Cycling durability and degradation behavior of La-Mg-Ni-Co-type metal hydride electrodes

Article abstract of DOI:10.1016/j.jallcom.2004.11.032

The cycling durability and degradation behavior of the La-Mg-Ni-based hydrogen storage alloys La_0.7Mg_0.3Ni _3.4-xCo_xMn_0.1 (x = 0, 0.75, 1.3) during charge/discharge cycling has been systematically studied by XRD, SEM, EIS, XPS and AES measurements. The reasons for the improvement of the cycling stability of the alloy electrodes with increasing Co content have also been analyzed and discussed. The results show that the pulverization of the alloy particles and the oxidation/corrosion of the active components of the alloys during charge/discharge cycling in the alkaline electrolyte are the two main factors responsible for the fast capacity degradation of the La-Mg-Ni-based alloy electrodes, and the capacity degradation mechanism can be decomposed into three consequent stages, i.e., the pulverization and Mg oxidation stage, the Mg and La oxidation stage and the oxidation and passivation stage. With the increase in Co content, the cell volume expansion ratio ΔV/V of the two main phases during hydrogenation/dehydrogenation was obviously decreased, which results in a reduction of the pulverization of the alloy particles and, consequently, in an increase in the charge and discharge efficiency and a decrease in the rate of contact of the fresh alloy surface with alkaline electrolyte and a subsequent lower rate of oxidation/corrosion. It is believed to be the most important reason responsible for the improvement of the cycling stability of the alloy electrodes with increasing Co content.

Full text of DOI:10.1016/j.jallcom.2004.11.032

Journal of Alloys and Compounds 395 (2005) 291–299
Cycling durability and degradation behavior of La–Mg–Ni–Co-type
metal hydride electrodes
Yongfeng Liu, Hongge Pan∗, Yuanjian Yue, Xuefeng Wu, Ni Chen, Yongquan Lei
Department of Materials Science and Engineering, Zhejiang University, Hangzhou 310027, P.R. China
Received 29 September 2004; received in revised form 16 November 2004; accepted 18 November 2004
Available online 7 January 2005
Abstract
The cycling durability and degradation behavior of the La–Mg–Ni-based hydrogen storage alloys La0.7Mg0.3Ni3.4−xCoxMn0.1 (x = 0, 0.75,
1.3) during charge/discharge cycling has been systematically studied by XRD, SEM, EIS, XPS and AES measurements. The reasons for the
improvement of the cycling stability of the alloy electrodes with increasing Co content have also been analyzed and discussed. The results show
that the pulverization of the alloy particles and the oxidation/corrosion of the active components of the alloys during charge/discharge cycling
in the alkaline electrolyte are the two main factors responsible for the fast capacity degradation of the La–Mg–Ni-based alloy electrodes, and
the capacity degradation mechanism can be decomposed into three consequent stages, i.e., the pulverization and Mg oxidation stage, the Mg
and La oxidation stage and the oxidation and passivation stage. With the increase in Co content, the cell volume expansion ratio ꢀV/V of
the two main phases during hydrogenation/dehydrogenation was obviously decreased, which results in a reduction of the pulverization of the
alloy particles and, consequently, in an increase in the charge and discharge efficiency and a decrease in the rate of contact of the fresh alloy
surface with alkaline electrolyte and a subsequent lower rate of oxidation/corrosion. It is believed to be the most important reason responsible
for the improvement of the cycling stability of the alloy electrodes with increasing Co content.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Hydrogen absorbing materials; Rare earth compounds; Metal hydrides; Electrochemical reactions; Degradation mechanisms
1. Introduction
[3–11]. For the gas–solid reaction, Kadir et al. [4] reported
that the hydrogen storage capacity of (Y0.5Ca0.5)(MgCa)Ni9
alloy was 1.98 wt.% at 3.3 MPa and 263 K. Chen et al. [6]
found that the substitutes of LaNi3 and CaNi3 all can ab-
sorb/desorb1.8 wt.%hydrogen. Fortheelectrochemicalreac-
tion, Kohno et al. [7] disclosed that the discharge capacity of
La0.7Mg0.3Ni2.8Co0.5 alloy reached 410 mAh/g. Lei et al. [8]
investigated the electrochemical properties of LaxMg3−xNi9
(x = 1.0–2.0) alloys and found that the La2MgNi9 alloy ex-
hibited the highest discharge capacity of 403 mAh/g. More-
over, our previous studies also revealed that the maximum
Hydrogen storage alloys have been developed and
commercialized as the negative electrode materials of
nickel/metal hydride (Ni/MH) secondary batteries due to
their high energy density, high rate capacity, long cycle
life and better environmental compatibility [1,2]. However,
further research and development are still essential to im-
prove the energy density, activation and cycling life. Re-
cently, R–Mg–Ni-based hydrogen storage alloys (R: rare
earth, Ca or Y) with PuNi3-type structure have been con-
sidered as promising candidates for negative electrode ma-
terials of Ni/MH batteries because of their high reversible
hydrogen storage capacities and fine electrode properties
discharge capacity of the La0.7Mg0.3(Ni0.85Co0.15 3.5 alloy
)
electrode is as high as 396 mAh/g [9], and the partial sub-
stitution of Ni by Mn can improve effectively the overall
properties of La–Mg–Ni-based alloys [10,11]. However, for
the practical application of La–Mg–Ni-based alloys as nega-
tive electrode materials, it is necessary to improve the cyclic
durability in the alkaline electrolyte.
