(Li/Ag)CoO2: A new intergrowth cobalt oxide composed of rock salt and delafossite layers

Article abstract of DOI:10.1021/ic200534x

A new ordered (Li/Ag)CoO₂ layered compound with an unusual oxygen packing combining rock salt and delafossite layers is obtained during the (Li⁺, Na⁺)/Ag⁺ ionic exchange from the OP4-(Li/Na)CoO₂ precursor. This compound is actually an intermediate step to the final D4-AgCoO₂ delafossite and can be isolated thanks to the kinetics difference between the Li⁺/Ag⁺ and Na ⁺/Ag⁺ exchange processes. It crystallizes in the P6 ₃/mmc space group with cell parameters a_hex. = 2.848(3) A and c_hex. = 21.607(7) A. The details of the structure as well as its thermal stability and transport properties are presented and discussed.

Full text of DOI:10.1021/ic200534x

ARTICLE
pubs.acs.org/IC
(Li/Ag)CoO2: A New Intergrowth Cobalt Oxide Composed of Rock Salt
and Delafossite Layers
R. Berthelot,†,‡ M. Pollet,*,† J.-P. Doumerc,† and C. Delmas†,§
†CNRS, Universitꢀe de Bordeaux, ICMCB, 87 avenue du Dr. A. Schweitzer, 33608 F-Pessac, France
‡CEA-Grenoble, DRT-LITEN, 17 rue des Martyrs, 38054 Grenoble, France
§CNRS, ENSCBP, ICMCB, 87 avenue du Dr. A. Schweitzer, 33608 F-Pessac, France
ABSTRACT: A new ordered (Li/Ag)CoO2 layered compound with an unusual oxygen packing
combining rock salt and delafossite layers is obtained during the (Liþ, Naþ)/Agþ ionic exchange from
the OP4-(Li/Na)CoO2 precursor. This compound is actually an intermediate step to the final
D4-AgCoO2 delafossite and can be isolated thanks to the kinetics difference between the Liþ/Agþ and
Naþ/Agþ exchange processes. It crystallizes in the P63/mmc space group with cell parameters ahex.
=
2.848(3) Å and chex. = 21.607(7) Å. The details of the structure as well as its thermal stability and
transport properties are presented and discussed.
and D4-AgCoO2,29 which regularly alternates O2/O3- and
1. INTRODUCTION
D2/D3-type blocks, respectively. The structure of such polytypes
with new oxygen stacking is directly related to the original struc-
ture of the OP4-(Li/Na)CoO2 precursor that alternates two dif-
ferent AO2 blocks: a P2-type sodium block and an O3-type lithium
one (Figure 1a).30 While the site symmetry could change through
the ionic exchange, the layer organization remains the same as in
the OP4 precursor, leaving unchanged the number of CoO2 slabs
necessary to describe the final hexagonal cell (i.e., four slabs). The
exchange process may induce or not a slab gliding in the a,b plane
but never a slab rotation, as such an operation would require a
CoꢀO bond breaking that only occurs at high temperature and
was never observed in the vicinity of room temperature.
During the OP4 f O4 transformation, which was also studied
by Komaba et al.,31 only Naþ ions are exchanged and the OP4 pre-
cursor directly gives the O4 stacking without any intermediate step. In
the case of the OP4 f D4 transformation, however, both Naþ and
Liþ ions are exchanged for silver ones. During this process, we noticed
the presence of traces of an intermediate phase OD4-(Li/Ag)CoO2
combining rock salt and delafossite blocks (Figure 1b). This report
details the structure of this new OD4-(Li/Ag)CoO2 intergrowth
layered cobalt oxide, explains its formation by the kinetics dif-
ference between Naþ/Agþ and Liþ/Agþ exchanges, and pre-
sents preliminary measurements of physical properties.
The very details of the structure of layered ACoO2 oxides like
stacking sequences or local symmetries are assumed to have a key
role on their remarkable properties, such as the electrochemical
behavior of the high-temperature polytype of LiCoO2, which is
the most commercialized positive electrode material for lithium
batteries,1 or the fascinating properties reported for P2-NaxCoO2
phases depending on their sodium content2 (promising thermo-
electric properties for high-sodium compositions,3ꢀ5 supercon-
ductivity for hydrated phases at x ∼ 0.3,6,7 or metal to insulator
8
transition for the half-filled Na1/2CoO2 ).
The structure of all these materials can be described from a packing
of CoO2 slabs, formed by edge-shared CoO6 octahedra, between which
Aþ monovalent cations are intercalated in the interslab space. Depend-
ing on the nature of the A element, its content (x), and the synthesis
conditions, various intercalation sites are available: octahedral in O3-,
O2-, and O4-LiCoO21,9ꢀ12 and O3-NaCoO2,13,14 trigonal prismatic in
P2- and P3-NaxCoO2 or -KxCoO2,13ꢀ16 tetrahedral in deintercalated
bilayered LixCoO2,17 or linear dumbbell in the case of the delafossite
structure, where A is typically a d10 or d9 monovalent cation like, for
instance, a noble metal.18ꢀ24 In order to standardize the descrip-
tion of the layered ACoO2 oxides, we follow the nomenclature
proposed by Delmas et al.25 using as a prefix of the chemical
formulas the combination of a letter that stands for the A cation
site symmetry (O for octahedron, P for trigonal prism, and D for
dumbbell coordination) and a figure indicating the number of
CoO2 layers necessary to describe the structure.
