Lithiation/delithiation performance of Cu6Sn5 with carbon paper as current collector

Article abstract of DOI:10.1149/1.1839466

Lithiation/delithiation of Cu₆Sn₅ both electrochemically and mechanochemically prepared on a carbon paper (CP) current collector was investigated by repeated galvanostatic cycles with lime-limited charges to insert different amounts of lithium per Sn atom (Li/Sn) into the alloy. The results demonstrated that, unlike Cu₆Sn₅ electrodeposited on Cu foil, both electrochemical and mechanochemical Cu ₆Sn₅ on CP can undergo hundreds of lithiation/delithiation cycles with Li/Sn > 2 at high current density (0.74 mA cm^-2), thereby delivering constant amounts of charge with coutombic efficiency near 100%. The Cu₆Sn₅/CP electrodes with a loading of 6.6 mg cm^-2 and charge limited to 2.43 Li/Sn yielded a specific capacity of 330 mAh g^-1 and a capacity per geometric area of 2.18 mAh cm ^-2. These values compare well with those of graphite-based anodes of commercial batteries. The significant improvement in cyclability performance of the Cu₆Sn₅/CP electrodes is due to the CP current collector's three-dimensional conductive matrix, which preserves the electric contact over cycling so that alloy cracking becomes less detrimental for the electric contact.

Full text of DOI:10.1149/1.1839466

Journal of The Electrochemical Society, 152 ͑2͒ A289-A294 ͑2005͒
A289
0013-4651/2004/152͑2͒/A289/6/$7.00 © The Electrochemical Society, Inc.
LithiationÕDelithiation Performance of Cu6Sn5 with Carbon
Paper as Current Collector
,z
*
*
Catia Arbizzani, Mariachiara Lazzari, and Marina Mastragostino
`
University of Bologna, Unita Complessa di Istituti di Scienze Chimiche, Radiochimiche e Metallurgiche,
40127 Bologna, Italy
Lithiation/delithiation of Cu6Sn5 both electrochemically and mechanochemically prepared on a carbon paper ͑CP͒ current collec-
tor was investigated by repeated galvanostatic cycles with time-limited charges to insert different amounts of lithium per Sn atom
͑Li/Sn͒ into the alloy. The results demonstrated that, unlike Cu6Sn5 electrodeposited on Cu foil, both electrochemical and
Ͼ
mechanochemical Cu6Sn5 on CP can undergo hundreds of lithiation/delithiation cycles with Li/Sn
2 at high current density
͑0.74 mA cmϪ2͒, thereby delivering constant amounts of charge with coulombic efficiency near 100%. The Cu6Sn5 /CP electrodes
with a loading of 6.6 mg cmϪ2 and charge limited to 2.43 Li/Sn yielded a specific capacity of 330 mAh gϪ1 and a capacity per
geometric area of 2.18 mAh cmϪ2. These values compare well with those of graphite-based anodes of commercial batteries. The
significant improvement in cyclability performance of the Cu6Sn5 /CP electrodes is due to the CP current collector’s three-
dimensional conductive matrix, which preserves the electric contact over cycling so that alloy cracking becomes less detrimental
for the electric contact.
© 2004 The Electrochemical Society. ͓DOI: 10.1149/1.1839466͔ All rights reserved.
Manuscript submitted March 30, 2004; revised manuscript received June 23, 2004. Available electronically December 20, 2004.
Lithium-ion battery technology is well consolidated in the field
of portable equipment.1 Nevertheless, research efforts in the lithium-
ion battery field have been exceedingly active in attempts to ame-
liorate the performance of these devices in terms of electrode mate-
rials, electrolyte media, and design. For negative electrodes, the
research is mainly focused in two directions: ͑i͒ the optimization of
the performance of carbonaceous materials,2-5 and (ii) the search
for alternative materials to the carbonaceous ones presently used,
which has led to the development of metal oxides6,7 and metal
alloys.8 While Li/metal alloys are very attractive for their specific
capacity, their high volume change, which is related to the insertion/
removal of lithium, causes pulverization of the electrode and, hence,
loss of electrical contact among particles, which is responsible for
the poor cycle life. For example, tin can sustain only a few charge/
discharge cycles before losing structural integrity, even when the
charge is limited to 1.5 Li/Sn atomic ratio.9 Several strategies to
reduce this problem have been pursued, including the preparation of
superfine and nanometric materials, intermetallic compounds, and
alloy/carbon composites, but none has been decisive.10-15
While the relationship between morphology and electrochemical
properties of alternative materials to graphite and of graphite itself
has been under study,14,16-18 minor studies have been devoted to the
nature and morphology of current collectors. Recently, Fujitani
et al.13,19 demonstrated the effect of the surface roughness of copper
foil, the current collector generally used for the negative electrode,
on improving cyclability data. This improvement is attributed to the
formation of a microcolumnar structure of the deposits, which
cracks periodically in accordance with the surface profile of the Cu
rough foil, providing spaces for the active material to swell either
upwards or sideways during charge. By contrast, flat surfaces crack
randomly, so that the active material pulverizes and delaminates
from the foil.
