Journal of New Materials for Electrochemical Systems 10, 95-99 (2007)
© J. New Mat. Electrochem. Systems
Mesoporous Gold as Anode Material for Lithium-Ion Cells
L. Yuan1, H. K. Liu1,2, A. Maaroof 3, K. Konstantinov1, J. Liu3 and *M. Cortie3
1
Institute for Superconducting & Electronic Materials, University of Wollongong, NSW 2522, Australia
2
ARC Centre for Nanostructured Electromaterials, University of Wollongong, NSW 2522, Australia
3
Institute for Nanoscale Technology, University of Technology Sydney, NSW 2007, Australia
Received: January 5, 2006, Accepted: November 22, 2006
Abstract: Mesoporous gold sponges were prepared by chemical removal of Al from thin films of an AuAl2 precursor that had been deposited on Cu sheet. The morphology of the Au was characterised by interconnected pores and channels of between 5 and 20 nm in diameter. Here we report an assessment of these films as the electrode in Li rechargeable cells. It was found that the Li alloying processes occurred in the voltage range of 0 to 0.25 V, while de-alloying occurred in two stages at about 0.15 and 0.45V. This is significantly lower
than in the Li–Sn or Li-Sn-Cu systems (0.2-1.0V) but comparable to that reported for thin, solid gold anodes. Overall, a multilayer
mesoporous Au film showed superior characteristics compared to an ordinary Au film, with a higher specific charge passed. Capacity of
all electrodes tested was of the order of 500 mA.h.g-1 during the initial discharge cycle, but was subject to a steep fade during subsequent
cycles. The capacity of the multilayer, mesoporous gold settled at about 80 mA.h.g-1 after 30 cycles, while that of the ordinary Au film fell
to about 10 mA.h.g-1.
Keywords: lithium-ion, negative electrode, anode, gold , mesoporous
but conductive matrix material for the anode, into which the Li
passes after reduction to form intermetallic compounds. The second criterion is facilitated when the phase diagram of the anode
alloy has a broad two phase region on the Li-rich side [8]. By consulting a standard compilation of phase diagrams [10], it can be
shown that Li-Au is suitable, with the field Li+Li15Au4 extending
from 12 to 100% wt. Li. In principle, voltage will be constant during alloying and de-alloying over this range since the activity of
the Li would be maintained at that pertaining to Li15Au4 as long as
any of this phase is in contact with electrolyte. The two phase
range in Li-Au is broader than in Li+Li22Sn5 (20.5 to 100 wt.% Li)
or Li-Ag (40 to 100 wt% Li). The third criterion is related to the
maximum theoretical capacity of the anode. Pure Li would yield
3854 mA.h.g-1 of anode, whereas graphite, tin and gold are expected to yield 372, 746 and 451 mA.h.g-1 respectively (based on
lithiated intermetallic compounds with the stoichiometries Li2C11,
Li22Sn5,and Li15Au4). However, since the intermetallic compounds
are significantly denser than the Li-C phases, their volumetric
energy density is competitive, with Li-Sn for example reported at
7200 mA.h.cm-3, which is significantly greater than the 840
mA.h.cm-3 of Li-C [7]. The fourth criterion, the cell voltage, is
determined by thermodynamics, in particular by the activity of the
Li in the anode. This should be as close to unity as possible to give
1. INTRODUCTION
Mesoporous metallic surfaces may be prepared by leaching an
active element from suitable precursor intermetallic compounds,
but are in general friable and readily oxidized. Gold is an exception, and stable, clean mesoporous surfaces and powders of it may
be prepared [1-3]. Porous metal coatings have found actual or
potential applications as electrodes in electrochemical devices [4]
and in sensors of various kinds [5,6], with the main attraction being that their increased surface area facilitates both faradaic and
capacitive processes. Recently, porous copper-tin has been investigated for application as the anode in Li batteries [7]. The properties desired of such anodes are that they should be stable after
repeated cycling, that they should have a wide range of stable cell
voltage and the lowest possible alloying/de-alloying potentials
relative to the Li/Li+ couple. This last point is important because
any deviation from the potential of the Li/Li+ couple will be at the
expense of the overall cell voltage, and the net effect will be a
reduction in the energy density of the battery [8,9].
