Journal of Power Sources, 8 (1982) 229 - 255 229

REVIEW OF THE STRUCTURE AND THE ELECTROCHEMISTRY OF

NICKEL HYDROXIDES AND OXY-HYDROXIDES

P. OLIVA, J. LEONARD1 and J. F. LAURENT

S.A.F.T., 111, Bld Alfred Daney, 33000 Bordeaux (France)
C. DELMAS and J. J. BRACONNIER
Laboratoire de Chimie du Solide du CNRS, Universite de Bordeaux I, 351, tours de la
Liberation, 33405, Talence Ce’dex (France)
M. FIGLARZ and F. FIEVET
Universite Paris VII, Place Jussieu, 75251, Paris Ckdex 05 (France)
A. de GUIBERT
C.G.E. (Laboratoire de Marcoussis), Route de Nosay, 91460, Marcoussis (France)

Summary

Particular emphasis is placed on the structural (and textural) character-

ization of the phases involved in nickel hydroxide electrodes as a prerequisite
to a rational approach to their redox behaviour.
Recent chemical and electrochemical papers are discussed with the
object of throwing some light on traditionally unexplored features of such
electrodes.

Introduction

The remarkable cycling durability of nickel hydroxide electrodes (NOE)

in alkaline cells has long been interpreted on the basis that a highly reversible
redox transfer occurs between well defined active species. This is the reason
why so much effort has been devoted to attempts to characterize the phases
present within the electrode during charge and discharge.
In fact, progress in the understanding of solid-phase properties and
transport mechanisms has now convinced experimentalists that continuous
slight changes occur on cycling both in the electrochemical mechanisms and
in the structures of the involved species, and that perfect control over NOE
redox processes would still require years of investigation.

0378-7753/82/0000-0000/$02.75 @ Elsevier Sequoia/Printed in The Netherlands

230

Hence the general reaction scheme initially proposed by Bode et al. [ l] *.

fl-Ni(OH), * fl-NiOOH + H+ + e-
g\ \
\.
1 \
+-La-.---
a-Ni(OH), -b r-NiOOH + H’ + e-

the validity of which is still unquestioned, must be acknowledged with suf-

ficient flexibility in order not to ignore the prevailing influence of such
parameters as electrolyte concentration and nature, material texture, charge
and discharge rates, etc. . . ., on the electrochemical system.
The purpose of this paper is thus to propose a synthetic view of the
subject through a close examination of all recent papers published both in
the chemical and the electrochemical literature, and to complement previous
reviews [ 2 - 61 by placing maximum emphasis on a structural interpretation
of the NOE behaviour.
The following subjects will be dealt with:
A - Physico-chemical and structural characteristics of Ni(I1) hydroxides.
B - Hydroxides and oxy-hydroxides with a nickel oxidation state
higher than II.
C - Structural approach to the NOE behaviour.
D - Thermodynamic and kinetic approach to the NOE behaviour.
E - Role of the electrolyte.
F - Role of foreign cations.

A. Physico-chemical and structural characteristics of Ni(I1) hydroxides

Early crystallographic studies established a marked distinction between
“well crystallized”p-Ni(OH)z, isomorphous with brucite, Mg(OH),, and com-
pounds which can still be considered as Ni( II) hydroxides but which contain a
variable excess of intersheet water (+ foreign ions) and exhibit low crystallinit!
The conventional a-Ni(OH), denomination used for such compounds
must then be taken as a general denomination for a large set of disordered
Ni(I1) hydroxides and does not represent a well defined polymorph of
Ni(OH), . However, the similarity of trends observed in the electrochemical
behaviour of all such disordered “(Y” hydroxides has induced electrochemists
to adopt the same a-p dichotomy as crystallographers, and we shall thus
adhere to such a convention in the present paper.
The different methods of preparation of nickel hydroxide have been re-
viewed by Oswald and Asper [ 71. Nickel hydroxide does not tend to form large,
well shaped crystals, and during its precipitation from aqueous salt solutions,
well crystallized Ni(OH), is not the primary product. Depending on experi-
mental conditions either hydroxy-salts (basic salts) or poorly crystallized
a-type hydroxides can be isolated as primary precipitation products. /3-
Ni(OH)z may then be obtained by ageing of such compounds [ 8 - 111, but
always contain adsorbed foreign ions and water.

*- - - Dashed lines indicate partial reactions.

 231

1. fi-NitOH),
To obtain pure nickel hydroxide, NHs proves to be a much better
precipitating reactant than strong alkali. Merlin’s method [ 121 consists of
adding an excess of ammonia to a nickel salt solution; on boiling, P-Ni(OH)z
precipitates slowly from the Ni( II) hexaammine solution thus obtained. A par-
ticularly pure nickel hydroxide has also been prepared by Fievet and Figlarz
[13, 141 via a two-step reaction: (i) precipitation of an a-type turbostratic
hydroxide from ammonia and a nickel solution, (ii) hydrothermal treatment
of this turbostratic hydroxide.
p-Hydroxide obtained by the latter method consists of hexagonal, plate-
like, submicronic particles (Fig. l(a)). Selected area electron diffraction
indicates that such hexagonal platelets are monocrystals lying on the (001)
plane (Fig. l(b)).

(a) (b)
Fig. 1. fl-Ni(OH)z particles obtained by hydrothermal growth of an aqueous suspension of
a turbostratic nickel hydroxide. (a) Electron microscopy; (b) selected area electron dif-
fraction by one particle of sample (a).

Crystal structure
@Ni(OH), crystallizes in the hexagonal system (P&zl-D33d) common
to several halogenides (CdIz type) and to other hydroxides M(OH)z (M = Ca,
Mg, Fe, Co, Mn, Cd). The unit cell parameters are a = 3.126 A and c = 4.605
A [ASTM, 14 - 1171.
The structure can be described as an hexagonal close-packed structure
of hydroxyl ions (AB oxygen packing) with Ni(II) occupying octahedral
interstices one plane out of two. It can also be visualized as a layered struc-
ture, each layer consisting of an hexagonal planar arrangement of Ni(I1) ions
with an octahedral coordination of oxygen, three oxygen atoms lying above
the nickel plane, three lying below. The layers are stacked up along the c
axis. By analogy with the structure of Mg(OH)z and Ca(OH)z the O-H bond
is thought to be parallel to the c axis [ 181.
As evidenced by IR there is no hydrogen bonding between the OH
groups of two adjacent layers [ 191.
232

Infrared and Raman spectra

A complete factor group analysis of crystals containing linear groups
has been done by Mitra [ 201. Each OH- ion in the brucite-type structure
occupies a C& site. Its internal and external normal vibrations form the
basis of the representation I = 2A, + 2E.
All the centrosymmetric g modes are Raman active, whereas the anti-
symmetric u modes are excited in the infrared. Thus four bands can be seen
both in the IR (Fig. 2) and the Raman spectra of P-Ni(OH),.
Ni(OH), IR absorption has been thoroughly studied by Cabannes-Ott
[21] in relation to the spectra of isomorphous M(OH)s hydroxides. The
narrow band located near 3650 cm-l is characteristic of the stretching vibra-
tion Van of non-hydrogen-bonded OH groups. The three other bands are
located in the 700 - 300 cm-’ range. They are attributed respectively to the
in-plane deformation hoi., (520 cm-‘), to the out-of plane deformation
vibration -yOH(345 cm-‘) and to the Ni-0 stretching vibration VNio (460
cm-l) [21].
Absorption bands related to adsorbed species are also often observed.
Thus, a broad band centered at 3450 cm-l and a weak absorption around
1630 cm-’ can be attributed to the respective stretching and bending modes
of adsorbed water molecules. Similarly, IR spectra often display absorption
bands due to nickel co-anions of the mother solution. For instance, adsorbed
NO, and CO,“- ions give rise to several bands in the range 1600 - 1000 cm-’
P3, Wa)l .
By contrast, Raman spectra have never been exhaustively analysed.

Fig. 2. Infrared spectra. (a) /3-Ni(OH),; (b) turbostratic Ni(OH),.

