1 s2.0 S0927796X02001043 Main

Materials Science and Engineering R 40 (2003) 47–102
Lithium intercalated compounds

Charge transfer and related properties
C.M. Julien*
Laboratoire des Milieux Désordonnés et Hétérogènes, CNRS-UMR 7603, Université Pierre et Marie Curie,
4 place Jussieu, case 86, 75252 Paris Cedex 05, France
Accepted 20 September 2002
Abstract
Numerous channel- or tunnel-structured compounds are interesting materials in which lithium intercalation
occurs primarily without destruction of the host lattice assuming a complete charge transfer. In many cases, the
rigid-band model (RBM) is a useful first approximation for describing the changes in electronic properties of the
host material with intercalation. The applicability of the rigid-band model is taken as a test for the properties most
desirable in a good intercalation material. This review paper presents experimental results, namely optical and
electrical measurements, obtained on various intercalation materials such as dichalcogenides, trichalcogenides and
oxides to probe the validity of the RBM. The impact of the lithium intercalation is discussed for various structural
forms, from well-crystallised to amorphous compounds. We observed, nevertheless, that the rigid-band model is not
applicable to all of the layered intercalation materials. This needs yet to be more extensively documented for their
promising application as insertion electrode in rechargeable lithium batteries. Compounds such as TiS2, TaS2, MoS2,
NiPS3, MoO3, V2O5, V6O13, LiCoO2, LiNiO2, LiMn2O4 and LiNiVO4 are tested in this work. Some guidelines are
established for improving the performances of these materials in their most eminent applications.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Intercalation compounds; Rigid-band model; Optical spectroscopy; Disordered materials; Lithium batteries
1. Introduction
Compounds exhibiting an open structure, in particular the transition-metal dichalcogenides

(TMDs) and transition-metal oxides (TMOs), can be intercalated with a wide range of both organic
and inorganic materials which may have a profound influence on the physical properties of the host
compound [1–3]. The intercalation reaction in these compounds is driven by charge transfer from the
intercalant to the host layered compound conduction band and thus electron-donating species can
take place in such a reaction. The reversible ion-electron transfer reaction is classically represented
by the scheme
xAþ þ xe þ hHi , Ax þ hHix (1)

in the usual case, where hHi is the host material, A an alkali metal and x the molar intercalation
fraction. The electronic transport plays an important role in such reaction towards the formation of
intercalation compounds. It also governs the phase transitions as the parameter expansion of the host
*
Tel.: þ33-144-274-561; fax: þ33-144-274-512.
E-mail address: cjul@ccr.jussieu.fr (C.M. Julien).
0927-796X/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 7 - 7 9 6 X ( 0 2 ) 0 0 1 0 4 - 3
48 C.M. Julien / Materials Science and Engineering R 40 (2003) 47–102
Fig. 1. The fundamental question concerning the evolution of the Fermi level.
structure has an electronic component. Consequently, it is possible to consider three classes of

intercalation reaction that correspond to the different steps in the delocalisation of the transferred
electrons. The level of acceptance can be either a discrete atomic state, or a molecular level of a
discrete polyatomic entity existing in the structure, or part of a conduction band.
The rigid-band model (RBM) is a useful approximation for describing the changes in electronic
properties of the host material with intercalation (Fig. 1). Sellmyer [4] distinguishes two versions of
the rigid-band model for dilute solid solutions which might be called the electron-gas RBM and the
screened-impurity RBM. In the former, due mainly to Jones [5], the valence electrons are regarded
essentially as in plane wave states and the only effect of alloying with an element having a valence
difference DZ, is to change the free-electron density to a new value, that obtained simply by scaling
the valences of the solvent and solute according to their atomic fractions in the alloy. In the screened-
impurity RBM of Friedel [6], it is recognised that the electron gas cannot support an electric field at
long distances from the charged impurity. It can be shown with a dielectric-screening or Thomas–
Fermi level that the conduction electrons will redistribute themselves to screen out the Coulomb
field. In this case, any charged impurity added to a solid will polarise the solid, and attract to itself.
The concept of rigid-bands implies a severe chemical stability of the intercalated host. From the
energetic point of view, this means the total energy of the substance is little affected by the addition
of intercalant electrons. The consequence is that the structure too is stable, and the only energy band
involved in intercalation is the narrow ‘‘d’’ conduction band in TMD materials for instance. These
are precisely the properties most desirable in a good cathode material which provide features such as
stable voltage against ageing and mechanical durability [7]. It is most important, therefore, to
investigate theoretically and experimentally how well the approximation can apply in either positive
or negative electrode materials employed for energy storage in rechargeable lithium batteries.
In this review, the validity (or the non-validity) of the rigid-band model is demonstrated in
Section 2. Optical and/or electrical experiments, which allow to investigate the band structure of
intercalation compounds such as dichalcogenides, trichalcogenides and oxides, are presented. To
probe the validity of the RBM and to examine the impact of the lithium intercalation, various
structural forms, from well-crystallised to amorphous compounds, are investigated using various
techniques such as X-ray diffractometry (XRD), Raman scattering (RS), Fourier transform infrared
(FTIR) spectroscopy and electrical conductivity measurements. Compounds such as TiS2, TaS2,
MoS2, NiPS3, MoO3, V2O5, V6O13, LiCoO2, LiNiO2, LiMn2O4 and LiNiVO4 are tested in this work.
Section 3 presents the physico-chemical properties of lithium intercalated compounds in relation
with their electrochemical features. The last part, Section 4, is devoted to the intercalation reaction in
disordered structures. Structure–composition relations are examined and some guidelines are
established for improving the performance of these materials in their most eminent applications.
C.M. Julien / Materials Science and Engineering R 40 (2003) 47–102 49
Fig. 2. Schematic band structures for all the TMD compounds with octahedral and trigonal prismatic coordination.
Sketches show the electronic structures before (a) and after (b) lithium intercalation with the respective position of the
Fermi level.
2. The reliability of rigid-band model
2.1. Lithium intercalation in TiS2
Electronic band structure of TMD materials have been generally calculated using simplest
molecular-orbital arguments [8]. The schematic band structure of the MX2 compounds with
octahedral and trigonal prismatic coordination are shown in Fig. 2. In the simple picture, during
intercalation, the donating-electrons will occupy one of the empty d-bands. The simplest
approximation to the band structure of an intercalation compound is just that of the parent host
compound with the Fermi level moved up to accommodate the extra electrons.
In LixTiS2, the magnitude of the Hall coefficient (RH), decreases with increasing lithium
content, confirming the occurrence of electron transfer from the intercalate to the host [9–11]. The
electron concentration in TiS2 before intercalation is 3:1 1020 cm3 indicating that this material is
of stoichiometry Ti1.0044S2 [11]. Upon lithium intercalation we observe a large decrease of the
resistivity as well as of the Hall coefficient. The carrier concentration in electro-intercalated samples
increases to 5 1021 and 9:6 1021 cm3 for x ¼ 0:25 and 0.5, respectively. The Hall coefficient
RH of all the samples is nearly temperature independent, as would be expected for a normal metal.
Metallic behaviour is also seen in the temperature dependence of the resistivity where, with
increasing x, a tendency towards an almost linear behaviour is noted. Indeed, the temperature
dependence of the resistivity in TiS2 and LixTiS2 can be fitted by the expression
rðTÞ ¼ r0 þ AT g ; (2)
where r0 is the temperature-independent residual resistivity and A is a constant. The power g is found
to vary from 2.3 in good-stoichiometry samples (with carrier concentration n ¼ 1:1 1020 cm3) to
1.6 in poor-stoichiometry samples (2:9 1021 cm3), with Tmin ranging from 20 to 110 K in the
corresponding cases.
Fig. 3. Room-temperature absorption spectra of TiS2 and LiTiS2 (after [10]). The schematic band structure shows that the
rigid-band model may be used for the intercalation complexes.
Klipstein et al. [9,10] explain this behaviour by a model involving the interplay between inter-
and intra-pocket scattering of electrons by longitudinal acoustic phonons, whereby the increase in
Fermi surface dimensions reduces the restriction on the wave vector of phonons that may take part in
the scattering process, implying that, as the carrier concentration increases, g should tend towards
unity and, simultaneously, the temperature Tmin below which the lnðr r0 Þ versus lnðTÞ curve starts
deviating from linearity, should increase. This model, which was originally based on studies in
pristine TiS2 with a varying degree of stoichiometry, was later verified to remain valid for higher
carrier concentrations, such as in TiS2 intercalated with lithium via the butyl-lithium technique [7] or
intercalated with hydrazine [12].
Electron transfer is also apparent in the optical properties of the Li–TiS2 system. Fig. 3 shows
the absorption spectra, in the energy range 0.5–6.0 eV of pure and Li-intercalated TiS2 [13]. In the
spectrum of the intercalated product, it is clearly seen that it shows free-carrier absorption, i.e. Drude
type, below 1 eV for LiTiS2. We also observe inter-band transitions which are the first direct allowed
transitions from the p-valence to the d-conduction band at the L point of the Brillouin zone. Moving
into the spectrum of LixTiS2, the onset of inter-band transitions is seen to have shifted to higher
energies and the oscillator strength under the absorption band is roughly halved. Beal and Nulsen
[13] argue that this is exactly that one would expect if the dz band is now half-filled following
saturation of the intercalation complexes.
Another optical experiment is the infrared reflectivity carried out on Li-electro-intercalated TiS2
[11]. Fig. 4 shows the temperature dependence of the reflectivity spectra of pure TiS2 and Li-
intercalated TiS2 single crystals. We observe a large shift in the plasma edge for Li1.0TiS2 with
respect to pure TiS2. According to the single carrier Drude model, the analysis of the dielectric
function gives the values as follows. In TiS2, the Drude edge lies around 1200 cm1, while in LiTiS2
Fig. 4. FTIR reflectivity spectra of (a) TiS2 and (b) LiTiS2 as a function of temperature.
the Drude edge occurs at about 4000 cm1 giving plasma frequencies of 1360 and 4100 cm1,
respectively, if we are taking into account that the high-frequency dielectric constant remains similar to
that of the pristine material and considering the electron effective mass as obtained by Isomaki et al.
[14]. At the L-point of the Brillouin zone, Isomaki et al. estimate ma ¼ 0:4m0 along the a-axis. This
assumption implies that the optical effective mass mopt has a value higher than 1.3m0. In the present
studies, the Drude analysis gives a carrier concentration of 1:7 1022 cm3 for complete intercalation
of TiS2 at x ¼ 1. This is in excellent agreement with the theoretically expected value of
1:75 1022 cm3 and very close to the value of 2:2 1022 cm3 determined from Hall measurements.
We assume in the spirit of the rigid-band model [7] that intercalation does not change
appreciably the conduction band effective mass, nor the high-frequency dielectric constant of the
host material. The charge transfer Dn from the alkali–metal atoms to the d-conduction band of the
host compound can be directly calculated from the difference between o2p before and after
intercalation. Here, Dn is expressed in terms of the number of electrons transferred per Ti atoms.
Using this method, we have obtained Dn ¼ 0:9
0:1 electrons. The uncertainty of 0.1 electrons is
thought to be a reasonable estimate in view of the assumptions made. It is interesting to note the
large increase of the plasma damping factor from 310 to 2160 cm1 in Ti1.005S2 and Li1.0Ti1.005S2,
respectively. This increase is observed in the energy-loss function by the broadening of the plasmon
peak. The damping factor can be expressed as follows:
1 q
G¼ ¼ ; (3)
t m mH
where mH is the Hall mobility of free carriers. The observed increase of g suggests a decrease of the
Hall mobility or a modification of the effective mass for the intercalated sample. In Li1.0Ti1.005S2 the
Fig. 5. Temperature dependence of the damping factor (inverse relaxation time) for (a) TiS2 and (b) LiTiS2.
electron mobility estimated by Hall measurements has a value of 1.9 cm2 V1 s1 at room
temperature [11]. This value can be related with those given in the literature of 13.5 and
0.35 cm2 V1 s1 for TiS2 and LiTiS2, respectively [9].
FTIR reflectivity spectra of the Li1.0Ti1.005S2 sample shows surprising departure from ordinary
Drude behaviour, and there is no strong change in the IR spectra as a function of the sample
temperature in comparison with that in pure material [15]. The dip in the reflectivity is close to the
plasma frequency op ¼ 4180 cm1 extracted from analysis of the data using a Drude-like model
with a frequency-dependent relaxation time, as
1
¼ xt0 þ a½ðpTÞ2 þ o2 : (4)
tðx; T; oÞ
A good fit to the optical data is achieved with the scattering rate given by Eq. (4), which reduces
to the ideal electron–electron scattering behaviour in the isotropic three-dimensional effective mass
model [16]. For Li-intercalated TiS2 sample, the temperature and frequency components of Eq. (3)
are strongly screened by the first term (xt0) as shown in Fig. 5. This may be due to the complete
filling of the d-band associated with a very low Hall mobility. In this case, it is difficult to evaluate
the optical mobility because the quantity ot @ 1 is no more valid. Considering that Hall
measurements on the Li1.0Ti1.005S2 sample give N H ¼ 1:8 1022 cm3 and that the Fermi energy
obtained by optical reflectivity measurements is EF ¼ 4180 cm1 ¼ 0:52 eV, we may estimate the
electron effective mass m ¼ 0:49m0. This value is very close to that in pure material reported by
Isomaki et al. [14]. In conclusion, it can be seen from the electrical and optical properties of the Li-
intercalated TiS2 presented above that, s-bands aside, they can all be explained in terms of the rigid-
band model. It is worth mentioning here that optical absorption results by Scholz and Frindt [17] on
Ag-intercalated TiS2 also agree with this model.
2.2. Lithium intercalation in TaS2
Among the group V TMDs, TaS2 has perhaps been the subject of greatest interest, because of
the fascination range of structural and electronic properties that this material exhibits. Due to valence
Fig. 6. Absorption spectra of pure 1T-TaS2 and 1T-LixTaS2. The schematic band structure shows the lowering of the dz
band upon Li intercalation (after [15]).
electron occupying their dz2 band, this metallic compound can exist in either 1T-, 2H-, or 4H-
structure [18]. As a consequence of the switching from octahedral (Oh) to trigonal prismatic (TP)
coordination, the shift of the dz2 band to lower energies occurs gradually. The absorption spectrum of
pure 2H-TaS2 shows a Drude edge below 1 eV associated with the free-carrier absorption in this
material owing to the half-filled dz2 band. After intercalation, the Drude edge disappears and the first
dz2 ! d transition shifts toward lower energy. These changes are attributed to the gradual filling of
the dz2 band by electron transferred from lithium.
The absorption spectrum of 1T-TaS2 is shown in Fig. 6. Above 1.5 eV, the absorption bands are
associated with the dz2 ! dz2 and dz2 ð1Þ ! d transitions, whereas the band above 3.5 eV owing to
ð1Þ ð2Þ
p ! dz2 ð2Þ transition. Charge transfer from Li to the host lattice increases the population of the dz2
band and raise the Fermi level EF to a new energy, E0 F. The displacement of the strong absorption
edge around 3 eV indicates a considerable lowering of the dz2 band with respect to its position in
the pristine material. The lowering of the dz2 band is attributed to the filling with electrons donated
by Li, as well as the modification of the crystal structure, e.g. an increase of the c/a ratio after
intercalation [19]. These results provide further support to the argument that, upon lithium
intercalation the rigid-band model is not entirely applicable in 1T-TaS2.
2.3. Lithium intercalation in 2H-MoS2
Among the group VI TMD, MoS2 is one of the semiconducting materials for which
intercalation reaction induces a transition of the host-related to local ligand field modification. In
that particular case, molybdenum presents a trigonal prismatic sulphur coordination which changes
Fig. 7. (a) Discharge curve of Li||LixMoS2 and (b) related incremental capacity, @x=@V.
to the octahedral one (TP ! Oh transition) [20–22]. The structure modification is accompanied by
an increase of the M–X band ionicity in agreement with the respective stability of the new atomic
arrangement, the Coulomb repulsion between partially charged ligands favouring the octahedral
form. Also, comparison of the d-band density of states for 2H-MoS2 and hypothetical 1T-LiMoS2
shows that the occupied bands, which contain six states, are lower in the case of the Oh phase
corresponding to the glide process between Mo and S atoms. This is a fine example of destabilisation
through lithium reduction.
The transformation from (TP) to (Oh) coordination is attributed to a process which is driven by
a lowering of the electronic energy for the octahedral structure when electrons are donated from Li to
the MoS2 layer upon intercalation [7]. The octahedral transformation starting at x 0:1 completes
around x ¼ 1:0 and is preserved on subsequent cycle of deintercalation–intercalation. Cell
discharge–charge occurs in the range 2.2–1.3 V with a mid-discharge voltage of 1.7 V. Electron
diffraction studies on the Li–MoS2 system have shown that this type of transformation is
accompanied by the formation of a 2a0 2a0 superlattice [23].
Fig. 7a and b shows the cell voltage versus composition of Li||LixMoS2 and the related variation
in the inverse derivative voltage, i.e. incremental capacity ð@x=@VÞ, at ambient temperature [22].
The natural sample has a 2H-MoS2 structure and upon Li intercalation behaves as a two-phase
system. The first phase is the initial material and the second one is the 1T-structure of Li1.0MoS2
which appears at x ¼ 1 [20,21]. The incremental capacity (Fig. 7b) shows that the electrode material
exhibits a complex intermediate behaviour; at least four states can be observed up to Li1.0MoS2, but
the analysis of such a feature is very difficult because the validity of the Fick law requires that the
host material remains a single phased compound. However, it is interesting to note that, in this case,
Fig. 8. Li intercalation of MoS2 in n-butyl-lithium. (a) After 0.2 h, defects are created near the edges and in the steps of the
specimen denoted by arrows. (b) After 2 h intercalation, superlattice spots appear (denoted by the letter s), which are
indexed as (1/2 0 1/2 0). Notice also the splitting of the main spots (denoted by the letter M). (c) A microcrograph taken
from the same area reveals that the specimen is heavily defected owing to intercalation. (d) Distribution of lithium vs.
distance from the edges of the specimen as revealed by SSNTD images.
the positive electrode is highly strained and each of the above states can be described approximately
with an interaction energy, U0,1, of the intercalant. At room temperature the difference in standard
potential U0,1 of the successive states is very small.
The structure observed at x 0:25 in the incremental capacity (Fig. 7b) may be related with the
formation corresponding to a 2a0 2a0 superlattice which is interpreted as a pseudo-staging on the
basal hexagonal lattice. These features were identified in the Raman scattering measurements of
Li0.3MoS2 by Sekine and co-workers [24,25] and in the electronic microscopy diffraction mode by
Chrissafis et al. [23]. Fig. 8 shows the TEM micrographs of Li-intercalated MoS2 in n-butyl-lithium.
(a) After 0.2 h, defects are created near the edges and in the steps of the specimen denoted by arrows.
(b) After 2 h intercalation, superlattice spots appear (denoted by the letter s), which are indexed as
(1/2 0 1/2 0). Notice also the splitting of the main spots (denoted by the letter M). (c) A
microcrograph taken from the same area reveals that the specimen is heavily defected owing to
intercalation. (d) Distribution of lithium versus distance from the edges of the specimen as revealed
by SSNTD images. However, we do expect that LixMoS2 is accompanied by the raising of the Fermi
level due to the charge transfer from Li intercalation. This is depicted in Fig. 9, where the
temperature dependence of the electrical conductivity of lithium intercalated MoS2. Undoped MoS2
is an n-type semiconductor with conductivity expressed by s expðEa =kB TÞ with Ea ¼ 0:05 eV,
while Li0.3MoS2 is a highly degenerate semiconductor with s T 1:4 . Here, we must be aware of
the limitations of the RBM. It is probably non-appropriate for LixMoS2 because the fully occupied
dz2 states in the pristine material imply new electrons to enter the next higher ‘‘d’’ band with a
change of the c/a ratio associated with the destabilisation of the host lattice.
Following the above structural transformation, phonon spectroscopy offers an excellent way of
quantifying the degree of anisotropy not only by distinguishing inter- and intralayer normal modes
but also determining the shear moduli in different directions. The Raman spectrum of 2H-MoS2 at
room temperature is shown in Fig. 10. It exhibits four Raman-active bands that are the intralayer A1g
Fig. 9. Temperature dependence of the electrical conductivity of lithium intercalated MoS2.
mode at 407 cm1 involving motion along the c-axis, the intralayer E2g mode at 382 cm1 involving
motion in the based plane, the E1g mode at 286 cm1 and the rigid-layer (RL) mode at 32 cm1 of
E2g symmetry. This last mode is of interlayer type involving rigid motion of neighbouring
sandwiches in opposite phase.
The Raman spectra of LixMoS2 with x 0:1 and 0.3 (Fig. 10) display the structural changes
from the b-phase (2H structure) to a a-phase (1T structure) upon Li intercalation [24,25]. This
transformation from trigonal prismatic to octahedral coordination has been attributed to a process
which is driven by a lowering of the electronic energy for the octahedral structure when electrons are
donated from Li to the MoS2 layer on intercalation [24,25]. The octahedral transformation in
LixMoS2 starts at x ¼ 0:1 and completes around x ¼ 1. For a degree of intercalation x 0:1, the
Raman intensity is considerably reduced (by a factor 5) and we observe two new bands: a broad peak
located at 153 cm1 (A-line) and a weak peak situated at 205 cm1 (B-line) and the intensity of the
Fig. 10. Raman spectra of 2H-MoS2 natural crystal as a function of Li content lithium.
RL mode is reduced. The two pristine intralayer modes can still be observed, with little shift in
frequency, but both are split to give weak additional side bands towards lower energies (C- and D-
lines). These Raman bands are attributed to the Davydov pairs of the optical phonon branches. For
x ¼ 0:3, the spectrum of LixMoS2 is modified compared with the former ones. The RL mode is no
longer recorded. All other lines are still observed. We remark the small shift in frequency of the
lattice modes of MoS2.
A simple model has been used to calculate the frequencies of the new modes appearing upon Li
intercalation. The intercalation mode is given by
1=2
kð2m1 þ m2 Þ
oint ð0Þ ¼ ; (5)
m1 m2
where m1 and m2 are the masses of the MoS2 molecule and of the Li atom, respectively, and k is the
force constant between the S and Li atoms. We estimate k ¼ 8:23 103 dyn cm1, which is much
smaller than the intralayer force constants.
For a degree of intercalation x ¼ 0:3, we assume that LixMoS2 is a two-phase system. The
following changes on the lattice dynamics can be expected. The RL mode disappears because the
elementary cell of the 1T-structure contains only one molecular unit (three atoms per sandwich). The
symmetry changes from D6h to D3d and the new symmetry allows only the two Raman-active A1g
and Eg modes, which are representative of the intralayer atomic motions. The weak spacing
expansion observed upon Li intercalation and the difference of the molybdenum coordination do not
modify significantly the frequency of these modes. A simplest calculation gives a change of about
6% in frequency. Thus, we can trust the validity of the lattice dynamics model using a 2H-structure
[24,25].
2.4. Lithium intercalation in MoO3
The molybdenum oxides display a varieties of structural types involving linked MoO6 octahedra
whose arrangements are favourable for intercalation process. These materials offer high voltages and
composition intervals accessible for lithium intercalation are wide. The interest of a-MoO3 arises
from its layered structure presenting open channels for fast Li-ions diffusion, a higher
electrochemical activity versus Li/Liþ than that of chalcogenides and the highest chemical stability
among the oxide lattices [26]. Fig. 11 shows the first discharge–charge curves of a Li||MoO3 cell
using a well-crystallised film. A film grown by sputtering technique on nickel substrate in oxygen
partial pressure of 100 mTorr displays the typical electrochemical features of the a-MoO3 phase
[27].
Fig. 12 shows the temperature dependence of electrical conductivity of LixMoO3
(0:0 x 0:3) intercalated by electrochemical titration method. One observes a clear departure
from semiconducting behaviour of pure MoO3 material to degenerate semiconductor of Li-
intercalated MoO3 even at low Li content. The metallic features are also observed in the temperature
dependence of the Hall coefficient. The observation of plasma absorption in this material implies
carrier concentration of at least 1018 electrons cm3 indicating a weak variation of the free-electron
effective mass and of the high-frequency permittivity. In MoO3, the bonding framework is composed
of five O(pp) and three Mo(t2g) orbitals which interact to form p and p bands [28]. As the
antibonding p states hold the extra electrons supplied by the inserted lithium, Li0.3MoO3 may be
expected to exhibit two-dimensional electronic conductivity. The narrowing of the conduction band
is expected to lead to an increase in the effective electron mass which can affect the position of the
Fig. 11. First discharge charge curves of a Li||MoO3 cell using a well-crystallised film grown by sputtering technique Ni
substrate (after [27]).
Drude edge in LixMoO3 phases. The conductivity of MoO3 is believed to exist because of the
electron hopping between Mo6þ and Mo5þ sites. The nature of the conductivity variation observed in
Fig. 12 corresponds to a steadily decrease with the addition of intercalants. Intercalation of lithium
ions in the MoO3 structure is believed to lower the valency state of molybdenum ions by transferring
electrons from lithium to molybdenum ions. The relative concentration of Mo6þ and Mo5þ ions
results in the lowering of the conductivity towards an insulator behaviour. Further experiments are
needed to elucidate the mechanism of the charge transfer occurring in transition-metal oxide
compounds but the rigid-band model seems adequate in LixMoO3 which found technological
application in electrochromic rear mirror in automotive industry.
Infrared absorption studies of LixMoO3 compounds revealed a transition from metallic to small-
polaron features [29]. After intercalation, the lattice vibration spectrum is completely screened by
the free electrons in the host material. The Drude edge contribution, i.e. plasmon feature, is
responsible for the metallic absorption due to high electron density in Li0.3MoO3 as shown in Fig. 13.
The free-carrier absorption coefficient can be expressed by
a ¼ o2p t½ncð1 þ o2 t2 Þ; (6)
Fig. 12. Arrhenius plot of the electrical conductivity of a-MoO3 and LixMoO3.
Fig. 13. FTIR absorption spectra of a-MoO3 and Li0.3MoO3.
where op is the plasma frequency, t the relaxation time of the free carrier, n the refractive index and c
the light velocity. Using Eq. (6), the fit of experimental data gives a carrier concentration of
5 1016 cm3 in Li0.3MoO3. This value is in good agreement with the Hall measurements. The
temperature dependence of the absorption coefficient shows a small increase of a with temperature,
which can be attributed to the fact that Li-intercalated MoO3 is a degenerate semiconductor for x ¼ 0:3.
Nadkarni and Simmons [30] studied the electrical properties of MoO3 and reported that there is
a donor band between the conduction and the valence bands due to oxygen vacancies. MoO3 has the
outer electron configuration 4s5 5s1. If MoO3 is considered to be ionic, i.e. composed only of Mo(VI)
and O2 ions, the valence band would be composed of oxygen 2p states and the conduction band of
empty molybdenum 4d and 5s states [31]. Upon Li intercalation the electrical conductivity of
LixMoO3 increases by two orders of magnitude and the temperature dependence of s shows
important changes in the conduction mechanism. The semiconducting character of MoO3 gradually
disappears and, for a degree of intercalation of x ¼ 0:3, the material exhibits a metallic behaviour.
2.5. Lithium intercalation in V6O13
V6O13 is a black colour material which derives from the ReO3 structure and is intermediate in
composition between VO2 and V2O5. In this family, V3O7 and V4O9 appear to have an intermediate
structure between that of V6O13 and V2O5. The monoclinic structure of V6O13 contains edge-shared
distorted VO6 octahedra forming single and double zigzag chains linked together by further edge
sharing corner-shared. The resulting sheets (single and double) are interconnected by corner sharing,
thus giving a three-dimensional framework [32]. This structure contains tri-capped cavities joined
through shared square faces. The three open faces of the cavity should permit lithium-ion diffusion
along (0 1 0) with the possibility of exchange between pairs of adjacent channels. Stoichiometric
V6O13 can be written as (V4þ)4(V5þ)2(O2)13 as far as the valence states of the vanadium ions are
concerned.
The structure of V6O13 is interesting from an electrochemical viewpoint due to the theoretical
maximum limit of lithium uptake giving an energy density of 890 mWh g1. The stoichiometric
V6O13 structure is believed to accommodate 8 Li per formula unit as determined by the available
electronic sites rather than the structural cavities [33]. The limit corresponds to a situation when all
the vanadium ions are present in the trivalent V3þ state. As a function of the stoichiometry, the
maximum uptake goes to 1.35 Li for VO2.18 oxide. Reversible chemical and electrochemical
Fig. 14. Arrhenius plot of the electrical conductivity of V6O13 and LixV6O13 (1 x 6).
insertion of lithium into V6O13 were first demonstrated by Murphy et al. [34,35], and its potential as
an active cathode material in practical batteries has since been more fully investigated [33]. The
discharge curve exhibits three distinct plateaus, reflecting the sequential filling of non-equivalent
sites in the host structure.
Fig. 14 shows the Arrhenius plot of the electrical conductivity of V6O13 and LixV6O13
(1 x 6). The pure material has a conductivity of 1 102 S cm1 at room temperature and
exhibits a semiconductor behaviour. The electronic conduction in V6O13 is due to the electron
hopping between V4þ and V5þ states. The nature of the conductivity observed in Fig. 14 corresponds
to a steadily decrease with addition of Li ions in the LixV6O13 framework. Intercalation of Li ions is
believed to lower the valence state of vanadium ions by transfer of electrons. The relative
concentration of reduced cations results in the lowering of the conductivity towards a poor electronic
semiconductor. Electronic conductivity of pressed V6O13 powders indicates a sharp fall in two steps
with increasing Li content [36,37]. For lithiated V6O13, we observe a continuous decrease of the
electrical conductivity. Lowest conductivity of 5 104 S cm1 has been measured in Li6V6O13.
This is also accompanied by an increase in activation energy, a general phenomenon observed in any
oxide with small-polaron conduction.
Infrared studies of LixV6O13 compounds revealed the transition from metal-like to small-
polaron features (Fig. 15). In pure V6O13, a Drude edge is observed around 200 cm1 which is the
contribution of the free-charge carriers. For lithiated V6O13, we record a continuous decrease of the
electrical conductivity which is also seen in the far-infrared spectrum by the disappearance of the
Drude absorption [29]. Both techniques confirm the charge transfer occurring upon Li insertion in
the V6O13 framework.
2.6. Lithium intercalation in LiCoO2
Lithiated transition-metal oxides with a layered, a-NaFeO2-type, structure such as LiMeO2

