Formation of LiAlyNi1-yO2 solid solutions under high and atmospheric pressure

Article abstract of DOI:10.1016/j.jssc.2006.06.011

Three methods were used for the synthesis of LiAl_yNi_1-yO₂ solid solutions with layered crystal structure: citrate and hydroxide precursor methods at atmospheric pressure and high-pressure synthesis in oxygen-rich atmosphere (3 GPa). Structural characterization of the oxides was performed by powder XRD analysis and electron paramagnetic resonance (EPR) spectroscopy. Irrespective of the different preparation techniques used, it was found that LiAl_yNi_1-yO₂ solid solutions can be formed in the limited concentration range of 0 ≤ y ≤ 0.5 and 0.75 ≤ y ≤ 1.0. The unit cell parameter a decreases linearly with the Al content whereas the unit cell parameter c increases sharper as compared to the linear interpolation of the c parameter calculated for the two end compositions LiNiO₂ and LiAlO₂. In these compositions, aluminum substitutes for Ni in the NiO₂-layer, the mean Al_yNi_1-y-O bond length decreasing. The extent of the trigonal distortion of Al_yNi_1-yO₆ and LiO₆-octahedra varies with the aluminum content and depends on the synthesis procedure used. The LiO₆-octahedra are more flexible to tolerate the increased trigonal distortion as compared to the Al_yNi_1-yO₆-octahedra. High-pressure synthesis favors the formation of oxides with a higher extent of trigonal distortion of both Al_yNi_1-yO₆ and LiO₆-octahedra. From EPR measurements, it was shown that local cationic distribution in LiAl_yNi_1-yO₂ depends on the synthesis temperature. At atmospheric pressure, higher synthesis temperatures promote the reaction of cation mixing between the layers.

Full text of DOI:10.1016/j.jssc.2006.06.011

ARTICLE IN PRESS
Journal of Solid State Chemistry 179 (2006) 3151–3158
www.elsevier.com/locate/jssc
Formation of LiAlyNi1ꢀyO2 solid solutions under high and
atmospheric pressure
Ã
E. Shinova, E. Zhecheva , R. Stoyanova
Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria
Received 15 March 2006; received in revised form 7 June 2006; accepted 10 June 2006
Available online 16 June 2006
Abstract
Three methods were used for the synthesis of LiAlyNi1ꢀyO2 solid solutions with layered crystal structure: citrate and hydroxide
precursor methods at atmospheric pressure and high-pressure synthesis in oxygen-rich atmosphere (3 GPa). Structural characterization
of the oxides was performed by powder XRD analysis and electron paramagnetic resonance (EPR) spectroscopy. Irrespective of the
different preparation techniques used, it was found that LiAlyNi1ꢀyO2 solid solutions can be formed in the limited concentration range of
0pyp0:5 and 0:75pyp1:0. The unit cell parameter a decreases linearly with the Al content whereas the unit cell parameter c increases
sharper as compared to the linear interpolation of the c parameter calculated for the two end compositions LiNiO2 and LiAlO2. In these
compositions, aluminum substitutes for Ni in the NiO2-layer, the mean AlyNi1ꢀy–O bond length decreasing. The extent of the trigonal
distortion of AlyNi1ꢀyO6 and LiO6-octahedra varies with the aluminum content and depends on the synthesis procedure used. The LiO6-
octahedra are more flexible to tolerate the increased trigonal distortion as compared to the AlyNi1ꢀyO6-octahedra. High-pressure
synthesis favors the formation of oxides with a higher extent of trigonal distortion of both AlyNi1ꢀyO6 and LiO6-octahedra. From EPR
measurements, it was shown that local cationic distribution in LiAlyNi1ꢀyO2 depends on the synthesis temperature. At atmospheric
pressure, higher synthesis temperatures promote the reaction of cation mixing between the layers.
r 2006 Elsevier Inc. All rights reserved.
Keywords: High-pressure synthesis; Lithium transition metal oxides; Cationic distribution; Cathode minerals for lithium ion batteries
1. Introduction
stoichiometric lithium nickelates, Li1ꢀxNi1+xO2 are usual-
ly obtained, where part of the Ni2+ ions occupy the Li-sites
[17,18]. It has been found that the synthesis conditions have
a strong effect on the composition and crystal structure of
lithium nickelates [17,18]. To obtain nearly stoichiometric
oxides, the temperature of 700 1C and oxygen atmosphere
are needed. LiAlO2 possesses three structural modifications
[19,20]: a-LiAlO2 is isostructural to layered LiNiO2 and is
converted, during heating above 600 1C, into g-LiAlO2,
where Li and Al occupy tetrahedral sites only. The layered
modification of of LiAlO2 (a-form) has been obtained at
intermediate pressure (3.5 GPa) [19,20] or by solid-state
reaction at atmospheric pressure and 600 1C between
boehmite AlOOH and Li2CO3 [22]. At pressure above
9 GPa, a third structural modification of LiAlO2 with
tetragonal structure and octahedral coordination of Al and
Li has been isolated [21].
