Reversible lithium insertion and copper extrusion in layered oxysulfides

Article abstract of DOI:10.1039/b605105g

All of the copper in the layered oxysulfides Sr₂MnO ₂Cu_1.5S₂ and Sr₂MnO ₂Cu_3.5S₃ may be extruded as the element and the copper ions replaced quasi-reversibly by lithiu

Full text of DOI:10.1039/b605105g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Reversible lithium insertion and copper extrusion in layered
oxysulfides{
Oliver J. Rutt, Gareth R. Williams and Simon J. Clarke*
Received (in Cambridge, UK) 10th April 2006, Accepted 10th May 2006
First published as an Advance Article on the web 6th June 2006
DOI: 10.1039/b605105g
Sr2MnO2Li3.8(1)S3 (2Li) respectively. The exchange of Cu+ by Li+
All of the copper in the layered oxysulfides Sr2MnO2Cu1.5S2
and Sr2MnO2Cu3.5S3 may be extruded as the element and the
copper ions replaced quasi-reversibly by lithium ions in
reductive topotactic ion exchange reactions; dramatic changes
in magnetic properties result.
and concomitant reduction of manganese suppresses long range
magnetic ordering.
All manipulations were carried out in an argon-filled dry box or
under nitrogen using Schlenk techniques. Black air-stable 1Cu and
2Cu5 were prepared{ on the 4 g scale by grinding together
stoichiometric amounts of SrS, Mn, MnO2, CuO and Cu2S and
heating pellets of the mixtures in alumina-lined evacuated silica
tubes at 1000 uC for 19 hours. Ion exchange was carried out by
stirring the powders with a 20-fold excess (Li : Mn ratio 20 : 1) of
0.25 M n-BuLi in a hexane/pentane mixture at 20 uC for 12 hours
(2Cu) or a 20-fold excess of 2.5 M n-BuLi in hexane at 50 uC for
4 days (1Cu). The reddish-black air-sensitive products were filtered
off, washed with pentane,{ dried under vacuum and stored in the
dry box. No reaction of 2Cu was observed under non-reductive ion
exchange conditions (28-fold excess of LiBr in refluxing hexanol at
160 uC for 8 hours).
The powder X-ray diffraction (PXRD)§ patterns of the lithiated
products showed mixtures of copper metal and compounds with
powder patterns resembling somewhat those of the starting
materials, but with different intensity distributions. The patterns
were indexed on similar tetragonal unit cells to those of 1Cu (body-
centred) and 2Cu (primitive),5 suggesting that the reactions are
topotactic. Refinement of the lattice parameters showed 1.45% and
0.96% increases in a and 4.0% and 3.4% increases in c on lithiation
of 1Cu and 2Cu respectively. The observations are consistent with
the replacement of Cu+ ions by the similarly-sized Li+ ions with the
extrusion of elemental copper and with reduction of manganese
resulting in longer Mn–O and Mn–S bonds.
So-called ‘‘soft’’ chemistry (chimie douce) in the solid state is
important in the synthesis of novel materials and in the control of
electron count, crystal structure and materials properties.1,2
Reductive insertion of alkali metals such as lithium into metal
oxides is the basis of recent advances in lithium-ion battery
technology.3,4 Non-redox ion exchange is also useful in the
synthesis ofnovel materials.2 These approachesenable the isolation
of products which are metastable with respect to the formation of
more thermodynamically stable products at elevated temperatures.