∗
Corresponding author. Tel.: +86 571 8795 2576;
fax: +86 571 8795 1152.
E-mail address: honggepan@zjuem.zju.edu.cn (H. Pan).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.11.032

292
Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
In our previous studies [12], we have also reported
To investigate the surface state of the electrode, XPS mea-
surements were conducted on a PHI-550 type spectrometer
at 10−8 Torr using Al K␣ radiation (1486.6 eV). In order to
study the element distribution on the electrode surface during
cycling, AES depth profiles were measured by using a PHI-
550 type electron spectrometer with an electron beam at 3 kV
and 10 ␮A. The electrode surface was sputtered with Ar+ on
an area of 1.5 mm × 1.5 mm at 4 kV and 15 mA. For inves-
tigating the pulverization of the alloy particles, the average
particle diameter was measured by Malvern particle analyzer
Mastersizer2000.
that the cycling stability of the La0.7Mg0.3Ni3.4−xMn0.1Cox
(x = 0–1.6) was markedly improved by increasing the Co
content. But the role of Co in the alloys has not been
fully understood yet. In this study, to obtain a further de-
tailed information concerning to the mode of action of
Co in the La–Mg–Ni-based alloys, the cycle behavior of
La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloy elec-
trodes was systematically investigated by XRD, SEM, XPS,
AES and EIS measurements, and the reasons for the improve-
ment of cycle stability with the increase in Co content were
discussed.
3. Results and discussion
2. Experimental
Fig. 1 shows discharge capacity versus cycle number
curves of the La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3)
alloy electrodes. In this figure, the preparation of test elec-
trodes and the cycling life measurement were carried out by
the same method, as reported previously [9–12]. It can be
seen that the maximum discharge capacity of the alloy elec-
trodes are 397.5, 403.1 and 386.1 mAh/g for x = 0, 0.75 and
1.3, respectively, which indicates that the discharge capac-
ity was not obviously reduced after partial substitution of Ni
by Co. However, the cycling stability of the alloy electrodes
is markedly improved with the increase in Co content. As
shown in Fig. 1, after 90 charge/discharge cycles, the capac-
ity retentions (C90/Cmax) for these three electrodes are 25.5,
36.9 and 61.3%, respectively.
Fig. 2(a) shows the XRD patterns of the La0.7Mg0.3
Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloys before charging and
after being fully charged. Fig. 2(b) illustrates as an exam-
ple, the XRD and Rietveld analyses pattern of the as-cast
alloy with x = 0.75. It is found that as-cast alloys consist of
a (La,Mg)Ni3 phase with rhombohedral PuNi3-type struc-
La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloys were
prepared by vacuum induction levitation melting on a water-
cooled copper crucible under argon atmosphere. The ingots
were turned over and remelted twice for homogeneity. Part of
the alloys were mechanically crushed and ground intopowder
of 300 mesh size.