 233

2. Hydrous, imperfectly ordered Ni(II) hydroxides

Hydrated “o-type” Ni(I1) hydroxides can be obtained either by
chemical precipitation or by cathodic deposition upon electrolysis of a nick-
el salt solution. Nickel nitrate solutions seem to be particularly suitable as
starting nickel solutions [ 1, 24 - 261, strong alkali or ammonia [ 14, 25, 261
being used as precipitating agents. In fact, a great variety of “(Y” hydroxides
can be obtained, ranging from the hydrated turbostratic hydroxide proposed
by Le Bihan [25] to a whole set of compounds for which it seems unrealistic
to establish a clear frontier between hydroxides and hydroxy-salts, since
they all represent unstable and non-stoichiometric intermediate compounds
in the formation process of the more stable fi-Ni(OH), form.
Louer, for instance, who devoted many studies to nickel hydroxy-
nitrates [27(a)] , advocates the idea that “(Y hydroxides” commonly prepared
by electrochemists [29 - 311 from nitrate baths might well be regarded as
highly hydrated, degenerate forms of Ni(I1) hydroxy-nitrates. The variations
in crystallinity of the hydroxide samples thus obtained can be illustrated by
the X-ray diffraction patterns of Fig. 3, which show a highly divided crystal-
line structure (Fig. 3(a)), microcrystalline solids (Fig. 3(b)) or turbostratic
layered structures (Fig. 3(c)) [14,25].

X-ray diffraction and crystalline structure

A detailed description of the imperfect crystalline organization of ar-
nickel hydroxide has been derived from X-ray diffraction spectra similar to
those of Fig. 3(c). By comparison with the spectra of well crystallized
/3-Ni(OH)s (Fig. 3(d)) the major characteristics of spectra 3(c) are:
(i) the existence of intensity maxima around 8.5 and 4.25 a instead of
the (001) line at 4.6 A;
(ii) the occurrence of asymmetric bands instead of the hk0 lines of the
brucite structure;
(3) the vanishing of the general hkl lines of the brucite structure.
To account for this type of spectra Bode et al. [l] proposed a structure
where the (001) layers of the brucite structure were still stacked up along
the c axis but were separated by intercalar water molecules. They assumed
a quasi-compact stacking of OH-HsO-OH planes, and they assigned a fixed
definite position for intercalar water molecules which would correspond to
the idealized formula 3Ni(OH),+ 2Hz0.
Such a well-organized structure has, however, been rejected by both
Le Bihan [25, 321 and McEwen [28] who proposed, instead, a turbostratic
structure similar to those observed with some layered minerals and carbon
black. The turbostratic nickel hydroxide consists of parallel and equidistant
Ni(OH)s layers, similar to those found in fl-Ni(OH), (as evidenced by
magnetic measurements [ 15 - 17]), but which are randomly oriented and
separated by interdalar water molecules bonded to the hydroxyl groups by
hydrogen bonds.
234

dL-
3’ Co Kal ,o
20 30 40
a

Fig. 3. X-ray diffraction spectra. a, Quasi amorphous nickel hydroxide precipitated by

ammonia and washed by decant&ion; b, microcrystalline nickel hydroxide precipitated
by NaOH; c, turbostratic nickel hydroxide precipitated by ammonia and washed by cen-
trifugation; d, pure P-Ni(OH)2 obtained by hydrothermal growth from an aqueous
suspension of turbostratic hydroxide.

The lines around 8.5 and 4.25 A displayed on the X-ray spectra are the
first and second order diffraction maxima corresponding to the regular
stacking of Ni(OH), layers along the c axis.
The asymmetric hk0 bands are explained by the randomisation of the
layers, c remaining constant. In his model McEwen assumes a very large
delocalization of water molecules in the intercalar layer. Le Bihan [33]
.theorized a model assuming an hexagonal quasi-close-packed arrangement of
intercalar water molecules, and interpreted the randomisation of the layers
 235

as arising from variations of the hydrogen bond length between intercalar

Hz0 molecules and OH ions of the layers. He found [ 251 that the crystal-
lites of the turbostratic hydroxide had a mean size of 30 A along [OO.l] ,
which would correspond to a stacking of 5 layers. The mean size of the crys-
tallites in the perpendicular direction was evaluated as 80 A.
By electron microscopy [ 33,351 turbostratic nickel hydroxides appear
as aggregates of thin crumpled sheets, without any definite shape (Fig. 4(a)).
This is a characteristic feature which has also been observed with turbostratic
clays. It is difficult to obtain microdiffraction patterns from a turbostratic
nickel hydroxide sample because it is easily dehydrated by the electron
beam. Nevertheless, it has been possible to observe diffraction rings corre-
sponding to the asymmetric bands MO observed by X-ray diffraction [ 331
(Fig. 4(b)).

(4 (b)
Fig. 4. Turbostratic nickel hydroxide. (a) Electron microscopy; (b) selected area electron
diffraction.

IR and Raman spectra

A detailed comparison of IR absorption bands for /3-Ni(OH), and (Y-
Ni(OH)s has been undertaken by Le Bihan and Figlarz [ 33, 34(a)] . The
sharp band at 3650 cm-’ corresponding to the stretching vibration Vou of
P-Ni(OH), is replaced by a broad and asymmetric band around 3500 cm-’
(Fig. 2), and the three bands corresponding to &on, yOH and Vnio are shifted
towards shorter wavelengths. In addition, the band at 1600 cm-l corre-
sponding .to the angular deformation of molecular water appears more pro-
nounced in the turbostratic hydroxide.
From these observations the authors concluded that the water mole-
cules present in the turbostratic hydroxide were essentially in intercalar
positions between the OH groups of the hydroxide layers and linked to these
OH groups by hydrogen bonds. Such conclusions support those inferred
from X-ray diffraction.
236

Turbostractic hydroxides also exhibit several bands in the range 1600 -

700 cm-’ which have been attributed to adsorbed nitrate and carbonate ions
[33, 34(a)]. Due to the high degree of division of the turbostratic hydroxide
the presence of adsorbed species cannot be avoided and is common to all
hydrous, poorly crystallized nickel hydroxides.
In the case of electrochemically prepared “(Y” hydroxides, the presence
of large quantities of nitrate ions (both adsorbed and intercalated) is con-
firmed unequivocally in Raman spectra [ 221.

Thermal stability
The thermal decomposition of nickel hydroxide with reference to the
thermal decomposition of well-crystallized Ni(OH)2 has been studied by
Dennstedt [26] and Figlarz [ 341. Finely divided /I-Ni(OH), loses adsorbed
water below 160 “C. Hydrothermally grown &Ni(OH), with no adsorbed
water does not exhibit this water loss and begins to dehydrate above 200 “C.
By constrast OL hydroxides (and particularly the turbostratic form) exhibit
a weight loss of 30% below 150 “C, corresponding to water evolution. Most
of this water is adsorbed water, but a part of it undoubtedly comes from the
intercalar,layers of the turbostratic structure.
This conclusion arises from the fact that after a thermal treatment at
150 “C the turbostratic structure is still observed by X-ray diffraction and IR
spectroscopy, but the interplanar distance between two Ni(OH)s layers is
lowered from 8.5 A to 7 W. This variation is fully reversible by rehydration
[34(b)], and thermoanalysis suggests that the - 7 A stacked structure would
correspond to the presence of monomolecular intersheet water layers.
Elimination of these intercalar water molecules and dehydroxylation of the
hydroxide sheets then proceed in one single step for temperatures slightly
higher than 200 “C, i.e., in the same temperature range as fl-Ni(OH), . This
confirms that intercalar water molecules of the turbostratic structure are
hydrogen bonded to OH groups.
NMR and impedance measurements [ 231 support thermal decomposi- .
tion results. First, ‘H NMR studies show unequivocally the existence of two
types of protons:
- one, associated with a broad absorption line, remains unchanged
when temperature is increased; it corresponds to Ni(I1) hydroxyl ions or to
“bonded” inter-sheetwater.
-the other, a narrow absorption line, disappears with increasing tem-
perature and corresponds to adsorbed or “mobile” intersheet water.
Furthermore, whereas the conductivity of turbostratic hydroxides is
around 10e5 R-l cm- l at room temperature, it falls to below lo-’ Sl-l
cm-l (at the same temperature) once samples have been heated at 100 “C or
dehydrated under high vacuum.

The transformation of “a”nicke1 hydroxides into p-Ni(OH),

The OL + 0 recrystallization reaction in water can be compared with the
complex structural evolution of nickel hydroxy-salts upon either ageing in
 237

their mother solutions [ 81 or hydrolysis [27(a)] .