The first criterion has been largely addressed by use of an inert
*To whom correspondence should be addressed: [email protected], ph. +612-9514-2208, fax +61-2-9514-8349, PO Box 123, University of Technology Sydney,
Broadway, NSW 2007, Australia
95
96
L. Yuan et al. / J. New Mat. Electrochem. Systems
Table 1. Details of the five coating types discussed
Sample
Top layer
Description
A
single layer of mesoporous Au on Cu sheet
B
multiple layers of mesoporous Au on Cu sheet
Middle layer
Bottom layer
200 nm Au + 1600 nm Al
-
-
50 nm Au + 320 nm Al
80 nm Au + 640 nm Al
50 nm Au+160 nm Al
C
thin film of solid Au on Cu sheet
100 nm Au
-
-
D
mesoporous Au on silicon wafer
100 nm Au + 795 nm Al
-
-
E
single layer of mesoporous Au on Cu sheet
20 nm Au + 160 nm Al
-
-
(a)
(b)
(c)
(d)
Figure 1. SEM images of mesoporous Au (a) Sample A, single layer containing the equivalent of 200 nm Au, (b) Sample B single layer
containing the equivalent of 100 nm of Au (c) cross-section of Sample B showing that coating is actually 165 nm thickness and (d) Sample E, surface of a coating containing equivalent of 30 nm Au after first discharge.
as big a cell voltage as possible. It is known that there are potential applications for high voltage rechargeable batteries (4.5V),
while the voltage for commercial lithium battery is 3.6V. Li-Sn
and Li(Cu,Sn) (with charging voltages that vary from 0.8 to 0V vs
Li/Li+ [7,8]) are inferior in this regard to Li-Au (0.4 to 0V), which
in turn is theoretically inferior to Li-Ag or Li-Zn (0.25 to
0V)[8,9]. We do not see the high cost of bulk gold as necessarily
being an insurmountable obstacle in this application since, if viable, only a small quantity of gold would be required per cell.
G. Taillades et al. reported that thin, solid, Au anodes showed a
capacity of 3400 mAh.cm-3 and a very negative and narrow voltage window [9]. However, there have been no reports to the best
of our knowledge on the usefulness of mesoporous Au as an electrode material for lithium rechargeable cells. The high surface area
(including occluded pores) of a mesoporous gold mass, taken in
combination with the high conductivity yet chemical stability of
Au, the low voltage for alloying/de-alloying in Li-Au, and the
form of the Li-Au binary phase diagram, suggested to the authors
that mesoporous Au anodes might have applications in respect of
specialized lithium rechargeable cells.
2. EXPERIMENTAL
Samples of mesoporous Au were prepared by the chemical re-
Mesoporous Gold as Anode Material for Lithium-Ion Cells / J. New Mat. Electrochem. Systems
moval of Al from thin films of the intermetallic compound AuAl2.
This intermetallic precursor material is commonly known as
‘purple gold’ [11] or ‘purple glory’ [12] on account of its attractive
purple color. The thin films of AuxAly were produced by simultaneously sputtering Au and Al from two elemental targets of 50 mm
diameter using a high vacuum DC magnetron sputtering apparatus.
The rates of sputtering were controlled by the varying the power
delivered, with the system having been pre-calibrated using a
quartz microbalance. Copper sheet of 1 mm thickness was used for
the substrate and was heated to 400°C during co-deposition using
an in situ heater. The sheets had been previously cleaned by sand
blasting following by etching in a citric acid solution. After deposition, the Al was removed by treatment with 0.5 NaOH. The conditions used to produce the five types of coating mentioned in this
paper are listed in Table 1. Au and Al (where applicable) were
simultaneously co-deposited. The actual thickness of the AuxAly
coating prior to removal of the Al would have been the sum of the
two thicknesses shown. However, the coatings actually densified
considerably after removal of the Al and the final thickness of the
Au coating was smaller.