(Me ¼ Ni, Co) have been a great interest as positive electrode materials for rechargeable lithium
batteries. This structure crystallises in the rhombohedral system (R3m space group) in which Liþ,
M3þ (M ¼ Co; Ni) and O2 occupy 3a, 3b and 6c sites (Wyckoff’s notations), respectively. LiCoO2
Fig. 15. FTIR absorption spectra of V6O13 and Li3.2V6O13.
has been proposed as cathode for lithium battery by Mizushima et al. [38] and currently being used
in commercial rechargeable lithium-ion batteries by Sony [39]. Fig. 16 shows the potential curve of
LiCoO2 during the first charge in the potential range 2.5–4.2 V versus Li/Liþ that corresponds to the
Li extraction from the host lattice. The charged capacity was 155 mAh g1 for the cathode and
matched well with published data [40].
Fig. 17 displays the FTIR absorption spectra of LixCoO2 cathode-active materials as a function
of the lithium content. As predicted from the factor-group analysis, one observes four distinct bands
(2A2u 2Eu modes) in the FTIR absorption spectrum of pristine LiCoO2. They are located at 269,
420, 539 and 602 cm1. The spectrum of LiCoO2 matches well with those reported previously [41–
45]. A closer examination of the shape of the high-wavenumber band at 602 cm1 indicates that a
shoulder exists at 646 cm1. We remark the shape of the IR band at 269 cm1 which corresponds to
an oscillator with a great strength. The infrared-active bands shown in Fig. 17 are generally broader
than those observed in Raman spectroscopy [44]. The broadening of IR bands is attributed to the
average oxidation state of Co, which are oxidised into the Co(IV) state during charge of the cell, and
to the random distribution of Liþ ions in the interlayer space.
Fig. 16. The potential curve of LiCoO2 during the first charge in the potential range 2.5–4.2 V vs. Li/Liþ (0.1 mA cm2
current density).
Fig. 17. FTIR absorption spectra of LiCoO2 as a function of the Li concentration.
To get a better understanding of the vibrational spectra of the layered LiCoO2 with R3m space
group, we consider a structure which consists of compressed CoO6 and elongated LiO6 octahedra,
that yields distinct vibrations in two different frequency regions. At 400–650 cm1 there are bands
due to CoO6 vibrations, while the LiO6 vibrations occur within the region 200–400 cm1. Infrared
bands located in the high-frequency region, i.e. at 602 and 539 cm1, are attributed to Co–O
stretching and O–Co–O bending motion, respectively. The low-frequency band situated at 269 cm1
involves the motion of Li atoms against their oxygen neighbours in an octahedral environment. This
peak is related to an asymmetric stretching vibration of LiO6 units [43].
There are obvious modifications of the FTIR spectrum of LiCoO2 upon lithium deintercalation
(Fig. 17) as follows. (a) We observed a decrease of the oscillator strength and a broadening of all the
infrared bands which can be associated to a disorder induced by the departure of Li ions located
between two CoO2 blocks. The broadening of the low-frequency band can be also attributed to the
random distribution of the Li ions remaining in the 8a sites of the host matrix. (b) No frequency
change is observed for the high-wavenumber bands which are assigned to the CoO6 vibrations. Thus,
as expected, we can conclude that the CoO2 layers are not affected significantly by the lithium
deintercalation process. (c) A significant shift of the low-frequency band is recorded. This band
shifts toward the low-energy side from 269 to 258 cm1 in Li0.5CoO2. The frequency shift
corresponds to the increase in the interlayer spacing due to an increase of the repulsive interactions
between two adjacent negatively charged CoO2 layers upon delithiation. Thus, the interlayer force
constant is reduced by about 8%. (d) The increase of the far-infrared absorption of Li0.4CoO2 sample
is attributed to the Drude edge due to the change in the electrical conductivity of the material. This
suggests the existence of collective delocalised electrons. These results agree well with the data of
electrical measurements which show that LiCoO2 has a semiconductor-like conductivity while
LixCoO2 exhibits almost a temperature-independent conductivity [45].
LiCoO2 is a p-type semiconductor (band-gap Eg ¼ 2:7 eV) [46] while LixCoO2 for x ¼ 0:75 has
a metal-like behaviour. LixCoO2 is predicted to have partially filled valence bands for x lower than
1.0 [47]. For every Li removed from LiCoO2 lattice, an electron hole is created within the valence
band. For x < 0:75, we expect that there are sufficient holes to allow for a significant degree of
screening, and in this regime, the hole states in the valence bands are likely to be delocalised such
that LixCoO2 exhibits metallic-like electronic properties. This behaviour is clearly observed in the
infrared absorption spectra where a strong absorption by holes occurs at low-wavenumbers. As
pointed out by Van der Ven et al. [47], the appearance of delocalised holes contribute to the free
energy of the electrode, influencing both the energetic and entropic terms. This could be at the origin
of the two-phase region observed in the potential curves of the Li||LiCoO2 cells for Li deintercalation
(charge) and/or intercalation (discharge) process.
From the FTIR data, it is also interesting to remark that LiCoO2 is less sensitive to lithium
content due to the higher bond covalency in the CoO2 slabs than LiNiO2 does. Consequently, the
strong bond covalency in LiCoO2, with reduced Co–O bond distance, results in stabilisation of
Co(III) in low-spin ground state, and reduces the electronic conductivity of the compound. Due to
the lithium deintercalation from the LixCoO2 crystal structure, the repulsion of the negatively
charged CoO2 layers increases and the Co4þ/Co3þ redox couple offers the possibility of the
electronic transfer. This cation oxidation results in an increasing electrical conductivity due to the
decrease of the covalent character of the CoO2 slabs. X-ray diffraction and infrared data seem to be
in good accordance with such a structural model.
3. Physico-chemical properties
Optical spectroscopy associated with transport measurements is a powerful tool for