Isovalent substitution of electrochemically inactive ions
for nickel ions has been reported to affect the electro-
chemical properties of lithium nickelates used as cathode
materials for lithium ion batteries [1,2]. Among electro-
chemically inactive ions, Al has been preferred due to lower
cost and environmental benignity [3–12]. In addition, it has
been found that Al substitution leads to an increase in the
electrode potential where the reversible Li intercalation
takes place [14,15]. Al-substituted oxides have been
reported to display better thermal stability in delithiated
state as compared to pure LiNiO2 [8–13].
The crystal structure of LiNiO2 is composed of LiO2-
and NiO2-layers comprising edge sharing LiO6 and NiO6
octahedra [16]. Due to the thermal instability of Ni3+, non-
Ã
Contrary to LiNiO2, Al ions in the LiO2-layers have not
been detected in a-LiAlO2 irrespective of the synthesis
Corresponding author. Fax: +359 2 870 50 24.
E-mail address: zhecheva@svr.igic.bas.bg (E. Zhecheva).
0022-4596/$ - see front matter r 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2006.06.011

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3152
method. Layered LiAlO2 and LiNiO2 have been shown to
react, forming LiAlyNi1ꢀyO2 solid solutions in the limited
concentration range 0pyp0:5, where Al substitutes for Ni
in AlyNi1ꢀyO2-layers [3–12]. To our knowledge, no data on
the formation of Al-rich compositions are reported in the
literature. In addition, the energy calculations of the phase
diagram of LiNiO2–LiAlO2 show miscibility gaps at low
temperature, but LiNiO2 forms solid solutions with LiAlO2
at room temperature [23]. To understand the electroche-
mical behavior of substituted oxides in an extended
concentration range, there is a need for detailed knowledge
of their crystal chemistry.
The aim of this paper is a detailed study on the
formation of LiAlyNi1ꢀyO2 solid solutions in the whole
concentration range. For the preparation of LiAlyNi1ꢀyO2
we have applied synthesis methods under atmospheric and
high pressures. At atmospheric pressure, the citrate and
hydroxide precursor methods were used. High-pressure
syntheses (3 GPa) were performed in oxygen-rich atmo-
sphere using one piston cylinder apparatus with the
intention to stabilize both Al3+ and Ni3+ ions in
octahedral coordination. Powder XRD analysis was used
for structural characterization of LiAlyNi1ꢀyO2. The
oxidation state of Ni ions and the local cationic distribu-
tion in LiAlyNi1ꢀyO2 were analyzed by electron paramag-
netic resonance (EPR) spectroscopy.
700 1C for 5 h in one piston cylinder type apparatus. When
the reaction took place in a Pt-capsule, an unusual splitting
of the (003) diffraction line of the layered structure was
observed for Al-rich compositions only. In the case of pure
LiAlO2 Chang and Margrave [24] have explained the
splitting of this diffraction line by the formation under the
high-pressure of the fourth modification of LiAlO2 with
monoclinic structure. For the samples prepared by us, the
EDAX and XRD analysis shows clearly that this splitting
is due to the appearance of Li2PtO3 as an impurity phase in
Al-rich compositions prepared in the Pt-capsule. To avoid
the formation of the impurity Li2PtO3 phase, the reaction
mixture of Li2O2 with NiO–Al2O3 was sealed in an Au-
capsule.
High-pressure synthesis was carried out in a 1/2 inch
end-loaded piston-cylinder apparatus at the Bayerisches
Geoinstitut. Experiments were performed using the ‘‘hot-
piston in’’ technique. Samples were encapsulated in 1 cm
long, 5 mm diameter welded Au capsules. Details of the
high-pressure experiments are given elsewhere [25,26].
The lithium content of the samples, the mean oxidation
state of nickel and the total nickel and aluminum content
were determined by atomic absorption analysis, iodometric
titration and complexometric titration, respectively.
X-ray phase analysis was carried out on a Philips X’Pert
powder diffractometer with CoKa-radiation, reflection
mode, and with a Si internal standard. The scan range
15p2Yp120 in a step increment of 0.021 was utilized. The
Fullprof computer program was used for the calculation
[27]. The diffractometer point zero, Lorentzian/Gaussian
fraction of the pseudo-Voigt peak function, scale factor,
lattice constants (a and c), oxygen parameter (z), thermal
factors for 3a, 3b and 6c positions, half-width parameters,
preferred orientation were refined. To gain stability during
the refinement, the Al/Ni-ratio was imposed by the
chemical composition of the oxides. Subsequently, the
cationic occupancy factors were refined taking into account
that the total occupancies of the 3a and 3b sites were equal
to unity.
2. Experimental
LiNi1ꢀyAlyO2 solid solutions were prepared by the
citrate precursor method. Lithium–nickel–aluminum citric
acid compositions were obtained by dissolution of
LiOH ꢁ H2O, Ni(OH)2, and Al(NO3)3 ꢁ 9H2O in aqueous
solutions of citric acid (0.1 M). The ratio between the
components was Li:(Ni+Al):Cit ¼ 1.05:1:1. Heating at
80 1C yielded a transparent solution which was concen-
trated to 0.5 M Li. After complexation, the solution
obtained was cooled down to room temperature, then
frozen instantly with liquid nitrogen and dried in vacuum
(20–30 mbar) at ꢀ20 1C with an Alpha-Crist Freeze-Dryer.