The crystal growth of microcrystalline nickel hydroxide has also been
investigated by Bagno and Longuet-Escard [lo], who described their pri-
mary hydroxide particles as hexagonal thin plates (thickness 10 A, size 100
a) and studied the ageing of these primary particles in the mother solution.
Crystal growth would proceed by joining of the primary particles.
Le. Bihan [ 351 proposed a different mechanism to explain the transfor-
mation of a turbostratic nickel hydroxide into fl-Ni(OH), in pure water.
X-ray diffraction, IR spectroscopy, electron microscopy and selected area
diffraction studies on the intermediary products support the hypothesis that
the slow, room temperature transformation proceeds by a biphasic mecha-
nism. The turbostratic hydroxide seems to dissolve slowly in water and /3-
Ni(OH)s appears by nucleation and growth from the solution.

Discussion - nickel hydroxides and NOE

The difficulty of obtaining well-crystallized fl-Ni(OH), by direct preci-
pitation explains why &Ni(OH), currently used in pocket plates or in sin-
tered plates cannot but exhibit structural and textural defects. This lack of per-
fection in crystallization must however be taken as the sine qua non reason for
Ni(OH), electroactivity. Unfortunately, to our knowledge, no correlation
has ever been drawn between crystallization advancement and electro-
activity, although studies such as Louer’s X-ray diffraction line profile
analyses on both hydroxy-salts [27(a)] and nickel hydroxides [27(b)] might
provide a useful tool for such an investigation.
As for cu-Ni(OH)z, even if many battery manufacturers have seriously
worked on the project (essentially by cathodic deposition), it has never been
really exploited industrially. Depending on both the electrolytic bath and
current densities, the phases obtained exhibit various salt contents and
various inter-layer distances. On the other hand, chemically prepared (Y
phases can be structurally controlled with more accuracy, but it becomes
difficult to follow their evolution once extra water may enter and leave the
intersheet space as soon as cu-Ni(OH)s is in contact with the electrolyte.
Therefore, in contrast with definite close-packed and turbostratic hydroxides
studied in this section, neither “0~” nor “0” hydroxides used by electro-
chemists can be considered as well-defined materials. This explains why
results obtained on NOE are usually presented over a range of values.

B. Hydroxides and oxy-hydroxides with a nickel oxidation state higher

than II
Although higher oxy-hydroxides have been as extensively investigated
as Ni( II) hydroxides, there still exist various controversies about their struc-
tural characterization, their oxidation state, and the similarity of phases
obtained through either chemical or electrochemical oxidation. However, in
233

an attempt to simplify their description, only two basic types of layered

oxyhydroxides, differing by their interlamellar distance, are usually con-
sidered and are conventionally written fl-NiOOH and y-NiOOH.
While fl-NiOOH is a relatively well defined material, r-NiOOH repre-
sents a whole family of compounds exhibiting a large inter-sheet distance,
of general formula A,H,(HzO),NiOz (x, y < 1). A represents alkali ions
(mainly K’ and Na+), and water molecules are intercalated between the NiO,
slabs. The oxidation state of nickel lies within the range 3 - 3.75. The diffi-
culty of characterizing such compounds is increased by the high degree of
division they usually exhibit and by disordered sheet stackings, which result
in a wide range of diffraction patterns.
In order not to present an extended list of all possible stoichiometric and
textural variations, we shall concentrate on what can be considered as well
established crystallographic grounds, bearing in mind that the purpose is to
propose the clearest possible approach to NOE mechanistic paths.
y and 0 materials can be obtained through either chemical or electro-
chemical reactions; chemical processes only will be considered in this sec-
tion.

Preparations and structures

p-NiOOH
This compound was first studied by Glemser and Einerhand [36], who
synthesized it by oxidation of nickel nitrate with K&OS in a 1N KOH
solution at room temperature.
fl-NiOOH crystallizes in the hexagonal system (a = 2.82 A, c = 4.85 A)
and can be regarded as deriving from @-Ni(OH), by a direct reaction remov-
ing one proton and one electron.
As in the case of oxides [37] the structure can be described as NiOz
sheets of edge-sharing NiOs octahedra, protons being intercalated between
the slabs, and oxygen atoms forming a hexagonal (AB) close-packing (idem
p-Ni(OH),). It is noticeable that the a parameter, which corresponds to the
Ni-Ni distance within the sheets, is significantly lower in NiOOH than in the
case of Ni(OH)z : 2.82 A us. 3.12 A. On the other hand, the increase in the c
parameter (inter-sheet distance), 4.85 W us. 4.60 A, results from the enhan-
cement of the repulsion between oxygen layers of adjacent NiOs sheets once
protons are removed.
In constrast to p-Ni(OH)z [34(a), 381, IR spectra indicate that NiOOH
is a hydrogen-bonded structure containing no free hydroxyl groups [38],
and Kober’s :H 2 fD exchange reaction tests on fl-Ni(OH), and fl-NiOOH
confirm [39] that protons are considerably more tightly bonded in the
higher oxide than in the divalent one.

y-NiOOH
The “7” denomination was first given by Glemser [36] to a compound
exhibiting a large intersheet distance and a high oxidation state obtained by
 239

hydrolysis of the products derived from melting of NaOH and NasOz in a

Ni crucible. Glemser’s -y-NiOOH crystallized in the rhombohedral system,
with corresponding hexagonal parameters a = 2.82 A and c = 20.65 A.
Years later, Bode similarly prepared “y type” oxyhydroxides by solid
state hydrolysis of NaNiO, in NaOH + Br, solutions [40]. In particular, he
obtained a phase of compositon Nao.s2H,zz(H,0)o.z5NiO~, corresponding to
a mean oxidation state of nickel equal to 3.46.
The repetition of the same procedure on a single crystal of NaNiO,
allowed Bode et al. [41] to obtain a single monoclinic crystal of Na,ss-
(HzO),s,NiO,, with parameters a = 4.90 A, b = 2.83 A, c = 7.17 A, p =
103.1”. The distortion from the ideal rhombohedral cell was small, and the
pseudo-hexagonal parameters (a = 2.83 A, c = 20.95 A) are close to those
given by Glemser for r-NiOOH [36]. In fact, depending on the experimental
conditions, a whole family of compounds having roughly similar X-ray dif-
fraction patterns, but oxidation states of nickel ranging from 3.3 to 3.75,
and different water and alkali cation contents within the sheets, can be ob-
tained. Bode called such phases “yl ” as opposed to the “yz” obtained from
direct oxidation (NaOH + Br,) of either Ni(NOs)z solutions or a-Ni(I1)
hydroxides.
Basically, y1 and y2 oxy-hydroxides belong to the same structural
family but exhibit different degrees of crystallinity, due essentially to dif-
ferent synthesis routes. While y1 phases are prepared from highly organized
sodium nickelate through a direct reaction which allows all 3 elementary
reactions (oxidation, cation exchange, water intercalation) to proceed
simultaneously, y2 phases are obtained at room temperature either from
Ni2+ solutions or from poorly crystallized a-Ni(OH)2.
Two highly oxidized potassium nickelates, K,SNiO,s and K,i4Ni02,
have also been prepared by Bityutskii [42] through oxidation of Ni(OH),
by bromine at 170 “C in concentrated KOH; their hexagonal cell parameters
are given in Table 1. These authors have also obtained the sodium nickelate
Nao.mNi1.s having the same structure as K,, Ni0i.s [42] .

TABLE 1

Cell parameters of “7-type” compounds

a C

(a (A)
Ko.23NiOl.8
-z%O 2.82 21.15
Ko.23NiOl.a 2.82 19.12
KoS14Ni02 - zH20 2.82 21.15
K0.14NiO2 2.82 19.15
“7-NiOOH” [ 361 2.82 20.65
240

Crystallographic discussion
All reported r-NiOOH phases correspond to layer structures and rhom-
bohedral cells, which can be described as NiO, sheets having an octahedral
nickel ion environment.
K+ and Na+ ions, protons, and water molecules are intercalated between
the slabs.
It is interesting to relate this result to the numerous studies which have
been devoted to anhydrous alkali layer oxides A,MOs [37, 43, 441, and
which show that in the rhombohedral symmetry only two types of oxygen
packing can be found: ABCABC or ABBCCA. While in the former, alkali
ions are inserted in octahedral environments, in the latter alkali ions exhibit
a trigonal prismatic surrounding (Fig. 5). Bityutskii found an octahedral
environment for potassium in Ke.,NiO,_s and a trigonal prismatic one for
K0.14Ni02, and indicated that the diffraction patterns of Kc_i4Ni02 and
-y-NiOOH are very close.
In addition, the crystal structure determination of Na,,ss(H20),,s7Ni02
has undoubtedly led to the assumption of an ABBCCA oxygen packing for the
atoms belonging to the NiOs slabs [41] ; Na’ ions and oxygen atoms belong-
ing to water molecules are in trigonal prisms. Braconnier and Delmas found
that the X-ray diffraction pattern of -y-NiOOH obtained by hydrolysis of
NaNiOs is very similar to those recorded for anhydrous alkali layer oxides
(A,MOs) exhibiting an ABBCCA oxygen packing [ 451. Thus all the available
data seem to corroborate the hypothesis of an ABBCCA oxygen packing for
7-NiOOH.

