Corrosion Science 58 (2012) 237–241
Contents lists available at SciVerse ScienceDirect
Corrosion Science
journal homepage: www.elsevier.com/locate/corsci
Chemical stability of layered lithium cobalt nitrides in air
E. Panabière a, N. Emery a,⇑, S. Bach a,b, J.P. Pereira-Ramos a, P. Willmann c
a
Institut de Chimie et des Matériaux Paris-Est, GESMAT, UMR 7182 CNRS-Université Paris Est Créteil, 2, rue Henri-Dunant 94320 Thiais, France
Université d’Evry Val d’Essonne, Département Chimie, Bd François Mitterrand, 91025 Evry Cedex, France
c
Centre National d’Etudes Spatiales, 118 Avenue Edouard Belin, 31401 Toulouse Cedex 9, France
b
a r t i c l e
i n f o
Article history:
Received 21 November 2011
Accepted 28 January 2012
Available online 6 February 2012
Keywords:
XRD
Atmospheric corrosion
Oxidation
a b s t r a c t
The chemical stability in air of a layered lithium nitridocobaltate Li2.13Co0.43N has been investigated using
in situ XRD diffraction as a function of time. A high reactivity of the nitride with air moisture is found
with the rapid emergence of lithium hydroxide. The ageing process finally leads to the decomposition
of the nitride into lithium carbonate (Li2CO3), cobalt hydroxide (Co(OH)2) and the release of gaseous
NH3. The effect of the ageing process on the electrochemical properties of this promising anode material
for Li-ion batteries is reported. The discharge–charge properties of the compound slightly deteriorate
after few hours while they are dramatically affected after 15 h. The present results suggest handling of
these anodic materials in dry air would be possible with satisfactory electrochemical properties.
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
In the frame of the intense research focused on new negative
electrode materials for Li-ion batteries, the use of layered ternary
nitridocobaltates has been proposed to substitute carbonaceous
composites [1–7]. In the 1.4 V/0 V vs Li/Li+ potential window, one
set of electrochemical data reports that Li2.6Co0.4N acts as a conversion electrode material with contradictory high capacity values in
the range 760–400 mA h g1 (211–111 mC g1) and a remarkable
stability up to 50 cycles [1–5]. Recently, lithium cobalt nitrides
Li32xCoxN (0.1 6 x 6 0.44) have been proposed as genuine Li intercalation materials instead of conversion electrode material in the
narrower 1 V/0.02 V vs Li/Li+ potential window [6,7]. Whatever
the Co content all these nitrides are electroactive with a single step
around 0.6 V/0.7 V vs Li/Li+ for the discharge and charge processes
respectively. Their electrochemical behaviour observed is typical of
a Li intercalation compound and involves the Co2+/Co+ redox couple in the interlayer plane combined with the reversible accommodation of Li+ ions in the cation vacancies located in Li2N layers.
This intercalation process explains the excellent capacity retention
found after 50 cycles at high and low rates. A specific capacity of
180 mA h g1 (50 mC g1) at C/20 rate (i.e. insertion of 1 Li by
chemical formula in 20 h) and 130 mA h g1 (36 mC g1) at C/5
rate has been achieved for Li2.23Co0.39N. As for most of lithiated anode and cathode active materials for Li-ion batteries, handling of
lithiated metallic nitrides requires the use of inert atmosphere.
The high reactivity in air of a cubic metallic nitride Li7MnN4
has been briefly reported as being decreased with the use of the
⇑ Corresponding author. Tel.: +33 1 49 78 11 63; fax: +33 1 49 78 11 95.
E-mail address: [email protected] (N. Emery).
0010-938X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.corsci.2012.01.032
corresponding oxinitride Li7.9MnN3.2O1.6 [8]. However no detailed
data on the chemical stability in air of the layered lithiated ternary
nitrides in the Li–Co–N system are available in spite of their potential interest as rechargeable negative electrode. This has prompted
us to investigate the reactivity of the lithiated nitridocobaltate
Li2.13(1)Co0.43(1)N towards air at room temperature. The corrosion
process of the compound is followed by in situ X-ray diffraction
as a function of exposure time in air. Influence of the corrosion process on the anode performance is observed from the evolution of
discharge–charge properties as a function of exposure time.
2. Experimental
2.1. Chemical synthesis and structural analysis
The lithium cobalt nitride Li2.13(1)Co0.43(1)N was prepared using
a solid state route from the reagent grade Li3N (Alfa-Aesar 99.5%,
10 lm) and Co metal (Alfa-Aesar 99.8%, 1.6 lm). A mixture of
Li3N and Co in a ratio R (R = Co/Li3N) of 0.5 was first mixed and
ground in agate mortar and then pressed into pellet (13 mm diameter and thickness of 6–8 mm). The pellet of about 1 g is then
transferred in an alumina crucible enclosed in a silica reactor.