After drying, the solid residues were decomposed at 500 1C
with a heating rate of 11/min and were further heated at
700, 750 and 800 1C for 20–120 h in oxygen.
Al-rich compositions were prepared following the
procedure for the preparation of layered LiAlO2 at
600 1C [22]. Mixed Al and Ni hydroxides were synthesized
by co-precipitation of Ni(NO3)2 ꢁ 6H2O and Al(NO3)3 ꢁ
9H2O by ammonia. The solid residue was dried under
vacuum at 100 1C. The co-precipitated Al and Ni hydro-
xides were mixed thoroughly with Li2CO3. The mixture
obtained was heated at 600 1C for 25 h in air.
EPR measurements at 9.23 GHz (X-band) were carried
out in an ERS 220/Q spectrometer within the temperature
range 85–410 K. The high-frequency EPR spectra were
recorded at a single-pass transmission EPR spectrometer
built in the High-Magnetic Filed Laboratory, Grenoble,
France. The frequencies were changed from 95 to 345 GHz
using Gunn diodes and their multipliers. The detection of
absorption was performed with an bolometer. The record-
ing temperatures were varied between 5 and 100 K using a
variable temperature insert (Oxford Instruments).
3. Results
For the preparation of LiAlyNi1ꢀyO2 under high
pressure we adopted the synthesis procedure described in
[24] for the preparation of layered LiAlO2. The mixture of
Li2O2 with NiO (obtained by thermal decomposition of
nickel hydroxide at 550 1C) and Al2O3 was sealed in an Au
or a Pt-capsule. The solid-state reaction proceeded at
3.1. Crystal structure of LiAlyNi1ꢀyO2 solid solutions
The solubility of Al into layered LiNiO2 depends on the
kind of synthesis procedure used. The citrate precursor
method allows preparing LiAlyNi1ꢀyO2 solid solutions

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3153
where Al content varies between 0% and 50%. Above
y40:5, a mixture between g-LiAlO2 and an Al-poor
composition (y ꢂ 0:5) is obtained when the citrate pre-
cursor method is used. Aluminum-rich compositions with
0:75pyp1:0 are successfully prepared under atmospheric
pressure using the hydroxide precursors. High-pressure
synthesis yields two type of compositions: LiAlyNi1ꢀyO2
with 0pyo0:5 and LiAlyNi1ꢀyO2 with 0:8oyp1:0
(Fig. 1). Any attempts to prepare the intermediate
compositions with 0:5oyo0:75 were unsuccessful
(Fig. 1). High-pressure synthesis of LiAlyNi1ꢀyO2 with
0:5oyo0:75 led to the phase separation into Al-poor and
Al-rich compositions (yo0:1 and 40.85, Fig. 1), while
mixtures of g-LiAlO2 and Al-poor compositions (y ꢂ 0:5)
were obtained for this composition range under atmo-
spheric pressure.
The XRD patterns of all the samples obtained are
indexed in the R-3m space group: Li and Ni occupy 3b and
3a positions and O is in the 6c position (Fig. 1). As in the
case of pure lithium nickelates, an improved fitting of the
XRD patterns was achieved when a small fraction of Ni
resides on the Li-site. The oxides prepared under atmo-
spheric pressure display a higher amount of Ni in the Li-
site as compared to oxides prepared under a high-pressure
(Table 1). The amount of Ni in the Li-site is sensitive
towards the Al-to-Ni ratio. Al-poor oxides display a higher
Fig. 1. XRD patterns and fitting by Rietveld procedure of LiAlyNi1ꢀyO2 synthesized at high oxygen pressure and 700 1C. The difference between the
observed and calculated profiles is plotted. Bragg reflections for layered LiAlyNi1ꢀyO2 and Si standard are indicated. The arrow indicates the second phase
with layered structure. Selective broadening of the (1 1 0) diffraction line as compared to (018) diffraction line is shown for LiNi0.5Al0.5O2 obtained under
high pressure and from citrate precursors.
Table 1
Structural parameters (unit cell parameters a and c, and nickel occupancy in the lithium layers, d) determined from Rietveld refinement of the XRD
patterns and the slope of the temperature dependence of the EPR line width dHpp/dT for Li1ꢀdNid[AlyNi1ꢀy]O2 oxides obtained from citrate (citrate) and
hydroxide precursors (OH) at atmospheric pressure and under high pressure in an oxygen atmosphere (H-P)
˚
a (A)
˚
c (A)
Y
Method of preparation
d (70.004)
dHpp/dT (mT/K)
0.0
H-P
H-P
2.8769
2.8742
2.8720
2.8454
2.8084
2.8009
2.8772
2.8685
2.8372
2.8158
2.8082
2.8053
14.1827
14.1880
14.2007
14.2516
14.2344
14.2253
14.1883
14.1997
14.2535
14.2620
14.2332
14.2108
0.014
0.006
0.009
0.0
0.421
0.368
0.360
—
0.05
0.10
0.42
0.90
1.0
H-P
H-P
H-P
H-P
0.0
0.0
—
—
0.0
Citrate
Citrate
Citrate
OH
0.054
0.035
0.009
0.014
0.011
0.0
0.37
0.18
—
0.12
0.50
0.76
0.94
1.0
—
—
OH
OH
—

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3154
Ni content in the LiO2-layers, while Al-rich oxides
obtained under a high pressure are characterized by Ni-
free LiO2-layers (Table 1). It is worth mentioning that
LiAlyNi1ꢀyO2 oxides with Ni amount in the Li-site less
than 4% were obtained in the whole Al concentration
range. Another feature of the XRD patterns of Al-poor
oxides is the selective broadening of the (110) diffraction
line as compared to that of the (108) diffraction line,
especially in the case of LiAlyNi1ꢀyO2 with y ¼ 0:50 (Figs.