Samples A to C were used as electrodes, and were assembled
into a Teflon test cell in an argon-filled glove-box (Mbraun,
Unilab, USA). The counter electrode was Li metal and the electrolyte was 1 M LiPF6 dissolved in a 50:50 (v/v) mixture of ethylene
carbonate (EC) and dimethyl carbonate (DMC) provided by Merck
KgaA, Germany. The cells were charge-discharged at room temperature at the constant current density of 30 μA.cm-2 over an area
of 1.4 cm2 and within the voltage range of 0.01-1.00 V vs Li/Li+.
Note that the potential of the couple is controlled by the lithiation/delithiation reactions in the test electrode. When the test sample contains no Li, ie. is completely de-alloyed, then the cell voltage relative to the Li metal counter electrode will be of the order of
a volt. However, as the sample is loaded with Li, ie. alloyed, the
activity of that species in it rises towards unity, and the cell voltage
vs Li/Li + decreases towards zero[7,8,13-16]. Cyclic
voltammograms (CV) measurements were carried out using an
EG&G potentiostat (Model M362) at a scanning rate of 0.2 mV.s1. Cycle life curves of mesoporous Au electrodes with different
thickness were measured using a current density of 15 mA cm-2 at
voltage window of 1.0–0.01 V vs Li.
The morphology of the coatings was examined using a LEO
scanning electron microscope (SEM) using in-lens imaging. A
satisfactory cross-sectional image of the coating on copper could
not be obtained due to the ductility of that element, but was found
to be possible for coatings deposited on silicon. This is the reason
why Sample D was made. While the material deposited on silicon
was not subjected to electrochemical testing, our experience has
shown that its morphology is nevertheless representative of that of
the coatings on the copper.
3. RESULTS AND DISCUSSION
A SEM image of the surface of the mesoporous Au coating of
Sample A is shown in Fig. 1 (a). In Fig 1(b) we show a high magnification image of the surface of Sample D while a high magnification cross-sectional view of that coating is shown in Fig. 1 (c). In
all cases it is clear that the coatings exhibit porous microstructures
with pore diameters of about 20 to 50 nm. However, while the
composite coating in Sample D was of the order of 900 nm thick
97
(a)
(b)
(c)
Figure 2. The second alloying/de-alloying cycle of Au electrodes,
(a) Sample A, single mesoporous layer, (b) Sample B, multilayer
of mesoporous Au, (c) Sample C, single, solid layer of Au
before chemical removal of the Al, the removal of that element
caused the thickness of the remaining material to shrink down to
165 nm. Since the equivalent of 100 nm Au had been co-deposited
with the Al, it follows that at a thickness of 165 nm the gold must
contain about 65% porosity and have an average density of about
12 g.cm-3. Fig. 1 (d) shows the cracks that develop on the surface
98
L. Yuan et al. / J. New Mat. Electrochem. Systems
(a)
(b) 0.00010
0. 0004
0.00005
0. 0002
0.00000
Current(A )
Curre nt(A)
0. 0000
-0. 0002
-0. 0004
-0.00005
-0.00010
-0.00015
-0. 0006
-0.00020
-0.00025
-0. 0008
0.0
0.5
1.0
1.5
2.0
2.5
3. 0
0.0
3.5
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Voltage(V)
Voltage(V)
Figure 3. The cyclic voltammograms curves of mesoporous Au electrode: (a) Sample A, (b) Sample B
Table 2. Alloying and de-alloying capacities in the second cycle (shown in Fig. 2) for each of the three samples.
Charge passed,
C(coulomb).cm-2
Charge passed, mA.h.cm-2
Charge passed, mA.h.g-1 Charge passed, mA.h.cm-3
Sample A alloying
1.204
0.348
1001
~18000
Sample A, de-alloying
0.602
0.169
501
~9000
Sample B, alloying
0.592
0.163
425
~8000
Sample B, de-alloying
0.328
0.089
212
~4000
Sample C, alloying
0.956
0.256
1372
~26000
Sample C, de-alloying
0.304
0.088
457
~8000
of the mesoporous coatings after their first charge/discharge cycle.
These cracks seem similar in nature to those found by Yang et al
[17] on Sn electrodes. In that case it was surmised that a drastic
increase in volume occurred in the first alloying half cycle. However, lithium removal and further cycling caused only minor effects.