The furnace temperature reaches 650 °C with a heating rate of
300 °C/h then 700 °C using a heating rate of 50 °C/h and remained
at this temperature for 7 h under a nitrogen gas stream. During the
synthesis procedure, handling of reactive species and products was
carried out in an argon glove box. A sample was also aged for
3 weeks in a dessicator filled with a large amount of silicagel to
evaluate the role played by moisture.
X-ray Diffraction (XRD) experiments were performed using a
panalytical XPert pro apparatus equipped with a X’Celerator
238
E. Panabière et al. / Corrosion Science 58 (2012) 237–241
Fig. 1. Crystal structure of Li32xCoxN compound (P6/mmm space group). Nitrogen
ions are in site, lithium ions in 1b and 2c sites, cobalt ions in 1b site.
detector and using Co Ka radiation (k = 1.789 Å). Prior the study of
the ageing process, a pattern was recorded for 1 h in air-tight sample-holder between 20 < 2h < 100° and analysed by a Rietveld
refinement using GSAS package with EXPGUI interface [9,10]. During air exposure, XRD patterns were recorded every hour until 26 h
with an acquisition time of 5 min in the 20–60° 2h angular range.
Additional diagrams collected after 32 h and 3 weeks were performed with an acquisition time of 1 h each.
Scanning Electron Microscopy (SEM) experiments were carried
out using a LEO 1530 instrument.
2.2. Electrochemical measurements
The electrolyte used was 1 Mol.L1 LiPF6 in Ethylene Carbonate
(EC), DiEthyl Carbonate (DEC) and DiMethyl Carbonate (DMC)
solution (1:1:1, v/v, EQ-Be-LiPF6, MTI Corp.). Electrochemical studies were carried out in two-electrode cells (SwagelokÒ type) where
the lithium disc acts as a reference and auxiliary electrode. The
working electrode consisted of a stainless steel grid (6 mm diam.,
0.2 mm thickness) with a geometric area of 25 mm2 on which
the cathode material was pressed (5000 kg cm2). The cathode
Fig. 3. SEM pictures of the lithiated nitridocobaltate Li2.13Co0.43N (a) before, (b)
after 60 h and (c) 3 weeks of air exposure.
was made of a mixture of active material (70% wt), acetylene black
(22% wt) and Teflon as binder agent (8%). Electrochemical
measurements were made with a Solartron 1470 apparatus.
Fig. 2. XRD pattern and Rietveld refinement of Li2.13Co0.43N. Lower and upper marks indicate Lithiated transition metal nitride and Li2O reflexions respectively. Inset: 36–48°
expanded view.
E. Panabière et al. / Corrosion Science 58 (2012) 237–241
239
Fig. 4. (a) XRD patterns of Li2.13Co0.43N exposed in air for 0, 2, 5, 10, 15 and 26 h. (b) evolution of the XRD pattern as a function of time in the 35–45° 2h range.
Galvanostatic discharge–charge tests between 1 and 0.02 V vs Li/
Li+ were carried out at C/10 rate (i.e. variation of 1 Li+ by chemical
formula in 10 h).
3. Results and discussion
Layered ternary lithiated nitridocobaltates crystallise in the aLi3N hexagonal type structure (space group P6/mmm). This structure is well described by a pure Li+ layer surrounded by Li2N
planes (Fig. 1). According to the chemical analysis, structural and
electrochemical data reported elsewhere on Li32xCoxN (0.1 6
x 6 0.44) [6,7], Co2+ ions substitute a part of the Li+ ions layered
(1b site) with the simultaneous emergence of an equivalent
amount of Li vacancies in the 2c sites of the Li2N plane.
X-ray powder diffraction pattern of Li2.13Co0.43N is shown in
Fig. 2. Rietveld refinement in the hexagonal space group P6/mmm
leads to satisfactory results. The refined cell parameters are
a = 3.7245(2) Å and c = 3.6748(2) Å and the agreement factors are
Rwp = 2.18% and Rp = 1.62% with goodness of fit v2 = 1.158. In
addition as can be seen from the expanded view of the 36–48°
2h range in the inset of Fig. 2, a low intensity peak located at
39.2° indicates the presence of a small amount of Li2O in the sample (ICSD pattern 00–012-0254). The chemical composition of the
lithiated metallic nitride was refined with a constraint on the site
occupancies according to the developed formula [Li1xCox]1b[Li2x/x]2cN1a (where / represent the lithium vacancies). The corresponding chemical formula, Li2.13(1)Co0.43(1)N (x = 0.43) is found,
associated with the cell parameters reported above, in excellent
agreement with the evolution of lattice parameters as a function
of the Co content in Li3-2xCoxN [6]. The Li2O amount was estimated
close to 2.8% wt (i.e. Li2.13Co0.43N, 0.05Li2O) from the Rietveld
refinement.
