Lithium storage properties of nanocrystalline eta-Cu6Sn5 alloys prepared by ball-milling

Article abstract of DOI:10.1016/S0925-8388(99)00804-X

Nanocrystalline eta-Cu₆Sn₅ alloy powders were prepared by high energy ball-milling. Lithium ions can be electrochemically inserted into Cu₆Sn₅ electrodes, alloying with Sn to form Li_xSn alloys. Simultaneously, an inactive Cu matrix is produced, which could buffer the volume expansion of the active Sn due to the formation of Li_xSn alloys. It appears that lithium can react reversibly with the nanocrystalline Cu₆Sn₅ alloy to some extent. A reversible capacity of approximately 400 mAh/g has been achieved in preliminary tests. The results of ex-situ X-ray diffraction support the above proposal.

Full text of DOI:10.1016/S0925-8388(99)00804-X

Journal of Alloys and Compounds 299 (2000) L12–L15
L
www.elsevier.com/locate/jallcom
Letter
Lithium storage properties of nanocrystalline eta-Cu6Sn5 alloys prepared
by ball-milling
*
G.X. Wang , L. Sun, D.H. Bradhurst, S.X. Dou, H.K. Liu
Energy Storage Materials Program, Institute for Superconducting and Electronic Materials, University of Wollongong, Northfileds Avenue,
Wollowong, NSW 2522, Australia
Received 5 November 1999; accepted 18 November 1999
Abstract
Nanocrystalline eta-Cu6Sn5 alloy powders were prepared by high energy ball-milling. Lithium ions can be electrochemically inserted
into Cu6Sn5 electrodes, alloying with Sn to form LixSn alloys. Simultaneously, an inactive Cu matrix is produced, which could buffer the
volume expansion of the active Sn due to the formation of LixSn alloys. It appears that lithium can react reversibly with the
nanocrystalline Cu6Sn5 alloy to some extent. A reversible capacity of |400 mAh/g has been achieved in preliminary tests. The results of
ex-situ X-ray diffraction support the above proposal.
2000 Elsevier Science S.A. All rights reserved.
Keywords: Cu–Sn alloys; Ball-milling; Nanocrystalline alloy; Li-cell electrode; Electrochemical capacity
1. Introduction
We think that lithium alloying with Sn in eta-Cu6Sn5
could be achieved by preparing nanosize eta-Cu6Sn5
powders, in which 1 mole of Cu6Sn5 could combine with
22 moles of Li to form Li22Sn5 and generate 6 moles of
Cu. The theoretical capacity for this process is 608 mAh/
g, which is almost twice that of the mechanism to form
Li13Cu6Sn5 alloy.
This paper describes the synthesis of nanocrystalline
eta-Cu6Sn5 alloy powders. Their electrochemical perform-
ance as electrodes in lithium cells was characterized. The
mechanism for lithium insertion in Cu6Sn5 alloy was
studied by ex-situ X-ray diffraction.
The tin based glass and intermetallic alloys recently
have attracted a worldwide attention as lithium storage
materials because of their high volumetric capacity. The
tin based amorphous composite oxide (TCO) demonstrated
a high reversible capacity of 650 mAh/g. However, a 37%
irreversible capacity in the initial cycle due to the forma-
tion of Li2O, prevented it from further commercial applica-
tion [1–4]. Theoretically, the intermetallic alloys seem not
to have this problem because there is no oxygen in the
alloy structure. The mechanism for reversible lithium
insertion and extraction in MSn alloys relies on lithium
alloying with Sn to form LixSn alloy. The nanosize
inactive M matrix is generated simultaneously, which
buffers the volume expansion of the active phase Sn due to
its alloying with Li. Nanosize or amorphous alloy powders
are usually required to realize the above process. Several
alloys such as SnSb, Sn2Fe and SnFe3C have been
investigated and the process of lithium alloying with Sn
has been confirmed by in-situ X-ray diffraction [5–9].
However, a different mechanism for lithium insertion into
the crystalline eta-Cu6Sn5 alloy has been demonstrated to
be a two phase reaction to yield LixCu6Sn5 (x513),
corresponding to a theoretical capacity of 358 mAh/g [10].
2. Experimental
Nanocrystalline eta-Cu6Sn5 alloy powders were pre-
pared by high energy ball-milling. The elemental powders
of Cu and Sn as well as hardened steel balls were put into
the steel vial. The vial was initially evacuated and then
filled with argon. The ball-milling was carried out in a
Pulverisette-5 ball-milling machine (LABTECHNICS,
Australia). For comparison, crystalline eta-Cu6Sn5 alloy
was also synthesized by sintering a pellet of a stoichio-
metric mixture of Cu and Sn powders in argon at 4008C
for 15 h. The sintered pellet was ground and sieved
through a 325-mesh stainless steel sieve. The alloy pow-
*Corresponding author. Fax: 161-2-4221-5731.
E-mail address: gw14@uow.edu.au (G.X. Wang)
0925-8388/00/$ – see front matter
PII: S0925-8388(99)00804-X
2000 Elsevier Science S.A. All rights reserved.