2. EXPERIMENTAL SECTION
2.1. Synthesis of the Precursors. Several authors have reported
Recently, we used the ordered OP4-Li∼0.42Na∼0.37CoO2 phase,
which was originally reported for its promising thermoelectric
properties,26ꢀ28 as a precursor for ionic exchanges, and we first
on the synthesis of the OP4-(Li/Na)CoO2 precursor with some nuances
Received: March 15, 2011
Published: June 16, 2011
12
reported on the two new layered polytypes obtained, O4-LiCoO2
r
2011 American Chemical Society
6649
dx.doi.org/10.1021/ic200534x Inorg. Chem. 2011, 50, 6649–6655
|

Inorganic Chemistry
ARTICLE
Figure 1. Perspective representation of (a) the ordered OP4-(Li/Na)CoO2 precursor and the three stacking that may theoretically be obtained from
(LiþþNaþ)/Agþ ionic exchanges: (b) the OD4-(Li/Ag)CoO2 studied in this article; (c) the PD4-(Na/Ag)CoO2, which is not evidenced due to
difference in the Liþ/Agþ and Naþ/Agþ exchange kinetics; and (d) the delafossite D4-AgCoO2 obtained after the complete exchange of the alkali
ions.29 In these structures, lithium and sodium ions occupy respectively octahedral (in yellow) or trigonal prismatic sites (in light purple), while silver
ions are intercalated in a dumbbell configuration. Sodium ions accommodate two kinds of trigonal prismatic sites depending on whether they share face
(Naf, in white) or edges (Nae, in purple) with surrounding CoO6 octahedra. For convenience, only one intercalated site is sketched in each interslab
space. The oxygen stacking sequence is written next to each structure.
in the synthesis protocol and in the chemical composition, but most of
them agree on the difficulty to obtain a pure phase.26ꢀ28,30 We recently
deeply reinvestigated the synthesis path and proposed a more pertinent
protocol.32 In brief, the precursors O3-LiCoO2 and P2-Na∼0.7CoO2
were first prepared from solid-state reactions using dry alkali carbonates
and cobalt oxide Co3O4 that were intimately ground. A stoichiometric
ratiowas used for thepreparationofO3-LiCoO2, while anexcessof5 wt%
of sodium carbonate was added for the preparation of P2-Na∼0.7CoO2 in
order to balance to sodium oxide volatility at high temperature. These
mixtures were heated at 850 and 900 °C for 24 h under oxygen flow
with heating and cooling rates set at þ2 and ꢀ5 °C minꢀ1, respectively.
Following our previous work,32 these two precursors were then inti-
mately mixed together in the nominal composition Li0.42Na∼0.41CoO2
in an argon-filled glovebox. The mixture was put into a gold tube that was
quickly sealed and then heated for at least 1 day at 920 °C in a preheated
furnace. The tubewas finallyquenchedin a water bath. Suchthermal treat-
ment avoids the decomposition of the OP4-(Li/Na)CoO2 phase.32 All
the compounds being moisture sensitive, they were stored in a dry atmo-
sphere. Rietveld refinement on XRD pattern (not shown) confirmed the
synthesis of pure OP4-Li∼0.42Na∼0.37CoO2 precursors with cell param-
eters in good agreement with the literature.30 The slight loss of sodium
may be explained by (i) the difficult control of the initial Na content in
the P2-Na∼0.7CoO2 precursor or (ii) a weak Na departure (as oxide)
in the free void of the sealed gold tube (indeed some white deposits
were sometimes noticed in the gold tube after synthesis). However, the
calculated chemical formula totally agrees with the already published
compositions.26,30
It was therefore necessary to increase the selectivity of the exchange, and
several kinetic and thermodynamic parameters were adjusted to maximize
the Naþ/Agþ exchange rate while the Liþ/Agþ exchange one was mini-
mized. Following our previous study on the D4-AgCoO2 compound, for
which wenoticedthatthe grain sizereduction was animportant parameter
since without ball-milling the obtaining of the pure D4 synthesis system-
atically failed (small amounts of OD4-(Li/Ag)CoO2 remained, as seen in
XRD patterns), we fixed the grain size to that of the as-synthesized OP4
powder, i.e., as large as possible. The other parameters were tuned as
follows: the temperature of the melt was varied from 300 °C down to
130 °C using a mixture of AgNO3/KNO3 at the eutectic molar ratio 3/235);
for lower temperature trials down to room temperature, the OP4 precursor
was immersed in a AgNO3 aqueous solution; this approach assumes a
two-step reaction:
OP4-ðLi=NaÞCoO2 þ Agþ T OD4-ðLi=AgÞCoO2 þ Naþ
ð1Þ
OD4-ðLi=AgÞCoO2 þ Agþ T D4-AgCoO2 þ Liþ
ð2Þ
We also tuned the silver and lithium concentrations in the molten
salts in order to create supplementary driving forces to hinder the second
step (i.e., the D4 formation): the nominal Naþ/Agþ ratio was tuned to 1,
while lithium nitrate was added up from 2 to 10 times excess (as com-
pared to the lithium amount in the OP4 precursor). Concerning the heat
treatment, the dwell time at high temperature was varied from 15 to 2 h.