The test electrodes were made by cold pressing 400 mg
pure alloy powder into a pellet under 20 MPa pressure. Then,
the pellet was sandwiched within two foamed Ni plates
(55 mm × 20 mm) with a Ni wire soldered on to form a nega-
tive electrode. Electrochemical measurements were carried
out in a standard tri-electrode cell, consisting of a work-
ing electrode (the MH electrode for studying), a sintered
Ni(OH)2/NiOOH counter electrode and a Hg/HgO refer-
ence electrode, and the 6 M KOH electrolyte. Each electrode
was charged at 60 mA/g for 7 h followed by a 10 min rest
and then discharged at 60 mA/g to the cut-off potential of
−0.6 VversustheHg/HgOreferenceelectrode. Electrochem-
ical impedance spectroscopy (EIS) studies were conducted
at 50% depth of discharge (DOD) using a Solartron SI1287
Electrochemical Interface with a 1255B Frequency Response
Analyzer, using the ZPLOT electrochemical impedance soft-
ware. The EIS spectra of the electrode after 5, 10, 25, 50, and
80 cycles were obtained in the frequency range from 10 to
5 mHz with an ac amplitude of 5 mV under the open-circuit
condition.
After a predetermined number of charging/discharging cy-
cles (10, 25, and 80 cycles), the electrode was taken out,
washed with distilled water and dried in vacuum. To investi-
gate the crystal structure, studies of the surface morphology
and concentration profiles of the degraded electrodes after a
definite number of cycles were made. We also performed X-
ray diffraction (XRD), scanning electron microscopy (SEM),
X-ray photoelectron spectroscopy (XPS) and auger electron
spectroscopy (AES) studies on the electrode after the spe-
cific cycling. The crystal structures of the alloys were de-
termined by XRD analyses, which were carried out by an
ARL X-ray diffractometer with Cu K␣ radiation at 45 kV
and 40 mA. The surface morphology of the alloy particle was
examined by SEM performed on a Hitachi S-4700 FESEM.
Fig. 1. Discharge capacity vs. cycle number for the La0.7Mg0.3Ni3.4−x
Mn0.1Cox (x = 0, 0.75, 1.3) alloy electrodes at 303 K.

Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
293
Fig. 2. XRD patterns of the La0.7Mg0.3Ni3.4Mn0.1 hydrogen storage alloy at difference states: (a) before charging and after fully charged, (b) example of the
XRD and Rietveld analyses pattern of the as-cast alloy with x = 0.75.
ture and a LaNi5 phase with hexagonal CaCu5-type structure
with a small amount of an impurity phase consisting of LaNi.
With the increase in Co content, the diffraction peaks shift
to lower angles, implying an increase in the lattice param-
eters. After being fully charged, it is found that the hydride
phases of the (La,Mg)Ni3 and LaNi5 phase retain the original
rhombohedral PuNi3-type structure and hexagonal CaCu5-
type structure, respectively. However, compared with the as-
cast alloys, the diffraction peaks of the hydride phases are
moved noticeably toward smaller angles, suggesting a large
volume expansion of the crystal lattice upon hydrogenation.
At the same time, we noticed that the diffraction peaks of
hydride phases shift toward larger angles with the increase
in Co content. This result indicates that the cell volume ex-
pansion of the hydride phases is reduced as a function of the
increase of Co content. Table 1 lists the change of both the
lattice parameters and unit cell volumes of the (La,Mg)Ni3
phase and LaNi5 phase before and after charging. It can be
seen that, with increasing Co content, the volume expansion
ratio ꢀV/V (%) of the (La,Mg)Ni3–H phase decreases from
21.58% (x = 0) to 16.61% (x = 1.3), and that of the LaNi5–H
phase decreases from 22.36% (x = 0) to 13.81% (x = 1.3). This
result indicates that, with the increase in Co content, the alloy
electrodes undergo a smaller cell volume expansion and con-
traction in each charge/discharge cycle, and a consequently
lower pulverization of the alloy particles on cycling. Thus,
the oxidation/corrosion of the active elements in the alloys
is decreased, and consequently the cycling durability of the
alloy electrodes is improved as shown in Fig. 1. This result is
in good agreement with studies on previous rare earth-based
alloys [1,13,14].
