Single-bath electrodeposition of a combinatorial library of binary Cu1-xSnx alloys

Article abstract of DOI:10.1149/1.1576769

Electrodeposition from a single bath in a single run was used to create a combinatorial library of binary Cu-Sn alloys. Electron microprobe and X-ray diffraction studies were used to determine the average stoichiometry and the phases present as a function of position on the substrate foil. The deposited film follows the Cu-Sn phase diagram, except that Cu₄₁Sn₁₁ was observed instead of Cu₃Sn. Electrodes with various average stoichiometry were punched from the film and tested as negative electrodes for Li-ion cells. Electrodes taken from the high Sn content area of the electrodeposited film display high capacity (～600 mAh/g) but poor capacity retention with cycling. Electrodes taken from the high Cu content area of the film display reduced capacity (～300 mAh/g) with dramatically improved capacity retention.

Full text of DOI:10.1149/1.1576769

Journal of The Electrochemical Society, 150 ͑7͒ C457-C460 ͑2003͒
C457
0
013-4651/2003/150͑7͒/C457/4/$7.00 © The Electrochemical Society, Inc.
Single-Bath Electrodeposition of a Combinatorial Library
of Binary Cu1ÀxSn Alloys
x
S. D. Beattie and J. R. Dahn
Department of Physics, Dalhousie University, Halifax, Nova Scotia, Canada B3H 3J5
Electrodeposition from a single bath in a single run was used to create a combinatorial library of binary Cu-Sn alloys. Electron
microprobe and X-ray diffraction studies were used to determine the average stoichiometry and the phases present as a function
of position on the substrate foil. The deposited film follows the Cu-Sn phase diagram, except that Cu Sn was observed instead
41
11
of Cu Sn. Electrodes with various average stoichiometry were punched from the film and tested as negative electrodes for Li-ion
3
cells. Electrodes taken from the high Sn content area of the electrodeposited film display high capacity ͑ϳ600 mAh/g͒ but poor
capacity retention with cycling. Electrodes taken from the high Cu content area of the film display reduced capacity ͑ϳ300
mAh/g͒ with dramatically improved capacity retention.
©
2003 The Electrochemical Society. ͓DOI: 10.1149/1.1576769͔ All rights reserved.
Manuscript submitted August 26, 2002; revised manuscript received February 5, 2003. Available electronically May 12, 2003.
Cu-Sn alloys have been proposed as possible anode materials for
power.16 A solution with good throwing power deposits a constant
thickness film regardless of macroscopic cathode irregularities. A
solution with poor throwing power is needed to obtain varying
thickness along the cathode in the Hull cell. The bath was operated
at room temperature without agitation.
Deposition was performed in a Hull cell ͑not ‘‘normal’’ speci-
fications͒ made from polyvinyl chloride ͑PVC͒. The cell is 7.5 cm
wide. The cathode is inclined at a 51.5° angle with respect to the
anode, which is 7.5 cm wide. The closest approach of anode to
cathode is 1.0 cm, and the furthest separating distance is 9.0 cm. The
cathode is 11.0 cm in length and the cell is 6.9 cm tall. The cell has
a volume of approximately 250 mL. A dimensionally stable titanium
anode was used.
All films described here were deposited on Ni foil. To ensure
minimal contamination of the bath and a clean deposition surface,
the Ni foil was pretreated. The foil was wiped with acetone, ethanol,
and then methanol. 3M plater’s tape was used to secure the foil to
the PVC backing around the immersed edges to prevent deposition
on the back of the foil. Note that the platers tape slightly decreases
the available area for deposition. The usable area is decreased from
1
Li-ion cells. Kepler et al. showed that Cu Sn can react reversibly
6
5
2
with lithium for a few tens of cycles. Larcher et al. studied the
Li-Cu Sn reaction using in situ X-ray diffraction ͑XRD͒. Tamura
6
5
3
et al. showed that Cu-Sn alloys could be prepared by electrodepos-
1
7
iting pure tin on copper foil, followed by an annealing step. Beattie
4
et al. showed that Cu Sn5 could be directly prepared by elec-
6
2
ϩ
trodeposition from a single bath containing both dissolved Cu and
2
ϩ
Sn ions. These films were shown to have approximately the same
electrochemical behavior in lithium cells as the powdered samples
described by Larcher et al. Electrodeposited films have the advan-
tage that they could be directly used as electrodes in lithium-ion
batteries without further processing.