G.X. Wang et al. / Journal of Alloys and Compounds 299 (2000) L12–L15
L13
ders were characterized by X-ray diffraction on a Philips
PW 1730 diffractometer with a Cu-Ka radiation.
5–10 nm, using the Scherrer formula: d 5 Kl
Y
b Cos u
The Cu6Sn5 electrodes were prepared by dispersing 97
wt% Cu6Sn5 powders and 3 wt% polyvinylidene fluoride
(PVDF) binder in dimethyl phthalate to form a slurry,
which was spread on to the copper foil. No carbon
additives were added because Cu6Sn5 alloy has good
electrical conductivity and also in order to eliminate the
ambiguity caused by the insertion of lithium into carbon.
The typical Cu6Sn5 electrode contains about 1 mg active
materials and 80 mm in thickness. Li/Cu6Sn5 coin cells
(CR2032) were assembled. The electrolyte was 1 M LiPF6
in EC (ethylene carbonate) and DMC (dimethyl carbonate)
(Battery grade, MERCK KGaA, Germany). The cells were
charged and discharged at a constant current density of
0.20 mAh/cm2.
[11]. However, a small amount of tin or copper oxides
were formed during the long term ball-milling, probably
due to the very active nature of the nanosize Cu6Sn5 alloy
and traces of oxygen present in the ball-milling vial.
The Li/Cu6Sn5 cells were cycled between 0 and 1.5 V
at a constant current density of 0.20 mA/cm2. The lithium
insertion into and extraction from Cu6Sn5 alloy are defined
as discharge and charge respectively. The first discharge
and charge profiles of Cu6Sn5 alloy electrodes are shown
in Fig. 2. The ball-milled nanosize Cu6Sn5 electrode
demonstrated markedly different behaviour for lithium
insertion compared to that of the sintered crystalline
Cu6Sn5 alloy. In the first discharge, nanosize Cu6Sn5 alloy
electrode delivered a capacity of 688 mAh/g, corre-
sponding to approximately 25 moles of lithium insertion
per mole of Cu6Sn5 alloy. Provided lithium alloys with Sn,
the maximum lithium insertion achievable theoretically is
that each mole of Cu6Sn5 alloy can combine with 22
moles of Li to form Li4.4Sn alloy, with a capacity of 608
mAh/g. As shown in Fig. 1, a small amount of Sn and Cu
oxides are generated during the long term ball-milling. The
extra lithium insertion capacity of 80 mAh/g observed
might be caused by lithium reacting with Sn or Cu oxides
to form Li2O. And this part of lithium cannot be extracted
again, which is similar to the behaviour of tin based glass.
A reversible capacity of 412 mAh/g was achieved in the
first charging which is 60% of the initial discharge
capacity. However, the first discharge capacity for sintered
Cu6Sn5 alloy was only 284 mAh/g. The theoretical
capacity for lithium insertion into Cu6Sn5 to form
Li13Cu6Sn5 (isostructural with Li2CuSn) is 358 mAh/g.
Here, we did not achieve the theoretical capacity, possibly
due to the relatively large particle size. With the scrutiny of
the differential capacity vs. voltage profiles (the insets in
Fig. 2), it might be seen that the ball-milled nanosize
Cu6Sn5 electrode shows several reduction peaks, corre-
sponding to the formation of Li2O and various LixSn
alloys, which is similar to the behaviours of Sn, SnO, SnO2
and SiSnO3 electrodes [12]. By contrast, the crystalline
Cu6Sn5 alloy electrode only shows one sharp reduction
peak, which corresponds to the topotactic insertion of
lithium into Cu6Sn5 alloy structure to form LixCu6Sn5.
This is in agreement with the report from K.D. Kepler et
al. [10]. Consequently, the two different mechanisms for
lithium reaction with ball-milled nanosize and sintered
crystalline Cu6Sn5 alloys are proposed, as follows:
After cycling, the cells were dismantled in an argon
filled glove-box (Unilab, Mbraun, USA) and the Cu6Sn5
electrodes were taken out. The Cu6Sn5 electrodes were
covered by a plastic film for ex-situ X-ray diffraction.
3. Results and discussion
The eta-Cu6Sn5 alloy has a hexagonal structure with a
space group of P63 /mmc. It can be synthesized by either
sintering the mixture of Cu and Sn powders or ball milling
Cu and Sn powders. Fig. 1 shows X-ray diffraction
patterns of Cu6Sn5 alloys. The sintering process produced
a typical crystal Cu6Sn5 alloy with a small amount of
impurity phase of the unreacted Sn, while the ball-milling
generated fine Cu6Sn5 alloy powders. As the ball-milling
time was increased, the diffraction peaks became
broadened, indicating a decrease of the particle size. When
ball-milled for 110 h, very broad diffraction peaks were
observed, from which the grain size was calculated to be
→
22Li 1 Cu Sn (nanosize) 5Li4.4Sn 1
←
6
5
6Cu (Alloying process)
(1)
13Li 1
Fig. 1. X-ray diffraction patterns of eta-Cu6Sn5 alloys. (a) Cu6Sn5 alloy
prepared by sintered process. (b) Cu6Sn5 alloy prepared by ball-milling
for 18 h. (c) Cu6Sn5 alloy prepared by ball-milling for 46 h. (d) Cu6Sn5
alloy prepared by ball-milling for 110 h. * a-Sn, . Cu2O or SnO2.
→
Cu Sn (crystalline) Li Cu Sn (Topotactic insertion p-
←
6
5
13
6
5
rocess)
(2)