Fig. 3 shows as an example the XRD patterns of the
La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloy electrode before and af-
ter cycling. It can be seen that after charge/discharge cycling,
the PuNi3-type structure and CaCu5-type structure were still
retained. With increasing the number of cycles, the intensity
of diffraction peaks decreases. On the other hand, after 80 cy-
cles, the broadening of the diffraction peaks decreased with
the increase in Co content, indicating that lattice defects and
refinement of the crystal grains caused by charge/discharge
cycling were reduced due to a decrease in the pulverization
of the alloy particles. Fig. 4 shows the average size of the
La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloy particles
before and after cycling. It is found that for each alloy elec-
Table 1
Characteristics of alloy phases in the La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloys and their hydrides
3
˚
˚
˚
Cell volume ꢀV (A ) ꢀV/V (%)
Alloy
Phases
Lattice parameter (A)
Cell volume Hydride phases Lattice parameter (A)
3
3
˚
˚
(A )
(A )
a
b
c
a
c
x = 0.00 (La,Mg)Ni3 5.039
24.232
3.997
4.458
533.07
87.78
183.89
(La,Mg)Ni3–H
LaNi5–H
5.394 25.714
5.389 4.269
648.117
107.409
115.047
19.629
21.58
22.36
LaNi5
LaNi
5.035
3.838 10.744
x = 0.75 (La,Mg)Ni3 5.052
24.262
4.002
4.457
536.98
88.22
183.65
(La,Mg)Ni3–H
LaNi5–H
5.361 25.608
5.382 4.198
637.610
105.339
100.63
17.119
18.73
19.40
LaNi5
LaNi
5.044
3.837 10.740
x = 1.30 (La,Mg)Ni3 5.065
24.287
4.0071
4.455
539.27
88.58
184.01
(La,Mg)Ni3
LaNi5
5.348 25.385
5.338 4.084
628.875
100.819
89.605
12.239
16.61
13.81
LaNi5
LaNi
5.052
3.837 10.761

294
Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
Fig. 3. XRD patterns of the La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloy electrode
before and after cycling.
Fig. 4. The average particle size of the La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0,
0.75, 1.3) alloys after cycling.
trode, the average particle diameter was decreased on cy-
cling due to the pulverization of the alloy particles, espe-
cially for the alloy with x = 0, the average particle diame-
ter was decreased from 27.33 to 8.56 ␮m. However, after
the specific cycles, the average diameter of the alloy par-
ticles was increased markedly with increasing Co content.
For example, after 80 cycles, the average particle diameter
increases from 8.56 ␮m (x = 0) to 21.17 ␮m (x = 1.3). With
the increase in Co content, the cell volume change on hy-
drogenation/dehydrogenation decreases as listed in Table 1.
Consequently, the pulverization of the alloy particles was
suppressed during cycling. This is an important factor for the
improvement of the cycling stability. From Fig. 3, it can be
also seen that the diffraction peaks of Mg(OH)2 and La(OH)3
appear on cycling due to the oxidation/corrosion of Mg and
La on the alloy surface in the alkaline electrolyte.
Fig. 5 shows as an example the SEM micrographs of the
La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0.75) alloy before and af-
ter cycling. With increasing cycle number, the alloy particles
are pulverized and the particle size is decreased due to the
cell volume expansion and contraction during hydrogena-
tion/dehydrogenation. Especially for the alloy without Co,
Fig. 5. SEM micrographs of the La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloys before and after cycling: (a) as-cast, (b) after 10 cycles, (c) after 25 cycles, (d) after 80
cycles.

Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
295
the pulverization of the alloy particles is very serious. How-
ever, with the increase in Co content, the pulverization of the
alloy particles becomes milder during cycling. At the same
time, it can be seen that the oxidization/corrosion of the ac-
tive constitutes increases for each alloy with increasing cycle
number. For instance, the surface of as-cast alloys was very
fresh and hence, easily activated, but, with increasing cycle
number, the active elements of the alloys were gradually ox-
idized, and a white and loose passive layer was formed on
the grey alloy matrix and then grew in thickness. In general,
the pulverization of the alloy is induced by strain based on
the lattice expansion and contraction during the hydrogena-
tion/dehydrogenation process. As a result, the fresh alloy sur-
face is exposed directly to the alkaline electrolyte, and then
the active components were easily oxidized, which leads to
the capacity degradation of the alloy electrodes. Moreover,
the pulverization induces a large number of cracks in the alloy
particles and interstices in the electrode. This increases the
contact resistance and decreases the conductivity between the
alloy particles. We believe that it is another reason for the ca-
pacity degradation of the La–Mg–Ni-based alloy electrodes
on cycling.
