Disproportionation of CaNi3 hydride: Formation of new hydride, CaNiH3

Article abstract of DOI:10.1016/S0925-8388(01)01739-X

The hydrogenation properties of CaNi₃ were investigated at temperatures ranging from 298 to 773 K by differential thermal analysis (DTA). This compound exhibited four exothermic peaks at 373, 498, 543 and 653 K followed by an endothermic peak at 748 K when the compound was heated to 773 K under a hydrogen atmosphere of 3 MPa, and its final products were CaH₂ and Ni. However, the formation of a new ternary hydride was observed in the X-ray diffraction (XRD) profiles below the endothermic reaction temperature at which the final products were formed. The Rietveld refinement of the obtained XRD profiles and a volumetric measurement of the hydrogen content based on the Sieverts' method indicated that this new phase was CaNiH₃ with a cubic CsCl type structure for the metal constituents. The experimental data obtained by XRD and DTA were discussed from the viewpoint of the disproportionation of CaNi₃ in the hydrogen atmosphere.

Full text of DOI:10.1016/S0925-8388(01)01739-X

Journal of Alloys and Compounds 333 (2002) 266–273
L
www.elsevier.com/locate/jallcom
Disproportionation of CaNi3 hydride: formation of new hydride, CaNiH3
b
a
Hiroyuki T. Takeshitaa , , Toshio Oishi , Nobuhiro Kuriyama
*
aOsaka National Research Institute, AIST, METI, 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan
bDepartment of Materials Science and Engineering, Faculty of Engineering, Kansai University, Yamada, Suita, Osaka 564-8680, Japan
Received 16 March 2001; received in revised form 25 June 2001; accepted 25 June 2001
Abstract
The hydrogenation properties of CaNi3 were investigated at temperatures ranging from 298 to 773 K by differential thermal analysis
(DTA). This compound exhibited four exothermic peaks at 373, 498, 543 and 653 K followed by an endothermic peak at 748 K when the
compound was heated to 773 K under a hydrogen atmosphere of 3 MPa, and its final products were CaH2 and Ni. However, the formation
of a new ternary hydride was observed in the X-ray diffraction (XRD) profiles below the endothermic reaction temperature at which the
final products were formed. The Rietveld refinement of the obtained XRD profiles and a volumetric measurement of the hydrogen content
based on the Sieverts’ method indicated that this new phase was CaNiH3 with a cubic CsCl type structure for the metal constituents. The
experimental data obtained by XRD and DTA were discussed from the viewpoint of the disproportionation of CaNi3 in the hydrogen
atmosphere.
2002 Elsevier Science B.V. All rights reserved.
Keywords: Hydrogen storage materials; Powder metallurgy; X-ray diffraction; Phase transitions; Thermal analysis
1. Introduction
known that some hydrides lose their crystallinity to form
amorphous ones by increasing temperature [10–12], which
is one of the phenomena observed for hydrogen-induced
amorphization (HIA) [13,14]. Thus it is very important and
interesting to investigate the hydrogenation properties of
materials at high temperatures. However, no information is
available for the high temperature hydrogenation properties
for CaNi3 and its related alloys, although it has been
reported for its stability for a long time that the hydro-
genated sample exposed to air for 1 year is attracted to a
magnet, which indicates the precipitation of metallic Ni
[1].
The objective of this study was to obtain more extensive
data for the hydrogenation characteristics of the CaNi3
compound at high temperatures up to 773 K (5008C).
Their hydrogenation properties and change in crystal
structure were analyzed by mainly using differential
thermal analysis (DTA) and X-ray diffraction (XRD),
respectively.
The hydrogenation properties of CaNi3 have been
studied by Oesterreicher et al. [1]. The authors reported
that CaNi3 absorbed hydrogen to form its hydride with the
composition of CaNi3H4.57, with retaining the PuNi3 type
structure [2]. In recent years, the hydrogenation properties
obtained by not only the gas–solid reaction but also an
electrochemical reaction have been reported for CaNi3 and
its related alloys [3–6]. Some of us have reported that,
under moderate conditions such as room temperature and
atmospheric pressure, these CaNi3-based alloys reversibly
absorb and desorb a large amount of hydrogen accom-
panied by hydrogen pressure plateaus [3,5]. These CaNi3-
based alloys are attractive as hydrogen storage materials in
the coming hydrogen energy era, because of their good
hydrogenation properties and relatively low material cost.
