrogenation monitoring
The purpose of our work is to investigate KAlH4 as a
possible reversible hydrogen storage material and this led
us to the discovery that this material has an easy and
reversible hydrogenation/dehydrogenation reaction over a
temperature range of 250–300 8C under ,10 bar of H2.
The emphasis is on reactions proceeding without any
assistance from the transition metal catalyst.
The degree of dehydrogenation of KAlH4 under ambient
pressure of hydrogen was measured using a TPD measure-
ment system of original design. The desorbed hydrogen
gas in the sample holder is vented through a mass flow
meter (Lintec MC-3102E 10SCCM) and a check valve to
outside of the system so that the internal pressure is kept
constant. The quantity of desorbed gas from the sample
was obtained by measuring the positive deviation from the
background flow rate, measured using the mass flow meter.
The sample was heated at a rate of 2 8C/min from room
temperature to 500 8C. The background data were accumu-
lated without sample.
2. Experimental
2.1. Synthesis of KAlH4
KAlH4 powder was synthesized based on the method
proposed by Dymova et al. [22]. Under high pressure of
hydrogen (.175 bar) at high temperature (270 8C), the
following reaction proceeds.
2.5. Rehydrogenation monitoring
KAlH4 was heated to 390 8C, so it decomposed into KH,
Al, and H2, and then naturally cooled down to room
temperature in H2 at 1 bar. For the hydrogenation, the
decomposed compound of KH and Al were exposed to
approximately 9 bar of hydrogen gas. The pressure change
in the sample holder was then monitored while the
compound was heated from room temperature to 500 8C at
rate of 0.5 8C/min. The negative deviation from blank data
indicates the onset of rehydrogenation.
270 8C
KH 1 Al → KAlH4
(4)
.175 bar
KH in mineral oil was purchased from Aldrich Co. and
was filtered and washed several times with hexane to
separate KH from the oil, followed by evaporation of the
solvent. The filtration process was done in a glove box
filled with argon. The mixture of KH and Al (.99.5%
Wako Co.) was mechanically milled in 35-ml milling
containers with three steel balls with a diameter of 10 mm
and 40 steel balls with a diameter of 5 mm for 30 min at
500 rev./min (Fritsch Pulverisette 7). The mixture was
then transferred to a 100-ml autoclave (Taiatsu techno
TVS-N2), and the autoclave was filled with approximately
135 bar of hydrogen gas at room temperature, followed by
heating to 270 8C. This resulted in increasing the pressure
of the autoclave to more than 175 bar. The temperature
was maintained for 3 days.
2.6. Cycle tolerance test
In the dehydrogenation cycle, KAlH4 was heated at 1
bar from room temperature to 370 8C at a rate of 2 8C/min
to dehydrogenate the sample to decompose it into KH, Al,
and H2. The amount of desorbed hydrogen gas was
recorded, as described in Section 2.4.
In the hydrogenation cycle, after each dehydrogenation
process, 10 bar of hydrogen gas were slowly induced into
the sample holder at room temperature, and the sample was
heated for 10 h at two different constant temperatures. In
the first three hydrogenation cycles, the sample was heated
at 250 8C, and in the fourth hydrogenation cycle, the
sample was heated at 330 8C.
2.2. Phase identification
X-ray diffraction measurement was performed with
RINT2000 (Rigaku Co.) using CuKa radiation to identify

312
H. Morioka et al. / Journal of Alloys and Compounds 353 (2003) 310–314
K3AlH6 → 3KH 1 Al 1 3/2H2 (H2: 1.4 wt.%)
3KH → 3K 1 3/2H2 (H2: 1.4 wt.%)
(6)
(7)
The first dehydrogenation occurred at approximately
300 8C to decompose to K3AlH6 and Al (reaction (5)), and
2.4 wt.% of hydrogen was released, which is comparable
to its theoretical value of 2.9 wt.%. The second and third
dehydrogenation steps, reactions (6) and (7), occurred
approximately at 340 and 430 8C, respectively, and 1.3 and
1.2 wt.% of hydrogen was released. These values agree
reasonably well with the theoretical values, 1.4 wt.% for
both steps.
It seems, therefore, that the decompositions followed
reactions (5)–(7) and the initial content of KAlH4 in the
sample was more than 80%. The rest of the content is
assumed to be aluminum, based on the XRD data shown in
Fig. 1.