1 and 2). The selective broadening is observed with both
types of oxides prepared under high- and atmospheric
pressure (Fig. 1). This has already been reported for
LiAlyNi1ꢀyO2 prepared under atmospheric pressure and
has been explained by short-range inhomogeneity, such as
a tendency to aluminum clustering [3,9]. In addition, at
atmospheric pressure, the selective broadening of the (110)
diffraction line slightly decreased when the oxides are
prepared at higher temperatures (Fig. 2). This observation
means that the preparation temperatures influence the
short-range Al inhomogeneity. It is worth mentioning that
short-range Al inhomogeneity does not appear for Al-rich
compositions with yX0:75 (Fig. 2).
4%). In addition, it is important that the synthesis method
does not affect the unit cell parameter a. In contrast, the
unit cell parameter c increases sharper with the Al content
compared to a linear interpolation of the c parameter
calculated from the two end members LiNiO2 and LiAlO2.
For Al-rich oxides, the unit cell parameter c is different
with oxides prepared under atmospheric or high pressures.
The observed variation in the unit cell parameters a and c
reveals the structural anisotropy of the layered structure.
For the trigonal structure, the a-parameter gives the
distance between the metal ions in the layers and can be
The effect of the Al content on the unit cell dimensions
of LiAlyNi1ꢀyO2 is shown on Fig. 3. The unit cell
parameter a decreases linearly with the Al substitution. It
is noticeable that the unit cell dimensions for solid
solutions with layered crystal structures have usually been
found to display the Vegard’s like behavior, with the
exception of compositions with a partial occupancy of
metal ions in the LiO2-layers [28–31]. In this context, the a-
parameter was shown in Fig. 3 for oxides with the
minimum amount of Ni ions in the Li-site (less than
Fig. 3. Unit cell parameters (a and c) vs. Al content for LiAlyNi1ꢀyO2
solid solutions synthesized under high oxygen pressure, as well as at
atmospheric pressure from citrate and hydroxide precursors.
Fig. 2. The (018) and (110) diffraction lines for LiAlyNi1ꢀyO2 prepared at atmospheric pressure (left): y ¼ 0:10, 0.25, 0.50 prepared at 700 1C from citrate
precursor and y ¼ 0:75 prepared at 600 1C from hydroxide precursor. The effect of the synthesis temperature on the (018) and (110) diffraction lines for
LiAl0.5Ni0.5O2 is shown (right). The difference in the line width of (018) and (110) diffraction lines for LiAl0.5Ni0.5O2 is also shown (DY20 1 8 ꢀ DY12 1 0
,
right).

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3155
expressed by
Ni1ꢀyAly–O, the mean Li–O bond length is more sensitive
towards the synthesis procedure used. This fact can be
related to the strength of the trigonal distortion of the unit
cell.
Fig. 4 also gives the deviation of the experimentally
determined extent of the trigonal distortion from that of
undistorted octahedron as a function of the Al-content:
1=2
a ¼ 21=2rM2Oð1 ꢀ cos aM2O2M
Þ
Þ
1=2
¼ 21=2rLi2Oð1 ꢀ cos aLi2O2Li
,
where rM–O and rLi–O denote the mean AlyNi1ꢀy–O and
Li–O bond lengths, aM–O–M and aLi–O–Li are the angles
between two neighboring ions via oxygen. The structural
anisotropy of the layered structure can be expressed by the
(jSAl Ni
=a ꢀ 0:816j and jSLiO =a ꢀ 0:816j). It is clear
O2
y
ð1ꢀyÞ
2
that high-pressure synthesis favors the formation of oxides
characterized by a higher extent of trigonal distortion of
both AlyNi1ꢀyO6 and LiO6-octaheda. It is noticeable that
the LiO6-octahedra are more flexible to tolerate the
increased trigonal distortion as compared to the Aly
Ni1ꢀyO6-octahedra. The increased trigonal distortion of
the LiO6-octahedra together with the increased thickness of
the LiO2-layers would have a positive effect on Li
diffusion. This feature would be favorable for the electro-
chemical properties of LiAlyNi1ꢀyO2 when used as cathode
materials in lithium ion batteries. However, using mole-
cular dynamics simulations, it has been demonstrated that
the extension of the LiO2-layer leads to destabilization of
the trigonal structure [32]. Moreover, according to Buta
et al. [23] the large difference in the ionic dimensions of
transition metal ions and Al3+ causes strains in the crystal
structure, resulting in a greater tendency for phase
separation. This agrees well with the observed lack of
LiAlyNi1ꢀyO2 solid solutions in the range 0:5oyo0:75.