0 (B)
0 (6)
M (C)
M (C)
0 (A)
0 IA)
A Ib)
A (bh (cl

0 (0
M I4

0 IB)
A Ia

0 (A)
M 0
0 (0

A 0)

0 (6) 0 (8)
M W M IO
0 (A) 0 (A)
-

Fig. 5. Structure of layer oxides with rhombohedral lattices. ABCABC and ABBCCA oxy-
gen packings.
 The overall redox reaction P-Ni(OH)z + r-NiOOH which is obse
under strong oxidation conditions in an electrochemical cell thus suppc:
transition from the AB oxygen packing of fl-Ni(OH), to the ABBCC
oxygen packing of r-NiOOH. Upon oxidation of fl-Ni(OH),, &NiOOI
which has the same AB packing, is first obtained. Further oxidation to
NiOOH implies that electrons are removed from Ni3+ ions (giving Ni*+)
protons from the intersheet space. At the same time water is intercalatc
protons are exchanged by alkali ions, and a sheet glide occurs, which is
trated in Fig. 6. In consequence of this structural rearrangement the in1
sheet distance increases (c P-NiOOH = 4.85 8, -+ c/3 y-NiOOH > 7 A). ’
minimum difference in the intersheet distance between both phases COI
sponds approximately to the size of a water molecule [28].

B-C-A

A-0 +C

B-G

rA-B
L
c B

Fig. 6. Transition between AB and ABBCCA oxygen packing by gliding of the seca
third sheets.

C. Structural approach of the NOE behaviour

The transitions which can take place in the nickel electrode can
summarized, as previously shown by the Bode diagram: (l), (2) (3).

/3-Ni(OH), &%? fl-NiOOH + H+ + e-

f
cu-Ni(OH)z r_NiOOH + H+ + e-
(3)
Reactions (l), (2) and (3) can now be discussed on a strictly structural

(a) Cycling between p-Ni(OH), and p-NiOOH

The NOE can be cycled between these two phases by avoiding ow
charge. To illustrate the involved topochemical mechanism the redox n
tion can be written:
/3-HzNiOa ?ft fl-H2_xNiOa + ZCH++ 3ce-.
Several studies have been carried out in order to determine whether tht
tion is homogeneous or heterogeneous. X-ray diffraction studies sugges
242

the reaction is heterogeneous. Only @-Ni(OH), and /3-NiOOH are apparently

present during charge or discharge and no continuous variation of cell param.
eters between both limits is observed [ 441. Kober’s IR analyses support this
hypothesis; that is, there is no shift of the vibration frequency of the OH
groups upon oxidation, and the IR spectrum of P-NiOOH appears continu-
ously during the charge [ 381.
Barnard’s investigation of the potential evolution of the nickel elec-
trode during charge indicates a variation of the potential between states 2
and 2.25, while it remains constant for oxidation states higher than 2.25.
This suggests that the reaction would be homogeneous between 2 and 2.25
with a material closely related to P-Ni(OH), . For further oxidation the reac-
tion would then be heterogeneous, in close agreement with X-ray and IR
data. However, direct determination of the oxidation state of the coexisting
phases proves extremely difficult because of either 0s evolution or the
formation of a small amount of y phase [46], though X-ray and IR studies
tend to indicate that stoichiometries are relatively close to HsNiO, and
HNiOs , respectively [ 36,38,47] . In consequence, it should be noted that
the intermediate phase H 1.33Ni02[Nis04(0H)4] suggested by several groups
of workers [36, 481 has never been observed during the electrochemical
process [ 441 .

(b) Oxidation of p-NiOOH and a-Ni(OH)2 to yNiOOH

High charge rates and prolonged overcharge of fl-Ni(OH), in KOH, as
well as the ageing of P-NiOOH, favour the formation of y-NiOOH.
To account for the lower discharge potential of r-NiOOH with respect
to B-NiOOH [46,49, 501, it may be of interest to compare this result with
those obtained with A,MOs layer oxides. Recent studies have been done on
A,MOs phases (M = 3d elements) used as cathodes in electrochemical cells
of the type: A/AClO, + propylene carbonate/A,MO, (A = Li, Na), [ 51, 521.
The potentials of such cells depend obviously on the nature of M and on the
intercalation ratio x. But the most important parameter influencing the
potential is the nature of the alkali ion.
Due to the ionic character of the M-O bonds, non-stoichiometric layer
oxides are very unstable. The repulsions between adjacent oxygen layers are
very important and the A,MOs phase stability varies with the intersheet
distance, i.e., the alkali radius.
Considering two cathodes differing by the nature of the alkali ion, the
smaller potential is observed (at the same oxidation state of M) for that with
the largest alkali ion.
On charge, the potential increases to the upper value fixed by the
stability of the electrolyte. This means that a cathode with a large intersheet
distance will reach higher oxidation levels than one with a small intersheet
distance [ 531.
Such results can be transposed to the case of nickel hydroxides. The
greater intersheet distance of r-NiOOH with respect to p-NiOOH (a7 A VS.
4.85 A) results in a smaller equilibrium potential and a larger capacity.
 243

Inevitably, this raises the question of why -y-NiOOH does not form directly
from P-Ni(OH), upon oxidation. It can reasonably be assumed that it is
mainly a matter of kinetics and steric hindrance. While the oxidation of fl-
Ni(OH), to P-NiOOH only implies the removal of one proton and one elec-
tron, with no other structural change except a slight cell parameter variation,
the oxidation to r-NiOOH requires the intercalation of Ha0 molecules, the
exchange of protons and alkali ions, and a sheet glide. As a consequence a
higher energy barrier exists and fl-NiOOH is first obtained, y phases appearing
ultimately upon overcharge or ageing in KOH.
This reaction scheme is corroborated by the fact that y-type oxy-hydro-
xides are formed directly from a-Ni(OH), since, in this case, the intersheet
distance is already expanded and there is no need to intercalate additional
water.

D. Thermodynamic and kinetic approach of the NOE behaviour

1. Thermodynamic properties
Since the pioneering work of Conway [ 541 who first established that
the potential of charged nickel hydroxide electrodes is well above the revers-
ible oxygen potential, and who derived (from e.m.f. decay measurements)
the first trustworthy results concerning the thermodynamic properties of Ni
hydroxides, considerable progress has been achieved in the field.
Reversible potentials have been measured for pn/pm and cll/y systems
over a wide range of KOH concentrations and for various states of charge.
Ageing of nickel hydroxide electrodes has also given rise to a number of
recent papers by Bramham [ 561, Prikryl [ 571 and Barnard [ 241.