2.3. Characterizations. XRD measurements were performed
using a PANalytical X’Pert Pro powder diffractometer in the Braggꢀ
Brentano geometry using either copper or cobalt KR radiation. Data were
collected using an X’Celerator detector in the 10°ꢀ80° window in the 2θ
range. High-temperature experiments were performed in an Anton Paar
HTK 1200N oven chamber under dry oxygen flow (þ1 °C minꢀ1 mean
rate until 700 °C and ꢀ10 °C minꢀ1 to cool until room temperature).
The chemical composition was determined by inductively coupled
plasma absorption electron spectroscopy (ICP-AES) on a Varian 720-ES.
For alkali lamellar oxides, samples were dissolved in HCl solution heated
to ∼70 °C For samples containing silver, HNO3 solution was sometimes
2.2. Ionic Exchanges. Following previous works dealing with
silver delafossites synthesis,21ꢀ23,33,34 ionic exchanges in molten salts
were performed by mixing the OP4-(Li/Na)CoO2 precursor with silver
nitrate and potassium nitrate. In the case of the pure D4-AgCoO2
synthesis, a preliminary ball-milling of the powder precursor as well as a
large silver nitrate excess (5 times compare to alkali nominal content), a
high temperature, and a long dwell time (15h at∼300 °C) were necessary
to complete the exchange.29 In this work, the goal was to only exchange
the Naþ ions with no departure of the Liþ ions from the O3-type blocks.
6650
dx.doi.org/10.1021/ic200534x |Inorg. Chem. 2011, 50, 6649–6655

Inorganic Chemistry
ARTICLE
additionally used as complement when the dissolution was difficult to
complete.
Table 1. Structural Data Used To Simulate the OD4 and PD4
sStacking
The average grain size of the powder samples was determined with a
Malvern Mastersizer 2000 particle size analyzer by immersing the powder
in ethanol bath, with a preliminary ultrasonic step to break up the biggest
aggregates.
OD4-(Li/Ag)CoO2 PD4-(Na/Ag)CoO2
space group
ahex. (Å)
P63/mmc
2.853b
2.845a
Thermogravimetric analysis (TGA) were performed with an STD
Q600 apparatus under Ar flow (initial dwell of 2 h at 100 °C, heating rate
þ5 °C minꢀ1, final dwell of 30 min at 700 °C, cooling rate ꢀ5 °C minꢀ1).
Transport properties were measured on nonsintered, compacted
pellets. Electrical dc resistivity was performed with the four-probe method
in the 4ꢀ300 K range, while thermoelectric power measurements were
collected with homemade equipment previously described.36
chex. (Å)
21.613
1.985a
23.126
1.944b
CoO2 slab thickness (Å)
AO2 interslab thickness (Å)
2.633 for A = Li
3.471 for A = Na
4.204 for A = Ag
a Obtained from P2-Na∼0.7CoO2 15 and D3-AgCoO2 37 average. b Obtained
from O3-LiCoO2 9 and D2-AgCoO2 37 average.
Table 2. Atomic Position of the OD4-(Li/Ag)CoO2 from the
3. ORDERED STACKING SIMULATION
Stacking Simulation in the P63/mmc Space Groupa
Two different layered structures can be sketched assuming
different selective exchanges for the lithium or sodium cations:
either an OD4-(Li/Ag)CoO2 phase that alternates O3-type lithium
layers and D2-type silver layers (Figure 1b, resulting from the selec-
tive exchange of only sodium ions) or a PD4-(Na/Ag)CoO2
phase that combines P2-type sodium layers and D3-type silver
layers (Figure 1c, resulting from the selective exchange of only
lithium ions). Actually, there are several arguments in favor of a
preferential exchange of Naþ ions compared to Liþ ones: (i) the
ions diffusion through face-sharing trigonal prisms is easier than
through octahedra, which requires the ions to pass through an
intermediate tetrahedral site; (ii) the interslab space AO2 (A = Li,
Na) increase that is necessary to accommodate the Agþ ions is
smaller for the Naþ/Agþ exchange (∼þ16%) than for the Liþ/Agþ
one (∼þ56%) (these values are calculated by comparing the inter-
slab thicknesses in OP4-Li∼0.42Na∼0.37CoO232 and in AgCoO221,37);
and (iii) a slab gliding is required during the Liþ/Agþ exchange,
while no oxygen packing change is needed for the Naþ/Agþ
exchange. As the Naþ ions should be removed more easily than
the Liþ ions, the theoretical PD4 phase should not appear during
the OP4 f D4 process.