As part of an investigation of mixed-anion compounds we are
probing a series of layered oxysulfides containing oxide slabs
which incorporate the ‘‘harder’’, more electropositive and
oxophilic metals, and sulfide slabs which incorporate the ‘‘softer’’,
more polarising, chalcophilic metals. Layered structures in oxy-
sulfides are common because this is a simple way for two anions
with different sizes and chemical requirements to order crystal-
lographically. We have recently characterised an homologous
series of oxysulfides of general formula Sr2MnO2Cu2m2dSm+1
(m 5 1, 2, 3)5 and the intergrowth of the m 5 1 and
m 5 2 members.6 These compounds contain Sr2MnO2 slabs
separated by copper sulfide layers of varying thickness which
resemble fragments of the Cu22dS anti-fluorite structure. These
semiconducting compounds are all copper deficient (d y 0.5) with
manganese oxidised to the +(2 + d) (i.e. y+2.5) state to maintain a
full valence band.5 Below 30 K the compounds are A-type
antiferromagnets with ferromagnetic alignment of manganese
moments in each layer and antiferromagnetic coupling between
layers. On application of sufficiently large external magnetic fields
the materials become bulk ferromagnets. The correlation of these
properties with complex low-temperature structural distortions5
are under investigation. Here we demonstrate that all of the Cu+
ions in the copper sulfide slabs in the m 5 1 and m 5 2 members of
the series, Sr2MnO2Cu1.5S2 (1Cu) and Sr2MnO2Cu3.5S3 (2Cu),
may be replaced topotactically with Li+ ions under reductive ion
exchange conditions,7 to form Sr2MnO2Li1.86(3)S2 (1Li) and
When the Cu ions in the structural models of 1Cu and 2Cu were
replaced by Li, the intensities in both the X-ray and neutron§
diffractograms of 1Li and 2Li were reproduced well. Rietveld
refinements of the structures of 1Li and 2Li against powder
neutron diffraction (PND) data included, as an additional phase,
elemental copper which we have been unable to remove
chemically7 or physically without destroying the lithiated products.
Refinement of the structure of 1Li proceeded without incident
yielding satisfactory agreement factors (x2 5 3.643, wRp 5 0.0178,
RF2 5 0.0314) with lithium ions replacing the copper ions in 1Cu in
tetrahedral coordination by S. Assuming that the natural 6Li : 7Li
isotope ratio was maintained in the product, the fractional
occupancy of these sites was 0.943(6) yielding a refined composi-
tion for 1Li of Sr2MnO2Li1.88(1)S2 (chemical analysis:{
Sr2MnO2Li1.86(3)S2). Refinement of the structure of 2Li with
lithium ions located only on the ideal tetrahedral sites in the sulfide
layer (Fig. 1) yielded satisfactory agreement factors (x2 5 1.832,
wRp 5 0.0170, RF2 5 0.061) and a refined composition
Department of Chemistry, University of Oxford, Inorganic Chemistry
Laboratory, South Parks Road, Oxford, UK OX1 3QR.
E-mail: simon.clarke@chem.ox.ac.uk; Fax: +44 1865 272690;
Tel: +44 1865 272600
{ Electronic supplementary information (ESI) available: Tables of
crystallographic data; kinetic parameters and magnetic susceptibility of
1Li. See DOI: 10.1039/b605105g
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 2869–2871 | 2869

View Article Online
Exposure of the products to moist air resulted in reversal of the
ion exchange reaction. PXRD patterns (Fig. 2) of a sample of 2Li
(plus copper metal) taken at intervals after initial air exposure
show the intensities of the Cu peaks diminishing and the intensities
and positions of the peaks due to the tetragonal phase changing
from those characteristic of 2Li to those characteristic of 2Cu.
After 24 hours the 2Li and Cu peak intensities were almost zero
and 2Cu was restored as the major phase. However, the presence
of small copper peaks even after a week’s air exposure suggests
incomplete reversibility under these conditions. The mean size of
the elemental copper particles in 1Li and 2Li estimated by Scherrer
analysis was y50 nm which may explain the reactivity of Cu with
2Li in air.
Fig. 1 Rietveld refinement of the structure of 2Li against PND data.