In practical use, hydrogen absorption and desorption of
these materials are considered to be performed by control-
ling the temperature rather than hydrogen pressure from
the viewpoint of the utilization of waste heat from indus-
try. Some of the alloys lose their reversibility for hydrogen
occlusion by decomposing to form stable elemental hy-
drides at high temperatures [7–9]. Moreover, it is well-
2. Experimental details
The starting materials included granular Ca (99.5%) and
powdered Ni (99.9%). CaNi2 was prepared in advance as a
raw material in order to synthesize the CaNi3 samples. The
Ca granules were cut into fine grains and mixed with Ni
*
Corresponding author. Fax: 181-727-51-9629.
E-mail address: hiroyuki-takeshita@aist.go.jp (H.T. Takeshita).
0925-8388/02/$ – see front matter
PII: S0925-8388(01)01739-X
2002 Elsevier Science B.V. All rights reserved.

H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
267
powder so that the total composition was CaNi2. The
mixtures were then cold-pressed into tablets. The tablets
were put into a Mo crucible and the crucible was placed in
a stainless steel vessel with a thermocouple inside. Rapid
heating of the tablets resulted in the melting of Ca at the
Ca–CaNi2 eutectic temperature (878 K [15]) as found in
the Ca–Ni binary phase diagram (Fig. 1), which caused
damage to the vessel. In order to avoid this problem, the
temperature of the tablets had to be slowly increased at
temperatures ranging from 850 to 900 K. As a result of the
preliminary evaluation for heating, the programmed tem-
perature control shown in Fig. 2a was applied to the heat
treatment of the tablets. The prepared tablets were crushed
into powder to check the homogeneity by XRD analysis.
The cold-press, heat treatment and pulverization were
repeated until samples with the single phase of CaNi2 were
obtained. The CaNi3 samples were prepared following the
same procedure as the CaNi2 ones, except for the starting
materials (powdered CaNi2 and Ni) and the heat treatment
condition (Fig. 2b). All the treatments described above
were conducted under an Ar atmosphere.
The crystal and metallographic structures of the samples
were examined by XRD analysis using Cu Ka radiation
and with a scanning electron microscope equipped with a
wavelength dispersive X-ray spectroscope (SEM-WDS),
respectively. The apparatus used was a Rigaku RU-200 for
the XRD and a Hitachi S-2500CX and Microspec WDX-
3PC for the SEM-WDS. Some of the XRD profiles
obtained with a step size of 0.028 in 2u at 298 K (258C)
were analyzed using the Rietveld refinement program
‘RIETAN97b’. The details for this program are described
in Refs. [16–18].
Fig. 2. Schematic diagrams for programmed temperature control on heat
treatment of (a) CaNi2 and (b) CaNi3 samples.
The measurements of the hydrogenation properties at
high temperature were carefully performed so that the
temperature of the alloy sample was not suddenly elevated
due to its exothermic reaction. Otherwise the direct
decomposition of CaNi3 into CaH2 and Ni occurred
without the formation of an intermediate phase, which was
quite different from the results described later.
The reaction of the samples with hydrogen gas were
examined by DTA. The DTA apparatus used was a Rigaku
DSC8230HP. For the DTA measurements, the samples,
each 50 mg, which were exposed to hydrogen gas of 3
MPa, were heated to 773 K (5008C) at the rate of 20 K per
min, cooled to 373 K at the same rate and finally cooled to
room temperature at a rate depending on the apparatus. A
conventional Sieverts’-type apparatus constructed by
LESCA was used for the volumetric determination of the
hydrogen content in the samples. For these measurements,
the samples, each 0.5 g, had been hydrogenated under the
conditions of 313 K (408C) and 2 MPa for 1 day. The
samples were carefully heated at 623 K (3508C) for several
hours and then cooled to 313 K again. After these
treatments, the hydrogen contents were measured at 313 K.