In our previous works, we simulated the O4-LiCoO2 and the
D4-AgCoO2 polytypes using as input parameters some previous
experimental crystallographic data for O2-LiCoO2,10,38,11 O3-
LiCoO2,9 and AgCoO2,18,21ꢀ23,37 like the cationꢀoxygen dis-
tances and the thicknesses of CoO2 slabs and AO2 interslab
spaces, to build up the new blocks of the O4 and D4 polytypes,
respectively.12,29 This approach was also extended to the descrip-
tion of the ordered OPP9-(Li/Na/Na)CoO2 packing that alter-
nates within the lamellar CoO2 stacking two P2 sodium layers for
one O3 lithium layer.32 We applied this method in the present study
to sketch the OD4-(Li/Ag)CoO2 and the PD4-(Na/Ag)CoO2
stacking.
atom
site
x
y
z
occ
1
1
Ag
Li
2b
2a
4f
0
0
/
/
1
4
2
0
0
0.84
2
1
Co
O
/
/
0.393
0.347
0.439
1
1
1
3
3
3
4e
4f
0
0
1
2
O
/
/
3
a The lithium occupancy factor is set to 0.84 assuming a theoretical
composition OD4-Li0.42Ag0.5CoO2.
4. RESULTS AND DISCUSSIONS
4.1. Ionic Exchanges. Figure 2a presents the XRD pattern
obtained after a first ionic exchange with the OP4-(Li/Na)CoO2
precursor (experimental conditions: (Liþ þ Naþ)/Agþ ratio close
to 1, thermal treatment of 15 h at ∼300 °C, final washing step).
The profile matching refinement evidences two phases that both
crystallize in the P63/mmc space group: the D4-AgCoO2 dela-
fossite according to our previous work29 and a new phase for
which the diffraction pattern nicely compares with the simulated
pattern of an OD4-(Li/Ag)CoO2 (Figure 2d). Moreover, there is
not any trace of the OP4-(Li/Na)CoO2 precursor. A possible
PD4-(Na/Ag)CoO2 phase formation is totally discarded here.
The ICP-AES analysis realized after washing of the product re-
veals that the sodium content is insignificant, clearly confirming
that the Naþ/Agþ exchange is completely achieved.
In order to roughly quantify the phase proportions, Rietveld
refinement was performed using the simulated structural data of
the D429 and the OD4 phases by only varying the cell parameters
as well as the profile parameters (the structural parameters were
fixed). The D4 delafossite appeared to be the majority phase
(∼5/6). In addition, the refinement gives OD4 lattice parameters
[ahex. = 2.848(3) Å and chex. = 21.607(7) Å], in rather good agree-
ment with our simulation work (Table 1).
The presence of this intermediate phase is related to the dif-
ferent mechanism implied for the lithium and sodium exchanges.
For these reasons, the lithium exchange requires high activation
energy and lower grain size as well as higher temperature, dwell
time and silver concentration are at the same time necessary to
make the exchange complete.
Decreasing the temperature to ∼130 °C while the dwell time
is shortened to only 2 h and the (Liþ þ Naþ)/Ag ratio is lowered
to ∼2 (i.e., Na/Ag ∼1) significantly improves the results, as the
OD4-(Li/Ag)CoO2 phase becomes the majority: the intensity of
the (004) diffraction peak of the D4-AgCoO2 delafossite clearly
The required input parameters needed are the thicknesses
of LiO2, NaO2, and AgO2 interslabs.9,15,37 These two stackings
are then built up by superimposing O3 and D2 blocks or P2 and
D3 blocks for respectively the OD4 and the PD4 compound. The
thicknesses ofthe CoO2 slabs are averagedbetweenthe two parent
phases, as well as ahex. parameters (Table 1); the z positions of the
cations are adjusted at the center of the slabs and interslabs.
A crystallographic symmetry operator determination led to
the P63/mmc space group in both cases, with lattice parameters
ahex. = 2.845 Å and chex. = 21.613 Å for OD4 and ahex. = 2.853 Å
and chex. = 23.126 Å for PD4. For the OD4-(Li/Ag)CoO2 phase,
the atomic positions are given in Table 2 and the corresponding
simulated XRD pattern is plotted in Figure 2d.
6651
dx.doi.org/10.1021/ic200534x |Inorg. Chem. 2011, 50, 6649–6655

Inorganic Chemistry
ARTICLE
the magnitude of the other parameters; this confirms that the
slab gliding (necessary in the Naþ/Liþ exchange) is more energy
consuming than the increase of the interslab thickness and that it
is the limiting factor for the exchange.