l(a). Reversible potentials of the fin/pm and a/y systems

Both @r//3’s and a/-y systems in their “activated” and “deactivated”
forms have been exhaustively investigated by Barnard [ 551 who also estab-
lished the influence of KOH and Ha0 activities on the reversible potentials.
To mention the “activated” forms only, he derived the equations
a+Eg= 0.3919 - 0.0139 log aKoH + 0.0386 log aH,$)
(w.r.t. Hg/HgO/ KOH),
/I + Eg = 0.4428 - 0.0028 log axon + 0.0315 log cn,o,

but an important contribution of his work was to establish that both (Yand p
systems can exist in a wide variety of forms associated, respectively, with the
potential ranges 0.392 - 0.440 V and 0.443 - 0.470 V, which confirms elec-
trochemically the great variations observed in the structure and the texture
of both systems.
As seen above, it was found that those potentials were independent of
the degree of oxidation of the nickel cation over an appreciable range of
244

oxidation states, which supports, in both cases, the hypothesis of a hetero-

geneous reaction.
One major difficulty in studying such systems is obviously the inter-
ference of oxygen evolution both on charge and self-discharge.
Conway and Bourgault [ 541 were the first to tackle the problem of
self-discharge. They proposed a scheme of two consecutive processes,
M+OH-+MOH+e-
MOH+OH-+MO+HzO+ e-
MO + MOH -+ MHOz + M (or MO.+ OH- + MHO, + e-)
MHO, + MOH + 2M + Ha0 + O2
(M refers to the surface oxides), to account for experimental evidence that
the rate controlling process in self-discharge of the nickel oxide electrode is
the anodic partial reaction of oxygen evolution.
From potentiodynamic investigations, however, Paszkiewicz [ 581
noticed that it was not until the electrode had stayed on open circuit for
24 h that its capacity decreased. By contrast, the electrode capacity increased
in the first hours following the charge and a potential decrease was observed.
This could be interpreted in terms of equalization of proton concentration
in the solid phase, with subsequent reduction of the most oxidized surface
oxides and increase of the oxidation state of the bulk oxyhydroxides
(together with possible structural rearrangements, as we have seen before).
The mechanism of oxygen evolution on overcharge has also been widely
investigated, particularly on thin oxide films [ 59 - 621 to minimize mass
transfer effects through the bulk hydroxides. For instance, Bronoel’s mecha-
nistic approach [61] led him to assume a first step of OH- discharge fol-
lowed by a chemical recombination of OH,, and OH- ions to give O&
which is then oxidized. The assumed mechanism takes account of the varia-
tion of the current as a function of the OH- activity in solution and of the
existence of two Tafel regions in the log i = f(a) curves.

1 (b). Electrode ageing

At least four major causes can account for the observed evolution per-
formance of nickel hydroxide electrodes upon ageing:
(a) the chemical transformation ar-Ni(OH), + fi-Ni(OH)z;
(b) the active mass textural evolution due to the lattice modifications
arising from the sequential redox processes fi-Ni(OH)z + P-NiOOH +
r-NiOOH;
(c) entropic order-disorder phenomena associated with metastable
states created by heterogeneous phase reactions;
(d) the structural and stoichiometric perturbations induced by the
introduction of alkaline ions in the active mass.
(a) The (Y+ 0 transformation, already studied in Part A (chemical evolu-
tion) has been followed by voltammetric sweeps by McArthur [63], Arvia
[ 641, and Barnard [ 241. McArthur and Arvia attributed the replacement of
 245

the anodic peak at 0.390 V by a peak at 0.460 V to the direct transforma-

tion cx + /3’. Barnard, on the other hand, advocated that cw-Ni(OH)2 first
transforms, by a first order reaction, into a “deactivated” (Yphase which
shows the D-type diffraction pattern but has the same marked dependence
of oxidation/reduction potentials on alkali and water activity as the OLphase,
and leads to r-NiOOH as the oxidation product.
The final transformation of cxto p occurs after a time or a number of
cycles. which depends on the initial way of precipitation of the OLphases,
electrochemically precipitated phases being much more stable than chemi-
cally precipitated phases.
(b) The transformation on cycling of electrochemically precipitated cu
phases has been followed by voltammetric sweeps in a range of KOH concen-
trations [ 651. In the absence of overcharge, the charge/discharge reactions
change from the u/y system to the /I-Ni(OH),/NiOOH system. The reaction
takes place in two steps (Fig. 7):

IImA

Fig. 7. Evolution from the a/‘y to the on/pm system. Voltammetric sweeps of electro-
chemically precipitated (Yphases in a sintered nickel electrode. 8N KOH; potential sweep
0.1 mV s-l ; Ref. Hg/HgO. Curve 1 -: cycle 1, al/y system; curve 2 -L -: cycle 3,
fi phase with low charge/discharge potentials; curve 3 . . .: cycle 65, typical fi phase.
246

(i) Change from the a/r system (curve 1) to a phase system with peak
shape characteristics of p phases but lower charge and discharge potentials
(curve 2). This step could be identified with the production of Barnard’s
[24] “deactivated” (Yphase in the discharged state. It may be added that the
same phenomenon exists in the charged condition.
(ii) Progressive increase of the charge/discharge potentials with corre-
sponding decrease of the chargeability (curve 3). This second step is a slow
one, and the charge potential finally remains lower than that reached upon
cycling with chemically precipitated fl-Ni(OH), . This transformation can be
associated with point (c).
(c) Order/disorder phenomena in the p-Ni(OH),/NiOOH system, which
have been studied voltammetrically on chemically precipitated hydroxides.
Depending on ageing, state of charge, and rate of charge, the discharge
potential of the phase may vary over a range of 40 mV (cf. Fig. 8).

0.4 E(V)

Fig. 8. Influence of the state of charge on the discharge potential of fl NOE; cathodic
voltammetric sweeps are recorded after galvanostatic charges at the C/5 rate in 5.2N
KOH. -:3hcharge;---:7hcharge;...: peaks of pm and y obtained after pro-
longed ageing. (Potential sweep 0.1 mV s-l ; ref. Hg/HgO.)

2. Electrochemical mechanisms and kinetics

E-Log i investigations for the nickel hydroxide electrode reveal com-
plex features for which there are not yet any completely satisfactory ex-
planations. However, it seems that the reaction kinetics in the solid phase
are certainly responsible for composition differences between the bulk
hydroxide and the surface.
The best documented approach, based on thermodynamic considera-
tions, is that of Barnard [ 551, which applies the thermodynamics of mixing
to account for the independence of the discharge potential of the NOE as a
function of nickel oxidation state. Constant potential regions could then
derive from heterogeneous equilibria between pairs of co-existing phases,
both containing nickel in upper and lower states of oxidation. However, such
a model remains essentially macroscopic and most electrochemists are more
interested in investigations of charge transport associated with the solid
state process of the overall reaction:

Ni(OH)s + xOH- =+NiOsHs, + xe- + xH,O.

 247

Three main approaches have been developed and tested (mainly on OL

hydroxides):
(a) McArthur [ 631 assumed that the charge-transfer was reversible and
that the reaction rate was controlled by proton diffusion. Linear sweep
techniques and potentiostatic pulses on u-Ni(OH)2 electrodes led him to
propose diffusion coefficients varying from 2 X lo-’ cmW2 s-l to 4.6 X
10-11 cm-2 s-1
and an activation energy of about 2.2 kcal-deg-1 mol-’ in
the temperature range 25 - 70 “C. Similar results were obtained by
Tysyachnyi [66] on thin 01 films, assuming a single phase model, and by
Briggs [67] .
(b) Takehara [68] considered that the rate determining step of the
reaction was the diffusion process of protons and/or defects in the hydroxide
layer. However, his model was based on the hypothesis of a variable surface
concentration of proton sites and vacancies set at the electrolyte interface by
the charge-transfer reaction. He supported his model with both potential
decay and electrode impedance measurements and derived diffusion coeffi-
cients which are quite similar to those obtained by McArthur.
(c) Feuillade [69], by contrast, claimed that anion (OH-) transfer was
predominant over proton transfer in the oxidation of electrodeposited nickel
hydroxide. He deduced such conclusions from 160 - 180 and ‘H - 3H isotopic
exchange measurements which made it possible to calculate anionic ex-
change currents of 60 PA/cm2 and 40 PA/cm2 for Ni(I1) and Ni(II1) hydrox-
ides, respectively.
Though direct OH- mobility and exchange may seem improbable in the
(supposedly) topochemical reaction fi-Ni(OH), ?? /3-NiOOH+ H’ + e-, the
hypothesis of oxygen exchange in the a/y system is worthy of consideration,
as this phenomenon is certainly plausible. Thus, Feuillade’s best contribution
may have been to establish the possibility of a dual type of transport (H’ and
OH-/H,O) in the solid phase, the relative importance of both types of trans-
port depending most probably on the interlayer distance and the subsequent
presence of “free” or “bound” water. It must be pointed out that, from a
kinetic point of view, the (Y/Yredox transfer appears more rapid than the
pn/psrtransfer. If we assume that proton mobility is the rate determining
step in the pn/pmreaction, the kinetics difference stresses the importance of
interlayer water as a charge transport factor in the a/r system.
In fact, a major weakness of all these models lies in the general implicit
assumption that a constant mechanistic path prevails all along the redox
process. This is undoubtedly a rough approximation if we accept the idea
that the NOE behaviour involves more than two oxidation states. Indeed,
both Labat’s [70] magnetic measurements and Aleshkevich’s [ 711 or
Dibrov’s [ 721 theoretical energy considerations would tend to confirm the
hypothesis that divalent nickel is directly converted to tetravalent nickel
(which then reacts in the solid phase), and that relatively stable high-valency
states of nickel can exist in charged electrodes 1731.
This raises the long debated arguments on NOE faradaic behaviour:
(i) maximum state of oxidation in the bulk?
248

(ii) role of the electrolyte concentration?