The experimental error for the obtained hydrogen content
was estimated to be 60.02 H/M (H/M; the atomic ratio of
hydrogen to metal). It was confirmed by XRD analysis that
Fig. 1. Binary phase diagram of Ca–Ni system [15].

268
H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
the sample prepared for the measurement of the hydrogen
content possessed the desired constituent phases.
identified as the PuNi3-type structure and the standard ones
of Si. The lattice constants were calculated to be a055.455
˚
˚
A and c0526.65 A for the PuNi3-type structure. These
values were also in agreement with those reported for
˚
˚
CaNi3H4.57 (a055.444 A, c0526.56 A) [1], and therefore,
the XRD profile in Fig. 4b indicated the existence of a
CaNi3 hydride. The lattice and volume expansion due to
the hydrogenation of CaNi3 was calculated to be 18.4,
19.8 and 128.6% for a0, c0 and V0, respectively. For the
sample heated to 523 K, the diffraction peaks from the
PuNi3-type structure disappeared and a few broad peaks
were observed (Fig. 4c). For the sample heated to 573 K,
several unidentified peaks appeared in addition to the ones
from metallic Ni (Fig. 4d). These peaks became larger and
sharper at 673 K (Fig. 4e). The XRD profile for the sample
heated to 773 K exhibited the existence of CaH2 and Ni
(Fig. 4f). The sample powders heated to temperatures
higher than 523 K were attracted to a magnet, which
indicated the existence of metallic Ni in the sample.
The detailed analysis of the unidentified peaks observed
in Fig. 4d,e indicated the existence of a new phase with the
structure of a simple body-centered cubic (b.c.c.) type. The
coexistence of this new phase and Ni was also observed in
the hydrogenated CaNi2 samples [20], which indicated that
the composition of the new phase could be expressed as
CaNi22x (0,x<2), where x denotes an unknown number
in the present study. One of the local compositions
analyzed with the SEM-WDS apparatus was Ca:Ni51:1.2,
although the others were Ca:Ni51:3. It was expected from
this information that this new phase had a simple b.c.c. or
CsCl type structure for its constituent metal atoms.
3. Results
Fig. 3 shows the DTA curve for the hydrogenation of
the CaNi3 sample. The DTA curve showed five peaks with
increasing temperature from room temperature to 773 K
but no peak during cool down to room temperature except
for the false ones due to the variation in the baseline. The
first peak (i) around 373 K was the largest and sharp
exothermic one. The second one (ii) observed around 498
K was also exothermic and sharp. The exothermic peak
(iii), which was small and broad, was observed around 543
K. Peak (iv), which was exothermic, the smallest and very
broad, was observed around 653 K. The fifth one (v)
around 748 K was endothermic.
In order to estimate the reactions corresponding to these
five peaks, the XRD analysis was performed for the
samples heated to 473, 523, 573, 673 and 773 K under a
hydrogen atmosphere of 3 MPa. The XRD profiles for
these samples are shown with the profile for the sample
before hydrogenation in Fig. 4. For the sample before the
DTA analysis, as shown in Fig. 4a, the diffraction peaks
were identified as a PuNi3-type structure and no peak from
impurity phases was observed, except Si which was mixed
with the sample powder as the internal standard material.
The lattice constants for the PuNi3-type structure were
˚
˚
calculated to be a055.036 A and c0524.32 A. These
values were in agreement with those reported by Buschow
(a055.030 A, c0524.27 A) [19]. For the sample heated to
473 K (Fig. 4b), all the diffraction peaks were also
We performed the Rietveld refinement for the XRD
profiles of the sample obtained at 673 K. The experimental
and calculated profiles are shown with their difference in
˚
˚
Fig. 3. Results obtained by differential thermal analysis ranging from room temperature to 773 K under hydrogen atmosphere of 3 MPa for CaNi3.