investigating the charge transfer in lithium intercalated layered compounds. Optical studies include
photoluminescence, Raman scattering, far-infrared reflectivity and absorption measurements. They
are carried out to determine the structural transformations of Li-intercalated materials [48–50] but
are also fruitful for the knowledge of charge transfer [51–53]. The small structural change due to the
replacement of relatively weak Van der Waals forces by stronger coulomb-type interactions has been
theoretically treated [24,25], and as an experimental evidence, in LixFePS3 the n-LiþS6 mode has
been clearly observed in the infrared spectrum [54]. Optical absorption studies have played a
significant role in establishing the charge transfer concept, and one of the earliest experiments in this
regard is the work of Acrivos et al. [55] on NaxMoS2. Upon intercalation, the host lattice is accepting
electrons in the conduction band which has a high density of states as in LixMoS2 [55] or in an empty
gap states as the 3d level of Ni2þ in the LixNiPS3 [56].
The aim of this section is to present a selection of appropriate layered materials, which illustrate the
modifications occurring upon lithium intercalation using different optical spectroscopies and transport
measurements. We shall discuss the modification of lattice dynamics and optical properties upon Li-ion
intercalation with the aim of a better understanding of the intercalation process and establishing some
guidelines for improving the performance of these materials in their most eminent applications, i.e.
namely their use as positive electrode materials in high-energy density lithium batteries.
3.1. Lithium intercalation in WS2
The TMDs are structurally quite similar in that they are all made up of X–M–X layer
sandwiches where X refers to a chalcogen (S, Se, Te) and M is one of the transition metal (Ti, Zr, Nb,
Ta, Mo, etc.). Each atomic plane forms a two-dimensional close packed structure, and the layers may
be stacked in a variety of ways resulting in a large number of polytypes. The group IV metal
compounds mainly adopt the octahedral coordination in which the sandwich layer is made up of
atomic planes arranged in AbC, whereas the group VI metal compounds adopt the trigonal prismatic
coordination with a layer arrangement of the form AbA [53]. The filling of a Van der Waals gap
between slabs introduces geometrical aspects that can be described by considering both local
structural effects, i.e. symmetry of the occupied site, ordering in the Van der Waals gap, and a more
global effect which is the parameter expansion in the direction perpendicular to the slabs. The alkali
metal can occupy either octahedral or trigonal prismatic sites between the slabs of TMD materials.
An octahedron can accommodate higher charges on the anions than a trigonal prism does [57].
One member of TMD materials is WS2, which belongs to the group VIb transition metal, and
crystallises into a layered structure identical to MoS2. Within each layer, there exists trigonal
prismatic coordination between the W and S atoms and the intralayer bonding is primarily covalent
[58,59]. Variations in the stacking sequence of the layers can lead to the formation polytypes [60].
The hexagonal (2H) and rhombohedral (3R) polytypes are commonly found among the materials
with trigonal prismatic coordination. Varieties of other polymorph structures have been described
such as o-WS2 and 2M-WS2 [61,62]. Orthorhombic b-WS2 (1T-structure) has been prepared by
removing Na from NaxWS2 [63].
WS2 is a grey-black indirect-gap semiconductor with an n-type character and an energy gap of
1.35 eV due to electronic transition of the d ! d type. In terms of the two-dimensional WS2 band
structure direct low energy transitions at about 1.9–2.5 eV are characterised by A and B exciton pairs
[64] and its narrow dz2 sub-band is completely filled. The empty conduction band based on dx2 y2 , dxy
orbitals of W is one of the higher energy bands which is expected to be populated by electron donors.
Several early works [65–67] on various forms of WS2 intercalation prompted us to examine more
generally the features of electrochemically LixWS2. In this part, we examine the optical and transport
properties of electrochemically and chemically lithium intercalated 3R-WS2 and 2H-WS2. In
particular, we have studied the ionic motion of Li into the layered structure and determined the
chemical diffusion coefficient of Liþ ions in LixWS2 (0 x 0:6) [68].
The results from electrochemical cells are presented as plots of cell voltage V (discharge and
next charge) versus Li content, x, in LixWS2 as well as plot of the inverse derivative ð@x=@VÞT
(first discharge plotted versus x) as shown in Fig. 18a and b. The electrochemical performance of
intercalation and deintercalation of lithium into WS2 microcrystalline cathode was investigated using
a current density of 0.1 mA cm2. The voltage-composition of discharge curve displays features as
follows. (a) The initial open-circuit voltage (OCV) was 3.0 V. (b) The cell voltage decreases
smoothly from 3.0 to 2.2 V in the range 0 x 0:2 and a break is clearly observed at about x ¼ 0:2.
(c) For x > 0:2, the discharge curve occurs with a pseudo-plateau which has a mid-value of 2.1 V. (d)
Finally, the cell voltage decreases down to 0.5 V at the composition x ¼ 0:6. (e) The charge curve
occurs without stepping and the potential of 3 V is recovered after deintercalation. The sloping
discharge of this electrode in the range 0:0 x 0:2 means that, for these compositions, LixWS2 is
a non-stoichiometric compound. The formation of the pseudo-plateau in the discharge curve is
attributed to the destabilisation of the structure of the 3R-WS2 microcrystalline cathode with the
formation of a 2H–LixWS2 phase as shown thereafter. Cathode formed from microcrystalline WS2
provides a practical power of 0.77 W kg1 and a practical energy density of 140 Wh kg1. The
charge transfer concerning the host reduction during the intercalation process can be described as
xLiþ þ xe þ ½W4þ S2 ! ½Lixþ Wð4xÞþ S2 : (7)
The primary phenomenon associated with the lithium intercalation in WS2 is the donation of nearly
0.6e from the guest to the host layers. The inverse derivative ð@x=@VÞT curve recorded at the first
Fig. 18. (a) Discharge (D) and charge (C) curves of a electrochemical Li||WS2 cell using the 3R-WS2 polymorph at room
temperature. (b) Inverse derivative voltage (@x=@V) of WS2 cell during the first discharge.
discharge (Fig. 18b) shows that the intercalated compound has an incremental capacity comparable
to a double-site energy system. Nagelberg and Worrell [69] have related the chemical potentials of
Liþ in layered transition-metal disulphides to several effects. The major contribution to the chemical
potential, m(Liþ), comes from the interaction energy of intercalated ions and the variation of the
electron chemical potential. Also the lattice expansion energy and the entropy of filling vacant sites
give substantial contributions—the latter especially at the composition extremes. So, the initial rapid
decrease of the cell voltage is mainly to be connected with the energy contribution needed to open
the structure—to separate the layers for a two-dimensional compound. As shown in Fig. 18a, the cell
voltage with microcrystalline cathode is stepped. This should be the evidence of either increasing
repulsive forces between inserted ions or formation of a new phase. During the insertion process, the
alkali metal transfers electrons to the empty conduction band based on dx2 y2 , dxy orbitals of W.
Intercalation with electron donors is expected to populate these higher energy bands providing
metallic properties. According to molecular-orbital calculations [19], with the dz2 band already full
and intercalation increasing the d-band population from d2 to d3, the band structure energy can be
lowered and a trigonal distortion can be achieved. However, for the W layered compounds, the
number of d electrons exceeds d2 on intercalation. The stability of the intercalates of WS2 is clearly
related to the ionisation potentials of the inserted metals and the affinity of the layer compound [67].
Thus, the prismatic coordination is destabilised on addition of lithium ions for x > 0:2. We suggest
that the transition appears as a structural transformation from 3R-WS2 to 2H–LixWS2 due to the
instabilities in the interlayer structure. Later in the discussion, it will be demonstrated that the
interlayer spacing increases with a small expansion upon Liþ insertion.
Previous studies have shown that transition-metal dichalcogenides can undergo a structural
transition when intercalated with lithium [20,21,70]. The driving mechanism for the structural
change has been discussed in terms of a charge transfer from Li to the host and in terms of the
respective energy-band diagrams for the 2H and 1T polytypes. For a high degree of Li insertion
(x > 1), the intercalation compound becomes unstable and decomposes to a mixture of W and Li2S.
This mixture was originally thought to be a third intercalated phase and so was identified as the g-
phase of LixWS2. Thus, upon lithium intercalation, the host lattice of WS2 will tend to adopt the
more stable configuration, which is the 2H-LixWS2 intercalate. This is consistent with the energetic
description of the 3R and 2H structure of the TP coordinated compound.
The equilibrium potential of an intercalation electrode proposed by Armand [71] can be
expressed as the sum of a standard electrode potential, a configurational term accounting for the
distribution over the ideal sites and a term accounting for the interactions between inserted ions such
as
RT x RT xm
V ¼ E0 þ ln g x ; (8)
F xm x F 2
where xm is the maximum lithium uptake and g is a temperature-dependent factor reflecting ion–ion
interactions. From a fit of the discharge curve in the region 0:0 x 0:2, one obtains an interaction
factor g ¼ 12. Typical discharge curves at higher temperatures in the range from 80 to 25 8C show
clearly that the cell voltage decreases with decreasing temperature. This can be attributed to the
change in the chemical potential at low temperature and to an important increase of the internal
resistance of the cell. The increase in internal resistance is mainly due to the decreasing ionic
conductivity in both electrolyte and cathode material and to a large decrease of the lithium
diffusivity in the positive electrode material producing a high voltage drop.
A major part of the structural analysis was performed on samples lithiated chemically using n-
butyl-lithium [72]. All diffraction peaks in the XRD patterns of WS2 could be indexed on the basis of a
rhombohedral unit cell with crystallographic parameters a ¼ 3:162 Å and c ¼ 3 6:1755 Å. A
continuous decrease of the (0 1 5) line (at 2y ¼ 48 ) and the growth of lines at low angle, are observed
in LixWS2, indicating the appearance of the 2H-LixWS2 phase. However, upon Li intercalation, the
(0 0 6) line is slightly shifted towards lower angles leading to a small increase of the interlayer spacing
of 0.6% in the direction perpendicular to the slabs (Fig. 19). In addition, the crystallinity of the
microcrystalline WS2 material does not suffer severely on lithium intercalation. Collongues [73] has
shown that an important expansion of the interlayer distances occurs in systems intercalated with other
alkali metals. An expansion of 4.03, 4.84 and 5.53 Å has been observed in KxWS2, RbxWS2 and
CsxWS2, respectively. The increasing interlayer spacing seems to be related with the ion size. This
effect can justify the weak charge transfer occurring in lithium intercalated WS2.
Transport properties of LixWS2 samples have been studied as a function of temperature and
lithium concentration. Results of temperature dependence of the electrical conductivity of pure 3H-
WS2 and LixWS2 after immersion during 20 h in n-butyl-lithium show that the mechanism of
conduction does not change drastically upon Li intercalation but an increase of conductivity is
observed. LixWS2 remains a semiconducting material. The lithium-ion intercalation occurs with a
decrease of the mobility of charge carriers. This should indicate that the electrons occupy states in
the narrow conduction band. From the Arrhenius plots, values of the activation energy is estimated to
be Ea ¼ 0:336 eV in pure material and Ea ¼ 0:152 eV in Li-intercalated WS2.
Kinetics, i.e. chemical diffusion coefficient and enhancement factor, of lithium intercalated
LixWS2 samples have been estimated as function of the degree of lithium intercalation in the domain
Fig. 19. Variation of the crystallographic d-spacing parameter of LixWS2 as a function of the intercalation time in n-butyl-
lithium.
of the galvanic cell reversibility [68]. Chemical diffusion coefficient (D) and enhancement factor (W)
of lithium have been determined by the GITT method from the variation of the cell voltage versus
time during the relaxation period following a long discharge or charge. D was calculated on the
assumption of uniform Liþ distribution at any composition. Fig. 20a and b show the variation in the
Fig. 20. (a) The chemical diffusion coefficient (D) of Liþ ions as a function of the degree of intercalation in LixWS2.
Dashed lines are the curves fitted using Eq. (10). (b) The thermodynamic factor (W) as a function of the degree of
intercalation in LixWS2.
chemical diffusion coefficient of Liþ-ions as well as in the enhancement factor as a function of the
degree of intercalation in LixWS2. The values of the chemical diffusion coefficients of lithium are
near 108 cm2 s1 in the range 0 < x < 0:25. This can be compared with other data reported for
other TMD compounds [74]. From the structural features it may be inferred that the crystal geometry
of 3R-WS2 allows fast Liþ diffusion.
If we consider that the diffusion of lithium ions into the host lattice has two limiting factors, the
ion hopping probability and the degree of occupancy of the sites, then the chemical diffusion
coefficient can be expressed as:
D ¼ kW½ðxmax xÞ2 þ bxðxmax xÞ; (9)

where k is the factor related with the probability for ion hopping and b represents an interaction
factor. For intercalated compound, in which different types of sites are available, Eq. (9) can be
generalised as
kW
D¼ 3
½ðx xmin Þðxmax xÞ2 þ bðx xmin Þ2 ðxmax xÞ: (10)
ðxmax xmin Þ
Full line in Fig. 20a is the curve fitted using Eq. (10) giving parameters k ¼ 2:6 106 and b ¼ 0:95
in the compositional range 0:00 < x < 0:25. The chemical diffusion coefficient drops down to
1011 cm2 s1 in the range x > 0:25 and is rather independent on composition, which is due to the
nature of the empty sites in the host structure. The initial rise in D gives evidence of the need to open
the crystalline network to allow fast Liþ-ion diffusion. The mobility drop at high intercalation
concentration is indicative of considerable interfacial blocking in microcrystalline material. This
could be also pointing to a difficult site occupancy due to charge repulsion.
Next, we present the optical properties of 2H-WS2 crystal, often referred to as the a-polytype,
grown directly from vapour phase using iodine as transport agent (P63/mmc space group with well-
defined lattice parameters a ¼ 0:316 nm, c ¼ 2 0:618 nm and c=a ¼ 1:955). Fig. 21a and b show
the absorbance spectra recorded at 5 K of pristine WS2 and Li-intercalated WS2 samples in the
energy range 1.4–2.5 eV. The results for the Li-intercalated crystal were obtained for a sample
containing a small amount of lithium (x 0:1). The transmission spectrum of pure WS2 exhibits an
Fig. 21. The absorption spectra recorded at 5 K of (a) pure WS2 and (b) Li-intercalated WS2 single crystal containing a
small amount of lithium.
absorption edge in the vicinity of 1.9 eV with a sharp band located at 1.93 eV. Essentially identical
absorption edge curve was obtained for Li-intercalated WS2 [75]. The spectrum exhibits an
absorption edge at about 1.7 eV with a broad band located at 1.88 eV. Our data give conclusive
evidence that modification occurs in the absorption spectrum on the optical band edge of WS2. We
found a red shift of the absorption edge as a result of lithium intercalation. Optical properties of WS2
can be interpreted on the inter-band transition model. In the energy range 1.5–2.5 eV, the general
features of the transmission spectrum is dominated by the presence of sharp absorption peaks having
an excitonic character. The nature and the character of these excitonic bands for 2H-WS2 and 3R-
WS2 have been widely discussed [76]. They originate from the inter-band d ! d transitions. The
fundamental absorption edges have been investigated by Baglio et al. [77]; direct and indirect band-
gap of WS2 were determined at 1.78 and 1.34 eV, respectively.
Upon lithium intercalation, there are two important effects on the optical properties of WS2, as
shown in Fig. 21. First, one observes the vanishing of the exciton peak (located at Eex on the curves)
due to free-carrier screening. In this process, the free carriers arise from band filling above the full
dz2 band. For LixWS2, the donated electrons should go into the empty conduction band based on the
dx2 y2 , dxy orbitals. The half-width-at-half-maximum (HWHM) of the exciton of pure WS2 at 5 K is
about 20 m eV, whereas one observes a HWHM of 55 meV for LixWS2. The charge transfer
concerning the cation reduction of the host lattice during the intercalation process can be described
as in Eq. (7). The primary phenomenon associated with the lithium intercalation in WS2 is the
donation of x electrons from the guest to the host matrix. The second effect is that the inter-band
absorption edge around 1.9 eV shifts to lower energies with Li insertion (Fig. 21). This result can be
understood in terms of a destabilization of the electronic band structure. The shift of the fundamental
absorption edge toward lower energies can be attributed to an interaction between the conduction
band made up of s and p orbital and the 2s orbital of lithium. Thus, the net effect is a decrease of the
optical band-gap of about 0.22 eV. Optical effects in Li-intercalated WS2 can be satisfactory
interpreted in terms of electron donation but the rigid-band model is essentially inadequate in such a
compound. One may note that, in the studied energy range (1.4–2.5 eV), intercalation of Li ions did
not give rise to extra absorption bands.
As WS2 is an indirect band-gap semiconductor, the variation of the optical absorption near the
fundamental absorption edge has allowed us to determine the energy gap. The classical theory of
inter-band absorption [78] shows that for semiconductor the absorption coefficient a varies with the
photon energy, E, according to the expression
aE ¼ AðE Eg Þn ; (11)
where A is a constant and Eg is the energy threshold. In Eq. (11), the exponent characterises the
transition mechanism and takes the values n ¼ 2 for indirect gap and n ¼ 1=2 for direct gap. The usual
method for the determination of the values of Eg involves plotting a graph of ½ða a0 ÞE1=n versus
photon energy. This curve gives a straight line and the intercept at ðaEÞ ¼ 0 gives the band-gap energy
with an accuracy of
0.02 eV. Results on band-gap energy of pure WS2 (shown in Fig. 22) are in good
agreement with values reported in [59]. WS2 exhibits a direct gap of 1.81 eV and an indirect gap of
1.39 eV. A red shift is observed in the band-gap energy of the Li-intercalated sample, as shown in
d
Fig. 22a and b. With EgðintÞ ¼ 1:58 eV, this shift is about 0.23 eV for Li0.1WS2. Similarly, optical
measurements on MoS2 have shown that the absorption spectrum changes drastically upon
intercalation with Na [55]. The sharp and prominent excitonic peaks as well as the absorption edge
of pure MoS2 are washed out. This indicates that the alkali metal intercalates as an ion, transferring
some of its charge to the crystal layer. This leads to partial filling of the next unoccupied band.
Fig. 22. Plot of ½Eða a0 Þ2 vs. photon energy for (a) pure WS2 and (b) Li-intercalated WS2 single crystal showing the
position of the direct gap energy at 300 K.
To explain the experimental data at relatively low Li concentration (x 0:1), we assume that
the conduction band WS2 is made up of metal s and p orbital and a non-bonding band made mostly
of metal by dx2 y2, dxy orbital (Fig. 23) as suggested by William and Shephered [79]. Modifications
of the optical properties are attributed to the charge transfer of electrons from intercalant species to
the host and to the destabilisation of the electronic band structure. A consequence of the band
structure may be the interaction between the 2s state of Li localised in the Van der Waals gap with
the (s, p) group of the host.
3.2. Lithium intercalation in MPS3 phases
The MPS3 compounds have layered structures formed by compact stacking of ABC.
Alternatively, the space between chalcogen atom sheets is either empty—forming a Van der Waals
gap—or filled with metallic and phosphorous ions. This family can be viewed as salts constituted by
M2þ cations and (P2S6)4 anions. An elementary cell of D3d symmetry corresponds to this
arrangement. The MPS3 materials are anti-ferromagnetic with strong anisotropy of the magnetic
Fig. 23. Electronic band structure of trigonal prismatic dichalcogenide (a) pure WS2 (as suggested by William and
Shephered [41]) and (b) Li0.1WS2 (this work).
Fig. 24. Raman spectra of (a) pure NiPS3, (b) Li0.85NiPS3, and (c) Li1.16NiPS3.
properties due to their layered structure. All observations published up to now converge to the
conclusion that the interlayer separation does not change upon Li intercalation. In addition, there is
no evidence for in-plane modification of the host material. Raman scattering and infrared absorption
spectra of lithium-intercalated NiPS3 crystal have been investigated to better understand the nature
of the intercalation process in this system. We attempted to explain the experimental results of
vibrational spectroscopy through the use of lattice dynamical models such as a simplified linear-
chain model and a simple local-symmetry approach [80].
3.2.1. Lithium intercalation in NiPS3