ICP-AES analyses were not appropriate to determine the pre-
cise composition of the OD4 phase, mainly because powders were
difficult to dissolve. However, one can notice that (i) the OP4
precursor is characterized by lithium vacancies30,26ꢀ28 and (ii)
previous reports on silver delafossites obtained by ionic exchanges
from P2-Na∼0.7CoO2 conclude the stoichiometric composition
Ag1CoO2.18,24 According to these results and assuming that (i) the
lithium layers remain unchanged during the ionic exchange pro-
cess, (ii) silver ions cannot enter into lithium interslab space as
the oxygen stacking is not suitable for accommodating the dumb-
bell coordination, and (iii) silver ions completely fill the sodium
interslab space, the chemical composition of the OD4 phase
should be close to Li∼0.42Ag0.5CoO2.
In conclusion of this experimental section, the OD4-(Li/
Ag)CoO2 formation is clearly evidenced to be an intermediate
step in the global OP4 f D4 ionic exchange. All our attempts to
isolate this phase pure were unsuccessful and it appeared impos-
sible from this method to replace all the sodium ions before
starting the exchange of lithium ions.
4.2. Thermal Stability. Ionic exchange reactions often lead to
metastable phases, such as the O4- and the O2-LiCoO2 poly-
types, which turn into the thermodynamically stable O3 polytype
when they are heated.11,12,38 Silver delafossites are also sensitive
to temperature as they decompose into metallic silver and cobalt
oxide Co3O4 below 600 °C.29 According to these remarks, we first
performed an in situ XRD experiment on an initial mixture of the
OD4 and the D4 phases obtained with the best above-mentioned
conditions (corresponding XRD in Figure 2c). Figure 3 shows the
XRD pattern evolution vs temperature in the 6.5°ꢀ60° 2θ range.
Up to 400 °C, the temperature increase causes a narrowing of
diffraction peaks [highlighted in the inset of Figure 3 focusing on
the (006) peak of the OD4 phase], which can be explained by an
enhancement of the crystallinity of each phase.
In the same temperature range, however, characteristic diffrac-
tion peaks of O3-LiCoO2, Co3O4, and metallic silver already
begin to grow. At ∼550 °C, all the (00l) diffraction peaks of the
OD4 phase have vanished. This experiment clearly demonstrates
the instability of the OD4 phase, which decomposes according to
the eq 3 (with the assumption of the initial chemical composition
Li0.42Ag0.5CoO2).
Figure 2. (aꢀc) Experimental XRD data obtained after different ionic
exchanges from the OP4-(Li/Na)CoO2 raw material. For comparison,
the calculated XRD pattern obtained from the simulation of (d) OD4-(Li/
Ag)CoO2 (from Table 2) and (e) D4-AgCoO2 (from ref 29) are shown.
decreases while the (006) and (008) diffraction peak intensities
of the OD4-(Li/Ag)CoO2 phase increase (Figure 2b). Quanti-
tatively, a Rietveld refinement gives approximately half of the
OD4 phase in the final product.
The addition of lithium nitrate in the salt (in 2 times molar
excess as compared to the initial lithium amount in the precursor
powder) significantly hinders the lithium exchange and finally
stabilizes the OD4 phase up to about 80% (Figure 2c); how-
ever, tuning further this parameter no longer improves the results,
and we could not get a complete elimination of the D4-AgCoO2
phase. In order to further decrease the reaction temperature sev-
eral exchange reactions in water solution were tested from boiling
point to room temperature, using equivalent [Agþ] and [Liþ]
concentrations as above. The OD4 formation is evidenced in all
cases, but OP4-(Li/Na)CoO2 still remains present in majority
while the D4-AgCoO2 delafossite diffraction peaks are clearly
appearing in all XRD patterns (not shown), even for stirring times
as short as 2 h, at room temperature. To conclude, we managed
to increase the proportion of the obtained OD4 phase by tuning
some reaction parameters; however, we did not succeed to
completely avoid the Liþ/Agþ ionic exchange.
The broadening of the diffraction lines of both OD4 and D4
phases might beattributed tostacking defects thatoccur duringthe
ionic exchange process. For example, one can reasonably assume
the presence of sporadic D4 blocks in the overall OD4 structure.
Note also that for all the ionic exchanges that use a high [Liþ]
concentration in the solution, a competition between silver and
lithium to replace sodium can be expected. In the latter case, the
O4-LiCoO2 polytype could be formed.12,31 The XRD patterns,
however, never showed any presence of this compound, whatever
Δ
Li0:42Ag0:5CoO2 f 0:42LiCoO2 þ 0:5Agmetal
ð1 ꢀ 0:42Þ
þ
ðCo3O4 þ O2Þ
ð3Þ
3
The D4 delafossite also decomposes just after the OD4 by
giving additional Co3O4 and metallic silver, in agreement with
our previous results.29
4.3. Morphology Characterization. The grain morphology
of a sample containing the OD4 and the D4 phases (correspond-
ing XRD in Figure 2c) are shown in Figure 4. The micrograph
analysis cannot distinguish the different phases, as the grain size
and morphology appear very homogeneous. The large grain size
(>10 μm) results from the absence of a preliminary milling step
before the ionic exchange. The lamellar shape is clearly visible,
especially with micrographs focused on the edges of the particles.