5
There are a number of binary intermetallic Cu-Sn phases, and
the electrochemical properties of each should be considered. How-
ever, some of these phases, like Cu Sn and Cu Sn contain little
41
11
3
tin so one would expect them to react with little lithium, and hence
6
be relatively ‘‘inactive’’ phases. On the other hand, Mao et al. have
shown that it can be advantageous to mix relatively inactive phases
like SnFe C with active phases like Sn Fe in efforts to obtain ma-
3
2
11.0 ϫ 7.5 cm to approximately 10.5 ϫ 7.0 cm. The PVC vessel
terials with good capacity retention and acceptable specific capacity.
Therefore, we decided to prepare mixed intermetallic phases in the
Cu-Sn system.
and anode were also wiped down prior to deposition.
Pulsed deposition was performed using a Keithley 236 source
measure unit. A Visual Basic program was written to apply pulsed
waveforms using the Keithley. All deposits were performed gal-
vanostatically.
Scanning electron microscopy ͑SEM͒ and energy dispersive
spectroscopy ͑EDS͒ studies were performed using a JEOL JXA-
8200 superprobe with a Noran energy dispersive spectrometer.
XRD measurements were performed using an INEL CPS120
curved position sensitive detector coupled to an X-ray generator
equipped with a Cu target X-ray tube. There is a monochromator in
the incident beam path that limits the wavelengths striking the
We decided to prepare Cu-Sn alloys with varying composition by
4
electroplating. Based on our previous work we believed that a wide
range of Cu-Sn alloy compositions could be electrodeposited from a
single bath using pulsed deposition in a Hull cell. Alloys prepared
using a Hull cell and its variations ͑i.e., rotating cylindrical Hull,
7
-10
RCH cell͒ have been discussed in the literature.
Our method is
similar to ones used in.1
1,12
However, our method is unique because
a low frequency on/off pulsed waveform and the geometry of the
Hull cell are used to exploit the exchange of deposited Sn by aque-
ous Cu to achieve a composition and structural spread as a function
of position on the substrate. To our knowledge such a novel prepa-
sample to Cu K . The incident angle of the beam with respect to the
␣
sample is about 6°. The detector measures the entire diffraction pat-
tern between scattering angles of 6° and 120° at once. The film
sample is placed on an x-y translating stage that allows measure-
ment and move operations to be sequentially programmed. Scans
were taken incrementally along the length of the film. To encourage
homogeneity of coexisting phases the film was annealed under argon
at 200°C for 2 h prior to XRD analysis. The film was not annealed
before EDS analysis.
1
3
ration scheme is1
a
new method in combinatorial
materials synthesis.
or
4,15
composition-spread
Figure 1 shows electron microprobe results for a Cu-Sn
composition-spread film prepared via electrodeposition in a single
run. Starting from the bottom, Sn-rich end of the film, the position in
centimeters is plotted vs. the Cu:Sn atomic ratio. The Cu:Sn atomic
ratio increases along the length of the film. This shows that a com-
binatorial library of Cu-Sn alloys has been prepared via electrodepo-
sition. Below, we will describe the preparation method and the be-
havior of a number of compositions as electrodes for Li-ion cells.
Cu-Sn film samples were tested in Li/Cu-Sn cells for their spe-
cific capacity and capacity retention. Standard 2325 ͑23 mm diam,
2.5 mm thick͒ coin cell hardware was used. The cells use a poly-
propylene microporous separator, 1 M LiPF dissolved in ethylene
6
Experimental
carbonate:diethyl carbonate ͑EC:DEC, 33:67 vol, Mitsubishi Chemi-
cal͒ electrolyte and lithium metal as the negative electrode. All cells
were assembled in an argon-filled glove box and tested using con-
stant charge and discharge currents of 30 mA/g.
Cu-Sn alloys were deposited from a pyrophosphate solution: 36
g/L Sn P O , 135 g/L2 K P O , 1 g/L Cu P O •3H O. Without
2
2
7
2
2
7
2
2
7
2
the addition of additives a pyrophosphate bath has poor throwing
Downloaded on 2014-07-07 to IP 129.130.252.222 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C458
Journal of The Electrochemical Society, 150 ͑7͒ C457-C460 ͑2003͒
Figure 1. Cu:Sn atomic ratio and atomic % Sn as determined by electron
microprobe analysis versus position on the electrodeposited film. The various
symbols correspond to six separate scan lines measured along the film. An
image of the deposited film is shown on the right.