The NiPS3 samples and the lithiated phases LixNiPS3 were prepared by immersing
microcrystalline powders in dilute hexane solutions of n-butyl-lithium for 1–2 weeks at room
temperature [80]. Fig. 24 shows the Raman scattering spectra of pure and Li-intercalated NiPS3
microcrystalline compounds recorded at room temperature. The FTIR spectra of NiPS3 and its
lithiated phases recorded at room temperature in the range 50–600 cm1 are shown in Fig. 25. In
order to investigate the effects of alkali–metal intercalation in NiPS3 samples, first we briefly analyse
the lattice dynamics of pure material. The NiPS3 compound is a layer-type semiconductor having the
CdCl2-type structure (compact stacking of sulphur layers of ABC type), in which one-third of the
transition metals are substituted by P2 pairs within the layers leading to the general formula
3½SNi2=3 ðP2 Þ1=3 S ¼ Ni2 P2 S6 . There is thus a succession of sandwiches composed of two sulphur
layers between which the P2 pairs and metal ions are located and the sandwiches are weakly bonded
together through Van der Waals S–S interactions. The crystal structure belongs to the C2/m (C32h )
space group with two NiPS3 entities per primitive cell. Vibrational spectra of MPS3 layered
compounds have been reconsidered by Sourisseau et al. [81]. A complete assignment of the observed
spectra was proposed and a normal coordinate analysis of the [P2S46] anion has been carried out. As
observed in Figs. 24 and 25, in the high-frequency range (200–600 cm1) all the bands are assigned in
terms of motions of PS3 groups; one notes that the n(P–P) bond stretching mode gives rise only to an
infrared band at ca. 438 cm1, which is in agreement with predictions. The few low frequency bands
observed at n ¼ 150 cm1 are thus attributed to translational motions of Ni2þ metal ions [80].
Fig. 25. FTIR spectra of (a) pure NiPS3, (b) Li0.85NiPS3, and (c) Li1.16NiPS3. No free-electron absorption is observed in
spectrum of intercalated samples.
The effect of lithium intercalation on the Raman scattering spectra of LixNiPS3 phases
(x ¼ 0:54, 0.85, 1.15) is three-fold. First, the normal modes of the host material vanish as the amount
of lithium increases, however, no significant frequency shift occurs; second, a new weak band is
observed which is located at 313 cm1; and third, the background increases as the wavenumber shift
increases, probably owing to the appearance of luminescence centres. These last features are
observed being sensitive to the sample temperature (Fig. 24). Infrared spectra recorded after
intercalation (Fig. 25) show the similar tendency than that reported previously [54]. Two new broad
bands are observed at 310–370 cm1.
In the LixNiPS3 phases, four types of sites are available for the Liþ ions inside the Van der
Waals gaps: 2d octahedral (symmetry C2h), 4h octahedral (symmetry C2), 4i tetrahedral (symmetry
Cs) and 8j tetrahedral (symmetry C1). Moreover, it has been shown that these ions can diffuse
without any expansion of the c parameter and cannot be localised on the 4i sites because of strong
electrostatic repulsion [82]. In infrared study of several chemically prepared lithium intercalates
LixFePS3 and LixNiPS3 (0:5 x 1:45), Barj et al. [54] showed that these compounds are
characterised essentially by a new broad absorption band at 336 and 310–330 cm1, respectively.
These bands increased in intensity with increase in lithium content and were thus assigned to
stretching vibrations of LiS6 units. This n3 mode has a F1u symmetry characteristic of Liþ in
octahedral sulphur environment, proving the nature of the intercalation sites. One of the most
interesting findings of the present investigation is the determination of modes highly sensitive to the
degree of intercalation. The Raman peak at 255 cm1 in the pure phase shifts to 236 cm1 in
Li0.85NiPS3. The two sulphur planes move rigidly in anti-phase along the z direction in this mode.
Also, it is worth mentioning that some changes are observed in the IR-active modes of NiPS3,
especially in the low-frequency region. The deformation modes of the PS3 groups are split into
doublets at 193–196 cm1 suggesting that distortions within the layers have occurred giving rise to
various P–S distances in the PS3 groups. The band attributed to the translational motions of Ni2þ
ions is shifted from 136 to 142 cm1 indicating strengthened Ni–S interactions.
The general trend in the spectra of lithiated phase is the appearance of a new mode at ca. 310–
320 cm1 and a decrease of the band intensities of the normal modes of NiPS3. Moreover, as far as
experiments were accurate, no significant frequency shift of the normal modes of the pristine
structure has been detected. This indicates that the elementary unit cell remains unchanged upon
intercalation and that no volume expansion takes place as suggested by structural analysis [82].
Considering the Raman spectra of LixNiPS3, we observed a vanishing of the deformation modes of
the PS3 group in the host lattice, of the n(P–P) mode, and of the low frequency translational motion
of Ni2þ ions. It is worth noting that the Raman-active mode at 436 cm1 disappears with a small
amount of intercalated Li ions.
The temperature dependence of the Raman spectra of LixNiPS3 has been investigated for
clarifying the presence of a spectrum background. The luminescence studies of NiPS3 at different
temperatures show that a broad band is observed at 2.412 eV at 5 K and shifts to 2.350 eV at 77 K.
Piacentini et al. [83] have examined the electronic transitions and XPS features of Ni2P2S6 in the
energy range 1–25 eV. The optical spectrum exhibits a weak absorption peak at 1.6 eV assigned to
the localised d ! d transition. The first strong band in the Ni2P2S6 spectrum occurs at 2.2 eV. This
band is assigned to a d ! V charge-transfer transition from sulphur valence band to the empty Ni 3d
states. Photoluminescence studies of MPS3 compounds show that the results are consistent with the
electronic band structure model, which predicts highly localised states associated with the transition-
metal d-orbitals and delocalised band-like states derived from sulphur and phosphorus orbitals.
When a small amount of lithium is intercalated onto NiPS3, the luminescence band is changed
substantially. At 5 K, the position of the band observed at 2.412 eV in pristine material shifts toward
the low-energy side in Li0.85NiPS3. This shift can be explained in terms of changes in interlayer
interactions upon lithiation. Similar to the case of LixTiS2, the complete ionisation of lithium upon
intercalation was determined through NMR of 7Li in LixNiPS3 [84], implying electron donation to
the host. ESR measurements [85] have shown that free carriers appear in the intercalated phase even
for a low lithium content. Thus, neither Raman scattering nor IR measurements exhibit the signature
of a charge transfer with, for instance, the appearance of a Drude-type contribution. This may be due
to the particular electronic structure around the Fermi level. Electronic bands arising from the
overlap of metal 3d and sulphur 3p orbitals very likely act as acceptor levels in the lithium
intercalation process.
The vibrational features of LixNiPS3 can be satisfactory explained using a simple local-
symmetry approach [80]. Considering a simple local-symmetry model, the octahedral LiS6 entities
have a Oh symmetry. Of the three stretches, n1(A1g) and n2(Eg) are Raman-active, whereas n3(F1u) is
infrared-active. In general, the order of the stretching frequencies is n1 > n3 @ n2 or n1 < n3 @ n2,
depending on the compound. From the above results it appears that the lithium intercalated phases,
whatever the reaction mechanism, give rise to a new infrared intercalation mode, which is derived in
part from the stretching vibration (n3, F1u) in LiS6 entities. This allows us to distinguish the
accommodation of Liþ ions in octahedral sites. As a matter of fact, LiS6 surroundings with Li–S
distances in the range of ca. 2.57–2.70 Å have been evidenced in Li4P2S6, Li2FeS2 and Li3HfS3
compounds giving rise to n3-infrared modes located at 335, 320–250 and 303 cm1, respectively.
These energies can be well compared to the 313 cm1 infrared band in LixNiPS3. All these results
suggest that similar Li–S distances and bonds are occurring. On the other hand, the octahedral
lithium accommodation was encountered in the wavenumber range 320–350 cm1. All these results
suggest that, whatever the intercalated phase in which it is observed, the LiS6 surrounding is always
associated with similar Li–S bonds. Also, considering all Liþ ions accommodated with an octahedral
LiS6 environment, the F1u modes are normally split into (Ag 2Bg ) Raman-active components and
(Au 2Bu ) IR-active component through the scheme
Oh C2 ðsiteÞ C32h ðcrystalÞ

F1u ! A 2B ! Ag 2Bg 1Au 2Bu : (12)
Therefore, the Au 2Bu IR-mode species are intense, while Raman modes having the Ag 2Bg
symmetry appear very weak. We have recorded a Raman-active intercalation mode which is very
weak and hardly detected at near 313 cm1 in LixNiPS3. In the partly amorphous intercalated
LixNiPS3 phases the F1u mode became Raman-active but is was difficult to evidence experimentally.
The modes observed at 313–320 cm1 correspond to the in-plane translations of lithium ions, while
the broad IR band at 363 cm1 is associated with the translation of lithium ions along the direction
perpendicular to the layers. They are not degenerate because the sulphur cage has trigonal symmetry.
A theoretical estimate of the phonon dispersion curves of LixNiPS3 has been done by
Bernasconi et al. [86]. They have used a force-constant model generated by a set of short-range two-
body potentials. The parameters of the dynamical model have been determined by fitting the IR-
absorption spectra reported by Barj et al. [54]. Accordingly, using a radial force constant
Fr ðLiSÞ ¼ 22 N m1, the whole phonon dispersion curves in LixNiPS3 (0:5 x 1:5) compounds
have been calculated. The Li–S force constant can be also calculated by using a simplified linear-
chain model. Considering one displacement direction, in this picture the chain axis corresponds to
the direction perpendicular to the layered plane. We assumed that there is a molecule of mass M in
the host unit cell which interacts with the nearest-neighbour molecules by a weak interlayer
coupling. The host atoms are positioned equidistantly and the intercalant of mass m are located on
the octahedral holes. At the centre (q ¼ 0) and the boundary (q ¼ p=2c) of the Brillouin zone, the
intercalation mode frequencies are given by [80]

2M þ m 1=2
oin ð0Þ ¼ k ; (13a)
Mm
p 1=2
2k
oin ; (13b)
2c m
where k is the Li–S force constant. Using the frequency oin ð0Þ ¼ 316 cm1, the average value
obtained from our Raman and IR measurements, one obtains the shear force constant
ks ¼ 19:3 N m1 which is very close to the value 22 N m1 obtained by Bernasconi et al. [86]. If
we consider that the IR band at 363 cm1 is associated with the translation of lithium ions along the
direction perpendicular to the layers, thus the compressional force constant is kc ¼ 25:3 N m1.
3.2.2. Lithium intercalation in FePS3

FePS3 has a layered structure isomorphic with NiPS3. Infrared studies of Li intercalated FePS3
phases have been early reported by Barj et al. [54]. FTIR spectra of FePS3 and lithium intercalate
Li1.1FePS3 are shown in Fig. 26. In the spectral range 50–700 cm1, all the fundamental modes of
FePS3 are expected. The FTIR spectrum of the intercalated compound Li1.1FePS3 does not show
tremendous changes with respect to that of the host lattice, but, mainly, a new band is clearly
observed at 338 cm1. The intensity of this band increases markedly with the lithium content. Thus,
this band is attributed to the Li ion vibrations. Its frequency is well compared with that reported at
335 cm1 for an intense infrared absorption in Li2PS3 [89]. It is a fact that in crystal structure of
Li2PS3, lithium on D3 sites is surrounded by six sulphur atoms defining distorted octahedra with Li–
S6 distances equal to 2.63 Å. Similar Li–S6 distances (2.58 Å) have been reported in LiTiS2 [90]. In
agreement with NMR measurements [91], it is concluded that Li ions are localised in octahedral 2d
sites in Li0.5FePS3 and are simultaneously occupying the 2d and 4h sites in LixFePS3 when the
concentration range is 0:5 < x < 1:5. Considering now the IR bands characteristic of the host
matrix, quite negligible frequency and intensity changes are observed for the stretching and
Fig. 26. FTIR spectra of (a) pure FePS3 and (b) LixFePS3. The arrow indicates the new mode due to intercalated lithium
ions.
deformation modes of the PS3 groups and for the low frequency translational motions of Fe3þ ions.
Also, the infrared band assigned to the n(P–P) mode appears at constant wavenumber 445 cm1 for
all the intercalates.
The low temperature (5 K) Raman spectrum of LixFePS3 shows drastic changes from that
obtained at room temperature. Three new bands appear for T T N at 88, 95 and 109 cm1. This
difference can be accounted for by the Brillouin zone folding effect due to formation of a magnetic
superstructure [87]. This superstructure is formed from the ferromagnetic linear chains which are
coupled to each other anti-ferromagnetically in a layer. The folded acoustic branches are Raman-
active modes below the Néel temperature, TN, corresponding to the paramagnetic/anti-ferromagnetic
transition temperature. FePS3 is well known as a two-dimensional Ising-type anti-ferromagnet with
the magnetic ions Fe2þ arranged in a honeycomb lattice and it undergoes an anti-ferromagnetic
phase at 118 K. In LixFePS3 obtained by n-butyl-lithium solution, a decrease of the Néel temperature
from 118 to about 56 K has been observed. This phenomenon has been explained in terms of indirect
exchange due to the large increase of the free carriers. Such a phenomenon has been also detected in
FePS3 intercalated by cobalticenium cations, where the ordered magnetic state occurs at lower
temperature (70 K) in the intercalates than in the pure material. Two possible origins have been
given: (i) an anti-symmetric exchange could be due to turn neighbour spins at right angles and (ii) a
local asymmetric anisotropy was attributed to intralayer distortion [88,89].
3.3. Lithium intercalation in SnS2
Some structural properties and thermodynamics of lithiated phases LixSnS2 have been reported
[90–94]. Chemical lithiation of SnS2 has been carried out by treating the host material with lithium
solutions in liquid ammonia [93]. Chemical lithiation has also been carried out in aqueous media
[92]. Recently, Morales et al. [94] have studied the Li intercalation in 2H-SnS2 using a chemical
process in n-butyl-lithium and an electrochemical titration technique. A new phase showing a staging
effect has been identified. Thus, the nature of the lithium site occupancy is still controversial in SnX2
compounds, vibrational spectroscopy studies can give some information about the position of the
intercalant species in the host structure [94].
Infrared absorption spectra of lithium-intercalated SnSySe2y (0 y 2) layered compounds
have been studied to better understand the nature of the intercalation process in this system. An
Fig. 27. FTIR absorption spectra of (a) pure SnS2 and (b) LixSnS2.
attempt to explain the experimental results of vibrational spectroscopy in LixSnSySe2y was made
through the use of lattice dynamical models [94]. The individual tightly bonded layer in SnS2 is a
S–Sn–S sandwich, in which the Sn atoms are octahedrally coordinated to sulphur atoms. The three-
atom basis of the elementary unit cell gives rise to nine vibrational modes; three doubly degenerate E
modes in which the atomic motions are parallel to the layer planes, and three non-degenerate A
modes in which the atomic motions are perpendicular to the layers. At the centre of the Brillouin
zone, the irreducible representations are two Raman-active modes (A1g Eg ,) and two IR-active
modes (A2u Eu ). The two IR-active modes are clearly observed in the spectrum (Fig. 27a). It is
worth noting that: (1) the frequencies of the A2u and Eu modes are not very far from each other, (2)
the Eu mode gives a very broad band with FWHM around 140 cm1. (3) the strength of the A2u mode
is roughly 20 times less than that of the Eu mode, which is a consequence of the preferential
orientation of the sample along the layer plane.
The effect of lithium intercalation on the vibrational spectra of SnS2 compounds is two-fold
(Fig. 27b). The normal modes of the host material remain almost unchanged; only a slight frequency
shift (few wavenumbers) of the lattice modes of the SnS2 occurs and new broad bands are observed
which are located in the high-frequency region at around 400 cm1. The general trend in the spectra
of lithiated phases is an increase of the frequency of the Eu mode, i.e. about 3 cm1 for the SnS2-like.
This frequency shift indicates an increase in the restoring force for this type of vibration, which is
probably due to an increase in the intralayer distance, a consequence of the tin reduction which
creates a negative charge in the intralayer space upon lithium intercalation. For the lithiated
products, the phase that is isostructural with SnS2 shows minor changes in the unit cell parameters
with the lithium content. On the other hand, the small decrease of the A2u mode frequency, i.e. about
5 cm1 for the LiSnS2, is in good agreement with the results of Morales et al. [94]. They observed a
decrease of the basal spacing, from c ¼ 5:884 Å in SnS2 to c ¼ 5:864 Å in Li1.98SnS2, which is
attributed to the strong attractions due to the presence of lithium cations between negatively charged
layers.
We attempt to explain the experimental results of vibrational spectroscopy in LixSnS2 through
the use of lattice dynamical models such as a simple local-symmetry approach and a simplified
linear-chain model. From the above results it appears that the lithium intercalated phases, whatever
the reaction mechanism, give rise to intense infrared new intercalation modes which are derived in
part from the stretching vibrations (n3) in LiS4 or LiS6 entities. This allows us to distinguish the
Fig. 28. The linear-chain model of lattice dynamics in the intercalated layered materials, where M and m are the masses of
the SnS2 molecule and of the lithium atom, respectively.
accommodation of Liþ ions in tetrahedral or octahedral sites. As a matter of fact, LiS4 surroundings
with Li–S distances ranging from 2.35 to 2.44 Å have been evidenced in Li2S, LixFeS2, LixMoS3 and
LixNbS3 compounds giving rise to n3-infrared modes at 465, 390–420, 390–450 and 430–460 cm1,
respectively. These wavenumbers can be well compared to the 444–420 cm1 infrared band in
LixSnS2. All these results suggest that similar Li–S distances and bonding are occurring. On the other
hand, the octahedral lithium accommodation was encountered in the wavenumber range 320–
350 cm1 [95].
Considering a simple local-symmetry model, the tetrahedral LiS4 entities have a Td symmetry.
The four normal modes of vibration of a tetrahedral XY4 molecule are all Raman-active, whereas
only n3 and n4 (with n4 < n3 ) are triply degenerate infrared-active modes. In the solid state n3 and n4
may split into two or three bands because of the site effect. The very strong peak in the region 444–
420 cm1 observed in the lithiated products is assigned to the n3(F2) mode and the weak peak
located at low wavenumbers (in the range 100–150 cm1) is attributed to the n4(F2) mode in the Td
symmetry. The temperature dependence of the FTIR spectra displayed that the n3(F2) mode is well
resolved at low temperature with a site effect for the LiSnS2 sample.
A simplified linear-chain model has been applied to SnS2 intercalated with lithium as shown by
the one-dimensional representation in Fig. 28. Considering one displacement direction, either
longitudinal or transverse, in this picture the chain direction corresponds to the direction
perpendicular to the layered plane [91]. We assumed that there is a diatomic molecule in the host
unit cell which interacts with the nearest-neighbour molecules by a weak interlayer coupling.
Moreover, we inserted a tetrahedrally accommodated intercalant between the layers and a repulsive
interaction with a force constant, k2, exists between two nearest-neighbouring intercalants. The host
atoms are positioned equidistantly and the intercalants are located on the tetrahedral holes. The
equations of motion are written as follows [95]:
2
d x3n
M ¼ Kðx3nþ2 x3n Þ þ k1 ðx3nþ1 x3n Þ k1 ðx3n x3n1 Þ k1 ðx3n x3n2 Þ; (14a)
dt2
2
d x3nþ1
m ¼ Kðx3nþ3 x3nþ1 Þ þ k2 ðx3nþ2 x3nþ1 Þ k1 ðx3nþ1 x3n Þ; (14b)
dt2
2
d x3n1
m ¼ k1 ðx3n x3nþ1 Þ k2 ðx3n1 x3n2 Þ k1 ðx3n1 x3n3 Þ; (14c)
dt2
where M and m are the masses of the host atom and the intercalant, respectively, xi are the
displacements of the atoms. We look for solutions to Eqs. (14a–c) in the form of a travelling wave
given by xi ¼ ai expfiðot þ qci Þg, where ci is the spacing between the host atoms, ai the amplitudes
of displacement of the host atoms and intercalant, and q is the wave vector. When q ¼ 0 (at the
Brillouin-zone centre), we obtain the three frequencies given by
oac ¼ 0; (14d)