148% with the same degree of lithiation and an expansion of only
59% when 2 Li are reversibly inserted to give Li2CuSn, according
to the following reactions22,23
4.4Liϩ ϩ 4.4eϪ ϩ Sn → Li4.4Sn
10Liϩ ϩ 10eϪ ϩ Cu6Sn5 → 5Li2CuSn ϩ Cu
22Liϩ ϩ 22eϪ ϩ Cu6Sn5 → 5Li4.4Sn ϩ 6Cu
͓1͔
͓2͔
͓3͔
The lithiation of Cu6Sn5 as in Reaction 2 occurs at 0.4 V vs. Li.
This potential value is an important feature of Cu6Sn5 , because, in
lithium-ion batteries using flammable organic electrolytes, anode
materials in which the lithiation process starts at a more positive
potential than in graphite are required to avoid the risk of Li metal
deposition. The theoretical specific capacity of Cu6Sn5 was esti-
mated to be 358 mAh gϪ1 ͑a value near that of graphite͒ when
lithiated to the composition Li13Cu6Sn5 , i.e., when the Li/Sn atomic
ratio is 2.6.24 Until now the cycling stability of Cu6Sn5 on conven-
tional metal current collectors ͑Cu,Ni͒ has not been satisfactory be-
cause the volume changes, still present, and the Cu extrusion lead to
a loss of electric contact among particles and delamination from the
current collectors.
We examined commercial carbon paper ͑CP͒ as a new current
collector for Cu6Sn5 negative electrodes for lithium-ion batteries.
CP is a porous material of high electrical conductivity and used as a
substrate/electrode in electrochemical devices like supercapacitors
and fuel cells capable of sustaining high currents per geometric area.
CP may be a suitable current collector for metals or intermetallic
compounds because it is a conducting three-dimensional matrix that
can host them, not only on the surface but also inside it, and it can
buffer the volume changes of the hosted material. While CP might
be involved in the electrochemical processes of lithiation/
delithiation, the advantage of Sn/carbonaceous materials composites
over Sn has been demonstrated.11,12 Electrodes of Cu6Sn5 electrode-
posited on CP and on Cu and of mechanochemical Cu6Sn5 perme-
ated through CP from a suspension (Cu6Sn5 96 wt % and poly͑vi-
nylidene fluoride͒ 4 wt %͒ in acetone were subjected to repeated
lithiation/delithiation galvanostatic cycles with time-limited charges
to insert different amounts of lithium, and the results using the two
current collectors are compared and discussed.
We focus on the electrochemical stability of Cu6Sn5 . The
Cu6Sn5 shows structural compatibility with its lithiated phase
Ͻ
Ͻ
13), which is isostructural with Li2CuSn, so
LixCu6Sn5 (0
x
that the sublattice of Sn, which is the active component of the inter-
metallic compound, suffers moderate distortion in the lithiated
phase.20 In addition, the Cu inactive matrix buffers the Sn volume
changes accompanying the reactions with Li, whereas Sn alone
reaches a volume expansion of up to 300% when 4.4 Li per Sn atom
are inserted.21 However, Cu6Sn5 displays a volume expansion of
Experimental
Cu-Sn alloys were galvanostatically deposited on copper foils
͑25 ␮m thick͒ and CP ͑Spectracorp 2050, 10 mils͒ sheets, both
pretreated in acid solution, from a 0.1 M SnCl2-0.025 M CuSO4-0.5
* Electrochemical Society Active Member.