However, it should be noticed that the oxidation/corrosion
process of the active elements was obviously decreased with
the increase in Co content. The alloy surfaces were covered
by passive layers after 25, 50 and 80 cycles for x = 0, 0.75
and 1.3, respectively. It can be ascribed to the fact that with
the increase in Co content, the cell volume expansion ratio
of two main phases during hydrogenation/dehydrogenation
was markedly decreased. This implies that the cell vol-
umes undergo a lower expansion and contraction during
charge/discharge cycling, and consequently the pulveriza-
tion of the alloy particles was gradually depressed. This
in turn causes an increase of the charge and discharge ef-
ficiency and a decrease of the rate of contact of the fresh
alloy surface with alkaline electrolyte and consequently a
lower rate of oxidation/corrosion. All these facts make us
believe that the decrease of the pulverization of the alloy par-
ticles during charge/discharge cycle with the increase in Co
content is an important factor responsible for the improve-
ment of the cycling durability of the La–Mg–Ni-based alloy
electrodes.
tinuous pulverization of the alloy particles during cycling.
The later increase of the charge-transfer resistance is sup-
posed to be caused by the surface degradation. With further
cycling, the active components of alloys were gradually ox-
idized due to the contact with the alkaline electrolyte, and
then an oxide/hydroxide passive film was formed and grew
in thickness, which decreased the electro-catalytic activity of
the alloy surface and inhibited the electrochemical hydrogen
reaction on the alloy surface.
Moreover, it should be noticed that for each alloy elec-
trode, the cycle number starting to increase in the radius of
the larger semicircle in the lower frequency region is differ-
ent. The cycle numbers for reaching the maximum value of
the electro-catalytic activity of the alloy surface are 10, 25
and 50 cycles for the alloy electrode with x = 0, 0.75 and 1.3,
respectively. This phenomenon can be ascribed to that the
rate of the alloy particle pulverization becomes slower with
the increase in Co content and, consequently, there is a de-
crease of the oxidation/corrosion of the active components.
Meanwhile, an interesting phenomenon can be found. For the
alloy electrode with x = 0, a new smaller semicircle (semicir-
cle C) in the low frequency region appears when the cycle
number exceeds 50. We ascribe this smaller semicircle to the
corrosion product resistance. According to the XRD analy-
ses and SEM observation, it can be concluded that, during
charge/discharge cycling, a passive oxide/hydroxide layer is
gradually formed. It increases in thickness on the alloy sur-
face due of the intense pulverization of the alloy particles and
the serious oxidation/corrosion of the active alloy materials.
This not only decreases the electro-catalytic activity of the
alloy surface, but also causes a new resistance, namely the
corrosion product resistance. As a hydrogen diffusion barrier
it prevents hydrogen to diffuse from the passive layer into
the fresh surface of the alloy since it cannot absorb/desorb
hydrogen. However, for the alloy electrodes with x = 0.75 and
1.3, the semicircle C did not appear within 80 cycles, which
is attributed to the decrease of the oxidation/corrosion rate
of the alloys caused by the suppression of the alloy particle
pulverization and the formation of a Co-based surface oxide
with the increase in Co content.
Fig. 7 shows as an example the core level spectra of the
La 3d, Mg 2p, Ni 2p, Mn 2p and Co 2p electrons in the
La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloy electrode. All component
elements are able to be detected on the alloy surface. In ad-
dition, it can be found that La, Mg, Ni, Mn and Co on the
alloy surface are all in the oxidized state due to the long ex-
posure to air of the alloy surface. After cycling, except for
the Mg 2p core level spectra, the binding energies for La 3d,
Ni 2p and Mn 2p core level spectra all increase. This result
indicates that the active components on the alloy electrode
surface are oxidized continuously. As to the Mg 2p spec-
tra, it can be seen that before cycling, the binding energy
is larger than that of metallic Mg (49.75 eV), implying that
Mg has been oxidized to form MgO due to its exposure to
air. However, after cycling, the binding energy for Mg 2p is
first decreased and then increased, which can be ascribed to
Fig. 6 shows the electrochemical impedance spectra (EIS)
of the La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloy
electrodes at the 50% DOD at 298 K. It can be seen that each
spectrum consists of several semicircles. According to the
analysis of Kuriyama et al. [15], the larger semicircle (semi-
circle B as shown in the figures) in the lower frequency region
is ascribed to the charge-transfer resistance of the alloy sur-
face. For three studiedalloy electrodes, the radiusof the larger
semicircle in the lower frequency region first decreases and
then increases on cycling, implying that the charge-transfer
resistance decreases first and then increases. The decrease
of the charge-transfer resistance is attributed to the exposure
of the fresh alloy surface and the consequent increase in the
electro-catalytic activity of the alloy surface due to the con-

296
Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
Fig. 6. Electrochemical impedance spectra of the La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) alloy electrodes at the 50% DOD at 298 K: (a) x = 0, (b) x = 0.75,
(c) x = 1.3.
the fact that, during charge/discharge cycling, Mg can eas-
ily be oxidized to form Mg(OH)2 due to the direct contact
with the alkaline electrolyte, and the standard binding energy
of Mg(OH)2 (49.5 eV) is lower than that of MgO (50.8 eV).