(iii) role of the cations?
(iv) cause of residual capacity?
One point is now well established about nickel hydroxide electrodes:
on cycling, electrodes do not revert to the Ni(I1) state. An average oxidation
state of 2.3 is generally accepted in the literature and is interpreted as the
basic level of Ni(II1, IV) defects required to give the hydroxide bulk suffi-
cient electronic conductivity and defect homogeneity. &Ni(OH), has a very
poor electronic conductivity due to the band structure of nickel in the di-
valent state and to the large intercation distance dNi__Ni = 3.12 A. By con-
trast, the shorter Ni-Ni bonds (2.86 A) of NiOOH favour a better orbital
overlapping.
The second (low potential) plateau observed when forcing the active
material to return to the divalent state is thus a non-equilibrium process that
might be compared with the behaviour of a mixed p.n. semiconducting
material under reverse bias conditions [ 491. The well-known “forming pro-
cess” of NOE is therefore related to the necessity to create enough stoichio-
metric and textural defects in the bulk hydroxide to ensure activation.
It is in connection with the oxidized states that most controversies
exist, however, both on the structure of the compounds and on the validity
of the oxidation values derived from either iodometric measurements of
active oxygen or electrode faradaic capacities. However, experimental data
undoubtedly show that a! and fl hydroxides exhibit different charging
abilities and can reach various oxidation states, depending on structural and
textural conditions as well as on charge characteristics (see Fig. 9, the in-
fluence of the electrolyte and of overcharge upon the capacity delivered by
a P-Ni(OH), electrode).

mAh g’N~(ll)

3 ia k KOH
Fig. 9. Influence of the KOH concentration upon the capacity of fl NOE. Discharges are
run at the C/5 rate. -: After 1.4 C overcharge (C/5 rate); - - -: after 4.8 C overcharge
(C/10 rate).
 249

More than 1 e- can be exchanged in an apparently reversible way (thus

involving at least a transition between states 2.3 and 3.3), and there is no
reason to believe that the excess capacity comes from the reduction of oxy-
gen or hydroperoxide molecules on the active material (cf. Section C).
Nevertheless, neither IR analysis [ 38, 741, X-ray diffraction, nor DTA [ 751
have been able to provide clear conclusions on the ultimate oxidation states
of nickel in NOE, part of the difficulty being the instability (evidenced by
self-discharge) of the highest states obtained upon overcharge.

E. Role of the electrolyte on the NOE behaviour

The electrolyte plays an essential part in the NOE working. Since it is in-
volved in the material transfer to and away from the electrodes by the exchange
of both OH- ions and alkali, it controls points as important as charge accep-
tance, charge and discharge potentials, structural changes, electrode ageing, . . .
Indeed, charge acceptance depends greatly on the oxygen evolution potential,
and the characteristics of the electrode redox transfers can be largely af-
fected by the possibility of alkali intercalation within the crystal lattice.

1. Oxygen evolution and charge acceptance

Rubin and Baboian [ 761 studied the electrochemical behaviour of nick-
el hydroxide in binary aqueous alkali hydroxides, MOH-Ha0 (M+ = Li+,
Na+, K+, Rb+,Cs+) for the temperature range -40 - +60 “C (at concentrations
corresponding to the maximum conductance), and observed maxima in the
“capacity vs. temperature” curves for each electrolyte.
The best capacities are obtained with sodium and lithium electrolytes
at temperatures higher than 30 “C, with potassium at intermediate tempera-
tures, and with rubidium or cesium at low temperatures. Such results were
explained as superimposed variations in electrode polarization and oxygen
evolution with temperature, both variations being partly interpreted in terms
of electrolyte interactions.
The oxygen evolution overpotential proves to be much more sensitive
to changes in the concentration or the nature of the ionic electrolyte popula-
tion than the Ni(II)/Ni(III, IV) system, and it is thus possible to optimize the
chargeability of NOE for a whole range of temperatures (-40 - +50 “C).
Hence, small quantities of lithium hydroxide have been added to standard
KOH electrolytes almost since Edison’s work, to shift the OER (Oxygen
Evolution Reaction) to more positive potentials and improve charge accep-
tance. Kelson et al. [77] found that the use of lithiated electrolytes simul-
taneously lowers the nickel hydroxide oxidation potential, but that the in-
crease of charge acceptance is partly compensated by the end of discharge at
a higher oxidation level. Bonnaterre [ 781 gave further evidence that the use
of lithiated electrolytes could increase the capacity of batteries up to 20%
but it also increased self-discharge.
250

2. Influence of the cations on the structure and the electrochemistry of the

oxidized phases
Harivel [79] showed that the -y phases are easily obtained on over-
charging in KOH and that their formation is promoted by the use of concen-
trated solutions.
Similar results are observed in NaOH, but y phases (defined as layered
compounds with intersheet distance >7 A) cannot be obtained in pure
lithium hydroxide electrolytes, and divergent results have been published in
the case of RbOH and CsOH [ 50, SO]. Such results underline the pre-
eminent role played by the alkali cations in the whole redox process, through
either their size, their acidity level, their stabilizing effect, etc.

The case of LP
Lithium nickelate is the only phase obtained upon oxidation in pure
LiOH electrolytes. Penetration of Li” in the crystal lattice is interpreted as an
exchange reaction between Li+ and protons, and Guliamov [81] brought
theoretical support to this hypothesis.
Lithium nickelate displays a low potential discharge peak and exhibits
a poor reversibility, especially at low temperatures [ 821. The discharge takes
place through a homogeneous reaction within a single solid phase, as shown
both on the X-ray diagrams (the structure of lithium nickelate remains ap-
parent, and no Ni(OH)2 is noticed until almost the end of discharge) and by
the shape of the discharge curve [82]. Due to the absence of y-phase, the
capacities of the electrodes remain low in pure LiOH. Thus lithium must be
used with extreme caution since, depending on experimental conditions, it
can either improve or poison NOE operation. In KOH solutions containing
lithium as an additive, capacity increases with the addition of LiOH up to
an atomic ratio Li/Ni = 0.06, and decreases beyond this point. The maximum
capacity is displaced towards higher values of the Li/Ni ratio with increasing
temperatures [ 831. At low Li+ concentrations, p and y phases are formed,
whereas lithium nickelate is obtained at higher Li+ contents.
All three phases have been obtained simultaneously in 7N KOH-1N
LiOH upon moderate overcharge, and characterized by their three voltam-
metric peaks [ 651 . Lithium nickelate discharges at a potential 50 mV lower
than the y phase. Its formation proceeds with a decrease in the apparent
degree of oxidation of nickel in the charged electrode [ 831 and an increase
in the discharged electrode [ 741. The rate of formation of lithium nickelate
is also increased with the total OH- concentration of the electrolyte.
The intersheet distance of lithium nickelate, c/3 = 4.7 A with respect to
hexagonal axes [ 841, is considerably smaller than that of y-NiOOH, and the
general observation (particularly true in the case of pocket plates) that the
addition of LiOH to KOH improves NOE longevity may well find its origin
in a mechanical swelling limitation. The beneficial effect of lithium against
NOE poisoning by iron (or ferrate ions) can (all the same) be interpreted on
the same basis, i.e., the impossibility of large radius species penetrating the
crystal lattice.
 251

From an electronic point of view the beneficial effect of LiOH addi-

tions to KOH is frequently attributed to p-type semiconduction enhance-
ment in the charged phase, as noted in lithium doped NiO (fl-NiOOH being
on the other hand an n-type semiconductor [ 85]), but a clear interpretation
of the conduction process of such disordered systems does not seem to be
at hand. However, Takehara [68] concluded from charge and discharge
activation energy measurements that the diffusion rate of protons (and/or
defects) in the crystal lattice was affected by the presence of Li’ ions, thus
showing a specific kinetic influence of lithium on the redox transfer.