z E-mail: marina.mastragostino@unibo.it
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A290
Journal of The Electrochemical Society, 152 ͑2͒ A289-A294 ͑2005͒
that can host active materials. SEM images of the electrochemical
CuxSny compound on CP in Fig. 1c,d show the CP fibers homoge-
neously covered with CuxSny . The covering appears to be particles
and aggregates of different sizes ͑the smallest are Ͻ1 ␮m͒, the big-
gest having a leaf shape. Figure 1e,f displays the SEM images of
mechanochemical Cu6Sn5 on CP and mechanochemical Cu6Sn5
powder. Here the Cu6Sn5 deposited on CP by permeation is more
localized where the fibers criss-cross, but it is present in the whole
volume of CP ͑as per SEM images, not shown, of the back side of
the Cu6Sn5 /CP electrodes͒. Most of the carbon fibers remained
bare, and the aggregates of different sizes are made of nanometric
particles. Figure 2 displays the ‘‘particle’’ size distribution of the
mechanochemical Cu6Sn5 . The abscissa is the particle or, as in our
case, the agglomerate size, which can be viewed as an ‘‘equivalent
diameter,’’ i.e., the diameter of a sphere which exhibits the same
physical properties as the measured, irregular particle or agglomer-
ate. The ordinate is the oversize cumulative distribution indicating
the total quantity of all agglomerates with equivalent diameter
smaller than or equal to the abscissa value. The size of 90% of the
agglomerates is Ͻ20 ␮m.
M H2SO4 degassed solution at room temperature. Adhesive tape
was used to prevent deposition on the back of the electrode and to
limit the exposed area to 1 cm2. Ultrapure water ͑Milli-Q, 18.6 M⍀
cm, Simplicity Water System, Millipore Co.͒ was employed for so-
lutions and rinsing. Electrodes with different amounts of Cu-Sn in-
termetallic compound ͑2-6 mg cmϪ2͒ were prepared at constant cur-
rent ͑Ϫ12 mA cmϪ2͒. As the efficiency of the electrodeposition was
low, alloy mass loading was evaluated by weighing. The potential
during galvanostatic electrodeposition was ca. Ϫ0.5 V vs. saturated
calomel electrode. The electrodes were dried under vacuum at 60°C
for 16 h before use. Nanometric ␩ -Cu Sn ͑Sn 45.5, Cu 54.5 atom
Ј
6
5
%͒ prepared mechanochemically was also used to prepare electrodes
by permeation of a suspension of Cu6Sn5 ͑96 wt %͒, poly͑vi-
nylidene fluoride͒ ͑4 wt %͒, and acetone through CP sheets. These
electrodes, with a Cu6Sn5 mass loading of 3-14 mg cmϪ2, were
dried 1 h at 70°C under vacuum at room temperature before use.
All the electrodes were characterized by X-ray diffraction ͑XRD͒
analysis with a Philips PW1050/81 powder diffractometer ͑Cu K␣
radiation, 40 mA, 40 kV͒. Scanning electron microscopy ͑SEM͒
analyses were performed with a Philips 515 SEM microscope
equipped with an EDAX PV9900 energy dispersion detector for
X-ray energy-dispersive spectroscopy ͑EDS͒. For secondary elec-
tron imaging and for EDS, 20 and 25 keV electrons were used. The
particle/aggregate dimensions were measured with a Fritsch Analy-
sette 22 Compact laser particle sizer in the range 0.3-300 ␮m ͑62
channels͒ with 5 scans/measurement.
Figure 3 shows the XRD patterns of different electrodes: bare
CP,
electrochemical
CuxSny /CP,
and
mechanochemical
Cu6Sn5 /CP; the pattern of Sn electrodeposited on CP is also re-
ported for comparison. The figure also indicates the main reflections
of the different materials ͑stars for the Cu6Sn5 and circles for the Sn
main peaks͒. The electrochemical CuxSny , unlike the pattern of
mechanochemical Cu6Sn5 , shows the coexistence of two phases,
All the electrodes were electrochemically characterized by gal-
vanostatic charge/discharge cycles at T ϭ 30 Ϯ 1°C in a T-shaped
cell sealed in a dry box ͑Mbraun Labmaster 130, O2 and
␩ Cu Sn and Sn. While the reflections attributable to tin oxides or
Ј
6
5
Ͻ
H2O
1 ppm͒. The working electrodes were cut with a hollow
copper oxides are absent in both the ͑b͒ and ͑c͒ spectra, the reflec-
tions of CP are recognizable in all the spectra. From the EDS analy-
sis of several portions of an electrochemical CuxSny /CP, we esti-
mated a mean value for the Cu/Sn atomic ratio of 1.19 Ϯ 0.11, and
we considered it reasonable to take the electrodeposited material as
Cu6Sn5 .