The morphology observed by SEM indicates the powder is made
of large aggregates (size 10 lm) made of smaller particles (0.5–
3 lm) (Fig. 3).
In Fig. 4a are compared selected XRD patterns in the 20–60° 2h
range recorded after exposure in air for 0, 2, 5, 10, 15 and 26 h. The
intensity of XRD peaks of Li2.13Co0.43N significantly decreases during the first five hours. This intensity loss for 100, 101, 110 peaks is
accompanied by a slight shift showing a small change in the unit
240
E. Panabière et al. / Corrosion Science 58 (2012) 237–241
Fig. 6. Comparison of the XRD patterns of Li2.13Co0.43N as prepared and exposed
3 weeks in ambient atmosphere and in dry air. Inset, SEM micrograph of the sample
exposed in dry air.
Fig. 5. Comparison of XRD patterns of Li2.13Co0.43N air exposure of 32 h and
3 weeks.
cell parameters. Indeed, a decreases from 3.725 to 3.686 Å after 2 h
and remains stable thereafter. At the same time, the interlayer c
parameter increases slightly from 3.675 to 3.716 Å. These variations cannot be only explained by the variation of the cobalt content and probably reflect the nitrogen loss.
The first degradation product observed is LiOH with the main
diffraction peaks at 23.8, 38.0 and 41.8° (ICSD pattern 01–0851064), even after 2 h. LiOH formation is a clear indication of the
moisture affinity of the nitride. Moreover, the Li2O reflexion identified at 39.2° on the starting material doesn’t grow up during all
the experiment, according to its relative stability in air. Hence,
Li2O is not involved in the degradation mechanism of the ternary
nitride. The LiOH amount rapidly increases to reach a maximum
after 5 h and then decreases with time to the benefit of lithium carbonate. Indeed, Li2CO3 can be clearly identified after 5 h of degradation (reflexions at 24.8, 35.7, 37.1 39.8, 43.2 and 46.3°, ICSD
pattern 00–022-1141).
Fig. 4b illustrates the XRD pattern evolution in the 35–45° 2h
range up to 26 h and provides an overall view of all the main
changes occurring with time. It illustrates the continuous augmentation of the Li2CO3 amount at the expense of LiOH which slightly
decreases and the correlative sharp decrease with time for the 101
reflexion of Li2.13Co0.43N.
After 26 h, no cobalt species has been identified. For longer
time, drastic changes take place as can be seen in XRD patterns
after 32 h and after 3 weeks (Fig. 5). Indeed, after 32 h, cobalt
hydroxide peaks located at 21.9, 44.3°, (ICSD pattern 00–030-
Fig. 7. Discharge–charge curves of Li2.13Co0.44N as a function of the exposure time
(C/10 in the 1 V/0.02 V vs Li/Li+ potential range; the capacity is expressed in lC per
mole of Li2.13Co0.44N).
0443) are observed for the first time. These two small and broad
peaks are indicative of a poor structural organisation of Co(OH)2.
A significant increase in the Co(OH)2 amount and its crystallinity
is observed after 3 weeks (Fig. 5). After 3 weeks, the lithiated
metallic nitride is deeply affected with the quasi disappearance
of the 100 and 101 reflexions to the benefit of Co(OH)2 and Li2CO3
peaks. A mixture of LiOH, Li2CO3, Co(OH)2 and metallic nitride with
a strongly modified chemical composition is then observed.
E. Panabière et al. / Corrosion Science 58 (2012) 237–241
SEM micrographs reported in Fig. 3a show the grain morphology prior air exposure and after 60 h and 3 weeks in air (Fig. 3b
and c respectively). After 60 h, the surface morphology significantly changes from dense particles to a porous material made of
small platelets 0.5 lm long with sharp edges, probably related
to the release of ammonia. The appearance of this porosity promotes the further degradation process. After 3 weeks, the same
kind of morphology is obtained but with smooth edges, which
could be an indication of the complete degradation of the material.
The present XRD study clearly indicates the high reactivity of
the lithiated metallic nitride Li32xCoxN towards moisture rapidly
leading to the formation of LiOH and Co(OH)2. The reaction could
be written:
Li32x Cox N þ 3H2 O ! ð3 2xÞLiOH þ xCoðOHÞ2 þ NH3 "
The ageing mechanism is supported by the XRD experiments
and by the characteristic pungent odour of ammonia emitted by
the sample during air exposure. The second step, leading to the formation of Li2CO3, is ascribed to the affinity of LiOH for carbon
dioxide.