The above result is in good agreement with the XRD mea-
surement.
Fig. 8 shows as an example the AES depth profiles of the
La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloy electrodes before and af-
ter cycling. It can be seen that the oxygen content is increased
constantly on the alloy surface on cycling. Although, the oxy-
gen concentration of the as-cast alloy surface is higher, it is
promptly decreased inside the alloy. The appearance of oxy-
gen on the alloy surface is ascribed to the adsorbed oxygen
due to its exposure to air. After cycling, the oxygen content
is strongly increased, especially inside the alloy. Combined
with the XPS study, it is believed that the alloy components
are oxidized first on the surface and then the oxidation grad-
ually penetrates into the inner of the alloy on cycling. Thus,
a passive layer is formed on the alloy surface and grows in
thickness with increasing the number of cycles. Meanwhile,
it is found that there exists a Ni-enriched layer on the surfaces
for all the as-cast alloys.
After 10 cycles, it can be also found that the Mg concentra-
tion is drastically increased and there is a La-enriched region
on the alloy surface layer, which indicates Mg was first ox-
idized and subsequently formed a gel-type Mg(OH)2 layer
during the initial cycling stages. This film is very loose and
offers little protection to the alloy from further corrosion [16].
Thus, La segregates easily to the alloy surface layer. However,
after 80 cycles, the La concentration was obviously increased
and the Mg concentration was decreased in the alloy surface
layer. The increase of La in the surface layer is due to the fact
that the segregated La is oxidized with further cycling, and
the decrease of Mg can be ascribed to the fact that part of
the Mg(OH)2 disperses into the electrolyte due to the loose
contact with alloy.

Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
297
Fig. 7. La 3d, Mg 2p, Ni 2p, Mn 2p and Co 2p XPS core level spectra of the La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloy electrode: (a) La 3d–Ni 2p, (b) Mg 2p, (c) Mn
2p, (d) Co 2p.
In addition, for these three alloy electrodes, it can be
found that the Ni content on the alloy surface is constantly
decreased in the whole cycling process. As Ni is electro-
catalytically very active in the alkaline electrolyte, the de-
crease in Ni content on the alloy surface decreases the elec-
trode kinetics of the hydriding/dehydriding. This is another
important reason for the capacity degradation of the elec-
trode in the present study. However, when we compare the
decline rate of the oxygen content from the surface to the
inner of the alloys after 80 cycles, it is interesting to no-
tice that the decrease in oxygen content becomes stronger
with the increase in Co content, which implies that the depth
of the passive oxide layer on the alloy surface is decreased
and the rate of oxidation of the alloy components is low-
ered due to the appearance of Co in the alloys. This result
makes us believe that the increase in anti-oxidation/corrosion
capability of the alloys with higher Co content is also an
important factor responsible for the improvement of the
cycling stability of the La–Mg–Ni-based hydrogen storage
alloys.
Accordingtoabovediscussion, itcanbeconcludedthatthe
capacity degradation mechanism of La–Mg–Ni-based alloy
electrodes can be classified into three stages, i.e., the pulver-
izationandMgoxidationstage, theMgandLaoxidationstage
and the oxidation and passivation stage. The pulverization of
the alloy particles and the oxidation/corrosion of the active
components of the alloys during cycling in the alkaline elec-
trolyte are two important reasons for the capacity degradation
of the La–Mg–Ni-based alloy electrodes. With the increase
in Co content, the improvement of the cycling stability of
the alloy electrodes can be ascribed to the diminution of the
cell volume expansion ratio ꢀV/V of the two main phases
during hydrogenation/dehydrogenation, which decreases the
pulverization of the alloy particles during charge/discharge

298
Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
Fig. 8. AES depth profiles of the La0.7Mg0.3Ni2.65Mn0.1Co0.75 alloy electrode before and after cycling: (a) as-cast, (b) after 10 cycles, (c) after 80 cycles.