brass punch. Due to the resharpening of the punch’s rim, the area
was 0.68 Ϯ 0.02 cm2. Given the inhomogeneous density of the CP
sheets, the masses of the CP current collectors, evaluated by weigh-
ing, are reported in all the figures related to electrochemical mea-
surements. The counter electrode was metallic Li in excess, and the
reference electrode for monitoring the electrode potentials was a Li
foil. The separator, a Whatman GF/D glass fiber disk, was imbibed
with the electrolyte solution ethylene carbonate:dimethyl carbonate
(EC:DMC2:1)-1 M LiPF6 ͑LP31 Merck, battery grade͒. The gal-
vanostatic cycles were carried out at 0.500 mA, i.e., 0.74
Ϯ 0.02 mA cmϪ2. The terms ‘‘charge’’ and ‘‘discharge’’ refer to the
alloy electrode being used as the anode in lithium-ion batteries, i.e.,
to lithiation and delithiation, respectively. The discharges were po-
tential limited ͑2.000 V vs. Li͒, and the charges were time limited
͑while maintaining a safe voltage cutoff of 0.005 V vs. Li͒ to insert
defined amounts of lithium per atom of tin. The amount of lithium to
be inserted ͑estimated as the product of current and time͒ was cal-
culated, in relation to the Li/Sn atomic ratio to be reached, on the
basis of the mass of the Cu6Sn5 loaded on the electrode. The specific
capacity of Cu6Sn5 was calculated by assuming that all the lithiation
charge involved the Cu6Sn5 . The cells with electrodes having high
alloy mass loading were disassembled and reassembled to change
the Li metal counter electrode ͑and in some cases even the separa-
tor͒ after a certain number of cycles, i.e., when the sum of their
charges was ca. 400 C cmϪ2, to prevent the Li metal deposition/
dissolution at the counter electrode from interfering with the cycling
stability of the Cu6Sn5 . The electrochemical deposition and charac-
terization were performed with a VMP multichannel potentiostat and
a PAR 273A potentiostat/galvanostat.
Cu6Sn5 was electrodeposited on both Cu and CP. As the effi-
ciency of the electrodeposition process was low, we considered the
increase of the electrode weight instead of the charge involved in the
electrodeposition to evaluate alloy mass loading. We first tested
electrodes with a Cu6Sn5 mass loading of ca. 4-6 mg cmϪ2. The
coulombic efficiency of the first galvanostatic cycle of Cu6Sn5 both
on Cu and CP ranged from 60 to 75%, being affected, just like the
carbonaceous materials, by irreversible capacity due to side reac-
tions at the electrode/electrolyte interface. Figure 4 reports the volt-
age profiles of repeated galvanostatic charge/discharge cycles of
Cu6Sn5 on copper, where the charge process was time limited to
insert lithium up to 1.88 Li/Sn atomic ratio. However, after the first
three cycles, in which the lithiation plateau at 0.4 V progressively
disappeared, the potential of the Cu6Sn5 /Cu electrode reached the
charge cutoff voltage ͑5 mV͒, and the amount of Li inserted ͑and
deinserted͒ became lower and lower than the 1.88 atom per Sn.
Figure 5 displays the discharge capacity from galvanostatic cycles of
two Cu6Sn5 /Cu electrodes of the same area with initial charges
further limited to 1.10 Li/Sn and 0.59 Li/Sn. In these cases, too, the
alloy electrode’s stability was low, as well as its coulombic effi-
ciency. The reason for this decay can be found in the formation of
cracks, as also reported in the literature.13 The Cu foil cannot sup-
port the stress of the Cu-Sn alloy during the lithiation/delithiation,
which results in pulverizing and delaminating from the Cu foil of
the material, with cracks up to 20 ␮m wide, as we observed by SEM
of the electrodes after 100 cycles. The low coulombic efficiency
may also lead to electrode material pulverization, which produces
new surface area able to react with the electrolyte, in addition to
interparticle contact loss.16
Results and Discussion
We started with the electrochemical synthesis of Cu6Sn5 on both
CP and Cu current collectors. Given that the copper electrodeposi-
tion potential is less negative than that of tin, a solution with a
concentration of tin higher than that of copper was selected for the
galvanostatic electrodeposition of the Cu-Sn intermetallic com-
pound. The SEM photographs at different magnification of CP in
Fig. 1a,b show an open and fibrous structure ͑fiber diam ca. 10 ␮m͒
Figure 5 also reports the capacity data from galvanostatic cycles
of a Cu6Sn5 /CP electrode, which indicate that the use of the CP
current collector is successful when compared to that of Cu. The
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑2͒ A289-A294 ͑2005͒
A291
Figure 1. SEM images of bare CP with ͑a͒ 131 times and ͑b͒ 2100 times magnification, of electrochemical Cu6Sn5 /CP with ͑c͒ 131 times and ͑d͒ 1200 times
magnification, and of ͑e͒ mechanochemical Cu6Sn5 /CP with 1050 times magnification, and ͑f͒ mechanochemical Cu6Sn5 powder with 2100 times mag-
nification.
