HAL Id: hal-03317546
https://hal.archives-ouvertes.fr/hal-03317546
Submitted on 6 Aug 2021
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Evidence for anionic redox activity in a
tridimensional-ordered Lirich positive electrode -Li2IrO3
Paul E Pearce, Arnaud J Perez, Gwenaelle Rousse, Matthieu Saubanère,
Dmitry Batuk, Dominique Foix, Eric Mccalla, Artem M Abakumov, Gustaaf
van Tendeloo, Marie-Liesse Doublet, et al.
To cite this version:
Paul E Pearce, Arnaud J Perez, Gwenaelle Rousse, Matthieu Saubanère, Dmitry Batuk, et al.. Evi-
dence for anionic redox activity in a tridimensional-ordered Lirich positive electrode -Li2IrO3. Nature
Materials, Nature Publishing Group, 2017, 16 (5), pp.580-586. �10.1038/nmat4864�. �hal-03317546�
1
Evidence for anionic redox activity in a tridimensional-ordered Li-
rich positive electrode -Li
2
IrO
3
Paul E. Pearce
1,2,3
, Arnaud J. Perez
1,2,3
, Gwenaelle Rousse
1,2,3
, Mathieu Saubanère,
2,4
Dmitry Batuk
1,5
, Dominique Foix
2,6
, Eric McCalla
1,2,7
, Artem M. Abakumov
8,5
, Gustaaf Van
Tendeloo
5
Marie-Liesse Doublet
2,4
and Jean-Marie Tarascon
1,2,3,*
1
Collège de France, Chimie du Solide et de l’Energie, UMR 8260, 11 place Marcelin Berthelot, 75231
Paris Cedex 05, France.
2
Réseau sur le Stockage Electrochimique de l’Energie (RS2E), FR CNRS 3459, France.
3
Sorbonne Universités - UPMC Univ Paris 06, 4 Place Jussieu, F-75005 Paris, France.
4
Institut Charles Gerhardt, UMR5253, CNRS and Université de Montpellier, Place Eugène Bataillon,
F34095 Montpellier
5
EMAT, University of Antwerp, Groenenborgerlaan 171, B-2020, Antwerp, Belgium.
6
IPREM/ECP (UMR 5254), University of Pau, 2 av. Pierre Angot, 64053 Pau Cedex 9, France.
7
CEMS, University of Minnesota, 421 Washington Ave., Minneapolis, MN, USA, 55455.
8
Skolkovo Institute of Science and Technology, Nobel str. 3, 143026 Moscow, Russia.
* Corresponding author: jean-marie.tarascon@college-de-france.fr
Abstract
Lithium-ion battery cathode materials have relied on cationic redox reactions until the recent
discovery of anionic redox activity in Li-rich layered compounds which enables capacities as high as
300 mAh g
-1
. In the quest for new high-capacity electrodes with anionic redox, a still unanswered
question was remaining regarding the importance of the structural dimensionality. The present
manuscript provides an answer. We herein report on a β-Li
2
IrO
3
phase which, in spite of having the Ir
arranged in a tridimensional (3D) framework instead of the typical two-dimensional (2D) layers seen
in other Li-rich oxides, can reversibly exchange 2.5 e
-
per Ir, the highest value ever reported for any
insertion reaction involving d-metals. We show that such a large activity results from joint reversible
cationic (M
n+
) and anionic (O
2
)
n-
redox processes, the latter being visualized via complementary
transmission electron microscopy and neutron diffraction experiments, and confirmed by density
functional theory calculations. Moreover, β-Li
2
IrO
3
presents a good cycling behaviour while showing
neither cationic migration nor shearing of atomic layers as seen in 2D-layered Li-rich materials.
Remarkably, the anionic redox process occurs jointly with the oxidation of Ir
4+
at potentials as low as
3.4 V versus Li
+
/Li
0
, as equivalently observed in the layered α-Li
2
IrO
3
polymorph. Theoretical
calculations elucidate the electrochemical similarities and differences of the 3D versus 2D polymorphs
in terms of structural, electronic and mechanical descriptors. Our findings free the structural
dimensionality constraint and broaden the possibilities in designing high-energy-density electrodes for
the next generation of Li-ion batteries.
2
Renewable energy sources and electric automotive transportation are popular topics in today’s
energy conscious society, hence placing electrochemical storage as one of the major technological
challenges in this new century. Rechargeable Li-ion batteries are the technology of choice for
powering electric vehicles and stand as a possible solution for grid applications provided that new
advances can be made for improving the performance/cost ratio and lifetime
1,2
. Today’s research on
positive electrodes is divided into two streams. One deals with polyanionic compounds that favour
safety and cost at the expense of energy density
3–5
; and the other with chemical substitutions within the
layered oxides which prioritize energy density
6
. The layered oxides are based on LiCoO
2
, where Co
and Li are each in 2D hexagonal arrays and Li intercalates/deintercalates via 2D migration along these
sheets. Research in layered oxides has led, with the partial substitution of Co by Ni and Mn in the
transition metal layer, to the presently commercialized layered oxide LiCo
1/3
Ni
1/3
Mn
1/3
O
2
phase
(referred to as Li-NMC) that delivers capacities ranging from 160 to 190 mAh g
−1
(ref. 6). Greater
capacities (>280 mAh g
−1
) were obtained by further substituting the transition metals by Li so as to
produce the Li[Li
x
Ni
y
Mn
z
Co
1−x−y−z
]O
2
layered phases denoted as Li-rich NMC
7,8
.