H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
269
The hydrogen contents of the samples heated at 473 and
623 K were measured at 313 K with a Sieverts’-type
apparatus. The hydrogen contents at 313 K for these
samples are listed in Table 3. The hydrogen content at 313
K and 2 MPa was 0.7660.02 H/M for the sample heated
at 623 K, which corresponded to CaNi3H3.0 in overall
composition. It was confirmed by XRD analysis that the
constituent phases of this sample were metallic nickel and
the CsCl-type phase which can be expressed as CaNiHx.
Based on the assumption that metallic Ni did not react with
hydrogen under the present conditions, the value of x was
found to be 3.0 from the following relation
CaNiHx 1 2Ni 5 CaNi3H3.0
Therefore, it was found that the new phase was CaNiH3
with a cubic CsCl type structure for the constituent metal
atoms.
The new hydride CaNiH3 has a large hydrogen capacity
for not only the atomic ratio (1.5 H/M) but also for the
weight ratio (3 wt%). Since the intermetallic compound
phase with the composition of Ca:Ni51:1 is not formed as
Table 1
Lattice constant of each constituent phase in the sample heated to 673 K
under hydrogen atmosphere of 3 MPa, obtained by Rietveld refinement of
its X-ray diffraction profile at 298 K
Phase
Space
group
a0
(A)
b0
(A)
c0
(A)
RB
(%)
RF
(%)
˚
˚
˚
]
Pm3m
]
CaNiH3
Ni
CaH2
3.5518(2)
3.5285(3)
5.936(9)
1.78
1.07
1.50
1.00
0.52
0.83
Fm3m
Pnma
3.594(4)
6.770(9)
Estimated standard deviations in parentheses; Rwp 58.10%, Rp 55.97%,
Re 56.37%.
Table 2
Crystallographic parameters for constituent metal atoms of CaNiH3
obtained by Rietveld refinement of X-ray diffraction profile
Atom
Site
x
y
z
Occupancy
Ca
Ni
1a
1b
0
1/2
0
1/2
0
1/2
1.00(1)
0.998(8)
Fig. 4. X-ray diffraction profiles of CaNi3 samples before differential
thermal analysis (a) and after heated to 473 K (b), 523 K (c), 573 K (d),
673 K (e) and 773 K (f) in differential thermal analysis.
Estimated standard deviations in parentheses; Biso values are fixed for
both Ca (0.17 A ) and Ni (1.13 A ).
2
2
˚
˚
Fig. 5. Tables 1 and 2 list the results of the refinement,
where a small amount of CaH2 was included in the sample
supplied to the Rietveld refinement. The results of the
refinement indicated that the CsCl type structure was much
more possible than the b.c.c. one, as could be easily found
from the (100) peak corresponding to the ordered lattice
peak of the CsCl type one, shown by an arrow in Fig. 5.
The 1a and 1b positions were fully occupied by Ca and Ni
atoms, respectively, so the composition of the new phase
for metal elements was Ca:Ni51:1.
Table 3
Hydrogen contents at 313 K and 2 MPa hydrogen pressure for samples
which are heated to 473 and 623 K and their constituent phases after heat
treatment
Constituent
phases before
hydrogenation
Treatment
temperature
(K)
Hydrogen
content
(H/M)
Constituent
phases after
hydrogenation
CaNi3
CaNi3
473
623
1.10(2)
0.76(2)
CaNi3H4.4
CaNiH3, Ni
Experimental errors in parentheses.

270
H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
Fig. 5. Rietveld profile refinement of X-ray diffraction profile measured at 298 K for sample heated to 673 K at 3 MPa; profile observed (top), calculated
CaNiH3, Ni and CaH2 (middle) and difference (bottom). Arrow indicates (100) ordered lattice peak of CaNiH3 with CsCl type structure for constituent
metal atoms.
an equilibrium one as seen in Fig. 1 [15], it can be said
that this new hydride is the phase stabilized by hydrogen
atoms, that is, the hydrogen-stabilized phase.
The compositions, Goldschmidt radii, crystal structures for
metal constituents and lattice constants for both hydrides
are listed in Table 4. The Goldschmidt radius of Ca is
approximately equal to that of Yb and both hydrides have
the same structures for the constituent metal atoms. The
obtained hydrogen content and lattice constant for the
hydride of CaNi are comparable to those for the YbNi
hydride, respectively.