0.4 V progressively decreases in length. After 40 lithiation/
delithiation cycles, the voltage profiles are smoothed, and none of
them reached the charge cutoff potential of 5 mV even when the
charge time was increased to 1 h 55 min ͑Fig. 6c͒. The voltage
profiles of bare CP at different cycle numbers and different time-
charge settings clearly indicate that CP can also be lithiated and
delithiated, the process being reversible with high coulombic effi-
ciency ͑near 100%͒ with the exception of the very first cycles. The
coulombic efficiency of the first cycle of different CP electrodes
ranged from 70-85%, indicating that CP even displays the irrevers-
ible capacity due to side reactions at the electrode/electrolyte inter-
Cu6Sn5 /CP electrodes display high reproducibility and high cycling
stability over 250 cycles, with coulombic efficiency near 100% both
when the charge is limited to 1.44 Li/Sn and 1.84 Li/Sn, and the
maximum specific capacity of the Cu6Sn5 /CP electrode in Fig. 5 is
253 mAh gϪ1
.
Figure 6 displays the voltage profiles of galvanostatic cycles of
the electrochemical Cu6Sn5 /CP electrode in Fig. 5 with time-
limited charges compared to the galvanostatic charge/discharge
cycles at the same current of bare CP current collector ͑0.68 cm2͒
with the time-limited charge set to 1 h 30 min ͑Fig. 6a,b͒ and 2 h
͑Fig. 6c͒. For the Cu6Sn5 /CP electrode, the plateau of lithiation at
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A292
Journal of The Electrochemical Society, 152 ͑2͒ A289-A294 ͑2005͒
Figure 4. Voltage profiles of different galvanostatic charge/discharge cycles
at 0.500 mA of electrochemical Cu6Sn5 ͑2.9 mg per 0.68 cm2͒ on Cu foil.
The initial lithium insertion was 1.88 Li/Sn mol ratio.
Figure 2. Agglomerate size distributions of mechanochemical Cu6Sn5 .
face. However, the lithiation process at the CP electrodes occurs at
lower potentials than that at Cu6Sn5 /CP, and with time-limited
charges set to 2 h the cutoff potential of 5 mV was reached after 40
cycles and the charge time of the 120th cycle became less than 1 h.
The specific capacity of CP was 140 mAh gϪ1 for the first 40 cycles,
but it decreased sharply and reached 65 mAh gϪ1 after 150 cycles.
All of this suggests that the lithiation process of Cu6Sn5 /CP elec-
trodes is mainly due to the Cu6Sn5 , and the role which may be
ascribed to CP is that of an excellent three-dimensional current col-
lector. The evolution of the voltage profiles of the Cu6Sn5 /CP elec-
trode over a hundred cycles is due to a deterioration of the crystal-
line structure of the alloy by gradual pulverization ͑also shown by a
slight decrease in the intensity of the characteristic peaks in XRD
patterns, not reported here͒. However, it does not result in capacity
loss because of the CP’s ability to prevent loss of electrical contact
with the active material. Even a low crystalline alloy hosted in the
three-dimensional CP current collector thus retains lithium insertion
capability.
amounts of inserted Li up to 3.65 Li/Sn are reported in Fig. 7. These
electrodes were stable over 260 cycles and never reached the charge
cutoff potential of 5 mV; the maximum specific capacity estimated
by the electrodes with light alloy mass loading is ca. 500 mAh gϪ1
.
The comparison of the voltage profiles of the light-loaded
Cu6Sn5 /CP electrodes and of bare CP seems to suggest an impor-
tant contribution of the lithiation of CP during the charge process of
the alloy, but only at the first tens of cycles.