cycling, and inhibits the oxidation/corrosion of the ac-
tive components and increases the charge and discharge
efficiency.
face is in direct contact with the alkaline electrolyte, and
increases the contact resistance between alloy particles. The
oxidation/corrosion of the active components causes a de-
crease of the hydrogen absorbing elements and an increase
of the surface reaction resistance. With the increase in Co
content, the improvement of the cycling stability of the alloy
electrodes can be ascribed to the decrease of the cell volume
expansion ratio ꢀV/V of two main phases during hydrogena-
tion/dehydrogenation, which decreases the pulverization of
the alloy particles during charge/discharge cycling, and in-
hibits the oxidation/corrosion of the active components and
increases the charge and discharge efficiency.
4. Conclusions
In this paper, the cycling capacity degradation be-
havior of the La–Mg–Ni-based hydrogen storage alloys
La0.7Mg0.3Ni3.4−xMn0.1Cox (x = 0, 0.75, 1.3) has been sys-
tematically studied. The results show that the capacity degra-
dation of the La–Mg–Ni-based alloy electrodes can be as-
cribed to two factors, i.e., the pulverization of the alloy
particles and the oxidation/corrosion of the active compo-
nents of the alloys during charge/discharge cycling in alka-
line electrolyte. The pulverization of the alloy particles leads
to large amount of cracks in the alloy particles and inter-
stices in the electrode, which implies that a fresh alloy sur-
Acknowledgement
This work was supported by the National Natural Science
Foundation of China (50131040).

Y. Liu et al. / Journal of Alloys and Compounds 395 (2005) 291–299
299
References
[9] H.G. Pan, Y.F. Liu, M.X. Gao, Y.Q. Lei, Q.D. Wang, J. Electrochem.
Soc. 150 (5) (2003) 565.
[10] Y.F. Liu, H.G. Pan, M.X. Gao, Y.F. Zhu, Y.Q. Lei, Q.D. Wang,
Electrochim. Acta 49 (4) (2004) 545.
[1] J.J.G. Willems, K.H.J. Buschow, J. Less-Common Met. 129 (1987)
13.
[11] Y.F. Liu, H.G. Pan, M.X. Gao, Y.F. Zhu, Y.Q. Lei, Q.D. Wang, Int.
J. Hydrogen Energy 29 (2004) 297.
[2] T. Sakai, H. Miyamura, N. Kuriyama, A. Kato, K. Oguro, H.
Ishikawa, J. Electrochem. Soc. 137 (1990) 795.
[12] Y.F. Liu, H.G. Pan, M.X. Gao, R. Li, Y.Q. Lei, J. Alloys Compd.
376 (2004) 304.
[3] K. Kadir, T. Sakai, I. Uehara, J. Alloys Compd. 257 (1997) 115.
[4] K. Kadir, T. Sakai, I. Uehara, J. Alloys Compd. 287 (1999) 264.
[5] K. Kadir, T. Sakai, I. Uehara, J. Alloys Compd. 302 (2000) 112.
[6] J. Chen, H.T. Takeshita, H. Tanaka, N. Kuriyama, T. Sakai, I. Uehara,
M. Haruta, J. Alloys Compd. 302 (2000) 304.
[7] T. Kohno, H. Yoshida, F. Kawashima, T. Inaba, I. Sakai, M. Ya-
mamoto, M. Kanda, J. Alloys Compd. 311 (2000) L5.
[8] B. Liao, Y.Q. Lei, G.L. Lu, L.X. Chen, H.G. Pan, Q.D. Wang, J.
Alloys Compd. 356–357 (2003) 746.
[13] J.J.G. Willems, Philips J. Res. 39 (Suppl. 1) (1984) 10.
[14] T. Sakai, K. Oguro, H. Miyamura, N. Kuriyama, A. Kato, H.
Ishikawa, J. Less-Common Met. 161 (1990) 193.
[15] N. Kuriyama, T. Sakai, H. Miyamura, I. Uehara, H. Ishikawa, T.
Iwasaki, J. Alloys Compd. 202 (1993) 183.
[16] W.H. Liu, Y.Q. Lei, D.L. Sun, J. Wu, Q.D. Wang, J. Power Sources
58 (1996) 243.