F. The role of foreign metallic cations on the NOE behaviour

A review of NOE cannot fail to devote some lines to this important

point. In fact, Harivel [ 861, Casey [87], Doran [88], and many others have
studied the influence of numerous foreign cations on NOE: Ag(I), Co(II),
Mn(II), Ba(II), Zn(II), Pb(II), Mg(II), Cd(II), Al(III), As(III), Sb(III), Bi(III),
Mn(IV), Pb(IV), Si(IV), . . . The results have been either totally negative or
selectively positive, with detrimental side effects on self discharge, faradaic
behaviour, charge and discharge potentials, or durability. Cobalt and cad-
mium, however, because of their practical interest, deserve special attention.
Cobalt hydroxide, which can syn-crystallize in the same OLand p forms
as nickel hydroxide, is commonly used by sintered plates manufacturers to
improve charge acceptance of nickel electrodes, particularly at high tempera-
tures. Cobalt (Co/Ni = 2 - 5%) shifts the reversible potential of NOE some
20 - 40 mV cathodically and improves the electrode stability. Voltammo-
grams [ 891 clearly indicate that the charging process occurs more reversibly
in the presence of coprecipitated cobalt hydroxide than in its absence. Since
charge transfer appears to be controlled primarily by mass transport through
a solid in which the conductivity depends upon the state of charge it can be
presumed that cobalt (which is certainly present in the oxidation state III)
optimizes the lattice imperfections in the active material so that conductivity
is substantially increased. However, large quantities of cobalt would tend to
inhibit the NOE behaviour since cobalt hydroxide exhibits a low redox rever-
sibility [ 31. Similar overall results are also claimed for barium.
In the case of cadmium, Ness [ 901 and Cittanova [ 911 have shown that
coprecipitation of cadmium hydroxide partially inhibits swelling and
promotes the NOE initial activity. The presence of Cd% ions thus seems to
act on the conductivity and maintain the cohesion of the NiOz slabs, while
forbidding the introduction of water and alkali within the crystal lattice. The
formation of y phases is thus delayed; consequently the oxidation of nickel
to states higher than III is made difficult, and there is a loss in charge accep-
tance.
Voltammograms obtained from electrodes containing coprecipitated
nickel and cadmium hydroxides have confirmed these results [65].
252

Conclusion

If we review the principle ideas resulting from a consideration of this pa-

per, we find that despite the recent gain of a deeper and closer insight into the
NOE electrochemical behaviour through, for instance, the measurement of Ea
potentials in different electrolytes, or the determination of the homogeneous
or heterogeneous character of the redox process, a number of basic questions
still remain partially or totally unanswered. In particular, it may seem sur-
prising that in spite of the tremendous amount of experimental data ac-
cumulated over past years, we are unable to characterize charged and un-
charged electrodes in a precise manner in the various possible electrolytes.
We have seen, however, that a strictly structural approach to the redox
reaction can throw some light on the different redox routes, and particularly
on the “7” phenomenon, by relating it to better known solid state proper-
ties. It is reasonable to assume that, in the next few years, solid state electro-
chemistry will help to find answers to the main problems posed by the redox
process and its interaction with the solvent, namely:
(i) The role of the electrolyte in the oxidation reaction: the importance
of the cation in the structural changes is well known, but how does it take
place?
(ii) Transport modes: how are these influenced by the respective
populations of protons, OH- ions and alkali ions?
(iii) Lithium, cobalt, barium . . .: how do they modify the conduction
process?
(iv) Controversy over the y phase(s): can the 7 phase(s) be defined by
a precise formula (an hypothesis supported recently by Barnard [ 80]), or at
least can we propose definitions in better conformity with reality?
Finally, this paper, which results from close collaboration between solid
state chemists and electrochemists, emphasizes two basic truths:
(a) As in any other chemical reaction, the redox reactions which occur
in the NOE obey fundamental steric and stoichiometric laws. The com-
plexity of all possible mechanisms, together with the existence of interfering
reactions (such as oxygen evolution), have engendered a belief among
electrochemists that NOE operation is not based on well defined reactional
grounds. By contrast, we have stressed that, despite an apparent state of con-
fusion on the nature and structure of all involved phases, general steric
mechanisms could be derived from the available data.
(b) Electroactivity and electrochemical characteristics cannot be inter-
preted other than in close connection with structural and textural param-
eters. This justifies the importance that we have given in this paper to the
early sections, to stress that there may exist as many structural and electro-
chemical differences within the same “(Y” or “0” family (e.g., between a
chemically prepared turbostratic hydroxide and an electrochemically
deposited “a-type” hydroxy-nitrate) as between (Yand /3hydroxides taken as
generic entities.
 253

A basic knowledge of the starting materials and of their chemical redox

reactivity is thus essential to a better understanding of the NOE behaviour.

Acknowledgements

The authors thank Drs C. Fouassier and G. Feuillade for interesting

discussions and helpful support.

References

1 H. Bode, K. Dehmelt and J. Witte, Electrochim. Actu, 11 (1966) 1079.

2 P. C. Milner and U. B. Thomas, in C. W. Tobias (ed.), Advances in Electrochemistry
and Electrochemical Engineering, Vol. 5, Interscience, New York, 1967, p. 1.
3 G. W. D. Briggs, Chem. Sot. Spec. Period. Rep., Electrochemistry, 4 (1974) 33.
4(a) S. U. Falk and A. J. Salkind, Alkaline Storage Batteries, Wiley, New York, 1969.
(b) S. U. Falk, in M. Barak (ed.), Electrochemical Power Sources, IEE Publications, P.
‘Peregrinus Ltd., 1980,~. 324.
5 A. J. Bard, Encyclopedia of Electrochemistry of the Elements, Dekker, New York,
1975, p. 349.
6 J. P. Hoare, The Electrochemistry of Oxygen, Interscience, New York, 1968.
7 H. R. Oswald and R. Asper, in R. M. A. Lieth (ed.), Preparation and Crystal Growth
of Materials with Layered Structures, Reidel, Dordrecht, 1977, Ch. 3, p. 109.
8 W. Feitknecht and A. Collet, Helu. Chim. Acta, 22 (1939) 1428; 23 (1940) 180.
9 A. Berger, Kolloid Z., 103 (1943) 13; 104 (1944) 24.
10 D. Bagno and J. Longuet-Escard,J. Chim. Phys., 51 (1954) 215.
11 E. Suoninen, T. Juntunen, H. Juslen and M. Pessa, Acta Chem. &and., 27 (1973)
2013.
12 A. Merlin and S. J. Teichner, C.R. Acad. Sci., 236 (1953) 1892.
13 F. Fievet and M. Figlarz, J. Catal., 39 (1975) 350.
14 F. Fievet, These Doctomt d%tat, Paris VII, 1980.
15 J. Deportes, P. Mollard, J. Penelon, S. Le Bihan and M. Figlarz, CR. Acad. Sci., Ser.
B, 272 (1971) 449.
16 J. Deportes, These 3eme cycle, Grenoble, 1971.
17 L. Neel, Nuovo Cimento, S-X, Suppl., 3 (1957) 942.
18 J. D. BernaI and H. D. Megaw, Proc. R. Sot. London, Ser. A, 151 (1935) 384.
19 W. R. Busing and H. A. LBvy,J. Chem. Phys., 26 (1957) 563.
20 S. S. Mitra, Z. Kristall., Bd 116 (1961) 149; Solid State Phys., 13 (1962) 1.
21 C. Cabannes-Ott, Ann. Chim., 5 (1960) 905.
22 M. Couzi, F. Cruege and P. Oliva, unpublished results.
23 J. C. Brethous, J. J. Braconnier, G. Villeneuve and P. Oliva, unpublished results.
24 R. Barnard, C. F. Randell and F. L. Tye, 12th Int. Symp. Power Sources, Brighton,
1980, Academic Press, London, 1981.
25 S. Le Bihan, J. Guenot and M. Figlarz, C.R. Acad. Sci., Ser. C, 270 (1970) 2131.
26 W. Dennstedt and W. Loser, Electrochim. Acta, 16 (1971) 429.
27 (a) D. Lou&, M. Lou&, D. Grandjean and A. Le Bail, Acta Crystallogr., B 29 (1973)
1696, 1703, 1707; Rev. Chim. Miner., 17 (1980) 522; J. Solid State Chem., 13
(1975) 319.
(b) D. Lou&, D. Weigel and J. I. Langford, J. Appl. Crystallogr., 5 (1972) 353.
28 R. S. McEwen,J. Phys. Chem., 75 (1971) 1782.
29 L. Kandler, Brit. Pat. 917,291 (1963).
254