Fig. 1. X-ray powder diffraction measurement of KAlH4, synthesized
from KH and Al at 270 8C at .175 bar of hydrogen (reaction (4)).
Fig. 3 shows the process of rehydrogenation to KAlH4
from KH and Al. The measurement was carried out under
9 bar of hydrogen while the temperature of the sample was
slowly increased to 500 8C starting from a room tempera-
ture. The gradual rehydrogenation of the sample started
below 200 8C. The rate of rehydrogenation increased until
the hydrogenation reaction stopped at around 340 8C.
The preceding experiments lead to the important conclu-
sion that, even without a catalyst, potassium alanate
exhibits reversible decomposition under much less hydro-
gen pressure than Ti-doped sodium alanate, whereas the
operating temperature is much higher (Ti/NaAlH4 150–
250 8C; KAlH4 300–400 8C).
The overall dehydrogenation and hydrogenation cycle
was repeated five times.
3. Results and discussion
Fig. 1 shows the result of X-ray diffraction measurement
of the synthesized compound indicating that KAlH4 was
formed by reaction (4), though a small amount of un-
reacted residual aluminum was included.
Fig. 2 shows the dehydrogenation behavior of KAlH4.
The expected decomposition processes are as in reactions
(5)–(7), in analogy with the decomposition process of
NaAlH4.
A repeated-cycle-test showed that the reaction is essen-
3KAlH4 → K3AlH6 1 2Al 1 3H2 (H2: 2.9 wt.%)
(5)
Fig. 3. The full line indicates the reduction of hydrogen gas pressure in
the sample holder, indicating the hydrogenation from KH and Al to
potassium alanate, and the dashed line is the result of a blank measure-
ment. After dehydrogenation of KAlH4 to KH and Al at 390 8C, the
sample was cooled down to room temperature under 1 bar of H2 gas and
heated at a rate of 0.5 8C/min to 500 8C under 9 bar of hydrogen, and the
pressure changes in the sample holder were monitored.
Fig. 2. TPD measurement of KAlH4 dehydrogenation. The sample was
heated from room temperature to 500 8C at the rate of 2 8C/min. Please
refer to reactions (5)–(7) for the decomposition reaction.

H. Morioka et al. / Journal of Alloys and Compounds 353 (2003) 310–314
313
in H2 pressure from 1 to 10 bar, at the temperature of
decomposition, the reaction can be driven to the reverse
direction for the potassium case. In addition, considering
Figs. 3 and 4, kinetic enhancement is abrupt at a higher
temperature (.300 8C). Such a situation makes KAlH4
active as a high-temperature reversible hydrogen storage
medium without the aid of an external catalyst and
rechargeable at low pressure. Further thermodynamic
investigations based on the data accumulated for the
Pressure–Composition–Temperature (PCT) diagrams will
lead to more quantitative understanding.
4. Conclusion
The reversible hydrogen decomposition of KAlH4 is
demonstrated. The hydrogen ability was above 3.5 wt.%.
Under 10 bar of hydrogen in a temperature range of
250–330 8C, the reversible reaction smoothly proceeds
without any catalyst, which is different from the reactions
of NaAlH4 and LiAlH4, in which homogeneous doping of
a transition metal catalyst is essential for reversibility and
good kinetics. These findings provide new guidelines for a
systematic understanding of the reversible hydrogen de-
composition of the alanate family.
Fig. 4. Repeated-cycle-test of KAlH4. For dehydrogenation, the sample
was heated to 370 8C at a rate of 2 8C/min. For hydrogenation the sample
was hydrogenated under 10 bar of hydrogen while heating the sample at a
constant temperature.
tially reversible but the amount of absorbed hydrogen
decreases as the number of cycle increases, except in the
fifth cycle (Fig. 4 and Table 1). The absorption tempera-
tures for the first through third hydrogenation cycles were
250 8C and for the fourth hydrogenation cycle, 330 8C. The
charging condition was 10 h under 10 bar of hydrogen for
all experiments. The only difference in the fifth dehydroge-
nation cycle was the charging temperature. Therefore, the
hydrogenation was completed at 330 8C but not at 250 8C,
which meant that the capacity decrease up to four cycles
was not critical but was a kinetic problem, e.g. 10 h of
hydrogenation at 250 8C was not long enough for the
hydrogenation to be completed.
Acknowledgements
The authors wish to express their gratitude to Dr M. Ata
and Dr T. Hatazawa for their helpful guidance and
discussions during the course of this research, and also to
Dr B. Pietzak who designed the measurement devices used
in this research.