The staggering increase in capacity was demonstrated, via complementary electron
paramagnetic resonance (EPR) and X-ray photoemission spectroscopy (XPS) measurements
performed on ruthenates Li
2
Ru
1−y
Sn
y
O
3
(refs 9,10), to be rooted in the anionic activity with the
reversible formation of (O
2
)
n−
peroxo-like species accompanied by a massive cationic migration to the
depleted Li layers. An increase in the covalence of the M–O bond, via the replacement of a 4d metal
(Ru
4+
) by a 5d metal (Ir
4+
) so as to form α-Li
2
IrO
3
, was shown to prevent this migration
11
. However, α-
Li
2
IrO
3
, which enlists a structural transition on charge (from cubic close-packed O3 to hexagonal
close-packed O1), presents lower cycling performances than Li
2
Ru
0.75
Sn
0.25
O
3
. Whether such a
difference is nested in the existence of a structural transition or the simultaneous, rather than
successive cationic and anionic redox processes for α-Li
2
IrO
3
as opposed to Li
2
RuO
3
, remains
unanswered. To address these questions, we decided to probe further the effect of modifying the
crystal structure on the anionic redox reactivity.
Searching for Li-rich ordered rocksalt structures with enhanced 3D character, we became
aware of the β-Li
2
IrO
3
polymorph
12,13
. Its structure exhibits an unusual topology that has recently
attracted a lot of interest for magnetism
12–16
. This hyperhoneycomb structure, which consists in a 3D
long-range ordered framework of the Ir–O bonds, differs drastically from that of layered honeycomb
α-Li
2
IrO
3
(Supplementary Fig. 1). It is now imperative to determine whether or not the 3D edge-
sharing connectivity of the IrO
6
octahedra in β-Li
2
IrO
3
can accommodate anionic distortions and
thereby show the enhanced capacity associated up to now with 2D-layered Li-rich oxides
Pure β-Li
2
IrO
3
powders were prepared by a high-temperature ceramic process from a mixture
of IrO
2
and Li
2
CO
3
, as described in the Methods. The synchrotron X-ray and neutron powder
3
diffraction patterns were fitted using the Rietveld method, confirming the previously reported Fddd
hyperhoneycomb structure
13
(Fig. 1a,b and Supplementary Table 1 and Supplementary Fig. 2). β-
Li
2
IrO
3
is built on edge-sharing IrO
6
octahedra that form a 3D matrix in which Li ions occupy all
available octahedral sites. The structure can therefore accommodate Li migration via corrugated
interconnected paths, as shown in Supplementary Fig. 3.
Figure 2a shows the electrochemical performances of β-Li
2
IrO
3
cycled versus Li in Swagelok
cells between 2 and 4.8 V at a C/10 rate. Remarkably, all Li can be removed, leading to an ‘IrO
3
’
phase which can then uptake nearly 1.8 Li in the following discharge. More specifically, the first
charge shows a staircase profile, with the presence of four successive plateaux at 3.45, 3.50, 4.40 and
4.55 V versus Li
+
/Li
0
, that can be visualized by four peaks in the dx/dV plots (Fig. 2b). Such peaks
remain well defined on the subsequent discharge. This contrasts with the typical behaviour of Li-rich
oxides: a staircase voltage profile on the first charge that converts to an S-shaped profile on the
following discharge, which is associated with inter/intra layer cationic migrations
9,17
. It also differs
from the electrochemical behaviour of the α-Li
2
IrO
3
polymorph, which shows only two oxidation
plateaux that progressively transform upon cycling into S-shaped profiles
11
. Fig. 2b also shows a
minor downshift in voltage of the anodic and cathodic peaks with continued cycling up to 4.8 V,
which can be avoided by limiting the cutoff voltage to 4.5 V (see Supplementary Fig. 4).