Despite some sporadic particles that keep the hexagonal shape
(highlighted by white dot lines), most of the particles are
6652
dx.doi.org/10.1021/ic200534x |Inorg. Chem. 2011, 50, 6649–6655

Inorganic Chemistry
ARTICLE
Figure 3. Temperature influence on the XRD powder pattern of a initial
mixture of OD4-(Li/Ag)CoO2 and D4-AgCoO2 phases in a 4:1 ratio
(corresponding to sample c in Figure 2). The O3-LiCoO2, Co3O4, and
metallic silver diffraction peaks begin to grow in the 350ꢀ450 °C tem-
perature range, reflecting the decomposition of the OD4-(Li/Ag)CoO2
phase. The delafossite D4-AgCoO2 decomposition occurs from 500 °C.
The inset focuses on the evolution of the (006) diffraction peak of the
OD4 phase and highlights the crystallinity enhancement at the begin-
ning of the heating sequence. The symbol (¤) marks the peak of the
polymer window of the airtight in situ cell.
characterized by a more rounded contour. A similar morphology
characterizes the OP4 precursor (Figure 4), but it is more stressed
after the ionic exchange that alters the shape of the grains.
4.4. Electrical Properties. Depending on the A element,
lamellar oxides ACoO2 exhibit different kinds of transport
properties: an insulator behavior is obtained in ACoO2 delafos-
sites with A = Cu, Ag, while a metallic one is obtained for A = Pd,
Pt; O3-LiCoO2 is insulating, but Lix<1CoO2 materials become
more and more conductive and even metallic as Liþ is further
deintercalated;39 the OP4-(Li/Na)CoO2 precursor exhibits in-
teresting thermoelectric properties with a large Seebeck coeffi-
cient coexisting with a relatively good electrical conductivity.26,27
The study of the electrical properties of OD4-(Li/Ag)CoO2 is
complicated because of (i) its poor thermal stability that prevent
any sintering, (ii) the presence of the secondary D4 phase that
could not be fully eliminated, and (iii) the anisotropy of the struc-
ture that could lead to a strong anisotropy of the electrical prop-
erties as in the case of delafossite oxides.20 The physical measure-
ments were finally performed on nonsintered compacted pellets
(bulk density close to 70%) containing a mixture of roughly 4:1
OD4:D4 (i.e., with the sample characterized by the XRD pattern
of Figure 2c). Figure 5 displays the temperature dependence of
the electrical resistivity and of the thermopower in the tempera-
ture range 5ꢀ300 K and compares them with the behavior of
pure D4-AgCoO2.29 Regarding the above three limitations, the
transport measurements are not expected to give access to
intrinsic properties but only to reflect some trends in the global
behavior of the new phase. The OD4/D4 sample displays a global
insulator behavior with a resistivity decreasing with temperature.
Figure 4. SEM micrographs of powder sample of the 4:1 mixture of the
OD4 and the D4 phases at three different magnifications and of the
OP4-(Li/Na)CoO2 for comparison. The white dotted lines highlight
the particle hexagonal shape.
The Seebeck coefficient is positive and its temperature depen-
dence follows a monotonous and roughly linear increase up to
room temperature. Such a behavior shows that charge carriers are
holes diffusing through a hopping mechanism. In comparison
with pure D4-AgCoO2, the OD4/D4 mixture sample exhibits a
lower resistivity and a lower thermopower, which is in agreement
with a larger charge-carrier concentration, as expected from the
nominal compositions. Determining whether holes are moving
within silver layers, like in most delafossite oxides,20 or in cobalt
6653
dx.doi.org/10.1021/ic200534x |Inorg. Chem. 2011, 50, 6649–6655

Inorganic Chemistry
ARTICLE
exchange of Naþ ions (as compared to Liþ ions) for Agþ explains
the existence of this OD4-(Li/Ag)CoO2 phase, which is the
second example of a mixed ordered (A/A0)CoO2 packing after
the OP4-(Li/Na)CoO2 and the first case of a intergrowth of rock
salt and delafossite layers. The space group is P63/mmc and the
cell parameters are ahex. = 2.848(3) Å and chex. = 21.607(7) Å, in
perfect agreement with simulation studies. Tuning different
thermodynamic and kinetic parameters enables one to impede
the D4 delafossite formation, but within our experimental setup
we could not avoid any Liþ/Agþ ionic exchange; the best result is
a mixture of ∼80% of the OD4 phase and ∼20% of the D4 phase.
As other ionic exchange products, the OD4-(Li/Ag)CoO2 phase
is not stable and tends to decompose above 400ꢀ450 °C, giving
O3-LiCoO2, Co3O4, and metallic silver. Electrical measurements
performed on a nonsintered, compacted pellet of a 4:1 mixture of
OD4 and D4 reveal a semiconductor behavior.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: pollet@icmcb-bordeaux.cnrs.fr.