K þ k1 1=2
oop ¼ ; (14e)
m

K þ k1 þ 2k2 1=2
oin ¼ ; (14f)
m
where m is the reduced mass. Of course, the frequency oac corresponds to that of the acoustic mode.
The frequency of the optical (intermolecular) mode oop shifts to a slightly higher frequency. The
new mode oin appears in the intercalated materials (Eq. (14f)). Therefore, we can easily guess that
the frequency of the intercalation mode is almost independent of the concentration of intercalant.
This result is consistent with the experimental data in many intercalated layered compounds [96].
Using the frequency oin ¼ 444 cm1 measured by infrared absorption, one obtains the interlayer
force constant K þ k1 ¼ 75 N m1 and the Li–Li repulsive force constant k2 ¼ 7:48 N m1.
Furthermore, it can be remarked that: (a) this model falls if the tetrahedral sites are not equally
occupied and (b) the frequency of the intercalation mode is always higher than oop even if k2 ! 0.
3.4. Lithium intercalation in V2O5
Layered oxides such as V2O5 and MoO3 are believed to have a better electrode performance
than chalcogenides do because of their high oxidation state. MoO3 crystallises with an orthorhombic
structure and can accommodate 1.5 Li per molybdenum atom, while V2O5 exhibits a high degree of
Li intercalation around xðLiÞ ¼ 3 per host molecule. Vanadium pentoxide is an interesting material
for technological applications because of its unique features such as high electrochemical activity,
high energy density and high stability towards lithium insertion [97,98]. V2O5, which is a functional
inorganic material, can also be used in gas sensors, electrochromic cells and as an oxidation catalyst
[99].
Structurally, the V2O5 host is unusual as it falls in a category somewhere between a layered
compound and a crystallographic shear structure. The structure of V2O5 was determined by
Bachmann et al. [100]. V2O5 crystallises in the orthorhombic system, Pmnm space group (D13 2h ). The
structure is built up from distorted trigonal bi-pyramidal coordination polyhedra of oxygen around
vanadium cations, which share edges to form (V2O4)n zigzag double chains along the (0 0 1)
direction and are cross-linked along the (1 0 0) axis through shared corners forming sheets in the xz
plane. The elementary structural unit shown in Fig. 29 designates the various oxygen atoms. The
V–Och bonds form puckered chains in the b direction; the chains are linked to each other in the a
direction by Obr bridging atoms, and the layers formed are connected by weak V–Ov (vanadyl) bonds
in the c direction. The short and intermediate length bonds are mainly covalent, whereas the long
vanadyl bond is mainly ionic [100].
There have been several studies of the phases formed on ambient temperature insertion of Li
into V2O5 [100–106] employing electrochemical and chemical synthesis. Fig. 30 reports these
phases with their domain of stoichiometry [100,101]. In the electrochemical studies of Li||V2O5
cells, it was observed that insertion of up to 1 Li per formula unit proceeds in several distinct steps,
each accounting for half the charge. The insertion reactions are however not reversible in the entire
Fig. 29. Coordination of vanadium in the V2O5 lattice. The oxygen atoms are represented as participating in bridges (br),
chains (ch) and vanadyl (v) bonds.
interval 0 < x < 3, as the a- and the g-phases are converted into a semi-amorphous o-phase when
more than 1.8 Li per formula unit is inserted [106].
The electrical properties of LixV2O5 were investigated as a function of the degree of
intercalation (Fig. 31) and the temperature in the range 200–450 K. A general condition for
semiconducting behaviour is that the transition-metal ion could exhibit several valence states such as
V5þ and V4þ in V2O5, so that electron hopping takes place between these two levels. If
C ¼ V4þ =Vtot is the ratio of the concentration of low valence ions with the total concentration of
transition-metal ions and A is the average hopping distance, one can estimate the number of charge
carriers n ¼ CA3 and the thermally activated electrical conductivity is given by the classical
Arrhenius law. Knowing the V4þ concentration and the average V–V distance, we can estimate the
room temperature electrical parameters of LixV2O5 using C 0:02 and A ¼ 0:384 nm. The carrier
mobility is quite low; it lies between 107 and 102 cm2 V1 s1. All experimental curves plotted as
lnðsTÞ versus 1/T are almost linear in the high-temperature range whereas the curves exhibit non-
linear features in the low temperature range. Such a behaviour is typical of a conduction mechanism
driven by a small-polaron hopping which has been widely reported in transition-metal oxides.
Fig. 30. Phase diagram of LixV2O5 in the range 0 < x < 2 (after Cocciantelli et al. [100]).
Fig. 31. The electrical conductivity of LixV2O5 in the range 0 < x < 1.
Electrical conductivity measurements are usually discussed in terms of Mott’s theory [107]