4. Discussion
4.1. Comparison of CaNiH3 with hydrides of MgNi and
Oesterreicher et al. described that the hydrogen uptake
of the compounds in the Ca–Ni system was given by the
weighted hydrogen uptake of the elemental Ca and Ni
hydrides [1], that is, the hydrogen uptake x of CaaNibHx
can be given to be 2a10.7b from the following equation,
YbNi
Orimo et al. have reported that an amorphous MgNi and
its hydride have local structures related to the CsCl type
structure [21,22]. Taking into account the fact that Mg is
an alkaline-earth element as well as Ca, it can be expected
that CaNi and MgNi hydrides have some similarities with
respect to their crystallographic structures. Therefore, the
present result for the CaNi hydride obtained by the
Rietveld refinement supports their results for the amor-
phous MgNi hydride. However, there are differences
between the hydrogenation properties of the CaNi and
MgNi hydrides. The hydrogen content of the CaNi hydride
is 1.5 H/M which is much higher than that of MgNi (about
1 H/M [21,23]). In addition, the CaNi does not show any
reversibility for the reaction with hydrogen gas, although
the MgNi exhibits a reversible hydrogenation and dehydro-
genation.
CaaNibHx 5 a ? CaH2 1 b ? NiH0.7
,
The hydrogen content and lattice constant of CaNiH3
were compared with those for the hydride of YbNi [24].
Table 4
Comparison of structural data for CaNiH3 with that for YbNiH2.7 [24]
Composition
Crystal
Goldschmidt
radius (A)
Lattice
constant (A)
structurea
˚
˚
b
CaNiH3.0
CsCl typeb
CsCl type
1.97 (Ca)
1.94 (Yb)
3.5518(2)b
3.55
YbNiH2.7
Fig. 6. Relation between hydrogen uptake and Ni content of Ca–Ni-
based compounds. Data are taken from Refs. [1,26–28] and present
results.
a For constituent metal atoms.
b Present work.

H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
271
where a and b are constants. Fig. 6 shows the change in
hydrogen uptake of the Ca–Ni compounds with their
compositions for the constituent metal elements. As seen
in Fig. 6, CaNiH3 obeys the rule that the maximum
hydrogen content of each compound lies close to the
straight line connecting the hydrogen contents of the two
elemental hydrides, CaH2 and NiH0.7. Also, this rule has
been reported for some rare earth(R)–transition metal(T)
systems [25].
Thus the hydrogenation and crystallographic properties
of CaNiH3 are similar to some extent to those of both
amorphous MgNi and crystalline RNi. This implies that
the substitution of a variety of elements for Ca can be
applied to the modification of the CaNiH3. In fact, the
effect of the substitution of rare earth elements and Mg for
Ca on the hydrogenation properties has been investigated
for CaNi3 [5]. The effect of the elemental substitution on
the hydrogenation properties of CaNiH3 will be described
in other papers.
amorphization of the crystalline CaNi3 hydride, so-called
hydrogen-induced amorphization (HIA), and the formation
of CaNiH3 and Ni from the amorphous hydride, respec-
tively. In this case, it can be explained that, since the
exothermic heat due to reaction (ii) triggers reaction (iii),
the crystalline phases are observed in Fig. 4c.