Observed: points; fit: solid line; difference: lower line. Peak positions are
marked for 2Li (bottom), Cu and V (sample container) (top). The
structure diagrams shows the unit cells with 99% anisotropic displacement
The kinetics of lithiation of 2Cu (1Cu reacts more slowly) was
investigated using time-resolved in situ energy dispersive PXRD
measurements.§ The integrated intensity of the 001 reflection,
which accidently has zero intensity in 2Cu, was measured as a
function of time in four separate runs carried out at 30, 45, 50 and
60 uC. This was converted into the extent of the reaction at time t,
a(t) according to a(t) 5 I001(t)/I001(‘). This was modelled using the
Avrami–Erofe’ev expression10,11 which relates a(t) to the rate
constant, k, and an exponent, m, characteristic of the rate of
nucleation and the mechanism of growth:12 a(t) 5 1 2
exp(2(kt)m). Sharp–Hancock13 plots: ln(2ln(1 2 a)) vs. ln(t)
(Fig. 3) were used to extract m (the gradient) and k (intercept is
mlnk). The plots were linear, confirming the Avrami–Erofe’ev
model as a valid description of this system. The values of m
between 30 and 60 uC lay in the range 1.16(2) , m , 1.7(1). These
values are consistent with a 2D diffusion controlled process
following deceleratory nucleation.12 The rate constant k at 50 uC
was 9.9(2) 6 1024 s21, and an Arrhenius plot (lnk vs. 1/T) (Fig. 3)
˚
ellipsoids and selected interatomic distances in A.
Sr2MnO2Li3.62(3)S3 (chemical analysis:{ Sr2MnO2Li3.79(1)S3). An
improvement in agreement factors (x2 5 1.779, wRp 5 0.0168,
RF2 5 0.059) was obtained for 2Li when further lithium ions were
located in off-tetrahedral general positions located at short, but not
8
˚
˚
unprecedented, distances of 2.19(2) A and 2.11(2) A from two
sulfur atoms. The refined fractional occupancy of these sites
(which are not available in 1Li, and which are not the same off-
tetrahedral sites as those occupied by Cu+ in 2Cu5) was 0.047(2),
and this analysis yielded an overall composition of
Sr2MnO2Li4.44(6)S3 for 2Li. The large range of values for the Li
content obtained by Rietveld refinement of 2Li is likely a
consequence of the relatively small absolute value of the scattering
length of Li compared with those of the other elements and
contamination of the diffractogram by more elemental copper
than in the case of 1Li. Further diffraction and solid state NMR
measurements are required to determine whether the suggested
occupancy of the off-tetrahedral sites in 2Li is real. Given the
excellent contrast between the scattering lengths of Li (21.90 fm)
and Cu (+7.72 fm), the refinements of both 1Li and 2Li do indicate
full replacement of Cu by Li. The refined molar ratio of 2Li to Cu
metal obtained from the PND refinement was 1 : 3.53(3), also
consistent with complete reductive extrusion of all the copper ions
from 2Cu. On lithiation, the four Mn–O distances increase from
produced an estimated activation energy of 58 ¡ 8 kJ mol21
.
Magnetic susceptibility" and PND measurements on 2Cu5 show
that in zero applied magnetic field the compound is an A-type
antiferromagnet below 30 K with ferromagnetic MnO2 sheets
coupled antiferromagnetically to each other. The weak antiferro-
magnetism is overcome by an applied magnetic field of about 1.2 T
leading to bulk ferromagnetic behaviour in large applied fields
(Fig. 4)(fieldsexceeding7 Tarerequiredtoaccomplishthis in 1Cu5).
On lithiation both the low field antiferromagnetic ordering and the
˚
2.0061(1) s in 1Cu to 2.03518(4) A in 1Li (+1.45%) and from
˚
˚
2.0076(1) A in 2Cu to 2.02708(5) A in 2Li (+0.96%). The two
˚
Mn–S distances increase from 2.915(1) A in 1Cu to 3.056(1) A in
˚
˚
˚
1Li (+4.8%) and from 2.877(1) A in 2Cu to 3.036(2) A in 2Li
(+5.5%). The long Mn–S distances in both 1Li and 2Li lead to
elongation of the Mn displacement ellipsoids along the Mn–S
bonds. Bond valence (BV) calculations using the EUTAX
package9 produced BV sums for Mn of 2.51 (1Cu), 2.55 (2Cu),
2.26 (1Li) and 2.32 (2Li). The structural similarities of the
compounds suggest that the changes in BV sum on lithiation are
meaningful measures of the degree to which Mn is reduced and
imply compositions Sr2MnO2Li1.75S2 (1Li) and Sr2MnO2Li3.73S3
(2Li). The compositions obtained from chemical analysis,
and inferred from BV analysis and supported in the case of 1Li
by the PND refinement results suggest that reasonable composi-
tions are Sr2MnO2Li1.86(3)S3 (1Li) and Sr2MnO2Li3.8(1)S3 (2Li)
with Mn reduced to the +2.14(3) and +2.2(1) oxidation states
respectively.