n0 e2 2aA W
s ¼ Cð1 CÞ exp ; (15)
AkB T kB T
where n0 is a phonon frequency, and a is the rate of the wave function decay. As a result, the thermal
activation energy can be expressed as
Wd
W ¼ Wh þ ; (16)
2
where Wh corresponds to the small-polaron formation and Wd to the disorder energy arising from the
energy difference of neighbours between two hopping sites. The separation of the activation energy
into a polaron term and a disorder term is a rather difficult problem which can be partly solved using
data taken at low temperatures.
Results in Fig. 31 show that evolution of the dc conductivity of LixV2O5 can be correlated to the
phase diagram (Fig. 30). We observe the following aspects: (a) a small increase of s in the range
0:0 < x < 0:2, (b) a large increase of s around x ¼ 0:3 corresponding to the appearance of the e-
Li0.3V2O5 phase. These two types of conduction can be related with the formation of lithiated
domains in the V2O5 network. For x > 0:2, enough lithiated domains exist to create continuous paths
for carriers even in the two-phase (e þ d) region.
The conductivity measurements s(T) of V2O5 show that the conduction is thermally activated
with W ¼ 0:21 eV, while a slight decrease of the activation energy W is observed for LixV2O5. A
value of W ¼ 0:17 eV is obtained for Li0.51V2O5 that could confirm the model of lithiated domains.
Each inserted lithium ion gives rise to the formation of a V4þ ion, which participates to the
conduction mechanism in V2O5 as verified by ESR measurements [108]. ESR spectra of V4þ ions
which form a vanadyl bond are therefore located in a typical region of the ESR spectrum, in the
range from about 300–400 mTesla in these experiments. It is clearly observed that the intensity of
the ESR spectrum of a lithiated V2O5 crystal increases upon the reduction of the vanadium ions. The
large linewidth observed in the spectra of LixV2O5 may arise from several effects: (i) a g parameter
distribution arising from the structural disorder, (ii) dipolar interactions between magnetic ions, and
(iii) a non-statistical distribution of magnetic V4þ ions in the samples.
Fig. 32. FTIR absorption spectra of lithiated V2O5 films and their LixV2O5 intercalates in the range 0:00 x 0:93.
Films were grown by the electron-beam method on silicon substrate. Lithium was intercalated galvanostatically at current
density 10 mA cm2.
Fig. 32 shows the FTIR spectra in the range 150–1200 cm1 for lithiated V2O5 films grown by
the electron-beam method on silicon substrate [109]. The as-deposited film yields three main
absorption bands at 1018, 790 and 450 cm1. We have found that the IR spectrum of LixV2O5 films
show significant changes in the range 0 x 1. Upon lithium intercalation, these high-frequency
spectral structures move towards lower wavenumber (lattice softening), while the bands at about
460 cm1 are displaced towards higher wavenumbers (lattice stiffening). The lithiated films have an
additional absorption band at 250 cm1 whose intensity increases with increased lithiation. The
infrared absorption data of V2O5 can be understood to a fair degree of detailed analyses by Abello
et al. [110], Clauws et al. [111] and Manukhin et al. [112]. The crystalline structure of V2O5 belongs
to the symmetry D13 2h. The unit cell can be written V4O10 and has 39 optical zone-centre modes. The
IR-active modes B1u, B2u and B3u are polarised along the c, b and a directions, respectively.
Absorption around 1000 cm1 is due to V–Ov bonds, absorption at 900–700 is due to V–Obr bonds,
and bands at 800–400 cm1 are due to V–Och bonds [112].
The infrared data display the displacement of the broad absorption band at 460 cm1, which has
the mix character of B3u modes due to stretching and deformation along the a-axis and of B1u modes
due to V–Obr bending along the c-axis. The significant lattice stiffening could be associated with a
lattice compression in the a direction [113]. The softening of the high-frequency bands is associated
with the weakening of the vanadyl bond with Li insertion into the interlayer channels. According to
Pigorsch and Steger, absorption at 400–300 cm1 is due to possible Li–O bonds [114].
3.5. Lithium intercalation in LiNiO2
The electrochemical properties of lithiated transition-metal oxides with a layered structure such
as LiNiO2 and LiCoO2 have been extensively studied as positive electrode materials for rechargeable
lithium batteries [115]. LiNiO2 have a rock-salt-based structure of a-NaFeO2-type (R3m space
Fig. 33. The first charge–discharge profile of a Li||LiNiO2 cell at 0.1 mA cm2. The LiNiO2 material was grown by solid-
state reaction of LiOH and NiO precursors.
group) with two interpenetrating close-packed fcc sublattices: one sublattice consists of oxygen
anions, and the other consists of Li and Ni cations on alternating (1 1 1) planes. The individual
coordinated octahedra are face-sharing. Ni cations are located in octahedral 3a (0 0 0) sites and
oxygen anions are in a cubic close-packing, occupying the 6c (0 0 z, 0 0 z) sites. Li cations reside at
Wyckoff 3b (0 0 1/2) sites. The Bravais unit cell contains one formula unit (Z ¼ 1) [116].
Fig. 33 shows the first charge–discharge profile of a Li||LiNiO2 cell at 0.1 mA cm2 using a
positive electrode material grown by solid-state reaction. During the charge process, the trivalent
nickel ions in the low spin configuration (t62 e1 ) are oxidised to the low spin tetravalent state (t62 ). The
LixNiO2 materials are small-polaron semiconductors with holes in the Ni4þ/3þ couple and a massive
Ni–O covalent mixing. Total reversibility of the first cycle is an indicative of the stoichiometric
compound with negligible Ni(II) in the lithium layer. However, the formation of low lithium content
LixNiO2 (x < 0:2) causes cycle life failure. In addition, the material become highly catalytic toward
electrolyte oxidation, and some of the nickel ions may migrate to lithium sites. The formation of
pure LiNiO2 is difficult, and residual Ni(II) (up to 1–2%) exist between the NiO2 slabs. In fact, the
first cycle irreversibility during charge–discharge is mainly related to the amount of Ni(II) between
the slabs of NiO2, which require extra charge for oxidation to higher valence state, when electrolyte
decomposition is controlled [117]. Through careful synthesis and adjustment of lithium
concentration in the material during heat-treatment, we obtained LiNiO2 very close to stoichiometry.
Fig. 34a shows the RS spectrum of LiNiO2 powders synthesised by solid-state reaction. The
Raman spectrum of LiNiO2, which is very weak compared with that of LiCoO2, is dominated by two
bands at ca. 545 and 465 cm1. By factor-group analysis, one obtains the irreducible representation of
degree of freedom (GT 0 ) for atoms such as GT 0 ¼ A2u Eu for one Li ion, GT 0 ¼ A1g Eg A2u Eu
for two oxygen atoms, and GT 0 ¼ A2u Eu for one Ni ion. Excluding the acoustic modes, at the
Brillouin-zone centre, the factor-group analysis of D53d yields that the (2A2u 2Eu ) modes are infrared-
active and the (A1g Eg ) modes are Raman-active [118]. It should be noted that the Raman-active
modes include only the oxygen vibrations in the direction parallel (A1g) or perpendicular (Eg) to the c-
axis [119]. The Raman band located at ca. 545 cm1 can be viewed as the symmetric O–Ni–O
stretching vibration of NiO6 units. This band is assigned to the A1g symmetry in the D53d spectroscopic
space group. The RS peak at 465 cm1 has the Eg symmetry.
Fig. 34. Raman scattering spectra of (a) rock-salt LiNiO2 crystal and (b) Li0.8NiO2 electrochemically delithiated.
Despite an almost identical mass of the transition metal in LiCoO2 and LiNiO2, one observes
a peak shift of ca. 50 cm1 for the A1g mode and a decrease in intensity of the Raman spectrum.
The frequency splitting d ¼ nðA1g Þ nðEg Þ decreases from d ¼ 110 cm1 in LiCoO2 to
d ¼ 80 cm1 in LiNiO2. The Raman scattering efficiency of LiNiO2 appears to be very weak
in comparison with that of other rock-salt compounds. The origin of these features may be due to
a reduction of the rhombohedral distortion or/and an increase in the electrical conductivity in
LiNiO2. A reduction of the rhombohedral distortion in LiNiO2 would degenerate the A1g and Eg
modes of the R 3m symmetry into a F2g mode, which is an inactive vibrational mode for the
Fm3m symmetry of the rock-salt structure. Thus, intensities of Raman peaks are very sensitive
to the long-range order in NiO2 slabs. As pointed out by Inaba et al. [120], another reason for
the decrease in intensity is an increase in electrical conductivity. It was reported that the
electrical conductivity of LiNiO2 is higher than that in LiCoO2 because of their electronic
structure [121]. The electrical conductivity of LiNi0.7Co0.3O2, which is considered to be similar to
that of LiNiO2, reduces the Raman scattering efficiency due to the weak optical skin depth of the
incident laser beam. The broad Raman feature of the Eg mode for LiNiO2 is attributed to the
tendency of non-stoichiometry as a result of the presence of an excess nickel leading to the
Li1zNi1þzO2 formula [122]. Despite the small departure of stoichiometry (z ¼ 0:02) detected by
Rietveld analysis of X-ray diffraction patterns of our LiNiO2 samples, the small amount of Ni
excess induces disorder in the Li predominant slabs which is responsible for the broadening of
the Eg mode.
Fig. 34b shows the Raman scattering spectrum of Li0.8NiO2 electrochemically delithiated.
These data were collected using 5 mW of the 476.5 nm laser line. When lithium ions are extracted
from the host matrix, we observe changes in intensity of the Raman peaks and a strong increase of
the spectrum background. As far as the data accuracy allows to collect the peak wavenumbers,
frequency shifts are observed. One might expect such frequencies to shift because the lattice
constants differ in the delithiated compound. For xðLiÞ < 0:8, the Raman signal was so low that we
could not obtain a spectrum. This suggests that the vanishing of the Raman intensity is correlated to
the increase in the electronic conductivity, which is driven by the charge transfer during the
deintercalation reaction. The appearance of a metal-like conduction in delithiated LiNiO2 has been
previously observed by Saadoune and Delmas [123]. The slight frequency shift toward the low-
wavenumber side corresponds to the increase in the interlayer spacing due to an increase of the
repulsive interactions between two adjacent negatively charged NiO2 layers upon delithiation. Thus,
the interlayer force constant is reduced by about 8%.
4. Disordered host structures
Until now, the materials investigated with a view to finding suitable intercalation host structures
for the positive electrode of electrochemical generators with an alkali metal or silver anode have
been essentially crystalline structures. Yet the discovery of the semiconducting properties of
phosphorus pentoxide-based glasses, a quarter of a century ago [124], has opened the area of
amorphous and disordered semiconductors. The amorphous material structures were considered
merely as ‘‘accidents’’ until only recently. Today, however, they are studied in their own right due to
certain interesting characteristics related to the disordered state. The higher capacity of the
amorphous over crystalline materials, in MoS2, for example, has been pointed out by Whittingham
et al. [125]. This presumably is associated with either the more open lattice in amorphous
compounds or the disordered framework which prevents the decomposition in some materials. The
energy-storage capacity of lithium insertion in amorphous materials is very high. Further, some
amorphous substances can be used and this area is receiving increasing attention. An intercalation
reaction is topotactic in nature, the structure of the host is changed only by atomic displacements; the
reaction does not involve diffusive rearrangement of the host atoms. As an example, for a-MoS3 the
initial energy density on discharge is 1 Wh g1; this may be compared with 0.48 and 0.8 Wh g1 for
crystalline TiS2 and V6O13, respectively. The counterpart of the amorphous state is the low mobility
of Liþ ions in the framework which is a limitation for high current densities.
The aim of this section is to present a selection of appropriate disordered materials and to
illustrate their properties upon lithium intercalation. We are going to show materials that belong to
the transition-metal dichalcogenide group (MoS2), to the layered transition-metal oxides (MoO3,
V2O5 and LiCoO2), and the disordered spinel structures (LiMn2O4 and LiNiVO4).
4.1. Disordered MoS2
The electrochemical properties of crystalline molybdenum disulphide (c-MoS2) have been early
reported [20,21]. The specific energy density of a Li||MoS2 cell is very low, about 100 Wh kg1 for
the utilisation as a primary cell. Haering and co-workers [20,21] have discovered that a lithium
molybdenum disulphide compound exhibits several distinct stages of operation when used as a
cathode in a battery having a lithium anode. In this case, the cell is reversible and the specific energy
density is twice. The electrochemical properties of a disordered MoS2 (d-MoS2) phase has been
Fig. 35. Electrical resistivity of (a) disordered MoS2 compared with that of (b) crystalline MoS2.
reported by Jacobson et al. [126] and Julien et al. [127,128]. It has been demonstrated, that by
including such a material in a lithium cell, it is possible to increase the specific energy density for its
possible application in a high-rate rechargeable batteries.
A highly disordered sample of MoS2 was prepared by heat-treatment of crystalline MoS2
(Ventron) at 400 8C for 4 h under 1 Pa pressure. X-ray diffraction patterns show an intense but rather
broad diffraction (0 0 2) line. The structure is a highly folded but disordered stacking pattern of layered
MoS2 with a tendency to grow in-plane and weak tendency for the layers to stack. The broadness of the
diffracted lines manifests the high disorder lattice of the sample. Fig. 35 shows the temperature
dependence of the electrical resistivity of d-MoS2. The room temperature resistivity is 0.3 O cm which
is a low value compared with 1.25 O cm of c-MoS2. Similar influence of the disorder on electrical
parameter has been reported on sputtered MoS2 films [129]. The difference between electrical
characteristics demonstrates that the transport mechanism is driven by the scattering at inter-crystallite
boundaries in d-MoS2. The band-bending at grain boundaries, which leads to the formation of potential
barriers, is responsible for the low resistivity of disordered samples.
Typical discharge curve for electrochemical cell containing d-MoS2 is shown in Fig. 36. An
initial open-circuit voltage (OCV) of 2.4 V was observed which dropped continuously to 1.1 V at
Li3.0MoS2. For comparison, the OCV of the same cell with crystalline MoS2 cathode is also shown.
An abrupt change in OCV versus x curve at x 1 is the indication of the phase transformation
discussed above. The constant part of the curve reflects the two-phase region of the material. The
disordered phase of MoS2 shows significantly higher storage charge capacity compared with that of
the crystalline form. The discharge curve of d-MoS2 resembles the b-phase discharge curve
described by Haering and co-workers [20,21]. We remark that the discharge is not rigorously smooth
but small plateaus appear at 1.9, 1.7, 1.5 and 1.3 V. However, in the compositional range
0:1 x 3:0, the discharge curve can be fitted by the linear expression E ¼ E kx, where
E ¼ 1:85 V and k ¼ 9RT=F have been obtained for a Li||d-MoS2 cell. It is also interesting to remark
that for disordered MoS2 the OCV versus x curve does not exhibit the two characteristic plateaus
observed for crystalline samples.
Fig. 37 displays the chemical diffusion coefficient in d-LixMoS2 as a function of the degree of
Li insertion. The apparent D in disordered MoS2 at room temperature has been calculated on the
assumption of uniform Liþ distribution at any composition in the solid solution electrode. We
observe a continuous decrease of the value of D with the increase of lithium content. Value of the
Fig. 36. Discharge profiles of Li||MoS2 cells as a function x of Li concentration in LixMoS2 electrodes (a) disordered MoS2
and (b) crystalline MoS2.
diffusion coefficient of 107 cm2 s1 was obtained at low lithium concentration, 0:0 x 0:2. This
high value for the diffusion coefficient may result from the smaller diffusion path length of the
lithium ion in the disordered phase compared to that of crystalline sample. The variation in the
diffusion coefficient can be approximated by a relationship of the form
D ¼ D0 expðbxÞ; (17)
with D0 ¼ 1:7 107 cm2 s1 and b ¼ 2:8 mol1 for d-MoS2 [127,128]. This behaviour reflects the
loss of structural integrity on repeated cycles of intercalation and deintercalation. Nevertheless, even
after long-term cycling the remaining Liþ mobility is still sufficient for practical application of small
particle size in d-MoS2 cathode materials. In fact this is a general feature of most of the insertion
compounds [130].
4.2. Disordered MoO3
Molybdenum is known to exist in a number of oxidation states and a variety of oxides, sub-
oxides, hydroxides and hydrated complexes [131]. These oxides and oxide-hydrates of molybdenum
Fig. 37. Chemical diffusion coefficient of lithium ions in d-MoS2 as a function of lithium concentration. The diffusion
coefficients were deduced from the potential step polarisation method in Li||MoS2 cells.
Fig. 38. Typical discharge–charge curves of anhydrous MoO3 materials. Data were collected at current density
0.1 mA cm2 in Li||MoO3 cell using a non-aqueous electrolyte 1 M LiPF6 in EC-DEC.
in their highest oxidation state display a variety of structural types involving linked MoO6 octahedra.
Of the anhydrous oxides, the well-known orthorhombic form, denoted a-MoO3 is stable at room
temperature and its structure can be described in terms of layered lattice in which distorted MoO6
octahedra share edges and vertices to form corrugated two-dimensional sheets separated by a Van
der Waals gap.
Molybdenum trioxide was found to react readily with lithium forming two well-defined
discharge products, according to the X-ray data, which are different, but similar to the known high-
temperature Li2MoO3 phase [131]. However, MoO3 only reacts with about 1.5 Li/Mo. Typical
discharge–charge curves of anhydrous MoO3 materials in Li cell are shown in Fig. 38. The
electrochemical lithium insertion into the MoO3 framework can be described assuming the reduction
from Mo(VI) to Mo(V) and Mo(IV) oxidation states. The capacity of MoO3 observed here is
approximately in agreement with the theoretical gravimetric capacity 280 mAh g1. Since the MoO3
reduction process is reversible, the construction of rechargeable cells is possible. Satisfactory
charge–discharge efficiency and storage capability are other favourable features of this cell. Upon
lithiation of LixMoO3, the electronic conductivity increases from 104 S cm1 for x ¼ 0 to ca.
101 S cm1 for 0:3 x 0:9 [132]. Li diffusion coefficients in LixMoO3 powder depend on x. A
maximum value of ca. 109 cm2 s1 has been reported for x 0:6. Li mobility was shown to slightly
decrease upon subsequent discharge–charge cycles due to irreversible structural and morphological
changes of the host matrix. However, when the cell was recharged, reoxidation of Mo produces a
resistive compound which induces a large polarisation of the cell for potential of 3.5 V (Fig. 38).
This is very advantageous from the technology point of view since this material acts as a self-
limiting voltage medium at the end of the charge.
The electrochemical discharge–charge reaction of non-stoichiometric Mo oxides shows
interesting features. This suggests that both channel/site size and electronic conductivity are the
predominant factors influencing the extent of reversible lithium incorporation by metal oxides with
framework structures shear-related to those of ReO3 and MoO3. Distortions and non-equivalence of
the sites available for Liþ accommodation should affect the thermodynamics of the reduction
processes. The non-stoichiometric molybdenum oxide Mo5O14 (MoO2.8) with framework structure
based on mixed networks of polyhedra is permeated by large open channels [32]. Mo5O14 was
obtained by dehydration and annealing-treatment of molybdenum monohydrate at 750 8C. Fig. 39
shows the discharge–charge profiles of a Li||MoO2.8 cell. The open-circuit voltage is 3.1 V versus
Fig. 39. Discharge–charge characteristics of a Li||Mo5O14 cell.
Li/Liþ. The first discharge displays a stepped behaviour with a voltage plateau at ca. 2.2 V followed
by a potential decline for x > 0:7Li=Mo. The limiting reversible lithium capacity is 1.5 Li/Mo for
Mo5O14. This compound can be assigned the following Mo valence distribution Mo3 6þ Mo2 5þ . If
reversible lithium incorporation produced all Mo4þ, the expected lithium uptake would be 8/5 (1.6)
Li/Mo; a value close to that measured by electrochemical titration (Fig. 39). The gravimetric
capacity was much higher than the 280 mAh g1 of anhydrous MoO3.
The suitability of molybdenum oxide-hydrates, MoO3–nH2O, as cathode materials for non-
aqueous lithium batteries has been assessed by several workers [133–138]. It was made clear as a
general trend that the discharge capacity of MoO3–nH2O increases with decreasing water content,
but cycle life increases with increasing water content in the compositional range 0:33 < n < 1:00.
Monoclinic monohydrate, MoO3–H2O, having only one coordinated water molecule, showed a
discharge capacity of about 200 mAh g1 of acid weight and a discharge potential around 2.5 V
versus Li/Liþ. This cathode material displayed good charge–discharge cyclic behaviour at a capacity
below 1 Li/Mo, while keeping the original layered lattice on cycling. The oxide-hydrate material
prepared by sol–gel [133] shows a good performance and could be a candidate for practical primary
lithium battery.
Fig. 40 shows typical discharge curves of amorphous MoO3–H2O and crystalline monoclinic
hemihydrate oxides [133]. Amorphous cathode exhibits a stepwise discharge behaviour, including
two plateaus; the first step up to 0.3e/Mo and the second one up to 1.0e/Mo. This amorphous
MoO3–1H2O delivers a discharge capacity of 260 mAh g1 corresponding to 1.5e/Mo. On the
other hand, the discharge cell voltage using the crystalline MoO3–(1/2)H2O cathode is rather
monotonous without any plateau appearance. This electrochemical behaviour is attributed to the
large cavities available for Li ions which prevent repulsive forces between inserted ions. Cathode
formed with hemihydrate oxides gave a practical energy density of 630 Wh kg1 for 1.5e/Mo. X-
ray diffraction patterns of LixMoO3–(1/2)H2O measured after electrochemical titration show that
the monoclinic structure remains largely unchanged upon lithiation with the strongest (0 0 1)
Bragg line shifted toward lower angles. This suggests that during the discharge, Liþ ions are
inserted between the layers leading to a small increase of the interlayer spacing which can be
explained by a model of rigid-hydrated-MoO3 layers in their original state separated by
intercalated Liþ cations.
Fig. 40. Typical discharge curves of (a) amorphous MoO3–H2O and (b) crystalline monoclinic hemihydrate MoO3–(1/2)H2O
oxides.
Molybdenum trioxide thin-films can be easily prepared providing various forms of crystal
chemistry by tuning the substrate temperatures, Ts, and/or the oxygen partial pressure in the
deposition chamber [139–141]. For example, MoO3 films grown by flash evaporation technique on
silicon substrates maintained in the temperature range 30–300 8C revealed a predominant a-phase
with a well-defined (0 k 0) orientation. The surface morphological studies show that films prepared
in the range 200–300 8C are crystalline with elongated crystal geometry. The layered nature of the
film is shown on each crystallite.
Discharge curves of Li||MoO3 cells are shown in Fig. 41. These microbatteries display
electrochemical features as follows. An initial voltage of 3.2 V was measured for MoO3 thin-film
cells which is higher than that recorded on the galvanic cell using crystalline material. The cell
voltage decreases continuously as a function of the degree of Li insertion and the steady behaviour is
a function of the structural arrangement in the film which depends on the substrate deposition
Fig. 41. Discharge curves of Li||MoO3 microbatteries including thin-films electrodes grown by the flash evaporation
technique. Experiments were carried out at current densities 2–10 mA cm2.
Fig. 42. Schematic representation of the electronic band structure of (a) a-MoO3 crystal and (b) MoO3 thin film. The
energy of the Fermi level EF lies with the localised electronic band upon Li insertion in the host material.
temperature [141]. The electrochemical process seems to be a classical intercalation mechanism for
the lithium ions with no voltage plateau occurrence during the discharge, and 1.5e transferred in the
host material. The reduction process is accompanied by coloration of MoO3 films. These results
suggest that Liþ diffusion may be anisotropic and limited by grain boundary effects which affect the
discharge curve. The discharge curve of the cell fabricated with the MoO3 film deposited at 250 8C
(Fig. 41, curve b) is quite stable. This may be attributed to the unique layered structure of a-MoO3
with large grain size in the film. A second possibility is the presence of the mixed a–b phase which
may reduce the standard potential. A third explanation may be oxygen-defects in the host structure
involving a lower Fermi level in such a semiconducting film.
In a simplest semiconductor model, electrochemical properties of LixMoO3 films can be
illustrated by an electronic band scheme as shown in Fig. 42 [142]. The cell voltage can be viewed as
the difference in the Fermi energies between lithium metal and positive electrode. In MoO3, there are
five O(pp) and three Mo(t2g) orbitals which interact to form p and p bands forming the valence and
conduction bands, respectively [143]. The antibonding p states hold the extra electrons supplied by
the inserted lithium ions. The narrowing of the conduction band is expected to lead to an increase in
the effective mass which will affect the position of the Fermi energy in LixMoO3 phases. In Li-
intercalated MoO3 crystal, the Fermi energy lies either with the donor level near the bottom of the
conduction band or in the conduction band itself (Fig. 42a), while in MoO3 films the density-of-
states might show a band in the middle of the gap and the Fermi level lies in the band. This last
description is supported by absorption measurements [141]. Thus, as a-MoO3 is a semiconductor
with a wide band-gap of 3.1 eV, the highest voltage displayed in thin-film cells is attributed to the
difference between Fermi levels (>0.5 eV) in the crystal and thin-film.
4.3. Polycrystalline V2O5
Vanadium pentoxide has been extensively studied for its potential application as material for
positive electrodes in advanced lithium batteries [130] and in electrochromic devices [144].
However, these applications depend on the techniques used to grow films and performances are
related with crystallinity and morphology of the films [145]. The open-circuit voltage is governed by
the V5þ/V4þ redox potential, but the capacity and reversibility of reaction mainly depend on the
structure of the oxide. Films of vanadium pentoxide can be prepared by various methods; in some
cases V2O5 possesses a non-stoichiometric character because of the oxygen-deficient nature of this
compound. Livage has reported that V2O5 prepared by sol–gel exhibits better properties than others
[146]. This advantage has been used to grow polycrystalline V2O5 films for lithium microbatteries.
Polycrystalline V2O5–nH2O films have been formed by the sol–gel method from neutral
precursors. Vanadium oxide ribbons behave as a polyvanadic acid H0.3V2O5–nH2O with a structure
Fig. 43. X-ray diffraction patterns of V2O5 films grown by sol–gel method and annealed in the temperature range 180–
400 8C.
in which lithium insertion is much easier and more reversible than in orthorhombic V2O5. Fig. 43
displays the XRD patterns of V2O5 films annealed at different temperatures in the range 180–450 8C.
These diagrams indicate that, after the water departure at 120 8C and the oxidation–decomposition
reaction started around 200 8C, the precursor transformed to the well-defined orthorhombic V2O5
structure at temperature as low as 350 8C. The layered-like structure of V2O5 is build up from
distorted trigonal bipyramidal coordination polyhedra of O atoms around V atoms, which share
edges to form (V2O4)n zigzag double chains along the (0 0 1) direction and are cross-linked along
(1 0 0) through shared corners.
When used as a cathode material in lithium rechargeable batteries, bulk vanadium pentoxide has
a theoretical energy density of about 500 Wh kg1, being one of the most important material for this
type of electrochemical cells. The electrochemical features of Li||V2O5 cells were analysed using
sol–gel films annealed in the temperature range 180–400 8C. Fig. 44 shows the discharge curves for
Fig. 44. Discharge curves of Li||V2O5 cells as a function of the degree of lithium insertion in the positive electrode.
LixV2O5 films grown by sol–gel method were annealed in the temperature range 180–4008C.
such cells. The discharge profile of polycrystalline V2O5 film (grown at 400 8C) presents the
appearance of the typical plateaus already observed in the bulk material (it could be compared with
the discharge curve of crystalline V2O5 shown in the inset of Fig. 44). Highly disordered V2O5 films
show a monotonously decreasing discharge profile. For all positive electrodes, the cell voltage is
higher than 2.0 V versus Li/Liþ for a lithium uptake of x ¼ 2:5, showing up that the performance of
film electrodes can be as promising as that of the bulk material. The monotonously decreasing curves
for disordered compounds, even for amorphous films, have been already observed for other materials
and have been associated to a preponderant entropic term in the thermodynamic voltage expression
[147]. The films can uptake about 3 Li at a discharge potential down to 1.9 V. The capacity of a cell
using a V2O5 film grown at T ¼ 350 8C (Fig. 44) is 110 mAh g1, which is nearby the value
measured for bulk material. The charging curve presents some hysteresis, probably linked to the
different kinetics on charge–discharge behaviour of films.
The kinetic parameters of Li insertion in the sol–gel LixV2O5 films, i.e. the chemical diffusion
coefficient and the thermodynamic factor, were evaluated during the relaxation period of the
discharge. The chemical diffusion coefficient is 7 1012 cm2 s1 at the beginning of the
intercalation process, and slightly decreases to 1 1012 cm2 s1 in the range 0 < x < 3. For
comparison, we remark that Baudry et al. [148] found a diffusion coefficient of 2:5 1012 cm2 s1
in thermally evaporated V2O5 thin-films and Bates et al. [149] reported diffusion coefficients varying
between 4 1015 and 2:6 1012 cm2 s1 for sputtered V2O5 films. The low diffusion coefficient
values are also attributed to the undefined conduction paths between grains of the films. The
potential barriers are more important in the inter-grain and the Liþ ions may be trapped by structural
defects.
The differences between materials of low and high crystallinity, as far as the microstructural and
macrostructural aspects are concerned, have been pointed out by Pereira-Ramos et al. [147]. At the
macroscopic level, materials of low crystallinity obtained by sol–gel technique are formed by small
particles. This experimental fact influences thermodynamics and kinetics as observed in the
electrochemical behaviour of the studied Li||V2O5 cells. When the microscopic level is taken into
account, several factors can distinguish poorly crystalline from highly crystalline materials, as
follows. For a reduced degree of long-range order, the presence of very small crystallites reduces the
probability that a diffusing ion impinges a channel in the V2O5 host structure. The polyhedra
constituting the unit cell may have some flexibility in the absence of a long-range order. The site
energies may not be the same as for the crystalline analogues, as a consequence of the different level
of symmetry of the basic polyhedra. The number of sites available for an intercalating ion may also
not be the same. In microcrystallite-based compounds, pathways for ion diffusion are shorter.
The absence of sudden change in the cell voltage for low-crystallinity materials such as V2O5
films denotes a reduced entropic contribution, in terms referring to the lack of the long-range order.
Conversely, the gradual variation of the cell voltage with composition seems to stem from a gradual
change in the enthalpy term. Instead, the unit cell possibly undergoes a modification which allows
intercalation of Liþ ions in more sites than permitted by the crystalline forms due to the flexibility of
the polyhedra constituting the unit cell of the host matrix in a local environment having a limited
range order. This seems to be connected to changing interactions among Li-intercalated ions and to
originate from an energy term related to minor modifications in the microstructure of V2O5 films.
As a result, it is clear that the morphology of the material plays an important role on the lithium
intercalation reaction in V2O5 films. Concerning the various interactions occurring in these
intercalated systems, one can notice that poorly crystalline V2O5 electrode does not show evidence
of constant repulsion forces among inserted neighbouring ions, the interaction between intercalated
ions and the host material seems to be more important for the low intercalation values; the
Fig. 45. X-ray diffraction patterns of LiCoO2 powder used as PLD target and LiCoO2 textured film grown into 50 mTorr
O2 partial pressure onto substrate maintained at T s ¼ 300 8C.
intercalation reaction may be also related with the density of the film, and the presence of impurities
should be considered due to the probability for a blocking effect of the ion diffusion.
4.4. LiCoO2 thin-films
Polycrystalline thin-films of lithium cobalt oxide were grown by pulsed-laser deposition (PLD)
technique [150,151]. Films of LiCoO2 were deposited onto Si substrates heated at temperature lower
than 300 8C from a sintered composite target (LiCoO2 þ Li2 O) irradiated with a Nd:YAG laser.
Under these deposition conditions, i.e. substrate temperature and partial oxygen pressure
temperature, LiCoO2 films grown from target without Li2O additive display the presence of cobalt
oxide impurities, namely Co3O4. Besides the peaks belonging to LiCoO2, two small peaks attributed
to Co3O4 are present at 2y ¼ 45 and 598. In addition, a large peak around 2y ¼ 70 is observed that
is mainly associated to the silicon substrate. As the amount of Li2O was added in the target, the XRD
patterns of the LiCoO2 films develop features expected for the regular layered phase. They are
indexed using the R 3m symmetry. Fig. 45 illustrates a highly textured (0 0 3) film obtained with the
use of a target with 15% Li2O additive. The X-ray diagram displays three sharp and intense peaks at
2y ¼ 19, 38 and 588, that are unambiguously attributed to the (0 0 3), (0 0 6) and (0 0 9) Bragg lines,
respectively [150]. The X-ray diagrams of LiCoO2 films formed at low substrate temperature
(T s < 150 8C) exhibit the amorphous nature of the layer. The typical peaks of the polycrystalline
trigonal layered phase of LiCoO2 films appear upon increasing the substrate temperature
(T s ¼ 300 8C) in oxygen partial pressure PðO2 Þ ¼ 50 mTorr using a lithium-rich target. The
cobalt–oxygen framework is well defined in this structure [151].
PLD LiCoO2 films obtained with a polycrystalline morphology were successfully used as
positive electrode materials in lithium microbatteries. Typical charge and discharge curves of a
Li||LiCoO2 cell using pulsed-laser deposited film grown at T s ¼ 300 8C are shown in Fig. 46.
Electrochemical measurements was carried out at a rate C/100 in the potential range 2.0–4.2 V; as
such, the voltage profile should provide a close approximation to the OCV. From the electrochemical
features, we may make some general remarks as follows. An initial voltage about 2.15 V versus Li/
Liþ was measured for LiCoO2 thin-film cathode cells, which is lower than that recorded on galvanic
cell using crystalline cathode [152]. The cell-voltage profiles display the typical profile currently
observed for LixCoO2 cathodes; the cell voltage is a function of the structural arrangement in the
Fig. 46. First charge–discharge curves of a Li||LiCoO2 microbattery using a LiCoO2 film grown by PLD technique in
50 mTorr oxygen pressure onto substrate heated at 300 8C.
film and thus depends on the substrate deposition temperature (Fig. 47). These potentials slightly
increased for films grown at high substrate temperature. This experimental fact is consistent with
many literature data and ensures that, at T s ¼ 300 8C, the material particles are electrochemically
active [152]. Specific capacity as high as 158 mC mm cm2 was measured on polycrystalline films
[151].
4.5. Disordered spinel LiMn2O4
Among the most promising compounds for Li-ion battery electrodes, the three-dimensional
spinel LiMn2O4 is able to reversibly uptake about 0.5 Li per transition-metal atom [153]. But there
are some problems such as difficulties encountered in controlling cationic order/disorder in these
systems during either their synthesis or cycling [154], that originate the delay of implement.
Problems associated to cationic distribution within LiMn2O4 appear to be severe because spinel
Fig. 47. Specific capacity and mid-voltage of Li||LiCoO2 cells as a function of the substrate temperature. PLD films were
grown onto silicon substrate at 25 T s 300 8C in PðO2 Þ ¼ 50 mTorr.
Fig. 48. The voltage-composition profile for a Li||LiMn2O4 using a spinel electrode synthesised by the combustion urea-
assisted method. The cell was charged and discharged at current density 0.05 mA cm2.
compounds of general formula AB2O4 are prone to cationic mixing between tetrahedral and octahedral
sites, leading to either normal [A]Tet.[B2]Oct.O4 or inverse [B]Tet.[AB]Oct.O4 spinels. LiMn2O4 was
shown to be a normal spinel, but deviations from such a cationic distribution exist depending on the
synthesis history, namely annealing temperature and cooling rate of the material. Here, we present the
properties of two LiMn2O4 spinel oxides, the first material was synthesised by the combustion urea-
assisted method [155] and the second one was obtained by milling of MnO2 and Li2O.
Using a wet-chemical technique assisted by urea as combustion agent, LiMn2O4 spinel oxides
exhibit grains with a quite small domain size of 0.5 mm The voltage composition profile for a
Li||LiMn2O4 using a spinel electrode synthesised by wet-chemical technique assisted by urea as
combustion agent is shown in Fig. 48. In the voltage domain between 3.0 and 4.5 V, the charge–
discharge curves correspond to the voltage profiles characteristic of the spinel LiMn2O4 cathode
material associated with lithium occupation of tetrahedral sites, in agreement with previous works.
From the variation of the cell potential for the complete cell, one can distinguish the presence of two
regions during the lithium insertion–extraction processes. The shape of the voltage curves indicates
whether the delithiated LiMn2O4 exists as a single- or a multiple-phase. In the latter case, the
potential is expected to be essentially invariant with composition. The first region (I) is characterised
by a S-shaped voltage curve, whereas the second region (II) corresponds to a plateau portion. In
region I, the charge voltage increases continuously in the voltage range of 3.80–4.05 V. In region II,
the charge voltage is stable around 4.10 V. The lithium extraction/insertion reactions in regions I and
II proceed in a LixMn2O4 matrix having a cubic symmetry. In the region II corresponding to the
upper voltage plateau, a two-cubic phase system is recognised, whereas the region I is attributed to a
single cubic phase characterised by a S-shaped voltage curve. The upper 4 V plateau provides over
110 mAh g1 based on the active material utilisation with an excellent cyclability [155].
Mechano-chemical technique is an efficient method which allows to obtain high dispersed
compounds directly during mechanical activation at room temperature [156,157]. LiMn2O4 was
synthesised by mechanical activation in plenary mill with steel balls and container of the mixture
Li2O and MnO2 electrochemically activated. The product shows a rather disordered spinel structure
as shown by the Raman scattering (RS) spectrum (Fig. 49). For this compound, the A1g mode at
630 cm1 is broadened in comparison with the RS spectrum of a l-LiMn2O4 ceramic.
This disordered structure is also at the origin of the electrochemical behaviour shown in Fig. 50.
The first charge–discharge curve recorded at the C/10 rate shows deviation from the standard
Fig. 49. Raman scattering spectra of l-LiMn2O4 spinel and LiMn2O4 disordered spinel grown by mechano-synthesis of
MnO2 and Li2O.
electrochemical profile of LiMn2O4 spinel. We do not observe the characteristic plateaus at 4.05 and
4.15 V occurring during in the charge reaction. This is attributed to the strong distortion of the MnO6
octahedra as evidenced by the red shift of the A1g Raman mode.
4.6. Disordered spinel LiNiVO4
Among the various vanadates, crystallised LiNiVO4 shows promising electrochemical