’ ACKNOWLEDGMENT
The authors want to thank J. Villot, S. Fourcade, C. Denage, L.
Etienne, E. Lebraud, and S. Pechev for technical assistance and
ANR OCTE for financial support. CEA is also thanked for a
scholarship to R.B.
’ REFERENCES
(1) Mizushima, K.; Jones, P.; Wiseman, P.; Goodenough, J. Mater.
Res. Bull. 1980, 15 (6), 783.
(2) Berthelot, R.; Carlier, D.; Delmas, C. Nat. Mater. 2011, 10 (1),
74.
Figure 5. Evolution vs temperature of the resistivity (triangles) and of
the Seebeck coefficient (circles) measured on nonsintered, compacted
pellets of the 4:1 mixture of OD4 and D4 phases (white marks) and of a
pure D4-AgCoO2 (black marks, from ref 29). (Inset) Evidence, with the
dashed line, of the variable range hopping conduction behavior in the
∼50ꢀ300 K temperature range.
(3) Terasaki, I.; Sasago, Y.; Uchinokura, K. Phys. Rev. B Condens.
Matter Mater. Phys. 1997, 56 (20), R12685.
(4) Lee, M.; Viciu, L.; Li, L.; Wang, Y.; Foo, M.; Watauchi, S.; Pascal,
R., Jr.; Cava, R.; Ong, N. Nat. Mater. 2006, 5 (7), 537.
(5) Lee, M.; Viciu, L.; Li, L.; Wang, Y.; Foo, M.; Watauchi, S.; Pascal,
R., Jr.; Cava, R.; Ong, N. Phys. B Condens. Matter 2008, 403 (5ꢀ9), 1564.
(6) Takada, K.; Sakurai, H.; Takayama-Muromachi, E.; Izumi, F.;
Dilanian, R.; Sasaki, T. Nature 2003, 422 (6927), 53.
(7) Schaak, R.; Klimczuk, T.; Foo, M.; Cava, R. Nature 2003, 424
(6948), 527.
layers is out of the scope of the present paper in view of the ex-
perimental limitations above-mentioned. In the second hypoth-
esis, the thermopower value at room temperature would be far
below the asymptotic value expected from the Heikes formula
(∼260 μV Kꢀ1) for the given sample composition and assuming
the usual electronic configuration for cobalt(III) and -(IV)
located in the relevant site symmetry.40
As shown in the inset of Figure 5 and as already observed for
pure D4-AgCoO2 delafossite,29 the resistivity data obey a Mott’s
law of the form F = A exp[(T0/T)1/4].41,42 However, below ca.
50 K, Mott’s law is no longer obeyed, and, below 15 K, the elec-
trical resistivity no longer decreases as the temperature is further
decreased. Such a behavior is rarely observed in experimental
studies of semiconducting oxides, but it was predicted as a pos-
sible transport mechanism at low temperature (T < ΘD/2, where
ΘD is the Debye temperature) when the zero point energy 1/2hν
associated with quantum fluctuations can take the place of thermal
vibrations.43
(8) Huang, Q.; Foo, M.; Lynn, J.; Zandbergen, H.; Lawes, G.; Wang,
Y.; Toby, B.; Ramirez, A.; Ong, N.; Cava, R. J. Phys.: Condens. Matter
2004, 16 (32), 5803.
(9) Levasseur, S.; Mꢀenꢀetrier, M.; Suard, E.; Delmas, C. Solid State
Ionics 2000, 128 (1ꢀ4), 11.
(10) Delmas, C.; Braconnier, J.-J.; Hagenmuller, P. Mater. Res. Bull.
1982, 17 (1), 117.
(11) Carlier, D.; Saadoune, I.; Croguennec, L.; Mꢀenꢀetrier, M.; Suard,
E.; Delmas, C. Solid State Ionics 2001, 144 (3ꢀ4), 263.
(12) Berthelot, R.; Carlier, D.; Pollet, M.; Doumerc, J.-P.; Delmas, C.
Electrochem. Solid-State Lett. 2009, 12 (11), A207.
(13) Fouassier, C.; Matejka, G.; Reau, J.-M.; Hagenmuller, P. J. Solid
State Chem. 1973, 6 (4), 532.
(14) Viciu, L.; Bos, J.; Zandbergen, H.; Huang, Q.; Foo, M.; Ishiwata,
S.; Ramirez, A.; Lee, M.; Ong, N.; Cava, R. Phys. Rev. B Condens. Matter
Mater. Phys. 2006, 73 (17), 174104.
(15) Huang, Q.; Foo, M.; Pascal, R., Jr.; Lynn, J.; Toby, B.; He, T.;
Zandbergen, H.; Cava, R. Phys. Rev. B Condens. Matter Mater. Phys. 2004,
70 (18), 184110.
(16) Blangero, M.; Decourt, R.; Carlier, D.; Ceder, G.; Pollet, M.;
Doumerc, J.-P.; Darriet, J.; Delmas, C. Inorg. Chem. 2005, 44 (25), 9299.