Figure 3. An SEM image of the deposited film. The size of the 40 ␮m
electron probe beam is indicated by the squares. Different numbers of large
crystallites are sampled from place to place along the film, causing scatter in
the electron microprobe data of Fig. 1.
Electroplating Strategy
In order to prepare the composition-spread library of Cu-Sn al-
loys, several requirements must be satisfied. First, copper deposits at
more positive potentials than tin, so the tin concentration in solution
should be much larger than that of copper. Second, the bath should
have a poor throwing power ͑like the pyrophosphate bath we
picked͒ so that the current density varies with the anode and/or
cathode distance in the Hull cell. This ensures that the thickness of
the deposit varies with position on the substrate foil. Finally, pulsed
deposition is used to create the composition spread. A series of
on/off pulses are used, because primarily Sn is deposited during the
on pulse, and copper from the plating solution ion exchanges with
the plated tin during the off pulse according to the reaction
40 ϫ 40 ␮m area. In that area many crystallites were sampled
simultaneously and each crystallite did not necessarily have the
same composition. Figure 3 shows an electron micrograph of the
surface of the film with squares indicating the approximate size of
the probe beam. For the three indicated positions of the squares
there were different numbers of large crystallites present. Large
crystallites are rich in Sn, so if the beam is analyzing an area with a
large number of crystallites, a large Sn concentration will be de-
tected. Instead, if the beam is sampling an area with very few crys-
tallites the Sn concentration will be low. Between adjacent points
the number of crystallites being analyzed can vary. This variation
can result in a significant difference in the Cu:Sn atomic ratio be-
tween adjacent points, causing scatter in the data, as seen in Fig. 1.
Nevertheless, it is clear that the average stoichiometry inferred from
Fig. 1 varies in a roughly linear manner.
ϩ2
ϩ2
Cu (aq) ϩ Sn͑s) → Cu͑s) ϩ Sn (aq). Then, a deposit like that
shown schematically in Fig. 2 is created.
Figure 2a shows the layers of tin rich material deposited during
the on pulses, whose thickness varies along the length of the sub-
strate electrode, and the layers of ion-exchanged copper deposited
during the off pulses, whose thickness is constant along the length of
the substrate electrode. In our depositions, the on pulse/off pulse
sequence is repeated hundreds of times to deposit a film several
micrometers thick. During the on pulse, the layer thickness depos-
ited is estimated to range between 4.7 and 1.3 nm along the length
of the substrate. Examinations of the film using small angle X-ray
scattering showed that a composition modulated superlattice struc-
ture was not formed. Presumably, this is because the copper-rich and
tin-rich layers are eliminated by interdiffusion. Therefore, Fig. 2b
shows a schematic of the deposited film after the Cu-Sn interdiffu-
sion occurs.
Results and Discussion
Figure 1 shows that a film where the Cu:Sn atomic ratio varies
along the length of the film has been deposited. Using the electron
microprobe results the atomic percent Sn can be determined along
the length of the film ͑Fig. 1, right ordinate͒. Then, the position of
the film in phase space can then be mapped onto the Cu-Sn phase
diagram as has been done in Fig. 4. The arrow in Fig. 4 shows the
range of compositions sampled by the deposited film. In theory,
using our deposition procedure, any stoichiometric range of phase
space could be easily sampled given the proper electrodeposition
parameters. Due to lack of control over temperature a structural
range would be more difficult to sample.
Figure 1 shows that the procedure described above deposits a
film where the average stoichiometry varies along the length of
the substrate. Each EDS data point in Fig. 1 represents a
Electrodeposition was performed at room temperature, however
the tabulated phase diagram only extends as low as 100°C. We as-
sume that the phase diagram at 21°C is identical to that at 100°C.
All samples in this paper were either prepared and studied at room
temperature or were annealed to 200°C under argon. Between 20
and 200°C, the arrow in Fig. 4 passes through five distinct stoichio-
metric regions: Sn ϩ Cu Sn , Cu Sn , Cu Sn ϩ Cu Sn, Cu Sn,
6
5
6
5
6
5
3
3
and Cu Sn, ϩCu.