thickness of the AlyNi1ꢀyO2 and LiO2-layers, SAl Ni
O2
ð1ꢀyÞ
and SLiO (i.e. the distance separating two opposite faces of
y
2
the octahedra along [001]), which are related to the c-
parameter as follows:
c ¼ SAl Ni
=ð2=3 ꢀ 2zÞÞ ¼ SLiO =ð2z ꢀ 1=3Þ,
O2
y
ð1ꢀyÞ
2
where z is the oxygen parameter. Furthermore, the extent
of the trigonal distortion of the AlyNi1ꢀyO6 and the LiO6
octahedra in the AlyNi1ꢀyO2 and LiO2-layers can be
evaluated by the ratio between the layer thickness,
SAl Ni
undistorted octahedron this value is 0.816.
and SLiO , and the unit cell parameter a. For
2
O2
y
ð1ꢀyÞ
Fig. 4 gives the variation in the mean Ni1ꢀyAly–O and
Li–O bond lengths. The mean Ni1ꢀyAly–O bond length
decreases with an increase in the Al-content. This trend is
consistent with the lower ionic radius of Al3+ as compared
to Ni3+ ions. In the same sequence the mean Li–O bond
length decreases slightly (Fig. 4). In addition, the mean
Ni1ꢀyAly–O and Li–O bond lengths are sensitive towards
the synthesis procedure used. When oxides are prepared
under a high pressure, the mean Ni1ꢀyAly–O bond length
decreases more rapidly as compared with the mean value of
Ni–O and Al–O bond lengths determined for two end
compositions LiNiO2 and LiAlO2. In comparison with
3.2. EPR spectroscopy of LiAlyNi1ꢀyO2 solid solutions
Nearly stoichiometric LiNiO2 has been shown to exhibit
an EPR spectrum consisting of a single Lorentzian line
with g ¼ 2:137 due to low spin Ni3+ ions [34]. The line
shape and line width depend on the registration tempera-
ture, Jahn–Teller effect and ferromagnetic interactions
between Ni3+ ions in the layers [35]. In the EPR spectra of
Al-substituted oxides studied by us, a single Lorentzian line
is still visible with the exception of LiAlO2 doped with 1%
Ni prepared by the hydroxide precursor methods. The
substitution of Al for Ni affects both the g-factor and the
line width. The g-factor decreases after replacement of Ni
by Al, indicating an increased covalency of the Ni–O bond.
The same effect has been found for Al-substituted lithium
nickelates with 0pyp0:25, as well as for layered
(1ꢀa)LiNi1ꢀyAlyO2 ꢁ aLi[Li1/3Ni2/3]O2 oxides obtained un-
der high-pressure [6,26]. Fig. 5 shows the variation of
the EPR line width (determined at 103 K) with Al
substitution. The EPR line width increases with the Al-
content, reaching a maximum at 0:25oyo0:5, and
decreases for Al-rich oxides. The observed changes in the
EPR line width can be related to magnetic dilution of
Ni3+-spin systems caused by Al3+ ions. The replacement
of paramagnetic Ni3+ by diamagnetic Al3+ ions will
suppress the development of exchange interactions between
Ni3+ ions leading to the increase in the EPR line width. In
the same sequence, the dipole–dipole interactions between
Ni3+ ions will be diminished resulting in a signal
Fig. 4. The mean AlyNi1ꢀy–O and Li–O bond lengths (upper) and the
deviation of the experimentally determined extent of the trigonal
distortion of AlyNi1ꢀyO6- and LiO6-octahedra from that of undistorted
octahedron as a function of the Al-content for LiAlyNi1ꢀyO2 solid
solutions synthesized under high oxygen pressure, as well as at atmo-
spheric pressure from citrate and hydroxide precursors.

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3156
narrowing. The competition between exchange and dipo-
le–dipole interactions determines the dependence of the
EPR line width on the Al-to-Ni ratio (Fig. 5).