Since the PuNi3 type structure of CaNi3 partially
contains the Laves one [29,30], we assume that the
information about interatomic diffusion in the Laves
compounds can be applied to the diffusion in CaNi3. It has
been reported for MgCu2 type RCo2 compounds that the
ratios of their crystalline hydride formation temperatures
(Th), amorphization ones (Ta) and crystallization (forma-
tion of elemental hydrides) ones (Tx) to the corresponding
melting (decomposition) points (Tm), that is Th /Tm, Ta /Tm
and Tx /Tm, are respectively, 0.28, 0.40 and 0.5 [12]. The
growth of crystallites from the mixture of very small ones
of two phases requires an interatomic diffusion over a long
distance and the growth of crystallites occurs at tempera-
tures higher than Tx [12]. Therefore, the temperature ratio
for the growth of crystallite is greater than 0.5. The ratio of
the temperature of peak (iv) (Tiv), corresponding to the
growth of CaNiH3 and Ni, to the peritectic melting point
4.2. Procedure for decomposition of CaNi3 under a
hydrogen atmosphere
9
9
The XRD profiles in Fig. 4a,b show the existence of the
CaNi3 compound and its hydride, respectively, which
indicate that the hydrogenation of the CaNi3 compound is
responsible for the largest exothermic peak (i) observed in
Fig. 3. From the comparison of the XRD profile in Fig. 4d
with the one in Fig. 4e, peak (iv) is attributed to the growth
of CaNiH3 and Ni crystallites. The XRD profiles in Fig.
4e,f show the coexistence of the CaNiH3 and Ni and the
coexistence of CaH2 and Ni, respectively, which indicates
that the decomposition of CaNiH3 into CaH2 and Ni is
responsible for peak (v) observed in Fig. 3. In this
decomposition, the total hydrogen amount in the sample
decreases from CaNi3H3 (CaNiH3 12Ni) to CaNi3H2
(CaH2 13Ni). Therefore, the endothermic heat of peak (v)
is attributed to the hydrogen desorption accompanying the
decomposition.
of CaNi3 (T m) (51287 K [31]), Tiv /T m, is 0.51 which is
in agreement with the reported information. On the other
9
9
hand, the ratios Tiii /T m and Tii /T m are only 0.42 and 0.39,
which are much smaller even compared with Tx /Tm. This
implies that the reaction based on the long-range inter-
atomic diffusion such as the growth of crystallites does not
9
easily occur at Tiii. Moreover, the value of Tii /T m (0.39) is
approximately equal to Ta /Tm at which the interatomic
diffusion could occur over a limited short range [12]. This
can indicate the disproportionation of CaNi3 hydride into
CaNiH3 and Ni that requires higher temperatures than Tii.
Thus the former proposal seems to be inadequate as a
possible explanation for the reactions causing peaks (ii)
and (iii). On the other hand, the amorphization at Tii is
9
possible because the value of the Tii /T m is about 0.4. The
Tiii /T m value (50.42) is intermediate between Ta /Tm and
9
From the comparison of the XRD profile in Fig. 4c with
the one in Fig. 4d, it can be judged that the broad peaks
observed in Fig. 4c correspond to the strong ones such as
those from the (110) and (211) planes for CaNiH3 and the
(111) plane for Ni in Fig. 4d. The existence of Ni in the
sample heated to 523 K is also confirmed by the fact that
the sample powder is attracted to a magnet. Therefore, the
precipitation of these phases is considered to start below
523 K. We have two possible explanations for the reactions
which are responsible for peaks (ii) and (iii). One is that
peaks (ii) and (iii) are attributed to the disproportionation
of the CaNi3 hydride into CaNiH3 and Ni and the growth
of these two phases, respectively. In this case, the crys-
tallites of CaNiH3 and Ni grow during two different steps
because both peaks (iii) and (iv) are responsible for the
growth of these crystallites. Another possible explanation
is that the reactions corresponding to these peaks are the
Tx /Tm. At this temperature, interatomic diffusion is ex-
pected to be still limited, but can occur over a longer range
than that in HIA. The decomposition into CaNiH3 and Ni
requires a longer-range interatomic diffusion than HIA and
is expected to occur by interatomic diffusion in the limited
range compared with that into CaH2 and Ni in which Ca
and Ni are completely separated from each other. Thus the
disproportionation of the amorphous CaNi3 hydride into
CaNiH3 and Ni can be considered to occur at Tiii.