Fig. 2 Evolution of the X-ray powder diffraction pattern following air
exposure over a 24 hour period of Sr2MnO2Li3.8(1)S3 (2Li). Lines marked
(*) are due to elemental copper. The diffractogram of Sr2MnO2Cu3.5S3
(2Cu) prior to lithiation is shown for comparison. The 001 reflection at
y8u 2h was used in the analysis of the in situ experiments.
2870 | Chem. Commun., 2006, 2869–2871
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Notes and references
{ MnO2 (ALFA 99.999%), Mn (Aldrich 99.99%), CuO (Aldrich 99.99%);
Cu2S was prepared by reacting Cu powder (ALFA 99.999%) with S
(ALFA 99.9995%) in an evacuated silica tube at 400 uC for 1 day then
700 uC for 4 days; SrS was prepared by reaction of SrCO3 (ALFA 99.99%)
with CS2 vapour (Aldrich 99.9%) in an Ar stream at 800 uC for 4 hours.
˚
Pentane for washing lithiated products was dried over 4 A molecular sieves,
refluxed over Na/K alloy under N2 and collected by distillation. Analysis
for Li was carried out using a Thermo Elemental Atomscan 16 ICP
analyser; Li ions were leached out by boiling the lithiated solid in 20% nitric
acid solution.
§ PXRD data were collected using a Philips X’Pert PW3209 diffractometer
operating with CuKa radiation in Bragg–Brentano geometry with samples
sealed inside air-tight cells. Time-of-flight PND data (POLARIS
diffractometer, ISIS, UK; detector banks at 35u, 90u and 145u 2h (0.5 ,
˚
d , 8 A)) were collected on 3 g samples sealed in vanadium cans. Rietveld
analysis was performed using the GSAS suite.16
Crystal data: Sr2MnO2Li1.86(3)S2 (1Li): T 5 295 K, tetragonal, space
Fig. 3 Arrhenius plot relating the rate of lithiation to reaction
temperature. Inset: the Sharp–Hancock13 plots from which the rate
constants were extracted (see text). In each case the lines are linear fits to
the data points.
˚
group I4/mmm (no. 139), a 5 4.07036(8), c 5 17.8721(4) A, Z 5 2;
Mn (0 0 0), Sr (0.5 0.5 0.09218(3)), O (0.5 0 0), S (0 0 0.17099(7)), Li (0.5 0
0.25; occupancy 0.943(6)); x2 5 3.643; wRp 5 0.0178.
Sr2MnO2Li3.8(1)S3 (2Li): T 5 295 K, tetragonal, space group P4/mmm
˚
(no. 123), a 5 4.05416(9), c 5 11.7962(3) A, Z 5 1; Mn (0 0 0), Sr (0.5 0.5
0.14101(8)), O (0 0.5 0), S1 (0 0 0.2574(2)), S2 (0.5 0.5 0.5), Li (0 0.5
0.3830(2); occupancy 0.904(7)); x2 5 1.832; wRp 5 0.0170. Note: this model
for 2Li has Li located only on the ‘‘ideal’’ tetrahedral sites; a second Li site
(0.103(5), 0.229(4), 0.415(2)) may be y5% occupied.
high field ferromagnetic ordering are apparently lost (Fig. 4). This
change is presumably a consequence of reduction of the Mn
oxidation state and changes in the between-layer superexchange
pathways effected by the substitution of Cu+ by Li+ and the large
increase in Mn–S bond length. On air exposure the magnetic
properties characteristic of 2Cu were not restored, confirming that
CCDC 604753–604760. For crystallographic data in CIF or other
electronic format see DOI: 10.1039/b605105g
In situ energy dispersive PXRD data were obtained on Station 16.4 of
the SRS, Daresbury Laboratory, UK using apparatus developed by
O’Hare and co-workers.17 0.3 g of Sr2MnO2Cu3.5S3 powder, 10 cm3 of
2.5 M n-BuLi solution (Li : Mn ratio 160 : 1) in hexanes and a glass-coated
magnetic follower were sealed inside a Pyrex ampoule. This was placed
inside a pre-heated aluminium heating block on Station 16.4 and the
reaction was stirred using a magnetic stirrer such that the amount of solid
exposed to the beam was constant with time. Slits in the heating block
allowed passage of the incident and diffracted beams. Scattered X-rays
were measured using three energy discriminating detectors located at 2–6u
the reaction is only quasi-reversible under these conditions.