In addition, for pure LiNiO2 the value of dDHpp/dT has
been found to depend on the preparation temperature of
pure LiNiO2 [33]: the extrapolated values for samples
obtained at 700 and 800 1C are 0.26 and 0.46 mT/K,
respectively. Based on this dependence, it has been inferred
that higher synthesis temperatures (800 1C) favor reactions
of intrinsic disorder, i.e. the cationic mixing between the
LiO2- and NiO2-layers [33]. Owing to the small scattering
factor for Li, the XRD method does not permit detection
of this type of reactions. In comparison with LiNiO2 oxides
prepared at atmospheric pressure, oxides prepared under a
high pressure display a higher value of the dDHpp/dT slope:
0.492 and 0.421 mT/K for Li1ꢀdNi1+dO2 with d ¼ 0:055
and 0.014, respectively [25]. This observation is in
agreement with the effect of high-pressure synthesis on
the cationic distribution in LiNiO2. It has been found that
the high pressure facilitates the insertion of Li into NiO2-
layers, Li[LixNi1ꢀx]O2 compositions being formed [25]. At
atmospheric pressure, the formation of Li[Li1/3Ni2/3]O2 has
not been reported. Returning to our samples, this
dependence of dDHpp/dT on the synthesis temperature
was also obeyed for oxides containing 10% Al. The
[Li1ꢀdNid][Al0.1Ni0.9]O2 oxides prepared at 700 1C display
a lower value of dDHpp/dT as compared to that for oxides
prepared at 800 1C, while the value of dDHpp/dT for oxides
prepared at 750 1C falls in the intermediate range: dDHpp/
dT is 0.18, 0.21 and 0.34 mT/K for samples obtained at
700, 750 and 800 1C and characterized with d ¼ 0:035,
0.043 and 0.049, respectively. In addition, high-pressure
synthesis affects also dDHpp/dT for [Li1ꢀdNid][Al0.1Ni0.9]O2
(Table 1). While the high-pressure synthesis allows
preparation of oxides with minimum amount of Ni3+ in
the Li-sites (determined by the XRD analysis), the value of
dDHpp/dT is higher: 0.46 and 0.36 mT/K for samples with
d ¼ 0:038 and 0.009 and prepared under high-pressure. As
in the case of pure LiNiO2, this observation can be related
to the favorable effect of the high pressure on the insertion
of lithium ions into the AlyNi1ꢀyO2-layers.
The temperature variation of the EPR line width, DHpp,
is also affected by the Al substitution (Fig. 6). For oxides
with 0pyp0:25, there is a linear decrease in the EPR line
width on cooling, while for oxides with yX0:5 the EPR line
width slightly depends on the registration temperature. For
pure LiNiO2, the dDHpp/dT slope has been shown to
depend on the strength of the 1801- and 901-Ni3+/2+
–O–Ni3+/2+ exchange interactions, the metal coordination
number of the exchange-coupled particles and on the
distance between them [34,35]. The slope of the linear
dependence, dDHpp/dT, decreases with the reduction of the
amount of Ni2+ in the lithium layers [35]. Further decrease
in dDHpp/dT is observed when Al substitutes for Ni,
Li1ꢀdNid[AlyNi1ꢀy]O2 (Fig. 6). This result also demon-
strates the formation of solid solutions between LiAlO2
and LiNiO2.
Fig. 5. EPR line width, DHpp vs. Al content for LiAlyNi1ꢀyO2 solid
solutions synthesized under high oxygen pressure, as well as at atmo-
spheric pressure from citrate and hydroxide precursors.
For the oxides with yX0:5, the slight dependence of the
EPR line width on the registration temperature indicates
Fig. 6. Temperature variation of the EPR line width (left) for LiAlyNi1ꢀyO2 obtained from citrate precursors at 700 1C with y ¼ 0:0, 0.10, 0.25 and 0.50.
Temperature variation of the EPR line width (right) for LiAl0.5Ni0.5O2 obtained from citrate precursors at 700 1C for 80 and 120 h and 750 1C for 40 and
80 h. The heating time is selected in order to compare LiAl0.5Ni0.5O2 samples with close values of the Ni amount in the Li-site (less than 4%).

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3157
that magnetic dipole–dipole interactions dominate over the
exchange interactions (Fig. 6). This is consistent with the
dependence of the EPR line width on the Al-content
(Fig. 5). In addition, the EPR line width for LiAl0.5Ni0.5O2
obtained from citrate precursor is higher for oxides
obtained at higher temperatures (Fig. 6). As in the case
of Al-poor compositions, one can suggest that high-
temperatures give rise to the reaction of cationic mixing
between the layers.
For the oxide with y ¼ 0:75 prepared from hydroxide
precursor at 600 1C, two low-intensity signals are super-
imposed on the main signal (Fig. 7). The two signals have
the Lorentzian line shape and a line width of 55.6 and
14.4 mT, which are much lower than that of the main EPR
signal (196.1 mT). These two signals have different
dependence on the registration temperature (Fig. 7). At
103 K, the intensities of both signals are much lower than
that of the main signal: 1% and 4% versus 95%,
respectively. The narrow signal exhibits EPR parameters
close to that of Ni-doped LiAlO2 (see Fig. 8). According to
the EPR study of Ni3+ in LiNixCo1ꢀxO2 solid solutions
[36], these three EPR signals can be assigned to Ni3+ ions
having different paramagnetic/diamagnetic metal ions
environments. Having in mind that in the layered structure,
each metal ion has six metal neighbors in the MO2-layers,
we can assume that the main EPR signal corresponds to
Ni3+ ions in a mixed Ni3+/Al3+ environment with a
statistical distribution, while the additional narrow signals
are due to Ni3+ in environment of Al3+ mostly. For a
statistical distribution of Ni and Al in the AlyNi1ꢀyO2-
layers, one can expect to register only one broad EPR
signal due to the convolution of the signals from different
Ni3+ ions located in a mixed Ni0.25Al0.75-environment. The
observation of two additional EPR signals in LiAl0.75
Ni0.25O2 does not contradict with the suggestion for a
statistical Ni/Al distribution since their intensities are much
lower as compared to that of the main signal. This means
that the additional low-intensity signals come, most
probably, from Ni3+ in defect crystal sites.