As is well-known, the materials with atomic size ratios
larger than 1.1 form their corresponding amorphous phases
by rapid quenching [32]. Aoki and Masumoto have shown
that all the MgCu2 type Laves RT2 compounds with
Goldschmidt radius ratio, RR /RT, values larger than 1.37
undertake HIA [13]. Thus a large atomic size ratio for the
constituent elements of the material is required for the
formation of its corresponding amorphous phase. Raj and

272
H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
Shashikala have described that the intermetallic com-
pounds showing HIA are either metastable or peritectic in
their corresponding phase diagram [14]. For CaNi3 show-
ing peritectic melting (see Fig. 1), the Goldschmidt radii of
Ca and Ni are 1.97 and 1.25, respectively, and the value of
RCa /RNi 51.58, which is much larger than 1.37. This
extremely large atomic size ratio for the Ca–Ni alloys may
cause the HIA behavior of CaNi3.
Based on the discussion described above, we propose
that the HIA and crystallization to form CaNiH3 and Ni as
the most possible reactions which are responsible for
exothermic peaks (ii) and (iii), respectively. The hydro-
genation of CaNi3 at elevated temperatures can be summa-
rized below, peak (i) at 373 K; hydrogenation to form
crystalline CaNi3Hx peak (ii) at 498 K; transformation of
crystalline to amorphous CaNi3Hx peak (iii) at 543 K;
crystallization of CaNi3Hx to form CaNiH3 and Ni peak
(iv) at 653 K; growth of crystallites of CaNiH3 and Ni
peak (v) at 748 K; disproportionation of CaNiH3 into CaH2
and Ni.
mately equal to the weighted hydrogen uptake of the
elemental hydrides (CaH2 1NiH0.7 5CaNiH2.7), which is
consistent with the suggestion of Oesterreicher et al. [1].
Acknowledgements
Authors would like to express their sincere thanks to Mr.
T. Nakahata for his advice and help in the preparation of
the CaNi3 samples. We wish to also thank Professor K.
Aoki of Kitami Institute of Technology for his useful
advice.
References
[1] H. Oesterreicher, K. Ensslen, A. Kerlin, E. Bucher, Mater. Res. Bull.
15 (1980) 275.
[2] K.H.J. Buschow, J. Less-Common Met. 38 (1974) 95.
[3] K. Kadir, N. Kuriyama, T. Sakai, I. Uehara, L. Eriksson, J. Alloys
Comp. 284 (1999) 145.
[4] K. Kadir, T. Sakai, I. Uehara, J. Alloys Comp. 287 (1999) 264.
[5] J. Chen, H.T. Takeshita, H. Tanaka, N. Kuriyama, T. Sakai, I.
Uehara, M. Haruta, J. Alloys Comp. 302 (2000) 304.
[6] J. Chen, N. Kuriyama, H.T. Takeshita, H. Tanaka, T. Sakai, M.
Haruta, Electrochem. Solid-State Lett. 3 (2000) 249.
[7] K. Yamanaka, H. Saito, M. Someno, J. Chem. Soc. Japan (in
Japanese) 8 (1975) 1267.
As far as we know, there is no report for the HIA
behavior of Ca-based intermetallic compounds and a few
reports for the materials with the PuNi3 type structure
except LaNi3 [25]. Therefore, further examinations are
required to confirm the hydrogenation and decomposition
procedure proposed in the present study.
[8] V.I. Mikheeva, M.E. Kost, A.L. Shilov, Russ. J. Inorg. Chem. 23
(1978) 657.
[9] V.N.Verbetsky, S.N. Klyamkin, A.Yu. Kovriga, A.P. Bespalov, Int. J.
Hydrogen Energy 21 (1996) 997.
5. Conclusion
[10] X.L. Yeh, K. Samwer, W.L. Johnson, Appl. Phys. Lett. 42 (1983)
242.
[11] K. Aoki, X.-G. Li, T. Hirata, E. Matsubara, Y. Waseda, T.
Masumoto, Acta Metall. Mater. 41 (1993) 1523.
[12] K. Mori, K. Aoki, T. Masumoto, Mater. Sci. Eng. A179–A180
(1994) 181.
[13] K. Aoki, T. Masumoto, J. Alloys Comp. 194 (1993) 251.
[14] P. Raj, K. Shashikala, Met. Mater. Processes 11 (1999) 329.