6
Preliminary
experiments
on
Sr4Mn2O4Cu5S5
and
5
Sr2MnO2Cu5.5S4 suggest that full lithiation is possible in these
compounds. Investigation of the reversibility of the processes using
electrochemical methods, the lithium ion mobility and the
relevance of these compounds to novel lithium ion battery
materials14 is in progress,15 together with further investigation of
the reaction kinetics and the magnetic and electronic properties of
both the copper- and lithium-containing materials.
˚
2h and covering a d-spacing range of 1–46 A.
" Magnetic susceptibility (Quantum Design MPMS-XL SQUID magnet-
ometer) measurements were made on warming in an applied field of 10 mT
after cooling in zero field (zero-field cooled) and after cooling in the
measuring field (field-cooled). Magnetisation isotherms were measured in
applied fields of up to ¡5 T.
We thank the CCLRC for access to ISIS and the SRS, Dr R. I.
Smith for assistance on POLARIS, Dr D. Taylor and Mr A. Neild
for assistance on Station 16.4, and the EPSRC for providing a
studentship for O. J. R.
1 J. Rouxel and M. Tournoux, Solid State Ionics, 1996, 84, 141.
2 J. Gopalakrishnan, Chem. Mater., 1995, 7, 1265.
3 P. G. Bruce, Chem. Commun., 1997, 19, 1817.
4 J. L. Tirado, Mater. Sci Eng., R, 2003, 40, 103.
5 Z. A. Ga´l, O. J. Rutt, C. F. Smura, T. P. Overton, N. Barrier,
S. J. Clarke and J. Hadermann, J. Am. Chem. Soc., 2006, DOI: 10.1021/
ja060892o.
6 N. Barrier and S. J. Clarke, Chem. Commun., 2003, 164.
7 A. C. W. P. James, B. Ellis and J. B. Goodenough, Solid State Ionics,
1988, 27, 45.
8 A. Grzechnik, A. Vegas, K. Syassen, I. Loa, M. Hanfland and
M. Jansen, J. Solid State Chem., 2000, 154, 603.
9 N. E. Brese and M. O’Keeffe, Acta Crystallogr., Sect. B: Struct. Sci.,
1991, 47, 192.
10 M. Avrami, J. Chem. Phys., 1939, 7, 1103.
11 B. V. Erofe’ev, Dokl. Akad. Nauk SSSR, 1946, 52, 511.
12 S. F. Hulbert, J. Br. Ceram. Soc., 1969, 6, 11.
13 J. D. Hancock and J. H. Sharp, J. Am. Ceram. Soc., 1972, 55, 74.
14 M. Morcrette, P. Rozier, L. Dupont, E. Mugnier, L. Sannier, J. Galy
and J. M. Tarascon, Nat. Mater., 2003, 2, 755.
15 S. Indris, J. Cabana, O. J. Rutt, S. J. Clarke and C. P. Grey, manuscript
in prepration.
Fig. 4 The zero-field cooled (ZFC) magnetic susceptibilities (measuring
field 10 mT) of Sr2MnO2Cu3.5S4 (2Cu) (N) and the lithiated product
Sr2MnO2Li3.8(1)S4 (2Li) (#). The field-cooled susceptibility of 2Li is
coincident with the ZFC susceptibility. The inset shows the magnetisation
isotherms at 5 K measured on the same samples.
16 A. Larson and R. B. von Dreele, The General Structure Analysis System,
Los Alamos National Laboratory, Los Alamos, NM, 1985.
17 S. M. Clark, A. Nield, T. Rathbone, J. Flaherty, C. C. Tang, J. S. O.
Evans, R. J. Francis and D. O’Hare, Nucl. Instrum. Methods, 1995, 97,
98.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 2869–2871 | 2871