behaviour as a 4 V electrode material [158–160]. Crystallised LiNiVO4 synthesised by the classical
Fig. 50. The first charge–discharge profile of disordered LiMn2O4 spinel grown by mechano-synthesis of MnO2 and Li2O
mixture.
Fig. 51. Raman scattering spectra of a LiNiVO4 positive electrode discharged at different potentials. (a) 1.0 V and (b)
0.6 V vs. Li/Liþ.
high-temperature route can reversibly react with about 7 Li ions per transition metal when
discharged to voltages lower than 0.2 V [161], so that specific capacities as 800–900 mAh g1 could
be achieved. This vanadate was shown to become amorphous upon the first electrochemical
discharge, which must be performed at a slow cycling rate to ensure a proper function of the
electrode material upon subsequent cycles.
LiNiVO4 has been synthesised by a wet-chemical method, which consists of a low-temperature
reaction assisted by glycine as combustion agent. The powder mass was annealed at 350 and 500 8C
in air for 6 h for improving the crystallinity of the LiNiVO4 final product. XRD data show that
LiNiVO4 belongs to inverse spinel structure (space group Fd3m-O7h ) with a cubic lattice parameter of
8.222 Å. The pentavalent vanadium is located on the tetrahedral 8a sites, whereas Li and Ni are
distributed on the octahedral 16d sites, the distribution being disordered.
Fig. 51 shows the RS spectra of Li intercalated LiNiVO4 phase. Spectra were recorded at the
beginning of the two plateaus at 0.9 and 0.4 V occurring during the Li||LiNiVO4 cell discharge (Li
intercalation). These plateau should be at the origin of biphasic transformation due to the high
reduction of vanadium ions. The Raman features of LiNiVO4 have been discussed elsewhere [160].
It is dominated by a broad band in the 700–850 cm1 region, which is assigned to a vibration
between the oxygen and the highest valency cation. This vibration corresponds to the stretching
mode of VO4 tetrahedron having the A1 symmetry, whereas the band situated at 335 cm1
corresponds to the bending mode of VO4 tetrahedron with an E symmetry. Therefore, one observes
either the bending vibrations of VO4 tetrahedron or the vibrations involving the NiO6, LiO6
octahedral environments. If one considers that all the Li ions are accommodated in octahedral LiO6
environments, the F1u modes are normally split into (A 2B) Raman-active and IR-active
components. Therefore, IR modes having (A 2B) symmetry are intense, whereas Raman modes
are expected to be very weak. These two modes are observed at 416 and 435 cm1, respectively.
As shown in Fig. 51, there are important modifications of the Raman spectra of LiNiVO4 upon
Li intercalation. When the cathode is discharged, the stretching mode of VO4 tetrahedra is shifted
toward the low-wavenumber side corresponding to the reduction of the highest valency cation. At
0.6 V the Raman efficiency is very weak. Thus, the host lattice is highly intercalated and the Raman
peaks of LiNiVO4 decrease and a strong luminescence band appears at about 1.8 eV for Li5NiVO4.
5. Concluding remarks
In this work, we have shown the importance of local probes such as vibrational spectroscopies
associated with transport experiments for a well-defined characterisation of the structural features
and the charge transfer occurring in lithium intercalated materials.
The microscopic features listed earlier determine substantial differences between positive
electrodes of quite different crystallinity in terms of thermodynamic and kinetic properties. A
striking difference in the compositional variation of the cell voltage is observed for the two classes.
Crystalline materials and disordered compounds show remarkable differences. It is typical of the
well-crystallised electrodes to show plateaus and knees, while amorphous materials have sloping
curves with more gradual changes of cell voltage with composition. From the Gibbs–Helmholtz
equation, the cell voltage is given by V ¼ F 1ðDH TDSÞ. One observes that both an enthalpic
and an entropic term contribute to the shape of the emf versus composition curve. The enthalpic
contribution, DH, mainly comes from changes in the electrochemical potential, interactions among
Liþ ions residing in neighbouring sites and energy required to modify the structure, e.g. volume
expansion. The entropic contribution is indicative of major structure reorganisations, such as two-
phase formation or structure destruction.
It is obvious that, during the lithium insertion–extraction reaction, the electrochemical
behaviour of positive electrode materials is linked to their structural characteristics of the host
structure. In some cases, the disorder can induce better thermodynamics due to the modification in
the electronic band structure of the host, but kinetics are always found worse than in well-crystallised
materials. Thus, the various aspects of electrode parameters must be considered for each case, i.e.
specific capacity, voltage limit, diffusion coefficient, cycle life, etc. to ensure that the system remains
nearly stable in the entire intercalation range.
Concerning the various interactions occurring in these intercalated systems, one can notice that
(1) poorly crystalline electrode material does not show evidence of constant repulsion forces among
inserted neighbouring ions, (2) the interaction between intercalated ions and the host material seems
to be more important for the low intercalation values, (3) the intercalation reaction may be related
with the local environment of cations in the host matrix, and (4) the presence of impurities should be
considered due to the probability for a blocking effect of the ion diffusion.
However, it is difficult to clearly establish relationships between factors such as crystallinity,
stoichiometry and discharge capability. It is evidence that the morphology of the materials plays an
important role on the lithium intercalation reaction. Work is in the course to better understand all
these points. Nevertheless, it is interesting to point out that disordered materials can be successfully
used in solid-state microbatteries using either polymeric or glassy electrolyte with large values of
specific storage capacity.
Acknowledgements
The developments presented in this paper have greatly benefited from numerous discussions
with the following, who made the collaboration both pleasant and productive, I. Kerrache, O.M.
Hussain, S. Rangan, M.A. Camacho-Lopez, E. Haro-Poniatowski. Special thanks are addressed to
Dr. S. Castro-Garcia and S. Ziolkiewicz for their assistance in experimental works on lithium
intercalation in oxides. Mr. M. Lemal is gratefully acknowledged for his contribution to XRD
measurements.
References
[1] M. Balkanski, Phys. Scr. T39 (1991) 9.