3
Three positions in the phase diagram ͑Fig. 4͒ were chosen for
focused study. A high Cu content ͑4a͒, a high Sn content ͑4c͒, and an
intermediate area ͑4b͒. According to the phase diagram, sample 4a
should correspond to coexisting phases of Cu Sn, ϩCu, sample 4b
3
Figure 2. ͑a͒ A schematic of film structure expected using on/off pulsed
deposition in a Hull cell with a Sn pyrophosphate solution containing a small
amount of Cu pyrophosphate in the absence of solid-state interdiffusion. ͑b͒
A schematic of the actual film stoichiometry due to interdiffusion.
should correspond to coexisting phases of Cu Sn ϩ Cu Sn and
6
5
3
sample 4c should correspond to coexisting phases of Sn
ϩ Cu Sn . The film was annealed under argon at 200°C for 2 h,
6
5
Downloaded on 2014-07-07 to IP 129.130.252.222 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, 150 ͑7͒ C457-C460 ͑2003͒
C459
Figure 4. The Cu:Sn phase diagram. The arrow indicates the region of phase
space sampled by the film ͑Fig. 1͒ deposited in the Hull cell using on/off
pulsed deposition. Three average compositions labeled 4a, 4b, and 4c are
considered for focused analysis.
Figure 6. 20 XRD patterns measured at equal spacings along the length of
the film. The Cu content increases from bottom to top. Sn rich phases are
observed at the bottom of the film, while Cu rich phases are observed at the
top. Bragg peaks from known phases are indicated.
then XRD patterns were taken along the entire length. XRD patterns
taken from the positions labeled 4a-c in Fig. 4 are presented in Fig.
5, correspondingly labeled 4a-c. XRD patterns taken from pure Sn
and known Cu-Sn alloys are superimposed onto those XRD patterns.
In Fig. 4a, coexisting Cu Sn, ϩCu phases are expected from the
3
phase diagram but they are not observed. Instead, the Cu Sn
41
11
phase, which does not extend down to room temperature in Fig. 4, is
present. Apparently Cu Sn is favored by the pulsed deposition
4
1
11
process described above, rather than Cu or Cu Sn. The peak near
3
42° in Fig. 4a is thought to arise from a copper phase containing a
small amount of dissolved tin. In Fig. 4b, the Cu Sn phase is
41
11
present once again, where Cu Sn was expected. The fact that films
3
fabricated by electrodeposition do not exactly follow the equilibrium
conditions present in a phase diagram at a specified temperature is
not surprising since electrodeposition does not occur at equilibrium
conditions. The Cu Sn phase is in the right stoichiometric range
4
1
11
in the Cu-Sn phase diagram to appear in samples 4a and 4b, but is
not expected at lower temperatures. The Cu Sn phase is not ex-
3
pected at high Cu content, Cu Sn is, therefore the XRD pattern at
4
1
11
position 4b in Fig. 4 should show coexisting Cu Sn and Cu Sn5
41
11
6
phases, as it does in Fig. 5b. Furthermore, at position 4c in Fig. 4,
coexisting Cu Sn and Sn is expected, as verified by the diffraction
6
5
pattern in Fig. 5c. Electron microprobe results and XRD results
agree with the Cu-Sn phase diagram as long as the Cu Sn phase is
41
11
expected at high Cu content.
Figure 6 shows XRD patterns along the length of the film start-
ing from the bottom of the film ͑that is, the part of the film closest to
the anode during deposition͒ up to the top of the film ͑furthest from
the anode during deposition and therefore Cu rich͒. The patterns that
were the subject of the focused study in Fig. 5 ͑labeled 4a, b and c͒
are indicated in Fig. 6. The XRD patterns at the bottom of the film
are rich in Sn. As the patterns progress up the film toward the Cu-
rich end of the film, the XRD pattern from Sn steadily decreases and
disappears approximately a third of the way up. Approximately two-
thirds up the film the Cu Sn phase also disappears. Correspond-
6
5
ingly, XRD patterns of Cu-rich alloys begin to emerge. There is a
smooth transition between Sn rich and Cu rich alloys.
Li/Cu-Sn coin cells were then constructed from circular elec-
trodes punched from various places on the film. Figure 7 shows
Figure 5. XRD patterns of portions of the film corresponding to the posi-
tions ͑a͒ 4a, ͑b͒ 4b, and ͑c͒ 4c. Known Sn and Cu-Sn alloy XRD patterns are
superimposed on the data for comparison.
Downloaded on 2014-07-07 to IP 129.130.252.222 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C460
Journal of The Electrochemical Society, 150 ͑7͒ C457-C460 ͑2003͒
metallic phases like Cu Sn exhibit lower capacities than Sn-rich
6
5
1,2,4
materials, but have dramatically improved capacity retention.