One can notice that this step is favourable to the propagation of
the degradation since some water is produced and can then react
with the metallic nitride. It must be outlined that the lithiated
transition metal nitride is still present after one day of exposure
(Fig. 4b and a). This can be explained by the large particles, inducing a slow diffusion of water from the surface to the core. The reactivity toward air moisture of a binary metallic nitride such as AlN
in specific conditions has been reported owing a similar reaction
process leading to the formation of gaseous ammonia and aluminium hydroxide [11].
The high reactivity of the lithiated metallic nitride with moisture is also supported by the comparison of XRD data obtained
for the same sample aged for 3 weeks in the dried air atmosphere
of a dessicator (Fig. 6) and those previously reported in Fig. 5. The
corresponding diffraction pattern shows that the sample is much
less affected in dry air; only a small amount of LiOH and Li2O is detected. Moreover ageing in dry air does not seem to change the
morphology of Li2.13Co0.43N particles, as seen in the SEM micrograph (inset in Fig. 6).
Fig. 7 shows the voltage profile of Li2.13Co0.43N electrodes discharged and then charged in the 1.0 V/0.02 V vs Li/Li+ voltage range
at C/10 rate. The second cycle is reported in order to avoid the SEI
formation to be included in the inserted ratio of lithium in the
material during the reduction. The discharge curve shows one
main process located at 0.55 V vs Li/Li+ which corresponds to the
insertion of 0.35 Li per mole of nitride (3.6 lC mol1), in good accord with previous studies on the Li–Co–N system [6]. The charge
process is located near 0.7 V vs Li/Li+ as a single signal and is practically quantitative which is well consistent with the reversible
241
reduction of Co (II) into Co (I) and the correlative Li intercalation
in the available Li vacancies. The discharge–charge profiles and
the insertion ratio are first little affected after few hours in air,
showing that only a small part of the nitride reacts with moisture.
However with increasing time (15 h), the discharge potential
strongly decreases and the specific capacity declines. At longer
time (32 h), the electrochemical behaviour dramatically changes
with a poor capacity of only 2.1 lC mol1 (i.e. 0.2 Li inserted) which
can be correlated with the significant appearance of Co(OH)2
(Fig. 5). After 3 weeks, the anode material is practically inactive
due to the destruction of the lithiated nitridocobaltate (Fig. 5).
4. Conclusions
We have reported here the ageing process in air of a lithium
nitridocobaltate which is a promising anode material for Li-ion
batteries. To our knowledge, this topic has never been addressed
in spite of the reactive character expected for such materials in
air and their potential interest as electrode material. In situ XRD
experiments vs. exposure time combined with the electrochemical
behaviour observed for aged compounds show the discharge–
charge properties are little affected after few hours, significantly
after 15 h and dramatically at longer times. The decrease both in
the lithium uptake and discharge potential is due to the reaction
of the nitride with air moisture leading to its significant destruction with LiOH and Co(OH)2 formation as well as NH3 release. This
study clearly indicates that lithiated nitridocobaltates present a
greater affinity for moisture than oxygen. These results suggest
handling of these anodic materials in dry air would be possible
with satisfactory electrochemical properties.
References
[1] T. Shodai, S. Okada, S. Tobishima, J. Yamaki, Solid State Ionics 86–88 (1996)
785–789.
[2] T. Shodai, S. Okada, S. Tobishima, J. Yamaki, J. Power Sources 68 (1997) 515–
518.
[3] M. Nishijima, T. Kagohashi, N. Imanishi, Y. Takeda, O. Yamamoto, S. Kondo,
Solid State Ionics 83 (1996) 107–111.
[4] M. Nishijima, T. Kagohashi, Y. Takeda, M. Imanishi, O. Yamamoto, J. Power
Sources 68 (1997) 510–514.
[5] T. Shodai, Y. Sakurai, T. Suzuki, Solid State Ionics 122 (1999) 85–93.
[6] J.B. Ducros, S. Bach, J.P. Pereira-Ramos, P. Willmann, J. Power Sources 175
(2008) 517–525.
[7] J.B. Ducros, S. Bach, J.P. Pereira-Ramos, P. Willmann Electrochem, Commun. 9
(2007) 2496–2500.
[8] J. Cabana, N. Dupré, C.P. Grey, G. Subias, M.T. Caldés, A.-M. Marie, M.R. Palacin,
J. Electrochem. Soc. 152 (2005) A2246–A2255.
[9] A.C. Larson, R.B. Von Dreele, ‘‘General Structure Analysis System (GSAS)’’, Los
Alamos National Laboratory Report LAUR 2004, 86–748 .
[10] B.H. Toby, J. Appl. Cryst. 34 (2001) 210–213.
[11] J. Li, M. Nakamura, T. Shirai, K. Matsumaru, C. Ishizaki, K. Ishizaki, J. Am. Ceram.
Soc. 89 (2006) 937–943.