[2] M.S. Whittingham, Prog. Solid State Chem. 12 (1978) 41.
[3] M.M. Thackeray, Prog. Solid State Chem. 25 (1997) 1.
[4] D.J. Sellmyer, Solid State Phys. 33 (1978) 83.
[5] H. Jones, Proc. R. Soc. London, Ser. A 144 (1934) 225.
[6] J. Friedel, Adv. Phys. 3 (1954) 446.
[7] W.Y. Liang, Mater. Sci. Eng. B 3 (1989) 139.
[8] E. Doni, R. Girlanda, in: V. Grasso (Ed.), Electronic Structure and Electronic Transitions in Layered Materials,
Riedel, Dordrecht, 1986, p. 1.
[9] P.C. Klipstein, C.M. Pereira, R.H. Friend, in: J.V. Acrivos, N.F. Mott, A.D. Yoffe (Eds.), Physics and Chemistry of
Electrons and Ions in Condensed Matter, NATO-ASI Series, Ser. C 130, Reidel, Dordrecht, 1984, p. 437.
[10] P.C. Klipstein, R.H. Friend, J. Phys. C 20 (1987) 4169.
[11] C. Julien, I. Samaras, O. Gorochov, A.M. Ghorayeb, Phys. Rev. B 45 (1992) 13390.
[12] A.M. Ghorayeb, R.H. Friend, J. Phys. C 20 (1987) 4181.
[13] A.R. Beal, S. Nulsen, Philos. Mag. B 43 (1981) 965.
[14] H. Isomaki, J. von Boehm, P. Krusius, J. Phys. C 20 (1979) 3239.
[15] C. Julien, J. Ruvalds, A. Virosztek, O. Gorochov, Solid State Commun. 79 (1991) 875.
[16] A. Virosztek, J. Ruvalds, Phys. Rev. B 42 (1990) 4064.
[17] G.A. Scholz, R.F. Frindt, Can. J. Phys. 61 (1983) 965.
[18] A.M. Ghorayeb, W.Y. Liang, A.D. Yoffe, in: M.S. Dresselhaus (Ed.), Intercalation in Layered Compounds, NATO-
ASI Series, Ser. B 148, Plenum Press, New York, 1986, p. 135.
[19] W.Y. Liang, in: M.S. Dresselhaus (Ed.), Intercalation in Layered Compounds, NATO-ASI Series, Ser. B 148,
Plenum Press, New York, 1986, p. 31.
[20] M.A. Py, R.R. Haering, Can. J. Phys. 61 (1983) 76.
[21] L.S. Selwyn, W.R. McKinnon, U. von Sacken, C.A. Jones, Solid State Ionics 22 (1987) 337.
[22] C. Julien, I. Samaras, I. Saikh, M. Balkanski, Mater. Res. Soc. Symp. Proc. 135 (1989) 461.
[23] K. Chrissafis, M. Zamani, K. Kambas, J. Stoemenos, N.A. Economou, I. Samaras, C. Julien, Mater. Sci. Eng. B 3
(1989) 145.
[24] T. Sekine, C. Julien, I. Samaras, M. Jouanne, M. Balkanski, Mater. Sci. Eng. B 3 (1989) 153.
[25] C. Julien, T. Sekine, M. Balkanski, Solid State Ionics 48 (1991) 225.
[26] J.B. Goodenough, in: J.R. Akridge, M. Balkanski (Eds.), Solid State Microbatteries, NATO-ASI Series, Ser. B 217,
[27] B. Yebka, C. Julien, Ionics 3 (1997) 83.
[28] S. Crouch-Baker, P.G. Dickens, Solid State Ionics 32/33 (1989) 219.
[29] C. Julien, G.A. Nazri, in: H.S. Nalwa (Ed.), Handbook of Advanced Electronic and Photonic Materials and Devices,
Academic Press, San Diego, 2000, p. 99.
[30] G.S. Nadkarni, J.G. Simmons, J. Appl. Phys. 41 (1970) 545.
[31] J.B. Goodenough, Prog. Solid State Chem. 5 (1971) 145.
[32] K.A. Wilhelmi, K. Waltersson, L. Kihlborg, Acta Chem. Scand. 25 (1971) 2675.
[33] K. West, B. Zachau-Christiansen, T. Jacobsen, Electrochim. Acta 28 (1983) 1829.
[34] D.W. Murphy, P.A. Christian, J.N. Carides, F.J. DiSalvo, in: P. Vashishta, J.N. Mundy, G.K. Shenoy (Eds.), Fast Ion
Transport in Solids, North-Holland, Amsterdam, 1979, p. 137.
[35] D.W. Murphy, P.A. Christian, F.J. DiSalvo, J.N. Carides, J.V. Waszczak, J. Electrochem. Soc. 128 (1981) 2053.
[36] N.C. Chaklanabish, H.S. Maiti, Solid State Ionics 21 (1986) 207.
[37] C. Julien, M. Balkanski, Mater. Res. Soc. Symp. Proc. 293 (1993) 27.
[38] K. Mizushima, P.C. Jones, J.B. Goodenough, Mater. Res. Bull. 15 (1980) 783.
[39] T. Nagaura, K. Tozawa, Prog. Batt. Solar Cells 9 (1990) 209.
[40] J.N. Reimers, J.R. Dahn, J. Electrochem. Soc. 139 (1992) 2091.
[41] W. Huang, R. Frech, Solid State Ionics 86–88 (1996) 395.
[42] C. Julien, A. Rougier, G.A. Nazri, Mater. Res. Soc. Symp. Proc. 453 (1997) 647.
[43] C. Julien, Solid State Ionics 136–137 (2000) 877.
[44] A. Rougier, G.A. Nazri, C. Julien, Ionics 3 (1997) 170.
[45] J. Molenda, A. Stoklosa, T. Bak, Solid State Ionics 36 (1989) 53.
[46] J. Van Elp, J.L. Wieland, H. Eskes, P. Kuiper, G.A. Sawatzky, F.M.F. De Groot, T.S. Turner, Phys. Rev. B 44 (1992)
6090.
[47] A. Van der Ven, M.K. Aydinol, G. Ceder, Mater. Res. Soc. Symp. Proc. 496 (1998) 121.
[48] C. Julien, M. Jouanne, P.A. Burret, M. Balkanski, Solid State Ionics 28–30 (1988) 1167.
[49] C.M. Pereira, W.Y. Liang, J. Phys. C 18 (1985) 6075.
[50] M.T. Ratajack, C.R. Kannewurf, J.F. Revelli, J.B. Wagner, Phys. Rev. B 17 (1978) 4674.
[51] A.R. Beal, W.Y. Liang, J. Phys. C 6 (1973) L482.
[52] R. Sudharsanan, K.K. Bordhan, B.P. Clayman, J.C. Irwin, Solid State Commun. 62 (1987) 563.
[53] W.Y. Liang, in: J.L. Beeby (Ed.), Condensed Systems of Low Dimensionality, NATO-ASI Series, Ser. B 253,
[54] M. Barj, C. Sourisseau, G. Ouvrard, R. Brec, Solid State Ionics 11 (1983) 179.
[55] J.V. Acrivos, W.Y. Liang, J.A. Wilson, A.D. Yoffe, J. Phys. C 4 (1979) L18.
[56] M.H. Whangbo, R. Brec, G. Ouvrard, J. Rouxel, Inorg. Chem. 24 (1985) 2459.
[57] J. Rouxel, J. Solid State Chem. 17 (1976) 223.
[58] J.A. Wilson, A.D. Yoffe, Adv. Phys. 18 (1969) 193.
[59] J.M. Vandenberg-Voorhoeve, in: P.A. Lee (Ed.), Optical and Electrical Properties of Layered Structures, Reidel,
Dordrecht, 1976, p. 426.
[60] A.R. Beal, J.C. Knights, W.Y. Liang, J. Phys. C 5 (1972) 3540.
[61] L.F. Mattheis, Phys. Rev. B 8 (1973) 3719.
[62] R. Huisman, R. De Jonge, C. Haas, F. Jellinek, J. Solid State Chem. 3 (1971) 56.
[63] R. Schöllhorn, U. Bethel, W. Paulus, Rev. Chim. Miner. 21 (1984) 545.
[64] C. Sourisseau, M. Fouassier, M. Alba, A. Ghorayeb, O. Gorochov, Mater. Sci. Eng. B 3 (1989) 119.
[65] E. Bayer, W. Rüdorff, Z. Naturforsch. 27b (1972) 1336.
[66] R.B. Somoano, V. Hadek, A. Rembaum, J. Chem. Phys. 58 (1973) 697.
[67] W.P.F. Omloo, F. Jellinek, J. Less-Common Met. 20 (1970) 121.
[68] B. Yebka, C. Julien, Solid State Ionics 90 (1996) 141.
[69] A.S. Nagelberg, W.L. Worell, J. Solid State Chem. 38 (1981) 321.
[70] W.R. McKinnon, J.R. Dahn, C. Levy-Clement, Solid State Commun. 50 (1984) 101.
[71] M.B. Armand, in: D.W. Murphy, J. Broadhead, B.C.H. Steele (Eds.), Materials for Advanced Batteries, Plenum
Press, New York, 1980, p. 145.
[72] M.B. Dines, Mater. Res. Bull. 10 (1975) 287.
[73] R. Collongues, La Non-Stoechiométrie, Masson, Paris, 1971.
[74] D.A. Dinn, J.M. Schemilt, B.C.H. Steele, Mater. Res. Bull. 11 (1976) 559.
[75] C. Julien, B. Yebka, C. Porte, Solid State Ionics 110 (1998) 29.
[76] A.R. Beal, W.Y. Liang, J. Phys. C 9 (1976) 2459.
[77] J.A. Baglio, G.S. Calabrese, K. Kamieniecki, R. Kershaw, C.P. Kubiak, A.J. Ricco, A. Wood, M.S. Wrighton, G.D.
Zoski, J. Electrochem. Soc. 129 (1982) 1461.
[78] B.G. Yacobi, F.W. Boswell, J.M. Corbett, J. Phys. C 12 (1979) 218.
[79] P.M. William, F.R. Shephered, J. Phys. C 6 (1973) L36.
[80] I. Kerrache, C. Julien, C. Sourisseau, Solid State Ionics 92 (1996) 37.
[81] C. Sourisseau, J.P. Forgerit, Y. Mathey, J. Solid State Chem. 49 (1983) 134.
[82] R. Brec, Solid State Ionics 22 (1986) 3.
[83] M. Piacentini, F.S. Khumalo, C.G. Olson, J.W. Anderegg, D.W. Lynch, Chem. Phys. 65 (1982) 289.
[84] M. Barj, C. Sourisseau, G. Ouvrard, R. Brec, Solid State Ionics 11 (1983) 179.
[85] P.J.S. Foot, T. Katz, S.N. Patel, B.A. Nevett, A.R. Piercy, A.A. Balchin, Phys. Status Solidi A 100 (1987) 11.
[86] M. Bernasconi, G.L. Marra, G. Benedek, L. Miglio, M. Jouanne, C. Julien, M. Scagliotti, M. Balkanski, Phys. Rev.
B 38 (1988) 12089.
[87] M. Balkanski, M. Jouanne, G. Ouvrard, M. Scagliotti, J. Phys. C 20 (1987) 4397.
[88] T. Sekine, M. Jouanne, C. Julien, M. Balkanski, Mater. Sci. Eng. B 3 (1989) 91.
[89] R. Mercier, J.P. Malugani, B. Fahys, J. Douglade, G. Robert, J. Solid State Chem. 43 (1982) 151.
[90] J. Bichon, M. Danot, J. Rouxel, C. R. Acad. Sci. (Paris) 276 (1973) 1283.
[91] C. Berthier, Y. Chabre, P. Segransan, Physica 99B (1980) 107.
[92] M. Danot, A. LeBlanc, J. Rouxel, Bull. Soc. Chim. Fr 8 (1969) 2670.
[93] A. Lerf, R. Schöllhorn, Inorg. Chem. 16 (1977) 2950.
[94] J. Morales, C. Perez-Vicente, J.L. Tirado, Solid State Ionics 51 (1992) 133.
[95] C. Julien, C. Perez-Vicente, Ionics 1 (1995) 387.
[96] C. Julien, H.S. Mavi, K.P. Jain, M. Balkanski, C. Perez-Vicente, J. Morales, Mater. Sci. Eng. B 23 (1994) 98.
[97] C. Julien, L. El-Farh, M. Balkanski, O.M. Hussain, G.A. Nazri, Appl. Surf. Sci. 65/66 (1993) 325.
[98] S.F. Cogan, N.M. Nyugen, S.J. Perroti, R.D. Rauh, J. Appl. Phys. 66 (1989) 1333.
[99] L. Fiermans, P. Clauws, W. Lambrecht, L. Vandeabroucke, J. Vennik, Phys. Status Solidi A 59 (1980) 485.
[100] H.G. Bachmann, F.R. Ahmed, W.H. Barnes, Z. Kristallogr. 115 (1961) 110.
[101] H.G. Bachmann, W.H. Barnes, Z. Kristallogr. 115 (1961) 215.
[102] J.M. Cocciantelli, J.P. Doumerc, M. Pouchard, M. Broussely, J. Labat, J. Power Sources 34 (1991) 103.
[103] A. Talledo, A.M. Andersson, C.G. Granqvist, J. Appl. Phys. 69 (1991) 3261.
[104] R.J. Cava, A. Santoro, D.W. Murphy, S.M. Zahurak, R.M. Fleming, P. Marsh, R.S. Roth, J. Solid State Chem. 65
(1986) 63.
[105] D.W. Murphy, P.A. Christian, F.J. DiSalvo, J.N. Carides, J.V. Waszczak, J. Electrochem. Soc. 128 (1981) 2053.
[106] C. Delmas, S. Brèthes, M. Menetrier, J. Power Sources 34 (1991) 113.
[107] N.F. Mott, J. Non-Cryst. Solids 1 (1968) 1.
[108] C. Julien, A. Khelfa, J.P. Guesdon, V. Tuncheva, F. Gendron, Ionics 2 (1996) 380.
[109] C.V. Ramana, O.M. Hussain, B.S. Naidu, C. Julien, M. Balkanski, Mater. Sci. Eng. B 52 (1998) 32.
[110] L. Abello, E. Husson, Y. Repelin, G. Lucazeau, Spectrochim. Acta A 39 (1983) 641.
[111] P. Clauws, J. Broeckx, J. Vennik, Phys. Status Solidi B 131 (1986) 459.
[112] A.V. Manukhin, O.A. Plaksin, V.A. Stepanov, Inorg. Mater. 24 (1988) 186.
[113] A. Talledo, C.G. Granqvist, J. Phys. D 27 (1994) 2445.
[114] E. Pigorsch, W.E. Steger, Phys. Status Solidi A 117 (1990) K189.
[115] K. Mizushima, P.C. Jones, J.B. Goodenough, Mater. Res. Bull. 15 (1980) 763.
[116] H.J. Orman, P.J. Wiseman, Acta Cryst. C 40 (1984) 12.
[117] J.P. Peres, C. Delmas, A. Rougier, M. Broussely, F. Perton, P. Biensan, P. Willmann, J. Phys. Chem. Solids 57 (1996)
1057.
[118] R.K. Moore, W.B. White, J. Am. Ceram. Soc. 53 (1970) 5679.
[119] W. Huang, R. Frech, Solid State Ionics 86–88 (1996) 395.
[120] M. Inaba, Y. Todzuka, H. Yoshida, Y. Grincourt, A. Tasaka, Y. Tomida, Z. Ogumi, Chem. Lett. (1995) 889.
[121] C. Julien, Solid State Ionics 136–137 (2000) 887.
[122] J.R. Dahn, U. von Sacken, C.A. Michal, Solid State Ionics 44 (1990) 87.
[123] I. Saadoune, C. Delmas, J. Mater. Chem. 6 (1998) 193.
[124] H. Rawson, P.L. Baynton, J.F. Stanworth, J. Electrochem. Soc. 4 (1957) 104.
[125] M.S. Whittingham, R.S. Chianelli, A.J. Jacobson, in: D.W. Murphy, J. Broadhead, B.C.H. Steele (Eds.), Materials
for Advanced Batteries, Plenum Press, New York, 1980, p. 291.
[126] A.J. Jacobson, R.R. Chianelli, M.S. Whittingham, J. Electrochem. Soc. 126 (1979) 2277.
[127] C. Julien, S.I. Saikh, G.A. Nazri, ISSI Lett. 1 (1990) 12.
[128] C. Julien, S.I. Saikh, G.A. Nazri, Mater. Sci. Eng. B 15 (1992) 73.
[129] R. Bichel, F. Levy, J. Phys. D 19 (1986) 1809.
[130] C. Julien, G.A. Nazri, Solid State Batteries: Materials Design and Optimization, Kluwer Academic Publishers,
Boston, 1994.
[131] C. Julien, B. Yebka, in: C. Julien, Z. Stoynov (Eds.), Materials for Lithium-ion Batteries, NATO-ASI Series, Ser. 3/
85, Kluwer Academic Publishers, Dordrecht, 2000, p. 263.
[132] C. Julien, G.A. Nazri, Solid State Ionics 68 (1994) 111.
[133] G. Guzman, B. Yebka, J. Livage, C. Julien, Solid State Ionics 86–88 (1996) 407.
[134] G.A. Nazri, C. Julien, Ionics 2 (1995) 1.
[135] F.W. Dampier, J. Electrochem. Soc. 121 (1974) 656.
[136] N. Margalit, J. Electrochem. Soc. 121 (1974) 1460.
[137] N. Kumagai, N. Kumagai, K. Tanno, J. Appl. Electrochem. 18 (1988) 857.
[138] M. Sugawara, Y. Kitada, K. Matsuki, J. Power Sources 26 (1989) 373.
[139] C. Julien, G.A. Nazri, J.P. Guesdon, A. Gorenstein, A. Khelfa, O.M. Hussain, Solid State Ionics 73 (1994) 319.
[140] C. Julien, A. Khelfa, O.M. Hussain, G.A. Nazri, J. Cryst. Growth 156 (1995) 235.
[141] C. Julien, B. Yebka, J.P. Guesdon, Ionics 1 (1995) 316.
[142] C. Julien, Ionics 2 (1996) 169.
[143] J.B. Goodenough, A. Manthiram, A.C.W.P. James, P. Strobel, Mater. Res. Soc. Symp. Proc. 135 (1989) 391.
[144] Y. Fujita, K. Miyazaki, C. Tatsuyama, Jpn. J. Appl. Phys. 29 (1990) L1708.

[145] A.M. Abo-El-Soud, B. Mansour, L.I. Soliman, Thin Solid Films 247 (1994) 140.
[146] J. Livage, Chem. Mater. 3 (1991) 578.
[147] J.P. Pereira-Ramos, N. Baffier, G. Pistoia, in: G. Pistoia (Ed.), Lithium Batteries, New Materials, Development and
Perspectives, Elsevier, Amsterdam, 1994, p. 281.
[148] P. Baudry, M.A. Aegerter, D. Deroo, B. Valla, J. Electrochem. Soc. 138 (1991) 460.
[149] J.B. Bates, G.R. Gruzalski, N.J. Dudney, C.F. Luck, X.H. Yu, Solid State Ionics 70–71 (1994) 619.
[150] L. Escobar-Alarcon, E. Haro-Poniatowski, M. Massot, C. Julien, Mat. Res. Soc. Symp. Proc. 548 (1999) 223.
[151] C. Julien, M.A. Camacho-Lopez, L. Escobar-Alarcon, E. Haro-Poniatowski, Mater. Chem. Phys. 68 (2001) 210.
[152] B. Garcia, J. Farcy, J.P. Pereira-Ramos, J. Perichon, N. Baffier, J. Power Sources 54 (1995) 373.
[153] J.M. Tarascon, D. Guyomard, Electrochim. Acta 38 (1993) 1221(and references therein).
[154] J.M. Tarascon, in: C. Julien, Z. Stoynov (Eds.), Materials for Lithium-ion Batteries, NATO-ASI Series, Ser.3/85,
Kluwer Academic Publishers, Dordrecht, 2000, p. 75.
[155] S. Chitra, P. Kalyani, T. Mohan, R. Gangadharan, B. Yebka, S. Castro-Garcia, M. Massot, C. Julien, M. Eddrief, J.
Electroceram. 3 (1999) 433.
[156] N.V. Kosova, I.P. Asanov, E.T. Devyatkina, E.G. Avvakumov, J. Solid State Chem. 146 (1999) 184.
[157] S. Soiron, A. Rougier, L. Aymard, J.M. Tarascon, J. Power Sources 97–98 (2001) 402.
[158] J.C. Bernier, P. Poix, A. Michael, C. R. Acad. Sci. (Paris) 253 (1961) 1578.
[159] G.T.K. Fey, W. Li, J.R. Dahn, J. Electrochem. Soc. 14 (1994) 2279.
[160] S.R.S. Prabaharan, M.S. Michael, S. Radhakrishna, C. Julien, J. Mater. Chem. 7 (1997) 1791.
[161] F. Orsini, E. Baudrin, S. Denis, L. Dupont, M. Touboul, D. Guyomard, Y. Piffard, J.M. Tarascon, Solid State Ionics
107 (1998) 123.

1 s2.0 S0927796X02001043 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S0927796X02001043 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S0927796X02001043 Main

Uploaded by

Copyright:

Available Formats

Materials Science and Engineering R 40 (2003) 47–102

Lithium intercalated compounds

Accepted 20 September 2002

Compounds exhibiting an open structure, in particular the transition-metal dichalcogenides

xAþ þ xe þ hHi , Ax þ hHix (1)

structure has an electronic component. Consequently, it is possible to consider three classes of

2. The reliability of rigid-band model

2.1. Lithium intercalation in TiS2

2.2. Lithium intercalation in TaS2

2.3. Lithium intercalation in 2H-MoS2

Fig. 9. Temperature dependence of the electrical conductivity of lithium intercalated MoS2.

2.4. Lithium intercalation in MoO3

a ¼ o2p t½ncð1 þ o2 t2 Þ ; (6)

Fig. 13. FTIR absorption spectra of a-MoO3 and Li0.3MoO3.

2.5. Lithium intercalation in V6O13

2.6. Lithium intercalation in LiCoO2

Lithiated transition-metal oxides with a layered, a-NaFeO2-type, structure such as LiMeO2

Fig. 15. FTIR absorption spectra of V6O13 and Li3.2V6O13.

Fig. 17. FTIR absorption spectra of LiCoO2 as a function of the Li concentration.

Optical spectroscopy associated with transport measurements is a powerful tool for

3.1. Lithium intercalation in WS2

xLiþ þ xe þ ½W4þ S2 ! ½Lixþ Wð4xÞþ S2 : (7)

D ¼ kW½ðxmax xÞ2 þ bxðxmax xÞ ; (9)

3.2. Lithium intercalation in MPS3 phases

3.2.1. Lithium intercalation in NiPS3

Oh C2 ðsiteÞ C32h ðcrystalÞ

3.2.2. Lithium intercalation in FePS3

3.3. Lithium intercalation in SnS2

3.4. Lithium intercalation in V2O5

3.5. Lithium intercalation in LiNiO2

4. Disordered host structures

4.1. Disordered MoS2

4.2. Disordered MoO3

Fig. 39. Discharge–charge characteristics of a Li||Mo5O14 cell.

4.3. Polycrystalline V2O5

4.4. LiCoO2 thin-films

4.5. Disordered spinel LiMn2O4

4.6. Disordered spinel LiNiVO4

Among the various vanadates, crystallised LiNiVO4 shows promising electrochemical

[1] M. Balkanski, Phys. Scr. T39 (1991) 9.

[144] Y. Fujita, K. Miyazaki, C. Tatsuyama, Jpn. J. Appl. Phys. 29 (1990) L1708.

You might also like

a ¼ o2p t½ncð1 þ o2 t2 Þ; (6)

D ¼ kW½ðxmax xÞ2 þ bxðxmax xÞ; (9)