[15] T.B. Massalski, J.L. Murray, L.H. Bennett, H. Baker (Eds.), Binary
Alloy Phase Diagrams, American Society for Metals, Metals Park,
OH, 1986.
Using differential thermal analysis, X-ray diffraction
and a volumetric measurement of the Sieverts method, the
decomposition procedure of CaNi3 under a hydrogen
atmosphere has been investigated in the range between
room temperature and 773 K. The obtained results are
summarized below.
(1) We suggest that CaNi3 reacts with hydrogen and
decomposes in the following five steps;
(i) Hydrogenation of CaNi3 to form its crystalline
hydride.
(ii) Amorphization of crystalline CaNi3 hydride (hydro-
gen-induced amorphization)
(iii) Decomposition of amorphous CaNi3 hydride to
form CaNi hydride and Ni
[16] F. Izumi, in: R.A. Young (Ed.), The Rietveld Method, International
Union of Crystallography, Oxford University Press, Oxford, 1993,
Ch. 13.
[17] F. Izumi, Rigaku-Denki J. (in Japanese) 34 (1996) 18.
[18] F. Izumi, http://www.nirim.go.jp/ |izumi/
[19] K.H.J. Buschow, J. Less-Common Met. 38 (1974) 95.
[20] H.T. Takeshita et al., to be published.
¨
[21] S. Orimo, K. Ikeda, H. Fujii, S. Saruki, T. Fukunaga, A. Zuttel, L.
(iv) Growth of crystallites of CaNi hydride and Ni
(v) Disproportionation of CaNi hydride into CaH2 and
Ni
Schlapbach, Acta Mater. 46 (1998) 4519.
[22] S. Orimo, K. Ikeda, H. Fujii, S. Saruki, T. Fukunaga, Mater. Trans.
JIM 63 (1999) 959.
[23] S. Orimo, H. Fujii, K. Ikeda, Y. Fujikawa, Y. Kitano, J. Alloys
Comp. 253–254 (1997) 94.
[24] K. Ensslen, E. Bucher, H. Oesterreicher, J. Less-Common Met. 92
(1983) 343.
[25] H. Oesterreicher, J. Clinton, H. Bittner, Mater. Res. Bull. 11 (1976)
1241.
[26] G.D. Sandrock, J.J. Murray, M.L. Post, J.B. Taylor, Mater. Res.
Bull. 17 (1982) 887.
(2) The crystal structure and hydrogen absorbing capaci-
ty are determined for a hydrogen-stabilized new ternary
hydride using the Rietveld refinement program and a
Sieverts’-type apparatus. The crystal structure and hydro-
gen capacity are a cubic CsCl type structure with the
˚
lattice constant of 3.5518(2) A and 1.5 H/M (3 wt%),
respectively. The hydrogen content of CaNiH3 is approxi-
[27] C.B. Magee, in: W.M. Mueller, J.P. Blackledge, G.G. Libowitz

H.T. Takeshita et al. / Journal of Alloys and Compounds 333 (2002) 266–273
273
(Eds.), Metal Hydrides, Academic Press, New York, 1968, pp.
165–240.
[30] B.D. Dunlap, P.J. Viccaro, G.K. Shenoy, J. Less-Common Met. 74
(1980) 75.
[28] B. Siegel, G.G. Libowitz, in: W.M. Mueller, J.P. Blackledge, G.G.
Libowitz (Eds.), Metal Hydrides, Academic Press, New York, 1968,
pp. 545–674.
[29] V.A. Yartys’, V.V. Burnasheva, K.N. Semenenko, N.V. Fadeeva, S.P.
Solov’ev, Int. J. Hydrogen Energy 7 (1982) 957.
[31] M. Notin, D. Belbacha, M. Rahmane, J. Hertz, G. Saindrenan, J.L.
Jorda, J. Less-Common Met. 162 (1990) 221.
[32] B.C. Giessen, in: T. Masumoto, K. Suzuki (Eds.), Proc. 4th Int.
Conf. on Rapidly Quenched Metals,Vol. 1, Japan Institute of Metals,
Sendai, 1982.