To verify the possibility of increasing the capacity per geometric
area of the Cu6Sn5 /CP, electrodes with high alloy mass were pre-
pared using mechanochemical Cu6Sn5 , given that the electrochemi-
cal deposition provided at maximum a mass loading of 4 mg. Figure
8a displays the capacity per geometric area of mechanochemical
Cu6Sn5 /CP electrodes with mass loading from 5.1 to 6.8 mg of
Cu6Sn5 and with the galvanostatic charges limited to a 2.02-2.08
Li/Sn ratio. It is evident from this figure that Cu6Sn5 /CP can sustain
a high number of galvanostatic cycles at high current density, deliv-
ering a capacity of up to 2.85 mAh cmϪ2 with a specific capacity of
ca. 285 mAh gϪ1 of Cu6Sn5 for all the tested electrodes. Electrodes
with a higher mass loading ͑9.5 mg͒ were also tested. While we
observed capacity values of 3.90 mAh cmϪ2 ͑corresponding to ca.
To increase specific capacity, we carried out galvanostatic
charge/discharge cycles of Cu6Sn5 /CP, where the charges were
Ͼ
time-limited to Li/Sn
2; so as not to overly affect the measure-
ment time, we also tested electrodes with Cu6Sn5 mass loading Ͻ3
mg cmϪ2. The capacity values of both electrochemical and mecha-
nochemical Cu6Sn5 /CP electrodes with time-limited charges for
Figure 5. Discharge capacity from galvanostatic charge/discharge cycles at
0.500 mA per 0.68 cm2 of electrochemical Cu6Sn5 /CP and Cu6Sn5 /Cu at
different amounts of lithium inserted. The Li/Sn mol ratio, time of charge for
Cu6Sn5 /CP, and the mass of Cu6Sn5 and CP are indicated. For Cu6Sn5 /Cu
electrodes, the initial Li/Sn was 0.59 ͑empty square͒ and 1.10 ͑empty tri-
angles͒, respectively.
Figure 3. XRD patterns of ͑a͒ bare CP, ͑b͒ electrochemical CuxSny /CP, ͑c͒
mechanochemical Cu6Sn5 /CP, and ͑d͒ Sn/CP electrodes.
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑2͒ A289-A294 ͑2005͒
A293
Figure 7. ͑a͒ Discharge capacity and ͑b͒ specific capacity from galvanostatic
charge/discharge cycles at 0.500 mA per 0.68 cm2 of electrochemical ͑empty
squares͒ and mechanochemical ͑solid triangles and circles͒ Cu6Sn5 /CP elec-
trodes at different amounts of lithium inserted. The mass of Cu6Sn5 and CP,
the Li/Sn value, and the time of charge are indicated.
Cu6Sn5 /CP electrode ͑6.8 mg of alloy mass͒ compared with the
40th cycle of a bare CP electrode. As in Fig. 6, it is evident that the
process of lithiation/delithiation is that of Cu6Sn5 , as its voltage
profiles during charge remain at higher voltages than those of CP. In
addition, the charge time of these heavy electrodes is always higher
than that which the bare CP electrodes can sustain.
The excellent stability performance of the submicrometric
Cu6Sn5 on CP, both for electrodes with high alloy mass loading up
Ͼ
cycles, is due to the CP matrix, which ensured a three-dimensional
electric contact even if alloy agglomeration and some cracks are
visible after 100 cycles, especially on the mechanochemical
Cu6Sn5 .
to 10 mg cmϪ2 and with lithium insertion Li/Sn
2 for over 100
Figure 6. Voltage profiles of galvanostatic charge/discharge cycles at 0.500
mA per 0.68 cm2 of electrochemical Cu6Sn5 ͑3.8 mg͒/CP and of bare CP ͑7.0
mg͒ at different cycle numbers. The charge time was 1 h 30 min ͑a and b͒
and 1 h 55 min ͑c͒, corresponding to 1.44 and 1.84 Li/Sn mol ratio, respec-
tively. The charge of bare CP was time-limited to 1 h 30 min ͑a and b͒ and
to 2 h ͑c͒. After 40 ͑2 h͒ cycles, the charge of CP was limited by the cutoff
potential ͑5 mV͒.