To better understand Li-driven structural changes in β-Li
2
IrO
3
, in situ X-ray powder
diffraction (XRD) was performed (Fig. 2c). We observe four successive biphasic regions, each of them
corresponding to a well-defined plateau in the voltage-composition curve (Fig. 2c). Each inflexion
point, at x = 1.5, x = 1, x = 0.35 and x = 0 in β-Li
x
IrO
3
, corresponds to a single phase. The TEM
analysis of these single phases reveals that the arrangement of the Ir atoms is preserved even when all
Li is extracted (Supplementary Fig. 5). Refined XRD patterns (shown in Supplementary Fig. 6)
demonstrate that when charging from x = 2 to x = 1, the Fddd space group is preserved with only a
change in the lattice parameters (Fig. 2c,d). When x reaches 0.35, the symmetry of the material is
lowered to the monoclinic, C2/c space group (note the corresponding reflection splitting in the
18.5◦−19.5◦ 2θ region in Fig. 2d and in Supplementary Fig. 6c and Supplementary Table 2). The fully
oxidized phase Li
x
IrO
3
(4.8 V, x ∼0) shows a significant broadening of the peaks, most likely
associated with increasing strain in the structure, and can also be indexed in the C2/c space group with
a smaller β angle as compared to Li
0.35
IrO
3
(Supplementary Fig. 6d). The feasibility of removing all
Li
+
ions from β-Li
2
IrO
3
, while preserving the structural framework, is somewhat spectacular as this is
normally seen only in polyanionic and spinel compounds; with the main difference that two atoms out
of six (that is, 33%) can be extracted from β-Li
2
IrO
3
, whereas it is only one out of seven (that is, 15%)
for LiFePO
4
or LiMn
2
O
4
. Upon discharge, the XRD pattern of the fully charged sample converts back
to that of the pristine phase via the same sequence of structure transformations. This demonstrates the
4
reversibility of the charge/discharge process but does not provide clues on the nature of the involved
redox species.
To grasp further insights into the mechanisms of the Li-redox processes in β-Li
x
IrO
3
, XPS
measurements were carried out on the pristine sample (x =2), on ex situ oxidized samples at x =1.5, 1,
0.35, 0, and on the compound that was charged to 4.8 V and discharged to 2 V versus Li
+
/Li
0
. Figure
3a,b shows a shift of the Ir 4f peak towards lower energies together with the appearance of an extra O
1s peak (red peak in Fig. 3a) associated with the formation of peroxo-like (O
2
)
n−
species during the
charging process. During the subsequent discharge, the opposite variations are seen, indicating a
reversible oxidation/reduction of both metal and oxygen upon cycling, as reported for the previously
investigated Ru- and Ir-based layered Li-rich compounds9–11. Moreover, Fig. 3d shows that the
portion of redox that can be attributed to peroxo-like species increases monotonously until the end of
charge at 4.8 V (x = 0), where it reaches ∼55%. Similarly, the evolution of the Ir 5d population based
on the Ir valence bands (Fig. 3c) indicates that the metal oxidation takes place preferentially during the
first part of charge. Overall, these results indicate that Li-deinsertion-driven cationic and anionic
oxidation processes in β-Li
2
IrO
3
occur conjointly, as seen in α-Li
2
IrO
3
(ref. 11).
At this stage, to explore the synergy between redox processes and structural transformations,
we exploit complementary neutron powder diffraction, high-resolution scanning transmission electron
microscopy (STEM) and density functional theory (DFT) calculations. An ex situ neutron powder
diffraction (NPD) pattern was recorded on the sample charged to 4 V (Li
x
IrO
3
with x = 1). The NPD
pattern can be indexed with the Fddd orthorhombic space group (Fig. 1d), with lattice parameters
reported in Table 1. Moreover, Rietveld refinements indicate that the remaining Li occupies a new
tetrahedral position (Fig. 1d,e), denoted Li3 in Table 1. The structure refinement also reveals
significant changes in the geometry of the IrO
6
octahedra (Fig. 1c and f). First, the average Ir–O
distance reduces from 2.028 Å in the pristine structure to 1.972 Å in the charged sample,
concomitantly with the decrease of the ionic radius of the Ir cations from 0.625 Å (Ir
4+
) to 0.57 Å (Ir
5+
)
18. Moreover, the shortest O–O bond in the IrO
6
octahedra is significantly shorter in the charged
material (2.53 Å, Fig. 1f) than in the pristine material (2.75 Å, Fig. 1c), which is indicative of the
formation of peroxo-like species on charge.
To track the deformation of the IrO
6
octahedra, we used [110] annular bright-field STEM
(ABF-STEM) images of the pristine and mid-charged, x = 1, samples (Fig. 4), where cationic and
anionic species form individual atomic columns (Supplementary Figs 7 and 8). In the image of the
pristine material, the O1 and O2 atomic columns are well resolved; the image also contains a weak
signal from the Li columns in the octahedral positions (see the intensity profiles). In the charged x =1
material, the signal from Li in the octahedral sites is drastically reduced in line with the NPD results.
The O1 positions remain clearly visible while the O2 positions become closely projected to the Ir
columns. The shortening of the Ir–O2 projected distances due to the shortening of O2–O2 distances
![Figure 4: Local structure of β-Li xIrO 3. [110] ABF-STEM images of the pristine β-Li 2IrO3 and two charged LixIrO3 materials with x = 1.0 and x = 0.35. The correspondence between the images and the projected structure of the materials is demonstrated on the magnified ABFSTEM in the middle (x 2). Intensity profiles highlight the structure changes in the material upon charging, i.e. the vanishing of the Li signal and formation of shortened projected Ir-O2 distances. The profiles are measured in the rectangul r regions indicated in the images and integrated over the width. The arrows indicate the approximate position of the atomic columns in the profiles.](/figures/figure-4-local-structure-of-b-li-xiro-3-110-abf-stem-images-21s0vwvp.webp)