30 E. J. McHenry, Electrochem. Technol., 5 (1967) 275.

31 R. L. Beauchamp, US. Patent 3,653,967.
32 S. Le Bihan and M. Figlarz, Electrochim. Acta, 18 (1973) 123.
33 S. Le Bihan, These, Paris, 1974, No. CNRS A0 9424.
34 (a) M. Figlarz and S. Le Bihan, C.R. Acad. Sci., Ser. C, 272 (1971) 50.
(b) S. Le Bihan and M. Figlarz, Thermochim. Acta, 6 (1973) 319.
35 S. Le Bihan and M. Figlarz, J. Cryst. Growth, 13/14 (1972) 458.
36 0. Glemser and J. Einerhand, 2. Anorg. Allg. Chem., 261 (1950) 26.
37 C. Delmas, C. Fouassier and P. Hagenmuller, Physica, 99 B (1980) 81.
38 F. P. Kober, J. Electrochem. Sot., 112 (1965) 1064.
39 F. P. Kober, J. Electrochem. Sot., 114 (1967) 215.
40 H. Bode, K. Dehmelt and J. Witte, 2. Anorg. Allg. Chem., 366 (1969) 1.
41 H. Bartl, H. Bode, G. Sterr and J. Witte, Electrochim. Acta, 16 (1971) 615.
42 P. N. Bityutskii and V. I. Khitrova, Sov. Phys.-Crystallogr., 13 (1968) 40, 867; 14
(1969) 99.
43 C. Delmas, C. Fouassier and P. Hagenmuller, Mater. Res. Bull., 11 (1976) 1483.
44 G. W. D. Briggs, E. Jones and W. F. K. Wynne-Jones, Trans. Faraday Sot., 51 (394)
(1955) 1433;52 (405) (1956) 1260.
45 C. Delmas, M. Devalette, C. Fouassier and P. Hagenmuller, Mater. Res. Bull., 10
(1975) 393.
46 R. Barnard, C. F. Randell and F. L. Tye, J. Electroanal. Chem., 119 (1981) 17.
47 D. Tuomi, J. Electrochem. Sot., 112 (1965) 1.
48 S. E. S. El Wakkard and E. S. Emara, J. Chem. Sot., 4 (1953) 3504.
49 R. Barnard, G. T. Crickmore, J. A. Lee and F. L. Tye, J. Appl. Electrochem., 10
(1980) 61.
50 N. Yu. Uflyand, A. M. Novakovski, Yu. M. Pozin and S. A. Rozentsveig, Elektro-
khimiya, 2 (1966) 234; 3 (1967) 537.
51 J. J. Braconnier, C. Delmas, C. Fouassier and P. Hagenmuller, Mater. Res. Bull., 15
(1980) 1797.
52 K. Mizushima, P. C. Jones, P. J. Wiseman and J. B. Goodenough, Mater. Res. Bull., 15
(1980) 783.
53 C. Delmas, unpublished results.
54 B. E. Conway and P. L. Bourgault, Can. J. Chem., 37 (1959) 292; 38 (1960) 1557;
40 (1962) 1690. Trans. Faraday Sot., 58 (1962) 593.
B. E. Conway and E. Gileadi, Can. J. Chem., 40 (1962) 1933.
55 R. Barnard, C. F. Randell and F. L. Tye, J. Appl. Electrochem., 10 (1980) 109,127.
56 R. W. Bramham, R. J. Doran, S. E. A. Pomroy and J. Thomson, in D. H. Collins (ed.),
Power Sources 1977, Academic Press, London, p. 129.
57 E. Prikryl, 0. Rademacher and K. Wiesener, 2. Phys. Chem., 258 (1977) 113.
58 M. Paskiewicz and I. Walas, Electrochim. Acta, 24 (1979) 629.
59 B. E. Conway, M. A. Sattar and D. Gilroy, Electrochim. Acta, 14 (1969) 677,695,
711. J. Electroanal. Chem., 19 (1968) 351.
60 P. W. T. Lu and S. Srinivasan, J. Electrochem. Sot., 125 (1978) 1416.
C. R. Davidson and S. Srinivasan, J. Electrochem. Sot., 127 (1980) 1060.
M. H. Miles,G. Kissel, P. W. T. Lu and S. Srinivasan, J. Electrochem. Sot., 123(1976)332.
61 G. Bronoel and J. Reby, Electrochim. Acta, 25 (1980) 973.
62 E. A. Khotskaya, A. S. Kolosov and V. V. Polishchuk, Elektrokhimiya, 7 (1971)
1064.
63 D. M. McArthur, in D. H. Collins (ed.), Power Sources, 3, Griel Press, Newcastle upon
Tyne, 1971, p. 91.
D. M. McArthur, J. Electrochem. Sot., 117 (1970) 422,729.
64 H. Gomez Meier, J. R. Vilche and A. J. Arvia, J. Appl. Electrochem., 10 (1980) 611.
65 A. de Guibert, unpublished results.
66 V. P. Tysyachnyi, 0. S. Ksenzhek and L. M. Pototskaya, Elektrokhimiya, 8 (1972)
1692.
 255

67 G. W. D. Briggs and M. Fleischmann, Trans. Faraday Sot., 62 (1966) 3217; 67 (1971)

2397.
68 Z. Takehara, M. Kato and S. Yoshizawa, Electrochim. Acta, 16 (1971) 833,
69 G. Feuillade and R. Jacoud, Electrochim. Acta, 14 (1969) 1297.
70 J. Labat, Th&e, Bordeaux, 1965.
71 S. A. Aleshkevich, E. I. Golovchenko, V. P. Morozov and L. N. Sagoyan, Elektro-
khimiya, 4 (1968) 1237.
72 I. A. Dibrov and T. V. Grigor’eva, Elektrokhimiya, 13 (1977) 979; 15 (1979) 281.
73 V. M. Vogel, Electrochim. Acta, 13 (1968) 1815.
74 0. G. Malandin, I. S. Shamina, S. M. Rakhovskaya, I. K. Kuchkaeva, L. N. SaI’kova,
A. V. Vasev, P. N. Bityutskii, G. V. Suchkova and L. A. Vereshchagina, Elektro-
khimiya, 10 (1974) 1571,1745; 12 (1976) 573; 14 (1978) 1380; 16 (1980) 1041.
75 M. A. Aia, J. Electrochem. Sot., 114 (1967) 418.
76 E. J. Rubin and R. Baboian, J. Electrochem. Sot., 118 (1971) 428.
77 P. Kelson, A. D. Sperrin and F. L. Tye, in D. H. Collins (ed.), Power Sources 4,
Academic Press, London, 1973.
78 R. Bonnaterre-Saft, unpublished results.
79 J. P. Harivel, B. Morignat, J. Labat and J. F. Laurent, in D. H. Collins (ed.), Power
Sources 1966, Academic Press, London, p. 239.
80 R. Barnard, C. F. Randell and F. L. Tye, J. Appl. Electrochem., 11 (1981) 517.
81 Yu. M. Guliamov, M. D. Dolgushin and L. N. Sagoyan, Elektrokhimiya, 7 (1971) 896.
82 N. Yu. Uflyand, S. V. Mendeleva and S. A. Rozentsveig, Elektrokhimiya, 6 (1970)
1312.
83 E. A. Kaminskaya, N. Yu. Uflyand and S. A. Rozentsveig, Elektrokhimiya, 7 (1971)
1839.
84 L. D. Dyer, B. S. Borie and G. P. Smith, J. Am. Chem. Sot., 76 (1954) 1499.
85 D. Tuomi and G. J. B. Crawford, J. Electrochem. Sot., 115 (1968) 450.
86 J. P. Harivel, ThBse, Strasbourg, 1969.
87 E. J. Casey, A. R. Dubois, P. E. Lake and W. J. Moroz, J. Electrochem. Sot., 112
(1965) 371.
88 R. J. Doran, in D. H. Collins (ed.), Batteries: Proc. 3rd Znt. Symp., Pergamon Press,
London, 1963, p. 105.
89 D. F. Pickett and J. T. Maloy, J. Electrochem. Sot., 125 (1978) 1026.
90 P. Ness, Conferences SZE, Keilkeim, 1973.
91 J. P. Cittanova, personal communication.