Conclusions
The results of this study demonstrate that the Cu6Sn5 alloy,
whether electrochemically prepared directly into the CP or mecha-
nochemically prepared and hosted into it, can undergo hundreds of
280 mAh gϪ1 of Cu6Sn5), they decreased to half after 100 cycles
and remained almost constant up to 200 cycles. There could be two
causes for this decrease. The involvement of a high charge per cycle,
due to the high alloy mass loading, required several changes of the
Li counter electrode, which involved disassembling and reassem-
bling the cell three times during the first 70 cycles; hence, the most
superficial alloy may have been removed from the current collector.
The second cause could be a more important bulk effect in the
electrode with very high mass loading with the alloy that does not
draw advantage from the CP support. These results thus indicate a
maximum feasible Cu6Sn5 mass loading of 10 mg cmϪ2 for CP.
Figure 8b shows the voltage profiles of the 50th cycle of the
Ͼ
cycles of lithiation to Li/Sn
2 and delithiation, delivering con-
stant amounts of charge with coulombic efficiency near 100% after
the first few cycles. The cycling test at 0.74 mA cmϪ2 of
Cu6Sn5 /CP electrodes with alloy loading of 3 mg cmϪ2 and charge-
limited to 3.50 Li/Sn provided specific capacity values of 480 mAh
gϪ1. Those of the electrodes with loading of 6.6 mg cmϪ2 and
charge-limited to 2.43 Li/Sn had specific capacity values of 330
mAh gϪ1 ͑and capacity values per geometric area of 2.18 mAh
cmϪ2 can be estimated͒, and those with loading of 10 mg cmϪ2 and
charge limited to 2.08 Li/Sn had capacity values of 285 mAh gϪ1
͑and an estimated capacity value per geometric area of 2.85 mAh
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A294
Journal of The Electrochemical Society, 152 ͑2͒ A289-A294 ͑2005͒
cycles.15,23 The excellent performance of the Cu6Sn5 /CP electrodes
to repeated lithiation/delithiation processes is due to the CP current
collector’s characteristics; ͑i͒ it is a three-dimensional conductive
matrix with micrometric, interconnected carbon fibers which can be
homogeneously covered with the electrochemical alloy or host the
mechanochemical one without addition of conducting agent ͑we did
not observe differences in performance of electrodes having the
same mass loading prepared by the two methods͒; ͑ii͒ it is a light-
weight current collector; and ͑iii͒ it is more ‘‘pliant’’ than Cu foil so
that alloy cracking is less detrimental.
It may also be interesting for materials such as Sn, and work is in
progress on this matter.
Acknowledgments
This research was funded by COFIN2002 ͑Nanostructured alloys
as anodes for Li-ion batteries͒. The authors thank the group of Pro-
fessor G. Cocco ͑University of Sassari͒, partner of the project, for
the mechanochemical preparation of the nanometric Cu6Sn5 alloy
and Dr. R. Berti ͑Department of Physics, University of Bologna͒ for
useful suggestions in the SEM and EDS analyses.
The University of Bologna assisted in meeting the publication costs of
this article.
References
1. J.-M. Tarascon and M. Armand, Nature (London), 414, 359 ͑2001͒.
2. M. Noel and V. Suryanarayanan, J. Power Sources, 111, 193 ͑2002͒.
3. J. S. Gnanaraj, M. D. Levi, E. Levi, G. Salitra, D. Aurbach, J. E. Fisher, and A.
Claye, J. Electrochem. Soc., 148, A525 ͑2001͒.
4. J. Yao, G. X. Wang, J. Ahn, H. K. Liu, and S. X. Dou, J. Power Sources, 114, 292
͑2002͒.
5. S. Hossain, Y.-K. Kim, Y. Saleh, and R. Loutfy, J. Power Sources, 114, 264 ͑2002͒.
6. Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, and T. Miyasaka, Science, 276,
1395 ͑1997͒.
7. S. Grugeon, S. Laruelle, L. Dupont, and J.-M. Tarascon, Solid State Ionics, 5, 895
͑2003͒.
8. M. Winter and J. O. Besenhard, Electrochim. Acta, 45, 31 ͑1999͒.
9. J. Yang, M. Winter, and J. O. Besenhard, Solid State Ionics, 90, 281 ͑1996͒.
10. W. X. Chen, J. Y. Lee, and Z. Liu, Electrochem. Commun., 4, 260 ͑2002͒.
11. Y. Liu, J. Y. Xie, Y. Takeda, and J. Yang, J. Appl. Electrochem., 32, 687 ͑2002͒.
12. Y. Liu, J. Y. Xie, and J. Yang, J. Power Sources, 119-121, 572 ͑2003͒.
13. N. Tamura, R. Ohshita, M. Fujimoto, M. Kamino, and S. Fujitani, J. Electrochem.
Soc., 150, A679 ͑2003͒.