L14
G.X. Wang et al. / Journal of Alloys and Compounds 299 (2000) L12–L15
Fig. 2. The first discharge/charge profiles of Li/Cu6Sn5 cells. The insets are the corresponding differential capacity vs. voltage. (a) Ball-milled Cu6Sn5
alloy electrode (ball-milling for 110 h). (b) Sintered crystalline Cu6Sn5 alloy electrode.

G.X. Wang et al. / Journal of Alloys and Compounds 299 (2000) L12–L15
L15
insertion into crystalline Cu6Sn5 to form the compound
LixCu6Sn5.
Fig. 4 shows the discharge capacity vs. cycle number for
the two different kinds of Cu6Sn5 alloys. The capacity for
both nanosize and normal crystal size Cu6Sn5 alloys
declines on cycling quite quickly. Further work needs to be
done to identify the mechanism for this phenomenon and
to improve the cyclability of Cu6Sn5 alloy electrodes.
4. Conclusions
It was found that there were two different mechanisms
for lithium reacting with ball-milled nanosize and sintered
crystalline eta-Cu6Sn5 alloys. The reaction of lithium with
nanosize Cu6Sn5 is an alloying process, in which lithium
alloys with Sn to form LixSn and generates Cu simul-
taneously. By contrast, an insertion process takes place in
the crystalline Cu6Sn5 alloy, in which Li ions are inserted
into the Cu6Sn5 structure to form LixCu6Sn5. Nanocrystal-
line Cu6Sn5 alloy demonstrated a much higher initial
lithium storage capacity than that of the crystalline
Cu6Sn5.
Fig. 3. The ex-situ X-ray diffraction patterns of Cu6Sn5 alloy electrodes
in the fully discharged state. (a) Ball-milled nanosize Cu6Sn5 alloys
(ball-milling for 110 h). (b) Sintered crystalline Cu6Sn5 alloys.
In order to confirm these mechanisms, ex-situ X-ray
diffraction was performed on the Cu6Sn5 electrode in the
fully discharged state. The ex-situ XRD patterns are shown
in Fig. 3. For the ball-milled nanosize Cu6Sn5 electrode in
the heavily lithiated state, Li4.4Sn is identified by the two
diffraction peaks at 2u523.58 and 39.98. The broad 2u5
39.98 peak suggests a very fine particle size of Li4.4Sn.
However, Li2CuSn phase was detected for sintered crys-
talline Cu6Sn5 in the fully discharged state, characterized
by the diffraction peaks at 2u524.28 and 40.78 respective-
ly. The unreacted Cu6Sn5 phase was also presented in the
XRD pattern. These unreacted Cu6Sn5 alloys could be
responsible for the low discharge capacity of crystalline
Cu6Sn5 alloy electrode due to the large particle size and
limited kinetics for lithium insertion. No any Li–Sn alloy
phase was found, which confirmed the process of lithium
Acknowledgements
Financial support from Department of Employment,
Education, Training and Youth Affairs, Australia is grate-
fully acknowledged.
References
[1] Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, T. Miyasaka,
Science 76 (1997) 1395–1399.
[2] Y. Idota, US Patent 5, 478, 671 (1995).
[3] I.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 (1997) 2943–
2948.
[4] W. Liu, X. Huang, Z. Wang, H. Li, L. Chen, J. Electrochem. Soc.
145 (1998) 59.
[5] J. Yang, M. Wachtler, M. Winter, O. Besenhard, Electrochem.
Solid-state Lett. 2 (1999) 131–133.
[6] O. Mao, R.L. Turner, I.A. Courtney et al., Electrochem. Solid-State
Lett. 2 (1999) 3–5.
[7] O. Mao, R.A. Dunlap, J.R. Dahn, J. Electrochem. Soc. 146 (1999)
405–413.
[8] O. Mao, J.R. Dahn, J. Electrochem. Soc. 146 (1999) 414–422.
[9] O. Mao, J.R. Dahn, J. Electrochem. Soc. 146 (1999) 423–427.
[10] K.D. Kepler, J.T. Vaughey, M.M. Thackeray, Electrochem. Solid-
State Lett. 2 (1999) 307–309.
[11] Traces Program, Diffraction Technology Pty. Ltd., Australia.
[12] I.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 (1997) 2045.
Fig. 4. Discharge capacity vs. cycle number for Li/Cu6Sn5 cells.