The V-t data for the second of the alloying/de-alloying cycles of
the three types of Au electrode are shown in Fig. 2. Fig. 2 (a)
shows the results for Sample A, the single layer of mesoporous
gold, Fig. 2 (b) shows data for Sample B, the multilayered
mesoporous coating, and Fig. 2(c) shows results for Sample C, the
single layer of solid gold with a nominal thickness of 100 nm. The
alloying and de-alloying reactions are especially visible in Fig.
2(c), and it can be seen that during the alloying process lithium
insertion occurred at two extended plateaus at about 0.23 and 0.11
V, while during de-alloying lithium removal occurred at two plateaus at 0.15 and 0.40 V. These figures are comparable to the 0.15
and 0.10 vs Li reported previously for alloying in Au, and to the
0.18 and 0.40 v vs Li reported for de-alloying on that element [9].
The voltage ranges associated with the single layer mesoporous
electrode, Sample A, are similar, being 0.20 and 0.10 for alloying,
and 0.20 and 0.40 V for de-alloying. However, only reactions at
0.15 and 0.45 V are well developed on the curves for the multilayer
Sample B. These results confirm that the electrochemical active
potential of Li-Au is lower than Li-Sn [9] and higher than Li-Ag
[18]. In Ag, substantial insertion of Li occurs below 0.07 V [18].
The cyclic voltammograms for Samples A and B (Fig. 3) also
confirm that the alloying/de-alloying processes occur in a very low
voltage range (0.4 to 0 V), compared to the Li–Sn system (0.2 to
1.0 V) [19]. In the first scanning cycle, two lithiation peaks appear,
which correspond to the formation of different LixAu alloys as Au
+ xLi+ + xe- ↔ LixAu. From the second scanning cycle, only one
pair of redox peaks occurs.
The total charges passed in the second cycle of alloying and dealloying the three types of electrodes are compared in Table 2. A
notional film density of 12 g.cm-3 has been used to obtain volumetric capacity. The first point to note is that the charge passed when
alloying is significantly larger than the charge passed when dealloying. This may be an indication of the relative inefficiency of
the alloying process of an anode material in comparison to the dealloying process. However, it is very clear that the processes are
more reversible on the mesoporous electrodes than on the solid
gold film. As far as drawing power from a battery is concerned, it
would be the charge passed during de-alloying of an anode that is
important. The multilayer (Sample B) and solid gold (Sample C)
electrodes start off far better than the single layer mesoporous film
in this regard. The calculated volumetric parameters are of the same
order of magnitude as those reported for Au by Taillades et al [9],
but have evidently been effected by uncertainty regarding the actual
density of the lithiated structures formed. They are reported here
only to indicate the approximate magnitude of energy density that
can be achieved. Fig. 4 shows the specific charge capacity and
cycle life numbers for Sample A, Sample B, and Sample C, as determined during their de-alloying cycles. It is clearly evident that
the capacity of the mesoporous Au on de-alloying is superior to
that of the equivalent layer of solid Au. The available data also
suggest that the method of synthesis of the mesoporous gold may
affect this capacity, with that of the multilayer mesoporous sample
Mesoporous Gold as Anode Material for Lithium-Ion Cells / J. New Mat. Electrochem. Systems
99
5. ACKNOWLEDGEMENTS
Authors from the University of Wollongong would like to thank
the Australian Research Council and the industry partners (Sons of
Gwalia Ltd. and OM Group) for their support through ARC Linkage project LP0219309 and for support through the ARC Centre for
Nanostructured Electromaterials.
REFERENCES
Figure 4. The specific charge storage capacity and cycle life response for mesoporous Au (Sample A - single mesoporous layer,
Sample B - multilayer of mesoporous gold, Sample C - single
layer of solid gold, 100 nm thick.)
being superior after 30 cycles to that of the single layer mesoporous
sample.