Figure 7c-h progressively illustrate the effect of increasing Cu con-
tent. Figure 7 shows that high Sn content films have higher capacity
and poor capacity retention and that high Cu content films have
lower capacity and dramatically improved capacity retention.
Conclusions
Using electrodeposition, a combinatorial library of binary
Cu1ϪxSn alloys has been successfully deposited from a single bath
x
in a single run. The Cu:Sn atomic ratio varies almost linearly along
the length of the film. Electron microprobe and XRD data agree with
the Cu-Sn phase diagram except that the high temperature Cu Sn
41
11
phase was observed in place of Cu Sn. Electrochemical studies of
3
Li/Cu-Sn cells show that Sn-rich regions of the film exhibit high
capacity ͑ϳ600 mAh/g͒ but poor capacity retention. Cu-rich com-
pounds exhibit lower capacity ͑ϳ300 mAh/g͒ with dramatically im-
proved capacity retention.
Acknowledgments
The authors acknowledge 3M, 3M Canada and NSERC for fi-
nancial support of this work.
Dalhousie University assisted in meeting the publication costs of this
article.
References
1
2
3
4
. K. D. Kepler, J. T. Vaughey, and M. M. Thackeray, J. Power Sources, 81-82, 383
1999͒.
. D. Larcher, L. Y. Beaulieu, D. D. MacNeil, and J. R. Dahn, J. Electrochem. Soc.,
47, 1658 ͑2000͒.
. N. Tamura, R. Ohshita, M. Fujimoto, S. Fujitani, M. Kamino, and I. Yonezu, J.
Power Sources, 107, 48 ͑2002͒.
͑
1
. S. D. Beattie and J. R. Dahn, J. Electrochem. Soc., Submitted.
Figure 7. Voltage vs. capacity and capacity vs. cycle number for five Li/
Cu-Sn cells with electrodes taken along the length of the film. The Cu con-
tent increases from bottom to top. The specific capacity decreases with in-
creasing Cu content with a corresponding increase in capacity retention.
Solid dots correspond to discharge capacity while empty dots correspond to
recharge capacity.
5. M. Hansen, Constitution of Binary Alloys, McGraw-Hill Book Company, New
York ͑1958͒.
6. O. Mao, R. L. Turner, I. A. Courtney, B. D. Fredericksen, M. I. Buckett, L. J.
Krausse, and J. R. Dahn, Electrochem. Solid-State Lett., 2, 3 ͑1999͒.
7
8
9
. H. Gao, A. Scheeline, and A. J. Pearlstein, J. Electrochem. Soc., 149, B248 ͑2002͒.
. C. Madore and D. Landolt, Plat. Surf. Finish., 80, 73 ͑1993͒.
. N. Zech, E. J. Podaha, and D. Landolt, J. Appl. Electrochem., 28, 1251 ͑1998͒.
1
0. Y. Liao, D. R. Gabe, and G. D. Wilcox, Plat. Surf. Finish., 85, 60 ͑1998͒.
11. D. Landolt, Plat. Surf. Finish., 88, 70 ͑2001͒.
1
1
1
2. S. Roy, M. Matlosz, and D. Landolt, J. Electrochem. Soc., 141, 1509 ͑1994͒.
3. X. D. Xiang, Annu. Rev. Mater. Sci., 29, 149 ͑1999͒.
4. J. J. Hanak, J. Mater. Sci., 5, 964 ͑1970͒.
specific capacity vs. cycle number and voltage vs. capacity for five
electrodes taken along the length of the film. Figure 7a and b cor-
respond to an electrode taken from the Sn-rich end of the film.
Figure 7i and j correspond to an electrode taken from the Cu-rich
end of the film. The trends in Fig. 7 are obvious and expected. It is
well known that Sn rich compounds exhibit large capacity but poor
capacity retention.18 It has also been shown that pure Cu-Sn inter-
15. R. B. van Dover, L. F. Schneemeyer, and R. M. Fleming, Nature (London), 392,
62 ͑1998͒.
1
1
6. M. Paunovic and M. Schlesinger, Fundamentals of Electrochemical Deposition,
John Wiley and Sons, New York ͑1998͒.
1
7. R. O. Hull, Trans. Am. Electroplat. Soc., 27, 52 ͑1939͒.
18. J. Yang, M. Winter, and J. O. Besenhard, Solid State Ionics, 90, 281 ͑1996͒.
Downloaded on 2014-07-07 to IP 129.130.252.222 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).