For LiAl0.9Ni0.1O2- and Ni-doped LiAlO2, the high-
temperature EPR spectra exhibit a single Lotrentzian line
(Fig. 8). On cooling, the symmetrical signal is transformed
into a signal with tetragonal symmetry. This behavior is
typical of the static Jahn–Teller effect. In order to improve
the resolution of the EPR spectrum, EPR at high frequency
and high magnetic field was undertaken. Fig. 8 shows the
EPR spectra at 115 GHz of LiAl0.9Ni0.1O2 and Ni-doped
LiAlO2. For the Ni-doped LiAlO2 composition, the EPR
spectrum consists of a tetragonal symmetry signal with a
small anisotropy in the perpendicular region: g1? ¼ 2:1700,
g2? ¼ 2:1610 and g ¼ 2:0379. In all cases, g is greater
?
k
than gJ, indicating that the Ni3+ ions are in a tetragonally
2
elongated octahedron with an A1g ground state. This
observation is consistent with the EPR data on Ni3+ in
LiAlyCo1ꢀyO2 solid solutions [37]. In the high-frequency
EPR spectrum of LiAl0.9Ni0.1O2, two overlapping signals
account for the EPR profile: one signal with tetragonal
symmetry and parameters close to that of Ni3+ in LiAlO2
and second signal with symmetrical shape and g ¼ 2:121.
These two signals can be assigned to Ni3+ ions in pure Al-
and in mixed Ni, Al-environment, respectively.
Fig. 7. The EPR spectra at 103 and 293 K of LiAl0.75Ni0.25O2 prepared at
600 1C from hydroxide precursor.
4. Conclusions
It was found that the incorporation of Al into the
layered structure of LiNiO2 is limited. At atmospheric
pressure, the citrate precursor method allows preparing
LiAlyNi1ꢀyO2 solid solutions in a limited concentration
range: 0pyp0:5. Above y40:5, a mixture between g-
LiAlO2 and Al-poor composition (y ꢂ 0:5) is obtained
when the citrate precursor method is used. Al-rich
compositions with yX0:75 are only formed under mild
conditions comprising the reaction between co-precipitated
Ni, Al hydroxides and LiCO3 at 600 1C. Even in the case
when high-pressure synthesis in oxygen-rich atmosphere is
used, layered LiAlyNi1ꢀyO2 solid solutions are formed in
Fig. 8. X-band and high-frequency EPR spectra (9.23 and 115 GHz) of
LiAl0.9Ni0.1O2 and Ni-doped LiAlO2 prepared at 600 1C from hydroxide
precursor.

ARTICLE IN PRESS
E. Shinova et al. / Journal of Solid State Chemistry 179 (2006) 3151–3158
3158
the limited concentration range of 0pyp0:5 and
0:75oyp1:0.
[8] S.H. Park, K.S. Park, Y.K. Sun, K.S. Nahm, Y.S. Lee, M. Yoshio,
Electrochim. Acta 46 (2001) 1215.
[9] M. Guilmard, A. Rougier, M. Grune, L. Croguennec, C. Delmas,
¨
J. Power Sources 115 (2003) 305.
When Al substitutes for Ni, there is a strong contraction
of the mean AlyNi1ꢀy–O bond length concomitant with an
increase of the LiO2-layer thickness, culminating in the
absence of LiAlyNi1ꢀyO2 solid solution for the 0:5oyo
0:75 concentration range. The extent of the trigonal
distortion of both AlyNi1ꢀyO6- and LiO6-octahedra de-
pends on the Al content and on the preparation method
used. High-pressure synthesis favors the formation of
oxides with a higher extent of trigonal distortion of both
AlyNi1ꢀyO6 and LiO6-octahedra. At atmospheric pressure,
higher synthesis temperature (800 1C) promotes the reac-
tion of cationic mixing between the layers.
[10] J. Kim, J. Liu, C. Chen, K. Amine, J. Electrochem. Soc. 150 (2003)
A1491.
[11] M.Y. Song, R. Lee, I. Kwon, Solid State Sci. 156 (2003) 319.
[12] P. Kalyani, N. Kalaiselvi, N.G. Renganathan, M. Raghavan, Mater.
Res. Bull. 39 (2004) 41.
[13] M. Guilmard, L. Croguennec, C. Delmas, Chem. Mater. 15 (2003)
4484.
[14] Y.-I. Jang, B. Huang, H. Wang, G.R. Maskaly, G. Ceder, D.R.
Sadoway, Y.-M. Chiang, H. Liu, H. Tamura, J. Power Sources 81
(1999) 589.
[15] T. Amriou, A. Sayede, B. Khelifa, C. Mathieu, H. Aourag, J. Power
Sources 130 (2004) 213.
[16] L.D. Dyer, B.S. Borie, G.P. Smith, J. Am. Chem. Soc. 76 (1954) 1499.