The formation energy for KH and NaH is similar (256.3
and 256.3 kJ/mol, respectively), while that for KAlH4
and NaAlH4 differs by about 30 kJ/mol (DHf for KAlH4
and NaAlH4 is 2183.7 and 2155.5 kJ/mol, respectively)
[23]. This shows that for the reaction MAlH4 →MH1Al1
3/2 H2, the enthalpy change with M5K would also be
about 30 kJ/mol more endothermic than with M5Na. This
accounts for the higher decomposition temperature of
KAlH4 (300–400 8C) than NaAlH4 under ambient H2
pressure. What is remarkable is that with a slight increase
References
´
[1] B. Bogdanovic, M. Schwickardi, J. Alloys Comp. 253 (1997) 1.
[2] C.M. Jensen, K.J. Gross, Appl. Phys. A 72 (2001) 213.
[3] C.M. Jensen, R. Zidan, N. Mariels, A. Hee, C. Hagen, Int. J.
Hydrogen Energy 24 (1999) 461.
[4] E. Akiba, Curr. Opin. Solid State Mater. Sci. 4 (1999) 267.
[5] R.A. Zidan, S. Tanaka, A.G. Hee, C.M. Jensen, J. Alloys Comp. 285
(1999) 119.
¨
[6] L. Zaluski, A. Zaluska, J.O. Strom-Olsen, J. Alloys Comp. 290
(1999) 71.
[7] J. Huot, S. Boily, V. Guther, R. Schulz, J. Alloys Comp. 383 (1999)
304.
Table 1
A summary of the cycle tests
[8] K.J. Gross, S. Guthrie, S. Takara, G. Thomas, J. Alloys Comp. 297
(2000) 270.
¨
[9] A. Zaluska, L. Zaluki, J.O. Strom-Olsen, J. Alloys Comp. 298
Peak
Dehydrogenation (wt.%)
(2000) 125.
´ ´
[10] B. Bogdanovic, R.A. Brand, A. Marjanovic, M. Schwickardi, J.
1st cycle
2nd cycle
3rd cycle
4th cycle
5th cycle
¨
Tolle, J. Alloys Comp. 302 (2000) 36.
First
Second
2.34
1.22
2.01
1.28
1.77
1.27
1.40
1.23
2.47
1.26
[11] V.B. Balema, V.K. Pecharsky, K.W. Dennis, J. Alloys Comp. 313
(2000) 69.
¨
The amount of hydrogen released for each peak is shown. The first
[12] A. Zaluska, L. Zaluski, J.O. Strom-Olsen, Appl. Phys. A 72 (2001)
peak started at about 300 8C and the second peak started at about 340 8C.
157.

314
H. Morioka et al. / Journal of Alloys and Compounds 353 (2003) 310–314
[13] J. Chen, N. Kuriyama, Q. Xu, H. Takeshita, T. Sakai, J. Phys.
Chem. B 105 (2001) 11214.
[19] D. Sun, T. Kiyobayashi, H.T. Takeshita, N. Kyriyama, C.M. Jensen,
J. Alloys Comp. 337 (2002) L8.
[14] K.J. Gross, G.J. Thomas, C.M. Jensen, J. Alloys Comp. 330 (2002)
683.
[20] G.P. Meisner, G.G. Tibbetts, F.E. Pinkerton, C.H. Olk, M.P. Balogh,
J. Alloys Comp. 337 (2002) 254.
[15] K. Gross, G. Sandrock, G.J. Thomas, J. Alloys Comp. 330 (2002)
691.
[21] G. Sandrock, K. Gross, G. Thomas, J. Alloys Comp. 339 (2002)
299.
[16] G. Sandrock, K. Gross, G. Thomas, C. Jensen, D. Meeker, S.
Takara, J. Alloys Comp. 330 (2002) 696.
[22] T.N. Dymova, N.G. Eliseeva, S.I. Bakum, Y.M. Dergachev, Dokl.
Akad. Nauk SSSR. 215 (1974) 1369.
[17] G.J. Thomas, K.J. Gross, N.Y.C. Yang, C. Jensen, J. Alloys Comp.
330 (2002) 702.
[23] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 80th
edition, CRC Press, Boca Raton, 2000.
[18] M. Fichtner, O. Fuhr, J. Alloys Comp. 345 (2002) 286.