5. CONCLUSION
A new lamellar phase alternating rock salt and delafossite
type layers is evidenced as an intermediate step during the global
OP4-(Li/Na)CoO2 f D4-AgCoO2 ionic exchange. The easier
6654
dx.doi.org/10.1021/ic200534x |Inorg. Chem. 2011, 50, 6649–6655

Inorganic Chemistry
ARTICLE
(17) Carlier, D.; Croguennec, L.; Ceder, G.; Mꢀenꢀetrier, M.; Shao-Horn,
Y.; Delmas, C. Inorg. Chem. 2004, 43 (3), 914.
(18) Shannon, R. D.; Rogers, D. B.; Prewitt, C. T. Inorg. Chem. 1971,
10 (4), 713.
(19) Prewitt, C.; Shannon, R.; Rogers, D. Inorg. Chem. 1971, 10 (4),
719.
(20) Rogers, D.; Shannon, R.; Prewitt, C.; Gillson, J. Inorg. Chem.
1971, 10 (4), 723.
(21) Shin, Y.; Doumerc, J.-P.; Dordor, P.; Pouchard, M.; Hagenmuller,
P. J. Solid State Chem. 1993, 107 (1), 194.
(22) Shin, Y.; Doumerc, J.-P.; Pouchard, M.; Hagenmuller, P. Mater.
Res. Bull. 1993, 28 (2), 159.
(23) Shin, Y.-J.; Kwak, J.-H.; Yoon, S. Bull. Korean Chem. Soc. 1997,
18 (7), 775.
(24) Muguerra, H.; Colin, C.; Anne, M.; Julien, M.-H.; Strobel, P.
J. Solid State Chem. 2008, 181 (11), 2883.
(25) Delmas, C.;Fouassier, C.;Hagenmuller, P.Phys. BþC(Amsterdam,
Neth.) 1980, 99 (1ꢀ4), 81.
(26) Ren, Z.; Shen, J.; Jiang, S.; Chen, X.; Feng, C.; Xu, Z.; Cao, G.
J. Phys.: Condens. Matter 2006, 18 (29), L379.
(27) Chen, X.; Xu, X.-F.; Hu, R.-X.; Ren, Z.; Xu, Z.-A.; Cao, G.-H.
Acta Phys. Sin. 2007, 56, 1627.
(28) Bos, J.; Hertz, J.; Morosan, E.; Cava, R. J. Solid State Chem.
2007, 180 (11), 3211.
(29) Berthelot, R.; Pollet, M.; Doumerc, J.-P.; Delmas, C. Inorg.
Chem. 2011, 50 (10), 4529.
(30) Balsys, R.; Lindsay Davis, R. Solid State Ionics 1994, 69 (1), 69.
(31) Komaba, S.; Yabuuchi, N.; Kawamoto, Y. Chem. Lett. 2009, 38
(10), 954.
(32) Berthelot, R.; Pollet, M.; Carlier, D.; Delmas, C. Inorg. Chem.
2011, 50 (6), 2420.
(33) Doumerc, J.-P.; Ammar, A.; Wichainchai, A.; Pouchard, M.;
Hagenmuller, P. J. Phys. Chem. Solids 1987, 48 (1), 37.
(34) Ammar, A. Contribution ꢁa l’ꢀetude des oxydes de type delafossite:
Corrꢀelations entre propriꢀetꢀes cristallochimiques et ꢀelectroniques. Ph.D.
thesis, University of Marrakech, 1988
(35) Ussow, A. Z. Anorg. Allg. Chem. 1904, 38 (1), 419.
(36) Dordor, P.; Marquestaut, E.; Villeneuve, G. Rev. Phys. Appl.
1980, 15, 1607.
(37) Seshadri, R.; Felser, C.; Thieme, K.; Tremel, W. Chem. Mater.
1998, 10 (8), 2189.
(38) Paulsen, J.; Mueller-Neuhaus, J.; Dahn, J. J. Electrochem. Soc.
2000, 147 (2), 508.
(39) Mꢀenꢀetrier, M.; Saadoune, I.; Levasseur, S.; Delmas, C. J. Mater.
Chem. 1999, 9 (5), 1135.
(40) Pollet, M.; Doumerc, J.-P.; Guilmeau, E.; Grebille, D.; Fagnard,
J.-F.; Cloots, R. J. Appl. Phys. 2007, 101 (8), 083708.
(41) Mott, N. F.; Davis, E. A. Electronic Process in Non-Crystalline
Materials; Clarendon Press: Oxford, 1979.
(42) Cox, P. Transition Metal Oxides: An Introduction to Their
Electronic Structure and Properties; Clarendon Press: Oxford, 1995.
(43) Mott, N. F. MetalꢀInsulator Transitions; Taylor & Francis:
London, 1990.
6655
dx.doi.org/10.1021/ic200534x |Inorg. Chem. 2011, 50, 6649–6655