[17] R. Alcantara, P. Lavela, J.-L. Tirado, E. Zhecheva, R. Stoyanova,
´
J. Solid State Electrochem. 3 (1999) 121.
[18] C. Delmas, M. Menetrier, L. Croguennec, I. Saadoune, A. Rougier,
Acknowledgments
C. Poillerie, G. Prado, M. Grune, L. Fournes, Electrochim. Acta 145
¨
(1999) 243.
The authors thank the National Science Fund of
Bulgaria (Contract no. Ch1304/2003) for financial support.
R.S. and E.Sh. are grateful to the EC for a grant of a EU
‘‘Research Infrastructures: Transnational Access’’ Pro-
gram (Contract no. 505320 (RITA)—High Pressure) for
performing high-pressure synthesis experiments at the
[19] M. Marezio, J.P. Remeika, J. Chem. Phys. 44 (1966) 3143.
[20] M. Marezio, Acta Crystallogr. 19 (1965) 396.
[21] X. Li, T. Kobayashi, F. Zhang, K. Kimoto, T. Sekine, J. Solid State
Chem. 177 (2004) 1939.
[22] K.R. Poeppelmeir, C.K. Chiang, D.O. Kipp, Inorg. Chem. 27 (1988)
4523.
Bayrisches Geoinstitut, Universitat Bayrewth. The high-
¨
[23] S. Buta, D. Morgan, A. Van der Ven, M.K. Aydinol, G. Ceder,
J. Electrochem. Soc. 146 (1999) 4335.
frequency EPR measurements carried out at High Mag-
netic Field Laboraty in Grenoble, France, were supported
by the European Commission within the Sixth framework
program ‘‘Transnational Access—Specific Support Ac-
tion’’ (Contract no. RITA-CT-2003-505474)—‘‘Access to
research in very high magnetic field’’. The authors are very
grateful to Dr. T. Boffa-Ballaran and Dr. C. McCammon,
from Bayerisches Geoinstitut, Universitat Bayreuth, and
¨
Dr. A.-L. Barra, from High Magnetic Field Laboratory in
Grenoble, for their help.
[24] C.H. Chang, J.L. Margrave, J. Am. Chem. Soc. 44 (1968) 2020.
[25] E. Shinova, E. Zhecheva, R. Stoyanova, G.G. Bromiley, J. Solid
State Chem. 178 (2005) 1661.
[26] E. Shinova, E. Zhecheva, R. Stoyanova, G.D. Bromiley, R.
Alca
´
ntara, J.L. Tirado, J. Solid State Chem. 178 (2005) 2692.
guez-Carvajal, in: Satellite Meeting on Powder Diffraction of
[27] J. Rodrı
´
the XV Congress of the IUCr, 1990, p. 127.
[28] E. Zhecheva, R. Stoyanova, Solid State Ion. 66 (1993) 143.
[29] R. Kano, T. Shirane, Y. Inaba, Y. Kawamoto, J. Power Sources 68
(1997) 145.
[30] G. Prado, E. Suard, L. Fournes, C. Delmas, J. Mater. Chem. 10
(2000) 2553.
´
[31] E. Gaudin, F. Taulelle, R. Stoyanova, E. Zhecheva, R. Alcantara,
References
P. Lavela, J.L. Tirado, J. Phys. Chem. B 105 (2001) 8081.
[32] G. Nuspl, M. Nagaoka, K. Yoshizawa, F. Mohri, T. Yamabe, Bull.
Chem. Soc. Jpn. 71 (1998) 2259.
[1] M. Broussely, P. Biensan, B. Simon, Electrochim. Acta 45 (1999) 3.
[2] M.S. Whittingham, Chem. Rev. 104 (2004) 4271.
[3] T. Ohzuku, A. Ueda, M. Kouguchi, J. Electrochem. Soc. 142 (1995)
4033.
[33] R. Alcantara, P. Lavela, J.-L. Tirado, R. Stoyanova, E. Kuzmanova,
E. Zhecheva, Chem. Mater. 9 (1997) 2145.
[34] R. Stoyanova, E. Zhecheva, C. Friebel, J. Phys. Chem. Solids 54
(1993) 9.
[4] O. Zhong, U. von Sacken, J. Power Sources 54 (1995) 221.
[5] T. Ohzuku, T. Yanagawa, M. Kouguchi, A. Ueda, J. Power Sources
68 (1997) 131.
[35] R. Stoyanova, E. Zhecheva, C. Friebel, Solid State Ion. 73 (1994) 1.
[36] R. Stoyanova, E. Zhecheva, R. Alcantara, P. Lavela, J.-L. Tirado,
Solid State Commun. 102 (1997) 457.
[6] R. Stoyanova, E. Zhecheva, E. Kuzmanova, R. Alca
J.L. Tirado, Solid State Ion. 128 (2000) 1.
[7] C. Julien, G.A. Nazri, A. Rougier, Solid State Ion. 135 (2000) 121.
´
ntara, P. Lavela,
´
[37] R. Stoyanova, E. Zhecheva, R. Alcantara, J.L. Tirado, J. Phys.
Chem. B 108 (2004) 4053.