Figure 8. ͑a͒ Discharge capacity from galvanostatic charge/discharge cycles
at 0.500 mA per 0.68 cm2 of mechanochemical Cu6Sn5 /CP electrodes. The
lithium insertion was Ͼ2 Li/Sn with time of charge of 3 h 54 min ͑solid
circles͒, 3 h ͑empty triangles͒, and 2 h 54 min ͑empty squares͒. The charge
cutoff potential of 5 mV was reached after the 60th and the 50th cycles for
the electrodes indicated by the solid circles and the empty triangles, respec-
tively. The mass of Cu6Sn5 and CP are indicated. ͑b͒ Voltage profiles of
galvanostatic charge/discharge cycles of mechanochemical Cu6Sn5 /CP elec-
trode with the charge time-limited to 3 h 54 min and of CP ͑7.0 mg͒.
14. J. Wolfenstine, S. Campos, D. Foster, J. Read, and W. K. Behl, J. Power Sources,
109, 230 ͑2002͒.
15. S. D. Beattie and J. R. Dahn, J. Electrochem. Soc., 150, A894 ͑2003͒.
16. J. Yang, Y. Takeda, N. Imanishi, T. Ichikawa, and O. Yamamoto, J. Power Sources,
79, 220 ͑1999͒.
17. K. Zaghib, X. Song, A. Guerfi, R. Kostecki, and K. Kinoshita, in New Trends in
Intercalation Compounds for Energy Storage Conversion, K. Zaghib, C. M. Julien,
and J. Prakash, Editors, PV 2004-20, p. 29, The Electrochemical Society Proceed-
ings Series, Pennington, NJ ͑2004͒.
18. N. Li, D. T. Mitchell, K.-P. Lee, and C. R. Martin, J. Electrochem. Soc., 150, A979
͑2003͒.
19. S. Fujitani, H. Yagi, K. Sayama, T. Yoshida, and H. Tarui, Abstract 1152, The
Electrochemical Society Meeting Abstracts, Paris, France, April 27-May 2, 2003.
20. D. Larcher, L. Y. Beaulieu, D. D. MacNeil, and J. R. Dahn, J. Electrochem. Soc.,
147, 1658 ͑2000͒.
21. L. Y. Beaulieu, S. D. Beattie, T. D. Hatchard, and J. R. Dahn, J. Electrochem. Soc.,
150, A419 ͑2003͒.
22. L. M. L. Fransson, J. T. Vaughney, K. Edstrom, and M. M. Thackeray, J. Electro-
chem. Soc., 150, A86 ͑2003͒.
23. M. M. Thackeray, J. T. Vaughney, C. S. Johnson, A. J. Kropf, R. Benedek, L. M. L.
Fransson, and K. Edstrom, J. Power Sources, 113, 124 ͑2003͒.
24. K. D. Kepler, J. T. Vaughey, and M. M. Thackeray, Electrochem. Solid-State Lett.,
2, 307 ͑1999͒.
25. B. A. Johnson and R. E. White, J. Power Sources, 70, 48 ͑1998͒.
26. R. Moshtev and B. Johnson, J. Power Sources, 91, 86 ͑2000͒.
cmϪ2͒. Given, too, the high current density used in the galvanostatic
tests, 330 mAh gϪ1 and 2.18 mAh cmϪ2 are values of particular
relevance from a practical point of view. That the Cu6Sn5 alloy
hosted into CP does not require a carbon-conducting additive, as the
graphite anodes generally do, and taking into account the addition of
4 wt % of binder, these capacity values became 317 mAh gϪ1 of
deposited mass and 2.09 mAh cmϪ2. These values compare well
with the values 358 mAh gϪ1 and 2.3 mAh cmϪ2 of graphite-based
anodes of high-performance commercial batteries25,26 and have the
safety advantage that most of the lithiation process occurs at more
positive potentials than in graphite. This comparison is still favor-
able when the mass of the current collectors is considered, being that
CP displays a mass of ca. 10 mg cmϪ2, a value lower than those of
the Cu foils of commercial batteries.
Our results represent a significant improvement with respect to
those reported in the literature on Cu6Sn5 on metal current collec-
tors, particularly in terms of cycling stability over a high number of
Downloaded on 2015-04-02 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).