In our case, among the three samples, the multilayer mesoporous
electrode demonstrates a significantly better electrochemical cyclability than the other types tested. The capacity decay of all three
types of electrode is a problem shared with many other metallic
alloy systems, see for example references [9,13,14]. After 30 cycles
the electrodes made of materials A, B and C retained 15, 10 and
2% of their capacity (in mA.h.g-1) respectively. The fading of the
capacity may be caused by two reactions: (1) insertion of Li atoms
leads to significant volume expansion of host structure, especially
in the beginning few cycles, which can cause cracking and loss of
electrical contact. However, removal of lithium from the LixAu
particles does not affect the size of the expanded particles very
much [17]. So the discharge capacities of the electrodes can remain
almost constant in subsequent cycles, (2) The formation of a passivating film, an electrically insulating layer on the battery electrodes also known as the solid electrolyte interphase (SEI) film,
might also be effecting the capacity [20]. When the thickness of
SEI layer increases, the ionic impedance of the SEI increases [21].
Since the lithium insertion process occurs on an electrode covered
with the SEI, the characteristics of lithation/delithiation, and stability of the interface are effected [22].
4. CONCLUSIONS
Thin films of mesoporous Au have been evaluated for use as the
anode of lithium rechargeable cells for the first time. The behaviour
was similar to that for thin, solid films of gold, with the alloying/de-alloying processes occurring in a very low voltage range vs
Li/Li+ (0.25 to 0 V, and 0.15 to 0.40V, compared to the 0.2 to 1.0V
of the Li–Sn system). This would minimize the reduction in cell
voltage in a battery that used this material as an anode. The multilayer mesoporous Au electrode showed superior discharge capacities and better cycle ability than the thin, solid gold film.
[1] E. Van der Lingen, M. B. Cortie, I. Glaner, International Patent
Application, WO 03/059507A1, 2003.
[2] M. B. Cortie, A. I. Maaroof, G. B. Smith, Gold Bulletin, 38(1),
15 (2005).
[3] M. B. Cortie, A. Maaroof, G. B. Smith, P. Ngoepe, Current
Applied Physics, 6, 440 (2006).
[4] B. E. Conway, V. E. Birss, J. Wojtowicz, J. of Power Sources,
66 1 (1997).
[5] D. van Noort, C. F. Mandenius, Biosensors & Bioelectronics,
15, 203 (2000).
[6] M. J. Natan and B. E. Baker, United States Patent 6,242,264,
2001.
[7] H.-C Shin and M. Liu, Advanced Functional Materials, 15(4)
582 (2005).
[8] R. A. Huggins, J. of Power Sources, 26 109 (1989).
[9] G. Taillades, N. Benjelloum, J. Sarradin, M. Ribes, Solid State
Ionics, 152-153, 119 (2002).
[10]H. Baker, (ed.), ASM Handbook. Volume 3. “Alloy Phase
Diagrams”, ASM INTERNATIONAL, Metals Park, Ohio,
1992.
[11]C. Cretu and E. Van der Lingen, Gold Bulletin, 32, 115 (1999).
[12]R. W. Cahn , Nature, 396, 523 (1998).
[13]J.-H. Ahn, G.X. Wang, J. Yao, H.K. Liu, S.X. Dou, J. of Power
Sources, 119–121, 45 (2003).
[14]H. Honda, H. Sakaguchi, I. Tanaka, T. Esaka, J. of Power
Sources 123, 216 (2003).
[15]M.S. Whittingham, Chem. Rev., 104, 4271 (2004).
[16]S-K. Chang, H-Jin Kim, S-Tae Hong, J. of Power Sources,
119–121, 69 (2003).
[17]J. Yang, M. Winter, J. O. Besenhard, Solid State Ionics, 90,
281 (1996).
[18]S-M Hwang, H-Y Lee, S-W Jang,a, S-M Lee, S-J Lee, H-K
Baik, and J-Y Lee Electrochemical and Solid-State Letters,
4(7), A97 (2001).
[19]W. Choi, J-Y. Lee, Mat. Res. Soc. Symp. Proc., 822, S3.3.1
(2004).
[20]M. Dollé, S. Grugeon, B. Beaudoin, L. Dupont, J.M. Tarascon,
J. Power Sources, 97-98, 104 (2001).
[21]B. M. Meyer, N. Leifer, S. Sakamoto, S. G. Greenbaum and C.
P. Greyb, Electrochemical and Solid-State Letters, 8(3), A145
(2005).
[22]B. V. Ratnakumar, M. C. Smart, S. Surampudi, J. Power
Sources, 97-98, 137 (2001).